Dietary fatty acid composition modifies hepatic lipid metabolism. To determine the effects of fatty acids on hepatic triglyceride storage, rats were fed diets enriched in carbohydrates (control), fish oil, or lard. After 4 weeks, the animals were fasted overnight. In the morning, the animals were either sacrificed or fed 8 g of their respective diets before sacrifice. Animals ingested more food calories with diets containing fish oil than with other diets. However, fish oil–fed animals weighed less and had less body fat. In fish oil–fed animals, liver triglyceride was lower by 27% (P < .05) and 73% (P < .01) than in control- and lard-fed animals, respectively. Fish oil altered the postprandial gene expression of hepatic regulators of fatty acid degradation and synthesis. Fish oil feeding blunted the normal postprandial decline in fatty acid degradation genes (PPARα, CPT1, and ACO) and blunted the normal postprandial rise in triglyceride synthesis genes (SREBP1-c, FAS, SCD-1). Therefore, the direct postprandial effect of fish oil ingestion decreases the propensity for hepatic triglyceride storage. In conclusion, n-3 polyunsaturated fatty acids decrease total body weight, total body fat, and hepatic steatosis. (HEPATOLOGY 2004;39:608–616.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Dietary fatty acid composition modifies hepatic lipid metabolism. Animals that ingest diets enriched in saturated fatty acids (the so-called “Western diet”) are more likely to have disorders in both glucose and lipid metabolism.1 Glucose disorders include insulin resistance and type 2 diabetes mellitus. Typical lipid disorders include dyslipidemia and excess lipid accumulation in adipose tissue and in nonadipose tissue (e.g., liver and muscle). It is not known if insulin resistance precedes lipid accumulation in insulin-sensitive tissue, or vice versa. Several investigators have proposed a lipotoxicity hypothesis, which states that insulin resistance develops because of excess lipid deposition in tissues.2, 3 Whatever the mechanism, there is a close association between insulin resistance and nonalcoholic fatty liver disease.4
Diets enriched in polyunsaturated fatty acids (PUFAs) sensitize both rodents5, 6 and humans7–9 to the actions of insulin. However, the mechanisms responsible for fatty acid effects on insulin sensitivity are not well understood. Some investigators have hypothesized that ingestion of PUFAs prevents the accumulation of lipotoxic triglycerides in insulin-sensitive target tissues and improves insulin sensitivity. Neschen and colleagues have recently demonstrated that fish oil, which has a high concentration of n-3 PUFAs, increases hepatic peroxisome content and lowers hepatic triglyceride content.10
PUFAs have been reported to regulate transcription factors and enzymes with key roles in the metabolism of fatty acids, which are substrates for triglyceride synthesis. For instance, n-3 PUFAs up-regulate peroxisomal proliferator-activated receptor-α (PPARα), which increases transcription of genes responsible for fatty acid degradation [e.g., mitochondrial carnitine palmitoyl transferase 1 (CPT1) and peroxisomal acyl-CoA oxidase (ACO)].11, 12 n-3 PUFAs down-regulate steroid-responsive element binding protein 1-c (SREBP1-c), which increases transcription of genes responsible for fatty acid synthesis [e.g., fatty acid synthase (FAS) and stearoyl Co-A desaturase (SCD-1)].13, 14 Fasting and refeeding greatly influence the gene expression of both PPARα and SREBP1-c. Fasting stimulates PPARα and inhibits SREBP1-c. Refeeding inhibits PPARα and stimulates SREBP1-c. Therefore, intrahepatic triglyceride content (and insulin sensitivity) may be determined, in part, by the balance between fatty acid degradation (PPARα activity) and fatty acid synthesis (SREBP1-c activity) in both the fasted and fed states.
