Apolipoprotein E effectively inhibits lipoprotein lipase-mediated lipolysis of chylomicron-like triglyceride-rich lipid emulsions in vitro and in vivo.

Apolipoprotein E (apoE) is an important determinant for the liver uptake of triglyceride-rich lipoproteins and emulsions by the remnant receptor. In the current study, we assessed an additional role of apoE as modulator of the metabolism of triglyceride-rich lipoproteins in vitro and in vivo. Glycerol tri[3H]oleate [14C]cholesteryl oleate double-labeled triglyceride-rich emulsions were injected into fasted rats. The serum half-life of glycerol tri[3H]oleate was 3-fold faster (5.4 min) than that of [14C]cholesteryl oleate (16.7 min), confirming lipoprotein lipase (LPL)-mediated processing. To establish a specific effect of apoE on emulsion lipolysis rather than liver uptake, rats were functionally hepatectomized, and hypo(apo)lipoproteinemia was induced by 17alpha-ethinyl estradiol treatment. An apoE concentration-dependent inhibition of emulsion-triglyceride hydrolysis was observed, reaching a 14.8-fold increased half-life of glycerol tri[3H]oleate as compared with that in the absence of exogenous apoE. The mechanism and specificity of the effect of apoE on emulsion lipolysis by purified LPL was assessed in vitro. Addition of apoE to glycerol tri[3H]oleate-labeled emulsions led to a concentration-dependent inhibition of [3H]oleate release (9.5% residual LPL activity at 60 microg/ml apoE), while apoA-I was ineffective. The inhibitory effect of apoE was not abolished by reductive methylation of lysine residues, whereas selective modification of arginine residues by 1,2-cyclohexadione completely cancelled the inhibitory effect of apoE. It is concluded that apoE can specifically inhibit the LPL-mediated hydrolysis of emulsion triglycerides both in vitro and in vivo, and that arginine residues in apoE are essential for this effect. We suggest that in addition to its role in receptor recognition, apoE also modulates the LPL-mediated processing of triglyceride-rich lipoproteins.

Lipoprotein lipase (LPL 1 ; EC 3.1.1.34), a 55-kDa glycopro-tein enzyme, is synthesized in various tissues such as adipose tissue and muscles, traverses the interstitial space and endothelial cell barrier, and subsequently binds to the luminal surface of capillary endothelial cells. LPL is highly expressed in adipose tissue, heart, and skeletal muscles, although it is also present in other tissues (1). Anchoring of LPL to endothelium is proposed to be mediated by 220-kDa heparan sulfate proteoglycans (2) and a 116-kDa heparin-releasable binding protein associated with proteoglycans (3)(4)(5). LPL, in its active dimeric form, is primarily responsible for the extrahepatic metabolism of very low density lipoproteins (VLDL) and chylomicrons (6). Through the concerted action of LPL and apolipoprotein (apo) C-II as activator (7), lipoprotein core triglycerides are hydrolyzed and the released fatty acids are transported into predominantly adipose tissue and muscle cells (8,9). Although LPL acts on mainly triglycerides, it can also hydrolyze the primary acyl bond of phospholipids in triglyceride-rich lipoproteins as a consequence of its phospholipase A-1 activity (10 -12). During lipolysis, VLDL and chylomicrons become enriched in apoE (13,14), and resulting remnants are subsequently taken up by the liver via apoE-specific receptors (15)(16)(17). Recently, nonenzymatic roles of LPL in lipoprotein metabolism have also been proposed, predominantly based on in vitro studies. LPL can form a bridge between lipoproteins and proteoglycans (18) or the low density lipoprotein (LDL) receptor-related protein (19), and it has been suggested that LPL facilitates the hepatic clearance of remnants through initial binding to proteoglycans, which is subsequently followed by internalization via the LDL receptor-related protein (20,21). This effect of LPL resides in its C-terminal domain and is independent of its catalytic activity (22,23).
The role of apolipoproteins in modulating LPL activity has been examined extensively using both native lipoproteins and triglyceride-rich emulsions. ApoC-II is a well-known activator of LPL (7, 24 -26). Windler et al. (27) suggested that an increased lysophosphatidylcholine concentration, resulting from progressive lipolysis, reduces the affinity of apoC-II for chylomicrons and results in the termination of triglyceride hydrolysis. Both apoH (␤ 2 -glycoprotein I) (28) and apoA-IV (29) have also been shown to increase LPL activity, although only in the presence of apoC-II. In contrast, apoC-III displayed noncompetitive inhibitory properties against both apoC-II and triolein in LPL-mediated triglyceride hydrolysis (10,30), an effect ascribed to the N-terminal domain of apoC-III (31). Jackson et al. (32) correlated the dose-dependent inhibition of LPL by apoC-III with the degree of apoC-II displacement from the substrates. Aalto-Setälä et al. (33) used VLDL from human apoC-III transgenic and control mice as substrates for purified LPL and observed no effect of the variation in apoC-III content on lipolysis rate. However, other variations in apolipoprotein composition of VLDL obtained from the transgenic mice (such as a reduced apoE content) could have masked such an effect (33).
