The typical western diet contains large quantities of PUFA that are heated or processed to varying degrees. In fast-food restaurants fat is heated in fryers for up to 18 h daily, at temperatures close to 180°C (Frankel et al. Reference Frankel, Smith, Hamblin, Creveling and Clifford1984). Several studies with animals have been performed to investigate the effects of oxidised fats on the metabolism (reviewed in Cohn, Reference Cohn2002). Recently, it has been shown in rats that oxidised fats are able to influence the lipid metabolism by activation of PPARα (Chao et al. Reference Chao, Chao, Lin and Huang2001, Reference Chao, Hsu, Lin, Li and Huang2004, Reference Chao, Yang, Tseng, Chang, Lu and Huang2005; Sülzle et al. Reference Sülzle, Hirche and Eder2004), a transcription factor belonging to the nuclear hormone receptor superfamily (Schoonjans et al. Reference Schoonjans, Staels and Auwerx1996). This is probably due to the occurrence of hydroxy- and hydroperoxy fatty acids such as hydroxy octadecadienoic acid and hydroperoxy octadecadienoic acid which are potent activators of PPARα (Delerive et al. Reference Delerive, Furman, Teissier, Fruchart, Duriez and Staels2000; Mishra et al. Reference Mishra, Chaudhary and Sethi2004; König & Eder, Reference König and Eder2006). Activation of PPARα leads to an increase in the transcription of genes related to fatty acid transport across the cell membrane, intracellular lipid trafficking, mitochondrial and peroxisomal fatty acid uptake, and both mitochondrial and peroxisomal fatty acid β-oxidation, gluconeogenesis and ketogenesis (Mandard et al. Reference Mandard, Muller and Kersten2004). Recently, it has been shown that PPARα activation influences also the expression or the proteolytic activation of sterol regulatory element-binding proteins (SREBP), transcription factors which control fatty acid synthesis and cholesterol homeostasis (Patel et al. Reference Patel, Knight, Wiggins, Humphries and Gibbons2001; Guo et al. Reference Guo, Wang, Milot, Ippolito, Hernandez, Burton, Wright and Chao2001; Knight et al. Reference Knight, Hebbach, Hauton, Brown, Wiggins and Patel2005; König et al. Reference König, Koch, Spielmann, Hilgenfeld, Stangl and Eder2006). Therefore, PPARα activation stimulates not only the degradation of fatty acids by enhancing β-oxidation but affects also the synthesis of cholesterol and TAG. Reduced liver and plasma concentrations of TAG and cholesterol are typical effects observed in animals treated with PPARα agonists, and such effects have been also observed in rats administered oxidised fats (Huang et al. Reference Huang, Cheung and Lu1988; Eder & Kirchgessner, Reference Eder and Kirchgessner1998; Eder, Reference Eder1999; Chao et al. Reference Chao, Chao, Lin and Huang2001, Reference Chao, Hsu, Lin, Li and Huang2004, Reference Chao, Yang, Tseng, Chang, Lu and Huang2005; Sülzle et al. Reference Sülzle, Hirche and Eder2004).
Regarding the expression of PPARα in tissues and the effects of PPARα activation on transcription of its target genes, there are great differences between various species. In rodents, PPARα is highly expressed, and activation of PPARα not only induces many genes involved in various metabolic pathways such as β-oxidation, ketogenesis and gluconeogenesis but also causes severe peroxisome proliferation in the liver (Peters et al. Reference Peters, Cheung and Gonzalez2005). In contrast to rodents, PPARα agonists do not induce peroxisome proliferation in the liver of many other species, such as guinea pigs, swine, monkeys and man (Holden & Tugwood, Reference Holden and Tugwood1999). These non-proliferating species have a lower expression of PPARα in the liver and the response of many genes to PPARα activation is much weaker than in proliferating species. For that reason, effects related to PPARα activation observed in rodents cannot be directly applied for non-proliferating species such as man. Therefore, it remains unknown whether oxidised fats are able to cause PPARα activation also in non-proliferating species.
The aim of the present study was to investigate whether a dietary oxidised fat, prepared by heating sunflower oil under usual deep-frying conditions (180°C) for 24 h in a deep fryer, is able to activate PPARα and to cause peroxisome proliferation in pigs. Pigs have been chosen as a model since they belong – like man – to the non-proliferating species (Yu et al. Reference Yu, Odle and Drackley2001; Peffer et al. Reference Peffer, Lin and Odle2005) and since pig liver cells show a similarity to human liver cells in the gene response to PPARα agonists (Cheon et al. Reference Cheon, Nara, Band, Beever, Wallig and Nakamura2005). We focused our analyses on liver and small intestine as both tissues exhibit a high expression of PPARα (Braissant et al. Reference Braissant, Foufelle, Scotto, Dauca and Wahli1996; Lemberger et al. Reference Lemberger, Braissant, Juge-Aubry, Keller, Saladin, Staels, Auwerx, Burger, Meier and Wahli1996). Moreover, both tissues play an important role in whole body lipid homeostasis, i.e. synthesis and secretion of lipoproteins rich in TAG and cholesterol (Lindsay & Wilson, Reference Lindsay and Wilson1965; Dietschy et al. Reference Dietschy, Turley and Spady1993). We examined the expression of various genes involved in lipid metabolism which have been already shown to be influenced by PPARα activation. Furthermore, in both tissues we determined gene expression of SREBP and important SREBP target genes involved in fatty acid synthesis and cholesterol uptake and synthesis.
