DHA attenuates postprandial hyperlipidemia via activating PPARα in intestinal epithelial cells.

It is known that peroxisome proliferator-activated receptor (PPAR)α, whose activation reduces hyperlipidemia, is highly expressed in intestinal epithelial cells. Docosahexaenoic acid (DHA) could improve postprandial hyperlipidemia, however, its relationship with intestinal PPARα activation is not revealed. In this study, we investigated whether DHA can affect postprandial hyperlipidemia by activating intestinal PPARα using Caco-2 cells and C57BL/6 mice. The genes involved in fatty acid (FA) oxidation and oxygen consumption rate were increased, and the secretion of triacylglyceride (TG) and apolipoprotein B (apoB) was decreased in DHA-treated Caco-2 cells. Additionally, intestinal FA oxidation was induced, and TG and apoB secretion from intestinal epithelial cells was reduced, resulting in the attenuation of plasma TG and apoB levels after oral administration of olive oil in DHA-rich oil-fed mice compared with controls. However, no increase in genes involved in FA oxidation was observed in the liver. Furthermore, the effects of DHA on intestinal lipid secretion and postprandial hyperlipidemia were abolished in PPARα knockout mice. In conclusion, the present work suggests that DHA can inhibit the secretion of TG from intestinal epithelial cells via PPARα activation, which attenuates postprandial hyperlipidemia.

system, a reporter plasmid (p4xPPRE-tk-luc) and pRL-CMV were transfected into Caco-2 cells. Transfection was performed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol. Four hours after transfection, transfected cells were cultured in medium containing DHA for an additional 24 h. Luciferase assays were performed using the dual luciferase system according to the manufacturer's protocol.

Real-time quantitative RT-PCR
Total RNA samples were prepared from Caco-2 cells, mouse intestinal epithelial cells, and hepatocytes using Sepasol Super-I (Nacalai Tesque) and Qiazol lysis reagent (Qiagen, Hilden, Germany) according to the manufacturer's instructions, respectively. Using M-MLV reverse transcriptase (Invitrogen), total RNA was reverse-transcribed following the manufacturer's protocol using a thermal cycler (Takara, Shiga, Japan). To quantify mRNA expression, real-time PCR was performed using a LightCycler system (Roche Diagnostics, Mannheim, Germany) using SYBR Green fl uorescence signals as described previously ( 22 ). Oligonucleotide primers of human and mouse 36B4 and PPAR ␣ target genes used in this study were designed using a PCR primer selection program found in the website of the Virtual Genomic Center from the Gen-Bank database, as previously described ( Table 1 ) ( 23 ). To compare mRNA expression levels among samples, copy numbers of all transcripts were divided by that of human and mouse 36B4, showing a constant expression level. All mRNA expression levels are represented relative to the control in each experiment.

Measurement of oxygen consumption rate in Caco-2 cells
The cellular oxygen consumption rate (OCR) was measured using a Seahorse Bioscience XF analyzer in 24-well plates at 37°C, with correction for positional temperature variations adjusted for the four empty wells in the plate ( 24,25 ). Caco-2 cells were cultured for 2 weeks after seeding on the plate and were treated with PPAR ␣ agonist, 50 M bezafi brate, or either 1 M or 25 M DHA. Immediately before the measurement, cells were washed, and 675 l of nonbuffered (sodium-carbonate-free, pH 7.4) DMEM medium supplemented with 0.2 mM palmitic acid, 0.2 mM eicosapentaenoic acid (EPA), are known to lower plasma TG; the mechanism responsible for their hypolipidemic action is thought to be involved in the regulation of TG clearance from circulation and TG synthesis in the liver (17)(18)(19). Recent studies have found that PUFAs increase the mRNA expression levels of genes involved in FA oxidation in intestinal epithelial cells ( 20 ). However, it is unknown whether dietary lipids, such as DHA could affect the intestinal lipid metabolism, resulting in improvement of postprandial hyperlipidemia.
In this study, we investigated whether DHA improves postprandial hyperlipidemia by altering the lipid metabolism in intestinal epithelial cells. DHA induced FA oxidation in intestinal epithelial cells by activating PPAR ␣ , which attenuated postprandial hyperlipidemia by directly reducing TG secretion from intestinal epithelial cells. Furthermore, we confi rmed that hepatic lipid metabolism is unlikely to contribute to these effects of DHA. These fi ndings suggest that activating intestinal PPAR ␣ by dietary lipids such as DHA may shed light on postprandial hyperlipidemiainduced cardiovascular diseases.

