High Affinity Saturable Uptake of Oxidized Low Density Lipoprotein by Macrophages from Mice Lacking the Scavenger Receptor Class A Type I/II*

Oxidation of low density lipoproteins (LDL) has been implicated as a causal factor in the pathogenesis of atherosclerosis. Oxidized LDL has been found to exhibit numerous potentially atherogenic properties in vitro, including receptor-mediated uptake by macrophages. Oxidized LDL is a ligand for the class A scavenger receptor type I/II (SR-AI/II), but cross-competition studies with cultured macrophages suggested that there is an additional receptor(s) that is specific for oxidized LDL and that does not interact with acetyl LDL or other chemically modified LDL. A number of macrophage membrane proteins, including CD36, FcγRII-B2, scavenger receptor BI, and macrosialin/CD68, have been found to bind to oxidized LDL in vitro and have been proposed as candidate oxidized LDL receptors. However, because of overlapping ligand specificity with the SR-AI/II, it has been difficult to evaluate the relative importance of these proteins in the uptake of oxidized LDL by macrophages. In the present report, we have studied the uptake and degradation of oxidized LDL by macrophages from mice in which the SR-AI/II gene had been disrupted. The uptake of acetyl LDL was reduced by more than 80% in macrophages from scavenger receptor knockout mice, confirming that most of the uptake of acetyl LDL by macrophages can be attributed to this receptor. In contrast, the uptake of extensively oxidized LDL was reduced by only 30% and showed high affinity, saturable uptake with apparent K m of about 5 μg/ml, similar to that of the SR-AI/II. This indicates that about 70% of the uptake of oxidized LDL in macrophages is attributable to an alternate oxidized LDL receptor(s). In contrast to findings reported with CD36, mildly oxidized LDL was internalized much more slowly than extensively oxidized LDL. Unlabeled oxidized LDL, polyinosinic acid, phosphatidylserine-rich liposomes, and LDL or bovine albumin modified by fatty acid oxidation products were effective competitors for the uptake of radioiodinated oxidized LDL by macrophages from knockout mice, whereas acetyl LDL and malondialdehyde-modified LDL were relatively poor competitors. This ligand specificity differs from that of CD36-related (class B) scavenger receptors but is similar to the reported specificity of macrosialin/CD68 in ligand blots. However, the rate of uptake of oxidized LDL by knockout macrophages was not increased by phorbol ester or in thioglycollate-elicited macrophages, both of which are expected to increase the amount of macrosialin on the cell surface. In macrophages from SR-AI/II knockout mice, ligand blots of membrane proteins with iodinated, oxidized, or acetylated LDL revealed several bands, with apparent molecular size on SDS-polyacrylamide gel electrophoresis of 60, 94, 124, and 210 kDa, but none of the bands were specific for oxidized LDL. These results provide direct evidence that a receptor other than SR-AI/II is responsible for most of the uptake of oxidized LDL in murine macrophages, but further studies are needed to identify the receptor(s) involved.

ported specificity of macrosialin/CD68 in ligand blots. However, the rate of uptake of oxidized LDL by knockout macrophages was not increased by phorbol ester or in thioglycollate-elicited macrophages, both of which are expected to increase the amount of macrosialin on the cell surface. In macrophages from SR-AI/II knockout mice, ligand blots of membrane proteins with iodinated, oxidized, or acetylated LDL revealed several bands, with apparent molecular size on SDS-polyacrylamide gel electrophoresis of 60, 94, 124, and 210 kDa, but none of the bands were specific for oxidized LDL. These results provide direct evidence that a receptor other than SR-AI/II is responsible for most of the uptake of oxidized LDL in murine macrophages, but further studies are needed to identify the receptor(s) involved.
