- Email: [email protected]
7-Ketocholesterol delivered to mice in chylomicron remnant-like particles is rapidly metabolised, excreted and does not accumulate in aorta
Biochimica et Biophysica Acta 1530 (2001) 209^218 www.elsevier.com/locate/bba
7-Ketocholesterol delivered to mice in chylomicron remnant-like particles is rapidly metabolised, excreted and does not accumulate in aorta Malcolm A. Lyons 1 , Andrew J. Brown * Cell Biology Group, Heart Research Institute, 145 Missenden Road, Camperdown, 2050 Sydney, N.S.W., Australia Received 29 August 2000; received in revised form 18 December 2000; accepted 21 December 2000
Abstract Cholesterol oxidation products (oxysterols) have been implicated in atherogenesis due to their presence in atherosclerotic tissue and their potent effects in vitro. One of the major oxysterols currently of interest is 7-ketocholesterol (7K) and it has been suggested that the diet is an important source of this oxysterol. This investigation tested the hypothesis that 7K, delivered in a physiologically relevant vehicle, chylomicron remnant-like emulsion (CMR), would be metabolised and excreted by mice in a similar manner and to a similar extent as previously observed in rats when delivered in a chemically modified lipoprotein, acetylated low-density lipoprotein (acLDL). Indeed, the metabolism of 14 C-7K delivered in CMR mirrored that of acLDL and was much more rapid than 3 H-cholesterol delivered simultaneously. The 7K-derived 14 C was cleared from the liver, appeared in the intestine and was excreted in the faeces. A substantial proportion of the 7K-derived 14 C in the intestine and faeces was aqueous-soluble, indicating metabolism to polar products, presumably bile acids. Moreover, while cholesterol-derived 3 H increased in the aorta, 14 C appeared transiently and there was no observable accumulation within 24 h. The data confirm our previous findings of rapid hepatic metabolism of 7K when delivered in acLDL and demonstrate that 7K delivered in a vehicle of dietary significance is similarly metabolised and excreted. Indeed, the data encourage further investigation into the contribution that dietary oxysterols may or may not make to atherogenesis. ß 2001 Elsevier Science B.V. All rights reserved. Keywords: Oxysterol; Metabolism; Chylomicron remnant; Acetylated low-density lipoprotein; Diet; Mouse
1. Introduction Abbreviations: 7K, 7-ketocholesterol; acLDL, acetylated lowdensity lipoprotein; apoE, apolipoprotein E; BHT, butylated hydroxytoluene ; CMR, chylomicron remnant-like emulsion ; DOPC, L-K-dioleoylphosphatidylcholine; EDTA, ethylenediaminetetraacetic acid; IDL, intermediate-density lipoprotein; LDL, low-density lipoprotein; PBS, phosphate-bu¡ered saline ; PL, phospholipid; TAG, triacylglycerol ; VLDL, very low-density lipoprotein * Corresponding author. Department of Molecular Genetics, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75235-9046, USA. Fax: +1-214-648-8804; E-mail: [email protected] 1 Present address: Box 213, The Jackson Laboratory, Bar Harbor, ME 04609, USA.
Oxysterols are oxidation products of cholesterol which are formed exogenously by autoxidation of cholesterol [1] and endogenously by free-radical attack upon cholesterol (e.g. [2]) or by enzymic production (e.g. [3]). Particular attention has been paid to oxysterols in the ¢eld of atherosclerosis research, mainly due to their presence in human atherosclerotic tissue and because they display many potent pro-atherogenic properties in vitro, and to a lesser extent in vivo (reviewed in [4]). Oxysterols have been shown to be concentrated, relative to plasma
1388-1981 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 8 - 1 9 8 1 ( 0 1 ) 0 0 0 8 3 - X
BBAMCB 55755 23-2-01
210
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
and normal tissue, in human macrophage-foam cells and atherosclerotic lesions [4]. The two major species are 27-hydroxycholesterol ((25R)-cholest-5-en-3L, 26-diol) and 7-ketocholesterol (cholest-5-en-3L-ol-7one, 7K) [3,5,6]. While it is known that 27-hydroxycholesterol is the ¢rst product of the alternative pathway of bile acid biosynthesis [7] and may be particularly important for the excretion of extrahepatic cholesterol [3,8], the origin of 7K remains equivocal [9]. It is believed that 7K is derived primarily by freeradical oxidation of cholesterol in vivo [10,11] but it has also been suggested that it may be derived by dietary absorption [12^15] from processed cholesterol-rich foodstu¡ (e.g. [16]). However, it is yet to be shown that exogenous oxysterols contribute to those found in the artery (aorta) wall. Recently, we employed acetylated low-density lipoprotein (acLDL), a chemically modi¢ed, high-uptake lipoprotein, administered intravenously to rats, to deliver oleoyl esters of 14 C-7K and 3 H-cholesterol simultaneously to investigate the metabolism of these sterols. In accordance with others using 3 H-cholesteryl ester-labelled acLDL injected into rats [17^19], the bulk of the radiolabelled steryl esters were cleared by the liver in a matter of minutes. Unlike cholesterol however, 7K underwent rapid hepatic metabolism and was excreted into the intestine mainly as aqueous-soluble products, presumably bile acids [20]. We suggested that such rapid clearance and excretion into the intestine may have implications for the purported role that 7K (and perhaps other dietary oxysterols) has in the development of atherosclerosis. A potential criticism of that work was that the vehicle (acLDL) used to deliver the radiolabelled sterols was not physiological. Dietary oxysterols including 7K have been shown to be absorbed and incorporated into chylomicrons in rats [13,14] and humans [12]. A recent experiment that fed oxysterols to rabbits demonstrated increased oxysterol concentrations in the combined chylomicron remnant plus very low-density lipoprotein (VLDL) fraction but importantly, 7K was not increased [15]. Chylomicron remnants are generated from chylomicrons after partial hydrolysis of the triacylglycerol (TAG) core and are removed from the circulation by hepatocytes [21^ 23]. In contrast, acLDL is removed mainly by hepatic endothelial cells and there then follows a net trans-
fer of the sterol content to hepatocytes within a matter of hours [19]. Therefore, the aim of the present study was to test the hypothesis that 7K delivered in a physiologically relevant vehicle, chylomicron remnant-like emulsion (CMR), would be metabolised in a similar fashion and to at least the same extent as its counterpart delivered in a modi¢ed lipoprotein, acLDL. To this end we utilised a well-characterised chylomicron remnant model developed by Redgrave and co-workers (e.g. [24^26]), radiolabelled with our sterols of interest and administered to mice. Additionally, use of this model lipoprotein allowed the complications of intestinal absorption and its quanti¢cation to be avoided. 2. Materials and methods 2.1. Materials Triolein, cholesterol, cholesteryl oleate and L-K-dioleoylphosphatidylcholine (DOPC) were of the highest commercial grade available ( s 98%) and were obtained from Sigma-Aldrich (Castle Hill, N.S.W., Australia). General reagents were of analytical reagent grade and were also purchased from SigmaAldrich or BDH (Kilsyth, Vic., Australia). Sodium pentobarbitone (Nembutal, 60 mg/ml) was supplied by Boehringer Ingelheim (Artarmon, N.S.W., Australia) and methoxy£urane (Penthrane) by Abbott Australasia (Sydney, N.S.W., Australia). Reagents for liquid scintillation were purchased from Canberra-Packard (Mount Waverly, Vic., Australia). [414 C]7-Ketocholesteryl oleate (14 C-7K oleate, speci¢c activity 2.0 MBq/Wmol) was synthesised as described elsewhere [20] and [1K,2K-3 H]cholesteryl oleate (speci¢c activity 1.7 GBq/Wmol) was purchased from Amersham-Pharmacia Biotech (Castle Hill, N.S.W., Australia). 2.2. acLDL For production of radiolabelled acLDL, human low-density lipoprotein (LDL) was isolated from plasma derived from a fasted, healthy donor, acetylated by the repeat addition of acetic anhydride and labelled as described previously [20]. The extent of modi¢cation was determined by relative electropho-
BBAMCB 55755 23-2-01
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
retic mobility (REMv3 compared to native LDL) [27]. Mice were injected with 100 Wl acLDL containing 43.6 Wg protein, 8.18 kBq 14 C-7K oleate and 29.0 kBq 3 H-cholesteryl oleate. The speci¢c activity of 3 H-cholesteryl oleate in this vehicle was estimated to be 2.0 MBq/Wmol based on a protein recovery of 59% and recovery of both radiolabelled and unlabelled lipid of 15%, as determined previously [20]. The speci¢c activity of 14 C-7K oleate was 2.0 MBq/ Wmol. 2.3. CMR CMR was prepared essentially as described previously [24^26] with the following minor modi¢cations: water containing 2.2% glycerol (w/v) was replaced with Tricine (pH 8.4) containing 2.2% glycerol (w/v); ethylenediaminetetraacetic acid (EDTA; 27 Wmol/l) was included in the preparation bu¡er; two radiolabelled steryl esters were included. Tricine was used to match the bu¡er used in the preparation of radiolabelled acLDL and EDTA was included to avoid autoxidation of the lipid components and generation of oxysterols by the titanium sonicator tip during preparation of the emulsion (A.J. Brown, unpublished observation). The emulsion was prepared from a mixture of triolein, DOPC, cholesteryl oleate and cholesterol (plus the two radiolabelled steryl esters) in the weight ratio of 4.5: 2.5: 0.8: 0.8, respectively, and was ¢lter-sterilised (0.2 Wm) prior to injection. Animals received an injection of 100 Wl containing 6.08 kBq 14 C-7K oleate and 18.3 kBq 3 H-cholesteryl oleate (assayed after ¢lter sterilisation). The CMR produced by the described method was reported to generate particles with diameter 55 þ 3 nm (n = 40) after puri¢cation by ultracentrifugation [25]. In this study, the availability of 14 C-7K oleate was limited as it was necessary to synthesise it from 14 C-cholesterol and thus the emulsion was used without ultracentrifugation. It has been reported previously that particles prepared by this method and isolated in three size classes by ultracentrifugation produced particles with diameters 47.8, 80.0 and 152 nm [28]. The CMR generated without ultracentrifugation therefore likely represents a continuum of sizes across this range. Furthermore, the particle size was correlated with the TAG/phospholipid (PL) ra-
211