Chemicals and cell culture
DHA and EPA were purchased from Nacalai Tesque (Kyoto, Japan) and dissolved in ethanol. Bezafi brate was purchased from Sigma (St. Louis, MO) and dissolved in dimethylsulfoxide (DMSO) as a stock solution. Decanoic acid and palmitic acid were purchased from Nacalai Tesque and Wako Pure Chemicals (Osaka, Japan), respectively. All other chemicals used were from Sigma or Nacalai Tesque and were guaranteed to be reagent or tissue-culture grade.
Human Caco-2 cells were purchased from American Type Culture Collection (ATCC, Manassas, VA) and were cultured in DMEM (100 mg/dl glucose) containing 10% fetal bovine serum, 1% nonessential amino acid solution, and 10 mg/ml penicillin/streptomycin at 37°C in 5% CO 2 /95% air under humidifi ed conditions. Caco-2 cells were seeded at a density of 1.2 × 10 6 cells/ml on 12-well Transwell plates (Corning Inc., Corning, NY) for 2 weeks for differentiation into intestinal epithelial-like cells. To evaluate differentiation of Caco-2 cells, we measured transepithelial electrical resistance. No signifi cance in transepithelial electrical resistance was detected in any experiment (data not shown). The apical medium was changed to DMEM containing either 1 M or 25 M DHA or 50 M bezafibrate and 600 M taurocholic acid Na salt hydrate and 500 M oleic acid. Additionally, the basolateral medium was changed to serumfree DMEM. After 48 h, the basolateral medium was collected to measure TG and apolipoprotein B (apoB) secretion. Cell viability was measured in Caco-2 cells treated with DHA and bezafi brate based on cell titers (Promega, Fitchburg, WI).

Animal experiments
DHA-rich oil containing 25.4% DHA and 8% EPA was a gift from NOF Corporation (Kanagawa, Japan). EPA-rich oil containing 28.4% EPA and 12.3% DHA was a gift from Nippon Suisan Kaisha, Ltd. (Tokyo, Japan). All other chemicals were from Sigma or Nacalai Tesque and were guaranteed to be reagent or tissueculture grade.
All mice were maintained separately in a temperature-controlled (23°C) facility under a constant 12 h light/dark cycle with free access to water. To analyze the effects of DHA on intestinal lipid metabolism and postprandial hyperlipidemia, 9-week-old male C57BL/6 mice (CLEA Japan, Tokyo, Japan) were fed a high-fat diet (HFD) consisting of 60% (kcal%) fat from dietary oil, 26% protein, and 14% carbohydrate for 1 week to induce postprandial hyperlipidemia ( 26 ), and were then divided into three groups with the same average serum TG level and body weight after 16 h fasting. Ten-week-old male C57BL/6 mice were maintained for 1 week either on a 60% HFD or on a diet containing 1.9% DHA or 3.7% L -carnitine, and 2% FA-free BSA was added to each well. After equilibration for 30 min, 2 min measurements were performed at 3 min intervals with inter-measurement mixing to homogenize the oxygen in the medium.

Measurement of TG and apoB secretion in Caco-2 cells
To measure TG secretion, we used the Triglyceride E Test Wako kit (Wako Pure Chemicals). To measure apoB secretion, an enzyme-linked immunosorbent assay (ELISA) was performed using an anti-human low density lipoprotein apoB antibody a Percent of weight from total food weight. b Percent of energy from total energy intake.