Materials and methods
Animals
For the experiment, eighteen male 8-week-old crossbred pigs ((German Landrace × Large White) × Pietrain) were kept in a room under controlled temperature at 23 ± 2°C and 55 ± 5 % relative humidity with light from 06.00 to 18.00 hours. One day before the beginning of the experimental feeding period, the pigs were weighed and randomly allocated to two groups with body weights of 12·0 (sd 1·1) kg in the control group and 12·2 (sd 0·9) kg in the treatment group. All experimental procedures described followed established guidelines for the care and use of laboratory animals and were approved by the local veterinary office.
Diets and feeding
Both groups of pigs received a nutritionally adequate diet for growing pigs containing (in g/kg) wheat (400), soyabean meal (230), wheat bran (150), barley (100), sunflower oil or test oil (90), and mineral premix including l-lysine, dl-methionine and l-threonine (30). This diet contained 14·4 MJ metabolisable energy and 185 g crude protein/kg. Diet intake was controlled, and each animal in the experiment was offered an identical amount of diet per day. During the feeding period, the amount of diet offered each day was increased continuously from 400 to 1200 g. The pigs had free access to water via nipple drinking systems. The experimental diets were administered for 28 d.
Preparation of the test fats
To prepare the oxidised fat, sunflower oil obtained from a local supermarket was heated at a temperature of 180°C for 24 h in a deep fryer. This treatment caused a loss of PUFA and tocopherols. The major fatty acids in the fresh and the oxidised fat, respectively, were (g/100 g total fatty acids): palmitic acid (16 : 0), 6·30 v. 6·70; stearic acid (18 : 0), 4·0 v. 4·2; oleic acid (18 : 1n-9), 22·8 v. 23·8; linoleic acid (18 : 2n-6), 63·6 v. 59·9. Other fatty acids were present only in small amounts ( < 0·5 g/100 g fatty acids). To equalise the fatty acid composition of the fresh and the oxidised fat, the fresh fat was composed of a mixture of sunflower oil and palm oil (93 : 7, w/w). To adjust dietary vitamin E concentrations, we analysed the native concentrations of tocopherols in the fresh fat and in the oxidised fats after the thermal treatment. With consideration of the native tocopherol concentrations of the dietary fats, the diets were supplemented individually with all-rac-α-tocopheryl acetate (the biopotency of all-rac-α-tocopheryl acetate is considered to be 67 % of that of α-tocopherol). The final vitamin E concentration was 620 mg α-tocopherol equivalents/kg in both fats. Concentrations of lipid peroxidation products were determined after the fats have been already included into the diets. Therefore, lipids of the diets were extracted by n-hexane and isopropanol (3 : 2, v/v; Hara & Radin, Reference Hara and Radin1978). Concentration of thiobarbituric acid-reactive substances (TBARS; Sidwell et al. Reference Sidwell, Salwin, Benca and Mitchell1954), conjugated dienes (Recknagel & Glende, Reference Recknagel and Glende1984), peroxide value (Deutsche Gesellschaft für Fettwissenschaft, 1994), acid value (Deutsche Gesellschaft für Fettwissenschaft, 1994) and concentration of total carbonyls (Endo et al. Reference Endo, Li, Tagiri-Endo and Fujimoto2001) were determined in the extracted fat.
Sample collection
After completion of the feeding period the animals were killed under light anaesthesia. Each pig was fed its respective diet 4 h before being killed. After killing, blood was collected into heparinised polyethylene tubes. Plasma was obtained by centrifugation of the blood (1100 g, 10 min, 4°C). Plasma lipoproteins were separated by step-wise ultracentrifugation (Mikro-Ultrazentrifuge; Sorvall Products, Bad Homburg, Germany) at 900 000 g at 4°C for 1·5 h. Plasma densities were adjusted by sodium chloride and potassium bromide and the lipoprotein fractions δ < 1·006 kg/l VLDL plus chylomicrons, 1·006 < δ < 1·063 kg/l LDL and δ>1·063 kg/l HDL were removed by suction. The liver was dissected and weighted and samples were stored at − 80°C until analysis. For preparation of liver homogenate, 1 g liver tissue was homogenised in PBS by TissueLyser (Qiagen, Haan, Germany), centrifuged at 600 g for 10 min at 4°C and the supernatant was stored at − 20°C until analysis. For isolation of intestinal epithelial cells, the abdomen was immediately opened after killing, and a 35 cm intestinal segment was dissected starting at 30 cm distal to the pyloric sphincter, and flushed twice with ice-cold wash buffer (PBS containing 0·2 mm-phenylmethylsulphonyl fluoride and 0·5 mm-dithiothreitol, pH 7·4). The isolation of porcine intestinal epithelial cells was performed by the modified distended intestinal sac technique according to Fan et al. (Reference Fan, Matthews, Etienne, Stoll, Lackeyram and Burrin2004). In brief, the intestinal segments were filled with 100 ml preincubation buffer (PBS containing 27 mm-sodium citrate, 0·2 mm-phenylmethylsulphonyl fluoride and 0·5 mm-dithiothreitol, pH 7·4), sealed with strings and filled intestinal segments were incubated in a saline bath (154 mm-NaCl) for 15 min at 37°C. Afterwards, the pre-incubation buffer was discarded, and the intestinal segments were filled with isolation buffer (PBS containing 1·5 mm-Na2EDTA, 0·2 mm-phenylmethylsulphonyl fluoride, 0·5 mm-dithiothreitol and 2 mm-d-glucose, pH 7·4). Two major cell fractions, consisting of the upper and the crypt cell fraction, were sequentially isolated from intestinal segments through two consecutive incubations with isolation buffer at 37°C for 40 (upper cell fraction) and 60 min (crypt cell fraction), respectively. Each cell fraction was collected separately, and washed twice with ice-cold PBS. Afterwards, cells were retained by centrifugation (400 g, 4 min, 4°C) and immediately frozen at − 80°C. For further analysis, we used the crypt cell fraction as it has been shown that these cells have a 6–8-fold higher capacity of lipid synthesis than villus cells (Shakir et al. Reference Shakir, Sundaram and Margolis1978).
