Up-regulation of Cholesterol Absorption Is a Mechanism for Cholecystokinin-induced Hypercholesterolemia*

Background: Intravenous injection of cholecystokinin (CCK) elevated mouse plasma cholesterol level. Results: CCK enhances cholesterol absorption in mice and cultured cells. Inhibition of CCK receptors and their downstream signaling proteins diminishes CCK-induced cholesterol absorption. Conclusion: Activation of enterocyte CCK receptors accelerates cholesterol absorption and induces hypercholesterolemia. Significance: Inhibition of CCK receptors could be a therapeutic strategy for treatment of hypercholesterolemia. Excessive absorption of intestinal cholesterol is a risk factor for atherosclerosis. This report examines the effect of cholecystokinin (CCK) on plasma cholesterol level and intestinal cholesterol absorption using the in vivo models of C57BL/6 wild-type and low density lipoprotein receptor knock-out (LDLR−/−) mice. These data were supported by in vitro studies involving mouse primary intestinal epithelial cells and human Caco-2 cells; both express CCK receptor 1 and 2 (CCK1R and CCK2R). We found that intravenous injection of [Thr28,Nle31]CCK increased plasma cholesterol levels and intestinal cholesterol absorption in both wild-type and LDLR−/− mice. Treatment of mouse primary intestinal epithelial cells with [Thr28,Nle31]CCK increased cholesterol absorption, whereas selective inhibition of CCK1R and CCK2R with antagonists attenuated CCK-induced cholesterol absorption. In Caco-2 cells, CCK enhanced CCK1R/CCK2R heterodimerization. Knockdown of both CCK1R and CCK2 or either one of them diminished CCK-induced cholesterol absorption to the same extent. CCK also increased cell surface-associated NPC1L1 (Niemann-Pick C1-like 1) transporters but did not alter their total protein expression. Inhibition or knockdown of NPC1L1 attenuated CCK-induced cholesterol absorption. CCK enhanced phosphatidylinositide 3-kinase (PI3K) and Akt phosphorylation and augmented the interaction between NPC1L1 and Rab11a (Rab-GTPase-11a), whereas knockdown of CCK receptors or inhibition of G protein βγ dimer (Gβγ) diminished CCK-induced PI3K and Akt phosphorylation. Inhibition of PI3K and Akt or knockdown of PI3K diminished CCK-induced NPC1L1-Rab11a interaction and cholesterol absorption. Knockdown of Rab11a suppressed CCK-induced NPC1L1 translocation and cholesterol absorption. These data imply that CCK enhances cholesterol absorption by activation of a pathway involving CCK1R/CCK2R, Gβγ, PI3K, Akt, Rab11a, and NPC1L.

proteins, has been suggested as a risk factor for the development of atherosclerosis (1). The apoB-carrying lipoproteins chylomicron and very low density lipoprotein (VLDL) are produced in the intestine and the liver, respectively (2). These circulating lipoproteins are partially metabolized by lipoprotein lipases, generating particles known as chylomicron remnants and low density lipoprotein (LDL). The LDL and chylomicron remnants are removed from the circulation mainly by an endocytic process mediated by hepatic LDL receptors (LDLRs) 3 and LDLR-related proteins. An increase in the generation of chylomicrons and/or VLDL and/or a decrease in the removal of LDL/ chylomicron remnants would induce hypercholesterolemia.
Absorption of cholesterol from the intestine involves multiple processes (3,4). Specifically, free cholesterol in the intestinal lumen is transported into enterocytes by the cholesterol transporter NPC1L1 (Niemann-Pick C1-like 1) (5). The internalized free cholesterol can be esterified by ACAT2 (acyl-CoA:cholesterol acyltransferase), forming esterified cholesterol, which is then transferred to chylomicrons by the microsomal triglyceride transport protein. Alternatively, free cholesterol can be exported by ATP-binding cassette A1 in the basolateral membrane and incorporated with apoAI to form high density lipoproteins (HDLs). The amount of intestinal cholesterol absorbed into the bloodstream varies from 20 to 80% among individuals (6). It has been shown that a change in the expression and/or activities of the aforementioned proteins alters the efficacy of enterocytes to absorb cholesterol (4).
A recent report from our laboratory showed that intravenous injection of [Thr 28 ,Nle 31 ]cholecystokinin (CCK) significantly elevated the plasma cholesterol level in fasting low density lipoprotein receptor knock-out (LDLR Ϫ/Ϫ ) mice (7). Ligation of the bile duct, blocking CCK receptors with proglumide, or inhibition of the Niemann-Pick C1-like 1 transporter with ezetimibe reduced the hypercholesterolemic effect of CCK on LDLR Ϫ/Ϫ mice (7). These findings suggest that CCK increased plasma cholesterol as a result of promoting reabsorption of biliary cholesterol from the intestine.
The present report examines the regulatory role of CCK on intestinal cholesterol absorption using both mouse models and cultured cells. Our data demonstrate that intravenous injection of CCK significantly enhances intestine absorption of radioactive cholesterol in both wild-type and LDLR Ϫ/Ϫ mice. Mouse primary intestinal epithelial cells (MPIECs) and human colonic adenocarcinoma Caco-2 cells express both CCK receptor 1 (CCK1R) and CCK2R. Our data suggest that CCK enhances cholesterol absorption by activation of these two isoforms of CCK receptors. Data from Caco-2 cells suggest that CCK-induced cholesterol absorption is associated with activation of the G protein ␤␥ dimer (G␤␥), phosphatidylinositide 3-kinase (PI3K), Akt, Rab-GTPase-11a (Rab11a), and NPLC1lL1. Further, inhibition or knockdown of these proteins diminishes CCK-induced transcellular cholesterol transport, suggesting a mechanistic role for them in CCK-induced cholesterol absorption.
