Ezetimibe Inhibits Hepatic Niemann-Pick C1-Like 1 to Facilitate Macrophage Reverse Cholesterol Transport in Mice

Abstract

Objective

Controversies have arisen from recent mouse studies regarding the essential role of biliary sterol secretion in reverse cholesterol transport (RCT). The objective of this study was to examine the role of biliary cholesterol secretion in modulating macrophage RCT in Niemann-Pick C1-Like 1 (NPC1L1) liver only (L1^LivOnly) mice, an animal model that is defective in both biliary sterol secretion and intestinal sterol absorption, and to determine if NPC1L1 inhibitor ezetimibe facilitates macrophage RCT by inhibiting hepatic NPC1L1.

Approach and Results

L1^LivOnly mice were generated by crossing NPC1L1 knockout (L1-KO) mice with transgenic mice overexpressing human NPC1L1 specifically in liver. Macrophage-to-feces RCT was assayed in L1-KO and L1^LivOnly mice injected intraperitoneally with [³H]-cholesterol-labeled peritoneal macrophages isolated from C57BL/6 mice. Inhibition of biliary sterol secretion by hepatic overexpression of NPC1L1 substantially reduced transport of [³H]-cholesterol from primary peritoneal macrophages to the neutral sterol fraction in bile and feces in L1^LivOnly mice without affecting tracer excretion in the bile acid fraction. Ezetimibe treatment for 2 weeks completely restored both biliary and fecal excretion of [³H]-tracer in the neutral sterol fraction in L1^LivOnly mice. HDL kinetic studies showed that L1^LivOnly relative L1-KO mice had a significantly reduced fractional catabolic rate without altered hepatic and intestinal uptake of HDL-cholesterol ether.

Conclusions

In mice lacking intestinal cholesterol absorption, macrophage-to-feces RCT depends on efficient biliary sterol secretion and ezetimibe promotes macrophage RCT by inhibiting hepatic NPC1L1 function.

Introduction

Reverse cholesterol transport (RCT) is classically defined as the movement of cholesterol from cells in peripheral tissues to circulating high density lipoproteins (HDLs) for hepatic uptake, biliary secretion, and fecal disposal.^{1, 2} This process is believed to explain, at least in part, why increased plasma HDL-cholesterol is atheroprotective.³ Recently, an intestinal route for mass cholesterol excretion in the feces has been reported,^{4, 5} which has promoted studies of the significance of this non-biliary route in RCT using genetic and surgical mouse models deficient in biliary cholesterol secretion.^{6, 7}

One genetic model used in macrophage RCT assays is the transgenic mice specifically overexpressing human Niemann-Pick C1-Like 1 (NPC1L1) in liver.^{6, 8} Unlike mice lacking the ATP-binding cassette (ABC) transporter B4 (ABCB4) that are deficient in both biliary cholesterol/phospholipids secretion and liver cholestasis,⁹ NPC1L1 liver transgenic mice exhibit reduced biliary cholesterol secretion without showing signs of liver cholestasis. NPC1L1 is almost exclusively expressed in the small intestine of mice, and its knockout in mice blocks intestinal cholesterol absorption.¹⁰ The function of NPC1L1 can be inhibited by ezetimibe,^10–13 a potent cholesterol absorption inhibitor developed to lower blood cholesterol.¹⁴ In humans, NPC1L1 is also expressed in liver in addition to intestine.^{8, 10, 15} We have previously shown that transgenic overexpression of human NPC1L1 in the mouse liver significantly reduces biliary cholesterol secretion without altering hepatic expression of the heterodimeric hepatobiliary cholesterol exporter ABCG5 and ABCG8, and this reduction in biliary cholesterol secretion can be rescued by ezetimibe treatment.⁸

Using Rader’s in vivo RCT assay protocol,¹⁶ Temel and colleagues showed that biliary sterol secretion is not required for macrophage RCT in NPC1L1 liver transgenic mice and in mice with acute biliary diversion, two mouse models deficient in biliary sterol secretion into the gut lumen.⁶ In striking contrast with this finding, Nijstad and associates reported almost simultaneously that biliary cholesterol secretion is required for functional RCT in mice using the similar protocol.⁷ Nijstad et al. showed that bile duct ligation in mice or genetic inhibition of biliary sterol secretion in ABCB4 knockout mice dramatically reduce macrophage-to-feces RCT. Further, they showed that pharmacological stimulation of macrophage RCT by a liver X receptor agonist depends on efficient biliary sterol secretion in mice. The mechanistic basis for different conclusions in these two studies is unclear.

