Studies were designed to test whether tyrosylation of high-density lipoprotein (HDLT) modifies its metabolic features. HDLT was less effective than native HDL in promoting cholesterol efflux from J774-AI macrophages. Cell association with fluorescent HDLT-apolipoprotein and the uptake of HDLT-[3H]cholesteryl hexadecyl ether were enhanced by 50% in comparison with native HDL. In addition, neutral cholesterol ester hydrolase (nCEH) activity in J774-AI, which controls the hydrolysis of cholesteryl ester stores to provide free cholesterol for cellular release, declined in the presence of HDLT. In vitro displacement experiments revealed the ability of HDLT to compete with oxidized and acetylated LDL, known as ligands of scavenger receptor (SR) class B type I/II. Similarly, treatment with a blocking antibody to SR-BI/II reduced the cell association of HDLT and native HDL by 50%. The addition of polyinosinic acid, an inhibitor of SR class A, reduced the cell association of HDLT without affecting that of native HDL. These findings provide evidence that HDLT can compete with modified LDL, bind SR-BI/BII and internalize cholesterol ester. Furthermore, the impaired capacity of HDLT in promoting cholesterol efflux from J774-AI was accompanied by diminished nCEH and enhanced recognition by SR-AI/II, which appears to involve the transport of cholesterol into cells.
The risk of cardiovascular disease is strongly and directly associated with the presence of low-density lipoprotein (LDL) (Brown and Goldstein, 1984; Kannel et al., 1971). The atherogenic effects of LDL particles may be related to their susceptibility to oxidative modifications, which affect the arterial wall (Quinn et al., 1987; Steinberg et al., 1989). The physicochemical properties of oxidized LDL include increased density, greater electrophoretic mobility, loss of esterified cholesterol, decreased polyunsaturated fatty acids, increased aldehydes and degraded apolipoprotein (apo) B (Gebicki et al., 1991). LDL oxidation results in the stimulation of macrophage cholesterol accumulation, foam cell formation, cytoxicity to cells of the arterial wall and the enhancement of inflammatory and thrombotic processes (Aviram and Fuhrman, 1998).
In contrast to LDL, elevated plasma levels of HDL lower the risk of clinically significant coronary events (Gordon et al., 1977; Gordon et al., 1989). Their beneficial anti-atherogenic actions stem from their involvement in the removal of excess cholesterol from extrahepatic cells for transport to and deposition in the liver, where cholesterol is excreted in the form of bile acids (Levy et al., 1996; Pieters et al., 1991). HDL has other potential anti-atherogenic properties, such as its ability to inhibit LDL oxidation (Parthasarathy et al., 1990), cytotoxicity of oxLDL (Hessler et al., 1979; Suc et al., 1997) and monocyte transmigration (Navab et al., 1991). However, HDL contains the majority of lipid peroxides in the plasma and is also susceptible to oxidative modifications (Bowry et al., 1992; Hahn and Subbiah, 1994; Salmon et al., 1992). Whether or not oxidative modification changes the antiatherogenic effects of HDL is under debate. On the one hand, it can alter the conformation and antigenicity of apo A-I, impair the ability of HDL to promote cholesterol efflux (Nagano et al., 1991; Morel, 1994; Salmon et al., 1992), reduce HDL binding to specific receptors (Mazière et al., 1993) and suppress cholesterol biosynthesis (Ghiselli et al., 1992). On the other hand, various investigators have reported that oxidative modification of HDL is less toxic than that of LDL (Alomar et al., 1992), reduces TNF-α secretion from human macrophages and, thus, the inflammatory process (Girona et al., 1997), and enhances cholesterol efflux (Francis et al., 1993; Wang et al., 1998). Such discrepancies in the aforementioned papers may depend on methodological variations in the isolation, storage and dosage of HDL or the cell model utilized.
Previously, it was emphasized that the generation of tyrosyl radicals by myeloperoxidase represents one physiologically plausible mechanism for the oxidation of lipoproteins in vivo. In fact, the likelihood of lipoprotein oxidation in vivo by this model is supported by the presence of high levels of myeloperoxidase and protein-bound dityrosine in human atherosclerotic lesions (Daugherty et al., 1994; Leeuwenburgh et al., 1997; Mazière et al., 1993), as well as the occurrence of micromolar concentrations of L-tyrosine in plasma (Mitchell et al., 1995) and tyrosyl-radical-modified LDL in the artery wall (Leeuwenburgh et al., 1997). Although tyrosyl-radical-oxidized HDL in vitro was shown to deplete cholesterol content in human skin fibroblasts and in mouse peritoneal macrophages (Francis et al., 1993; Wang et al., 1998), our own data demonstrated its ability to hamper in vivo reverse cholesterol transport, that is, the transfer of cholesterol to the liver and its transformation into bile acids (Guertin et al., 1997) in a similar fashion to malondialdehyde-modified HDL (Guertin et al., 1994). Therefore, it is plausible that tyrosylation changes in HDL apoA-I may be detrimental to cholesterol efflux.
