In Vitro Interactions of Oxidatively Modified LDL With Type I, II, III, IV, and V Collagen, Laminin, Fibronectin, and Poly-d-Lysine
Abstract The accumulation of LDL in the arterial intima is considered a key event in atherogenesis. We investigated the binding of oxidized LDL (ox-LDL) to microtiter plates coated with type I or II collagen, laminin, fibronectin, or poly-d-lysine. Oxidation of LDL, 125I-LDL, or Eu3+-LDL was performed with CuCl2, varying the time of oxidation. Bound lipoprotein was assessed by counting radioactivity or fluorescence in the wells. Binding of highly ox-LDL in PBS followed the order: type I collagen>poly-d-lysine>type II collagen>laminin>fibronectin. Comparing various collagen types, the binding of ox-LDL followed the order: type I>type V and, type III>type IV>type II collagen. Binding of ox-LDL in PBS was dependent on an increase in negative charge of ox-LDL. Testing certain amino acids as competitors for binding of highly ox-LDL to type I collagen put lysine first, followed by arginine and histidine. On laminin, histidine competed most, followed by lysine and arginine. When studying the influence of Na+, K+, Ca2+, Mg2+ (equivalent to their concentrations in the interstitial fluid), native LDL, moderately ox-LDL, and highly ox-LDL showed the same affinity to type I collagen. However, a fivefold dilution of the buffer increased the affinity of moderately and highly ox-LDL 3.9- and 10-fold compared with native LDL. Application of the F(ab′)2 from a monoclonal antibody to ox-LDL revealed a strong competition of the binding of highly ox-LDL to type II collagen (60%), laminin (35%), type I collagen (20%), and poly-d-lysine (15%), whereas the binding to fibronectin was not affected.
- Received February 19, 1996.
- Accepted July 17, 1996.
Binding of LDL to connective tissue of the arterial intima, such as various types of collagen, elastin, and glycosaminoglycans, was found to play an important role in the formation of atherosclerotic lesions.1 2 3 4 It was suggested that this binding mechanism leads to a deposition of LDL in the vascular wall during atherogenesis, independent of cellular uptake and foam cell formation. LDL is assumed to become mildly oxidized when entering the arterial wall and to recruit monocytes, which become resident macrophages.5 These cells, as well as smooth muscle cells and endothelial cells, are able to generate free-radical species, which in turn oxidatively modify LDL. On the one hand, this modified form of LDL was shown to be recognized by scavenger receptors on macrophages.6 7 Accumulation of lipids in macrophages leads to an enhanced transformation of these cells to so-called foam cells, which are typical cells in early atherosclerotic lesions.8 On the other hand, ox-LDL was shown in vitro to bind to collagen gels.9 10 11 The structure of two scavenger receptors, recently explored by cloning cDNA from bovine lung library and designated type I and type II, was found to contain a collagenous domain, which plays a key role in binding chemically modified proteins, such as acetylated LDL, ox-LDL, and maleylated bovine serum albumin.12 13 Thus, the scenario was suggested that LDL bound to collagen, became oxidized, and would afterwards be taken up by macrophages gliding along collagen fibrils.14
So far, binding of native or ox-LDL to collagen has been estimated by measuring the diffusion of iodinated LDL in collagen gels.9 10 11 As to the influence of the negative charge of ox-LDL on its binding to various types of collagen, one sort of ox-LDL was used in one investigation,10 whereas in another study the affinity of type I collagen to ox-LDL modified to different degrees was reported.11 In the present study, a special form of sandwich assay was used to study the binding of 125I- or Eu3+-labeled LDL—either in their native form or after copper-mediated oxidation—to type I, II, III, IV, and V collagen and to poly-d-lysine. In addition to these collagen phenotypes and poly-d-lysine, we investigated for the first time the binding of ox-LDL to laminin and fibronectin. These two proteins are also important components of connective tissue. They enable endothelial cells to attach to underlying extracellular matrix by binding to collagen and they interact with fibrin, heparin, and integrins. The experiments were performed either in PBS only or in presence of divalent cations of calcium and magnesium corresponding to their concentrations in the interstitial fluid. Since histidine was reported to inhibit the binding of ox-LDL to collagen I,11 this amino acid, as well as lysine and arginine, was used as competitor. Furthermore, a mAb to ox-LDL was used to test whether the interactions of ox-LDL with the selected proteins can be interfered with by masking the epitopes newly generated on ox-LDL.
