Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The fatty streak is the first recognizable lesion in the development of atherosclerosis and arises from cholesteryl esters maintained as lipid droplets inside macrophage-derived foam cells, which are found throughout life (Stary et al., 1994). Human studies as well as atherosclerotic animal models have identified the initiating role of the macrophage in atherogenesis (Gerrity, 1981; Stary, 1987; Schwartz et al., 1992). Differentiating monocytes in arterial walls express mRNA and undergo new protein synthesis for the macrophage scavenger receptor that recognizes modified forms of low-density lipoprotein (LDL) (Goldstein et al., 1979; Brown and Goldstein, 1983). The class A scavenger receptors (SR-As) (see Freeman, 1997, for classification) are helically coiled homotrimers of a single gene product and exist in three forms (SR-AI, II, and III) (Matsumoto et al., 1990; Krieger, 1992; Krieger and Herz, 1994; Gough et al., 1998). The SR-As from four species have been cloned and found to be 64 to 81% identical for SR-AI and 60 to 81% identical for SR-AII (Matsumoto et al., 1990; Ashkenas et al., 1993). The larger SR-AI form has a molecular mass of 220 kDa, arising from 77-kDa monomers (453 amino acids in the human), which use Cys83 to self-assemble in the endoplasmic reticulum (Penman et al., 1991). SR-AII lacks most of the cysteine-rich terminal region and is composed of 349 amino acids. However, the known ligand-binding portion of each receptor is identical, is a collagen-like domain (Penman et al., 1991; Acton et al., 1993; Doi et al., 1993; Tanaka et al., 1996; Gough et al., 1998), and, in humans, consists of 23 triplet repeats of the amino acid sequence Gly-X-Y. SR-AIII is a novel splice variant that does not endocytose ligand, but remains trapped within the endoplasmic reticulum, and may exhibit a dominant-negative effect when coexpressed with functional receptors (Gough et al., 1998). SR-As I and II bind oxidized and acetylated forms of LDL, which become more negatively charged upon derivatization of critical lysine residues (Steinberg et al., 1989; Zhang et al., 1993). Such a particle is no longer a ligand for the LDL receptor, but becomes, instead, a ligand for the macrophage scavenger receptor (Krieger, 1992). The relative importance of SR-As to atherogenesis was demonstrated recently when it was reported that SR-A/apolipoprotein E (apoE) double-knockout mice developed 60% smaller atherosclerotic lesions than did single-apoE knockout littermates (Suzuki H et al., 1997). This was the first direct in vivo evidence that suggested that elevated plasma cholesterol may not be as important as arterial wall macrophages for the initiation of atherogenesis and validated the SR-A as a potential target for interventional therapy.

The known ligands of the macrophage scavenger receptor, besides the modified forms of LDL (molecular mass, 2500 kDa), consist of large, polyanionic polymers of indeterminate molecular mass (Brown et al., 1980). Some widely recognized polyanionic ligands include dextran sulfate, fucoidin, and polyinosine (poly I) and polyguanosine but not polyadenine or polycytosine. Polynucleotide binding to SR-As is determined by the ability to form a base-quartet-stabilized, four-stranded helix, which produces an array of negatively charged phosphate groups (Pearson et al., 1993). To date, no small molecules have been discovered that can antagonize the binding of modified LDLs to the scavenger receptor. We report here the identification of a novel, small-molecule, macrophage scavenger receptor antagonist by using permanently transfected human embryonic kidney (HEK) 293 cells as a screening tool.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Native LDL, acetylated LDL (AcLDL), OxLDL (Cu²⁺-oxidized LDL), 1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate-labeled LDL (DiI-AcLDL), and [¹²⁵I]AcLDL (SA 0.15-0.25 µCi/µg protein) were obtained from Biomedical Technologies, Inc. (Stoughton, MA). Medium for tissue culture and geneticin were from GIBCO/BRL. Fetal bovine serum was from Hyclone (Logan, UT). Penicillin, streptomycin, human AB serum, poly I, fucoidin, dextran sulfate, and BSA (fraction V powder) were from Sigma Chemical Co. (St. Louis, MO). Human recombinant granulocyte-macrophage colony-stimulating factor (10⁷ reference units/mg protein) was obtained from Promega (Madison, WI).

