The experimental protocol was approved by the Institutional Animal Care and Use Committee of Maastricht University and complied with the principles of laboratory animal care (approval number 2006-024). During the experiments all efforts were made to minimize suffering of the animals.

Twenty-two, eight-week old male Wistar rats were obtained from Charles River Laboratories and were housed individually in an environmentally controlled laboratory (temperature 22±1°C and relative humidity 55±2%). Rats were randomly divided in two groups and were allowed ad libitum access to food and tap water. One group (n = 10) received standard CHOW (10 energy % as fat) (Ssniff, Germany) and the other group (n = 12) received a high-fat diet (HFD) that contained 45 energy % of fat derived from lard (Research Diets, D01060502, Wijk bij Duurstede, The Netherlands). Diets were provided for three weeks.

Two weeks before the end of the experiment the rats were sedated under isoflurane anaesthesia and permanent indwelling catheters were placed in the right jugular vein (for the sampling of blood) and the left carotid artery (for the infusion of insulin and glucose) under aseptic conditions. Both catheters were tunnelled subcutaneously and exteriorized at the top of the head, where they were attached. To maintain patency, the catheters were filled with a mixture of 85% glycerol (Merck, Darmstadt, Germany) and 500 U/ml heparin (Leo Pharma BV, Breda, The Netherlands). All rats regained pre-surgery body mass within 3 days.

One-week before the hyperinsulinemic euglycemic clamp overexpression of mouse DGAT1 in the left Tibialis anterior (TA) muscle of the rat was obtained, with the right TA as sham-electroporated internal control. DNA electroporation was done under isoflurane anaesthesia. Left TA muscles were transcutaneously injected with 150 µg (2 µg/µl) pcDNA sport DGAT1-construct (pCMV-sport6, Invitrogen, Breda, The Netherlands) in 0.9% sterile NaCl. Right TA muscles were injected with 150 µg (2 µg/µl) empty vector. Within 15 seconds after the last injection 5 electric pulses were applied by two stainless steel plate electrodes placed at each side of the leg to obtain transient permeabilization of the cell membrane, facilitating the passage of the DNA. One high voltage pulse of 800 V/cm and four low voltage pulses of 80 V/cm at 1 Hz were generated by an ECM 830 electroporator (BTX, San Diego) as described before [19], [20], [21]. Shaving the leg and the application of conductive gel minimized impedance.

Hyperinsulinemic euglycemic clamping of conscious rats was performed after a 6 h fast. A primed continuous infusion of insulin (Actrapid HM; Novo Nordisk, Copenhagen, Denmark) was administered at a rate of 13 mU/kg/min for 120 min. Blood glucose values were monitored at 10-min intervals throughout the clamp. The glucose infusion rate (GIR) was adjusted to maintain blood glucose concentration within the range of 4.5–5.5 mmol/L. Insulin stimulated glucose disposal in vivo within the individual TA muscles was studied according to a previously described method [22]. Upon completion of the clamp, the rats were killed with an overdose of sodium pentobarbital and TA muscles were quickly dissected and frozen in liquid nitrogen-cooled isopentane and stored at −80°C until further analyses.

Tibialis anterior muscles were cut in three parts and the middle part was ground with a N₂-cooled mortar and pestle. After homogenisation, the samples were heated to 99°C for 30 min, cooled and centrifuged for 5 min at 13,000 rpm. 2-deoxy-[1-3H]-D-glucose-phosphate was separated from 2DG in the supernatant by binding to a Dowex column (Dowex 1*8-100, ion exchange resin, Sigma). Bound 2-deoxy-[1-3H]-D-glucose-phosphate was eluted from the column by using the eluate buffer (formic acid, ammoniumacetate in mQ, pH 4.9). 14 ml LCD scintillation fluid was added to 1 ml eluate, and this cocktail was counted in the scintillation counter for 5 min.

Cryosections (5 µm) from the midbelly region of the TA muscle were stained for neutral lipids with Oil Red O and DGAT1 overexpresssion was quantified as previously described [15].

