THEJOURNAL OF B I O ~ G I CCHEMISTRY AL 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 269, No.21, Issue of May 27, pp. 15274-15279,1994
Printed in U.S.A.
A Monoclonal Antibodyagainst Oxidized Lipoprotein Recognizes Foam Cellsin Atherosclerotic Lesions COMPLEX FORMATION OF OXIDIZED PHOSPHATIDYLCHOLINES AND POLYPEPTIDES* (Received for publication, January 31, 1994) Hiroyuki ItabeSS, Eiji TakeshimaS, Hitomi Iwasakin, Junji Kimuran, Yoji Yoshidall, Tsuneo ImanakaS, and Tatsuya Takano+ From the $Department of Microbiology and Molecular Pathology, Faculty of Pharmaceutical Sciences, Teikyo University, Sagamiko, Kanagawa 199-01, the llVessel Research Laboratory Znc., Kanagawa 259-01, and the IDepartment of Pathology, Yamanashi Medical University, Yamanashi 409-38, Japan
In this study we have used homogenates of human plexes with polypeptides. The antigenic materials are atheromatous plaque as immunogen to establish a mu- detected in foam cellsin atherosclerotic lesions. rine monoclonalantibodywhichrecognizesoxidized low density lipoprotein (LDL). This monoclonal antibody, FOHlalDLH3,reacted with oxidized LDL, but enOxidative modification of low density lipoprotein (LDL)’ is zyme-linked immunosorbent assay showedthat it had believed to be involved in atherogenesis (1-3). Uptake of oxino reaction with native, acetylated, or malonaldehyde- dized LDL by macrophages and smooth muscle cells leads to treated LDL. The antibody cross-reacted with oxidized the formationof foam cells which accumulate lipid droplets. It high density lipoprotein, suggesting that specific se- is well known that a variety of modified lipoproteins, including quences of the apolipoprotein B are not essential for acetylated LDL, malonaldehyde-treated LDL (MDA-LDL), and antigen recognition by the antibody.Immunohistooxidized LDL, are effectively incorporated into macrophages, chemical analysis of thin paraffin sections from human resulting in a massive accumulationof cholesteryl ester in vitro coronary arteries showed that foam cells derived from (4-9). Transformation of macrophages tofoam cells is induced macrophages in atherosclerotic lesionswereheavily stained by this antibody. Severalother structures in the by the incorporation of modified lipoproteins via specific receplesions, including swollen collagen fibers, cellular de- tors. Scavenger receptor, or acetylated LDL receptor, has been bris in necrotic cores, and endothelial cells, were mod- isolated and the genes cloned from both mouse and human erately stained. The epitope of this antibody was char- tissue (10-12). Two other receptors specific for oxidized LDL, LDL, havebeen identified acterized by a model antigen-generating system using which do notbindacetylated ferrous ion-induced peroxidationof lipids. When the to- recently (13, 14). Atherosclerotic lesions have been shown to to LDL, partal lipid fraction extracted from LDL was treated with contain materials with similar propertiesoxidized the ferrous ion-induced peroxidation system,the reac- ticularly with respect to their physical properties, immunoretion mixture was recognizedby the antibody. Antigenic activity, and enhanced uptakeby macrophages (2). Specific anproduct(s) was produced only when phosphatidylchotibodies against MDA-conjugated lysine, 4-hydroxynonenalline (PC) wastreated with the ferrous ion-induced per- conjugated lysine, or oxidized LDL gave positive reactions in oxidation, other lipids failed to react to the antibody. To Watanabe heritable hyperlipidemea(WHHL) rabbits (15-17). investigate the possible formationof a complex of antiThe structureof oxidized LDL has not yet been well characgenic product with a polypeptide, a synthetic peptide terized, largely due to the technical difficulties of analyzing and a rabbit antiserum against the peptide were used. oxidized LDL. Apolipoprotein B (apoB) is a huge protein andis Reaction mixture of ferrous ion-induced peroxidationof extensively modified during the oxidation reaction. FragmenPC in thepresence of the peptide was addedto a micro- tation and aggregation of apoB during oxidation of LDL have titer wellprecoatedwith the monoclonalantibody been reported (18-20). Several lipid products, however, have FOHlalDLH3. After washing, the peptide remaining in been isolated from oxidized LDL (21-241, including cytotoxic the well wasdetected with rabbit antiserum against the peptide, whereas no reactivity was observed when pep- products suchas 4-hydroxynonenal(21),epoxycholesteryl ester tide alone was added to the well. Binding of the anti- (22), andlysophosphatidylcholine (8, 24) formed by enzymatic hydrolysis of peroxidized phosphatidylcholine (PC) (25). Hygenic complex to the precoated monoclonal antibody was competed by oxidized PC produced in the absence droxy fatty acids and oxysterols have been found in human of any polypeptide.We conclude that oxidized phospho- atheroma andaorta of cholesterol-fed rabbits (26,27).However, lipid product(s) is the epitope of the monoclonal anti- little is known about the distribution of oxidized LDL and/or body and that the oxidized phospholipid forms com- peroxidation products in atherosclerotic lesions. Monoclonal antibodies could be used to characterize individual components of this complicated particle. We, and other * This study was supported in part by %search Funds from Uehara Foundation and OnoFoundationand by grants-in-aid for Scientific The abbreviations used are: LDL, low density lipoprotein; HDL, Research from the Ministry of Education,Science andCulture of Japan. The costs of publication of this article were defrayed in part by the high density lipoprotein; apoB, apolipoprotein B; ELISA, enzyme-linked payment of page charges. This article must therefore be hereby marked immunosorbent assay;BHT, butylated hydroxytoluene;DAB, diamino“aduertisenent” in accordancewith 18 U.S.C. Section 1734 solely to benzidine;PBS,phosphate-bufferedsaline;TBS,Tris-HC1-buffered saline; BSA,bovine serumalbumin;PC,phosphatidylcholine;PE, indicate this fact. 8 To whom correspondence shouldbe addressed:Dept. ofMicrobiology phosphatidylethanolamine;TBARS, thiobarbituric acid reactive suband Molecular Pathology, Faculty of Pharmaceutical Sciences, Teikyo stance; MDA, malonaldehyde; MDA-LDL, malonaldehyde-treated low coenzyme University, Sagamiko, Kanagawa 199-01,Japan. Tel.: 81-426-85-3739; density lipoprotein; HMG-CoA, 3-hydroxy-3-methylglutaryl A , WHHL, Watanabe heritable hyperlipidemea. Fax: 81-426-85-3738.