The role of apoE in LPL-mediated lipolysis has been subject to controversy and was mainly assessed in vitro (34 -37). Functional apoE was postulated to be required for the metabolic conversion of VLDL into LDL (34), and Yamada and Murase (35) observed that the activity of LPL toward triolein emulsions was enhanced by apoE. However, an inhibitory role of apoE in LPL-mediated lipolysis was also suggested (36,37). Using a synthetic peptide containing the receptor binding domain of apoE (amino acids 139 -153), McConathy and Wang (36) observed an inhibition of the lipolysis of VLDL. A K i of 50 M toward triolein emulsions was estimated, and therefore the physiological importance of apoE in modulating plasma triglyceride levels was suggested. Gómez-Coronado et al. (37) fractionated VLDL and showed that the apoE content, but not the apoC-II content, was inversely correlated with the LPL-mediated lipolysis rate of VLDL in vitro. The potential effect of apoE on LPL is thus subject to discussion, as well as its relevance for the in vivo modulation of LPL.
Recently, we utilized a triglyceride-rich apolipoprotein-free emulsion model of native chylomicrons with well-defined physicochemical characteristics. Upon intravenous injection into rats, the emulsion acquired apolipoproteins, e.g. apoE and apoCs, and was subsequently taken up by liver parenchymal cells via lactoferrin-sensitive apoE-specific receptors just like endogenous chylomicrons (38). The rate of liver uptake was greatly increased by preloading the emulsion with rec-apoE. In the present study, we used this emulsion model to evaluate the role of apoE in LPL-mediated lipolysis and the mechanism of interaction between apoE and LPL. A specific inhibition of LPL by apoE is observed, and Arg residues and structural elements (possibly the lipid binding domain) of apoE are essential for this effect. Since these emulsions mimic chylomicrons, we anticipate that enrichment with apoE during lipolysis of lipoproteins may be a physiological determinant for the release of remnants from LPL in vivo.
Preparation of Emulsions-Emulsions were prepared according to the sonication and ultracentrifugation procedure of Redgrave and Maranhao (40) from 100 mg of total lipid at a weight ratio triolein:egg yolk phosphatidylcholine:lysophosphatidylcholine:cholesteryl oleate:cholesterol of 70:22.7:2.3:3.0:2.0, using a soniprep 150 (MSE Scientific Instruments, UK) at 18 output that is equipped with a water bath for temperature (54°C) maintenance (38). The emulsion was fractionated by consecutive density gradient ultracentrifugation steps in a Beckman SW 40 Ti rotor. After centrifugation for 22 min at 20000 rpm at 20°C, an emulsion fraction containing large emulsion particles was removed from the top of the tube by aspiration and replaced by NaCl buffer of similar density (i.e. 1.006 g/ml) (40). The emulsion fraction obtained after a second centrifugation for 22 min at 40,000 rpm was used for subsequent studies. This emulsion contained 78.1 Ϯ 0.6% triolein, 15.4 Ϯ 0.7% phospholipid (14.4 Ϯ 0.5% egg yolk phosphatidylcholine and 1.0 Ϯ 0.3% lysophosphatidylcholine), 4.2 Ϯ 0.3% cholesteryl oleate, and 2.4 Ϯ 0.1% cholesterol (w/w) (mean Ϯ S.E.; n ϭ 6) as determined with Boehringer Mannheim (Mannheim, Germany) enzymatic kits. The obtained emulsions were homogeneous with respect to size (low polydispersity of 0.07-0.13), and mean particle diameters were 82.2 Ϯ 2.9 nm (mean Ϯ S.E.; n ϭ 8) as determined by photon correlation spectroscopy using a Malvern 4700 C system (Malvern Instruments, UK). Measurements were performed at 25°C and a 90°angle between laser and detector. Emulsions were stored at 20°C under argon and used for characterization and metabolic studies within 5 days following preparation, in which period no physicochemical changes occurred. For synthesis of emulsions containing both radioactively labeled cholesteryl oleate and triolein, 10 Ci of [ 14 C]cholesteryl oleate and 75 Ci of glycerol tri[ 3 H]oleate were added to 100 mg of total lipid. Further preparation and isolation of the emulsions were similar as described above.