Cloning of cDNAs Encoding Human SR-AI and SR-AII Receptors. Human SR-AI and SR-AII cDNAs were isolated from human placenta based on the known sequence information of the human SR-AI and SR-AII receptors (Matsumoto et al., 1990). Oligonucleotide primers were synthesized corresponding to the amino and carboxyl termini and used to obtain the full-length clones of SR-AI and SR-AII receptors from the human placental RNA by polymerase chain reaction. The full-length clones were completely sequenced from both DNA strands, and the sequence identity of both clones was confirmed. The deduced polypeptides consist of 453 and 349 amino acids for human SR-AI and SR-AII, respectively (Matsumoto et al., 1990).

Construction of a Permanently Transfected Cell Line. The SR-AI or SR-AII cDNAs were subcloned into the mammalian expression vector pCDN in the correct orientation (Elshourbagy et al., 1993). The resulting constructs containing the SR-AI or SR-AII cDNAs were used to transfect HEK 293 cells by calcium phosphate transfection. To initiate selection, fresh supplemented Eagle's minimal essential medium (EMEM) containing 0.4 mg/ml of geneticin was added. After approximately 2 to 3 weeks, each plate was examined under the microscope for small patches of growing cells, which were trypsinized and expanded.

Cell Culture. Transfected HEK 293 cells were routinely passaged in EMEM containing 0.4 mg/ml geneticin and 10% fetal bovine serum (FBS). Human monocytes were isolated from monocyte-enriched leukopacks (leukopheresis product; Biological Specialty Corp., Colmar, PA) and purified by standard procedures (Colotta et al., 1984). Monocytes were grown for 2 weeks in RPMI 1640 medium containing 1 ng/ml granulocyte-macrophage colony-stimulating factor, 100 U/ml penicillin, 100 µg/ml streptomycin, 2 mM glutamine, and 5% human AB serum to promote differentiation into macrophages. Rat peritoneal macrophages were obtained by peritoneal lavage, washed, and plated with RPMI 1640 medium containing 100 U/ml penicillin, 100 µg/ml streptomycin, 2 mM glutamine, and 20% FBS into 96-well plates at a concentration of 4 × 10⁵/well for analysis. Cells were incubated in a CO₂ incubator at 37°C for at least 2 h to facilitate attachment; nonadherent cells were aspirated, and macrophages were assayed as described below.

DiI-AcLDL Uptake Assay. The fluorescent compound DiI-AcLDL has been shown to be a useful tool in assessing activity of the macrophage scavenger receptor (Freeman et al., 1991; Penman et al., 1991). We developed a high-throughput assay for SR-A antagonists based on uptake of DiI-AcLDL by the transfected HEK 293 cells. For most assays, HEK 293 cells transfected with SR-AI were used, although both SR-AI and SR-AII appeared to have equivalent activity in all studies performed. Briefly, HEK 293 cells were seeded at 2 × 10⁴ cells/well in a 96-well plate in EMEM with 2 mM glutamine, 10% FBS, and 0.4 mg/ml geneticin. The cells showed essentially 100% attachment and achieved confluency by 3 days. Testing was performed in serum-free EMEM containing 2 mg/ml BSA. Confluent cells were incubated with DiI-AcLDL (final concentration, 2 µg/ml) in the absence and presence of SR-A ligands (quadruplicate determinations) for 4 h at 37°C. During aspiration of solutions and a Locke's buffer wash, there was minimal detachment of cells because of their enhanced attachment relative to unmodified 293 cells. Results were quantified with a CytoFluor 2350 fluorescence plate reader at 530 nm excitation/590 nm emission (PerSeptive Biosystems, Framingham, MA). Background fluorescence of unlabeled cells was subtracted from DiI-AcLDL fluorescence values (quadruplicate determinations), and results were expressed as percent total DiI fluorescence of cells incubated in the absence of inhibitor.