Anti-human Oxidized LDL Monoclonal Antibody
and nativeLDL as antigen (1 pg of proteidwell). After incubationof 50 p1 of hybridoma supernatants, and with intervening washes with Trisbuffered saline, pH 7.8, containing 0.05% Tween 20 (TBS-Tween), the wells were incubated with alkaline phosphatase-conjugated goat antimouse Ig(G+M)antibody, followed by a substrate solution containing1 mg/ml p-nitrophenyl phosphate. Hybridoma cells corresponding tosupernatants that werepositive on the oxidized LDL and negativeon the native LDL were thencloned by limiting dilution. After repeated screening of 1750 colonies, eight clones were obtained. The isotype of the monoclonal antibodies was determined on a n ELISA using Mouse Typer [email protected]
(Bio-Rad). Ascites fluids were obtained by injecting hybridomas into mice previously primed with pristane. The monoclonal antibodies were partially purified from the ascites fluid by ammonium sulfate precipitation followed by gelfiltration column chromatography on Sephacryl S-300. ELISA-In the standard assay, 100 pl of PBS containing 1 pg of antigen protein was placed in eachwell of a 96-well microtiter plate and incubated for 18 h at 4 "C. After removing the antigen fluid, thewells were blocked with TBS containing2% skim milk (TBS-skim milk) for 2 EXPERIMENTAL PROCEDURES h at room temperature. After washing thewells three times withTBSTween, the monoclonal antibody (1:5000 diluted in PBS) was reacted for Materials-Protein A-agarose, egg yolk PC, l-palmitoyl-2-oleoyl-PC, l-palmitoyl-2-linoleoyl-PC, l-stearoyl-2-arachidonoyl-PC, 1-0-hexade- 2 h a t room temperature, followed by alkaline phosphatase-conjugated cyl-2-arachidonoyl-PC, bovine brainsphingomyelin,trilinolein, cho- goat anti-mouse Ig(G+M) antibody (1:5000 diluted in TBS-skim milk) lesteryl linoleate, and angiotensin I1 were purchased from Sigma. Egg for 2 h at room temperature. Finally, the remaining alkaline phosphatase activity was determined using p-nitrophenyl phosphate as subyolk phosphatidylethanolamine (PE), N-monomethyl-PE, and N,N-dimethyl-PE were purchased from Avanti Polar Lipids Inc. (Pelham, AL). strate. The results were read spectrophotometrically as optical density on an ELISA plate reader (Bio-Rad, model 2550) with the filterat 405 Alkaline phosphatase-conjugated goat anti-mouse IgG antibody and alkalinephosphatase-conjugatedgoatanti-mouse Ig(G+M) antibody nm. were purchased from Tag0 Inc. (Burlingame, CAI. Sheep anti-human In the sandwich ELISA for the detection of a complex of the antigen apoB antibody was purchased from Binding Site Inc. (Birmingham, with a polypeptide, each well of a 96-well plate was initially coated with monoclonal antibody for 2 United Kingdom). Mouse anti-muscleactin monoclonal antibody 0.5 pg of the partially purified FOHlalDLH3 (HHF35) and mouse anti-human macrophage antibody (HAM561 were h at room temperature. Afterblocking the well with TBS-skim milkfor 2 h a t room temperature, the reaction mixture of ferrous ion-induced purchased from ENZO Diagnosis Inc. (New York). Biotinylated rabbit anti-mouse Ig antibody, peroxidase-conjugated streptavidin, and alka- peroxidation of PC in the presence of a peptide was added to well the and line phosphatase-conjugated anti-alkaline phosphatase murine mono- incubated for 18 h at 4 "C. After washing three times with TBS-Tween clonal antibody were purchased from DAKO Japan Co. Ltd. (Kyoto, rabbit antiserum against the peptide (1:5000 dilutedin PBS) was reacted for 2 h at room temperature, followed by alkaline phosphatase-conjuJapan). Malonaldehydebiddimethylacetal)waspurchasedfrom (1:5000 diluted in TBS-skim milk) Aldrich. Acetic anhydride, p-nitrophenyl phosphate, and 3,3'-diamino- gated goat anti-rabbit IgG antibody phosbenzidine tetrahydrochloride (DAB) were purchased from Wako Pure for 2 hat room temperature. After washing the remaining alkaline as subChemical Co. (Osaka, Japan). A peptideof 3-hydroxy-3-methylglutaryl phatase activity was determined usingp-nitrophenyl phosphate coenzyme A (HMG-CoA) reductaseCterminus(HNRSKINLQNLQstrate on a n ELISA plate reader with the filtera t 405 nm. GTC) and anHMG-CoA synthetase C-terminal peptide(CTAAESESAVImmunohistochemical Analysis-Human thoracic aortae and coroISNGEH) were donated by Dr. Motojima of TohoUniversity, Funabashi, nary arteries were obtained from 18 autopsy cases where the donors Japan. Rabbit antisera against these synthetic peptides and angioten-had died of various diseases in hospital, witha