In a subset of 4 CHOW and 4 high-fat fed mice both the DGAT1 overexpressing muscle as well as the empty-vector control condition were processed for routine transmission electron microscopy. In brief, small blocks of muscle tissue (2*2*2 mm) from both legs were fixed by glutaraldehyde and post-fixed by osmiumtetroxide and epon embedded. Semi-thin sections (4 microns) were cut to confirm longitudinal fiber orientation by light microscopy. Subsequently, ultrathin (80–90 nm) sections were cut and stained and contrasted with uranyl acetate and lead-citrate. Then, sections were examined in a Philips CM100 transmission electron microscope (Philips, Eindhoven, The Netherlands). Per muscle 15–20 images were grabbed and analyzed by an experienced electron microscopist blinded for the conditions. Using the NIH image analysis program all individual lipid droplets were identified, counted and encircled to compute lipid droplet area and derived morphometric parameters. Frequency distributions were computed and plotted using Excel for Windows.

Total lipids were extracted from frozen muscle using the method of Folch et al. [23]. The extracts were filtered and lipids recovered in the chloroform phase. DAG was isolated using thin-layer chromatography on Silica Gel 60 A plates developed in petroleumbenzin-diethyl ether-acetic acid (120∶25∶1.5 by volume) and visualized by rhodamine 6G. The DAG band was scraped from the plate and methylated using a mixture of toluene-methanol- (BF3/methanol 14%) (20%-55%-25% by volume), as described by Morrison and Smith [24]. The methylated fatty acids were extracted with hexane and analyzed by capillary gas liquid chromatography using a 50 m ×0.25 mm CP-sil 88 silica column (varian) with helium as carrier gas at a flow of 130 kPa. The column oven was maintained at 165°C for 10 min and increased at a rate of 5°C/min to 190°C. This temperature was maintained for 15 min. Then temperature was increased to 230°C for 22 min with a flow rate of 2°C/min. Fatty acid methyl esters were identified by comparing retention times to those of known standards. Inclusion of an internal standard, 19∶0 (dinonadecanoic acid), and an average molecular mass for each fatty acid methyl ester permits quantification of the amount of DAG in the sample, expressed as micromolar per gram of wet tissue weight.

Ceramide content was determined as previously described (25) with slight modifications. Briefly, 50 µl of muscle homogenate was extracted with 600 µl of CHCl₃/MeOH 1/2 (vol/vol). The extract was centrifuged for 10 min at 14,000 g and the pellet discarded. 500 µl of CHCl₃/MQ-H₂O 1/1.5 (vol/vol) was added, mixed, and centrifuged for 3 min at 14,000 g to separate the phases. The lower phase was collected and the upper phase re-extracted with 400 µl of CHCl₃. The combined lower phases were dried under N₂ flow, taken up in 500 µl of freshly prepared 0.1 M NaOH in MeOH, and deacetylated in a microwave (SAM-155, CEM corp.) for 60 min. 50 µl of this solution was derivatized with 25 µl o-phtaldehyde (OPA) reagent. The OPA-derivatized lipids were separated by HPLC and ceramides were quantified as previously described, using C17 sphinganine as an internal standard [25].

Muscle samples were homogenized as described previously [26] and processed for standard SDS-PAGE and Western blotting. Protein concentration was assessed, and equal amounts of protein were loaded per lane. Actin was used as a loading control. Membranes were incubated with antibodies against OXPAT (GP31, Progen, Heidelberg, Germany), OXPHOS (MS601; MitoSciences, Eugene, OR, USA), PGC1α (516557; Calbiochem; VWR International BV, Amsterdam, The Netherlands) and mUCP3 [27], ADRP (GP40, Progen), ATGL (2138; Cell Signalling Technology; Bioké; Leiden, The Netherlands) and CGI58 (NB110-41576; Novus Biologicals, Littleton, CO, USA). Blots incubated with OXPAT, OXPHOS, PGC1α and mUCP3 were probed with IRDye800-conjugated or IRDye700-conjugated secondary antibodies (Rockland, Gilbertsville, PA, USA and LICOR Biosciences, Westburg, Leusden, The Netherlands), and bands at a molecular weight corresponding to the control samples were quantified using the Odyssey infrared imaging system (LICOR Biosciences, Wateringen, the Netherlands). Blots detecting ADRP, ATGL and CGI58 were probed with the appropriate horse radish-conjugated antibodies and after incubation specific protein bands were visualized by chemiluminescence and analyzed by Chemidoc XRS system (Bio-Rad, Veenendaal, The Netherlands).

