Intestine-specific MTP and global ACAT2 deficiency lowers acute cholesterol absorption with chylomicrons and HDLs.

Intestinal cholesterol absorption involves the chylomicron and HDL pathways and is dependent on microsomal triglyceride transfer protein (MTP) and ABCA1, respectively. Chylomicrons transport free and esterified cholesterol, whereas HDLs transport free cholesterol. ACAT2 esterifies cholesterol for secretion with chylomicrons. We hypothesized that free cholesterol accumulated during ACAT2 deficiency may be secreted with HDLs when chylomicron assembly is blocked. To test this, we studied cholesterol absorption in mice deficient in intestinal MTP, global ACAT2, and both intestinal MTP and global ACAT2. Intestinal MTP ablation significantly increased intestinal triglyceride and cholesterol levels and reduced their transport with chylomicrons. In contrast, global ACAT2 deficiency had no effect on triglyceride absorption but significantly reduced cholesterol absorption with chylomicrons and increased cellular free cholesterol. Their combined deficiency reduced cholesterol secretion with both chylomicrons and HDLs. Thus, contrary to our hypothesis, free cholesterol accumulated in the absence of MTP and ACAT2 is unavailable for secretion with HDLs. Global ACAT2 deficiency causes mild hypertriglyceridemia and reduces hepatosteatosis in mice fed high cholesterol diets by increasing hepatic lipoprotein production by unknown mechanisms. We show that this phenotype is preserved in the absence of intestinal MTP in global ACAT2-deficient mice fed a Western diet. Further, we observed increases in hepatic MTP activity in these mice. Thus, ACAT2 deficiency might increase MTP expression to avoid hepatosteatosis in cholesterol-fed animals. Therefore, ACAT2 inhibition might avert hepatosteatosis associated with high cholesterol diets by increasing hepatic MTP expression and lipoprotein production.

High plasma cholesterol levels increase risk for atherosclerosis. Absorption of dietary and biliary cholesterol by the intestine is a major determinant of plasma cholesterol levels. Therefore, cholesterol absorption has been studied extensively (1)(2)(3) and reduction in cholesterol absorption is a valid target to lower plasma cholesterol concentrations ( 4 ). Cholesterol absorption involves uptake of free cholesterol by enterocytes from the apical side by a mechanism involving Niemann-Pick C1-like 1 (NPC1L1), conversion to cholesteryl esters by ACAT2 in the endoplasmic reticulum (5)(6)(7)(8), and assembly and secretion with chylomicrons by microsomal triglyceride transfer protein (MTP) (9)(10)(11)(12).
ACAT2 expression is restricted to the small intestine and liver ( 13,14 ), and its defi ciency reduces cholesterol absorption rendering mice resistant to diet-induced hypercholesterolemia, gallstone formation, and atherosclerosis ( 15,16 ). Cholesterol esterifi ed by ACAT2 is packaged into chylomicrons and secreted toward the basolateral side. Chylomicron assembly and secretion is critically dependent on MTP and apoB (9)(10)(11)(12). Ablation of both apoB and MTP results in embryonic lethality in mice, as assembly of apoB-containing lipoproteins by the yolk sac appears to be essential for the survival of the embryo (17)(18)(19). Intestine-specifi c MTP (gene, Mttp ) ablation signifi cantly reduces cholesterol absorption in mice ( 20,21 ). Thus, both ACAT2 and MTP play an important role in cholesterol absorption via the chylomicron pathway. Here, we Abstract Intestinal cholesterol absorption involves the chylomicron and HDL pathways and is dependent on microsomal triglyceride transfer protein (MTP) and ABCA1, respectively. Chylomicrons transport free and esterifi ed cholesterol, whereas HDLs transport free cholesterol. ACAT2 esterifi es cholesterol for secretion with chylomicrons. We hypothesized that free cholesterol accumulated during ACAT2 defi ciency may be secreted with HDLs when chylomicron assembly is blocked. To test this, we studied cholesterol absorption in mice defi cient in intestinal MTP, global ACAT2, and both intestinal MTP and global ACAT2. Intestinal MTP ablation signifi cantly increased intestinal triglyceride and cholesterol levels and reduced their transport with chylomicrons. In contrast, global ACAT2 defi ciency had no effect on triglyceride absorption but signifi cantly reduced cholesterol absorption with chylomicrons and increased cellular free cholesterol. Their combined defi ciency reduced cholesterol secretion with both chylomicrons and HDLs. Thus, contrary to our hypothesis, free cholesterol accumulated in the absence of MTP and ACAT2 is unavailable for secretion with HDLs. Global ACAT2 defi ciency causes mild hypertriglyceridemia and reduces hepatosteatosis in mice fed high cholesterol diets by increasing hepatic lipoprotein production by unknown mechanisms. We show that this phenotype is preserved in the absence of intestinal MTP in global ACAT2-defi cient mice fed a Western diet. Further, we observed increases in hepatic MTP activity in these mice. Thus, ACAT2 defi ciency might increase MTP expression to avoid hepatosteatosis in cholesterol-fed animals. Therefore, ACAT2 inhibition might avert hepatosteatosis associated with high cholesterol diets by increasing hepatic MTP expression and lipoprotein production. and Western diets to compensate for lower dietary fat and cholesterol absorption in intestinal MTP-defi cient mice.

Generation of intestine-specifi c MTP, global ACAT2, and MTP/ACAT2 double knockout mice
Three different sets of mice were used in this study. Generation of global Soat2 Ϫ / Ϫ ( 15 ) and intestine-specifi c MTP knockout ( 21 ) mice have been described previously. To  . To induce intestinal Mttp gene ablation, tamoxifen (0.5 mg/mouse) was injected intraperitoneally three times on alternate days in 200 l of corn oil as described before ( 21 ). For a matching control, Soat2 +/+ and ER T2 -villin-Cre;Mttp f/f mice were injected with corn oil (200 l) used to dissolve tamoxifen. No signifi cant differences were found among these two controls; therefore, they have been combined and used as WT controls. Mice were fed either chow diet (LabDiet 5001) or Western diet (TD88137, Harlan Teklad) containing 17, 48.5, 21.2, and 0.2% by weight of protein, carbohydrate, fat, and cholesterol, respectively, after the fi rst tamoxifen injection until they were euthanized 7 days after the last tamoxifen injection. All studies were approved by the institutional animal care and use committee of SUNY Downstate Medical Center.