tio [28]. Therefore, as a control based on this correlation, the TAG/PL ratio was inverted to 2.5/4.5 (TAG/PL, w/w) so as to produce smaller particles. The behaviour of these smaller particles was compared with that of the two main vehicles (acLDL and CMR) in the delivery of 14 C-7K oleate and 3 H-cholesteryl oleate as an indication of the dependency of the sterols' metabolism on particle size. Apart from the inversion of the TAG/PL ratio, the emulsion (labelled PL/TAG to indicate the greater proportion of PL) was generated in an identical manner to the CMR. The speci¢c activity of 3 H-cholesteryl oleate was estimated to be 0.25 MBq/Wmol and that of 14 C-7K oleate determined to be 2.0 MBq/ Wmol in both the CMR and PL/TAG vehicles. 2.4. Animals Male C57BL/6J mice were obtained from Biological Resources Centre, University of New South Wales (Sydney, N.S.W., Australia) and kept under a 12 h light/dark cycle and allowed free access to water and chow. All procedures were conducted in accordance with the Central Sydney Area Health Service Animal Welfare Committee. The mice were selected at random for each condition (n = 21 each condition) and time-point (n = 3) and weighed 23.6 þ 1.9 g (mean þ S.D., n = 42). Mice were mildly anaesthetised using methoxy£urane, an inhalant anaesthetic administered via a nose cone. They were then injected with one of the three radiolabelled particles via the tail vein after its immersion in warm water to promote vasodilation. The animals were then allowed to recover under a heat lamp with supervision until fully conscious. 2.5. Tissues and sample preparation At the appropriate time, animals were injected intraperitoneally with sodium pentobarbitone (400 Wl, 10 mg/ml) in conjunction with administration of methoxy£urane. When the animals were fully anaesthetised, blood was taken by cardiac puncture and added to microfuge tubes containing EDTA (20 Wl, 200 mmol/l) for collection of plasma. The animals were then perfused for 5 min with phosphate-bu¡ered saline (PBS) containing EDTA (1 mmol/l) and butylated hydroxytoluene (BHT; 20 Wmol/l) by a
BBAMCB 55755 23-2-01
212
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
gravity-feed system. The liver and intestine, from immediately below the stomach to the rectum, were then excised. The aorta, from the heart to immediately below the iliac-femoral bifurcation, was freed of connective tissue in situ and then removed. All tissues were weighed immediately and stored at 380³C until further analysis. Livers were homogenised with one volume and intestines with two volumes of the same bu¡er used for perfusion but also containing chloramphenicol (100 mg/l; Boehringer Mannheim, Castle Hill, N.S.W., Australia) using an Ultra-Turrax T8 hand held disperser (IKA, Crown Scienti¢c, Moorebank, N.S.W., Australia). Homogenates were also stored at 380³C prior to analysis. 2.6. Determination of radioactivity All determinations of radioactivity were conducted using a TriCarb 2100TR Liquid Scintillation Analyser (Canberra-Packard) and a dual-label protocol. Results (mean þ S.D., n = 3 for each time-point) are presented as the quantity of isotope present in any given sample as a percentage of the total injected dose. The total volume of plasma (ml) was estimated
by multiplication of the weight of the mouse (g) by 49.2 and division of the product by 1000 [29]. Plasma (200 Wl) was analysed by the addition of scintillant (4.5 ml; Ultima Gold). Tissue homogenates were extracted by the method of Folch et al. [30] and the aqueous-soluble fraction assayed for radioactivity essentially as described previously [20] using approximately 200 mg liver homogenate or 300 mg intestinal homogenate. The aqueous-soluble radioactivity was determined as an indication of metabolism to bile acids or bile acid-like products. Total radioactivity in homogenates and aorta was determined by solubilisation using a tissue solubiliser (Soluene 350, Canberra-Packard) according to the manufacturer's instructions and as detailed previously [20]. The total dissected aorta and similar amounts of homogenates as used for the determination of aqueous-soluble radioactivity were used for determination of total radioactivity. Lipid-soluble radioactivity was calculated as the di¡erence between the total and aqueous-soluble radioactivities. 3. Results 3.1. Plasma
Fig. 1. Clearance of esters of 14 C-7K and 3 H-cholesterol from plasma of mice when administered incorporated into acLDL or CMR. AcLDL (open symbols) or CMR (¢lled symbols) was administered intravenously and the cholesterol-derived 3 H (circles) and 7K-derived 14 C (squares) were measured in plasma over the time-course. Radioactivities were determined by the addition of scintillation £uid and analysis by liquid scintillation counting with a dual-label protocol. Data are expressed as a percentage of the total dose injected and are reported as mean+S.D. (n = 3).
Clearance of radioactivity contained within acLDL from the plasma was essentially complete within a matter of minutes (Fig. 1), consistent with results obtained in other studies [17^19]. 14 C derived from 7K reappeared transiently then decreased to very low levels from 12 h onwards while 3 H derived from cholesterol increased to 6 h then reached a plateau. These data for acLDL are in excellent agreement with those derived from rats that we have published previously [20]. In contrast, the radioactivity contained in CMR was contained wholly within the plasma compartment at the earliest time-point, 2 min. The two labels were cleared from the plasma in an identical fashion up to 1 h and had diverged by 3 h. This pattern of clearance up to 1 h is consistent with that demonstrated by Rensen et al. [28] using three di¡erently sized populations of CMR injected into in mice. 14 C was then cleared to a greater extent than 3 H and was indistinguishable from the acLDL result for 14 C after 12 h. A slightly lower level of 3 H was retained in the plasma when delivered in CMR
BBAMCB 55755 23-2-01
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
213
Fig. 2. Clearance of 14 C-7K and to 3 H-cholesterol from the liver of mice when the steryl esters were delivered in either acLDL (A) or CMR (B). Cholesterol-derived 3 H (circles) and 7K-derived 14 C (squares) were determined for total liver radioactivity (¢lled symbols, solid line) and aqueous-soluble radioactivity (open symbols, broken line). Radioactivity was determined by liquid scintillation counting using a dual-label protocol from samples of solubilised tissue homogenate (total) and the aqueous phase of a tissue homogenate lipid extraction [30]. Data are expressed as a percentage of the total dose injected and are reported as mean+S.D. (n = 3).
compared with acLDL. The disappearance of the labels from the plasma is consistent with particle clearance. The plasma data derived from the PL/ TAG-treated animals were essentially the same as those derived from the CMR-treated animals (data not shown). 3.2. Liver Uptake of radioactivity by the liver (Fig. 2) was rapid for acLDL and CMR vehicles with approximately 45% of the injected dose appearing after 30
min for both isotopes, although it was initially more rapid for acLDL. Over the 24 h time-course, the level of radioactivity detected in the liver was very similar for both isotopes independent of the delivery particle. Aqueous-soluble products of both 3 H-cholesterol and 14 C-7K were detected in the liver. However, aqueous-soluble 14 C was found at a greater level than aqueous-soluble 3 H across the time-course for both vehicles. The hepatic data derived from the PL/ TAG-treated animals were essentially the same as those derived from the CMR-treated animals (data not shown).
Fig. 3. Appearance in the intestine of 7K-derived 14 C and cholesterol-derived 3 H when the steryl esters were administered to mice incorporated into either acLDL (A) or CMR (B). 3 H (circles) and 14 C (squares) were determined for total liver radioactivity (¢lled symbols, solid line) and aqueous-soluble radioactivity (open symbols, broken line). Radioactivity was determined by liquid scintillation counting using a dual-label protocol from samples of solubilised tissue homogenate (total) and the aqueous phase of a lipid extraction [30] of tissue homogenate. Data are expressed as a percentage of the total dose injected and are reported as mean+S.D. (n = 3).
BBAMCB 55755 23-2-01
214
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
3.3. Intestine A much larger proportion of 14 C-7K-derived radioactivity appeared in the intestine for both acLDL and CMR than the respective level of 3 H-cholesterolderived radioactivity (Fig. 3). The response for each of the sterol-derived radioactivities was similar between delivery vehicles although the 14 C delivered in acLDL (Fig. 3A) that was detected at 6 h was less in comparison with CMR (Fig. 3B). The aqueous-soluble 14 C fraction was greater than the aqueous-soluble 3 H in both cases. These results are consistent with our previous ¢ndings although the aqueous-soluble 14 C portion was found to be the major component of the intestinal pool of 14 C in our earlier study [20]. The intestinal data derived from the PL/TAG-treated animals were essentially the same as those derived from the CMR-treated animals (data not shown).
the aortic levels were approximately 0.005% of the injected dose of 14 C (Fig. 4B) after 2 min whereas mice administered PL/TAG emulsion had approximately 0.018% of the injected dose at the same time-point (Fig. 4C). The same was observed for cholesterol-derived 3 H reaching levels of 0.008 and 0.02% of the injected dose, respectively.
3.4. Aorta Greater levels of both 3 H derived from cholesterol and 14 C derived from 7K were detected in the aortas of animals administered acLDL (Fig. 4A) than the respective isotope in animals given CMR (Fig. 4B). Most importantly, while 14 C was detected in both treatment groups at earlier time-points, its appearance was transient and no accumulation was demonstrated. After 24 h, the raw 14 C radioactivity (dpm) was at the limit of detection (910 dpm). On the other hand, 3 H increased over the 24 h period. Consistent with CMR-treated mice (Fig. 4B), no accumulation of 14 C could be detected in aortas of animals treated with PL/TAG emulsion (Fig. 4C), whereas cholesterol-derived 3 H increased to similar levels in both of these groups. There was a noticeable di¡erence in the level of radioactivity detected in the aorta at early time-points. When treated with CMR, C
Fig. 4. Determination of 7K-derived 14 C and cholesterol-derived 3 H in the aortas of mice when the steryl esters were administered incorporated into either acLDL (open symbols, A), CMR (shaded symbols, B) or PL/TAG (¢lled symbols, C). 3 H (circles) and 14 C (squares) were determined by liquid scintillation counting using a dual-label protocol from the solubilised whole aorta. Data are expressed as a percentage of the total dose injected multiplied by 1000 and are reported as mean+S.D. (n = 3).
