Cholesterol-loading of peripheral tissues alters the interconversion of high density lipoprotein subfractions in rabbits

ht. J. Biochem.

1357-2725@5)00127-1

Cell Bid. Vol. 28, No. 2, pp. 151-163,

1996 Copyright 0 1996 Else&r Science Ltd Printed in Great Britain. All rights reserved 1357-2725/96 $15.00 + 0.00

Cholesterol-loading of Peripheral Tissues Alters the Interconversion of High Density Lipoprotein Subfractions in Rabbits YARA D. FRAGOSO, E. ROY SKINNER* Department of Molecular and Cell Biology, University of Aberdeen, Marischal College, Aberdeen AB9 IAS, U.K. Higb density lipoprotein (HDL) has been implicated in the process of reverse choleste.rol transport, by which surplus cholesterol is removed from peripheral tissuesaad tramqorkd to the liver for excretion. It has been saggestedthat some sabfractlons of HDL may have a particnlar role ln thii process, though the onderl*g meclmnism remains anclear. The present stady was aimed at investigating the role of specik subfractions of HDL in reversecholesterol transport. The hterconverslon of HDL sobfractions in normal and chokterol-loaded rabbits was studled in ho. Rabbit HDL was separated by hepari&kpharose afIbdty chromatography into six subfractions (HDL&IDL&, which were progreaslvelyenriched with apollpoproteln E (apo E), and varied in diameter and composition. Total HDL and its sabfractk+ were hullvkally labeiled with “C sucrose and h&c&l in the rabbits. When rabbits wblch were sot acutely loaded with [%QAdesterol were injected with “C-HDQn, 70% of the iabel remained la &is fraction while lessthan 5% was recovered in HDLWo, containh8g the hugest particles and those most enriched in apo E. No label was detectable in the liver of these animals. In rabbits which had received a prior loading of cholesterol, an average of only 18.3% of the “C label was present ia HDLo, while approx. 40% of the label was recovered in HDL,,. On average, 5.1% of the’ total “C injected in these rabbits was recovered in the liver. It ls concluded that two alternative roates for reverse cholesterol transport may be operative. While a contimmas cholesterol-clearanceroute may be provided by particles of HDL of intermediate size, another roate may be operative for clearance of excesscholesterol loaded into peripheral endothelial cells. Keywords: High density lipoprotein Cholesterol Apolipoprotein E Int. J. Biochem.

Reverse cholesterol transport

HDL subfractions

Cell Biol. (1996) 23, 151-163

INTRODUCTION

Evidence is accumulating that reverse cholesterol transport (Glomset, 1968; Barter, 1993), the process by which excess cholesterol is removed from peripheral tissues and transported to the liver for excretion, is dependent on high density lipoprotein (HDL). HDL may therefore provide the only pathway for the removal of cholesterol from arterial walls, thus preventing the development of atherosclerotic lesions. Several epidemiological investigations have established an inverse relationship between *To whom all correspondence should be addressed. Received 24 Mar .h 1995; accepted 18 September 1995.

plasma levels of HDL and the incidence of coronary heart disease (Gordon and Bifkind, 1989; Wilson et al., 1990). More recently, specific subfractions of HDL have been implicated in the epidemiological findings of coronary and carotid protection. These are the larger HDL particles (Cheung et al., 1991; Johansson et al., 1991; Williams et al., 1992), rich in apolipoprotein E (apo E) (Wilson et al., 1993). Using an in vitro system, the efhuxed cholesterol from peripheral cells has been reported to be taken up by the small, pre-beta migrating HDL particles (Castro and Fielding, 1988; Francone et al., 1989). After esterification by the 151

152

Yara D. Fragoso and E. Roy Skinner

action of lecithin : cholesterol acyltransferase (LCAT), the cholesterol is transferred by the action of cholesteryl ester transfer protein (CETP) to low density lipoprotein (LDL) for hepatic uptake by the LDL receptor. Although widely accepted, this theory does not include the evidence of the direct binding of HDL to the hepatocytes for delivery of cholesteryl esters. Current evidence suggests that hepatocytes are capable of direct uptake of the cholesterol moiety of HDL particles by receptor(s) other than the LDL receptor (Chacko, 1982; Glass et al., 1985; Fidge and Nestel, 1985; Richard and Pittman, 1993). Such uptake may be mediated by apo E (Mahley, 1988) and this pathway may represent an alternative route for the delivery of cholesteryl ester to the liver, independent of LDL. Under normal circumstances, the clearance of cholesteryl esters from HDL particles has been reported to be relatively independent of the content of apo E in HDL of rabbits (Goldberg et al., 1991). However, the cholesterol uptake in an in vivo system may be volatile, responding strongly to metabolic changes such as acute cholesterol loading. The formation of large, apo E-rich HDL in response to cholesterol feeding in animals may reflect an attempt at greater removal of cholesterol from an overloaded system (Quarfordt et al., 1980). Indeed, large, apo E-rich HDL may arise independently of the general HDL pool, with the specific role of returning cholesterol effluxed from the peripheral tissues to the liver (Mahley, 1988). In vivo models employing subfractions of HDL are crucial in the understanding of the underlying mechanism in this complex process, yet few in vivo studies have been reported. Gavish et al. (1987) have shown that radiolabelled cholesteryl ester incorporated into small, apo E-poor HDL injected into rats was detectable in large, apo E-rich HDL after 18 hr. A more recent report (Richard and Pittman, 1993), however, has shown that the hepatic uptake of HDL appears to be independent of apo E over 24 hr. These studies were designed to use an animal which lacked CETP. Also the rat has a very different lipoprotein profile from the human and is not prone to atherosclerosis. The present study was carried out to investigate the effect of cholesterol loading on the distribution of HDL subspecies in relation to reverse cholesterol transport. The rabbit was selected for the study because it has a plasma

