Eicosapentaenoic acid (20:5 n-3) increases fatty acid and glucose uptake in cultured human skeletal muscle cells.

This study was conducted to evaluate the chronic effects of eicosapentaenoic acid (EPA) on fatty acid and glucose metabolism in human skeletal muscle cells. Uptake of [14C]oleate was increased >2-fold after preincubation of myotubes with 0.6 mM EPA for 24 h, and incorporation into various lipid classes showed that cellular triacylgycerol (TAG) and phospholipids were increased 2- to 3-fold compared with control cells. After exposure to oleic acid (OA), TAG was increased 2-fold. Insulin (100 nM) further increased the incorporation of [14C]oleate into all lipid classes for EPA-treated myotubes. Fatty acid β-oxidation was unchanged, and complete oxidation (CO2) decreased in EPA-treated cells. Basal glucose transport and oxidation (CO2) were increased 2-fold after EPA, and insulin (100 nM) stimulated glucose transport and oxidation similarly in control and EPA-treated myotubes, whereas these responses to insulin were abolished after OA treatment. Lower concentrations of EPA (0.1 mM) also increased fatty acid and glucose uptake. CD36/FAT (fatty acid transporter) mRNA expression was increased after EPA and OA treatment compared with control cells. Moreover, GLUT1 expression was increased 2.5-fold by EPA, whereas GLUT4 expression was unchanged, and activities of the mitogen-activated protein kinase p38 and extracellular signal-regulated kinase were decreased after treatment with OA compared with EPA. Together, our data show that chronic exposure of myotubes to EPA promotes increased uptake and oxidation of glucose despite a markedly increased fatty acid uptake and synthesis of complex lipids.


INTRODUCTION
Skeletal muscle is the major site of insulin-stimulated glucose disposal, and insulin resistance of skeletal muscle is strongly linked to the development of type 2-diabetes.
Increased lipid availability is thought to play a role for insulin resistance in skeletal muscle (1,2). It has long been recognized that increased plasma free fatty acids (FFAs) as well as triacylglycerol (TAG) are associated with insulin resistance in vivo in humans, and several studies show that the intramyocellular content of lipids (e.g. TAG) is increased (1)(2)(3)(4)(5). Both the amount and type of fatty acids are important. Diets high in saturated fats or rich in linoleic acid and n-6 fatty acids lead to insulin resistance due mainly to effects in oxidative skeletal muscle (6,7). The physiological alterations in metabolic flux induced by high-saturated fat-feeding mimic those reported in patients with type 2-diabetes (8). However, replacement of some of the fat with long-chain n-3 polyunsaturated fatty acids (PUFAs) prevents the development of insulin resistance caused by high-fat feeding (9)(10)(11). The mechanism by which fish oil improve peripheral insulin resistance is unclear, but it may involve changes in glucose (12) as well as fatty acid metabolism in skeletal muscle. Mechanisms may be linked to membrane incorporation (13), changes in the level of muscle lipid species such as acyl-CoA (14), diacylglycerol (DAG) and lipid storage (TAG) (2,5), altered lipid oxidation (15,16), and/or interference with insulin signaling (17). Reduced plasma concentrations of TAG and FFA observed after dietary intake of n-3 fatty acids may also contribute to less lipid flux to skeletal muscle, decreased muscular fatty acid oxidation and lipid storage (9).
The present study was conducted to determine the effects and explore mechanisms of long-term exposure to eicosapentaenoic acid (EPA, a n-3 PUFA from fish oil) on glucose and lipid metabolism in human myotubes. The actions of EPA were by guest, on August 15, 2017 www.jlr.org Downloaded from compared to oleic acid (a common monounsaturated fatty acid) and a fatty acid-free control (containing BSA, bovine serum albumin). by guest, on August 15, 2017 www.jlr.org

Downloaded from
Groningen, The Netherlands. All other chemicals used were standard commercial, high purity quality.