On average, ~50% of cholesterol in the gut lumen is absorbed in humans and rodents.^{17, 18}, and the remainder excreted in feces. Inhibiting intestinal cholesterol absorption by ezetimibe has been shown to dramatically increase macrophage RCT in wild-type mice,^{19, 20} a model that does not express NPC1L1 in liver.¹⁰ Altered biliary cholesterol secretion was reported to influence intestinal cholesterol absorption rates.^{21, 22} Acute biliary diversion or bile duct ligation reduces intestinal cholesterol absorption and profoundly alters intestinal metabolism, including increases in intestinal cholesterol synthesis.^{23, 24} To eliminate effects of cholesterol absorption changes on fecal excretion of bile-derived cholesterol, we crossed cholesterol absorption-deficient NPC1L1 knockout (L1-KO) mice¹⁰ to liver-specific NPC1L1 transgenic mice⁸ and generated mice expressing no endogenous NPC1L1, but human NPC1L1 in liver only (L1^LivOnly mice)²⁵. We have previously shown that ezetimibe treatment increases biliary sterol excretion by inhibiting hepatic NPC1L1.^{8, 25} This observation raised an interesting question: can ezetimibe facilitate macrophage RCT by inhibiting hepatic NPC1L1? L1^LivOnly mice provided us a unique opportunity to address this question. In the present study, we performed macrophage RCT assays in L1^LivOnly mice using the mouse primary peritoneal macrophages. We found that the macrophage-to-feces RCT was dramatically reduced in L1^LivOnly mice. The reduction in macrophage RCT in these animals was completely restored by ezetimibe treatment.

Results

Hepatic Overexpression of NPC1L1 Inhibits Biliary Cholesterol Secretion and Increases Cholesterol Levels in Plasma and Liver of L1-KO mice

In a recent study using L1^LivOnly mice, we found that liver-specific overexpression of human NPC1L1 in mice of NPC1L1 knockout background almost abolished biliary cholesterol secretion as evidenced by results from bile duct cannulation studies, and significantly increased plasma and hepatic cholesterol levels.²⁵ Consistently, in the present study using mice of the same genotypes, we found that overexpression of human NPC1L1 in the L1-KO liver remarkably reduced biliary cholesterol concentrations and molar ratios (Figure 1A) without significantly altering biliary concentrations and molar ratios of phospholipids (Figure 1B) and bile acids (Figure 1C). The effect of hepatic NPC1L1 on biliary cholesterol was completely reversed by ezetimibe treatment for 2 weeks (Figure 1A). Ezetimibe treatment had no effects on biliary concentrations and molar ratios of phospholipids and bile acids (Figure 1B and 1C). In contrast to previous studies using the same liver-specific NPC1L1 transgenic mice on the wild-type background,⁸ or on NPC1L1 knockout background but on a 0.2% cholesterol diet (about 10 times higher than that used in the current study), we observed a significant 35.1% reduction in fecal excretion of neutral sterols (a sum of cholesterol and its bacterial metabolites coprostanol and cholestanone) in L1^LivOnly relative to L1-KO mice on the 0.015% cholesterol diet (Figure 1D). This observation suggests the importance of the use of low cholesterol diet in delineating the role of biliary sterol secretion in fecal sterol excretion in rodents. Interestingly, ezetimibe treatment for 2 weeks in L1^LivOnly mice significantly increased fecal neutral sterol excretion by 46.1% (from 30.92 µmol/day/100 g BW in L1^LivOnly group to 45.17 µmol/day/100 g BW in L1^LivOnly&Eze group) (Figure 1D), suggesting a substantial contribution of biliary cholesterol secretion to fecal cholesterol excretion in our animals on the low (0.015%) cholesterol diet.

