Inhibition of intestinal cholesterol absorption with ezetimibe increases components of reverse cholesterol transport in humans

Atherosclerosis 230 (2013) 322e329

Contents lists available at ScienceDirect

Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis

Inhibition of intestinal cholesterol absorption with ezetimibe increases components of reverse cholesterol transport in humans Michael H. Davidson a, Jason Voogt b, Jayraz Luchoomun b, Julie Decaris b, Salena Killion b, Drina Boban c, Alexander Glass b, Hussein Mohammad b, Yun Lu b, Deona Villegas b, Richard Neese d, Marc Hellerstein c, d, David Neff e, Thomas Musliner e, Joanne E. Tomassini e, Scott Turner b, * a

Radiant Research and The University of Chicago Pritzker School of Medicine, 515 North State Street, Suite 2700, Chicago, IL, USA KineMed, Inc., 5980 Horton Street, Suite 400, Emeryville, CA, USA c Department of Medicine, Division of Endocrinology and Metabolism, University of California at San Francisco, 505 Parnassus Avenue, Room M994, San Francisco, CA, USA d Department of Nutritional Sciences and Toxicology, University of California at Berkeley, 309 Morgan Hall, University of California at Berkeley, CA, USA e Merck & Co., Inc., One Merck Drive, Whitehouse Station, NJ, USA b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 January 2013 Received in revised form 3 August 2013 Accepted 5 August 2013 Available online 13 August 2013

Objective: Reverse cholesterol transport (RCT) can be defined as a pathway of flux of cholesterol from peripheral tissues to the liver for potential excretion into feces. This prospective, placebo-controlled, double-blind crossover study assessed the effect of ezetimibe on several RCT parameters in hyperlipidemic patients. Methods: Following 7 weeks of treatment (ezetimibe 10 mg/day or placebo), 26 patients received 24-h continuous IV infusions of [3,4-13C2]-cholesterol, then took heavy water (2H2O) by mouth. Cholesterol excretion was measured by quantification of neutral/acid sterols in stool and blood samples during 7 days post-infusion with continued treatment. Plasma de novo cholesterol synthesis was assessed by 2Hlabeling from 2H2O. Results: Ezetimibe significantly reduced levels of low-density lipoprotein cholesterol (22%, P < 0.001) without significant changes in triglycerides and high-density lipoprotein cholesterol and significantly increased the flux of plasma-derived cholesterol into fecal neutral sterols by 52% (P ¼ 0.04) without change in flux into fecal bile acids. Total fecal neutral sterol output increased by 23% (P ¼ 0.02). Plasma de novo cholesterol synthesis increased by 57% (P < 0.001). The fractional clearance rate (FCR) of plasma cholesteryl-ester trended higher (7%; P ¼ 0.055) with a reduction in absolute cholesteryl-ester production rate (9%, P < 0.01). Whole-body free cholesterol efflux rate from extra-hepatic tissues into plasma was not measurably changed by ezetimibe. Conclusion: Ezetimibe treatment approximately doubled the flux of plasma-derived cholesterol into fecal neutral sterols, in association with increases in total fecal neutral sterol excretion, FCR of plasma cholesterol ester, and plasma de novo cholesterol synthesis. These effects are consistent with increased cholesterol transport through the plasma compartment and excretion from the body, in response to ezetimibe treatment in hyperlipidemic humans. Clintrials.gov: NCT00701727. Ó 2013 Elsevier Ireland Ltd. All rights reserved.

Keywords: Ezetimibe Reverse cholesterol transport Cholesterol efflux Cholesterol absorption Fecal sterols

1. Introduction

* Corresponding author. KineMed, Inc., 5980 Horton Street, Emeryville, CA 94720, USA. Tel.: þ1 510 655 6525x101. E-mail address: [email protected] (S. Turner). 0021-9150/$ e see front matter Ó 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2013.08.006

Ezetimibe is an inhibitor of intestinal cholesterol absorption acting through inhibition of the Niemann-Pick C1-like1 (NPC1L1) transporter [1]. In humans, once-daily treatment with ezetimibe monotherapy (10 mg) inhibits cholesterol absorption on average by 54e65% and results in about a 20% reduction of plasma low-density lipoprotein cholesterol (LDL-C) [2,3]. A similar incremental

M.H. Davidson et al. / Atherosclerosis 230 (2013) 322e329

reduction in LDL-C with ezetimibe treatment is observed when coadministered with other cholesterol-lowering therapies, including statins, bile acid sequestrants, and fenofibrate, but not dietary plant sterols [4]. The LDL-C lowering effect of ezetimibe treatment is usually attributed primarily to increased catabolism of LDL-C, presumably due to the up-regulation of the LDL-receptor, in response to reduced cholesterol delivered to the liver via the chylomicron pathway [3,5,6]. Studies in mice have suggested that ezetimibe may also augment parameters of reverse cholesterol transport (RCT) [7,8]. Although increased macrophage-to-feces cholesterol transport has been demonstrated with ezetimibe in animals [7e9], the role of this effect in contributing to a reduction in atherosclerosis as reported in mouse and rabbit models is uncertain [10]. In humans, ezetimibe alone and combined with simvastatin significantly decreased fractional cholesterol absorption compared with placebo (P < 0.001) [2,3] and also significantly increased fecal sterol excretion, while simvastatin had no effect on absorption rates and slightly decreased fecal sterol excretion [3]. Treatment of a sitosterolemic subject with ezetimibe is reported to have resulted in complete regression of xanthomatosis [11]. These findings raise the hypothesis that ezetimibe may reduce the cholesterol content of peripheral tissues (including the vascular wall) by increasing RCT as well as lowering plasma LDL-C levels. Inhibition of NPC1L1 reduces the intestinal absorption of cholesterol, both endogenous and exogenous [1,12]. In humans, only about 25e35% of intestinal cholesterol is derived from diet, while the remaining w2/3 originates from the liver in bile, from sloughed intestinal cells, and perhaps an additional direct transintestinal cholesterol efflux (TICE) route [13,14]. Biliary cholesterol can be derived from at least three distinct sources: hepatic de novo cholesterol synthesis, hepatic cholesterol stores, or cholesterol