LPL-mediated Processing of Emulsions in Vivo-Overnight fasted male Wistar rats of mass 225-245 g were anesthetized by intraperitoneal injection of sodium pentobarbital (15 mg/kg body weight), and the abdomens were opened. Glycerol tri[ 3 H]oleate [ 14 C]cholesteryl oleate double-labeled emulsions (0.50 mg of emulsion triglyceride) were injected via the vena cava inferior. When indicated, rats received a preinjection of heparin (500 units/kg) via the left arteria carotis that had been cannulated, 10 min prior to injection of the radiolabeled emulsion. At the indicated times, blood samples of 300 l were taken from the vena cava inferior and allowed to clot for 30 min. 3 H and 14 C radioactivity in 100 l of serum, obtained after centrifugation for 3 min at 16,000 ϫ g, were counted following combustion in a Packard Tri-Carb 306 Sample Oxidizer (recovery Ͼ97%). The total amount of radioactivity in the serum was calculated using the equation: serum volume (ml) ϭ [0.0219 ϫ body weight (g)] ϩ 2.66 (41). In order to determine liver uptake, liver lobules were tied off, excised, and weighed at the indicated times. The amount of liver tissue tied off during the experiment did not exceed 15% of the total liver weight. At 30 min, rats were killed, and organs were excised and weighed. Radioactivity in liver and other tissue samples was counted after combustion and corrected for the serum radioactivity in the tissues at the time of sampling (38).
Source of LPL-Purified bovine milk LPL (5800 units/mg), suspended in 3.8 M ammonium sulfate, 0.02 M Tris-HCl at pH 8.0, was obtained from Sigma. Before use, aliquots were centrifuged (10 min at 8000 ϫ g, 4°C), the resulting pellet was resuspended in 0.02 M Tris-HCl, 0.1 M NaCl at pH 8.0 and dialyzed extensively against repeated changes of buffer (42). The LPL solution was stored at 4°C at a final protein concentration of 0.1 mg/ml and used within 3 days after dialysis.
LPL Activity Assays-Glycerol tri[ 3 H]oleate [ 14 C]cholesteryl oleate double-labeled emulsions were incubated with purified LPL based on earlier described methods (8,43). Hereto, 1.0 mg/ml emulsion triglycerides were incubated at 37°C in 60 mg/ml FFA-free BSA as [ 3 H]oleate acceptor, 5% (v/v) heat-inactivated (30 min at 56°C) rat serum as a source of coactivator apoC-II, and 0.1 M Tris-HCl pH 8.5, at a total volume of 1.2 ml. At t ϭ 0, 4.0 g/ml LPL was added, and the scattered photon count rate was determined during lipolysis by photon correlation spectroscopy using a Malvern 4700 C system (Malvern Instruments, UK). Measurements were performed at 37°C and a 90°angle between laser and detector.
Purification of ApoA-I-Human HDL was obtained from the blood of healthy volunteers by differential ultracentrifugation (1.063 Ͻ d Ͻ 1.21 g/ml) as described by Redgrave et al. (46). HDL was depleted from apoE by passage through a heparin Sepharose affinity column (47), dialyzed against water, and freeze-dried. Lipids were removed by repeated extraction at 4°C (ethanol:diethyl ether ϭ 3:1, v/v) and centrifugation (5 min at 2000 ϫ g). The final pellet was dried, stored under N 2 at Ϫ20°C, and dissolved in PBS, pH 7.4, prior to use. The purity of apoA-I was confirmed by 5-20% gradient SDS-polyacrylamide gel electrophoresis and subsequent Coomassie Brilliant Blue R-250 staining.
Modification of Lysine (Lys) and Arginine (Arg) Residues of ApoE-Lys residues of apoE were reductively methylated according to the method of Weisgraber et al. (48) for LDL and HDL c , as described in detail for lactoferrin (49). 2.0 mg of apoE in 1 ml of 0.15 M NaCl, 0.3 mM EDTA, pH 7.0, was mixed with 0.75 ml of 0.3 M Na 2 B 4 O 7 , pH 9.0, at 0°C. At t ϭ 0, 1 mg of Na(BH 4 ) was added, directly followed by six additions of 1 l of formaldehyde (37% aqueous) at 6-min intervals. The reaction was terminated by elution from a Sephadex G-25 medium column with 10 mM Tris-HCl, 0.15 M NaCl, pH 7.4. Methylated apoE (Me-apoE) was dialyzed against repeated changes of 0.15 M NaCl, 0.3 mM EDTA, pH 7.0 at 4°C. The extent of Lys modification was Ͼ95%, as checked by determination of residual free amino groups using 2,4,6trinitrobenzenesulfonic acid exactly according to the method of Habeeb (50). The linearity range of the reaction was assessed using BSA as a standard.
Arg residues of apoE were selectively modified with CHD according to the procedure of Patthy and Smith (51). 2.0 mg/ml apoE in 0.15 M NaCl, 0.01% EDTA, was diluted three times in 0.15 M CHD, 0.2 M Na 2 B 4 O 7 at pH 8.1, and incubated at 35°C. The extent of modification of Arg residues was monitored by subjecting 20-l samples (withdrawn from the incubate and stored in liquid N 2 ) to 0.75% (w/v) agarose gel electrophoresis at pH 8.8 using 0.075 M Tris-HCl, 0.080 M hippuric acid, 0.65 mM EDTA buffer. R f values of Coomassie Brilliant Blue R-250 visualized bands were determined relative to the front marker bromphenol blue. After 14 h of incubation, the reaction mixture was extensively dialyzed against repeated changes of PBS, 1 mM EDTA, pH 7.4 at 4°C. The effect of both Lys and Arg modifications on apoE integrity was monitored by 5-20% SDS-polyacrylamide gel electrophoresis, and subsequent staining with 0.2% Coomassie Brilliant Blue R-250.