Human skeletal muscle cell cultures
A cell bank of satellite cells was established from muscle biopsy samples of the m. vastus lateralis of 6 healthy volunteers, age 24.7 ± 0.7 years, body mass index 23.6 ± 1.1 kg/m 2 , fasting glucose and insulin within normal range and with no family history of diabetes. The biopsies were obtained with informed consent and approval by the National Committee for Research Ethics, Oslo, Norway. Muscle cell cultures free of fibroblasts were established by the method of Henry et al. (18) with minor modifications. Briefly, muscle tissue was dissected in Hams F-10 media at 4°C, dissociated by three successive treatments with 0.05% trypsin/EDTA, and satellite cells were resuspended in SkGM with 2% FCS, 50 U/ml penicillin, 50 g/ml streptomycin, 1.25 g/ml amphotericin B and no added insulin. The cells were grown on culture wells or flasks coated with ECM gel (19). After 1-2 weeks, at about 80% confluence, growth medium was replaced by αMEM with 2% FCS, 50 U/ml penicillin, 50 g/ml streptomycin and 1.25 g/ml amphotericin B in order to induce differentiation of myoblasts into multinucleated myotubes. The cells were cultured in humidified 5% CO 2 atmosphere at 37°C, and medium was changed every 2-3 days.
All cells used in the experiments were subcultured 3 to 6 times. All myotube cultures were used for analysis on day 8 after onset of differentiation.

Pre-treatment of myotubes with fatty acids
On day 7, differentiated myotubes were pre-treated with 0.6 mM of fatty acids, either oleic acid (OA) or eicosapentaenoic acid (EPA), in αMEM with 2% FCS for 24 h. Fatty acids 0.6 mM were bound to 0.24 mM fatty acid-free albumin (BSA) (ratio FA/BSA 2.5/1). BSA (0.24 mM) was used as fatty acid-free control.

Lipid distribution
Myotubes were incubated in 6 well plates for 4 h in αMEM with [1-14 C]oleic acid (0.5 µCi/ml, 0.6 mM), 0.24 mM fatty acid-free albumin (BSA) ± 100 nM insulin to study basal and insulin-mediated cellular oleate distribution, respectively. After incubation, myotubes were placed on ice, washed twice with ice-cold PBS, and harvested into a tube in two additions of 250 µl distilled water and stored at -20°C.
The cells were later sonicated and assayed for protein content, and cellular lipids were extracted. The homogenized cell fraction (400 µl) was mixed with 20 volumes of chloroform:methanol (2:1, v/v) (20) with FCS (50 µl) as a carrier. After 30 min, 4 volumes of 0.9% NaCl (pH 2) was added, and the mixture was centrifuged (1000g, 5 min). The organic phase was evaporated under a steam of nitrogen at 45°C. The residual lipid extract was redissolved in 200 µl n-hexane and separated by thin-layer chromatography (TLC), using hexane:diethylether:acetic acid (65:35:1, v/v/v) as mobile phase. The various lipid bands were visualized by iodine, excised and the radioactivity quantified by liquid scintillation.
After centrifugation (1800g, 10 min), 500 µl of the supernatant was counted by liquid scintillation. No-cell controls were included.
Carbon dioxide. Cells were cultured in 12.5 cm 2 flasks, and the myotubes were

Glycogen synthesis
Myotubes were incubated for 60 min in serum-free αMEM ± insulin (1-100 nM) at 37°C, before addition of D-[ 14 C(U)]glucose (1 or 2 µCi/ml, 5.5 mM) ± insulin (1-100 nM). After 60-120 min the cells were washed three times with ice-cold PBS and lysed with 1 M KOH. Synthesized glycogen was measured as described by Franch et al. (23). The response to insulin (100 nM) was a more than 2-fold increase in glycogen formation when compared to unstimulated cells, and glycogen synthesis was linear within 4 h.

Glucose oxidation
Myotubes were incubated in 12.5 cm 2 flasks made airtight by stopper tops in serumfree αMEM with [1-14 C]D-glucose (0.5 µCi/ml, 5.5 mM), 20 mM HEPES ± 100 nM insulin. After 4 h, 300 µl phenyl ethylamine:methanol (1:1, v/v) was added with a syringe to a centre well containing a folded filter paper. Subsequently, 300 µl 1 M perchloric acid was added to the cells through the stopper tops by use of a syringe.
The flasks were placed for a minimum of 4 h at room temperature to trap labelled CO 2 . The filter paper was counted by liquid scintillation. No-cell controls were included to correct for unspecific CO 2 trapping.

