Fluorometric determination of total retinyl esters in ... - CiteSeerX

2 downloads 0 Views 325KB Size Report
pemic samples interferes with the fluorescence mea- surement, all samples were incubated for 10 min with a clearing buffer containing esterase and detergents.
Clinical Chemistry 44:7 1459 –1465 (1998)

Lipids and Lipoproteins

Fluorometric determination of total retinyl esters in triglyceride-rich lipoproteins ¨ ¨ ru¨mez,1 Matthias Orth,1,2* Monika Hanisch,2 Gerd Kro¨ning,1 Mustafa Porsch-Ozcu 2 1,2 Heinrich Wieland, and Claus Luley

A time-consuming sample preparation and measuring procedure is required for the quantitation of retinyl palmitate by HPLC. We developed a fluorometric method for the determination of total retinyl esters in chylomicrons, chylomicron remnants, and VLDL. This method is precise, sensitive, rapid, simple, and particularly useful for large-scale studies of postprandial lipid metabolism. Because the turbidity of postprandial lipemic samples interferes with the fluorescence measurement, all samples were incubated for 10 min with a clearing buffer containing esterase and detergents. This buffer eliminates the turbidity and hydrolyzes all retinyl esters to retinol. The fluorescence signal (excitation wavelength, 330 nm; emission wavelength, 490 nm) was linear from 0.1 mg/L up to 4 mg/L retinyl palmitate, and the CVs were 3.6% within-run and 5.1% within-series. A first application studied postprandial lipoproteins, which were first separated by ultracentrifugation and then subjected to size exclusion chromatography. Fluorescence analysis revealed that the chylomicron density fraction contains large amounts of chylomicron remnants. Several observations indicate that chylomicron remnants can cause and sustain atherogenesis. In cell culture, chylomicron remnants are taken up by smooth muscle cells (1 ) and induce the transformation of macrophages to foam cells (2, 3 ). In cholesterol-fed rabbits, increased remnant lipoproteins cause marked hypercholesterolemia associated with rapidly growing atherosclerosis (4 – 6 ), and in humans, the hereditary disease “type III hyperli-

1 Institut fu¨r Klinische Chemie, Klinikum der Otto-von-Guericke-Universita¨t, 39120 Magdeburg, Germany. 2 Abteilung Klinische Chemie, Albert-Ludwigs Universita¨t, 79106 Freiburg, Germany. * Address correspondence to this author at: Institut fu¨r Klinische Chemie, Klinikum der Otto-von-Guericke-Universita¨t, Haus 39, Leipziger Str. 44, D-39120 Magdeburg, Germany. Fax 49-391-671-3902; e-mail matthias.orth @medizin.uni-magdeburg.de. Received December 30, 1997; revision accepted April 2, 1998.

poproteinemia”, an accumulation of remnant lipoproteins, is accompanied by severe atherosclerosis (7 ). Hence, it may be important for the assessment of atherosclerotic risk to test individuals for the presence of increased concentrations of postprandial triglyceride-rich lipoproteins, in particular of chylomicron remnants (8, 9 ). The plasma concentrations of chylomicron remnants cannot be assessed by the plasma triglyceride concentration alone, because this value is composed of triglycerides transported in several lipoprotein classes secreted from different organs. Exogenous triglycerides are transported in chylomicrons and chylomicron remnants, whereas the bulk of endogenous triglycerides are transported in VLDL. To differentiate between triglyceride-rich lipoproteins of intestinal origin (chylomicrons and their remnants) and lipoproteins of hepatic origin (VLDL), only two methods are currently available. One method is the specific determination of apolipoprotein B48 (a marker for the intestinal origin of the lipoprotein), and the alternative method is the labeling of lipoproteins of intestinal origin with retinyl palmitate (10 ). Because the former determination is difficult to obtain by immunological techniques due to the identity of apolipoprotein B48 with the amino terminus of apolipoprotein B100 (a marker for lipoproteins secreted by the liver), several investigators have decided to determine retinyl esters after a fatty meal containing high doses of retinyl palmitate as a marker of chylomicrons and chylomicron remnants (10 –15 ). This is possible because retinyl esters are incorporated into chylomicrons during intestinal assembly, and they remain an integral part of the lipoprotein particle during the intraplasmatic degradation of chylomicrons to chylomicron remnants until the chylomicron remnants are cleared by liver and bone marrow. Retinyl ester concentrations are very low under physiological conditions, and the liver does not secrete retinyl esters. Therefore, the plasma concentration of retinyl esters after the ingestion of a fatty meal containing high doses of retinyl palmitate is a good estimate of the plasma concentration of chylomicrons and chylomicron remnants (16 ).