BBAMCB 55755 23-2-01
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
Fig. 5. Faecal and urinary excretion of 7K-derived 14 C and cholesterol-derived 3 H after administration to mice of the steryl esters incorporated into acLDL, CMR or PL/TAG. Faecal (A) and urinary (B) samples from three animals for each treatment were pooled over 24 h and subjected to solubilisation to determine total radioactivity and to lipid extraction [30] to determine the aqueous-soluble radioactivity (open). Lipid-soluble radioactivity (¢lled) was calculated as the di¡erence between total and aqueous-soluble radioactivity. Radioactivity was assayed by liquid scintillation counting using a dual-label protocol. Data were calculated for the total, pooled sample, averaged between three animals and expressed as a percentage of the total dose injected per animal per 24 h.
3.5. Faeces and urine Pooled samples of faeces and urine were collected from animals in each of the 24 h time-point treatment groups (n = 3 per group). A large proportion of 14 C-7K-derived radioactivity, amounting to approximately one-third of the dose, was detected in the faeces of animals treated with any vehicle after 24 h (Fig. 5A). Of this, 20^30% was aqueous-soluble, presumably having undergone complete conversion to
215
bile acids. Approximately twice the amount of 3 Hcholesterol-derived radioactivity originating from acLDL was measured in the faeces compared with CMR but in both cases the total amount was much less than the faecal 14 C. The proportion of aqueoussoluble products was similar in the two groups. In comparison, the excretion of radiolabels into the faecal material (Fig. 5A) was similar if not slightly higher for the PL/TAG emulsion animals but 14 C excretion remained substantially greater than that of 3 H. Thus, a substantial proportion of the injected dose of 14 C was excreted into the faeces in the initial 24 h period, irrespective of the delivery vehicle, indicating that 7K undergoes rapid excretion. Low levels of radioactivities, each amounting to less than 1% of the injected dose, were detected in the urine (Fig. 5B). 3 H was found to be similar in all cases but slightly more 14 C was detected when 7K was administered in acLDL. The bulk of the radioactivity was aqueous-soluble, but the lipid-soluble portion was detectable. Approximately two- to three-fold the amount of 14 C compared with 3 H was excreted by all groups of animals in the urine. This level comprised only 0.3^0.5% of the injected dose of 14 C. It is evident from the appearance of 14 C in the faeces that 7K undergoes much more e¤cient excretion from the body than cholesterol. 4. Discussion Oxysterols have been implicated in atherogenesis (reviewed in [4]). Non-enzymically formed oxysterols, such as 7K, are often assumed to originate from the diet and exert their e¡ects locally in the artery wall subsequent to their accumulation. We have previously reported that 7K undergoes rapid hepatic metabolism and does not accumulate in the aorta when delivered to rats in a high-uptake modi¢ed lipoprotein (acLDL) [20]. The aim of the present study was to extend this ¢nding to mice and more signi¢cantly, to employ a more physiologically relevant vehicle, CMR, to investigate the metabolism of 7K. Similar to cholesterol, oxysterols are absorbed and packaged into chylomicrons [12^14] (although absorption data vary widely, ranging from 6% [14] to 93% [31]). Following lipolysis by lipoprotein lipase, the resultant chylomicron remnants are taken up by
BBAMCB 55755 23-2-01
216
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
hepatocytes [22]. In this study, CMR and acLDL, both of which are removed from the circulation by the liver and each containing oleoyl esters of 14 C-7K and 3 H-cholesterol, were administered intravenously to mice and the fate of the radiolabels monitored in plasma, liver, intestine and aorta over 24 h. Pooled faeces and urine were also collected over the 24 h period. For 3 H-cholesterol, the kinetics of appearance and disappearance in the circulation and liver were similar to those reported previously when delivered in acLDL to rats [19,20] or CMR to mice [28]. Reappearance in the plasma of 3 H originally incorporated into acLDL and its incomplete clearance when delivered in CMR, may indicate reverse cholesterol transport from endothelial cells to hepatocytes for 3 Hcholesterol delivered in acLDL [19] or resynthesis into lower density lipoproteins for either or both particles. Using smaller particles with a lower TAG/PL ratio (or higher PL/TAG ratio), we have demonstrated the validity of the CMR vehicle. In essence, the results obtained using the PL/TAG emulsion were indistinguishable from the CMRtreated animals although slightly higher levels of both 3 H and 14 C radioactivity were found in the aortas of animals treated with the PL/TAG emulsion. This may be explained by the greater arterial permeability a¡orded the PL/TAG emulsion particles due to their smaller diameter compared with CMR in a manner analogous to the greater permeability of chylomicron remnants versus chylomicrons in the artery wall [32]. The demonstration of rapid hepatic metabolism and excretion of 14 C-7K into the faeces and its lack of accumulation in the aorta in all three delivery vehicles, points to the robustness of this ¢nding. The role of dietary oxysterols in atherosclerosis remains equivocal. Of 16 oxysterol-feeding studies identi¢ed in the literature using various animal models, seven were pro-atherogenic, eight anti-atherogenic and one showed no clear e¡ect [33]. For dietary oxysterols to play a role in the development and/or initiation of atherosclerosis, they should accumulate within the artery wall. By injecting the sterols in CMR, we have avoided the complications of intestinal absorption and yet there is still no evidence that 7K accumulated in the aorta. Thus, we have consistently observed that while 3 H-cholesterol accumu-
lated in the aorta, 14 C derived from 7K appeared transiently and its removal was essentially complete by 24 h. Furthermore, there was no indication of enterohepatic recycling, thereby ruling out the possibility of accumulation later than 24 h. While mice do not generally possess cholesteryl ester transfer protein activity [34], it may be argued that in species that do, such as man, prolonged exposure to oxysterols contained in chylomicron remnants may allow transfer of the oxysterols to lower density, potentially atherogenic lipoproteins. For example, a study of 25-hydroxycholesterol (cholest-5en-3L,25-diol) administered orally to squirrel monkeys suggested that this oxysterol may contribute to atherogenesis because the bulk was detected in the VLDL/LDL fraction, however quantitative and excretion data were not presented [35]. Furthermore, 25-hydroxycholesterol seldom contributes to the overall oxysterol content in foods (e.g. [16]). Another study in which pigeons were fed cholesterol plus cholestan-3L,5K,6L-triol found increased coronary artery lumenal stenosis but no signi¢cant di¡erence in arterial cholesterol or cholesteryl ester content [36]. More recent, purportedly pro-atherogenic oxysterol-feeding studies have presented data highly consistent with 7K data from the present study and a previous study [20]. Three studies with rabbits that have investigated the role of oxysterols in atherosclerosis have found increased cholesterol contents in the aorta [37^39]. However, oxysterols were either not measured [39], were not elevated above the control group [38] or were not detected, despite being detected in circulating TAG-rich lipoproteins [37]. It would be interesting to determine if any of this accumulated aortic cholesterol was derived from 7K, especially in light of the ¢nding that we [20] and others [40,41] have shown that 7K can be converted to cholesterol. In another study [15], rabbits were fed in addition to cholesterol, an amount of oxysterols typical of levels of total sterols consumed by humans (on a per kg body weight basis) [4]. Despite this extreme intake, no signi¢cant increase was found in the concentration of 7K in the combined VLDL and chylomicron fraction although increases were observed in three other oxysterols [15]. In addition, there was no signi¢cant increase in any of the oxysterols in LDL [15]. It is conceivable that the increased oxysterol content in the VLDL/chylomicron fraction could be
BBAMCB 55755 23-2-01
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
solely due to newly ingested oxysterols packaged in chylomicrons which then undergo lipolysis, hepatic uptake and subsequent metabolism. Indeed, the authors demonstrated increased levels of oxysterols in the liver, including 7K. It was claimed that levels of 7K in LDL were not increased because it is transported by serum albumin rather than lipoproteins [15]. However, others have shown that 7K distribution among lipoproteins re£ects cholesterol distribution [42^44] even at radiotracer concentrations [4]. The aortic area covered by fatty streaks was found to be greater in the group of rabbits fed oxysterols but notably, neither the cholesterol nor the oxysterol concentrations of the aorta were reported [15]. Another investigation by the same group [45], examined the e¡ect of feeding oxysterols to two murine models of atherosclerosis, apolipoprotein E (apoE)de¢cient and LDL receptor-de¢cient mice. Due to impaired lipoprotein uptake, these mice have increased levels of chylomicron remnants and VLDL [46,47] and LDL and intermediate-density lipoprotein (IDL) [47,48], respectively. It may be argued that these animals better represent models to test atherogenic parameters because both are hypercholesterolaemic [46,48]. In addition, the apoE-de¢cient mice develop spontaneous atherosclerosis [46] which is exacerbated by cholesterol-feeding [49] and the LDL receptor-de¢cient mice develop atherosclerosis when fed a high-cholesterol diet [49]. Chylomicron remnant uptake is dependent upon apoE recognition [22,47,50] by the LDL receptor pathway and a second, slower endocytosis mechanism [47,50]. Both of these animal models demonstrate impaired cellular metabolism of chylomicron remnants [50]. This is pertinent to potential criticism of the present study because it is our contention that under normal circumstances chylomicron remnant clearance is rapid and thus 7K metabolism will be correspondingly swift. Under conditions in which clearance mechanisms are perturbed, dietary 7K could potentially contribute to atherogenesis. In this context it should be noted that remnant particles derived from both chylomicrons and VLDL are an emergent risk factor for the development of atherosclerosis [32,51]. Chylomicron remnants are known to penetrate arterial tissue and can lead to focal accumulation of these particles in the subendothelial space [32]. Patients with famil-
217
ial hypercholesterolaemia and Watanabe Heritable Hyperlipidaemic (WHHL) rabbits, both LDL receptor-de¢cient, also display impaired clearance of chylomicron remnants [32]. It also appears that chylomicron remnant metabolism may be perturbed in most cases of atherogenic risk, for example diabetes [52] and even normolipidaemic patients with coronary artery disease [32]. Therefore, due to the risk that postprandial lipoproteins represent, dietary oxysterols such as 7K may still play a role in atherogenesis but that role may have been overstated. This is the ¢rst study to directly compare a commonly used, chemically modi¢ed lipoprotein with a more physiologically relevant vehicle, CMR, regarding the metabolism of 7K versus cholesterol. 7K was metabolised and excreted primarily in the faeces more rapidly than cholesterol, independent of its delivery vehicle. Moreover, the data indicate that 7K does not accumulate in the aorta to any signi¢cant extent over this period of time. The present investigation con¢rms our previous ¢ndings and raises doubts about the assertion that dietary oxysterols, 7K in particular, accumulate in the artery wall and contribute to atherogenesis. Further studies are necessary in which various animal models of atherosclerosis are fed oxysterols (including a labelled oxysterol) at levels realistically encountered in the human diet. Acknowledgements We thank Dr Wendy Jessup for critically reading this manuscript and Prof. Trevor Redgrave for scienti¢c discussion. This work was supported by a grant from the American Egg Board. References [1] L.L. Smith, Cholesterol Autoxidation, Plenum, New York, 1981. [2] H.N. Hodis, A. Chauhan, S. Hashimoto, D.W. Crawford, A. Sevanian, Atherosclerosis 96 (1992) 125^134. [3] I. Bjo«rkhem, O. Andersson, U. Diczfalusy, B. Sevastik, R.-J. Xiu, C. Duan, E. Lund, Proc. Natl. Acad. Sci. USA 91 (1994) 8592^8596. [4] A.J. Brown, W. Jessup, Atherosclerosis 142 (1999) 1^28. [5] M. Crisby, J. Nilsson, V. Kostulas, I. Bjo«rkhem, U. Diczfalusy, Biochim. Biophys. Acta 1344 (1997) 278^285.
