The ddY mouse: a model of postprandial hypertriglyceridemia in response to dietary fat.

Postprandial hyperlipidemia (lipemia) is a risk factor for atherosclerosis. However, mouse models of postprandial hyperlipidemia have not been reported. Here, we report that ddY mice display marked postprandial hypertriglyceridemia in response to dietary fat. In ddY mice, the fasting serum total triacylglyceride (TG) concentration was 134 mg/dl, which increased to 571 mg/dl after an intragastric safflower oil load (0.4 ml/mouse). In C57BL/6J mice, these concentrations were 57 and 106 mg/dl, respectively. By lipoprotein analysis, ddY mice showed increases in chylomicron- and VLDL-sized TG fractions (remnants and VLDL) after fat load. In C57BL/6J mice, post-heparin plasma LPL activity after fat load was increased 4.8-fold relative to fasting. However, in ddY mice, the increase of LPL activity after fat load was very small (1.2-fold) and not significant. High fat feeding for 10 weeks led to obesity in ddY mice. A difference in LPL amino acid composition between C57BL/6J and ddY mice was detected but was deemed unlikely to cause hypertriglyceridemia because hypertriglyceridemia was not evident in other strains harboring the ddY-type LPL sequence. These findings indicate that postprandial hypertriglyceridemia in ddY mice is induced by decreased LPL activity after fat load and is associated with obesity induced by a high-fat diet.

oral gavage (0.4 ml/mouse) as described previously ( 23 ), and the other mice were kept under fasting. The fasted mice and saffl ower oil-loaded mice were euthanized 3 h later, and body and tissue weights were measured (n = 8 per group). In an experiment to examine the effects of saffl ower oil loading on the TG increase after fat load in various strains, the same experiments described above were conducted in C57BL/6J, ddY, C57BL/6N, DBA/1JN, DBA/2N, BALB/c, C3H/He, and CBA/JN mice (n = 6 per group). In a time-course study, after 24 h fasting, mice were loaded with 0.4 ml of saffl ower oil by gavage (n = 8 per group). Serum was drawn from the tail vein at 0, 1, 3, 6, 9, and 12 h (for TG, total cholesterol [TC], non-esterifi ed fatty acid [NEFA], insulin, and glucose-dependent insulinotropic polypeptide [GIP]) and at 0.5, 1, 1.5, 2, 3, and 4 h (for glucose) after administration of saffl ower oil.

Serum chemistry
Serum TG, TC, and NEFA were assayed enzymatically with TG E, T-Cho E, and NEFA C test colorimetric kits (Wako Pure Chemicals Industries, Osaka, Japan), respectively. Serum insulin was assayed with a mouse insulin ELISA kit (Morinaga, Kanagawa, Japan). Serum glucose was measured on an Ascensia autoanalyzer (Bayer Medical Ltd., Tokyo, Japan). Serum GIP was assayed with a Rat/Mouse GIP ELISA kit (Millipore, Billerica, MA). Serum adiponectin was assayed with a Mouse/Rat adiponectin ELISA kit (Otsuka Pharmaceutical Co., Ltd., Tokyo, Japan).

Plasma lipoprotein analysis
Plasma lipoproteins were analyzed by an on-line dual enzymatic method for quantifi cation of TG and TC by HPLC at Skylight Biotech (Akita, Japan) according to the procedure described by Usui et al. ( 24 ). The amount of TG and TC in separated lipoproteins by two tandem TSK columns were estimated by hydrogen peroxide production by glycerol-3-phosphate oxidase and cholesterol oxidase, respectively ( 24 ).

Estimation of lipoprotein production by Triton WR 1339 injection
Lipoprotein production was estimated by measuring TG concentration after intravenous injection of Triton WR 1339, which inhibits LPL activity ( 25 ). To estimate lipoprotein production under fasting, C57BL/6J and ddY mice were given Triton WR 1339 (Sigma-Aldrich, St. Louis, MO) at 500 mg/kg body weight as a 15% solution in saline after a 24 h fast. Blood was obtained before injection and at 1, 2, 3, and 4 h after injection (n = 8 per strain). To estimate lipoprotein production after saffl ower oil loading, C57BL/6J and ddY mice were intravenously injected with Triton WR 1339 (500 mg/kg body weight as a 15% solution in saline) after a 24 h fast, and 30 min later, saffl ower oil (0.4 ml/ mouse) was administrated (n = 8 mice per strain). Mice were bled before Triton WR 1339 injection and at 1, 2, 3, and 4 h after saffl ower oil administration. Plasma TG concentration over the 4 h period was measured by enzymatic colorimetric methods using the TG E test as described above.

Liver lipid analysis
Lipids in the liver were extracted with an ice-cold mixture of chloroform and methanol (2:1, v/v) according to the method of Folch et al. ( 26 ). Total TG concentrations in liver homogenates were measured by enzymatic colorimetric methods using the TG E test.