n age range of 51 t o 84 sin I1 were prepared by immunization of the peptides in the popliteal years. Aortae and coronary arteries were cut longitudinally and translymph nodes of rabbits. versely, respectively, and rinsed with cold PBS, and specimens with Preparation and Modification of Lipoproteins-LDL and other li- atherosclerotic plaques were takenfor immunohistochemical investigapoproteins were separated from human plasma by a sodium bromide tion. The specimens werefmed in 4% paraformaldehyde in PBSfor 4 h stepwise density gradient centrifugation (31). All sodium bromide stock at room temperature and processed for paraffin-embedding following a solutions contained 0.25 m~ EDTA. After centrifugation the fractions standard procedure. Serialthin sections of 3 pm thickness were stained with a density of 1.019-1.063 were pooled as LDL. Prior t o oxidation of with hematoxylin and eosin and elastica-Masson's trichrome solution. Adjacent sections were stained immunohistochemically the LDL, it was passed through a [email protected]
column(Bio-Rad) to as follows. After remove the EDTA. LDL (0.2 mg/ml) was then oxidized by incubation blocking endogenous peroxidase activity with 0.3% hydrogen peroxide by addition in methanol for 30 min at room temperature, the specimens were prewith 5 p~ CuSO, at 37 "C for 3 h. The reaction was stopped of EDTA (final 0.25 mM). MDA-LDL was prepared by the method of incubated with 10% normal rabbit serum in PBS for 1 h and then Fogelman et al. (6). To 1 ml of LDL solution (0.5 mg/ml) in 50 mM incubated with the antibody FOHlalDLH3 (1:lOO diluted in PBS)for 2 phosphate buffer, pH 6.5, was added5 pl of 1M MDA solution (prepared h. They were carefully washed in PBS three times and exposed to from malonaldehyde biddimethyl acetal) by acidic hydrolysis just be- biotinylated rabbit anti-mouse Ig antibody(1:700 diluted in PBS)for 1 fore use) and the mixture incubatedfor 3 h at 37"C. The reaction was h. After washing in PBS the specimens were incubated with peroxidasestopped by extensive dialysis of the reaction mixture against phosconjugated streptavidin (1:500 diluted in PBS) for 1 h and deposition phate-buffered saline (PBS) containing 0.25 m~ EDTA at 4 "C. Acety- was visualized usingDAB. In the case of double immunostaining using lated LDL was prepared by the method of Basu etal. (32). Briefly, after monoclonal antibodies, HAM56 and HHF35, cobalt DAB and alkaline 1mg/ml of LDL solution was mixed with a n equal volume of saturated phosphatase methods were performed,respectively. sodium acetate, 1 p1 of acetic anhydride/mg of LDL was added to the Preparation of Various Oxidized Lipids-Total lipids of LDL were solution with continuous mixing. The reaction was carried out for 90 extracted using chloroform and methanol by the method of Bligh and min a t 0 "C, then the sample was dialyzed against PBS containing 0.25 Dyer (34). Lipids in chloroform were placed in plastic sample tubes. mM EDTA a t 4 "C. After the solvent was removed under a stream of argon gas,the sample Preparation of Monoclonal Antibodies-Monoclonal antibodies were was dispersed in PBS with a vortex mixer followed by a sonicator to prepared according to the method describedby Kimura et al.(33) with make 2m~ suspension. Ferrous sulfate and ascorbic acid were dissolved a slight modification. FemaleBALB/c mice were immunizedthree times in distilled water justbefore use. The reaction mixture in PBS containwith homogenates of human atheromatous plaquesof aortae, obtained ing 0.4 m~ lipids, 200 pg/ml BSA or peptides, 0.2 mM ferrous sulfate, 2 from autopsy cases over aperiod of 2 months. The human aortae from m~ ascorbic acid was incubated for 3 h at 37 "C. The reaction was autopsies were thegift from Dr. Numano of Tokyo Medical and Dental stopped by addition of butylated hydroxytoluene (BHT, final 0.2 mM). University, Tokyo, Japan. Spleen cells from the immunizedmice were Other AnalyticalMethods-The level of phosphorous from phosphofused with P3m1 murine myeloma cells and cultured in HAT (hypox- lipids were determined by the method of Zhou and Arthur (35). Thioantine, amethopterin, and thymidine) selection medium. Culture su- barbituricacid-reactivesubstances (TBARS) weremeasured by the pernatants of the hybridoma were screened using a n ELISA, employing method of Beuge and Aust (36). Protein concentration was measured by pairs of wells of microtiter plateson which were absorbedoxidized LDL the bicinchoninic acid (BCA) method (37).
