A novel approach to measuring macrophage-specific reverse cholesterol transport in vivo in humans

Reverse cholesterol transport (RCT) is thought to be an atheroprotective function of HDL, and macrophage-specific RCT in mice is inversely associated with atherosclerosis. We developed a novel method using 3H-cholesterol nanoparticles to selectively trace macrophage-specific RCT in vivo in humans. Use of 3H-cholesterol nanoparticles was initially tested in mice to assess the distribution of tracer and response to interventions known to increase RCT. Thirty healthy subjects received 3H-cholesterol nanoparticles intravenously, followed by blood and stool sample collection. Tracer counts were assessed in plasma, nonHDL, HDL, and fecal fractions. Data were analyzed by using multicompartmental modeling. Administration of 3H-cholesterol nanoparticles preferentially labeled macrophages of the reticuloendothelial system in mice, and counts were increased in mice treated with a liver X receptor agonist or reconstituted HDL, as compared with controls. In humans, tracer disappeared from plasma rapidly after injection of nanoparticles, followed by reappearance in HDL and nonHDL fractions. Counts present as free cholesterol increased rapidly and linearly in the first 240 min after nadir; counts in cholesteryl ester increased steadily over time. Estimates of fractional transfer rates of key RCT steps were obtained. These results support the use of 3H-cholesterol nanoparticles as a feasible approach for the measurement of macrophage RCT in vivo in humans. Reverse cholesterol transport (RCT) is thought to be an atheroprotective function of HDL, and macrophage-specific RCT in mice is inversely associated with atherosclerosis. We developed a novel method using 3H-cholesterol nanoparticles to selectively trace macrophage-specific RCT in vivo in humans. Use of 3H-cholesterol nanoparticles was initially tested in mice to assess the distribution of tracer and response to interventions known to increase RCT. Thirty healthy subjects received 3H-cholesterol nanoparticles intravenously, followed by blood and stool sample collection. Tracer counts were assessed in plasma, nonHDL, HDL, and fecal fractions. Data were analyzed by using multicompartmental modeling. Administration of 3H-cholesterol nanoparticles preferentially labeled macrophages of the reticuloendothelial system in mice, and counts were increased in mice treated with a liver X receptor agonist or reconstituted HDL, as compared with controls. In humans, tracer disappeared from plasma rapidly after injection of nanoparticles, followed by reappearance in HDL and nonHDL fractions. Counts present as free cholesterol increased rapidly and linearly in the first 240 min after nadir; counts in cholesteryl ester increased steadily over time. Estimates of fractional transfer rates of key RCT steps were obtained. These results support the use of 3H-cholesterol nanoparticles as a feasible approach for the measurement of macrophage RCT in vivo in humans. Although an association between plasma concentrations of HDL cholesterol (HDL-C) levels and CVD has been found in epidemiological studies, the causal basis of this association has been recently questioned by the negative results observed in human genetics studies and clinical trials (1Voight B.F. Peloso G.M. Orho-Melander M. Frikke-Schmidt R. Barbalic M. Jensen M.K. Hindy G. Holm H. Ding E.L. Johnson T. Plasma HDL cholesterol and risk of myocardial infarction: a mendelian randomisation study.Lancet. 2012; 380: 572-580Abstract Full Text Full Text PDF PubMed Scopus (1648) Google Scholar, 2Rader D.J. Hovingh G.K. HDL and cardiovascular disease.Lancet. 2014; 384: 618-625Abstract Full Text Full Text PDF PubMed Scopus (414) Google Scholar). However, substantial interest remains in the concept that measures of HDL function may be more closely tied mechanistically to atherosclerosis (3Rosenson R.S. Brewer Jr., H.B. Davidson W.S. Fayad Z.A. Fuster V. Goldstein J. Hellerstein M. Jiang X.C. Phillips M.C. Rader D.J. Cholesterol efflux and atheroprotection: advancing the concept of reverse cholesterol transport.Circulation. 2012; 125: 1905-1919Crossref PubMed Scopus (681) Google Scholar). One of the biological functions of HDL is its ability to promote cellular cholesterol efflux and the return of that cholesterol to the liver for excretion in the feces in a process known as reverse cholesterol transport (RCT) (4Cuchel M. Rader D.J. Macrophage reverse cholesterol transport: key to the regression of atherosclerosis?.Circulation. 