Lipid analysis
Lipids from liver were extracted with a mixture of n-hexane and isopropanol (3 : 2, v/v; Hara & Radin, Reference Hara and Radin1978). For determination of the concentrations of lipids in liver, aliquots of the lipid extracts were dried and the lipids were dissolved using Triton X-100 (De Hoff et al. Reference De Hoff, Davidson and Kritchevsky1978). Concentrations of TAG and cholesterol in plasma and lipoproteins and those of liver were determined using enzymatic reagent kits (cat. no. 113009990314 for cholesterol and cat. no. 157609990314 for TAG; Ecoline S+, DiaSys, Holzheim, Germany).
Preparation of liver microsomal and cytosolic fractions
Liver (1 g) was homogenised in 10 ml 0·1 m-phosphate buffer, pH 7·4, containing 0·25 m-sucrose using a Potter-Elvehjem homogeniser. Homogenates were centrifuged at 1000 g for 10 min at 4°C, and the supernatant was centrifuged at 15 000 g for a further 15 min. The microsomal pellet was obtained by centrifugation of the 15 000 g supernatant at 105 000 g for 60 min. The resulting cytosolic fraction in the supernatant was separated, microsomal pellets were resuspended in the homogenisation buffer and all samples were stored at − 20°C for further analysis. The protein concentrations of cytosolic and microsomal fractions were determined with the BCA reagent according to the protocol of the supplier (Interchim, Montelucon, France) using bovine serum albumin as standard.
RT–PCR analysis
Total RNA from liver tissue and enterocytes, respectively, was isolated by the TissueLyser (Qiagen) using Trizol reagent (Invitrogen, Karlsruhe, Germany) according to the manufacturer's protocol. RNA concentration and purity were estimated from the optical density at 260 and 280 nm (SpectraFluor Plus; Tecan, Crailsheim, Germany). The quality of all RNA samples was furthermore assessed by agarose gel electrophoresis. Total RNA (1·2 μg) was used for cDNA synthesis as described previously (König & Eder, Reference König and Eder2006). The mRNA concentration of genes was measured by real-time detection PCR using SYBR® Green I and the Rotor Gene 2000 system (Corbett Research, Mortlake, Australia). Real-time detection PCR was performed with 1·25 U Taq DNA polymerase, 500 μm-dNTP and 26·7 pmol of the specific primers. For determination of mRNA concentration a threshold cycle (C t) and amplification efficiency was obtained from each amplification curve using the software RotorGene 4·6 (Corbett Research). Calculation of the relative mRNA concentration was made using the ΔΔC t method as previously described (Pfaffl, Reference Pfaffl2001). The housekeeping gene glyceraldehyde-3-phosphate dehydrogenase was used for normalisation. The PCR primers used for real-time RT–PCR were obtained from Operon (Köln, Germany) and Roth (Karlsruhe, Germany), respectively, and are listed in Table 1.
Table 1 Characteristics of the specific primers used for RT–PCR analysis
ACC, acetyl-CoA carboxylase; ACO, acyl-CoA oxidase; CPT-1, carnitine palmitoyltransferase-1; CYP7, cholesterol 7α-hydroxylase; FAS, fatty acid synthase; FATP, fatty acid transport protein; FDPS, farnesyl diphosphate synthase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HMG-CoA-R, 3-hydroxy-3-methylglutaryl-CoA reductase; I-FABP, intestinal fatty acid binding protein; Insig, insulin-induced gene; L-FABP, liver fatty acid binding protein; mAAT, mitochondrial aspartate aminotransferase; mHMG-CoA-S, mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase; MTP, microsomal TAG transfer protein; NPC, Niemann-Pick type C; SCD, stearoyl-CoA desaturase; SOD, superoxide dismutase; SREBP, sterol regulatory element-binding protein.