investigators (15-17,28-301, have generated monoclonal antibodies t o oxidized LDL using a number of different experimental designs.Antigenic materials recognized by these monoclonal antibodies have been shown betopresent in homogenate of atheroma andthin section of atheromatous lesions of WHHL rabbits. In the present study, we used human atheromatous plaque as immunogen to obtain a monoclonal antibody that reacts with oxidized LDL. This was used to investigate whether materials deposited in human atherosclerotic lesions are immunologically related to oxidized LDL. The epitope of this monoclonal antibody was revealed to beoxidized product(s) of PC, in complex structures together with proteins. This monoclonal antibody also recognized foam cells derived from macrophages in human atherosclerotic intima.
LDL Monoclonal Antibody
15276 LDL OxLDL 0.5
Ac-LDL MDA-LDL HDL
Immunoreactivity (OD405 at 10 min)
FIG.1. Reactivity of the monoclonal antibodyFOHlaDLH3 to various modified lipoproteins. Microtiter wells were coated with 1 1.18 each of native and modified lipoproteins. F O H l a L H 3 (ascites fluid) was diluted 1:10,000. The direct ELISA was performed as described under “Experimental Procedures.” OxLDL 0.5, OxLDL 3 , and OxLDL 24 represent the LDL treated with CuSO, at 37 “C for 30 min, 3 h, and 24 h, respectively. OxHDL 3 represents the HDL treated with CuSO, at 37 “C for 3 h. Ac-LDL represents the acetylatedLDL. Values are mean f S.D. of three experiments. RESULTS
Specificity of the Monoclonal Antibody FOHla lDLH3During the preparation of the monoclonal antibodies, hybridomas were selected by comparing the reactivities of the culture supernatant tooxidized LDL and nativeLDL. Among the eight clones obtained, the clone FOHla/DLH3 showed the most distinctive recognition of oxidized LDL against native LDL. Fig. 1 shows the specificity of this monoclonal antibody to various native and modified forms of lipoproteins. Although there was a strong reaction to oxidized LDL, almost no reactivity was observed against other forms of LDL, including native LDL, acetylated LDL, and MDA-LDL. The LDL treated with CuSO, for only 30 min was not reactive, although a small amount of TBARS was, detected (data not shown). The antibody crossreactedwith oxidized HDL but not with native HDL. This suggests that a specific structure of the apoB is not essential for antigen recognition by the antibody and that peroxidation products of lipids in LDL and HDL may be the feature recognized by the antibody. The reactivity of the oxidized HDL to the antibody was notaltered when apoB was eliminatedcompletely by preincubation with anti-human apoB antibody and protein A-agarose before CuSO, treatment (data not shown). The isotype on the monoclonal antibody was IgM with K light chain. Localization of the Antigenic Material in Foam CellsSequential thin sections of human atheromatousplaques from the left coronary arteries were subjected to immunohistochemical analysis. FOHldDLH3 recognized some histological compositions in atheromatous plaques, such as monocyte-derived macrophages, debris in atheromatous necrotic cores, swollen collagen fibers, and endothelial cells (Fig. 2C). The adjacent section was doubly stained with two monoclonal antibodies HAM56 and HHF35, specific for macrophages and smooth muscle cells, respectively (Fig. 2, D and F). Fine eosinophilic membranous structures surrounding lipid droplets in the cytoplasm of lipid-laden macrophages (foam cells) were stained conspicuously with the antibody. A higher magnification staining revealed that fine eosinophilic membranous structures surrounding variously sized lipid droplets in foam cells were strongly and exclusively stained (Fig. 2E). The foam cells, which are positive against HAM56 monoclonal an-
tibody, appeared and accumulated in fibrousregions in the vicinity of atheroma. FOHla/DLH3also stained cellular debris dispersed in the necrotic cores. Swollen collagen fibers, either in a fibrous cap on atheromas or in atheromas,were faintly but positively reactive to the antibody. Spindle-shaped intimal smooth muscle cells scattered in the plaque, about 60% of medial smooth muscle cells, and endothelial cells were moderately stained. These observations suggest that the antigen recognized by FOHlaDLH3 is located in foam cells derived from macrophages and in several other locations within the human atherosclerotic lesions. Involvement of Oxidized Phosphatidylcholine-As a first