Figure 1.

Ezetimibe treatment restores biliary cholesterol output in L1^LivOnly mice. Lipid concentrations and molar ratios in the gallbladder bile of L1-KO mice (n = 5), L1^LivOnly mice (n = 6), and L1^LivOnly mice treated with ezetimibe (L1^LivOnly&Eze) (n = 4). (A) Biliary cholesterol. (B) Biliary phospholipids. (C) Biliary bile acids. (D) Fecal neutral sterol excretion in L1-KO (n = 8), L1^LivOnly (n = 9), and L1^LivOnly&Eze (n = 7) mice. P < 0.05 between^a and^b groups for each measurement (ANOVA).

Hepatic overexpression of human NPC1L1 in L1-KO mice on the low cholesterol diet increased plasma and hepatic levels of total cholesterol, free cholesterol and cholesterol ester (Table 1). The increase of blood cholesterol was mainly distributed in the large HDL fraction (Figure 2), consistent with our previous study showing that increased blood cholesterol is in the apolipoprotein E-rich HDL fraction.⁸ Ezetimibe treatment for 2 weeks virtually reversed all of these alterations.

Table 1.

Lipid Contents in Plasma (mg/dl) and Liver (µg/mg proteins) (Mean ± SEM)

	L1-KO	L1LivOnly	L1LivOnly&Eze
Plasma TC	209 ± 4b	274 ± 8a	222 ± 10b
Plasma FC	44 ± 3c	72 ± 2a	56 ± 4b
Plasma CE	257 ± 6b	315 ± 10a	259 ± 10b
Hepatic TC	39.0 ± 2.2b	47.6 ± 2.0a	40.5 ± 1.5b
Hepatic FC	26.5 ± 1.6a.b	31.6 ± 1.6a	25.8 ± 1.3b
Hepatic CE	21.1 ± 1.7b	26.9 ± 1.4a	24.7 ± 1.8a,b

L1-KO mice, L1^LivOnly mice, and L1^LivOnly mice treated with ezetimibe (L1^LivOnly&Eze) (n = 7–9 mice per group) were fasted for 4 h during the daytime cycle prior to collections of blood and liver for analyses of plasma and liver contents of total cholesterol (TC), free cholesterol (FC), and cholesterol ester (CE). Cholesterol ester contents were calculated by multiplying the difference between total and free cholesterol mass by 1.67.

P < 0.05 among^{a, b, c} groups for each measurement (ANOVA).

Figure 2.

Plasma lipoprotein-cholesterol profile in L1-KO mice, L1^LivOnly mice and L1^LivOnly mice treated with ezetimibe (L1^LivOnly&Eze). An equal amount of plasma from each mouse in each group (n = 4) was pooled. The same amount of pooled plasma was subjected to FPLC for monitoring lipoprotein-cholesterol distribution. VLDL, very low density lipoprotein; LDL, low density lipoprotein; HDL, high density lipoprotein.

Hepatic Overexpression of NPC1L1 in L1-KO mice Increases Plasma and Tissue [³H]-Tracer

To measure macrophage RCT, we first determined how hepatic NPC1L1 overexpression influences homeostasis and distribution of [³H]-cholesterol derived from intraperitoneally injected primary macrophages that were isolated from the peritoneal cavity of C57BL/6 mice and loaded with radiolabeled cholesterol. Plasma and tissue levels of [³H]-tracer were measured during in vivo macrophage RCT studies. Consistent with elevated plasma cholesterol in L1^LivOnly mice, plasma [³H]-tracer levels were significantly higher in L1^LivOnly relative to L1-KO mice at 6 h, 24 h and 48 h post peritoneal injection of [³H]-cholesterol-labeled primary macrophages (Figure 3A). [³H]-tracer contents were also increased in several tissues of L1^LivOnly mice, including liver, small intestine (SI), lung and spleen (Figure 3B). In this RCT experiment, we included a group of L1-KO mice treated with ezetimibe to determine whether ezetimibe alters plasma and tissue levels of [³H]-tracer during macrophage RCT assays when NPC1L1 is disrupted. Consistent with NPC1L1 being the target of ezetimibe,^{12, 13} ezetimibe treatment had no effects on plasma and tissue [³H]-tracer levels in L1-KO mice (Figure 3).