Enzyme assays
Superoxide dismutase (SOD) activity in liver cytosol was determined according to the method of Marklund & Marklund (Reference Marklund and Marklund1974) with pyrogallol as the substrate. One unit of SOD activity is defined as the amount of enzyme required to inhibit the autoxidation of pyrogallol by 50 %. The activity of glutathione peroxidase in liver cytosol was determined with t-butyl hydroperoxide as substrate according to the method of Paglia & Valentine (Reference Paglia and Valentine1967). One unit of glutathione peroxidase activity is defined as 1 μmol reduced β-nicotinamide adenine dinucleotide phosphate oxidised/min. The activity of glutathione S-transferase was determined using 1-chloro-2,4-dinitrobenzene as substrate as described by Habig et al. (Reference Habig, Pabst and Jakoby1974). One unit of glutathione S-transferase is defined as one nmol substrate consumed/min. Catalase activity in liver homogenate was determined using H2O2 as substrate according to the method of Aebi (Reference Aebi and Bergemeyer1986). One unit of catalase activity is defined as the amount consuming 1 mmol H2O2/min.
Determination of conjugated dienes, thiobarbituric acid-reactive substances and α-tocopherol
Lipids from liver were extracted using a mixture of n-hexane and isopropanol (3 : 2, v/v; Hara & Radin, Reference Hara and Radin1978). After drying the lipid extracts, 1 mg extract was dissolved in 1 ml n-hexane. The concentrations of conjugated dienes were calculated by using the molar extinction coefficient for conjugated dienes at 234 nm (ε = 29 500 mol/cm). The concentrations of TBARS were measured in liver homogenates as described (Brandsch et al. Reference Brandsch, Ringseis and Eder2002). The concentration of α-tocopherol in liver tissue was determined by HPLC (Brandsch et al. Reference Brandsch, Ringseis and Eder2002).
Determination of H2O2
To determine the H2O2 content in liver homogenates, the method for cell culture systems described by Royall & Ischiropoulos (Reference Royall and Ischiropoulos1993) was modified, using dihydrorhodamine 123 as substrate. Homogenates were incubated with 27·5 μm-dihydrorhodamine 123 for 1 h at 37°C in a final volume of 400 μl. After incubation, the fluorescence of rhodamine 123, the oxidation product of dihydrorhodamine 123, was measured (excitation wavelength 485 nm, emission wavelength 538 nm). As previously shown by Walrand et al. (Reference Walrand, Valeix, Rodriguez, Ligot, Chassagne and Vasson2003), dihydrorhodamine 123 is specifically oxidised by H2O2.
Determination of 3-hydroxybutyrate
Concentration of 3-hydroxybutyrate in plasma was determined using an enzymatic assay (cat. no. 10907979035; R-Biopharm AG, Darmstadt, Germany).
Transmission electron microscopy
Liver tissues were fixed in 3 % sodium cacodylate-buffered glutaraldehyde (pH 7·2) and post-fixed with 1 % osmium tetroxide. After washing three times, probes were dehydrated in an ethanol series and embedded in Spurr's epoxy resin. For observations with an EM 900 transmission electron microscope (Carl Zeiss SMT, Oberkochen, Germany), ultrathin sections (80 nm) were mounted on copper grids. Catalase is known to be located in peroxisomes specifically and was marked for a better visualisation of peroxisomes. For immunohistochemistry, ultrathin sections were blocked for 30 min with 1 % bovine serum albumin and 0·1 % Tween in PBS and incubated overnight with sheep polyclonal anti-catalase serum (1 : 50; Biotrend, Köln, Germany). For detection of primary antibody, sections were incubated for 1 h with a gold-marked donkey–anti-sheep antibody (1 : 25; Biotrend) and finally stained with uranyl acetate/lead citrate. Peroxisomes were counted in 1000 different prints per liver sample for each animal with a magnification of 12 000 × .
Statistics
The results were analysed using Minitab (State College, PA, USA) statistical software (release 13). Statistical significance of differences of the mean values of the two groups of pigs was evaluated using Student's t test. Mean values were considered significantly different for P < 0·05.
Results
Fatty acid composition and concentration of lipid peroxidation products in the dietary fats
Palmitic, stearic, oleic and linoleic acid were the major fatty acids in the dietary fats. The sum of these fatty acids accounted for about 95 g/100 g total fatty acids in the fats (Table 2). Amounts of stearic, oleic and linoleic acid were nearly identical in both fats; the amount of stearic acid was slightly higher in the fresh fat than in the oxidised fat. Peroxide value, acid value and concentration of conjugated dienes were 4–5-fold higher in the oxidised than in the fresh fat included in the diet (Table 2). The concentration of total carbonyls was 10-fold higher and that of TBARS was 30-fold higher in the oxidised than in the fresh fat (Table 2).
Table 2 Major fatty acids and concentrations of some lipid peroxidation products in the fresh and the oxidised fat after inclusion into the diet
TBARS, thiobarbituric acid-reactive substances.
Body weights, antioxidant status and concentrations of lipid peroxidation products in the liver
Body weights of the pigs at the end of the experiment on day 28 did not differ between the two groups (25·6 (sd 1·4) v. 26·0 (sd 1·5) kg in pigs fed the oxidised fat v. pigs fed the fresh fat; nine pigs per group). Pigs fed the oxidised fat had a higher mRNA concentration and a higher activity of SOD and a lower activity of microsomal glutathione S-transferase in the liver than pigs fed the fresh fat (P < 0·05; Table 3). Activities of glutathione peroxidase and cytosolic glutathione S-transferase as well as mRNA concentrations of these enzymes in the liver did not differ between both groups of pigs (Table 3). Concentrations of total, reduced and oxidised glutathione in the liver also did not differ between the two groups of pigs whereas the concentration of α-tocopherol was lower in pigs fed the oxidised fat than in pigs fed the fresh fat (P < 0·05; Table 3). Concentration of TBARS in the liver did not differ between the two groups of pigs whereas the concentration of conjugated dienes was slightly but significantly higher in pigs fed the oxidised fat than in pigs fed the fresh fat (P < 0·05; Table 3).
