by the acetyl-LDL receptor

Proc. Natl. Acad. Sci. USA Vol. 84, pp. 537-540, January 1987 Medical Sciences

Recognition of solubilized apoproteins from delipidated, oxidized low density lipoprotein (LDL) by the acetyl-LDL receptor (atherosclerosis/lipid peroxidation/apoprotein B/macrophage)

SAMPATH PARTHASARATHY, LOREN G. FONG, DEBORAH OTERO, AND DANIEL STEINBERG* Department of Medicine, University of California, San Diego, La Jolla, CA 92093

Contributed by Daniel Steinberg, September 17, 1986

ABSTRACT Macrophages express a specific receptor that recognizes acetylated low density lipoprotein (LDL) and certain other chemically modified forms of LDL but not native LDL. LDL oxidatively modified either by incubation with endothelial cells in Ham's F-10 medium or by incubation with 5 IAM copper(ll) ion in the absence of cells is recognized by this same receptor. This oxidative modification, whether cell-induced or copper-catalyzed, is accompanied by many changes in the physical and chemical properties of LDL, including an increase in density, conversion of phosphatidylcholine to lysophosphatidylcholine, generation of lipid peroxides, and degradation of apolipoprotein B-100. Which changes are essential for eliciting the recognition by the receptor is not known. In the present paper it is shown that fragments of the degraded apolipoprotein from delipidated, oxidized LDL can be almost quantitatively resolubilized using n-octyl f3-D-glucopyranoside. These 125Ilabeled, solubilized apoproteins were degraded rapidly by mouse peritoneal macrophages, and that degradation was competitively inhibited by unlabeled acetyl-LDL and endothelial cell-modified LDL but not by native LDL. These results show that the acetyl-LDL receptor recognizes an epitope on the apoprotein moiety, either newly generated or exposed as a result of oxidative modification, rather than some oxidized lipid moiety. Further, the results suggest that the lipids of oxidatively modified LDL do not play an obligatory role in determining the conformation of that epitope.

breakdown of apolipoprotein B-100 (apo B-100) into smaller peptides (11, 12). In an attempt to characterize the receptorbinding site on the biologically modified LDL, we have developed a method for resolubilizing the delipidated apoproteins of oxidatively modified LDL using octyl glucoside. We show that mouse peritoneal macrophages take up and degrade these apoproteins at almost the same high rate at which they degrade intact oxidatively modified LDL. This uptake and degradation of the apoprotein moiety is inhibited competitively by acetyl-LDL and malondialdehyde-conjugated LDL (MDA-LDL). The results strongly suggest that, during oxidative modification, one or more new epitopes are generated on the apoprotein (or newly exposed); that these are recognized by the acetyl-LDL receptor; and that the LDL lipids are not obligatorily involved in determining the configuration recognized by the receptor.

MATERIALS AND METHODS Materials. Carrier-free Na125I was purchased from Amersham. Ham's F-10 and Dulbecco's modified Eagle's medium (DMEM) were obtained from GIBCO. Fetal bovine serum was from HyClone (Logan, UT). n-Octyl 3-D-glucopyranoside (octyl glucoside) was purchased from Sigma. Lipoproteins. LDL (density, 1.019-1.063 g/cm3) was isolated by preparative ultracentrifugation from fresh human plasma collected in EDTA (1 mg/ml) and was radioiodinated as previously described (13). Unlabeled and 1251I-labeled acetyl-LDL were prepared as described by Basu et al. (14). MDA-LDL was prepared as described (15, 16) by incubation of LDL (4 mg) with 0.15 M MDA at 370C for 3 hr. MDA was freshly generated by hydrolysis of 0.5 M tetramethoxypropane with 0.4 M HCl at 370C for 10 min. 125I-labeled MDA-LDL was similarly prepared, using labeled LDL. All lipoprotein samples were dialyzed against phosphatebuffered saline (137 mM NaCl/2.7 mM KCl/9.5 mM phosphate, pH 7.4) containing 0.01% EDTA. Cells. Rabbit aortic EC provided by V. Buonassissi (Alton Jones Cell Science Center, Lake Placid, NY) were grown in Ham's F-10 medium containing 15% fetal bovine serum. The cells were used at confluency for the modification of LDL. Mouse peritoneal macrophages were harvested from female Swiss Webster mice (25-30 g) by peritoneal lavage and were cultured overnight in 12-well plastic dishes (2.5 X 10' cells per dish) in DMEM containing 10% fetal bovine serum. LDL Modification. LDL (100 ug/ml) was incubated with confluent EC cultures or with 5 gM copper sulfate in the absence of cells in 2 ml of Ham's F-10 medium at 370C for 24 hr in 60-mm plastic dishes (11, 12). The medium containing