BBAMCB 55755 23-2-01
218
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
[6] A.J. Brown, S.-l. Leong, R.T. Dean, W. Jessup, J. Lipid Res. 38 (1997) 1730^1745. [7] N.B. Javitt, FASEB J. 8 (1994) 1308^1311. [8] E. Lund, O. Andersson, J. Zhang, A. Babiker, G. Ahlborg, U. Diczfalusy, K. Einarsson, J. Sjovall, I. Bjo«rkhem, Arterioscler. Thromb. Vasc. Biol. 16 (1996) 208^212. [9] M.A. Lyons, A.J. Brown, Int. J. Biochem. Cell Biol. 31 (1999) 369^375. [10] H.N. Hodis, D.W. Crawford, A. Sevanian, Atherosclerosis 89 (1991) 117^126. [11] L.L. Smith, B.H. Johnson, Free Radic. Biol. Med. 7 (1989) 285^332. [12] H.A. Emanual, C.A. Hassel, P.B. Addis, S.D. Bergmann, J.H. Zavoral, J. Food Sci. 56 (1991) 843^847. [13] K. Osada, E. Sasaki, M. Sugano, Lipids 29 (1994) 555^559. [14] D.F. Vine, K.D. Croft, L.J. Beilin, J.C. Mamo, Lipids 32 (1997) 887^893. [15] I. Stapra¬ns, X.-M. Pan, J.H. Rapp, K.R. Feingold, Arterioscler. Thromb. Vasc. Biol. 18 (1998) 977^983. [16] J.H. Nielsen, C.E. Olsen, C. Duedahl, L.H. Skibsted, J. Dairy Res. 62 (1995) 101^113. [17] T.J.C. van Berkel, J.F. Nagelkerke, L. Harkes, J.K. Kruijt, Biochem. J. 208 (1982) 493^503. [18] R. Blomho¡, C.A. Drevon, W. Eskild, P. Helgerud, K.R. Norum, T. Berg, J. Biol. Chem. 259 (1984) 8898^8903. [19] H.F. Bakkaren, F. Kuipers, R.J. Vonk, T.J.C. van Berkel, Biochem. J. 268 (1990) 685^691. [20] M.A. Lyons, S. Samman, L. Gatto, A.J. Brown, J. Lipid Res. 40 (1999) 1846^1857. [21] P.H.E. Groot, T.J.C. van Berkel, A. van Tol, Metabolism 30 (1981) 792^797. [22] T.G. Redgrave, Int. Rev. Physiol. 28 (1983) 103^130. [23] M.C.M. van Dijk, M. Pieters, T.J.C. van Berkel, Eur. J. Biochem. 211 (1993) 781^787. [24] T.G. Redgrave, R.C. Maranha¬o, Biochim. Biophys. Acta 835 (1985) 104^112. [25] I.J. Martins, T.G. Redgrave, J. Lipid Res. 39 (1998) 691^ 698. [26] B.-J. Zeng, B.-C. Mortimer, I.J. Martins, U. Seydel, T.G. Redgrave, J. Lipid Res. 39 (1998) 845^860. [27] A.J. Brown, R.T. Dean, W. Jessup, J. Lipid Res. 37 (1996) 320^335. [28] P.C.N. Rensen, N. Herijgers, M.H. Netscher, S.C.J. Meskers, M. van Eck, T.J.C. van Berkel, J. Lipid Res. 38 (1997) 1070^1084. [29] M.D. Buckland, Mowlem, Whatley, A Guide to Laboratory Animal Technology, Heinemann Medical, London, 1981. [30] J. Folch, M. Lees, G.H. Sloane-Stanley, J. Biol. Chem. 226 (1957) 497^509.
[31] J. Bascoul, N. Domergue, J. Mourot, G. Dery, A. Crastes de Paulet, Lipids 21 (1986) 744^747. [32] J.C.L. Mamo, S.D. Proctor, D. Smith, Atherosclerosis 141 (1998) S63^S69. [33] M.A. Lyons, A.J. Brown, in: R.T. Dean, D.Kelly (Eds.), Atherosclerosis, Oxford University Press, Oxford, 2000, in press. [34] S. Jiao, T.G. Cole, R.T. Kitchens, B. P£eger, G. Schonfeld, Metabolism 39 (1990) 155^160. [35] S.-K. Peng, C.B. Taylor, E.H. Mosbach, W.Y. Huang, J. Hill, B. Mikkelson, Atherosclerosis 41 (1982) 395^402. [36] M.S. Jacobson, M.G. Price, A.E. Shamoo, F.P. Heald, Atherosclerosis 57 (1985) 209^217. [37] D.F. Vine, J.C.L. Mamo, L.J. Beilin, T.A. Mori, K.D. Croft, J. Lipid Res. 39 (1998) 1995^2004. [38] J. Rong, S. Rangaswamy, L. Shen, R. Dave, Y. Chang, H. Peterson, H. Hodis, G. Chisolm, A. Sevanian, Arterioscler. Thromb. Vasc. Biol. 18 (1998) 1885^1894. [39] J.X. Rong, L. Shen, Y.H. Chang, A. Richters, H.N. Hodis, A. Sevanian, Arterioscler. Thromb. Vasc. Biol. 19 (1999) 2179^2188. [40] J.S.M. Sarma, R.J. Bing, S. Ikeda, R. Fischer, Artery 2 (1976) 153^160. [41] S.A. Erickson, A.D. Cooper, S.M. Matsui, R.G. Gould, J. Biol. Chem. 252 (1977) 5186^5193. [42] C.J.W. Brooks, R.M. McKenna, W.J. Cole, J. MacLachlan, T.D.V. Lawrie, Biochem. Soc. Trans. 11 (1983) 700^701. [43] I.C. Gelissen, A.J. Brown, E.L. Mander, L. Kritharides, R.T. Dean, W. Jessup, J. Biol. Chem. 271 (1996) 17852^ 17860. [44] R.G. Dyer, M.W. Stewart, J. Mitcheson, K. George, M.M. Alberti, M.F. Laker, Clin. Chim. Acta 260 (1997) 1^13. [45] I. Stapra¬ns, X.-M. Pan, J.H. Rapp, C. Grunfeld, K.R. Feingold, Arterioscler. Thromb. Vasc. Biol. 20 (2000) 708^714. [46] S.H. Zhang, R.L. Reddick, J.A. Piedrahita, N. Maeda, Science 258 (1992) 468^471. [47] S. Ishibashi, J. Herz, N. Maeda, J.L. Goldstein, M.S. Brown, Proc. Natl Acad. Sci. USA 91 (1994) 4431^4435. [48] S. Ishibashi, M.S. Brown, J.L. Goldstein, R.D. Gerard, R.E. Hammer, J. Herz, J. Clin. Invest. 92 (1993) 883^893. [49] J.L. Breslow, Science 272 (1996) 685^688. [50] B.C. Mortimer, D.J. Beveridge, I.J. Martins, T.G. Redgrave, J. Biol. Chem. 270 (1995) 28767^28776. [51] M.L. Brown, M.P. Ramprasad, P.K. Umeda, A. Tanaka, Y. Kobayashi, T. Watanabe, H. Shimoyamada, W.-L. Kuo, R. Li, R. Song, W.A. Bradley, S.H. Gianturco, Proc. Natl Acad. Sci. USA 97 (2000) 7488^7493. [52] S.D. Proctor, C.K. Pabla, J.C.L. Mamo, Atherosclerosis 149 (2000) 315^322.