1459

1460

Orth et al.: Fluorometry of retinyl esters

The conventional method for the determination of retinyl esters in plasma is HPLC. This method is specific and precise but requires time-consuming sample preparation and measuring procedures, and is therefore not appropriate for the determination of retinyl esters in large sample numbers. We therefore strove to develop an alternative procedure that allows easier and more rapid determination of retinyl esters in triglyceride-rich lipoproteins. For this purpose, we took advantage of the fluorescent properties of retinol (17 ). If used with an appropriate clearing buffer, this new procedure turned out to be precise, sensitive, rapid, and very easy to perform.

Materials and Methods The protocol for all studies involving human subjects was approved by the local ethics committee.

oral fat loads After an overnight fast, male volunteers received a liquid fatty meal consisting of fresh dairy cream and vitamin A (AROVITt, Hoffmann-La Roche). The fatty meal (1 g fat per kg body weight and 60 000 IU vitamin A (retinyl palmitate) per m2 body surface area) was consumed within 10 min (18 ). Similar fatty meals without retinyl palmitate were administered to obtain lipoproteins that are not labeled with retinyl esters.

clearing buffer Samples were mixed 1:1 (by volume) with a clearing buffer [1.2 g/L polyoxyethylene 9-lauryl ether (Boehringer Mannheim), 220 mmol/L magnesium aspartate (Fluka), 40 mmol/L sodium cholate (Merck), 2000 U/L cholesterolesterase (E.C. 3.1.1.13; Boehringer Mannheim), and 100 mmol/L Tris buffer, pH 7.7] and incubated for 10 min at room temperature before being submitted to the fluorescence measurement. The clearing buffer was stable at 4 °C for up to 1 week.

fluorescence measurement Fluorescence was determined in a Hitachi spectrofluorometer with a flow-through cuvette (Model F 1050, Merck) or with a LS-50 spectrofluorometer with a standard cuvette (Perkin–Elmer). Wavelengths were 330 nm for excitation and 490 nm for emission. Because the substance producing the fluorescence signal in the sample was retinol, the method was calibrated with retinol dissolved in ethanol (20 ml of retinol diluted into 500 ml with saline 1 500 ml of clearing buffer). The concentration of retinol can be measured by spectrophotometry. The calibration (in mg/L) was carried out using the equation: Fluorescence signal(retinol) 5 Fluorescence signal(retinyl palmitate) 3 0.5549 This factor is derived from the molecular weights of retinol and retinyl palmitate. We also compared the calibration with retinol with the

calibration with chylomicron samples containing known concentrations of retinyl palmitate to rule out the possibility that differences in background signal and quenching of the calibrators interfere with the calibration. (The retinyl palmitate concentrations of these chylomicron preparations from volunteers obtained 4 h after a retinyl palmitate-containing fatty meal had been measured before by HPLC.) Because dilution curves show linear signals over a wide fluorescence range, calibration was carried out as a single-point calibration.