The presence of lipid-laden foam cells is a characteristic feature of early atherosclerotic lesions (1-3). Recent evidence strongly suggests that a substantial number of these foam cells are derived from monocyte/macrophages (1-3). However, these cells possess few receptors for native low density lipoprotein (LDL), and those receptors are downregulated in the presence of high concentrations of LDL. Thus, these cells degrade native LDL poorly (4). Acetylation and certain other chemical modifications of LDL that block e-amino groups of lysine residues generate modified forms of LDL that are degraded avidly by macrophages. These negatively charged forms of LDL are degraded by way of another receptor, the acetyl-LDL or "scavenger" receptor

(5).

Incubation of LDL with endothelial cells (EC) (6-8), smooth muscle cells (8, 9), or macrophages (10) generates a biologically modified, negatively charged LDL that is also

degraded much faster than native LDL and also by way of the same acetyl-LDL receptor. Incubation of LDL in the presence of high concentrations of copper(II) ion, even in the absence of cells, mimics these cell-induced changes in the lipoprotein (11). The cell- and copper-induced modifications of LDL include (i) peroxidation of the lipids, (it) conversion of lecithin to lysolecithin, (iii) an increase in density, and (iv)

Abbreviations: LDL, low density lipoprotein; apo B, apolipoprotein B; EC, endothelial cell(s); MDA, malondialdehyde. *To whom reprint requests should be addressed at: Division of Endocrinology and Metabolism, Department of Medicine, M-013D, University of California, San Diego, La Jolla, CA 92093.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 537

538

Medical Sciences: Parthasarathy et al.

the modified LDL was harvested and used for the determination of lipid peroxidation, macrophage degradation, and extent of-apoprotein solubilization by octyl glucoside. Determination of Lipid Peroxidation. Lipid peroxidation was determined by quantitation of thiobarbituric acid-reactive materials, as described (17). Tetramethoxypropane was used as standard and the results are expressed as nmol equivalents of MDA. MacrophageDegradation. Macrophages were cultured overnight, washed twice with DMEM (2 ml) and incubated at 37°C for 5 hr in a total volume of 1 ml of serum-free DMEM containing 5 ug of the 125I-labeled lipoprotein or apoprotein samples. The medium was then harvested and the trichloroacetic acid-soluble radioactivity was determined as described (6). Protein was determined by the method of Lowry et al. (18). Samples containing octyl glucoside were assayed for protein using Lowry reagent containing 6 mg of sodium deoxycholate per ml. Solubilization of Lipid-Free Apoproteins of Oxidatively Modified LDL. Medium (1.8 ml) containing =180 ,ug of the modified LDL (expressed in terms of the protein moiety) was extracted by adding 2 ml of ice-cold methanol followed by 2 ml of chloroform (19). The mixture was centrifuged at 1800 x g for 10 min to separate the phases and to allow the protein to form a dense interphase. The two phases were removed carefully, leaving the protein band adhering to the walls ofthe tube. The protein was washed twice with 2 ml of ice-cold water and once with cold acetone. After a final wash with 2 ml of water, 1 ml of an aqueous solution of octyl glucoside (6.0 mg/ml, 30 times the protein mass) was added to the protein and the mixture was mixed vigorously on a Vortex for about 5 min. Any turbidity was usually cleared with the addition of 5 ,l of 1 M NaOH. The solution was immediately dialyzed overnight against phosphate-buffered saline. Delipidated apo B from native LDL could not be completely solubilized under these conditions and was used as a turbid suspension in the presence of octyl glucoside. RESULTS Solubilization of Lipid-Free Apoproteins of Oxidatively Modified LDL. As is the case for intact apo B, after delipidation essentially none of the apoprotein of copper-oxidized LDL could be dissolved in simple aqueous buffers. These insoluble polypeptide fragments are still quite large, all >70 kDa (L.G.F., S.P., J. L. Witztum, and D.S., unpublished results). However, octyl glucoside was very effective in solubilizing the peptide mixture (Table 1). The delipidated 1251-labeled apoproteins of copper-oxidized LDL were treated with increasing concentrations of octyl glucoside. The resultant solution was cleared by centrifugation, and the radioactivity in the supernatant was determined. As seen in Table 1, essentially complete solubiiization was achieved when the mass ratio of detergent to protein was 30 or greater. In the presence of octyl glucoside at these concentrations, the rates of macrophage degradation of intact copper-oxidized LDL and of the solubilized apoproteins derived from it were almost identical (Table 1). Even when the solubilization was incomplete (at a 10:1 detergent/protein ratio), the proteins that were solubilized were degraded at the same rate as the intact copper-oxidized LDL or the totally solubilized proteins, showing that there was probably no preferential solubilization of certain fragments. The apo B of delipidated acetyl-LDL or MDA-LDL was not solubilized even at high concentrations of octyl glucoside. The possibility was considered that the degradation of the resolubilized apoproteins might have occurred in the macrophage incubation medium (e.g., as a result of cell damage due to the presence of the detergent). However, the degradation