In Vivo Cholesterol Absorption-Male wild-type (C57BL/6) and LDLR Ϫ/Ϫ mice were obtained from the Jackson Laboratory and fed a chow diet containing ϳ5% fat and 19% protein by weight (Harlan Teklad). At ϳ4 months of age, mice were starved overnight and placed into one of two different but related studies. In the first study, wild-type and LDLR Ϫ/Ϫ mice were injected with either 50 ng/kg [Thr 28 ,Nle 31 ]CCK or 50 l of phosphate-buffered saline (PBS) via the tail vein under ketamine anesthesia as described previously (7). About 20 l of blood was collected by retro-orbital bleeding before and 2 h after the injection of CCK. In the second study, wild-type and LDLR Ϫ/Ϫ mice were first injected with either 50 ng/kg [Thr 28 ,Nle 31 ]CCK or 50 l of PBS via the tail vein as above but were then gavage-fed with 0.1 Ci of [ 3 H]cholesterol (8); 2 h after the injection of CCK, ϳ0.5 ml of blood was collected via cardiac puncture under anesthesia. The mouse livers were perfused with PBS through the portal vein, and the livers were collected for determining absorbed radioactive cholesterol. All procedures were approved by the Institutional Animal Care and Use Committee at Meharry Medical College.
Plasma cholesterol was measured by spectrophotometric quantification using reagents obtained from Modern Laboratory Services (7). Plasma was fractionated by fast protein liquid chromatography (FPLC). The radioactivity in the liver, plasma, and FPLC fractions was determined by liquid scintillation counting (9). The radioactivity in various lipoproteins was calculated using the FPLC fractions, as we described previously (7,9).
Cholesterol Micelles Preparation-A mixture containing 1.4 mM sodium cholate, 1.5 mM sodium deoxycholate, 1.7 mM phosphatidylcholine, 1.9 mM monopalmitoylglycerol, 2.2 mM oleic acid, and 0.02 Ci of [ 3 H]cholesterol or 0.1 M unlabeled cholesterol was dissolved in chloroform/methanol (1:1, v/v) (10) and dried using a Labconco CentriVap concentrator. The pellet was resuspended in an equal volume of DMEM by sonication. The micellar solution was passed through a 0.20-m filter and kept at 37°C until use. The micelle stock was diluted 10 times with DMEM for treatment of cells.
Cholesterol Association and Secretion in MPIECs-MPIECs were isolated using an EDTA treatment method (8). Briefly, the jejunum and ileum obtained from wild-type mice were cleaned with a washing solution containing 117 mM NaCl, 5.4 mM KCl, 0.96 mM NaH 2 PO 4 , 26.19 mM NaHCO 3 , and 5.5 mM glucose and then filled with the washing solution supplemented with 1.5 mM EDTA and 0.5 mM dithiothreitol. The intestine was bathed in oxygenated 0.9% NaCl at 37°C for 10 min. The luminal contents were collected and centrifuged at 1,500 ϫ g for 5 min. The cell pellet was resuspended in 1 ml of DMEM supplemented with 10% FBS in a 6-well plate and cultured at 37°C for 30 min. The purity of the isolated cells was examined by flow cytometry (FACSCalibur, BD Biosciences) after the cells were sequentially stained with a primary antibody against the enterocyte marker protein SI and a FITC-conjugated second antibody. About 86% of the cells were SI-positive. Trypan blue staining repeatedly demonstrated that ϳ90% of the isolated cells remain viable, and the viability was not changed within 6 h (data not shown).
To study the effect of CCK on cholesterol association, MPIECs were treated for 10 min with 10 nM [Thr 28 ,Nle 31 ]CCK with or without 1 M lorglumide or L365260. Thereafter, 100 l of a 0.02-Ci [ 3 H]cholesterol micellar solution was added to the culture and incubated for an additional 5, 10, 15, 30, or 60 min.
Radioactivity associated with the cells was determined by liquid scintillation counting. To study the effect of CCK on cholesterol secretion, MPIECs were first incubated with the [ 3 H]cholesterol micellar solution for 1 h as described above. After removal of the [ 3 H]cholesterol micellar solution, cells were incubated in fresh DMEM for an additional 2 h in the presence or absence of CCK with or without lorglumide or L365260 as described above. The radioactivity in the enterocyte-conditioned medium was counted for determination of cholesterol secretion. The [ 3 H]cholesterol micellar solution was generated following the method described by Iqbal et al. (10).
Caco-2 Cell Cultures-Caco-2 cells were seeded in 6.5-and 24-mm transwell inserts and cultured in 24-and 6-well plates in DMEM supplemented with 20% FBS for 21 days. The medium in the apical and basolateral compartments was then replaced, respectively, with 10% FBS and 0.5% bovine serum albumin (BSA), unless otherwise stated.