2006; 113: 2548-2555Crossref PubMed Scopus (439) Google Scholar). By using an in vivo assay in mice that we developed to measure the rate of RCT from macrophages to liver and feces, a large number of studies from multiple laboratories have demonstrated that the rate of in vivo macrophage RCT is a much better predictor of the impact of a genetic or pharmacologic perturbation on atherosclerosis than is the simple measure of HDL-C itself (5Rader D.J. Alexander E.T. Weibel G.L. Billheimer J. Rothblat G.H. Role of reverse cholesterol transport in animals and humans and relationship to atherosclerosis.J. Lipid Res. 2009; 50: S189-S194Abstract Full Text Full Text PDF PubMed Scopus (452) Google Scholar). Furthermore, using an ex vivo method developed to assess the cholesterol efflux capacity of HDL from a specific individual, we showed that this was strongly inversely associated with prevalent atherosclerosis, even after adjusting for HDL-C levels (6Khera A.V. Cuchel M. de la Llera-Moya M. Rodrigues A. Burke M.F. Jafri K. French B.C. Phillips J.A. Mucksavage M.L. Wilensky R.L. Cholesterol efflux capacity, high-density lipoprotein function, and atherosclerosis.N. Engl. J. Med. 2011; 364: 127-135Crossref PubMed Scopus (1532) Google Scholar). Similarly, cholesterol efflux capacity of HDL has been shown to be inversely associated with incident CVD (7Rohatgi A. Khera A. Berry J.D. Givens E.G. Ayers C.R. Wedin K.E. Neeland I.J. Yuhanna I.S. Rader D.R. de Lemos J.A. HDL cholesterol efflux capacity and incident cardiovascular events.N. Engl. J. Med. 2014; 371: 2383-2393Crossref PubMed Scopus (969) Google Scholar, 8Saleheen D. Scott R. Javad S. Zhao W. Rodrigues A. Picataggi A. Lukmanova D. Mucksavage M.L. Luben R. Billheimer J. Association of HDL cholesterol efflux capacity with incident coronary heart disease events: a prospective case-control study.Lancet Diabetes Endocrinol. 2015; 3: 507-513Abstract Full Text Full Text PDF PubMed Scopus (348) Google Scholar) These results suggest that intervention to promote macrophage efflux and RCT might be an effective approach to prevent and treat atherosclerotic CVD. An approach to assess RCT in humans was reported (9Turner S. Voogt J. Davidson M. Glass A. Killion S. Decaris J. Mohammed H. Minehira K. Boban D. Murphy E. Measurement of reverse cholesterol transport pathways in humans: in vivo rates of free cholesterol efflux, esterification, and excretion.J. Am. Heart Assoc. 2012; 1: e001826Crossref PubMed Google Scholar), but this method is not macrophage-specific. A method to measure macrophage-specific RCT in humans would be useful, not only in assessing the role of key proteins regulating RCT in humans, but also in evaluating the potential of novel therapeutic approaches targeting cholesterol efflux and RCT. Nilsson and Zilversmit demonstrated that intravenous (i.v.) injection of rats with a saturated solution of radiolabeled unesterified cholesterol mixed with albumin rapidly disappeared from the blood compartment, primarily due to the uptake by reticuloendothelial (RE) cells (10Nilsson A. Zilversmit D.B. Fate of intravenously administered particulate and lipoprotein cholesterol in the rat.J. Lipid Res. 1972; 13: 32-38Abstract Full Text PDF PubMed Google Scholar). Schwartz et al. used a similar preparation to study cholesterol metabolic pathways in humans. Their results were consistent with the rapid clearance of the cholesterol-albumin complexes from the blood compartment and subsequent reappearance of the tracer on circulating HDL as free cholesterol (FC), suggesting that this approach may specifically measure the efflux of cholesterol from macrophage cells to HDL as sole acceptor (11Schwartz C.C. Berman M. Vlahcevic Z.R. Halloran L.G. Gregory D.H. Swell L. Multicompartmental analysis of cholesterol metabolism in man. Characterization of the hepatic bile acid and biliary cholesterol precursor sites.J. Clin. Invest. 