Radioligand Assay. Assays of SR-A degradation and binding/internalization of [¹²⁵I]AcLDL were adapted from previous studies (Goldstein et al., 1979, 1983; Ashkenas et al., 1993). Briefly, HEK 293 cells transfected with SR-AI or II were seeded at 10⁵ cells/ml/well in a 24-well dish in EMEM supplemented with 2 mM glutamine, 10% FBS, and 0.4 mg/ml geneticin. After 3 days, medium was replaced with 500 µl of fresh serum-free medium containing 2 mg/ml BSA and [¹²⁵I]AcLDL at 5 µg/ml, and cells were incubated at 37°C for 5 h. After this suitable period for ligand degradation, cells were removed to a 4°C room. Supernatant was removed into trichloroacetic acid, and the mixture was centrifuged. The supernatant was chloroform-extracted to isolate [¹²⁵I]monoiodotyrosine, the degradation product of [¹²⁵I]AcLDL, and portions were counted to determine degradative activity. Blank plastic wells containing equivalent amounts of ligand without cells were incubated and processed in the same fashion to determine background ligand degradation, which was subtracted from the total counts. Nonspecific binding was determined by incubation with poly I at 100 µg/ml. To determine cell-associated ligand, cell monolayers were washed and incubated at 4°C with ice-cold buffer A containing 150 mM NaCl, 50 mM Tris-HCl, and 2 mg/ml BSA, pH 7.4, to eliminate nonspecifically bound counts. Cells were washed three times rapidly with 1 ml, incubated twice for 10 min each in 1 ml of buffer A, and then washed twice rapidly in 1 ml of buffer A without BSA. After aspiration of all wash buffer, cells were lysed in 0.1 N NaOH and removed to counting vials for determination of binding/uptake and subsequent protein determination (Pierce BCA assay). Unless otherwise noted, triplicate determinations were made for each data point within an experiment, and n equals the number of experiments performed.

Data Analysis. Data from all ligand-uptake and binding-competition assays were analyzed by nonlinear regression analysis and curve-fitting, using Graphpad Prism, version 2.0 (Graphpad Software, San Diego, CA). Binding and degradation data from saturation [¹²⁵I]AcLDL experiments were analyzed by MacLIGAND (version 4.97) (Munson and Rodbard, 1980).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

DiI-AcLDL Assay. To characterize the specificity of the transfected receptors with known ligands, several polyanionic ligands were tested. As shown in Fig. 1A, poly I was the most effective inhibitor of DiI-AcLDL uptake, with an IC₅₀ of 0.74 µg/ml, followed closely by fucoidin (sulfated polyfucose), with an IC₅₀ of 4.7 µg/ml, and AcLDL, with an IC₅₀ of 11.7 µg/ml. OxLDL was not very potent, with an IC₅₀ of 84 µg/ml. Others (Freeman et al., 1991; Doi et al., 1993) also have seen that OxLDL has less affinity for SR-A, the acetyl LDL receptor. Unmodified LDL showed no inhibition at up to 100 µg/ml, in agreement with studies that have shown that native LDL does not bind to SR-A (Brown et al., 1980). To determine the relationship of molecular weight to inhibitory potency, a variety of molecular weights of the known SR-A ligand, dextran sulfate (DexSO₄), were also tested. As shown in Fig. 1B, the ability of DexSO₄ to inhibit uptake of DiI-AcLDL improved with increasing molecular weight, with DexSO₄ of molecular weight <10,000 being virtually ineffective at up to 100 µg/ml. These results confirm those obtained for DexSO₄ inhibition of macrophage uptake of modified LDL (Basu et al., 1979). In contrast, DexSO₄ with molecular weights of 50,000 and 500,000 inhibited well and was nearly as potent as poly I, with an IC₅₀ of 1.1 µg/ml (Fig. 1B). Screening of putative novel antagonists was performed in the same 96-well format, with transfected HEK 293 cells being incubated with DiI-AcLDL in the presence of test compounds at a concentration of approximately 30 µM. In this fashion, we identified a novel antagonist (Fig. 2) that closely matched the known inhibitors in potency, with an IC₅₀ of 13 µM (approximately 6 µg/ml) (Fig. 3). In comparison with the polyanionic ligands, the novel antagonist was twice as potent as AcLDL.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Inhibition of DiI-AcLDL uptake into HEK 293 cells transfected with SR-AI. A, transfected cells in 96-well dishes were incubated for 4 h at 37°C with DiI-AcLDL at a final concentration of 2 µg/ml in the presence of increasing concentrations of known polyanionic macrophage scavenger receptor ligands. Fluorescence was quantified with a CytoFluor 2350 plate reader, and the inhibition curves are expressed as percent total uptake of cells incubated in the absence of competitor. Curves are averaged data ± S.E.M. from four to six experiments performed with quadruplicate determinations. B, inhibition of DiI-AcLDL uptake into transfected HEK 293 cells by dextran sulfate of various molecular weights. Results show that dextran sulfate of a molecular weight <10,000 is an ineffective polyanionic ligand (n = 3 experiments).

