NPC1L1-dependent transport of 27-alkyne cholesterol in intestinal epithelial cells

Abstract

Niemann-Pick C1 Like-1 (NPC1L1) mediates the uptake of micellar cholesterol by intestinal epithelial cells and is the molecular target of the cholesterol-lowering drug ezetimibe (EZE). The detailed mechanisms responsible for intracellular shuttling of micellar cholesterol are not fully understood due to the lack of a suitable NPC1L1 substrate that can be traced by fluorescence imaging and biochemical methods. 27-Alkyne cholesterol has been previously shown to serve as a substrate for different cellular processes similar to native cholesterol. However, it is not known whether alkyne cholesterol is absorbed via an NPC1L1-dependent pathway. We aimed to determine whether alkyne cholesterol is a substrate for NPC1L1 in intestinal cells. Human intestinal epithelial Caco2 cells were incubated with micelles containing alkyne cholesterol in the presence or absence of EZE. Small intestinal closed loops in C57BL/6J mice were injected with micelles containing alkyne cholesterol with or without EZE. Alkyne cholesterol esterification in Caco2 cells was significantly inhibited by EZE and by inhibitor of clathrin-mediated endocytosis Pitstop 2. The esterification was similarly reduced by inhibitors of the acyl-CoA cholesterol acyltransferase (ACAT). Alkyne cholesterol efficiently labeled the apical membrane of Caco2 cells and the amount retained on the membrane was significantly increased by EZE as judged by accessibility to exogenous cholesterol oxidase. In mouse small intestine, the presence of EZE reduced total alkyne cholesterol uptake by ∼75%. These data show that alkyne cholesterol acts as a substrate for NPC1L1 and may serve as a nonradioactive tracer to measure cholesterol absorption in both in vitro and in vivo models.

INTRODUCTION

Intestinal cholesterol absorption is a major source of cholesterol in the body and is important for the maintenance of cholesterol homeostasis (1, 2). Previous studies have demonstrated a positive correlation between the efficiency of cholesterol absorption and the level of plasma cholesterol (3, 4). Also, the pharmacological inhibition of intestinal cholesterol absorption by ezetimibe efficiently reduces plasma low-density lipoprotein (LDL) cholesterol (5–7). The combination of ezetimibe with other cholesterol-lowering drugs such as statins, which inhibit cholesterol biosynthesis, was more effective as compared to statin alone in reducing plasma cholesterol and decreasing the risk of cardiovascular diseases (8, 9). Despite the advent of several therapeutics, decreasing plasma cholesterol in patients with a high risk of developing cardiovascular diseases to stringent low levels in accordance with current guidelines remains challenging (6, 10). Therefore, additional novel approaches for a further effective decrease in plasma cholesterol are needed (6, 10). In this regard, ezetimibe inhibits cholesterol absorption only by ∼54% (7), and the search for more effective inhibitors is warranted. It is crucial, therefore, to develop novel advanced methods to unravel additional therapeutic targets to inhibit intestinal cholesterol absorption.

Cholesterol absorption is a complex process in which the Niemann-Pick-C1 Like 1 (NPC1L1) cholesterol transporter, the molecular target for ezetimibe, plays a key role (11–13). NPC1L1 mediates the initial step of cholesterol entry into epithelial cells from the intestinal lumen by the transfer of cholesterol from the luminal membrane to endosomes of the endocytic recycling compartment (ERC) (1, 14). Cholesterol is then shuttled to the endoplasmic reticulum to be esterified by the Acyl-CoA cholesterol acyltransferase (ACAT2) and packaged into chylomicrons (1, 15). The molecular mechanisms by which cholesterol is endocytosed from the luminal membrane to the ERC have been extensively investigated (1). Recent studies described the role of a protein complex formed by the interaction between NPC1L1 and the clathrin adaptor proteins NUMB and AP2 to facilitate the movement of clathrin-coated vesicles along actin filaments to deliver cholesterol to the ERC (14). However, many other aspects of the cellular itinerary of cholesterol are not yet known. For example, the molecular basis for transferring the absorbed cholesterol from the ERC to the ER remains unknown (15). Notably, discerning the mechanisms involved in the distribution of the absorbed cholesterol into different organelles remains challenging, as it requires advanced methods to trace the absorbed molecules using biochemical and complex imaging techniques. Such an experimental approach may not be currently available. For example, although cellular cholesterol can be fluorescently labeled by filipin (16), this approach does not differentiate between the endogenously synthesized cholesterol from that exogenously absorbed via NPC1L1. It should be noted that the appropriate tracer for cholesterol absorption must be a substrate for NPC1L1. The fluorescently labeled NBD-cholesterol was shown to be absorbed from the intestine independent of NPC1L1, limiting its value as an appropriate tracer for cholesterol absorption (17). One potential candidate tracer is 27-alkyne cholesterol, which can be linked to a fluorophore via a click chemistry-based reaction for subsequent detection by imaging and/or biochemical methods (18). Indeed, previous studies have shown that alkyne cholesterol could serve as a substrate for a number of cellular processes including cholesterol esterification (18). However, how alkyne cholesterol is absorbed and handled by intestinal epithelial cells has not been investigated.