323

cleared from plasma lipoproteins. If only hepatic-derived biliary cholesterol were responsible for the increased excretion of cholesterol seen with ezetimibe treatment, it would be anticipated that this would have no impact on cholesterol flux from extrahepatic tissues, including the vascular wall/atherosclerotic plaque [12]. On the other hand, if the source of the excreted cholesterol is at least partially derived from plasma lipoproteins, this may represent a component of the RCT pathway by which cholesterol that is effluxed from body tissues can exit the body, thereby reducing the cholesterol burden of extra-hepatic tissues, potentially including the vascular wall/atherosclerotic plaque [12,14]. The RCT pathway involves the net movement of cholesterol from peripheral tissues to the liver and can be conceptually viewed as comprising several major components: 1) the efflux of free cholesterol from tissues into plasma, 2) cholesterol transport with associated esterification in plasma, 3) hepatic uptake of plasma sterols, and 4) excretion into fecal sterols [12]. As the techniques used to measure macrophage-to-stool RCT in rodents [8,10] are not generally applicable to humans, alternative approaches to assessing RCT are needed. A method that measures several in vivo elements of RCT components in humans, namely efflux, plasma transport and excretion, using stable, non-radioactive isotopic tracers was recently developed and published by several authors of this study (Fig. 1) [15,16]. Here, we carried out a randomized, placebocontrolled, crossover study, investigating the effect of ezetimibe on RCT parameters in 26 hyperlipidemic human subjects using this approach. Thus, barring a specific effect of ezetimibe treatment on macrophage cholesterol metabolism without an effect on wholebody cholesterol fluxes, the methods applied here in humans would be anticipated to yield results similar to those observed in the RCT studies in rodents, assuming the effect of ezetimibe treatment is consistent across species.

Fig. 1. Efflux, plasma transport and excretion components of reverse cholesterol transport measured in this study. FC is effluxed from peripheral tissues via transporter-dependent and -independent mechanisms. Once on the HDL-particle, FC can be esterified to CE and transferred to LDL-particles via CE-transfer protein. FC and CE are delivered to the liver for secretion in bile as FC or BA, and then to the intestine with the possibilities of absorption and “recycling” or excretion in the feces. Cholesterol flux is measured with an intravenous infusion of 13C2-cholesterol. Sterol masses and enrichments in plasma FC, CE, and fecal neutral and acidic sterols are determined. Fluxes though each arm of RCT were calculated as described in Methods. FC ¼ free cholesterol, CE ¼ cholesterol ester, CETP ¼ CE-transfer protein, NS ¼ neutral sterols, BA ¼ bile acids, LCAT ¼ lecithin-cholesterol acyltransferase; SRBI = Scavenger Receptor type B1; LDLr = LDL receptor. Adapted from Expert Review of Cardiovascular Therapy, April 2008, Vol. 6, No. 4, Pages 447-470 with permission of Expert Reviews Ltd [40].

324

M.H. Davidson et al. / Atherosclerosis 230 (2013) 322e329

Fig. 2. Study design.

2. Methods 2.1. Materials [3,4-13C2] (13C2)-cholesterol (99%) was purchased from Cambridge Isotope Labs (Andover, MA) and 2H2O from Isotec Inc (Miamisburg, OH). Neutral sterol standards were purchased from Steraloids (Newport, RI). D4-Sitostanol and other reagents were from SigmaeAldrich (St. Louis, MO).

2.2. Study design This prospective, placebo-controlled crossover trial compared the effects of ezetimibe (10 mg/day) versus placebo during 7 weeks of treatment on several parameters of RCT in subjects with mild hypercholesterolemia (Fig. 2). The study was conducted in accordance with the principles of the Declaration of Helsinki, International Council on Harmonization, and Good Clinical Practice Guidelines. All subjects provided written informed consent and met all other entry qualifications based on medical history, physical examination, and laboratory tests. The protocol was approved by the appropriate institutional review boards, and the study was conducted between June 2008 and March 2009. The pre-specified primary outcome measure was fecal excretion of endogenous (plasma-derived) cholesterol, measured over the 10-day period following the cholesterol infusion. Secondary outcome measures included whole-body de novo cholesterol synthesis, total cholesterol efflux rate from extra-hepatic tissues into plasma (elsewhere named “rate of appearance” [15]), and percentage of efflux from tissues that is excreted as fecal sterols. Other measures included the rate of esterification of free cholesterol to cholesteryl-ester in plasma, the rate of cholesteryl-ester clearance from plasma, and fasting lipids during the 7-week treatment period. Safety variables assessed included adverse events, physical examinations, and laboratory measurements including increases in aspartate aminotransferase/alanine aminotransferase (AST/ALT) >3X upper limit of normal (ULN). Eligible subjects were non-smoking males and females, 40e75 years of age inclusive with a diagnosis of hypercholesterolemia (i.e., LDL-C between 3.4 and 5.2 mmol/L [130 mg/dl and 200 mg/dl] and with triglyceride concentrations less than <4.0 mmol/L [350 mg/dl], high-density lipoprotein cholesterol [HDL-C] values between 0.8 and 1.6 mmol/L [30 mg/dl and 60 mg/dl] for men, respectively, and between 1.0 and 1.8 mmol/L [40e70 mg/dl] for postmenopausal women). Subjects on lipid-lowering therapy within two months prior to the study or with a known history of coronary heart disease

