Rats selectively bred for high capacity running (HCR) or low capacity running (LCR) display divergence for intrinsic aerobic capacity and hepatic mitochondrial oxidative capacity, both factors associated with susceptibility for nonalcoholic fatty liver disease. Here, we tested if HCR and LCR rats display differences in susceptibility for hepatic steatosis after 16 wk of high-fat diets (HFD) with either 45% or 60% of kcals from fat. HCR rats were protected against HFD-induced hepatic steatosis, whereas only the 60% HFD induced steatosis in LCR rats, as marked by a doubling of liver triglycerides. Hepatic complete fatty acid oxidation (FAO) and mitochondrial respiratory capacity were all lower in LCR compared with HCR rats. LCR rats also displayed lower hepatic complete and incomplete FAO in the presence of etomoxir, suggesting a reduced role for noncarnitine palmitoyltransferase-1-mediated lipid catabolism in LCR versus HCR rats. Hepatic complete FAO and mitochondrial respiration were largely unaffected by either chronic HFD; however, 60% HFD feeding markedly reduced 2-pyruvate oxidation, a marker of tricarboxylic acid (TCA) cycle flux, and mitochondrial complete FAO only in LCR rats. LCR rats displayed lower levels of hepatic long-chain acylcarnitines than HCR rats but maintained similar levels of hepatic acetyl-carnitine levels, further supporting lower rates of β-oxidation, and TCA cycle flux in LCR than HCR rats. Finally, only LCR rats displayed early reductions in TCA cycle genes after the acute initiation of a HFD. In conclusion, intrinsically high aerobic capacity confers protection against HFD-induced hepatic steatosis through elevated hepatic mitochondrial oxidative capacity.
- fatty liver
- fat oxidation
reports over the last decade have strongly suggested that a sedentary lifestyle and physical inactivity (38–40) increase susceptibility to nonalcoholic fatty liver disease (NAFLD), a relatively new member of the ever-growing list of metabolic abnormalities associated with obesity (39). NAFLD is composed of a spectrum of liver disease states, including simple lipid accumulation (steatosis), steatohepatitis, and cirrhosis. Up to two-thirds of individuals with obesity or type 2 diabetes have NAFLD (46). Most of the hepatic lipids are stored as inert triglycerides; however, excessive triglyceride storage is commonly associated with the accumulation of lipid species that putatively trigger hepatic insulin resistance (31). Therefore, NAFLD likely plays a key role in the development of insulin resistance and type 2 diabetes (37, 52). In addition, NAFLD diagnosis has been linked to a greater risk for cardiovascular disease and early mortality in both adolescent and adult populations (2, 12). All told, NAFLD is a serious health condition, and currently the only accepted treatment is diet and exercise. Moreover, we are only beginning to understand the mechanisms by which lifestyle prevents and treats NAFLD.
Recent attention has focused on factors driving susceptibility for NAFLD. Beyond inactivity and sedentarism, aerobic capacity has also been linked to NAFLD (40). Aerobic capacity is the maximal capacity of the cardiopulmonary system to deliver, and the peripheral tissues to use, oxygen during maximal exercise and is also commonly called “aerobic or cardiopulmonary fitness.” Aerobic capacity is determined by physical activity, genetics, and age (9). Higher daily physical activity levels and/or regular endurance exercise increases and/or aids in maintaining aerobic capacity, whereas inactivity leads to lower fitness across the lifespan (7). Approximately 50% of aerobic capacity is due to genetics (9).
Multiple groups have shown that low aerobic capacity is perhaps the most powerful predictor of early mortality (6, 30) and the development of numerous diseases, including type 2 diabetes (23) and NAFLD (10). However, defining the biological mechanism(s) by which aerobic capacity impacts susceptibility for metabolic disease is difficult in humans for a number of logistical and ethical reasons. In response to this need, Britton and Koch (51) created a polygenetic rodent model specifically designed to test the mechanism(s) by which aerobic capacity impacts disease. Rats were selectively bred over multiple generations for high or low endurance running capacity, resulting in high (HCR) and low capacity running (LCR) rats with dramatically different aerobic capacities. These traits are intrinsic and thus occur in caged rats that have never been exposed to exercise training. One of the key traits of HCR and LCR rats is that they display robust differences in oxidative capacity in muscle tissues (34). Interestingly, we have also shown that they display large differences in hepatic mitochondrial phenotypes (27, 48), where LCR rats display lower hepatic mitochondrial oxidative capacity as defined by reduced hepatic fatty acid oxidation (FAO), enzyme activity, and mitochondrial respiratory capacity compared with HCR rats. Moreover, we have shown that HCR rats were protected against but that LCR rats were susceptible to acute (3-day) high-fat diet (HFD)-induced hepatic steatosis (27).