In the current studies, we aimed to examine the mechanism of 27-alkyne cholesterol uptake by intestinal epithelial cells. Our results showed that the uptake and esterification of 27-alkyne cholesterol in intestinal epithelial cells occurs via an NPC1L1-dependent pathway in both cell culture and animal models. These findings indicate that alkyne cholesterol represents a novel tool to investigate the molecular pathways responsible for cholesterol absorption using in vitro and in vivo models.

MATERIALS AND METHODS

Cell Culture and Materials

Human colon cancer cells (Caco2) were purchased from ATCC (Manasses, VA) and grown in plastic flasks at 37°C in an atmosphere consisting of 5% CO₂ and 95% air. Cells were cultured in Eagle’s minimum essential medium (EMEM) containing 4 mM l-glutamine, 4.5 g/L glucose, 1 mM sodium pyruvate, and 1.5 g/L sodium bicarbonate, supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 µg/mL streptomycin (Gibco, Waltham, MA). For experiments on plastic plates, Caco2 cells were seeded on six-well plates and grown to ∼90% confluence. For experiments performed in Transwells, Caco2 cells were plated at 50,000 cells/well in six-well plates and allowed to differentiate for 15 days. All chemicals and reagents were purchased from Sigma-Aldrich unless otherwise specified.

Incubation of Caco2 Cells with Micelles Containing Alkyne Cholesterol

Micelles were prepared similarly to previously described (19, 20) with 10 mM l-α-phosphatidylcholine, 0.3 mM oleic acid, 5 mM taurocholate, and 10 µM 27-alkyne cholesterol (either synthesized in our laboratory or purchased from Avanti Polar Lipids, Alabaster, AL). Briefly, lipids dissolved in ethanol were mixed in a glass vial and the solvent was evaporated under nitrogen. EMEM containing 5% lipoprotein-deficient serum (LPDS) was added to the dried lipids and incubated on a magnetic stirrer overnight at room temperature. Cells were pretreated for 1 h at 37°C with 5% LPDS in EMEM with either 50 µM ezetimibe, 25 µM Pitstop 2, or vehicle. Cells were then incubated at 37°C with alkyne cholesterol-containing micelles supplemented with 50 µM ezetimibe, 25 µM Pitstop 2, or vehicle for 2–6 h. Cells were washed with ice-cold PBS, scraped in 0.4 mL PBS, and transferred to glass vials. Lipids were extracted as previously described (21). Briefly, the cells were vortexed sequentially with 1 mL methanol, 0.5 mL chloroform (twice), and 1 mL water. The vials were centrifuged at 1,700 g for 10 min and the lower (organic) phase was collected. Lipids were dried under nitrogen gas.

Measurement of NPC1L1-Dependent Alkyne Cholesterol Uptake and Esterification

Dried lipids were reconstituted in 7-µL chloroform and subjected to a click reaction with 3-azido-7-hydroxycoumarin as previously described (18). Briefly, a click reaction mixture comprised of CuBF4 (250 µL of 10 mM in acetonitrile), 3-azido-7-hydroxycoumarin (5 µL of 22.25 mM in ethanol; Jena Bioscience, Jena, Germany), and ethanol (1 mL). Thirty microliter of the reaction mixture was added to the reconstituted lipids. As a control, 30 µL of the reaction mixture was added to a separate tube containing 1 µg alkyne cholesterol in 7 µL chloroform. The reaction was incubated on a heating block at 55°C for 4 h. The mixture was centrifuged, and the full volume was loaded on 20-cm silica thin layer chromatography (TLC) plates. TLC plates were developed in chloroform:methanol:water:acetic acid (65:25:4:1) for 4–5 cm, dried, and developed further in hexanes:ethyl acetate (1:1) to the top of the plate. After drying, the plate was briefly covered in 4% N,N-diisopropylethylamine in hexanes and dried. Fluorescence images (excitation: 460–490 nm, emission: 518–546 nm) were captured using a ChemiDoc MP imager (BioRad, Des Plaines, IL), and image analysis was performed in Fiji ImageJ (22).

