161
Biochimica et Biophysics Acta, 528 (1978) 161-175 @ Elsevier/North-Holland Biomedical Press
BBA 57119
THE CATABOLISM OF HUMAN AND RAT VERY LIPOPROTEINS BY PERFUSED RAT HEARTS
LADISLAV
DORY, DOROTHY
Department
of Biochemistry,
LOW DENSITY
POCOCK and DAVID RUBINSTEIN
McGill University, Montreal, Quebec (Canada)
(Received May 25th, 1977) (Revised manuscript received October 18th, 1977)
Summary The catabolism of human and rat 1251-labelled very low density lipoproteins (VLDL) was compared by perfusing the lipoproteins through beating rat hearts. Triacylglycerol was removed from the VLDL to a greater extent than the protein moiety, leaving remnants containing relatively more apo-B and less apo-C. The change in apo-C content of the remnants correlated with the loss of triacylglycerol. The extent of removal of triacylglycerol from the rat and human VLDL was similar and in most cases appeared to saturate the heart lipoprotein lipase. The remnants were slightly smaller in size than the VLDL, and included particles which appeared to be partially emptied. In addition to remnants of d < 1.019 g/ml, iodinated lipoproteins derived from rat and human VLDL were recovered at d 1.019-1.063 and 1.063-1.21 g/ml. The former contained largely cholesterol and cholesteryl esters, while phospholipids were the dominant lipid in the latter. An average of 40% of the ‘2SI-labelled apoprotein lost from the VLDL was associated with the perfused hearts. Very little d 1.019-1.063 g/ml lipoprotein was produced from low (physiological) concentrations of rat VLDL, most of the lipoprotein being removed by the heart. However, lipoproteins of density 1.019-1.063 g/ml were formed from human VLDL at all concentrations in the perfusate, as well as from higher concentrations of the rat VLDL. Agarose gel filtration of lipoproteins following heart perfusion with human VLDL revealed large aggregates containing particles which resemble low density lipoproteins (LDL) in electron microscopic appearance and apoprotein composition, since they contain largely apo-B. These data suggest that at normal concentrations rat VLDL are almost completely catabolised and taken up by the heart without the formation of LDL, while LDL is produced from human VLDL at all concentrations. Abbreviations: VLDL. very low density lipoproteins (d < 1.006 g/ml); IDL. intermediate density lipoproteins (d 1.006--1.019 g/ml); LDL. low density lipoproteins (d 1.019-1.063 s/ml): HDL, high density lipoproteins (d 1.063-1.21 g/ml) [6.321: SDS, sodium dodecyl sulphate.
162
Introduction The catabolism of the triacylglycerol-rich serum lipoproteins, chylomicrons and VLDL, appears to follow a common pathway [l-3] requiring at least two stages. The first involves the hydrolysis of the triacylglycerol by lipoprotein lipase, leading to the formation of remnants relatively deficient in triacylglycerol. Chylomicron remnants remain in the density range Sf > 400 and are removed from circulation by the liver [ 4,5]. The fate of the VLDL remnants is less clear. They probably exist at the denser end of the VLDL range (S, 20-40) in both man [6] and rats [ 71. It has been suggested that with more complete hydrolysis each rat VLDL particle gives rise to a single remnant [8]. While it is possible that rat VLDL remnants may be cleared rapidly by the liver [9], there is considerable evidence that human VLDL is converted to LDL and HDL in the circulation [6,10,11], possibly after passing through an IDL stage [6]. However, studies of the production of remnants and their subsequent fate are complicated by the difficulty of isolating individual steps of lipoprotein catabolism. In intact animals the liver may remove or modify the remnants. Even when this is limited by the use of hepatectomized [5] or supradiaphragmatic [ 121 rats or post-heparin plasma [8], difficulties are presented by the presence of plasma which contains lecithin-cholesterol acyltransferase and various substances which may act as acceptors for lipid or protein. In addition apo-C and possibly the arginine-rich protein readily exchange between HDL and VLDL [ 131, making it difficult to follow the fate of radioactively labelled apopro’teins. The use of purified lipoprotein lipase also has disadvantages since the system is unphysiological, the enzyme unstable [ 3,141, and albumin-bound free fatty acids accumulate in the incubation medium. In order to overcome these difficulties our laboratory devised a two stage in vitro model of chylomicron metabolism [15]. Remnants were prepared by perfusion through beating rat hearts, isolated, and perfused through rat livers. We set out to determine if this model is applicable to VLDL by first investigating the heart perfusion system. The kinetic characteristics of the hydrolysis of triacylglycerol of rat VLDL by the heart lipoprotein lipase have been extensively described [2,16,17], but little attention has been paid to the lipoprotein products. This communication describes and compares some of the products formed during the catabolism of human and rat VLDL. A preliminary report has been presented [ 181. Methods Hearts, obtained from male hooded rats, weighing 200 g and fasted overnight, were perfused by the method of Bleehen and Fisher [ 191 in a modification of the apparatus of Miller et al. [20] with a Silastic tubing oxygenator [21]. All glassware was siliconized prior to use. 60 ml of perfusate, consisting of KrebsRinger bicarbonate buffer, pH 7.4, containing 0.22 mM Ca’+ and 60 mg glucose was used with a gas phase of 95% O2 and 5% CO*. Rat hearts were perfused to remove residual blood prior to their installation in the perfusion apparatus. The ‘2SI-labelled VLDL was added to the perfusate in the apparatus. The hearts were changed after every 40 min of perfusion and beat steadily (>175 beats/ min). Albumin was omitted from the gerfusate, producing a more rapid and
163
regular heart beat. The omission of albumin did not reduce the extent of hydrolysis of the triacylglycerol from the VLDL,. the free fatty acids being taken up by the heart, rather than accumulating in the perfusate. After perfusion the hearts were flushed with 30 ml of Krebs-Ringer solution under pressure to remove any lipoproteins trapped in the vascular system. Controls consisted of recycling the perfusate containing labelled VLDL through the apparatus for 120 min. Isolation and preparation of lipoproteins. Plasma was obtained from 400-g hooded male rats by aortic puncture or from pooled freshly drawn titrated human blood. Chylomicrons were removed by cen~fugation for 1 h at 33 000 rev./min. VLDL were isolated by a modification [ 131 of the method of Have1 et al. 1221 at d < 1.006 g/ml. To isolate remnants after perfusion the per&ate was adjusted to 1.019 g/ml by the addition of NaCl (final concentration 0.46 M) and EDTA (final concentration 0.01%) and centrifuged at 10°C. Iodination of VLDL was carried out by a modification 1231 of the method of McFarlane [24], Approx. 80 and 60% of the “‘1 was bound to the protein moieties of the human and rat VLDL, respectively. To remove free iodine the labelled VLDL was filtered through a Sephadex G-25 column, dialysed overnight and re-isolated by ultracentrifugation. Higher density lipoproteins were isolated from the perfusate by centrifugation after the addition of solid NaCl (d < 1,963 g/ml) and solid NaBr (d < 1.21 g/ml). The lipoproteins in the perfusate were also separated by agarose gel filtration using a 2.2 X 60 cm column containing Bio-Gel A-150m (W-100 mesh) in the bottom 25 cm, Bio-Gel A-50m (10~200 mesh) in the next 25 cm and 8 cm of BioGel P-10 polyacrylamide (50-100 mesh) at the top. The latter retained free iodine as the lipoproteins entered the agarose column. The observed void volume of the agarose column was 82 ml and the bed volume was 200-210 ml. 4-ml aliquots of the perfusate were loaded on the column and eluted with a 0.18 M Tris buffer, pH 8.5, containing 0.02% EDTA. The hydrostatic pressure was 13 cm. 2-ml fractions of the eluate were collected and their radioactivity monitored. In several experiments the lipoproteins in each fraction were precipitated and delipidated and the protein moiety counted. This did not significantly change the shape of the elution pattern or the location of the peaks. Ana~~ticaf procedures. Radioactivity bound to un~p~ted apoproteins was determined by applying aliquots of lipoprotein solutions to 2-cm filter paper discs. The lipoproteins were air dried, precipitated with 10% trichloroacetic acid, delipidated overnight in ethanol/ether (3 : 1, v/v) at -10°C and counted [25]. To analyse the apoproteins by disc gel electrophoresis, aliquots of the lipoproteins were first delipidated in at least 50 volumes of ethanol/ether (3 : 1, v/v) at -10°C and the apoproteins solubilized in 0.04 M Tris, 0.05 M glycine buffer, pH 8.9, containing 7 M urea and 1% SDS. The apoproteins were then separated by electrophoresis in 7 M urea on 10% polyacrylamide gel with a 3% stacking gel “as previously described f26 J. SDS was not present in the polyacrylamide gel or the running buffer. The gels were stained with amid0 black and the mdioactivity of the protein bands and in~~ening areas counted in a gamma spectrometer. The ‘251-labelled protein in the cardiac tissue was estimated by homogenizing all hearts used in a single perfusion in 10% trichloro-