Fatty acids that enter the liver have two fates: storage or oxidation. This has led to the hypothesis that reduced hepatic FAO due to ineffective oxidative machinery increases susceptibility for NAFLD, a link that we have made in previous rodent studies (17, 41, 42). Despite reduced hepatic mitochondrial respiratory function (15, 36), NAFLD also elevates energetic requirements (45) and fatty acid metabolism (18) in humans and HFD-fed rodents (44). Importantly, there is also evidence that hepatic mitochondrial respiratory function is influenced by the acute versus chronic nature of hepatic insulin resistance and the feeding status studied (13, 19). It has been suggested that increased demand on poorly functioning mitochondria might precipitate some of the pathophysiology of NAFLD (43), an effect that may be apparent in humans who have elevated or impaired hepatic mitochondrial respiratory capacity depending on disease severity (22). Hence, individuals with superior mitochondrial respiratory capacity, such as HCR rats, may be inherently protected from the metabolic overload of a HFD. Therefore, the purpose of this study was to test the role of divergent baseline aerobic capacity and associated differences in hepatic mitochondrial oxidative capacity for susceptibility to NAFLD after feeding of both 45% and 60% (kcal) chronic HFDs.
The HCR/LCR rat model was developed and characterized as previously described (21, 27, 34, 48, 51). At 25–30 wk of age, animals were singly housed (12:12-h light-dark cycle, 76–79°F) and acclimatized to the control low-fat diet (LFD; D12110704: 10% kcal fat, 3.5% kcal sucrose, and 3.85 kcal/g, Research Diets, New Brunswick, NJ) for at least 7 days before the initiation of two separate HFD's (D12451: 45% kcal fat, 17% kcal sucrose, and 4.73 kcal/g or D12492: 60% kcal fat, 7% kcal sucrose, and 5.24 kcal/g, Research Diets). Two different sets of HCR/LCR rats were used for the two HFD interventions, resulting in n = 16 for each strain on the LFD and n = 8 for each HFD in each strain. Food intake and body weight were monitored weekly during the 16-wk diets. Animals were fasted overnight before euthanasia and tissue collection. The animal protocols were approved by the Institutional Animal Care and Use Committee of the University of Missouri and the Subcommittee for Animal Safety at the Harry S. Truman Memorial Veterans Affairs Hospital.
Liver triacylglycerol concentration was determined as previously described using a commercially available kit (F6428, Sigma, St. Louis, MO) (48).
Mitochondria were isolated from rat liver tissue, as previously described (27). Briefly, tissue was homogenized (Teflon on glass) in cold liver mitochondrial isolation buffer (220 mannitol, 70 sucrose, 10 mM Tris, and 1 mM EDTA, adjusted to pH 7.4 with KOH) and centrifuged (1,500 g, 10 min, 4°C). The supernatant was serially centrifuged (8,000 g, 6,000 g, and 4,000 g, 10 min, 4°C), with the pellet resuspended (glass on glass) in liver mitochondrial isolation buffer after each centrifugation. The protein concentration was determined by the BCA assay.
Mitochondrial oxygen consumption was measured using a Clark type electrode system (Strathkelvin Instruments, North Lanarkshire, Scotland) as previously described (27). The consumption of oxygen (in nmol/min) was normalized to mitochondrial protein in the respirometer cell. One animal for each group was analyzed daily. Data are presented as a relative value compared with the daily average to control for day-to-day and between-set variance.
Palmitate oxidation capacity pyruvate oxidation by liver homogenate and isolated mitochondria.
The oxidation of [1-14C]palmitate was measured in fresh liver homogenates and isolated mitochondria as previously described (5, 27, 48). FAO was assessed by measuring the production of 14CO2 (complete FAO) and 14C acid-soluble metabolites (incomplete FAO) in a sealed trapping device. Complete FAO to CO2 represents the entry oxidation of acetyl-CoA produced by β-oxidation by flux through the tricarboxylic acid (TCA) cycle to CO2. Liver homogenates or isolated mitochondria were incubated in reaction buffer (125 mM sucrose, 10 mM Tris·HCl, 12.5 mM KPO4, 100 mM KCl, 1.25 mM MgCl2·6H2O, 1.25 mM l-carnitine, 0.125 mM malate, 2 mM ATP, 0.0625 mM CoA, and 1.25 mM DTT; pH 7.4) at 37°C containing [1-14C]palmitate bound to 0.5% BSA (200 μM for homogenates and 20 μM for mitochondria) for 1 h. In the liver, incomplete FAO is primarily the production of ketones. To approximate carnitine palmitoyltransferase (CPT)-1-mediated and non-CPT-1-mediated FAO in the liver homogenate, parallel incubations were included with the CPT-I inhibitor etomoxir (100 μM). Etomoxir-inhibitable FAO was calculated as the difference in parallel FAO and FAO in the presence of etomoxir incubations. The high concentration of etomoxir used in these experiments could potentially inhibit CPT-2 as well, impacting the FAO in these samples. However, the production of 14CO2 in the etomoxir samples suggests maintenance of mitochondrial capacity to uptake and oxidize acetyl units produced from peroxisomal FAO. The oxidation of 5 mM pyruvate ([2-14C]pyruvate) to 14CO2 by isolated liver mitochondria in the reaction buffer was used to monitor pyruvate oxidation and approximate TCA cycle flux (5, 28). The rate of pyruvate dehydrogenase complex and pyruvate carboxylase activities are responsible for the [2-14C]pyruvate-derived carbon entry into the TCA cycle and could be impacted by differential intrinsic aerobic capacity and chronic HFD. [2-14C]pyruvate oxidation as an assessment of TCA flux is limited by these factors. Data are presented as a relative value compared with the daily average to control for day-to-day variance.
