PPARα, δ and FOXO1 Gene Silencing Overturns Palmitate-Induced Inhibition of Pyruvate Oxidation Differentially in C2C12 Myotubes

Simple Summary Frequent high-dietary fat intake increases muscle lipid use and reduces muscle carbohydrate use, thereby inducing metabolic inflexibility. The latter term can be described as a poor muscle biochemical and molecular response to increased availability of insulin, which in the long term results in chronically excessive-high glucose levels in blood. Chronic hyperglycaemia is associated with many pathological conditions, including type 2 diabetes mellitus, which can cause severe health damages in humans. Here, we attempt to unravel the underlying mechanism and its associated factors behind the inhibition of muscle glucose use by a high-fat diet, thereby providing evidence for appropriate therapeutic intervention. Abstract The molecular mechanisms by which free fatty acids (FFA) inhibit muscle glucose oxidation is still elusive. We recently showed that C2C12 myotubes treated with palmitate (PAL) presented with greater protein expression levels of PDK4 and transcription factors PPARα and PPARδ and lower p-FOXO/t-FOXO protein ratios when compared to control. This was complemented with the hallmarks of metabolic inflexibility (MI), i.e., reduced rates of glucose uptake, PDC activity and maximal pyruvate-derived ATP production rates (MAPR). However, the relative contribution of these transcription factors to the increase in PDK4 and reduced glucose oxidation could not be established. Therefore, by using a similar myotube model, a series of individual siRNA gene silencing experiments, validated at transcriptional and translation levels, were performed in conjunction with measurements of glucose uptake, PDC activity, MAPR and concentrations of metabolites reflecting PDC flux (lactate and acetylcarnitine). Gene silencing of PPARα, δ and FOXO1 individually reduced PAL-mediated inhibition of PDC activity and increased glucose uptake, albeit by different mechanisms as only PPARδ and FOXO1 silencing markedly reduced PDK4 protein content. Additionally, PPARα and FOXO1 silencing, but not PPARδ, increased MAPR with PAL. PPARδ silencing also decreased FOXO1 protein. Since FOXO1 silencing did not alter PPARδ protein, this suggests that FOXO1 might be a PPARδ downstream target. In summary, this study suggests that the molecular mechanisms by which PAL reduces PDC-mediated glucose-derived pyruvate oxidation in muscle occur primarily through increased PPARδ and FOXO1 mediated increases in PDK4 protein expression and secondarily through PPARα mediated allosteric inhibition of PDC flux. Furthermore, since PPARδ seems to control FOXO1 expression, this may reflect an important role for PPARδ in preventing glucose oxidation under conditions of increased lipid availability.


Introduction
Frequent nutrient overload, especially in the form of fat, contributes to the development of skeletal muscle metabolic inflexibility (MI), a state associated with mitochondrial

C2C12 Cell Culture, Differentiation into Myobutes and Small Interfering RNA (siRNA) Transfection
The C2C12 skeletal muscle cells used in the present experiments were acquired from ATCC (Manassas, VA, USA). The cells were cultured in collagen-coated 6-well plates and maintained in a growth media containing High Glucose Dulbecco's Modified Eagle Media (DMEM), 10% Foetal Bovine Serum (FBS) and a 1% Pen/Strep. Once the myoblasts reached 80-100% confluency (after four days), they were differentiated into myotubes by using differentiating media containing high glucose DMEM, 2% Donor Horse Serum and 1% Pen/Strep. Two days into differentiation (day 6), the myotubes were transfected with 100 nM siRNA and 3 µL of Lipofectamine (ThermoFisher, Loughborough, UK) in an antibiotic-free medium in a 6-well plate according to the manufacturer's instructions and allowed to grow for two additional days. At the end of day 8, the media were removed, and standard media (see above) with or without 500 µM palmitate were added to the cells, which were allowed to grow for a further 24 h. At the end of day 9, the media were removed, and the myotubes were harvested and used for different experimental measurements, as will be further described. The same concentrations of non-specific siRNA (scramble) with Lipofectamine were transfected in parallel plates as controls. The cells were constantly incubated at 37 • C in 5% CO 2 throughout all phases and experiments.