Measurement of LPL activity in plasma and tissues
For measurement of LPL activity in plasma, 8 week-old C57BL/6J and ddY mice were fasted for 24 h (n = 32 in each

Animals
Six week-old male C57BL/6J, DBA/1JN, DBA/2N, and CBA/ JN mice were obtained from Jackson Laboratories (Bar Harbor, ME); 6 week-old male BALB/c, C3H/He, and ddY mice were obtained from Japan SLC, Inc. (Hamamatsu, Japan); and 6 week-old male C57BL/6N mice were obtained from CLEA Japan, Inc. (Tokyo, Japan). Mice were fed a normal laboratory diet (CE2; CLEA, Japan) for at least 1 week to stabilize metabolic conditions before being used in the study. Mice were exposed to a 12 h light/12 h dark cycle and maintained at a constant temperature of 22°C. Four mice were housed per plastic cage. Each cage was equipped with plastic partitions to separate individual mice. Mice were cared for in accordance with the NIH Guide for the Care and Use of Laboratory Animals. All animal procedures were reviewed and approved by the National Institute of Health and Nutrition (Japan).

ddY mice
The ddY strain is outbred and has been maintained as a closed colony. There is only one repository in Japan (Japan SLC, Inc). Mice of this strain show good reproductive performance and superior growth. The ddY strain is derived from a mouse colony at the Institute of Infectious Diseases of Tokyo University (Denken; presently, the Institute of Medical Sciences of Tokyo University). This colony was introduced from Germany in the 1910s and 1920s and established as a strain at the National Institute of Health (Yoken, presently, the National Institute of Infectious Diseases). The strain name ddY stands for Deutschland, Denken, and Yoken. The ddY strain differs from the DDY strain because the DDY strain has been established as an inbred strain from the ddY colony at Yoken (note that the inbred strain DDY is spelled with all uppercase letters and with lowercase letters in the name of outbred strain ddY). Some disease models have been developed from the ddY strain, including HIGA mice (IgA nephritis model) and obese/diabetic mice ( 21,22 ).

Dietary experiments
To study the acute effects of diet, 9 week-old C57BL/6J and ddY mice were fasted for 24 h and then fed a HF diet containing 60 energy% (en%) saffl ower oil, 20 en% casein, and 20 en% starch (n = 6 in each strain) for 3 h. Saffl ower oil (high-oleic type) contained 45% (wt/wt) oleic acid (18:1n-9) and 46% linoleic acid (18:2n-6). Serum was drawn from the tail vein after 24 h fasting and 3 h feeding. Two weeks later, these C57BL/6J and ddY mice were fasted for 24 h and then introduced to a fat-free diet containing 80 en% starch and 20 en% casein (n = 6) for 3 h. Serum was drawn from the tail vein after 24 h fasting and 3 h feeding.
Food intake per day was estimated by subtracting the remaining food weight from the initial food weight of the previous day. With these data, average energy intake per each group of mice was calculated for the total experimental period.

Oral saffl ower oil loading
Eight week-old C57BL/6J and ddY mice were fasted for 24 h (n = 16 per mouse line). The next morning (10:00 AM), half of the mice in each mouse line were loaded with saffl ower oil by After 2 h incubation at room temperature, the plates were washed four times with PBS containing 0.05% Tween-20, and 100 l of TMB One Solution (Promega, Madison, WI), which turns blue in the presence of peroxidase, was added to the plates. After 10-min incubation in the dark, 1 M H 2 SO 4 was used to stop the reaction. The plates were read by plate reader using a 450-nm fi lter.

Other assays
Protein concentration was determined using the Bio-Rad Protein Assay (Bio-Rad Laboratories, Hercules, CA) using bovine IgG as a standard.
DNA was extracted from mouse tails using proteinase K and phenol/chloroform ( 31 ) with an automated DNA isolation system (PI-50␣; KURABO, Osaka, Japan) and treatment with RNase. Exon 8 and exon 9 of the LPL gene were amplifi ed by PCR. Purifi ed PCR products were sequenced using a 3730xl DNA Analyzer (Applied Biosystems, Foster City, CA) and the recommended sequencing standards (BigDye Terminator v3.1 Cycle Sequencing Kit; Applied Biosystems).

Statistical analysis
Two-way ANOVA was used to examine the two main effects of diet or saffl ower oil and mouse strain and their interaction (Stat-View 5.0; Abacus Concepts, Inc., Berkeley, CA). When differences were signifi cant with respect to main or interaction effects, each group was compared with the others by Fisher's protected least signifi cant difference test. Repeated measures analysis was used for comparisons of the serial data from the same individual mice, and, where statistically signifi cant, each group was compared with the others by Fisher's protected least signifi cant difference test. Data from two groups were compared by Student's paired t -test. Statistical signifi cance was set at P < 0.05. Values are shown as mean ± SEM.