Association of (Modified) Apolipoproteins to Emulsions-ApoE and apoA-I were radioiodinated at pH 10.0 with carrier-free 125 I according to a modification (52) of the ICl method (53). Free 125 I was removed by Sephadex G-25 gel filtration and extensive dialysis against 8 mM PBS containing 1 mM EDTA, pH 7.4, with repeated changes of buffer. More than 98.5% of the label in the proteins were trichloroacetic acid-precipitable. The specific activities of 125 I-apoE and 125 I-apoA-I were 473 and 1668 dpm/ng, respectively. Subsequently, Lys and Arg residues in 125 I-apoE were modified as described above. Radioiodination did not affect the extent of modification in both procedures. Radiolabeled (modified) apolipoproteins were then incubated with emulsions at various ratios for 30 min at 37°C, and emulsion-bound activity was separated from free activity using density gradient ultracentrifugation. The mixtures were placed at the bottom of a Kontron centrifuge tube and overlaid with 2.8-ml KBr solutions (all including 0.2 M NaCl and 0.3 mM EDTA, pH 7.4) with densities of 1.063, 1.019, and 1.006 g/ml, respectively. After centrifugation for 21-22 h at 40,000 rpm, 4°C, top fractions were aspirated following tube slicing. The remaining volumes were fractionated at a flow rate of 1.2 ml/min using an LKB Bromma 2132 Microper-pex peristaltic pump, starting at the bottom of the tube. All fractions were assayed for 125 I activity.
Effect of (Apolipo)proteins on LPL Activity-LPL activity was assessed by the methods described above. To determine the effect of apoE on particle size reduction, the emulsion (1.0 mg/ml triglyceride) was preincubated with 60 g/ml apoE (triglyceride:apoE ϭ 500:30, w/w) for 30 min at 37°C before addition of LPL. Similarly, emulsions were preincubated with varying concentrations of (modified) apoE, apoA-I, and/or lactoferrin to monitor the effect on LPL-mediated [9,10-3 H]oleate generation. When appropriate, relative reaction velocities were calculated as percent [ 3 H]oleate/min and related to those of concomitantly incubated control emulsions (100%).

Effect of ApoE on Emulsion Lipolysis in Hepatectomized Rats-Male
Wistar rats of mass 260 -320 g were anesthetized by diethyl ether, and the abdomens were opened. Functional hepatectomy was performed as described (54, 55) and checked by the absence of bleeding upon liver incision. Glycerol tri[ 3 H]oleate [ 14 C]cholesteryl oleate double-labeled emulsions (0.50 mg of emulsion triglyceride) were injected via the vena cava inferior, with or without previous incubation with apoE for 30 min at 37°C. At 1, 11, and 21 min after injection, a bolus injection of 0.33 ml of 5% glucose in PBS, pH 7.4, was administered. Blood samples were taken and processed as described above, and a liver sample was checked for the absence of radioactivity following combustion. When indicated, rats were pretreated with 17␣-ethinyl estradiol (EE) dissolved in propylene glycol at 5 mg/kg body weight for 3 successive days (56).
Effect of Serum on Dissociation of ApoE from Emulsions-The amount of apoE that, upon preincubation with emulsions and subsequent injection into control and 17␣-EE pretreated rats, remains associated with the emulsions was assessed as follows. Emulsions (100 g of triglyceride) were successively incubated with 125 I-apoE (triglyceride: apoE ϭ 500: 2.5 Ϫ 100, w/w; 30 min at 37°C) and with serum obtained from control or 17␣-EE pretreated rats (1.0 ml; 30 min at 37°C). Incubation mixtures were then subjected to density gradient ultracentrifugation, tubes were fractionated, and fractions were assayed for 125 I activity.

Liver Uptake and Serum Decay of the Emulsion-Previous
experiments have shown that, upon intravenous injection into rats, the emulsion acquires apolipoproteins, and that the cholesteryl oleate core is taken up by apoE-specific receptors on liver parenchymal cells (38). In order to verify that the emulsion also resembles native chylomicrons with respect to in vivo hydrolysis of core triglycerides by LPL, the glycerol tri[ 3 H]oleate [ 14 C]cholesteryl oleate double-labeled emulsion was injected into rats (Fig. 1) (Fig. 1, right panel). Uptake of glycerol tri[ 3 H]oleate by the liver (9.6 Ϯ 5.2% at 30 min after injection) was lower than that of the emulsion core label [ 14 C]cholesteryl oleate (58.0 Ϯ 0.1%) (Fig. 1, left panel). Preinjection of 500 units/kg heparin, which results in release of LPL into the plasma reaching sustained maximal levels between 10 and 60 min after injection (57), accelerated the processing of the emulsion. The serum half-life of glycerol tri[ 3 H]oleate was decreased (t1 ⁄2 Ͻ 1 min) whereas the uptake rate of [ 14 C]cholesteryl oleate by the liver was increased (84.2 Ϯ 0.4% at 10 min after injection) (results not shown).