Western blot analysis.
Triton X-100 cell lysates were used to produce Western blot-ready Laemmli samples. Samples were separated by SDS-PAGE, transferred to a polyvinylidene difluoride membrane and probed with primary antibodies. SOD2, sirtuin (Sirt)3, Sirt5, and acetylated lysine antibodies were purchased from Cell Signaling Technologies (Danvers, MA). 4-Hydroxynonenal (4-HNE) antibody was purchased from Alpha Diagnostics (San Antonio, TX). Oxpat antibody was purchased from American Research Products (Waltham, MA). Representative Western blots for each antibody are shown in Figs. 6C and 7C. Individual protein bands were quantified using a densitometer (Bio-Rad), and protein loading was corrected by 0.1% amido black (Sigma) staining to determine total protein, as previously described (48).
Acylcarnitines in the liver and plasma were measured on by API 3200 triple quadrapole liquid chromotography tandem mass spectroscopy as previously described (44). Liver and plasma free carnitine and acylcarnitines were extracted and derivatized. Individual acylcarnitine peaks were then quantified by comparison with a 13C internal standard (Cambridge Isotopes, Andover, MA). Metabolites were normalized to liver tissue weight (in g; Thermo Scientific, Rockford, IL).
Rat (liver) Genome 430 2.0 array.
The Rat (liver) Genome 430 2.0 array was performed by the Genome Technology Access Center at Washington University in St. Louis, MO. RNA was isolate using TRI reagent followed by cleanup using Qiagen mini columns including an on-column DNAse digestion step. Biotinylated cRNAs were prepared according to the standard Affymetrix protocol from 100 ng total RNA (Expression Analysis Technical Manual, Affymetrix, 2008). After fragmentation, 12.5 μg cRNA was hybridized for 16 h at 45°C on Rat Genome 430 2.0 microarrays. GeneChips were washed and stained using Affymetrix Fluidics Station 450 and scanned using Affymetrix GeneArray Scanner 7G06. We used the ranked gene list method to perform testing for enrichment of t-statistics in Gene Onocology biological proces terms, and gene sets with a nominal P < 0.001 and a false discovery rate q < 5% were considered as significant per the recommendations of the GSEA software manual.
The main effects of phenotype and diet were tested by two-way ANOVA. Where significant main effects were observed, post hoc analysis was performed using least significant difference to test for any specific pairwise differences. Two-way analysis of covariance (ANCOVA) was performed to test the impact of differences in animal size on the main effects of phenotype and diet for various outcome measures. All statistical analysis was performed with SPSS (SPSS, Armonk, NY). Statistical significance was set at P < 0.05.
Animal characteristics and blood measures.
Animal characteristics, including body weight and food/energy intake, are shown in Table 1. Body mass was 18% higher in LCR rats compared with HCR rats in all groups before the initiation (data not shown) and after 16 wk of HFDs. Furthermore, LCR rats had significantly increased weight gain with both 45% and 60% HFDs compared with the LFD. In contrast, neither of the HFDs altered weight gain in HCR rats, confirming previous reports that they are protected from HFD-induced weight gain (32, 34). The longitudinal changes in weight gain demonstrated that differences in HFD-induced weight gain in LCR rats largely occur during the first 3 wk of the HFD, after which weight gain occurs in parallel with the LFD (Fig. 1B). Retroperitoneal fat pad mass was higher in LCR than HCR rats across all diets, but both groups displayed an increase in fat pad mass on the 45% and 60% HFDs (Table 1).