Type I and II collagen were prepared at the Department for Ophthalmology as described below. Type III, IV, and V collagen, laminin (free of enactin), fibronectin, and poly-d-lysine hydrobromide were purchased from Becton Dickinson. Type III and V collagen were from human placenta and were electrophoretically homogenous, according to the manufacturer. Type IV collagen and laminin were from Engelbreth-Holm-Swarm mouse tumor and of 99% and >96% purity in SDS-PAGE, according to the manufacturer. Fibronectin was from human plasma and of >85% purity in 5% SDS-PAGE under reducing conditions, according to the manufacturer. Anti-rabbit IgG and anti-mouse IgG were purchased from Sigma, rabbit antiserum against human apoB was from Behring AG, and the Eu3+-labeling kit (DELFIA, No. 1244-302) was from Wallac Oy.
Preparation of Connective Tissue Protein Components
Preparation of type I collagen was performed as described.15 16 Briefly, bovine cornea was treated twice with 0.5 mol/L sodium acetate at 4°C (10 mL acetate per gram wet weight) for 24 hours, washed with water, and extracted with two changes of 0.5 mol/L citrate buffer, pH 3.7 (10 mL citrate per gram wet weight), for 48 hours. After centrifugation, the pellet was solubilized with 5 mg pepsin per milligram wet weight (2500 U/mg; Boehringer) at 4°C for 24 hours and centrifuged again. The pepsin-solubilized collagen was precipitated by adding solid NaCl to a final concentration of 0.9 mol/L, centrifuged for 10 minutes at 3000g (4°C), and redissolved in 50 mmol/L Tris-HCl, pH 7.5, containing 1 mol/L NaCl. The collagen solution was kept at 4°C for 4 days to inactivate pepsin, dialyzed against 0.167 mol/L acetic acid several times, and dried in air after acetone precipitation. For preparation of type II collagen, bovine articular cartilage was treated with sodium acetate and citrate buffer, respectively, as described above. After centrifuging the 0.5 mol/L citrate buffer, extraction of type II collagen was performed as described by Miller.16 The purity of both collagen preparations was checked by SDS-PAGE. The content of glycosaminoglycans was <1%, as determined by the method of Bitter and Muir.17
For this study, type I, II, III, IV, and V collagen and poly-d-lysine were dissolved in 10 mmol/L PBS, pH 7.4, containing 0.5 mol/L acetic acid. For laminin and fibronectin, 10 mmol/L PBS, pH 7.4, was used as solvent. The concentration of these proteins was 1 mg/mL.
LDL was isolated from the plasma of normolipemic, fasting (12 to 14 hours) young male donors with serum lipoprotein(a) levels <1 mg/dL. Kallikrein inactivator (aprotinin, 100 000 U/L; Bayer), Pefabloc (11.2 mg/L; Merck), and EDTA (1 g/L; Merck) were present during LDL preparation by differential ultracentrifugation at a density range between 1.020 and 1.063 g/mL.
Protein of LDL was measured by the method of Lowry et al,18 sterile filtered, and stored at 4°C.
Labeling of LDL With 125I and Eu3+
LDL (6 mg protein) was labeled by the sodium 125I radioiodination method, using N bromosuccinimide.19 After purification on a Sephadex PD 10 column (Pharmacia), LDL was dialyzed against 10 mmol/L PBS (pH 7.4). The final specific activities ranged from 100 to 200 cpm/ng LDL protein; free radioactivity was less than 3% of the total radioactivity. 125I-LDL was stored in the dark at 4°C and used within 3 weeks.
