In addition to decreased insulin sensitivity, diabetes is a pathological condition associated with increased inflammation. The ω-3 fatty acids have been proposed as anti-inflammatory agents. Thus, the major goal of this study was to analyze the effects of fatty acid supplementation on both insulin sensitivity and inflammatory status in an animal model of type 2 diabetes. Diabetic rats (Goto-Kakizaki model) were treated with eicosapentaenoic acid (EPA) or linoleic acid at 0.5 g/kg body weigh (bw) dose. In vivo incorporation of 14C-triolein into adipose tissue was improved by the ω-3 administration. In vitro incubations of adipose tissue slices from EPA-treated rats showed an increase in 14C-palmitate incorporation into the lipid fraction. These observations were linked with a decreased rate of fatty acid oxidation. EPA treatment resulted in a decreased fatty acid oxidation in incubated strips from extensor digitorum longus (EDL) muscles. The changes in lipid utilization were associated with a decrease in insulin plasma concentration, suggesting an improvement in insulin sensitivity. These changes in lipid metabolism were associated with an activation of AMP-activated protein kinase (AMPK) in white adipose tissue. In addition, EPA treatment resulted in a decreased content of peroxisome proliferator-activated receptor-α (PPARα) and PPARδ and in increased GLUT4 expression in skeletal muscle. Moreover, EPA increased 2-deoxy-D-[14C]glucose (2-DOG) uptake in C2C12 myotubes, suggesting an improvement in glucose metabolism. Concerning the inflammatory status, EPA treatment resulted in a decreased gene expression for both tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6) both in skeletal muscle and adipose tissue. The data suggest that EPA treatment to diabetic rats clearly improves lipid metabolism although the evidences on insulin sensitization are less clear.
Type 2 diabetes mellitus is a disease characterized by both insulin resistance and alterations in insulin secretion. The decrease in the insulin sensitivity is reflected in peripheral tissues, particularly in skeletal muscle and adipose tissue. The overall picture is decreased glucose utilization and a tendency to hyperglycemia, which results in metabolic alterations (1). These alterations are associated with a mild inflammatory status, suggesting that inflammatory mediators (i.e. cytokines) may be involved in both glucose and lipid metabolic alterations (2). The pharmacological treatment of type 2 diabetes is based on the administration of antidiabetogenic agents, which basically lower circulating glucose by favoring the uptake of the sugar by peripheral tissues (3).
Different diabetic animal models have been developed in order to study insulin sensitivity and lipid metabolism. Thus, the Otsuka Long-Evans Tokushima Fatty rats, a genetic model of the spontaneous development of type 2 diabetes mellitus, show innate polyphagia, which causes rapid body-weight gain (obesity) resulting in glucose intolerance with hyperinsulinemia (4). Another model is the Goto-Kakizaki rat, which is a spontaneous model of type 2 diabetes. This animal model displays profoundly defective insulin secretion leading to basal hyperglycemia (5). We chose this second model to perform our studies as we wanted to focus on the type 2 diabetes rather than obesity.
Currently, the approach to mitigate and control type 2 diabetes is a combination between pharmacotherapy and lifestyle. The latter focuses on weight reduction, decreased total and saturated fat consumption, and increased physical activity (6). From this point of view, healthy diets improve the lifestyle management in type 2 diabetes (7). Dietary fish oils are known to improve insulin sensitivity in non-insulin-dependent diabetes mellitus animal models (8). However, the mechanism of fish oils' effects on decreasing insulin resistance is unclear. Fish oils contain long-chain polyunsaturated ω-3 fatty acids such as eicosapentaenoic (EPA) and docosahexaenoic (DHA) fatty acids. Dietary intake of ω-3 fatty acids decreases the plasma concentration of blood lipids including cholesterol and triacylglycerol (9). Storlien et al. (8) observed that fish oil supplements prevented the insulin resistance due to a high-fat diet. In addition, ω-3 fatty acids increase bleeding time, decrease platelet aggregation (10), blood viscosity, and fibrinogen (11), and increase erythrocyte deformability (12), thus decreasing the tendency to thrombus formation. They therefore have a positive effect on diseases such as hypertension (13), arthritis (14), atherosclerosis (15), myocardial infarction (16), thrombosis (17), and some cancers (18). Bearing all this in mind the aim of this investigation was to study the effects of long-term EPA treatment on substrate utilization and inflammatory status in diabetic rats.