An increase in postprandial TG concentrations in ddY mice
As suggested by our previous studies (18)(19)(20), a marked increase in postprandial TG concentrations after being fed a HF diet was observed in ddY mice but not in C57BL/6J mice. Mice were fed a standard laboratory chow and fasted for 24 h and then allowed to freely eat a HF diet (60% energy from saffl ower oil). Blood was obtained before and at 3 h after the initiation of refeeding. Average intake of energy during the 3 h period was 7.9 ± 1.3 kcal and 11.8 ± 2.2 kcal for C57BL/6J and ddY, respectively (n = 8; p = 0.146). Fasted TG and TC concentrations in ddY mice were 2.3-and 1.8-fold larger than in C57BL/6J mice, respectively ( Fig. 1A ). There were no differences between fasted and postprandial TG concentration in C57BL/6J mice, whereas postprandial TG concentration was 2.3-fold higher than that of fasted TG in ddY mice. There was no signifi cant difference in TC concentration between fasting and postprandial states in either strain of mice. Fasting and postprandial glucose concentrations did not differ between the two strains of mice; however, the postprandial insulin concentration was 2.6-fold higher in ddY mice.
When mice were fed a fat-free, high-starch diet, there were no signifi cant increases in postprandial TG concentrations between fasting and refeeding conditions in either mouse strain). To assess the post-heparin effect on plasma, a quarter of the mice were injected intraperitoneally with porcine heparin (1,000 U/kg body weight) (n = 8 in each group) ( 27 ). Cardiac blood was collected 5 or 20 min later and subjected to an LPL activity kit (Roar Biomedical, New York, NY). There are two main heparin-releasable types of LPL, one that is relevant to the lipolysis of lipoproteins and one that is largely irrelevant ( 28 ). To distinguish LPL activity in these two compartments, measurements at two points (5 and 20 min) were conducted as different experiments (see Results). A total of 2 µl of 100-fold diluted heparinized plasma was mixed with 600 µl of substrate emulsion and incubated at 25°C for 1 h, and the fl uorescence intensity was measured at 370 nm excitation/450 nm emission ( 29 ). The rest of the mice were loaded with saffl ower oil (0.4 ml/mouse), and cardiac blood was drawn at 1.5, 3, and 4.5 h after administration of saffl ower oil. Mice were injected intraperitoneally with porcine heparin (1,000 U/kg body weight; n = 8 per group) 5 or 20 min before blood collection at each time point. Hepatic lipase activity was determined by incubating the reaction in the presence of 1 M NaCl. LPL activity was calculated as the difference between total lipase activity and hepatic lipase activity.
For tissue LPL assays, an independent experiment using a similar animal protocol for measurement of plasma LPL activity was conducted without intraperitoneal heparin injection (n = 8 per group). At the end of the experiment, 0.1 g of WAT, gastrocnemius, and brown adipose tissue (BAT) was fl ash frozen in liquid nitrogen and stored at Ϫ 70°C until used. Tissues were transferred to ice-cold tubes containing 1 ml of DMEM containing 2% (w/v) BSA and 2 units/ml of heparin ( 30 ). The tissue was homogenized with a Kontes Dounce Tissue Grinder (Kimble Chase, Vineland, NJ) and incubated at 37°C for 1 h. After centrifugation, cleared supernatant was used for detection of LPL activity as described above.  Serum TG, TC, glucose, and insulin concentrations of fasted mice and those fed a HF diet or a fat-free, high-starch diet for 3 h. Nine week-old male C57BL/6J and ddY mice were fasted for 24 h. After serum was collected, a HF diet (A) was fed for 3 h to the fasted mice, and serum was again collected. Two weeks later, these mice were fasted for 24 h. After serum was collected, a fat-free diet (B) was fed for 3 h to the fasted mice, and serum was again collected. C57BL/6J mice are represented by black bars; ddY by white bars. Statistical signifi cance as determined by ANOVA is shown in the boxes. Data represent means ± SEM (n = 6). The interaction of saffl ower oil and mice strain on serum TG and insulin concentrations in a HF diet is signifi cant by ANOVA, suggesting that changes in these metabolites are unique responses to dietary fat in ddY mice. Means that share a common letter are not signifi cantly different, whereas means without a common letter are signifi cantly different by a post hoc test ( P < 0.05). strain ( Fig. 1B ), suggesting that the increase in postprandial TG concentrations in ddY mice after a HF diet was not due to feeding itself but to saffl ower oil ingestion. Glucose concentrations did not markedly differ between the two strains. However, the increase in insulin concentration in response to a high-starch diet was larger in ddY mice, which was similar to that seen after being fed a HF diet.

ddY mice have a marked increase in serum TG concentrations after an intragastric saffl ower oil load
To avoid the effects caused by the difference in the amount and time of consumed dietary fat between the two strains, C57BL/6J and ddY mice were given an intragastric saffl ower oil bolus (0.4 ml per mouse). Mice were euthanized at 24 h fasting or at 3 h after the bolus, and their phenotypes, including blood metabolites, were examined. Because low levels of adiponectin are reportedly associated with increased plasma TG ( 32 ), the serum adiponectin concentration was measured. Due to the facts that the average body weight of ddY mice (32.0 ± 0.3 g; n = 16) was larger than that of C57BL/6J mice (18.6 ± 0.3 g; n = 16) ( Table 1 ) and that both strains were given a 0.4 ml dose per mouse, the ddY dose can be considered lower relative to total body weight. Tissue weights (liver, WAT, gastrocnemius, and BAT) were also larger in ddY mice.
Under fasting, ddY mice had higher serum TG and TC concentrations ( Table 1 ). In C57BL/6J mice, TG was 57 mg/ dl compared with 134 mg/dl in ddY mice (not statistically signifi cant). In C57BL/6J mice, TC was 115 mg/dl compared with 257 mg/dl in ddY mice (statistically signifi cant). There were no differences among serum glucose and NEFA concentrations. At 3 h after a saffl ower oil load, ddY mice had a 4.3-fold increase (signifi cant) in serum TG concentration, whereas C57BL/6J mice had only a 1.9-fold increase (not signifi cant) compared with fasting values. There was no significant difference in TC concentration between fasting and post saffl ower load in either strain. After a saffl ower oil load, the NEFA concentration was decreased in C57BL/6J mice but was increased in ddY mice by 1.8-fold. In contrast, blood glucose concentrations in C57BL/6J mice were increased but were decreased in ddY mice. There was no difference in adiponectin concentrations between fasting and postprandial state in either strain; however, serum adiponectin concentrations in ddY mice were about 50% lower than in C57BL/6J mice in fasting and postprandial states.
Considering the concordance in the amount of fat load and similarity in metabolite changes between the two different methods of fat loading, oil loading by gavage was used in the following experiments.