lipid and lipoprotein distribution which is similar to that of man (Chapman, 1980), it has similar profile of HDL subfractions as determined by heparin-Sepharose affinity chromatography (Fragoso and Skinner, 1992, 1993) to that of humans (Wilson et al., 1992) and possesses CETP with higher activity than that of humans (Ha and Barter, 1982). It has also been demonstrated that rabbits whose tissues were cholesterol-loaded by injection of [‘H]cholesteryl ester incorporated into acetyl LDL showed enrichment of the radiolabel in the specific large apo E-containing HDL 4 hr after the injection (Fragoso and Skinner, 1993). In the present study, rabbits were injected with total 14C-HDL and with 14C-HDL subfractions separated according to their contents of apo E. Some animals were submitted to a previous cholesterol loading of peripheral and hepatic endothelial cells with [3H]cholesteryl ester (CE). The distribution of radiolabels among HDL subfractions and the tissue uptake of 14C and [3H] were measured in order to identify the relationship between HDL interconversion and cholesterol efflux from peripheral tissues in both cases. The process was studied within a time-course of 45 and 90 min in order to assess the acute metabolic response rather than the basal situation achieved after many hours that has already been investigated by others. MATERIALS

AND

METHODS

Animals New Zealand White rabbits, aged 6-12 months and with an average weight of 3.8 kg, were caged individually, in a room with controlled temperature (20°C) and cycled with 12 hr light/darkness. Water and standard laboratory rabbit chow (R14, Biosure, Cambridgeshire, U.K.) were available ad libitum. Blood samples were drawn from the right marginal ear vein of calm, unfasted, unanaesthetized animals into vials containing EDTA to give a final concentration of 1.3 mg/ml. Vials were immediately placed on ice and plasma and lipoproteins were obtained without delay. Radiolabelled lipoproteins were injected in the left marginal ear vein and plasma decay of radioactivity was assessed from samples of blood drawn from the right ear. At the end of each study, the animals were given an overdose of Sag&al (pentobarbitone sodium, BP) by injection in the left marginal ear vein. During

HDL and reverse cholesterol transport

deep anaesthesia, blood was collected by cardiac puncture and placed in a vial containing EDTA to give a final concentration of 1.5 mg/ml. After death was confirmed the abdomen was opened and the liver removed. Bile was collected directly from the gall bladder. The manipulation of rabbits in the present study was approved by The British Home Office.

153

Chemical modtjication of human LDL and incorporation of [‘Hlcholesteryl ester

These were performed as previously described (Fragoso and Skinner, 1993). Briefly, human LDL isolated by ultracentrifugation was chemically modified by acetylation (Fraenkel-Conrat, 1957; Basu et al., 1976). A sample of 1 ml of LDL containing 3.&4.0mg of protein was placed in a vial in an ice bath, and 1 ml of a saturated solution of sodium acetate was added. Preparation of plasma and isolation of The contents were kept under continuous stirlipoproteins ring and small aliquots (2 ~1) of acetic anhyPlasma was separated by centrifugation at dride were added over a period of 1 hr. After the 800g for 15 min, at 4°C. The supernatant was addition of a total mass of acetic anhydride re-centrifuged under the same conditions. equivalent to 1.5 times the total mass of protein, In order to reduce ultracentrifugation time to the contents were kept under continuous stirring avoid alterations that occur in the HDL par- for an additional 30 min and then dialysed for ticles at high g values (Kunitake and Kane, 24 hr at 4°C against a minimum of 6 1 of 1982), total rabbit plasma was submitted to one 150 mM NaCl, 0.3 mM EDTA, pH 7.4. No single ultracentrifugation procedure. The den- traces of the slower migrating native LDL were sity of the plasma was adjusted to 1.25 g/ml by detectable in the chemically modified LDL, as addition of solid NaBr and the lipoprotein assessed by agarose gel electrophoresis (Noble, fraction was isolated by flotation in a Beckman 1968). preparative ultracentrifuge model XL-80, with Previously modified human LDL was the 50Ti rotor, at 105,400g at 16°C for 40 hr radiolabelled with rH]cholesteryl linoleate (Skinner, 1992). Lipoproteins were isolated by ([la,2a(n)-3H]cholesteryl linoleate 1.7 TBq/ tube slicing and immediately applied to a gel mmol, 2.61 GBq/mg, Amersham International permeation column (Rude1 et al., 1986). The Ltd, Little Chalfont, U.K.) (Brown et al., 1975). 100 x 1 cm glass column was packed with An aliquot of IOOpCi (3.7 x IO6 Bq) of Sepharose CG6B (Pharmacia LKB) and the [3H]cholesteryl linoleate in toluene was evaporchromatography was performed with 150 mM ated to dryness and redissolved in 0.5 ml of NaCI, I mM EDTA, 5 mM Tris-HCl, pH 7.4, 150mM NaCl, 0.3mM EDTA, pH 7.0, 10% at lOml/hr. No traces of apo A-I were de- dimethylsulphoxide (DMSO). The vial containtectable in the first eluted peak, which contained ing this solution was placed in a water bath at very low density lipoprotein (VLDL) and 37°C. After lOmin, 0.5 ml of human acetylLDL. LDL containing 600-800 pg of protein was The region of the elution profile containing added. The incubation was allowed to proceed HDL was concentrated by means of Microsep for 6 hr, at 37°C and then dialysed overnight 10 K concentrator tubes (Filtron Techn. Co., against 6 1 of 150 mM NaCl, 0.3 mM U.S.A.) coated with 10% glycerol, and the EDTA, pH 7.0, at 4°C. The absence of free concentrated sample submitted to immunoaffin[3H]cholesteryl linoleate in the [3H]acetyl-LDL ity chromatography on Sepharose coupled to sample was assessed by thin layer chromatoganti-whole rabbit serum depleted of lipoproteins raphy on silica gel plates developed in pet(Sigma Chemical Co., U.K.). This procedure acetate : acetic acid roleum ether: ethyl removed the small amounts of contaminating (90: 10: 1). plasma proteins. The final HDL sample was dialysed against 25 mM NaCl, 5 mM Tri-HCl, pH 7.4, prior to Heparin-Sepharose afinity chromatography Subfractions of HDL were isolated according heparin-Sepharose affinity chromatography as previously described (Fragoso and Skinner, to their contents of apo E by affinity chromato1993). graphy on a column of heparin coupled to Rabbit HDL which was not to be injected activated Sepharose (March et al., 1974; Kloer in vivo as well as human LDL were obtained by et al., 1976; Wilson et al., 1992). For the ultracentrifugation procedures as described purpose of the present paper, subfractions of elsewhere (Fragoso and Skinner, 1993). HDL isolated by heparin-Sepharose affinity Bc*w--8