RNA isolation and analysis of gene expression by real time PCR
RNA was either obtained as total RNA (totRNA) or messenger RNA (mRNA) according to the following protocols. For isolation of totRNA, human skeletal muscle cells grown on 60 mm dishes were washed, trypsinized and pelleted before totRNA was isolated by RNeasy Mini kit according to the supplier totRNA isolation protocol. Total and messenger RNA was reversely transcribed with oligo primers using a Perkin-Elmer Thermal Cycler 9600 (25°C for 10 min, 37°C for 1h, 99°C for 5 min) and a TaqMan reverse-transcription reagents kit. One µg total RNA or 1 µl mRNA per 10 µl total TaqMan reaction solution was added. Real time PCR was performed using an ABI PRISM ® 7000 Detection System (Applied Biosystems). RNA expression was determined by SYBR ® Green, and primers were designed using Primer Express® (Applied Biosystems). Each target were quantified in triplicates and carried out in 25 µl reaction volume according to the supplier protocol. All assays were run for 40 cycles (95°C for 12 s followed by 60°C for 60 s). The housekeeping control genes α-tubulin and β-actin were both measured, and transcription levels are presented relative to β-actin.

Statistics and data presentation
All values are reported as means ± SEM. The value n represents the number of different donors used. All experiments were run with duplicate or triplicate samples unless otherwise stated. Statistical analyses were performed with StatView ® (SAS Institute Inc., USA). Comparison of different treatments was evaluated by ANOVA and Fisher's PLSD. A p value < 0.05 was considered significant.

Effect of 24 h pre-treatment with fatty acids on 14 C-oleic acid uptake, distribution into cellular lipids and oxidation
Skeletal muscle cells from young, healthy subjects were differentiated for 8 days prior to measurement of oleate (OA) uptake, esterification to lipids and oxidation. Lipid metabolism was examined after 24 h exposure of myotubes to various fatty acids, after which the cells were incubated with [1-14 C]oleic acid (OA, 0.6 mM) for another 4 h. Uptake of labelled OA (sum of cell-associated lipids and oxidized oleate) was significantly increased by 84% after pre-treatment with 0.6 mM EPA as compared to the fatty acid-free BSA control and by 44% when compared to OA pre-treatment (Fig.   1A). Pre-treatment with OA did not affect subsequent uptake of labelled OA (Fig.   1A). EPA pre-treatment dose dependently increased OA uptake demonstrating that EPA also significantly increased fatty acid uptake at lower concentrations (Fig. 1B).
Incorporation of labelled oleic acid into various lipid classes after treatment with EPA showed that cellular triacylglycerol (TAG), and phospholipids (PL) were 2.5-3-fold increased when compared to the BSA control ( Fig. 2A). In addition, cellular free OA (FFA) was also increased by 87% versus the BSA control after EPA treatment ( Fig.   2A). After exposure of myotubes to OA for 24 h, a more than 2-fold increase in TAG was observed, whereas the other lipid classes remained unchanged ( Fig. 2A).
Oxidation of oleate after the 24 h fatty acid exposure was determined by measurement of acid soluble metabolites (ASM, β-oxidation products) ( Fig. 2A) and carbon dioxide (CO 2 ) released from the cells. There was no change in formation of ASM in EPAtreated myotubes when compared to the BSA control, despite a much higher oleate uptake ( Fig. 2A). Production of CO 2 was on the other hand decreased by 50% after pre-treatment with EPA (4.9 ± 0.3 for EPA vs. 9.9 ± 0.4 nmol/mg cell protein/4 h for BSA, respectively). Finally, ASM formation was decreased by 25% for OA compared to the BSA control ( Fig. 2A).

Short-term effects of insulin on oleic acid uptake and metabolism
Insulin-mediated 14 C-oleate metabolism after fatty acid pre-treatment of myotubes was also studied. In the presence of insulin (100 nM) for 4 h, total 14 C-oleate uptake was similarly increased above basal level by 22% for OA and the BSA control, whereas it was 38% above basal for EPA (Fig. 1). The net insulin effect was thus 2.5-3-fold higher for EPA than the BSA control and OA (Fig. 1). Figure 2B shows the acute effects of insulin on metabolism of oleate into various cellular lipids and ASM.
Insulin markedly increased incorporation of labelled oleate into all lipid classes in EPA-treated myotubes when compared to control and OA-treated cells, which had similar insulin responses as control cells. Fatty acid oxidation (ASM) on the other hand, was not significantly changed by short-term insulin treatment (Fig. 2B).