Postprandial TG and apoB secretion
To measure plasma TG concentration, mice were administrated an oral gavage of 300 l olive oil after a 16 h fast, and blood samples were collected every 30 min to 240 min after olive oil administration from the tail vein of nonanesthetized mice.
To measure TG secretion from intestinal epithelial cells, mice were injected with 500 mg/kg body weight tyloxapol (T0307, Sigma) into the intraperitoneal cavity to block serum lipase activity after a 16 h fast ( 28 ). After 30 min, mice were administrated an oral gavage of 300 l olive oil. Blood samples were obtained before tyloxapol injection and every 30 min for 240 min after olive oil administration. Plasma TG concentration was determined using the Triglyceride E-Test Wako kit (Wako Pure Chemicals).
To measure postprandial apoB48 secretion, plasma collected at 120 min was mixed with Laemmli sample buffer (Bio-Rad) (1:8) and boiled for 5 min at 95°C. Plasma samples were subjected to SDS-PAGE on a 5% gel. Separated proteins were transferred electrophoretically to polyvinylidene fl uoride membranes (Millipore DHA, maintaining the total amount of fat at 60%. The detailed composition of the experimental diets is described in Table 2 ( 27 ). As shown in supplementary Fig. III, EPA-rich oil containing 28.4% EPA and 12.3% DHA was diluted with corn oil to prepare a HFD with fi nal concentrations of 3.4% EPA and 1.5% DHA, maintaining the total amount of fat at 60%. The energy intake of all mice was adjusted by pair feeding, and food intake was determined daily for seven consecutive days. Anesthesia was induced using sevofl urane in all experiments. The procedures for animal care were approved by the Animal Research Committee of Kyoto University.
To clarify whether the effects of DHA-rich oil on intestinal lipid metabolism and postprandial hyperlipidemia involves PPAR ␣ , we used PPAR ␣ Ϫ / Ϫ mice with a C57BL/6 genetic background. PPAR ␣ Ϫ / Ϫ mice were fed a HFD consisting of 60% (kcal%) fat for 1 week, and were then divided into two groups with the same average serum TG level and body weight after 16 h fasting. Ten-week-old male PPAR ␣ Ϫ / Ϫ mice were maintained for 1 week either on a 60% HFD or on a 60% HFD containing 3.7% DHA or 0.2% bezafi brate. For RNA analysis, the proximal intestine and the liver were harvested from the mice. After washing, intestinal epithelial cells were collected using a slide glass. Collected tissue was stored in RNAlater (Ambion, Austin TX; Applied Biosystems, Foster City, CA) at Ϫ 80°C until use.

Measurement of FA oxidation
FA oxidation with isolated intestinal epithelial cells and hepatocytes was analyzed as previously described ( 15,25 ). Briefl y, collected intestinal epithelial cells and hepatocytes were washed with 1% FBS/DMEM three times and used for experiments. Cells were incubated with a piece of fi lter paper containing 200 l 3 N NaOH in DMEM containing 200 M palmitic acid, 0.1% FA-free BSA, The mRNA expression levels of FA oxidation-related genes ( Acs , Cpt1a , and Aox ) and other PPAR ␣ target genes ( Ucp2 and Fabp ) were quantifi ed. F: The OCR was determined using extracellular fl ux analysis as described in the Materials and Methods section. Values of controls were set at 100% and the relative values were represented as fold induction relative to that of control. Values are means ± SEM of six tests. Cont, control; Beza, bezafi brate. * P < 0.05 and ** P < 0.01 compared with each control. levels of genes involved in FA oxidation in DHA-treated Caco-2 cells. DHA treatment induced mRNA expression of genes involved in FA oxidation, such as acyl-CoA synthetase ( Acs ), carnitine palmitoyltransferase 1A ( Cpt1a ), and acyl-CoA oxidase ( Aox ) and other PPAR ␣ target genes such as uncoupling protein-2 ( Ucp2 ) and fatty acid binding protein ( Fabp ) ( Fig. 2A-E ). Moreover, the OCR, determined using extracellular fl ux analysis, was enhanced following DHA treatment as shown in Fig. 2F . In contrast, decanoic acid (C10), which had little activity toward PPAR ␣ , did not affect mRNA expression of Cpt1a , and palmitic acid (C16), which showed lower PPAR ␣ activity than DHA, did not significantly induce Cpt1a expression in Caco-2 cells (supplementary Fig. IIA, B). These fi ndings suggest that DHA enhances FA oxidation in Caco-2 cells.

