Apolipoprotein E Mediates the Retention of High-Density Lipoproteins by Mouse Carotid Arteries and Cultured Arterial Smooth Muscle Cell Extracellular Matrices
Lipoprotein retention in the vascular extracellular matrix (ECM) plays a critical role in atherogenesis. Previous studies demonstrated the presence of apo A-I and E in atherosclerotic lesions, suggesting that HDL may be trapped by the artery wall. We sought to determine mechanisms by which HDL could be bound and retained by the arterial wall, and whether apo E was a principal determinant of this binding. We evaluated in situ accumulation of fluorescently labeled DiI-human HDL±apo E in perfused carotid arteries from apo E–null mice. Apo E was important in mediating HDL binding to the vascular wall, with a 48±16% increase in accumulation of DiI-labeled apo E–containing HDL (HDL3+E) compared with DiI-apo E–free HDL (HDL3−E) (P=0.003). To investigate possible mechanisms responsible for retention, we assessed binding of unlabeled HDL3−E and HDL3+E to ECM generated by cultured arterial smooth muscle cells. Similar to the in situ carotid artery data, HDL3+E bound better to the ECM than did HDL3−E (3-fold lower Ka and 3.5-fold higher Bmax for HDL3+E versus HDL3−E). These differences were eliminated after either neutralization of arginine residues on apo E or digestion of matrix with chondroitin ABC lyase, suggesting that chondroitin and/or dermatan sulfate proteoglycans were responsible for apo E–mediated increased binding. These findings demonstrate that HDL can bind to both intact murine carotid arteries and smooth muscle cell–derived ECM, and that apo E is a principal determinant in mediating the ability of HDL to be trapped and retained via its interaction with ECM proteoglycans.
The retention of lipoproteins in the arterial extracellular matrix (ECM) is a critical step in atherogenesis.1–3⇓⇓ Apolipoproteins (apo) B and E, found on triglyceride-rich, atherogenic lipoproteins, have been found extensively in atherosclerotic lesions, often associated with ECM proteoglycans.4 However, immunohistochemical studies show that apo A-I, which is the primary apolipoprotein on HDL, is present in human atherosclerotic lesions as well.4–7⇓⇓⇓ Further, apo A-I and E were found to colocalize with biglycan,4 a major arterial extracellular proteoglycan, suggesting that HDL can be trapped by ECM proteoglycans. We also have shown that apo E–containing HDL (HDL+E) binds to purified matrix proteoglycans in vitro, whereas apo E–free HDL (HDL−E) does not bind.4,8⇓ Because HDL−E contains a large amount of apo A-I, this suggests that apo A-I itself does not bind to proteoglycans. This hypothesis is supported by reports that mammalian apo A-I does not contain heparin binding domains.9 Therefore, these data indicate that apo E may play an important role in the retention of HDL by ECM proteoglycans.
Because the artery wall contains ECM components in addition to proteoglycans, we conducted studies to evaluate the hypothesis that the presence of apo E on HDL could affect its retention by the intact vascular ECM. To evaluate this hypothesis, we used a system to measure in situ accumulation of fluorescently labeled HDL3 in perfused carotid arteries from apo E–null (−/−) mice. To determine mechanisms responsible for this retention, we used an in vitro binding assay of HDL3 to ECM generated by cultured arterial smooth muscle cells. Based on data from both models, our studies show that HDL3−E does bind to the arterial ECM, although HDL3+E binds significantly more, and is mediated in part by ECM proteoglycans. Thus, the presence of apo E on HDL3 is important in the mechanism(s) mediating the binding of HDL3 to the arterial ECM and may play a role in HDL retention in the artery wall.
Materials and Methods
HDL3 (d=1.125−1.21 g/mL) was isolated by sequential density ultracentrifugation from plasma obtained from a pool of healthy human volunteers, as described previously.10 Each of the 10 HDL preparations used in this study were obtained from a pool of 6 individuals, and the majority had an apo E3/E3 genotype. HDL3−E and HDL3+E fractions were prepared, with the absence of apo B confirmed by SDS-PAGE and Western blotting, as described previously8 (detailed in the expanded Materials and Methods section, which can be found in the online data supplement available at http://www.circresaha.org). The content of apo A-I and E was quantified by immunonephelometry (Behring Nephelometer Analyzer, Germany).