Methods and Procedures
Male 3-week-old Goto-Kakizaki rats (Taconic, Ry, Denmark) were used. This is a rat model with spontaneous type 2 diabetes (5). The animals were maintained on a regular light-dark cycle (light on from 08:00 am to 08:00 pm) at an ambient temperature of 22 ± 2 °C and had free access to food and water. The diet (B.K. Universal, Barcelona, Spain) consisted of 455–485 g carbohydrate/kg diet, 185 g protein/kg diet, and 310 g fat/kg diet (the remainder was nondigestible material, vitamins, and minerals). The weight of the animals and food and water intakes were measured daily. All animal manipulations were made in accordance with the European Community guidelines for the use of laboratory animals.
Animal groups and treatments
Rats were divided into one of three groups: control, EPA and linoleic acid (LINO), respectively. Control group: animals fed a control diet ad libitum (normal laboratory chow diet). These animals received only the vehicle (carboxymethylcellulose 1%). EPA group: animals fed the same as control but supplemented daily with 0.5 g/kg body weight (bw) of EPA mixed in carboxymethylcellulose 1% given by a single gavage. LINO group: animals fed the same as control but supplemented daily with 0.5 g/kg bw of LINO mixed in carboxymethylcellulose 1% given by a single gavage. The treatment took place for 4 weeks (28 days) before the animals were killed.
Circulating glucose and insulin levels
Blood glucose concentrations were measured using Accutrend GCT device (Roche, Basel, Switzerland). Plasma levels of insulin were obtained by a quantitative sandwich enzyme immunoassay technique (Linco Research, St Charles, MO). The enzyme activity is measured spectrophotometrically by the increased absorbance at 450 nm, corrected from the absorbance at 590 nm.
First-strand complementary DNA was synthesized from total RNA with oligo dT15 primers and random primers pdN6 by using a complementary DNA synthesis kit (Transcriptor Reverse Transcriptase; Roche, Basel, Switzerland). Analysis of mRNA levels for each gene was performed with specific primers designed to detect each product (Table 1). To avoid the detection of possible contamination by genomic DNA, primers were designed in different exons. The real-time PCR was performed using a commercial kit (LightCycler FastStart DNA Master SYBR Green I; Roche, Indianapolis, IN). The relative amount of all mRNA was calculated using comparative CT method. 18S mRNA was used as the invariant control for all studies.
Table 1. Primer sequences
Total muscle and adipose tissue protein level in each sample were determined with the Bradford technique. Crude muscle homogenates (100 µg protein/sample) were separated by electrophoresis, transferred to polyvinylidene difluoride membranes, blocked with nonfat milk, and incubated overnight with the selective anti-AMP-activated protein kinase (AMPK) antibody (provided by A. Zorzano, University of Barcelona, Spain) and anti-Na+/K+ ATPase antibody (Transduction Laboratories, Lexington, KY). Specific proteins were detected with horseradish peroxidase-conjugated secondary antibodies and a chemiluminescence kit. Blots were scanned with an imaging densitometer and optical densities were quantified with the software Diversity database 2.1.1 (BioRad Laboratories, Philadelphia, PA) for all of the samples.
Measurement of lipid oxidation and tissue lipid accumulation
About 0.5 µCi of [1−14C]-triolein (dissolved in 0.5 ml of triolein) per rat was given enterally by gastric intubation, without anesthetic but with minimal stress to the animal, after 28 days of EPA/LINO fatty acid administration. The metabolic fate of an orally administered [1−14C]-lipid load was examined as described by Oller do Nascimento and Williamson (19). After the collection period, animals were weighed and anesthetized with ketamine/xylazine mixture. Blood was collected from the abdominal aorta into heparinized tubes and centrifuged (3500g, 10 min, 4 °C) to obtain plasma. Samples were taken of gastrocnemius, liver, heart, dorsal white adipose tissue, epididymal adipose tissue, and brown adipose tissue. Tissues were rapidly excised, weighed, and frozen in liquid nitrogen. The gastrointestinal tract (plus contents) was homogenized in 150 ml of 3% (wt/vol) HClO4. Samples of tissues and plasma were saponified and the lipid was extracted (20). The extracted fatty acids were dissolved in 8 ml of liquid-scintillation fluid for determination of 14C-lipid formation. The amount of absorption was calculated by subtracting total gastrointestinal radioactivity from the amount that had been administered.