154

Yara D. Fragosoand E. Roy Skinner

chromatography will be designated according to their order of elution as HDL,,,-HDL,,,,. HDL,,, contains the smallest particles of HDL (7.7-8.6 nm) isolated by this chromatographic method; they are poor in apo E and contain relatively low levels of cholesterol. The subfractions subsequently eluted are progressively larger, and contain more apo E and cholesterol (Wilson et al., 1992). HDL,,,, contains particles of 11.6-l 1.8 nm dia. (Fragoso and Skinner, 1993), with apo E accounting for approx. 7.5% of the total apoprotein content.

HDL was also delipidised and apolipoproteins were separated by SDS-PAGE. The different bands were excised from the gel and taken for radioactivity measurements in order to assess the distribution of the label among the proteins of HDL. Distribution of 14C label was proportional to the relative percentage of the apoproteins, as described by others (Wishart and Mackinnon. 1990). Radioactivity

measurements

All samples were dissolved in Optiphase Hisafescintillation fluid (Pharmacia LKB) Preparation of “C-sucrose-labelled HDL and counted for 3 min in a LKB 1219 Rackbeta Total HDL from rabbits, as well as subfracliquid scintillion counter, the data being protions of HDL separated by heparin-Sepharose cessed on the Ultroterm 2, which was proaffinity chromatography, were labelled with 14C- grammed to count [3H]/‘4C and to correct for sucrose by a modification of the method orig- quenching. There was an excess [3H] in all inally described by Pittman et al. (1979). Briefly, double-labelled samples. [U-‘4C]sucrose 25 PCi dried (1 MBq; In vivo studies in the rabbit 20.74 GBq/mmol; Amersham International Ltd, Little Chalfont, U.K.), were dissolved in 200 ~1 A total of 12 rabbits was used in the first of 10 mM 2,4,6-trichloro-1,3,5-triazine (cyan- phase of the study, in which each subfraction of uric chloride) in acetone. The commercial prep- HDL labelled with *4C-sucrose was injected into aration of cyanuric chloride was recrystallized two animals. Each animal received an injection in toluene prior to use in these experiments. To of 1.5 ml containing 2 mg of HDL protein. They the reaction vial, 150 p 1 of 20 mM NaOH were were sacrificed after 45 and 90 min, respectively, added and 15 set later 150 ~1 of 2.5 mM acetic and the radioactivity was measured in tissues, acid were used to quench the reaction. To the plasma, lipoproteins and subfractions of HDL. HDL resulting “C-sucrose-dichlorotriazine, One further rabbit was injected with total 14Ccontaining 2 mg of protein in 1 ml of 150 mM sucrose-HDL and sacrificed after 90 min. NaCl, 20 mM NaH2P04, 1 mM EDTA, pH 7.4, In the second phase of this study, two rabbits were added and the reaction was allowed to received an injection of [3H]CE acetyl LDL 2 hr proceed at room temperature for 6 hr in the prior to the injection of “C-sucrose-HDL,,. presence of 200 ~1 of 20 mM NaH,PO,, 1.5 mM Two further rabbits were injected with rqCE 5,5’-dithio-bis (2-nitrobenzoic acid) (DTNB), acetyl LDL 2 hr prior to the injection of total pH 7.4. The sample was then dialysed over a 14C-sucrose-HDL. The animals were sacrificed period of 12 hr with several changes of the same after 90 min and radioactivity was measured as phosphate buffer used for the incubation. Total above. HDL, as well as subfractions of HDL, were labelled with an average 5000 cpm (0.17 Bq)/pg RESULTS protein. No adverse reaction was observed in any of Efficiency of labelling was assessed by lipid extraction with chloroform: methanol (2 : 1, the animals injected with radiolabelled lipov/v), where less than 0.05% of the label was proteins. Following injection of E3H]acetyl-LDL, radioaccounted for in the lipid phase. Precipitation by 10% trichloroacetic acid yielded more than activity was rapidly cleared from plasma, on average only 3.7% of the label being detectable 99.6% of the label in the precipitate. Additionin plasma after 2 hr. ally, labelled HDL was submitted to heparinWhen rabbits were injected with subfraction Sepharose affinity chromatography to check on (I) of HDL containing small, apo E-poor HDL distribution of the label among subfractions. The results were comparable for labelled and particles labelled with 14C, 69.1 and 70.7% of unlabelled HDL, suggesting that covalent link- the label remained in this subfmction after 45 and 9Omin, respectively. Less than 4% of the ing with i4C-sucrose does not alter the binding properties of HDL subfractions. The labelled 14C label was accounted for in subfraction (VI)

