Long term highly saturated fat diet does not induce NASH in Wistar rats

Our study was performed following the recommendations provided by the European Convention for the protection of Vertebrate Animals used for Experimental and Scientific purposes (Council of Europe N° 123, Strasbourg, 1985). Male Wistar rats, purchased from Charles River Laboratories France, were housed four per box on a 12-h light/dark cycle. Animals were allowed to acclimatise to their new conditions for a week before the study. At the age of 21 days the animals were allocated to four different groups. The first group received a standard diet (A04 – Scientific Animal Food & Engineering, France) (std, n = 10). The second group had the choice between standard diet and coconut diet (coco, n = 10), for 14 weeks. The third group had the choice between standard diet and butter diet (butter, n = 4) for 14 weeks as well. The methionine and choline deficient diet (MCDD n = 10; ICN Pharmaceuticals France SA, ref n° 960439), was given for 6 weeks from the age of 9 weeks. Diets components are shown in Table 1. The coconut or butter diets had 67% of energy derived from fat compared with the standard diet (7%). High fat diets were weekly prepared in our laboratory in pellets and stored at +4°C. The food was renewed every day during the treatment and the standard and high fat diet intakes are monitored daily independently. Rats had free access to food and water and were weighed every week for the duration of the study.

Histological analyses were achieved in the gastroenterology service of Caen hospital (France). Liver specimens were fixed in 4% buffered formalin for 24 to 48 h. They were then embedded in paraffin, cut at 5 μm, and routinely stained with hematoxylin-eosin (H&E) and reticulin. Severity of steatosis was evaluatedusing the percentage of macrovesicular fat within hepatocytes: mild (5 to 30 %); moderate (30 to 60 %); severe (more than 60%) [23].

The liver was rapidly removed and finely minced and washed with ice-cold isolation medium containing 250 mM sucrose, 2 mM KH₂PO₄, 1 mM EGTA and 20 mM Tris-HCl (pH 7.2). Liver mitochondria were prepared by standard differential centrifugation procedures [24]. Mitochondrial protein content was determined by the Biuret method [25] with serum albumin as standard. For the determination of oxygen consumption, mitochondria were incubated at a concentration of 2 mg/mL in an oxygraph vessel with a Clark electrode, thermostatically controlled at 37°C, in a medium containing 125 mM KCl, 1 mM EGTA, 2 mM KH₂PO₄, 20 mM Tris-HCl with 0.1% fatty acid-free Bovin Serum Albumin (pH 7.2). The control state of respiration was initiated by the addition of 5 mM succinate/0.5 mM malate, in the presence of rotenone (1.25 μM) while the addition of 1 mM ADP initiated the active state of respiration (state 3). Oligomycin (1.25 μg/mg protein) was then added to the mitochondrial suspension to determine the non-phosphorylating respiratory rate (state 4).

Hepatocytes were isolated from 20–24 h starved rats as previously described by Berry and Friend (Berry, 1969) and modified by Groen et al. [26], by a two-step in situ collagenase perfusion technique. Hepatocytes (10 mg/mL dry weight) were incubated at 37°C in closed vials containing a Krebs-bicarbonate buffer (120 mM NaCl, 4.8 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 24 mM NaHCO₃, pH 7.4) saturated with 95% O₂/5% CO₂ containing BSA (2%, w/v) and 2.4 mM CaCl₂. Experiments were performed with 20 mM dihydroxyacetone, with or without fatty acids, 4 mM octanoate or 2 mM oleate. After 20 min, oxygen uptake (J O₂) was measured polarographically at 37°C with a Clark electrode before and after the addition of 6 μg/ml oligomycin. At t = 0 and 30 min, samples of hepatocyte suspension were taken, quenched in HClO₄ (4% v/v final concentration) and neutralized with 2 M KOH/0.3 M MOPS for later enzymatic measure of 3-hydroxybutyrate, acetoacetate and glucose as described by Bergmeyer [27].

Adipocytes were prepared from retroperitoneal white fat pads. 1.5 g of fat tissue was digested for 30 min at 37°C with 2 mg/mL of type II collagenase. The digestion medium was a Krebs-Ringer medium (139 mM NaCl, 5.4 mM KCl, 1 mM NaH₂PO₄, 1 mM MgSO₄, 2.2 mM CaCl₂, pH 7.4) buffered with 20 mM Hepes, containing 2% (w/v) BSA with 7 mM glucose. Isolated cells were obtained by filtration through a coarse nylon mesh (250 μm) before being washed twice with a 2% BSA buffer. Adipocytes were then observed under a microscope and the determination of the diameter was measured on pictures of roughly 200–250 cells using a calibrated scale.