hplc of retinyl esters All solvents were of HPLC quality and purchased from Merck. All assays were carried out in subdued light to prevent chemical alterations. Retinyl palmitate (0.05–1.2 mg/L; Sigma Chemical Co.) dissolved in sample solution [acetonitrile/ethanol; 6:4 (by volume)] served as a standard. The concentrations of retinyl palmitate and the internal standard retinyl margarinate (Hoffmann-La Roche) dissolved in ethanol were measured by spectrophotometry. Lipoproteins (200 ml) were first extracted with four volumes of ethanol containing the internal standard (0.25 mg/L retinyl margarinate) and then with 20 volumes of hexane. The hexane layer was evaporated under nitrogen, and the residue was dissolved in 200 ml of acetonitrile/ ethanol 6:4 (by volume) containing 0.5 mL/L butylated hydroxytoluene. The sample (50 ml) was injected into a Merck HPLC system, and retinyl esters were separated by reversed-phase HPLC using an analytical C8-bonded silica Lichrospher RP 8-100 HPLC column (Bischoff) with RP18 Lichrocart 4-4 guard columns (Merck) and a mobile phase of 940 mL/L acetonitrile–50 mL/L tetrahydrofuran–10 mL/L water at a flow rate of 1.5 mL/min. The absorbance of the retinyl esters (oleate, margarinate, palmitate, and stearate) and of retinol was recorded at 325 nm. The detection limit for retinyl palmitate was 10 mg/L. Extraction recovery was 90.7% 6 5.5% for retinyl palmitate and 89.0% 6 6.4% for retinyl margarinate (mean 6 SD).

isolation of lipoproteins from plasma With the subjects in a standardized sitting position, EDTA blood samples were collected by clean venipuncture 4 h after the fatty meal, when chylomicrons and chylomicron remnants reach their maximum concentrations. Chylomicrons (Sf .1000) were isolated by preparative ultracentrifugation of postprandial plasma, using an L5 ultracentrifuge (Beckman) and a Ti 50.3 rotor (1.6 3 106 g, 10 °C, 30 min). VLDL, intermediate density lipoprotein (IDL), and LDL were prepared by sequential ultracentrifugation of the chylomicron infranatant at densities of 1.006, 1.019, and 1.063 kg/L, respectively, for 18 h each. The investigation of unspecific fluorescence was car-

Clinical Chemistry 44, No. 7, 1998

ried out with lipoproteins obtained after gel filtration of postprandial plasma. This separation procedure was chosen because it allows the continuous visualization of the entire lipoprotein size range. Because the separating properties of a single Sepharose product do not cover the entire size range of human lipoproteins, a combination of two different Sepharose media was used. The lower 53 cm of glass columns (1000 mm length, 10 mm i.d.) were filled with Sepharose CL-6B, the upper 43 cm with Sepharose CL-2B (Pharmacia). Plasma (2.5 mL) was applied at a flow rate of 100 mL/min (filtration buffer: 145 mmol/L NaCl, 0.21 mmol/L Na2EDTA, and 25 mmol/L Tris, pH 7.5). To avoid decay of retinyl esters by light, the separation was carried out in the dark. Eluates were collected in 2-mL fractions and were analyzed for triglyceride and cholesterol concentrations by enzymatic tests (Boehringer Mannheim) with an automatic analyzer (Hitachi 747, Boehringer Mannheim) and for fluorescence by fluorometry (excitation, 330 nm; emission, 490 nm) after incubation with the clearing buffer. For fluorometry, 200 ml of each fraction was mixed with 200 ml of clearing buffer. To compare lipoprotein separation by this method with results obtained after separation by ultracentrifugation, isolated lipoprotein fractions after ultracentrifugation (chylomicrons, VLDL, IDL, and LDL) were separated over the same column, and triglyceride and cholesterol concentrations and fluorescence were measured in the eluate fractions.

1461

Fig. 1. Decrease of unspecific fluorescence after incubation with the clearing buffer within 5 min. One milliliter of chylomicrons [triglyceride concentration, 12.9 mmol/L (11.4 g/L)] was isolated by ultracentrifugation from plasma obtained 4 h after the ingestion of a fatty meal (without retinyl palmitate added) and mixed with 1 mL of clearing buffer at 0 min; the decay of fluorescence was monitored during incubation at room temperature (excitation, 330 nm; emission, 490 nm).

plot (19 ), both agree well at low and high concentrations of retinyl palmitate (Fig. 3, right panel). To test for linearity, dilutions of a retinyl ester-labeled