BBAMCB 55755 23-2-01
7-Ketocholesterol delivered to mice in chylomicron remnant-like particles is rapidly metabolised, excreted and does not accumulate in aorta Malcolm A. Lyons 1 , Andrew J. Brown * Cell Biology Group, Heart Research Institute, 145 Missenden Road, Camperdown, 2050 Sydney, N.S.W., Australia Received 29 August 2000; received in revised form 18 December 2000; accepted 21 December 2000
Abstract Cholesterol oxidation products (oxysterols) have been implicated in atherogenesis due to their presence in atherosclerotic tissue and their potent effects in vitro. One of the major oxysterols currently of interest is 7-ketocholesterol (7K) and it has been suggested that the diet is an important source of this oxysterol. This investigation tested the hypothesis that 7K, delivered in a physiologically relevant vehicle, chylomicron remnant-like emulsion (CMR), would be metabolised and excreted by mice in a similar manner and to a similar extent as previously observed in rats when delivered in a chemically modified lipoprotein, acetylated low-density lipoprotein (acLDL). Indeed, the metabolism of 14 C-7K delivered in CMR mirrored that of acLDL and was much more rapid than 3 H-cholesterol delivered simultaneously. The 7K-derived 14 C was cleared from the liver, appeared in the intestine and was excreted in the faeces. A substantial proportion of the 7K-derived 14 C in the intestine and faeces was aqueous-soluble, indicating metabolism to polar products, presumably bile acids. Moreover, while cholesterol-derived 3 H increased in the aorta, 14 C appeared transiently and there was no observable accumulation within 24 h. The data confirm our previous findings of rapid hepatic metabolism of 7K when delivered in acLDL and demonstrate that 7K delivered in a vehicle of dietary significance is similarly metabolised and excreted. Indeed, the data encourage further investigation into the contribution that dietary oxysterols may or may not make to atherogenesis. ß 2001 Elsevier Science B.V. All rights reserved. Keywords: Oxysterol; Metabolism; Chylomicron remnant; Acetylated low-density lipoprotein; Diet; Mouse
1. Introduction Abbreviations: 7K, 7-ketocholesterol; acLDL, acetylated lowdensity lipoprotein; apoE, apolipoprotein E; BHT, butylated hydroxytoluene ; CMR, chylomicron remnant-like emulsion ; DOPC, L-K-dioleoylphosphatidylcholine; EDTA, ethylenediaminetetraacetic acid; IDL, intermediate-density lipoprotein; LDL, low-density lipoprotein; PBS, phosphate-bu¡ered saline ; PL, phospholipid; TAG, triacylglycerol ; VLDL, very low-density lipoprotein * Corresponding author. Department of Molecular Genetics, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75235-9046, USA. Fax: +1-214-648-8804; E-mail: [email protected] 1 Present address: Box 213, The Jackson Laboratory, Bar Harbor, ME 04609, USA.
Oxysterols are oxidation products of cholesterol which are formed exogenously by autoxidation of cholesterol [1] and endogenously by free-radical attack upon cholesterol (e.g. [2]) or by enzymic production (e.g. [3]). Particular attention has been paid to oxysterols in the ¢eld of atherosclerosis research, mainly due to their presence in human atherosclerotic tissue and because they display many potent pro-atherogenic properties in vitro, and to a lesser extent in vivo (reviewed in [4]). Oxysterols have been shown to be concentrated, relative to plasma
1388-1981 / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 8 - 1 9 8 1 ( 0 1 ) 0 0 0 8 3 - X
BBAMCB 55755 23-2-01
210
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
and normal tissue, in human macrophage-foam cells and atherosclerotic lesions [4]. The two major species are 27-hydroxycholesterol ((25R)-cholest-5-en-3L, 26-diol) and 7-ketocholesterol (cholest-5-en-3L-ol-7one, 7K) [3,5,6]. While it is known that 27-hydroxycholesterol is the ¢rst product of the alternative pathway of bile acid biosynthesis [7] and may be particularly important for the excretion of extrahepatic cholesterol [3,8], the origin of 7K remains equivocal [9]. It is believed that 7K is derived primarily by freeradical oxidation of cholesterol in vivo [10,11] but it has also been suggested that it may be derived by dietary absorption [12^15] from processed cholesterol-rich foodstu¡ (e.g. [16]). However, it is yet to be shown that exogenous oxysterols contribute to those found in the artery (aorta) wall. Recently, we employed acetylated low-density lipoprotein (acLDL), a chemically modi¢ed, high-uptake lipoprotein, administered intravenously to rats, to deliver oleoyl esters of 14 C-7K and 3 H-cholesterol simultaneously to investigate the metabolism of these sterols. In accordance with others using 3 H-cholesteryl ester-labelled acLDL injected into rats [17^19], the bulk of the radiolabelled steryl esters were cleared by the liver in a matter of minutes. Unlike cholesterol however, 7K underwent rapid hepatic metabolism and was excreted into the intestine mainly as aqueous-soluble products, presumably bile acids [20]. We suggested that such rapid clearance and excretion into the intestine may have implications for the purported role that 7K (and perhaps other dietary oxysterols) has in the development of atherosclerosis. A potential criticism of that work was that the vehicle (acLDL) used to deliver the radiolabelled sterols was not physiological. Dietary oxysterols including 7K have been shown to be absorbed and incorporated into chylomicrons in rats [13,14] and humans [12]. A recent experiment that fed oxysterols to rabbits demonstrated increased oxysterol concentrations in the combined chylomicron remnant plus very low-density lipoprotein (VLDL) fraction but importantly, 7K was not increased [15]. Chylomicron remnants are generated from chylomicrons after partial hydrolysis of the triacylglycerol (TAG) core and are removed from the circulation by hepatocytes [21^ 23]. In contrast, acLDL is removed mainly by hepatic endothelial cells and there then follows a net trans-
fer of the sterol content to hepatocytes within a matter of hours [19]. Therefore, the aim of the present study was to test the hypothesis that 7K delivered in a physiologically relevant vehicle, chylomicron remnant-like emulsion (CMR), would be metabolised in a similar fashion and to at least the same extent as its counterpart delivered in a modi¢ed lipoprotein, acLDL. To this end we utilised a well-characterised chylomicron remnant model developed by Redgrave and co-workers (e.g. [24^26]), radiolabelled with our sterols of interest and administered to mice. Additionally, use of this model lipoprotein allowed the complications of intestinal absorption and its quanti¢cation to be avoided. 2. Materials and methods 2.1. Materials Triolein, cholesterol, cholesteryl oleate and L-K-dioleoylphosphatidylcholine (DOPC) were of the highest commercial grade available ( s 98%) and were obtained from Sigma-Aldrich (Castle Hill, N.S.W., Australia). General reagents were of analytical reagent grade and were also purchased from SigmaAldrich or BDH (Kilsyth, Vic., Australia). Sodium pentobarbitone (Nembutal, 60 mg/ml) was supplied by Boehringer Ingelheim (Artarmon, N.S.W., Australia) and methoxy£urane (Penthrane) by Abbott Australasia (Sydney, N.S.W., Australia). Reagents for liquid scintillation were purchased from Canberra-Packard (Mount Waverly, Vic., Australia). [414 C]7-Ketocholesteryl oleate (14 C-7K oleate, speci¢c activity 2.0 MBq/Wmol) was synthesised as described elsewhere [20] and [1K,2K-3 H]cholesteryl oleate (speci¢c activity 1.7 GBq/Wmol) was purchased from Amersham-Pharmacia Biotech (Castle Hill, N.S.W., Australia). 2.2. acLDL For production of radiolabelled acLDL, human low-density lipoprotein (LDL) was isolated from plasma derived from a fasted, healthy donor, acetylated by the repeat addition of acetic anhydride and labelled as described previously [20]. The extent of modi¢cation was determined by relative electropho-
BBAMCB 55755 23-2-01
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
retic mobility (REMv3 compared to native LDL) [27]. Mice were injected with 100 Wl acLDL containing 43.6 Wg protein, 8.18 kBq 14 C-7K oleate and 29.0 kBq 3 H-cholesteryl oleate. The speci¢c activity of 3 H-cholesteryl oleate in this vehicle was estimated to be 2.0 MBq/Wmol based on a protein recovery of 59% and recovery of both radiolabelled and unlabelled lipid of 15%, as determined previously [20]. The speci¢c activity of 14 C-7K oleate was 2.0 MBq/ Wmol. 2.3. CMR CMR was prepared essentially as described previously [24^26] with the following minor modi¢cations: water containing 2.2% glycerol (w/v) was replaced with Tricine (pH 8.4) containing 2.2% glycerol (w/v); ethylenediaminetetraacetic acid (EDTA; 27 Wmol/l) was included in the preparation bu¡er; two radiolabelled steryl esters were included. Tricine was used to match the bu¡er used in the preparation of radiolabelled acLDL and EDTA was included to avoid autoxidation of the lipid components and generation of oxysterols by the titanium sonicator tip during preparation of the emulsion (A.J. Brown, unpublished observation). The emulsion was prepared from a mixture of triolein, DOPC, cholesteryl oleate and cholesterol (plus the two radiolabelled steryl esters) in the weight ratio of 4.5: 2.5: 0.8: 0.8, respectively, and was ¢lter-sterilised (0.2 Wm) prior to injection. Animals received an injection of 100 Wl containing 6.08 kBq 14 C-7K oleate and 18.3 kBq 3 H-cholesteryl oleate (assayed after ¢lter sterilisation). The CMR produced by the described method was reported to generate particles with diameter 55 þ 3 nm (n = 40) after puri¢cation by ultracentrifugation [25]. In this study, the availability of 14 C-7K oleate was limited as it was necessary to synthesise it from 14 C-cholesterol and thus the emulsion was used without ultracentrifugation. It has been reported previously that particles prepared by this method and isolated in three size classes by ultracentrifugation produced particles with diameters 47.8, 80.0 and 152 nm [28]. The CMR generated without ultracentrifugation therefore likely represents a continuum of sizes across this range. Furthermore, the particle size was correlated with the TAG/phospholipid (PL) ra-
211