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Structure of the novel SR-A antagonist (E)-methyl 4-chloro--[4-(4-chlorophenyl)-1,5-dihydro-3-hydroxy-5-oxo-1-(2-thiazolyl)-2H-pyrrol-2-ylidene]benzeneacetate.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Inhibition of DiI-AcLDL uptake into transfected HEK 293 cells and cultured human and rat macrophages by a novel, low-molecular-weight, SR-A antagonist. Assays were performed as in Fig. 1. Results show that the novel antagonist is equally effective against transfected receptors as well as against native human and rat SR-As. Curves are averaged data from five experiments performed with quadruplicate determinations.

Verification of Novel Antagonist in Macrophages. To verify the use of this transfected HEK 293 cell assay, other cell types that normally express SR-As, such as rat peritoneal macrophages and human macrophages, were also tested. Cultured human macrophages were allowed to mature for 2 weeks, during which time spontaneously floating cells were subcultured into 96-well dishes, where they quickly attached, and were maintained as confluent cultures until assayed as described. We confirmed the novel compound's activity against cultured human macrophages and demonstrated an IC₅₀ of 21 µM (Fig. 3). Because animal models of atherosclerosis ultimately will be used to verify in vivo activity of any novel compound, SR-A antagonist activity against rat peritoneal macrophages was also determined. Figure 3 also shows the inhibition of DiI-AcLDL uptake into rat macrophages with an IC₅₀ of 17 µM, comparable to those obtained with the transfected HEK 293 cells and with human macrophages.

Radioligand Assay. To determine saturability of SR-As in transfected HEK 293 cells, cells were incubated at 37°C with increasing concentrations of [¹²⁵I]AcLDL up to 50 µg/ml in serum-free medium. Data from representative assays are shown in Figs. 4 and 5. Binding/uptake and degradation of [¹²⁵I]AcLDL in the transfected 293 cells were saturable processes that showed high affinity and a low nonspecific component. Replicate experiments and MacLIGAND analysis determined an apparent dissociation constant (K_D) of 11.4 µg/ml with a B_max of 6525 ng/5 h/mg protein for binding/uptake (Fig. 4) and an apparent K_D of 5.1 µg/ml with a B_max of 2680 ng/5 h/mg protein for degradation (Fig. 5). Calculated on a nanomolar basis, using 514 kDa as the molecular mass of the labeled apolipoprotein B (Snyder et al., 1994), the values are an apparent K_D of 22.3 ± 2.3 nM with a B_max of 12.7 ± 1.3 pmol/mg protein for binding/uptake and an apparent K_D of 10 ± 1.5 nM with a B_max of 5.2 ± 0.5 pmol/mg protein for degradation. These values compare favorably with those reported previously in the literature for both normal cells and for those expressing transfected receptors (Goldstein et al., 1979; Freeman et al., 1991; van Berkel et al., 1991; Ashkenas et al., 1993). The apparent K_D values listed above should be viewed with caution, however, because previous authors have noted that these binding studies are not conducted under equilibrium conditions, but, rather, are conditions of steady-state (see Ashkenas et al., 1993). In addition, because the ligand is large (2500 kDa) relative to the receptor (220 kDa), there may be steric hindrance as saturation is approached, causing curvilinear Scatchard plots (Chappell et al., 1991). We have noted this effect in the transfected cells, yet MacLIGAND analysis usually gives only a single-site fit. Nonetheless, it is felt that these apparent K_D values are comparable for both degradation and binding/uptake (Goldstein et al., 1979; Ashkenas et al., 1993).