164
acetic acid. The precipitated tissue was washed three times in 10% trichloroacetic acid and delipidated three times with chloroform/methanol (2 : 1, v/v). The residue was then dried, and the radioactivity of a weighed aliquot determined. A second ahquot was dissolved in NaOH. The dissolved protein was applied to filter paper, dried and counted. These two procedures gave similar results. Protein was determined by the method of Lowry et al. [27]. Lipids were extracted for analysis by the method of Folch et al. [28], and triacylglycerol was determined by the method of van Handel [ 291. Lipid classes were separated by glass paper chromatography and visualized by sulphuric acid char using a modification of the method of Pocock et al. [30]. The chromatograms were developed for 14 cm in isooctane (2,2,4_trimethylpent~e)/benzene/ glacial acetic acid/acetone (100 : 30 : 0.10 : 0.60, v/v), air dried and again developed for 16 cm in 100% isooctane, with the result that the cholesteryl esters remained behind the second solvent front. The lipoproteins were visualized by electron microscopy following negative staining with phosphotungstic acid at pH 7.2 [31]. Materials. Rats were obtained from Canadian Breeding Farms, St. Constant, Quebec, Nai2’I from Charles E. Frosst and Co., Montreal, Que., agarose and polyacrylamide for gel filtration from Bio-Rad Laboratories, Toronto, Ont. Acrylamide and bis-acrylamide were purchased from Eastman Kodak Co., Rochester, N.Y. and recrystallized from chloroform. Urea, purchased from Fisher Scientific Co., Montreal, Que., was recrystallized to remove ~~b~ylating impurities before use. Silastic tubing was supplied by DowCorning Co., Midland Mich. U.S.A. Glass fibre paper (ITLC-SG) was obtained from Gelman Instrument Co., Ann Arbor, Mich., standard cholesteryl palmitate and monopalmitoylglycerol from Sigma Chemical Co., St. Louis, MO., tripalmitoylglycerol and dipalmitoylglycerol from the Hormel Institute, Austin, Minn., cholesterol from ICN Biochemicals, Cleveland, Ohio and phosphatidylPa. Mixed free fatty acids and triacylcholine from Supelco Inc., Bellefonte, glycerols were prepared from corn oil. Pesticide grade solvents for glass paper chromatography were purchased from Fisher Scientific Co., Montreal, Que. Results The removaf of ~iacylglycerol and protein from the ‘Z51-labelled human VLDL during heart perfusion, shown in Fig. 1, continued for approx. 90 min, after which there was little activity in spite of the insertion of a fresh heart into the perfusion system after 80 min. Nevertheless, to insure maximal hydrolysis by lipoprotein lipase subsequent experiments were carried out for 120 min. The loss of triacylglycerol was significantly greater than that of protein, suggesting the formation of a remnant relatively poor in triacylglycerol. In early experiments, the VLDL and the IDL were isolated separately after perfusion. However, no accumulation of triacylglycerol or labelled lipoproteins was found in the IDL. IDL, when present in the ‘251-labelled VLDL at the start of the perfusion (usually 1.21 g/ml infranatant and about 40-45% in the heart. In order to rule out the possibility that the uptake of apoproteins by the heart is simply due to ultr~~tration and trapping of VLDL, hearts were perfused for 3 min with 200 pg heparin dissolved in 20 ml of non-recycling perfusate, releasing the lipoprotein lipase [ 16 3, These hearts were then perfused for 90 min with perfusate containing human 12SI-labelled VLDL. The resulting uptake of the labelled apoprotein was negligible ( 1.21 g/ml
21 + l(13) 11 + l(12) 5 + 2 (11)
21 ?r 7 (7) 15 k 4 (7) 8 r 4 (7)
Heart
44 f 5
42 k 12 (4)
-
(7)
protein of d 1.019-1.063 g/ml was produced from rat VLDL when the initial perfusate concentration was low. As the latter rose, the recovery of ‘*‘Ilabelled apoprotein in the d 1.019-1.063 g/ml lipoproteins equalled or exceeded that derived from human VLDL.
0
.
0
0.
0
‘0
4-I 0
.
.
-‘-i;*” . no
0.2
0.6 0.4 T R I G L Y C E R I D E (r&ml
.
l
0.6 pcrfurote
)
I.0
Fig. 5. Recovery of apoproteins in the d 1.019-1.063 g/ml fraction following perfusion with human (m) and rat (0) 1*5I-labelled VLDL. Each point represents a separate pool of VLDL. The line representing the was drawn following analysis by least squares, while the curve reprehuman VLDL experiments ( -) senting those with rat VLDL (- - - - - -) is a logarithmic curve fit. Fig. 6. Glass paper chromatographic separation of the lipid components of human VLDL and their products following heart perfusion. STD. mixture of lipid standards. 1. VLDL prior to perfusion: 24. products after perfusion; 2. d < 1.019 g/ml lipoproteins; 3. d 1.019-1.063 g/ml lipoproteins; 4, d 1.0631.21 g/ml lipoproteins. The quantity of lipid spotted on the chromatogram is equal to the lipid extracted if the lipoproteins had been isolated from the following volumes of Perfusate: 1, 0.03 ml; 2, 0.03 ml; 3, 0.07 ml; 4.0.06 ml.
169
The lipid composition of human VLDL and its products isolated are illustrated in Fig. 6. Following the heart perfusion there is a noticeable decrease in all lipids with the possible exception of the cholesteryl esters. A trace of free fatty acids is also discernible. The d 1.019--1.063 g/ml fraction has virtually no triacylglycerol with cholesterol, cholesteryl esters and phospholipid as the dominant lipids. Small quantities of monoacylglycerol and free fatty acids are also seen. The d 1.063-1.21 g/ml fraction contains largely phospholipids with small amounts of monoacylglycerols, cholesteryl esters and triacylglycerol. The lipid compositions of the d 1.019-1.063 and 1.063-1.21 g/ml fractions correspond to those normally associated with human LDL and HDL, respectively 1321.
Changes in the distribution of the apoproteins of VLDL following perfusion were studied by polyacrylamide gel electrophoresis. Typical gels of rat and human VLDL are shown in Fig. 7. The gels are divided into four areas whose major apoproteins correspond to apo-B, the a&nine-rich protein, apo-D or apo-A, and apo-C-II and C-III. The locations of the major apoproteins (except for apo-D and apo-A) were identified by comparison with the mobility of purified apoproteins. The proportions of the labelled apoproteins before and after perfusion are shown in Table II. The d < 1.019 g/ml remnants are richer in apo-B and poorer in apo-C. These effects are more obvious in rat VLDL
-1
OOV
loo TO REMOVED
Fig. 7. Typical polyacrylamide gel electrophoreds patterns into the zones dedgnated l-4 for assay of radioactivity.
of human
t?W
and rat VLDL. The gals were cut
Fig. 8. Relationship between the loss of triacylglycerol and the decrease in the proportion of ape-C-II and ape-C-III among the apoprotains of the d < 1.019 g/ml remnants isolated after perfusion of human (a) and rat (0) VLDL through rat hearta. The correlations coefficient (r) was calculated using all points from both rat and human VLDL. The triacylglycerol concentration ranged from 0.1 to 1.0 me/ml perfusate.