Effect of ApoE on LPL Activity in Vitro-The effect of apoE on LPL-mediated lipolysis was first investigated in vitro, using two independent methods based on emulsion size reduction and [ 3 H]oleate release, respectively.
In photon correlation spectroscopy, the scattered radiation intensity from large particles is more intense than that from the same quantity of small particles, and could be used as an objective relative measure of size reduction. Indeed, addition of LPL to the emulsion in the presence of an LPL coactivator (apoC-II) and FFA acceptor (BSA) resulted in a time-dependent decrease in scattered photon count rate (Fig. 2). The fact that LPL activity can be suppressed by a high salt concentration (43) has been adopted to validate this method. Indeed, addition of 1.0 M total concentration NaCl after 45 min of incubation led to a complete blockade of further count rate reduction (Fig. 2). Preincubation of the emulsion with apoE at a triglyceride:apoE weight ratio of 500:30 resulted in an increased initial count rate, which was caused by association of apoE to the emulsion particle shell. ApoE clearly retarded the velocity of count rate reduction. As shown later, binding of apoE to the emulsion was nearly saturated at this ratio (63.1 Ϯ 2.4 molecules of apoE per emulsion particle). The scattered photon count rate resulting from a similar apoE solution in the absence of emulsion was negligible. As an equilibrated particle solution is required for optimal calculation of particle size distribution, the data on photon scattering have not been transformed into average particle diameters.
Incubation of emulsion with LPL in the presence of apoC-II and BSA resulted in rapid generation of [ 3 H]oleate from emulsion-incorporated glycerol tri[ 3 H]oleate, obeying initial linear enzyme kinetics. Lipolysis was not terminated before complete conversion of glycerol tri[ 3 H]oleate was achieved (Fig. 3). Preincubation of emulsion with apoE led to a dose-dependent inhibition of the lipolysis rate. At an emulsion triglyceride: apoE ϭ 500:14 weight ratio, 63.6% inhibition of [ 3 H]oleate release was observed. A further increase in the apoE concentration led to a further inhibition of [ 3 H]oleate generation reaching a value of 94.2% at a triglyceride:apoE ϭ 500:51 weight ratio (Fig. 3).
Specificity Effect of ApoE on LPL Inhibition in Vitro-Lys and Arg residues in apoE were modified by reductive methylation (Me-apoE) (48) and 1,2-cyclohexanedione modification (CHD-apoE) (51), respectively. Reductive methylation did not alter the net surface charge of apoE (Fig. 4) nor the ability of apoE to associate with the emulsion. 69.1 Ϯ 5.5 molecules of Me-apoE associated per emulsion particle (emulsion triglyceride:Me-apoE ϭ 500:30, w/w) (Fig. 5). CHD modification resulted in an increase in the electrophoretic mobility of apoE (R f ϭ 0.45) on agarose gel toward the anode, reaching an R f value of 0.59 after 14 h of incubation (Fig. 4). The ability of 14-h modified apoE to associate with the emulsion was approximately 6-fold reduced, as only 11.7 Ϯ 6.2 molecules of CHD-apoE associated per emulsion particle (emulsion triglyceride: CHD-apoE ϭ 500:30, w/w) (Fig. 5). Both modification procedures did not result in degradation of apoE as determined by 5-20% SDS-polyacrylamide gel electrophoresis, but a very slight increase (Ͻ0.5 kDa) in apparent molecular mass was observed for 14-h modified CHD-apoE (results not shown).
A concentration-dependent inhibition of LPL-mediated [ 3 H]oleate release was still observed upon methylation of apoE, although less effective (Fig. 6, left panel). A 33.1% inhibition of [ 3 H]oleate release was observed at an emulsion triglyceride: Me-apoE ϭ 500:14 weight ratio, reaching 87.5% at a 500:51 weight ratio. In contrast, neutralization of positive charges of Arg residues completely cancelled the inhibitory effect of apoE on [ 3 H]oleate release (Fig. 6, right panel). In fact, a stimulatory effect on lipolysis rate was observed (52.9% at an emulsion triglyceride:CHD-apoE ϭ 500:51 weight ratio) as compared with that of the protein-deficient emulsion.
As positively charged amino acids of apoE appeared to be involved in LPL inhibition, involvement of the receptor binding domain of apoE (residues 141-150) was suspected. The N terminus of lactoferrin possesses an Arg/Lys cluster similar to that of the binding domain of apoE and therefore effectively inhibits the apoE-mediated interaction of the emulsion with the remnant receptor on rat liver parenchymal cells in vivo (38). However, lactoferrin was unable to inhibit LPL-mediated hydrolysis of glycerol tri[ 3 H]oleate at protein concentrations similar to those used for apoE (Fig. 7). Lactoferrin appeared also unable to modulate the inhibiting effect of apoE (emulsion triglyceride:apoE ϭ 500:30, w/w), even at a 55.5-fold weight excess over apoE (results not shown).