HDL and reversecholesteroltransport

of HDL containing large, apo E-rich HDL particles at either time (Table 1). After injection of subfractions (II)of “C-sucrose, some redistribution of label among subfractions was observed (Table 1). Although the degree of redistribution of the 14C label was variable depending on the time point and on the subfraction injected, no other subfraction of HDL retained the label as subfraction HDL,,, had done. Most of the 14Clabel injected with each of the HDL subfractions (IHVI) remained in the HDL density range after 45 and 90min (Table 2). There was no marked uptake in the tissues studied of 14C from HDL(,,-HDL(v,, while injection of HDL,‘, resulted in hepatic recovery of 4.6 and 7.2% of the injected 14C label after 45 and 90 min, respectively (Table 2). In order to observe whether the apo E-poor HDL particles (HDL,,,) would remain as a relatively independent pool of particles in a rabbit acutely loaded with cholesterol, two animals were injected with “C-sucrose HDL,‘, 2 hr after an injection of [3H]CE acetyl-LDL. The proportion of 14Cin subfraction (I) after 45 and 90 min fell to 20.6 and 15.9%, respectively, in the two rabbits (Table 3). Most of the 14C label in the HDL of these rabbits was accounted for in subfraction (VI). The recovery of 14C in subfraction (VI) was more than 20 times higher than in the animals not pre-loaded with cholesterol. An appreciable proportion of [3H] was also recovered in the large, apo E-rich subfraction (VI) of HDL in both rabbits submitted to acute cholesterol loading (Table 3). At the same time, the distribution of 14C label between subfractions of intermediate size (HDL,,,,,,) was not, in general, markedly different from that when “C-sucrose-HDL,, was injected without prior cholesterol loading. Also the [3H] label was distributed fairly evenly between these subfractions. Injection of 14C HDL,,, into rabbits under normal conditions did not produce any detectable uptake of 14C by the liver after 45 or 90 min. However, the recovery of the 14C label in the liver of the two rabbits previously injected with [3H]CE acetyl-LDL was 4.9 and 5.3% after 90 min. In all animals, over 90% of the 14Clabel and most of the [3H] label were recovered within the HDL density range at the end of the experiments (Table 2). To investigate the effects of injecting total HDL into this system, rabbits received 14C-total HDL with and without acute cholesterol load-

155

ing, Heparin-Sepharose affinity chromatography of the “C-sucrose-HDL sample showed the following relative distribution of 14C label among the subfractions HDL,,,-HDL,‘,, respectively:. 18.5, 14.5, 29.5, 25.6, 5.8 and 6.3%. The re-distribution of 14C among the subfractions of HDL, when total HDL labelled with “C-sucrose was injected into the animals, is shown in Table 4. The 14C label tended to be concentrated in subfractions of smaller (HDL,,) and larger (HDL,,, and possibly HDL,,) particle size at the expense of those of intermediate size (HDLa’H’v,). The redistribution of 14C into larger particles of HDL was more marked for rabbits pre-loaded with [3H]cholesteryl linoleate prior to the “C-sucrose-HDL injection. In these animals, an average of 29.9% of the total 14C recovered was present in subfractions (V) and (VI), in contrast to 18.7% in the rabbits not pre-loaded with cholesterol (Table 4). Furthermore, subfractions III and IV, which contained 55.1% of 14C of the HDL sample injected, contained only 24.1% of the 14C recovered in HDL after 90 min in the rabbit without cholesterol load. When the rabbits were pre-loaded with [3H]cholesteryl linoleate, a similar proportion of 14C, i.e. 21.4%, was still recovered in these subfractions (Table 4). The relative proportion of 14Clabel recovered in subfraction (II) in rabbits with and without cholesterol loading was similar to that initially injected in the “C-sucrose-HDLK sample (Tables 4 and 5). A large proportion (average 47.9%) of [3H]CE was recovered in subfractions (I) and (VI) in the cholesterol-loaded rabbits, with a relatively even distribution of the label in the fractions of intermediate size (Table 4). Cholesterol preloading also resulted in a greater hepatic uptake of 14C than without cholesterol preloading when [‘4C]-total HDL was injected, while the bulk of the label remained within the HDL subfraction (Table 5). When calculated on the basis of protein concentration, the distribution of HDL subfractions in the plasma samples withdrawn after 90 min was similar to that of the HDL sample injected for both rabbits (Table 3). This suggests that the treatment did not appreciably perturb the distribution of subfractions over the time interval studied. It is also of importance to observe that the injection of acetyl-LDL produced an increase of only 5.2% of LDL-cholesterol/rabbit. Such increase was unlikely to alter the lipoprotein system beyond its expected physiological response.

I II III IV V VI

Subfraction injected

69.1 22.7 50.8 29.2 9.5 13.2

Table 1, Redistribution of t4C label recovered

[36.5] [25.4] [28.71 [32.1] [26.4] [24.6]

45

I

70.7 51.8 24.1 14.5 41.1 18.9

[37.0] [38.2] [20.5] [23.2] [33.0] rl6.71

90

10.5 17.4 13.1 18.0 12.0 10.3

[14.1] [19.8] [21.1] [16.8] [10.4] [15.4]

45’

II

9.8 16.5 20.2 13.5 11.5 10.7

[16.8] [14.5] [9.6] [15.0] [11.2] [17.4]

90 5.9 18.7 12.2 17.0 24.6 16.9

of “‘C into the six isolated subfractions of HDL following in each subfraction, when total 14C eluted from the column

[8.0] [15.3] [15.0] [15.5] [14. l] [14.4]

45’

III 4.4 10.2 17.2 17.0 9.3 14.0

[11.6] [l 1.31 [15.71 [14.2] [12.1] [17.2]

90

Subfraction

4.8 14.0 9.3 30.7 21.1 21.4

[15.1] [14.8] [lO.O] [14.9] [15.9] [16.2]

45’

IV 6.4 6.9 8.3 10.9 8.6 14.6

[10.4] [9.9] [17.1] [13.5] [14.1] [16.7]

90 6.4 18.6 7.0 9.3 18.7 16.8

[15.3] [9.6] [12.4] [ll.O] [15.7] [14.2]

45’

V 7.6 6.6 18.8 21.7 11.6 9.0

112.51 [IO.11 [17.6] [16.7] [11.6] [13.8]

90 3.2 8.6 7.7 19.4 14.0 21.4

[10.9] [15.0] [12.5] [IO.21 [17.5] [15.2]

45’