Transcellular Cholesterol Transport in Caco-2 Cells-Caco-2 cells grown in transwell inserts were cultured in 24-well plates. Transcellular cholesterol transport was determined as described by Iqbal et al. (10), with brief modifications. Specifically, the apical medium was replaced with 200 l of 0.002-Ci [ 3 H]cholesterol micelles solution supplemented with 10% FBS. The basolateral medium was replaced with 1.2 ml of 5% BSA with or without 10 nM CCK, 100 nM A71623, or 100 pM pentagastrin. In experiments using CCK antagonists or G protein ␤␥ dimer, PI3K and Akt inhibitors, 50 nM lorglumide, 200 nM L365260, 10 M gallein (11), 20 M LY294002, or 600 nM Akt inhibitor XI was added into the basolateral medium. After an 18-h incubation, the basolateral-conditioned medium was centrifuged at 13,000 ϫ g for 10 min. The supernatant was filtrated through a 0.2-m PVDF membrane by using a Bio-Rad microfiltration blotting device to eliminate the cholesterol that was not associated with lipoprotein. The radioactivity on the membrane was determined by liquid scintillation analysis for determination of transcellularly transported [ 3 H]cholesterol.
Small Interfering RNA (siRNA) Knockdown-Caco-2 cells seeded in transwell inserts were transfected with scrambled siRNA or specific siRNA against CCK1R, CCK2R, or Rab11a using Fugene HD reagent and Opti-MEM medium according to the manufacturer's instructions. After 6 h, cells were replenished with fresh medium containing 10% FBS and cultured for an additional 24 h. These transfection procedures were repeated one more time (12). In the experiments for determination of protein levels, transfected cells grown in 24-mm transwell inserts were treated with 10 nM CCK or DMEM alone (control) in the basolateral compartment for time periods as indicated in the figure legends and harvested for Western blot analysis. In the experiments for determination of cholesterol absorption, transfected cells grown in 6.5-mm transwell inserts were incubated for 18 h with 0.002 Ci of [ 3 H]cholesterol micelles in the apical compartment and 10 nM CCK or DMEM alone as a control. The radioactivity in the basolateral medium was determined by liquid scintillation analysis.
Immunoprecipitation and Biotin Precipitation-Caco-2 cells grown in transwell inserts were treated with 10 nM CCK or DMEM (control) in the basolateral compartment for 1 h. In the experiments involving PI3K and Akt inhibitors, cells were incu-bated with 10 M gallein, 20 M LY294002, or 600 nM Akt inhibitor XI for 30 min prior to CCK treatment. For immunoprecipitation, the cells were lysed with a lysis buffer containing 50 mM Tris HCl (pH 8), 150 mM NaCl, 1% Triton, 2 mM EDTA, and a proteinase inhibitor mixture. A 20-l aliquot of the cell lysate was taken for input control in Western blot analysis. The rest of the lysate was incubated with protein A/G Plus-agarose beats on ice for 1 h (13). The agarose was removed by centrifugation at 2,000 ϫ g for 5 min. Precleaned lysates were incubated with antibodies against Rab11a, CCK1R, or CCK2R at 4°C for 2 h and then incubated with Protein A/G Plus-agarose beads at 4°C for 14 h. The pelleted beads were washed once with lysis buffer and then three times with washing buffer (lysis buffer minus Triton) and resuspended in Western blot loading buffer. The agarose beads were discarded by centrifugation at 10,000 ϫ g for 1 min.
For biotin precipitation of membrane proteins, the cells treated with CCK with or without LY294002 or Akt inhibitor XI were incubated with the Sulfo-NHS-Biotin reagent in the apical side at 4°C for 2 h (14). After the cells were lysed, a 20-l aliquot of lysate was taken for an input control in a Western blot analysis. The rest of the lysate was incubated with Streptavidinagarose at 4°C for 45 min. The pelleted beads were washed once with lysis buffer and then three times with washing buffer and resuspended in Western loading buffer. The agarose beads were discarded by centrifugation at 10,000 ϫ g for 1 min. The biotin precipitants were analyzed by Western blots with NPC1L1, SI, or CYP1a1 antibodies. SI is highly expressed in the apical membrane and used as an apical marker protein in Caco-2 cells (15), whereas CYP1a1 is expressed mainly in the cytosol and endoplasmic reticulum (16).
Western Blot Analysis-Caco-2 cells were treated as described for the immunoprecipitation and biotin precipitation studies. Cells were lysed in M-PER mammalian protein extraction reagent. Cell lysates as well as the immunoprecipitated and biotin-precipitated proteins were resolved on 10% SDS-polyacrylamide gels. The proteins were transferred to a PVDF membrane. After blocking with 5% fat-free milk, the membranes were incubated with antibodies as indicated in the figure legends. Immunoreactive bands were visualized using ECLplus chemiluminescence reagent (GE Healthcare) and analyzed with a GS-700 Imaging Densitometer (Bio-Rad) (16).
Quantitative Real-time RT-PCR Assay-Caco-2 cells grown in a transwell insert were treated with 10 nM CCK or culture medium alone in the basolateral compartment for the time periods indicated in the figure legends. Total RNA was extracted using TRIzol reagent and subjected to reverse transcription using a high capacity cDNA reverse transcription kit (Applied Biosystems). The resulting cDNAs were subjected to quantitative real-time PCR with an iCycler system (Bio-Bad) using primers synthesized by Qiagen for amplification of NPC1L1, CCK1R, CCK2R, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (16). The specific primers used for amplification of these genes are presented in Table 1.
