Triglyceride enrichment of HDL does not alter HDL-selective cholesteryl ester clearance in rabbits

Transcription

1 Triglyceride enrichment of HDL does not lter HDL-selective cholesteryl ester clernce in rits Shiry Rshid,* Kristine D. Uffelmn,* P. Hugh R. Brrett, Polo Vicini, Khosrow Adeli,** nd Gry F. Lewis*,1 Deprtment of Medicine, * Division of Endocrinology, University of Toronto, Toronto, Ontrio, Cnd M5G 2C4; Deprtment of Medicine, University of Western Austrli, Perth 6847, Austrli; Resource Fcility for Popultion Kinetics, Deprtment of Engineering, University of Wshington, Settle, WA 98195; nd Deprtment of Lortory Medicine nd Pthoiology, ** Hospitl for Sick Children, University of Toronto, Toronto, Ontrio, Cnd M5G 1X8 Astrct Triglyceride (TG) enrichment of high density lipoprotein (HDL), which occurs in hypertriglyceridemic sttes, significntly enhnces the rte t which HDL polipoprotein (po)a-i is clered from the circultion of helthy humns. In the New Zelnd White (NZW) rit, species nturlly deficient in heptic lipse (HL), TG enrichment of HDL requires prior lipolytic modifiction to enhnce poa-i clernce. However, the effect of TG enrichment of HDL on the susequent clernce of HDL cholesteryl ester (CE) hs not previously een exmined in vivo. Therefore, we investigted, in the NZW rit, the effect of ex vivo TG enrichment of rit HDL (y incution with humn very low density lipoprotein) on the clernce of HDL CE nd poa-i rdioleled with 3 H-cholesteryl oleyl ether nd with 131 I, respectively. In nine experiments, TG enrichment of rit HDL resulted in n 87% verge increse in HDL TG nd corresponding 31% reduction in HDL CE content. The clculted poa-i nd CE frctionl ctolic rtes ssocited with TG-rich versus fsting HDL trcers were not significntly different (poa-i: vs pools per h, P 0.68; CE: vs pools per h, P 0.20). In n niml model deficient in HL, TG enrichment of HDL did not lter the rtes of HDL poa-i or selective CE clernce. Further studies re needed to determine whether, in the presence of HL, TG enrichment of HDL lters selective HDL CE clernce. Rshid, S., K. D. Uffelmn, P. H. R. Brrett, P. Vicini, K. Adeli, nd G. F. Lewis. Triglyceride enrichment of HDL does not lter HDL-selective cholesteryl ester clernce in rits. J. Lipid Res : chnges in HDL lipid composition secondry to hypertriglyceridemi re n importnt determinnt of plsm HDL-C levels (6). Indeed, we hve previously demonstrted in humns tht triglyceride (TG) enrichment of HDL mrkedly enhnces the clernce of HDL-ssocited polipoprotein (po)a-i from the circultion (7). Most in vivo studies to dte tht hve exmined the effect of hypertriglyceridemi on HDL metolism hve focused on the kinetics of HDL-ssocited poa-l, the mjor protein component of the HDL prticle (6). However, up to 65% of the cholesteryl ester (CE) component of HDL is currently elieved to e selectively removed from HDL prticles y tissues such s liver nd drenl, process tht is independent of HDL whole-prticle uptke (8). The scvenger receptor clss B, type 1 (SR-B1) ws recently discovered to ply key role in the process of selective CE uptke from HDL (9, 10). The hypertriglyceridemirelted determinnts of HDL CE metolism my therefore not e the sme s those for poa-i, mking it impertive tht the effects of hypertriglyceridemi on HDL CE metolism e studied directly. Recent in vitro studies investigting the effect of TG enrichment of HDL, s occurs in hypertriglyceridemic sttes, on the process of selective CE uptke hve shown opposing results. In one study, TG enrichment of humn HDL ws found to impir selective uptke of HDL CE in HepG2 nd drenl YI cells (11). In nother study, TG-enriched Supplementry key words polipoprotein A-I selective cholesterol clernce high density lipoprotein There is strong inverse correltion etween plsm levels of high density lipoprotein cholesterol (HDL-C) nd the development of therosclerotic crdiovsculr disese (1, 2). Hypertriglyceridemi is commonly ssocited with low plsm levels of HDL-C (3 5). Although the moleculr mechnism ccounting for this ssocition is not entirely cler, we nd other reserchers hve postulted tht Arevitions: po, polipoprotein; CE, cholesteryl ester; CETP, cholesteryl ester trnsfer protein; FCR, frctionl ctolic rte; HDL, high density lipoprotein; HDL-C, high density lipoprotein cholesterol; HL, heptic lipse; PAGGE, polycrylmide grdient gel electrophoresis; SDS-PAGE, sodium dodecyl sulfte polycrylmide gel electrophoresis; SR-B1, scvenger receptor type B, clss 1; TCA, trichlorocetic cid; TG, triglyceride; TLC, thin-lyer chromtogrphy; NZW, New Zelnd White; VLDL, very low density lipoprotein. 1 To whom correspondence should e ddressed t Toronto Generl Hospitl, 200 Elizeth St., Room EN11-229, Toronto, Ontrio, Cnd M5G 2C4. e-mil: gry.lewis@uhn.on.c Journl of Lipid Reserch Volume 42,