VI 1.5 7.9 11.4 22.3 17.9 32.8

[11.6] [16.0] [19.5] [17.5] [16.1] [18.1]

90

labelled subfraction after 45 and 90 min. Results are presented as relative percentage The relative proportion of each subfraction, quantified as protein (OD280), is shown

recovered

injection of each individual is considered to be 100%. in square brackets

$ r S’ 3 8

E m

71 I% 8

;

e g

HDL and reverse cholesterol transport

157

Table 2. Recovery of i4C in tissues and fluids of rabbits injected with each specific subfraction of HDL labelled with i4C-sucrose. Fractions (I)-(W) are as for Table 1. L lipoprotein protein (2 mg) was injected in each case. Results for tissues and fluids are presented as percentage of total label injected, and values lower than 0.1% were considered to be not detectable (n.d.). Values for lipoprotein fractions are expressed as percentages of total lipoprotein recovered Liver Bile Aorta VLDL + LDL HDL Plasma Subfraction 45’ 90’ 45’ 90’ 45’ 90’ 45’ 90’ 45’ 90’ 45’ 90 I 30 18 n.d. n.d. n.d. nd. n.d. nd. 2 3 98 97 II 51 39 0.5 0.3 n.d. n.d. n.d. n.d. 2 3 98 9-i III 29 26 0.6 1.0 n.d. n.d. nd. n.d. 4 5 96 95 IV 42 38 0.6 1.0 n.d. n.d. n.d. 0.1 3 3 91 97 V 71 75 0.3 0.5 nd. 0.1 0.1 0.4 5 6 95 94 VI 19 17 4.6 7.2 0.5 0.5 0.1 0.2 7 7 93 93

Recovery of the labels in plasma, lipoproteins and liver of rabbits 90 min after injection with 14C-total HDL, with and without pre-loading of cholesterol, is shown in Table 5. Over 90% of i4C and the major proportion of the [3H] recovered in plasma lipoproteins were found within the HDL density range in all animals. While in the rabbit without cholesterol preloading, only 3.2% of the injected 14C was recovered in the liver after 90 min, rabbits pre-loaded with cholesterol had 5.4 and 15.0% of the injected 14C accounted for in the liver after this time period. An appreciable proportion of the injected [3H]CE was also present in the liver. DISCUSSION

The role of HDL as an acceptor of cholesterol effluxed from peripheral cells has been supported by several in vitro studies, though the mechanism by which HDL mediates the process of reverse cholesterol transport (Barter, 1993) is not fully understood. HDL comprises a variety of particles of different sizes and composition, their metabolic roles being largely determined by their apolipoprotein composition (Skinner, 1994). In vivo studies are required to clarify the role of specific subfractions of HDL in this complex pathway. The present report provides evidence that more than one pathway is operative for the interconversion of subfractions of HDL in response to acute cholesterol loading, using the rabbit as an animal model. This study has concentrated on the short-term response of the lipoprotein system to cholesterol-loading, which has not been previously reported. In this investigation, peripheral tissues were loaded with [3H]cholesteryl ester by injecting labelled acetyl-LDL which was taken up by endothelial cells by modified LDL receptors (Miller et al., 1985). The subsequent association of the effluxed cholesterol with different HDL subfractions was observed. This design was

adopted as it represented a physiological approach in that it did not involve the formation of HDL particles which were not normally present in the plasma. This contrasts with the situation which pertains to the feeding of a high cholesterol diet and leads to the formation of specific HDL particles (HDLJHDL,) whose size greatly exceeds that of the normal range of HDL and contains a predominantly greater content of apo E than any HDL subfraction normally present in plasma (Gordon et al., 1983; Koo et al., 1985). One group of rabbits received an injection of an apo E-poor HDL subfraction of small particle size (HDL,,,), labelled with “C-sucrose. These animals responded to the acute cholesterol loading by a marked and rapid increase in the formation of large, apo E-rich HDL particles with little alteration in the radioactivity of the fractions of intermediate particle size. However, when the rabbits were not pre-loaded with cholesterol, the injection of labelled small, apo E-poor HDL particles did not produce ‘any significant formation of the large HDL particles. These results demonstrate that the apo E-rich pool of large HDL particles arises independently of the HDL population as a whole in response to a system acutely loaded with cholesterol, as proposed by others (Quarfordt et al., 1980; Goldberg et al., 1991). This is supported by the observation that the label associated with subfractions II-V, when injected separately, is redistributed independently between these subfractions. It is not altogether clear why injection of subfraction VI should result in the appearance of the label in fractions II-V as well as in fraction I. This may arise through the selective uptake of the cholesterol moiety of these particles by the kidneys in a manner analogous to that which is suggested to occur in the liver (Glass et al., 1985; Goldberg et al., 1991; Richard and Pittman, 1993). The kidneys have been shown to be the major site of

15.9

29.1

B

Rabbit 90

21.3

19.9

11.0

12.0

Prot

14.4

6.4

r4C

rH]

i4C

20.6

35.6

Prot

A

Rabbit 90

II

I

15.4

13.8

[3H]

14.7

12.9

Prot

10.0

18.9

i4C

III

10.0

20.4

[‘HI

12.1

12.0

__Prot

12.5

5.6

14C

9.6

9.5

[3H]

Subfraction IV

14.3

18.8

Prot

recovered

10.0

6.3

14C

V

12.3

10.0

[‘HI

18.8

13.7

Prot

37.1

42.2

14C

VI

31.4

26.4

[‘H]

5.3

4.9

14C

Liver

2.9

7.2

[3H]

“‘C

0.1

nd.

Bile

0.5

0.4

[‘HI

n.d.

nd.

14C

Aorta

nd.

nd.