The expression level of CCK1R and CCK2R mRNAs was determined as described by Jin et al. (17) with a slight modification. Briefly, standard CCK1R and CCK2R cDNA fragments were synthesized using the primers described above. The PCR products generated using Caco-2 cell cDNA as a template included 242 and 224 bp of CCK1R and CCK2R cDNAs, respectively, whereas the PCR products from MPE cDNA included 97 and 135 bp of CCK1R and CCK2R cDNAs, respectively. The molar concentration of the cDNA products was calculated on the basis of their absorbance values and molecular weights. The cDNA products were serially diluted to prepare standard cDNA solutions and subjected to PCR amplification as per test samples (17). Standard curves for each standard cDNA were constructed by plotting the cDNA concentration (fM) versus the PCR threshold cycle (C t ) (i.e. the fraction cycle number at which a PCR product was first detected). The concentration of the mRNA in the test samples was determined by fitting the measured C t to the standard curve. The level of the CCK1R and CCK2R mRNAs in Caco-2 cells and MPIECs was expressed relative to the total cellular protein level.
Statistical Analysis-Data are reported as the mean Ϯ S.E. Differences between treatment and control groups were analyzed by Student's unpaired t test or analysis of variance followed by Tukey's post hoc test. Statistical significance was considered when p was less than 0.05. Statistix software was used for statistical analyses.

CCK Elevates Plasma Cholesterol Levels and Accelerates
Intestinal Cholesterol Absorption in Mice-Plasma CCK concentrations have been reported to be 1.5-10 pM in mice (18 -20) and 1.2-26 pM in humans at the fasting state (21,22) and elevate from 50% to 400% after a meal. We previously reported that the fasting plasma CCK concentrations were 10.7 Ϯ 0.9 and 11.7 Ϯ 1.2 pM in 4-month-old male wild-type and LDLR Ϫ/Ϫ mice, respectively (7). The CCK concentrations in these mice were elevated ϳ30% at 30 min after injection of 50 ng/kg [Thr 28 ,Nle 31 ]CCK and returned to the basal level 2 h after injection (7). We also observed that the plasma cholesterol levels in CCK-treated wild-type and LDLR Ϫ/Ϫ mice were ϳ13 and 31% higher, respectively, than their vehicle-treated controls (7). Similarly, data from this report show that the plasma cholesterol level was elevated by ϳ21 and 37% in wild-type and LDLR Ϫ/Ϫ mice, respectively, after an intravenous injection of 50 ng/kg [Thr 28 ,Nle 31 ]CCK (Fig. 1A). These results suggest that CCK is able to elevate plasma cholesterol in mouse models, and the hypercholesterolemic activity of CCK is more prominent in LDLR Ϫ/Ϫ mice than in wild-type mice. To study the mechanism responsible for an apparent CCK-induced hypercholesterolemia, we examined the impact of CCK on intestine cholesterol absorption by measuring the plasma and liver radioactivity following a [ 3 H]cholesterol tracer dose fed by gavage. Data in Fig. 1, B and C, show that the radioactivity in the plasma was significantly higher in CCK-treated mice than in untreated control mice. This suggests that the CCK-induced hypercholesterolemia results, at least in part, from an increase in intestinal cholesterol absorption. In addition, the plasma radioactivity was significantly higher in LDLR Ϫ/Ϫ mice than in wild-type mice, whereas the accumulated radioactivity in the liver was significantly higher in wild-type mice than in LDLR Ϫ/Ϫ mice under both CCK-treated and control conditions. Because the liver is the primary organ for removal of LDL and chylomicron remnants from the plasma, a reduced clearance of remnant lipoproteins by the liver might account for the stronger hypercholesterolemic effect of CCK in LDLR Ϫ/Ϫ mice than in wild-type mice.