Plasma lipid measurements
Plasma and tissue total cholesterol and triglyceride (Thermo-Fisher Scientifi c), free cholesterol (Wako Chemicals), and glycerol (Sigma) levels were measured using commercial kits. Esterifi ed cholesterol was calculated by subtracting free cholesterol from total cholesterol. Plasma free glycerol levels were subtracted from triglyceride levels. HDL lipid levels were measured after precipitating apoB lipoproteins using phosphotungstate/ MgCl 2 reagent ( 22 ). Plasma lipoproteins were separated by fast protein liquid chromatography (FPLC) using a Superose 6 column (fl ow rate of 0.2 ml/min) and 200 l fractions were collected to measure cholesterol and triglycerides.

Short-term lipid absorption studies
Age-matched male mice (n = 3 per group) on a chow or Western diet were fasted overnight and injected intraperitoneally with poloxamer 407 (P407) (

H] cholesterol by primary enterocytes
To study uptake, primary enterocytes isolated from overnight fasted mice (n = 3) as previously described ( 22,23 ) Besides being converted to cholesteryl esters and secreted with the chylomicron pathway, free cholesterol taken up by the enterocytes can be secreted with HDLs (22)(23)(24)(25). This pathway appears to be dependent on ABCA1. Ablation of intestinal ABCA1 reduces acute cholesterol absorption by ‫ف‬ 28%, mainly via the HDL pathway ( 21,24 ). Using intestine-specifi c MTP and ABCA1-defi cient mice, we showed that these two pathways are responsible for ‫ف‬ 95% of cholesterol absorption ( 21 ).
The role of ACAT2 in the metabolism of free cholesterol in the intestine has been studied using ACAT2 [gene, sterol O -acyltransferase ( Soat ) 2 ] knockout mice. Early studies indicated that ACAT2 defi ciency increases free cholesterol levels in the intestine ( 15 ). Subsequently, it was shown that ACAT2 defi ciency is associated with increases in intestinal ABCA1 expression, most likely a secondary consequence of increases in cellular free cholesterol. Augmentation of intestinal ABCA1 expression in the absence of ACAT2 was hypothesized to facilitate absorption of cholesterol via the ABCA1/HDL pathway ( 26 ). Indeed, combined defi ciency of ACAT2 and ABCA1 additively reduced cholesterol absorption compared with that observed individually in Abca1 Ϫ / Ϫ and Soat2 Ϫ / Ϫ mice ( 27 ). Thus, it appears that ACAT2 defi ciency might increase free cholesterol absorption by the HDL pathway.
In this study, we examined whether increasing cellular free cholesterol by ACAT2 ablation and curtailing chylomicron assembly by intestinal MTP ablation would further enhance absorption of cholesterol via the HDL pathway. Chylomicrons transport both free and esterifi ed cholesterol. In the presence of ACAT2, cholesterol is esterifi ed and is packaged into chylomicrons. In the absence of ACAT2, free cholesterol may be absorbed via the chylomicron and HDL pathways. However, if the chylomicron pathway is inhibited by ablation of the Mttp gene, then free cholesterol might be secreted with the HDL pathway. Therefore, we hypothesized that ACAT2 knockout mice would increase free cholesterol transport via the HDL pathway when chylomicron assembly was curtailed. To test this hypothesis, we studied acute cholesterol absorption in mice defi cient in intestinal MTP and global ACAT2. Further, we hypothesized that the phenotype of increased cholesterol secretion with HDLs related to the defi ciencies of ACAT2 and MTP might be enhanced when a Western diet rich in cholesterol was fed to these mice.
Another aim of this study was to identify compensatory changes that might occur in mice fed chow and Western diets in the absence of intestinal MTP and global ACAT2. The hypothesis tested was that in the absence of intestinal MTP, the liver might upregulate transport of lipids via VLDLs and these changes might be further affected in the absence of ACAT2. It is known that ACAT2 defi ciency increases free cholesterol in the intestine but not in the liver. One possibility is that VLDL assembly might be increased to alleviate free cholesterol accretion in ACAT2-defi cient hepatocytes. Therefore, we examined whether ACAT2 deficiency affects hepatic lipid metabolism in mice fed chow the loss of ACAT2 does not upregulate ACAT1 in either the intestine or the liver. Further, ablation of ACAT2 had no effect on intestinal and hepatic MTP mRNA ( Fig. 1A, B ) and activity ( Fig. 1C, D ) indicating that ACAT2 deficiency also does not affect MTP expression. ACAT2 deficiency had no signifi cant effect on intestinal triglyceride ( Table 1 ) and on lipid accumulation in enterocytes as determined by Oil Red O staining ( Fig. 1E ). Furthermore, ACAT2 ablation had no signifi cant effect on total cholesterol in the intestine, but it increased free cholesterol by 46% and decreased cholesteryl esters by 43% ( Table 1 ). ACAT2 defi ciency had no effect on hepatic triglyceride, reduced hepatic total cholesterol, had no effect on free cholesterol, and decreased esterifi ed cholesterol ( Table 1 ) consistent with other studies ( 15,31 ). ACAT2 defi ciency had no effect on plasma triglyceride, but reduced total cholesterol concentrations by 18%, mainly due to reductions in esterifi ed cholesterol ( Table 1 ). FPLC analysis showed that ACAT2 defi ciency had no effect on triglyceride and cholesterol in VLDL/LDL fractions ( Fig. 1F, G ), but reduced cholesterol in the HDL fraction ( Fig. 1G ). Thus, ACAT2 defi ciency reduces esterifi ed cholesterol in tissues and plasma. However, it increases free cholesterol in the intestine, but not in the liver, of chow-fed mice.