Effect of pre-treatment with fatty acids on 14 C-oleic acid partitioning
For further insight into how fatty acid-treated myotubes metabolize labelled OA, partitioning of OA between oxidation (ASM) and complex lipids was calculated.
Fatty acid pre-treatment changed fatty acid distribution in favor of lipid synthesis and decreased fatty acid oxidation (Fig. 3). Insulin-mediated effects on partitioning showed a further significant change from fatty acid oxidation towards synthesis of complex cellular lipids for EPA and control myotubes (Fig. 3).

Effect of 24 h pre-treatment with fatty acids on glucose transport and metabolism
Glucose transport in the absence of insulin (basal) was 2.4-fold increased after preincubation of the myotubes for 24 h with 0.6 mM eicosapentaenoic acid (EPA) when compared to the fatty acid-free control (BSA) and oleic acid (OA) (Fig. 4A). Acute treatment of the myotubes with insulin (100 nM) further increased (p<0.05) glucose transport for control and EPA, whereas insulin failed to increase glucose transport after OA treatment (Fig. 4A). Glucose transport was also significantly (p<0.05) increased after pre-treating myotubes with a lower concentration of EPA (0.1 mM) for 24 h; basal by 19 ± 7% and insulin-stimulated (100 nM) by 70 ± 8% of the BSA control, respectively. Glucose transport measured after shorter pre-treatment (4 h) of myotubes with 0.6 mM EPA was 113 ± 16% for basal and 114 ± 6% with insulin of the BSA control, respectively, demonstrating that the effect on glucose transport was dependent on long-term exposure to this fatty acid.
In order to further study intracellular glucose metabolism, glycogen synthesis and glucose oxidation were determined after fatty acid pre-treatment. Basal glycogen synthesis was significantly decreased by 52% with EPA compared to the BSA control, whereas OA had no effect (Fig. 4B). Acute insulin treatment increased (p<0.05) glycogen synthesis more than 2-fold in both control, EPA and OA treated cells (Fig. 4B). However, the net insulin effects were markedly lower after exposure of the myotubes to EPA, a reduction of 53% was observed (Fig. 4B). Glycogen synthesis was also measured after exposure to 0.1 mM of EPA for 24 h; basal as well as insulin-stimulated (100 nM) were significantly (p<0.05) decreased by 33 ± 2 and 35 ± 2 % , respectively of the BSA control.
Basal glucose oxidation (CO 2 formation) was 2-fold increased after pre-incubation of the myotubes with 0.6 mM of EPA for 24 h when compared to control and OAtreated cells (Fig. 4C). Glucose oxidation in the presence of insulin (100 nM) was further increased (p<0.05) to the same extent for both EPA and control myotubes, and totally abolished by OA pre-treatment (Fig. 4C).

Expression of genes involved in fatty acid and glucose metabolism
Since pre-treatment with EPA had profound effects on oleic acid uptake and lipid synthesis we studied the effect on some genes of central role in lipid metabolism. The expression of the fatty acid transporter CD36/FAT was significantly increased by 65% after EPA and 2-fold after OA pre-incubation compared to the BSA control (Fig.   5). Diacylglycerol acyltransferase-1 (DGAT-1), the terminal enzyme involved in TAG biosynthesis, was not changed by any of the treatments (Fig. 5). Since pretreatment with EPA had profound effects on glucose transport and glucose oxidation, we also studied mRNA levels of the glucose transporters GLUT1 and GLUT4. After 24 h pre-treatment with EPA, GLUT1 mRNA expression was 2.5-fold increased, whereas the levels of GLUT4 were not changed after exposure of the myotubes neither to EPA nor OA (Fig. 5).

Effect of pre-treatment with fatty acids on PKB/Akt, ERK and p38 activity
The signalling pathways affected by EPA are poorly explored. Previously, we found that pre-treatment with the saturated fatty acid palmitate for 24 h reduced the activation of protein kinase B (PKB)/Akt of the PI3K pathway and increased the activation of the p38 mitogen activated protein kinase (24) indicating that these signalling pathways may be influenced by long-term exposure to fatty acids. Pre-by guest, on August 15, 2017 www.jlr.org Downloaded from treatment with EPA for 24 h did not affect the activity of PKB/Akt (data not shown).
Oleic acid compared to EPA and the BSA control, decreased the phosphorylation (e.g. activation) of both the extracellular signal-regulated kinase (ERK) and p38 mitogen activated protein kinases (Fig. 6).