Time course of metabolic blood markers after saffl ower oil load
To examine the changes in metabolite and hormones in detail, blood samples were drawn at 0, 1, 3, 6, 9, and 12 h after the bolus. Serum TG, NEFA, glucose, insulin, and GIP concentrations were determined in both strains of mice. GIP was measured in this experiment because dietary fat has been found to markedly stimulate GIP secretion and may increase insulin secretion ( 33,34 ).
The increase in TG concentration in ddY mice was largest at 3 h after an intragastric saffl ower oil load and returned to a baseline level at 12 h ( Fig. 2A ). Similar to the TG changes, the increase in NEFA concentration was largest at 3 h, suggesting that NEFA might be derived from lipolysis of lipoproteins ( Fig. 2B ). In C57BL/6J mice, NEFA concentrations were decreased after a saffl ower oil load, possibly by the inhibition of lipolysis of TG in adipocytes. Glucose concentrations at fasting and after a saffl ower oil load were lower in ddY mice than in C57BL/6J mice ( Fig. 2C ). This might be mediated by increased serum insulin concentration because an increase in insulin concentration was observed at 3 h after a saffl ower oil load ( Fig. 2D ). GIP concentration was increased by 9.2-fold and could possibly stimulate insulin release from pancreatic ␤ -cells ( Fig. 2E ).
In the dietary experiment shown in Fig. 1 , there was no difference in glucose levels between ddY and C57BL/6J mice, and both strains had a similar increase 3 h after a HF diet. This differs from the saffl ower oil loading experiments where glucose levels increased in C57BL/6J but decreased in ddY ( Table 1 ; Fig. 2 ). This was apparently due to the gavage because the increase of glucose concentration after saline loading by gavage was observed in C57BL/6J mice but not in ddY mice (data not shown); however, the mechanism of this is unclear.
ddY mice have increased TG in postprandial CM, VLDL-sized fractions, and free glycerol TG and TC concentrations in plasma lipoproteins before and 3 h after a saffl ower oil load in both strains of

Other mouse strains did not show postprandial hypertriglyceridemia
To examine whether other mouse strains displayed postprandial hypertriglyceridemia, six other strains of male mice (C57BL/6N, DBA/1JN, DBA/2N, BALB/c, C3H/He, and CBA/JN) were given an intragastric saffl ower oil bolus (0.4 ml per mouse), and their TG concentrations were compared with those of C57BL/6J and ddY mice. Among these six strains, BALB/c mice showed the largest increase in TG at 3 h after the saffl ower oil load. However, the BALB/c postprandial TG concentration was 136 mg/dl ( Table 2 ), which was much lower than that of ddY mice (309 mg/dl), suggesting that the ddY mouse is a unique strain that shows hypertriglyceridemia in response to dietary fat. Female ddY mice showed a similar increase in postprandial TG concentration (303.9 ± 30.4 mg/dl; n = 6) (data not shown) to that of male ddY mice. Fasting TC concentrations did not differ between strains (data not shown). Nine week-old C57BL/6J and ddY mice were fasted for 24 h, at which point half of the mice were euthanized, and serum and tissues were collected. The remaining mice were loaded with 0.4 ml of saffl ower oil and 3 h later were euthanized, and serum and tissues were collected. Values are means ± SEM (n = 8 in each group). The interaction of saffl ower oil and mice strain on serum TG, NEFA, and glucose is signifi cant by ANOVA, suggesting that changes in these metabolites are unique responses to saffl ower oil loading in ddY mice. Means that share a common letter are not signifi cantly different, whereas means without a common letter are signifi cantly different according to a post hoc test ( P < 0.05). All strains of 8 week-old male mice were fasted for 24 h and loaded with 0.4 ml of saffl ower oil. Serum was collected under fasting and at 3 h after saffl ower oil loading. Values are means ± SEM (n = 6 in each mouse strain). * P < 0.05, ** P < 0.01, and *** P < 0.001 compared with fasted mice (paired t -test). did not differ between fasting and postprandial states ( Table 3 ).