Results are presented as means ± SEM. Differences between legs within one rat were analysed by paired samples t-test. Between groups, differences were analysed with an independent t-test. The accepted level of statistical significance was p<0.05 for all analyses. Al calculations were done using the Statistical Package for the Social Sciences (SPSS 16.0 software).

Body mass gradually increased over the three weeks of dietary intervention in both the CHOW- and HFD-fed rats. However, this short-term dietary intervention did not result in differences in body mass between diets (300±10 g and 290±10 g in HFD- vs. CHOW-fed rats, p = 0.70).

One week after electroporation, immunohistochemistry revealed a widely distributed DGAT1 overexpression in the muscle fibres of left TA muscle of rats; both in rats receiving the HFD as well as the CHOW. The mean green intensity of images from the DGAT1 electroporated leg was significantly higher than observed in the sham-electroporated contralateral control leg (25.77±2.09 AU vs. 8.32±1.68 AU in HFD-DGAT1 vs. HFD-control, p<0.001, and 25.73±3.17 AU vs. 7.37±1.79 AU in CHOW-DGAT1 vs. CHOW-control, p<0.001), indicating significant overexpression of DGAT1. The DGAT1 overexpression was associated with a ∼52% increase in intramyocellular lipids in the HFD-DGAT1 overexpressing leg and a ∼66% increase in muscle triglyceride content in the CHOW-DGAT1 overexpressing leg, compared to control legs of the animals (p = 0.018, HFD-DGAT1 vs. HFD-control leg, and p = 0.044, CHOW-DGAT1 vs. CHOW-control leg, figure 1).

To determine whether the increased IMCL content in TA muscle overexpressing DGAT1 was accompanied by a reduction of DAG, we measured the latter in the legs overexpressing DGAT1 as well as in their empty vector control legs (figure 2). Three weeks of high-fat feeding was associated with a ∼75% increase in DAG content (p = 0.039). Remarkably, overexpressing DGAT1 led to a ∼30% increase in DAG content in the HFD-DGAT1 overexpressing leg and a ∼35% increase in DAG content in CHOW-DGAT1 overexpressing leg, compared to control legs of the animals (p = 0.073, HFD-DGAT1 vs. HFD-control leg, and p = 0.045, CHOW-DGAT1 vs. CHOW-control leg).

Ceramide content was measured in TA muscle overexpressing DGAT1 and in the contralateral empty vector electroporated legs to determine whether the increase in IMCL content is linked to changes in muscle ceramide levels (figure 3). Compared the CHOW-fed animals, three weeks of high-fat feeding did not increase ceramide content in TA muscle. Furthermore, also DGAT1 overpression did not affect ceramide levels both in CHOW-DGAT1 overexpressing legs and in HFD-DGAT1 overexpressing legs compared to their controls, respectively.

One week after electroporation of DGAT1 whole-body and tissue insulin sensitivity were assessed in vivo by hyperinsulinemic-euglycemic clamps in combination with radioisotope-labelled infusion of 2-deoxy-[3H]glucose (2DG). Compared to CHOW-fed animals, three weeks of high-fat feeding induced whole body insulin resistance as reflected by a significantly lower glucose infusion rate (30.4±1.2 mg/min/kg vs. 35.2±1.3 mg/kg/min in HFD- vs. CHOW-fed animals, p = 0.02, figure 4A). In CHOW fed rats, 2DG uptake was significantly elevated in DGAT1 overexpressing TA muscle compared to their empty vector control legs (p = 0.036, figure 4B). Also upon high-fat feeding, DGAT1 overexpression significantly increased 2DG uptake in TA muscle compared to the empty vector control legs of these rats (p = 0.05, figure 4C). These results were not confounded by altered basal GLUT4 expression, as this was not affected by DGAT1 overexpression (data not shown).