Figure 3.

Ezetimibe treatment reverses [³H]-tracer accumulation in plasma and tissues in L1^LivOnly mice injected with mouse primary macrophages. (A) Plasma [³H]-tracer recovery in L1-KO (n = 4), L1^LivOnly (n = 4), L1-KO mice treated with ezetimibe (L1-KO&Eze) (n = 4), and L1^LivOnly&Eze (n = 4) mice at different time points after peritoneal injections of [³H]-cholesterol-labeled mouse primary peritoneal macrophages. (B) [³H]-tracer recovery in liver, small intestine (SI), lung and spleen in the same mice 48 h post macrophage injection. P < 0.05 between^a and^b groups for each measurement, time point, and tissue (ANOVA).

Ezetimibe Treatment Restores Hepatic NPC1L1-Induced Inhibition of Biliary and Fecal [³H]-Cholesterol Excretion in L1^LivOnly Mice

To determine if hepatic NPC1L1 overexpression-mediated inhibition of biliary cholesterol secretion influences macrophage-to-feces RCT in mice lacking endogenous NPC1L1 (deficient in intestinal cholesterol absorption), we measured [³H]-tracer concentrations in bile and excretion in feces in our models. We have previously shown that L1^LivOnly mice are deficient in biliary secretion of cholesterol mass.²⁵ Consistently, biliary [³H]-cholesterol output was dramatically reduced by 72.7% in L1^LivOnly mice (0.84 ± 0.07 % of dose/ml) relative to L1-KO mice (3.08 ± 0.19 % of dose/ml) at 48 h post peritoneal injection of [³H]-cholesterol-labeled primary macrophages, which was almost completely restored in the ezetimibe-treated L1^LivOnly mice (3.03 ± 0.20 % of dose/ml) (Figure 4A). Ezetimibe treatment had no effects on biliary [³H]-cholesterol output in L1-KO mice (3.11 ± 0.13 % of dose/ml) (Figure 4A), consistent with L1-KO mice being ezetimibe-insensitive.¹⁰ The biliary [³H]-bile acid output was comparable among 4 groups (Figure 4B).

Figure 4.

Hepatic overexpression of NPC1L1 inhibits macrophage RCT in L1-KO mice injected with mouse primary macrophages. L1-KO (n = 4), L1^LivOnly (n = 4), L1-KO&Eze (n = 4), and L1^LivOnly&Eze (n = 4) mice were peritoneally injected with [³H]-cholesterol-laden primary macrophages isolated from C57BL/6 mice. Bile and feces were collected 48 h post injection. (A) [³H]-cholesterol recovery in the gallbladder bile. (B) [³H]-bile acid recovery in the gallbladder bile. (C) [³H]-cholesterol recovery in the feces. (D) [³H]-bile acid recovery in the feces. P < 0.05 between^a and^b groups for each measurement (ANOVA).

The fecal excretion represents the final step of RCT. We measured fecal excretion of [³H]-tracer in the neutral and acidic sterol fractions in mice after peritoneal injection of labeled mouse primary macrophages. Transgenic overexpression of human NPC1L1 in the liver of L1-KO mice dramatically reduced fecal [³H]-neutral sterol excretion by ~60% (4.04 ± 0.45 % of dose/day/100 g BW in L1^LivOnly mice vs. 10.09 ± 0.42 % of dose/day/100 g BW in L1-KO mice), which was virtually reversed by ezetimibe treatment for 2 weeks (a significant 3-fold increase to 10.13 ± 1.31 % of dose/day/100 g BW) (Figure 4C). Ezetimibe treatment didn’t alter fecal [³H]-neutral sterol excretion in L1-KO mice (9.98 ± 0.51 % of dose/day/100 g BW) (Figure 4C). Hepatic overexpression of human NPC1L1 or its inhibition by ezetimibe didn’t change fecal [³H]-bile acid excretion in L1-KO mice (Figure 4D), which was in agreement with our previous finding that the mass bile acid excretion via feces is similar between L1-KO and L1^LivOnly mice.²⁵