To monitor the accumulation of HDL in mouse carotid arteries, HDL3−E and HDL3+E each were labeled with the fluorescent hydrocarbon probe DiI (No. D-282, Molecular Probes) using a modification of a procedure described previously.11 Briefly, HDL3−E and HDL3+E were incubated separately in the presence of DiI (final concentration, 0.053 mmol/L DMSO) and human lipoprotein-deficient serum for 6 hours at 37°C under sterile conditions. DiI-labeled HDL3−E and DiI-HDL3+E were reisolated by ultracentrifugation and concentrated using Centricon-100 concentrators (Amicon). Free DiI was removed using a Sephadex G-25 PD-10 column (Pharmacia) equilibrated in 150 mmol/L NaCl and 1 mmol/L EDTA (pH 7.4). Samples were filtered with a 0.22-μm filter before perfusion in mouse carotid arteries (detailed in online data supplement).
In Situ Perfusion System
To evaluate the binding of HDL to the vasculature in situ, 5- to 7-month-old apo E−/− mice bred on a C57BL/6 background were obtained from Jackson Laboratories (Bar Harbor, Maine). Apo E–null mice were chosen to avoid any interference of endogenous apo E with the binding of apo E present on HDL particles. Mice were placed on a high-fat diet (15% (wt/wt) fat, 1.25% (wt/wt) cholesterol, 0.5% (wt/wt) cholate; Ralston-Purina) for 8 weeks to increase vascular matrix deposition. Mice were anesthetized with Nembutal (50 mg/kg), and carotid arteries were isolated and removed for the perfusion experiments, as described previously12,13⇓ (detailed in online data supplement). The use of animals in the present study was approved by the Animal Use Committee and the University of California, Davis.
Lipoprotein flux was measured using fluorescence microscopy as described in detail previously,12,13⇓ except that DiI-labeled HDL was perfused and its flux monitored. The in situ perfusion process was performed using both DiI-HDL3−E and DiI-HDL3+E with the same mouse carotid artery. Experiments alternated the order of perfusion of DiI-HDL3−E or DiI-HDL3+E. Photons emitted from the DiI-HDL3−E and DiI-HDL3+E were detected and quantified by a photometer that converts the number of detected photons to millivolts (online Figure 1). The fluorescence in millivolts then was converted into nanograms of HDL protein that accumulated per minute per centimeter2 arterial surface area, using the apo A-I concentration of DiI-HDL3−E and DiI-HDL3+E in the perfusate solution and the size of the photometric window, in the following relationship13:
Lumen concentration was calculated by determining lumen volume and concentration of DiI-HDL3−E and DiI-HDL3+E (as apo A-I) in the perfusate. Fractional accumulation (If accumulation) then was calculated by dividing the HDL accumulation by the amount of HDL in the perfusate. This additional step was necessary to normalize for the range of apo A-I concentrations in the samples.
The results obtained with this method are significant in that they are “real-time” measurements of HDL flux in the artery wall where vascular flow, hydrostatic pressure, and blood components, such as transfer proteins, are held constant. This enabled us to better study the actions of apo E–free and apo E–containing HDL with the artery wall, uncomplicated by other parameters.
To capture photographic images showing the differential arterial accumulation of DiI-HDL3−E and DiI-HDL3+E, some of the mouse carotid arteries were prepared for fluorescence microscopy. Immediately after the washout of DiI-HDL3−E or DiI-HDL3+E from the artery with unlabeled buffer, the carotid arteries were embedded in OCT compound (Tissue Tek, Sakura) and frozen at −80°C. The arteries were cut into 10-μm sections, fixed for 10 minutes in paraformaldehyde, washed with PBS, and coverslipped with polyvinyl alcohol (PVA) aqueous mounting media. Photographic images were captured using a Zeiss Axioplan fluorescence microscope, a Hamamatsu fast-cooled CCD camera, and an MCID M2 image analysis system (Imaging Research). To control for background fluorescence, control sections were prepared in the same manner from arteries that were perfused with nonfluorescent buffer solution alone.