One of the properties that distinguishes oxidized LDL from native LDL is the ability to bind to "scavenger" receptors on macrophages (34,35). It has been postulated that this could lead to unregulated cholesterol accumulation in macrophages, resulting in the formation of foam cells (1,36). The term scavenger receptor has been applied to a number of different cellsurface proteins that can bind to chemically or biologically modified lipoproteins and/or senescent cells. The first scavenger receptor to be fully characterized is termed the class A type I/type II scavenger receptor, or SR-AI/II (37). Two forms of the receptor are produced from a single gene through alternative splicing of mRNA (38,39). The SR-AI has six structural domains, including a collagen-like domain (V) that mediates ligand binding (40). The cysteine-rich extracellular domain VI is deleted in the type II receptor, without evident effect on binding of most ligands (41). Various negatively charged substances including polyinosinic acid, maleylated albumin, and acetyl LDL can bind to this receptor, suggesting that ligand binding may be mediated principally by ionic interactions (40,42). It has been shown that a particular conformation of charges is required for receptor binding, because only polynucleotides that are capable of forming base-quartet-stabilized helices act as ligands (43).
Several groups have reported evidence of a receptor for oxidatively modified LDL distinct from the SR-AI/II on macrophages (44,45) or Kupffer cells (46 -48). In several of these studies, the evidence for multiple receptors was the finding that acetyl LDL was able to compete for only about 40% of oxidized LDL uptake (44 -47). More direct evidence for a separate receptor for oxidized LDL was the finding that in rats, intravenously injected oxidized LDL was cleared preferentially by Kupffer cells, whereas acetyl LDL was cleared by sinusoidal endothelial cells (46). Endemann and co-workers (49, 50) used a mouse cDNA library and an expression cloning strategy with fluorescently labeled oxidized LDL to identify two cell surface proteins that could mediate the binding and internalization of oxidized LDL but not acetyl LDL. The first was found to be the murine Fc␥II-B2 receptor, but blocking antibodies to the Fc␥II-B2 receptor failed to inhibit the uptake of oxidized LDL in mouse peritoneal macrophages; hence, it is unlikely that this receptor contributes significantly to oxidized LDL uptake in these cells (49). The second protein was shown to be the murine homologue of CD36 (50). It was found that only a mild degree of oxidation of LDL was required for maximal binding to the murine homologue of CD36, whereas extensive oxidation of LDL is required for receptor-mediated uptake and degradation in mouse peritoneal macrophages (35). This is indirect evidence that CD36 does not account for the SR-AI/II-independent uptake of oxidized LDL in murine macrophages. However, in human macrophages, anti-CD36 antibody has been shown to partly inhibit the degradation of oxidized LDL (50,51), and monocyte-macrophages from subjects with inherited deficiency of CD36 were found to have a reduced rate of uptake and degradation of oxidized LDL (52). Hence, it is possible that CD36 is a mediator of oxidized LDL uptake and degradation in human cells. Acton et al. (53) identified a scavenger receptor (termed SR-BI) homologous to CD36 that binds oxidized or acetylated LDL but also native LDL (53). It also binds high density lipoprotein, and it has been proposed that one function of SR-BI may be to promote adhesion of high density lipoprotein to cells and facilitate cholesterol ester transfer (54). More recently, a 94 -97-kDa macrophage plasma membrane protein identified by ligand blotting has been proposed to be a specific receptor for oxidized LDL (55,56). This protein has been identified as macrosialin, the murine homologue of CD68 (57). Although macrosialin is targeted to lysosomes and only a small percentage of the total appears on the plasma membrane, it is so abundantly expressed in macrophages that a significant number of molecules are detectable on the cell surface (58). Antibodies to CD68 partially inhibited the uptake of oxidized LDL by THP-1 cells (58).