Glucose Uptake Assay
Glucose uptake was determined as previously reported [12]. Briefly, C2C12 myotubes attached to the plate were incubated in 1 mL of media buffer supplemented with 138 mM NaCl, 1.85 mM CaCl 2 , 1.3 mM MgSO 4 , 4.8 mM KCl, 50 mM HEPES pH 7.4 and 0.2% (w/v) BSA for 2 h. Then, 250 µL of 2-deoxy-H 3 -glucose (1 mCi/mL) was added to the well, and the cells were incubated for 15 min. After incubation, the media were removed, and the well was washed three times with cold PBS buffer. After removing the last wash content, 500 µL of NaOH 50 mM and SDS 0.1% was added to solubilise the cells. A 500 µL aliquot of the solubilised cell mixture was removed and pipetted into a vial containing 5 mL of scintillation liquid, and radioactivity was measured with a scintillation counter (make and model).

Cell Metabolite Levels and PDC Activity Measurements
Medium lactate concentration was determined fluorometrically using a modified spectrophotometric method [13]. Acetylcarnitine (as an index of PDC flux; [14,15]) was measured in the cell extract using an enzymatic radioactive assay as previously described [16]. The cellular PDC activity was measured as described previously [7]. Briefly, C2C12 cells were extracted in a buffer containing sodium fluoride (NaF) and dichloroacetate (DCA). PDC activity in its dephosphorylated active form (PDCa) was assayed and expressed as the rate of acetyl-CoA formation (pmol acetyl-CoA/min/mg protein) at 37 • C.

Determination of Maximal Mitochondrial ATP Production Rates in Intact C2C12 Myotubes
Mitochondrial maximal ATP production rates (MAPR) from glucose-derived pyruvate were measured in C2C12 myotubes as described previously [17]. Of note, the MAPR measurements were conducted in the absence of palmitate, which was removed after cell pelleting and resuspension in the cell permeabilisation buffer, followed by resuspension in the ATP measuring buffer. Briefly, 3 µL of permeabilised cells was added to each well of a 96-well luminometer plate. Each well contained 200 µL of ATP monitoring reagent, 12.5 µL of 12 mM ADP and 35 µL of 375 mM pyruvate + 143 mM malate. Bioluminescence was recorded in duplicates continuously for 10 min (BMG LABTECH Ltd., Ortenberg, Germany), after which an injection of 150 pmoles ATP standard took place. The increase in the luminescence elicited by the ATP standard was used to calculate the values for absolute MAPRs. Subsequently, individual MAPRs were normalised for protein content by using the Bradford assay [18].

Western Blotting
The Near-Infrared (NIR) Western Blot Detection was employed [19], which is more dynamic, covers a more comprehensive linear range and has a higher sensitivity than compared to traditional chemiluminescent Western blotting. Briefly, C2C12 myotubes were lysed with a buffer containing a cocktail of protease and phosphatase inhibitors. The total protein concentration was determined using a protein assay (Bio-Rad, Hercules, CA, USA). In total, 25 µg of protein from each lysate was denatured, dissolved and separated with sodium dodecyl sulfate-polyacrylamide using gel electrophoresis (SDS-PAGE) in an XCell4 SureLock™ Midi-Cell system (Invitrogen, Inchinnan, UK). The separated proteins were transferred to a polyvinylidene fluoride (PVDF) or nitrocellulose membrane and blocked with 5% milk in Tris-buffered saline with Tween-20 (TBST) for 1 h at room temperature. After washing three times with TBST, the membrane was probed with the primary antibodies corresponding to the proteins of interest in TBST overnight at 4 • C. The following day, Licor IRDye 800CW (for α-actin) and 680RD (for PPARα, δ and FOXO1) fluorescent secondary antibodies were applied in TBST for 1 h at room temperature. The intensity of the protein bands was quantified using a near infra-red Licor imaging system (Odyssey ® CLx instrument; Cambridge, UK).