(CHD), stroke or prior revascularization procedure, or peripheral vascular disease were excluded, as were those with diabetes mellitus. Other exclusions included baseline elevations in AST/ALT >2X ULN, fasting glucose levels 7.0 mmol/L [126 mg/dl], abnormal thyroidstimulating hormone (TSH), or laboratory evidence of renal impairment. Subjects were randomized to an initial treatment arm of placebo or ezetimibe and underwent measurements of cholesterol transport and de novo cholesterol synthesis parameters after 6 weeks. After completion of the first cholesterol transport measurements, subjects were crossed over at the end of day 53 to receive either placebo or ezetimibe, depending on what they received the first 6 weeks. Subjects underwent measurements of cholesterol transport and de novo cholesterol synthesis again after another 6 weeks of treatment. Treatment was continued for one additional week during each measurement period. Subjects were centrally randomized to receive either ezetimibe 10 mg or placebo by Radiant Development (Chicago, IL) in accordance with the manufacturer study product kit. All study personnel, including the investigators, study site personnel, patients, monitors, and central laboratory personnel, remained unaware of the treatment allocation throughout the study period. At baseline, subjects were admitted to the clinical research center for a 24-h, in-clinic stay for stable isotope infusions and oral heavy water administration. An early supper was served with ad-lib food intake allowed. After 5:00 PM food and caloric-containing beverages were withheld. Subjects continued treatment with study drug or placebo throughout the inpatient stay and subsequent 10 days. Subjects had an IV catheter placed for cholesterol infusion and another for blood draws. A stable isotope infusion of 13C2-cholesterol mixed in 10% IntralipidÒ and 10% ethanol was given piggy-backed into normal saline over 24 h. A total of 200 mg of labeled cholesterol was given. After the conclusion of the IV infusion and prior to discharge, subjects were administered 240 ml, divided into 6 doses over 18 h, of 70% enriched 2H2O in 40 ml aliquots during the in-clinic stay and were instructed to continue drinking the water (40 ml BID) for 10 days post-discharge. Subjects also received 2H4-sitostanol to take, 3 mg PO TID with meals, and collected a sample stool each day from the in-clinic period through 10 days post-discharge.

2.3. Outcome measures 2.3.1. Excretion of plasma-derived cholesterol into fecal sterols Total fecal excretion of sterols was measured from stool samples utilizing the measurement of H4-sitostanol in the same samples, the latter derived from the administration of 2H4-sitostanol

M.H. Davidson et al. / Atherosclerosis 230 (2013) 322e329

capsules, 3 mg given three times daily for 10 days. Sterol excretion (excretion) of neutral sterols or bile acids was calculated as:

Excretion ¼ ½Sterol ðmg=sampleÞ=H4 -Sitostanolðmg=sampleÞ  H4 -Sitostanol ðmg=dayÞ Total isotope recovery in stool over the 7-day period following the 13C2-cholesterol infusion was determined from the neutral sterol and bile acid enrichments multiplied by total sterol excretion rates. Neutral sterol and bile acid enrichments were determined by isotope ratio/mass spectrometry (IR/MS; Thermo Finnegan MAT 253 IR-MS, Bremen, Germany). Pre-infusion stool samples were used to determine baseline enrichments in each fraction, which were subtracted from post-infusion samples to determine 13C2-content arising from administered label. Total sterol excretion was calculated from the area under the curve of neutral sterols or bile acids multiplied by the average daily excretion of the subject. Excretion of plasma-derived cholesterol into fecal sterols was also calculated as the % recovery of total administered 13C2-cholesterol into fecal neutral sterols or bile acids, summed over 7 days. 2.3.2. De novo cholesterol synthesis The appearance of newly synthesized cholesterol in plasma was measured after 7 days of 2H2O intake. The isotopic enrichment in cholesterol was compared to body water enrichment and fraction of new cholesterol determined as described [17,18]. Some residual de novo synthesized (2H-labeled) cholesterol was present during the second study period; for these samples the background enrichment was determined in a plasma sample collected prior to the 2H2O administration and the enrichment in this sample was subtracted from the values on day 7. 2.3.3. Efflux of free cholesterol from tissues into plasma Free cholesterol efflux and cholesteryl-ester production were determined during the in-patient period in the fasting state. As described in detail elsewhere, a multi-compartment model was used to describe the flux of free cholesterol into and out of plasma, including flux from the liver, peripheral tissue, and red blood cells (RBCs) [15]. Labeled free 13C2-cholesterol was infused into human subjects over a period of 24 h, and the label concentration (isotopic enrichment) in, and the pool sizes of, plasma and RBC-free cholesterol pools and cholesteryl-ester pools were measured. The model was designed to incorporate data from all major, reasonably accessible pools of cholesterol, including 1) the rapidly exchanging pool of cholesterol that includes both plasma and liver free cholesterol [19,20], 2) erythrocyte-free cholesterol, and 3) plasma cholesteryl-ester pools. The model assumes that subjects are at metabolic steady state (i.e. constant weight and cholesterol concentrations), each pool is at steady state (i.e. flux in ¼ flux out), esterification of plasma free cholesterol is irreversible, and there is no direct removal of RBC-free cholesterol other than through the plasma free cholesterol pool. The percentage of cholesterol efflux from extra-hepatic tissues that was excreted as fecal sterols was determined by calculation of tissue efflux into plasma x % cholesterol excreted per day (13C2recovery). 2.3.4. Cholesteryl-ester production and fractional catabolic rates Direct in vivo determination of the rate of cholesteryl-ester production from free cholesterol by lecithin-cholesterol acyltransferase (LCAT) was measured during constant infusion of 13 C2-cholesterol and the appearance of label in plasma free cholesterol and cholesteryl-ester. Cholesteryl-ester fractional clearance rate (FCR, pools/hour), and plasma cholesteryl-ester