Eu3+ labeling of LDL (1 mg protein) was performed in 50 mmol/L NaHCO3, pH 8.3 to 8.5, containing 20 μmol/L Trolox (Hofmann LaRoche). LDL was incubated with 0.2 mg of Eu3+ chelate of N1-(p-isothiocyanatobenzyl)-diethylenetriamine-N1, N2, N3, N3-tetraacetic acid (DELFIA europium-labeling kit; Wallac Oy) at 25°C in the dark for 12 hours. Sephadex G-25 chromatography (Pharmacia Biotech) was used for separation of labeled protein from free chelate in 50 mol/L Tris-HCl, pH 7.8, containing 0.05% NaN3 and 20 μmol/L Trolox. The labeling yield of Eu3+-LDL was between 4 and 22 Eu3+/protein (mol/mol). Eu3+-LDL was used within 3 weeks.
Cu2+-Mediated Oxidation of LDL, Eu3+-LDL, and 125I-LDL
Before oxidation, LDL, Eu3+-LDL, and 125I-LDL were dialyzed against 0.01 mol/L PBS, pH 7.4, which was carefully degassed and then saturated with nitrogen. Cu2+-mediated oxidation of LDL (1 mg/mL) was performed at 37°C with 20 or 30 μmol/L CuSO4. At intervals between 0 and 24 hours, the reaction was terminated by adding a stop solution to achieve a final EDTA concentration of 2.7 mmol/L. The samples were saturated with nitrogen and stored at 4°C.
The degree of modification of LDL was estimated as the REM, ie, relative to the nonmodified and nonlabeled native LDL, on agarose gels (1%) at pH 8.05, using the Lipidophor system (Immuno AG). In some samples, lipid peroxides were estimated by a spectrophotometric assay with CHOD-iodide color reagent (Merck) at 365 nm, as developed recently in this laboratory.20
To check for possible changes of the fluorescence intensity of Eu3+-LDL occurring during oxidation, the samples (1 mg/mL LDL protein) were applied on polystyrene microtitration plates (Maxisorb, Nunc), diluted 1:10 000 with enhancement solution. The fluorescence counts were measured with the DELFIA research fluorometer (Wallac Oy). The influence of oxidation of 125I-LDL was monitored by application of the oxidized 125I-LDL samples, diluted in a 1:50 scintillation cocktail, on polystyrene microtitration plates and counting the radioactivity on a microplate scintillation counter (Packard Instruments).
Monoclonal and Polyclonal Antibodies Against LDL, Ox-LDL, 4-Hydroxynonenal-Modified (HNE)-LDL, Malondialdehyde-Modified (MDA)-LDL, Chromogranin, and Human IgG1
A mAb OB/04 against ox-LDL, which was recently generated and characterized in this laboratory,21 was purified by protein A Sepharose CL-4B (Pharmacia Biotech) affinity chromatography and then lyophilized. Fragmentation of OB/04 to F(ab′)2 was established by incubation of 1 mg OB/04 in 10 μg pepsin (Sigma) in 0.1 mol/L acetic acid, pH 4.5, for 30 hours at 37°C. Digestion was stopped by adding 0.1 vol 2 mol/L Tris-HCl, pH 8.0. F(ab′)2 was separated from intact OB/04 and Fc by protein A Sepharose CL-4B (Pharmacia Biotech).
Two other monoclonal antibodies from mice were used for control experiments: anti-chromogranin, from Immunotech SA, and anti-human IgG1, from The Binding Site Limited. Their fragmentation was performed essentially as described above for F(ab′)2 from OB/04.
Purification of antisera against apoB (Behring AG), HNE-LDL, and MDA-LDL,22 23 all from rabbit, was performed on a DEAE Sepharose fast flow (Pharmacia Biotech). After dialyzing the antisera against 50 mmol/L Tris-HCl, pH 8.2, and 40 mmol/L NaCl, they were applied on the column and eluted with 50 mmol/L Tris-HCl, pH 8.2, and 40 mmol/L NaCl, monitoring at 280 nm. IgG fractions were collected, dialyzed against 50 mmol/L Tris-HCl, pH 8.2, containing NaN3 at a final concentration of 0.01%, and stored at 4°C. The purity of all antisera was checked by SDS-PAGE.