Due to the higher energy density, both 45% and 60% HFDs led to reduced average weekly food intake (in g) compared with the control diet (Table 1). ANCOVA of food intake adjusting for the difference in body weight resulted in the same findings but also revealed that LCR rats consumed less food on both HFDs. Weekly energy intake (in kcal) was increased in both strains by 45% and 60% HFDs but only reached statistical significance in the 45% HFD (Table 1). Covariate body weight adjustment of energy intake was increased by only the 45% HFD in both strains. Unexpectedly, the 60% HFD in LCR rats led to a similar weekly rate of energy intake as the LFD. A further examination of these data longitudinally (Fig. 1A) revealed that LCR rats displayed dramatic swings in energy intake, which peaked at 600 kcal at the end of week 1 and was reduced to 325 kcal at the end of week 7, after which it reached steady-state levels from weeks 10 to 16 at ∼400 kcal per week.
Fasting glucose was higher in LCR rats than in HCR rats in all diet conditions, and only LCR rats increased glucose after the 60% HFD (Table 2). Plasma insulin and serum triglycerides did not differ among groups, and circulating free fatty acids (FFAs) and β-hydroxybutyrate were significantly lower in LCR than HCR rats with all diets (Table 2). In addition, only LCR rats displayed a reduction in β-hydroxybutyrate on the 60% HFD. The 60% HFD lowered FFAs in both HCR and LCR rats compared with other diet conditions.
LCR rats had higher liver triglycerides across all diets. Liver triglycerides were increased by the chronic 60% HFD in only LCR rats, whereas HCR rats displayed no change in liver triglycerides compared with the control diet (Fig. 2B). This was confirmed by representative hematoxylin and eosin-stained images clearly showing that only the 60% HFD LCR livers experienced a significant accumulation of lipid droplets (Fig. 2A).
Liver FAO capacity and mitochondrial content.
We first measured FAO capacity in the whole liver homogenate, a condition in which both mitochondrial and extramitochondrial contributions to FAO are assessed. Complete FAO to CO2 in the liver homogenate was significantly lower in LCR rats than in HCR rats in all conditions (Fig. 3A). Additionally, complete FAO was suppressed with 45% and 60% HFDs in both HCR and LCR rats. Both groups increased liver homogenate acid-soluble metabolites (ASM) production on the 60% HFD, an effect primarily driven by the increase in the LCR group (Fig. 3B). FAO in the liver homogenate was also assessed in the presence of a CPT-1 inhibitor, etomoxir, to examine the impact of non-CPT-1-mediated FAO. It is presumed that oxidation of palmitate through this process occurs after long-chain fatty acid shortening occurring in peroxisomes before acetyl-CoA entry into the mitochondria. Interestingly, HCR rats maintained a significantly higher rate of both complete FAO and ASM production compared with LCR rats in the presence of etomoxir (Fig. 3, C and D). We also deduced “etomoxir-inhabitable” complete FAO and ASM production by subtracting the difference for FAO with and without etomoxir. Again, there was a trend (P < 0.07) for HCR rats to have higher complete FAO across all diets, whereas the 60% HFD lowered this measure in only HCR rats (Fig. 3E). However, there was no difference between strains for etomoxir-inhabitable ASM production, suggesting that long-chain fatty acids entering the mitochondria through CPT-1 were incompletely oxidized at the same rate between HCR and LCR rats (Fig. 3F).
Liver mitochondrial metabolism.
To assess hepatic mitochondrial content, we measured the mitochondrial marker electron transport chain complex IV-subunit I protein content (Fig. 3G). As we have previously shown (48), the ∼20% lower hepatic complex IV protein content in LCR compared with HCR rats suggests lower liver mitochondrial content. An additional marker of mitochondrial content, citrate synthase enzyme activity, showed a similar trend (data not shown).
We also measured FAO capacity, pyruvate oxidation, and respiration in isolated hepatic mitochondria. 2-[14C]pyruvate oxidation to 14CO2 is an index of TCA cycle flux as the second labeled carbon is cycled twice through the TCA cycle before conversion to CO2 (1, 28). 2-Pyruvate oxidation was reduced in LCR compared with HCR rats across all diets, and this was further suppressed with the 60% HFD in LCR rats only (Fig. 4A). Hepatic mitochondrial complete FAO to CO2 was also significantly lower in LCR compared with HCR rats (Fig. 4B), with LCR rats being much more susceptible to HFD-diet induced reduction in complete FAO to CO2. Interestingly, these differences in mitochondrial complete FAO were not due to difference in mitochondrial CPT-1α protein content (Fig. 6A), where no effect of strain was observed but diet increased CPT-1α protein content in both strains.