Intestinal MTP ablation increases intestinal lipids and reduces plasma lipoproteins
Intestine-specifi c Mttp gene ablation was obtained by injecting tamoxifen on three alternate days in chow-fed male ER T2 -Villin-Cre;Mttp f/f mice as previously described ( 21 ). All studies were performed 7 days after the last injection. Tamoxifen injection reduced MTP mRNA ( Fig. 1A ) and activity ( Fig. 1C ) by ‫ف‬ 80% in the intestine, but had no signifi cant effect on intestinal and hepatic ACAT1 and ACAT2 mRNA ( Fig. 1A, B ) and hepatic MTP mRNA ( Fig.  1B ) and activity ( Fig. 1D ). To determine the consequences of MTP deletion on tissue homeostasis, lipids in the intestine of these mice were quantifi ed and Oil Red O staining was performed on the frozen intestinal sections. Consistent with previous reports ( 20,21 ), conditional intestinal ablation of MTP was associated with signifi cant increases in triglycerides, cholesterol, and free cholesterol by 42fold, 16 and 29%, respectively; with no signifi cant change in esterifi ed cholesterol levels ( Table 1 ). As expected, MTP ablation was associated with enhanced Oil Red O staining of the intestinal sections ( Fig. 1E ). Lipids were mostly present in the absorptive epithelial cells of the villi. Intestine-specifi c MTP ablation reduced triglycerides and increased cholesterol, mainly esterifi ed cholesterol, in the livers of these mice ( Table 1 ). Next, the effects of intestinal MTP ablation on plasma lipids were assessed. I-Mttp Ϫ / Ϫ mice had 46 and 55% less plasma triglyceride and cholesterol, respectively ( Table 1 ). FPLC analysis of plasma showed lower triglycerides in VLDL/LDL fractions ( Fig. 1F ) and reduced cholesterol concentrations in both VLDL/ LDL and HDL fractions ( Fig. 1G ). These studies showed that intestine-specifi c MTP ablation is associated with signifi cant lipid accumulation in the intestine and reduced plasma lipids and lipoproteins. radiolabeling experiments, lipids were extracted from cells and media and separated by thin layer chromatography to quantify incorporation of [ 3 H]oleic acid into triglycerides, phospholipids, and cholesteryl esters. In the [ 3 H]cholesterol radiolabeling experiment, media were subjected to density gradient ultracentrifugation to determine radiolabeled cholesterol distribution among different lipoprotein classes ( 22 ).

Histology
Aliquots of intestines were fi xed overnight in 10% formalin, dehydrated in 30% sucrose, embedded in M1 cryo-preservation media at Ϫ 20°C, and stored at Ϫ 70°C. Sections (7 m) were placed on Tissue-Tack (Polysciences) slides, dehydrated in 60% isopropanol, immersed in 1% Oil Red O for 30 min at 22°C, washed in 60% isopropanol, rinsed with tap water for 10 s, counterstained with Gills hematoxylin for at least 20 min, rinsed with tap water until clear, acidifi ed in alcohol (0.4% HCl in 95% ethanol), rinsed with tap water again, and dipped in basic solution (0.03 N NaOH) until sections visibly darkened. Images were taken with a SPOT RT3 digital camera. Image analysis was performed using SPOT software from Imaging Diagnostics.

mRNA quantifi cations
Total RNA from tissues was isolated using TriZol TM (Invitrogen). The purity of RNA was assessed by the A260/A280 ratio and preparations with ratios more than 1.7 were used for cDNA synthesis. The fi rst strand cDNA was synthesized using Omniscript RT (Qiagen) kit. Each reaction of quantitative (q)PCR was carried out in a volume of 20 l, consisting of 5 l cDNA sample (1:100 dilution of the fi rst strand cDNA sample) and 15 l of PCR master mix solution containing 1× PCR reaction buffer (qP-CR TM Core Kit for SYBR Green I, Eurogentec) and specifi c primers ( 21 ). The PCR was carried out by incubating the reaction mixture fi rst for 10 min at 95°C followed by 40 cycles of 15 s incu bations at 95°C and 1 min at 60°C in an ABI 7000 SDS PCR machine. Data were analyzed using the ⌬ ⌬ C T method according to the manufacturer's instructions and presented as arbitrary units and were normalized to ARPp0 mRNA.

Statistics
Data are presented as mean ± SD. Statistical signifi cance ( P < 0.05) was determined using either Student's t -test, one-way ANOVA and comparisons between groups were analyzed using the Newman-Keuls posttest, or two-way ANOVA with Bonferroni's posttest (GraphPad Prism 5). For all knockout mice, WT mice served as controls. However, for I-DKO mice there were two more controls; Soat2 Ϫ / Ϫ and I-Mttp