To further ascertain the specificity of apoE in LPL inhibition and to exclude the possibility that apoE inhibits LPL through protein occupation of the emulsion shell leading to diminished access of LPL to triglycerides, the effect of apoA-I on lipolysis was also examined. ApoA-I resembles apoE with respect to its 22-amino acid amphipathic ␣-helical segments (58,59) and therefore readily associates with emulsions (60, 61), but does not contain the typical Arg/Lys sequence. Indeed, apoA-I effectively associated with the emulsion and to a higher extent than apoE (Fig. 5), However, apoA-I showed almost no effect on lipolysis in a concentration range similar to that used for apoE. For example, at an emulsion triglyceride:apoA-I molar ratio comparable to an emulsion triglyceride:apoE ϭ 500:51 weight ratio, 87.8% residual LPL activity was still observed in contrast to only 5.8% in case of apoE (Fig. 7).
Effect of ApoE on Emulsion Lipolysis in Hepatectomized Rats-Previous experiments have shown that enrichment of emulsions with apoE greatly increases the rate of liver uptake of the emulsion in rats (38). To establish the effect of apoE on emulsion lipolysis rather than on liver uptake, rats were functionally hepatectomized. Subsequent injection of the glycerol tri[ 3 H]oleate [ 14 C]cholesteryl oleate double-labeled emulsion resulted in a rapid serum decay of glycerol tri[ 3 H]oleate (t1 ⁄2 ϭ 5.8 min). [ 14 C]Cholesteryl oleate did not disappear from the serum as was expected in the absence of additional specific recognition sites for the emulsion on extrahepatic tissues (38). Preincubation of the emulsion with apoE up to an emulsion triglyceride:apoE ϭ 500:21 weight ratio did not alter the serum decay of glycerol tri[ 3 H]oleate (Fig. 8, left panel).
To minimize the pool of apo(lipo)proteins that may facilitate exchange or interference with the effect of apoE, rats were pretreated with 17␣-EE, which induces a profound hypo(apo)lipoproteinemia (56,62). Pretreatment resulted in a 6.9 Ϯ 2.3% (mean Ϯ S.D.; n ϭ 6) loss of body weight, whereas all untreated rats gained weight. Total serum cholesterol showed a 4.1-fold reduction from 0.575 Ϯ 0.053 mg/ml (mean Ϯ S.D.; n ϭ 6) to 0.140 Ϯ 0.035 mg/ml (mean Ϯ S.D.; n ϭ 6), which was mainly caused by a pronounced reduction of the high density lipoprotein (HDL) level.
To estimate exchange of apoE from emulsions to lipoproteins in vivo, emulsions were preincubated with apoE and subsequently incubated with control or 17␣-EE rat serum. The number of apoE molecules that, upon reisolation, still associated with the emulsion was determined (Fig. 9). Within the concentration range of apoE used for the assessment of its effect on emulsion lipolysis in vivo (triglyceride:apoE ϭ 500:0 -100, w/w), it appeared that the number of apoE molecules that was recovered with the emulsion was reduced approximately 2-fold upon incubation with control serum (Fig. 9), while the lost apoE was mainly found in the HDL density range. In contrast, a low transfer of apoE to HDL and hence a higher emulsion association (75-90% of the maximal amount of apoE) were observed upon incubation with serum from 17␣-EE pretreated rats.
Upon injection of emulsions into 17␣-EE pretreated rats, the serum decay of emulsified glycerol tri[ 3 H]oleate was again rapid (t1 ⁄2 ϭ 3.3 min), indicating that LPL is still active under these hypo(apo)lipoproteinemic conditions, whereas [ 14 C]cholesteryl oleate remained in the serum (Fig. 8, right panel). Under these conditions, apoE showed a profound concentrationdependent inhibiting effect on emulsion lipolysis. At an emulsion triglyceride:apoE ϭ 500:52 weight ratio, a 5.0-fold increased serum half-life of glycerol tri[ 3 H]oleate was observed (t1 ⁄2 ϭ 16.5 min), which increased up to t1 ⁄2 ϭ 48.8 min at a 500:111 weight ratio. (modified) apolipoprotein weight ratios. Subsequently, emulsion-associated and free apolipoproteins were separated by density gradient ultracentrifugation, and the binding of apolipoprotein to the emulsion (molecules per particle) was calculated.