Solid-Phase Sandwich Fluorescence Assay
For the solid-phase fluorescence sandwich assay, polystyrene microtitration plates were coated with 200 μL per well of the proteins (10 μg/mL type I, II, III, IV, and V collagen, laminin, fibronectin, and poly-d-lysine) in 10 mmol/L PBS, pH 7.4 at room temperature overnight. After washing twice with washing buffer (10 mmol/L PBS, pH 7.4; 0.9 g/L NaCl; 0.05% Tween 20; and 0.02% NaN3), 200 μL blocking buffer (10 mmol/L PBS, pH 7.4, and 1 g/dL bovine serum albumin) was added to the wells to block remaining nonspecific binding sites. The wells were then washed three times, and 200 μL of Eu3+-LDL oxidized to different degrees (5 to 10 μg/mL LDL) was incubated in 10 mmol/L PBS, pH 7.4, for 90 minutes at room temperature on a shaker. After six washes with washing buffer, Eu3+ was released with 200 μL per well enhancement solution (Wallac Oy), and fluorescence was measured with a DELFIA research fluorometer (Wallac Oy). In the case of type I collagen, poly-d-lysine, and laminin, the wells were also washed eight times with washing buffer before counting fluorescence. This additional washing did not weaken the binding of oxidized Eu3+-LDL to the wells, indicating a stable binding of the lipoprotein with the proteins used for coating the wells.
To check whether the same amount of each protein was bound to the wells, the microtitration plates were coated with 2 μg per well protein. After washing, blocking, and washing essentially as described above, the proteins were resolubilized in 10% SDS. Protein estimation was performed according to Lowry et al18 using the protein standard from Sigma (bovine serum albumin to which 10% SDS was added).
To investigate the formation of oxidation-specific epitopes or the loss of native epitopes on apoB, an assay was performed as follows: 200 μL of oxidatively modified LDL samples (10 μg/mL) per well was incubated on microtitration plates in 10 mmol/L PBS, pH 7.4, plus 1 g/L EDTA at 4°C overnight. The amount of bound LDL was monitored using IgGs from anti-HNE-apoB, anti-MDA-apoB, and OB/04 as detection antibodies (1 μg per well) in 10 mmol/L PBS, pH 7.4, for 1 hour at room temperature. After washing the plates three times, Eu3+-labeled rabbit anti-IgG (11 Eu3+ mol/mol IgG) or Eu3+-labeled mouse anti-IgG (14 Eu3+ mol/mol IgG) was used as reporting antibody, with a concentration of 50 ng per well in 10 mmol/L PBS, pH 7.4, for 1 hour at 25°C. Anti-apoB (50 ng per well in 10 mmol/L PBS, pH 7.4) was applied as its labeled form (16 Eu3+ mol/mol IgG).
Solid-Phase Sandwich Radioassay
Polystyrene microtitration plates were coated with 200 μL per well of the oxidized 125I-LDL fractions (10 μg/mL) in 10 mmol/L PBS, pH 7.4, + 1 g/L EDTA. After washing six times, binding of oxidized 125I-LDL to the surface of the plates was measured on a microplate scintillation counter (Packard Instruments) adding 200 μL per well scintillation cocktail (Beckman Instruments). Binding studies of oxidized 125I-LDL on tissue proteins and poly-d-lysine were established similar to the solid-phase sandwich fluorescence assay described above.