The fate of lipoproteins is largely determined by the interaction between their constituent apolipoprotein moieties and the cellular receptors and enzymes involved in their lipid metabolism. Although a number of HDL-binding proteins has been described (Johnson et al., 1991; Oram and Yokoyama, 1996), their physiological role in reverse cholesterol transport remains uncertain, since none has been shown to be a functional HDL receptor that facilitates cholesterol efflux. Only recently has the scavenger receptor type B class I (SR-BI) been convincingly shown to bind HDL with high affinity and to mediate the selective cellular uptake and efflux of HDL cholesterol (Acton et al., 1996; Ji et al., 1997). Interestingly, scavenger receptors are unique with respect to their ability to interact with both native and oxidized LDL (Rigotti et al., 1995). Whether or not they can act as receptors for tyrosylated HDL has never been assessed.
In the current study, we examined the effects of enzymatically generated tyrosylation on the physiochemical and metabolic properties of HDL. More specifically, we tested cholesterol removal from cultured mouse macrophages of the J774-A1 cell line as well as the uptake and internalization of HDL-cholesterol. We also explored the receptors involved in the binding of tyrosylated HDL and the mechanism for cell cholesteryl ester depletion.
Materials and Methods
J774-A.1 cells (ATCC # TIB67) were obtained from American Type Culture Collection (Rockville, MD). Multiwell culture plates and flasks were from Falcon labware (Becton Dickinson Co.). Fetal Calf Serum (FCS) was purchased from Wisent Inc. and was heat-inactivated before use. DMEM, Penicillin-Streptomycin and phosphate-buffered saline (PBS) were obtained from GIBCO-BRL, whereas bovine serum albumin (BSA) fraction V, fatty-acid free, was from ICN Biochemicals. The Bio-Rad protein assay was from Bio-Rad Laboratories. Oregon Green 514 carboxylic acid, succinimidyl ester was purchased from Molecular Probes, Inc. (Eugene, OR). [1α, 2α(n)-3H]cholesteryl oleate (43.0 Ci/mmol) and cholesteryl [1-14C] oleate (56 mCi/mmol) were from Amersham Life Science, whereas [cholesteryl-1,2-3H(N)]-hexadecyl ether (40-60 Ci/mmol) was from DuPont-New England Nuclear Research Inc. Blocking anti-scavenger receptor BI and BII polyclonal antibody from Novus Biologicals Inc. was a mixture of IgGs purified on a sepharose column, and non-immune rabbit IgG was utilized as a negative control. All the other products and reagents were purchased from Sigma Chemical Co.
Isolation and modification of lipoproteins
Human LDL (d=1.019 to 1.063 g/ml), total HDL (d=1.063 to 1.210 g/ml) and HDL3 fraction (d=1.125 to 1.210 g/ml) were prepared from plasma of healthy human subjects, isolated by differential ultracentrifugation (Havel et al., 1955) and dialyzed intensively against PBS (pH 7.4) containing 150 mM NaCl and 0.3 mM ethylene diaminetetraacetic acid (EDTA). HDL3 (1 mg apoA-I/ml) was tyrosylated in the presence of horseradish peroxidase (100 nM), H2O2 (100 μM) and L-tyrosine (100 μM) for 24 hours at 37°C according to the method Francis et al. (Francis et al., 1993). Acetyl LDL (AcLDL) was prepared by the addition of acetic anhydride to normal LDL as described elsewhere (Basu et al., 1976), and its modification was verified by its mobility on agarose gel electrophoresis (Paragon, Beckman Instruments). In order to generate oxidized LDL (oxLDL), plasma LDL (3 mg apo B/ml) was extensively dialyzed against PBS (pH 7.4) containing 150 mM NaCl and 5 μM EDTA and then incubated with 10 μM CuSO4 for 18 hours at 37°C. The peroxidation level of oxLDL was determined by the measurement of its TBARS content (Yagi, 1987). All lipoproteins were filtered through a 0.2 μM Millipore membrane and stored at 4°C. ApoB and apoA-I concentrations were determined by immunonephelometry (Behring system) (Havel et al., 1955).
Labeling of lipoproteins
AcLDL was labeled with [3H]cholesteryl oleate (10 μCi/mg of apoB) as described previously (Guertin et al., 1994; Guertin et al., 1997; Terpstra et al., 1989). The labeling of normal and tyrosylated HDL3 with [3H]cholesteryl hexadecyl ether (10 μCi/mg of apoA-I) was carried out as reported previously (Guertin et al., 1994; Guertin et al., 1997; Terpstra et al., 1989). Normal and tyrosylated HDL3 (10 mg apoA-I) were labeled with the fluorescent probe Oregon Green 514 carboxylic acid, succinimidyl ester (0.5 mg) (Oregon-HDL3) according to the manufacturer's instructions (Molecular Probes, Inc Eugene, OR). The labeling efficiency between normal and tyrosylated HDL3 was evaluated by spectrofluorimetry, which revealed no marked differences in their Oregon Green 514 carboxylic acid content. Succunimidyl ester reacted with the primary amines of the two major HDL apolipoproteins (A-I and A-II). By contrast, [3H]cholesteryl hexadecyl ether was incorporated in HDL3 and served to measure cellular cholesteryl ester-HDL3 uptake. All lipoproteins were sterilized by filtration (0.2 μM Millipore membrane), stored at 4°C in the dark and used within 2 weeks of their preparation. Native and tyrosylated Oregon-HDL3, as well as HDL3-[3H]cholesteryl hexadecyl ether showed no evidence of oxidative modifications compared with non-labeled HDL3 since (a) the electerophoretic mobility of nascent and tyrosylated HDL was similar on agarose gel; (b) there was no apparent degradation of apolipoproteins when nascent and tyrosylated HDL were applied on SDS-PAGE; and (C) the lipid and protein composition of nascent and tyrosylated was not altered.