Statistical Analysis
All data are expressed as mean ± SEM of 6 individual experiments in each study group. The results for the cell glucose uptake are presented as the mean percentage (%) from the value of the first experiment in the glucose control group. This procedure provided a statistic (SEM) for the spreading of the dataset (including control) over dayto-day experimental runs. Two-way repeated-measures analysis of variance (ANOVA) was applied to identify treatment effects. When a significant F-ratio was obtained, a least significant difference (LSD) post hoc test was used to locate specific between treatment differences. Significance was set at the p < 0.05 level of confidence. The siRNA transfections were validated at transcriptional and translational levels by real-time PCR and Western blotting techniques, respectively. Figure 1A shows that PPARα mRNA expression with or without PAL was significantly lower with PPARα siRNA treatment than their corresponding CON (both p < 0.01). Typical Western blots displayed in Figure 1B show that the mean PPARα protein level was significantly lower with PPARα gene silencing with or without PAL than CON ( Figure 1C; both p < 0.01). Figure 1D shows that PPARδ mRNA expression with or without PAL was significantly lower with PPARδ siRNA treatment than CON (both p < 0.01). Typical Western blots displayed in Figure 1E show that the mean PPARδ protein level was significantly lower with PPARδ gene silencing with or without PAL than CON ( Figure 1F; both p < 0.01). Figure 1G shows that FOXO1 mRNA expression with or without PAL was significantly lower with FOXO1 siRNA treatment than CON (both p < 0.01). Typical Western blots displayed in Figure 1H show that the mean total FOXO1 protein content with FOXO1 siRNA intervention was significantly lower than CON with or without PAL ( Figure 1I; both p < 0.01).
In summary, these data show that all siRNA transfections were successful. Figure 1G shows that FOXO1 mRNA expression with or without PAL was significantly lower with FOXO1 siRNA treatment than CON (both p < 0.01). Typical Western blots displayed in Figure 1H show that the mean total FOXO1 protein content with FOXO1 siRNA intervention was significantly lower than CON with or without PAL (Figure 1I; both p < 0.01).
In summary, these data show that all siRNA transfections were successful.    Figure 2A shows glucose uptake expressed as % difference from CON after individual PPARα, PPARδ and FOXO1 gene silencing with or without PAL. In the absence of PAL, siRNA silencing did not affect glucose uptake in any group compared to CON.
in the absence (left panels) or the presence of palmitate (PAL; right panels). Significantly different from control without PAL (CON, ** p < 0.01). Significantly different from CON + PAL ( ## p < 0.01). Data represent mean ± SEM of 6 individual experiments for each intervention group. Nsi-silencer negative control. Figure 2A shows glucose uptake expressed as % difference from CON after individual PPARα, PPARδ and FOXO1 gene silencing with or without PAL. In the absence of PAL, siRNA silencing did not affect glucose uptake in any group compared to CON. However, following siRNA gene silencing, glucose uptake in all groups with PAL was significantly greater than CON with PAL (p < 0.01; all groups). The glucose uptake in CON with PAL was significantly lower than CON without PAL (p < 0.01). Figure 2B shows media lactate concentration in C2C12 myotubes after individual PPARα, PPARδ and FOXO1 gene silencing with or without PAL. Irrespective of PAL's However, following siRNA gene silencing, glucose uptake in all groups with PAL was significantly greater than CON with PAL (p < 0.01; all groups). The glucose uptake in CON with PAL was significantly lower than CON without PAL (p < 0.01). Figure 2B shows media lactate concentration in C2C12 myotubes after individual PPARα, PPARδ and FOXO1 gene silencing with or without PAL. Irrespective of PAL's absence or presence, media lactate concentration with PPARδ silencing was significantly lower than their corresponding CON (both p < 0.01). Although FOXO1 silencing did not affect lactate levels without PAL, lactate levels were significantly lower than CON when PAL was present (p < 0.05).  Figure 3A shows PDC activity in CON and after individual PPARα, PPARδ and FOXO1 gene silencing with or without PAL. In the absence of PAL, only siRNA PPARδ silencing significantly upregulated PDC activity (p < 0.01). With PAL, all gene silencing significantly upregulated PDC activity compared to CON (p < 0.01 in all conditions).