[3H]

Table 3. Recovery of radiolabels in each of the six subfractions of HDL, liver, bile and aorta after injection of [‘H]cholesteryl linoleate-acetyl LDL followed by 14C-sucrose-HDL,,. In subfractions of HDL, results are presented as relative percentage of radiolabel recovered in each subfraction, when total 14C eluted from the column is considered to be 100%. In liver, bile and aorta, results are presented as percentage total injected label. Results presented in the first column (Prot) represent the relative percentage of the subfraction of HDL eluted from the column on a protein basis. Values lower than 0.1% were considered to be not detectable (n.d.)

i? zi q

3 c z 5 e pl ;FI s

P2 r(

Prot 26.5

24.4

40.0

Rabbit C

Rabbit D

Rabbit E

33.6

30.0

31.6

i4C

31.2

24.3

[‘HI

13.9

14.4

Prot 15.8

12.2

12.8

19.6

i4C

13.6

13.5

[‘HI

13.6

13.7

Prot 15.5

9.6

14.0

14.5

i4C

11.3

13.7

[‘HI

8.8

14.6

Prot 14.3

5.7

13.4

9.6

i4C

10.6

13.4

[‘HI

8.8

14.7

Prot 13.5

11.5

13.6

6.1

i4C

14.7

13.5

[3H]

18.6

15.2

21.6

18.1

Prot 14.4

18.4

16.2

12.6

i4C

[‘HI

Table 4. Recovery of 14Cwhich had been injected with total HDL in six subfractions of HDL. Results are presented as relative percentage of radiolabel recovered in each subfraction, when total 14Celuted from the column is considered to be 100%. Results presented in the first column (Prot) represent the relative percentage of the subfraction of HDL eluted from the column, on a protein basis. Results are presented for a rabbit under normal conditions (C) and for two animals which received a prior injection of [‘H]CE-acetyl-LDL (D and E) Subfraction recovered I II III IV V VI

160

Yara D. Fragoso and E. Roy Skinner

Table 5. Recovery of 14Cwhich had been incorporated into total HDL in tissues and fluids of New Zealand White rabbits 90 min after injection. Heparin-Sepharose affinity chromatography of the “C-sucrose-HDL sample showed the following relative distribution of 14C label among the subfractions HDLrn-HDL(v,,, respectively: 18.5, 14.5, 29.5, 25.6, 5.8 and 6.3%. Results are presented for a rabbit under normal conditions (C) and for two animals which received a prior injection of [‘H]CE-acetyl-LDL (D and E). Values for tissues are expressed as percentages of total label injected. The values for lipoprotein fractions are percentages of total lipoprotein recovered Under normal [3H]CE followed conditions by 14C-HDL Rabbit C 14C 37.0 3.2 0.3 n.d. 2.8 5.5 91.7

Plasma Liver Bile Aorta VLDL LDL HDL

Rabbit D 14C 1.7 22.4 7.2 5.4 0.4 nd. nd. nd. 9.8 0.7 31.2 7.1 58.9 92.2

[‘HI

Rabbit E [‘HI ‘T 4.0 27.8 3.0 15.0 0.1 n.d. 0.1 n.d. 5.4 2.0 18.1 6.7 76.5 91.3

apo A-I uptake following injection of doublelabelled HDL (Glass et al., 1985) and a receptor for apo A-I on the kidney cell membrane has also been reported (Fidge, 1986). During such a process, the selective removal of cholesterol would lead to a remodelling of the lipoprotein

particle to produce HDL particles of smaller size and within the range of fractions I-V. Furthermore, the uptake of some of the labelled apolipoprotein by the kidneys, and possibly by some other tissues, would explain, at least in part, the relatively low recovery of the label in the plasma. Another group of rabbits received an injection of total HDL labelled with 14C. Whether or not these animals were loaded with cholesterol, there was an increase in the concentration of label in the HDL subfractions of smallest and largest particle sizes. However, this was more marked in the animals which received acute cholesterol loading. In both cases there was little change in the distribution of particles of intermediate size. These particles may be continuously formed by pathways such as lipolysis and CETP-mediated exchange, giving rise to large apo E-rich HDL particles at a slow rate. Based on the results obtained in the present investigation, we suggest that two alternative routes for cholesterol transport in HDL may be operational in the rabbit (Fig. 1). The results of the present investigation are in accordance with our previous observation that cholesterol loading induced a significant

Route A

Nascent HDL Lipolysis products \

LDL LDL

Cholesterol

receptor

apoE-HDL receptor LDL receptor, LRP?

Loading LCAT,

apoE

Route B

Fig. 1. Schematic representation of two proposed pathways for reverse cholesterol transport. In the normal situation (Route A), small discoidal particles of HDL are released from the liver and intestine and undergo LCAT-catalysed esterification to form small spherical HDL. Small HDL are also considered to originate from pre-beta HDGlike particles containing cholesterol eIIIuxed from peripheral cells. Unidentified fusion and transfer processes, including the addition of surface phospholipid, cholesterol and apohpoproteins from triglyceride-rich lipoproteins during lipolysis, result in the production of a series of HDL particles of increasing particle sixe (represented by HDLeu-HDL,). Part of the cholesteryl ester content of these particles is then exchanged for triglycerides from LDL (mediated by CETP) and taken up by hepatic LDL receptors. With cholesterol-loading (Route B), there is a rapid conversion of smaB, ape B-poor HDL,,, to large, apo Erich HDLo,,, particles, without significant participation of fractions of intermediate size (HDLuu-HDLo+ Pathway A could be dependent on CETP activity and may be operative at a slow rate at all times, whereas pathway B represents the acute response to cholesterol-loading of peripheral cells.