J774-A1 cells were grown and maintained in 25 cm2 flasks with DMEM containing 10% (vol/vol) Fetal Calf Serum, penicillin (100 U/ml) and streptomycin (100 μg/ml) in 5% CO2 and humidified air at 37°C. For neutral cholesteryl ester hydrolase (nCEH) activity, cells were seeded on 25 cm2 flask (2×106 cells/flask). In order to assess cellular HDL3-apolipoprotein association and cholesterol uptake, cells were seeded in 12-well plates at 0.2×106 cells per well. Similar experimental conditions were applied while testing the blocking effects of anti-scavenger receptor BI and BII antibody, except that 24-well plates were employed. On day two, all cells were cholesterol loaded during 24 hours in DMEM containing antibiotics, 5% LPDS and AcLDL (50 μg apoB /ml). On day three, medium was removed and cells were washed three times with DMEM (containing 5% LPDS and antibiotics) and then incubated in the same medium overnight to equilibrate intracellular cholesterol pool. After two washes with DMEM (containing antibiotics and 5% LPDS), cells were used for the different experiments.
J774-A1 cells were cultured in six-well plates at 0.8×106 cells per well and loaded with radiolabeled cholesterol by incubation for 24 hours in 1 ml of DMEM containing antibiotics, 5% LPDS and 7.4×105 DPM/ml [3H]cholesteryl oleate AcLDL (50μ g apoB/ml). Under the conditions used, approximately 0.6 pmol of labeled cholesterol was incorporated per incubation dish. Cells were then treated as mentioned above and incubated for 24 hours in DMEM supplemented with 5% LPDS in the absence (control) or presence of different concentrations of normal or tyrosylated total HDL or HDL3. The media were centrifuged at 4000 g for 10 minutes to remove any suspended or dead cells. The cells were washed three times with PBS, resuspended in 1 ml of distillated water and homogenized by sonication on ice. Cell and medium lipids were extracted by the method of Folch (Folch et al., 1957). The organic phase was dried down under nitrogen, and the radioactivity resuspended in a small volume of chloroform and fractionated by thin layer chromatography (TLC) on silica gel plates using a solvent system composed of hexane:ethyl-ether:acetic-acid (80:20:3, vol/vol). Cholesteryl ester and free cholesterol radioactivity in cells and media were measured in a liquid scintillation β counter (Beckman, LS 5000 TD). Protein concentration was determined in cell extracts by Bio-Rad Protein Assay kit with BSA as a standard. The net efflux of [3H] cholesterol mediated by normal or tyrosylated HDL types was expressed as a percentage of the amount of labeled cholesteryl ester and free cholesterol released to the medium, relative to total label in each well (medium plus cells), per amount of cell protein. All determinations were carried out in triplicate.
Assay of nCEH activity
An assay of nCEH activity was carried out as described by Holm and Osterlund (Holm and Osterlund, 1999). Cells were incubated in DMEM containing antibiotics and 5% LPDS in the absence (control) or presence of different concentrations of normal or tyrosylated HDL3 for 24 hours. After three washes with cold PBS, cells were homogenized in Tris-HCl buffer (50 mM, pH 7.0) containing sucrose (250 mM), EDTA (1 mM), dithiothreitol (DTT) (1 mM), leupeptine (20μ g/ml) and pepstatine (1 μg/ml) with a Sonifier microtip at the lowest power setting (3×10 seconds on ice). Cytosolic cell extracts were obtained by ultracentrifugation for 30 minutes at 43,000 g at 4°C. The supernatant was used as the nCEH enzyme solution. Protein content was determined with the Bio-Rad protein assay kit (Bio-Rad Laboratories) with BSA as a standard. The 0.2 ml nCEH reaction contained 50 mM potassium phosphate, pH 7.0, 0.5 mM EDTA, 0.5 mM DTT, 25 mg/ml fatty acidfree BSA, 8.04μ M cholesteryl [1-14C] oleate (2.105 dpm, 1.61 nmol), 177.5 μg/ml phosphatidylcholine/phosphatidylinisitol (3:1, wt/wt) and freshly sonicated cytosolic cell extract (∼4 mg protein/ml, 100 μl). After incubation for 60 minutes at 37°C, the [14C] oleate released was extracted by the addition of 3.25 ml of the mixture methanol:chloroform:heptane (10:9:7) containing non labeled oleic acid as a carrier (3 g/l) and 1.05 ml of 0.1 M potassium carbonate, 0.1 M boric acid, pH 10.5 (Belfrage and Vaughan, 1969). The reaction tubes were shaken vigorously for 1 minute and centrifuged at 1500 g for 10 minutes at room temperature. Radioactivity of the upper layer was counted. One unit of enzyme hydrolyzes 1μ mol of cholesterol ester in 1 minute. Results are expressed as a percentage of the control without HDL3.