ACAT2 ablation reduces total plasma cholesterol
As anticipated, ACAT2 mRNA levels were very low in the intestine ( Fig. 1A ) and liver ( Fig. 1B ) of Soat2 Ϫ / Ϫ mice. Deletion of ACAT2 had no effect on the relative expression of ACAT1 in both the intestine and liver ( Fig. 1A, B ), in agreement with other reports ( 15,26 ), confi rming that and Soat2 Ϫ / Ϫ mice, and lower esterifi ed cholesterol levels compared with WT mice ( Table 1 ). However, they were similar to those in I- Ϫ / Ϫ , but these mice did not register any change in ACAT1 mRNA levels compared with WT mice ( Fig. 1A ). I-DKO mice, similar to I-Mttp Ϫ / Ϫ mice, had signifi cantly reduced MTP mRNA , and I-DKO (n = 5) male mice fed a chow diet was used to quantify mRNA levels of ACAT1, ACAT2, and MTP. Intestinal (C) and hepatic (D) tissues were also used to measure MTP activity. Data are presented as mean ± SD. ** P < 0.01 and *** P < 0.001 compared with WT as determined by Student's t -test. Statistically signifi cant differences in different parameters in the four groups were evaluated by one-way ANOVA with Newman-Keuls multiple comparison test. Different letters above bars for each component indicate statistically signifi cant differences in the mean values in different groups ( P < 0.05) as determined by one-way ANOVA. E: Proximal intestinal sections were used for lipid staining by Oil Red O. A higher magnifi cation image of the boxed area is shown under each picture to show the presence of lipids in the absorptive epithelial cells. F, G: Plasma was separated by gel fi ltration to determine mass of triglycerides (F) and cholesterol (G) in different lipoproteins.  Fig. 2A ) indicating that MTP, but not ACAT2, contributes to plasma triglycerides. To gain a better understanding of the contribution of these genes in the transport of newly absorbed lipids, we followed radiolabeled triolein. Appearance of [ 14 C]triolein-derived lipids in the plasma of ACAT2-defi cient mice was not different compared with WT mice after 2 h of gavage ( Fig. 2B ). In contrast, defi ciency of intestinal MTP reduced the appearance of [ 14 C]triolein-derived lipids by >90% compared with WT mice ( Fig. 2B ). Similarly, plasma of I-DKO mice had signifi cantly fewer triolein-derived counts and these counts were not signifi cantly different than those in I-Mttp Ϫ / Ϫ mice. These studies indicate that ACAT2 does not, but MTP does, play a signifi cant role in triglyceride absorption.
Soat2 Ϫ / Ϫ mice ( Table 1 ). One-way ANOVA analysis revealed that total cholesterol in the livers of I-DKO mice was signifi cantly higher than WT and Soat2 Ϫ / Ϫ mice, but not higher than I-Mttp Ϫ / Ϫ mice. These data also showed that accumulation of cholesterol in the liver of I-DKO mice was due to increased free cholesterol. I-DKO mice had signifi cantly lower plasma levels of total triglyceride, as well as total, free, and esterifi ed cholesterol ( Table 1 ). Further, they had reduced lipids in both VLDL and HDL fractions ( Fig. 1F, G ). The plasma lipid levels were not signifi cantly different from I-Mttp Ϫ / Ϫ mice. These studies suggest that the major effect in I-DKO mice on plasma and tissue lipids is due to MTP defi ciency.
Lower plasma lipids in intestine-specifi c MTP-defi cient and global ACAT2 knockout mice are due to reduced lipid absorption To understand the physiological and biochemical mechanisms for lower plasma lipids in I-DKO mice, we hypothesized that reductions in plasma triglyceride and cholesterol concentrations in I-DKO mice might be secondary to lower lipid absorption during the postprandial Intestinal MTP and global ACAT2 defi ciency does not affect fatty acid uptake, but reduces cholesteryl ester secretion by enterocytes Next, we investigated the role of ACAT2 and MTP on the uptake and secretion of fatty acids by the enterocytes.

Soat2 Ϫ / Ϫ and I-Mttp
Ϫ / Ϫ enterocytes took up similar amounts of [ 3 H]oleic acid compared with WT enterocytes ( Fig. 2C ). Moreover, combined defi ciencies of ACAT2 and MTP also had no effect on oleic acid uptake. These studies suggest that both ACAT2 and MTP do not affect fatty acid uptake by enterocytes. We then studied the secretion of [ 3 H]oleic acid-labeled lipids by enterocytes using a pulse-chase protocol. At the end of the chase, radiolabeled lipids in Soat2 Ϫ / Ϫ enterocytes were not statistically different from WT enterocytes ( Fig. 2D ). In contrast, I-Mttp Ϫ / Ϫ and I-DKO enterocytes contained signifi cantly higher labeled lipids, indicating increased retention of lipids in these enterocytes compared with controls. Next, we measured amounts of radiolabeled lipids secreted during the 2 h chase period. Soat2 Ϫ / Ϫ enterocytes secreted lesser amounts of oleic acid-derived labeled lipids ( Fig. 2E ). Similarly, I-Mttp Ϫ / Ϫ enterocytes secreted signifi cantly lower amounts of radiolabeled lipids compared with WT enterocytes. These studies suggest that individual absence of ACAT2 and MTP reduces secretion of 3 H-oleic acid-labeled lipids. I-DKO enterocytes secreted signifi cantly fewer lipids than those secreted by I-Mttp Ϫ / Ϫ and Soat2 Ϫ / Ϫ enterocytes, indicating that ACAT2 and MTP additively contribute to lipid secretion.