2 HDL isolted from CE trnsfer protein (CETP)/poA-I trnsgenic mice showed significnt increse in SR-B1- medited HDL CE uptke in Chinese hmster ovry cells compred with control HDL isolted from poa-i trnsgenic mice (12). The uthors of the second study ttriuted the increse in HDL CE uptke to remodeling of HDL y CETP-medited CE-TG interchnge. An importnt question, which hs yet to e nswered, is whether TG enrichment of HDL lso modultes the process of selective CE clernce in vivo. Therefore, in the present study, we exmined the influence of ex vivo TG enrichment of rit HDL prticles in determining the susequent clernce of oth HDL poa-l nd HDL CE from the circultion of norml (i.e., wild-type) rits. We chose the rit s the model to study the metolism of HDL in hypertriglyceridemic sttes ecuse rits re deficient in the enzyme heptic lipse (HL) (13). The lipolytic nd nonlipolytic effects of HL on HDL metolism hve recently een extensively reviewed (14). In the present study, we wished to investigte the effect of TG enrichment of HDL per se on the clernce of HDL CE, independent of the effects of HL. Our specific im ws to determine whether TG enrichment of HDL, similr to tht oserved in hypertriglyceridemic sttes, without sustntil in vivo hydrolysis nd enhncement of uptke y HL, is sufficient to enhnce the clernce of HDL CE, s compred with the clernce of poa-i in the HDL prticle. MATERIALS AND METHODS Animls New Zelnd White (NZW) mle rits, 3.95 to 5.5 kg in weight, were used in totl of nine experiments. All rits received stndrd chow-fed diet (17% protein, 2.1% ft, 18% fier). The study protocol comprised two phses: In phse 1, HDL from fsting rit serum nd from serum tht hd een enriched with TG y incution with humn very low density lipoproteins (VLDL; see elow) ws isolted nd rdioleled. In the second phse of the study, the two HDL trcers (fsting nd TG-enriched) were injected into two weight-mtched recipient rits nd the clernce of poa-i nd CE ssocited with ech trcer ws determined. Phse 1: Isoltion nd rdioleling of fsting nd TG-enriched rit HDL Blood ws otined vi crdic puncture from mle NZW rits following n overnight fst (16 h) nd sedtion with 0.2 ml/ kg ody weight of ketmine (Rogr/STB Phrmceuticls, London, Ontrio, Cnd) nd 0.2 ml/kg ody weight of xylzine (Byer, Etoicoke, Ontrio, Cnd). Individul donor rits were used for ech experiment. Next, 13 ml of rit serum ws incuted ex vivo with humn VLDL for 4 h t 37 C to enrich the HDL prticles with triglycerides. VLDL ws otined immeditely prior to ech experiment from helthy normolipidemic donor humn suject. The mount of VLDL included in the incution mix ws sufficient to contriute 1 mm of triglycerides. Fsting serum nd serum incuted with VLDL were then incuted for 2 h t 37 C with 3 H-cholesteryl oleyl ether (2 5 Ci/ml of serum) to whole-lel HDL CE. CETP within the rit serum medited the exchnge of 3 H-cholesteryl oleyl ether from VLDL to HDL CE. Fsting nd TG-rich HDL trcers were isolted y sequentil ultrcentrifugtion of whole serum t consecutive densities of g/ml for 20 h t 40,000 rpm nd 1.21 g/ml for 24 h t 55,000 rpm t 4 C in Beckmn 70.1 Ti rotor, nd then dilyzed overnight in uffer contining 0.15 M NCl, 0.01 M Trisse, nd 5 mm ethylenediminetetrcetic cid, ph 8.0 (herefter referred to s Tris uffer). Approximtely 1 3 mg of fsting protein nd TG-rich HDL were iodinted y modifiction of the iodine monochloride method of McFrlne (15) using 250 Ci of 131 I. Briefly, HDL ws first dded to mixture of 20 mol ICl/ mol HDL protein nd 250 Ci of rdioctive iodine. The mixture ws then eluted through 0.2 M Glycine-Sephdex G-50 column, ph 10 (Phrmci Biotech, Uppsl, Sweden) to llow for the iodintion of HDL nd the seprtion of the HDL-ound rdioctive iodine from the free iodine. The rdioleled HDL ws then eluted through sline-sephdex G-50 column to ring the ph of the solution contining the rdioleled HDL ck to physiologicl rnge (ph 7.4). An equl mount of unleled TG-rich nd fsting HDL ws dded to the rdioiodinted HDL frctions s cold crrier ( mg HDL protein) nd HDL trcers were wshed t d 1.21g/ml, 55,000 rpm, 4 C, in 70.1 Ti rotor for 24 h, followed y dilysis in Tris uffer t 4 C. The