Hepatic Overexpression of NPC1L1 Does Not Affect Hepatic and Intestinal Uptake of [³H]Cholesteryl Oleyl Ether (CEt)-Labeled HDL in L1-KO Mice

The transport of cholesterol from peripheral tissues to HDL particles is the first step of RCT. Alterations in HDL turnover and hepatic/intestinal uptake thus have the potential to influence fecal disposal of cholesterol disposal. To determine if hepatic overexpression of NPC1L1 affects HDL turnover and tissue uptake, we injected [³H]CEt-HDL into our mice via the tail vein, and then followed the plasma decay of radioactivity. Hepatic and intestinal uptake was assessed 48h after the injection. Despite a significant reduction in the fractional catabolic rate (FCR), hepatic and intestinal uptake of [³H]CEt-HDL was not reduced in L1^LivOnly mice (Figure 5), suggesting that reduced macrophage-to-feces cholesterol transport in L1^LivOnly mice may not be a result of decreased availability of liver cholesterol for biliary disposal.

Figure 5.

Hepatic overexpression of NPC1L1 reduces HDL-cholesterol ether fractional catabolic rate (FCR) in L1-KO mice. FCR and tissue uptake of radiolabels were assessed as described under Methods in male L1-KO, L1^LivOnly, L1-KO treated with ezetimibe (L1-KO&Eze) and L1^LivOnly treated with ezetimibe (L1^LivOnly&Eze) mice. P < 0.05 (n = 3-4) between^a and^b groups for each measurement (ANOVA).

Discussion

In this study, we have shown that, after peritoneal injection of [³H]-cholesterol-labeled mouse primary peritoneal macrophages, L1-KO mice expressing hepatic NPC1L1 (L1^LivOnly) accumulated more [³H]-tracer in blood and tissues, and secreted significantly reduced amounts of [³H]-neutral sterols in gallbladder bile and feces, when compared to L1-KO mice expressing no hepatic NPC1L1. Ezetimibe treatment reversed the accumulation of [³H]-tracer in blood and tissues, and restored biliary and fecal excretion of [³H]-neutral sterols in L1^LivOnly mice. Our results demonstrated an essential role of biliary sterol secretion in mediating macrophage-to-feces RCT in mice deficient in intestinal cholesterol absorption. Given that human livers express NPC1L1,^{8, 10, 15} our findings suggest that ezetimibe may have a previously unappreciated action: promoting macrophage RCT via direct inhibition of hepatic NPC1L1 in humans.

Recent studies on mice genetically or surgically deficient in biliary cholesterol secretion have shown that fecal excretion of mass cholesterol is independent of biliary cholesterol secretion, therefore the trans-intestinal cholesterol efflux (TICE) concept was developed as an alternative explanation for these observations.^{4, 5} This concept was subsequently used to challenge the classic view on RCT and it was hypothesized that macrophage RCT may occur via the intestinal route.^{6, 26} While one study performed by Temel et al. using the liver-specific NPC1L1 transgenic mice on the wild-type background, or using bile diverted mice appears to support the non-biliary route hypothesis on RCT,⁶ the opposite conclusion was obtained from a study conducted by Nijstad and associates using ABCB4 knockout mice, scavenger receptor class B type I (SR-BI) knockout mice, mice treated with the liver × receptor, and mice that were subjected to bile duct ligation.⁷ Our present work is in agreement with the conclusion of Nijstad and associates that “biliary cholesterol secretion represents the major pathway relevant for RCT” in mice,⁷ but in contrast with the finding by Temel et al.⁶ Although the same liver-specific NPC1L1 transgenic mice were used in both Temel’s study and the present work, it should be emphasized that there was a fundamental difference between the two animal models. The present study used the liver-specific NPC1L1 transgenic mice that are deficient in intestinal cholesterol absorption as a result of NPC1L1 knockout while Temel et al. used the liver-specific NPC1L1 transgenic mice that have normal cholesterol absorption due to the presence of endogenous NPC1L1 in the intestine. On average, ~50% of cholesterol in the intestinal lumen is absorbed.¹⁸ This would suggest that ~50% of bile-derived cholesterol is re-absorbed without loss in the feces when NPC1L1 is present. In the present study, we eliminated this re-absorption factor that has been shown to dramatically inhibit macrophage RCT.^{19, 20} Thus we believe our animal model is a more reliable system relative to the one used by Temel for examining how biliary cholesterol secretion modulates in vivo RCT. The data from our model clearly demonstrated that biliary cholesterol secretion is essential for macrophage RCT, thereby supporting the classic view on RCT.^{1, 2} In addition, our data strongly argue against a role of the TICE pathway in promoting macrophage RCT, at least in our animal model.