ACAT2 and MTP gene ablations reduce cholesterol secretion by enterocytes via the chylomicron pathway
After evaluating the role of ACAT2 and MTP on fatty acid and triglyceride absorption, we studied the effect of these gene deletions on the acute absorption of cholesterol in mice injected with P407 to inhibit plasma lipases. Plasma cholesterol mass remained unchanged in Soat2 Ϫ / Ϫ mice after the gavage of cholesterol and olive oil, but was signifi cantly reduced in I-Mttp Ϫ / Ϫ and I-DKO mice compared with WT controls ( Fig. 3A ). However, the appearance of [ 3 H]cholesterol-derived lipids in the plasma was signifi cantly reduced in both ACAT2-and MTP-defi cient mice ( Fig. 3B ). Decrease in I-DKO acute cholesterol absorption was not signifi cantly different from I-Mttp Ϫ / Ϫ mice, but both these groups showed reduced cholesterol absorption compared with Soat2 Ϫ / Ϫ and WT mice. Thus, ACAT2 defi ciency reduces the appearance of radiolabeled cholesterol in the plasma, whereas MTP defi ciency reduces both radiolabeled and unlabeled cholesterol. We interpret these data to suggest that ACAT2 plays a significant role in the transport of newly absorbed cholesterol, whereas MTP is critical for the transport of both newly absorbed and previously stored cholesterol.
Next, we evaluated the role of MTP and ACAT2 in cholesterol absorption by enterocytes. Absorption involves uptake and subsequent secretion. ACAT2 defi ciency insignifi cantly decreased cholesterol uptake by 15% compared with WT mouse enterocytes ( Fig. 3C ). This was surprising, as ACAT2 defi ciency has been shown to reduce NPC1L1 expression in cholesterol-fed mice ( 26,27 ). Therefore, we measured mRNA levels of NPC1L1. Levels of NPC1L1 were reduced by 23%, but the difference did not reach statistical signifi cance ( Fig. 3D ). No signifi cant reductions in NPC1L1 and cholesterol uptake might be related to the chow-fed mice used in this study. Indeed, feeding of a Western diet reduced cholesterol uptake in these mice (described later). Hence, ACAT2 defi ciency and a high cholesterol diet are needed to see reductions in NPC1L1 expression and cholesterol uptake.
As reported earlier ( 21 ), uptake of cholesterol ( Fig. 3C ) and expression of NPC1L1 ( Fig. 3D ) were reduced in the absence of intestinal MTP. Combined defi ciency of ACAT2 and MTP reduced cholesterol uptake by 49% compared with WT mice ( Fig. 3C ). This reduction was also significantly higher compared with I-Mttp Ϫ / Ϫ and Soat2 Ϫ / Ϫ enterocytes. Gene expression analysis ( Fig. 3D ) revealed that I-DKO enterocytes had reduced expression of NPC1L1 and SR-B1, and these changes might have contributed to signifi cant decreases in the uptake of cholesterol. As opposed to variable results of gene ablations on cholesterol uptake, cholesterol secretion was consistently reduced by 46 and 71% in ACAT2-and MTP-defi cient enterocytes, respectively, and their combined defi ciency reduced cholesterol secretion by 86% compared with controls ( Fig. 3E ). The decrease in cholesterol secretion by the enterocytes from I-DKO was also signifi cantly different from Soat2 Ϫ / Ϫ and I-Mttp Ϫ / Ϫ enterocytes. These studies indicate that both ACAT2 and MTP additively contribute to cholesterol secretion by enterocytes.
It is known that enterocytes secrete cholesterol via the chylomicron (apoB-dependent) or HDL (apoB-independent) pathways ( 22 ). The chylomicron pathway transports both free and esterifi ed cholesterol, whereas the HDL pathway is mainly involved in free cholesterol transport. Therefore, we hypothesized that MTP and ACAT2 defi ciencies that accrete cellular free cholesterol might increase free cholesterol export via the HDL pathway. To test this hypothesis, we subjected media to ultracentrifugation. Individual defi ciencies of ACAT2 and MTP signifi cantly reduced secretion of cholesterol with chylomicrons but had no signifi cant effect on HDL compared with controls ( Fig. 3F-H ).
Thus, reduced secretion of cholesterol with HDL might be secondary to lower intestinal expression of ABCA1 in I-DKO mice.

Effect of intestine-specifi c MTP and global ACAT2 defi ciency on the expression of hepatic lipid metabolism genes
Signifi cant changes in hepatic cholesterol metabolism have been reported in cholesterol-fed Soat2 Ϫ / Ϫ mice ( 15,26,27,32 ), but these changes have not been quantifi ed in chow-fed animals. Intestinal MTP defi ciency has been shown to affect hepatic gene expression ( 20,21 ). Thus, defi ciencies of ACAT2 and MTP have been shown to affect hepatic lipid metabolism. There is no data about the effects of I-Mttp and global ACAT2 defi ciencies on hepatic lipid However, combined defi ciency of MTP and ACAT2 reduced cholesterol secretion with chylomicrons by 91% and with HDLs by ‫ف‬ 49% ( Fig. 3G, H ). Thus, individual defi ciencies of ACAT2 and MTP reduce secretion of cholesterol with chy lomicrons only, but their combined deficiency additionally reduces secretion of cholesterol with HDLs.
To understand the reasons for decreased cholesterol secretion with HDL by enterocytes defi cient in both ACAT2 and MTP, we measured mRNA levels of ABC transporters involved in cholesterol effl ux ( Fig. 3I ). ACAT2 defi ciency increased ABCG5, ABCG8, and ABCA1 expression. MTP defi ciency reduced ABCG5 and ABCG8, but had no effect on ABCA1 expression. Surprisingly, combined deficiency of ACAT2 and MTP reduced expression of ABCA1. , and I-DKO (n = 5) male mice fed a chow diet was used to quantify mRNA levels of different cholesterol absorption (I). Each measurement was done in triplicate with three mice per group. Data in (C) and (E-H) were normalized to cellular protein and are representative of two separate experiments. Data are presented as mean ± SD. * P < 0.05, ** P < 0.01, and *** P < 0.001 compared with WT as determined by Student's t -test. Statistically signifi cant differences in different parameters in the four groups were evaluated by one-way ANOVA with Newman-Keuls multiple comparison test. Different letters above bars indicate statistically signifi cant differences ( P < 0.05) as determined by one-way ANOVA. metabolism in Western diet-fed mice. We hypothesized that reduced delivery of lipids from the intestine might preclude increases in hepatic VLDL secretion.
To determine the effect of diet enriched in fat and cholesterol on lipid absorption, mice were fed a Western diet for 12 days starting with the fi rst tamoxifen injection. First, we looked at the changes within the Western diet-fed mice. Western diet had no signifi cant effect on intestinal triglyceride and total cholesterol, increased free cholesterol by ‫ف‬ 65%, and decreased esterifi ed cholesterol by 50% in ACAT2-defi cient mice compared with WT mice ( Table 1 ). Thus, ACAT2 defi ciency affects percent distribution of free and esterifi ed cholesterol in the intestine. Intestinal MTP defi ciency alone and in combination with ACAT2 defi ciency increased intestinal triglycerides, but had variable effects on intestinal cholesterol. Thus, ACAT2 deficiency decreases intestinal cholesterol esters, whereas MTP defi ciency increases triglycerides.
ACAT2 defi ciency reduced hepatic triglycerides and cholesterol consistent with an earlier study ( 32 ). Intestinal MTP defi ciency had no effect on hepatic triglyceride and total cholesterol in Western diet-fed mice. However, I-DKO mice had signifi cantly reduced hepatic triglyceride and cholesterol . Thus, intestinal MTP and total ACAT2 defi ciencies reduce hepatic lipids.
Soat2 Ϫ / Ϫ mice had higher plasma triglyceride but lower cholesterol levels. I-Mttp Ϫ / Ϫ mice had lower plasma triglyceride and cholesterol levels. I-DKO mice had signifi cantly higher plasma triglyceride but reduced plasma cholesterol ( Table 1 ). Thus, total defi ciency of ACAT2 appears to have a dominant effect on plasma and hepatic triglyceride levels than intestinal MTP defi ciency has on plasma and hepatic triglycerides. These results suggest that reduced delivery of lipids from the intestine in I-DKO mice might not affect increases in hepatic VLDL secretion that occur as a consequence of ACAT2 defi ciency in cholesterolfed mice.
Second, we compared the interactions of diets and genes by two-way ANOVA. Except for intestinal total and free cholesterol, all lipid parameters showed signifi cant interactions ( Table 1 ). Third, we compared the effects of the Western diet on these parameters in different types of mice by applying Bonferroni posttest. Although intestinal triglyceride content tended to increase in Western diet-fed WT and Soat2 Ϫ / Ϫ mice, they did not reach statistical signifi cance compared with chow-fed animals. Surprisingly, intestines from the Western dietfed I-Mttp Ϫ / Ϫ and I-DKO mice had signifi cantly lower amounts of triglycerides compared with the chow-fed animals. Intestinal cholesteryl esters increased after Western diet feeding in WT and I-Mttp Ϫ / Ϫ mice. Hepatic triglycerides were signifi cantly increased in all four different types of mice fed a Western diet compared chowfed animals. Total hepatic cholesterol increased in WT and I-Mttp Ϫ / Ϫ mice, but not in Soat2 Ϫ / Ϫ and I-DKO mice.