composition of ech rdioleled trcer ws mesured using commercilly ville enzymtic ssy kits s descried elow. The size nd integrity of the fsting nd TG-rich HDL trcers were lso nlyzed y 4 30% nondenturing polycrylmide grdient gel electrophoresis (PAGGE). HDL prticle size ws determined from the migrtion of stndrd moleculr weight proteins of known dimeter (HMW Clirtion Kit, Phrmci, Pisctwy, NJ). Pek nd men rdius of smll nd lrge HDL prticles were determined y densitometric scnning (Imge Mster DTS densitometer with Imgemster computer softwre, Phrmci LKB, Uppsl, Sweden) sed on reltive migrtion distnce (R f ) of stndrds nd using two distinct pproches. First, the estimted rdius of the mjor pek in ech scn ws identified s the HDL pek prticle size. A men (or weighted) HDL prticle size ws lso clculted using modifiction of the method descried y Li et l. (16). The men HDL prticle size for ech suject ws otined y multiplying the size of ech nd y its frctionl re. This men HDL prticle rdius (in nm) therefore comines the HDL size distriution s well s the reltive concentrtion of ech HDL nd (16). Phse 2: HDL turnover study An liquot of ech trcer contining mg HDL protein, ( 10 7 ) cpm of 131 I-HDL, nd ( 10 6 ) dpm of 3 H-cholesteryl ether ws injected simultneously into the right mrginl er vein, the TG-rich trcer into one mle NZW rit, nd the fsting HDL trcer into nother. Blood smples (2 ml) were otined over the next 3 dys from vein in the opposite er t the following time intervls: 10 min, nd 1, 2, 3, 4, 6, 24, 27, 30, 48, 51, nd 54 h. Isoltion of HDL nd poa-i for rdioctivity counting Serum ws immeditely seprted from red cells y centrifuging t 3,000 rpm for 20 min t 4 C. HDL ws isolted (from 1) ml serum) y sequentil ultrcentrifugtion s follows: 1) the serum ws first djusted to density of g/ml nd centrifuged for 5 h t 100,000 rpm t 4 C in n Optim TLX ultrcentrifuge (Beckmn Instruments, Plo Alto, CA) using TLA rotor, 2) the ottom two-thirds frction from this spin ws djusted to density of 1.21 g/ml nd centrifuged for 5 h t 100,000 rpm, nd 3) the top one-third frction from the centrifugtion ws wshed for further 5 h t 100,000 rpm. Aliquots of totl serum nd of the d g/ml HDL nd d 1.21 g/ml 266 Journl of Lipid Reserch Volume 42, 2001

3 serum frctions were tken to mesure the proportion of 131 I protein nd 3 H-cholesteryl ether rdioctivity found in ech of these frctions. To determine 3 H-cholesteryl oleyl ether rdioctivity, ech smple ws counted in Redy Solve liquid scintilltion cocktil (Beckmn, Fullerton, CA) using Beckmn 6500 liquid scintilltion counter. All 3 H-cholesteryl oleyl ether rdioctivity counts were djusted ck to totl serum rdioctivity in dpm/ml. In preliminry experiment, delipidted HDL trcers were nlyzed using thin-lyer chromtogrphy (TLC). It ws found tht greter thn 97% of oth fsting nd TG-rich HDL lipid trcer rdioctivity co-migrted with cholesteryl ether on TLC. To mesure the 131 I rdioctivity specificlly ssocited with poa-i, 500- l liquots of the dilyzed HDL were delipidted s previously descried (17) nd dissolved in 0.5 M phosphte uffer with 1% SDS, 1% mercptoethnol, ph 7.2, overnight. The smples were then run on 15% sodium dodecyl sulfte polycrylmide gel electrophoresis (SDS-PAGE) to isolte poa-i. The polipoproteins were stined overnight with R250 Comssie lue colloidl stin nd destined for 2 h in wter, nd the A-I nd ws sliced from the gel. Rdioctivity in ech serum frction nd in the poa-i nd on the gel ws counted in Beckmn 5500 gmm counter. The poa-i rdioctivity in the gel ws tken s percentge of the rdioctivity found in tht gel nd djusted ck to totl serum rdioctivity in cpm/ml. All rdioctivity counts were corrected for the hlf-life of the isotope nd djusted ck to the time of injection. In ddition to the use of 15% SDS-PAGE, the injecttes were lso nlyzed for the percentge of 131 I counts in the vrious polipoproteins using isoproprnol precipittion nd isoelectric focusing techniques. The percentge of rdioctivity of the injecttes present in ound versus free form ws lso nlyzed using trichlorocetic cid (TCA) precipittion. Lortory mesurements Cholesterol ws mesured using the CHOD-PAP enzymtic colorimetric kit (Boehringer Mnnheim GmH Dignostic, Montrel, Cnd). CE ws mesured using the cholesterol kit fter degrding free cholesterol in the smples with mixture of cholesterol oxidse, peroxidse, phenol, triton-x, nd phosphte uffer (18). Protein ws mesured y the technique descried y Lowry et l. (19). Triglycerides were mesured