The plasma cholesterol in wild-type mice is distributed mostly in HDL, whereas knockout of LDLR mainly elevates LDL and VLDL cholesterol (7,23). The data in Fig. 1, D and E, illustrate that the plasma radioactivity was distributed mainly to the HDL fraction in wild-type mice and to the LDL/VLDL fractions in LDLR Ϫ/Ϫ mice. These findings are consistent with the distribution of cholesterol in lipoproteins in wild-type mice and LDLR Ϫ/Ϫ mice. Fig. 1, D and E, also shows that CCK increases radioactive cholesterol in all of the lipoprotein fractions. Specifically, intravenous injection of 50 ng/kg  [Thr 28 ,Nle 31 ]CCK to WT mice increased the VLDL, LDL, and HDL radioactivity by 65, 95, and 104 dpm/ml of plasma, respectively (Fig. 1D). The same dose of CCK increased VLDL, LDL, and HDL radioactivity by 85, 210, and 58 dpm/ml plasma, respectively, in LDLR Ϫ/Ϫ mice (Fig. 1E). These data suggest that the greater hypercholesterolemic effect of CCK in LDLR Ϫ/Ϫ mice than in wild-type mice results mainly from accumulation of absorbed cholesterol in the LDL fraction. CCK Enhances Transcellular Cholesterol Transport in Caco-2 Cells by Activation of Both CCK1R and CCK2R-The Caco-2 cell line has been used as a tissue culture model for intestinal absorption (5,24). These cells spontaneously form polarized cell monolayers in culture that exhibit many of the morphological and functional characteristics of normal enterocytes, including the production and polarized secretion of lipoproteins synthesized from absorbed lipids (24). We observed that Caco-2 cells expressed CCK1R and CCK2R, with a greater expression of CCK2R. Specifically, the amounts of CCK1R and CCK2R mRNAs were 2.3 Ϯ 0.3 and 18.2 Ϯ 1.4 fM/g of cellular proteins, respectively. The protein level of CCK2R was also ϳ1.9-fold higher than CCK1R in Caco-2 cells. The data in Fig.  2, A and B, showed that CCK1R and CCK2R can be co-immunoprecipitated by antibodies against their isoforms and that CCK treatment increases the amount of the co-immunoprecipitated receptor. For example, the levels of CCK2R protein precipitated by CCK1R antibody and the CCK1R protein precipitated by CCK2R antibody were ϳ2.3and 2.5-fold higher, respectively, in cells treated with CCK for 2 min than in those without CCK treatment. Prolongation of CCK treatment to 15 min slightly elevated the amount of precipitated CCK receptors; however, the difference of the precipitated proteins between 2 and 15 min did not reach statistical significance. The immunoblotting images of the input controls in Fig. 2A illustrate that CCK treatment did not affect the cellular expression level of these receptors. These results suggest that CCK1R and CCK2R form heterodimers in Caco-2 cells and that binding with CCK enhances their heterodimerization.
This study determined the impact of CCK on cholesterol absorption using Caco-2 cells as a model of the intestinal barrier. The dose-response curve in Fig. 2C shows that the transcellular transport of [ 3 H]cholesterol was increased ϳ46% when [Thr 28 ,Nle 31 ]CCK was raised from 1 to 10 nM, with no further rise at higher CCK concentrations. It has been suggested that intestinal cholesterol is absorbed by formation of both chylomicrons and HDL (8). We next studied the effect of CCK receptor agonists and antagonists on cholesterol absorption in Caco-2 cells. Fig. 2D shows that the CCK A71623 and pentagastrin mimicked the effect of CCK with regard to the stimulation of cholesterol absorption. Pentagastrin is ϳ968-fold selective for CCK2R over CCK1R (25), whereas A71623 is 1200-fold selective for CCK1R over CCK2R (26). The data in Fig. 2D also indicate that inhibition of CCK receptors by the CCK1R-preferring antagonist lorglumide or the CCK2R-preferring antagonist L365260 diminished CCK-induced cholesterol absorption. These findings were confirmed by knockdown of CCK receptors. Fig. 2, E and F, illustrates that transfection of Caco-2 cells with both CCK1R and CCK2R siRNAs or either one of them reduced CCK1R and/or CCK2R proteins by ϳ50 -60%.   MAY 9, 2014 • VOLUME 289 • NUMBER 19 JOURNAL OF BIOLOGICAL CHEMISTRY 12993 fected with scrambled control siRNA but did not induce further decrease in cholesterol absorption in cells transfected with CCK1R and/or CCK2R siRNAs. These findings imply that activations of CCK receptors and NPC1L1 are two essential steps in a pathway through which CCK induces cholesterol absorption. Inhibition of either of these two steps could diminish CCKinduced cholesterol absorption; however, simultaneous inhibition of both of them cannot cause greater suppressive effect than inhibition of one of them alone.

CCK Up-regulates Cholesterol Absorption
CCK Increases Membrane NPC1L1 in Caco-2 Cells-It is known that NPC1L1 is primarily located in recycling endosomes in the basal state (27,28). Translocation of the NPC1L1 to the apical surface of enterocytes is a critical step in cholesterol absorption. Fig. 3, A and B, data show that treatment of Caco-2 cells with 10 nM [Thr 28 ,Nle 31 ]CCK did not alter NPC1L1 total cellular protein and mRNA levels. However, this dose of CCK increased the membrane-associated NPC1L1 protein levels by ϳ2.1and 2.3-fold at 30 and 120 min post-treatment, respectively (Fig. 3, B and C). In this study, cell surface membrane proteins were isolated by biotin precipitation. The purity of the isolated proteins was confirmed by the absence of the cytoplasmic protein CYP1a1 (16) and the presence of apical marker protein SI (15) in the precipitants (Fig. 3B).
This study determined the impact of underexpressing NPC1L1 on CCK-induced cholesterol absorption. Fig. 3, D and E, data show that transfection of Caco-2 cells with NPC1L1 siRNA reduced total NPC1L1 proteins by ϳ55 and 47%, respectively, in cells treated with or without CCK. CCK treatment did not alter the total NPC1L1 protein level in cells transfected with scrambled or NPC1L1 siRNA. In contrast, CCK significantly increased membrane NPC1L1 in these two lines of cells, whereas the protein levels of membrane NPC1L1 under the control conditions and after CCK treatment were ϳ41 and 46% less, respectively, in cells transfected with NPC1L1 siRNA than in those transfected with scrambled siRNA (Fig. 3, D and E). Furthermore, CCK significantly increased cholesterol absorption in these two lines of cells, whereas the levels of absorbed cholesterol under the control conditions and after CCK treatment also were ϳ33 and 28% less, respectively, in cells transfected with NPC1L1 siRNA than in those transfected with scrambled siRNA (Fig. 3F). These data suggest that the cholesterol absorption efficiency in Caco-2 cells is positively related to the NPC1L1 protein level in the cell surface. This suggests that CCK increases cholesterol absorption by promoting NPC1L1 translocation to the cell surface.