Matrix Binding Assay
The ability of HDL3−E and HDL3+E to bind to vascular ECM in vitro was assessed using a modification of an ELISA plate assay.14 Nonfluorescent, unlabeled HDL was used in this assay because bound HDL was detected using colorimetric development. A goat polyclonal anti-human apo A-I antibody (a kind gift from Dr Jack Oram, University of Washington, Seattle, Wash) was used to detect bound HDL (detailed in online data supplement).
To evaluate the role of ECM proteoglycans in the interaction, in some experiments chondroitin and dermatan sulfate glycosaminoglycans were removed by incubating the matrices with chondroitin ABC lyase (0.2 U per well) (ICN Biomedicals) for 3 hours at 37°C15 and the matrices washed 3 times with Earl’s Balanced Salt Solution/1% (wt/vol) bovine serum albumin before the addition of HDL. This method has been shown to digest, as well as liberate glycosaminoglycans but not matrix proteins, from the ECM.14,15⇓
We have previously reported that neutralization of arginine residues on apo E abolishes the ability of HDL3+E to bind smooth muscle cell matrix-derived proteoglycans, which are negatively charged.8 We chose to use this modified HDL to further test the idea that apo E is a principal determinant in mediating the ability of HDL bind the ECM via its interaction with matrix proteoglycans. Arginine residues were neutralized by treating HDL3+E samples with cyclohexanedione as described previously16 (detailed in online data supplement). Immediately after these modification procedures, HDL was dialyzed extensively into PBS, containing 25 μmol/L butylated hydroxytoluene, at 4°C and analyzed for the ability to bind to the vascular ECM in vitro as described above. The extent of derivatization was assessed by measurement of electrophoretic mobility on agarose gels in barbital buffer at pH 8.610 or by amino acid analysis.17
Detection of Lipases
The presence or absence of hepatic lipase (HL) and lipoprotein lipase (LpL) in HDL3−E, HDL3+E, and arterial smooth muscle cell–generated ECM was determined by SDS PAGE (10% acrylamide) and Western blot analysis using a monospecific polyclonal rabbit anti-human HL antiserum (titer 1:5,500; a kind gift from Dr Helén Dichek, University of Washington, Seattle, Wash) or a monoclonal mouse anti-human LpL antibody (5D2; titer 1:1000; a kind gift from Dr John Brunzell, University of Washington). After incubation with the appropriate IgG-peroxidase–linked secondary antibody (Boehringer-Mannheim), the peroxidase reaction was developed using a chemiluminescence detection system (Pierce). HL-positive and -negative controls were plasma samples from a human HL-transgenic mouse and a non–HL-transgenic mouse, respectively (from Dr Helén Dichek, University of Washington). Purified bovine milk LpL (from Dr John Brunzell, University of Washington) was included as the positive control for LpL, which is known to cross-react with the 5D2 antibody.18
Statistical significance was determined by the Student’s t test. A value of P<0.05 was considered statistically significant.