Muscle and adipose tissue preparations and incubations
Rats previously treated with either EPA, LINO, or vehicle were used for in vitro whole muscle and adipose slices incubations. The dissection and isolation of the extensor digitorum longus (EDL) muscles (21) were carried out under ketamine/xylazine mixture anesthesia. The isolated muscles were fixed to a stainless-steel clip in order to maintain the muscle under slight tension (making it comparable to resting length) during the incubation. Such muscles are able to maintain normal ATP and phosphocreatine concentrations during a 3-h incubation period (22). The tissues (both EDL muscles and dorsal white adipose tissue slices) were incubated in a shaking-thermostated water bath at 35 °C for 3 h in 2 ml of Krebs-Henseleit physiological saline pH 7.4, containing 5 mmol/l glucose, branched-chain amino acids (170 µmol/l l-leucine, 100 µmol/l l-isoleucine, 200 µmol/l l-valine), and 20 mmol/l HEPES. After the addition of the muscles to the vials, the incubation started at a shaking rate of 45 cycles/min. Vials were gassed with O2/CO2 (19:1) during the entire incubation period. The tissues were incubated for 3 h: the first 30 min in supplemented medium and then for 150 min in fresh supplemented medium containing 0.2 mmol/l palmitic acid.
Metabolic fate of 14C-palmitate
The palmitic acid was incubated with albumin at 37 °C for 90 min in order to form a complex with the protein. EDL muscles and dorsal white adipose tissue slices were incubated for 1 h in supplemented medium containing 0.2 mmol/l palmitic acid and 0.2 µCi [1−14C]palmitic acid and albumin. The oxidation of the tracer was assessed by measuring the 14CO2 produced. Following incubation, the medium was acidified using HClO4 (60%) and the 14CO2 released was collected by means of hyamine hydroxide. Total radioactivity was measured in 4 ml of liquid-scintillation fluid. The muscles and adipose tissue slices incubated were homogenized in liquid nitrogen, and the lipid fraction was extracted using a chloroform:methanol mixture. It was later dissolved in 8 ml of liquid-scintillation fluid and total radioactivity was measured. For measuring the 14C-palmitate incorporation to protein fraction, the hydrosoluble fraction of lipid extraction was incubated with Protosol (Dupont, Boston, MA) at 50 °C for 2 h and then, hydrogen peroxide was added. It was later dissolved in 10 ml of liquid-scintillation fluid and total radioactivity was determined.
C2C12 mouse skeletal muscle cells were obtained from the American Type Culture Collection (Manassas, VA). Cells were passaged in high-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 100 U/ml penicillin, 100 µg/ml streptomycin, 25 ng/ml fungizone, 110 µg/ml sodium pyruvate, and 2 mmol/l l-glutamine in a humidified atmosphere of 5% CO2 and 95% air at 37 °C. For experimental analyses, cells were seeded at 3.7 × 104 cells/cm2 in 10% fetal bovine serum/DMEM until they reached 90–100% confluence 24 h later. At this time, the medium was replaced by DMEM containing 10% horse serum for induction of differentiation. Abundant myotube formation, monitored microscopically, occurred after 4 days in 10% horse serum/DMEM. Such fused myotube cultures were utilized for experimental analyses 6 days after transferring cells to 10% horse serum/DMEM.
Measurement of in vitro 2-DOG uptake
For the estimation of glucose uptake, 2-deoxy-D-[14C]glucose (2-DOG) was used as tracer. In order to test the effects of EPA and LINO treatments, myotubes (day 6 after differentiation) were incubated for 180 min in 2 ml of Krebs-Henseleit saline solution pH 7.4, containing 2 mmol/l pyruvate, 20 mmol/l HEPES, 25 mmol/l 2-DOG, and 0.01 µCi [1−14C]2-DOG, containing 500 µmol/l of EPA, LINO, or none. To obtain the cells, we added 1 ml of 0.5 mol/l NaOH—0.1% Triton X-100 and incubated then at 37 °C for at least 1 h. Later, 250 µl of the lysate was placed in scintillation vials containing 10 ml of scintillation cocktail with 25 µl of 1.74 N acetic acid and processed for 14C counting. Protein content was determined using the BCA assay (Pierce Chemical, Rockford, IL). The results were expressed as nmol 2-DOG/mg protein.