s esterified glycerol using n enzymtic colorimetric kit (Boehringer Mnnheim GmH Dignostic). Free glycerol ws eliminted from the smple in preliminry rection followed y enzymtic hydrolysis of triglyceride with susequent determintion of the lierted glycerol y colorimetry. Phospholipids were mesured using kit (Test-Comintion Phospholipids; Boehringer Mnnheim GmH Dignostic). Kinetic nlysis The rdioctivity die-wy curves were nlyzed using twopool model s previously descried (17, 18). This model ssumes the existence of n intrvsculr pool in equilirium with nonvsculr pool, nd ll losses from the system re ssumed to occur from the intrvsculr pool. For ech rdioleled trcer, popultion kinetic prmeters were computed y popultion kinetic nlysis using n itertive two-stge method (20) sed on using the popultion men t ech itertion s priori for improving the individul prmeter estimtes. Popultion kinetic nlysis tkes into considertion intersuject vriility to estimte kinetic prmeters for group or popultion of individuls (17). The popultion kinetic prmeters clculted s ove were then used to generte men die-wy curves for the fsting nd TG-enriched poa-i nd CE HDL trcers. As previously descried (17), computing popultion kinetics is the process of fitting model to n individul s dt while using sttisticl informtion from ll sujects to help generte the prmeters of the popultion. Popultion kinetic prmeters re similr to men kinetic prmeters otined from individul fitting of sujects in group (for dt-dense situtions) (17), ut unlike men kinetic prmeters, they re otined y tking into ccount the errors ssocited with the individul prmeter estimtes, thus reducing the vriility of the estimtes (17). The estimtes of prmeters otined y the popultion nlysis were compred to the individul fit [simultion, nlysis, nd modeling II (SAAMII)] nd determined to e similr, ut were ssocited with lower coefficients of vrition. The two-pool model provided good fits to HDL poa-i nd CE trcer dt to yield estimtes of the frctionl ctolic rte (FCR). Estimtes of the precision of the individul FCRs, coefficients of vrition (CV), re determined during the fitting process. These CV re functions of the noise in the experimentl dt nd the structure of the model. The verge CV for poa-l FCRs were 11 1% for TG-rich HDL nd 20 4% for fsting HDL. The verge CV for CE FCRs were 15 2% for TG-rich HDL nd 27 4% for fsting HDL. The proportion of CE ctolized vi selective clernce from the HDL frction ws clculted s the difference etween the FCR of HDL CE nd the FCR of 131 I poa-l (21). Sttistics Results re presented s men SEM. Unpired t-tests were performed to compre popultion FCR vlues etween fsting nd TG-rich HDL trcers. Pired t-tests were used to test differences in trcer composition nd size. RESULTS In totl of nine HDL turnover studies, the physiologicl vrition in HDL TG, cholesterol, CE, phospholipid, nd protein verged 9%, 7%, 11%, 9%, nd 6%, respectively, in the nimls receiving fsting trcer; nd 8%, 11%, 17%, 8%, nd 4%, respectively, in the nimls receiving TG-rich trcer. Tle 1 presents the men lipid nd protein composition of the fsting nd TG-rich rit HDL trcers. TG enrichment of rit HDL with humn VLDL resulted in lrge (87%) increse in HDL TG content (P ) in the TG-rich HDL trcer compred with the fsting HDL trcer. TG enrichment lso produced reciprocl reductions TABLE 1. Men lipid nd protein chrcteriztion of the TG-rich HDL nd fsting HDL trcers TG-Rich HDL Fsting HDL % Difference % of HDL weight Triglycerides Cholesterol Esters c Free Phospholipids Proteins Vlues re men SEM. % Difference indictes the reltive chnge in the TG-rich HDL ove or elow the fsting HDL vlues. P compred with fsting HDL. c P Rshid et l. Selective HDL cholesteryl ester clernce in rits 267

4 in HDL totl cholesterol nd CE content [33% (P ) nd 31% (P 0.05), respectively] of the TG-rich HDL. Finlly, HDL phospholipid nd protein concentrtions were not modified to significnt extent y incution with humn VLDL. Anlysis of the HDL trcers pre- nd postiodintion showed tht the iodintion procedure hd no significnt effect on the composition of the prticles (results not shown) or on the integrity of the rdioleled trcers when monitored on 4 30% PAGGE. There ws no significnt difference in either the weighted men or pek sizes of the fsting nd TG-rich HDL trcers (weighted size: vs , respectively; pek size: vs nm, respectively). Both trcers were found to hve more thn 92% of their 131 I rdioctivity in protein-ound form, s