The experimental groups consisting of male Wistar rats (aged 9–10 weeks; weighing 329 ± 2 g, n = 32) were obtained from The University of Queensland Biological Resources unit and individually housed at the University of Southern Queensland’s Animal House Facility. All experimental protocols were approved by the Animal Experimentation Ethics Committee of the University of Southern Queensland, under the guidelines of the National Health and Medical Research Council of Australia. Rats were divided into 4 groups: (i) corn starch (C, n = 8), (ii) high carbohydrate, high fat (H, n = 8), (iii) C + tocotrienol-rich fraction (TRF) (CT, n = 8), (iv) H + TRF (HT, n = 8). This TRF contained α-tocotrienol (31.9%), β-tocotrienol (2.1%), γ-tocotrienol (24.8%) and δ-tocotrienol (18.3%) together with α-tocopherol (22.9%). 120 mg/kg/day palm TRF dissolved in vitamin E-stripped palm olein (240 mg TRF/mL palm olein, Malaysian Palm Oil Board, Malaysia) was given for the final 8 weeks of the 16 weeks protocol via once-daily oral gavage. This dose was chosen as the reported oral no-observed-adverse-effects level in male rats given a similar tocotrienol-tocopherol mixture [15]. All experimental groups were housed in a temperature-controlled, 12-h light/dark cycle environment with ad libitum access to water and food. Measurements of body weight and food and water intakes were taken daily to monitor the day-to-day health of the rats. Feed conversion efficiency (%) was calculated as:

All group-specific diets were prepared in our laboratory. Corn starch diet was prepared by thorough mixing of corn starch, powdered rat feed (meat-free rat and mouse feed; Speciality Feeds, Glen Forrest, WA, Australia), Hubble, Mendel and Wakeman salt mixture (MP Biochemicals, Seven Hills, NSW, Australia) and water, while the corn starch and part of water were replaced with condensed milk, fructose and beef tallow in the high carbohydrate, high fat diet [14]. The drinking water in all high carbohydrate, high fat-fed rats was augmented with 25% fructose.

Echocardiography was performed by trained cardiac sonographers at the Medical Engineering Research Facility, The Prince Charles Hospital, Brisbane, Australia. Rats were anaesthetised via intraperitoneal injection with Zoletil (tiletamine 15 mg/kg, zolazepam 15 mg/kg) and Ilium Xylazil (xylazine 10 mg/kg). Echocardiographic images were obtained using the Hewlett Packard Sonos 5500 (12 MHz frequency fetal transducer) at an image depth of 3 cm using two focal zones. Measurements of left ventricular posterior wall thickness and internal diameter were made using two-dimensional M-mode taken at mid-papillary level [13,16].

Dual-energy X-ray absorptiometric (DXA) measurements using a Norland XR36 DXA instrument (Norland Corp., Fort Atkinson, WI, USA) were performed on the rats after 16 weeks of feeding, 2 days before rats were killed for pathophysiological assessments. DXA scans were analysed using the manufacturer’s recommended software for use in laboratory animals (Small Subject Analysis Software, version 2.5.3/1.3.1; Norland Corp.) as previously described [17]. The precision error of lean mass for replicate measurements, with repositioning, was 3.2%. Visceral adiposity index (%) was calculated from wet weights of fat pads at euthanasia as: and expressed as adiposity per cent [18].

Systolic blood pressure was measured after 0, 4, 8, 12 and 16 weeks under light sedation with i.p. injection of Zoletil (tiletamine 15 mg/kg, zolazepam 15 mg/kg), using an MLT1010 Piezo-Electric Pulse Transducer (ADInstruments) and inflatable tail-cuff connected to a MLT844 Physiological Pressure Transducer (ADInstruments) and PowerLab data acquisition unit (ADInstruments, Sydney, Australia). Abdominal circumference was measured using a standard measuring tape under light sedation. Rats were killed with an intraperitoneal injection of pentobarbitone sodium (100 mg/kg).

Oral glucose tolerance tests were performed after 0, 8 and 16 weeks of diet. After 12 h of fasting, blood glucose concentrations were measured in blood samples taken from the tail vein. Subsequently, each rat was treated with glucose (2 g/kg) via oral gavage. Tail vein blood samples were taken every 30 min up to 120 min following glucose administration. The blood glucose concentrations were analysed with a Medisense Precision Q.I.D glucose meter (Abbott Laboratories, Bedford, MA, USA).

For insulin tolerance testing, basal blood glucose concentrations were measured after 4–5 h of food deprivation as above. The rats were injected i.p. with 0.33 IU/kg insulin-R (Eli Lilly Australia, West Ryde, NSW, Australia), and tail vein blood samples were taken at 0, 30, 60, 90 and 120 min. Rats were withdrawn from the test if the blood glucose concentrations dropped below 1.1 mmol/L, and 4 g/kg glucose was administered immediately by oral gavage to reverse hypoglycaemia.

Changes in the responsiveness of thoracic aorta were defined using organ bath studies. Thoracic aortic rings (4 mm in length) were suspended in an organ bath chamber with a resting tension of 10 mN. Cumulative concentration-response (contraction) curves were measured for noradrenaline (Sigma-Aldrich Australia); concentration-response (relaxation) curves were measured for acetylcholine (Sigma-Aldrich Australia) or sodium nitroprusside (Sigma-Aldrich Australia) in the presence of a submaximal contraction to noradrenaline [14].