1978; 61: 408-423Crossref PubMed Scopus (133) Google Scholar, 12Schwartz C.C. Vlahcevic Z.R. Berman M. Meadows J.G. Nisman Swell L. role of lipoprotein in plasma free cholesterol Clin. Invest. PubMed Scopus Google Scholar, C.C. Cholesterol in subjects with bile relationship between of the bile acid precursor and bile acid Clin. Invest. PubMed Scopus Google Scholar, C.C. cholesteryl ester and in vivo in Lipid Res. Full Text Full Text PDF PubMed Scopus Google Scholar). on we that cholesterol nanoparticles of 3H-cholesterol be used to 3H-cholesterol selectively to macrophages in vivo and the of macrophage-specific cholesterol efflux in vivo in humans. We present the results of studies and a study in humans that were to the and of studies were in mice a Cholesterol nanoparticles were to that for human studies and solution were Plasma was and was after by using to the HDL was as M. S. de la Llera-Moya M. D. Rothblat G.H. Phillips M.C. Rader D.J. by reconstituted high-density lipoprotein free cholesterol from and from PubMed Scopus Google Scholar) and administered a of to and studies using the liver X receptor agonist were and and mice on were for and the of the and were mice of or were the and or and a for by after animals were with and of 3H-cholesterol nanoparticles were intravenously the of Cholesterol nanoparticles were by of in with of of in for min was to after injection and the study in a with were plasma by from and min by plasma was to and was by cells were from liver as In vivo and in uptake and of lipoprotein by liver and Full Text PDF PubMed Google Scholar). mice were and the was the was and the liver was for min with the liver in with for cells were after the liver in and for min of cells as by were for min in to the By of the cells and were to the method of et al. In vivo and in uptake and of lipoprotein by liver and Full Text PDF PubMed Google Scholar). were in mice by using as A. by of and PubMed Google Scholar). mice were with of is known to or in was assessed by using to liver and were by levels of the mice were with of cholesterol nanoparticles and were and or min after was in plasma, and blood cells as cholesterol efflux of from mice was assessed to an using acid and to the and or the HDL by of using as an acceptor B.F. de la Llera-Moya M. Rothblat G.H. of HDL on cellular and cholesterol Lipid Res. Full Text Full Text PDF PubMed Scopus Google Scholar). of or the and Use the and were in with the on and Use of 3H-cholesterol was in the of the of that to complexes with cholesterol was from and 3H-cholesterol was from Cholesterol nanoparticles were by of human albumin to a solution of cholesterol in a in and solution was for min and a to the 3H-cholesterol and was administered as a of the of the cholesterol nanoparticles, were on an solution by using a distribution and analysis for and A of the cholesterol nanoparticles was to be We and in and with a of HDL-C subjects were only of Use of the or the study or was subjects were to the clinical of were to for a of and in the for a of of the was in was used for radiolabeled cholesterol and the for blood were with a the of the tracer and a and after the injection of the Cholesterol nanoparticles were as and administered of preparation as a rapid were in and after tracer and on were from the after the blood and to the and after for blood and stool collection. study was and Administration and by the of and subjects the were in and A of plasma was in a of solution acid and was and in a on and nonHDL was with of and was and the was in the plasma was in and for analysis and as Measurement of a was to the of cholesterol the study and to be used for of specific blood was and were were by using E.G. A rapid method of and J. PubMed Scopus Google Scholar). HDL, and nonHDL were to and cholesterol ester with acid blood cells were with and and were as were and of from were and a was were with and the cholesterol was by acid of bile the of tracer in the stool bile acid in cholesterol was by using Lipid and were by using vivo cholesterol efflux capacity of the HDL was as (6Khera A.V. Cuchel M. de la Llera-Moya M. Rodrigues A. Burke M.F. Jafri K. French B.C. Phillips J.A. Mucksavage M.L. Wilensky R.L. Cholesterol efflux capacity, high-density lipoprotein function, and atherosclerosis.N. Engl. J. Med. 2011; 364: 127-135Crossref PubMed Scopus (1532) Google Scholar). multicompartmental analysis was by using the on the cholesterol developed by Schwartz et al. C.C. cholesteryl ester and in vivo in Lipid Res. Full Text Full Text PDF PubMed Scopus Google Scholar), efflux of to HDL from cells of the system as a key measure of the RCT used a tracer and not bile the Schwartz was that be with results of the the process C.C. Cholesterol in subjects with bile relationship between of the bile acid precursor and bile acid Clin. Invest. PubMed Scopus Google Scholar, M. S. de la Llera-Moya M. D. Rothblat G.H. Phillips M.C. Rader D.J. by reconstituted high-density