Increased lipoprotein production in ddY mice
The increase in CM and VLDL-sized TG fractions (CM remnant, VLDL remnant, or VLDL) in ddY mice after a saffl ower oil load could possibly be due to an increase in lipoprotein production and/or impaired lipoprotein clearance. To answer this question, the rate of lipoprotein production was estimated by monitoring serum TG concentrations in the presence of Triton WR 1339, which blocks lipolysis of CM and VLDL.
Under fasting, gradual TG accumulation was observed in C57BL/6J mice ( Fig. 4 ). ddY mice showed a 1.9-fold increase in TG accumulation relative to C57BL/6J mice, suggesting that VLDL production in ddY mice was increased by 1.9-fold relative to C57BL/6J mice.
After a saffl ower oil load, 4.2-and 5.4-fold increases of TG accumulation relative to fasting were observed in C57BL/6J and ddY mice, respectively, suggesting that production of lipoproteins in both lines of mice markedly increased after a saffl ower oil load. However, lipoprotein production (accumulated amount of TG) in ddY mice was mice were analyzed by HPLC. Under fasting, ddY mice showed a slight increase of TG peaks in VLDL-sized fractions and free glycerol compared with C57BL/6J mice ( Fig. 3A ; Table 3 ), suggesting that increased VLDL secretion and/or decreased VLDL clearance might occur in ddY mice. Glycerol might be released from adipose tissues and VLDL. Under fasting, the amounts of cholesterol in VLDL, LDL, and HDL fractions in ddY mice were larger than those in C57BL/6J mice ( Table 3 ).
At 3 h after saffl ower oil loading, there was no significant change of TG in lipoprotein profi les in C57BL/6J mice relative to fasting; however, the amounts of TG in all fractions and free glycerol were markedly elevated in ddY mice ( Fig. 3B ; Table 3 ). This suggests that increased production or decreased clearance of CM, CM remnant, VLDL remnant, and/or VLDL might occur in ddY mice with increased lipolysis of these lipoproteins (estimated by an increase in free glycerol). In ddY mice, the amounts of cholesterol in CM and VLDL fractions were increased after saffl ower oil loading, whereas those in LDL and HDL fractions were decreased relative to the fasting state, suggesting that formation of LDL from VLDL might be impaired in ddY mice. The total amount of cholesterol respectively. These two time points were chosen because there are two main heparin-releasable LPL pools, one relevant to the lipolysis of lipoproteins and one that is largely irrelevant ( 28 ). The fi rst pool is on the surface of capillaries in muscle and adipose tissues and released within 1-10 min after an intravenous heparin injection ( 28 ). The second pool is extravascular, perhaps in subendothelial compartments or on the surface of adipocytes or myocytes, and released within 20-30 min after an intravenous heparin injection.
There was no signifi cant difference in fasting LPL activity between the two mouse strains ( Tables 4 and 5 ). However, at 5 min after heparin injection, saffl ower oil-loaded, C57BL/6J mice had 5.0-, 4.5-, and 4.9-fold increases in LPL activity at the 1.5, 3, and 4.5 h time points, respectively, compared with fasting, whereas ddY mice had 1.2-, 1.5-, and 0.8-fold increases ( Table 4 ). This lower LPL activity after saffl ower oil loading in ddY mice was not due to a decrease in the amount of LPL protein because there was no signifi cant change in the amount of LPL protein between the two strains. The lower specifi c LPL activity suggested that modifi cation of LPL to reduce its activity might occur in ddY mice in vivo and could be detected by an in vitro assay. Hepatic LPL activity also increased after a saffl ower oil load in both strains of mice, and these increases were similar between C57BL/6J and ddY mice.
At 20 min after heparin injection, saffl ower oil-loaded, C57BL/6J mice had 3.9-, 4.0-, and 3.2-fold increases in LPL activity at the 1.5, 3, and 4.5 h time points, respectively, compared with fasting, whereas ddY mice had 2.3-, 2.4-, and 2.2-fold increases ( Table 5 ). The increase in LPL activity induced by saffl ower oil loading was severely 2.4-fold higher than in C57BL/6J mice. In C57BL/6J mice, the TG concentration at 3 h after saffl ower oil load in the absence of Triton WR 1339 was 106 mg/dl ( Table 1 ), whereas in the presence of Triton WR 1339 the level at 3 h was 1,547 mg/dl, suggesting that 1,441 mg/dl (= 1,547 Ϫ 106) of TG was hydrolyzed by LPL during this 3 h period (if we assumed that Triton WR 1339 completely inhibited LPL activity). In contrast, the TG concentration in ddY mice at 3 h after saffl ower oil loading in the absence of Triton WR 1339 was 571 mg/dl ( Table 1 ), whereas in the presence of Triton WR 1339 the level at 3 h was 3,784 mg/dl, suggesting that 3,213 mg/dl (= 3,784 Ϫ 571) of TG was hydrolyzed by LPL during this 3 h period. The total amount of TG that was hydrolyzed by LPL was 2.2-fold larger in ddY mice (3,213 mg/dl) than in C57BL/6J mice (1, 441 mg/dl). However, the hydrolysis of lipoprotein was possibly incomplete in ddY mice because CM and VLDLsized TG fractions were increased in ddY mice ( Fig. 3 ). These data suggest that the marked increase in TG after a saffl ower oil load in ddY mice might be due to increased lipoprotein production and insuffi cient lipolysis.