170 TABLE II PERCENT DISTRIBUTION OF l25-I-LABELLED FOLLOWING HEART PERFUSION
APOPROTEINS
IN VLDL AND THEIR PRODUCTS
Initial triacylglycerol concentrations ranged from 0.1 to 1.0 mg/ml perfusate. Each figure is the mean t S.E. Zone (cf. Fig. 7)
HUIllaIl
Rat
VLDL
1 2 3 4 II
d 1.019-1.063
Intact
Remnant
47 r 5 Sk2 521 38 r 3
51 +4 10 * 2 622 29 + 3
7
7
59 ?r3 11 t 2 7+2 20 + 3 4
g/ml
d 1.063-1.21
38 10 10 33 4
+ t + i
4 4 3 1
g/ml
VLDL Intact
Remnant
45 i 6 8+2 10 ? 3 31 + 1
66 + 5 8r2 9+3 St4
4
4
-
remnants. In some experiments tetramethylurea was used to delipidate and dissolve an aliquot of the lipoproteins [33]. The relative proportions of the precipitated apo-B and soluble apo-C following electrophoresis were similar to those noted when ethanol/ether delipidation and solubilization in urea was utilized. Because of some variation in the extent of iodination of apo-B and apo-C between different VLDL preparations, changes in the proportion of these apoproteins can best be seen by analysis of the paired differences in each experiment. It was found that in the human remnants the proportion of apo-B increased by 4.0 k 2% (P < 0.05, it = 7) and that of the apo-C decreased by 8.6 + 1.2% (P < 0.001). Greater changes were observed in rat VLDL remnants, where the proportion of apo-B increased by 20.5 5 6.5% (P < 0.05, n = 4) while that of apo-C decreased by 22.8 + 4.8% (P< 0.01). A closecorrelation (r = 0.75, P < 0.001) between the loss of triacylglycerol from VLDL and the decrease in apo-C during formation of remnants can be seen when the data from the individual experiments are plotted in Fig. 8. The data in Table II also indicate that the products of human VLDL in the d 1.019-1.063 and 1.063-1.21 g/ml ranges do not have the apoprotein composition commonly associated with serum LDL and HDL, respectively. Thus, in spite of the lower percent of apo-C in the d 1.019-1.063 g/ml fraction, the proportion of non-B apoproteins far exceeds that found in circulating LDL [34,35]. The d 1.063-1.21 g/ml fraction contains a larger amount of labelled apoprotein at the top of the gel (region 1) than is usually found when the apoproteins of circulating HDL are analyzed. Although apo-A tends to polymerize after centrifugation in dilute protein solution [ 361, apo-A is found only in minute quantities in the original VLDL. Thus, the lipoproteins isolated at d 1.019-1.063 and 1.063-1.21 g/ml may represent mixtures of product8 of VLDL catabolism. About 20% of the material in the d 1.019-1.063 g/ml fraction was not absorbed by a concanavalin A column and therefore represents lipoprotein deficient in apo-B [37]. This was confirmed by polyacrylamide gel electrophoresis of the non-absorbed 1251-labelledlipoprotein which contained 75% of the radioactivity in apo-C and relatively very little at the top of the gel.
171
The possibility that some of the higher density products of VLDL catabolism are mixtures of different lipoproteins having similar hydrated densities led to an attempt to separate them by agarose gel filtration without prior centrifugation. The distribution of the lipoproteins in typical control and experimental perfusates following perfusion with human VLDL are shown in Fig. 9. It, will be seen that except for a very small peak at the void volume, usually smaller than in the preparation depicted in Fig. 9, the control VLDL elutes as a single peak. Following perfusion three lipoprotein peaks were discernible. The first was found in the void volume. The major peak corresponds to the control VLDL, although in some experiments a slight shift to the right was observed. A new peak appeared as the bed volume was approached. Material from each peak was pooled, dialysed, concentrated by lyophilization, delipidated and analyzed by 1polyacrylamide gel electrophoresis: The ratios of the radioactivity in regions i : 4 (apo-B : apo-C) of the gels (cf. Fig. 7) for each peak is shown in Fig. 9. It will be noted that the perfusion did not produce a significant change in the ratios of peak II. Peak I contained lipoproteins predominantly labelled in ap&B, resembling circulating LDL. On the other hand, the lipoproteins in peak 01 contained very little if any labelled apoB relative to 1251-labelled apo-C, ai pattern consistent with HDL. Electron microscopy (Fig. 10) of the material eluted in peak I showed large aggregates containing many particles resemblin LDL in size as well as some VLDL. It thus appears likely that during hy %rolysis of the triacylglycerol of human VLDL by heart lipoprotein lipase duting perfusion, some LDL is formed.
100
140
180
220
200
ELUENT Iml,
Fig. 9. Agarose gel column filtration of lipopro human VLDL in the absence (control - - - - - -) radioactfvitv in ape-B : a~04 regions of the represent the mean of three experiments f S.E.
ins in perfusatcs following perfusion of 12SX-labelled of beating rat hearts. The ratio ,oi lipoproteins in the three peaks
Fig. 10. Negatively stained lipoprotein from the void volume peak (If of the agarose gel filtration (cf. Fif. 9). Magnification 93 000 X. Note the presence of numerous particles of LDL size as well as some larger VLDL particles. AU particles appear to be emeshed in a large aggregate.
Discussion The use of perfused hearts for the study of the first stage of the catabolism of VLDL has produced a number of interesting observations. The heart is a source of immobilized lipoprotein lipase which has similar characteristics to those of heparin-solubilized heart lipoprotein lipase [ 161. However, the beating heart is a more physiologic~ preparation since some of the lipid and protein products of lipoprotein lipase activity are taken up by the tissue. In general there appears to be little difference in the extent of removal of the triacylglycerol between human and rat VLDL. The relatively constant amount of triacylglycerol removed from rat VLDL over a wide range of perfusate concentrations suggests that the enzyme system is close to saturation even at the
lowest concentrations used in these experiments. This would be predicted from the data of Fielding and Higgins [16]. Although the greater variation in the extent of breakdown of human VLDL (cf. Fig. 2) may be due to individual differences among blood donors, it is also possible that the kinetics of catabolism of human and rat VLDL differ. There are differences between the products of rat and human VLDL recovered in the d 1.019-1.063 g/ml fraction. At low levels of rat VLDL in the perfusate (corresponding to the normal rat serum concentrations) very little LDL is formed, most of the lipoprotein being taken up by the heart. This correlates with the low concentration of LDL found in rat serum. At higher concentrations lipoproteins of d 1.019-1.063 g/ml accumulate. In contrast, a constant proportion of the catabolized human VLDL is consistently found in the d 1.019-1.063 g/ml range regardless of the initial perfusate concentrations. These lipoproteins appear to be a mixture of true LDL and triacylglycerol-depleted remnants which retain some of the non-B apoproteins. The production of the LDL, even at low concentrations of VLDL, may correlate with the relatively high concentration of LDL in human circulation. It is possible, however, that a species-specific receptor of limited capacity leads to the removal of rat LDL at low perfusate VLDL levels. Hydrolysis of triacylglycerol-rich lipoproteins is associated with a loss of apo-C [7,15,38]. The present data show a significant correlation between the loss of triacylglycerol and apo-C from VLDL during catabolism. A similar correlation has recently been reported by Glangeaud et al. [ 391 using purified milk LDL and VLDL labelled by absorption of exogenous isolated apo-C. In view of the relatively small (and in some experiments negligible) recovery of the apoprotein in the d > 1.21 g/ml fraction of the perfusate, it is likely that much of the native apo-C lost from