DHA decreased the secretion of TG and apoB from Caco-2 cells
To determine the effects of PPAR ␣ activation by DHA on lipid secretion from Caco-2 cells, we examined the amounts of lipid secreted from DHA-treated Caco-2 cells. TG secretion from DHA-treated Caco-2 cells was signifi cantly decreased (to 77 and 72% with either 1 or 25 M DHA treatment, respectively), as shown in Fig. 3A . DHA treatment reduced Corporation, Billerica, MA), which were blocked with 5% nonfat dried milk in phosphate-buffered saline. The membranes were incubated with the anti-mouse apoB48/100 antibodies (Meridian Life Science, Memphis, TN), and then with peroxidase-conjugated anti-rabbit IgG antibodies (Santa Cruz Biotechnologies, Santa Cruz, CA), respectively. Protein bands were detected using an enhanced chemiluminescence (ECL) Western blotting detection system (Millipore Corporation). The bands were quantitatively evaluated using National Institutes of Health Image J software.

Measurement of intestinal TG in mice
Lipids in intestinal mucosa were extracted using the hexane/ isopropanol (3:2) extraction methods ( 29 ). Briefl y, intestinal mucosa were homogenized using hexane/isopropanol (3:2) for 1 min, the suspension was centrifuged, and the pellet was rinsed with the same solvent. The entire liquid phase was evaporated, the dried extract dissolved in isopropanol, and TG content was measured as above. Triolein dissolved in isopropanol was used as the standard for TG. The effi ciency of extraction was measured by comparing the recovery of triolein in samples that had been spiked and samples that had not been spiked with known quantities of triolein standard ( 30 ). The assessed recovery was 81.2 ± 4.65%.

Measurement of TG in feces of mice
The feces were dried at 60°C overnight and the lipids were extracted using the Folch method ( 31 ). This analysis enables measurement of lipids extracted per gram of dried fecal samples. Briefl y, lipids present in the feces were extracted using chloroform/methanol (2:1), dissolved in isopropanol, and TG content was measured as above.

Statistical analysis
Data are presented as means ± SEM. For analyses of two groups, unpaired Student's t -test was used. To analyze three or more groups, ANOVA was used along with Tukey-Kramer's multiple comparison tests to determine statistical signifi cance. Differences were considered signifi cant at P < 0.05.

DHA activated PPAR ␣ in CV-1 cells and Caco-2 cells
First, we investigated whether DHA activated PPAR ␣ based on a luciferase assay using the GAL4/PPAR ␣ chimera system. DHA activated luciferase activity of PPAR ␣ in CV-1 cells in a dose-dependent manner ( Fig. 1A ). Furthermore, DHA stimulated PPAR-response element (PPRE)-luciferase activity in Caco-2 cells ( Fig. 1B ). DHA also activated luciferase activity of PPAR ␣ in Caco-2 cells ( Fig. 1C ). The effects of DHA on PPAR ␣ activation were higher than those of EPA under our experimental conditions (approximately 5.9-and 2.6-fold increases at 25 M DHA and EPA, respectively), as shown in Fig. 1C . Moreover, DHA enhanced the activation of PPAR ␥ by approximately 1.7-fold ( Fig. 1D ). However, DHA did not increase PPAR ␦ activation in Caco-2 cells ( Fig. 1E , supplementary  Fig. I). Cytotoxicity was not observed following 25 M DHA treatment of Caco-2 cells (data not shown). These results suggest that DHA induces PPAR ␣ activation in intact cells.