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Saturation assays of [125I]AcLDL binding/uptake in transfected HEK 293 cells. Transfected cells in 24-well dishes were incubated for 5 h at 37°C with [125I]AcLDL at increasing concentrations up to 50 µg/ml in the absence (triplicate wells) or presence of 100 µg/ml of poly I to define nonspecific binding (duplicate wells). The above data are from a representative experiment. Replicate experiments and MacLIGAND analysis determined an apparent KD of 22.3 ± 2.3 nM with a Bmax of 12.7 ± 1.3 pmol/5 h/mg protein for binding/uptake (n = 3).

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Saturation assays of [125I]AcLDL degradation in transfected HEK 293 cells. Transfected cells in 24-well dishes were incubated for 5 h at 37°C with [125I]AcLDL at increasing concentrations up to 50 µg/ml in the absence (triplicate wells) or presence of 100 µg/ml of poly I to define nonspecific binding (duplicate wells). The above data are from a representative experiment. Replicate experiments and MacLIGAND analysis determined an apparent KD of 10 ± 1.5 nM with a Bmax of 5.2 ± 0.5 pmol/ 5 h/mg protein for degradation (n = 3).

View larger version (45K): [in this window] [in a new window]   Fig. 6.   Comparison of SR-AI with SR-AII activity in transfected HEK 293 cells. Transfected cells in 24-well dishes were incubated for 5 h at 37°C with [125I]AcLDL at a final concentration of 5 µg/ml, degradation counts were determined by processing aliquots of supernatant, and cell-associated counts were obtained from well-washed cell monolayers (see Experimental Procedures). Degradation (D) averaged 1122 ± 176 ng/5 h/mg protein for SR-AI and 1191 ± 126 ng/5 h/mg protein for SR-AII (n = 18 experiments). Binding (B) averaged 1895 ± 186 ng/5 h/mg protein for SR-AI and 1767 ± 133 ng/5 h/mg protein for SR-AII (n = 18 experiments). Differences between receptors were not statistically significant (Student's t test).

Activity of Known Ligands. As with the fluorescence-based assay, known SR-A ligands were used to characterize further the specificity of the transfected HEK 293 cell receptor in this assay (Fig. 7). In both the binding and degradation assays, poly I again was an effective inhibitor, with a K_i of 1 µg/ml. DexSO₄ (molecular weight, 500,000), shown in the functional assay to be nearly as potent as poly I, showed similar inhibitory ability here, with K_i values of 1.1 and 0.7 µg/ml, respectively; fucoidin was also effective, with K_i values of 2.6 and 7.8 µg/ml, respectively (Fig. 7).

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Inhibition of [125I]AcLDL binding and degradation in transfected HEK 293 cells by polyanionic ligands. Transfected cells in 24-well dishes were incubated for 5 h at 37°C with [125I]AcLDL at a final concentration of 5 µg/ml, in the presence of increasing concentrations of known polyanionic macrophage scavenger receptor ligands. The inhibition curves are expressed as percentage total binding/uptake (A) or percentage of total degradation (B) of cells incubated in the absence of competitor. Curves are averaged data ± S.E.M. from three experiments performed in triplicate.