Radiochemicals, [1−14C]-triolein, [1−14C]-palmitate, and 2-deoxy-D-[14C]glucose were obtained from Amersham International (Amersham, UK). EPA and LINO were provided by Novartis Nutrition (Hopkins, MN).
Statistical analysis of the data was performed by means of the Student's t-test (STATS, Oxford, UK).
Animal model: effects of EPA on weight and food intake
Administration of EPA for 28 days at a dose of 0.5 g/kg bw had no effects on body weight, food or water intakes (Table 2). The lack of effect was observed both in relation with control and LINO-treated animals. Similarly, neither EPA nor LINO treatment had any effect on the weight of skeletal muscle, adipose tissue, liver, heart, kidneys, spleen, lungs, gastrointestinal tract, or carcass (data not shown). Interestingly, EPA treatment induced a significant decrease in brain weight (12%). This decrease was not observed in the LINO-treated group.
Table 2. Animal weights and food intake
Effects of EPA on glucose metabolism
Although EPA treatment did not modify the glucose concentrations (Figure 1) or glucose tolerance curves (data not shown), ω-3 fatty acid treatment had important effects on both insulin sensitivity and glucose transport. In this way, administration of EPA resulted in a significant decrease in circulating insulin (48%); this was not observed in the LINO group (Figure 1). In addition, administration of EPA resulted in an increased GLUT4 mRNA content in both skeletal muscle (41%) and adipose tissue, although in this last tissue, the increase (51%) did not reach statistical significance (Figure 2a,b). Interestingly, LINO treatment significantly increased GLUT4 expression in adipose tissue but not in skeletal muscle (Figures 2a,b). Moreover, EPA treatment resulted in changes in AMPK content in adipose tissue (see Supplementary Figure S1a,b online). In addition, EPA increased 2-deoxyglucose uptake (Table 3) in C2C12 myotubes whereas LINO did not.
Table 3. 2-DOG uptake in C2C12 myotubes
Effects of EPA on lipid metabolism
The results observed in relation with carbohydrate uptake lead us to evaluate whether EPA treatment also resulted in changes in lipid oxidation and/or accumulation. Bearing this in mind, we studied the in vivo metabolic fate of 14C-triolein load. The animals were given a triolein bolus and absorption of 14C-lipid deposition was studied 6 h after the gavage. The results presented in Table 4 clearly show that neither EPA nor LINO treatments had any effects on the intestinal absorption of the exogenous lipid load. In relation with the incorporation of lipid of the tracer into 14C-lipid fraction, no changes were observed in gastrocnemius muscle, liver, heart, or plasma (data not shown). However, EPA treatment resulted in an important increase in 14C-lipid accumulation in perigenital white adipose tissue (both expressed per gram of total tissue) compared to the LINO group but not to the control. The LINO treatment resulted in a significant decrease in 14C-lipid accumulation in dorsal white adipose tissue (26%) compared to the control.
Table 4. In vivo incorporation of 14C-triolein into adipose tissue 14C-lipid
Figure 3a,b shows the effects of EPA treatment on the metabolic fate of 14C-palmitate in isolated preparations of skeletal muscle and adipose tissue of Goto-Kakizaki rats. In addition, we decided to see whether the exogenous uptake and accumulation of fatty acids was also affected by ω-3 fatty acid treatment. The results obtained (Figure 3b) clearly show that EPA treatment resulted in an increase in lipid accumulation in adipose tissue (38%). This increased fatty acid accumulation was associated with a decreased fatty acid oxidation (50%) as measured as 14CO2 production (Figure 3b). Similarly, EPA treatment induced a very important decrease in 14C-palmitate oxidation (77%) in skeletal muscle (Figure 3a). However, none of these changes were observed in the LINO-treated group. Supplementary Figure S2a clearly shows that EPA treatment resulted in a decreased mRNA content of PPARα (55%) and PPARδ (94%) in skeletal muscle. No changes were observed in PPARγ gene expression as a result of either EPA or LINO treatment.