VLDL, but remaining in the perfusate, is in the form of a lipoprotein containing considerable phospholipid. Although the major objective of the present work was to characterize the products of VLDL catabolism, the incidental finding that apoproteins of VLDL are removed from the perfusate by the hearts is of interest. While the present data does not completely rule out the possibility that some VLDL is trapped by the heart, it seems unlikely that this is significant. The uptake of VLDL is markedly diminished if the heart has been treated with heparin to remove lipoprotein lipase. In addition, chylomicron apoprotein is not taken up by the perfused rat heart [15] in spite of the larger size of these particles. The utilization of triacylglycerol-fatty acid from chylomicrons (Dolphin, P.J., unpublished results) and VLDL [40] for cardiac acylglycerol synthesis and oxidation also suggests that the lipoproteins are not simply filtered or trapped during passage through the heart. It is not presently known whether the iodinated VLDL protein removed from the perfusate is simply bound to the endothelium or is internalized into either the endothelial or myocardial cells. Bierman et al. [38] have noted that cultured cells take up VLDL and their remnants, but these are catabolized very slowly. Similarly, the recovery of 8085% of the ‘*‘I-labelled apoprotein removed from VLDL as protein, of which approx. 50% is in perfused hearts (cf. Table I) suggests a relatively slow catabolic rate of protein in the present experiments. The present results raise some questions as to whether the in vitro model of
174
successive heart and liver perfusions devised for chylomicrons can be applied to VLDL catabolism. A considerable amount of the apoprotein and lipid is taken up directly by the heart. There is evidence in the swine that the liver plays a minor role in the catabolism of LDL [41]. In the rat, radioactivity from injected labelled VLDL appears primarily in the liver [42]. However, labelling of the lipid as well as the protein moiety of rat VLDL makes it difficult to evaluate fully the fate of remnant apoproteins. Preliminary experiments carried out in our laboratory show that there is considerably less uptake of human or rat VLDL remnants by rat liver than the corresponding chylomicron remnants. Thus there appears to be a more complex disposal mechanism for the products of VLDL catabolism than for chylomicron remn~~. The possibility must also be considered that the remnant formed by the action of heart lipoprotein lipase are incomplete, requiring at least a second step, probably involving lecithin-cholesterol acyltransferase. Schumaker and Adams [43] have suggested that lecithin-cholesterol acyltransferase may play a role in modifying the lipoprotein surface coat to accomodate the continuous removal of triacylglycerol. The observation that LDL derived from human VLDL tends to occur in an aggregated state may be a reflection of an inability to remodel the membrane. Norum et al. [44] observed a tendency of some LDL isolated from lecithin-cholesterol acyltransferasedeficient patients to aggregate. The large partially emptied remnants noted following perfusion resemble the very large particles found in the LDL of lecithin-cholesterol acyltr~sfer~deficient subjects [ 451,. Further studies are therefore being carried out on the role of this enzyme, HDL, and other factors which may play a role in the catabolism of VLDL. The results which have been obtained with the perfused rat heart suggest that there are some small but significant differences between the catabolism of rat and human VLDL which may be related to the accumulation of LDL in man. Acknowledgements This work was supported by grants from the Medical Research Council of Canada and the Quebec Heart Foundation. We wish to thank the Montreal Transfusion Centre of the Canadian Red Cross Blood T~sfusion Service and Mme. M. Caisse for the supply of human blood. References 1 2 3 4 6 6 7 8 9 10
Brunzell, J.D., Hazzard, W.R., Porte, Jr.. D. and Bierman, E.L. (1973) J. Clin. Invest. 62,1678-1586 Higgins, J.M. and Fielding, C.J. (1975) Biochemistry 14, 2288-2293 Dolphin, P.J. and Rubinstein, D. (1974) Biochem. Biophys. Res. Commun. 57. 808-814 Nestel. P.J.. Havel, R.J. and Bezman, A. (1963) J. Clin. Invest. 42,1313-1321 Redgrave. T.G. (1970) J. Clin. Invest. 49, 466-471 Eisanberg, S., Bilheimer. D.W.. Levy, R.1. and Lindgren. F.T. (1973) Biochim. Biophys. Acta 326, 361-377 Mjgs, O.D., Faergeman, 0.. Hamilton, R.L. and Havel, R.J. (1975) J. Ciin. Invest. 56,603-615 Eisenberg. S. and Rachm6ewitz. D. (1975) J. Lipid Res. 16, 341-351 Faergeman, 0. and Have& R.J. (1975) J. Clin. Invest. 65.1216-1218 GitIin, D., Cornwell, D.G.. Nakasato, D., Oncley, J.L., Hughes, Jr.. W.L. and Janeway, C.A. (1968) J. Clin. Invest. 37.172-184
175 11 12 13 14 15
Wilson. D.E. and Lees, R.S. (1972) J. Chn. Invest. 51, 1051-1057 Bezmsn-Tarcher. A. and Robinson, D.S. (1965) Proc. R. Sot. Lond. Ser. B. 162,406410 Rubenstein. B. and RubInstein. D. (1972) J. Lipid Res. 13.317-324 Twu. J-S., Garfinkel. A.S. and Schots. M.C. (1976) Atherosclerosis 24.119-128 Noel. S-P., Dolphin, P.J. and Rubinstein. D. (1975) Biochem. Biophys. Res. Commun. 163, 764-772
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
Fielding, C.J. and Higgins, J.M. (1974) Biochemistry 13.4324-4330 Fielding. C.J. (1976) Biochemistry 15.879-884 Dory, L. and Rubinstein, D. (1976) Clin. Res. 26, 680A Bleehen. N.M. and Fisher, R.B. (1954) J. Physiol. Lond. 123, 260-276 Miller, L.L.. Bly. C.G.. Watson, M.L. and Bale. W.F. (1951) J. Exp. Med. 94, 431-453 Hamilton, R.L.. Berry, M.N., Williams. M.C. and Seveiinghaus. E.M. (1974) J. Lipid Res. 15. 182-186 Have]. R.J., Eder. H.A. and Bragdon, J.H. (1955) J. Clin. Invest. 34, 1345-1353 Fidge. N.H. and Pouhs, P. (1974) CIin. Chim. Acta 52.15-26 McFarlane. A.S. (1958) Nature 182. 53 Mans, R.J. and NoveIIi. G.D. (1961) Arch. Biochem. Biophys. 94.48-53 Nestruck. A.C. and Rubinstein, D. (1976) Can. J. Biochem. 54,617-628 Lowry, O.H.. Rosebrough. N.J.. Farr, A.L. and RandaB. R.J. (1951) J. Biol. Chem. 193.265-275 Folch. J.. Lees, M. and SloaneStanley. G.H. (1957) J. Biol. Chem. 226. 497-509 Van Handel, E. (1961) CIin. Chem. 7.249-251 Pocock. D.M.-E., RafaI. S. and Vost. A. (1972) J. Chromatogr. Sci. 10, 72-76 Blanchette-Mackie, E.J. and Scow. R.O. (1973) J. Ceil Biol. 58. 689-708 Eisenberg. S. and Levy, R.I. (1975) Adv. Lipid Res. 13, l-89 Kane, J.P. (1973) Anal. Biochem. 53,350-364 Lee, D.M. and Alaupovic. P. (1970) Biochemistry 9.2244-2252 Rubenstein, B. and Steiner, G. (1975) Can. J. Biochem. 53.128-134 Noel, S-P. and Rubinstein. D. (1974) 3. Lipid Res. 15. 301-308 McConathy, WJ. and Alaupovic. P. (1974) FEBS Lett. 41.174-178 Bierman. E.L., Eisenberg, S., Stein, 0. and Stein, Y. (1973) Biochim. Biophys. Acta 329.183-169 Glangeaud. M.C.. Eisenberg. S. and Olivecrona. T. (1977) Biochim. Biophys. Acta 486, 23-35 Delcher. ILK., Fried, M. and Shipp. J.C. (1965) Biochim. Biophys. Acta 106.10-18 Sniderman, A.D., Carew. T.E.. Chandler. J.G. and Steinberg, D. (1974) Science 183.526-528 Risenberg, S. and RachmiIewitz. D. (1974) Biochhn. Biophys. Acta 326. 378-390 Schumaker. V.N. and Adams, G.H. (1970) J. Theor. Biol. 26, 89-91 Norum. K.R., Glomset, J.A.. Nichols. A.V. and Forte, T. (1971) J. CIin. Invest. 50.1131-1140 Forte. T.. Norum. K.R., Glomset. J.A. and Nichols, A.V. (1971) J. CIin. Invest. 50.1141-1148
Biochimica et Biophysics Acta, 528 (1978) 161-175 @ Elsevier/North-Holland Biomedical Press
BBA 57119
THE CATABOLISM OF HUMAN AND RAT VERY LIPOPROTEINS BY PERFUSED RAT HEARTS
LADISLAV
DORY, DOROTHY
Department
of Biochemistry,
LOW DENSITY
POCOCK and DAVID RUBINSTEIN
McGill University, Montreal, Quebec (Canada)
(Received May 25th, 1977) (Revised manuscript received October 18th, 1977)