Results Because the turbidity of lipemic samples causes considerable unspecific fluorescence, which makes the analysis of specific fluorescence impossible, a clearing buffer must be used. This clearing buffer eliminates unspecific fluorescence in chylomicron samples within 2– 6 min, depending on the sample triglyceride concentration (Fig. 1). Therefore, an incubation period of 10 min was chosen for all following analyses. The clearing buffer reliably cleared lipemic samples with triglyceride concentrations up to 11.3 mmol/L (10 g/L). The clearing buffer hydrolyzes ester bonds not only in triglycerides and cholesterol esters but also in retinyl esters. This was demonstrated by HPLC analyses of retinyl ester-containing chylomicrons before and after incubation with the clearing buffer (Fig. 2). Without clearing buffer, the major peak was retinyl palmitate; after incubation with the clearing buffer, this peak virtually disappeared, whereas a corresponding retinol peak appeared. The correlation of the results of the new fluorometric method and the HPLC method was examined with 20 chylomicron samples obtained from different individuals after the ingestion of a fatty meal containing retinyl palmitate that have been isolated by gel chromatography (Fig. 3, left panel). Results from both methods correlate well (r 5 0.989), and as is evident from the Bland-Altman

Fig. 2. HPLC analysis reveals complete hydrolysis of retinyl palmitate to retinol by the clearing buffer. Retinyl esters and retinol were measured in retinyl palmitate-labeled chylomicrons before (left) and after (right) incubation with clearing buffer. Upon incubation with the clearing buffer, the former retinyl palmitate peak virtually disappeared and a corresponding retinol peak appeared. The left chromatogram also displays retinyl oleate and retinyl stearate (incomplete peak) as minor peaks; both peaks also disappear upon incubation with the clearing buffer.

1462

Orth et al.: Fluorometry of retinyl esters

Fig. 3. (Left) Scatter plot of retinyl palmitate concentrations determined by the HPLC method (x-axis) vs the fluorescence method (y-axis) in 20 isolated chylomicron preparations from plasma samples obtained 4 h after the ingestion of a fatty meal containing retinyl palmitate. The dotted lines indicate the 95% confidence interval of the regression line. (Right) Bland-Altman plot; horizontal lines represent the mean differences and 6 SD.

chylomicron preparation (which had been isolated by gel chromatography) showed a good linearity up to 4 mg/L (r 5 0.989) (Fig. 4). If higher concentrations of retinyl ester were present, i.e., especially in samples of hypertriglyceridemic patients, the samples were diluted with saline. The CVs were tested by measuring aliquots of one chylomicron sample labeled with retinyl palmitate 10 times in series and on 10 consecutive days. The CVs within-run and between-days were 3.6% and 5.1%, respectively. To examine whether plasma lipoproteins yield unspecific fluorescence, two plasma samples obtained from the same individual were separated by gel filtration, and the fluorescence and the triglyceride concentrations in the eluates were compared (Fig. 5). Both samples were obtained 4 h after a fatty meal; however, retinyl palmitate was added to the second fatty meal only. In the sample

without retinyl palmitate, fluorescence was not detected in the eluate in the size range of chylomicrons and chylomicron remnants, and little autofluorescence was detected in the range of small VLDL and IDL (Fig. 5). High autofluorescence was present in the LDL size region and particularly in the size range of HDL and of plasma proteins. Upon addition of retinyl palmitate to the fatty meal (which specifically labels chylomicrons and chylomicron remnants), two large fluorescent peaks in the size

Fig. 5. Test for unspecific fluorescence of plasma lipoproteins by comparison of the elution profiles of postprandial plasma samples after gel filtration.

Fig. 4. Test of the linearity of the fluorescence procedure by analysis of dilutions of chylomicron remnants.

Both samples were obtained from the same subject 4 h after a fatty meal. Because retinyl palmitate (RP) was added to the second test meal only, fluorescence (FL) in the first sample indicates unspecific fluorescence, and the difference between the second and the first sample indicates specific labeling of chylomicrons and chylomicron remnants. Retinyl palmitate does not change the triglycerides (TG) elution profile. This figure demonstrates that chylomicron remnants can be identified easily by the fluorescence pattern of their retinyl ester label but not by their triglyceride content.