In the present study, we fed diets enriched with carbohydrate, fish oil, or lard to rats that were prone to fatty liver and impaired glucose metabolism. After 4 weeks, we measured indices of glucose and lipid metabolism in both the fasting and postprandial state. Furthermore, we measured fasting and postprandial messenger RNA (mRNA) levels of PPARα, SREBP1-c, and their regulated genes. We hypothesized that the n-3 PUFAs in the fish oil diet would prevent the intrahepatic accumulation of triglycerides by stimulating intrahepatic lipid degradation and inhibiting lipid synthesis in both the fasting and postprandial state. The regulation of triglyceride content within the liver is complex and includes the synthesis, packaging, and secretion of triglycerides and apoproteins in lipoprotein particles. We did not study these important pathways. Our results showed that n-3 PUFAs indeed prevented fatty liver. The mechanisms for prevention of fatty liver by n-3 PUFAs are likely multifactorial. Compared with fasting, the postprandial effects of n-3 PUFA on intrahepatic gene expression appear to play a greater role in reducing intrahepatic lipid content.
PUFAs, polyunsaturated fatty acids; PPARα, peroxisomal proliferator-activated receptor-α; CPT1, carnitine palmitoyl transferase 1; ACO, acyl-CoA oxidase; SREBP1-c, steroid-responsive element binding protein 1-c; FAS, fatty acid synthase; SCD-1, stearoyl Co-A desaturase; F344 rat, Fischer 344 rat; DXA, dual-energy X ray absorptiometry; cDNA, complementary DNA; PCR, polymerase chain reaction.
Animals and Diets.
All animals were humanely treated, and the experimental protocols were reviewed and accepted by the Institutional Animal Care and Use Committee at Virginia Commonwealth University. Recently, we characterized a rodent model—the Fischer 344 (F344) rat—of type 2 diabetes mellitus.15, 16 These animals are insulin-resistant at a young age and develop diabetes mellitus with aging; they are also leptin-resistant. At young ages (70 days), F344 rats are hyperleptinemic, but they develop moderate obesity and hepatic steatosis. They have dyslipidemia, as in other models of insulin and leptin resistance. F344 rats 9 to 10 weeks old were purchased from Harlan Sprague-Dawley Inc. (Indianapolis, IN) and housed in individual cages at 22° C. Lighting was controlled on a natural dark–light cycle (lights out at 1800, lights on at 0600). The animals were allowed free access to food and water. The quantity of food was measured daily. The animals were fed either a low-fat diet (ICN Biomedical, Aurora, OH) or they were fed one of two high-fat diets (fish oil or lard from ICN Biomedical). The fractions of total calories from carbohydrates in the control and high-fat diets were 78% and 37.6%, respectively; the fractions of total calories from fat in these same diets were 5.1% and 45%, respectively. The fatty acid and cholesterol contents of each of the two fat diets are shown in Table 1.
Table 1. Fatty Acid and Cholesterol Composition, g/100 g Edible
NOTE. Values are based on the US Department of Agriculture Nutrient Database for Standard Reference. Values less than 2 g/100 g edible are not shown.
After 4 weeks, the animals were fasted overnight and either sacrificed (fasted: n = 5), or refed 8 g of their respective diets (refed: n = 5); they were then sacrificed 3 hours after the introduction of food. Invariably, all food was ingested. The 3-hour refeeding period was chosen because the rises in serum lipid and leptin concentrations reach a peak by approximately 3 hours after the introduction of food.17, 18 Body composition was measured and blood and tissue were collected to assess various glucose and lipid metabolic parameters as discussed below.
Total body fat was quantified using dual-energy X ray absorptiometry (DXA; Hologic QCR 1000 W, Waltham, MA) at time 0 (3 days after arrival in the animal facility) and 3 days before sacrifice, as previously described.15 To perform the DXA procedure, the animals were anesthetized with isoflurane inhalation, combined with intraperitoneal pentobarbital (30 mg/kg) and intramuscular ketamine (40 mg/kg) injections. The animals were placed in a prone position on an animal platform. We used Hologic ultra high-resolution software to determine whole body composition of the rats. Values for percent fat were recorded. In addition, at the conclusion of the experiment, the animals were sacrificed, and the epididymal fat depots were excised and weighed. DXA measurements of body fat correlate well with other body fat measurements, such as chemical analysis and fat depot weight.19–21
Serum and Tissue Measurements.
All animals were anesthetized with isoflurane before they were decapitated. Blood was collected and cooled in ice. After spinning in a centrifuge, the serum was isolated and saved in a freezer at −80° C. Liver and epididymal fat pads were quickly dissected, weighed, frozen in liquid nitrogen, and then stored in a freezer at −80° C.