Proc. Natl. Acad. Sci. USA 84

(1987)

Table 1. Solubilization of lipid-free apoproteins of copper-oxidized, '251-labeled LDL with octyl glucoside Mass ratio Percentage of of octyl '25I-labeled Degradation glucoside protein by macrophages,t Sample to protein solubilized* Ag/mg Intact LDL 5.52 Apoproteins from delipidated LDL 10 5.83 NDt 20 62.70 5.68 30 98.00 5.77 40 99.30 4.64 *Determined from the 125J radioactivity remaining in the supernatant after centrifugation of the sample at 1800 x g for 10 min. tAmount degraded by macrophages was determined directly by using 5 ,ug equivalent of solubilized proteins, without dialysis. The actual concentration of octyl glucoside in the macrophage incubation system was between 200 and 300 Ag/ml. Values (Qg degraded in 5 hr, per mg of macrophage protein) are averages of duplicate incubations from a set of three separate trials. tNot determined.

of another ligand for the acetyl-LDL receptor, '251-labeled MDA-LDL, was not appreciably affected by the presence of up to 300 ,g of octyl glucoside per ml of medium; even if no octyl glucoside had been removed from the resolubilized apoproteins by dialysis, the final concentration during the macrophage incubation could not have been >150 ,g/ml. Nor did the presence of octyl glucoside increase the degradation of 1251I-labeled native LDL. A lipid-free suspension of apo B from native LDL was not degraded more rapidly than intact native LDL in the presence of the detergent. Finally, there was no evidence that octyl glucoside damaged or removed macrophages from the dishes; even when octyl glucoside was present at 250 ,g/ml in the medium, recovery of macrophage protein was at least 86% that in the absence of detergent. The macrophage degradation of intact EC-modified LDL was competitively inhibited by the resolubilized apoproteins derived from delipidated EC-modified LDL (Fig. 1). The degradation of 125I-labeled EC-modified holo-LDL (not delipidated) was inhibited by the unlabeled apoprotein mixture just as effectively as it was by unlabeled EC-modified holo-LDL (Fig. LA). Native LDL did not inhibit significantly. Reciprocally, the degradation of 251I-labeled apoproteins derived from delipidated 1251I-labeled, EC-modified LDL was inhibited almost to the same extent either by unlabeled apoproteins from EC-modified LDL or by unlabeled ECmodified holo-LDL (Fig. 1B). Again, native LDL did not inhibit significantly. The results show that the resolubilized apoproteins prepared from EC-modified LDL are recognized and metabolized by the macrophage in much the same way as is the EC-modified intact LDL. The degradation of EC-modified or copper-modified LDL is inhibited competitively by acetyl-LDL and MDA-LDL, as well as by other ligands for the acetyl-LDL receptor (7, 8). As shown in Fig. 2, the macrophage degradation of the labeled solubilized apoproteins was also inhibited by unlabeled acetyl-LDL. As shown above, native LDL did not compete for the degradation of the solubilized proteins. Reciprocally, the solubilized apoproteins competed for the degradation of 1251I-labeled acetyl-LDL (Fig. 3) and also for the degradation of 125I-labeled MDA-LDL (results not shown).

DISCUSSION The several chemical modifications of LDL that convert it to a form recognized by the acetyl-LDL receptor have in common that they modify the e-amino groups of lysine residues on the apoprotein. Conjugation with MDA has been

Medical Sciences:

Proc. Natl. Acad. Sci. USA 84 (1987)

Parthasarathy et al.

