In an experimental study, fifty male weaning Wistar rats (Pasture Institute, Iran), 21 days of age on arrival were housed individually in wire bar-floor cages. Body weight (BW) in grams was recorded on arrival and weekly thereafter. Food intake was also monitored twice a week. The animals (weight: 35.59 ± 0.39) were allowed 1 week of acclimatization in a standard environment at 22°C, 50% humidity and 12-h light/dark cycles with free access to food and water. During the first week, all animals were fed a standard laboratory chow (Pasture Institute, Iran) and afterwards, they were randomly assigned to five groups. Diet composition and energy value are given in Table 1. The composition of standard diet (SD) was completely the same as the commercial AIN-93G. Diet treatments included high fat diets (HFD) with soy oil (HF-S), butter (HF-B), fish oil (HF-F) or olive oil (HF-O). High fat diets were prepared to provide equal vitamins and minerals per calorie, and contained equal percentages by weight of fiber. After an 8-week feeding period, blood samples were collected from the retro-orbital veins into polypropylene tubes containing sodium EDTA (1-mg/mL blood) and aprotinin (500-unit/mL blood, Bayer) in the non-fasted and 24 h-fasted state for measurement of serum parameters and were then centrifuged for 15 min at 3000x g. Plasma samples were stored at -80°C. Animals were anesthesized with CO₂ to minimize the potential impact of anesthesia on hormone levels.

The experimental HFDs were almost isocaloric (Table 1), and composed of a fat-free basal diet based on American Institute of Nutrition Rodent Diets -Growth purified diet-(AIN93-G),²⁵ containing carbohydrate (corn starch, dextrinized corn starch, Dyets company, USA), protein (98.5% casein hydrolisate and 1.5% L-cyctine, choline bitartrate, Dyets companay, USA), fiber (cellulose, Dyets company, USA), vitamin-mineral mix, and tertbutylhydroquinone (Dyets company, USA) corn starch. Dietray fats were also added in order to provide a balanced diet and included fish oil and olive oil which were a generous gift from Nooshdarooye Darya Institute, Iran, soy oil (purchased from Ladan Institute, Iran) and butter (obtained locally). The diets were prepared weekly and stored as vacuum packed (500 g) at –20°C. Packs taken for use were thawed in the refrigerator at 4°C. The food was offered daily at the beginning of the dark phase, and the remains were weighed and removed after an 48 hours.

All experiments were carried out in accordance with standards approved by the local ethics committee of the Research Institute for Endocrine Science of Shahid Behesti University of Medical Sciences. The ethics code was 291EC1388.12.11.

Plasma glucose was determined by an enzymatic (Glucose Oxidase) colorimetric method (Pars Azmoun Co, Tehran, Iran). The assay sensitivity was 1 mg/dl, the intra- and interassay coefficients of variation were 1.2% and 1.8%, respectively. Plasma Insulin was determined by an ELISA method (Mercodia AB, Uppsala, Sweden). The assay sensitivity was 0.07 μg/L, the intra- and interassay coefficients of variations were 2.5% and 4.1%, respectively.

Blood samples were immediately transferred to chilled tubes containing Na2-EDTA (1 mg/ml) and aprotinin (500-unit/mL blood, Bayer) centrifuged in 2000 rpm for 15 min at 4°C. Hydrogen chloride was added to the samples at a final concentration of 0.1 N immediately after separation of the plasma. Unacylated form of ghrelin was measured using DAG ELISA kit according to the manufacturer's protocol (Mitsubishi Kagaku Iatron, Inc). The minimal detection limit of DAG in this assay system was 12.5 fmol/ml. The assay used to detect DAG has less than 0.1% of cross-reaction with acylated ghrelin. The intra- and interassay coefficients of variation were 3.7% and 8.1%, respectively.