Activities of the Novel Antagonist. DiI-AcLDL uptake measures ligand that has been accumulated within the lysosomal compartments during the 4-h incubation and not just ligand that has been bound to cell surface receptors or merely endocytosed. This is an important distinction in the search for a small-molecule scavenger receptor antagonist, because weakly basic compounds such as chloroquine or trifluoperazine have the ability to alkalize the endosomal and lysosomal compartments of the cell. In doing so, they can effectively inhibit the cellular mechanism for degradation of AcLDL and prevent SR-A recycling (Goldstein et al., 1979; van Berkel et al., 1981, 1991). Ideally, an authentic antagonist should have the ability to inhibit ligand binding to cell surface receptors, and, as a consequence, one would see less degradation of AcLDL. By performing dose responses on putative hits identified as potential SR-A antagonists in the DiI-AcLDL assay, we eliminated many compounds that inhibited degradation without inhibiting binding/uptake. Most false positives were found to be basic in nature, including compounds such as phenothiazines, suggesting that they were acting as lysosomotropic amines to inhibit only ligand degradation, but not binding/uptake. This is illustrated in Fig. 8, in which the phenothiazine trifluoperazine was found to inhibit only degradation of radioiodinated ligand with a K_i of 9 µM, whereas binding/uptake actually increased with trifluoperazine concentration. However, the novel antagonist inhibited both degradation and binding/uptake in the transfected HEK 293 cells (Fig. 9). In the degradation assay, the average K_i was calculated to be 18 µM (8.6 µg/ml), similar to that seen in the DiI-AcLDL assay, and against binding/uptake, the average K_i was calculated to be 40 µM (19 µg/ml).

View larger version (21K): [in this window] [in a new window]   Fig. 8.   Inhibition of [125I]AcLDL degradation but not binding in transfected HEK 293 cells by trifluoperazine. Assays were performed as in Fig. 7. Results show that trifluoperazine is effective against degradation of [125I]AcLDL with a Ki of 9 µM.

View larger version (22K): [in this window] [in a new window]   Fig. 9.   Inhibition of [125I]AcLDL binding and degradation in transfected HEK 293 cells by a novel, low-molecular-weight, SR-A antagonist. Assays were performed as in Fig. 7. Results show that the novel antagonist is effective against transfected receptors in this confirmatory assay, with a Ki of 18 ± 0.3 µM against degradation of [125I]AcLDL and a Ki of 40 ± 0.7µM against [125I]AcLDL binding/uptake. Curves are averaged data from three experiments performed in triplicate.

View larger version (16K): [in this window] [in a new window]   Fig. 10.   Inhibition of [125I]AcLDL binding at 4°C in transfected HEK 293 cells. Transfected cells in 24-well dishes were incubated for 2 h at 4°C with [125I]AcLDL at a final concentration of 5 µg/ml in the presence of increasing concentrations of either the novel SR-A antagonist (SR-AA) or poly I. The inhibition curves are expressed as percent specific binding for the SR-A antagonist (n = 4) (A) or percent total binding for poly I (n = 5) (B).

View larger version (21K): [in this window] [in a new window]   Fig. 11.   Saturation assays of [125I]AcLDL binding/uptake in transfected HEK 293 cells in the presence of monensin. Transfected cells in 24-well dishes were incubated for 5 h at 37°C with [125I]AcLDL at increasing concentrations up to 30 µg/ml in the absence (triplicate wells) or presence of 100 µg/ml of poly I to define nonspecific binding (duplicate wells). The ionophore monensin was included at a final concentration of 10 µM and inhibited ligand degradation by >95%. The above data are from a representative experiment. Replicate experiments and MacLIGAND analysis determined an average apparent KD of 1.9 ± 0.17 nM with a Bmax of 2726 ± 354 fmol/5 h/mg protein for binding/uptake (n = 3).