We have previously shown that hepatic overexpression of human NPC1L1 inhibits biliary cholesterol secretion and increases plasma cholesterol without altering hepatic expression of many proteins involved in liver cholesterol homeostasis, including several cholesterol transporters (ABCG5, ABCG8, and ABCA1), HDL receptor scavenger receptor class B type I (SR-BI), and low density lipoprotein receptor (LDLR).⁸ We have also shown that hepatic NPC1L1 inhibits biliary cholesterol secretion in L1^LivOnly mice.²⁵ Consistently, the present work showed that hepatic overexpression of human NPC1L1 inhibited biliary [³H]-cholesterol excretion (Figure 4A) and raised blood/tissue [³H]-tracer concentrations (Figures 3). Ezetimibe treatment completely reversed blood/tissue [³H]-tracer accumulation in L1^LivOnly mice (Figure 3).

Ezetimibe was developed to inhibit intestinal cholesterol absorption to lower blood cholesterol.¹⁴ In our previous studies, we found that the drug also inhibits hepatic NPC1L1 function to promote biliary cholesterol secretion.^{8, 25} The present work showed that ezetimibe increases macrophage-to-feces RCT by inhibiting hepatic NPC1L1. It has been previously reported by others that ezetimibe treatment stimulates macrophage RCT in wild-type mice.^{19, 20} In those studies, the effect of ezetimibe on RCT should be considered as a result of inhibition of intestinal cholesterol absorption because mice do not normally express measurable NPC1L1 in the liver.¹⁰ However, human and monkey livers express NPC1L1.^{8, 10, 11, 15} A clinical trial [NCT00701727, Ezetimibe Reverse Cholesterol Transport (RCT) Pilot Study. Weblink: http://clinicaltrials.gov/show/NCT00701727] was initiated to examine the effect of ezetimibe on RCT in humans. Based on the results posted on the website of this trial, ezetimibe treatment at 10 mg/day for 7 weeks significantly increased fecal excretion of plasma-derived cholesterol in hypercholesterolemic humans. Our present work suggests that ezetimibe may promote RCT in humans through inhibiting NPC1L1 function in both intestine and liver.

In conclusion, biliary sterol secretion plays a key role in promoting macrophage-to-feces RCT. Ezetimibe promotes macrophage RCT via inhibiting hepatic NPC1L1 function to stimulate biliary cholesterol secretion, at least, in mice. Additionally, the data from our animal model strongly argue against a role of the TICE pathway in promoting macrophage RCT.

Materials and Methods

Mice and Diets

L1^LivOnly mice¹ were generated by crossing L1-KO mice² with liver-specific NPC1L1 transgenic mice.³ The genetic background of L1^LivOnly and control mice was 93.75% C57BL/6. All mice were housed in a specific pathogen-free animal facility in plastic cages at 22°C, with a daylight cycle from 6 AM to 6 PM. The mice were provided with water and standard chow diet (Prolab RMH 3000; LabDiet, Brentwood, MO) ad libitum, unless stated otherwise. All animal procedures were approved by the Institutional Animal Care and Use Committee at Wake Forest University Health Sciences and at University of Maryland.