DISCUSSION
Binding of LPL both to endothelial surface sites and substrates can be affected by a variety of structurally unrelated compounds. LPL is readily displaced from binding sites on endothelial cells by generation of an excess of free fatty acids, providing a potential feedback mechanism to regulate the extent of lipolysis (63)(64)(65). Injection of heparin that can bind to Arg/Lys clusters positioned in the N-terminal domain of LPL (26,66) leads to the release of LPL as LPL-heparin complexes (9) and an increased lipolysis rate of plasma lipoproteins (67).
Concomitant with the hydrolysis of core triglycerides by LPL (8,9), chylomicrons acquire apolipoproteins such as apoE (13,14), both circulating in a free tetrameric form (68) or in HDL (69,70). Functional apoE has been postulated to be required for the LPL-mediated metabolic conversion of VLDL into LDL (34). Yamada et al. (35) observed an activation of LPL by apoE toward triolein emulsions. Clark and Quarfordt (63) also showed an enhancing effect of apoE on emulsion triolein hydrolysis by LPL that was immobilized to heparin-Sepharose, although resulting from apoE-mediated binding of emulsions to the support rather than a direct LPL-activating effect of apoE. Borensztajn et al. (71) were unable to observe any effect of apoE on LPL-mediated lipolysis of apolipoprotein-free chylomicrons, but suboptimal conditions were used in which the LPL-coactivator apoC-II was absent. An increasing body of evidence, however, points at a potentially inhibitory role of apoE in LPL-mediated lipolysis (36,37,72).
In this study we assessed the effect of apoE on lipolysis both in vitro and in vivo using a defined emulsion model of chylomicrons. Previously, we have shown that the emulsion acquires apoE and apoCs upon injection into rats, and that it is taken up by the liver via apoE-specific receptors on parenchymal cells, similarly as chylomicrons (38). Lactoferrin that specifically inhibits the apoE-mediated uptake of chylomicron remnants and ␤-VLDL by the remnant receptor on parenchymal cells through an Arg/Lys cluster similar to the receptor binding domain (amino acids 141-150) of apoE (16,73), was also very effective in inhibiting the liver uptake of the emulsion (38). Redgrave and Maranhao (40) showed that endogenously radiolabeled rat chylomicrons displayed a 3.7-fold faster serum decay and 2.9-fold reduced liver uptake of their triacylglycerols compared to the cholesteryl esters. The emulsion showed a comparable in vivo behavior. Experimentally increased plasma LPL levels by preinjection with heparin also strongly accelerated emulsion processing. These data further substantiate the relevance of the emulsion as a model for chylomicrons, including affinity for and extensive processing by LPL. This emulsion thus forms a well-defined system to study the role of individual pure apolipoproteins on LPL activity.
In the absence of apoE, emulsions were rapidly processed by LPL as determined by photon correlation spectroscopy and release of free fatty acids, suggesting that apoE is not an absolute requisite for the emulsion to be processed by LPL. ApoE that was preincubated with the emulsion appeared to inhibit LPL-mediated emulsion lipolysis dose dependently in vitro, up to 94.2% at an emulsion triglyceride:apoE ϭ 500:51 weight ratio. Theoretically, the observed effect could be a general effect of apolipoproteins as binding to the emulsion particle surface could lead to a diminished access of LPL to the triglyceride substrate. With the assumption that, within an air-water interface monolayer, the limiting surface area per amino acid of apoE is 20.8 Å 2 (68), it can be calculated from our data that at saturation of the emulsion particle shell with apoE (73.2 Ϯ 2.2 molecules per particle), maximally 23.8% of the emulsion particle shell is occupied by apoE. This value corresponds well to a maximal 20.5% shell occupation as can be determined from previous data on the binding of maximally 7 molecules of apoE per 26 nm sized phosphatidylcholine/triolein emulsion particle (68). It can thus be concluded that even at the highest apoE/ lipid ratio, still more than 75% of the shell is accessible for LPL. Both apoA-I and apoE bind to the emulsion in a similar concentration-dependent manner, which can be explained by their similar 22-amino acid amphipathic segments (58,59). ApoA-I did not exert any effect on LPL-mediated lipolysis, and we can therefore suggest that apoE exerts a specific inhibitory effect on LPL.
It was subsequently attempted to characterize the element(s) of apoE responsible for LPL inhibition. One of the main features that discriminates apoE from other apolipoproteins is the presence of an Arg/Lys-rich sequence in the receptor binding domain (residues 141-150) (15). Previously, a synthetic peptide of apoE (residues 139 -153) has been shown to inhibit VLDL lipolysis in vitro (36). Selective modification of Arg residues in apoE by CHD strongly reduced its ability to associate with the emulsion (approximately 6-fold) and completely cancelled the LPL-inhibitory properties of apoE. At a similar protein association to the emulsion (14.4 molecules of apoE per particle at a triglyceride:protein ϭ 500:3 weight ratio versus 12.2 molecules of CHD-apoE at a 500:51 weight ratio), apoE caused inhibition of LPL (16.0%) whereas CHD-apoE stimulated lipolysis (52.9%). It can thus be concluded that in addition to a reduction in emulsion association, CHD modification of apoE also directly affects the inhibitory effect on LPL.