Clinical Chemistry 44, No. 7, 1998

1463

range of large lipoproteins appeared, indicating that chylomicrons and chylomicron remnants had now been labeled by retinyl esters. The addition of retinyl palmitate to the fatty meal did not alter the elution profile of triglycerides in lipoproteins (Fig. 5). Both triglyceride elution profiles overlapped well, indicating that the triglyceride response after repeated standardized fatty meals is very reproducible. The first application of this method was to study the distribution of the fluorescence label among lipoprotein density fractions in plasma obtained 4 h after a fatty meal containing retinyl palmitate. These fractions had been first isolated by sequential ultracentrifugation and were then further separated by gel filtration (Fig. 6). The bulk of the fluorescence was found in the chylomicron fraction (Sf .1000), which contained two lipoprotein peaks of different sizes. In the first peak, which contained large lipoproteins, the distribution of fluorescence and triglycerides corresponded well. In the second peak, however, the maxima of fluorescence and triglycerides did not correspond. The fluorescence maximum was eluting before the triglycerides maximum. Apolipoprotein B48 could be demonstrated in both fluorescence peaks by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Coomassiet staining (data not shown). Size and composition (lipids, fluorescence, and protein) of the second peak imply that this peak is composed of chylomicron remnants. Less fluorescence than in the chylomicron fraction was found in VLDL, and only minute amounts were detectable in IDL. Fluorescence in LDL was pronounced and was probably caused by retinoids (Fig. 5). Similar results were obtained when lipoprotein fractions were analyzed by fluorometry directly without additional fractionation by gel filtration [chylomicron fraction: 27.7 arbitrary units (AU); VLDL fraction, 5.8 AU; IDL fraction, 0.9 AU; LDL fraction, 32 AU].

Discussion This procedure was developed to measure retinyl esters in chylomicrons and chylomicron remnants by a rapid and easy fluorometric method (17 ) as an alternative to HPLC. Before retinyl esters in postprandial plasma can be measured by fluorometry, two problems had to be solved: the interference of lipemic turbidity with fluorometry and the interference of other fluorescing substances in plasma with the fluorescence of retinyl palmitate. The problem with the turbidity was resolved by the use of a clearing buffer, which was derived from a reagent mixture used in a commercial test kit for cholesterol determination (cat. no. 290319, Boehringer Mannheim). The cholesterol esterase contained in this kit serves two purposes. First, it liberates free cholesterol from cholesterol esters for subsequent reactions, and second, it also cleaves other ester bonds, in particular those in triglycerides, because this enzyme is rather unspecific. Thus, this cleavage of triglycerides, as well as the detergents contained in the buffer, clears the lipemic sample and allows cholesterol

Fig. 6. Comparison of the elution profiles of plasma and isolated lipoprotein density fractions (chylomicrons, VLDL, IDL, and LDL) after gel filtration. Lipoproteins were isolated by sequential ultracentrifugation from plasma obtained 4 h after a fatty meal containing retinyl palmitate. The fractions were separated by gel filtration, and cholesterol and triglyceride concentrations and fluorescence were measured in the eluate fractions.

determination even in samples in which turbidity is caused by lipemia. In our procedure, this enzyme again serves two purposes. First, it clears the turbidity of the sample by cleaving esters bonds in triglycerides, and second, it hydrolyzes all retinyl esters completely to retinol. Therefore, this procedure is highly sensitive, with