325

production rate were determined from the three-pool model fit performed as described (SAAM II) [15]. Cholesteryl-ester pool size was measured by chemical assay (Wako Chemicals USA, Richmond VA), as the difference between total and free cholesterol, times plasma volume estimated as 4.5% of body weight. 2.4. Analysis of cholesterol metabolites Plasma free cholesterol was extracted with ethanol-acetone, acetylated with toluene/pyridine/acetyl chloride, and dissolved in toluene for analysis by mass spectrometry (MS). Free cholesterol from RBCs was analyzed after homogenization with silicate beads and extraction in chloroform-methanol. For measurement of cholesteryl-ester enrichment, extracted plasma free cholesterol and cholesteryl-ester were first separated on an amino-propyl solid-phase extraction (SPE) cartridge, the fatty-acid moiety of the cholesteryl-ester cleaved by methanolic HCl, and the resultant free cholesterol was subsequently acetylated. Isotopic enrichments were determined by gas chromatography (GC)-MS for de novo cholesterol synthesis (2H, see above) and GC-combustion-IR-MS for 13 C2 (see below). Stool samples were first homogenized with an equal volume of water, then neutral sterol and bile acids were extracted separately under basic and acidic conditions, respectively, in the presence of the internal standards 5-a-cholestane and 5-a-cholanic acids. The bile acid extract was split into two fractions. The first, used for compositional analysis by flame ionization detection FID, was directly subjected to a two-step derivatization: butylation with butanolic-HCl followed by silylation by BSTFA-pyridine. The second half was further purified for MS analysis with an octadecyl-SPE cartridge, selectively eluting primarily deoxycholic acid with a 20% aqueous-methanol solution prior to butylation and silylation. The neutral sterol fraction was silylated directly for both compositional and isotope analysis. Isotopic enrichments of 13C2-cholesterol were measured using GC/C-IR/MS (Thermo Finnegan MAT 253 IR-MS, Bremen, Germany). Enrichments were determined as atom percent excess (APE) by comparison of the unknown samples to a standard curve generated with gravitametrically prepared working lab standards with known enrichments. Molar percent excess (MPE) is calculated as 14.5 or 15  APE for the acetyl- or silyl-derivative of cholesterol, respectively, and by 17  APE for the butyl-silylderivative of deoxycholic acid. Compositional analysis and excretion measurement of neutral sterols and bile acids were performed by GC/flame-ionization detection by comparison to the internal standards and sitostanol. GC peak areas of cholesterol, coprostanol, epicoprostanol, coprostan-3-one, and cholestanol were used to calculate neutral sterol mass. GC peak areas of isolithocholic, isodeoxycholic, lithocholic, deoxycholic, cholic, chenodeoxycholic, ursodeoxycholic, and 7-ketolithocholic were used to calculate acidic-sterol mass. Table 1 Plasma lipid levels.

Placebo mmol/L mg/dl Ezetimibe mmol/L mg/dl P-value

Total cholesterol

Triglycerides

Calculated LDL-C

HDL-C

Mean  SD

Mean  SD

Mean  SD

Mean  SD

5.67  0.7 219  26

3.31  1.2 128  48

3.83  0.6 148  24

1.2  0.2 46  9

4.84  0.6 187  24 <0.0001

3.13  1.3 121  50 >0.05

3.00  0.5 116  20 <0.0001

1.16  0.3 45  11 0.95

SD = standard deviation.

326

M.H. Davidson et al. / Atherosclerosis 230 (2013) 322e329

Table 2 Effect of ezetimibe on RCT parameters. RCT parameter

Neutral sterol excretion mg/day % infused 13C2-cholesterol recovered as neutral sterols Bile acid excretion mg/day % infused 13C2-cholesterol recovered as bile acids Total infused 13C2-cholesterol recovered in stoola CE pool size (mg/kg) CE FCR (pools/hr) CE PR (mg/kg/hr) FC efflux from extrahepatic tissues (mg/kg/hr) FC efflux from extrahepatic tissues excreted as fecal sterols (mg/kg/day) DNCS (%/day)

Placebo n ¼ 26

Ezetimibe n ¼ 26

Mean  SD

Mean  SD

1593  1287 19  16

1950  915 29  15

468 2.7 22 73 0.018 1.37 4.6 3.6 3.4

        

391 2.0 16 19 0.003 0.25 0.5 2.3 0.1

478 2.7 32 63 0.019 1.23 4.4 4.9 4.7

        

600 2.5 16 7 0.003 0.25 0.7 3.0 0.1

P-value

0.019 0.035 0.65 0.71 0.027 0.015 0.055 0.008 0.12 0.14 <0.001

CE ¼ cholesteryl-ester; DNCS ¼ de novo cholesterol synthesis; FC ¼ free cholesterol; FCR ¼ fractional clearance rate; PR ¼ production rate; SD ¼ standard deviation. a % recovery of infused 13C2-cholesterol as neutral sterols þ bile acids.

2.5. Statistical analyses The pre-specified primary hypothesis was that ezetimibe treatment would increase the excretion of endogenous (plasmaderived) cholesterol as fecal sterols expressed as the percent of fecal sterols (neutral and acidic) derived from the circulating plasma free cholesterol pool and excreted over 10 days following

the IV infusion of 13C2-cholesterol. The secondary hypotheses were that ezetimibe treatment would result in a significant increase in de novo cholesterol synthesis, increase cholesterol efflux from tissues into the bloodstream, and increase efflux from tissues excreted as fecal sterols. Summary statistics were provided for measurement of endogenous (plasma-derived) cholesterol excretion (%/day), de novo

Fig. 3. Individual changes in neutral sterol and bile acid excretion.