ssessed with TCA precipittion. To determine percentge of 131 I counts in the vrious polipoproteins, the injecttes were lso nlyzed using 15% SDS-PAGE, isopropnol precipittion, nd isoelectric focusing techniques. Thirty-seven percent of this rdioctivity ws ound to poa-l [n 9 experiments; P not significnt (NS) etween TG-rich nd fsting HDL trcers], with the reminder of the counts distriuted in pob, poe, nd poc (n 3 experiments; P NS etween TG-rich nd fsting HDL trcers). The proportion of injected 131 I rdioctivity tht ws recovered during the turnover studies within the HDL frction ws similr whether derived from the fsting (50 8%) or the TG-rich HDL (43 5%) trcers. The proportion of serum rdioctivity in the d g/ml nd the d 1.21 g/ml frctions were lso similr (P 0.41 for oth) for the fsting nd the TG-enriched HDL turnover studies (d g/ml: 13 4% nd 12 3%; d 1.21 g/ml: 37 5% nd 46 3%, respectively). The percentge of 3 H-cholesteryl ether rdioctivity recovered in the different lipoprotein frctions ws lso clculted during the turnover studies. For oth fsting HDL nd TG-rich HDL, the mjority of 3 H-cholesteryl ether rdioctivity ws recovered in HDL (69 6% nd 71 6%, respectively) nd the recoveries were similr for oth trcers in the HDL (P 0.79) nd the d g/ml frctions (31 6% nd 29 6%, P 0.79). Effect of TG enrichment of HDL on the FCR of HDL poa-i Fig. 1A shows the die-wy clernce curves of rdioleled HDL poa-i from fsting HDL nd from the TGenriched HDL for one representtive experiment. As descried in Mterils nd Methods, the isotope clernce curves were constructed using the counts in the isolted poa-i nd. The kinetic prmeters (FCR) derived from the clernce curves re presented in Tle 2. Results from nine experiments showed no systemtic difference in the rte of clernce of fsting compred with TG-rich HDL poa-i ( vs pools per h), with n 11.6% greter men FCR of TG-rich versus fsting HDL CE, P 0.68). Fig. 2A illustrtes the die-wy curves for rdioleled fsting nd TG-rich HDL poa-l generted from the estimte of the men popultion prmeters. TABLE 2. Kinetic prmeters of poa-i from TG-rich nd fsting HDL in rits Experiment TG-Rich HDL Fsting HDL FCR h 1 (CV, %) Fig. 1. Die-wy curves of HDL poa-i (A) nd CE (B) rdioctivity from one representtive experiment. Results for the TG-enriched HDL trcer (oserved dt: open tringles; modeled dt: solid line) nd fsting HDL trcer (oserved dt: closed circles; modeled dt: dshed line) re shown. The dt presented on the y xis of grph A represent the rdioctivity on HDL poa-i isolted y gel electrophoresis of the delipidted HDL frction (see Mterils nd Methods). The dt presented on the y xis of grph B represent the CE rdioctivity in the totl HDL frction. The rdioctivity dt for ech trcer were normlized s function of the rdioctivity mesured t the first time intervl (10 min) (8.3) (11.2) (8.0) (17.9) (10.7) (35.0) (15.5) (11.6) (19.4) (37.1) (10.3) (23.3) (8.0) (9.9) (11.1) (7.6) (7.8) (30.6) Men SEM No significnt difference etween fsting nd TG-rich HDL, P CV, %: coefficient of vrition is the stndrd devition divided y the estimted FCR, expressed s percent. The CV is n estimte of the degree of precision of the FCR. 268 Journl of Lipid Reserch Volume 42, 2001

5 TABLE 3. Kinetic prmeters of CE from TG-rich nd fsting HDL in rits Experiment TG-Rich HDL Fsting HDL FCR h 1 (CV, %) (22.3) (26.9) (12.1) (27.4) (22.9) (32.9) (13.3) (52.4) (22.2) (25.5) (11.5) (25.6) (10.4) (19.4) (13.5) (11.2) (10.8) (21.4) Men SEM No significnt difference etween fsting nd TG-rich HDL, P CV, %: coefficient of vrition is the stndrd devition divided y the estimted FCR, expressed s percent. The CV is n estimte of the degree of precision of the FCR. Fig. 2. Simulted die-wy curves of HDL poa-i (A) nd CE (B) rdioctivity. The die-wy curves for the TG-enriched HDL trcer (solid line) nd fsting HDL trcer (dshed line) were generted from the estimted popultion nlysis prmeters. The rdioctivity dt for ech trcer were normlized s function of the rdioctivity mesured t the first time intervl (10 min). Insets show the FCR ssocited with TG-rich nd fsting HDL. Effect of TG enrichment of HDL on the FCR of HDL cholesteryl ether nd on selective CE clernce The individul kinetic dt for HDL cholesteryl ether derived from ech trcer re presented in Tle 3. The die-wy curves of HDL cholesteryl ether from fsting nd TG-enriched HDL re shown in Fig. 1B (one representtive experiment) nd Fig. 2B (simulted curve derived y using men popultion prmeters). The curves were