CCK Induces PI3K and Akt Phosphorylation in Caco-2 Cells-It has been suggested that the PI3K-Akt pathway regulates the recycling of vesicles to the cell surface (29). This experiment studied the impact of CCK on PI3K and Akt expression and phosphorylation. As shown in Fig. 4, A and B, CCK treatment did not alter the levels of total PI3K and Akt proteins but increased PI3K and Akt phosphorylation in a time-dependent manner. Specifically, CCK-induced PI3K phosphorylation started at 2 min post-treatment, reached a peak at 15 min, started to decline at 30 min, and returned to the basal level within 60 min. CCK-induced Akt phosphorylation occurred at 5 min after treatment, started to decline at ϳ60 min, and returned to the basal level within 120 min. Knockdown of CCK1R or CCK2R by siRNAs or inhibition of G protein ␤␥ dimer (G␤␥) with gallein diminished CCK-induced PI3K and Akt phosphorylation (Fig. 4, C-G). It has been established that activation of CCK receptors triggers the PI3K-Akt pathway via the G protein ␤␥ dimer (G␤␥). Data in Fig. 4, C and D, indicate that inhibition of G␤␥ with gallein attenuated CCK-induced PI3K and Akt phosphorylation. In addition, inhibition of PI3K and Akt by LY294002 and Akt inhibitor XI, respectively, diminished the CCK-induced NPC1L1 translocation (Fig. 4H) and cholesterol absorption (Fig. 4I). These results indicate that CCK enhances PI3K and Akt phosphorylation by activation of G␤␥ and both CCK1R and CCK2R and that activation of the PI3K-Akt pathway is a mechanism by which CCK induces NPC1L1 translocation and cholesterol absorption.
CCK Increases Rab11a-NPC1L1 Interactions in Caco-2 Cells-Translocation of NPC1L1 to the cell surface is mediated by a protein complex containing Rab11 and its effectors (30,31). Activation of PI3K-Akt has been suggested as a means to increase the capability of Rab11 to drive recycling vesicle translocation to the cell surface (32). Here we studied the impact of CCK on the physical interaction of Rab11a with NPC1L1 in Caco-2 cells. The data in Fig. 5, A and B, indicate that NPC1L1 is co-immunoprecipitated with Rab11a. CCK treatment did not alter the expression levels of Rab11a and NPC1L1 proteins but increased the co-precipitated NPC1L1 by ϳ2.2-fold. Knock- down of CCK1R or CCK2R reduced the amount of NPC1L1 co-precipitated with Rab11a in CCK-treated cells (Fig. 5, A and  B). Fig. 5C shows that inhibition of PI3K or Akt by LY294002 and Akt inhibitor XI diminished CCK-induced Rab11a-NPC1L1 interaction. These results suggest that CCK enhances the physical interaction of Rab11a and NPC1L1 via activation of CCK1R/CCK2R, PI3K, and Akt.
Having demonstrated the interaction of Rab11a and NPC1L1, we then studied the regulatory role of Rab11a in CCKinduced NPLC1 translocation and cholesterol absorption by knockdown of Rab11a expression. Fig. 5, D and E, shows an ϳ56% reduction in Rab11a protein in cells transfected with Rab11a siRNA as compared with the cells transfected with scrambled control siRNAs. This level of Rab11a underexpression attenuated CCK-induced cholesterol absorption (Fig. 5F) and NPC1L1 translocation (Fig. 5, G and H). These suggest that Rab11a mediates NPC1L1 translocation to the cell surface (30,31). Here we also confirmed the regulatory effect of PI3K on CCK-induced NPC1L1 translocation and cholesterol absorption by knockdown of PI3K. We observed that transfection of Caco-2 cells with siRNA specific for PI3K regulatory subunit p85␣ reduced the protein level of this gene by ϳ62% (Fig. 5, D  and E). This level of PI3K underexpression did not significantly alter the basal level of NPC1L1 translocation and cholesterol absorption but decreased CCK-induced NPC1L1 translocation  and cholesterol absorption by ϳ67 and 75%, respectively (Fig. 5, E and F). This agrees with the finding in Fig. 4, H and I, that inhibition of PI3K by LY294002 diminished CCK-induced NPC1L1 translocation and cholesterol absorption.