Evaluation of the role of these candidate receptors for oxidized LDL is rendered difficult because of the presence of scavenger receptors in macrophages, because any effect of these postulated "specific" receptors for oxidized LDL is observed against a high background level of uptake via the SR-AI/II. To date, no normal cell type has been identified that expresses oxidized LDL receptors and not SR-AI/II. However, it has been suggested that Kupffer cells may exhibit a relatively high level of "oxidized LDL receptor" activity. Nevertheless, these cells also have scavenger receptors, and the results of competition experiments remain somewhat equivocal. Because of this "interference" by scavenger receptors, identification of specific oxidized LDL receptors would be greatly facilitated in animals where scavenger receptors were absent. Suzuki et al. (59) have recently generated transgenic mice in which the SR-AI/II (msr) gene was inactivated by targeted gene disruption. The homozygous mutants have complete absence of SR-AI/II protein, but their phenotype is virtually normal. The present studies were done to characterize the uptake and degradation of modified lipoproteins by macrophages from these SR-AI/II-deficient animals.

MATERIALS AND METHODS
Carrier-free 125 I and 32 P were purchased from Mandel Scientific (Guelph, Ontario, Canada). Gentamicin, ␣-minimal essential medium, ultra pure agarose, proteinase K, ethidium bromide, BglII, XbaI, 1-kb DNA ladder, and high molecular weight DNA markers were from Canadian Life Technologies (Burlington, Ontario, Canada). Hyclone-defined fetal bovine serum and Stratagene NucTrap probe purification columns were purchased from Professional Diagnostics (Edmonton, Alberta, Canada). Arachidonic acid, linoleic acid, bovine serum albumin, n-octyl ␤-D-glucopyranoside, phorbol myristate acetate, cholesterol, tetramethoxypropane and phenylmethylsulfonyl fluoride were obtained from Sigma. Phosphatidylserine and phosphatidylcholine were from Avanti Polar Lipids (Birmingham, AL). Nitrocellulose membrane and prestained SDS-polyacrylamide gel electrophoresis standards were purchased from Bio-Rad. Hybond-N hybridization transfer membrane was obtained from Amersham Corp. Kodak X-Omat AR film was from MedTec Marketing (Richmond, British Columbia, Canada). Other reagents were purchased from Fisher Scientific (Vancouver, British Columbia, Canada) or VWR Canlab (Edmonton, Alberta, Canada).
Animals-The procedure for disruption of the msr gene (which encodes SR-AI/II) has been reported recently (59). A targeting vector was introduced into exon 4 (the ligand-binding region) of the gene. Brothersister mating of heterozygotes was carried out to generate homozygous mutants. Normal littermates were bred as controls. To verify that there had been no accidental intermingling of wild-type and homozygous mice during the study, genotyping of selected experimental animals throughout the study as well as of all breeders was performed by Southern blotting using a 600-base pair probe from a region adjacent to exon 4 of the msr gene. Digestion of genomic DNA with BglII generated a single 23-kb band in wild-type mice and a single 10-kb band in homozygous knockout mice. None of the mice showed both the 23-kb band and the 10-kb band characteristic of heterozygotes.
Lipoprotein Isolation and Labeling-LDL (d ϭ 1.019 Ϫ 1.063) was isolated by sequential ultracentrifugation of EDTA-anticoagulated fasting plasma obtained from healthy normolipidemic volunteers (60). Radioiodination was performed using a modification of the iodine monochloride method of MacFarlane (61). Specific radioactivities were 100 -250 cpm/ng. Iodination was performed prior to modification of LDL.
Lipoprotein Modification-The concentration of EDTA in LDL preparations was reduced prior to oxidation by dialysis against Dulbecco's phosphate-buffered saline containing 10 M EDTA. Standard conditions for LDL oxidation were: 200 g/ml LDL in Dulbecco's phosphatebuffered saline containing 5 M CuSO 4 incubated at 37°C for 20 h (62). This typically resulted in electrophoretic mobility of 0.85 relative to albumin. For some experiments, the extent of oxidation was controlled by varying the incubation time. Mildly oxidized LDL was prepared by incubating for 3.5 h and resulted in electrophoretic mobility 0.45 relative to albumin. Extensively oxidized LDL was prepared by incubating for 30 h and resulted in electrophoretic mobility of 1.06 relative to albumin. Acetylation, malondialdehyde modification, arachidonic acid oxidation product modification or linoleic acid oxidation product modification of LDL, or bovine serum albumin was performed as described previously (63).