Oregon-HDL3 binding and association assays
Cells were incubated for 3 hours at 37°C in 0.5 ml of DMEM supplemented with antibiotics and 5% LPDS containing the indicated final concentrations of native or tyrosylated Oregon-HDL3 or with 25 μg/ml of native or tyrosylated Oregon-HDL3 for the indicated times. To determine the level of non-specific binding, cells and Oregon-HDL3 (10 μg/ml to 50 μg Apo A-I/ml) were incubated with a 50-fold excess of the corresponding unlabeled HDL3. After incubation, cells were washed twice with cold DMEM-5% LPDS and two times with cold PBS, detached from the plate by gentle pipetting and immediately subjected to fluorescence flow cytometry using a FACScan from Becton Dickinson. The forward-angle light-scatter gates were established to exclude debris and cellular aggregation. At least 10,000 cells were analyzed for each sample. Cell-associated and bound lipoprotein were expressed as a median intensity of fluorescence (MIF) assessed using fluorescence windows (channel numbers). Values for Oregon-HDL3 binding and association were obtained by subtracting non-specific background. For inhibition experiments, unlabeled lipoproteins, polyinosinic acid (poly-I) and the labeled-HDL3 were added simultaneously to the cells. A blocking anti-scavenger receptor BI and BII polyclonal antibody was preincubated for 30 minutes with cells before the addition of labeled-HDL3. Rabbit IgG was used at the same concentration as a negative control. Experiments were done several times using different lipoprotein preparations.
Following the incubation of cells for 3 and 6 hours with 100 μg apo A-I/ml (Oregon-HDL3) at 37°C, the coverslips were rinsed, fixed for 15 minutes with 3% formaldehyde and photographed using a fluorescence microscope equipped with a fluorescein isothiocyanate filter set.
Cells were incubated for 3 hours at 37°C in 0.5 ml of DMEM supplemented with antibiotics and 5% LPDS containing the indicated final concentrations of native or tyrosylated [3H]cholesteryl hexadecyl ether-HDL3 or with 25 μg apoA-I/ml of native or tyrosylated [3H]cholesteryl hexadecyl ether-HDL3 for the indicated times. To determine the non-specific binding, cells and labeled-HDL3 (10-50 μg/ml of apoA-I) were incubated in the presence of a 50-fold excess of the corresponding unlabeled HDL3. After incubation, cells were washed twice with cold DMEM containing 5% LPDS and twice with cold PBS, detached from the plate by gentle pipetting with 1 ml of distillated water containing 0.25% Triton and homogenized by sonication on ice (3×10 seconds, lowest power setting). For each well, radioactivity was counted, and protein was determined by Bio-Rad protein assay kit with BSA as a standard. Results are expressed in DPM/mg cell protein. The specific uptake of native or tyrosylated [3H]cholesteryl hexadecyl ether-HDL3 by cells was calculated by subtracting the non-specific value. For inhibition experiments, unlabeled lipoproteins, Poly-I and the labeled-HDL3 were added to the cells. Only the blocking anti-scavenger receptor BI and BII polyclonal antibody was preincubated for 30 minutes with cells before the addition of labeled HDL3. Experiments were done several times using different lipoprotein preparations.
SDS-polyacrylamide gel electrophoresis
The apolipoprotein content of nascent and tyrosylated HDL was examined by electrophoresis on sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE), as described previously (Levy et al., 1994). The gels were stained with Coomassie blue and destained in 70% acetic acid.
Data from the experiments were analyzed by using a student's t-test. Reported values are expressed as means±s.e. Statistical significance was accepted at P<0.05.
Physicochemical features of normal and tyrosylated HDL fractions
Before assessing the ability of tyrosylated HDL to extract cholesterol from J774-A1 macrophages, we first examined the structural modifications of HDL induced by tyrosyl radical oxidation. As illustrated in Fig. 1A, HDL developed a maximal fluorescent emission following its exposure to horseradish peroxidase, H2O2 and L-tyrosine at a 410 nm wavelength characteristic of dityrosine. About 2.5-5% of the tyrosyl residues in peroxidase-modified HDL were converted to dityrosine, which represented 90% of the fluorescence. Furthermore, a modest increase in apparent apolipoprotein molecular masses was revealed by SDS-PAGE in non-reducing conditions (Fig. 1B), which probably was generated by the reactive tyrosyl-apolipoprotein radical crosslinking. The apparent molecular masses of 47 kDa, 61 kDa and 83 kDa probably resulted from the formation of apoA-I-(apoA-II)2 complexes as well as from apoA-I trimers. One could not detect any significant alterations in the chemical composition of control and tyrosylated HDL on the basis of the relative distribution of their components (triglycerides, free and esterified cholesterol, phospholipids and proteins) (data not shown). Similarly, no aggregation was noted in the total HDL fraction and HDL3 subfraction following tyrosylation treatment. Finally, only limited amounts of TBARS, a marker of lipid peroxidation, were detectable in tyrosylated HDL (0.7 nmol/mg apo A-I) and native HDL (0.55 nmol/mg apo A-I).