DISCUSSION
The present study shows that pre-treatment of differentiated human myotubes with eicosapentaenoic acid (EPA) promote increased uptake and metabolism of 14 C-oleate to cellular lipids, and the increased oleate uptake was not accompanied by an increased 14 C-oleate oxidation. Interestingly, after EPA pre-treatment the net insulin responses on 14 C-oleate metabolism were enhanced despite an elevated lipid synthesis when compared to control cells (BSA) and cells exposed to oleic acid (OA).
Moreover, EPA also increased basal as well as insulin-mediated glucose uptake and oxidation in comparison with myotubes pre-treated with OA, which showed an abolished insulin action on these two prosesses.
Exposure of myotubes to EPA increased basal as well as insulin-mediated 14 C-oleate uptake, and also the expression of the fatty acid transporter CD36/FAT when compared to control cells. CD36/FAT is an important factor involved in fatty acid uptake and intracellular transport. This protein is ubiquitously expressed and its expression correlates with long-chain fatty acid uptake into both heart and muscle (25,26). It was shown that peroxisome proliferator-activated receptor gamma (PPARγ) activation by thiazolidinedione treatment increased palmitate uptake and oxidation in cultured human skeletal muscle cells in concert with upregulation of CD36/FAT expression (26). EPA has previously been shown to induce PPARγ mRNA levels in isolated human adipocytes (27). We also observed a 2.3-fold increase in PPARγ expression after exposure of myotubes to EPA (0.6 mM, 24 h) when compared to OA (data not shown). PPARγ shows a strong preference for binding to PUFAs, such as EPA, as compared to monounsaturated or saturated fatty acids (28).
The upregulation of CD36/FAT mRNA levels seen after EPA treatment observed in this study may be important for the increased 14 C-oleate uptake. Pre-treatment of the myotubes with OA did not significanly increase fatty acid uptake, but the expression of CD36/FAT increased similarly to EPA-exposed cells. It has been observed that PUFAs and their oxidation products stimulate the expression of CD36/FAT in human macrophages (29), most likely through activation of PPARγ, and also activation of p38 MAPK has been shown to be involved in the regulation of CD36/FAT through PPARγ (30). On the other hand, activation of CD36/FAT has been shown to induce phosphorylation of both ERK and p38 (31)(32)(33). Activities of p38 and ERK were significantly decreased in cells treated with OA compared to EPA and control cells; however whether this is a consequence of CD36 activation or contributes to CD36 regulation and fatty acid uptake is unknown. In cardiac myotubes, PKB/Akt is reported to be involved in insulin-induced regulation of CD36/FAT (34), but pretreatment with EPA did not affect the activity of this kinase in human myotubes (data not shown).
There is evidence that insulin can activate fatty acid uptake in rat skeletal muscle by translocation of CD36/FAT to the plasma membrane (35). In human myotubes, insulin acutely increases uptake of the saturated fatty acid palmitate and formation of triacylglycerol (TAG) (26,36). We examined oleate uptake and subsequent metabolism in myotubes under basal conditions and in the presence of short-term high concentration of insulin (100 nM). EPA pre-treatment changed basal 14 C-oleate distribution with an increased level of free OA and enhanced incorporation into TAG and phospholipids (PL) compared to control cells, whereas DAG was unaltered. On the other hand, pre-treatment with OA caused increased synthesis of TAG and a decreased fatty acid β-oxidation. Moreover, after EPA pre-treatment, insulin-induced changes for 14 C-oleate uptake and incorporation into various lipids were enhanced compared to the BSA control and OA. On the contrary β-oxidation of 14 C-oleate as well as complete oxidation to CO 2 were decreased by EPA. Thus, after exposure of myotubes to EPA, cellular partitioning of oleate showed a distinct change from fatty acid oxidation towards increased incorporation into complex cellular lipids, especially PL, also independent of short-term addition of insulin. In accordance with this, a similar change in fatty acid distribution has been observed in rat hepatocytes after overexpression of mitochondrial glycerol-3-phosphate acyltransferase (mtGPAT), which markedly decreased fatty acid oxidation and increased lipid formation, especially PL synthesis (37). The role of an increased formation of PL, and probably changed fatty acid composition, for modulating insulin action in myotubes after exposure to EPA needs to be further elucidated, but it could be related to changes in cellular membrane structure and function and lipid-derived signalling pathways (13).