Assays with Cultured Macrophages-Resident peritoneal macrophages were obtained from wild-type mice or SR-AI/II knockout mice by peritoneal lavage with ice-cold Ca 2ϩ -free Dulbecco's phosphate-buffered saline. Cells were suspended in ␣-minimal essential medium with 10% fetal bovine serum and plated in 12-well plastic culture plates at a density of 1 ϫ 10 6 cells/well. Adherent macrophages were cultured overnight in a humidified CO 2 incubator and then washed with serumfree ␣-minimal essential medium. Modified LDLs with or without competitors were added to the cells in ␣-minimal essential medium supplemented with 2.5 mg/ml lipoprotein-deficient serum to minimize cytotoxicity. For studies with polycytidylic acid, 10 units/ml recombinant ribonuclease inhibitor (Promega Corp.) was added to prevent degradation of this polynucleotide (43). After 5 h incubation at 37°C, media were removed and assayed for trichloroacetic acid-soluble noniodide degradation products. Cells were washed twice with Dulbecco's phosphate-buffered saline, dissolved in 0.1 N NaOH, scraped from the plates, counted, and assayed for protein content.
Liposome Preparation-To obtain phosphatidylserine liposomes, 12 mol of bovine brain phosphatidylserine, 12 mol of egg phosphatidylcholine, and 12 mol of cholesterol were suspended in 1.5 ml of 150 mM NaCl, 10 mM Hepes, and 0.1 mM EDTA, pH 7.5. The mixture was passed 10 times though a 0.1-m polycarbonate membrane using a Lipex mini-extruder. Phosphatidylcholine liposomes were prepared with the same procedure, except that 20 mol of egg phosphatidylcholine and 10 mol of cholesterol were used.
Macrophage Membrane Preparation-Resident peritoneal macrophages from 35-40 mice were suspended in cold 5 mM Tris-HCl, pH 8.0, containing 0.25 M sucrose and 1 mM phenylmethylsulfonyl fluoride and disrupted by two cycles of nitrogen cavitation at 460 kPa for 30 min. Nuclei were pelleted by centrifugation at 1000 ϫ g for 10 min at 4°C. The suspension was centrifuged at 10,000 ϫ g for 15 min, and the resulting supernatant was centrifuged at 100,000 ϫ g for 120 min at 4°C. The crude membrane pellet was solubilized in 100 l of 10 mg/ml octyl glucoside containing 1 mM phenylmethylsulfonyl fluoride and 0.1 mM EDTA and assayed for protein content. Typical yields of macrophage membrane protein were 0.3-1 mg. Solubilized membrane was stored at 4°C and used within 1 day for ligand blotting.
Ligand Blotting-Ligand blotting was performed according to Ottnad et al. (55). Crude membrane proteins were electrophoresed in SDS/8% polyacrylamide mini gels under nonreducing conditions and electroblotted onto nitrocellulose membranes. Membrane strips were preincubated for 1 h with 5% nonfat dry milk in 10 mM Tris-HCl, 90 mM NaCl, and 1 mM EDTA, and then 10 g/ml of native or modified 125 I-labeled LDL (150 -250 cpm/ng) were added for an additional 1.5 h. The nitrocellulose strips were washed eight times for 10 min with buffer containing 1% nonfat dry milk, air-dried, and exposed to Kodak X-Omat AR film for 10 days at Ϫ70°C.
Analytic Procedures-Protein determination was done by the method of Lowry et al. (64) in the presence of 0.05% sodium deoxycholate to minimize turbidity. Bovine serum albumin was used as the standard. Lipoprotein electrophoresis was done using a Corning apparatus and Universal agarose film in 50 mM barbital buffer, pH 8.6. Lipoprotein bands were visualized by staining with fat red.