Although no differences were seen, as a result of EPA treatment, on the circulating levels of either TNF-α and IL-6 (data not shown), administration of EPA resulted in an important decrease in the mRNA content of both TNF-α (49%) and IL-6 (75%) in skeletal muscle (Figure 4). Interestingly, LINO treatment had similar effects on this tissue. Concerning adipose tissue, EPA treatment clearly resulted in a decreased IL-6 gene expression; the levels of expression of TNF-α were also lower in the EPA-treated group, although the results did not reach statistical significance.
In western society, the rate of obesity is increasing dramatically, in some countries, such as the United States, reaching >30% of the total population (23). Interestingly, obesity is very strongly associated with type 2 diabetes, which is a disease that is also growing at an exponential rate. The rate of disease is also increasing in young individuals. This is being accompanied by alarmingly higher health-care costs associated with poorer quality of life in the individuals affected (24). It is therefore essential to find, in addition to the pharmacological management of the disease, nutritional supportive measures. From this point of view, fish oil-rich diets have been suggested to ameliorate some of the metabolic disturbances associated with type 2 diabetes. Therefore, the aim of the present investigation was to determine the potential benefit of a high ω-3 fatty acid diet on an animal model of experimental type 2 diabetes.
In order to study the effects of EPA in type 2 diabetes, we chose an animal model, the Goto-Kakizaki rat, which is a spontaneous model of this pathology; the animals display profoundly defective insulin secretion leading to increased hyperglycemia.5 For the purpose of comparison, the study design included a LINO-treated group (LINO; an ω-6 fatty acid) at 0.5 g/kg bw dose. From this point of view, different studies (25) have suggested that administration of ω-6 polyunsaturated fatty acids to diabetic animals results in a decrease in insulin sensitivity. In addition, rather than anti-inflammatory, several studies (26) indicate that this type of fatty acid may behave in a proinflammatory manner.
Neither EPA nor LINO treatment affected tissue weights with the exception of brain, which decreased as a consequence of EPA treatment. Although it is somehow difficult to interpret this observation, a recent publication has shown that the Goto-Kakizaki model has very high triglyceride content in brain as compared with Wistar rats; surprisingly this high triglyceride content is associated with null EPA levels in brain (27). It may then be possible to speculate that EPA treatment could, to some extend, induce changes in brain lipid composition and weight.
In our study, EPA (but not LINO) also decreased circulating insulin levels. Previous studies using long-term EPA administration in both humans (28) or Otsuka Long-Evans Tokushima Fatty rats (29) have not observed any differences in circulating insulin levels following EPA administration. This discrepancy may be explained by a difference in EPA dosing or because the animal model used was both diabetic and obese whereas the Goto-Kakizaki rat is diabetic but normal weight. In any case, the reduction in circulating insulin is an indication that EPA may contribute to improved insulin sensitivity.
In addition, EPA resulted in an increased GLUT4 mRNA content in both skeletal muscle and adipose tissue. Other studies have also observed an improvement in GLUT4 gene expression in skeletal muscle in an animal model of diabetes (30). Similarly, Delarue et al. (31) have put forward the idea that EPA can prevent the decrease in GLUT4 mRNA content in muscle and adipose tissue induced by a high-fat diet. Concerning AMPK, this enzyme is considered to be one of the best indicators for insulin sensitivity, as its activation in vivo improves blood glucose homeostasis, cholesterol concentrations, and blood pressure in insulin-resistant rodents (32). This last result therefore reinforces the idea that EPA may be improving insulin sensitivity, at least at the level of adipose tissue. Previous work has revealed a reciprocal regulation of muscle and fat depots (33). In addition, EPA increased 2-deoxyglucose uptake in C2C12 myotubes, suggesting that the effects of ω-3 polyunsaturated fatty acids on glucose metabolism are direct on the muscle cells. In summary, either GLUT4 gene expression or AMPK protein content—perhaps the best mediators of insulin sensitivity—are increased by EPA either in fat or muscle tissue suggesting that ω-3 fatty acid treatment may result in an improvement of insulin resistance and its effects are direct, at least in muscle cells. It could be interesting to perform further research into the involvement of different proteins related to insulin action, such as PI3K, mTOR, or alternatively those that interfere with the hormone signaling pathways, such as SOCS-3, JNK, and IKKβ. Further work will concentrate on these aspects.