DHA induced the genes involved in FA oxidation and OCR in Caco-2 cells
To investigate the effects of PPAR ␣ activation by DHA on intestinal lipid metabolism, we measured mRNA expression Fig. 3. DHA decreased TG and apoB secretion from Caco-2 cells. TG (A) and apoB (B) secretion from Caco-2 cells into the basolateral side were measured as described in the Materials and Methods section. Control (Cont) values were set at 100% and the relative values were represented as fold induction relative to that of control. Values are means ± SEM of six tests. Beza, bezafi brate. * P < 0.05 and ** P < 0.01 compared with each control. difference in food intake between groups. The mRNA expression levels of FA oxidation-related genes such as Acs , Cpt1a , and Aox and other PPAR ␣ target genes such as Ucp2 , Fabp , and Cd36 were increased in C57BL/6 mice fed a HFD containing DHA-rich oil for one week ( Fig. 4A-F ). When the cells were incubated with [ 14 C]palmitic acid for 2 h, oxidation of [ 14 C]palmitic acid to CO 2 and ASMs were enhanced in intestinal epithelial cells of DHA-rich oil-fed mice compared with control mice ( Fig. 4G, H ). However, surprisingly, DHA-rich oil-fed mice showed no increase in mRNA expression levels of FA oxidation-related genes in the liver under the same conditions as shown in Fig. 5A-C . Moreover, the production of CO 2 and ASM were not augmented in the liver of DHA-rich oil-fed mice compared with control mice ( Fig. 5D, E ). These fi ndings suggest that DHA-rich oil enhances FA oxidation in intestinal epithelial cells of mice. the secretion of apoB, which is the primary apolipoprotein of chylomicrons, to 67 and 59% with either 1 or 25 M DHA treatment, respectively ( Fig. 3B ). The effects of DHA on secretion were similar to those of bezafi brate, a potent PPAR ␣ agonist ( Fig. 3A, B ). While C10 did not inhibit TG secretion, C16 did decrease TG secretion from Caco-2 cells. However, the effect of C16 on decrease of TG secretion was lower than that of DHA (supplementary Fig. IIC). These results suggest that lipid secretion from intestinal epithelial cells is related to PPAR ␣ activity.

DHA-rich oil enhanced FA oxidation in intestinal epithelial cells of C57BL/6 mice
Next, we examined whether the effects of DHA in vitro also occurred in vivo. Because PPAR ␣ agonists are known to reduce food intake in rodents ( 32 ), all mice were housed in pair-fed conditions in each experiment; there was no

Effects of DHA on postprandial lipid metabolism were mediated by the activation of intestinal PPAR ␣
To clarify the involvement of PPAR ␣ in the effects of DHA on postprandial lipid metabolism, we examined the effects of DHA in PPAR ␣ Ϫ / Ϫ mice. The baseline characteristics of PPAR ␣ Ϫ / Ϫ mice compared with control mice are shown in supplementary Table I. The mRNA expression levels of genes involved in FA oxidation ( Acs , Cpt1a , and Aox ), and the production of CO 2 and ASMs did not change signifi cantly in intestinal epithelial cells of DHA-rich oilfed PPAR ␣ Ϫ / Ϫ mice or bezafi brate-fed PPAR ␣ Ϫ / Ϫ mice ( Fig. 7A-E , supplementary Fig. IVA-C). Moreover, there was no difference in intestinal TG levels between DHArich oil-fed PPAR ␣ Ϫ / Ϫ mice and control mice ( Fig. 7F ). Finally, the effects of DHA-rich oil on plasma TG and apoB levels after olive oil administration were abolished in PPAR ␣ Ϫ / Ϫ mice without and with tyloxapol, similar to the results of bezafi brate ( Fig. 7G-I , supplementary Fig. IVD-F), suggesting that lipid secretion from intestinal epithelial cells is related to PPAR ␣ activity. These fi ndings suggest that the activation of intestinal PPAR ␣ is a key factor for attenuating postprandial hyperlipidemia by decreasing TG secretion from intestinal epithelial cells.