Hepatic free cholesterol decreased in Soat2
Ϫ / Ϫ and I-DKO mice, increased in I-Mttp Ϫ / Ϫ mice, and did not change in WT mice. Hepatic esterifi ed cholesterol content increased in all mice after Western diet feeding. Plasma metabolism. It would be interesting to know changes in hepatic lipid metabolism associated with reduced intestinal lipid absorption. We hypothesized that reduced delivery of lipids in the absence of intestinal MTP and ACAT2 might alter hepatic lipid metabolism. Therefore, we studied the effect of intestinal MTP and global ACAT2 defi ciency on the expression of hepatic genes involved in lipid metabolism. First, we studied the effects of ACAT2 defi ciency on hepatic expression of genes involved in fatty acid and triglyceride synthesis. ACAT2 defi ciency did not affect hepatic I-FABP, L-FABP, DGAT1, SCD1, and ACC1, but increased MGAT2, DGAT2, FAS, and SREBP-1c ( Fig. 4A ). Intestinal MTP defi ciency had no effect on hepatic L-FABP, DGAT2, and SCD1, but increased MGAT2, DGAT1, FAS, ACC, and SREBP-1c mRNA levels. Combined defi ciencies of ACAT2 and MTP reduced the expression of I-FABP, L-FABP, and DGAT2; increased the expression of MGAT2 and SREBP-1c; and had no effect on DGAT1, SCD1, and ACC. These studies indicate variable effects of ACAT2 and MTP gene ablations on the expression of genes involved in fatty acid and glycerolipid synthesis. Analysis of mRNA levels of genes involved in fatty acid oxidation showed that the livers of Soat2 Ϫ / Ϫ mice had signifi cantly increased expression of PPAR ␣ , PPAR ␥ , and CPT1 ␣ ( Fig. 4B ). I-Mttp Ϫ / Ϫ mice had reduced expression of hepatic PPAR ␣ , PPAR ␥ , and CPT1 ␣ ( Fig. 4B ). In I-DKO mice, only PPAR ␣ mRNA levels were signifi cantly lower.
These studies indicate modest, if any, effects on fatty acid oxidation in I-DKO mice.
Next, we studied the effect of ACAT2 defi ciency on genes involved in cholesterol metabolism. We did not see signifi cant changes in the expression of genes involved in cholesterol metabolism, except for increases in SR-B1, consistent with no signifi cant changes in hepatic cholesterol ( Table 1 ). We observed signifi cant increases in hepatic cholesterol levels in I-Mttp Ϫ / Ϫ mice and in the expression of ABCG5, ABCG8, ABCA1, HMGR, LDL receptor, SREBP-2, and SR-B1 ( Fig. 4C ), but not in ACAT1 and ACAT2 ( Fig. 1B ). These studies suggest signifi cant effects of intestinal MTP defi ciency on cholesterol transport. In I-DKO mice, hepatic expression of ABCG5, ABCG8, HMGR, and LDL receptor was increased, but that of SR-B1 was unchanged ( Fig. 4C ). These fi ndings suggest modest compensatory alterations in hepatic lipid metabolism in I-DKO mice fed a chow diet.

Effect of ACAT2 and MTP defi ciency on intestinal, hepatic, and plasma lipids in Western diet-fed mice
It is known that ACAT2 defi ciency increases free cholesterol in the intestine, but not in the liver, in cholesterol fed mice. Further, it has been shown that VLDL assembly is increased in these mice. It is possible that increases in VLDL assembly occur to avoid toxicity associated with hepatic free cholesterol assimilation. If this is true, then there might not be any need to increase hepatic VLDL assembly when intestinal cholesterol absorption is curtailed. Therefore, we examined whether intestinal MTP defi ciency in combination with global ACAT2 defi ciency affects hepatic lipid triglycerides, total cholesterol, and free cholesterol were increased in all four types of mice fed a Western diet.