Glucose Uptake and Media Lactate Accumulation
absence or presence, media lactate concentration with PPARδ silencing was significantly lower than their corresponding CON (both p < 0.01). Although FOXO1 silencing did not affect lactate levels without PAL, lactate levels were significantly lower than CON when PAL was present (p < 0.05). Figure 3A shows PDC activity in CON and after individual PPARα, PPARδ and FOXO1 gene silencing with or without PAL. In the absence of PAL, only siRNA PPARδ silencing significantly upregulated PDC activity (p < 0.01). With PAL, all gene silencing significantly upregulated PDC activity compared to CON (p < 0.01 in all conditions). Significantly different from CON + PAL ( # p < 0.05; ## p< 0.01). Data represent mean ± SEM of 6 individual experiments for each intervention group. Nsi-silencer negative control. Figure 3B shows the naïve cellular acetylcarnitine concentration (used as an index of the flux through the PDC reaction) and cells with individual PPARα, PPARδ and FOXO1 gene silencing with or without PAL. In the absence of PAL, acetylcarnitine concentration  Figure 3B shows the naïve cellular acetylcarnitine concentration (used as an index of the flux through the PDC reaction) and cells with individual PPARα, PPARδ and FOXO1 gene silencing with or without PAL. In the absence of PAL, acetylcarnitine concentration following PPARα gene silencing was significantly lower than CON (p < 0.01). In the presence of PAL, acetylcarnitine concentrations following PPARδ and FOXO1 gene silencing were significantly greater than CON + PAL (p < 0.05 in both conditions). resulted in significantly lower MAPR than CON irrespective of the absence or the presence of PAL (p < 0.01 in both conditions), while PPARα and FOXO1 gene silencing significantly increased MAPR relative to CON in the absence (p < 0.01 and p < 0.05, respectively) or the presence of PAL (p < 0.01 and p < 0.05, respectively). ence of PAL, acetylcarnitine concentrations following PPARδ and FOXO1 gene silencing were significantly greater than CON + PAL (p < 0.05 in both conditions). The results presented in Figure 4 show mean MAPR values in CON and after individual PPARα, PPARδ and FOXO1 gene silencing with or without PAL. PPARδ silencing resulted in significantly lower MAPR than CON irrespective of the absence or the presence of PAL (p < 0.01 in both conditions), while PPARα and FOXO1 gene silencing significantly increased MAPR relative to CON in the absence (p < 0.01 and p < 0.05, respectively) or the presence of PAL (p < 0.01 and p < 0.05, respectively).

PDK4 mRNA and Protein Expression
The effects of individual PPARα, PPARδ and FOXO1 silencing on the mean PDK4 gene expression in the absence (left panels) or the presence of palmitate (PAL; right panels) are presented in Figure 5A,D,G, respectively. The results show that PPARα silencing did not prevent upregulation of PDK4 mRNA expression by PAL. However, PPARδ and FOXO1 silencing prevented upregulation of PDK4 mRNA by PAL (both p < 0.01). Figure 5B,E,H show typical Western blots of PDK4 and their corresponding α-actin protein bands after PPARα, δ and FOXO1 gene silencing interventions, respectively, with or without PAL. The mean PDK4 protein expression is presented in Figure 5C,F,I, respectively. Figure 5C shows that PPARα silencing did not prevent upregulation of PDK4 protein by PAL. However, both PPARδ and FOXO1 silencing prevented upregulation of PDK4 protein by PAL (both p < 0.01). Significantly different from control without PAL (CON, ** p < 0.01). Significantly different from CON + PAL ( # p < 0.05; ## p < 0.01). Data represent mean ± SEM of 6 individual experiments for each intervention group. Nsi-silencer negative control.

PDK4 mRNA and Protein Expression
The effects of individual PPARα, PPARδ and FOXO1 silencing on the mean PDK4 gene expression in the absence (left panels) or the presence of palmitate (PAL; right panels) are presented in Figure 5A,D,G, respectively. The results show that PPARα silencing did not prevent upregulation of PDK4 mRNA expression by PAL. However, PPARδ and FOXO1 silencing prevented upregulation of PDK4 mRNA by PAL (both p < 0.01). Figure 5B,E,H show typical Western blots of PDK4 and their corresponding α-actin protein bands after PPARα, δ and FOXO1 gene silencing interventions, respectively, with or without PAL. The mean PDK4 protein expression is presented in Figure 5C,F,I, respectively. Figure 5C shows that PPARα silencing did not prevent upregulation of PDK4 protein by PAL. However, both PPARδ and FOXO1 silencing prevented upregulation of PDK4 protein by PAL (both p < 0.01).