CCK Enhances Cholesterol Association and Secretion and Induces PI3K and Akt Phosphorylation in MPIECs-Because of the difficulty of directly measuring transcellular cholesterol transport in a cellular suspension of MPIECs, this report studied the effect of CCK on cholesterol association with and secretion from these cells. Fig. 6A shows that cell-associated radioactivity increased with time. In the absence of CCK, cellassociated radioactivity reached a maximum value at 30 min after incubation with [ 3 H]cholesterol and remained at a plateau value thereafter. CCK treatment did not alter the maximum value of cell-associated radioactivity but accelerated the association of MPIECs with cholesterol. Specifically, CCK-treated cells achieved their maximum radioactivity association within 20 min and manifested a greater association at 10 -20 min after incubation with [ 3 H]cholesterol compared with untreated control cells (Fig. 6A). Because the level of cell-associated radioactivity is a result of the equilibrium between cell uptake and secretion of [ 3 H]cholesterol, cholesterol uptake was dominant at the beginning of the incubation time, whereas cholesterol secretion increased as cholesterol accumulated within the cells. When cholesterol secretion came into balance with uptake, cell-associated radioactivity reached a maximum value and remained on a plateau. Thus, the ability of CCK to shorten the time period for achievement of the maximal radioactivity association and elevate the level of cell-associated radioactivity at the early incubation time periods suggests an up-regulatory effect of CCK on MPIEC cholesterol uptake. Pretreatment of MPIECs with the CCK receptor antagonist lorglumide or L365260 diminished the up-regulatory effect of CCK on cholesterol association in MPIECs (Fig. 6B). It has been reported that lorglumide is ϳ2,300-fold more selective for CCK1R over CCK2R, whereas L365260 is ϳ140-fold selective for CCK2R over CCK1R (25). Our data thus suggest that CCK enhances cholesterol association by activation of both CCK1R and CCK2R.
Secretion of internalized cholesterol across the basolateral membrane is another critical step of cholesterol absorption (8). The data in Fig. 6C show that treatment of MPIECs with 10 nM [Thr 28 ,Nle 31 ]CCK elevated cholesterol secretion by 33% and that pretreatment of MPIECs with lorglumide or L365260 significantly diminished CCK-elevated cholesterol secretion in MPIECs. These data suggest that activation of CCK1R and CCK2R also enhances cholesterol secretion.
Similar to the results seen in Caco-2 cells, CCK treatment also induced a time-dependent increase in PI3K and Akt phosphorylation in MPIECs. Specifically, CCK-induced PI3K and Akt phosphorylation started at 2 and 5 min, respectively, and returned to the basal level within 90 and 120 min post-treatment, respectively (Fig. 6, D and E).

DISCUSSION
Results from our previous (7) and current studies are consistent in that they both demonstrate that intravenous injection of CCK elevates plasma cholesterol in wild-type and LDLR Ϫ/Ϫ mice and that the response is greater in LDLR Ϫ/Ϫ mice. The greater hypercholesterolemic effect of CCK in LDLR Ϫ/Ϫ mice appears to be due to a decreased clearance of plasma cholesterol by the liver. Specifically, we observed that a [ 3 H]cholesterol feeding-induced accumulation of radioactivity was higher in the plasma but lower in the liver in LDLR Ϫ/Ϫ mice than in wild-type mice; this was true under control and CCK treatment conditions. It has been known that absorbed cholesterol is removed from the plasma mainly by the liver and that the rate of hepatic LDL cholesterol clearance in wild-type mice (ϳ500 ml/day/kg) is ϳ20-fold greater in LDLR Ϫ/Ϫ mice (ϳ25 ml/day/ kg) (33). Thus, it is highly likely that the high rate of hepatic LDL cholesterol clearance in wild-type mice forestalls accumulation of absorbed cholesterol in the plasma and therefore prevents hypercholesterolemia. It is worthwhile to note that the activity of human liver to clear remnant lipoprotein cholesterol is relatively low. Specifically, the rate of hepatic remnant lipoprotein clearance in humans (12 ml/day/kg) is ϳ40-fold less than in wild-type mice (33). In contrast, the rate of hepatic clearance of remnant lipoprotein cholesterol in LDLR Ϫ/Ϫ mice is only ϳ2-fold greater than in humans (33). It is therefore rational to postulate that CCK could induce an even greater rise in hypercholesterolemia in human subjects than in LDLR Ϫ/Ϫ mice, and the data obtained from LDLR Ϫ/Ϫ mice would be more physiologically relevant than those from wild-type mice. It has been reported that human plasma cholesterol and CCK concentrations as well as intestinal cholesterol absorption rates elevate with age (34 -36). Therefore, it will be interesting to determine whether an increase in plasma CCK is at least in part responsible for age-related hypercholesterolemia.
Enterocyte-mediated cholesterol uptake and secretion are essential steps for cholesterol transport across the intestinal epithelial cell barrier (8). This report demonstrates that mouse enterocytes express both CCK1R and CCK2R. Treatment of mouse enterocytes with CCK enhanced cholesterol uptake and secretion. Blockade of CCK1R or CCK2R with their antagonists attenuated the up-regulatory role of CCK on cholesterol uptake and secretion in these cells. These findings suggest that direct stimulation of enterocytes is a mechanism by which CCK enhances intestinal cholesterol absorption and that both CCK1R and CCK2R contribute to the regulatory activity of CCK in enterocyte cholesterol absorption. This concept is supported by the experimental data from Caco-2 cells, which also express both CCK1R and CCK2R. Treatment of Caco-2 cells with CCK and agonists selective for CCK1R or CCK2R enhanced transcellular cholesterol transport. Blockade of CCK1R or CCK2R with antagonists or knockdown of these receptors with siRNAs diminished CCK-induced transcellular cholesterol transport in Caco-2 cells. In addition, we observed that CCK stimulation induced CCK1R/CCK2R heterodimerization. It has been reported that CCK1R and CCK2R are able to form homo-and heterodimers (37,38). Heterodimerization of these receptors enhances agonist-induced activities, such as cellular signaling and cell proliferation (38). It is highly likely that binding of CCK with CCK1R and/or CCK2R triggers heterodimerization of these receptors, activates the downstream signaling pathways, and enhances cholesterol absorption. Thus, underexpression of either CCK1R or CCK2R reduces the formation of the heterodimers and diminishes CCK-induced cholesterol absorption, whereas underexpression of both CCK1R and CCK2R will not induce greater suppressive effects than underexpression of either one of them alone.