lipoprotein free cholesterol from and from PubMed Scopus Google Scholar). on HDL transfer to nonHDL blood or cholesterol of cholesterol from or be to nonHDL or is by Cholesterol by liver is plasma as HDL or nonHDL, for of the labeled cholesterol tracer or in analysis of in the human study was with were by using of between and 240 min was for by using a for was used to assess between of was as the of cholesterol in mice, the tracer rapidly disappeared from a min followed by reappearance in plasma that by and after the cholesterol tracer was primarily in the liver and the By and after the counts in the liver and were and counts in as and were of liver and cells showed on a of the of tracer in cells was that of cells after the tracer present in the cells to that in cells 3H-cholesterol as counts of in liver and cells and min after of a solution were with or and with 3H-cholesterol were or min and counts were in plasma, blood and Data a of We that the liver cells for uptake of the cholesterol nanoparticles were cells of the were from mice by using of were by as by in of macrophage-specific and not of cholesterol from the plasma of mice that received is rapid with a of min and with of plasma counts after min of in mice that received in a clearance of the with of plasma counts after min was by a to increase in tracer in the blood min and min the uptake by the liver min was by of and min ability to efflux cholesterol to known demonstrated cholesterol efflux to HDL and that was to with to with that in macrophages the we studies using interventions to increase cholesterol efflux and RCT. of a known to RCT in was mice were treated with the agonist R. D. J. J. R. J. of a novel liver X receptor agonist with and a in 2015; PubMed Scopus Google Scholar) or for followed by the of of the was from the plasma in in the and this rate was increased in animals results were in animals However, reappearance of in plasma was in or in and mice that in this and agonist macrophage cholesterol efflux is but not mice were with complexes or as we M. S. de la Llera-Moya M. D. Rothblat G.H. Phillips M.C. Rader D.J. by reconstituted high-density lipoprotein free cholesterol from and from PubMed Scopus Google Scholar). in a increase in cholesterol in plasma also in a increase in the rate of reappearance of cholesterol tracer after with plasma counts in the compared with the A of healthy subjects were the and shown in and of study subjects as or not Cholesterol ex vivo cholesterol efflux capacity is reported as to the of a as D. Scott R. Javad S. Zhao W. Rodrigues A. Picataggi A. Lukmanova D. Mucksavage M.L. Luben R. Billheimer J. Association of HDL cholesterol efflux capacity with incident coronary heart disease events: a prospective case-control study.Lancet Diabetes Endocrinol. 2015; 3: 507-513Abstract Full Text Full Text PDF PubMed Scopus (348) Google Scholar). blood in a Data as or not Cholesterol ex vivo cholesterol efflux capacity is reported as to the of a as D. Scott R. Javad S. Zhao W. Rodrigues A. Picataggi A. Lukmanova D. Mucksavage M.L. Luben R. Billheimer J. Association of HDL cholesterol efflux capacity with incident coronary heart disease events: a prospective case-control study.Lancet Diabetes Endocrinol. 2015; 3: 507-513Abstract Full Text Full Text PDF PubMed Scopus (348) Google Scholar). blood the of particulate the tracer counts in plasma a and a rate over the after and steadily over of counts present in plasma as in the tracer is rapidly the tracer present as increased an of the tracer in results were observed in HDL and nonHDL fractions. of reappearance and was and the specific was for counts present as than and Counts were initially in the HDL than in the nonHDL suggesting that the tracer first in of increase in in nonHDL and HDL for increase in as as as in a Data as subjects to over a counts in the were and of the and tracer from plasma, HDL, and nonHDL as as from blood cells and fecal were used to fractional transfer and transport rates using multicompartmental modeling. transfer rates for key steps of the the efflux of from macrophage to HDL, the of to and the transfer of from HDL to nonHDL, were and shown in on the by the macrophage on labeled nanoparticles was to HDL an rate of of the in the cells was HDL the of to and the transfer of to nonHDL In and as the transfer of from HDL to nonHDL rapidly In the transfer of from HDL to nonHDL was In to was transfer of from HDL to nonHDL, was for the transfer of