Oxidation of Alkyne Cholesterol and MβCD Treatment

Alkyne cholestenone standards were generated by oxidation of alkyne cholesterol as previously described (18). Briefly, cholesterol oxidase was dissolved in 50 mM phosphate buffer (KH₂PO₄, pH 7.5) with 1 mg/mL lipid-free BSA. In glass vials, 2 ng of cholesterol oxidase was incubated with 800 ng of alkyne cholesterol in a total volume of 100 µL of phosphate buffer supplemented with 0.1% Triton X-100. The reaction mixture was vortexed at 1,100 rpm for 10 min at room temperature. The reaction was stopped by diluting with a 3:1 mixture of chloroform and methanol. The mixture was then centrifuged at 1,700 g for 10 min and the lower (organic) phase was collected and dried under nitrogen gas.

For oxidation experiments performed in Caco2 cells, micelles were prepared as above except for containing 0.2 mM unlabeled cholesterol instead of alkyne cholesterol. Caco2 cells were incubated with 5 µM alkyne cholesterol in EMEM containing 1% FBS for 18 h to label membranes as previously described (19). Cells were pretreated for 1 h at 37°C with 5% LPDS in EMEM supplemented with 50 µM ezetimibe, 25 µM Pitstop 2, or vehicle. Cells were then incubated for 4 h at 37°C with unlabeled cholesterol-containing micelles supplemented with either 50 µM ezetimibe or vehicle. Cells were washed and fixed with 1% glutaraldehyde on ice for 30 min. Cells were then incubated with 7 U/mL cholesterol oxidase in PBS (pH 7.4) for 5 min at room temperature, washed with ice-cold PBS, and processed as described in Incubation of Caco2 Cells with Micelles Containing Alkyne Cholesterol.

For alkyne cholesterol oxidation over a time series, Caco2 cell membranes were prelabeled with alkyne cholesterol as described above. Cells were treated with 10 µM methyl-β-cyclodextrin (MβCD) to deplete the membranes of cholesterol. Cells were washed and incubated with 7 U/mL cholesterol oxidase in PBS (pH 7.4) for 0, 1, 2, 10, 30, and 120 min at room temperature and processed as above.

Uptake of 27-Alkyne Cholesterol by Small Intestinal Closed Loops

All animal experiments were approved by the IACUC at the Jesse Brown VA Medical Center. 27-Alkyne cholesterol micelles were prepared as above using PBS. Three male and three female 8–20-wk-old C57BL/6J mice (Jackson Laboratory, Bar Harbor, ME) were anesthetized with isoflurane (2.5% induction → 1.5% maintenance), the abdomen surgically opened, and the small intestine exposed. Two segments of jejunum, 1–2 cm in length, were closed by tying with suture. Each closed loop was injected with ∼300 µL of solution containing 10 µM 27-alkyne cholesterol micelles supplemented with either 50 µM ezetimibe or vehicle, alternating between animals which proximal and distal loop received each treatment. The abdomen was closed by sutures and the mice were maintained under anesthesia for 3 h, at which point they were euthanized, and the intestine harvested and rinsed thoroughly. Approximately half of each loop was snap-frozen in an optimal cutting temperature (OCT) compound for staining. Intestinal epithelial cells (IECs) were isolated from the remainder of the loops as previously described (23). Lipids were extracted from IECs and processed as above.

Immunostaining of Small Intestinal Closed Loops

Five micrometer-thick cryostat sections were cut from the snap-frozen intestinal tissues in OCT and were mounted on the microscopic slides. Sections were fixed in 4% paraformaldehyde (PFA) for 15 min by placing the slides in a dark chamber. Tissue sections were washed with 1× PBS for 5 min followed by washing with the click reaction buffer (100 mM HEPES/KOH PH 7.4) three times for 10 min each as previously described (18). Sections were incubated with 500 µL of prewarmed (43°C) click reaction buffer (100 mM HEPES/KOH pH 7.4) containing 50 µM of BODIPY FL azide (Lumiprobe, Hunt Valley, MD). The click reaction was initiated by addition of 10 µL of 10 mM CuBF₄ in acetonitrile and incubated for 15 min at 43°C on a heating block. An additional 5 µL of 10 mM CuBF₄ was added to the reaction and incubated further for 15 min. Sections were washed with the click reaction buffer two times for 10 min each and were permeabilized with 0.5% NP-40 for 5 min. Blocking was performed with 1% normal goat serum at room temperature. Tissues were incubated with monoclonal mouse anti-villin antibody (1:100 dilution; Thermo Fisher, Waltham, MA; Cat No. MA5-12227, RRID: AB_10980388) in a humid chamber followed by incubation with Alexa Fluor 568-conjugated goat anti-mouse IgG (Thermo Fisher, Waltham, MA; Cat No. A-11004, RRID: AB_2534072). Anti-villin antibodies were validated using Western blot by Thermo Fisher demonstrating expression of the expected ∼92-kDa protein in epithelial cell lines but not in other cell lines. Slides were washed and mounted with antifade mounting media with DAPI by putting one drop on each section and were viewed using a confocal microscope (Zeiss LSM 510 Meta).