RESULTS
Assuming that the expression of other receptors is not altered by the disruption of the SR-AI/II gene, then the uptake of acetyl LDL or oxidized LDL by macrophages via "alternate" receptors should be equivalent to saturable uptake in macrophages from knockout mice, whereas the uptake via scavenger receptors should be the difference in uptake between wild-type and knockout macrophages. Accordingly, resident peritoneal macrophages from wild-type mice or scavenger receptor knockout mice were incubated with increasing concentrations of 125 Ilabeled acetyl LDL or 125 I-labeled oxidized LDL. As shown in Fig. 1, the uptake and degradation of acetyl LDL was reduced by more than 80% in macrophages from scavenger receptor knockout mice. This is direct evidence that most of the uptake of acetyl LDL by macrophages can be attributed to the SR-AI/ II. In contrast, the uptake and degradation of extensively oxidized LDL was reduced by only 30% in knockout mice and showed high affinity, saturable kinetics with an apparent K m of about 5 g/ml, similar to that of SR-AI/II. This suggests that about 70% of the uptake of oxidized LDL in macrophages is attributable to an alternate oxidized LDL receptor(s). Fig. 1 also shows that after 5 h incubation with acetyl LDL, the amount of cell-associated LDL was about 20% of the amount degraded, whereas with oxidized LDL, this was similar to or even greater than the amount degraded. This is not surprising, because oxidized LDL has been shown previously to resist lysosomal degradation, and hence tends to accumulate within cells (65,66).
To define the degree of oxidative modification of LDL that is required for high affinity uptake in SR-AI/II knockout macrophages, the uptake of LDL preparations that differed in degree of oxidation was compared. Results shown in Fig. 2 indicate that the uptake of mildly oxidized LDL was the same as native LDL, and that at least a moderate degree of oxidation is required for recognition by the alternate receptor(s). This oxidized LDL had electrophoretic mobility 3.4-fold that of native LDL, which corresponds to derivatization of about 40% of lysine residues (35). This degree of modification is similar to the "threshold" degree of modification required for uptake of oxidized or chemically modified LDL by the SR-AI/II (35, 67).
Scavenger receptors have a rather broad and often overlapping ligand specificity but can be distinguished on the basis of

FIG. 2. Extensive oxidation of LDL is required for rapid uptake in SR-AI/II-deficient macrophages. Cultured resident peritoneal macrophages from control mice (solid symbols) or knockout mice (open symbols) were incubated for 5 h at 37°C with the indicated concentrations of radioiodinated native LDL (A), mildly oxidized LDL (B), extensively oxidized LDL (C), or very extensively oxidized LDL (D).
Cumulative LDL uptake (sum of cell-associated and degraded LDL) was then measured. The electrophoretic mobility of mildly oxidized LDL was 1.9 times that of native LDL, that of extensively oxidized LDL was 3.4 times that of native LDL, and that of very extensively oxidized LDL was 4.6 times that of native LDL. Values shown are means of triplicates that varied by less than 10%. Similar results were obtained in two replicate experiments.
binding to polynucleotides in that class A receptors are capable of interacting with polyinosinic acid and certain other polynucleotides, whereas class B receptors do not (53). To further characterize the uptake pathway for oxidized LDL in SR-AI/II knockout macrophages, several different modified LDLs and polynucleotides were tested for their ability to compete for the degradation of radiolabeled oxidized LDL. Fig. 3 shows that as expected, unlabeled oxidized LDL was an effective competitor, whereas acetyl LDL was a poor competitor. Polyinosinic acid completely blocked oxidized LDL uptake, indicating that the uptake of oxidized LDL in knockout macrophages cannot be attributed to SR-B. We have shown previously that reaction of LDL with fatty acid peroxidation products (under conditions where oxidation of LDL itself was prevented) results in high affinity saturable uptake by macrophages (63,68). To determine if oxidation product-modified LDL was a ligand for the "alternate" scavenger receptor in SR-AI/II knockout macrophages, we tested its ability to compete for the uptake of oxidized LDL in these cells. As shown in Fig. 4A, LDL that had been modified by arachidonic acid oxidation products competed for the uptake of oxidized LDL, and the efficiency of oxidation product-modified LDL depended on its extent of derivatization as reflected by electrophoretic mobility. Bovine serum albumin modified by arachidonic acid oxidation products was also a good competitor, indicating that neither apolipoprotein B nor phospholipids were required for ligand activity. Similar results were obtained when oxidation products from linoleic acid were used to modify LDL, except that a much lower degree of modification (as judged by electrophoretic mobility in agarose) was required to fully compete for oxidized LDL uptake. This suggests that this receptor can distinguish between different lipid peroxidation products, and that those derived from linoleic acid may generate a higher affinity ligand.