View larger version (22K): [in this window] [in a new window]   Fig. 12.   Noncompetitive inhibition of [125I]AcLDL binding to transfected HEK 293 cells by the novel SR-A antagonist. Matched saturation assays were performed as in Fig. 11, but in the presence and absence of 50 µM the SR-A antagonist (SR-AA). The above data are from a representative experiment. Replicate experiments and MacLIGAND analysis determined an average apparent KD of 2.1 ± 0.28 nM (nonsignificant) with a significantly lower (p = 0.01) Bmax of 1964 ± 287 fmol/5 h/mg protein for binding/uptake in SR-A antagonist-treated cells (n = 3).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The novel antagonist gave qualitatively similar dose/response curves in the transfected 293 cells for both the DiI-AcLDL- and ¹²⁵I-AcLDL-binding assays, with IC₅₀ and K_i values that were close to those of AcLDL and fucoidin. Furthermore, the antagonist was an effective inhibitor of SR-A activity in both human monocyte-derived macrophages and in rat peritoneal macrophages. These data suggest that novel antagonists that are identified by transfected cells expressing SR-As can be anticipated to maintain a functional antagonism in macrophages. In addition, both binding/uptake and degradation of ¹²⁵I-AcLDL were sensitive to the novel antagonist, whereas weak bases such as trifluoperazine only inhibited degradation. These data suggest an antagonism at the level of the receptor, which is similar to the known polyanionic ligands. Although these results demonstrate that a small-molecule SR-A antagonist is possible, more development obviously is needed to attain more potency, as well as bioavailability and tolerability, before testing in animal models of atherosclerosis.

Nevertheless, it is not clear that we have identified an antagonist that competes at the level of the collagen-like binding site for AcLDL. Evidence from truncation mutants or from point mutations of the SR-A established that the collagenous domain is necessary for AcLDL binding (Acton et al., 1993; Doi et al., 1993). As few as six Gly-X-Y repeats of the collagenous domain have been made into a functional, synthetic SR-A receptor, although AcLDL binds with diminished affinity (Tanaka et al., 1996). However, the collagenous domain may not be the only site to which ligands bind or the only site necessary for ligand binding. Mutational analysis of the -helical coiled-coil domain also has shown a necessary role for this domain in SR-A function and in ligand dissociation (Doi et al., 1994; Suzuki K et al., 1997). The discovery that the collagenous domain of the SR-A lies alongside the -helical coiled-coil region at physiological pH (Resnick et al., 1996) also implies a broader potential binding site than what had been predicted formerly by molecular and biochemical analysis alone (Krieger, 1992; Krieger and Herz, 1994). Perhaps many small molecules would have to interact at these regions to block with the same efficacy as a larger, multivalent polyanionic ligand. It is possible that our novel antagonist may have binding sites within the -helical coiled-coil region as well as within the collagenous domain. Alternatively, it could prevent proper interaction between the two domains. An allosteric type of inhibition could explain our findings of noncompetitive inhibition by the novel antagonist.