DISCUSSION
Activation of PPAR ␣ is well-known to decrease plasma TG levels through FA oxidation in the liver and skeletal muscle (9)(10)(11). Although the role of PPAR ␣ expressed in intestinal epithelial cells remained obscure ( 8,33 ), we and

DHA-rich oil attenuated postprandial TG levels by reducing TG secretion from intestinal epithelial cells in mice
To investigate whether DHA-rich oil decreases postprandial TG levels in mice, we measured plasma TG levels every 30 min to 240 min after oral administration of olive oil. Plasma TG levels were signifi cantly lower in DHArich oil-fed mice than those in control mice from 120 to 240 min after administration ( Fig. 6A ). It was also confi rmed that postprandial triglyceridemic response, determined based on the area under the curve (AUC), was lowered to 66 and 45% in 1.9 and 3.7% DHA-rich oil-fed mice, respectively ( Fig. 6B ). In addition, plasma apoB48 was also reduced at 120 min in 3.7% DHA-rich oil-fed mice ( Fig. 6C ). To clarify whether DHA-rich oil altered postprandial TG secretion from intestinal epithelial cells, we measured plasma TG levels after oral administration of olive oil in the presence of tyloxapol, an inhibitor of TG clearance. Plasma TG levels were signifi cantly decreased after 150 min and AUC was reduced to 66% in 3.7% DHArich oil-fed mice compared with control mice ( Fig. 6D, E ). Furthermore, TG accumulation in intestinal epithelial cells was lower in DHA-rich oil-fed mice than in control mice and there was no difference in the weight of feces and fecal TG levels between control and DHA-rich oil-fed mice ( Fig. 6F-H ). In contrast, EPA did not affect postprandial lipid metabolism compared with DHA (supplementary Fig. III). These results suggest that DHA attenuates postprandial hypertriglyceridemia by decreasing TG secretion from intestinal epithelial cells. Although it has been shown that PPAR ␣ activation in intestinal epithelial cells reduces postprandial hyperlipidemia, it was unknown whether postprandial hyperlipidemia is also improved by dietary lipids, which generally show lower PPAR ␣ activation than synthesized PPAR ␣ agonists ( 34,35 ). Previous studies have indicated that DHA increases mRNA expression levels of FA oxidation-related genes in intestinal epithelial cells ( 36,37 ) and that PUFAs including DHA enhance FA oxidation in hepatocytes ( 38 ). The present others have recently demonstrated that PPAR ␣ agonists improve postprandial hyperlipidemia through increasing FA oxidation in intestinal epithelial cells ( 15,16 ). It is suggested that PPAR ␣ activation reduces TG secretion from intestinal epithelial cells, which attenuates postprandial hyperlipidemia ( 15,16 ) (supplementary Fig. IV). To clarify the contribution of intestinal PPAR ␣ activation to postprandial systemic lipid metabolism, further investigation is necessary, including studies involving intestinal epithelial cell-specifi c PPAR ␣ knockout mice. However, these fi ndings indicate that intestinal PPAR ␣ activation plays a critical Fig. 6. Postprandial hyperlipidemia was attenuated by decreasing TG secretion from intestinal epithelial cells of DHA-rich oil-fed C57BL/6 mice. A, B: Plasma TG levels every 30 to 240 min and plasma apoB48 levels at 120 min after oral administration of olive oil were measured in control (Cont) and DHA-rich oil-fed C57BL/6 mice. Plasma TG level was measured by enzymatic colorimetric assay. C: Plasma apoB48 protein levels at 120 min were visualized by Western blotting and band density was determined using National Institutes of Health Image J software . D, E: Postprandial TG secretion in control and DHA-rich oil-fed mice that had been administered tyloxapol, an inhibitor of TG clearance, was examined. AUC is shown as relative values and is represented as fold