Although many potentially reactive products have been found after autooxidation of polyunsaturated fatty acids, 2-unsaturated aldehydes and malondialdehyde are thought to be among the more important (69). To assess the specificity of the alternate receptor(s) for aldehyde-modified proteins, malondialdehyde and simple 2-unsaturated aldehydes of varying chain length were used to modify LDL, and the resulting modified LDLs were tested for their ability to compete for the degradation of radioiodinated oxidized LDL by macrophages from SR-AI/II knockout mice. Fig. 5A shows that acrolein-modified LDL and malondialdehyde-modified LDL failed to compete for oxidized LDL uptake. Heptenal-LDL was a moderately effective competitor but was much less potent than LDL modified by fatty acid peroxidation products. This suggests that the lipid peroxidation products that cause binding to the alternate receptor(s) are unlikely to be simple short-chain aldehydes. Several scavenger receptors, including macrosialin, CD36, and SR-BI are reported to interact with phosphatidylserine-rich membranes (57,70). To determine if the uptake of oxidized LDL in SR-AI/II-deficient macrophages might involve one of these receptors, we tested the effect of phosphatidylserine-rich liposomes on oxidized LDL degradation. As shown in Fig. 5B, phosphatidylserine liposomes were effective competitors, whereas phosphatidylcholine liposomes had no effect.
As a preliminary step to identifying membrane proteins that might mediate the binding of oxidized LDL in the SR-AI/II FIG. 3. Specificity of the oxidized LDL uptake pathway in knockout macrophages. Cultured resident peritoneal macrophages from SR-AI/II-deficient mice were incubated for 5 h at 37°C with 5 g/ml 125 I-labeled oxidized LDL and the indicated concentration of unlabeled oxidized LDL (q), acetyl LDL (E), polyinosinic acid (f), or polycytidylic acid (Ⅺ). LDL degradation products were then measured. The value in the absence of competitor was 6.4 g/mg. The electrophoretic mobility of the labeled and unlabeled oxidized LDL or acetyl LDL preparations ranged between 4.1 and 4.3 times that of native LDL. Each point is the mean of four determinations from two separate experiments that varied by less than 15%.
FIG. 4. Competition for the uptake of oxidized LDL in knockout macrophages by proteins modified with fatty acid peroxidation products. Cultured resident peritoneal macrophages from SR-AI/II-deficient mice were incubated for 5 h at 37°C with 5 g/ml 125 I-labeled oxidized LDL and the indicated concentration of competitor. A, LDL derivatized with arachidonic acid oxidation products to an extent that increased its electrophoretic mobility to 2.8 (E), 4.6 (Ⅺ), and 6.3 (Ç) times that of native LDL and bovine serum albumin extensively modified with arachidonic acid oxidation products (q). B, native LDL (E), LDL aged for 8 months at 4°C (q), LDL modified with linoleic acid oxidation products to an extent that increased its electrophoretic mobility to 1.7 (Ⅺ) and 2.75 (f) times that of native LDL. The electrophoretic mobility of aged LDL was 1.3 times that of native LDL, and absorbance at 234 nm of a 50 g/ml solution was 2.3 times that of native LDL. The 125 I-labeled oxidized LDL had electrophoretic mobility 4.2 times that of native LDL. Each point is the mean of duplicates from one of three experiments with almost identical results.