M.H. Davidson et al. / Atherosclerosis 230 (2013) 322e329

cholesterol synthesis, the cholesterol efflux rate (mg/kg/hr). All data presented are mean þ/ standard deviation. Paired t-test was performed to determine statistical significance, with P < 0.05 considered significant. Safety and tolerability were assessed by statistical and/or clinical review of all safety parameters, including adverse events, laboratory results, physical exams, and vital signs. Based on an evaluable sample size of 30 subjects, this study was estimated to have power of >90% to detect a 20% difference in the primary endpoint of plasma-derived cholesterol excretion (%/day), assuming a standard deviation of 30% (a ¼ 0.05, 2-sided). Based on historical intra-individual variability of 10% in the measurement of the cholesterol efflux (mg/kg/hr), this study was expected to provide >90% power to detect a 6.5% effect on cholesterol efflux. Intraindividual variability data were not available to estimate study power for de novo cholesterol synthesis (%). McNemar’s test was used to assess within-subject differences in categorical variables. 3. Results A total 26 of the 31 patients who were enrolled, completed this randomized, controlled, crossover study. Of those who discontinued the study, 1 patient withdrew consent, 1 was lost to follow-up, 1 subject was off-drug for 1 week, and 2 experienced adverse events and withdrew from the study (1 swelling of arm secondary to IV; 1 bruising bilateral arms at blood draw sites). There were 10 women in the study ranging in age from 50 to 71 years (mean 60.9 years), and 16 men ranging from 29 to 72 years (mean 55.8 years). Mean baseline LDL-C levels were 3.9 mmol/L (150 mg/dl), HDL-C levels were 1.2 mmol/L (47 mg/dl) and triglyceride levels were 1.7 mmol/L (151 mg/dl). Plasma lipid levels were measured on four occasions during the final 10 days of each of the two treatment phases, and the mean values are given in Table 1. Total plasma cholesterol and LDL-C were significantly reduced with ezetimibe treatment compared with placebo. Plasma triglycerides and HDL-C were not significantly changed. Ezetimibe treatment significantly increased total neutral sterol excretion by 23% (from 1593 to 1950 mg/d; P ¼ 0.019) and endogenous, plasma-derived neutral sterol excretion (% recovery of administered 13C2-cholesterol as neutral sterol) by 52% (from a mean of 19% to a mean of 29%; P ¼ 0.035) compared with placebo (Table 2). Total bile acid excretion (mg/day) and endogenous (plasma-derived) bile acid excretion (% recovery of administered 13 C2-cholesterol as bile acids) were not significantly changed. Individual changes in fecal sterols are shown in Fig. 3. In the majority of subjects, ezetimibe treatment increased total fecal neutral sterol excretion values and percent recovery of endogenous (plasmaderived) neutral sterols. Bile acid values were lower or unchanged in most subjects following ezetimibe treatment. Plasma cholesterol kinetic results are also presented in Table 2. There was no detectable effect of treatment on 13C2-cholesterol free cholesterol efflux from extra-hepatic tissues into the rapidly exchanging compartment in blood (P ¼ 0.12). The percentage of extra-hepatic tissue effluxed cholesterol recovered in feces per day, however, was significantly increased. The FCR of cholesteryl-ester (pools/hr), trended higher (P ¼ 0.055), while the production rate of cholesteryl-ester (FCR x cholesteryl-ester pool size) was significantly reduced with treatment (P ¼ 0.008), indicating that the reduction in plasma cholesteryl-ester concentration was due to both increased clearance and reduced input of cholesteryl-esters. Ezetimibe treatment significantly increased plasma de novo cholesterol synthesis measured on day 7 post-2H2O labeling compared with placebo (P < 0.001) (Table 2). This finding presumably reflects adaptation to the net loss of whole-body cholesterol with ezetimibe treatment. Directly measured changes in

327

cholesterol synthesis were in general agreement with plasma lathosterol levels (data not shown). Ezetimibe monotherapy was generally well-tolerated in this study. There were no serious adverse events leading to study discontinuations and no reported hepatic or muscle adverse events. 4. Conclusions In this study, the cholesterol absorption inhibitor, ezetimibe, had potentially favorable effects on several components of RCT in patients with hyperlipidemia. Ezetimibe significantly increased the efficiency of plasma cholesterol excretion into fecal neutral sterols and increased total neutral sterol output. Notably, bile acid excretion was unaffected. De novo synthesized cholesterol appearing in plasma was significantly increased. The FCR of plasma cholesterylesters trended higher with ezetimibe treatment and cholesterylester production rate was decreased, resulting in a reduced cholesteryl-ester pool size. Whole-body cholesterol efflux rate from tissues into plasma was not detectably changed with ezetimibe treatment, although the calculated percentage of tissue effluxed cholesterol that appeared in feces was significantly increased. Overall, these results indicate that ezetimibe stimulates the flux of plasma cholesterol into fecal neutral sterols in hyperlipidemic patients. There are several important conclusions to be drawn from this study. First, ezetimibe treatment enhanced the excretion of endogenous (plasma-derived) cholesterol as neutral sterols, resulting in an increased removal of cholesteryl-esters from plasma and a compensatory increase in de novo cholesterol synthesis to maintain cholesterol balance. These data suggest that the observed increase in cholesterol flux from plasma-to-stool involved at least 2 of the 3 components of RCT, and are consistent with previous studies that have shown ezetimibe-induced increases in various parameters of RCT in both rodents [7,8,10,21e23] and in humans [2,3]. Further, while inhibition of cholesterol absorption is not commonly considered to be a mechanism that increases RCT, these findings and prior studies in animals suggest that this may occur [7,8]. Cholesterol flux is perhaps best understood as a cycle. In this process, free cholesterol continually fluxes between extra-hepatic tissues and plasma in amounts recently estimated to be in the range of 8 g/d [15]. Free cholesterol that is associated with HDLparticles in plasma can become esterified via LCAT and then be partially redistributed to LDL (and potentially other apoBcontaining) particles. Both HDL- and LDL-particles can either deliver cholesterol back to the tissues or present it to the liver for clearance from the circulation. From the liver, it can be excreted into bile or secreted back into plasma within VLDL [24]. Biliary cholesterol can then be either excreted in feces or taken up by the enterocyte and returned to the plasma via chylomicrons. Ezetimibe inhibition of cholesterol absorption reduces the cholesterol content of chylomicrons and consequently the amount of cholesterol delivered to the liver [1,25]. One way the body compensates for this diminished cholesterol delivery to the liver is by up-regulation of LDL-receptors with resultant increased hepatic uptake of cholesterol-rich apoB-containing lipoproteins [26]. The latter phenomenon accounts for the LDL-lowering effect of ezetimibe. Another compensatory mechanism seen in the current study, as well as other ezetimibe studies, is a significant increase in de novo cholesterol synthesis, which also presumably contributes to the increased plasma-to-feces cholesterol flux that we observed [2,3,27e29]. Finally, as demonstrated for the first time in humans in the current study, ezetimibe significantly increases the flux of cholesterol from plasma into feces, reflecting a potentially important effect on this component of the RCT pathway.