It has been acknowledged that binding kinetics for the SR-As are complicated, and there well could be positive cooperativity displayed by a small-molecule antagonist, as additional molecules bind to a stretch of positively charged residues within the putative binding domain (Acton et al., 1993; Doi et al., 1993; Tanaka et al., 1996). Analysis of SR-A binding is also complicated by possible steric hindrance as a very large ligand binds to a receptor that is clustered in coated pits for the purpose of receptor-mediated endocytosis (Chappell, 1991). An allosteric effector would engender an additional level of complexity to these studies. The loss of efficacy by the novel antagonist noted in ligand-binding experiments performed at 4°C may be a result of the physical nature of the charged, collagenous binding domain. This region is temperature-sensitive, because collagen is more soluble at 4°C, whereas it aggregates at 37°C because of strong hydrophobic interactions (Doi et al., 1993). Site-directed mutagenesis of a critical lysine in this region, analogous to Lys335 in humans, is enough to seriously abrogate [¹²⁵I]AcLDL binding at 37°C, presumably because the mutant receptor cannot form the correct binding array to display its positive charges. The above mutated receptor bound [¹²⁵I]AcLDL at 4°C, however, because a decreased temperature would cause disaggregation of the tight trimer formation, extending the distance between the helices. This altered shape then may have displayed appropriate charges to accommodate the AcLDL particle or large polyanionic molecules for binding at 4°C. However, a small-molecule antagonist of limited molecular mass may not be able to find a suitable binding pocket among the disordered helices, which may only need to change shape to a small extent. As a consequence of the above results, those investigators have emphasized the importance of performing binding studies at 37°C versus 4°C (Doi et al., 1993). A recent report has expanded on the above-cited studies with a broader mutational analysis of basic residues along the entire collagenous domain (Andersson and Freeman, 1998). In this new study, multiple charged amino acid interactions within both proximal and distal regions of the receptor were demonstrated to be required for ligand binding. These mutated residues appeared to be critical for stabilizing receptor conformation at physiological temperatures. As recognized by Andersson and Freeman (1998), multiple interactive residues within the collagenous domain may be essential for cooperativity with the -helical coiled-coil domain to adopt a closed-jackknife conformation that seems essential for binding (Resnick et al., 1996). A small-molecule allosteric effector could interact among these residues to open the jackknife conformation and inhibit ligand binding. It is expected that the availability of a purified preparation of secreted SR-As would be extremely useful to pursue these questions pertaining to receptor-ligand interactions (Resnick et al., 1996; Andersson and Freeman, 1998).

Plaque regression is a function of the dynamic balance among initiation, progression, stabilization, and removal of plaque constituents (Schwartz et al., 1992). Although lowering of plasma cholesterol and LDL cholesterol has demonstrated significant patient benefits in a number of clinical trials, the magnitude of angiographic regressive changes is relatively small despite aggressive lipid-lowering regimens. Because atherosclerosis is a progressive disease initiated early in life, even subjects with normal LDL-cholesterol levels eventually will be at risk. Likewise, subjects with normal LDL-cholesterol levels but low high-density lipoprotein (HDL)-cholesterol levels are now thought to be at risk. There is an emerging need for alternative or complementary therapeutic regimens, such as scavenger receptor antagonists, which target a pivotal cellular and molecular mechanism involved in initiation and progression of atherogenesis. These new modalities will augment the success already achieved with cholesterol-lowering agents alone.

One hypothesis is that scavenger receptor antagonists will prevent the formation of macrophage-derived foam cells and the development of the fatty streak, the initial lesion of atherogenesis. With the plethora of macrophage receptors that are available to bind modified LDLs such as FcRII-B2 (Stanton et al., 1992), CD36 (Endemann et al., 1993), SR-B1 (Acton et al., 1994), CD68 (Ramprasad et al., 1995), and LOX-1 (Sawamura et al., 1997), elimination of only the SR-A class of receptor has been shown to be of benefit in knock-out animal models of atherosclerosis (Suzuki K et al., 1997; Sakaguchi et al., 1998). Furthermore, elimination of SR-A may account for the loss of as much as 80% of the degradation of modified LDL, as observed in vitro with isolated peritoneal macrophages from these knock-out animals (van Berkel et al., 1998). If we can block lipid accumulation within macrophage-derived foam cells by utilizing scavenger receptor antagonists, we may retard plaque progression and reduce vulnerability by initiating plaque regression through the process of "reverse cholesterol transport" to acceptor HDL particles (Brown and Goldstein, 1983). Therefore, this inhibition may halt further cholesteryl ester accumulation, arrest the progression of atherosclerotic lesions, and favor their regression. Even advanced lesions can regress over time with dietary changes and therapeutic intervention. Macrophage scavenger receptor antagonists may hasten plaque regression in patients that need it most, such as those suffering from ischemic heart disease or stroke, where further lipid accumulation within arterial macrophages would decrease further lumenal diameter and possibly precipitate a fatal event.