Reductive methylation of Lys residues did not have any effect on binding of apoE to the emulsion and did not greatly influence the inhibitory effect of apoE on emulsion lipolysis. The fact that reductive methylation of Lys residues of apoE in HDL c totally abolished the binding affinity toward human fibroblasts (48) suggests a different mechanism of interaction of apoE with LPL as compared to receptor interaction. Lactoferrin does also possess an Arg/Lys-rich cluster, similarly to the binding domain of apoE, although it does not contain a hydrophobic domain to enable interaction with emulsions or lipoproteins. Since lactoferrin did not modulate LPL, nor the inhibiting effect of apoE on LPL, it is suspected that besides the positively charged Arg/Lys sequence, direct interaction of the C-terminal lipid binding domain in apoE with the emulsion and/or LPL is necessary to exert an effect on lipolysis.
Initially, an effect of apoE on emulsion lipolysis was not detected in hepatectomized control rats in vivo. As apoE should exert a local effect on LPL at the level of the endothelium, it should remain associated with the emulsion upon injection. However, approximately 50% of the emulsion-associated apoE was redistributed to HDL upon in vitro incubation with serum, as can be expected from its well-known property to exchange between lipoproteins (68 -70). To minimize the interference with circulating lipoproteins, a hypo(apo)lipoproteinemic rat model was therefore used. Indeed, redistribution of apoE to HDL was reduced (10 -25% of emulsion-bound apoE), leaving approximately 2-fold more exogenous apoE on the emulsion for interaction with LPL as compared to the situation in normolipidemic rats. The apoE concentrations at the level of LPL  3 H and 14 C activities were determined in the serum. In order to enable comparison between rats, the values represent the relative 3 H activity (compared to the injected dose) per ml of serum divided by that of the 14 C activity.
FIG. 9. Effect of serum on dissociation of apoE from emulsions. Emulsions (100 g of triglyceride) were incubated (30 min at 37°C) with 125 I-apoE at various emulsion triglyceride (TG):apoE weight ratios, without (E) or with subsequent incubation with serum obtained from control (Ç) or 17␣-EE pretreated (É) rats. The emulsions were reisolated by density gradient ultracentrifugation, and the association of 125 I-apoE (molecules per particle) was calculated. could thus explain the dose-dependent inhibitory effect of apoE on LPL activity as observed in vivo in hepatectomized hypo(apo)lipoproteinemic rats. The serum half-life of emulsiontriglycerides increased up to 14.8-fold at a triglyceride:apoE ϭ 500:111 weight ratio compared to that in the absence of exogenous apoE.
The impossibility to measure an inhibitory effect of apoE on emulsion lipolysis in the control rats does not necessarily mean that a local LPL inhibition does not occur. Under physiological conditions, chylomicrons are gradually released from the lymph into the blood, and sufficient apoE is expected to be available for efficient association during remnant formation. In our experimental set-up, however, emulsions must be administered as a bolus injection and apoE will initially redistribute to the available HDL pool. We showed that 50 -80 molecules of exogenous apoE per particle led to a marked inhibition of lipolysis in vivo. As normal rat VLDL contains approximately 25 apoE molecules (74), while 45-50 apoE molecules are present in VLDL remnants, 2 this is thus in the physiological range.
Saxena et al. (75)(76)(77) also investigated the effect of apoE on LPL, although not related to the enzymatic role of LPL. In their in vitro system, LPL appeared to increase LDL retention to glycosaminoglycans of the subendothelial matrix (75). Purified apoE and apoE-rich high density lipoproteins (HDL), but not apoA-I or apoE-free HDL, were able to displace LDL from LPL (76). Studies with chemically modified LDL-apoB and apoE indicated the predominant role of the Arg/Lys-rich domain of apoE in the inhibition of LDL retention (77). Although based on lipoprotein binding to LPL instead of lipoprotein lipolysis by LPL, these data are in agreement with our present observations. The similarity in results may also imply that apoE inhibits lipolysis by substrate dissociation from LPL rather than by inactivation of substrate-associated LPL. Physiologically, it can be envisioned that enrichment of lipoproteins with apoE during LPL-mediated lipolysis results in dissociation of the lipoprotein remnants from endothelium-bound LPL, leading to the accessibility of the remnants for uptake by hepatic apoEspecific receptors. The physiological importance of the inhibition of LPL by apoE might also be concluded from the characterization of lipoproteins in transgenic apoE-deficient animals. Indeed, VLDL isolated from homozygous apoE-knockout mice appears to contain less triglycerides compared to VLDL from control mice (78), suggesting more extensive processing in the absence of apoE.
It is concluded that apoE can effectively inhibit the LPLmediated hydrolysis of emulsion triglycerides in vitro and in vivo. Since these emulsions mimic chylomicrons, we anticipate that in addition to its role in receptor recognition, apoE might also be an important factor in the modulation of the LPLmediated hydrolysis of triglycerides.