1464

Orth et al.: Fluorometry of retinyl esters

a detection limit of 0.1 mg/L retinyl palmitate. Because the substance producing the fluorescence signal is retinol, the method can be calibrated easily with retinol. The results from the new procedure agree well with the results obtained by HPLC. The method is linear up to 4 mg/L retinyl palmitate and shows satisfactory CVs. The problem with interference by other fluorescing substances in plasma was resolved by separation of triglyceride-rich lipoproteins from other plasma proteins. Two methods may be used for this separation. One method, separation by ultracentrifugation, is easy to carry out but does not separate chylomicrons and chylomicron remnants well, as we could we show in a first application of this method (Fig. 6). A large proportion of chylomicron remnants appeared in the chylomicron fraction (Sf .1000) and only a small proportion in the chylomicron infranatant. These data indicated that the analysis of retinyl esters in samples that have been separated by ultracentrifugation might underestimate the concentration of chylomicron remnants and overestimate the concentration of chylomicrons. Another method for the separation of lipoproteins, gel filtration, can separate chylomicrons and chylomicron remnants well (Figs. 5 and 6). Because gel filtration generates a large number of samples, the procedure might be further simplified by using a continuous flow system. If the eluate of a fast protein liquid chromatography column or a gel filtration column and a continuous supply of clearing buffer would be connected with the flow cell of a spectrofluorometer, both the retinyl ester concentration and the size of retinyl ester-labeled lipoproteins could be obtained simultaneously. One limitation of the method is its applicability only to lipoproteins with a density of VLDL and lower, i.e., chylomicrons and chylomicron remnants. The high unspecific fluorescence of LDL (Figs. 5 and 6) is probably caused by both the protein moiety and by retinoids in LDL. HDL and, of course, unfractionated plasma samples cannot be measured by this procedure because the unspecific fluorescence of proteins and other retinoids far exceeds the fluorescence signal of retinyl esters in the sample (17 ). This limitation, however, does not diminish the usefulness of the procedure. Retinyl esters that have been added to a fatty meal are transported almost exclusively in chylomicrons and chylomicron remnants (16 ), and both lipoproteins have only a negligible unspecific fluorescence (Fig. 5). As long as these lipoproteins are separated from other plasma proteins, the determination of their retinyl ester label by fluorescence is reliable (Fig. 3) and specific for retinyl esters (Fig. 5). It should be kept in mind that the fluorescence signal is caused by retinol derived from all retinyl esters present in a given sample and not only from retinyl palmitate. Most investigators of postprandial lipid metabolism determine only retinyl palmitate (11, 15, 20, 21 ), which normally makes up about 80% of all retinyl esters after a fatty meal containing retinyl palmitate. Because this percentage is assumed to be fairly constant, retinyl palmitate is re-

garded as representative for all retinyl esters in the sample (16 ). Because the new method determines all retinyl esters, no assumption of a constant reesterification to palmitate must be made. This might be advantageous, especially when the effects of fatty meals containing different fatty acids are to be compared. When fatty acids other than palmitate are present in high concentrations, intestinal reesterification of retinol derived from retinyl palmitate may occur with these fatty acids, and a higher proportion of retinol may be found in retinyl esters other than retinyl palmitate. The procedure is easy to carry out. Sample preparation and measurement are rapid because all that must be done is to add clearing buffer to the isolated sample and mix briefly, wait 10 min, and transfer the mixture into a spectrofluorometer, either by pipetting into a cuvette or by pumping through a flow cell. The alternative method, HPLC, requires considerably more time for sample preparation (extraction with organic solvents, phase separation, drying, and resuspension) and the measuring process; this limits the use of HPLC to studies with small sample numbers. Therefore, the new method will be particularly suitable whenever large sample numbers of retinyl ester-containing lipoproteins must be analyzed or for laboratories without HPLC equipment. Applications of this method, for example, are large-scale investigations of postprandial lipoprotein metabolism in which lipoproteins have been isolated by ultracentrifugation or studies involving the separation of lipoproteins by gel filtration. We thank Isolde Friedrich and Andrea Grasser for excellent technical assistance.