To evaluate the role of apo E in the retention of HDL by the artery wall, the in situ accumulation of fluorescently labeled (DiI) HDL3+E and HDL3−E was measured in perfused carotid arteries from apo E−/− mice. This method allowed us to determine how the presence of apo E on HDL could affect the accumulation by the ECM as it exists ex vivo. After 8 weeks on the high-fat diet, the intima was thickened, but no fatty streaks or atheroma were detected in the carotid arteries (data not shown). Perfusion with DiI-HDL3+E resulted in significantly higher amounts of fluorophores remaining in the arterial wall after initial washout (If accumulation) compared with DiI-HDL3−E (Figure 1). By obtaining measurements after the initial washout, we were able to correct for any differences in DiI labeling efficiency among the HDL preparations. There was greater arterial accumulation of DiI-HDL3+E compared with DiI-HDL3−E, as noted by the fact that for each time point after the washout, there was greater fluorescence intensity for DiI-HDL3+E. In addition, it can be seen that the area under the curve is greater for DiI-HDL3+E accumulation than for DiI-HDL3−E (Figure 1). Fluorescence intensity did not reach baseline during the washout period following perfusion of DiI-HDL3+E, but returned to baseline before the end of the washout period following DiI-HDL3−E perfusion, which can be seen at the far right side of the graph in Figure 1 (see also online Figures 2 and 3⇓). This further indicates that the presence of apo E on HDL resulted in greater arterial wall accumulation. The fraction of the total amount of DiI-HDL3 that remained in the artery (fractional accumulation) compiled from 10 independent experiments demonstrated that DiI-HDL3+E accumulated to a significantly greater degree (48±16%) in apo E−/− mouse carotid arteries compared with DiI-HDL3−E (P=0.003) (Figure 2).
Consistent with these data, fluorescence microscopy showed that DiI-HDL3+E accumulated in the artery wall to a greater extent than DiI-HDL3−E (Figure 3). This is shown by the greater amount of red fluorescent DiI from HDL3+E compared with HDL3−E and the control (nonfluorescent buffer-perfused) within the artery wall. There was a significant amount of autofluorescence present in the internal elastic lamina for each of the arteries, whether they were perfused with DiI-HDL or nonfluorescent buffer alone. The presence of apo E on HDL3 resulted in the presence of DiI within the endothelium, whereas perfusion with HDL−E or the nonfluorescent buffer alone (no HDL) resulted in an absence of endothelial fluorescence retention. There also was greater fluorescence intensity within the subendothelium in those arterial sections that were perfused with DiI-HDL3+E, compared with the DiI-HDL3−E and buffer-alone treatments.
In order to address mechanisms responsible for the greater retention of HDL3+E compared with HDL3−E, experiments were conducted using ECM generated by cultured arterial smooth muscle cells. With this system we could identify specific components within the arterial wall that were responsible for the apo E–mediated increase in HDL binding. Similar to the in situ carotid artery data, HDL3+E bound significantly better to smooth muscle cell ECM than HDL3−E (Figure 4). Consistent with this, binding constants for the interaction indicated that HDL3+E had an approximately 3-fold lower affinity constant (Ka) (1.2×10−7 versus 3.4×10−7 mmol/L, respectively), and thus, 3-fold greater affinity and 3-fold higher maximum binding constant values (Bmax) (159.5 versus 50.8 relative absorbance units, respectively) compared with HDL3−E. Therefore, the apo E–mediated difference was greater for these in vitro ECM studies than the in situ carotid data. This may be due to differences in the types and amounts of proteoglycan and non-proteoglycan components, including LpL, that constitute the ECM generated by monkey arterial smooth muscle cells compared with the mouse vascular wall.
To evaluate the role of matrix dermatan and chondroitin sulfate proteoglycans in the retention of HDL by the arterial ECM, some of the matrices were digested with chondroitin ABC lyase before incubation with HDL. Removal of chondroitin and dermatan sulfate glycosaminoglycans from the matrix abolished the apo E–mediated enhanced binding of HDL3 (Figure 4), such that ECM that were devoid of these glycosaminoglycans bound HDL3+E and HDL3−E to the same extent. Matrices that were depleted of glycosaminoglycans demonstrated low-affinity binding of HDL3−E and HDL3+E, consistent with the idea that ECM proteoglycans play a major role in arterial lipoprotein retention.2,3⇓ Thus, these experiments demonstrated that the presence of apo E on HDL3 resulted in greater binding to ECM generated by cultured arterial smooth muscle cells. In addition, this binding was mediated in part by matrix dermatan sulfate and chondroitin sulfate proteoglycans.