Evidence suggests that HDL promotes cholesterol efflux from cells, which constitutes the rate-limiting step in the reverse cholesterol pathway (Pieters et al., 1991; Levy et al., 1996). Thus, in the next series of experiments, we assessed the capacity of tyrosylated HDL for cholesterol removal from J774-A1 macrophages pre-loaded with radiolabeled AcLDL-[3H]cholesteryl oleyl ester in 5% LPDS-DMEM for 24 hours. The findings shown in Fig. 2 indicate that tyrosine modification resulted in impaired tyrosylated HDL3- or total HDL-mediated cholesterol efflux from J774-A1 cells. It was evident that at all concentrations between 50-200 μg apoA-I/ml a diminished efflux occurred regardless of whether the tyrosylated total HDL (Fig. 2A) or HDL3 (Fig. 2B) subfraction was used as the cholesterol acceptor. Starting from 11.2±1.4% (control values in the absence of lipoprotein acceptors), the maximal efficacy was obtained at nearly 100 μg/ml of apoA-I for both native (28.86±0.51%) and tyrosylated (21.14±0.84%) HDL3. For the same apoA-I concentration, native and tyrosylated total HDL appeared to extract more cholesterol. Calculated from the specific activity of cholesterol loading of the cells, the difference in the efflux rate between native and modified total HDL at 100 μg apoA-I/ml was 0.12 pmol cholesterol efflux per 24 hours, whereas that between native and tyrosylated HDL3 was 0.17 pmol cholesterol efflux per 24 hours. Therefore, only the HDL3 fraction was used for the following studies. From these experiments, we concluded that the capacity of tyrosylated HDL for cholesterol efflux from J774-A1 macrophages was weaker than that of native HDL.
In our experiments, both free (FC) and esterified (EC) cholesterol were measured in the medium for efflux studies. The average ratio of CE/FC was 0.21 for controls (absence of lipoprotein acceptors), 0.15 for native HDL and 0.07 for tyrosylated HDL.
HDL3 cell association
Since interaction of apolipoprotein with specific membrane receptors is required for the mobilization of cellular cholesteryl ester, we verified whether tyrosylation altered HDL3 recognition by J774-A1 macrophage membrane-binding domains. Indeed, defective interaction between tyrosylated HDL3 and cells could affect the effluxing capacity of HDL3. To this end, we covalently labeled apolipoprotein moieties of native and tyrosylated HDL3 with the fluorescent dye Oregon Green 514 to study HDL3 cell association by flow cytometry, after cellular cholesterol loading with AcLDL. Fig. 3A represents the results of the time course studies carried out with J774-A1 macrophages incubated with 25 μg apoA-I/ml of Oregon-HDL3. Under our experimental conditions, the estimation of HDL3 cell association by the Oregon technique revealed an increased association of tyrosylated HDL3 with cells compared with native HDL3 at the time points analyzed. In fact, two- to fourfold enhancement occurred at all incubation periods. Cell association was also tested as a function of increasing HDL3 concentrations (Fig. 3B) and found to be directly proportional to Oregon intensity. Again, similar results were obtained showing that tyrosylated HDL3 cell association was two- to threeold increased in comparison with native HDL3, which confirmed the pronounced association of tyrosylated particles with macrophages. In addition, fluorescence microscopy was used to visualize the association of Oregon-HDL3 with J774-A1 cells after the incubation periods of 3 and 6 hours and confirmed the augmented association of tyrosylated HDL3 monitored with flow cytometry. Therefore, tyrosylation did not impair the cell-tyrosylated HDL3 interaction.
HDL3 cholesteryl ether cell uptake
With regard to these data, cholesterol ether-HDL3 uptake was monitored in a dose- and time-dependent manner. Cells were incubated with 25μ g apoA-I/ml of [3H]cholesteryl hexadecyl ether-HDL3 for up to 300 minutes at 37°C (Fig. 4A) or for 3 hours with increasing concentrations of the labeled HDL3 (Fig. 4B). As for the Oregon studies, the uptake of radiolabeled tyrosylated HDL3 was more pronounced than that of native [3H]cholesterol ether-HDL3 (Fig. 4). We concluded that, in contrast to its reduced ability to remove cholesterol from cholesterol-enriched cells, tyrosylated HDL3 displayed both a more efficient interaction with and more cholesteryl ether transfer to J774-A1 macrophages than native HDL3. Therefore, low ability to extract cholesterol could not be attributed to defective cell association.