This report also demonstrates that CCK increased cell surface NPC1L1 but did not alter the total NPC1L1 protein levels in Caco-2 cells. Under steady state conditions, NPC1L1 is located predominantly in transferrin-rich recycling vesicles in the perinuclear region (27,28). Absorption of cholesterol requires translocation of NPC1L1 to the cell membrane, a process mediated by a protein complex containing Rab-GTPase-11 (Rab11) and its partner Rab11 family of interacting protein-2 (FIP2) (30,31) or other Rab11-interacting proteins (RIPs), such as Rip11 and Rab-coupling protein (32). The data from this report document the regulatory role of Rab11a in CCK-induced NPC1L1 translocation and cholesterol absorption. Specifically, we observed that CCK enhanced the interaction between NPC1L1 and Rab11a and that knockdown of Rab11a inhibited CCK-induced NPC1L1 surface translocation and transcellular cholesterol transport in Caco-2 cells. These data suggest that CCK stimulation enhances the physical interaction of Rab11a and NPC1L1, resulting in NPC1L1 translocation and cholesterol absorption.
Two CCK receptor subtypes have been identified in mammalian cells (i.e. CCK1R and CCK2R) (39). Both of them are G-protein-coupled receptors. Stimulation of GPCRs, such as CCK receptors, is able to trigger a PI3K-dependent pathway via activation of G␤␥ (40). In this pathway, the G␤␥ dimer activates PI3K, which phosphorylates phosphoinositides to form phosphatidylinositol 3-phosphate and other phosphorylated phosphoinositides. These lipids bind both Akt and phosphoinositide-dependent protein kinase 1 (PDK1). This co-localization allows PDK1 to phosphorylate Akt (41). Indeed, data from this report demonstrate that activation of either CCK1R or CCK2R was able to activate the PI3K-Akt pathway (i.e. treatment of Caco-2 cells with CCK augmented PI3K and Akt phosphorylation, and knockdown of CCK1R or CCK2R or inhibition of the G␤␥ dimer weakened CCK-induced PI3K and Akt phosphorylation).
Activation of the PI3K-Akt pathway has been suggested as a mechanism for activation of Rab11 complex and the sequential cell surface translocation of recycling vesicles (29). For instance, activation of PI3K-Akt pathway by insulin has been shown to induce phosphorylation of Rab GTPase-activating proteins (GAPs). The Akt-phosphorylated GAPs dissociate with the Rab11-RIP protein complex. This dissociation enables the GTP-bound form of Rab11 to drive the recycling vesicle translocation to the cell membrane (32). Insulin also has been shown to induce NPC1L1 translocation from recycling vesicles to the cell surface (28). The data from this report demonstrate that inhibition of PI3K or Akt diminished CCK-induced Rab11a-NPC1L1 interaction and NPC1L1 translocation and reduced CCK-induced transcellular cholesterol transport. These data suggest that activation of a PI3K-Akt pathway is a mechanism by which CCK activates Rab11a and induces NPC1L1 translocation to the cell surface, leading to cholesterol absorption. The data in this report also showed that the NPC1L1 protein level in the cell membrane continued to remain at a high level after 2 h of CCK treatment, whereas PI3K and Akt phosphorylation had returned to the basal level at this time point. Currently, the mechanism for triggering NPC1L1 relocation to the storage compartment from the cell surface remains unclear.
In summary, this report demonstrates that CCK increased cholesterol absorption in vitro and in vivo. Treatment of Caco-2 cells with CCK also induced NPC1L1 translocation, Rab11a-NPC1L1 interaction, and PI3K and Akt phosphorylation. Inhibition of CCK1R/CCK2R, Rab11a, PI3K, or Akt diminished CCK-induced NPC1L1 translocation and cholesterol absorption. In addition, CCK treatment enhanced CCK1R/CCK2R heterodimerization in Caco-2 cells. Based on the findings from the current work and the known functions of the above mentioned molecules (29,32,40,41), we propose the following model to describe how CCK up-regulates cholesterol absorption (Fig. 7). CCK1R and CCK2R form heterodimers in response to ligand binding. Such a receptor complex triggers a pathway involving G␤␥, PI3K, phosphatidylinositol 3-phosphate, PDK1, Akt, GAPs, and RIPs, which enhances the interaction of Rab11a with NPC1L1 and induces NPC1L1 translocation to the cell surface, leading to an increase in cholesterol absorption. Further studies are required to confirm the involvement of PDK1, GAPs, and RIPs in this CCK-activated pathway.
It is known that activation of G-protein-coupled receptors is able to activate other signaling pathways, such as the phospho-lipase C-dependent pathway (42) and cyclic AMP-dependent pathway (43). Activation of CCK receptors has been shown to regulate cellular function through these pathways (43,44). Further studies are required in order to understand if these pathways also contribute to CCK-induced cholesterol absorption.