To further test the idea that apo E on HDL binds ECM via its interaction with matrix proteoglycans, we neutralized arginine residues on apo E of HDL3+E using cyclohexanedione. We previously have shown that this treatment abolishes the ability of HDL3+E to bind matrix-derived proteoglycans in vitro.8 In the current set of experiments, the HDL3+E was modified with cyclohexanedione to a similar extent as that reported previously by Olin et al8 (data not shown). If apo E is important for HDL binding to the ECM via proteoglycans, then cyclohexanedione-modified HDL3+E should not bind as well to ECM as native HDL3+E, and should bind to a similar degree as HDL3−E. Indeed, cyclohexanedione-modified HDL3+E bound to the ECM significantly less than native HDL3+E, and even a little less than HDL3−E (Figure 5). This latter observation may be due to the modification of arginine residues on HDL3+E that are involved in non–proteoglycan-mediated binding to the ECM.
HL and LpL have previously been shown to facilitate the binding of lipoproteins to proteoglycans and the ECM.19–21⇓⇓ Thus, we tested if HL and LpL were present on HDL samples or within the ECM because these lipases could potentially act as mediators of HDL binding in addition to apo E. Using Western blot analysis, there was no detectable HL in either HDL3+E, HDL3−E, or within the ECM (Figure 6A). There was a strong signal for the positive control (plasma from a human HL-transgenic mouse) and no signal from a non–HL-transgenic mouse. Western blot analysis also indicated that LpL was not present in either HDL3+E or HDL3−E despite a strong signal from just 88 ng purified bovine milk LpL (Figure 6B). In addition, there was no detectable LpL in the arterial smooth muscle cell–generated ECM (Figure 6B). Therefore, these data suggest that neither HL nor LpL contribute to HDL binding to the ECM in vitro. Furthermore, the absence of these lipases on HDL3−E and HDL3+E provides additional support that apo E is the principal determinant of HDL binding to both the vascular wall and the ECM.
This study demonstrates that the presence of apo E on HDL3 plays an important role in the ability of HDL3 to be bound and retained by the vascular ECM, which is mediated in part by the binding of apo E by ECM proteoglycans. This was shown in experiments in which HDL3+E was retained to a greater extent than HDL3−E in a model that can measure the retention of fluorescently labeled (DiI) lipoproteins in perfused mouse carotid arteries in situ, as well as by ECM generated by cultured arterial smooth muscle cells.
We previously have shown that apo E was required for the binding of HDL to purified extracellular arterial proteoglycans, such as biglycan, in vitro,4 and data presented in the present study represent a progression of those studies. Biglycan is a dermatan sulfate proteoglycan that appears to be especially important in the retention of lipoproteins in human atherosclerotic lesions.4 However, it is only one of several components that constitute the ECM, the physiologically relevant and complex meshwork that lipoproteins encounter when they enter the subendothelial intima in vivo. Thus, the results presented in the present article are significant in that with both an in situ and an in vitro model of the complexities of the artery wall, the presence of apo E on HDL is an important determinant of its retention.
There have been multiple reports describing the presence of apo A-I, which is the major apolipoprotein found on HDL, in human atherosclerotic lesions.4–7⇓⇓⇓ Further, the study by O’Brien et al4 demonstrated that apo A-I colocalized with apo E and biglycan in human atherosclerotic lesions. Taken together, these findings suggest that the artery wall may retain a subset of HDL, and the present investigation provides possible mechanisms responsible for this retention.