Summary The catabolism of human and rat 1251-labelled very low density lipoproteins (VLDL) was compared by perfusing the lipoproteins through beating rat hearts. Triacylglycerol was removed from the VLDL to a greater extent than the protein moiety, leaving remnants containing relatively more apo-B and less apo-C. The change in apo-C content of the remnants correlated with the loss of triacylglycerol. The extent of removal of triacylglycerol from the rat and human VLDL was similar and in most cases appeared to saturate the heart lipoprotein lipase. The remnants were slightly smaller in size than the VLDL, and included particles which appeared to be partially emptied. In addition to remnants of d < 1.019 g/ml, iodinated lipoproteins derived from rat and human VLDL were recovered at d 1.019-1.063 and 1.063-1.21 g/ml. The former contained largely cholesterol and cholesteryl esters, while phospholipids were the dominant lipid in the latter. An average of 40% of the ‘2SI-labelled apoprotein lost from the VLDL was associated with the perfused hearts. Very little d 1.019-1.063 g/ml lipoprotein was produced from low (physiological) concentrations of rat VLDL, most of the lipoprotein being removed by the heart. However, lipoproteins of density 1.019-1.063 g/ml were formed from human VLDL at all concentrations in the perfusate, as well as from higher concentrations of the rat VLDL. Agarose gel filtration of lipoproteins following heart perfusion with human VLDL revealed large aggregates containing particles which resemble low density lipoproteins (LDL) in electron microscopic appearance and apoprotein composition, since they contain largely apo-B. These data suggest that at normal concentrations rat VLDL are almost completely catabolised and taken up by the heart without the formation of LDL, while LDL is produced from human VLDL at all concentrations. Abbreviations: VLDL. very low density lipoproteins (d < 1.006 g/ml); IDL. intermediate density lipoproteins (d 1.006--1.019 g/ml); LDL. low density lipoproteins (d 1.019-1.063 s/ml): HDL, high density lipoproteins (d 1.063-1.21 g/ml) [6.321: SDS, sodium dodecyl sulphate.
162
Introduction The catabolism of the triacylglycerol-rich serum lipoproteins, chylomicrons and VLDL, appears to follow a common pathway [l-3] requiring at least two stages. The first involves the hydrolysis of the triacylglycerol by lipoprotein lipase, leading to the formation of remnants relatively deficient in triacylglycerol. Chylomicron remnants remain in the density range Sf > 400 and are removed from circulation by the liver [ 4,5]. The fate of the VLDL remnants is less clear. They probably exist at the denser end of the VLDL range (S, 20-40) in both man [6] and rats [ 71. It has been suggested that with more complete hydrolysis each rat VLDL particle gives rise to a single remnant [8]. While it is possible that rat VLDL remnants may be cleared rapidly by the liver [9], there is considerable evidence that human VLDL is converted to LDL and HDL in the circulation [6,10,11], possibly after passing through an IDL stage [6]. However, studies of the production of remnants and their subsequent fate are complicated by the difficulty of isolating individual steps of lipoprotein catabolism. In intact animals the liver may remove or modify the remnants. Even when this is limited by the use of hepatectomized [5] or supradiaphragmatic [ 121 rats or post-heparin plasma [8], difficulties are presented by the presence of plasma which contains lecithin-cholesterol acyltransferase and various substances which may act as acceptors for lipid or protein. In addition apo-C and possibly the arginine-rich protein readily exchange between HDL and VLDL [ 131, making it difficult to follow the fate of radioactively labelled apopro’teins. The use of purified lipoprotein lipase also has disadvantages since the system is unphysiological, the enzyme unstable [ 3,141, and albumin-bound free fatty acids accumulate in the incubation medium. In order to overcome these difficulties our laboratory devised a two stage in vitro model of chylomicron metabolism [15]. Remnants were prepared by perfusion through beating rat hearts, isolated, and perfused through rat livers. We set out to determine if this model is applicable to VLDL by first investigating the heart perfusion system. The kinetic characteristics of the hydrolysis of triacylglycerol of rat VLDL by the heart lipoprotein lipase have been extensively described [2,16,17], but little attention has been paid to the lipoprotein products. This communication describes and compares some of the products formed during the catabolism of human and rat VLDL. A preliminary report has been presented [ 181. Methods Hearts, obtained from male hooded rats, weighing 200 g and fasted overnight, were perfused by the method of Bleehen and Fisher [ 191 in a modification of the apparatus of Miller et al. [20] with a Silastic tubing oxygenator [21]. All glassware was siliconized prior to use. 60 ml of perfusate, consisting of KrebsRinger bicarbonate buffer, pH 7.4, containing 0.22 mM Ca’+ and 60 mg glucose was used with a gas phase of 95% O2 and 5% CO*. Rat hearts were perfused to remove residual blood prior to their installation in the perfusion apparatus. The ‘2SI-labelled VLDL was added to the perfusate in the apparatus. The hearts were changed after every 40 min of perfusion and beat steadily (>175 beats/ min). Albumin was omitted from the gerfusate, producing a more rapid and
163
regular heart beat. The omission of albumin did not reduce the extent of hydrolysis of the triacylglycerol from the VLDL,. the free fatty acids being taken up by the heart, rather than accumulating in the perfusate. After perfusion the hearts were flushed with 30 ml of Krebs-Ringer solution under pressure to remove any lipoproteins trapped in the vascular system. Controls consisted of recycling the perfusate containing labelled VLDL through the apparatus for 120 min. Isolation and preparation of lipoproteins. Plasma was obtained from 400-g hooded male rats by aortic puncture or from pooled freshly drawn titrated human blood. Chylomicrons were removed by cen~fugation for 1 h at 33 000 rev./min. VLDL were isolated by a modification [ 131 of the method of Have1 et al. 1221 at d < 1.006 g/ml. To isolate remnants after perfusion the per&ate was adjusted to 1.019 g/ml by the addition of NaCl (final concentration 0.46 M) and EDTA (final concentration 0.01%) and centrifuged at 10°C. Iodination of VLDL was carried out by a modification 1231 of the method of McFarlane [24], Approx. 80 and 60% of the “‘1 was bound to the protein moieties of the human and rat VLDL, respectively. To remove free iodine the labelled VLDL was filtered through a Sephadex G-25 column, dialysed overnight and re-isolated by ultracentrifugation. Higher density lipoproteins were isolated from the perfusate by centrifugation after the addition of solid NaCl (d < 1,963 g/ml) and solid NaBr (d < 1.21 g/ml). The lipoproteins in the perfusate were also separated by agarose gel filtration using a 2.2 X 60 cm column containing Bio-Gel A-150m (W-100 mesh) in the bottom 25 cm, Bio-Gel A-50m (10~200 mesh) in the next 25 cm and 8 cm of BioGel P-10 polyacrylamide (50-100 mesh) at the top. The latter retained free iodine as the lipoproteins entered the agarose column. The observed void volume of the agarose column was 82 ml and the bed volume was 200-210 ml. 4-ml aliquots of the perfusate were loaded on the column and eluted with a 0.18 M Tris buffer, pH 8.5, containing 0.02% EDTA. The hydrostatic pressure was 13 cm. 2-ml fractions of the eluate were collected and their radioactivity monitored. In several experiments the lipoproteins in each fraction were precipitated and delipidated and the protein moiety counted. This did not significantly change the shape of the elution pattern or the location of the peaks. Ana~~ticaf procedures. Radioactivity bound to un~p~ted apoproteins was determined by applying aliquots of lipoprotein solutions to 2-cm filter paper discs. The lipoproteins were air dried, precipitated with 10% trichloroacetic acid, delipidated overnight in ethanol/ether (3 : 1, v/v) at -10°C and counted [25]. To analyse the apoproteins by disc gel electrophoresis, aliquots of the lipoproteins were first delipidated in at least 50 volumes of ethanol/ether (3 : 1, v/v) at -10°C and the apoproteins solubilized in 0.04 M Tris, 0.05 M glycine buffer, pH 8.9, containing 7 M urea and 1% SDS. The apoproteins were then separated by electrophoresis in 7 M urea on 10% polyacrylamide gel with a 3% stacking gel “as previously described f26 J. SDS was not present in the polyacrylamide gel or the running buffer. The gels were stained with amid0 black and the mdioactivity of the protein bands and in~~ening areas counted in a gamma spectrometer. The ‘251-labelled protein in the cardiac tissue was estimated by homogenizing all hearts used in a single perfusion in 10% trichloro-