Serum glucose was measured with an automated colorimetric glucose oxidase system (Vitros 700 system, Johnson & Johnson, Somerville, NJ). Serum leptin and insulin were measured using radioimmunoassay kits (Linco Research, St. Charles, MO). The limits of sensitivity and linearity for the rat leptin assay were 0.5 ng/mL and 50 ng/mL, respectively. The intra-assay variation for the leptin assay at 1.6 ng/mL (Quality Control 1) was less than 2%. Liver triglycerides and nonesterified fatty acids in serum were measured using kits as described by the supplier (Wako Chemicals USA, Inc., Richmond, VA). In brief, 100 mg of liver were homogenized with a saw-tooth generator (OMNI TH, Warrenton, VA) in 2 mL of homogenizing buffer [18 mM Tris-HCl (pH 7.5), 300 mM D-mannitol, 5 mM EGTA (ethylenebis(oxyethylene nitrilo)tetraacetic acid]). The solution was centrifuged (3,500 rpm for 20 minutes) to remove debris. Isopropyl alcohol (5 mL) was added to 0.2 mL of supernatant (experimental sample) or standard solutions (0–2,000 mg/dL). After vigorous shaking for 10 minutes, the solutions were centrifuged (3,000 rpm for 10 minutes) and 3.0 mL of color reagent solution of Triglyceride E (Wako #432-40201) was added to 0.3 mL of supernatant. The absorbance at 600 nm was measured with a spectrophotometer.
Gene Expression by Quantitative Reverse-Transcriptase Polymerase Chain Reaction.
RNA was extracted from liver using the TRIzol Reagent method as described by the suppliers (Invitrogen, Carlsbad, CA). Complementary DNA (cDNA) synthesis was performed with an oligo(dT) primer and Thermoscript reverse transcriptase as described by the supplier (Invitrogen, Frederick, MD). Duplex polymerase chain reaction (PCR) was performed in a solution with primer pairs from a gene of interest and a housekeeping gene. The sense and antisense primers for each gene are shown in Table 2. The PCR conditions for each duplex are shown in Table 3. PCR conditions were carefully chosen to optimize amplification of the experimental and housekeeping cDNA and to limit the amplification of nonspecific cDNA.
Table 2. Primers for PCR Amplification of Indicated Genes
As shown in Fig. 1, amplification of each duplex cDNA was relatively linear within a range of cycle numbers. The range of linearity depended on the gene of interest and the conditions of PCR. Cycle number was chosen so that the amplification of the cDNA pair was well within the linear portion of the curve of cycle number versus the quantity of DNA. When we compared the gene expression among experimental groups, all duplex PCR reactions contained the same master mix, were simultaneously amplified in a thermocycler (MJ Research, Boston, MA), and were analyzed on the same agarose gel. Solutions for PCR differed only by the addition of various cDNAs.
An aliquot (8 μL) of the PCR reaction (total volume = 25 μL) was analyzed on a 1.5% to 2.0% agarose gel with ethidium bromide. The gel was placed in a luminescent image analyzer (LAS-1000, Fujifilm, Valhalla, NY) and the densities (light arbitrary units per mm2) of the amplified DNA were assessed by the Advanced Image Data Analyzer software (AIDA version 2.3.401; Raytest USA, Wilmington, NC). The abundance of the expressed gene was calculated by dividing the density of the amplified experimental gene by the density of the amplified housekeeping gene in the same sample.
During the 4-week feeding period, the animals that were fed fish oil and lard ingested more total food energy than did the control-fed animals. As shown in Table 4, fish oil–fed animals ingested more kcal per day, particularly in the first 3 weeks, than did the animals in the other dietary groups. Despite ingesting more calories, the fish oil–fed animals gained less body weight than did the control- and lard-fed animals (see Table 4). The weight difference between fish oil– and control-fed animals was apparent after 1 week, and the difference in weight increased with time.