In the present investigation, the difference in HDL3−E and HDL3+E binding to ECM generated by cultured arterial smooth muscle cells was eliminated when the matrices were first digested with chondroitin ABC lyase, removing chondroitin and/or dermatan sulfate glycosaminoglycans. This suggests that matrix chondroitin and dermatan sulfate proteoglycans were responsible for the majority of the apo E–mediated increase in HDL binding to ECM in vitro, despite the presence of multiple other matrix components. These data are consistent with the idea that proteoglycans can bind lipoproteins via their glycosaminoglycan side chains.22,23⇓ This interaction is thought to be primarily ionic in nature, such that negatively charged glycosaminoglycans bind to positively charged amino acid residues on apo B and E of lipoproteins.22,24⇓ The fact that HDL3+E has a greater positive charge than HDL3−E, due to the presence of clusters of positive amino acids on apo E,25 would be consistent with the observation that the increased retention of HDL3+E is mediated by its interaction with negatively charged glycosaminoglycans. It has been reported that apo E can bind to the core protein of biglycan in vitro, an interaction also hypothesized to be ionic.26 Thus, there also may be non–glycosaminoglycan chain–mediated mechanisms responsible for the interaction of lipoproteins with proteoglycans. These may be occurring in the in situ experiments presented in this study in which HDL3+E accumulates in the mouse artery wall to a greater degree than HDL3−E. However, for the in vitro ECM experiments, it appears that glycosaminoglycans are the primary determinants for apo E–mediated binding of HDL because chondroitin ABC lyase digestion of the ECM specifically eliminated the differences in binding between HDL3−E and HDL3+E. Furthermore, when we neutralized arginine residues on HDL3+E with cyclohexanedione to the extent that it would no longer bind matrix-derived proteoglycans in vitro,8 this modified HDL3+E bound significantly less than native HDL3+E to the smooth muscle cell–generated ECM. Cyclohexanedione-treated HDL3+E bound to the ECM to a similar extent as HDL3−E, which does not bind matrix proteoglycans.4,8⇓ Therefore, this further supports the idea that apo E is the principal determinant of HDL binding to the ECM via its binding to matrix proteoglycans.
LpL has been demonstrated to enhance the binding and retention of apo B– and E–containing lipoproteins to matrix proteoglycans, by acting as a “bridge” molecule.19,20⇓ This cannot explain the greater binding of HDL3+E to matrices generated from cultured smooth muscle cells presented in this study, because neither the HDL nor the cultured ECM contain LpL or HL, as detected by Western blot analysis. However, LpL and/or HL could have been present in the vascular wall in situ, which could contribute to the observations that HDL3+E was bound and retained to a greater extent than HDL3−E in mouse carotid arteries. Thus, there may be other component(s) in addition to proteoglycans that could contribute to apo E–mediated increases in HDL retention in vivo.
In the present study, HDL−E also was bound and retained by the arterial wall in situ, as well as by ECM in vitro, although the binding capacity and affinity were low compared with HDL3+E. Because HDL3−E binding to the ECM was not affected by chondroitin ABC lyase treatment, this suggests that the binding was mediated by non-glycosaminoglycan components of the matrix. This is consistent with our previous report demonstrating that HDL−E does not bind to purified proteoglycans in vitro.4 It may be that apo A-I, which is abundant on HDL3−E, interacts with non-proteoglycan ECM components, because previous studies have shown that apo A-I can bind fibronectin and collagen, for example.27 LpL also could potentially function within the mouse arterial wall to facilitate non–proteoglycan-mediated HDL3−E binding, because LpL may bind to HDL through lipophilic interactions that are independent of apo E.28
Apo B– and E–containing atherogenic lipoproteins (VLDL, IDL, and LDL) are thought to be the primary lipoproteins retained by arterial proteoglycans, as a means of initiating atherosclerosis. However, we show that HDL, typically an antiatherogenic lipoprotein, also binds to the arterial ECM. Albeit only a small percentage (approximately 5% to 6%) of HDL3 contain apo E,8 this fraction could have an increased chance of being retained by the artery wall via ECM proteoglycans. The majority of HDL does not contain apo E. However, a recent article reports that the amount of apo E on HDL was a significant independent predictor for a recurrent cardiovascular event.29 Thus, it may be that the presence of apo E on HDL reduces some of the antiatherogenic functions of HDL by increasing its ability to be trapped by vascular ECM proteoglycans.