step identifying the antigenic material in the oxidized lipoproteins, model antigens were prepared by peroxidation of various lipid standards, by incubating with ferrous ion and ascorbic acid in the presence of BSA at 37 “C for 3 h. Since apoB does not seem to be essential for the specificity of the antibody, we replaced apoB in this model system with other polypeptides such as BSA. When the total lipid fraction extracted from native LDL was incubated with FeSO, and ascorbic acid, antigenic material was produced (Fig. 3A). A combination of ferrous ion and ascorbic acid is known to induce lipid peroxidation(36-38), and indeed the production of TBARS was detectedduring this treatment (see Fig. 3B). Incubation of the total lipid extracts without FeSO, resulted in no reactivity to the antibody. The antigenic material(s) was produced when egg yolk PC was peroxidized in the reaction system, whereas other lipids such as trilinolein, cholesteryl linoleate, egg yolk PE, and bovine brain sphingomyelin did not react with theantibody after treatment (Fig. 3A). Production of the antigenic material(s) was formed from the PC species containing polyunsaturated fatty acids. The production of the antigen was not correlated with their TBARS values (Fig. 3B). N-Monomethyl-PE and N,N-dimethyl-PE also failed t o produce the antigen (data not shown). Since these phospholipids were prepared from egg yolk PC by the head-exchange reaction,their fattyacid compositions are essentially the same as thatof egg yolk PC. Therefore the choline head group is likely to be required for antigen recognition. Formation of a Complex with Polypeptide-The effects on the antigen production of BSA added t o the reaction system was examined. 1-Palmitoyl-2-linoleoyl-PCwas incubated with FeSO, and ascorbic acid in the presence or absence of BSA at 37 “C for 3 h and the reaction stopped by addition ofBHT. Then, to thereaction mixture without BSA, the corresponding amount of BSA was added and incubated for a further 30 min (Fig. 4).ELISA analysis showed that theantibody reacted with the incubation mixture in the presence ofBSA in a dose-dependent manner. The oxidized PC mixed with BSA after the peroxidation reaction had been stopped showed much lower immunoreactivity. Prolonged incubation of the pre-oxidized PC and BSA did not increasethe immunoreactivity (data not shown). It is probable that a polypeptide such asBSA enhances the immunoreactivity of the oxidized PC through modification of the polypeptide and that this modification takes place during the lipid peroxidation reaction (see “Discussion”). Since apoB is not required for antigen recognition of the antibody FOHldDLH3, model antigen may be formed from an oxidized PC and polypeptide. To test whether an antigencomplex is formed from oxidized PC and a peptide, a synthetic 15-mer peptideof HMG-CoA reductase C terminus wasutilized as a probe of this model system. When 1-palmitoyl-2-linoleoyl-PC was oxidized with FeSO, and ascorbic acid in the presence of the peptide, the product recognized by both F O H l d DLH3 antibody and the antiserum against the HMG-CoA reductase C-terminalpeptide was clearly detected on a sandwich ELISAassay (Fig. 5, PC-Fe-Red). When PC was incubatedwith
FIQ. 2. Immanohistoohemicaldetection of the FOHla/DIBtI antigen in foam cells in human atherosclerotic lesiom.Serial t e sections of an atheromatous plaque of the leR coronary artary of an Sdyear-old male were stained with hematoxylin and eosin (A), elastica-
Masson's trichrome (B),anti-oxidized LDL monoclonal antibodyFOHldDLH3 (Cand E), and monoclonal antibodiesHAM66 ( s m c for human (sped% for m d e actin, red) (D and F).E and F are higher magnificationsof C add D, rekpectively. macrophages, bh&) and -35
'(A) ELlSA LDL lipids
brain SM Trlllnolain Cholestryl linoleate
160.18:l PC 16:O-182 PC
1 6 0 4 8 2 PC
0.4 0.8 0.6
Immunoreactivity (OD405 at 30 min)
TBARS (nmol I pnol lipids)
FIG. 3. Formation of the FOHla/DLE8 antigen during peroxidation of various lipids with ferrous ion and ascorbic acids in the pr88enc.eof BSA.Variow lipid eamples suspended in PBS (0.4 m ~were ) i n d a t e d with or without FeSO, ([email protected]
)in the presence of =corbie acid (0.4m ~and ) BSA (1 pg of proteinJnmol of lipid) at 37 "C for 3 h. Concentration ofthe "LDL lipids* represents that of lipid phosphorous. Reactivity to the FOHla/DLH3on EJJSA (A) and TBARS values ( B )weredetermined. Values are mean S.D.of three experiments.