328

M.H. Davidson et al. / Atherosclerosis 230 (2013) 322e329

Although these findings do not directly tell us whether the flux of cholesterol from peripheral tissues to the liver and ultimately out of the body (Fig. 1) is increased by ezetimibe in humans, studies in rodents suggest that this may be the case [7,8]. Increased flux of labeled cholesterol from intraperitoneally injected macrophages into the feces has been demonstrated in rodents treated with ezetimibe [7,8]. Total bidirectional flux of free cholesterol between tissues and plasma would be expected to be balanced or nearly so in the steady state. The net effect of ezetimibe after w7 weeks of treatment has been shown here to include an increase in a component of RCT, namely the flux of plasma-derived cholesterol into feces, approximately balanced by increased whole-body de novo cholesterol synthesis. In this study, the inhibition of intestinal transport by ezetimibe produced no detectable change in efflux of extra-hepatic tissue free cholesterol into plasma as measured by the 13C2-cholesterol infusion method, although the study was not powered to detect small changes. However, these data do not support the speculation put forth by some that ezetimibe might inhibit free cholesterol transport from tissues to the liver [30]. Specifically, it has previously been suggested that ezetimibe may inhibit mediators of RCT including adenosine triphosphate-binding cassette transporter and scavenger receptor type B1, based on in vitro tissue culture studies [31e33], although others have reported that this is not the case [34,35]. In this study, the absence of any significant reduction in tissue cholesterol efflux into plasma and its clear effect to increase plasma-to-feces cholesterol flux provides strong evidence that ezetimibe treatment does not adversely inhibit the RCT pathway in vivo. Rather, along with recent rodent findings demonstrating increased macrophage-to-feces cholesterol flux with ezetimibe treatment, our findings are consistent with the hypothesis that reduction of cholesterol absorption and the consequent decrease in chylomicron cholesterol reaching the liver leads to compensatory mechanisms that include increased cholesterol plasma-to-feces flux and increased de novo synthesis [36]. The enhanced plasmato-feces flux may in turn be associated with enhanced cholesterol flux from some tissues to plasma, as suggested by the stimulation of appearance of intraperitoneal macrophage-labeled cholesterol in feces with ezetimibe treatment in rodents. Because neither an increase nor reduction in free cholesterol efflux from tissues-to-plasma was observed in our study, it appears likely that increased de novo cholesterol synthesis in the liver (which rapidly equilibrates with the plasma compartment) leaves the measured extra-hepatic tissue cholesterol efflux essentially unchanged or too small to detect. This does not preclude the possibility that a degree of negative cholesterol balance, with potentially favorable effects on atherosclerosis, occurs with initiation of ezetimibe therapy associated with stimulation of RCT, or possibly even at a low level in the steady state. There are some methodological limitations to the study worth noting. In our analysis, we have assumed that each subject is being evaluated at metabolic steady state relative to the flux rates reported; while we believe this to be a reasonable assumption, a nonsteady-state in the rates of cholesterol fluxes over the 24 h infusion, or the subsequent days of 2H2O labeling could affect the accuracy of our results. Additionally, the efflux of free cholesterol as measured herein represents non-hepatic whole-body free cholesterol flux into plasma. As such, the degree to which atherosclerotic plaque may, or may not be affected by changes is not addressed. The rate of cholesteryl-ester formation is assumed to be dominated by plasma LCAT activity; however, if a plasma free cholesterol molecule was esterified by acyl-CoA:cholesterol acyltransferase (ACAT) and then subsequently released onto a lipoprotein particle, this would be included in the ester formation rate reported here. Based on our understanding of the relative contributions of the two pathways we

