Low-fat diet, and medium-fat diets containing coconut oil and soybean oil exert different metabolic effects in untrained and treadmill-trained mice

Fat as a bioactive compound influences metabolism by inducing muscle and liver phenotypic remodeling through transcriptional activation of PPARs [15, 17, 3840]. We hypothesized that diets varying in fat source and proportion together with training would lead to different adaptations in the muscle and liver consequently affecting whole-body metabolism and endurance. Comparative studies on the effects of different diets with training have been conducted [20, 41, 42]. However, data on energy expenditure at rest and during exercise, substrate utilization, and gene transcription are scarce. Coconut oil is a good source of MCFAs which are rapidly metabolized relative to LCFAs [43]. MCFAs with training improve endurance in swimming [20] but other aspects of adaptation required further investigation. We aimed to update the current knowledge on the effects of MCFAs and fat types. We show that fat source and content in the diet exert variably influence different aspects of basal and treadmill training adaptations particularly on endurance, exercise whole-body metabolism, energy substrate storage and utilization, and genetic and biochemical characteristics of the muscle and liver.

Medium-fat diets increased FAT regardless of fat type and training promoted CHO at rest without affecting energy expenditure. Utilizing a lower intensity training protocol with the same soybean oil diet increased CHO, but it accompanied increased energy expenditure without affecting FAT [23] suggesting that different training intensities differentially affect resting metabolism [44, 45]. Our data in relation to [15] suggest that higher absolute MCFAs content may increase VO2 even without training relative to LCFAs.

Whole-body metabolism during exercise under different diets is associated with changes in VO2max [42]. Unfortunately, we could not measure VO2max because of technical limitations. During exercise under slight food deprivation, training but not diet influenced whole-body metabolism suggesting that at rest with ad libitum feeding, diet composition determined differences in resting energy metabolism while general effects were due to training. Moreover, training lowered VO2 (and energy expenditure) implying that exercise economy increased in trained groups during exercise [46]. In our previous study using a lower training intensity, decreased RER without changes in VO2 was observed suggesting higher fat utilization [23]. These observations further indicate that changes in energy metabolism at rest or during exercise is influenced by training intensity. It is important to note that inaccurate calculations in CHO and FAT, especially in C with ketosis during exercise, may exist because the complete oxidation of acetoacetate and β-HB give respiratory quotient values of 1.0 and 0.89, respectively [47]. Unfortunately, corrections for ketone body oxidation and its relative contribution to substrate metabolism could not be performed because a time-course ketone body profile in the blood was not available.

Our group and others show that diets high in fat induce muscle mitochondrial biogenesis and also imply that MCFAs promote mitochondrial biogenesis better than LCFAs at the same absolute concentration [15, 17]. β-HAD activity increased in C without training. Also, in C2C12 muscle cells, C10 and C12 fatty acids increased succinate dehydrogenase activity [15]. However, changes in the activities of other enzymes may not be entirely dependent on fatty acid species but on their amount as seen with SCOT and CS. We also showed that training universally increased mitochondrial function suggesting mitochondrial biogenesis. Therefore, elevated oxidative capacity in C, albeit small, likely improved endurance in the untrained state as measured by work while robust increase in mitochondria improved endurance in trained mice regardless of diet [48, 49]. In contrast with the swimming modality, MCFAs increased swimming time attributed to increased CS and SCOT relative to LCFAs [20] underscoring the notion that different training modalities variably influence adaptation, and potentially, endurance.

Exercise increases PGC1A and this co-activates or potentiates the PPARs and ERRs in the control of mitochondrial biogenesis [9]. We did not observe changes in PGC1A mRNA and protein despite increased mitochondrial function suggesting differential effects of training intensity on their half-lives [23, 50, 51]. However, we show that diets influenced basal and training-induced changes in mRNA expression of PPARs and ERRs. ERRs had decreased expression in medium-fat diets. This did not negatively affect mitochondrial enzyme activities, Pgc1a or Erra in the untrained state suggesting that at the basal level, homeostatic control and/or other ERR isotypes likely compensated for decreased mRNA expression of ERRγ [52, 53]. Training increased Erra and Errb, which could explain the increased mitochondrial biogenesis in the muscle [49].