EPA treatment had no effects on the intestinal absorption of an exogenous lipid load; moreover, it resulted in an increased lipid accumulation in white adipose tissue but not in skeletal muscle. However, Kusunoki et al. have shown that administration of EPA to diabetic rats results in an accumulation of lipid in skeletal muscle (29). The increased uptake of 14C-lipid in adipose tissue could well be related to an improved triglyceride-rich lipoprotein metabolism. Indeed, EPA seems to improve the low-density lipoprotein metabolism (34), favoring lipid deposition in peripheral tissues. Interestingly, Singer et al. (35) have shown a negative correlation of EPA and lipid accumulation in hepatocytes of individuals with diabetes. This fact also suggests that EPA may be able to improve triglyceride-rich lipoprotein metabolism, in this case facilitating liver fatty acid exportation and, consequently, the low-density lipoprotein secretion (35). Again these results support the idea that insulin sensitivity may be improved by EPA treatment.
The treatment with EPA also resulted in an increase in lipid accumulation in isolated preparations of adipose tissue, associated with a decreased fatty acid oxidation (50%). In addition, EPA decreased palmitate oxidation in skeletal muscle preparations incubated in vitro. However, none of these changes were observed in the LINO-treated group. These data, together with the results obtained in the in vivo experiment, clearly suggest that EPA favors tissue lipid accumulation and clearly there is an effect on an improvement in insulin sensitivity stated above. Concerning the decreased levels of both PPARα and PPARδ mRNA in skeletal muscle, this observation could possibly be associated with the decrease in fatty acid oxidation previously observed. Indeed, these two transcription factors are regulators of lipid and glucose metabolism, allowing adaptation to the prevailing nutritional environment (36). Indeed, PPARδ has been associated with fuel preference in skeletal muscle; Brunmair et al. (37) have shown that its activation allows skeletal muscle to switch fuel preference from glucose to fatty acids. The results found here suggest that EPA, by decreasing PPARδ in skeletal muscle, may favor glucose utilization rather than fatty acids.
It is well known that type 2 diabetes is associated with an inflammatory status, partially induced by humoral mediators, such as cytokines (38). We, therefore, decided to measure both the circulating levels and tissue mRNA content for both TNF-α and IL-6, both cytokines are well-known markers of inflammation (39). The results obtained in this study indicate that EPA treatment is clearly promoting a decrease in inflammatory status. The decrease in IL-6 in skeletal muscle, in addition to suggesting a reduced inflammatory state, may act to regulate muscle substrate utilization, being related with the decreased lipid oxidation. From this point of view Al-Khalili et al. (40) have demonstrated, using human skeletal muscle cells in culture, that IL-6 directly promotes skeletal muscle differentiation and regulates muscle substrate utilization promoting glycogen storage and lipid oxidation.
As stated above, in addition to the different pharmacological approaches for type 2 diabetes, nutritional management of the disease seems essential. The results presented in this investigation reinforce the idea that supplemental marine fatty acid may be beneficial for the diabetic animal (patient). Indeed, EPA treatment, as given here, seems to improve insulin sensitivity, mainly by increasing GLUT4 expression, AMPK protein content and glucose uptake in myotube cultures, as previously commented. Because type 2 diabetes is an inflammatory disease it is also interesting to note that the data presented provide an indication that there may be a decrease in inflammation as measured by expression of the well-known inflammatory mediators, TNF-α and IL-6. In conclusion, future disease management of the diabetic patient may well benefit from the inclusion of the ω-3 fatty acid supplements in the diet.
This work was supported by Novartis Nutrition Corporation (project: DIAB-RES-01-04 ES). We are particularly grateful to Norman A. Greenberg, Anne Falk, and Fariba Roughead for their comments and suggestions. We are grateful to A. Zorzano (University of Barcelona) for providing anti-AMPK antibody.