The fatty acid composition of skeletal muscle cell membrane PL, especially phosphatidylcholine (PC), is known to influence insulin responsiveness in man (38).
An increased intramyocellular content of lipids, especially TAG, is often associated with impaired insulin action (1-5), however, information on the regulatory processes involved in synthesis and degradation of lipids within the muscle and their putative link to insulin resistance, is at present fragmentary. In this study we demonstrate that pre-treatmernt with EPA positively influenced glucose metabolism and insulin action despite an increased fatty acid uptake and synthesis of complex lipid. In support of this, it has been demonstrated in rats that administration of EPA caused a reduction in plasma cholesterol and TAG, but increased cholesterol and TAG contents in skeletal muscle without causing insulin resistance (39). Furthermore, it has also been shown in L6 myotubes pre-treated with palmitic acid that overexpression of carnitine palmitoyltransferase-1 (CPT-1), which accelerates fatty acid β-oxidation, exerts an insulin sensitizing effect independent of changes in intracellular lipid content (40). It was also shown in C2C12 myocytes that overexpression of CD36 increased fatty acid uptake and channeling to a lipase-accessible TAG pool. This TAG pool was related to lipid content in the cells and insulin responsiveness depending on the degree of futile cycling of fatty acids (41). In contrast to EPA, exposure of myotubes to OA showed a decreased fatty acid β-oxidation and a enhanced TAG synthesis only, and also failed to respond to acute insulin stimulation on glucose transport and oxidation. In accordance with this, we have recently shown that exposure of human myotubes to palmitate caused an increased fatty acid uptake and TAG formation, decreased fatty acid oxidation and impaired insulin-mediated glucose uptake (42). Taken together, our data from human myotubes exposed to various fatty acids suggest that there is a complex relationship between metabolism of intracellular lipids and glucose transport and utilisation in skeletal muscle that needs to be addressed in future studies.
Pre-treatment of myotubes with EPA for 24 h promoted increased basal glucose transport, whereas insulin-mediated transport was similar to control cells. An increased basal glucose uptake after prolonged incubation with other fatty acids has also been observed (42,43). Moreover, the mRNA expression of the glucose transporter GLUT1 was 2.5-fold increased after EPA, whereas GLUT4 mRNA expression was unaltered. The increased expression of GLUT1 may explain the markedly increased basal uptake and oxidation of glucose after EPA treatment. The mechanisms behind the EPA-induced GLUT1 expression and increased basal glucose uptake are unknown.
It was also observed that EPA pre-treatment decreased complete oleic acid oxidation (CO 2 ) by the cells. Lower fatty acid oxidation might be a consequence of a markedly increased glucose transport and oxidation as well as an increased de novo lipogenesis. This process where glucose suppress fatty acid oxidation is often referred to as the by guest, on August 15, 2017 www.jlr.org Downloaded from inverse glucose-fatty acid cycle, and is thought to occur via an increased formation of malonyl-CoA and inhibition of CPT-1 (44). It has recently shown that oleic acid oxidation can be reduced during acute hyperglycemia in human myotubes (45). In addition, although de novo lipogenesis is low in human skeletal muscle, we have demonstrated that chronic glucose oversupply leads to increased lipid synthesis by human myotubes that might interfere with fatty acid metabolism (24).
In summary, our data show that chronic exposure of human myotubes to EPA promotes increased uptake of oleic acid and uptake and oxidation of glucose, and the mechanisms may involve elevated expression of CD36/FAT and GLUT1.
Interestingly, despite an enhanced fatty acid uptake and synthesis of complex lipids, the insulin responses after EPA pre-treatment were maintained for glucose uptake and oxidation, and even increased for oleate uptake and distribution into complex lipids.
Thus, our data show that EPA directly modulates lipid and glucose metabolism in human skeletal muscle, and the observed changes may contribute to the beneficial effects of n-3 PUFA in relation to peripheral insulin resistance and type 2-diabetes.