Decreased whole body plasma LPL activity in ddY mice
LPL is the primary enzyme responsible for hydrolysis of CM-and VLDL-TG. After intraperitoneal heparin injection, which releases LPL from attached endothelial cells into blood, LPL activity in plasma and the amount of LPL protein (mass) were measured at four points: 24 h fasting and 1.5, 3, and 4.5 h after saffl ower oil loading in C57BL/6J and ddY mice. LPL activity was measured in plasma collected at 5 min and 20 min after an intravenous heparin injection, and these results are shown in Tables 4 and 5 ,  Table 3 . The TG levels in CM, VLDL, LDL, HDL-TG, and free glycerol fractions are shown in the graph. difference in increase in body weight or tissue weight (except liver weight) between the two strains of mice when fed a LF diet.

Difference of two amino acids in LPL among mice strains
From our collected samples, we were able to detect changes in LPL activity in vitro, suggesting that the modulations of LPL activity in vivo could be sustained in the in vitro environment. This in vitro activity was evident despite the fact that most of the serum had been removed. This implies that there were intrinsic alterations to the LPL itself that caused changes in its activity. We therefore analyzed the gene encoding LPL. Genomic sequencing revealed that there were differences between C57BL/6J and ddY mice in two amino acids in exons 8 and 9. In C57BL/6J mice, the amino acids in residues 327 and 383 were aspartic acid (Asp) and isoleucine (Ile), respectively. However, in ddY mice, those residues were asparagine (Asn) and methionine (Met), respectively ( Fig. 6 ). We sequenced exons 8 and 9 in the LPL gene in all mouse strains described in Table 2 and found that each mouse strain could be inhibited at 5 min but not at 20 min after heparin injection in ddY mice, suggesting that LPL activity was predominantly impaired in endothelial cells.
To examine which tissues are responsible for lowering LPL activity in ddY mice, LPL activity in epididymal WAT, gastrocnemius, and BAT was measured. Both on the basis of change per tissue mass or per whole tissue, an increase in LPL activity in each tissue in response to a saffl ower oil load was observed in C57BL/6J but not in ddY mice ( Table 6 ). This suggests that the increase in LPL activity after a saffl ower oil load in the whole body might be reduced in ddY mice.

Increased body weights after a HF diet in ddY mice
To examine the chronic effects of dietary fat, C57BL/6J and ddY were fed a HF diet or a LF diet, and their phenotypes were compared after 10 weeks of feeding. At 10 weeks after a HF diet, ddY mice showed larger increases in body weight, subcutaneous and mesenteric WATs, and liver weights than C57BL/6J mice without a signifi cant difference in energy intake ( Fig. 5 ). There was no signifi cant Eight week-old C57BL/6J and ddY mice were fasted for 24 h and loaded with 0.4 ml of saffl ower oil. At fasting or 3 h after saffl ower oil administration, serum was collected, lipoprotein fractions were separated by HPLC, and TG and TC in each fraction were measured. A representative graph of TG measurement is shown in Fig. 3 . Values are means ± SEM (n = 8 per group). The interaction of saffl ower oil and mouse strain on total TG and free glycerol, but not total cholesterol, is statistically signifi cant by ANOVA, suggesting that alterations in TG metabolism after fat load is a major phenotype of ddY mice. Means that share a common letter are not signifi cantly different, whereas means without a common letter are signifi cantly different according to a post hoc test ( P < 0.05). postprandial hypertriglyceridemia ( Table 2 ), we conclude that these two differences in amino acids of LPL could not be a cause of postprandial hypertriglyceridemia.