PPARδ and FOXO1 Translational and Post-Translational (Phosphorylation) Levels
after PPARα siRNA Gene Silencing in C2C12 Myotubes Treated with or without Palmitate (PAL) Figure 6A shows typical PPARδ and FOXO1 protein Western blots after PPARα silencing with or without PAL. PPARα siRNA gene silencing did not prevent the marked upregulation of PPARδ protein in all groups with PAL ( Figure 6B). PPARα gene silencing with PAL significantly reduced both phosphorylated FOXO1 ( Figure 6C, p < 0.05) and total-FOXO1 ( Figure 6D, p < 0.01) protein levels, albeit by different magnitudes in such a manner that p-FOXO1/t-FOXO1 ratio increased significantly compared to PAL alone ( Figure 6E, p < 0.01).  lencing with or without PAL. PPARα siRNA gene silencing did not prevent the marked upregulation of PPARδ protein in all groups with PAL ( Figure 6B). PPARα gene silencing with PAL significantly reduced both phosphorylated FOXO1 ( Figure 6C, p < 0.05) and total-FOXO1 ( Figure 6D, p < 0.01) protein levels, albeit by different magnitudes in such a manner that p-FOXO1/t-FOXO1 ratio increased significantly compared to PAL alone (Figure 6E, p < 0.01).  Figure 7A shows typical Western blots after PPARδ gene silencing with or without PAL. PPARδ gene silencing with PAL decreased PPARα protein significantly compared to PAL alone ( Figure 7B; p < 0.01). PPARδ gene silencing with PAL reduced both p-FOXO1 and t-FOXO1 protein expression significantly compared to PAL alone ( Figure 7C,D, both p < 0.01), but the p-FOXO1/t-FOXO1 ratio remained unchanged from PAL alone ( Figure 7E). Figure 7A shows typical Western blots after PPARδ gene silencing with or without PAL. PPARδ gene silencing with PAL decreased PPARα protein significantly compared to PAL alone ( Figure 7B; p < 0.01). PPARδ gene silencing with PAL reduced both p-FOXO1 and t-FOXO1 protein expression significantly compared to PAL alone ( Figure 7C,D, both p < 0.01), but the p-FOXO1/t-FOXO1 ratio remained unchanged from PAL alone ( Figure  7E). Significantly different from control without PAL (CON, ** p < 0.01). Significantly different from CON + PAL ( # p < 0.05; ## p < 0.01). Data represent mean ± SEM of 6 individual experiments for each intervention group. Nsi-silencer negative control.

PPARα and δ Translational Levels after FOXO1 siRNA Gene Silencing in C2C12
Myotubes Treated with or without Palmitate (PAL) Figure 8A shows typical Western blots after FOXO1 silencing intervention with or without PAL. Figure 8B shows FOXO1 gene silencing decreased PPARα protein expression with PAL compared to PAL alone (p < 0.01). However, the FOXO1 gene silencing with PAL treatment did not change PPARδ protein expression ( Figure 8C).
3.3.4. PPARα and δ Translational Levels after FOXO1 siRNA Gene Silencing in C2C12 Myotubes Treated with or without Palmitate (PAL) Figure 8A shows typical Western blots after FOXO1 silencing intervention with or without PAL. Figure 8B shows FOXO1 gene silencing decreased PPARα protein expression with PAL compared to PAL alone (p < 0.01). However, the FOXO1 gene silencing with PAL treatment did not change PPARδ protein expression ( Figure 8C).