RESULTS

Esterification of Exogenous 27-Alkyne Cholesterol in Intestinal Epithelial Cells

Absorbed cholesterol is delivered to the ER for esterification via a NPC1L1-dependent pathway in intestinal epithelial cells (14, 19). The esterification of cholesterol was previously shown to be highly sensitive to the NPC1L1 inhibitor ezetimibe, and the ezetimibe-sensitive esterification of the absorbed cholesterol was considered as a reliable method to assess NPC1L1 function (19, 20). We first investigated if model intestinal epithelial Caco2 cells process 27-alkyne cholesterol in the same manner as natural cholesterol when provided to the cells in the form of micelles. Micelles containing 10 µM of 27-alkyne cholesterol were prepared as previously described (20) and added to the luminal side of Caco2 cells for 4 h in the presence or absence of 50 µM ezetimibe. Cells were then harvested for total lipid extraction. The alkyne cholesterol in the lipid extracts was then subjected to Cu-mediated cycloaddition reaction (click reaction) in the presence of azido-fluorophore (3-azido-7-hydroxycoumarin) as previously described (18). The lipids were then resolved by TLC and imaged by a fluorometric scanner. The band for alkyne cholesterol was clearly detected in the standard sample containing only alkyne cholesterol (Fig. 1A). Similar bands were detected representing free alkyne cholesterol in each sample of total lipid extracts from Caco2 cells. A second band on the TLC was also detected that corresponds to alkyne cholesterol esters. As shown in the figure, the band for alkyne cholesterol esters was diminished in the presence of ezetimibe. Densitometric analysis showed that the rate of alkyne cholesterol esterification was significantly reduced by ∼90% by ezetimibe in Caco2 cells.

Figure 1.

Esterification of 27-alkyne cholesterol in intestinal epithelial cells is ezetimibe (EZE)-sensitive and time-dependent. Caco2 cells were incubated with mixed micelles containing 10 µM 27-alkyne cholesterol in the absence (CT) or presence of 50 µM ezetimibe (EZE). Lipids were extracted, subjected to copper-catalyzed cycloaddition reaction with 3-azido-7-hydroxycoumarin, resolved by thin layer chromatography, and imaged by fluorometric scanner. Controls of click reactions without alkyne cholesterol (Rxn) or with alkyne cholesterol (ST) were added to the plate as standards. A: a representative fluorescence image and densitometric analysis from Caco2 cells grown on Transwells and incubated for 4 h with mixed micelles. The relative rate of cholesterol esterification was measured as a ratio of alkyne cholesterol ester (aCE) to the sum of alkyne cholesterol ester and alkyne cholesterol (aChol) detected on TLC. Data are presented as means ± SE from three values obtained on separate occasions; **P < 0.01 compared with control. B and C: a representative fluorescence image (B) and relative rate of alkyne cholesterol esterification (C) from Caco2 cells grown on plastic plates and incubated for 2–6 h with mixed micelles. Data are presented as mean ± SE from three values obtained on separate occasions; n.s: not significant; **P < 0.01, ***P < 0.001 compared with 2 h; †††P < 0.001, ††††P < 0.0001 compared with control. Caco2, colon cancer cells; TLC, thin layer chromatography.

We also examined the time course for the esterification of micellar alkyne cholesterol in Caco2 cells. As shown in Fig. 1B, when Caco2 cells were incubated with 27-alkyne cholesterol-containing micelles for different periods (2, 4, and 6 h), there was an apparent increase in the levels of alkyne cholesterol esters. It is important to note that while there are additional bands detected in this image, they also appear in the reaction mixture control lane that does not contain any lipids and thus likely represent byproducts of the reaction. Figure 1C shows that the rate of micellar alkyne cholesterol esterification in intestinal epithelial Caco2 cells was time-dependent and was significantly inhibited by ezetimibe at all time points. These results strongly suggest that micellar 27-alkyne cholesterol is a substrate for intestinal NPC1L1 that is transferred to the ER for esterification similar to previously used radiolabeled natural cholesterol (19, 20).