constructed from the totl counts in the HDL pool. In nine experiments, the men percentge difference in clernce of poa-i from TG-rich HDL versus fsting HDL ws 27.8%. This trend towrd greter HDL poa-l FCR resulting from TG enrichment of HDL ws not significnt ( vs pools per h, P 0.20) compred with fsting HDL. The proportion of CE ctolized vi selective clernce from the HDL frction ws clculted s the difference etween the FCR of HDL cholesteryl ether nd the FCR of 131 I poa-l (21) (Tle 4). The rtes of selective CE clernce were not significntly different etween fsting nd TG-rich HDL (P 0.52). DISCUSSION The present study shows tht TG enrichment of HDL, common metolic consequence of hypertriglyceridemi, does not significntly enhnce the metolic clernce rtes of HDL poa-l nd CE in rits, species nturlly deficient in HL (13). Similrly, selective clernce of HDL CE ws not ltered y TG enrichment of HDL. Indeed, n 87% verge increse in the TG content of HDL did not significntly lter HDL poa-l nd CE clernce rtes or selective CE clernce from the HDL frction. There ws trend towrd greter poa-l nd CE FCR with TG-rich compred with fsting rit HDL prticles (11.6% nd 27.7%, respectively), ut this trend did not rech sttisticl significnce. Power clcultions indicte tht the experimentl smple size would need to e incresed to more thn 170 for HDL poa-i nd to 26 for HDL CE to result in sttisticlly significnt greter clernce in the TG-rich compred with fsting HDL. This is in contrst to the highly significnt increse in poa-i FCR in TGenriched HDL, which we recently oserved in study of six humn sujects (7). Whether TG enrichment of HDL TABLE 4. Kinetic prmeters of selective CE clernce from TG-rich nd fsting HDL in rits Experiment TG-Rich HDL Fsting HDL FCR h Men SEM Selective CE clernce FCR cholesteryl ether FCR poa-i. No significnt difference etween fsting nd TG-rich HDL, P Rshid et l. Selective HDL cholesteryl ester clernce in rits 269

6 enhnces selective CE clernce from HDL in humns is currently not known. The finding from the present study indictes tht the increse in HDL TG content lone does not fully explin the increse in HDL poa-i FCR nd ssocited reduction in poa-i nd HDL-C levels oserved in hypertriglyceridemic sttes. There re mny differences in lipoprotein metolism etween rits nd humns, ut given the importnt function of HL in the metolism of HDL (14, 22), we speculte tht HL deficiency in the rit my hve ccounted for lck of oserved difference in the clernce of poa-i nd CE from TG-rich versus fsting HDL. This hypothesis will need to e tested in future experiments. Further indirect evidence tht the sence of HL in the rit my hve ccounted for the lck of effect of HDL TG enrichment in enhncing HDL poa-i nd CE clernce comes from recent study y Lmert et l. (23). These uthors showed tht HL promotes the selective uptke of HDL CE vi SR-B1. Furthermore, the smll clculted difference we oserved etween HDL poa-i FCR nd HDL CE FCR (selective CE clernce) in the present study for oth TG-rich nd fsting HDL trcers my hve een due to the nturl deficiency of HL in the rit. In previous study, Golderg, Beltz, nd Pittmn (21) estimted tht only 20% of HDL CE is clered y selective uptke in the rit, fr smller percentge thn hs een oserved in other species. In ddition, the uthors of tht study, similr to the present study, oserved wide vriility in the contriution of the selective uptke pthwy to cellulr uptke of HDL CE (from 0% to 47% of the totl). They ttriuted the vriility in selective uptke to widely vrying CETP ctivities nd TG levels in rits, oth of which correlted positively with selective HDL CE uptke (21). In ddition to the mrked increse in TG content of TG-rich compred to fsting HDL trcers used in the rit kinetic studies, there were significnt reciprocl reductions in totl cholesterol, CE, nd free cholesterol mss in the TG-rich HDL. CETP in serum medites the idirectionl exchnge of TG nd CE etween HDL nd the more uoynt lipoproteins (VLDL, LDL, nd CM) (24), process tht is enhnced in hypertriglyceridemi (25). Overll, this study ws designed to test the effect of the TG content of HDL prticles on HDL metolism. Besides the norml physiologicl depletion of HDL CE ssocited with TG enrichment of HDL, other mesured components of HDL composition remined unltered. The fsting nd TG-rich HDL trcers were similr in size (oth men pek nd weighted men sizes) nd contined similr percentge mss of totl protein nd phospholipid. Use of intrlipid, synthetic TG