164
acetic acid. The precipitated tissue was washed three times in 10% trichloroacetic acid and delipidated three times with chloroform/methanol (2 : 1, v/v). The residue was then dried, and the radioactivity of a weighed aliquot determined. A second ahquot was dissolved in NaOH. The dissolved protein was applied to filter paper, dried and counted. These two procedures gave similar results. Protein was determined by the method of Lowry et al. [27]. Lipids were extracted for analysis by the method of Folch et al. [28], and triacylglycerol was determined by the method of van Handel [ 291. Lipid classes were separated by glass paper chromatography and visualized by sulphuric acid char using a modification of the method of Pocock et al. [30]. The chromatograms were developed for 14 cm in isooctane (2,2,4_trimethylpent~e)/benzene/ glacial acetic acid/acetone (100 : 30 : 0.10 : 0.60, v/v), air dried and again developed for 16 cm in 100% isooctane, with the result that the cholesteryl esters remained behind the second solvent front. The lipoproteins were visualized by electron microscopy following negative staining with phosphotungstic acid at pH 7.2 [31]. Materials. Rats were obtained from Canadian Breeding Farms, St. Constant, Quebec, Nai2’I from Charles E. Frosst and Co., Montreal, Que., agarose and polyacrylamide for gel filtration from Bio-Rad Laboratories, Toronto, Ont. Acrylamide and bis-acrylamide were purchased from Eastman Kodak Co., Rochester, N.Y. and recrystallized from chloroform. Urea, purchased from Fisher Scientific Co., Montreal, Que., was recrystallized to remove ~~b~ylating impurities before use. Silastic tubing was supplied by DowCorning Co., Midland Mich. U.S.A. Glass fibre paper (ITLC-SG) was obtained from Gelman Instrument Co., Ann Arbor, Mich., standard cholesteryl palmitate and monopalmitoylglycerol from Sigma Chemical Co., St. Louis, MO., tripalmitoylglycerol and dipalmitoylglycerol from the Hormel Institute, Austin, Minn., cholesterol from ICN Biochemicals, Cleveland, Ohio and phosphatidylPa. Mixed free fatty acids and triacylcholine from Supelco Inc., Bellefonte, glycerols were prepared from corn oil. Pesticide grade solvents for glass paper chromatography were purchased from Fisher Scientific Co., Montreal, Que. Results The removaf of ~iacylglycerol and protein from the ‘Z51-labelled human VLDL during heart perfusion, shown in Fig. 1, continued for approx. 90 min, after which there was little activity in spite of the insertion of a fresh heart into the perfusion system after 80 min. Nevertheless, to insure maximal hydrolysis by lipoprotein lipase subsequent experiments were carried out for 120 min. The loss of triacylglycerol was significantly greater than that of protein, suggesting the formation of a remnant relatively poor in triacylglycerol. In early experiments, the VLDL and the IDL were isolated separately after perfusion. However, no accumulation of triacylglycerol or labelled lipoproteins was found in the IDL. IDL, when present in the ‘251-labelled VLDL at the start of the perfusion (usually 1.21 g/ml infranatant and about 40-45% in the heart. In order to rule out the possibility that the uptake of apoproteins by the heart is simply due to ultr~~tration and trapping of VLDL, hearts were perfused for 3 min with 200 pg heparin dissolved in 20 ml of non-recycling perfusate, releasing the lipoprotein lipase [ 16 3, These hearts were then perfused for 90 min with perfusate containing human 12SI-labelled VLDL. The resulting uptake of the labelled apoprotein was negligible ( 1.21 g/ml
21 + l(13) 11 + l(12) 5 + 2 (11)
21 ?r 7 (7) 15 k 4 (7) 8 r 4 (7)
Heart
44 f 5
42 k 12 (4)
-
(7)
protein of d 1.019-1.063 g/ml was produced from rat VLDL when the initial perfusate concentration was low. As the latter rose, the recovery of ‘*‘Ilabelled apoprotein in the d 1.019-1.063 g/ml lipoproteins equalled or exceeded that derived from human VLDL.
0
.
0
0.
0
‘0
4-I 0
.
.
-‘-i;*” . no
0.2
0.6 0.4 T R I G L Y C E R I D E (r&ml
.
l
0.6 pcrfurote
)
I.0
Fig. 5. Recovery of apoproteins in the d 1.019-1.063 g/ml fraction following perfusion with human (m) and rat (0) 1*5I-labelled VLDL. Each point represents a separate pool of VLDL. The line representing the was drawn following analysis by least squares, while the curve reprehuman VLDL experiments ( -) senting those with rat VLDL (- - - - - -) is a logarithmic curve fit. Fig. 6. Glass paper chromatographic separation of the lipid components of human VLDL and their products following heart perfusion. STD. mixture of lipid standards. 1. VLDL prior to perfusion: 24. products after perfusion; 2. d < 1.019 g/ml lipoproteins; 3. d 1.019-1.063 g/ml lipoproteins; 4, d 1.0631.21 g/ml lipoproteins. The quantity of lipid spotted on the chromatogram is equal to the lipid extracted if the lipoproteins had been isolated from the following volumes of Perfusate: 1, 0.03 ml; 2, 0.03 ml; 3, 0.07 ml; 4.0.06 ml.
169
The lipid composition of human VLDL and its products isolated are illustrated in Fig. 6. Following the heart perfusion there is a noticeable decrease in all lipids with the possible exception of the cholesteryl esters. A trace of free fatty acids is also discernible. The d 1.019--1.063 g/ml fraction has virtually no triacylglycerol with cholesterol, cholesteryl esters and phospholipid as the dominant lipids. Small quantities of monoacylglycerol and free fatty acids are also seen. The d 1.063-1.21 g/ml fraction contains largely phospholipids with small amounts of monoacylglycerols, cholesteryl esters and triacylglycerol. The lipid compositions of the d 1.019-1.063 and 1.063-1.21 g/ml fractions correspond to those normally associated with human LDL and HDL, respectively 1321.
Changes in the distribution of the apoproteins of VLDL following perfusion were studied by polyacrylamide gel electrophoresis. Typical gels of rat and human VLDL are shown in Fig. 7. The gels are divided into four areas whose major apoproteins correspond to apo-B, the a&nine-rich protein, apo-D or apo-A, and apo-C-II and C-III. The locations of the major apoproteins (except for apo-D and apo-A) were identified by comparison with the mobility of purified apoproteins. The proportions of the labelled apoproteins before and after perfusion are shown in Table II. The d < 1.019 g/ml remnants are richer in apo-B and poorer in apo-C. These effects are more obvious in rat VLDL
-1
OOV
loo TO REMOVED
Fig. 7. Typical polyacrylamide gel electrophoreds patterns into the zones dedgnated l-4 for assay of radioactivity.
of human
t?W
and rat VLDL. The gals were cut
Fig. 8. Relationship between the loss of triacylglycerol and the decrease in the proportion of ape-C-II and ape-C-III among the apoprotains of the d < 1.019 g/ml remnants isolated after perfusion of human (a) and rat (0) VLDL through rat hearta. The correlations coefficient (r) was calculated using all points from both rat and human VLDL. The triacylglycerol concentration ranged from 0.1 to 1.0 me/ml perfusate.