Discussion
The present study consolidated previous findings, which showed that gene and protein expressions of PPARα δ and FOXO1 transcription factors in C2C12 skeletal muscle myotubes are markedly augmented when the cells are treated with PAL compared to the absence of PAL [10]. Additionally, these molecular outcomes were associated with hallmarks of MI/IR, i.e., lower rates of cellular glucose uptake, PDC activity and flux through PDC reaction and maximal glucose-derived pyruvate ATP production (MAPR).
In order to gain further mechanistic insight, we executed siRNA gene silencing of PPARα, δ and FOXO1 transcription factors individually. All three transcription factors

Discussion
The present study consolidated previous findings, which showed that gene and protein expressions of PPARα δ and FOXO1 transcription factors in C2C12 skeletal muscle myotubes are markedly augmented when the cells are treated with PAL compared to the absence of PAL [10]. Additionally, these molecular outcomes were associated with hallmarks of MI/IR, i.e., lower rates of cellular glucose uptake, PDC activity and flux through PDC reaction and maximal glucose-derived pyruvate ATP production (MAPR).
In order to gain further mechanistic insight, we executed siRNA gene silencing of PPARα, δ and FOXO1 transcription factors individually. All three transcription factors have been shown by us and others to be associated with an increase in PDK4 protein content in cells treated with palmitate [10,[20][21][22]. Therefore, we hypothesized that their knockdown would overcome PAL induced inhibition of PDC activity and increase glucose-derived pyruvate oxidation.
It was found that PPARα, δ and FOXO1 siRNA gene silencing reversed the PALmediated decline in PDC activity in all three instances. However, it appears that this was achieved through different mechanisms, as only PPARδ and FOXO1 siRNA gene silencing reduced the expression of PDK4 protein markedly. Therefore, it seems that PPARα may downregulate glucose oxidation independently from the levels of PDK4 protein [23]. This latter finding bears resemblance to the result that overexpression of PPARα lowers glucose oxidation in rodent heart tissue by decreasing flux through PDC reaction exclusively via the allosteric inhibition of PDC by the products of FFA oxidation, rather than through levels of PDK4 protein [24]. Of note here, the p-FOXO1/t-FOXO1 ratio increased significantly after PPARα gene silencing with PAL in the present study ( Figure 6E). Although FOXO1 robustly upregulates muscle PDK4 protein during states associated with elevated circulating FFA levels (e.g., starvation) [21], conversely, an increase in p-FOXO1 (phosphorylated form), such as that observed after PPARα gene silencing, would induce inactivation of FOXO1 [25], thereby providing an even weaker molecular drive for downregulating PDK4 mRNA and protein ( Figure 5A,C).
Our investigation also showed that PPARα, PPARδ and FOXO1 silencing with PAL improved glucose uptake and glucose-derived pyruvate oxidation compared with naïve cells. Furthermore, PPARδ and FOXO1 silencing with PAL lowered cell media lactate and increased acetylcarnitine concentrations, which collectively point towards improved flux through PDC reaction compared to PPARα silencing.
Usually, there is a positive association between the rate of flux through the PDC reaction and the magnitude of intracellular acetylcarnitine accumulation. When the rate of PDC mediated acetyl-CoA formation exceeds the rate of acetyl-CoA entry into the citric acid cycle, the excess acetyl-CoA is buffered by carnitine, which causes the cellular content of acetylcarnitine rise [14,15]. Collectively, these findings point to flux through PDC reaction after PPARδ and FOXO1 silencing with PAL being significantly greater than after PPARα silencing. Therefore, the latter finding adds more weight to the contention that the effect of functional PPARδ and FOXO1 on PDC activity is more impactful than that of PPARα.
Additionally, PPARα and FOXO1, but not PPARδ silencing, rescued MAPR from the effects of PAL. It was also found that PPARδ silencing reduced FOXO1 protein, although FOXO1 silencing did not change PPARδ protein, suggesting that FOXO1 may potentially be a PPARδ downstream target.