HDL and reverse cholesterol transport

increase in the relative proportion of label associated with a specific large HDL subfraction, here called HDL,,, in rabbits (Fragoso and Skinner, 1993). Such large HDL particles may represent an acute alternative for the clearance of surplus cholesterol from peripheral cells, while transfer of cholesteryl ester from HDL to VLDL/LDL via CETP might represent a longer-term response. CETP activity becomes increased due to increased CETP gene expression in peripheral and central tissues (Tall, 1993) and such response might require longer than the times used in the present experiment to fully participate in the reverse cholesterol transport process. It must be noted that cholesteryl linoleate as used in the present work is very effectively transferred from HDL to VLDL (Morton, 1986). An increased recovery of 14C in the liver was observed in the rabbits pre-loaded with cholesterol. Nevertheless, the transfer of 14C from HDL to VLDL and/or LDL was still similar to that observed in the untreated animals, suggesting that HDL may be capable of direct delivery of cholesteryl esters to the liver, even in animals with plasma CETP activity. Previous in viva studies in the rabbit (Goldberg et al., 1991) used animals which were not pre-loaded with cholesterol and the tracer had been followed for much longer periods, i.e. 24 hr. These authors observed that the major pathway for clearance of HDL cholesteryl esters was via CETP-mediated transfer to VLDL and LDL and their eventual hepatic uptake via the LDL receptor, although a substantial role was played by selective uptake of cholesteryl esters directly from HDL by the liver. The present study suggests that, as an acute response to cholesterol loading, the rate of delivery of cholesteryl ester to the liver by direct uptake of HDL is increased. This process appears to be dependent on the large, apo E-rich HDL particles, although apo E may not necessarily be the ultimate ligand for the binding of this lipoprotein to the liver. In the present investigation, the quantities of HDL-cholesterol injected into each rabbit was less than 1% of the total amount of circulating HDL and it is therefore unlikely that the addition of this lipoprotein would produce any perturbation of the system. Furthermore, it was observed that 45 and 90 min after injection of 14C-HDL,,, there was essentially no change in the distribution of HDL subfractions on the basis of their protein content. This suggests that

161

cholesterol pre-loading has altered the flux through the system without appreciably altering the concentration of the individual HDL subspecies that constitute the pathway. A full quantitative appraisal of the routes involved in reverse cholesterol transport would require measurement of turnover rates, plasma clearance and rate of hepatic uptake for each subspecies of HDL. As pointed out by others (Goldberg et al., 1991), this would require a vast number of animals and is therefore not readily amenable to experimental determination. In the present study, a variety of measurements was made on each rabbit in order to properly investigate the processes involved in the rapid response to cholesterol-loading of endothelial cells. The expected interindividual variation cannot be underestimated. However, they represent only small, occasional fluctuations when compared with the consistent differences observed in fraction VI in cholesterol-primed and non-primed rabbits. The observed increase in the formation of large, apo E-rich HDL in cholesterol-loaded animals may be relevant to an understanding of the atherogenic process in humans. The significantly reduced concentration of this subfraction in patients with coronary heart disease (Cheung et al., 1991; Johansson et al., 1991; Wilson et al., 1993) and in subjects of high coronary risk (Griffin et al., 1988) may arise through a deficiency in Route B (see Fig. 1) which may not be fully compensated for by Route A for the removal of surplus cholesterol from the arterial wall. In conclusion, the recent development of methods for the quantitative isolation of a number of subfractions of HDL has enabled the movement of cholesterol through interconversion of HDL subspecies to be examined in greater detail than has previously been possible. Using this approach, it can be suggested that two pathways exist for the transfer of cholesterol via HDL, the fluxes of which differ according to whether or not the peripheral tissues of the rabbit were pre-loaded with cholesterol. These findings may be of value in understanding the process of reverse cholesterol transport and the mechanism underlying the protective effect of high concentrations of HDL against the development of coronary heart disease. Acknowledgements-This work was supported by The Wellcome Trust. We are grateful to Mr S. McBain and Mrs C. Watt for skilled technical assistance.

162

Yara D. Fragoso and E. Roy Skinner REFERENCES

Barter P. J. (1993) High density lipoprotein and reverse cholesterol transport. Curr. Opin. Lipidol. 4, 210-217. Basu S. K., Goldstein J. L., Anderson R. G. W. and Brown M. S. (1976) Degradation of cationized low density lipoprotein and regulation of cholesterol metabolism in homozygous familial hypercholesterolaemia. Proc. Nat1 Acad. Sci. U.S.A. 73, 3178-3182. Brown M. S., Dana S. E. and Goldstein J. L. (1975) Receptor-dependent hydrolysis of cholesteryl ester contained in plasma low density lipoprotein. Proc. Nat1 Acad.

Sci.

U.S.A.

72, 2925-2929.

Castro G. R. and Fielding C. (1988) Early incorporation of cell-derived cholesterol intro pre-beta-migrating high density lipoprotein. Biochemistry 27, 25-29. Chacko G. K. (1982) Human high density lipoprotein (HDL3) binding to rat liver plasma membranes. Biochim. Biophys. Acta 712, 129-141. Chapman J. M. (1980) Animal lipoproteins: chemistry, structure and comparative aspects. J. Lipid Res. 21, 789-853. Cheung M. C., Brown B. G., Wolf A. C. and Albers J. J. (1991) Altered particle size of apolipoprotein A-Icontaining lipoproteins in subjects with coronary heart disease. J. Lipid Res. 32, 383-394. Fidge N. H. (1986) Partial purification of a high density lipoprotein-binding protein from rat liver and kidney membranes. FEBS Let?. 199, 265-268. Fidge N. H. and Nestel P. J. (1985) Identification of apolipoproteins involved in the interaction of human high density lipoprotein3 with receptors on cultured cells. J. Biol.

Chew

260, 357&3575.

Fraenkel-Conrat H. (1957) In Methods in Enzymology (Edited by Colowick S. P. and Kaplan N. O.), Vol. 4, pp. 247-269. Academic Press, New York. Fragoso Y. D. and Skinner E. R. (1992) The effect of gammalinolenic acid on the subfractions of plasma high density lipoprotein of the rabbit. Biochem. Pharmacol. 44, 1085-1090. Fragoso Y. D. and Skinner E. R. (1993) The role of subfractions of high density lipoprotein in the in vivo transport of cholesterol from cholesterol-loaded hepatic and peripheral cells in the New Zealand White rabbit. Camp.

Biochem.

Physiol.