Table 3 mRNA concentrations and activities of antioxidant enzymes and concentrations of antioxidants and lipid peroxidation products in livers of pigs fed a diet with a fresh fat or an oxidised fat (Mean values and standard deviations)
TBARS, thiobarbituric acid-reactive substances.
Mean values were significantly different from those of the fresh fat group: *P < 0·05.
Indices of peroxisome proliferation
Liver weights of the pigs were not different between the two groups but pigs fed the oxidised fat had a higher peroxisome count and a higher activity of catalase in the liver than pigs fed the fresh fat (P < 0·05; Table 4). Relative mRNA concentration of acyl-CoA oxidase (ACO), a peroxisomal enzyme, in the liver, was 34 % higher in pigs fed the oxidised fat than in control animals (P = 0·062; Table 4). The concentration of H2O2 which is mainly released from peroxisomal oxidases was not different between the two groups of pigs (Table 4).
Table 4 Indices of peroxisome proliferation in livers of pigs fed a diet with a fresh fat or an oxidised fat (Mean values and standard deviations)
Mean values were significantly different from those of the fresh fat group: †P < 0·1; *P < 0·05.
mRNA concentrations of genes in liver and intestine
In liver, mRNA concentrations of PPARα and genes involved in fatty acid transport and oxidation [liver fatty acid binding protein (L-FABP), carnitine palmitoyltransferase-1 (CPT-1)], fatty acid and cholesterol synthesis [SREBP-1 and -2, insulin-induced gene-1 and -2, fatty acid synthase, acetyl-CoA carboxylase (ACC), stearoyl-CoA desaturase (SCD), 3-hydroxy-3-methylglutaryl-CoA reductase (HMG-CoA-R)], cholesterol uptake (LDL receptor), bile acid synthesis [cholesterol 7α-hydroxylase (CYP7)], lipoprotein assembly and secretion [microsomal TAG transfer protein (MTP)], inhibition of lipoprotein lipase (apo CIII) and ketogenesis [mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase (mHMG-CoA-S)] were determined (Fig. 1). Pigs fed the oxidised fat had significantly higher mRNA concentrations of mHMG-CoA-S, a classical PPARα target gene, SREBP-1 and its target genes ACC and SCD, and SREBP-2 and its target genes HMG-CoA-R and LDL receptor than control pigs fed the fresh fat (P < 0·05). mRNA concentration of CYP7 was lower in pigs fed the oxidised fat than in pigs fed the fresh fat (P < 0·05). mRNA concentrations of CPT-1 and MTP, two other PPARα target genes, tended to be increased in pigs fed the oxidised fat (P = 0·074 and P = 0·065, respectively) compared to pigs fed the fresh fat whereas mRNA concentrations of PPARα, L-FABP, apo CIII, insulin-induced gene-1 and -2, and fatty acid synthase were not different between pigs fed the fresh fat and those fed the oxidised fat (Fig. 1).
Fig. 1 Relative mRNA concentrations (- represents 1·00) of various genes involved in hepatic lipid metabolism in enterocytes of pigs fed a diet with a fresh fat or an oxidised fat. Values were determined by real-time detection RT–PCR using the mRNA concentration of glyceraldehyde-3-phosphate dehydrogenase for normalisation Values are means with their standard deviations depicted by vertical bars (n 9) obtained for the pigs fed the oxidised fat relative to the values of the control group fed fresh fat. ACC, acetyl-CoA carboxylase; CPT-1, carnitine palmitoyltransferase-1; CYP7, cholesterol 7α-hydroxylase; FAS, fatty acid synthase; HMG-CoA-R, 3-hydroxy-3-methylglutaryl-CoA reductase; Insig, insulin-induced gene; L-FABP, liver fatty acid binding protein; mHMG-CoA-S, mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase; MTP, microsomal TAG transfer protein; SCD, stearoyl-CoA desaturase; SREBP, sterol regulatory element-binding protein. Mean values were significantly different from those of the fresh fat group: *P < 0·05.
In enterocytes, relative mRNA concentrations of PPARα and of proteins involved in fatty acid uptake (L-FABP, intestinal fatty acid binding protein, fatty acid transport protein, mitochondrial aspartate aminotransferase), fatty acid oxidation (ACO, CPT-1), intracellular trafficking of cholesterol (Niemann-Pick type C1 and 2) and fatty acid synthesis (SREBP-1, fatty acid synthase) were not different between pigs fed the oxidised fat and those fed the fresh fat (Fig. 2). However, mRNA concentration of SREBP-2 and its target genes HMG-CoA-R and LDL receptor, involved in cholesterol synthesis and uptake, were higher in pigs fed the oxidised fat than in pigs fed the fresh fat (P < 0·05; Fig. 2). mRNA concentration of farnesyl diphosphate synthase did not differ between the two groups of pigs (Fig. 2).
