All experimental procedures involving animals were approved by the cantonal veterinary office in Switzerland and revised by the internal ethical committee of the Nestlé Research Center. Male Wistar rats (n = 30, 270 ± 5 g) were housed in independent cages (Euro standard Type III H: 425 × 266 × 185 mm, floor area 800 cm², Techniplast, Switzerland), received tap water and diets ad libitum for 21 days. The animals were randomly divided into five groups.

Fish Oil (FO) diet consisted of AIN 93M base diet supplemented with fish oil (Sofinol S.A., Manno, Switzerland), high oleic sunflower oil (HOSO, Sofinol S.A., Manno, Switzerland) and cocoa butter (Gerkens Cacao^®, Deventer, The Netherlands) as a lipid source. Total lipid content was 20 g/100 g diet (approximately 40% of the energy), see Table 1. Diets Fish oil + Orlistat (FO + O), free MAG + O, sn-2 MAG vanillin acetal + Orlistat (Vanil + O) and Acetylated sn-2 MAG + Orlistat (Acetyl + O) were AIN 93M base diet supplemented with FO, HOSO and either MAG Vanillin Acetal (Stepan Co., Northfield, IL, USA), Diacetylated MAG (Stepan Co., Northfield, IL, USA) or free MAG (Cognis GmbH, Illertissen, Germany) + Orlistat (MAG + O) containing 95.8% of α-MAG, as previously characterized [9] (Figure 1), plus 400 mg of Orlistat/kg of diet as listed in Table 1. The diets were prepared by mixing all the ingredients except Orlistat. The homogenized powders were dried at a low temperature and stored in a small sachet under vacuum at −20 °C. These special precautions were taken to avoid oxidative degradation of n-3 LC-PUFA. All animals were fed with a paste (water/powder diet, 1:1 w/w) prepared freshly every day. For the groups receiving Orlistat, the drug was added to the diet at 400 mg/kg every time the paste was prepared. FA composition of diets is described in Table 2.

Body weight was recorded twice a week for 21 days. Food intake was recorded five times a week. Body composition (fat and lean mass) was determined in animals in triplicate by quantitative nuclear magnetic resonance spectroscopy (Echo MRI-4in 1/500™; Echo Medical Systems, Houston, TX, USA), before the diet intervention (Day 0) and the day of the necropsy (Day 21). For determination of the fecal lipid excretion, animals were placed in metabolic cages for 48 h starting at day 17. Apparent lipid absorption was calculated using lipid intake and excretion data after feces were collected for 48 h and the food intake was monitored over this time period.

Blood was collected into heparinized tubes from the caudal vein after 6 h of food restriction at days 3, 7 and 14. The day of the necropsy (Day 21), animals were anesthetized with isoflurane. Blood was collected from the abdominal aorta, and then RBC and plasma were separated by centrifugation at 626× g for 2 min. Plasma and RBC were stored at −80 °C until lipid analyses were carried out. Liver, brain, retina and spleen were removed and transferred to different vials, flash frozen with liquid nitrogen and stored at −80 °C until further analyses.

Lipid extraction from organs (liver, spleen and brain) was described in details elsewhere [8]. Briefly, frozen tissues were pulverized, placed in chloroform, methanol, and water (2:1:0.8), homogenized and followed by a final Bligh and Dyer step to finish the lipid extraction. Total lipids were used to prepare fatty acid methyl esters (FAME) as described in the next section. A combination of methods was used to extract total lipids from feces [8]. Briefly, 1 g of dry frozen feces was macerated and lipids were extracted by the Bligh and Dyer method while heated at 60 °C to ensure complete extraction of total lipids. Lipids weight was obtained gravimetrically for fat absorption determination.

FAMEs in plasma (60 μL) were prepared by mixing sample with a methanolic solution of hydrochloric acid (2 mL, 1.5 N, Supelco, Bellefonte, Palo Alto, CA, USA), methanol (2 mL) and n-hexane (1 mL) in a test tube. Tubes were then heated at 100 °C for 60 min [10]. After cooling down to room temperature, water (2 mL) was added and tubes were centrifuged at 2000 rpm for 4 min. The organic phase was collected for GC analyses. Direct methylation of RBC (100 μL) was performed as described for plasma having RBC previously washed with PBS buffer. Tissue fat was methylated from the fat extracted, as described in the previous section.

Analysis of total FAMEs was performed on a 7890 Agilent gas chromatograph (Agilent Technologies, Palo Alto, CA, USA), equipped with a fused-silica BPX-70 capillary column (10 m × 0.1 mm I.D., 0.2 μm film thickness; SGE, Melbourne, Australia). Split injector (25:1) and flame ionization detection (FID) systems were operating at 250 °C. Oven temperature programming was 50 °C isothermal for 0.2 min, increased to 180 °C at 120 °C/min, isothermal for 1 min at this temperature then increased to 220 °C at 20 °C/min and then to 250 °C at 50 °C/min. The carrier gas (H₂) was maintained at a constant 1 mL/min and the acquisition of the FID signal at 100 Hz [13].

The main comparisons assessed are the comparisons between the positive control FO + O and the treatment groups Vanil + O, Acetyl + O, and MAG + O. In order to ensure that the addition of Orlistat lowered the absorption of fatty acids, the positive control was also compared to the negative control in this case FO.

Data is presented as means ± standard error of the mean (S.E.M.), except for FA relative percentage in RBC, plasma, liver, spleen, retina and brain where median ± robust SEM are provided. For food intake, fat intake, fecal lipid excretion and apparent lipid absorption (%), an ANOVA and two-sided appropriate contrasts were calculated. For body weight parameters, for days 7, 14 and 21 on body weight, mixed models were performed with body weight at day 3 as covariate, day as continuous variable and subject as random variable. One-sided tests were calculated. For fat mass and lean mass parameters, an ANOVA with parameters at D-3 as covariate was performed. Appropriate one sided contrasts were calculated. For FA in different organs, Kruskal Wallis and exact Wilcoxon tests were performed. One-sided tests were calculated. Statistical calculations were performed using SAS software, version 9.1. Differences were considered as significant at P < 0.05.