Effect of ACAT2 and MTP defi ciency on hepatic lipid accumulation in Western diet-fed mice
It has been reported previously that ACAT2 defi ciency prevents hepatic steatosis in cholesterol-fed mice by increasing the mobilization of hepatic triglycerides ( 32 ). In this study, we observed that ACAT2 defi ciency increases plasma triglyceride and decreases hepatic triglyceride even when triglyceride absorption from the intestine is signifi cantly curtailed due to MTP defi ciency ( Table 1 ). To explain the mechanisms, we measured expression of genes involved in lipogenesis, ␤ -oxidation, and lipoprotein production. ACAT2 defi ciency had no effect on I-FABP, MGAT2, and DGAT2, but reduced L-FABP, FAS, and SREBP-1c compared with WT mice ( Fig. 6A ). I-DKO mice had reduced expression of I-FABP and FAS. These studies suggest that there might not be consistent reductions in lipogenesis in the absence of ACAT2.
Next, we measured expression of genes involved in ␤oxidation. ACAT2 defi ciency had no signifi cant effect on PPAR ␣ , PPAR ␥ , and CPT1 ␣ mRNA levels ( Fig. 6B ). Livers of I-Mttp Ϫ / Ϫ and I-DKO mice had lower levels of PPAR ␣ and CPT1 ␣ compared with WT and Soat2 Ϫ / Ϫ mice. Therefore, changes in the expression of genes involved in ␤oxidation do not explain consistent reductions in hepatic triglyceride in Western diet-fed Soat2 Ϫ / Ϫ and I-DKO mice.
However, we did observe a signifi cant increase in hepatic Mttp mRNA levels in ACAT2-defi cient and I-DKO mice ( Fig. 6B ). Further, MTP activity was increased by 20-44% in Western diet-fed Soat2 Ϫ / Ϫ and I-DKO mice ( Fig. 6C ). Hence, it is likely that increased MTP expression might be associated with enhanced lipoprotein production in the absence of ACAT2, explaining reduced hepatosteatosis and increased plasma triglyceride in Soat2 Ϫ / Ϫ mice. These results suggest that increases in hepatic MTP expression and lipoprotein production in I-DKO mice is independent of any changes in the intestinal lipid absorption.

DISCUSSION
Cholesterol secretion by enterocytes occurs by apoBdependent (chylomicron) and apoB-independent (HDL) pathways. A hypothesis tested in this paper was that free cholesterol absorption might be increased via the HDL pathway after the inhibition of chylomicron assembly by ablating Mttp gene in the intestines of ACAT2-defi cient mice. Our studies show that accumulation of free cholesterol after ACAT2 ablation and failure to assemble chylomicrons due to MTP defi ciency does not increase the transport of free cholesterol with HDLs; instead, we found signifi cant reduction in cholesterol secretion with HDLs in Western diet-fed mice. Our studies indicate that a reason for reduced cholesterol secretion with HDLs in I-DKO mice might be related to diminished expression of ABCA1. The reasons for these unexpected fi ndings are not clear. It is possible that free cholesterol that accumulates in the absence of ACAT2 is not available for transport via the ABCA1 pathway, suggesting the existence of different pools specifi c for these two pathways. However, this hypothesis does not mice ( Table 1 ). These studies indicate signifi cant gene/ diet interactions in these mice.