from of and from fractional transfer rates for cholesterol by multicompartmental of the cholesterol tracer to to to to transfer transfer of compartment not transfer rates to the transfer rates in the shown in 1: to to to nonHDL to to to HDL to to in a Data not transfer rates to the transfer rates in the shown in 1: to to to nonHDL to to to HDL to to We a method to trace macrophage efflux and RCT in humans in studies demonstrated after radiolabeled cholesterol tracer administered as cholesterol nanoparticles rapidly disappeared from and was primarily by cells of the with efflux the over time. Furthermore, we demonstrated that were to efflux cholesterol with pathways that similar to observed in a used macrophage (6Khera A.V. Cuchel M. de la Llera-Moya M. Rodrigues A. Burke M.F. Jafri K. French B.C. Phillips J.A. Mucksavage M.L. Wilensky R.L. Cholesterol efflux capacity, high-density lipoprotein function, and atherosclerosis.N. Engl. J. Med. 2011; 364: 127-135Crossref PubMed Scopus (1532) Google Scholar, B.F. de la Llera-Moya M. Rothblat G.H. of HDL on cellular and cholesterol Lipid Res. Full Text Full Text PDF PubMed Scopus Google Scholar) and be a for of by the radiolabeled cholesterol nanoparticles in the liver by was than the on the uptake by Zilversmit (11Schwartz C.C. Berman M. Vlahcevic Z.R. Halloran L.G. Gregory D.H. Swell L. Multicompartmental analysis of cholesterol metabolism in man. Characterization of the hepatic bile acid and biliary cholesterol precursor sites.J. Clin. Invest. 1978; 61: 408-423Crossref PubMed Scopus (133) Google Scholar), was due to liver uptake may have been due to the increase in tracer observed in the blood of mice with that a of the present in the nanoparticles not may have been in the plasma, with blood and with with uptake by the is that in the of of the cholesterol nanoparticles may have been by liver with this et al. have recently demonstrated that nanoparticles, similar in to the cholesterol nanoparticles used in studies, primarily by but of is an increase in the uptake by T. A. distribution of in the PubMed Scopus Google Scholar). rapid reappearance of the tracer that of the radiolabeled cholesterol by the system cells not and and is the and rapidly to be to acceptor is of tracer between the cells and the as by the that the of tracer in the liver cells over the of the but in the suggesting that the tracer was and with cholesterol et al. macrophage-specific RCT in rats using cholesteryl in were preferentially by the was and in to be in the bile S. D. A. from in vivo rapidly in cells and to liver cells and Google Scholar). We the use of the preparation of for human use is not to use cholesterol nanoparticles for the of the RCT was by the of preparation for human use and the of not only from studies by Nilsson and Zilversmit (10Nilsson A. Zilversmit D.B. Fate of intravenously administered particulate and lipoprotein cholesterol in the rat.J. Lipid Res. 1972; 13: 32-38Abstract Full Text PDF PubMed Google Scholar), but also from human by Schwartz et al. (11Schwartz C.C. Berman M. Vlahcevic Z.R. Halloran L.G. Gregory D.H. Swell L. Multicompartmental analysis of cholesterol metabolism in man. Characterization of the hepatic bile acid and biliary cholesterol precursor sites.J. Clin. Invest. 1978; 61: 408-423Crossref PubMed Scopus (133) Google Scholar, 12Schwartz C.C. Vlahcevic Z.R. Berman M. Meadows J.G. Nisman Swell L. role of lipoprotein in plasma free cholesterol Clin. Invest. PubMed Scopus Google Scholar, C.C. Cholesterol in subjects with bile relationship between of the bile acid precursor and bile acid Clin. Invest. PubMed Scopus Google Scholar, C.C. cholesteryl ester and in vivo in Lipid Res. Full Text Full Text PDF PubMed Scopus Google Scholar). support the concept that the to albumin in nanoparticles was with blood the use of cholesterol nanoparticles as a method to assess RCT in we used interventions that known to promote R. D. J. J. R. J. of a novel liver X receptor agonist with and a in 2015; PubMed Scopus Google Scholar, M. Billheimer Rothblat G.H. Rader D.J. of liver X reverse cholesterol transport in 2006; 113: PubMed Scopus Google Scholar) and J. S. S. cholesterol and rapidly and macrophage in for PubMed Scopus Google Scholar, S. S. Murphy J. A. D. high-density lipoprotein plasma high-density lipoprotein and cholesterol efflux capacity in with Am. 2009; PubMed Scopus Google Scholar). Administration of cholesterol nanoparticles in mice that received was associated