539

Competitor, Ag/ml FIG. 1. Competition for degradation of intact EC-modified 125I-labeled holo-LDL (A) or of apoproteins derived from EC-modified 125I-LDL (B) by unlabeled EC-modified LDL (o), apoproteins from EC-modified LDL (*), or native LDL (i). Five micrograms of EC-modified I251-LDL (A) or 5 /ig of solubilized apoprotein derived from EC-modified '"I-LDL (B) was incubated with macrophages in the presence either of unlabeled intact EC-modified LDL (o), unlabeled solubilized apoproteins from EC-modified LDL (e), or intact native LDL (m) at the concentrations specified, in a total volume of 1 ml of DMEM at 37TC for 5 hr. Trichloroacetic acid-soluble radioactivity was determined at the end of the incubation. Values are means of duplicates from a representative experiment.

extensively investigated (16, 20), and it appears that a critical number of lysine residues must-be conjugated before the new recognition site is established. Exactly what that site is remains unknown. While it seems likely that the recognition site is protein-associated, even that has not previously been explicitly established. Oxidative modification, whether induced by incubation with cells or by chemical oxidation in the absence of cells, affects both the lipid and protein moieties, as previously described (6-8, 10-12). Since oxidatively modified LDL competes with both acetyl-LDL and MDALDL (7, 8), it is presumed that a similar structural modification is produced by each of these treatments. The present findings show that the new recognition site is associated with one or more of the apoprotein fragments generated during oxidative modification. Thus, the new epitope or epitopes generated do not depend upon interaction with LDL lipids to

maintain the configuration recognized by the acetyl-LDL receptor. We can also conclude tentatively that the oxidation of LDL lipids and the hydrolysis of lecithin associated with oxidative modification do not directly generate a recogniton site. These results limit the possible explanations for the appearance of a site on oxidized LDL that is recognized by the acetyl-LDL receptor. One possibility is that the oxidative modification, accompanied by breakdown of apo B, exposes a "buried" sequence in apo B. A second possibility is that there is a significant conformational change in the LDL protein associated with oxidative modification. Studies by Haberland et al. (21, 22) have shown that maleoylation of serum albumin does in fact induce a change in conformation 24

10 rto

-o

20 16 -o

8

1 16

-0

3o

-0

6

O

-o 12

-o

0 ,IC

la 0 r_ =

4

2_ 0

4 I

I

I

I

I

10

20

30

40

50

Competitor, ug/ml FIG. 2. Degradation by mouse peritoneal macrophages of 125I_ labeled apoproteins solubilized from EC-modified LDL in the presence of various concentrations of unlabeled acetyl-LDL as competitor. Macrophages were incubated with 5 ug of apoproteins derived from 125I-labeled, EC-modified LDL and various concentrations of unlabeled acetyl-LDL in a total volume of 1 ml at 37TC. After 5 hr of incubation, trichloroacetic acid-soluble radioactivity was determined. Each point represents the mean of a duplicate set of

experiments.

8

O0 CpVetVg/

Competitor, tug/ml FIG. 3. Effect of unlabeled apoproteins derived from EC-modified LDL (o) or unlabeled apoproteins derived from copper-oxidized LDL (e) on macrophage degradation of l25l-labeled acetyl-LDL. Five micrograms of 125I-labeled acetyl-LDL was incubated at 370C with macrophages in a volume of 1 ml, in the presence of unlabeled apoproteins at the concentration specified. After 5 hr of incubation, trichloroacetic acid-soluble radioactivity in the medium was determined. Values are means of duplicate determinations from a representative set of incubations.

540

Medical Sciences: Parthasarathy et al.

recognized by the acetyl-LDL receptor. In the present instance, the fact that the solubilized apoproteins are recognized just as well as the oxidatively modified holo-LDL makes it somewhat less likely (but does not rule out) that a change in conformation is involved. The apo B fragments of oxidized LDL have molecular weights greater than 70,000 (L.G.F., S.P., J. L. Witztum, and D.S., unpublished results) and are therefore large enough to retain conformations present in the holo-LDL from which they are ddrived, even in the presence of the detergent. Whichever of these possibilities is operative, it is possible that the alteration leading to