References 1. Floren CH, Albers JJ, Bierman EL. Uptake of chylomicron remnants causes cholesterol accumulation in cultured human arterial smooth muscle cells. Biochim Biophys Acta 1981;663:336 – 49. 2. Floren CH, Chait A. Uptake of chylomicron remnants by the native LDL receptor in human monocyte-derived macrophages. Biochim Biophys Acta 1981;665:608 –11. 3. Van Lenten BJ, Fogelman AM, Jackson RL, Shapiro S, Haberland ME, Edwards PA. Receptor-mediated uptake of remnant lipoproteins by cholesterol-loaded human monocyte-macrophages. J Biol Chem 1985;260:8783– 8. 4. Prior JT, Kurtz DM, Ziegler DD. The hypercholesterolemic rabbit. An aid to understanding atherosclerosis in man? Arch Pathol 1961; 71:672– 84. 5. Yla¨-Herttuala S, Lipton BA, Rosenfeld ME, Goldberg IJ, Steinberg D, Witztum JL. Macrophages and smooth muscle cells express lipoprotein lipase in human and rabbit atherosclerotic lesions. Proc Natl Acad Sci U S A 1991;88:10143–7. 6. Proctor SD, Mamo JC. Arterial fatty lesions have increased uptake of chylomicron remnants but not low-density lipoproteins. Coron Artery Dis 1996;7:239 – 45. 7. Mahley RW, Weisgraber KH, Innerarity TL, Rall SC Jr. Genetic defects in lipoprotein metabolism. Elevation of atherogenic lipoproteins caused by impaired catabolism. JAMA 1991;265:78 – 83. 8. Havel RJ. Postprandial hyperlipidemia and remnant lipoproteins. Curr Opin Lipidol 1994;5:102–9.

Clinical Chemistry 44, No. 7, 1998

9. Mack WJ, Krauss RM, Hodis HN. Lipoprotein subclasses in the Monitored Atherosclerosis Regression Study (MARS). Treatment effects and relation to coronary angiographic progression. Arterioscler Thromb Vasc Biol 1996;16:697–704. 10. Hazzard WR, Bierman EL. Delayed clearance of chylomicron remnants following vitamin-A-containing oral fat loads in broadbeta disease (type III hyperlipoproteinemia). Metabolism 1976; 25:777– 801. 11. Weintraub MS, Eisenberg S, Breslow JL. Dietary fat clearance in normal subjects is regulated by genetic variation in apolipoprotein E. J Clin Invest 1987;80:1571–7. 12. Cohn JS, Johnson EJ, Millar JS, Cohn SD, Milne RW, Marcel YL, et al. Contribution of apoB-48 and apoB-100 triglyceride-rich lipoproteins (TRL) to postprandial increases in the plasma concentration of TRL triglycerides and retinyl esters. J Lipid Res 1993;34:2033– 40. 13. Boerwinkle E, Brown S, Sharrett AR, Heiss G, Patsch W. Apolipoprotein E polymorphism influences postprandial retinyl palmitate but not triglyceride concentrations. Am J Hum Genet 1994; 54:341– 60. 14. Reznik Y, Pousse P, Herrou M, Morello R, Mahoudeau J, Drosdowsky MA, et al. Postprandial lipoprotein metabolism in normotriglyceridemic non-insulin-dependent diabetic patients: influence of apolipoprotein E polymorphism. Metabolism 1996;45: 63–71.

1465

15. Lopez-Miranda J, Ordovas JM, Ostos MA, Marin C, Jansen S, Salas J, et al. Dietary fat clearance in normal subjects is modulated by genetic variation at the apolipoprotein B gene locus. Arterioscler Thromb Vasc Biol 1997;17:1765–73. 16. Norum KR, Blomhoff R. McCollum Award Lecture, 1992: Vitamin A absorption, transport, cellular uptake, and storage. Am J Clin Nutr 1992;56:735– 44. 17. Bubb FA, Murphy GM. Determination of serum phytofluene and retinol. Clin Chim Acta 1973;48:329 –34. 18. Orth M, Wahl S, Hanisch M, Friedrich I, Wieland H, Luley C. Clearance of postprandial lipoproteins in normolipemics: role of the apolipoprotein E phenotype. Biochim Biophys Acta 1996; 1303:22–30. 19. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. 20. Groot PH, van Stiphout WA, Krauss XH, Jansen H, van Tol A, van Ramshorst E, et al. Postprandial lipoprotein metabolism in normolipidemic men with and without coronary artery disease. Arterioscler Thromb 1991;11:653– 62. 21. Brown SA, Chambless LE, Sharrett AR, Gotto A Jr, Patsch W. Postprandial lipemia: reliability in an epidemiologic field study. Am J Epidemiol 1992;136:538 – 45.