Hepatic mitochondrial respiration of palmitoyl-carnitine was lower in LCR compared with HCR rats in state 2, state 3, and state 3 + succinate conditions (Fig. 4, C–E). Interestingly, the 60% HFD reduced respiration only in HCR rats in the state 3 + succinate condition. Finally, maximal uncoupled mitochondrial respiration of palmitoyl-carnitine (with FCCP) was lower in LCR rats than in HCR rats, demonstrating a reduced capacity of long-chain FFAs to be used for respiration (Fig. 4F). In addition, there was a main effect for the HFDs to lower maximal uncoupled respiration in the mitochondria of both strains. Overall, the mitochondrial respiratory experiments suggested that both steady-state (state II and III) and uncoupled respiratory rates were lower in LCR rats than in HCR rats.
Transcriptional responses to an acute HFD.
We performed rat genome arrays followed by gene set enrichment analysis in livers from HCR/LCR rats that were fed an acute (3-day) HFD (45% kcal from fat) as previously reported (27). Gene set enrichment analysis revealed unique transcriptional differences in HCR/LCR rats at the onset of HFD conditions that would be predictive of changes that occurred with chronic HFD exposure. LCR rats displayed a concerted reduction in the expression of TCA cycle genes, results that were not witnessed in HCR rats (Fig. 5). These included significant reductions in phosphoenolpyruvate carboxykinase 1 (Pck1), aconitase 1 (Aco1), ATP citrate lyase (Acly), succinate-CoA ligase α-subunit (Sucgl1), and succinate-CoA ligase GDP-forming β-subunit (Suclg2).
Liver and plasma acylcarnitine profiles.
Acylcarnitine profiling of the serum and liver (Table 3) was quantified to further assess the effects of HFD feeding on hepatic lipid metabolism. Acylcarnitine esters represent byproducts of mitochondrial substrate metabolism and have been previously used to estimate changes in metabolic flux (3). Unlike what has previously been shown in the skeletal muscle of HCR/LCR rats (33), hepatic-free carnitine levels did not decrease with chronic HFD feeding; rather, the 60% HFD increased hepatic-free carnitine levels in both strains. Plasma carnitine levels were higher across all groups in LCR rats than in HCR rats, and only HCR rats displayed a decrease on the 45% and 60% HFDs. Butyryl-carnitine (C4) was decreased by the 45% and 60% HFDs in both the plasma and liver of both strains, but HCR rats displayed a greater decline in plasma butyryl-carnitine after the 60% HFD. Hepatic octanoyl-carnitine (C8), myristoyl-carnitine (C14), and palmitoyl-carnitine (C16) were all lower in LCR rats than in HCR rats regardless of diet. Increased long-chain acylcarnitines could be indicative of greater FFA uptake in HCR rats. In addition, only HCR livers displayed an increase in octanoyl-carnitine (both 45% and 60% HFDs) and myristoyl-carnitine (60% HFD only) in response to the HFD. There were no differences between strains for plasma medium- and long-chain acylcarnitines (C8, C14, and C16; Table 2). Overall, HFD feeding lowered plasma long-chain acylcarnitine levels (C14 and C16) in both strains, and LCR rats displayed overall lower plasma C14 acylcarnitine across all diets compared with HCR rats.
Comparison of short-chain and long-chain acylcarnitines within a tissue can also provide insights into β-oxidation. HCR rats displayed higher levels of hepatic medium- and long-chain acylcarnitines (C8, C14, and C16) than LCR rats but maintained similar levels of hepatic acetylcarnitine levels. These data were not explained by a similar profile in plasma acylcarnitines. Ratios between shorter- and longer-chain acylcarnitines in tissues have been previously used to estimate flux through β-oxidation (47). The C2-to-C4 ratio was greater in HCR rats compared with LCR rats regardless of diet. Additionally, there was a main effect for both 45% and 60% HFDs increasing the C2-to-C4 ratio. Interestingly, an opposite relationship was observed for the C4-to-C16 ratio, with HCR rats being lower than LCR rats regardless of diet and both diets resulting in significant reductions. Increased C2-to-C4 and reduced C4-to-C16 ratios are indicative of increased complete FAO with a subsequent reduction in C4. While the interpretation of acylcarnitine profiles is subjective and open to interpretation, when combined with our other outcome measures, these data suggest higher rates of fatty acid uptake, β-oxidation, and TCA cycle flux in HCR rats than in LCR rats, supporting the divergent patterns that emanated from the ex vivo oxidation assays in both strains.
Mitochondrial oxidative stress and acetylation state.
We further examined if hepatic mitochondria from the two models had different levels of oxidative stress by measuring 4-HNE protein in isolated mitochondria. No difference was observed in protein 4-HNE content; however, LCR rats fed the 60% HFD tended to have higher 4-HNE mitochondrial content compared with the LFD (Fig. 6B). SOD protein content in isolated mitochondria was increased by the HFDs in both strains (Fig. 6D), suggesting that the HFD increased processes to mitigate increased mitochondrial oxidative stress, a response that was likely sufficient given the insignificant effects upon 4-HNE. Finally, we also measured hepatic mitochondrial uncoupling protein 2 (UCP2) protein content to assess if the state II respiratory differences could be explained by increased uncoupling capacity in HCR rats. On the LFD, UCP2 protein was lower in LCR rats than in HCR rats; however, both HFD conditions increased UCP2 in only LCR rats (Fig. 6E).