with an increase in plasma counts as compared with animals that received results were observed in mice treated with an agonist as compared with animals that received an increase in was not observed in and mice, suggesting that the of cholesterol of macrophages is primarily by in this is have been shown to cholesterol efflux in response to in macrophages in L. J.A. X receptor in for therapeutic 2006; PubMed Scopus Google Scholar, E.G. J. S. M. of liver X receptor and in of an and a agonist in mice in 2006; PubMed Scopus Google Scholar). the of cholesterol efflux in mice be in of the of and and efflux and However, in studies in to and the efflux pathways in with and more Billheimer Rader D.J. Phillips M.C. Rothblat G.H. of pathways in the of cholesterol from Lipid Res. Full Text Full Text PDF PubMed Scopus Google Scholar). that have cholesterol nanoparticles in vivo may also efflux cholesterol primarily and of by the agonist and studies support the concept that this method is to assess in RCT. is also that a increase in macrophage RCT by an agonist was by using this method and that the for of atherosclerosis in J. J. G. D. R. and of an agonist on human and lipoprotein metabolism and circulating Full Text Full Text PDF PubMed Scopus Google Scholar). Data from the study in healthy consistent with reported by Schwartz et al. (11Schwartz C.C. Berman M. Vlahcevic Z.R. Halloran L.G. Gregory D.H. Swell L. Multicompartmental analysis of cholesterol metabolism in man. Characterization of the hepatic bile acid and biliary cholesterol precursor sites.J. Clin. Invest. 1978; 61: 408-423Crossref PubMed Scopus (133) Google Scholar, 12Schwartz C.C. Vlahcevic Z.R. Berman M. Meadows J.G. Nisman Swell L. role of lipoprotein in plasma free cholesterol Clin. Invest. PubMed Scopus Google Scholar, C.C. Cholesterol in subjects with bile relationship between of the bile acid precursor and bile acid Clin. Invest. PubMed Scopus Google Scholar, C.C. cholesteryl ester and in vivo in Lipid Res. Full Text Full Text PDF PubMed Scopus Google Scholar). the tracer after the specific of the 3H-cholesterol and was initially in the HDL than in suggesting that tracer reappearance in is initially by the HDL These consistent with a in the tracer is in HDL is with C.C. cholesteryl ester and in vivo in Lipid Res. Full Text Full Text PDF PubMed Scopus Google Scholar). of the tracer showed a to that was observed in the of reappearance may on the ability of the to efflux as shown in mice by the increase in the after the of the as as the of the acceptor as shown by the increase in the plasma tracer after the of of of the rate of reappearance was due to the rapid of by and the with of the tracer as in HDL increased rapidly and linearly the of may an of the ability to efflux HDL and to the in the HDL was and the specific was in the HDL as compared with the nonHDL Multicompartmental an of the RCT process and an of the results obtained. used is consistent with the cholesterol developed by Schwartz et al. C.C. cholesteryl ester and in vivo in Lipid Res. Full Text Full Text PDF PubMed Scopus Google Scholar), and the results consistent with has been reported C.C. Cholesterol in subjects with bile relationship between of the bile acid precursor and bile acid Clin. Invest. PubMed Scopus Google Scholar, C.C. cholesteryl ester and in vivo in Lipid Res. Full Text Full Text PDF PubMed Scopus Google Scholar). interest is the multicompartmental we an in vivo of the key steps of RCT that the reappearance of the tracer in the fractional transfer rate of the tracer as from the system compartment to HDL by the ability of the macrophage to efflux cholesterol and the of the the fractional transfer rate of the tracer present in HDL from to is and the fractional transfer rate of the nonHDL cholesteryl ester transfer of RCT steps be an for the of novel therapeutic approaches to RCT. We demonstrated after the of cholesterol nanoparticles, radiolabeled tracer is and be in Although counts in fecal were over the we observed a for stool as as in may have a role and to be In the the use of cholesterol nanoparticles as a method to assess RCT in in vivo and a for the use of this approach in humans. results of the human study support the and of this approach and suggest that this method may be used to measure the rate of macrophage RCT in vivo in humans. cholesteryl ester free cholesterol HDL cholesterol intravenous liver X receptor reverse cholesterol transport reticuloendothelial