tio [28]. Therefore, as a control based on this correlation, the TAG/PL ratio was inverted to 2.5/4.5 (TAG/PL, w/w) so as to produce smaller particles. The behaviour of these smaller particles was compared with that of the two main vehicles (acLDL and CMR) in the delivery of 14 C-7K oleate and 3 H-cholesteryl oleate as an indication of the dependency of the sterols' metabolism on particle size. Apart from the inversion of the TAG/PL ratio, the emulsion (labelled PL/TAG to indicate the greater proportion of PL) was generated in an identical manner to the CMR. The speci¢c activity of 3 H-cholesteryl oleate was estimated to be 0.25 MBq/Wmol and that of 14 C-7K oleate determined to be 2.0 MBq/ Wmol in both the CMR and PL/TAG vehicles. 2.4. Animals Male C57BL/6J mice were obtained from Biological Resources Centre, University of New South Wales (Sydney, N.S.W., Australia) and kept under a 12 h light/dark cycle and allowed free access to water and chow. All procedures were conducted in accordance with the Central Sydney Area Health Service Animal Welfare Committee. The mice were selected at random for each condition (n = 21 each condition) and time-point (n = 3) and weighed 23.6 þ 1.9 g (mean þ S.D., n = 42). Mice were mildly anaesthetised using methoxy£urane, an inhalant anaesthetic administered via a nose cone. They were then injected with one of the three radiolabelled particles via the tail vein after its immersion in warm water to promote vasodilation. The animals were then allowed to recover under a heat lamp with supervision until fully conscious. 2.5. Tissues and sample preparation At the appropriate time, animals were injected intraperitoneally with sodium pentobarbitone (400 Wl, 10 mg/ml) in conjunction with administration of methoxy£urane. When the animals were fully anaesthetised, blood was taken by cardiac puncture and added to microfuge tubes containing EDTA (20 Wl, 200 mmol/l) for collection of plasma. The animals were then perfused for 5 min with phosphate-bu¡ered saline (PBS) containing EDTA (1 mmol/l) and butylated hydroxytoluene (BHT; 20 Wmol/l) by a
BBAMCB 55755 23-2-01
212
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
gravity-feed system. The liver and intestine, from immediately below the stomach to the rectum, were then excised. The aorta, from the heart to immediately below the iliac-femoral bifurcation, was freed of connective tissue in situ and then removed. All tissues were weighed immediately and stored at 380³C until further analysis. Livers were homogenised with one volume and intestines with two volumes of the same bu¡er used for perfusion but also containing chloramphenicol (100 mg/l; Boehringer Mannheim, Castle Hill, N.S.W., Australia) using an Ultra-Turrax T8 hand held disperser (IKA, Crown Scienti¢c, Moorebank, N.S.W., Australia). Homogenates were also stored at 380³C prior to analysis. 2.6. Determination of radioactivity All determinations of radioactivity were conducted using a TriCarb 2100TR Liquid Scintillation Analyser (Canberra-Packard) and a dual-label protocol. Results (mean þ S.D., n = 3 for each time-point) are presented as the quantity of isotope present in any given sample as a percentage of the total injected dose. The total volume of plasma (ml) was estimated
by multiplication of the weight of the mouse (g) by 49.2 and division of the product by 1000 [29]. Plasma (200 Wl) was analysed by the addition of scintillant (4.5 ml; Ultima Gold). Tissue homogenates were extracted by the method of Folch et al. [30] and the aqueous-soluble fraction assayed for radioactivity essentially as described previously [20] using approximately 200 mg liver homogenate or 300 mg intestinal homogenate. The aqueous-soluble radioactivity was determined as an indication of metabolism to bile acids or bile acid-like products. Total radioactivity in homogenates and aorta was determined by solubilisation using a tissue solubiliser (Soluene 350, Canberra-Packard) according to the manufacturer's instructions and as detailed previously [20]. The total dissected aorta and similar amounts of homogenates as used for the determination of aqueous-soluble radioactivity were used for determination of total radioactivity. Lipid-soluble radioactivity was calculated as the di¡erence between the total and aqueous-soluble radioactivities. 3. Results 3.1. Plasma
Fig. 1. Clearance of esters of 14 C-7K and 3 H-cholesterol from plasma of mice when administered incorporated into acLDL or CMR. AcLDL (open symbols) or CMR (¢lled symbols) was administered intravenously and the cholesterol-derived 3 H (circles) and 7K-derived 14 C (squares) were measured in plasma over the time-course. Radioactivities were determined by the addition of scintillation £uid and analysis by liquid scintillation counting with a dual-label protocol. Data are expressed as a percentage of the total dose injected and are reported as mean+S.D. (n = 3).
Clearance of radioactivity contained within acLDL from the plasma was essentially complete within a matter of minutes (Fig. 1), consistent with results obtained in other studies [17^19]. 14 C derived from 7K reappeared transiently then decreased to very low levels from 12 h onwards while 3 H derived from cholesterol increased to 6 h then reached a plateau. These data for acLDL are in excellent agreement with those derived from rats that we have published previously [20]. In contrast, the radioactivity contained in CMR was contained wholly within the plasma compartment at the earliest time-point, 2 min. The two labels were cleared from the plasma in an identical fashion up to 1 h and had diverged by 3 h. This pattern of clearance up to 1 h is consistent with that demonstrated by Rensen et al. [28] using three di¡erently sized populations of CMR injected into in mice. 14 C was then cleared to a greater extent than 3 H and was indistinguishable from the acLDL result for 14 C after 12 h. A slightly lower level of 3 H was retained in the plasma when delivered in CMR
BBAMCB 55755 23-2-01
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
213
Fig. 2. Clearance of 14 C-7K and to 3 H-cholesterol from the liver of mice when the steryl esters were delivered in either acLDL (A) or CMR (B). Cholesterol-derived 3 H (circles) and 7K-derived 14 C (squares) were determined for total liver radioactivity (¢lled symbols, solid line) and aqueous-soluble radioactivity (open symbols, broken line). Radioactivity was determined by liquid scintillation counting using a dual-label protocol from samples of solubilised tissue homogenate (total) and the aqueous phase of a tissue homogenate lipid extraction [30]. Data are expressed as a percentage of the total dose injected and are reported as mean+S.D. (n = 3).
compared with acLDL. The disappearance of the labels from the plasma is consistent with particle clearance. The plasma data derived from the PL/ TAG-treated animals were essentially the same as those derived from the CMR-treated animals (data not shown). 3.2. Liver Uptake of radioactivity by the liver (Fig. 2) was rapid for acLDL and CMR vehicles with approximately 45% of the injected dose appearing after 30
min for both isotopes, although it was initially more rapid for acLDL. Over the 24 h time-course, the level of radioactivity detected in the liver was very similar for both isotopes independent of the delivery particle. Aqueous-soluble products of both 3 H-cholesterol and 14 C-7K were detected in the liver. However, aqueous-soluble 14 C was found at a greater level than aqueous-soluble 3 H across the time-course for both vehicles. The hepatic data derived from the PL/ TAG-treated animals were essentially the same as those derived from the CMR-treated animals (data not shown).
Fig. 3. Appearance in the intestine of 7K-derived 14 C and cholesterol-derived 3 H when the steryl esters were administered to mice incorporated into either acLDL (A) or CMR (B). 3 H (circles) and 14 C (squares) were determined for total liver radioactivity (¢lled symbols, solid line) and aqueous-soluble radioactivity (open symbols, broken line). Radioactivity was determined by liquid scintillation counting using a dual-label protocol from samples of solubilised tissue homogenate (total) and the aqueous phase of a lipid extraction [30] of tissue homogenate. Data are expressed as a percentage of the total dose injected and are reported as mean+S.D. (n = 3).
BBAMCB 55755 23-2-01
214
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
3.3. Intestine A much larger proportion of 14 C-7K-derived radioactivity appeared in the intestine for both acLDL and CMR than the respective level of 3 H-cholesterolderived radioactivity (Fig. 3). The response for each of the sterol-derived radioactivities was similar between delivery vehicles although the 14 C delivered in acLDL (Fig. 3A) that was detected at 6 h was less in comparison with CMR (Fig. 3B). The aqueous-soluble 14 C fraction was greater than the aqueous-soluble 3 H in both cases. These results are consistent with our previous ¢ndings although the aqueous-soluble 14 C portion was found to be the major component of the intestinal pool of 14 C in our earlier study [20]. The intestinal data derived from the PL/TAG-treated animals were essentially the same as those derived from the CMR-treated animals (data not shown).
the aortic levels were approximately 0.005% of the injected dose of 14 C (Fig. 4B) after 2 min whereas mice administered PL/TAG emulsion had approximately 0.018% of the injected dose at the same time-point (Fig. 4C). The same was observed for cholesterol-derived 3 H reaching levels of 0.008 and 0.02% of the injected dose, respectively.
3.4. Aorta Greater levels of both 3 H derived from cholesterol and 14 C derived from 7K were detected in the aortas of animals administered acLDL (Fig. 4A) than the respective isotope in animals given CMR (Fig. 4B). Most importantly, while 14 C was detected in both treatment groups at earlier time-points, its appearance was transient and no accumulation was demonstrated. After 24 h, the raw 14 C radioactivity (dpm) was at the limit of detection (910 dpm). On the other hand, 3 H increased over the 24 h period. Consistent with CMR-treated mice (Fig. 4B), no accumulation of 14 C could be detected in aortas of animals treated with PL/TAG emulsion (Fig. 4C), whereas cholesterol-derived 3 H increased to similar levels in both of these groups. There was a noticeable di¡erence in the level of radioactivity detected in the aorta at early time-points. When treated with CMR, C
Fig. 4. Determination of 7K-derived 14 C and cholesterol-derived 3 H in the aortas of mice when the steryl esters were administered incorporated into either acLDL (open symbols, A), CMR (shaded symbols, B) or PL/TAG (¢lled symbols, C). 3 H (circles) and 14 C (squares) were determined by liquid scintillation counting using a dual-label protocol from the solubilised whole aorta. Data are expressed as a percentage of the total dose injected multiplied by 1000 and are reported as mean+S.D. (n = 3).