Table 4. Weight and Average Daily Energy Consumption
Energy Consumed (kcal/d)
NOTE. F344 rats were fed diets enriched in carbohydrate (control), fish oil, or lard as described in the Experimental Procedures Section. After each week, the total grams of food ingested were measured. The control diet contained 3.77 kcal/g and the high-fat diets contained 4.89 kcal/g. Weekly energy consumption was calculated by multiplying the grams of food consumed in a week by 3.77 (control) or by 4.89 (high fat). The average daily energy consumption was calculated by dividing the weekly energy consumption by 7 days. The data represent the mean daily energy consumption kcal/d ± S.E.M. (n = 10).
Weights (in grams) were measured weekly. Each point represents the mean weight ± S.E.M. (n = 10).
P < .05
P < .01
P < .001 (compared with control, unpaired t test).
Percent body fat was measured using DXA. The initial fraction of body fat in all animals was approximately 16%. After 4 weeks, lard-fed animals had more body fat (25.1 ± 0.6%) than did the control-fed animals (22.1 ± 0.8%, P < .05). Fish oil–fed animals (18.8 ± 0.5%) had less body fat than did control-fed animals (P < .05).
The effect of dietary fatty acid composition on the intrahepatic and serum lipid metabolism was measured. Intrahepatic triglyceride concentrations were 27% less in fish oil–fed animals than in control-fed animals (3.8 ± 0.4 mg/mg wet tissue versus 5.2 ± 0.9 mg/mg wet tissue, P < .05). Intrahepatic triglyceride concentrations in lard-fed animals (14.0 ± 3.4 mg/mg wet tissue) was approximately three times higher than in control-fed animals (P < .01). Fasting and postprandial serum triglyceride concentrations and fasting free fatty acid concentrations were less in fish oil–fed animals than in control-fed animals (Table 5). Fasting triglyceride and free fatty acid concentrations were less in lard-fed animals than in control-fed animals, but the postprandial rises in these lipids were markedly greater in the lard-fed animals. Finally, fasting serum leptin concentration was lowest in the fish oil–fed animals and highest in the lard-fed animals. These values reflect the relative body fat in these animals. After the animals had eaten, serum leptin concentrations increased in the animals that were fed a control diet, but not in animals that were fed high-fat diets.
Table 5. Parameters of Lipid Metabolism at 4 Weeks
NOTE. Each value represents the mean ± S.E.M. (n = 5).
Abbreviations: T0, serum obtained from fasted animals; T3, serum obtained from animals 3 hours after food was presented.
The effect of dietary fatty acids on glucose metabolism is shown in Table 6. Fish oil–fed animals had lower postprandial glucose concentrations, and lower fasting and postprandial insulin concentrations than did the control-fed animals. We estimated insulin sensitivity by the Quantitative Insulin Sensitivity Check Index (QUICKI; see Table 6 legend for formula). Insulin sensitivity was significantly higher in fish oil–fed animals than in control-fed animals. By contrast, fasting glucose concentrations were significantly higher in lard-fed animals than in control-fed animals. However, there were no differences in the fasting or postprandial insulin concentrations or in the insulin sensitivity index in the control- and lard-fed animals.
Table 6. Fasting and Refed Serum Glucose, Insulin, and Insulin Sensitivity at 4 Weeks
NOTE. Each value represents the mean ± S.E.M. (n = 5).
Abbreviations: T0, serum obtained from fasted animals; T3, serum obtained from animals 3 hours after food was presented; QUICKI, Quantitative Insulin Sensitivity Check Index, which is calculated by the following formula: 1/[loginsulin + logglucose] where insulin concentration is measured in μU/mL and glucose concentration in mg/dL.
The effects of dietary fatty acid composition on fatty acid oxidation genes are shown in Fig. 2. Fasting PPARα gene expression was similar in the animals that were fed control, fish oil, and lard diets (see Fig. 2, top panel). In response to refeeding, PPARα gene expression decreased by approximately 50% in animals fed a control or lard diet. The postprandial decline in PPARα gene expression was markedly blunted in fish oil–fed animals. CPT1 (see Fig. 2, middle panel) and ACO (see Fig. 2, bottom panel) gene transcription are, in part, regulated by PPARα. As with PPARα mRNA, fasting levels of CPT1 and ACO mRNA were similar in the high-fat and control diets. After refeeding, CPT1 and ACO gene expression fell by 50% to 80% in the control- and lard-fed animals but fell by only 20% to 25% in fish oil–fed animals.