To test if the increase in lipid content in DGAT1 overexpressing muscle was accompanied by a compensatory increase in mitochondrial capacity, we determined protein levels of major mitochondrial markers. DGAT1 overexpression was associated with a ∼32% increase in PGC1α in the HFD-DGAT1 overexpressing leg compared to the control leg of the animals (p = 0.04, figure 5A), suggesting an increased mitochondrial biogenesis. High fat feeding did not affect the PGC1α content in TA muscle. DGAT1 overexpression in CHOW animals had no significant effect on PGC1α expression (p = 0.64, figure 5A). Western blotting of the OXPHOS complexes revealed that DGAT1 overexpression, in rats fed a HFD, led to an increase in protein levels of the individual complexes I, III and V, suggesting an increased mitochondrial density (Complex I: 1.2±0.4 AU vs. 0.86±0.26 AU, p = 0.007, Complex II: 0.97±0.31 AU vs. 0.98±0.31 AU, p = 0.89 Complex III: 1.0±0.32 AU vs. 0.77±0.23 AU, p = 0.0023, Complex V: 1.0±0.30 AU vs. 0.89±0.26 AU, p = 0.014 in HFD-DGAT1 vs. HFD-control leg, figure 5B). DGAT1 overexpression did not enhance mitochondrial density in rats receiving CHOW represented as differences in protein expression of the individual complexes (Complex I: 0.95±0.39 AU vs. 0.92±0.37 AU, p = 0.72, Complex II: 0.97±0.31 AU vs. 0.96±0.31 AU, p = 0.96, Complex III: 0.83±0.29 AU vs. 0.97±0.34 AU, p = 0.35, Complex V: 0.94±0.30 AU vs. 1.06±0.34 AU, p = 0.20 in CHOW-DGAT1 vs. CHOW-control leg, figure 5B). Western blotting of UCP3 revealed a marked ∼33% decrease in protein content after overexpression of DGAT1 in HFD-fed rats, which was nearly significant (p = 0.09, HFD-DGAT1 vs. HFD-control leg, figure 5C). UCP3 protein levels were not significantly different between the HFD-DGAT1 overexpressing leg and both the CHOW-DGAT1 and CHOW-control legs (p = 0.88, CHOW-DGAT1 vs. CHOW-control leg, figure 5C).

Oil Red O staining showed a marked increase in the number of intramyocellular lipid droplets upon DGAT1 overexpression, both in CHOW- and HFD-fed animals. However, electron microscopy revealed that DGAT1 overexpression on CHOW resulted in a massive increase in the number of lipid droplets of submicron size, which do not exceed the detection limit of light microscopy (figure 6A, upper panel). Upon high-fat feeding however, DGAT1 overexpression mainly increased the size of the lipid droplets, rather than the number (figure 6A lower panel). In contrast, in CHOW-fed rats, DGAT1 overexpression did not further increase the diameter of the lipid droplets (Figure 6A, upper panel). The increased size of the lipid droplets in the HFD-DGAT overexpressing leg, compared to the empty-vector control leg, corresponded with an increase in the lipid droplet coating protein OXPAT (p = 0.08, HFD-DGAT1 vs. HFD-control leg, and p = 0.14, CHOW-DGAT1 vs. CHOW-control leg, figure 6B).

The finding that IMCL, DAG and markers of oxidative capacity were increased in DGAT1 overexpressing muscle may suggest that the turnover of IMCL is increased in the insulin-sensitive DGAT1 overexpressing legs. To test this hypothesis we determined the protein expression of ATGL and CGI58 as markers of intramyocellular triglyceride lipolysis. DGAT1 overexpression was associated with a ∼35% increase in ATGL content in the HFD-DGAT1 overexpressing leg compared to the control leg of the animals (p = 0.04, HFD-DGAT1 vs. HFD-control leg, figure 7A). DGAT1 overexpression in CHOW-fed rats tended to further enhance the ATGL protein content (p = 0.06, CHOW-DGAT1 vs. CHOW-control leg, figure 7A). Western blotting of CGI58, a cofactor of ATGL showed similar results. DGAT1 overexpression led to a ∼25% increase in CGI58 protein in the HFD-DGAT1 overexpressing leg compared to the control leg of the animals (p = 0.009, HFD-DGAT1 vs. HFD-control leg, figure 7B). DGAT1 overexpression in CHOW animals did not affect CGI58 protein expression (p = 0.16, CHOW-DGAT1 vs. CHOW-control leg, figure 7B). Accordingly, ADRP, a lipid droplet coating protein suggested to be involved in the regulation of lipolysis, was ∼25% increased in the HFD-DGAT1 overexpressing leg compared to the control leg of the animals (p = 0.05, HFD-DGAT1 vs. HFD-control leg, figure 7C). DGAT1 overexpression in CHOW animals had no effect on ADRP expression (p = 0.90, CHOW-DGAT1 vs. CHOW-control leg, figure 7C).