FIG. 5. Competition for the uptake of oxidized LDL in SR-AI/ II-deficient macrophages by aldehyde-modified LDL and liposomes.
A, cultured resident peritoneal macrophages from knockout mice were incubated for 5 h at 37°C with 5 g/ml 125 I-labeled oxidized LDL and the indicated concentration of LDL modified with malondialdehyde (E), acrolein (q), or heptenal (Ⅺ). Relative to native LDL, the electrophoretic mobility of 125 I-labeled oxidized LDL was 4.1, that of malondialdehyde-modified LDL was 4.3, that of acrolein LDL was 2.3, and that of heptenal LDL was 1.9. B, cultured resident peritoneal macrophages from knockout mice were incubated for 5 h at 37°C with 5 g/ml 125 I-labeled oxidized LDL and the indicated concentration of phosphatidylcholine liposomes (E) or phosphatidylserine-rich liposomes (q). Each point is the mean of triplicates that varied by less than 10%. knockout animals, we separated membrane proteins from knockout macrophages by SDS-polyacrylamide gel electrophoresis, transferred these to nitrocellulose membranes, and blotted the membranes with radioiodinated extensively oxidized LDL, mildly oxidized LDL, native LDL, or acetyl LDL. As shown in Fig. 6, the pattern of bands was similar with all lipoproteins; the two most intense were at 94 and 210 kDa, with minor bands at 60 and 124 kDa. Binding to the 60-, 94-, and 210-kDa bands was completely blocked by polyinosinic acid. The same bands were seen in membrane proteins from control macrophages, but in addition, a band at 220 -240 kDa was visualized with acetyl LDL that probably represents the trimeric SR-AI. The 94-kDa band has been identified as macrosialin, and according to Ramprasad et al. (57), the 210-kDa band may be a dimer of macrosialin. The identity of the other bands is at present unknown. The finding that native LDL interacted on ligand blotting with a protein at 94 kDa, whereas native LDL had no effect on the rate of oxidized LDL uptake, would suggest that this band is not the receptor. However, SR-BI is known to bind LDL and, in its glycosylated form, is similar in size to macrosialin; hence, the band at 94 kDa seen with native LDL might represent SR-BI rather than macrosialin. Because macrosialin is said to be up-regulated in thioglycollate-elicited macrophages (57), we compared the uptake of oxidized LDL in elicited macrophages from knockout mice with that in resident macrophages and found no increase in the elicited macrophages. We also tested the effect of preincubating macrophages from knockout mice with 15-100 nM phorbol myristate acetate, which is reported to increase cell-surface macrosialin in THP-1 cells (58), but found no increase in the rate of oxidized LDL uptake (data not shown). However, these negative results do not exclude a possible role for macrosialin in the uptake of oxidized LDL in these cells because we were unable to obtain an antibody to murine macrosialin and hence could not verify that these interventions had actually increased macrosialin expression. DISCUSSION The existence of a specific receptor(s) for oxidized LDL was initially proposed on the basis of studies showing incomplete competition by acetyl LDL for the uptake and degradation of oxidized LDL (44,45,49). However, the extent of competition of acetyl LDL for oxidized LDL uptake depends greatly on the extent of oxidation and aggregation of oxidized LDL, and hence it is difficult to assess the relative importance of this "specific" receptor(s) in the uptake of oxidized LDL compared with uptake via the SR-AI/II (71). The findings in the present study with SR-AI/II knockout mice provide direct evidence confirming the existence of a separate, high affinity saturable pathway for the uptake and degradation of oxidized LDL. Furthermore, the results indicate that this pathway accounts for about 70% of the uptake of extensively oxidized LDL in murine macrophages, and that only 30% is attributable to SR-AI/II. These findings differ significantly from those reported recently by Sakai et al. (72), who found that the degradation of oxidized LDL was decreased by 70% in SR-AI/II macrophages. This difference is probably a result of differences in the degree of LDL oxidation and in the assay conditions. For instance, the rate of uptake in our studies was an order of magnitude greater than that of Sakai's, suggesting that the extent of oxidation in their studies may not have been sufficient to obtain an optimal ligand. Competition studies in knockout mice revealed some characteristics in common with the SR-AI/II, in that extensive LDL oxidation was required for uptake, and polyinosinic acid was a good competitor for oxidized LDL uptake whereas polycytidylic acid, which does not form a base-quartet-stabilized helix (43), was ineffective. The ligand specificity of the uptake pathway in knockout mice differed from that of the SR-AI/II in that the apparent affinity of this alternate pathway toward acetyl LDL and malondialdehyde-modified LDL was at least an order of magnitude lower than that for oxidized LDL.