Given the clear differences in hepatic mitochondrial FAO and respiration found between HCR and LCR rats, we also examined mitochondrial sirtuin and mitochondrial acetylation states, factors that putatively modulate mitochondrial function. Hepatic mitochondrial Sirt3 protein, a deacetylase found exclusively in mitochondria, was differentially altered by the HFD in HCR and LCR rats, with the 60% HFD lowering Sirt3 protein in HCR rats while increasing Sirt3 in LCR rats (Fig. 7A). Sirt5, an enzyme with deacetylase, succinylase, and malonylase capacities that is also located in the mitochondria, was found to be lower in LCR rats than in HCR rats, primarily due to an effect of the 45% and 60% HFDs to reduce its expression (Fig. 7D). In HCR rats, only the 60% HFD reduced Sirt5 in hepatic mitochondria. Finally, we also examined the acetylation and succinylation status of isolated mitochondria due to a report (4) showing that HFD-induced increases in both factors alter mitochondrial function. Changes in mitochondrial Sirt3 and Sirt5 protein content were not associated with similar changes in mitochondrial protein acetylation or succinylation status (Fig. 7, B and E). No differences in mitochondrial protein acetylation was observed between strains, and the 60% HFD resulted in similar increases in mitochondrial acetylation levels in both HCR and LCR rats (Fig. 7B). No significant differences in mitochondrial protein succinylation were observed between strains or among diets (Fig. 7E).
These results confirm that rats bred for high aerobic capacity are protected against hepatic steatosis after chronic HFD feeding, whereas rats bred for low aerobic capacity are susceptible, specifically with the chronic consumption of a 60% kcal HFD. The protection against dietary-induced hepatic steatosis witnessed in HCR rats is associated with higher hepatic complete FAO capacity and higher hepatic mitochondrial respiratory capacity that was largely unchanged after chronic consumption of a HFD. In comparison, LCR rats displayed lower initial complete FAO capacity in both whole homogenates and isolated mitochondria. In addition, only LCR rats displayed a robust reduction in hepatic mitochondrial complete FAO capacity (both 45% and 60% HFDs) and an index of TCA cycle flux (2-pyruvate oxidation in isolated mitochondria) on the 60% HFD. These chronic HFD-induced reductions in complete FAO capacity and TCA cycle flux were further confirmed by evidence that acute exposure to a 3-day HFD lowered TCA cycle genes selectively only in LCR rats. Finally, HCR livers also displayed approximately twofold higher levels of long- and medium-chain acylcarnitines despite maintaining the same levels of acetylcarnitine, further supporting the higher rates of hepatic fatty acid uptake and β-oxidative flux.
Aerobic capacity is impacted by lifestyle (regular exercise or activity vs. physical inactivity), genetics, and age. There is a large degree of genetic heterogeneity for aerobic capacity at any given age. Bouchard et al. (8) showed that sedentary, nonexercising young men at an average age of 30 yr had a large heterogeneous range of aerobic capacity (from 29 to 53 ml·kg−1·min−1) with a mean of 41 ml·kg−1·min−1 and a large SD (±9 ml·kg−1·min−1). These human data highlight the importance of genetic predisposition for aerobic capacity and validate our chose of the HCR/LCR model to investigate mechanisms by which intrinsic aerobic capacity impacts susceptibility for metabolic conditions and other disease states.
The development of NAFLD has been linked to a number of factors, including obesity, race, sex, sedentary lifestyle, physical inactivity, diet composition, insulin resistance, or diabetes status (39, 40). Several reports have indicated that aerobic capacity uniquely impacts liver health and liver fat. Baseline cardiorespiratory fitness was shown to correlate with the degree to which exercise and weight loss lowers liver fat (20) or the degree to which obesity tracks with liver fat (10). Additionally, an examination of monozygotic twins with disparate aerobic capacity levels revealed that the lower fit sibling had greater liver fat, despite no significant difference in body mass index (14). Finally, liver failure patients who display the lowest aerobic capacity have the lowest survivability after liver transplants (11). The human evidence that whole body aerobic capacity impacts liver health is noteworthy and again signifies the needs to better understand underlying mechanisms.