What is the significance of this apo E–mediated retention of HDL? Apo E that is retained in the vascular intima could potentially contribute to lipid retention, and thereby, promote atherosclerosis. Apo E–containing HDL that is bound to and retained by the arterial wall may be susceptible to oxidation, after which it could be taken up by macrophages and be potentially atherogenic.30 In fact, a recent article describes the presence of oxidized HDL in atheromatous plaques.31
On the other hand, apo E within the ECM and artery wall could be atheroprotective. Apo E is thought to possess numerous antiatherogenic functions (reviewed by Mahley32 and Curtiss33), including the maintenance of plasma cholesterol,34,35⇓ participation in reverse cholesterol transport,36 regulation of T-lymphocyte and smooth muscle cell–mediated inflammatory responses,37 and antioxidant functions.38 Apo E also could compete with apo B for binding sites within the ECM, which might prevent the more atherogenic LDL from accumulating.
Whether apo E is atheroprotective or atherogenic or a combination of the two may depend on its localization and the level of apo E, among other factors. The data presented in this article suggest that the apo E present in a subset of HDL particles may contribute to the binding and retention of apo E–containing HDL by the ECM within the artery wall, where they might exert either anti- or pro-atherogenic effects. Further studies are needed to resolve these important biological questions.
This work was supported in part by National Institutes of Health Grants HL07028, HL30086, HL18645, HL03045, DK17047, DK02456, and DK07247. We gratefully acknowledge Joyce Murphy and Mohamed Omer for expert technical assistance.
Original received July 26, 2001; resubmission received October 1, 2001; second resubmission received April 29, 2002; revised resubmission received May 23, 2002; accepted May 23, 2002.
- ↵Williams KJ, Tabas I. The response-to-retention hypothesis of early atherogenesis. Arterioscler Thromb Vasc Biol. 1995; 15: 551–561.
- ↵O’Brien KD, Olin KL, Alpers CE, Chiu W, Hudkins K, Wight TN, Chait A. Comparison of apolipoprotein and proteoglycan deposits in human coronary atherosclerotic plaques: colocalization of biglycan with apolipoproteins. Circulation. 1998; 98: 519–527.
- ↵Hoff HF. Apolipoprotein localization in human cranial arteries, coronary arteries, and the aorta. Stroke. 1976; 7: 390–393.
- ↵Olin KL, Potter-Perigo S, Barrett PH, Wight TN, Chait A. Biglycan, a vascular proteoglycan, binds differently to HDL2 and HDL3: role of apoE. Arterioscler Thromb Vasc Biol. 2001; 21: 129–135.
- ↵Pitas RE, Innerarity TL, Weinstein JN, Mahley RW. Acetoacetylated lipoproteins used to distinguish fibroblasts from macrophages in vitro by fluorescence microscopy. Arteriosclerosis. 1981; 1: 177–185.
- ↵Rutledge JC, Woo MM, Rezai AA, Curtiss LK, Goldberg IJ. Lipoprotein lipase increases lipoprotein binding to the artery wall and increases endothelial layer permeability by formation of lipolysis products. Circ Res. 1997; 80: 819–828.
- ↵Chang MY, Potter-Perigo S, Wight TN, Chait A. Oxidized LDL bind to nonproteoglycan components of smooth muscle cell extracellular matrices. J Lipid Res. 2001; 42: 824–833.
- ↵Saito HT, Yamagata T, Suzuki S. Enzymatic methods for the determination of small quantities of isomeric chondroitin sulfates. J Biol Chem. 1968; 243: 1536–1542.
- ↵Innerarity TL, Pitas RE, Mahley RW. Lipoprotein-receptor interactions.In: Albers JJ, Segrest JP, eds. Plasma Lipoproteins. San Diego, Calif: Academic Press; 1986: 548–552.
- ↵Hazen SL, Gaut JP, Hsu FF, Crowley JR, d’Avignon A, Heinecke JW. p-Hydroxyphenylacetaldehyde, the major product of l-tyrosine oxidation by the myeloperoxidase-H2O2-chloride system of phagocytes, covalently modifies epsilon-amino groups of protein lysine residues. J Biol Chem. 1997; 272: 16990–16998.
- ↵Voyta J, Via D, Kinnunen P, Sparrow J, Gotto AJ, Smith L. Monoclonal antibodies against bovine milk lipoprotein lipase: characterization of an antibody specific for the apolipoprotein C-II binding site. J Biol Chem. 1985; 260: 893–898.