The left ventricular function of the rats in all treatment groups was assessed using the Langendorff heart preparation. Terminal anaesthesia was induced via i.p. injection of pentobarbitone sodium (100 mg/kg). Once anaesthesia was achieved, heparin (1000 IU) was injected into the right femoral vein. After removal of the heart, isovolumetric ventricular function was measured by inserting a latex balloon into the left ventricle connected to a Capto SP844 MLT844 physiological pressure transducer and Chart software on a Maclab system. All left ventricular end-diastolic pressure values were measured by pacing the heart at 250 beats per minute using an electrical stimulator. End-diastolic pressure was obtained starting from 0 mmHg up to 30 mmHg. The right and left ventricles were separated and weighed. Diastolic stiffness constant (κ, dimensionless) was calculated [16].

Following euthanasia, the heart, liver, kidneys, visceral fat pads and spleen were removed and blotted dry for weighing. All organ weights were normalised relative to tibial length at the time of removal and presented in mg/mm.

Immediately after removal, heart and liver tissues were fixed in 10% buffered formalin with change of formalin every 3 days for 10 days to remove traces of blood from the tissue. The samples were then dehydrated and embedded in paraffin wax. Thin sections (5 μm) of left ventricle and the liver were cut and stained with haematoxylin and eosin stain for determination of inflammatory cell infiltration. Collagen distribution was observed in the left ventricle following picrosirius red staining. Laser confocal microscopy (Zeiss LSM 510 upright Confocal Microscope) was used to determine the extent of collagen deposition in selected regions.

Blood was collected from the abdominal aorta following euthanasia and centrifuged at 5000× g for 15 min within 30 min of collection into heparinised tubes. Plasma was separated and transferred to Eppendorf tubes for storage at −20 °C before analysis. Plasma concentrations of total cholesterol, triglycerides, non-esterified fatty acids (NEFA), activities of plasma alanine transaminase (ALT) and aspartate transaminase (AST) were determined using kits and controls supplied by Olympus using an Olympus analyser (AU 400 Tokyo, Japan) [14].

All data sets were represented as mean ± standard error of mean (SEM). Comparisons of findings between groups were made via statistical analysis of data sets using one-way and two-way analysis of variance (ANOVA). When interaction and/or the main effects were significant, means were compared using Newman-Keuls multiple-comparison post hoc test. A p-value of <0.05 was considered as statistically significant. All statistical analyses were performed using Graph Pad Prism version 5.00 for Windows.

Feeding of the high carbohydrate, high fat (H) diet for 16 weeks increased systolic blood pressure compared with cornstarch (C) diet. With TRF supplementation for 8 weeks, blood pressure was normalised in rats fed with H diet (Figure 1). H feeding diminished noradrenaline contraction in isolated thoracic aortic rings (Figure 2A) and vascular relaxation responses to sodium nitroprusside (SNP) and acetylcholine (ACh) compared with C rats (Figure 2B,C). With TRF, thoracic aortic contractions to noradrenaline were improved (Figure 2A) but relaxation responses were unchanged (Figure 2B,C).

Compared to C group, H rats demonstrated eccentric hypertrophy, defined as an increased left ventricular weight and internal diameter in diastole (LVIDd) without any changes in relative wall thickness, with increased stroke volume and cardiac output (Table 1). H rats showed impaired systolic function seen as reduced fractional shortening. H rats showed decreased contractility, measured as maximal rate of positive rise of pressure (+dP/dt) and negative rise of pressure (−dP/dt), and LV developed pressure, in the isolated heart. Compared with H rats, relative wall thickness of HT rats was reduced. Systolic function was improved in TRF-treated rats, as shown by the increased fractional shortening, ejection fraction and cardiac output and decreased ejection time. Functionally, the increased diastolic stiffness in H rats was decreased in HT rats with improved −dP/dt. Histology of the H heart showed marked inflammatory cells infiltration (Figure 3C) and collagen deposition (Figure 3G) in the left ventricle compared with C rats. These changes were normalised with TRF treatment (Figure 3D,H).

After 16 weeks, rats fed the H diet had increased body weight compared with C. H-fed rats had increased abdominal circumference and their visceral adiposity index was significantly higher than that of C rats (6.8% ± 0.6% compared to 3.6% ± 0.2%, respectively, Table 2). Consistent with the body weight results, fat mass and total visceral adipose tissue (retroperitoneal, epididymal and omental fat pads) in H rats were higher than in C rats. TRF treatment in H rats reduced omental, but not epididymal or perirenal fat pads, reduced abdominal circumference and increased food intake and lean mass. H rats had higher total cholesterol, triglyceride and NEFA plasma concentrations compared with C rats; TRF reduced triglyceride and NEFA but not total cholesterol concentrations (Table 2).

H rats had higher fasting blood glucose concentration than C rats (Table 2) TRF treatment decreased the blood glucose concentrations in HT rats. The plasma glucose response to oral glucose loading was greater in H rats than C rats (Figure 4A). At 120 min, HT, CT and C rats had lower plasma glucose concentrations than H rats. In the insulin tolerance test, C, CT and HT rats had lower area under the curve than H rats (Figure 4B).

H rats had 1.8-fold greater plasma ALT and AST activities compared with C rats (Table 2). Liver weights of H rats were higher than C rats. With TRF, plasma AST and ALT activities were reduced by 41.2% and 40%, respectively, in HT rats. There were no differences in liver weight of HT and H rats. H rats developed fat vacuoles within the hepatocytes, showing ballooned hepatocytes with portal inflammatory cell infiltration compared with C rats (Figure 5E,G). Livers from the HT group showed a markedly attenuated degree of fatty change, with decreased fat vacuole size and number compared with the H group (Figure 5H). In addition, the HT rats displayed normalised portal inflammatory cells infiltration compared with H group (Figure 5C,D).