ACAT-2 Inhibitors Block the Esterification of Micellar 27-Alkyne Cholesterol

Ezetimibe blocks the delivery of micellar cholesterol to the ER reducing its availability for ACAT2 in intestinal epithelial cells (24). We examined if the direct inhibition of ACAT2 in intestinal epithelial cells affects the esterification of micellar alkyne cholesterol. Caco2 cells were incubated with micelles containing 27-alkyne cholesterol alone or with 25 µM of either Sandoz-58-035 (25) or avasimibe (26), inhibitors of ACAT2. As depicted in Fig. 2A, the alkyne cholesterol esterification was largely reduced in the presence of ACAT2 inhibitors. The data presented in Fig. 2B demonstrate that the direct inhibition of ACAT2 significantly reduced the rate of alkyne cholesterol esterification to the same extent as achieved by ezetimibe. These data confirmed that micellar alkyne cholesterol transferred to the ER by NPC1L1-dependent pathway is esterified by ACAT2 in intestinal epithelial cells.

Figure 2.

ACAT inhibitors abolish esterification of 27-alkyne cholesterol. Caco2 cells grown on plastic plates were incubated for 4 h with micelles containing alkyne cholesterol (10 µM) in the absence (CT) or presence of 25 µM Sandoz 58-035 (San) or 25 µM Avasimibe (Ava). A: a representative fluorescence image from three separate experiments. B: the relative rate of alkyne cholesterol esterification as determined by densitometric analysis. Data are presented as means ± SE from three values obtained on separate occasions; ****P < 0.0001 compared with control. ACAT, acyl-CoA cholesterol acyltransferase.

The Esterification of the Micellar 27-Alkyne Cholesterol Is Reduced by the Inhibition of Endocytosis

Since endocytic internalization was shown to be critical for cholesterol absorption by NPC1L1 (14, 19), we next examined the effect of blocking endocytosis on the esterification of micellar 27-alkyne cholesterol. Endocytosis was inhibited by incubating the cells with 25 µM Pitstop 2, which has been previously used to block NPC1L1 endocytosis (27). The cells were then exposed to micelles containing 27-alkyne cholesterol in the presence or absence of ezetimibe. As shown in Fig. 3A, the esterification of 27-alkyne cholesterol was abolished as expected in the presence of ezetimibe. Interestingly, 27-alkyne cholesterol esters were diminished by Pitstop 2, suggesting that the inhibition of endocytosis remarkably reduced NPC1L1-mediated uptake of 27-alkyne cholesterol. Figure 3B shows that the extent of inhibiting the esterification of 27-alkyne cholesterol by Pitstop 2 was similar to that of ezetimibe. These observations strongly suggest that the uptake of 27-alkyne cholesterol and trafficking to the ER for esterification occurs in the same manner as previously described for natural cholesterol (14, 19).

Figure 3.

Inhibition of clathrin-mediated endocytosis abolishes esterification of 27-alkyne cholesterol. Caco2 cells grown on plastic plates were incubated for 4 h with micelles containing alkyne cholesterol (10 µM) in the absence (CT) or presence of 50 µM ezetimibe (EZE) or 25 µM Pitstop 2. A: a representative fluorescence image from three separate experiments. B: the relative rate of alkyne cholesterol esterification as determined by densitometric analysis. Data are presented as means ± SE from five values obtained on separate occasions; ****P < 0.0001 compared with control. Caco2, colon cancer cells.

The Accessibility of 27-Alkyne Cholesterol Incorporated in the Apical Plasma Membrane of Intestinal Epithelial Cells to Cholesterol Oxidase

Cholesterol incorporated into plasma membranes is susceptible to exogenously added cholesterol oxidase to intact fixed cells (28). We next investigated the accessibility of 27-alkyne cholesterol that is incorporated into the apical membrane of Caco2 cells to exogenously added cholesterol oxidase. Well-differentiated monolayers of Caco2 cells were first incubated for 18 h with 27-alkyne cholesterol to label cellular and plasma membrane. The cells were then treated for 1 h with or without methyl-β-cyclodextrin (MβCD) to sequester cholesterol from the plasma membrane. The monolayers were then fixed with glutaraldehyde and incubated for different periods with cholesterol oxidase that was added to the apical membrane of fixed cells. Total lipids were then extracted, alkyne cholesterol was fluorescently labeled by click chemistry reaction, and lipids were resolved on TLC plates followed by fluorography. Figure 4A shows that the two bands for alkyne cholestenone and free alkyne cholesterol were detected as judged by their corresponding standards. The amount of alkyne cholesterol on the plasma membrane of Caco2 cells accessible to cholesterol oxidase was measured as a ratio of alkyne cholestenone to total alkyne cholesterol. Figure 4B shows that the oxidation of apical membrane alkyne cholesterol was time-dependent and was significantly decreased by treatment with MβCD. These data clearly indicate that 27-alkyne cholesterol is suitable as a surrogate to track apical membrane cholesterol in intestinal epithelial cells.

Figure 4.