Alterations of LDL, Eu3+-LDL, and 125I-LDL During Cu2+-Mediated Oxidation
LDL, Eu3+-LDL, and 125I-LDL were oxidized in the presence of 20 μmol/L CuSO4 at 37°C. The oxidation was terminated at different intervals between 1 and 24 hours by adding EDTA. The degree of modification of ox-LDL was estimated as its EM, relative to the nonmodified or nonlabeled native LDL. Comparing the EM of nonoxidized 125I- and Eu3+-labeled LDL with their unlabeled forms showed that the labeling itself caused modification of LDL. Estimation of the content of lipid hydroperoxides (data not shown) in 125I- and Eu3+-LDL underlined the assumption that LDL was slightly oxidized during the labeling procedure with iodine. While native and Eu3+-labeled LDL demonstrated almost an equal increase in the relative EM during the oxidation at certain time intervals, the oxidative modification of 125I-LDL occurred much faster.
To test whether oxidation of either Eu3+- or 125I-labeled LDL had any influence on counting fluorescence or radioactivity, the samples (1 mg/mL LDL protein) were applied on polystyrene microtitration plates. Oxidized fractions of Eu3+-LDL were diluted 1:10 000 with enhancement solution and the fluorescence was measured with the DELFIA research fluorometer. 125I-LDL oxidized fractions were diluted 1:50 with scintillation cocktail and the radioactivity was measured on a microplate scintillation counter. The fluorescence of Eu3+-LDL oxidized for 24 hours decreased to nearly 60% compared with the fluorescence of nonoxidized Eu3+-LDL, while the counts of radioactivity of 125I-LDL oxidized for 24 hours increased by about 32% (Table⇓).
To study any influence of oxidation on the affinity of 125I-ox-LDL to microtitration plates, 125I-LDL samples, oxidized for different time periods, were applied on polystyrene microtitration plates in 10 mmol/L PBS plus 1 g/100 mL EDTA, pH 7.4, for 16 hours at 4°C. After washing, radioactivity was measured. The oxidation of 125I-LDL led to a slight decrease in the affinity of ox-LDL to microtitration plates.
Recognition of Ox-LDL by Anti-HNE-LDL, Anti-MDA-LDL, Anti-Ox-LDL, and Anti-ApoB Using a Fluorescence Immunoassay
Polyclonal antibodies against HNE- and MDA-modified LDL and a mAb OB/04 against ox-LDL were generated and characterized.21 22 23 We studied the recognition of LDL oxidized for different periods of time up to 24 hours. Epitopes recognized by anti-HNE-LDL or anti-MDA-LDL were formed after 2 hours of oxidation, while OB/04 recognized epitopes only after 5 hours of oxidation. Furthermore, the recognition of ox-LDL by anti-apoB decreased dramatically within the first 3 hours of oxidation (Fig 1⇓).
Binding of Oxidized Eu3+-LDL and Oxidized 125I-LDL to Tissue Proteins and Poly-d-Lysine
Binding of oxidatively modified Eu3+-LDL and 125I-LDL to tissue proteins was studied on microtitration plates coated with type I, II, III, IV, and V collagen, fibronectin, laminin, or poly-d-lysine. To check whether the same amounts of each protein was bound to the wells, microtitration plates were coated with 2 μg per well of each protein and the proteins were resolubilized in 10% SDS. The recovery (given in micrograms per well of protein based on Lowry protein estimation) was as follows: type I collagen, 1.56; type II collagen, 1.56; type III collagen, 1.62; type IV collagen; 1.66; type V collagen, 1.66; laminin, 1.64; fibronectin, 1.60; and poly-d-lysine, 1.62. Thus, all proteins had bound to a similar and comparable extent to the wells (variation <5%).
Binding of the differently modified samples of ox-LDL was strongly dependent on the negative charge of the ligand as expressed in terms of the REM. While Eu3+-LDL, oxidized at different time intervals, showed similar strong binding to type I collagen and poly-d-lysine (Fig 2A⇓), oxidized 125I-LDL bound best by far to type I collagen and weaker but to a similar extent to type II collagen and poly-d-lysine (Fig 2B⇓).
Neither native Eu3+-LDL nor native 125I-LDL bound to laminin or fibronectin. Furthermore, their oxidatively modified forms bound much weaker to these two connective tissue proteins than to collagen I or poly-d-lysine. However, by increasing the concentration of laminin for coating the wells, the binding of ox-LDL could be enhanced (data not shown).