Several macrophage membrane proteins that bind oxidized LDL have recently been identified including Fc␥RII-B2 (49), CD36 (50), SR-BI (54), and macrosialin, the mouse homologue of human CD68 (57). Recent studies have shown that SR-BI binds native LDL with high affinity and that both SR-B1 and CD36 are unable to bind to polyanionic ligands such as poly I (54). In addition, CD36 binds mildly oxidized LDL at least as well as extensively oxidized LDL (50). These characteristics are clearly different from those associated with oxidized LDL uptake in knockout macrophages, indicating that neither of these class B scavenger receptors are likely to be responsible for oxidized LDL uptake in these cells. However, there is evidence that CD36 may play a role in human macrophages, in that monocyte-derived macrophages from CD36-deficient subjects have about a 50% reduction in binding and uptake of oxidized LDL, and anti-CD36 blocks 22-50% of oxidized LDL binding and uptake in normal human macrophages (50 -52).
Macrosialin/CD68 has recently been proposed as a receptor for oxidized LDL (57). The interaction of oxidized LDL with macrosialin/CD68 has been characterized mostly on the basis of ligand blot analyses (55,57). However, the ligand specificity appears strikingly similar to that we found for the uptake of oxidized LDL by SR-AI/II knockout macrophages in that there is no interaction with native LDL, a higher affinity for oxidized LDL than acetyl LDL (55), inhibition of ligand binding with polyinosinic acid but not polycytidylic acid, and binding with phosphatidylserine-rich but not with phosphatidylcholine liposomes. This concordant ligand specificity is consistent with the hypothesis that macrosialin may be an important mediator of oxidized LDL uptake. We found no increase in oxidized LDL uptake with thioglycollate-elicited macrophages or with cells pretreated with phorbol ester (both of which are expected to increase cell-surface macrosialin), but Ramprasad et al. (58) showed recently that antibodies to CD68 specifically inhibited about 20% of the uptake of oxidized LDL by human THP-1 macrophage-like cells. Definitive assessment of the role of macrosialin/CD68 may require inhibiting its expression by targeted gene disruption. Nishikawa et al. (73) described a receptor for acidic phospholipid vesicles that shares many properties with macrosialin and may in fact be the same protein (74,75). Elomaa et al. (76) have described a new receptor structurally similar to the SR-AI that is expressed only on macrophages in the marginal zone of the spleen and in the medullary cord of lymph nodes. It binds acetyl LDL, but its relative affinity for oxidized LDL and the level of expression by peritoneal macrophages have not yet been reported. This receptor is not detectable in macrophages from lung or liver, and hence its tissue distribution cannot account for the fact that in SR-AI/II knockout animals, 60 -80% of intravenously injected oxidized LDL or acetyl LDL is cleared by the liver, whereas only about 1% is taken up by the spleen. 2 A number of other scavenger receptors have been described, including a 125-kDa glycoprotein that binds glycosylated or aldehyde-modified proteins (77,78), a 90-kDa receptor for proteins modified by advanced glycosylation end-products (79), receptors for conformationally modified albumins (80 -83), and a "novel" acetyl-LDL receptor (84). Although not all of these receptors have been fully characterized, none has a ligand specificity that is the same as that of the uptake pathway for oxidized LDL in SR-AI/II knockout macrophages.