Alkyne cholesterol in the plasma membrane is accessible to cholesterol oxidase. Caco2 cells were incubated for 18 h with free alkyne cholesterol (5 µM) in cell culture media, followed by 1 h incubation with 10 µM methyl-β-cyclodextrin (MβCD). Cells were then fixed with glutaraldehyde and incubated with 7 U/mL of cholesterol oxidase for 0, 1, 2, 10, 30, and 120 min at room temperature. Alkyne cholestenone standard (OxST) was generated via in vitro oxidation of alkyne cholesterol. A: a representative fluorescence image after 30 min incubation with cholesterol oxidase. B: the rate of alkyne cholesterol oxidation was measured as a ratio of alkyne cholestenone to the sum of alkyne cholesterol and cholestenone detected on TLC. Data are presented as means ± SE. ***P < 0.001 compared with control. TLC, thin layer chromatography.

We next examined the effects of ezetimibe and the inhibition of endocytosis on the pool of alkyne cholesterol that is incorporated in the apical membrane and is accessible to the exogenous cholesterol oxidase. The membranes of Caco2 cells were first labeled with alkyne cholesterol and then the cells were incubated with micelles containing natural cholesterol to stimulate NPC1L1-dependent endocytosis. The experiments were performed in the absence or presence of ezetimibe. The surface cholesterol was then oxidized by the addition of cholesterol oxidase to the apical membrane. Both free alkyne cholesterol and alkyne cholestenone were visualized by fluorography. As shown in Fig. 5A, the levels of alkyne cholestenone were higher in the presence of ezetimibe. Figure 5B shows that the ratio of alkyne cholestenone to alkyne cholesterol was significantly increased in the presence of ezetimibe, indicating that the endocytosis of alkyne cholesterol incorporated in the plasma membrane was blocked when NPC1L1 was inhibited.

Figure 5.

Ezetimibe (EZE) treatment increases plasma membrane alkyne cholesterol that is accessible to cholesterol oxidase. Caco2 cells were incubated for 18 h with free alkyne cholesterol (5 µM) in cell culture media, followed by 4 h incubation with mixed micelles containing unlabeled cholesterol (0.2 mM) in the absence (CT) or presence of 50 µM EZE. Cells were fixed and treated with cholesterol oxidase for 5 min. A: a representative fluorescence image from three separate experiments. B: alkyne cholesterol oxidation as determined by densitometric analysis, normalized to control. Data are presented as means ± SE. ****P < 0.0001 compared with control. Caco2, colon cancer cells.

The Uptake of Micellar 27 Alkyne Cholesterol in Mouse Intestinal Closed Loops

We next investigated the uptake of 27-alkyne cholesterol by native mouse intestinal epithelial cells. Two intestinal closed loops were prepared in each animal while under anesthesia: one loop was injected with a solution containing 27-alkyne cholesterol micelles without ezetimibe, and the other was injected with the 27-alkyne cholesterol micelles with 50 µM ezetimibe. The intestinal loops were harvested, and tissue blocks were made with OCT freezing media. Tissue sections were then subjected to click chemistry reaction as previously described (18) to fluorescently label the micellar alkyne cholesterol that was absorbed by the intestine. Confocal images shown in Fig. 6A indicate that the uptake of micellar alkyne cholesterol was fluorescently detected in epithelial cells and was reduced by ezetimibe. Quantification revealed a 52 ± 9% decrease in alkyne cholesterol fluorescence (normalized to villin fluorescence) with ezetimibe treatment (P < 0.01; n = 6 images). Additionally, intestinal epithelial cells were isolated from intestinal loops, and lipids were extracted, labeled with 3-azido-7-hydroxycoumarin, and imaged as described above. Quantification of the fluorescence intensity of the alkyne cholesterol band in Fig. 6B clearly shows that the uptake of micellar alkyne cholesterol was significantly inhibited in the presence of ezetimibe. These observations strongly suggest that the uptake of alkyne cholesterol in the native mouse intestine is mediated by NPC1L1 and is sensitive to ezetimibe.

Figure 6.

Uptake of alkyne cholesterol in the mouse intestine. Jejunal closed loops were injected with mixed micelles containing alkyne cholesterol (10 µM) in the absence (CT) or presence of 50 µM ezetimibe (EZE), maintained for 3 h, and harvested. A: intestinal tissue was fixed and stained by copper-dependent click reaction using BODIPY FL azide (green), anti-villin (red), and DAPI (blue). A representative confocal image from one of six animals. Scale bars: 100 μm. B: lipids were extracted from isolated intestinal epithelial cells, subjected to copper-catalyzed cycloaddition reaction with 3-azido-7-hydroxycoumarin, resolved by thin layer chromatography, and imaged by fluorometric scanner. Fluorescence intensity of the band representing alkyne cholesterol was determined by densitometric analysis. Data are presented as paired measurements from each animal. *P < 0.05 compared with control.