- ↵Edwards IJ, Goldberg IJ, Parks JS, Xu H, Wagner WD. Lipoprotein lipase enhances the interaction of low density lipoprotein with artery-derived extracellular matrix proteoglycans. J Lipid Res. 1993; 34: 1155–1163.
- ↵Ji ZS, Dichek HL, Miranda RD, Mahley RW. Heparan sulfate proteoglycans participate in hepatic lipase and apolipoprotein E–mediated binding and uptake of plasma lipoproteins, including high density lipoproteins. J Biol Chem. 1997; 272: 1285–1292.
- ↵Camejo G, Rosengren B, Olson U, Lopez F, Olofson S-O, Westerlund C, Bondjers G. Molecular basis of the association of arterial proteoglycans with low density lipoproteins: Its effect on the structure of the lipoprotein particle. Eur Heart J. 1990; 11: 164–173.
- ↵Camejo G, Fager G, Rosengren B, Hurt-Camejo E, Bondjers G. Binding of low density lipoproteins by proteoglycans synthesized by proliferating and quiescent human arterial smooth muscle cells. J Biol Chem. 1993; 268: 14131–14137.
- ↵Camejo G, Olofsson S-O, Lopez F, Carlsson P, Bondjers G. Identification of apo B-100 segments mediating the interaction of low density lipoproteins with arterial proteoglycans. Arteriosclerosis. 1988; 8: 368–377.
- ↵Cardin AD, Weintraub HJR. Molecular modeling of protein-glycosaminoglycan interactions. Arteriosclerosis. 1989; 9: 21–32.
- ↵Klezovitch O, Formato M, Cherchi GM, Weisgraber KH, Scanu AM. Structural determinants in the C-terminal domain of apolipoprotein E mediating binding to the protein core of human aortic biglycan. J Biol Chem. 2000; 275: 18913–18918.
- ↵Choi SY, Sivaram P, Walker DE, Curtiss LK, Gretch DG, Sturley SL, Attie AD, Deckelbaum RJ, Goldberg IJ. Lipoprotein lipase association with lipoproteins involves protein-protein interaction with apolipoprotein B. J Biol Chem. 1995; 270: 8081–8086.
- ↵Sacks FM, Alaupovic P, Moye LA, Cole TG, Sussex B, Stampfer MJ, Pfeffer MA, Braunwald E. VLDL, apolipoproteins B, CIII, and E, and risk of recurrent coronary events in the Cholesterol and Recurrent Events (CARE) trial. Circulation. 2000; 102: 1886–1892.
- ↵Nakajima T, Origuchi N, Matsunaga T, Kawai S, Hokari S, Nakamura H, Inoue I, Katayama S, Nagata A, Komoda T. Localization of oxidized HDL in atheromatous plaques and oxidized HDL binding sites on human aortic endothelial cells. Ann Clin Biochem. 2000; 37: 179–186.
- ↵Mahley RW. Apolipoprotein E: cholesterol transport protein with expanding role in cell biology. Science. 1988; 240: 622–630.
- ↵Curtiss LK. ApoE in atherosclerosis: a protein with multiple hats. Arterioscler Thromb Vasc Biol. 2000; 20: 1852–1853.
- ↵Thorngate FE, Rudel LL, Walzem RL, Williams DL. Low levels of extrahepatic nonmacrophage ApoE inhibit atherosclerosis without correcting hypercholesterolemia in ApoE–deficient mice. Arterioscler Thromb Vasc Biol. 2000; 20: 1939–1945.
- ↵Boisvert WA, Black AS, Curtiss LK. ApoA1 reduces free cholesterol accumulation in atherosclerotic lesions of ApoE-deficient mice transplanted with ApoE-expressing macrophages. Arterioscler Thromb Vasc Biol. 1999; 19: 525–530.
- ↵Pepe MG, Curtiss LK. Apolipoprotein E is a biologically active constituent of the normal immunoregulatory lipoprotein, LDL-In. J Immunol. 1986; 136: 3716–3723.