Several previous studies showed that overexpression of PPARδ and FOXO1 via agonism or intralipid infusion in in vivo models upregulate PDK4 gene or protein expression, thereby decreasing skeletal muscle PDC activity [26][27][28]. Conversely, it was shown here that PPARδ and FOXO1 silencing downregulate PDK4 gene and protein expression and improved PDC activity and glucose-derived pyruvate oxidation with PAL. Among the genes investigated here, it was found that PPARδ silencing had the highest potency of all three for improving PDC activity (45%), suggesting that along with the well-established role of PPARδ in mobilising fat oxidation, it also has, although not uniquely, the ability to reduce glucose-derived pyruvate oxidation. After each gene silencing, the PDK4 protein expression levels recorded revealed that PPARδ and FOXO1 genes, rather than PPARα, mediated the rise in PDK4 protein with PAL, thereby confirming that PPARδ and FOXO1 transcription factors are de facto central modulators controlling PDC activity.
Since both PPARδ and FOXO1 silencing resulted in lower PDK4 protein expression levels, this may signify similar roles for these two transcription factors in PDK4 modulation. However, since PPARδ silencing had the most significant recovery effect on PDC activity with PAL, this may suggest that PPARδ should also have had the most significant impact in modulation glucose-derived pyruvate oxidation out of any of the transcription factors investigated. However, our additional data contradict this. Thus, from a mitochondrial perspective, the MAPR responses with PPARδ silencing and PAL seemed to follow a different pattern from that of PPARα and FOXO1 silencing, which was to depress MAPR further rather than rescuing it. This finding may seem counterintuitive in hindsight also goes against a previous in vivo study, which showed that PPARδ agonism inhibited MAPR [29]. However, it is worth noting that PPARδ also controls the expression of skeletal muscle PGC-1α (peroxisome-proliferator-activated receptor γ activator-1α) [30,31]. PGC-1α is a co-transcriptional regulation factor that induces mitochondrial biogenesis by activating different transcription factors, including nuclear respiratory factor 1 and nuclear respiratory factor 2, which activate mitochondrial transcription factor A. The latter drives the transcription and replication of mitochondrial DNA (mtDNA) [32].
Given the above background information, it can be cautiously assumed that a decline of the mtDNA encoded mitochondrial proteins (n = 13, e.g., the cytochrome oxidase subunits 1-3, components of complex I (ND1 and ND2)) could have occurred with PPARδ gene silencing. Consequently, the latter could have reduced the mitochondrial pyruvate derived MAPR. Indeed, the expression of PGC-1α markedly decreased in PPARδ deficient heart, concomitant with a decreased mitochondrial DNA copy number [33]. Moreover, glucose oxidation rates were markedly depressed in the cardiomyocytes of PPARδ knockout hearts [33].
Interestingly, similarities in the pattern of PDC activity and its relative function after PPARδ and FOXO1 silencing were observed. Therefore, we explored whether an interaction between PPARδ and FOXO1 may exist. The protein results showed that PPARδ silencing reduced total FOXO1 protein, but FOXO1 silencing did not change the expression levels of PPARδ protein. The latter finding along with the reported greater significance of PPARδ silencing on rescuing PDC activity and reducing media lactate accumulation, even in the presence of PAL compared with FOXO1 silencing, suggests that PPARδ may be an upstream modulator of FOXO1.
Study limitation. During the experiments undertaken in the present study, PPARδ siRNA gene silencing in the presence of PAL depressed pyruvate derived MAPR responses, contrasting with those from PPARα and FOXO1. This finding is not in accordance with a previous in vivo finding [29]. Although we provided a cautious interpretation earlier based on evidence from the literature of this in vitro finding [30,32,33], it should be acknowledged that this could have also been attributable to the in vitro results not being extrapolatable to an in vivo environment.

Conclusions
In summary, the findings of this molecular investigation suggest that transcription factors PPARδ, FOXO1 and PPARα play, in decreasing order, essential roles in cellular fuel selection in skeletal muscles by downregulating glucose-derived pyruvate oxidation during conditions of chronically elevated levels of FFAs. It appears that this occurs due to the ability of PPARδ and FOXO1 to increase muscle PDK4 protein levels and through PPARα allosteric inhibition of flux through PDC reaction. Furthermore, since PPARδ seems to control FOXO1 expression, this may reflect an important role for PPARδ in preventing glucose oxidation under conditions of increased lipid availability.

Conflicts of Interest:
The authors declare no conflict of interest.