Weight parameters, dietary intake, lipid absorption and calculated intake of EPA and DHA results for rats fed sample diets are shown in Table 2. Body weight increased for all groups with no significant differences observed. When groups were compared with changes obtained in group FO + O, no significant differences were observed except for body weight of the Acetyl + O group which was higher through time although not significantly different at day 20 (P = 0.0163, Table 2).

Significantly higher lipid and daily food intake were observed in the group receiving Vanil + O, compared to the FO + O group. All the groups fed with Orlistat had significantly higher lipid excretion levels and reduced apparent lipid absorption relative to the FO receiving no Orlistat. Comparison of the data obtained in rats fed FO or FO + O diets demonstrates the efficiency of Orlistat to lower lipid absorption (−44.8% intake reduction), and therefore the consistency of the experimental approach. None of the MAG diets modified the action of Orlistat. Lipid absorption was reduced in all groups, compared to the control group receiving fish oil in absence of Orlistat. No significant differences were found for EPA intake between groups, while higher levels of DHA were found for the Acetyl + O and Vanil + O groups.

At baseline, the EPA level in RBC was 0.2% of total FA and no differences were found among groups. The level of EPA in the groups fed Vanil + O, Acetyl + O and MAG + O were significantly higher at Day 21 when compared with groups fed control diet (FO + O), as shown in Table 3. Incorporation of EPA in RBC at different time points is shown in Figure 2 for all treatments. For rats fed FO + O, EPA level in RBC was significantly lower when compared with levels of EPA of rats fed FO diet with no Orlistat. These data are consistent with results obtained on lipid absorption and validate the fact that Orlistat, taken in conjunction with a dietary source of EPA (fish oil), lowers the incorporation of EPA in RBC.

MAG + O diet showed the highest incorporation of EPA in RBC throughout the study, when compared to FO + O, followed by the Vanil + O and Acetyl + O groups as seen in Figure 2. These results show that MAG Vanillin Acetal, Diacetylated MAG and MAG are suitable carriers to efficiently deliver EPA to RBC compared with fish oil, even if Orlistat is present in the diets (Table 3, Figure 2). When Orlistat was not directly present in fish oil diets, evaluation of the level of EPA in groups fed with the different MAG diets, indicates that free MAG and MAG derivatives are efficient carriers of EPA when Orlistat is given orally. Indeed, the level of EPA was significantly higher in RBC of animals fed with MAG + O diet through time, than in the FO group and similar differences were observed for animals fed with the Vanil + O diet.

In plasma, the EPA level at baseline was not significantly different among groups (Table 3). Figure 3 shows differences between animals fed FO diet and those fed with the control (FO + O) diet. Supplementation of fish oil diet with Orlistat tends to reduce EPA, especially at day 7, although this change was not significantly lower. For animals fed with MAG + O, Vanil + O and the Acetyl + O diets, the level of EPA was significantly higher, compared to the control diet. These results clearly corroborate the data obtained on RBC and show the potential of MAG and MAG derivatives to deliver EPA when lipid digestive function is impaired.

The level of EPA in plasma lipids found in animals fed with MAG and MAG derivatives + Orlistat diets compared to animals fed the FO + O diet demonstrate as previously observed in RBC that free MAG and MAG derivatives are superior to FO to supply LC-PUFAs in this condition. Overall, at the end of the experiment (Day 21) levels of incorporation of EPA in plasma were higher in all groups when compared with the control group. However, at the last sampling point (Day 21) a trend towards decreased EPA is observed, potentially a consequence of the faster turnover that the plasma has in comparison with the RBC.

As shown in Table 3, EPA levels in liver, spleen, retina and brain were evaluated. The EPA level of incorporation in liver, the central organ for lipid metabolism, was significantly lower for FO and FO + O groups. For the Vanil + O, Acetyl + O and MAG + O diets, levels were significantly higher when compared with control diet, with the highest level of EPA incorporation observed for MAG + O. A similar pattern of FAs was found for lipids of spleen with lower levels of EPA accumulation than for liver. In retina and brain, dietary composition is reflected by the incorporation of EPA. When compared with FO + O treatment, retina showed significantly higher levels for Vanil + O, Acetyl + O and MAG + O groups, whereas brain showed significant increments with Vanil + O, and MAG + O but not with Acetyl + O.

The diets were made such that the EPA content was similar in all groups (approximately 4.4 g/kg of diet) with DHA levels in the range of 2.9–7.4 g/kg diet. However, results on DHA show it as one of the major LC-PUFA identified. Incorporation of DHA in RBC through time is shown in Table 4 for all treatments. The Vanil + O treatment produced a significantly higher DHA level in RBC when compared with the control diet. The level of DHA in plasma lipids was significantly higher in animals fed the FO diet compared to those fed the FO + O diet. Similar results were observed for levels of EPA. DHA level was significantly higher in groups that received MAG or MAG derivatives in diets, compared to animals fed the control diet.

In liver, levels of DHA were significantly higher for Vanil + O, Acetyl + O and MAG + O treatments when compared with the control treatment. The level of incorporation of DHA in spleen, was significantly higher for the Vanil + O and the MAG + O treatment, although a lower level, when compared with the FO + O treatment. DHA level of incorporation with Acetyl + O treatment was also higher, although not significant. In retina, similar levels of DHA were observed (Table 4).