At 6 weeks of age, male L1-KO and L1^LivOnly mice were fed a synthetic diet containing 10% energy from palm oil and 0.015% (w/w) cholesterol. The diet was prepared at the institutional diet core and used in our previous studies.⁸ After being fed the diet for 14 days, a subgroup of L1^LivOnly mice were switched to the same diet supplemented with 0.005% (w/w) of ezetimibe.

In Vivo Macrophage RCT Studies

The in vivo macrophage-to-feces RCT assay was performed according to the protocol developed by Rader and colleagues.⁴ Thioglycollate-elicited peritoneal macrophages isolated from adult wild-type C57BL/6 mice were cultured in DMEM medium supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 µg/ml streptomycin. The cells were loaded with [³H]-cholesterol for 24 h in 10% FBS-containing DMEM medium supplemented with a mixture of 5 µCi/ml [³H]-cholesterol and 100 µg/ml acetylated low-density lipoprotein (LDL). Acetylated LDL converted peritoneal macrophages to lipid-laden macrophage foam cells. These cells were washed twice with PBS and equilibrated for an additional 12 h period in serum-free RPMI-1640 containing 0.2% BSA. After equilibration, the cells were harvested and resuspended in serum-free DMEM containing 0.2%BSA immediately before injection. [³H]-cholesterol labeling efficiency was measured by extracting lipids from an aliquot of cells using the method of Bligh and Dyer⁵ and by separating extracted lipids on thin layer chromatography using a solvent system (70:30:1; hexane:diethyl ether:acetic acid).

The mice were individually housed on wire bottom cages with ad libitum access to food and water, and then intraperitoneally injected with 500 µl of the cell suspension (~3×10⁶ cells/ml at ~2×10⁶ dpm/ml). Feces were collected for 48 h. At 6 h and 24 h post injection, blood samples were collected via the submandibular vein. At 48 h post injection, the mice were sacrificed to collect blood, bile, liver, small intestine, lung and spleen for analysis.

[³H]-Tracer Measurements in Bile, Feces, and Tissues

Lipids in ~10 µl of gallbladder bile from each mouse were extracted by sequentially adding/vortexing 1 ml H₂O, 3 ml methanol:chloroform (2:1), 2 ml chloroform, and 1 ml H₂O in a 16×100 mm glass tube. To separate aqueous and organic phases, the tube was centrifuged at 2,700 rpm for 10 min. The upper aqueous phase contained [³H]-bile acids, and the lower organic phase contained [³H]-cholesterol. A known volume of each phase was dried under N₂ and resuspended in 5 ml scintillation cocktail (Bio-Safe II, Order#:111195, Research Product International Corp., Mount Prospect, IL) for the determination of biliary recoveries of [³H]-bile acids and [³H]-cholesterol. The data were expressed as the percentage of [³H]-tracer recovered from the total dpm injected.

Feces were collected for 48 h after injection, dried at 70°C in a vacuum oven overnight, and then weighed. The entire fecal sample from each mouse was rehydrated in 20 ml of 95% ethanol. A total of 2 ml rehydrated fecal sample was transferred into a new glass tube and saponified by adding 400 µl of 10N NaOH and heating on a 60°C heating block for 2 h. Lipids in saponified fecal sample were extracted by 6 ml hexane for 3 times. All hexane phase was dried down under N₂ and the lipid extract was resuspended in scintillation cocktail for the determination of [³H]-cholesterol recovery. The remaining saponified fecal sample in aqueous phase was acidified by adding ~200 µl concentrated HCl to adjust pH to <1, and then extracted with 6 ml hexane for 3 times. All hexane phase was dried down under N₂ and the extract was resuspended in 5 ml scintillation cocktail for the determination of fecal [³H]-bile acid recovery.