Liver triacylglycerol concentration was estimated from glycerol release after ethanolic KOH hydrolysis [28], using a commercial colorimetric kit (Biomerieux, France).

Mitochondria from brown adipose tissue (BAT) were prepared as described previously. 40 μg of BAT mitochondrial proteins were separated by SDS-PAGE (12.8% acrylamide) and transferred to polyvinylidene fluoride membranes (Immobilon-P, Millipore). Immunological detection was performed using a rabbit antiserum against UCP1 (α-diagnostics UCP11-S (1:15000), USA). The detection was realized with a horseradish peroxidase-coupled anti-rabbit (Bio-Rad 170–6516 (1:5000)) secondary antibody and an enhanced chemiluminescence (ECL) detection kit (Amersham, UK). Quantification of autoradiographs was performed by scanning densitometry.

BAT samples (40–50 mg) were immediately homogenized at +4°C in 0.3 M phosphate buffer containing 0.05% bovine serum albumin (pH 7.7) using a glass Potter-Elvehjem homogenizer. Then, they were frozen at -80°C and thawed three times to disrupt the mitochondrial membrane. 3-hydroxyacyl-CoA dehydrogenase (HAD, EC 1.1.1.35) was spectrophotometrically determined as previously described by Lowry & Passonneau [29]. Enzyme activity was determined at 25°C and expressed as micromoles of substrate per minute per milligram of protein.

Differences between groups were determined using non-parametric Mann & Whitney tests. Data were expressed as mean values ± SEM and differences between means were considered significant when P < 0.05.

The rats received either a standard diet with 7% of energy derived from fat, or had free choice between the standard diet and high fat diets with 67% of energy derived from fat. The first high fat diet contained coconut oil, composed of 86% of saturated fatty acid and especially lauric acid (C12:0). The second high fat diet was realised with milk butter that contains 51% of SFA, the most abundant being palmitic acid (C16:0). The respective composition of the two diets in the main fatty acids is presented in Table 2. Because the rats had free access to either standard (2900 kcal/kg) or high fat diets (5500 kcal/kg) we first estimated total calorie ingestion per day and monitored resulting body mass. Animals were housed four per cage. Rats could freely choose between the standard and the high fat diet which were both available within the cages. The amount of food consumed per day from each respective diet was calculated and then that amount was divided by four to estimate the consumption of each rat. As shown in Fig. 1A, the total caloric intake was higher in both groups fed with high fat diets than in rats fed the standard diet. These caloric intakes at the end of the experiment were 143% higher for coconut group (178 kcal/day/rat) and 30% higher for butter group (95 kcal/day/rat), as compared with standard group (73 kcal/day/rat). The profile of lipids, proteins and carbohydrates ingestion at weeks 3 and 16 of experiment is reported in Table 3 for each group. The proportion of energy derived from lipids was increased 1.5 fold for the coconut group to reach 45% by the end of the protocol, while it reached 42% at week 16 of diet in the butter group. At the end of the study period, total lipid ingestion, expressed as a percentage of the calories ingested, was higher in both high fat diet groups than in the standard diet (coconut 45% vs butter 42% vs standard 7% kcal _total). In the coconut group, this energy came mostly from SFA (41.3%) whereas in the butter group energy was shared between SFA and MUFA (28% and 12% respectively) (Table 3). Surprisingly, in spite of a larger energy intake, body mass was not affected in rats fed the high fat diets (Fig. 1B). However, fat mass was largely affected. Hence, retroperitoneal white adipose tissue depot (Fig. 2) was heavier in high fat diet rats (+62% in coconut and +93% in butter) than in standard rats. To investigate which factors were implicated in this fat modification, adipocytes were isolated from retroperitoneal white adipose tissue, from coconut and standard groups and their size and volume measured. A two-fold increase in cell volume was observed with the coconut diet, demonstrating that the change in white adipose tissue mass was partly due to a change in the volume of individual adipocytes (Fig. 3). Hence, the high fat diet groups displayed the expected feature related to high lipid ingestion, i.e. retroperitoneal fat accumulation and adipocyte growth.