BBAMCB 55755 23-2-01
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
Fig. 5. Faecal and urinary excretion of 7K-derived 14 C and cholesterol-derived 3 H after administration to mice of the steryl esters incorporated into acLDL, CMR or PL/TAG. Faecal (A) and urinary (B) samples from three animals for each treatment were pooled over 24 h and subjected to solubilisation to determine total radioactivity and to lipid extraction [30] to determine the aqueous-soluble radioactivity (open). Lipid-soluble radioactivity (¢lled) was calculated as the di¡erence between total and aqueous-soluble radioactivity. Radioactivity was assayed by liquid scintillation counting using a dual-label protocol. Data were calculated for the total, pooled sample, averaged between three animals and expressed as a percentage of the total dose injected per animal per 24 h.
3.5. Faeces and urine Pooled samples of faeces and urine were collected from animals in each of the 24 h time-point treatment groups (n = 3 per group). A large proportion of 14 C-7K-derived radioactivity, amounting to approximately one-third of the dose, was detected in the faeces of animals treated with any vehicle after 24 h (Fig. 5A). Of this, 20^30% was aqueous-soluble, presumably having undergone complete conversion to
215
bile acids. Approximately twice the amount of 3 Hcholesterol-derived radioactivity originating from acLDL was measured in the faeces compared with CMR but in both cases the total amount was much less than the faecal 14 C. The proportion of aqueoussoluble products was similar in the two groups. In comparison, the excretion of radiolabels into the faecal material (Fig. 5A) was similar if not slightly higher for the PL/TAG emulsion animals but 14 C excretion remained substantially greater than that of 3 H. Thus, a substantial proportion of the injected dose of 14 C was excreted into the faeces in the initial 24 h period, irrespective of the delivery vehicle, indicating that 7K undergoes rapid excretion. Low levels of radioactivities, each amounting to less than 1% of the injected dose, were detected in the urine (Fig. 5B). 3 H was found to be similar in all cases but slightly more 14 C was detected when 7K was administered in acLDL. The bulk of the radioactivity was aqueous-soluble, but the lipid-soluble portion was detectable. Approximately two- to three-fold the amount of 14 C compared with 3 H was excreted by all groups of animals in the urine. This level comprised only 0.3^0.5% of the injected dose of 14 C. It is evident from the appearance of 14 C in the faeces that 7K undergoes much more e¤cient excretion from the body than cholesterol. 4. Discussion Oxysterols have been implicated in atherogenesis (reviewed in [4]). Non-enzymically formed oxysterols, such as 7K, are often assumed to originate from the diet and exert their e¡ects locally in the artery wall subsequent to their accumulation. We have previously reported that 7K undergoes rapid hepatic metabolism and does not accumulate in the aorta when delivered to rats in a high-uptake modi¢ed lipoprotein (acLDL) [20]. The aim of the present study was to extend this ¢nding to mice and more signi¢cantly, to employ a more physiologically relevant vehicle, CMR, to investigate the metabolism of 7K. Similar to cholesterol, oxysterols are absorbed and packaged into chylomicrons [12^14] (although absorption data vary widely, ranging from 6% [14] to 93% [31]). Following lipolysis by lipoprotein lipase, the resultant chylomicron remnants are taken up by
BBAMCB 55755 23-2-01
216
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
hepatocytes [22]. In this study, CMR and acLDL, both of which are removed from the circulation by the liver and each containing oleoyl esters of 14 C-7K and 3 H-cholesterol, were administered intravenously to mice and the fate of the radiolabels monitored in plasma, liver, intestine and aorta over 24 h. Pooled faeces and urine were also collected over the 24 h period. For 3 H-cholesterol, the kinetics of appearance and disappearance in the circulation and liver were similar to those reported previously when delivered in acLDL to rats [19,20] or CMR to mice [28]. Reappearance in the plasma of 3 H originally incorporated into acLDL and its incomplete clearance when delivered in CMR, may indicate reverse cholesterol transport from endothelial cells to hepatocytes for 3 Hcholesterol delivered in acLDL [19] or resynthesis into lower density lipoproteins for either or both particles. Using smaller particles with a lower TAG/PL ratio (or higher PL/TAG ratio), we have demonstrated the validity of the CMR vehicle. In essence, the results obtained using the PL/TAG emulsion were indistinguishable from the CMRtreated animals although slightly higher levels of both 3 H and 14 C radioactivity were found in the aortas of animals treated with the PL/TAG emulsion. This may be explained by the greater arterial permeability a¡orded the PL/TAG emulsion particles due to their smaller diameter compared with CMR in a manner analogous to the greater permeability of chylomicron remnants versus chylomicrons in the artery wall [32]. The demonstration of rapid hepatic metabolism and excretion of 14 C-7K into the faeces and its lack of accumulation in the aorta in all three delivery vehicles, points to the robustness of this ¢nding. The role of dietary oxysterols in atherosclerosis remains equivocal. Of 16 oxysterol-feeding studies identi¢ed in the literature using various animal models, seven were pro-atherogenic, eight anti-atherogenic and one showed no clear e¡ect [33]. For dietary oxysterols to play a role in the development and/or initiation of atherosclerosis, they should accumulate within the artery wall. By injecting the sterols in CMR, we have avoided the complications of intestinal absorption and yet there is still no evidence that 7K accumulated in the aorta. Thus, we have consistently observed that while 3 H-cholesterol accumu-
lated in the aorta, 14 C derived from 7K appeared transiently and its removal was essentially complete by 24 h. Furthermore, there was no indication of enterohepatic recycling, thereby ruling out the possibility of accumulation later than 24 h. While mice do not generally possess cholesteryl ester transfer protein activity [34], it may be argued that in species that do, such as man, prolonged exposure to oxysterols contained in chylomicron remnants may allow transfer of the oxysterols to lower density, potentially atherogenic lipoproteins. For example, a study of 25-hydroxycholesterol (cholest-5en-3L,25-diol) administered orally to squirrel monkeys suggested that this oxysterol may contribute to atherogenesis because the bulk was detected in the VLDL/LDL fraction, however quantitative and excretion data were not presented [35]. Furthermore, 25-hydroxycholesterol seldom contributes to the overall oxysterol content in foods (e.g. [16]). Another study in which pigeons were fed cholesterol plus cholestan-3L,5K,6L-triol found increased coronary artery lumenal stenosis but no signi¢cant di¡erence in arterial cholesterol or cholesteryl ester content [36]. More recent, purportedly pro-atherogenic oxysterol-feeding studies have presented data highly consistent with 7K data from the present study and a previous study [20]. Three studies with rabbits that have investigated the role of oxysterols in atherosclerosis have found increased cholesterol contents in the aorta [37^39]. However, oxysterols were either not measured [39], were not elevated above the control group [38] or were not detected, despite being detected in circulating TAG-rich lipoproteins [37]. It would be interesting to determine if any of this accumulated aortic cholesterol was derived from 7K, especially in light of the ¢nding that we [20] and others [40,41] have shown that 7K can be converted to cholesterol. In another study [15], rabbits were fed in addition to cholesterol, an amount of oxysterols typical of levels of total sterols consumed by humans (on a per kg body weight basis) [4]. Despite this extreme intake, no signi¢cant increase was found in the concentration of 7K in the combined VLDL and chylomicron fraction although increases were observed in three other oxysterols [15]. In addition, there was no signi¢cant increase in any of the oxysterols in LDL [15]. It is conceivable that the increased oxysterol content in the VLDL/chylomicron fraction could be
BBAMCB 55755 23-2-01
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
solely due to newly ingested oxysterols packaged in chylomicrons which then undergo lipolysis, hepatic uptake and subsequent metabolism. Indeed, the authors demonstrated increased levels of oxysterols in the liver, including 7K. It was claimed that levels of 7K in LDL were not increased because it is transported by serum albumin rather than lipoproteins [15]. However, others have shown that 7K distribution among lipoproteins re£ects cholesterol distribution [42^44] even at radiotracer concentrations [4]. The aortic area covered by fatty streaks was found to be greater in the group of rabbits fed oxysterols but notably, neither the cholesterol nor the oxysterol concentrations of the aorta were reported [15]. Another investigation by the same group [45], examined the e¡ect of feeding oxysterols to two murine models of atherosclerosis, apolipoprotein E (apoE)de¢cient and LDL receptor-de¢cient mice. Due to impaired lipoprotein uptake, these mice have increased levels of chylomicron remnants and VLDL [46,47] and LDL and intermediate-density lipoprotein (IDL) [47,48], respectively. It may be argued that these animals better represent models to test atherogenic parameters because both are hypercholesterolaemic [46,48]. In addition, the apoE-de¢cient mice develop spontaneous atherosclerosis [46] which is exacerbated by cholesterol-feeding [49] and the LDL receptor-de¢cient mice develop atherosclerosis when fed a high-cholesterol diet [49]. Chylomicron remnant uptake is dependent upon apoE recognition [22,47,50] by the LDL receptor pathway and a second, slower endocytosis mechanism [47,50]. Both of these animal models demonstrate impaired cellular metabolism of chylomicron remnants [50]. This is pertinent to potential criticism of the present study because it is our contention that under normal circumstances chylomicron remnant clearance is rapid and thus 7K metabolism will be correspondingly swift. Under conditions in which clearance mechanisms are perturbed, dietary 7K could potentially contribute to atherogenesis. In this context it should be noted that remnant particles derived from both chylomicrons and VLDL are an emergent risk factor for the development of atherosclerosis [32,51]. Chylomicron remnants are known to penetrate arterial tissue and can lead to focal accumulation of these particles in the subendothelial space [32]. Patients with famil-
217
ial hypercholesterolaemia and Watanabe Heritable Hyperlipidaemic (WHHL) rabbits, both LDL receptor-de¢cient, also display impaired clearance of chylomicron remnants [32]. It also appears that chylomicron remnant metabolism may be perturbed in most cases of atherogenic risk, for example diabetes [52] and even normolipidaemic patients with coronary artery disease [32]. Therefore, due to the risk that postprandial lipoproteins represent, dietary oxysterols such as 7K may still play a role in atherogenesis but that role may have been overstated. This is the ¢rst study to directly compare a commonly used, chemically modi¢ed lipoprotein with a more physiologically relevant vehicle, CMR, regarding the metabolism of 7K versus cholesterol. 7K was metabolised and excreted primarily in the faeces more rapidly than cholesterol, independent of its delivery vehicle. Moreover, the data indicate that 7K does not accumulate in the aorta to any signi¢cant extent over this period of time. The present investigation con¢rms our previous ¢ndings and raises doubts about the assertion that dietary oxysterols, 7K in particular, accumulate in the artery wall and contribute to atherogenesis. Further studies are necessary in which various animal models of atherosclerosis are fed oxysterols (including a labelled oxysterol) at levels realistically encountered in the human diet. Acknowledgements We thank Dr Wendy Jessup for critically reading this manuscript and Prof. Trevor Redgrave for scienti¢c discussion. This work was supported by a grant from the American Egg Board. References [1] L.L. Smith, Cholesterol Autoxidation, Plenum, New York, 1981. [2] H.N. Hodis, A. Chauhan, S. Hashimoto, D.W. Crawford, A. Sevanian, Atherosclerosis 96 (1992) 125^134. [3] I. Bjo«rkhem, O. Andersson, U. Diczfalusy, B. Sevastik, R.-J. Xiu, C. Duan, E. Lund, Proc. Natl. Acad. Sci. USA 91 (1994) 8592^8596. [4] A.J. Brown, W. Jessup, Atherosclerosis 142 (1999) 1^28. [5] M. Crisby, J. Nilsson, V. Kostulas, I. Bjo«rkhem, U. Diczfalusy, Biochim. Biophys. Acta 1344 (1997) 278^285.