nCEH activity in the presence of HDL3
Previous evidence supported an association between nCEH activity and cholesterol ester accumulation in macrophages. Therefore, we intended to determine whether the limited cholesterol efflux by tyrosylated HDL3 could be accounted for by a corresponding reduction in the activity of nCEH, the main enzyme responsible for the bulk hydrolysis of cytosolic CE droplets (Contreras and Lasuncion, 1994; Graham et al., 1996). For this purpose, experiments were carried out under similar conditions of cholesterol efflux measurement. Cells were first cholesterol loaded with AcLDL-[3H]cholesteryl oleyl ester before their incubation for 24 hours in the presence or absence of native or tyrosylated HDL3 (0-200 μg apoA-I/ml). Then, nCEH activity in cell homogenate supernatant was assayed. As seen in Fig. 5, the incubation of J774-A1 macrophages with tyrosylated HDL3 resulted in lower nCEH activation compared with native HDL3. Furthermore, no significant differences were observed when nCEH activity was measured in the presence or absence of tyrosylated HDL3, indicating that the latter did not enhance nCEH activity.
HDL3 cell association competition studies
By which mechanisms does HDL3 interact with J774-A1? In order to answer this question, we first defined the ability of lipoproteins to compete for native or tyrosylated Oregon-labeled HDL3 cell association. To this end, 25 μg apoA-I/ml of Oregon-HDL3 was co-incubated with increasing concentrations of unlabeled native HDL3, tyrosylated HDL3, AcLDL or oxLDL for 3 hours at 37°C. The non-specific binding was obtained by the addition of a 50-fold excess of corresponding unlabeled HDL3 and was subtracted from the other values. As shown in Fig. 6, native HDL3 cell association was markedly (70 and 90%) inhibited by unlabeled native HDL3 (125 and 250 μg apoA-I/ml, respectively). A complete inhibition (98%) was reached with 375 μg apoA-I/ml. A similar pattern was observed with modified lipoproteins in decreasing competition efficiency order: tyrosylated HDL>oxLDL>AcLDL. Tyrosylated Oregon-HDL3 was also incubated with increasing concentrations of unlabeled native HDL, tyrosylated HDL3, AcLDL or oxLDL for 3 hours at 37°C. The most competitive effect was obtained with AcLDL: 49, 94 and 100% at 125, 250 and 375 μg apoA-I/ml, respectively. The competition efficiency declined from AcLDL to native HDL3 and OxLDL. Additionally, we noted in few studies that native LDL (at the concentrations of 250-375 μg apo B/ml) was able to compete with tyrosylated HDL and to inhibit its cell association by 30-40%. These findings suggest that (1) tyrosylated HDL3 efficiently interacts with cell-surface receptors for native HDL3 and modified LDL and (2) the involved receptors have a broad ligand specificity.
HDL3-cholesteryl ether cell uptake competition studies
In a second step, we evaluated the transfer of [3H]cholesteryl ether-labeled native and tyrosylated HDL3 (10 μg/ml apoA-I) co-incubated with 50, 100 and 150 μg/ml apoA-I of unlabeled native HDL3 and tyrosylated HDL3, as well as with AcLDL or OxLDL for 3 hours at 37°C. As illustrated in Fig. 7, a decrease in the cellular uptake of labeled cholesteryl ether was noted when the lipoproteins were incubated with native HDL3 (Fig. 7A) or tyrosylated (Fig. 7B). Only tyrosylated HDL was unable to compete with native HDL3. Overall, our data confirm the cell association competition studies and suggest common binding site(s) for cholesterol deposition in macrophages.
Identification of tyrosylated HDL3 cell receptor
We attempted to define the mechanism of receptor-mediated cell association with and lipid uptake from tyrosylated HDL3. We first examined the class B type I/II scavenger receptors (SR-BI/BII) that are expressed in various mammalian cells, bind HDL with high affinity and also serve as receptors for modified lipoproteins (Acton et al., 1996; Ji et al., 1997). Using an antibody raised against the residues 230-380 of the extracellular domain of SR-BI/BII, we could detect about 50% inhibition of cell association with native or tyrosylated Oregon-HDL3 (Fig. 8A). Similarly, SR-BI/BII-specific blocking antibody profoundly inhibited the cellular uptake of native and tyrosylated [3H]cholesteryl ether-HDL3 (Fig. 8B). These studies, therefore, support the suggestion that the association and the CE internalization of tyrosylated HDL3 are substantially mediated by SR-BI/BII.
Since class A scavenger receptors are mainly expressed in macrophages and recognize and internalize modified lipoproteins, we assessed their contribution to cell association and cholesterol ester uptake of tyrosylated HDL3. Thus, we employed polyinosinic acid (poly-I), an established inhibitor of class A scavenger receptors, that bind tightly to SR-AI/II because it is capable of forming base quarted-stabilized four-stranded helices. J774-A1 cells were incubated for 3 hours at 37°C with 10 μg apoA-I/ml of native or tyrosylated Oregon HDL3 (Fig. 9A) or with native or tyrosylated [3H]cholesterol ether-HDL3 (Fig. 9B) in the presence of increasing concentrations of Poly-I. As shown in Fig. 9A, this ligand slightly reduced the cell association with native Oregon-HDL3 (20% of control), whereas that of tyrosylated Oregon-HDL3 was dramatically hampered (∼50-60% of control). By contrast, Poly-I enhanced [3H]cholesterol ether cell uptake from native HDL3, whereas tyrosylated [3H]cholesterol ether-HDL3 cell uptake was inhibited by Poly-I in a concentration-dependent manner (Fig. 9B). These findings indicated clearly that type I and II class A scavenger receptors mediated macrophage binding to and cholesterol ester uptake from tyrosylated HDL3.