induction relative to that of the control, which was set at 100%. TG content in intestinal epithelial cells (F), the weight of feces (G), and fecal TG levels (H) in control and DHA-rich oil-fed C57BL/6 mice were determined as described in the Materials and Methods section. The values are means ± SEM of 5-10 tests. * P < 0.05 and ** P < 0.01 compared with each control. rich oil-fed mice, plasma TG levels were decreased after olive oil administration with tyloxapol, which inhibits plasma lipoprotein lipase, suggesting that TG secretion from intestinal epithelial cells was reduced ( Fig. 6D ). This was supported by the results that DHA reduced TG and apoB secretion in Caco-2 cells, as shown in Fig. 3 . Moreover, we observed that TG accumulation in intestinal epithelial cells was generally decreased ( Fig. 6F ) and the level in the weight of feces and fecal TG did not change in DHA-rich oil-fed mice ( Fig. 6G, H ). These fi ndings suggest that DHA is a potent factor to reduce TG secretion from intestinal epithelial cells via FA oxidation by PPAR ␣ activation, resulting in attenuating postprandial hyperlipidemia.
In this study, mRNA expression levels of intestinal FA oxidation-related genes in DHA-rich oil-fed PPAR ␣ Ϫ / Ϫ mice study showed that DHA enhanced FA oxidation and decreased TG secretion in Caco-2 cells and intestinal epithelial cells ( Figs. 2-4, 6 ), resulting in reduction of postprandial hyperlipidemia via PPAR ␣ activation in mice ( Figs. 6, 7 ). However, surprisingly, no induction of the genes involved in FA oxidation was observed in the liver of DHA rich oil-fed mice under our experimental conditions ( Fig. 5 ). Our fi ndings presented here strongly indicate that effects of DHA in attenuating postprandial hyperlipidemia are attributed to the decrease of TG secretion from intestinal epithelial cells. During early stages after a meal, most TG secretion into circulation is thought to be derived from dietary fat absorbed in intestinal epithelial cells because they are directly exposed to dietary fat, while insulin prevents hepatic VLDL secretion during the postprandial state ( 39,40 ). In DHA were increased, although the increases were not significant ( Fig. 7A, C ). Previous reports have indicated that PPAR ␦ compensates for the lack of PPAR ␣ in the skeletal muscles of PPAR ␣ Ϫ / Ϫ mice ( 41 ) and that PPAR ␦ activates FA oxidation ( 42 ). DHA and bezafi brate did not activate PPAR ␦ in our luciferase assays ( Fig. 1E , supplementary  Fig. I). However, the concentration of DHA exposed to intestinal epithelial cells may have been much higher than that used in Caco-2 cells. Therefore, the increase in intestinal FA oxidation-related genes in Fig. 7A and C may be related to the PPAR ␦ effect. The present study showed higher mRNA expression levels of Cd36 ( Fig. 4 ), which is involved in FA transport in intestinal epithelial cells of DHA-rich oil-fed mice. Cd36 is thought to be involved in regulating chylomicron production ( 43,44 ). Interestingly, Cd36 knockout mice showed both fasting and postprandial hyperlipidemia and have been used as a model of postprandial hyperlipidemia ( 45 ). A recent study showed that Cd36 critically regulates FA oxidation in skeletal muscle ( 46 ). Additionally, Cd36 is one of PPAR ␣ target genes ( 47 ). Therefore, an increase of Cd36 may contribute to reduction of postprandial hyperlipidemia via intestinal FA oxidation in DHA-rich oil-fed mice.
In conclusion, we found that DHA directly reduced TG secretion from intestinal epithelial cells by activation of PPAR ␣ -induced FA oxidation, resulting in improving postprandial hyperlipidemia. The present work suggests that a dietary lipid such as DHA, which activates PPAR ␣ , is a promising factor to attenuate postprandial hyperlipidemia via intestinal FA oxidation.