The effects of fatty acid composition on fatty acid synthetic genes are shown in Fig. 3. Fasting SREBP-1c gene expression was similar in animals fed control or high-fat diets (see Fig. 3, top panel). In response to refeeding, the SREBP-1c mRNA increased approximately fivefold in the control-fed animals, threefold in the lard-fed animals, and not at all in the fish oil–fed animals. The results were similar with FAS mRNA (see Fig. 3, middle panel), an SREBP1-c–regulated gene. The FAS response to refeeding was markedly blunted in animals that were fed a high-fat diet. The gene expression of SCD-1 was not as responsive to feeding as that of SREBP1-c or FAS (see Fig. 3, bottom panel). SCD-1 mRNA levels in animals that were fed diets enriched with fat were lower than in animals that were fed diets enriched with carbohydrates. Fish oil feeding resulted in the lowest levels of SCD-1 mRNA of all the dietary groups.
In F344 rats, which serve as a model for prediabetes mellitus, we have shown that dietary fatty acids greatly influence total body and hepatic fat composition. Animals that consumed diets enriched in saturated and monounsaturated fatty acids (lard) gained more weight (see Table 4), body fat, and hepatic triglycerides than the animals that consumed diets high in carbohydrate. By contrast, animals that consumed diets enriched in n-3 polyunsaturated fatty acids (fish oil) had gained less weight (see Table 4), body fat, and hepatic triglycerides than did the animals in the other dietary groups. Furthermore, animals that ingested fish oil had lower postprandial glucose concentrations, lower fasting and postprandial insulin concentrations, and a higher index of insulin sensitivity than did the animals in the other dietary groups (see Table 6).
Steady-state concentrations of intrahepatic triglycerides depend on the balance between triglyceride uptake and secretion and between triglyceride synthesis and degradation. To determine mechanisms for changes in hepatic triglyceride concentrations, we measured both fasting and postprandial nutrient fluxes. We propose that dietary n-3 fatty acids reduce intrahepatic stores of triglycerides by several mechanisms. First, dietary n-3 fatty acids decrease fasting and postprandial serum triglycerides (see Table 5). This response likely reduces the hepatic uptake of triglycerides. Second, a fish oil diet decreases fasting (relative to a control diet) and postprandial (relative to a lard diet) plasma concentrations of free fatty acids (see Table 5). Therefore, diets enriched in fish oil may decrease the hepatic influx of free fatty acids, which are important precursors to triglyceride synthesis. Fasting free fatty acid concentrations depend, in part, on how well insulin inhibits lipolysis in a given animal. The lower fasting free fatty acid concentrations in fish oil–fed animals likely reflect their enhanced insulin sensitivity (see Table 6). We do not have a good explanation for why the free fatty acid concentrations in lard-fed animals are lower than in control-fed animals, except for biologic variation in a relatively small study. Third, n-3 polyunsaturated fats that enter the liver are more likely degraded (see Fig. 2) and are less likely synthesized (see Fig. 3) to triglycerides than are saturated fats. Although we did not directly measure the degradation and synthesis of fatty acids and triglycerides, we did measure the gene expression of key proteins that are involved in fatty acid degradation and synthesis. The gene expression of these proteins is profoundly regulated by fasting and refeeding. We found that dietary fatty acids differentially regulate gene expression of these proteins within hours after the animals have eaten (see Figs. 2 and 3). However, dietary fatty acids have little effect on the gene expression of these proteins after a fast.