another peptide (an HMGCoAsynthetase C-terminal peptide), no reactivity wasobserved, indicating that theantipeptide ant i s 0 m was specific, and nonspecific binding of the antiserum was minimal (Fig. 5, PC-Fe-Syn). No reaction was observed when the peptide alonewas added to the assay system (Fig.5, Red and Syn). These results demonstrate the production of a complex of antigenic material of FOHlalDLH3and the peptide. Formation of the complex of the antigenic oxidized PC and the peptide was again detected using other peptides, a synthetic 16-mer HMGCoA synthetase C-terminal peptide and angiotensin 11, together with the appropriate antisera against these peptides (datanot shown). The Oxidized PC MoietyIs Suficient for Antigen Binding of the Monoclonal Antibody FOHlaIDLH3"We carried out a competition assay using the sandwich ELISA methodto 5ee if the peptide portionof the complex is required forantigen recognition (Fig. 6). As a competitor 1-palmitoyl-2-linoleoyl-PC
was incubated with FeSO, in the absence of any polypeptide. When the complex of oxidized PC and theHMGCoA reductase C-terminal peptide was addedto the wells of microtiter plates together withexcess amount of the oxidized PC as a competitor, the immunoabsorption of the antigen complex was inhibited in a dodependent manner. No inhibition was observed whenPC incubated without FeSO, was added as competitor. These results suggest that theantigenic oxidized PC product(s) forms a complex with peptides, but the antibody does not necessarily require the complex formation for its antigen recognition. DISCUSSION
In this study, we have obtained a new murine monoclonal of the antibody that recognizesoxidizedlipoproteins.One monoclonal antibodies, F O H l a L H 3 (named after the key LDL, H-human atheroma) words,F-foamcells,0-oxidized clearly distinguishedoxidized lipoproteinsfrom acetylated and
Anti-human Oxidized LDL Monoclonal Antibody I.4
1.2 1 .o
0.8 0.6 0.4 0.2
Amount of antiqen coated (nmol PC) Amount of competitor (fold excess of the OxPC-peptidecomplex ) FIG.4. Effect of BSA on the antigen production during the peroxidationof PC with ferrousion and ascorbic acids. 1-PalmiFIG.6. Oxidized PC inhibited the bindingof the oxidizedPCtoyl-2-linoleoyl-PC suspended in PBS (0.4 mM) wasincubatedwith peptide complex to the monoclonal antibody FOHla/DLH3. The (closed circle) or without (open circle) BSA (1 pghmol lipid) in the antigen complex of oxidized PC and theHMG-CoA reductase C-termipresence of ascorbic acid (0.4 mM) and FeSO, (40 PM) at 37 “C for 3 h. nal peptide was prepared as in the legend of Fig. 5. Competitors were The reaction was stoppedby adding BHT. Then to the reaction mixture prepared by incubating 1-palmitoyl-2-linoleoyl-PC liposome in PBS(0.4 without BSA, the corresponding amount of BSA was added and incu- mM) with or withoutFeSO, (40 p ~ in) thepresence of ascorbic acid (0.4 bated for a further 30 min (closed triangle).Reactivity to the antibody mM) at 37 “C for 3 h.To 1nmol PCof the antigencomplex, 1, 10, and40 FOHlaiDLH3 on ELISA. nmol of oxidized PC (closed circle)or PC(open circle)were addedbefore samples were transferred to the microtiter plates precoated with the antibody FOHlaiDLH3. Values are mean f S.D. of three experiments.
FIG.5. Formation of a complexof antigenic oxidized PC and an HMG-CoA reductase C-terminal peptide. 1-Palmitoyl-2-linoleoyl-PC liposome in PBS(0.4 mM) was peroxidized with FeSO, (40 PM) and ascorbic acid (0.4 mM) in the presence of an HMG-CoA reductase C-terminal peptide(1pghmol lipid) at 37 “C for 3 h. A sandwich ELISA was performed using FOHlaiDLH3 and antiserum against the HMGCoA reductase C-terminal peptide as described under “Experimental Procedures.” As a control another peptide with totally different sequence, an HMG-CoA synthetaseC-terminalpeptide,wasused. Samples used as antigen in the sandwich ELISA were (from the left column): PC-Fe-Syn,PC peroxidized in the presence of the HMG-CoA synthetase C-terminal peptide;PC-Fe-Red,PC peroxidized in the presRed, the HMG-CoA ence of the HMG-CoAreductase C-terminal peptide; reductase C-terminal peptide alone;Syn, the HMG-CoA synthetase Cterminal peptide alone. Values are mean * S.D. of three experiments.
MDA-treated ones. Moreover, it appears that the antigen is primarily located in the foam cells derived from macrophages, with littlefound in extracellularcomponents of atherosclerotic lesions, except in the cellular debris of necrotic cores. It is noteworthy that the extracellular space of the lesion was not strongly stained with the antibody, showing that it has a different distribution from that of apoB in the lesion (39, 40). We developed a model antigen-producing system to characterize the epitope of the monoclonal antibody. The observation that the oxidized HDL and oxidized LDL were almost equally