High-fat diets increases Ppara but not the other isotypes in rats [54]. We did not observe significant elevations in Ppara in the untrained state possibly due to a relatively lower fat content of our diets. On the other hand, only Pparb/d increased with training in contrast to increased Ppara and unchanged Pparb/d when trained at a lower intensity even with the same soybean oil diet [23] suggesting that PPARs respond differently with training intensity. Interestingly, coconut oil impaired the training-induced upregulation of Pparb/d. Consistent with increased Pparb/d with training, target genes related to glucose and fat utilization, and fiber type remodelling (Glut4, Cpt1b, Fatp1 and Myh2) responded similarly especially in S [5558]. Because exercise increases fatty acid uptake by translocation of CD36 to the sarcolemma, not only increased expression of PPARβ/δ but also increased fatty acid-induced activation and availability could upregulate these targets [23, 59] which may explain some of the differences between S and L with training. On the other hand, these changes could not be attributed to increased CD36 as its mRNA and protein did not respond as expected [23] indicating that training intensity affects specific genetic adaptations.

The expression of Cpt1b and Lpl with medium-fat diets in the untrained state is probably related to PPARα as this isotype also controls their transcription in the muscle [60]. Because functions of PPAR isotypes overlap in some of these genes, we could not discount the contribution of PPARγ. Although PPARγ is abundant in adipose tissues, it is also present in skeletal muscle and MCFAs and LCFAs strongly activate PPARγ [1114, 61].

Total caloric intake was similar among diet groups. This means that the minimum amount of consumed soybean oil was similar among diet groups suggesting that coconut oil inhibited training-induced upregulation of PPARβ/δ and some downstream targets. Whether MCFAs inhibited LCFAs by competitive binding in PPARβ/δ activation requires further research. While competitive binding assays between fatty acids and synthetic agonists have been performed [11, 12, 14], competitive binding assay among fatty acids to PPARs has yet to be undertaken. Overall, PPAR-related gene transcription as a training adaptation was influenced by the type of fat in the diet. Also, these adaptations reflect the route of catabolism of energy substrates within these diets during exercise.

In the untrained state, diets did not affect pre-exercise muscle glycogen possibly due to similar circulating lipids or serum β-HB as these influence glycogen storage [6265]. Training increases muscle glycogen but ketone bodies, particularly acetoacetate, inhibit insulin-stimulated glucose uptake that occurs during feeding after training [66]. This may explain the inhibited glycogen accumulation in C. Furthermore, while GLUT4 is not essential for glycogen repletion per se it could influence glycogen accumulation with insulin post-exercise by increasing the rate of glucose transport [6769]. Liver glycogen accumulation was also impaired in C and to a lesser extent in S which was emphasized by training. This could be partly explained by exhaustion of hepatic glycogen reserves with MCFAs and the glycogen replenishing effect carbohydrates [70, 71].

Increased muscle glycogen and glycogen sparing improves endurance by slowing the utilization of circulating glucose and liver glycogen [72, 73]. Muscle glycogen is spared by improved utilization of fatty acids and ketone bodies [7477] linking glycogen sparing with high serum β-HB, decreased serum NEFA and intramuscular TG especially in C with training post-exercise. These suggest the existence of compensatory mechanisms to preserve endurance despite low pre-exercise muscle glycogen in this group. Nevertheless, muscle glycogen availability and utilization together with increased oxidative and mitochondrial functions likely promoted robust endurance improvement in all trained groups.

In the liver, unlike LCFAs, MCFAs can bypass fatty acid transport proteins to enter the cell and mitochondria for oxidation and these undergo β-oxidation for complete oxidation or ketogenesis [36, 59, 7881]. Improved oxidative capacity with increased upstream β-HAD activity suggested increased capacity to produce ketones particularly in C [82] despite higher downstream ketogenic enzyme activities in S than C. Furthermore, because MCFAs are undetectable in the serum at rest, the overwhelming increase in β-HB during exercise in C suggests that liver and adipose tissues stored MCFAs, released them to circulation during exercise and were immediately catabolized [6, 20, 79, 8385].

Fatty acid oxidation and ketogenesis in the liver is controlled by PPARα [10]. Diets did not influence Ppara but training increased its expression in C. Fibroblast growth factor 21 (FGF21) is induced by acetoacetate via an upstream regulator and this upregulates PPARα [38, 86] thus connecting the link between training, coconut oil, increased ketone bodies during training, and increased Ppara. While non-significant, increased Cpt1a and β-HAD activity in this group suggest PPARα activation in the liver [82]. Hmgcs2, the gene that encodes the first enzyme of ketogenesis [36], was unaffected by diet nor training suggesting that high β-HB observed post-exercise in C was primarily caused by increased supply of ketogenic precursors for β-oxidation in C rather than changes in ketogenic activity. On the other hand, relatively lower β-HB with training at post-exercise is likely because of increased muscular utilization accompanying increased SCOT activity. Whether increased oxidative capacity prevented increase of liver weight in C with training was not investigated. Overall, coconut oil with training promoted liver remodeling to an oxidative phenotype without influencing mitochondrial biogenesis.