An opposite picture was observed with MCD diet. Rats fed a MCD diet showed a smaller food intake (Fig. 1C) and a drop in body mass (Fig. 1D) as classically observed with this diet [30‐32]. The mass of retroperitoneal white adipose tissue was dramatically reduced (-44%) in MCD rats (Fig. 2) which clearly indicates no fat accumulation in white adipose tissue. Conversely, the liver mass of MCD rat was higher than that of controls (4.5 ± 0.2 vs 3.4 ± 0.1 g/100g; P < 0.05).

Liver histology (Fig. 4A) from rats fed a standard diet showed no steatosis, inflammatory cells or fibrosis. The rats which were fed coconut oil or butter display a very mild macrovacuolar steatosis (affecting less than 5% of hepatocytes), whereas steatosis was severe (affecting more than 60% of hepatocytes) in the MCDD rats and associated to inflammation. This massive steatosis in the MCDD group was confirmed by triacylglycerol content, which was 22-fold higher in MCDD rats than in controls. High fat diet did not significantly increase liver triacylglycerol content (Fig. 4B). Thus, MCDD rats demonstrated the typical liver feature of non alcoholic steatohepatitis (steatosis and inflammation), whereas both high fat diets (coconut oil or butter) induced only very mild steatosis.

In order to understand why rats fed a high fat diet did not develop liver injury, we searched for an increased capacity for fatty acid oxidation in liver cells. Fatty acid oxidation takes place in mitochondria and can be either partial, leading to ketone body synthesis (3-hydroxybutyrate and acetoacetate), or complete, leading to a large supply of reducing equivalents to the oxidative phosphorylation, thus resulting in a higher oxygen consumption. Fatty acid oxidation was therefore assessed by the measurement of both ketone body formation and hepatocyte respiration. We used either a long chain fatty acid (oleate), which is oxidized after transfer into the mitochondria by the carnitine-palmitoyl transferase (CPT), or a medium chain fatty acid (octanoate), that enters the mitochondria without any transporters. As expected, ketone body synthesis was higher with octanoate than with oleate, due to the direct entry of this fatty acid into the mitochondria. There was no significant change in ketone body formation in response to high fat diets, whatever the fatty acid used as substrate (Fig. 5A). This was corroborated by similar respiratory rate of hepatocytes from the three group studied (Fig. 5B).

Changes in dietary lipid can modify cell membrane composition that can, in turns, alter cell integrity resulting in a metabolism decrease. Hence, we studied gluconeogenesis which is controlled by, and therefore reflects, hepatocyte energy state. Using dihydroxyacetone as substrate, glucose production from isolated hepatocytes was not different in high fat diet groups as compared with standard group (Fig. 5C).

To verify that whole cell respiration measurement was not biased by the oxygen consumption of other organelles such as peroxisomes, the respiration capacities of isolated liver mitochondria were measured. Bovine serum albumin was added to eliminate any artefactual uncoupling effect of free fatty acids. No difference was found between mitochondrial oxygen consumption from high fat and standard diet fed rats (Fig. 5D). Therefore, resistance to steatosis was not explained by an enhanced ability of the liver to oxidize fatty acids.

Despite enhanced lipid ingestion in the high fat diet animals, their body mass did not increase and fatty acids were not stored in the liver. Then, we searched for an other site of fatty acid utilisation and energy wasting. In rats, BAT is a well-known site for energy wastage due to the abundance of mitochondria and the profusion of an uncoupling protein (UCP1) in these mitochondria. Interscapular BAT depot was weighed and we estimated lipid oxidation by HAD activity and quantified the UCP1 protein content in high fat diet and standard diet fed rats. Intrascapular BAT depot was heavier (+23%) in coconut diet fed animals (Fig. 6A) while there was only a non significant trend in butter diet fed rat (+12%), possibly due to both little variations and a small number of rats in this group. A greater mitochondrial HAD activity was observed in coconut diet fed rats as compared with standard rats but not in butter fed rats (+55%; P < 0.05). A higher UCP1 protein content was observed in mitochondria from both coconut and butter diet fed rat (in arbitrary unit: coconut: +25%, butter: +28%; P < 0.05). Taken together, these results suggest a thermogenic activation of BAT that may result in fat oxidation contributing to the maintenance of body weight and the protection of liver.