Fig. 2 Relative mRNA concentrations (- represents 1·00) of various genes involved in intestinal lipid metabolism in livers of pigs fed a diet with a fresh fat or an oxidised fat. Values were determined by real-time detection RT–PCR using the mRNA concentration of glyceraldehyde-3-phosphate dehydrogenase for normalisation. Values are means with their standard deviations depicted by vertical bars (n 9) obtained for the pigs fed the oxidised fat relative to the values of the control group fed fresh fat. ACO, acyl-CoA oxidase; CPT-1, carnitine palmitoyltransferase-1; FAS, fatty acid synthase; FATP, fatty acid transport protein; FDPS, farnesyl diphosphate synthase; HMG-CoA-R, 3-hydroxy-3-methylglutaryl-CoA reductase; I-FABP, intestinal fatty acid binding protein; Insig, insulin-induced gene; L-FABP, liver fatty acid binding protein; mAAT, mitochondrial aspartate aminotransferase; NPC, Niemann-Pick type C; SREBP, sterol regulatory element-binding protein. Mean values were significantly different from those of the fresh fat group: *P < 0·05.
Concentrations of TAG and cholesterol in liver, plasma and lipoproteins
Concentrations of TAG in liver, plasma and TAG-rich lipoproteins did not differ between pigs fed the fresh fat and those fed the oxidised fat. Concentrations in pigs fed the oxidised fat v. pigs fed the fresh were (nine pigs per group): liver, 88 (sd 20) v. 91 (sd 19) μmol/g; plasma, 0·96 (sd 0·26) v. 1·09 (sd 0·17) mmol/l; chylomicrons+VLDL, 0·80 (sd 0·25) v. 0·93 (sd 0·16) mmol/l. Concentrations of cholesterol in liver, plasma, LDL and HDL were also not different between the two groups of pigs. Concentrations in pigs fed the oxidised fat v. pigs fed the fresh were: liver, 73 (sd 14) v. 69 (sd 10) μmol/g; plasma, 2·63 (sd 0·32) v. 2·83 (sd 0·22) mmol/l; LDL, 0·96 (sd 0·16) v. 0·97 (sd 0·15) mmol/l; HDL, 1·02 (sd 0·18) v. 1·13 (sd 0·11) mmol/l.
Concentration of 3-hydroxybutyrate in plasma
Pigs fed the oxidised fat had a higher concentration of 3-hydroxybutyrate in plasma than pigs fed the fresh fat (1·23 (sd 0·58) v. 0·52 (sd 0·27) mmol/l; P < 0·05).
Discussion
In the present study, pigs were fed a diet containing an oxidised fat prepared under usual deep-frying conditions. The relatively low concentrations of lipid peroxidation products (conjugated dienes, TBARS, peroxides and carbonyls) in the oxidised fat indicate that this fat was mildly oxidised. Concentrations of peroxidation products in this fat were indeed even lower than in soyabean oil or hydrogenated animal–vegetable oil blends used for frying of potatoes at 190°C over a period of 24 h (Frankel, Reference Frankel1998). The reason for the relatively low degree of oxidation is that we did not add foodstuffs to be fried during the preparation of the oil as we wanted to avoid contamination of the oil with food ingredients. It is well known that ingredients of foodstuffs, i.e. metal ions, enhance the lipid peroxidation process during frying of fats (Kubow, Reference Kubow1992). The concentration of conjugated dienes which include the potent PPARα activators hydroxy- and hydroperoxy fatty acids (Delerive et al. Reference Delerive, Furman, Teissier, Fruchart, Duriez and Staels2000; König & Eder, Reference König and Eder2006) was approximately four times higher in the oxidised fat than in the fresh fat. The finding of an increased activity of SOD and a slightly elevated concentration of conjugated dienes, together with the observation of a slightly reduced concentration of α-tocopherol, indicates that the oxidised fat produced oxidative stress in the liver of the pigs. It has been demonstrated that under oxidative stress, expression of SOD is stimulated and concentration of α-tocopherol is reduced due to an enhanced consumption (Liu & Huang, Reference Liu and Huang1995; Ruiz-Gutierrez et al. Reference Ruiz-Gutierrez, Perez-Espirosa, Vazquez and Santa-Maria1999; Atalay et al. Reference Atalay, Laaksonen, Khanna, Kaliste, Hanninen and Sen2000). However, the oxidative stress produced by the oxidised fat was very moderate as concentrations of TBARS and glutathione remained completely unchanged. In rodents treated with PPARα agonists such as fibrates or WY-14,643, production of H2O2 is largely increased due to a strong up-regulation of peroxisomal oxidases, and this causes oxidative stress and contributes to hepatocarcinogenesis in these species (Peters et al. Reference Peters, Cheung and Gonzalez2005). In the present study, feeding the oxidised fat did not increase the concentration of H2O2 in the liver. This was probably due to two reasons: first, there was only a slight increase in the mRNA concentration of ACO, one of the enzymes producing H2O2; second, activity of catalase, the key enzyme of decomposition of H2O2 in peroxisomes was increased. Therefore, generation of H2O2 did not contribute to oxidative stress in animals treated with oxidised fat. The reason for the moderate oxidative stress may be that a part of the dietary lipid peroxidation products is absorbed in the intestine and reaches the liver via lipoproteins (Staprans et al. Reference Staprans, Pan, Rapp and Feingold2005). Production of oxidative stress by intake of strongly oxidised fats has been shown several times in rodents (Yoshida & Kajimoto, Reference Yoshida and Kajimoto1989; Liu & Huang, Reference Liu and Huang1996; Liu & Lee, Reference Liu and Lee1998; Ammouche et al. Reference Ammouche, Rouaki, Bitam and Bellal2002; Eder et al. Reference Eder, Keller and Brandsch2004; Keller et al. Reference Keller, Brandsch and Eder2004a, Reference Keller, Brandsch and Ederb). The present study shows for the first time that even a mildly oxidised fat, as used in human nutrition, can induce moderate oxidative stress in pigs as a non-proliferating species.