DISCUSSION

Our data provide strong evidence that alkyne cholesterol is a substrate for NPC1L1 in intestinal epithelial cells. We also demonstrated that the esterification of micellar alkyne cholesterol was abolished by ezetimibe or by ACAT2 inhibitors. We further showed that the esterification of micellar alkyne cholesterol and its levels on the apical membrane were sensitive to ezetimibe and the inhibition of endocytosis.

The structure of 27-alkyne cholesterol closely resembles that of natural cholesterol with only slight modification at the side chain of the molecule (18). Alkyne cholesterol was previously shown to support the survival and proliferation of an auxotrophic yeast strain suggesting that it suffices for essential functional roles that are usually played by natural cholesterol at the cellular level (18). Consistent with this work, our studies revealed no detrimental effect of 27-alkyne cholesterol on Caco2 cellular processes or viability. In line with this notion, the studies by Hofmann et al. (18) also showed that alkyne cholesterol is a suitable substrate for many of the enzymes that are involved in cholesterol modification and metabolism including for an example the acyltransferases and oxidases. However, cholesterol is also transferred between different organs and tissues by processes mediated via transport proteins and receptors (1). Our main objective in the current study was to investigate the uptake and processing of alkyne cholesterol by intestinal epithelial cells. The epithelial cells in the intestine are unique in their ability to acquire cholesterol from an additional pool that is not available to any other cell type in the body, namely, the cholesterol that is present in mixed micelles in the intestinal lumen (29). Therefore, it is of particular importance to investigate the uptake of alkyne cholesterol by intestinal epithelial cells to determine its suitability as an analogue to study intestinal cholesterol absorption.

Cholesterol absorption is a complex process involving multiple steps and key luminal and cellular factors (29). For example, it was established that the cholesterol transporter NPC1L1 expressed on the brush border membrane of epithelial cells plays a central role in cholesterol absorption. Indeed, ezetimibe, the cholesterol-lowering drug, was shown to block cholesterol absorption by targeting NPC1L1 (11). Therefore, for an analogue to be suitable to study cholesterol absorption, its uptake by intestinal epithelial cells should be mediated by an ezetimibe-sensitive, NPC1L1-dependent pathway.

To evaluate the uptake of alkyne cholesterol, we utilized the intestinal epithelial Caco2 cell line to model the intestinal epithelium (30). Previous studies from our laboratory and others supported the suitability of Caco2 cells to investigate the cellular pathways responsible for cholesterol absorption (20). Using Caco2 cells as a model, previous studies from our laboratory and others showed that the esterification of micellar radioactive cholesterol in Caco2 cells was remarkably sensitive to ezetimibe (19, 20). The data presented in this manuscript also showed that alkyne cholesterol esterification was also greatly inhibited by ezetimibe, indicating that alkyne cholesterol follows the same cellular pathway that transfers natural cholesterol to the ER for esterification by ACAT2. Inhibition and genetic deletion of ACAT2 were shown to block the esterification of micellar cholesterol and inhibit intestinal cholesterol absorption (4, 31). We have used two known ACAT1/2 inhibitors, Sandoz 058-08 and avasimibe (25, 26), to block ACAT2 in Caco2 cells. Our data showing that the direct inhibition of ACAT2 abolishes the esterification of micellar cholesterol confirms that this analogue serves as a substrate for this enzyme. Since the esterification of micellar cholesterol depends on trafficking mechanisms (19), our findings supported the notion that the esterification of alkyne cholesterol was also reduced to a great extent by Pitstop 2, an inhibitor of cellular endocytosis. This small molecule was developed as an inhibitor of clathrin-mediated endocytosis (32), though it blocks clathrin-independent pathways as well (33) and should not be used to delineate the endocytic pathway(s) responsible for NPC1L1-dependent cholesterol absorption. Collectively, our current data strongly support that alkyne cholesterol is recognized by NPC1L1, endocytosed, and transferred to the ER for esterification similar to natural cholesterol. Therefore, these data validate the use of 27-alkyne cholesterol as a surrogate for cholesterol absorption.

Radioactive cholesterol was used previously to label plasma membranes and assess its accessibility to exogenous cholesterol oxidase (34). Previous studies in Caco2 cells showed that blocking NPC1L1-dependent endocytosis by ezetimibe significantly increased the level of radioactive cholesterol accessible to exogenous cholesterol oxidase in the apical membranes (19). The results of our studies also showed that the plasma membrane of well-differentiated, polarized Caco2 cells could be labeled by alkyne cholesterol and that alkyne cholesterol in the apical membranes of these cells is also accessible to exogenously added cholesterol oxidase. Our results clearly showed that oxidized alkyne cholesterol was decreased after removing cholesterol from the membrane by MβCD. Further, our data with alkyne cholesterol corroborated previous results showing that the pool of plasma cholesterol accessible to the oxidase was increased by treatment with ezetimibe.