emulsion, to enrich HDL prticles with TG in our previous kinetic studies in humns nd rits led to ltertions in the phospholipid content of HDL (7). Hence, to void this chnge in phospholipid content ssocited with intrlipid, we used VLDL from normotriglyceridemic humn sujects to enrich HDL with TG in the present study. Thus, the physiologicl relevnce of the trcers ws n importnt considertion in our present investigtion. A numer of mechnisms hve een proposed to explin the reduction in serum HDL cholesterol levels nd serum poa-i in hypertriglyceridemic sttes (6). These mechnisms include thermodynmic instility of TG enriched nd CE-depleted HDL prticles, irreversile loss of poa-i from HDL during lipolysis of the TG-rich HDL prticle y HL, nd rpid clernce from the circultion of the lipolytic end-product smller HDL prticle, to mention few (6). Nonetheless, s the present study demonstrtes, TG enrichment of HDL prticles, y incuting rit serum with humn VLDL, ws not sufficient to enhnce serum HDL poa-l clernce in this model. Our previous work in rits (17, 18) s well s tht of others (26) hs indicted tht TG enrichment, without sustntil lipolysis of the prticle, is insufficient to enhnce the clernce of HDL-ssocited poa-l. Our previous studies were confounded, however, y the fct tht TG-enriched HDL ws derived from hypertriglyceridemic humn sujects, nd their ctolism susequently trced in rit model. The present study voids tht confounding fctor, nd yet it confirms our previous findings tht TG enrichment per se, without sustntil susequent lipolytic modifiction of the HDL prticle, does not enhnce HDLssocited poa-i clernce. Moreover, the effect of TG enrichment of HDL on selective HDL CE clernce hs not een exmined in vivo prior to this study. Selective HDL CE clernce hs een identified s mjor pthwy in the reverse cholesterol trnsport of cholesterol from peripherl tissues to the liver nd drenl (21), nd therey is elieved to ply n importnt role in the crdioprotective effect of HDL. Recently, it hs een shown in vitro in HepG2 cells tht TG enrichment of HDL, with concomitnt reduction in HDL CE content, impirs the process of selective CE uptke in dose-dependent mnner (11). However, tht study, in contrst to the present study, ws conducted in the presence of HL ctivity. HL is well known to ply key role in the process of selective HDL CE uptke. HL ound to proteoglycn sulfte t heptocytes promotes the selective uptke of HDL CE y numer of mechnisms (14), including vi the SR- B1 receptor (23, 27). Indeed, Collet et l. (12) demonstrted tht HDL CE uptke in vitro is enhnced in TGenriched humn HDL upon tretment with HL. CONCLUSION Results of the present study suggest tht TG enrichment of HDL without susequent ction y HL is not sufficient to lter the rtes t which HDL poa-i nd CE re clered from the circultion. In ddition, selective HDL CE clernce is not ltered with TG enrichment of HDL in n HLdeficient species. Therefore, chnges in the TG content of HDL do not, y themselves, ccount for the deleterious effects of hypertriglyceridemi on HDL cholesterol metolism. Future studies re required to determine more precisely how the structure nd composition of HDL, HDLmodifying enzymes, nd HDL receptors interct to lter HDL poa-l, nd CE clernce in hypertriglyceridemic sttes. 270 Journl of Lipid Reserch Volume 42, 2001

7 This work ws supported y n operting grnt to G.F.L., Creer Investigtor of the Hert nd Stroke Foundtion of Cnd nd Cnd Reserch Chir, from the Hert nd Stroke Foundtion of Ontrio. S.R. is Ph.D. student supported y the Medicl Reserch Council of Cnd. P.H.R.B. nd P.V. re supported y Ntionl Institutes of Helth grnt NCRR Mnuscript received 12 July 2000 nd in revised form 20 Octoer REFERENCES 1. Miller, G. J., nd N. E. Miller Plsm high density lipoprotein concentrtion nd development of ischemic hert disese. Lncet. 1: Durrington, P. N How HDL protects ginst therom. Lncet. 342: Lmrche, B., J. P. Despres, M. C. Pouliot, D. Prud homme, S. Moorjni, P. J. Lupien, A. Ndeu, A. Tremly, nd C. Bouchrd Metolic heterogeneity ssocited with high plsm triglyceride or low HDL cholesterol levels in men. Arterioscler. Throm. 13: Lmrche, B., J. P. Despres, S. Moorjni, B. Cntin, G. R. Dgenis, nd P. J. Lupien Triglycerides nd HDL-cholesterol s risk fctors for ischemic hert disese. Results from the Queec crdiovsculr study. Atherosclerosis. 