Liver, small intestine, lung, and spleen were collected from each mouse and the organ weight was recorded. The small intestine was equally separated into 5 segments. A piece of liver, each segment of small intestine, and the whole lung and spleen were placed in 16×100 mm glass tubes and extracted in 9 ml of chloroform:methanol (2:1) until the tissue sank to the bottom of the glass tube (indicative of complete extraction of lipids). After centrifugation at 2,700 rpm for 10 min, an aliquot of 5 ml of chloroform:methanol extract was dried down under N₂ and resuspended in 5 ml scintillation cocktail for the determination of [³H]-cholesterol recovery in each tissue.

Lipid Analyses in Plasma, Liver and Bile

Plasma concentrations of total cholesterol, free cholesterol, and triglyceride were analyzed by enzymatic assay as described previously.⁶ For analysis of hepatic lipid contents, the lipids were extracted from ~80 mg of liver tissues and measured enzymatically as described previously.³ Biliary concentrations of free cholesterol, phospholipids and bile acids were determined as described previously.³

Measurements of Fecal Neutral Sterol Excretion

The mice were individually housed. The feces were collected for 48 h and dried in a 70°C vacuum oven. The dried feces were weighed and crushed. About 100 mg of feces were placed into a 16×100 mm glass tube containing 100 µg of 5α-cholestane as an internal standard. The feces were saponified in 2 ml of 95% ethanol and 200 µl of 50% KOH (w/v in water) on a 65°C; heating block for 2 h. The lipids were extracted by adding 2 ml hexane and 2 ml H₂O. After centrifugation at 2,700 rpm for 10 min at room temperature, 1 ml of hexane phase was transferred to a gas chromatography vial for the determination of neutral sterols (cholesterol and its bacterial metabolites coprostanol and cholestanone) by gas-liquid chromatography.

In Vivo HDL Turnover Studies

HDL was isolated from wild-type mice and labeled with [³H]cholesteryl oleyl ether exactly as described previously.⁷ [³H]CEt-HDL was dialyzed with PBS and radioactivity was then counted. [³H]CEt-HDL solution (0.5 million of dpm) was injected into each mouse via the tail vein. After injection, blood was taken from the tail vein at 5 min, 30 min, 1 h, 3 h, 8 h, 12 h, 24 h, and 48 h. Plasma decay curves were generated by dividing the plasma radioactivity at each time point by the radioactivity at the initial 5 min time point after [³H]CEt-HDL injection as described by Nijstas E, et al.⁸ The fractional catabolic rate (FCR) was calculated from the decay curves according to the Matthews method.⁹ The organ uptake of [³H]CEt-HDL was assessed 48 h after injection and the value was expressed as a percentage of the injected dose calculated by multiplying the initial plasma counts (5 min) with the estimated plasma volume (3.5% of total body weight).⁸

Statistical Analysis

All data are presented as Mean ± SEM (Standard Error of Mean). Significance of differences was determined for each group of values by One-way ANOVA (Tukey-Kramer honestly significant difference). A P value less than 0.05 was considered significant.

Significance.

Reverse cholesterol transport (RCT) is classically viewed as cholesterol movement from peripheral tissues/cells such as macrophages to feces via biliary secretion. Recently, a trans-intestinal cholesterol efflux (TICE) pathway was proposed to promote RCT, challenging the classical view of RCT. However, published data on this topic are very controversial due to lack of an ideal animal model. Here we created a novel genetically altered mouse model to address this issue. Our model minimized many confounding factors affecting RCT assays. Data from this model strongly argue against a role of TICE, and support the classical view in RCT. Additionally, our findings are the first to demonstrate that liver NPC1L1 inhibits macrophage RCT, and that ezetimibe can inhibit liver NPC1L1 to promote macrophage RCT. Given a critical role of macrophage RCT in atheroprotection and the world-wide use of ezetimibe as a cholesterol-lowering drug, the impact of our findings on human health is substantial.

Abbreviations

ABC

ATP-binding cassette

L1-KO

NPC1L1 knockout

NPC1L1

Niemann-Pick C1-Like 1

RCT

reverse cholesterol transport

TICE

trans-intestinal cholesterol efflux