DISCUSSION
In this study, we found that ddY mice showed marked hypertriglyceridemia (5.4-fold) in response to a saffl ower oil load relative to C57BL/6J mice. The ddY strain also showed mild fasting hypertriglyceridemia (2.4-fold) and hypercholesterolemia (2.2-fold). However, the increase in TG concentration in the postprandial state was much larger than in the fasting state (the interaction of saffl ower oil and strain was found to be statistically signifi cant by ANOVA).
In ddY mice, postprandial hypertriglyceridemia was, at least in part, caused by an incomplete hydrolysis of lipoproteins, which might be mediated by a decrease in whole body LPL activity. CM-and VLDL-sized TG fractions were elevated at 3 h after a saffl ower oil load ( Fig. 3 ). VLDLsized TG fractions may contain CM remnant, VLDL remnant, and VLDL. In humans, it has been reported that approximately 80% (or more) of the remnants observed in the postprandial state constitute VLDL-remnant (apoB-100 particles) ( 35 ). However, because rodents produce apoB48-containing lipoproteins in the liver, it was diffi cult to distinguish their origin by the size of apoB. It is possible that VLDL might be hydrolyzed to remnants but not to LDL due to decreased LPL activity. It is also possible that the VLDL fraction contained signifi cant amounts of VLDL. VLDL could originate from the liver and/or small intestine. FFA released from CM-TG by LPL is converted efficiently to VLDL-TG in the liver (36)(37)(38). In rats, VLDL has been found in mesenteric lymph ( 39 ), and CM and VLDL are packed differently in the intestine ( 40 ), suggesting that increased VLDL production in the small intestine may occur in ddY mice.
Triton WR 1339 experiments suggested that CM and VLDL production was increased in ddY mice, if we assumed that Triton WR 1339 could suffi ciently inhibit LPL classifi ed into C57BL/6J type or ddY type, according to LPL sequence. C57BL/6N harbored the C57BL/6J type (Asp 327 and Ile 383), whereas DBA/1JN, DBA/2N, BALB/c, C3H/He, and CBA/J harbored the ddY type (Asn 327 and Met 383). Considering that DBA/1JN, DBA/2N, BALB/c, C3H/He, and CBA/J did not show Fig. 4. Serum TG concentration after Triton WR 1339 injection with or without a saffl ower oil load. Serum TG concentration was measured before and 1, 2, 3, and 4 h after Triton WR 1339 injection (500 mg/kg body weight) in C57BL/6J (black squares) and ddY mice (white squares) under fasting state. In other groups, fasted mice were injected with Triton WR 1339 (500 mg/kg body weight) and loaded 30 min later with saffl ower oil. Serum TG concentration was measured before and 1, 2, 3, and 4 h after saffl ower oil loading (0.4 ml/mouse) (circles). The x axis indicates the time after Triton WR 1339 injection. The arrow indicates the time of saffl ower oil loading. Data represent means ± SEM (n = 8). *** P < 0.001 versus C57BL/6J mice loaded with saffl ower oil after WR 1339 injection. ++ P < 0.01, +++ P < 0.001 versus WR 1339 injected C57BL/6J mice under fasting. Eight week-old male C57BL/6J and ddY mice were fasted for 24 h and re-fed with a 0.4 ml saffl ower oil load. At fasting or 1.5, 3, and 4.5 h after saffl ower oil loading, plasma was collected at 5 min after heparin injection. Values are means ± SEM (n = 8 per group; eight groups). The interaction of saffl ower oil and mouse strain in LPL activity is signifi cant by ANOVA, suggesting that a decrease in LPL activity after saffl ower oil loading is unique in ddY mice. Means that share a common letter are not signifi cantly different, whereas means without a common letter are signifi cantly different according to a post hoc test ( P < 0.05).
in the SREBP-1c promoter and show hepatic steatosis when fed sucrose supplementation or a HF diet ( 19 ). Mice with guanine at this site show increased liver SREBP-1c mRNA in response to a high-fructose diet, whereas mice with adenine, such as C57BL/6J mice, do not ( 47 ). It is unlikely, however, that a mutation of Ϫ 468 bp in the SREBP-1c promoter contributes solely to postprandial hypertriglyceridemia because other mouse strains (BALB/c, C3H/He, and CBA/JN) that harbored the same mutation of the SREBP-1c promoter did not show marked postprandial hypertriglyceridemia. It is possible that this mutation contributes to postprandial hypertriglyceridemia because BALB/c and C3H/He mice showed small increases in blood TG concentrations after saffl ower oil loading ( Table 2 ).
By experiments with and without Triton WR 1339, it was estimated that the amount of serum TG hydrolyzed by LPL during the 3 h period after saffl ower oil loading was about 2.2fold larger in ddY mice (3,213 mg/dl) than in C57BL/6J mice (1,441 mg/dl) (see Results). However, the plasma LPL activity of ddY mice after a saffl ower oil load was 39% that of C57BL/6J activity. There are several possible mechanisms for increased CM and VLDL secretion from the small intestine. First, increased insulin concentration may increase intestinal CM secretion. A higher apoB-48 production rate (intestinal origin) has been observed in men with hyperinsulinemia compared with lower insulin levels, suggesting that insulin may stimulate CM secretion ( 41 ). Increased insulin secretion after a saffl ower oil load in ddY mice might be mediated by an increase in GIP concentration. A HF diet increases GIP concentration in humans ( 33,34,42,43 ). Second, elevated FFA could stimulate CM secretion. It has been reported that acute elevation of FFA stimulates intestinal and hepatic lipoprotein production in hamsters and humans ( 44,45 ). Third, sterol regulatory element-binding protein (SREBP)-1c expression in enterocytes may increase in ddY mice, and this leads to increased lipoprotein secretion. Elevated SREBP-1c expression has been demonstrated in insulin-resistant enterocytes and may increase lipogenesis and therefore enlarge the intracellular pool of fatty acid available for assembly into lipoprotein-TG by intestinal cells ( 46 ). The ddY mice possess guanine Ϫ 468 bp Eight week-old male C57BL/6J and ddY mice were fasted for 24 h and re-fed with a 0.4 ml saffl ower oil load. At fasting or 1.5, 3, and 4.5 h after saffl ower oil loading, plasma was collected at 20 min after heparin