To find out whether the mildly oxidised fat caused activation of PPARα in the liver of pigs, we determined mRNA concentrations of the classical PPARα target genes ACO, CPT-1 and mHMG-CoA-S as well as peroxisome count, activity of catalase and plasma concentration of 3-hydroxybutyrate. Recent studies in pigs have shown that activation of PPARα in pigs, by either treatment with clofibrate or by fasting, leads to an increased expression of these PPARα target genes, and in turn stimulates mitochondrial and peroxisomal β-oxidation and ketogenesis (Yu et al. Reference Yu, Odle and Drackley2001; Peffer et al. Reference Peffer, Lin and Odle2005; Cheon et al. Reference Cheon, Nara, Band, Beever, Wallig and Nakamura2005). The finding of an increased peroxisome count together with increased activity of catalase, a peroxisomal enzyme, a significantly increased mRNA concentration of mHMG-CoA-S and an increased plasma concentration of 3-hydroxybutyrate strongly indicate that the oxidised fat caused PPARα activation in the liver of the pigs. The finding that mRNA concentrations of ACO and CPT-1, two other classical PPARα target genes were also increased by 34 and 29 %, although not significantly different to control, supports the assumption that the oxidised fat induced hepatic PPARα activation in the pigs. It has been shown that these two enzymes are only moderately up-regulated in pig liver by PPARα agonists. For instance, in pigs treated with clofibrate, a strong PPARα agonist, hepatic gene expression of CPT-1 and ACO was only 1·89- and 1·42-fold, respectively, increased over control while gene expression of mHMG-CoA-S was increased 3·32-fold (Cheon et al. Reference Cheon, Nara, Band, Beever, Wallig and Nakamura2005). This presents an explanation for the observations that mHMG-CoA-S was significantly increased in pigs treated with oxidised fat and that ACO and CPT-1 were only slightly increased. The finding that mRNA concentration of MTP, a gene recently shown to be up-regulated by PPARα activation (Ameen et al. Reference Ameen, Edvardsson, Ljungberg, Asp, Akerblad, Tuneld, Olofsson, Linden and Oscarsson2005), tended to be increased in the liver of pigs fed the oxidised fat also indicates that the oxidised fat caused PPARα activation in the liver. Recently, studies in rats have already shown that oxidised fats are able to activate PPARα in the liver (Huang et al. Reference Huang, Cheung and Lu1988; Chao et al. Reference Chao, Chao, Lin and Huang2001; Sülzle et al. Reference Sülzle, Hirche and Eder2004). In these rat studies, up-regulation of PPARα target genes in the liver was much stronger than in pigs of the present study. This may have two different reasons: first, most PPARα target genes respond stronger to PPARα activation in rats than in non-proliferating species such as pigs or man; second, fats used in the rat studies were more strongly oxidised than the mildly oxidised fat used in the present study. The present study shows for the first time that even a mildly oxidised fat causes activation of PPARα in pigs which are, as man, less sensitive to PPARα agonists than rodents.
To study whether the oxidised fat caused PPARα activation in small intestine, we considered in addition to the classical PPARα target genes ACO and CPT-1, several genes involved in fatty acid transport (L-FABP, intestinal fatty acid binding protein, fatty acid transport protein and mitochondrial aspartate aminotransferase) and cholesterol trafficking (Niemann-Pick type C1 and 2) in intestinal tissue. All these genes have been shown to be up-regulated by PPARα activation (Darimont et al. Reference Darimont, Gradoux, Cumin, Baum and De Pover1998; Motojima et al. Reference Motojima, Passilly, Peters, Gonzalez and Latruffe1998; Mochizuki et al. Reference Mochizuki, Suruga, Yagi, Takase and Goda2001; Chinetti-Gbaguidi et al. Reference Chinetti-Gbaguidi, Rigamonti, Helin, Mutka, Lepore, Fruchart, Clavey, Ikonen, Lestavel and Staels2005). The finding that none of these genes was up-regulated in cells of small intestine indicates that oxidised fat caused no or even weak PPARα activation and does not influence intestinal fatty acid transport and cholesterol trafficking.