High plasma levels of HDL are associated with a decreased risk of developing atherosclerosis, an effect commonly attributed to their central role in reverse cholesterol transport (Eisenberg, 1984; Fielding and Fielding, 1995). In this process, the HDL fraction clears excess cholesterol from peripheral cells and delivers it to the liver where cholesterol removal occurs via bile acid synthesis and biliary secretion (Levy et al., 1996; Pieters et al., 1991; Terpstra et al., 1989). Since oxidative modification can occur in HDL, as in LDL (Hahn and Subbiah, 1994; Quinn et al., 1987; Salmon et al., 1992; Steinberg et al., 1989), several investigators attempted to examine the ability of modified HDL to mediate cellular cholesterol efflux. In most of the studies, oxidation or modification of HDL resulted in diminished cholesterol efflux (Marcel et al., 1989; Mazière et al., 1993; Salmon et al., 1992). We, therefore, initiated a comparative study of the efficiency of native and tyrosylated HDL (devoid of lipid peroxidation) in promoting cholesterol efflux from J774-A1 macrophages with special respect to the mechanisms involved. Our kinetic studies clearly established that tyrosylated HDL was much less capable of removing cholesterol from macrophages. A number of facts add strength to this finding: one, in line with our observations, oxidative tyrosylation of HDL resulted in a loss of the hepatobiliary stem capacity to normally process this modified HDL in vivo (Guertin et al., 1997); two, less radioactivity was found in bile and bile acids following the intravenous injection of tyrosylated HDL-[14C]cholesterol in rats, suggesting abnormal reverse cholesterol (Guertin et al., 1997); and three, there was evidence that high levels of protein-bound dityrosine along with myeloperoxidase were capable of generating tyrosyl radicals in human atherosclerotic lesions (Daugherty et al., 1994; Leeuwenburgh et al., 1997; Mazière et al., 1993).
J774 cells represent a remarkable macrophage model that are frequently employed for the investigation of cholesterol metabolism. They have largely been used to study cholesterol efflux mechanisms triggered by HDL particles and their apolipoprotein content. Indeed, this cell line does not express apo E (Langer et al., 2000), nor caveolin-1 (Matveev et al., 1999), which constitutes an evident advantage for the assessment of cholesterol efflux by native or tyrosylated HDL given that apo E may interfere and disguise the effects of tyrosalated apo A-I by compensatory mechanisms
Previously, Francis et al. reported that oxidative tyrosylation of HDL enhanced cholesterol removal from cultured fibroblasts and macrophage foam cells (Francis et al., 1993). Possible explanations for the discrepancy between our observations and those of Francis et al. include the use of different cell lines, the degree of oxidative tyrosylation that can produce intermolecular dytirosine crosslinks of apo A-I and apo A-II as well as the size, density, apolipoprotein composition and lipid content of the lipoprotein acceptor, which were shown to independently affect cholesterol release (Johnson et al., 1991). Our data are consistent with numerous studies reporting a lessened effect of oxidized HDL on cholesteryl ester depletion in foam cells by various agents, such as hypochlorite (Bergt et al., 1999), malondialdehyde (Guertin et al., 1994; Salmon et al., 1992), Cu2+ (Gesquière et al., 1997) or even cell-HDL association (Cogny et al., 1996).
We attempted to elucidate the mechanisms linked to the impaired cholesterol efflux abilities of tyrosylated HDL. One hypothesis raised in our studies was that tyrosylation might be involved in reducing the binding affinity of HDL, resulting in the reduced capacity of the particle to remove cholesterol from J774-AI macrophages. However, tyrosylation of HDL particles did not compromise their association nor their CE transfer to J774-A1 macrophages. This indicates that the reduction in cholesterol efflux was not due to the disruption of tyrosylated HDL interaction with regulatory domains of plasma membrane lipids or critical receptors. On the contrary, our observations unambiguously demonstrated that the interaction between cells and tyrosylated HDL, as well as the uptake of tyrosylated HDL by J774-A1 macrophages, is increased. In the second step, it was reasonably assumed that the hydrolysis and mobilization of intracellular cholesteryl ester stores, leading to cellular cholesterol efflux through an HDL cell-surface receptor (Oram et al., 1991; Slotte et al., 1987), were affected. We thus focused on neutral cholesterol esterase involved in the degradation of CE, since several investigators have suggested that the mobilization of CE in foam cells by HDL is mediated by this cytosolic enzyme (Contreas and Lasuncion, 1994; Goldstein et al., 1983). Indeed, we found that incubation with native HDL raised nCEH activity in J774-A1 cells. The post stimulation values were slightly lower than those described before by Miura et al. (Miura et al., 1997), probably because J774-AI cells were cholesterol loaded with AcLDL prior to the addition of HDL3. Accordingly, an increase in intracellular cholesterol was previously shown to downregulate nCEH (Tomita et al., 1997). On the other hand, the exposure of J774-A1 macrophages to tyrosylated HDL resulted in reduced nCEH activity, which may be related to the decrease in cholesterol efflux. Since CE stored in the cytoplasm undergoes a continual cycle of hydrolysis and re-esterification, additional studies are necessary to assess the contribution of acyl CoA:cholesterol acyltransferase in the slow desorption of cholesterol from the plasma membrane by tyrosylated HDL.