While hepatic FAO, mitochondrial respiration, TCA cycle flux, gluconeogenesis, and overall fatty acid handling are accelerated in hepatic steatosis under some conditions (22, 44, 45), we hypothesized that higher or lower liver complete FAO and mitochondrial respiratory capacity associated with contrasting fitness levels would impact susceptibility for hepatic steatosis. Our results support our previous studies showing that complete FAO in whole liver homogenates and isolated mitochondria are dramatically different between HCR and LCR rats (27, 48). Increased complete FAO to CO2 represents more efficient coupling of acetyl-CoA produced via β-oxidation to flux through the TCA cycle, which has been proposed to have metabolic benefits (29). As such, we have also previously reported that complete FAO capacity is increased when we effectively prevent or treat fatty liver with exercise training in hyperphagic, obese Otsuka Long-Evans Tokushima fatty (OLETF) rats with hepatic steatosis (24, 25, 41). Moreover, we have shown that complete FAO capacity is reduced in young OLETF rats before hepatic steatosis or other metabolic factors that may drive risk for fatty liver, suggesting that hepatic complete FAO likely plays a critical role in the development of steatosis (42). In addition, we previously have reported that the liver-specific overexpression of peroxisome proliferator activated receptor-γ coactivator-1α (PGC-1α) increased hepatic mitochondrial complete FAO capacity in accordance with reduced hepatic triglyceride storage and secretion both in vivo and in primary hepatocytes (28). Furthermore, chronic overexpression of a mutated constitutively active enzyme controlling long-chain fatty acid entry into mitochondria (CPT-1) via adeno-associated virus reduced HFD-induced steatosis and weight gain (35). In addition, livers with overexpression of constitutively activated CPT-1 showed characteristics that matched HCR rats, including higher rates of complete FAO, elevated circulating ketones, and reduced epididymal fat pad mass. Hence, our results in the HCR/LCR model support previous work that hepatic complete FAO impacts susceptibility for hepatic steatosis.
Beyond the baseline differences in capacity for complete FAO capacity, LCR rats also displayed a reduction in mitochondrial complete FAO capacity on 45% and 60% HFDs. In addition, the 60% HFD induced a near collapse of 2-pyruvate oxidation, a marker of TCA cycle flux, in only LCR rats. It is likely that these HFD-induced alterations in LCR rats were primarily driven by transcriptional alterations as only 3 days on a 45% HFD reduced TCA cycle genes in LCR rats but not HCR rats. A great deal of work has examined changes in mitochondrial content and function after HFDs in the liver and skeletal muscle. Changes in both tissues have yielded different results likely attributed to the differences in the sources of fatty acids in the diet, the length of feeding, and the type of outcome measures used (in vitro vs. in vivo, etc.) (13, 19). Importantly, expression of enzymes that control posttranslational modifications including deacetylases (Sirt3 and Sirt5) and mitochondrial acetylation and succinylation levels were not different between HCR and LCR rats, despite increasing with the 60% HFD. A previous report (4) has suggested that increased acetylation leads to decreased functionality of mitochondrial proteins, but if this is the case, they only tracked with changes in LCR rats as there was no decline in mitochondrial FAO capacity or respiration observed in HCR rats. It is possible that acetylation or other posttranslational modifications may impact mitochondrial machinery differently between strains, and this deserves further examination in the future.
It is conceptually important to understand why endurance or aerobic capacity would impact hepatic mitochondrial respiratory and FAO capacity. Ex vivo mitochondrial FAO and respiration measures in the presence of maximal substrates assess the maximal capacity of oxidative phosphorylation. In contrast, rates of mitochondrial FAO and respiration in vivo are driven by systemic energy demand. Mitochondrial density and oxidative capacity are typically adaptations to the chronic level of energy demand within a cell or tissue. We have shown that resting energy expenditure is higher in HCR than LCR rats, providing a link between strain differences for whole body energy metabolism and hepatic mitochondrial oxidative capacity (27). Prolonged endurance exercise stimulates a large energy demand in the liver due to the increased rate of hepatic glucose production (gluconeogenesis and glycogenolysis) so that euglycemia is maintained (49). Liver FAO is the primary catabolic process producing the necessary ATP needed to fuel the energy costly process of gluconeogenesis. As proof of the link between hepatic FAO and gluconeogenesis, previous research has shown that disruption of hepatic FAO leads to impairments in gluconeogenic capacity (50). Therefore, by selectively breeding for high or low endurance/aerobic capacity, alterations in mitochondrial respiratory and FAO capacity and presumably gluconeogenic flux have also been altered over time in both strains. As evidence of this, we have consistently observed higher PGC-1α gene expression in HCR over LCR livers (unpublished observations). PGC-1α is a transcriptional coactivator that dually regulates both mitochondrial respiratory function, FAO, and gluconeogenesis (26), linking these two important pathways at a transcriptional level.