reactive suggested that theepitope might be derived from lipids, so ferrous ion-induced lipid peroxidation was used in the model system. It wasshown that the antigenic material(s) was produced during ferrous ion-induced peroxidation of PC in the presence of BSA, whereas none of the other lipids formed the antigenic product. We then used a sandwich ELISA t o investigate the relationshipbetween the lipid peroxidation product(s) and polypeptides. If the antigenic product from PC forms a complex with a peptide, the peptidewill be associated with the monoclonal antibody, although it does not recognize the peptide. The sandwich ELISA experiment clearly showed the formation of a complex of oxidized PC with thepeptide. Finally, a competition experiment was performed to determine whether the oxidized PC molecule is sufficient for recognition by the antibody. This is the first demonstration of modification of apolipoproteins by oxidized phospholipids formed in oxidized LDL and the detection of such material in atherosclerotic lesions, specifically in foam cells. Several PC derivatives formed during peroxidation have been identified (41-44).When hydroperoxides are cleaved t o form short chain products, PCs containing a short chain aldehyde group or a short chain carboxyl group are produced. PC containing a short chaincarboxyl group hasbeen identified as a cytotoxic product generated duringhemoglobin-induced peroxidation of PC (43).PCs containing short chain aldehyde and short chain carboxylic acids have also beenshown to have platelet-activating activity (41, 42). Those products are considered to be possible epitopes of the antibody. We are currently workingon the isolation and identification of the chemical structure of the epitope derived from PC. Although the monoclonal antibody bindst o oxidized PC molecule(s), the addition of BSA enhanced the immunoreactivity. BSA may increase the absorption of the antigenon the surface of microtiter wells and the accessibility of the antibody to the oxidized PC molecules. Immunoreactivitywasrather poor when the complexes of oxidized PC and small peptides were directly coated onto the wells (data not shown). Extraction of the antigenic oxidized PC from the oxidized LDL with organic solventswasunsatisfactory(datanot
Anti-human Oxidized LDL Monoclonal Antibody shown). The antigenic oxidized PC may not necessarily be extractable, since it can form a complex with apoB. The free form of the oxidized PC, which is presumed to be the extractable form, may be less efficient forthe direct ELISA method, as PC oxidized in theabsence of BSA was not effectively detected (see above). It may be that hydrophobic interaction between the oxidized PCand thesurface of the microtiter wells is too weak for tight binding or that possibly absorption of the oxidized PC on the plastic surface may cause significant steric hindrance to antibody binding. Since specimens subjected to immunohistochemical analysis were embedded in paraffin, the lipids deposited in foam cells were washed out during the procedure, suggesting that the antigen(s) detected in these specimens is in the form of an oxidized PC-polypeptide complex. It has been reported that acetylated LDL is degraded almost completely after incorporation in mouse peritoneal macrophages, whereas oxidized LDL is rather resistant to lysosomal hydrolysis (45-47). "he modification of lipoproteins by oxidized PC may alter its susceptibility to lysosomal hydrolytic enzymes and result in accumulation of the antigen fragments. "he possibility cannot be ruled out that the antigen(s) in the foam cells is not necessarily derived from oxidized LDL. When cells die and their antioxidant systems fail, lipid peroxidation may occur. Under these conditions complexesof antigenic oxidized PC and polypeptides other than oxidized LDL recognized by the monoclonal antibody might be formed. We and several other investigators (15-17,26-28) have generated monoclonal antibodies to modified LDL by a variety of experimental design. Rather diffuse staining patterns were demonstrated in WHHL atheroma using monoclonal antibodies against MDA-LDL (15,17) andoxidized LDL (16). Witzumand co-workers (17, 30) obtained a monoclonal antibody against delipidated oxidizedLDL fragments. "his antibody stained foam cells in aortas of WHHL rabbits, although they have not characterized the antigen further. "he present study revealed the occurrence and localization of a novel type of oxidized LDLrelated modification in human atherosclerotic lesions. "he monoclonal antibody FOHla/DLH3 would be a good tool to investigate the occurrence of oxidized LDL in vitro and in vivo. Further study is required to understand the metabolic fate of oxidized LDL in foam cells. Acknowledgments-We thank Dr. F. Numano of Tokyo Medical and Dental University for the kind gift of human atheromatous plaques from autopsies and Dr. Motojima of Toho University for the synthetic peptides of HMG-CoA reductase and synthetase. REFERENCES 1. Witzum, J. L., and Steinberg, D. (1991) J. Clin. Znuest. 88, 1785-1792 2. Yla-Herttuala, S., Palinski, W., Rosenfeld, M. E., Parthasarathy, S., Carew, T. E., Butler, S., Socher, S . A, Witzum, J. L., and Steinberg, D. (1989) J. Clin. Znuest. 84,1086-1095 3. Brown, M. S.,and Goldstein, J. L. (1983)Annu.Reu. Biochem. 52,223-261 4. Mahley, R. W., Innerarity, T. L.,Weisgraber,K. H., and Oh, M. 5. (1979) J. Clin. Znuest. 61, 743-750 5. Graham, A., Hogg, N., Kalyanaraman, B., OLeary, V.,Darley-Usmar, V., and Moncada, S . (1993) FEES Lett. 330,181-185 6. Fogelman, A.M., Shechter, I., Seager, J., Hokom,M., Child, J. S., and