recognition may produce a unique epitopic change that could also be distinguished by specific antibodies. Previous attempts in this laboratory to generate monoclonal antibodies that recognize oxidatively modified LDL but not native LDL have been unsuccessful (J. L. Witztum, unpublished results). It may now be possible to utilize the delipidated protein fragments as antigens and obtain the desired specific monoclonal antibody. The studies of Goldstein et al. (4) and Brown et al. (5) have shown that the acetyl-LDL receptor is unusual in several ways. A number of high molecular weight polyamines, including poly(inosinic acid) and fucoidin, compete very effectively with acetyl-LDL for the receptor (5), but not all polyanions do so. Conjugation of several different ligands to lysine generates a more negatively charged LDL that is recognized by the acetyl-LDL receptor, showing that the specificity is not structurally unique (i.e., not uniquely acetyllysine or acetoacetyllysine or MDA-lysine, etc.). It is possible that the change in charge, while not sufficient in itself, may be necessary in that it exposes or creates the new structural feature uniquely recognized. The normal function and the physiological ligand or ligands for the acetyl-LDL receptor remain unknown. Acetylation and the other in vitro chemical modifications probably do not occur in vivo to any extent. The extent to which oxidative modification of LDL occurs in vivo also remains to be established. The acetyl-LDL receptor might function to remove oxidized and therefore "damaged" lipoproteins from the circulation and from interstitial fluids. That might represent a protective function under ordinary circumstances and yet become, inadvertently, a contributing factor in atherogenesis. Atherosclerosis is almost uniquely a human disease, and probably there has not been any significant evolutionary genetic pressure against the acetyl-LDL receptor. Furthermore, even in humans the disease expresses itself significantly only after the reproductive period and thus should exert little if any selective pressure. The ubiquity of the acetyl-LDL receptor on macrophages from many species

Proc. Natl. Acad. Sci. USA 84 (1987) strongly implies some general physiological function. Ascertaining the nature of the ligand that it normally recognizes could be of general biological significance. This work was supported in part by Grant HL14197 and by Training Grant HL07276 from the National Heart, Lung, and Blood Institute. 1. Fowler, S., Shio, H. & Haley, W. J. (1979) Lab. Invest. 41, 372-378. 2. Schaffner, T., Taylor, K., Bartucci, E. J., Fisher-Dzoga, K., Beeson, J. M., Glagov, S. & Wissler, R. W. (1980) Am. J. Pathol. 100, 57-73. 3. Gerrity, R. (1981) Am. J. Pathol. 103, 181-190. 4. Goldstein, J. L., Ho, Y. K., Basu, S. K. & Brown, M. S. (1975) Proc. Natl. Acad. Sci. USA 76, 333-337. 5. Brown, M. S., Basu, S. K., Flack, J. R., Ho, Y. K. & Goldstein, J. L. (1980) J. Supramol. Str. 13, 67-81. 6. Henriksen, T., Mahoney, E. M. & Steinberg, D. (1981) Proc. Natl. Acad. Sci. USA 78, 6499-6503. 7. Henriksen, T., Mahoney, E. M. & Steinberg, D. (1982) Ann. N. Y. Acad. Sci. 401, 102-116. 8. Henriksen, T., Mahoney, E. M. & Steinberg, D. (1983) Arteriosclerosis 3, 149-159. 9. Heinecke, J. W., Baker, L., Rosen, H. & Chait, A. (1986) J. Clin. Invest. 77, 757-761. 10. Parthasarathy, S., Printz, D. J., Boyd, D., Joy, L. & Steinberg, D. (1986) Arteriosclerosis, in press. 11. Steinbrecher, U. P., Parthasarathy, S., Leake, D. S., Witztum, J. L. & Steinberg, D. (1984) Proc. Natl. Acad. Sci. USA 81, 3883-3887. 12. Parthasarathy, S., Steinbrecher, U. P., Barnett, J., Witztum, J. L. & Steinberg, D. (1985) Proc. Natl. Acad. Sci. USA 82, 3000-3004. 13. Weinstein, D. B., Carew, T. E. & Steinberg, D. (1976) Biochim. Biophys. Acta 424, 404-421. 14. Basu, S. K., Goldstein, J. L., Anderson, R. G. W. & Brown, M. S. (1976) Proc. Natl. Acad. Sci. USA 73, 3178-3182. 15. Fogelman, A. M., Shechter, I., Seager, J., Hokom, M., Childs, J. S. & Edwards, P. A. (1980) Proc. Natl. Acad. Sci. USA 77, 2214-2218. 16. Haberland, M. E., Fogelman, A. M. & Edwards, P. A. (1982) Proc. Natl. Acad. Sci. USA 79, 1712-1716. 17. Yagi, K. (1976) Biochem. Med. 15, 212-216. 18. Lowry, 0. H., Rosebrough, N. H., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193, 265-275. 19. Bligh, E. G. & Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37, 911-917. 20. Haberland, M. E., Olch, C. L. & Fogelman, A. M. (1984) J. Biol. Chem. 259, 11305-11311. 21. Haberland, M. E. & Fogelman, A. M. (1985) Proc. Natl. Acad. Sci, USA 82, 2693-2697. 22. Haberland, M. E., Rasmussen, R. R., Olch, C. L. & Fogelman, A. M. (1986) J. Clin. Invest. 77, 681-689.