170 TABLE II PERCENT DISTRIBUTION OF l25-I-LABELLED FOLLOWING HEART PERFUSION
APOPROTEINS
IN VLDL AND THEIR PRODUCTS
Initial triacylglycerol concentrations ranged from 0.1 to 1.0 mg/ml perfusate. Each figure is the mean t S.E. Zone (cf. Fig. 7)
HUIllaIl
Rat
VLDL
1 2 3 4 II
d 1.019-1.063
Intact
Remnant
47 r 5 Sk2 521 38 r 3
51 +4 10 * 2 622 29 + 3
7
7
59 ?r3 11 t 2 7+2 20 + 3 4
g/ml
d 1.063-1.21
38 10 10 33 4
+ t + i
4 4 3 1
g/ml
VLDL Intact
Remnant
45 i 6 8+2 10 ? 3 31 + 1
66 + 5 8r2 9+3 St4
4
4
-
remnants. In some experiments tetramethylurea was used to delipidate and dissolve an aliquot of the lipoproteins [33]. The relative proportions of the precipitated apo-B and soluble apo-C following electrophoresis were similar to those noted when ethanol/ether delipidation and solubilization in urea was utilized. Because of some variation in the extent of iodination of apo-B and apo-C between different VLDL preparations, changes in the proportion of these apoproteins can best be seen by analysis of the paired differences in each experiment. It was found that in the human remnants the proportion of apo-B increased by 4.0 k 2% (P < 0.05, it = 7) and that of the apo-C decreased by 8.6 + 1.2% (P < 0.001). Greater changes were observed in rat VLDL remnants, where the proportion of apo-B increased by 20.5 5 6.5% (P < 0.05, n = 4) while that of apo-C decreased by 22.8 + 4.8% (P< 0.01). A closecorrelation (r = 0.75, P < 0.001) between the loss of triacylglycerol from VLDL and the decrease in apo-C during formation of remnants can be seen when the data from the individual experiments are plotted in Fig. 8. The data in Table II also indicate that the products of human VLDL in the d 1.019-1.063 and 1.063-1.21 g/ml ranges do not have the apoprotein composition commonly associated with serum LDL and HDL, respectively. Thus, in spite of the lower percent of apo-C in the d 1.019-1.063 g/ml fraction, the proportion of non-B apoproteins far exceeds that found in circulating LDL [34,35]. The d 1.063-1.21 g/ml fraction contains a larger amount of labelled apoprotein at the top of the gel (region 1) than is usually found when the apoproteins of circulating HDL are analyzed. Although apo-A tends to polymerize after centrifugation in dilute protein solution [ 361, apo-A is found only in minute quantities in the original VLDL. Thus, the lipoproteins isolated at d 1.019-1.063 and 1.063-1.21 g/ml may represent mixtures of product8 of VLDL catabolism. About 20% of the material in the d 1.019-1.063 g/ml fraction was not absorbed by a concanavalin A column and therefore represents lipoprotein deficient in apo-B [37]. This was confirmed by polyacrylamide gel electrophoresis of the non-absorbed 1251-labelledlipoprotein which contained 75% of the radioactivity in apo-C and relatively very little at the top of the gel.
171
The possibility that some of the higher density products of VLDL catabolism are mixtures of different lipoproteins having similar hydrated densities led to an attempt to separate them by agarose gel filtration without prior centrifugation. The distribution of the lipoproteins in typical control and experimental perfusates following perfusion with human VLDL are shown in Fig. 9. It, will be seen that except for a very small peak at the void volume, usually smaller than in the preparation depicted in Fig. 9, the control VLDL elutes as a single peak. Following perfusion three lipoprotein peaks were discernible. The first was found in the void volume. The major peak corresponds to the control VLDL, although in some experiments a slight shift to the right was observed. A new peak appeared as the bed volume was approached. Material from each peak was pooled, dialysed, concentrated by lyophilization, delipidated and analyzed by 1polyacrylamide gel electrophoresis: The ratios of the radioactivity in regions i : 4 (apo-B : apo-C) of the gels (cf. Fig. 7) for each peak is shown in Fig. 9. It will be noted that the perfusion did not produce a significant change in the ratios of peak II. Peak I contained lipoproteins predominantly labelled in ap&B, resembling circulating LDL. On the other hand, the lipoproteins in peak 01 contained very little if any labelled apoB relative to 1251-labelled apo-C, ai pattern consistent with HDL. Electron microscopy (Fig. 10) of the material eluted in peak I showed large aggregates containing many particles resemblin LDL in size as well as some VLDL. It thus appears likely that during hy %rolysis of the triacylglycerol of human VLDL by heart lipoprotein lipase duting perfusion, some LDL is formed.
100
140
180
220
200
ELUENT Iml,
Fig. 9. Agarose gel column filtration of lipopro human VLDL in the absence (control - - - - - -) radioactfvitv in ape-B : a~04 regions of the represent the mean of three experiments f S.E.
ins in perfusatcs following perfusion of 12SX-labelled of beating rat hearts. The ratio ,oi lipoproteins in the three peaks
Fig. 10. Negatively stained lipoprotein from the void volume peak (If of the agarose gel filtration (cf. Fif. 9). Magnification 93 000 X. Note the presence of numerous particles of LDL size as well as some larger VLDL particles. AU particles appear to be emeshed in a large aggregate.
Discussion The use of perfused hearts for the study of the first stage of the catabolism of VLDL has produced a number of interesting observations. The heart is a source of immobilized lipoprotein lipase which has similar characteristics to those of heparin-solubilized heart lipoprotein lipase [ 161. However, the beating heart is a more physiologic~ preparation since some of the lipid and protein products of lipoprotein lipase activity are taken up by the tissue. In general there appears to be little difference in the extent of removal of the triacylglycerol between human and rat VLDL. The relatively constant amount of triacylglycerol removed from rat VLDL over a wide range of perfusate concentrations suggests that the enzyme system is close to saturation even at the
lowest concentrations used in these experiments. This would be predicted from the data of Fielding and Higgins [16]. Although the greater variation in the extent of breakdown of human VLDL (cf. Fig. 2) may be due to individual differences among blood donors, it is also possible that the kinetics of catabolism of human and rat VLDL differ. There are differences between the products of rat and human VLDL recovered in the d 1.019-1.063 g/ml fraction. At low levels of rat VLDL in the perfusate (corresponding to the normal rat serum concentrations) very little LDL is formed, most of the lipoprotein being taken up by the heart. This correlates with the low concentration of LDL found in rat serum. At higher concentrations lipoproteins of d 1.019-1.063 g/ml accumulate. In contrast, a constant proportion of the catabolized human VLDL is consistently found in the d 1.019-1.063 g/ml range regardless of the initial perfusate concentrations. These lipoproteins appear to be a mixture of true LDL and triacylglycerol-depleted remnants which retain some of the non-B apoproteins. The production of the LDL, even at low concentrations of VLDL, may correlate with the relatively high concentration of LDL in human circulation. It is possible, however, that a species-specific receptor of limited capacity leads to the removal of rat LDL at low perfusate VLDL levels. Hydrolysis of triacylglycerol-rich lipoproteins is associated with a loss of apo-C [7,15,38]. The present data show a significant correlation between the loss of triacylglycerol and apo-C from VLDL during catabolism. A similar correlation has recently been reported by Glangeaud et al. [ 391 using purified milk LDL and VLDL