PPARα is the major transcription factor that regulates fatty acid degradation within the liver.22, 23 Free fatty acids, particularly the 20- and 22-carbon n-3 polyunsaturated fatty acids found in fish oil, stimulate PPARα activity.24, 25 In fasting subjects, lipolysis is stimulated, and increasing amounts of endogenous free fatty acids are delivered to the liver and stimulate PPARα activity. During carbohydrate ingestion, insulin-mediated inhibition of lipolysis reduces the free fatty acid flux to the liver, and PPARα activity decreases (see Fig. 2, control diet). However, high-fat feeding elevates the subsequent concentration of circulating free fatty acids (see Table 5), presumably because of the hydrolysis of triglycerides by lipoprotein lipase. The increased postprandial free fatty acids are delivered to the liver and bind to PPARα. Because n-3 polyunsaturated fatty acids stimulate PPARα activity more effectively than saturated and monounsaturated fatty acids,24, 25 PPARα-regulated gene expression, such as that for CPT1 and ACO, is higher in fish oil–fed animals than in lard-fed animals (see Fig. 2). Therefore, n-3 polyunsaturated feeding probably stimulates fatty acid oxidation, even when the subject is in the postprandial state, when fatty acid oxidation is normally inhibited.
SREBP1-c is the major transcription factor that regulates the synthesis of fatty acids within the liver.13 In control-fed animals, SREBP1-c gene expression is low in fasted animals but is high in refed animals (see Fig. 3); insulin mediates most of this response.26, 27 Because the high-fat diets contain less carbohydrate than the control diets, the postprandial insulin response is reduced and the SREBP1-c gene expression is blunted (see Fig. 3). Polyunsaturated fatty acids also regulate SREBP1-c gene expression by insulin-independent mechanisms. For instance, polyunsaturated fats diminish SREBP1-c gene expression by accelerating transcript decay28 and by antagonizing the ligand-dependent activation by liver X receptor.29, 30 We have found that polyunsaturated fats inhibit SREBP1-c and FAS gene expression in postprandial subjects (see Fig. 3). The effect of polyunsaturated fats on the low fasting SREBP1-c mRNA levels is minimal. The n-3 polyunsaturated fat mediated reduction in postprandial gene expression of SREBP1-c and FAS probably diminish the synthesis of fatty acids and triglycerides in subjects in the postprandial state.
Other factors induced by dietary n-3 polyunsaturated fatty acids likely contribute to a reduction in intrahepatic triglyceride concentrations. Animals that ingested fish oil had less total body fat and improved insulin sensitivity compared with animals that ingested carbohydrates and lard (see Table 6). Interventions that improve insulin sensitivity have been shown to reduce intrahepatic triglyceride storage; such interventions include weight loss,31, 32 metformin therapy,33 and thiazolidenedione therapy.34 Previous interventional studies in animals5, 6 and epidemiologic studies in humans7–9 have shown that ingestion of fish oil and other polyunsaturated fats improves insulin sensitivity. Storlien and colleagues have assessed insulin action by the euglycemic clamp technique.5 They found that fish oil improves insulin action in the liver, which is the major organ of endogenous glucose production, and in skeletal muscle, which is the major site of glucose disposal. Fish oil may improve insulin sensitivity by decreasing the deposition of excess lipids in muscle and liver, as proposed in the lipotoxicity hypothesis.3, 35 The lipotoxicity hypothesis has been strengthened by findings that insulin sensitivity improves when triglyceride stores are reduced by pharmacologic and hormonal therapies in nonadipose tissue.35–40 In the present study, fish oil lowered intrahepatic triglyceride content and increased insulin sensitivity.
Fish oil feeding may improve insulin sensitivity by reducing total body fat. Animals that were fed fish oil were leaner than control- and lard-fed animals, even if the caloric intakes were equal (see Table 4). Other investigators have observed less body fat in animals that were fed diets enriched with fish oil than in animals that were fed isocaloric diets enriched with other types of fatty acids.41 Because various fatty acids have the same energy values when they are metabolized, the differences in fat deposition among dietary fatty acids must be due to their differential ability to regulate total energy expenditure. Therefore, animals that ingest fish oil must have an increased basal metabolic rate (thermogenesis), energy with activity, or energy associated with food digestion (thermal effect of food). Baillie and coworkers found that dietary fish oil reduces fat deposition by increasing the expression of mitochondrial uncoupling proteins, which thereby increase thermogenesis.42 Other pathways—such as stimulation of hypothalamic sympathetic nerve pathways by polyunsaturated fats—may also reduce fat accumulation in fish oil–fed animals.43