assume that ACAT activity is minor here. The tissue source of de novo synthesized cholesterol is not determined here, therefore both hepatic and extra-hepatic cholesterol synthesis may contribute to the percent per day value we report. Finally, the patient population studied had a wide range of baseline lipid values which may have limited our ability to detect changes in lipid fluxes; however, the variability in this cohort may provide real-world relevance. In summary, our findings demonstrate that inhibition of NPC1L1 with ezetimibe treatment increases flux of plasmaderived cholesterol from plasma-to-feces, compared with placebo. This effect was associated with increases in components of the RCT pathway in humans. Whether increased cholesterol elimination and increased RCT flux from plasma-to-feces translates to clinical benefit beyond the plasma LDL-C lowering effect of ezetimibe is unknown and remains speculative at this time. The recent Study of Heart and Renal Protection (SHARP) [37] demonstrated that ezetimibe/simvastatin (10/20 mg) reduced cardiovascular events in patients with chronic kidney disease. While the observed risk reductions for major atherosclerotic or vascular events in SHARP for the observed LDL-C lowering were consistent with expectations from the Cholesterol Treatment Trialists metaanalysis of statin trials [38], chronic kidney disease patients consist of a complex population confounded by a high incidence of non-atherosclerotic heart disease that makes it difficult to assess the magnitude of relevant risk reduction. The incremental effect of ezetimibe on cardiovascular outcomes beyond statin therapy is the specific focus of the ongoing Improved Reduction of Outcomes: Vytorin Efficacy International Trial (IMPROVE-IT) [39]. The concept that ezetimibe may affect RCT in a manner that is clinically beneficial is one that should be considered if the results of IMPROVE-IT were to suggest a treatment effect exceeding that anticipated from the LDL-C reduction achieved.

Funding source This Investigator-Initiated Research Study was sponsored by Radiant Research and funded by Merck, Sharp & Dohme Corp., a subsidiary of Merck & Co. Inc., Whitehouse Station, NJ, USA.

Disclosures Drs. A Glass, J Luchoomun, J Voogt, and S Turner, and J Decaris, S Killion, D Villegas, Y Lu and H Mohammad are employees of KineMed and may hold stock/stock options in the company. Dr. Hellerstein has ownership interest in KineMed and is Chief of KineMed’s Scientific Advisory Board. Dr. S Turner has also received a research grant from Merck. Dr. M H Davidson has ownership interest in Omthera Pharma and sits on the Advisory Boards of Amgen, Merck, Roche, and Sanofi-Aventis. Drs. Musliner, Neff, and Tomassini are employees of Merck, Sharp & Dohme Corp., a subsidiary of Merck & Co. Inc., Whitehouse Station, NJ, USA.

Acknowledgments and author contributions Martha C. Vollmer, M.A., Merck & Co., Inc. Whitehouse Station, NJ, USA is acknowledged for editorial assistance. All authors jointly developed the manuscript content and were involved in at least one of the following: conception, design, data acquisition, analysis, statistical analysis, interpretation of data, drafting the manuscript, and/or revising the manuscript for important intellectual content. All authors provided final approval of the version to be published.

M.H. Davidson et al. / Atherosclerosis 230 (2013) 322e329

References [1] Davis Jr HR, Tershakovec AM, Tomassini JE, Musliner T. Intestinal sterol transporters and cholesterol absorption inhibition. Curr Opin Lipidol 2011;22: 467e78. [2] Sudhop T, Lutjohann D, Kodal A, et al. Inhibition of intestinal cholesterol absorption by ezetimibe in humans. Circulation 2002;106:1943e8. [3] Sudhop T, Reber M, Tribble D, et al. Changes in cholesterol absorption and cholesterol synthesis caused by ezetimibe and/or simvastatin in men. J Lipid Res 2009;50:2117e23. [4] Zetia [product insert]. Whitehouse Station, NJ: Merck & Co., Inc; 2012. [5] Tremblay AJ, Lamarche B, Cohn JS, Hogue JC, Couture P. Effect of ezetimibe on the in vivo kinetics of apoB-48 and apoB-100 in men with primary hypercholesterolemia. Arterioscler Thromb Vasc Biol 2006;26:1101e6. [6] Tremblay AJ, Lamarche B, Hogue JC, Couture P. Effects of ezetimibe and simvastatin on apolipoprotein B metabolism in males with mixed hyperlipidemia. J Lipid Res 2009;50:1463e71. [7] Sehayek E, Hazen SL. Cholesterol absorption from the intestine is a major determinant of reverse cholesterol transport from peripheral tissue macrophages. Arterioscler Thromb Vasc Biol 2008;28:1296e7. [8] Briand F, Naik SU, Fuki I, et al. Both the peroxisome proliferator-activated receptor delta agonist, GW0742, and ezetimibe promote reverse cholesterol transport in mice by reducing intestinal reabsorption of HDL-derived cholesterol. Clin Transl Sci 2009;2:127e33. [9] Xie P, Jia L, Ma Y, et al. Ezetimibe inhibits hepatic Niemann-Pick C1-Like 1 to facilitate macrophage reverse cholesterol transport in mice. Arterioscler Thromb Vasc Biol 2013;33:920e5. [10] Davis Jr HR, Lowe RS, Neff DR. Effects of ezetimibe on atherosclerosis in preclinical models. Atherosclerosis 2011;215:266e78. [11] Salen G, Starc T, Sisk CM, Patel SB. Intestinal cholesterol absorption inhibitor ezetimibe added to cholestyramine for sitosterolemia and xanthomatosis. Gastroenterology 2006;130:1853e7. [12] Turley SD. The role of Niemann-Pick C1-Like 1 (NPC1L1) in intestinal sterol absorption. J Clin Lipidol 2008;2:S20e8. [13] Kruit JK, Groen AK, van Berkel TJ, Kuipers F. Emerging roles of the intestine in control of cholesterol metabolism. World J Gastroenterol 2006;12:6429e39. [14] van der Velde AE, Brufau G, Groen AK. Transintestinal cholesterol efflux. Curr Opin Lipidol 2010;21:167e71. [15] Turner S, Voogt J, Davidson M, et al. Measurement of reverse cholesterol transport pathways in humans: in vivo rates of free cholesterol efflux, esterification, and excretion. J Am Heart Assoc 2012;1:e001826. [16] Holleboom AG, Jakulj L, Franssen R, et al. In vivo tissue cholesterol efflux is reduced in carriers of a mutation in APOA1. J Lipid Res 2013. http://dx.doi.org/ 10.1194/jlr.P028449. [17] Lee WN, Bassilian S, Guo Z, et al. Measurement of fractional lipid synthesis using deuterated water (2H2O) and mass isotopomer analysis. Am J Physiol 1994;266:E372e83. [18] Previs SF, Mahsut A, Kulick A, et al. Quantifying cholesterol synthesis in vivo using (2)H(2)O: enabling back-to-back studies in the same subject. J Lipid Res 2011;52:1420e8. [19] Schwartz CC, Zech LA, Vandenbroek JM, Cooper PS. Cholesterol kinetics in subjects with bile fistula. Positive relationship between size of the bile acid precursor pool and bile acid synthetic rate. J Clin Invest 1993;91: 923e38. [20] Schwartz CC, Vandenbroek JM, Cooper PS. Lipoprotein cholesteryl ester production, transfer, and output in vivo in humans. J Lipid Res 2004;45:1594e607. [21] Gomez-Garre D, Munoz-Pacheco P, Gonzalez-Rubio M, Aragoncillo P, Granados R, Fernandez-Cruz A. Ezetimibe reduces plaque inflammation in a