Effect of ACAT2 and MTP defi ciency on lipid absorption in Western diet-fed mice
Next, we studied the effects of intestinal MTP and ACAT2 defi ciency on the acute absorption of triglycerides and cholesterol in mice fed a Western diet and injected with P407 to inhibit plasma lipases. Similar to chow-fed mice, the appearance of [ 14 C]triolein-derived lipids was unaffected by ACAT2 defi ciency ( Fig. 5A ). However, I-Mttp Ϫ / Ϫ and I-DKO mice showed a significant decrease of 80 and 86%, respectively, in the absorption of [ 14 C]triolein ( Fig. 5A ). The appearance of [ 3 H]cholesterol-derived lipids in the plasma of Soat2 Ϫ / Ϫ , I-Mttp Ϫ / Ϫ , and I-DKO mice was signifi cantly reduced by 59, 76, and 87%, respectively, compared with WT mice ( Fig. 5B ). The reduction in cholesterol absorption in I-DKO mice was not statistically different from that in I-Mttp Ϫ / Ϫ mice, suggesting that ACAT2 defi ciency does not affect cholesterol absorption in the absence of MTP. Cholesterol-derived counts were lower in apoB-containing nonHDL lipoproteins ( Fig. 5C ). Individual ablation of these genes had no effect on cholesterol transport via HDLs, but I-DKO mice showed 42% decreased cholesterol absorption with HDLs ( Fig. 5D ). These studies indicate that both ACAT2 and MTP play a signifi cant role in cholesterol absorption via nonHDL pathways. However, combined defi ciency of these genes also reduces cholesterol transport with HDLs.
We then evaluated the role of MTP and ACAT2 in the uptake of cholesterol by enterocytes isolated from Western diet-fed gene-ablated mice. Cholesterol uptake was reduced by 25, 29, and 44% in Soat2 Ϫ / Ϫ , I-Mttp Ϫ / Ϫ , and I-DKO mice, respectively ( Fig. 5E ). To understand the reasons for reduced uptake, we measured mRNA levels of genes involved in cholesterol uptake and transport. Individual and combined defi ciencies of ACAT2 and MTP reduced NPC1L1 mRNA levels ( Fig. 5F ). These studies indicate that Western diet feeding may reduce cholesterol uptake in ACAT2-and MTP-defi cient enterocytes by reducing NPC1L1 expression.
Furthermore, we studied the secretion of cholesterol by enterocytes. ACAT2 defi ciency decreased secretion of cholesterol by 27%, whereas MTP defi ciency decreased it by 55% and their combined defi ciency decreased cholesterol secretion by 67% ( Fig. 5G ). The decrease in cholesterol secretion by I-DKO enterocytes was not statistically different from I-Mttp Ϫ / Ϫ mouse enterocytes. This decrease in cholesterol secretion was mainly with chylomicrons in Soat2 Ϫ / Ϫ and I-Mttp Ϫ / Ϫ enterocytes ( Fig. 5H, I ). However, enterocytes from I-DKO mice showed reduced cholesterol secretion with both chylomicrons and HDLs ( Fig. 5H-J ). To fi nd out the reasons for reduced cholesterol secretion with HDLs, we measured mRNA levels of ABCA1, ABCG5, and ABCG8 ( Fig. 5F ). In I-DKO mice, expression of these transporters was signifi cantly reduced. These data suggest that feeding a Western diet to ACAT2-and MTP-defi cient mice reduces uptake and secretion of cholesterol by the enterocytes. show that ACAT2 defi ciency reduces secretion of cholesterol with chylomicrons, indicating that ACAT2 is a major enzyme contributing to the secretion of cholesterol with intestinal lipoproteins. These studies confi rm earlier observations that MTP defi ciency has no effect on fatty acid uptake but reduces glycerolipid and cholesterol secretion , and I-DKO (n = 3) male mice fed a Western diet for 12 days was used to quantify mRNA levels of different genes involved in lipid synthesis (A) and fatty acid oxidation (B). Further MTP mRNA (B) and activity (C) was measured. Data are presented as mean ± SD. * P < 0.05, ** P < 0.01, and *** P < 0.001 compared with WT as determined by Student's t -test. Statistically signifi cant differences in different parameters in the four groups were evaluated by one-way ANOVA with Newman-Keuls multiple comparison test. Different letters above bars indicate statistically signifi cant differences ( P < 0.05) as determined by one-way ANOVA. uptake and triglyceride secretion. Further, ACAT2 deficiency has no effect on cholesterol uptake in chow-fed mice, but feeding a Western diet reduces intestinal cholesterol uptake. Our studies are in agreement with earlier observations that ACAT2 defi ciency increases intestinal cellular free cholesterol and decreases esterifi ed cholesterol. We of hepatic free cholesterol. It is interesting to note that this accommodation happens in the liver, but not in the intestine, probably refl ecting the possibility that the liver regulates intracellular free cholesterol levels more stringently than the intestine. It is known that the liver can also convert cholesterol to bile acids. It is not known why bile acid synthesis is not increased to avoid reduced intracellular free cholesterol in ACAT2-defi cient mice. We speculate that secretion of free cholesterol via VLDL biogenesis might be more effi cient to reduce hepatic free cholesterol levels. Thus, ACAT2 defi ciency increases hepatic MTP expression after feeding high cholesterol diets, but mechanisms involved in this regulation remain to be explained.
In summary, these studies show that combined deficiency of ACAT2 and MTP reduces cholesterol secretion compared with WT mice and mice defi cient in individual genes. Biochemical studies suggest that this reduction is related to signifi cantly reduced secretion of cholesterol by chylomicrons in chow-fed animals. In Western diet-fed mice, cholesterol secretion via both the chylomicron and HDL pathways is reduced. Further, these studies show that ACAT2 defi ciency in animals fed high cholesterol diets increases hepatic MTP expression to avoid hepatosteatosis and cause hypertriglyceridemia. This might be a mechanism to avoid hepatic free cholesterol accumulation. Hence, it is likely that ACAT2 inhibitors might be able to avoid hepatosteatosis associated with high cholesterol diets. Further, they may act in combination with MTP inhibitors to lower hepatosteatosis.
with chylomicrons resulting in increased accumulation of triglyceride and cholesterol in the intestine and signifi cant reductions in plasma lipids. Here we show that combined defi ciencies of ACAT2 and MTP reduce glycerolipid and cholesterol secretion by enterocytes without affecting uptake of free fatty acids and cholesterol in chow-fed animals.
ACAT2 defi ciency reduces acute cholesterol absorption, but this effect is not seen when absorption studies are performed over a longer time period. Previous studies have shown that cholesterol absorption, studied over a period of 72 h using dual label gavage of 3 H-cholesterol and 14 C-sitosterol, is not different in Soat2 Ϫ / Ϫ and WT mice ( 15,26 ). However, Soat2 Ϫ / Ϫ mice show signifi cantly lower fractional cholesterol absorption when fed diets that contain higher cholesterol content ( 15,26 ). These data have been interpreted to suggest that cholesterol absorption is not impaired when mice eat a low-cholesterol-containing chow diet, but is impaired when fed high cholesterol diets. In our studies, we used low-cholesterol-containing chowfed mice and studied acute cholesterol absorption by providing a bolus of cholesterol along with tracer. Under these acute conditions, Soat2 Ϫ / Ϫ mice fed a chow diet did absorb signifi cantly less cholesterol than WT mice. Similarly, signifi cant reductions in the acute absorption of cholesterol are also observed in enterocytes obtained from mice fed high cholesterol diets similar to those in longterm cholesterol absorption studies in mice fed diets high in cholesterol ( 15,26 ). Hence, we suggest that Soat2 Ϫ / Ϫ enterocytes do show signifi cant defects in the acute cholesterol absorption and these effects are signifi cantly enhanced when mice are fed high cholesterol diets. However, these effects are not apparent when studied over a period of 72 h, owing to adequate absorption of cholesterol over a longer time period. Apart from the role ACAT2 plays in cholesterol absorption, Alger et al. ( 32 ) have shown that feeding high cholesterol diets to ACAT2-deficient mice causes mild hypertriglyceridemia, prevents hepatosteatosis, and increases hepatic lipoprotein production. Here, we observed a similar phenotype in Western diet-fed Soat2 Ϫ / Ϫ mice which led to our hypothesis that increased intestinal lipid absorption in Western diet-fed mice resulting in increased delivery of fat to the liver might enhance hepatic lipoprotein production in the absence of ACAT2. However, we observed that Western diet-fed I-Mttp Ϫ / Ϫ Soa t2 Ϫ / Ϫ mice develop hypertriglyceridemia and have less hepatosteatosis in the absence of intestinal MTP simi lar to Soat2 Ϫ / Ϫ mice. Therefore, increased intestinal lipid absorption is not a cause for increased hepatic lipoprotein production in these mice. Instead, we observed that Western diet increases hepatic MTP expression in ACAT2deficient Soat2 Ϫ / Ϫ and I-Mttp Ϫ / Ϫ Soat2 Ϫ / Ϫ mice ( Fig. 6 ).
Thus, it is possible that increased hepatic MTP expression, independent of any changes in intestinal lipid absorption, could facilitate lipoprotein production, reduce hepatosteatosis, and cause hypertriglyceridemia. We speculate that increased hepatic MTP expression and lipoprotein production might be a mechanism to avoid accumulation