119: Brinton, E. A., S. Eisenerg, nd J. L. Breslow Humn HDL cholesterol levels re determined y poa-i frctionl ctolic rte, which correltes inversely with estimtes of HDL prticle size. Effects of gender, heptic nd lipoprotein lipses, triglyceride nd insulin levels, nd ody ft distriution. Arterioscler. Throm. 14: Lmrche, B., S. Rshid, nd G. Lewis HDL metolism in hypertriglyceridemic sttes: n overview. Clin. Chim. Act. 286: Lmrche, B., K. D. Uffelmn, A. Crpentier, J. S. Cohn, G. Steiner, P. H. R. Brrett, nd G. F. Lewis Triglyceride enrichment of HDL enhnces in vivo metolic clernce of HDL po A-I in helthy men. J. Clin. Invest. 103: Glss, C., R. C. Pittmn, D. B. Weinstein, nd D. Steinerg Dissocition of tissue uptke of cholesterol ester from tht of poprotein A-I of rt plsm high density lipoprotein: selective delivery of cholesterol ester to liver, drenl, nd gond. Proc. Ntl. Acd. Sci. USA. 80: Rigotti, A., B. Trigtti, J. Bit, M. Penmn, S. Xu, nd M. Krieger Scvenger receptor B1 cell surfce receptor for high density lipoprotein. Curr. Opin. Lipidol. 8: Lndschulz, K. T., R. K. Pthk, A. Rigotti, M. Krieger, nd H. H. Hos Regultion of scvenger receptor, clss B, type I, high density lipoprotein receptor, in liver nd steroidogenic tissues of the rt. J. Clin. Invest. 98: Greene, D. J., J. W. Skeggs, nd R. E. Morton. Nov. 6, Elevted triglyceride content diminishes the cpcity of HDL to deliver cholesteryl esters vi the scvenger receptor BI (SR-BI). J. Biol. Chem /jc.M Collet, X., A. R. Tll, H. Serjuddin, K. Guendouzi, L. Royer, H. Oliveir, R. Brrs, X. C. Xing, nd O. L. Frncone Remodeling of HDL y CETP in vivo nd y CETP nd heptic lipse in vitro results in enhnced uptke of HDL CE y cells expressing scvenger receptor B-I. J. Lipid Res. 40: Fn, J., J. Wng, A. Bensdoun, S. J. Luer, Q. Dng, R. W. Mhley, nd J. M. Tylor Overexpression of heptic lipse in trnsgenic rits leds to mrked reduction of plsm high density lipoproteins nd intermedite density lipoproteins. Proc. Ntl. Acd. Sci. USA. 91: Sntmrin-Fojo, S., C. Hudenschild, nd M. Amr The role of heptic lipse in lipoprotein metolism nd therosclerosis. Curr. Opin. Lipidol. 9: McFrlne, A. S Efficient trce lelling of proteins with iodine. Nture. 182: Li, Z., J. R. McNmr, J. M. Ordovs, nd E. J. Schefer Anlysis of high density lipoproteins y modified grdient gel electrophoresis method. J. Lipid Res. 35: Lewis, G. F., B. Lmrche, K. D. Uffelmn, A. C. Hetherington, L. W. Szeto, M. A. Honig, nd P. H. R. Brrett Clernce of postprndil nd lipolyticlly-modified humn HDL in rits nd rts. J. Lipid Res. 38: Lmrche, B., K. D. Uffelmn, G. Steiner, P. H. R. Brrett, nd G. F. Lewis Anlysis of prticle size nd lipid composition s determinnts of the metolic clernce of humn high density lipoproteins in rit model. J. Lipid Res. 39: Lowry, O. H., N. J. Roserough, A. L. Frr, nd R. J. Rndll Protein mesurement with the Folin phenol regent. J. Biol. Chem. 193: Steimer, J. L., A. Mllet, J. L. Golmrd, nd J. F. Boisvieux Alterntive pproches to estimtion of popultion phrmcokinetic prmeters: comprison with the nonliner mixed-effect model. Drug Met. Rev. 15: Golderg, D. I., W. F. Beltz, nd R. C. Pittmn Evlution of pthwys for the cellulr uptke of high density lipoprotein cholesterol esters in rits. J. Clin. Invest. 87: Cohen, J. C., G. L. Veg, nd S. M. Grundy Heptic lipse: new insights from genetic nd metolic studies. Curr. Opin. Lipidol. 10: Lmert, G., M. B. Chse, K. Dugi, A. Bensdoun, H. B. Brewer, nd S. Sntmrin-Fojo Heptic lipse promotes the selective uptke of high density lipoprotein-cholesteryl esters vi the scvenger receptor BI. J. Lipid Res. 40: Morteon, R. E., nd D. B. Zilversmit Interreltionship of lipids trnsferred y the lipid trnsfer protein isolted from humn lipoprotein-deficient plsm. J. Biol. Chem. 258: Murkmi, T., S. Michelgnoli, R. Longhi, G. Ginfrnceschi, F. Pzzucconi, L. Clresi, C. R. Sirtori, nd G. Frnceschini Triglycerides re mjor determinnts of cholesterol esterifiction/ trnsfer nd HDL remodeling in humn plsm. Arterioscler. Throm. Vsc. Biol. 15: Horowitz, B. S., I. J. Golderg, J. Mer, T. M. Vnni, R. Rmkrishnn, nd H. N. Ginserg Incresed plsm nd renl clernce of n exchngele pool of polipoprotein A-I in sujects with low levels of high density lipoprotein cholesterol. J. Clin. Invest. 91: Amr, M. J. A., K. A. Dugi, C. C. Hudenschild, R. D. Shmurek, B. Foger, M. Chse, A. Bensdoun, R. F. Hoyt, H. B. Brewer, nd S. Sntmrin-Fojo Heptic lipse fcilittes the selective uptke of cholesteryl esters from remnnt lipoproteins in poedeficient mice. J. Lipid Res. 39: Rshid et l. Selective HDL cholesteryl ester clernce in rits 271