Synthesis of lipids in mammalian cells is controlled by a network involving the action of insulin-induced genes and SREBP, and it has been recently shown in several studies that this network is influenced by PPARα activation (Guo et al. Reference Guo, Wang, Milot, Ippolito, Hernandez, Burton, Wright and Chao2001; Patel et al. Reference Patel, Knight, Wiggins, Humphries and Gibbons2001; Knight et al. Reference Knight, Hebbach, Hauton, Brown, Wiggins and Patel2005; König et al. Reference König, Koch, Spielmann, Hilgenfeld, Stangl and Eder2006). The present study shows that feeding a mildly oxidised fat increased the mRNA concentration of SREBP-1 and its target genes ACC and SCD, two key enzymes of de novo fatty acid synthesis, in the liver. These alterations may be caused by activation of PPARα in the liver. Knight et al. (Reference Knight, Hebbach, Hauton, Brown, Wiggins and Patel2005) found that treatment with WY 14,643, a synthetic PPARα agonist, causes a strong up-regulation of enzymes involved in hepatic fatty acid synthesis and stimulates fatty acid synthesis in wild-type mice but not in PPARα null mice. Knight et al. (Reference Knight, Hebbach, Hauton, Brown, Wiggins and Patel2005) suggest that up-regulation of hepatic fatty acid synthesis is a compensatory response on the increased fatty acid oxidation to maintain a constant cellular TAG level. The finding that TAG levels in liver and plasma were not reduced in pigs fed the oxidised fat compared to control pigs indeed suggests that an increased β-oxidation of fatty acids was compensated by an increased fatty acid synthesis. As there is no evidence for a direct action of PPARα on the promoter regions of SREBP-1 and ACC genes, it is likely that the increased mRNA concentrations of these genes are an indirect result of PPARα activation. In contrast, SCD is not only dependent on SREBP-1 but has also a PPAR response element in its promoter (Miller & Ntambi, Reference Miller and Ntambi1996). Therefore, its transcription may have been in part directly stimulated by PPARα activation. An up-regulation of SCD which catalyses the formation of MUFA from SFA has also been observed in pigs treated with clofibrate (Cheon et al. Reference Cheon, Nara, Band, Beever, Wallig and Nakamura2005). These findings of the effects of the oxidised fat on gene expression of lipogenic enzymes observed in pigs are opposite to those observed in rats in which a dietary oxidised fat causes a down-regulation of lipogenic enzymes and a strong reduction of liver and plasma TAG (Eder & Kirchgessner, Reference Eder and Kirchgessner1998; Eder et al. Reference Eder, Schleser, Becker and Körting2003).
It is moreover shown that feeding the mildly oxidised fat led to a moderate but significant up-regulation of SREBP-2, and its target genes HMG-CoA-R and LDL receptor, in both liver and small intestine. The present findings suggest that the oxidised fat could have stimulated synthesis and uptake of cholesterol in these tissues. As this effect occurs not only in the liver but also in the small intestine where no PPARα activation was found in pigs fed the oxidised fat, it is questionable whether these effects are linked to PPARα activation. The finding that hepatic genes involved in cholesterol synthesis were not altered in pigs treated with clofibrate indeed suggests that PPARα activation does not influence SREBP-2 controlled transcription of genes involved in cholesterol homeostasis (Cheon et al. Reference Cheon, Nara, Band, Beever, Wallig and Nakamura2005). On the other hand, treatment with the PPARα agonist WY 14,643 caused an up-regulation of genes involved in hepatic cholesterol synthesis in wild-type mice but not in PPARα null mice, indicating that PPARα activation indeed could directly stimulate cholesterol synthesis (Knight et al. Reference Knight, Hebbach, Hauton, Brown, Wiggins and Patel2005). It should be noted, however, that there is also another study that found a suppression of gene expression and proteolytic activation of SREBP-2, and a strong down-regulation of its target genes accompanied by reduced cholesterol synthesis in rats (König et al. Reference König, Koch, Spielmann, Hilgenfeld, Stangl and Eder2006). The effect of PPARα activation on SREBP-2-dependent cholesterol synthesis is not yet clear and may also be different between various species. Besides an up-regulation of genes involved in synthesis and uptake of cholesterol, we found a down-regulation of CYP7, the key enzyme of bile acid formation, in the liver. It has been shown in human and rat liver cells that PPARα agonists lower CYP7 expression probably by reducing the availability of hepatic nuclear factor 4α which is required for binding to a DR-1 in CYP7 promoter (Marrapodi & Chiang, Reference Marrapodi and Chiang2000; Patel et al. Reference Patel, Knight, Soutar, Gibbons and Wade2000). Therefore, we assume that down-regulation of CYP7 in the liver of pigs fed the oxidised fat was caused by PPARα activation induced by the oxidised fat. Increased hepatic cholesterol synthesis and uptake of cholesterol into the liver, together with a decreased bile acid synthesis, is expected to increase hepatic cholesterol concentration. In contradiction to this, liver and plasma cholesterol concentrations were unchanged in pigs fed the oxidised fat compared to pigs fed the fresh fat. We assume that the changes in gene expression were too small to induce phenotypical alterations of cholesterol concentrations.
In conclusion, the present study shows that a mildly oxidised fat causes PPARα activation in the liver of pigs as indicated by an increased peroxisome count, a moderate up-regulation of PPARα target genes and a stimulation of ketogenesis. Moreover, the oxidised fat led to an up-regulation of the expression of SREBP-1 and SREBP-2 and their target genes involved in TAG and cholesterol synthesis, suggesting a stimulation of lipid synthesis. As the fat used in the present study was even less oxidised than fats used for deep-frying of foods, and as there exists a similarity in the gene response to PPARα agonists between pig and human liver cells, deep-fried fats could exert similar effects in man.
Acknowledgements
Sebastian Luci and Bettina König contributed equally to this work.