The accumulation of oxidised LDL has been related to the transformation of monocyte-derived macrophages into foam cells (Brown and Goldstein, 1984; Steinberg et al., 1989). The uptake of these modified LDLs is mediated by scavenger receptors (SR) located on the surface of macrophages (Acton et al., 1996; Ji et al., 1997; Stangl et al., 1998). The major function of SR-BI is thought to allow cholesterol transport, and in particular the selective uptake of HDL cholesterol ester by the liver and steroidogenic tissues, without having to internalize the entire HDL macromolecule (Acton et al., 1996; Ji et al., 1997; Rigotti et al., 1995). Our results clearly demonstrated that, besides recognizing native HDL and oxidatively modified LDL, J774-A1 macrophage cells interacted with tyrosylated HDL particles. The latter were shown to be effective competitors of scavenger-receptor-mediated binding of native HDL as well as acetylated and oxidized LDL, all ligands ascribed to SR-BI/BII (Acton et al., 1996; Ji et al., 1997; Rigotti et al., 1995; Stangl et al., 1998). Direct evidence for SR-BI/II involvement is provided by our in vitro experiments in which the antibody to the extracellular domain of SR-BI/II blocked HDL-CE uptake and delivery to cultured J774-A1 cells. In view of the comparable competence of SR-BI/BII to bind native and tyrosylated HDL3, noted with the specific antibody, it is possible to exclude its participation in defective cholesterol efflux in J774-A1 cells.
With regard to the scientific literature, SR-BI recognizes native and modified LDL in addition to HDL (Acton et al., 1996; Ji et al., 1997; Rigotti et al., 1995). However, acetylated LDL are preferentially captured by SR-AI/AII, whereas oxidized LDL are probably taken by SR-BI (Lougheed et al., 1997; Lougheed et al., 1999). Furthermore, according to Murao et al. and Fluiter et al., the affinity of SR-BI respects the following pattern HDL>LDLox>LDL>LDLac (Murao et al., 1999; Fluiter et al., 1997). The data in Fig. 8 fit this profile especially when we consider the important role of SR-BI in the recognition of native HDL (Fig. 8).
SR-AI/AII was originally identified on the basis of its ability to mediate the internalization of chemically modified LDL and cause massive lipid accumulation in macrophages (Brown et al., 1980; Goldstein et al., 1979). It is present on the surface of macrophages, Kupffer cells and endothelial cells. Polynucleotides can bind tightly to the SR-AI/II given their ability to form base quartet-stabilized four-stranded helices (Pearson et al., 1993). Using polyionsinic acid in our in vitro experiments, we observed that cell association with tyrosylated HDL was reduced by 50% compared with that of native HDL. Similar results were obtained with regard to [3H]cholesteryl ether uptake from native and tyrosylated HDL.
These data suggest that SR-AI/II may be responsible for the enhanced uptake of CE from tyrosylated HDL. Since the main function ascribed to the SR-AI/II involves promoting the clearance of microbial pathogens, senescent cells or altered plasma proteins by phagocytic cells (Pearson et al., 1993; Pearson, 1996; Platt et al., 1996; Terpstra et al., 1997), it is possible that the modification of apos A-I and A-II by tyrosylation makes them more suitable for recognition and internalization by SR-AI/AII. Moreover, the affinity of SR-AI/AII for tyrosylated HDL would favor cholesterol uptake, which could explain the decreased cholesterol efflux by these lipoprotein particles. Finally, there may be additional, as yet unidentified, cell-surface molecules that bind to tyrosylated HDL and underlie some of the actions mentioned in this study.
In conclusion, the current data highlight the efficiency of tyrosylated HDL for competing with oxidatively modified LDL, binding to scavenger receptors AI/II and BI/II and transferring their cholesterol moiety to J774-A1 macrophages. By contrast, tyrosylated HDL particles were less efficient in effluxing cholesterol and more competent in accumulating cholesterol in J774-A1 cells, two processes that were accompanied by diminished nCEH activity and enhanced SR-AI/II recognition and cholesterol ester uptake.
We wish to thank Louise Brissette for helpful discussions and Schohraya Spahis for assistance in the preparation of the manuscript. This work was supported by the MRC and NSERC grants.
- Accepted August 28, 2002.
- © The Company of Biologists Limited 2003