We also assessed measures of oxidative stress to determine if HCR or LCR rats differently responded to chronic HFD intake. Only LCR rats showed a trend for an increase in 4-HNE on the 60% HFD, whereas no changes were found in HCR rats. However, both strains had increased SOD2 protein content in isolated mitochondria after both HFDs. Moreover, the response was graded such that the 45% HFD caused a 30–40% increase while the 60% HFD caused a 60–70% increase. These data suggest that the HFD did indeed induce increased mitochondrial oxidative stress in the livers of both strains, but that compensatory antioxidant pathway activation was sufficient to counter significant hepatic changes in oxidative stress.
Despite the clear differences in hepatic mitochondrial oxidative capacity between HCR and LCR rats, it is also likely that whole body differences in metabolic function drive differences in susceptibility for hepatic steatosis. These whole body differences in HCR and LCR metabolic phenotypes have been clearly outlined in our previous report (27) examining hepatic steatosis susceptibility with an acute HFD. Major differences in LCR rats compared with HCR rats included a higher energy intake, lower energy expenditure, lower whole body dietary fat oxidation, and higher rate of dietary fatty acids trafficked to adipose stores rather than skeletal muscle. Similar differences in energy intake are reported here, where LCR rats on both 45% and 60% HFDs had higher energy intake than HCR rats during the first 4 wk of the diet, after which energy intake was similar between strains and among diet groups. As a result, the change in weight gain from baseline is only greater in HFD-fed LCR rats during the first 2 wk of the HFD. Importantly, previous work has suggested that hepatic FAO and energy state (ATP) can impact control of energy intake (16), effects that we are currently examining in HCR/LCR rats (unpublished observations). Therefore, it appears that selection for intrinsic aerobic capacity influences a number of in vivo factors that also likely alter susceptibility for NAFLD.
In conclusion, rats bred for high and low intrinsic aerobic capacity display different susceptibility for hepatic steatosis after chronic HFD feeding. Moreover, baseline differences in hepatic mitochondrial complete FAO and respiratory capacity, which are a result of selective breeding for running capacity, appear to impact the differences in susceptibility for chronic HFD-induced hepatic steatosis. Furthermore, rats with low aerobic capacity display transcriptional declines in TCA cycle genes at the onset of a HFD and pronounced reductions in mitochondrial oxidative capacity after chronic exposure to HFDs. Overall, these data suggest that baseline hepatic complete FAO and overall oxidative capacity as well as the maintenance of these factors are critical for protection against hepatic steatosis.
This work was partially supported by National Institutes of Health (NIH) Grants DK-088940 (to J. P. Thyfault), 5-T32-AR-48523-8 (to E. M. Morris), and RO1-DK-078184 (to S. C. Burgess), American Heart Association Grant 14POST20110034 (to E. M. Morris), NIHVHA-CDA2 IK2BX001299 (to R. S. Rector), Veterans Affairs (VA) Merit Review Grant 1I01BX002567-01 (to J. P. Thyfault), Robert A. Welch Foundation Grant I-1804 (to S. C. Burgess), and United States Department of Agriculture Grant CRIS 6206-51000-010-05S (to K. Shankar). This work was supported with resources and the use of facilities at the Harry S. Truman Memorial VA Hospital in Columbia, MO. The LCR-HCR rat model system was funded by the Office of Research Infrastructure Programs (NIH Grant P40-OD-021331 to L. G. Koch and S. L. Britton).
We thank the Genome Technology Access Center in the Department of Genetics at Washington University School of Medicine for help with genomic analysis. The center is partially supported by NIH Cancer Center Support Grant P30-CA-91842 to the Siteman Cancer Center and by Grant UL1-TR-000448, a component of NIH and NIH Roadmap for Medical Research.
No conflicts of interest, financial or otherwise, are declared by the author(s).
E.M.M., L.G.K., S.L.B., J.I., R.S.R., and J.P.T. conception and design of research; E.M.M., G.M.M., J.A.F., K.S., and X.F. performed experiments; E.M.M., G.M.M., J.A.F., K.S., X.F., S.C.B., and J.P.T. analyzed data; E.M.M., K.S., S.C.B., R.S.R., and J.P.T. interpreted results of experiments; E.M.M. and K.S. prepared figures; E.M.M. and J.P.T. drafted manuscript; E.M.M., J.A.F., X.F., and J.P.T. edited and revised manuscript; E.M.M., G.M.M., L.G.K., S.L.B., J.A.F., K.S., X.F., S.C.B., J.I., R.S.R., and J.P.T. approved final version of manuscript.
We acknowledge the expert care of the rat colony provided by Molly Kalahar and Lori Heckenkamp. Contact L. G. Koch () or S. L. Britton ( ) for information on the LCR and HCR rats: these rat models are maintained as an international resource with support from the Department of Anesthesiology at the University of Michigan (Ann Arbor, MI).