injection. Values are means ± SEM ( n = 8 in each group; eight groups). The interaction of saffl ower oil and mouse strain in LPL activity is not signifi cant by ANOVA, suggesting that the impaired response in LPL activity after saffl ower oil loading in ddY mice is not evident at 20 min after heparin injection. Means that share a common letter are not signifi cantly different, whereas means without a common letter are signifi cantly different according to a post hoc test ( P < 0.05). Eight week-old male C57BL/6J and ddY mice were fasted for 24 h, at which point half of the mice were euthanized, and tissues were collected. The remaining mice were loaded with 0.4 ml of saffl ower oil, and 3 h later mice were euthanized, and tissues were collected. LPL activity in tissues was measured and expressed by (g) tissue weight and by total tissue. Values are means ± SEM (n = 8 in each group). The interaction of saffl ower oil and mouse strain in LPL activity in gastrocnemius and WAT is signifi cant by ANOVA, suggesting that the impaired response in LPL activity after saffl ower oil loading in ddY mice might be present in the whole body. Means that share a common letter are not signifi cantly different, whereas means without a common letter are signifi cantly different according to a post hoc test ( P < 0.05).
Because we could detect changes in LPL activity in vitro, even after the removal of most of the serum, it appears that the modulation of LPL activity in ddY mice relates to an intrinsic alteration in LPL. As such, there may be several mechanisms that could account for the decrease in LPL activity in ddY mice. There were two amino acid differences in LPL between C57BL/6J and ddY mice. However, these differences may not affect LPL activity because other mice strains harboring the same amino acid LPL sequence as ddY mice did not show marked postprandial mice, suggesting that although LPL activity in the whole body was decreased in ddY mice, LPL in ddY mice could hydrolyze most of the CM-TG. It is also conceivable that the inhibition of lipolysis by Triton WR 1339 was insuffi cient. In this case, the production rate of lipoproteins was underestimated (the actual rate of lipoprotein production could be larger than the estimated value). If the residual LPL activity after Triton WR 1339 treatment in ddY mice was lower than that in C57BL/6J mice, the actual production rate of lipoprotein might not be increased in ddY mice relative to C57BL/6J mice.  5. Body weight gain and tissue weights, total energy intake, and liver TG of C57BL/6J and ddY mice after 10 weeks on LF and HF diets. Body weight gain (A), total energy intake (B), the tissue weights of subcutaneous and mesenteric WAT and liver and liver TG concentration (C) after 10 weeks feeding are shown. The experiment was initiated on 7 week-old male mice. Statistical signifi cance as determined by ANOVA is shown in the boxes. Values are mean ± SEM (n = 4 for LF diet; n = 8 for HF diet). The interaction of saffl ower oil and mice strain on body weight gain and liver weight is signifi cant by ANOVA, suggesting that changes in fat accumulation are unique responses to dietary fat in ddY mice. Means that share a common letter are not signifi cantly different, whereas means without a common letter are signifi cantly different according to a post hoc test ( P < 0.05). In panel B, *** P < 0.001 versus C57BL/6J mice fed a HF diet. ++ P < 0.01, +++ P < 0.001 versus C57BL/6J mice fed a LF diet. 3 and 4 ( 55 ). Alterations in activities of these proteins may also affect LPL activity.
In humans, postprandial hypertriglyceridemia is associated with visceral fat deposition (56)(57)(58). In obese men, post-heparin LPL activity has been reported to be lower in subjects with postprandial hypertriglyceridemia ( 59 ). However, it is unknown whether postprandial hypertriglyceridemia mediated by decreased LPL activity is causative in the development of obesity. HF feeding for 10 weeks in ddY mice resulted in more fat accumulation in subcutaneous and mesenteric adipose tissues and liver than in C57BL6/6J mice ( Fig. 5 ), suggesting that hypertriglyceridemia might be causal to intra-abdominal obesity. However, it cannot be ruled out that other genetic factors in ddY mice (ddY mice featured a larger body size including adipose tissues than C57BL/6J mice before the dietary experiment), rather than postprandial hypertriglyceridemia, may promote obesity after a HF dietary challenge. The ddY mouse is the fi rst murine model by which we can examine the mechanisms and effects of postprandial hypertriglyceridemia on obesity and atherosclerosis. Further studies are required to examine the mechanism of this decrease in LPL activity and the effect of long-term, HF diets on atherosclerosis.
hypertriglyceridemia. Decreases in serum adiponectin concentration may contribute to lowering LPL activity. Adiponectin concentrations have been reported to account for 23% of the variation in LPL activity ( 48 ), and adenovirus-mediated increases in blood adiponectin result in increased post-heparin LPL activity ( 49 ). Increased FFA after a saffl ower oil load may inhibit LPL activity. FFA has been reported to inhibit LPL activity (50)(51)(52). Mild hypertriglyceridemia after an intragastric fat load (2-fold increase relative to wild-type mice) has been observed in fatty acid translocase (CD36) knockout mice ( 53 ). CD36 is abundant in peripheral tissues (e.g., adipose tissue, skeletal muscle, and heart) and plays an important role in FFA uptake in these tissues ( 54 ). In CD36 knockout mice, it was hypothesized that the loss of CD36 increased plasma FFA concentrations, which leads to decreased LPL activity and then to an increase in TG-rich lipoproteins ( 53 ). In ddY mice, there are no mutations in CD36 protein (data not shown); however, it is possible that activities of CD36 and/or proteins related to fatty acids transport might be decreased in peripheral tissues. Regulation of LPL activity is complex and controlled by several modulators, such as apolipoproteins (apoC-1, apoC-II, apoC-III, and apoA-V), glycosylphosphatidylinositol-anchored high-density lipoprotein-binding protein 1, and angiopoietin-like proteins