105B,

373-376.

Francone 0. L., Fielding C. J. and Fielding P. E. (1989) Distribution and functions of 1ecithin:cholesterol acyltransferase and cholesteryl ester transfer in plasma lipoproteins. J. Biol. Chem. 31, 2195-2220. Gavish D., Oschry Y. and Eisenberg S. (1987) In vivo conversion of human HDL, to HDL, and apoE-rich HDL, in the rat: effects of lipid transfer protein. J. Lipid Rex

28, 257-267.

Glass C., Pittman R. C., Civen M. and Steinberg D. (1985) Uptake of high density lipoprotein associated with apoprotein A-I and cholesterol esters by 16 tissues of the rat in vivo and by adrenal cells and hepatccytes in vitro. J. Biol.

Chem.

260, 744-750.

Glomset J. A. (1968) The plasma lecithin: cholesterol acyltransferase reactions. J. Lipid Rex 9, 155-167. Goldberg D. I., Belz W. F. and Pittman R. C. (1991) Evaluation of pathways for the cellular uptake of high density lipoprotein cholesterol esters in rabbits. J. Clin. Invest.

87, 331-346.

Gordon D. J. and Rifkind

B. M. (1989) High density

lipoprotein-the clinical implications of recent studies. N. Engl. J. Med. 321, 1311-1316. Gordon V., Innerarity T. L. and Mahley R. W. (1983) Formation of cholesterol- and apoprotein E-enriched high density lipoproteins. J. Bioi. Chem. 258, 620226212. Griffin B. A., Skinner E. R. and Maughan R. J. (1988) Plasma high density lipoprotein subfractions in subjects with different coronary risk indices as assessedby plasma lipoprotein concentrations. Atherosclerosis 70, 1655169. Ha Y. C. and Barter P. J. (1982) Differences in plasma cholesteryl ester activity in sixteen vertebrate species. Comp.

Biochem.

Physiol.

71B,

265-269.

Johansson J., Carlson L. A., Landou C. and Hamsten A. (1991) High density lipoproteins and coronary atherosclerosis: a strong inverse relation with the largest particle is confined to normotriglyceridemic subjects, Arteriosclerosis Thromb. 11, 174-l 82. Kloer H. U., Schmidt J. W., Ditschuneit H. and Glomset J. (1976) Affinity chromatography of lipoproteins with heparin-agarose. Protides Biol. Fhdds 23, 581-588. Koo C., Innerarity T. L. and Mahley R. W. (1985) Obligatory role of cholesterol and apolipoprotein E in the formation of large cholesterol enriched and receptoractive high density lipoprotein. J. Biol. Chem. 260, 1193411943. Kunitake S. T. A. and Kane J. P. (1982) Factors affecting the integrity of high density lipoproteins in the ultracentrifuge. J. Lipid. Res. 23, 936940. Mahley R. W. (1988) Apolipoprotein E: cholesterol transport protein with expanding role in cell biology. Science

240, 622-630.

March S. C., Parikh I. and Cuatrecasas P. (1974) A simplified method for cyanogen bromide activation of agarose for affinity chromatography. Anal. Biochem. 60, 149-152. Miller N. E., LaVille A. and Crook D. (1985) Direct evidence that reverse cholesterol transport is mediated by high density lipoprotein in the rabbit. Nature 314, 109-111. Morton R. E. (1986) Specificity of lipid transfer protein for molecular species of cholesteryl ester. J. Lipid Res. 27, 523-529. Noble R. P. (1968) Electrophoretic separation of plasma lipoproteins in agarose gel. J. Lipid Res. 9, 693-700. Pittman R. C., Green S. R., Attie A. D. and Steinberg D. (1979) Radiolabeled sucrose covalently linked to protein. J. Biol.

Chem.

W,

687&6879.

Quarfordt S., Hanks J., Jones R. S. and Shelburne F. (1980) The uptake of high density lipoprotein cholesteryl ester in the perfused rat liver. J. Biol. Chem. 235, 2934-2937. Richard B. M. and Pittman R. C. (1993) Role of HDL, in cholesteryl ester uptake in rats. J. Lipid Res. 34, 571-578. Rude1 L. L., Marzetta C. A. and Johnson F. L. (1986) In Methodr in Enzymology (Edited by Segrest J. P. and Albers J. J.), Vol. 129, pp. 4547. Academic Press, London. Skinner E. R. (1992) In Lipoprotein Analysis, a Practical Approach (Edited by Converse C. A. and Skinner E. R.), pp. 85-l 18. IRL Press, Oxford. Skinner E. R. (1994) High density lipoprotein subclasses. Curr.

Opin.

Lipidol.

5, 241-247.

Tall A. R. (1993) Plasma cholesteryl ester transfer protein. J. Lipid

Res. 34, 1255-1274.

Williams P. T., Krauss R. M., Vranizan K. M., Stafanick M. L., Wood P. D. S. and Lindgren F. T. (1992) Familial

HDL and reverse cholesterol transport correlations of HDL subclasses based on gradient gel electrophoresis. Arteriosclerosis Thromb. 12, 14671474. Wilson H. M., Pate1 J. C. and Skinner E. R. (1990) The distribution of high density lipoprotein subfractions in coronary survivors. Biochem. Sot. Trans. 18, 117551176. Wilson H. M., Griffin B. A., Watt C. and Skinner E. R. (1992) The isolation and characterization of high density

163

lipoprotein subfractions containing apolipoprotein E from human plasma. Biochem. J. 284, 477-48 1. Wilson H. M., Pate1 J. C., Russell D. and Skinner E. R. (1993) Alterations in the concentration of an apolipoprotein E-containing subfraction of high density lipoprotein in coronary heart disease. Clinica Chim. Acra 228, 175-187. Wishart R. and Mackinnon M. (1990) Uptake and metabolism of high density lipoproteins by cultured rabbit hepatocytes. Biochim. Biophys. Acta 1844, 375-381.