Edwards, P. A. (1980) Pmc. Natl. Acad. Sci. U.S. A. 77,2214-2218 7. Vijayagopal, P., Srinivasan, S . R., Jones, K. M., Radhakrishnamurthy, B., and Berenson, G. S . (1985) Biochim. Bwphys. Acta 837,251-261 8. Steinbrecher, U.P., Parthasarathy, S . , h a k e , D. S., Witzum, J. L., and Steinberg, D. (1984) Proc. Natl. had.Sci. U.S. A. 81,3883-3887 9. Goldstein, J. L., Ho, Y. K., Basu, S . K., and Brown, M. S . (1979) Proc. Natl. Acad. Sci. U.S. A. 76, 333-337 10. Kodama, T., Reddy, P., Kishimoto, C., and Krieger, M. (1988) Proc. Natl. Acad. Sci. U.S. A 86,9238-9242 11. Kodama, T., Freeman, M., Rohrer, L., Zabrecky, J., Matsudaira, P., and Krieger, M. (1990) Nature 345,531635 12. Matsumoto, A,, Naito, M., Itakura, H., Ikemoto, S . , Asaoka, H., Hayakawa, I., Kanamori, H., Aburatani, H.,Takaku, F., Suzuki, H., Kobari, Y., Miyai, T., Takahashi, K., Cohen, E. H.,Wydro,R., Housman, D., and Kodama, T. (1990) h. Natl. Acad. Sei. U.S. A. 87,9133-9137 13. Stanton, L.W., White, R. T., Bryant, C. M., Rotter, A. A,, and Endemann, G. (1992) J. Biol. Chem. 267,2244€-22451 14. Endemann, G., Stanton, L. W., Madden, K. S., Bryant, C. M., White, R.T., and Protter, A. A. (1993) J. Biol. Chem. 268,11811-11816 15. Harberland, M. E., Fong, D., and Cheng, L. (1988) Science 241,215-218 16. Boyd, H. C., Gown, A. M., Wolfbauer, G., and Chait, A. (1989)Am. J. Pathol. 135,815-825 17. Rosenfeld, M. E., Palinski, W., Yla-Herttuala, S., Butler, S., and Witzum, J. L. (1990) Arteriosclerosis 10, 336-349 18. Sattler, W., Kostner, G.M.,Wang,G., and Esterbauer, H. (1991) Biochim. Biophys. Acta 1081, 65-74 19. Hoff, H. F., and O'Neil, J. (1991)AtherosclerosisThromb. 11, 1209-1222 20. Hoff, H. F., Whitaker, T. E., and O'Neil, J. (1992) J. Biol. Chem. 267,602-609 21. Bhadra, S.,Arshad, M. A. Q., Rymaszewski, 2.. Norman, E., Wherley, R.,and Subbiah, M.T.R. (1991) Biochem. Biophys. Res. Commun. 176,431-440 22. Wang, T., Yu, W.-G., and Powell, W. S . (1992) J. Lipid Res. 33, 525-537 23. Kamido, H., Kuksis, A., Marai, L., and Myher, J. J. (1992) FEES Lett. 304, 269-272 24. Mangin, E. L., Jr., Kugiyama, K, Nguy, J. H., Kerns, S . A,, and Henry, P.D. (1991) Circ. Res. 72, 161-166 25. Steinbrecher, U. P., and Pritehard, P. H. (1989) J. Lipid Res. 30,305315 26. Wang, T., and Powell, W. S . (1991) Biochim. Biophys. Acta 1084,129-138 27. Carpenter, K. L. H., Taylor, S . E., Ballantine, J. A,, Fussell, B., Halliwell, B., and Mitchinson, M. J. (1993) Biochim. Biophys. Acta 1167,121-130 28. Mowri, H., Ohkuma, S., and Takano, T. (1988) Biochim. Bwphys. Acta 963, 208-214 29. Gonen, B., Fallon, J. J., and Baker, S. A. (1987)Atherosclerosis66,265-272 30. Plainski, W., Yl8-Herttuala, S., Rosenfeld, M. E., Butler, S., Socher, 5. A,, Curtiss, L. K., and Witzum, J. L. (1990) Arteriosclerosis 10, 325335 31. Redgrave, T. G., R o b e r t a , D.C., and West, C. E. (1985) Anal. Biochem. 66, 42-49 32. Basu, S . K., Goldstein, J. L., Anderson, R. G. W., and Brown, M. S . (1976)P m . Natl. Acad. Sci. U.S. A. 73, 31783182 33. Kimura, J.,Nakagami, K., Akanuma, K , Ohkuma, S., Yoshida, Y.,and Takano, T. (1986) Virchows Arch.A 410,159-164 34. Bligh, E.A., and Dyer, W. J. (1959) Can. J. Biochem. Physiol. 37,911-917 35. Zhou, X.,and Arthur, G. (1992) J. Lipid Res. 33,1233-1236 36. Buege, J. A,, and Aust, S . D. (1978) Methods Enzymol. 52,302-310 37. Smith, D. K., Krohn, R. I., Hermandon, G.T., Mallia, A. K., Gartner, F. H., Provenzano, M. D., Fqjimoto, E. K., Goeke, N. M., Olsen, B. J., and Klerk, D.C. (1985) Anal. Biochem. 160,76-85 38. O'Brien (1969) Can. J. Biochem. 47,485-492 39. Galis, 2. S., Ghitescu, L. D., Li, Z., Misbahuddin, Z., and Moore, S . (1992) Lab. Znuest. 66, 624-638 40. Trakht, I. N.,Kovaleva, K A., Janushevskaya, E. V., Balmukhanov, T. S., F'rintseva, 0.Y., Preobrazhensky, 5. N., Peklo, M. M., Tsibulsky, V. P., and Prokrovsky, S . N. (1986) Atherosclerosis Reu. 17, 51-66 41. Tanaka, T., Minamino, H., Unezaki, S., Tsukatani, H., and lbkumura, A. (1993) Bioehim. Biophys. Acta 1166,264-274 42. lbkumura, A., Asai, T., Takauchi, K., Kamiyasu, K., Ogawa, T.,and Tsukatani, H. (1988) Biochem. Biophys. Res. Commun. 155,863469 43. Itabe, H., Kushi, Y., Handa, S., and Inoue, K. (1988) Biochim. Biophys. Acta s62,8-15 44. Stremler, K. E., StaEorini, D.M., F'rescott, S . M., Zimmerman, G. A,, and McIntyre, T.M. (1989) J. Biol. Chem. 264, 5331-5334 45. Lougheed, M.,Zhang, H., and Steinbrecher, U.P. (1991) J. Biol. Chem. 266, 14519-14525 46. Jessup, W., Mander, E. L., and Dean, R. T. (1992) Biochim. Biophys. Acta1126, 167-177 47. Roma, P., Bernini, F., Fogliatto, R., Bertulli, S . M., Negri, S., Fumagalli, R., and Catapano, A. L. (1992) J. Lipid Res. 33, 819-829