The unique advantage of alkyne cholesterol is the ability to track its cellular distribution by fluorescent microscopy after in situ attachment to an azide-reporter. Using fluorescent imaging, our data in this manuscript clearly show that the fluorescence intensity of alkyne cholesterol in native mouse intestinal tissues was significantly inhibited by ezetimibe. This observation supports the superiority of alkyne cholesterol over the fluorescent NBD-cholesterol, which is not a substrate for NPC1L1, to be used as a surrogate for studying cholesterol absorption (17). Although other sterols with auto-fluorescent activity such as dehydroergosterol (DHE) could be used as alternatives, the intensity of the fluorescent signal from these compounds is relatively weak and requires sensitive imaging cameras (16). Given its structural similarity to natural cholesterol and the availability of sensitive detection methods, alkyne cholesterol remains a better choice. Nevertheless, these imaging results should be interpreted with caution as the alkyne moiety in this probe could be masked in certain cellular compartments. Careful evaluation of the imaging analysis along with concomitant biochemical approaches is necessary when using alkyne cholesterol as a surrogate to study cholesterol absorption.

It should be noted that the ability of alkyne cholesterol to recapitulate native cholesterol absorption may depend on luminal factors that have not been fully investigated in this manuscript. Luminal cholesterol is first emulsified in the presence of bile acids and phospholipids to facilitate its absorption (29). The solubilization of cholesterol depends on certain physiochemical variables such as hydrophobicity index related to different types of bile acids (35–37). For the current experiments, we have prepared micelles containing 5 µM taurocholic acid along with 10 µM of l-phosphatidylcholine as we and others have previously used with radioactively labeled cholesterol for the in vitro studies (19, 20). Our findings in this manuscript showed that the rate of esterification for micellar alkyne cholesterol by Caco2 cells was comparable to that previously measured for radioactive cholesterol. This suggests that alkyne cholesterol was efficiently solubilized in the current experimental conditions similar to radioactive cholesterol. However, future studies should focus on investigating the degree of alkyne cholesterol solubility in the presence of different types of bile acids and phospholipids to determine the physiochemical indices that may affect its emulsification and, hence, the efficiency of intestinal absorption.

Our current findings suggest that the use of click chemistry-based approaches using alkyne cholesterol is superior to other currently available methods to investigate the molecular mechanisms responsible for cholesterol absorption at the level of intestinal epithelial cells. Future studies will focus on addressing other critical questions related to alkyne cholesterol: 1) Does alkyne cholesterol serve as a substrate for other cholesterol transporters? 2) Is alkyne cholesterol transferred in the blood via lipoproteins? 3) Is alkyne cholesterol secreted in the bile similar to natural cholesterol? Nevertheless, our current studies represent a crucial proof of concept that is necessary to pave the way for future investigations to further validate the use of alkyne cholesterol as a surrogate for cholesterol absorption and metabolism utilizing experimental animal models.

GRANTS

These studies were supported by the Veterans Affairs Research Career Scientist Award No. BX005243 (to W. A. Alrefai) and a Merit Award No. BX000152 (to W. A. Alrefai). The studies are also partially supported by NIH Grants: DK109709 (to W. A. Alrefai), DK117535 (to A. L. Ticho), DK054016 (to P. K. Dudeja), DK098170 (to R. K. Gill), and VA Merit Awards: BX002011 (to P. K. Dudeja) and BX002867 (to S. Saksena). P. K. Dudeja is also supported by a VA Senior Research Career Scientist Award (BX005242).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.L.T., N.C., D.L., R.K.G., and W.A.A. conceived and designed research; A.L.T., N.C., P.M., H.L., and A.N.A. performed experiments; A.L.T., N.C., P.M., H.L., S.S., P.K.D., D.L., R.K.G., and W.A.A. analyzed data; A.L.T., N.C., S.S., P.K.D., D.L., R.K.G., and W.A.A. interpreted results of experiments; A.L.T., N.C., R.K.G., and W.A.A. prepared figures; A.L.T., N.C., R.K.G., and W.A.A. drafted manuscript; A.L.T., N.C., P.M., H.L., A.N.A., S.S., P.K.D., D.L., R.K.G., and W.A.A., edited and revised manuscript; A.L.T., N.C., P.M., H.L., A.N.A., S.S., P.K.D., D.L., R.K.G., and W.A.A. approved final version of manuscript.