Insulin resistance was estimated by homeostasis model assessment (HOMA), according to the formula: insulin resistance index = fasting insulin (μg/L) × fasting glucose (mg/dl)/405.²⁶

Gas chromatography analyses were carried out on gas chromatography (Younglin Instrument, 6000 series, South Korea) equipped with a split-injector and flame ionization detector. Methyl-esters of fatty acids (FAME) were prepared using methanolic KOH, according to the standard method (ISO5509:2000). The fatty acid profile was determined by gas chromatographic separation of their methyl esters (ISO 5508: 1990) on a capillary column (J&W Scientific DB-23, 30 m × 0.25mm × 0.25 μm). Chromatography software (Unicam 4880 chromatography data system) was used for data collection and processing. Data are presented in Table 2.

Statistical analyses were performed using SPSS 16.0 software (Chicago, IL, USA). The Kolmogorov–Smirnov test was applied to determine the normality of the distribution of the data to be used in the parametric test . One-way ANOVA, paired t test and Tukey test were used to compare the diet groups and p value<0.05 was considered statistically significant. Nonparametric tests (Kruskal-Wallis, Wilcoxon Signed-Rank and Mann-Whitney U) were also employed for insulin level and HOMA-IR variables and p value<0.05 was considered significant. The Spearman's rank correlation coefficient was also used to determine whether a significant relationship existed between insulin and DAG concentrations.

Food, calorie intakes and body weights of the groups are shown in Table 3. Food intake was significantly different among the groups fed with different dietary fats. High fat butter and HF-S groups had significantly higher food and calorie intake in comparison with SD, HF-F and HF-O groups (F=6.38, p=0.00). Initial body weights were not different among dietary groups but the weight gains in the HF-B and HF-S were 26.6 and 17.74 g higher, respectively, than the weight gain of SD group animals (p=0.01 and p=0.04, respectively). High fat butter had also 21.2 and 21.7 g higher weight gain compared with the HF-F and HF-O groups, respectively (p=0.00 and p=0.00, respectively). Weight gain of HF-F and HF-O groups were 12.3 and 12.8 g higher than the weight gain of HF-S group, respectively (p=0.04 and p=0.01, respectively). Final body weight in HF-S and HF-B groups was higher than that of DS, HF-F and HF-O groups (p=0.04).

The plasma DAG concentration was significantly decreased in SD, HF-F, HF-O and HF-S diets in the ad libitum fed status, compared with fasting condition (p=0.00, Figure 1A). Plasma DAG levels at fed state in HF-B and HF-S diet were higher than those of SD, HF-F and HF-O diets; however, these findings were not significant.

High fat fish, SD and HF-O diet rats showed 47.3, 42.7 and 38.6 ρg/ml higher DAG concentrations than the HF-B diet rats at fasted state, respectively (p=0.00); meanwhile, plasma DAG level in the HF-S diet rats was 16.4 ρg/ml lower than the HF-F diet rats (p=0.00).

Ad libitum fed plasma insulin concentration in the HF-B diet was significantly higher than that of the SD, HF-F and HF-O diets (p=0.00) and the increase in the HF-S diet group was 0.35, 0.37 and 0.31 μ g/l higher than that of the SD, HF-O and HF-F diets, respectively (p=0.01, Figure 1B). Furthermore, plasma insulin levels were decreased by fasting in all diet groups, although the decreases were not significant. Rats on the SD, HF-F and HF-O diets had significantly lower fasting insulin plasma levels than those on HF-B (p=0.04).

Plasma glucose levels were significantly decreased by fasting in all of diet groups, compared with fed state (p=0.00, Figure 1C). However, plasma glucose levels were not significantly different among dietary groups.

After 24 h fasting, HF-B diet had significantly higher HOMA-IR in comparison to the HF-F, HF-O and SD diets (p=0.05, Figure 1D).

There was no significant correlation between changes in insulin and changes in DAG levels in the feeding (Spearman correlation=0.16, p=0.3) and 24 hour fasting status (Spearman correlation=-0.34, p=0.06).