BBAMCB 55755 23-2-01
218
M.A. Lyons, A.J. Brown / Biochimica et Biophysica Acta 1530 (2001) 209^218
[6] A.J. Brown, S.-l. Leong, R.T. Dean, W. Jessup, J. Lipid Res. 38 (1997) 1730^1745. [7] N.B. Javitt, FASEB J. 8 (1994) 1308^1311. [8] E. Lund, O. Andersson, J. Zhang, A. Babiker, G. Ahlborg, U. Diczfalusy, K. Einarsson, J. Sjovall, I. Bjo«rkhem, Arterioscler. Thromb. Vasc. Biol. 16 (1996) 208^212. [9] M.A. Lyons, A.J. Brown, Int. J. Biochem. Cell Biol. 31 (1999) 369^375. [10] H.N. Hodis, D.W. Crawford, A. Sevanian, Atherosclerosis 89 (1991) 117^126. [11] L.L. Smith, B.H. Johnson, Free Radic. Biol. Med. 7 (1989) 285^332. [12] H.A. Emanual, C.A. Hassel, P.B. Addis, S.D. Bergmann, J.H. Zavoral, J. Food Sci. 56 (1991) 843^847. [13] K. Osada, E. Sasaki, M. Sugano, Lipids 29 (1994) 555^559. [14] D.F. Vine, K.D. Croft, L.J. Beilin, J.C. Mamo, Lipids 32 (1997) 887^893. [15] I. Stapra¬ns, X.-M. Pan, J.H. Rapp, K.R. Feingold, Arterioscler. Thromb. Vasc. Biol. 18 (1998) 977^983. [16] J.H. Nielsen, C.E. Olsen, C. Duedahl, L.H. Skibsted, J. Dairy Res. 62 (1995) 101^113. [17] T.J.C. van Berkel, J.F. Nagelkerke, L. Harkes, J.K. Kruijt, Biochem. J. 208 (1982) 493^503. [18] R. Blomho¡, C.A. Drevon, W. Eskild, P. Helgerud, K.R. Norum, T. Berg, J. Biol. Chem. 259 (1984) 8898^8903. [19] H.F. Bakkaren, F. Kuipers, R.J. Vonk, T.J.C. van Berkel, Biochem. J. 268 (1990) 685^691. [20] M.A. Lyons, S. Samman, L. Gatto, A.J. Brown, J. Lipid Res. 40 (1999) 1846^1857. [21] P.H.E. Groot, T.J.C. van Berkel, A. van Tol, Metabolism 30 (1981) 792^797. [22] T.G. Redgrave, Int. Rev. Physiol. 28 (1983) 103^130. [23] M.C.M. van Dijk, M. Pieters, T.J.C. van Berkel, Eur. J. Biochem. 211 (1993) 781^787. [24] T.G. Redgrave, R.C. Maranha¬o, Biochim. Biophys. Acta 835 (1985) 104^112. [25] I.J. Martins, T.G. Redgrave, J. Lipid Res. 39 (1998) 691^ 698. [26] B.-J. Zeng, B.-C. Mortimer, I.J. Martins, U. Seydel, T.G. Redgrave, J. Lipid Res. 39 (1998) 845^860. [27] A.J. Brown, R.T. Dean, W. Jessup, J. Lipid Res. 37 (1996) 320^335. [28] P.C.N. Rensen, N. Herijgers, M.H. Netscher, S.C.J. Meskers, M. van Eck, T.J.C. van Berkel, J. Lipid Res. 38 (1997) 1070^1084. [29] M.D. Buckland, Mowlem, Whatley, A Guide to Laboratory Animal Technology, Heinemann Medical, London, 1981. [30] J. Folch, M. Lees, G.H. Sloane-Stanley, J. Biol. Chem. 226 (1957) 497^509.
[31] J. Bascoul, N. Domergue, J. Mourot, G. Dery, A. Crastes de Paulet, Lipids 21 (1986) 744^747. [32] J.C.L. Mamo, S.D. Proctor, D. Smith, Atherosclerosis 141 (1998) S63^S69. [33] M.A. Lyons, A.J. Brown, in: R.T. Dean, D.Kelly (Eds.), Atherosclerosis, Oxford University Press, Oxford, 2000, in press. [34] S. Jiao, T.G. Cole, R.T. Kitchens, B. P£eger, G. Schonfeld, Metabolism 39 (1990) 155^160. [35] S.-K. Peng, C.B. Taylor, E.H. Mosbach, W.Y. Huang, J. Hill, B. Mikkelson, Atherosclerosis 41 (1982) 395^402. [36] M.S. Jacobson, M.G. Price, A.E. Shamoo, F.P. Heald, Atherosclerosis 57 (1985) 209^217. [37] D.F. Vine, J.C.L. Mamo, L.J. Beilin, T.A. Mori, K.D. Croft, J. Lipid Res. 39 (1998) 1995^2004. [38] J. Rong, S. Rangaswamy, L. Shen, R. Dave, Y. Chang, H. Peterson, H. Hodis, G. Chisolm, A. Sevanian, Arterioscler. Thromb. Vasc. Biol. 18 (1998) 1885^1894. [39] J.X. Rong, L. Shen, Y.H. Chang, A. Richters, H.N. Hodis, A. Sevanian, Arterioscler. Thromb. Vasc. Biol. 19 (1999) 2179^2188. [40] J.S.M. Sarma, R.J. Bing, S. Ikeda, R. Fischer, Artery 2 (1976) 153^160. [41] S.A. Erickson, A.D. Cooper, S.M. Matsui, R.G. Gould, J. Biol. Chem. 252 (1977) 5186^5193. [42] C.J.W. Brooks, R.M. McKenna, W.J. Cole, J. MacLachlan, T.D.V. Lawrie, Biochem. Soc. Trans. 11 (1983) 700^701. [43] I.C. Gelissen, A.J. Brown, E.L. Mander, L. Kritharides, R.T. Dean, W. Jessup, J. Biol. Chem. 271 (1996) 17852^ 17860. [44] R.G. Dyer, M.W. Stewart, J. Mitcheson, K. George, M.M. Alberti, M.F. Laker, Clin. Chim. Acta 260 (1997) 1^13. [45] I. Stapra¬ns, X.-M. Pan, J.H. Rapp, C. Grunfeld, K.R. Feingold, Arterioscler. Thromb. Vasc. Biol. 20 (2000) 708^714. [46] S.H. Zhang, R.L. Reddick, J.A. Piedrahita, N. Maeda, Science 258 (1992) 468^471. [47] S. Ishibashi, J. Herz, N. Maeda, J.L. Goldstein, M.S. Brown, Proc. Natl Acad. Sci. USA 91 (1994) 4431^4435. [48] S. Ishibashi, M.S. Brown, J.L. Goldstein, R.D. Gerard, R.E. Hammer, J. Herz, J. Clin. Invest. 92 (1993) 883^893. [49] J.L. Breslow, Science 272 (1996) 685^688. [50] B.C. Mortimer, D.J. Beveridge, I.J. Martins, T.G. Redgrave, J. Biol. Chem. 270 (1995) 28767^28776. [51] M.L. Brown, M.P. Ramprasad, P.K. Umeda, A. Tanaka, Y. Kobayashi, T. Watanabe, H. Shimoyamada, W.-L. Kuo, R. Li, R. Song, W.A. Bradley, S.H. Gianturco, Proc. Natl Acad. Sci. USA 97 (2000) 7488^7493. [52] S.D. Proctor, C.K. Pabla, J.C.L. Mamo, Atherosclerosis 149 (2000) 315^322.
BBAMCB 55755 23-2-01