labelled by absorption of exogenous isolated apo-C. In view of the relatively small (and in some experiments negligible) recovery of the apoprotein in the d > 1.21 g/ml fraction of the perfusate, it is likely that much of the native apo-C lost from VLDL, but remaining in the perfusate, is in the form of a lipoprotein containing considerable phospholipid. Although the major objective of the present work was to characterize the products of VLDL catabolism, the incidental finding that apoproteins of VLDL are removed from the perfusate by the hearts is of interest. While the present data does not completely rule out the possibility that some VLDL is trapped by the heart, it seems unlikely that this is significant. The uptake of VLDL is markedly diminished if the heart has been treated with heparin to remove lipoprotein lipase. In addition, chylomicron apoprotein is not taken up by the perfused rat heart [15] in spite of the larger size of these particles. The utilization of triacylglycerol-fatty acid from chylomicrons (Dolphin, P.J., unpublished results) and VLDL [40] for cardiac acylglycerol synthesis and oxidation also suggests that the lipoproteins are not simply filtered or trapped during passage through the heart. It is not presently known whether the iodinated VLDL protein removed from the perfusate is simply bound to the endothelium or is internalized into either the endothelial or myocardial cells. Bierman et al. [38] have noted that cultured cells take up VLDL and their remnants, but these are catabolized very slowly. Similarly, the recovery of 8085% of the ‘*‘I-labelled apoprotein removed from VLDL as protein, of which approx. 50% is in perfused hearts (cf. Table I) suggests a relatively slow catabolic rate of protein in the present experiments. The present results raise some questions as to whether the in vitro model of
174
successive heart and liver perfusions devised for chylomicrons can be applied to VLDL catabolism. A considerable amount of the apoprotein and lipid is taken up directly by the heart. There is evidence in the swine that the liver plays a minor role in the catabolism of LDL [41]. In the rat, radioactivity from injected labelled VLDL appears primarily in the liver [42]. However, labelling of the lipid as well as the protein moiety of rat VLDL makes it difficult to evaluate fully the fate of remnant apoproteins. Preliminary experiments carried out in our laboratory show that there is considerably less uptake of human or rat VLDL remnants by rat liver than the corresponding chylomicron remnants. Thus there appears to be a more complex disposal mechanism for the products of VLDL catabolism than for chylomicron remn~~. The possibility must also be considered that the remnant formed by the action of heart lipoprotein lipase are incomplete, requiring at least a second step, probably involving lecithin-cholesterol acyltransferase. Schumaker and Adams [43] have suggested that lecithin-cholesterol acyltransferase may play a role in modifying the lipoprotein surface coat to accomodate the continuous removal of triacylglycerol. The observation that LDL derived from human VLDL tends to occur in an aggregated state may be a reflection of an inability to remodel the membrane. Norum et al. [44] observed a tendency of some LDL isolated from lecithin-cholesterol acyltransferasedeficient patients to aggregate. The large partially emptied remnants noted following perfusion resemble the very large particles found in the LDL of lecithin-cholesterol acyltr~sfer~deficient subjects [ 451,. Further studies are therefore being carried out on the role of this enzyme, HDL, and other factors which may play a role in the catabolism of VLDL. The results which have been obtained with the perfused rat heart suggest that there are some small but significant differences between the catabolism of rat and human VLDL which may be related to the accumulation of LDL in man. Acknowledgements This work was supported by grants from the Medical Research Council of Canada and the Quebec Heart Foundation. We wish to thank the Montreal Transfusion Centre of the Canadian Red Cross Blood T~sfusion Service and Mme. M. Caisse for the supply of human blood. References 1 2 3 4 6 6 7 8 9 10
Brunzell, J.D., Hazzard, W.R., Porte, Jr.. D. and Bierman, E.L. (1973) J. Clin. Invest. 62,1678-1586 Higgins, J.M. and Fielding, C.J. (1975) Biochemistry 14, 2288-2293 Dolphin, P.J. and Rubinstein, D. (1974) Biochem. Biophys. Res. Commun. 57. 808-814 Nestel. P.J.. Havel, R.J. and Bezman, A. (1963) J. Clin. Invest. 42,1313-1321 Redgrave. T.G. (1970) J. Clin. Invest. 49, 466-471 Eisanberg, S., Bilheimer. D.W.. Levy, R.1. and Lindgren. F.T. (1973) Biochim. Biophys. Acta 326, 361-377 Mjgs, O.D., Faergeman, 0.. Hamilton, R.L. and Havel, R.J. (1975) J. Ciin. Invest. 56,603-615 Eisenberg. S. and Rachm6ewitz. D. (1975) J. Lipid Res. 16, 341-351 Faergeman, 0. and Have& R.J. (1975) J. Clin. Invest. 65.1216-1218 GitIin, D., Cornwell, D.G.. Nakasato, D., Oncley, J.L., Hughes, Jr.. W.L. and Janeway, C.A. (1968) J. Clin. Invest. 37.172-184
175 11 12 13 14 15
Wilson. D.E. and Lees, R.S. (1972) J. Chn. Invest. 51, 1051-1057 Bezmsn-Tarcher. A. and Robinson, D.S. (1965) Proc. R. Sot. Lond. Ser. B. 162,406410 Rubenstein. B. and RubInstein. D. (1972) J. Lipid Res. 13.317-324 Twu. J-S., Garfinkel. A.S. and Schots. M.C. (1976) Atherosclerosis 24.119-128 Noel. S-P., Dolphin, P.J. and Rubinstein. D. (1975) Biochem. Biophys. Res. Commun. 163, 764-772
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
Fielding, C.J. and Higgins, J.M. (1974) Biochemistry 13.4324-4330 Fielding. C.J. (1976) Biochemistry 15.879-884 Dory, L. and Rubinstein, D. (1976) Clin. Res. 26, 680A Bleehen. N.M. and Fisher, R.B. (1954) J. Physiol. Lond. 123, 260-276 Miller, L.L.. Bly. C.G.. Watson, M.L. and Bale. W.F. (1951) J. Exp. Med. 94, 431-453 Hamilton, R.L.. Berry, M.N., Williams. M.C. and Seveiinghaus. E.M. (1974) J. Lipid Res. 15. 182-186 Have]. R.J., Eder. H.A. and Bragdon, J.H. (1955) J. Clin. Invest. 34, 1345-1353 Fidge. N.H. and Pouhs, P. (1974) CIin. Chim. Acta 52.15-26 McFarlane. A.S. (1958) Nature 182. 53 Mans, R.J. and NoveIIi. G.D. (1961) Arch. Biochem. Biophys. 94.48-53 Nestruck. A.C. and Rubinstein, D. (1976) Can. J. Biochem. 54,617-628 Lowry, O.H.. Rosebrough. N.J.. Farr, A.L. and RandaB. R.J. (1951) J. Biol. Chem. 193.265-275 Folch. J.. Lees, M. and SloaneStanley. G.H. (1957) J. Biol. Chem. 226. 497-509 Van Handel, E. (1961) CIin. Chem. 7.249-251 Pocock. D.M.-E., RafaI. S. and Vost. A. (1972) J. Chromatogr. Sci. 10, 72-76 Blanchette-Mackie, E.J. and Scow. R.O. (1973) J. Ceil Biol. 58. 689-708 Eisenberg. S. and Levy, R.I. (1975) Adv. Lipid Res. 13, l-89 Kane, J.P. (1973) Anal. Biochem. 53,350-364 Lee, D.M. and Alaupovic. P. (1970) Biochemistry 9.2244-2252 Rubenstein, B. and Steiner, G. (1975) Can. J. Biochem. 53.128-134 Noel, S-P. and Rubinstein. D. (1974) 3. Lipid Res. 15. 301-308 McConathy, WJ. and Alaupovic. P. (1974) FEBS Lett. 41.174-178 Bierman. E.L., Eisenberg, S., Stein, 0. and Stein, Y. (1973) Biochim. Biophys. Acta 329.183-169 Glangeaud. M.C.. Eisenberg. S. and Olivecrona. T. (1977) Biochim. Biophys. Acta 486, 23-35 Delcher. ILK., Fried, M. and Shipp. J.C. (1965) Biochim. Biophys. Acta 106.10-18 Sniderman, A.D., Carew. T.E.. Chandler. J.G. and Steinberg, D. (1974) Science 183.526-528 Risenberg, S. and RachmiIewitz. D. (1974) Biochhn. Biophys. Acta 326. 378-390 Schumaker. V.N. and Adams, G.H. (1970) J. Theor. Biol. 26, 89-91 Norum. K.R., Glomset, J.A.. Nichols. A.V. and Forte, T. (1971) J. CIin. Invest. 50.1131-1140 Forte. T.. Norum. K.R., Glomset. J.A. and Nichols, A.V. (1971) J. CIin. Invest. 50.1141-1148