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30] [31]

[32] [33]

[34]

[35] [36]

[37]

[38]

[39]

[40]

329

rabbit model of atherosclerosis and inhibits monocyte migration in addition to its lipid-lowering effect. Br J Pharmacol 2009;156:1218e27. Jakulj L, Trip MD, Sudhop T, von Bergmann K, Kastelein JJ, Vissers MN. Inhibition of cholesterol absorption by the combination of dietary plant sterols and ezetimibe: effects on plasma lipid levels. J Lipid Res 2005;46:2692e8. Jakulj L, Vissers MN, van Roomen CP, et al. Ezetimibe stimulates faecal neutral sterol excretion depending on abcg8 function in mice. FEBS Lett 2010;584: 3625e8. deGoma EM, Degoma RL, Rader DJ. Beyond high-density lipoprotein cholesterol levels evaluating high-density lipoprotein function as influenced by novel therapeutic approaches. J Am Coll Cardiol 2008;51:2199e211. Van Heek M, Compton DS, Davis HR. The cholesterol absorption inhibitor, ezetimibe, decreases diet-induced hypercholesterolemia in monkeys. Eur J Pharmacol 2001;415:79e84. Ginsberg HN. REVIEW: efficacy and mechanisms of action of statins in the treatment of diabetic dyslipidemia. J Clin Endocrinol Metab 2006;91: 383e92. Assmann G, Kannenberg F, Ramey DR, Musliner TA, Gutkin SW, Veltri EP. Effects of ezetimibe, simvastatin, atorvastatin, and ezetimibe-statin therapies on non-cholesterol sterols in patients with primary hypercholesterolemia. Curr Med Res Opin 2008;24:249e59. Repa JJ, Turley SD, Quan G, Dietschy JM. Delineation of molecular changes in intrahepatic cholesterol metabolism resulting from diminished cholesterol absorption. J Lipid Res 2005;46:779e89. Salen G, von Bergmann K, Lutjohann D, et al. Ezetimibe effectively reduces plasma plant sterols in patients with sitosterolemia. Circulation 2004;109: 966e71. Taylor AJ, Villines TC, Stanek EJ, et al. Extended-release niacin or ezetimibe and carotid intima-media thickness. N Engl J Med 2009;361:2113e22. During A, Dawson HD, Harrison EH. Carotenoid transport is decreased and expression of the lipid transporters SR-BI, NPC1L1, and ABCA1 is downregulated in Caco-2 cells treated with ezetimibe. J Nutr 2005;135:2305e12. Kasza I, Hegyi Z, Szabo K, et al. Model system for the analysis of cell surface expression of human ABCA1. BMC Cell Biol 2009;10:93. Labonte ED, Howles PN, Granholm NA, et al. Class B type I scavenger receptor is responsible for the high affinity cholesterol binding activity of intestinal brush border membrane vesicles. Biochim Biophys Acta 2007;1771:1132e9. Field FJ, Watt K, Mathur SN. Ezetimibe interferes with cholesterol trafficking from the plasma membrane to the endoplasmic reticulum in CaCo-2 cells. J Lipid Res 2007;48:1735e45. Field FJ, Watt K, Mathur SN. Origins of intestinal ABCA1-mediated HDLcholesterol. J Lipid Res 2008;49:2605e19. Dietschy JM. Theoretical considerations of what regulates low-densitylipoprotein and high-density-lipoprotein cholesterol. Am J Clin Nutr 1997;65:1581Se9S. Baigent C, Landray MJ, Reith C, et al. The effects of lowering LDL cholesterol with simvastatin plus ezetimibe in patients with chronic kidney disease (Study of Heart and Renal Protection): a randomised placebo-controlled trial. Lancet 2011;377:2181e92. Baigent C, Blackwell L, Emberson J, et al. Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170 000 participants in 26 randomised trials. Lancet 2010;376:1670e81. Califf RM, Lokhnygina Y, Cannon CP, et al. An update on the IMProved reduction of outcomes: Vytorin Efficacy International Trial (IMPROVE-IT) design. Am Heart J 2010;159:705e9. Bays HE, Neff D, Tomassini JE, Tershakovec AM. Ezetimibe: cholesterol lowering and beyond. Expert Rev Cardiovasc Ther 2008;6:447e70.