Fig 3⇓ shows the dependence of the concentration of highly oxidized Eu3+-LDL (REM, 3.5, 24 hours’ oxidation) on its binding to type I and type II collagen, poly-d-lysine, laminin, and fibronectin.
In another set of experiments, the binding of highly oxidized Eu3+-LDL (REM 3.4, 24 hours’ oxidation) to collagen was studied, comparing type I, II, III, IV, and V collagen. LDL was prepared from another donor and contained less Eu3+-labeling yield as the material used for the experiments shown in Figs 2B⇑ and 3⇑. The binding of ox-LDL followed the order: type I>type V, type III>type IV>type II collagen, as shown in Fig 4⇓.
Kinetics and Competition of Binding of Oxidized Eu3+-LDL to Type I Collagen
The binding of Eu3+-LDL oxidized to different degrees (EM, 2.2, 4 hours’ oxidation; EM 3.2, 6 hours’ oxidation; EM 3.7, 8 hours’ oxidation; EM 3.9, 24 hours’ oxidation) to type I collagen was measured after 15, 30, 60, 90, and 120 minutes of incubation (Fig 5⇓). Binding of native Eu3+-LDL (EM, 1.2) to type I collagen was not dependent on the incubation time, because the fluorescence activity counted after 15 minutes of incubation of native Eu3+-LDL showed a similar value to that after 210 minutes. Eu3+-LDL oxidized for 4 hours showed a strong increase in binding to type I collagen during 210 minutes, yet it did not reach saturation. However, with the other, more strongly oxidized forms of LDL, a maximal binding was reached after 90 minutes.
Additionally, competition experiments were performed using type I collagen (2 μg per well) for coating, 5 μg/mL oxidized Eu3+-LDL (REM 3.4, 24 hours’ oxidation), and 1, 5, 10, 50, 125, 500, 750, and 1000 μg/mL ox-LDL (REM 3.1, 24 hours’ oxidation) (200 μL per well). The incubation lasted 90 minutes. Under these conditions, a maximal competition of 63% of the binding of oxidized Eu3+-LDL to type I collagen was obtained at 750 to 1000 μg/mL ox-LDL. The fact that the binding of oxidized Eu3+-LDL could not be totally competed by an excess of unlabeled ox-LDL is probably due to the lower negative charge of the latter.
Competition of l-Arginine, l-Histidine, and l-Lysine With the Binding of Oxidized Eu3+-LDL to Tissue Proteins
It has been reported previously that histidine was able to compete with ox-LDL for binding to collagen gels.11 We used l-histidine as well as l-arginine and l-lysine in a competitive fluorescence assay as competitor for oxidized Eu3+-LDL binding to type I or type II collagen and laminin. l-Arginine was able to compete with oxidized Eu3+-LDL (REM 3.2, 24 hours’ oxidation) for binding to these three proteins. Binding to type I collagen was inhibited by more than 30% at 0.1 μg/mL, whereas 1 μg/mL and 100 μg/mL concentrations of the competitor were necessary to inhibit binding to type II collagen and laminin to a similar extent. l-Histidine was able to prevent binding of ox-LDL to laminin and type II collagen by about 65% and 30%, respectively. Binding of oxidized Eu3+-LDL to type I collagen was inhibited only weakly, and binding to fibronectin was not at all affected by l-histidine at the concentrations applied. l-Lysine was the strongest inhibitor. It prevented binding of oxidized Eu3+-LDL on type I collagen to almost 50% at only 0.01 μg/mL. From 10 to 100 μg/mL, l-lysine was necessary to obtain a similar competition on laminin. On type II collagen, the competition was weaker than on laminin. Similar results were obtained with other preparations of oxidized Eu3+-LDL or with biotinylated ox-LDL, when ox-LDL was modified to a comparable degree.
Influence of Monovalent and Divalent Cations and the Ionic Strength on Binding of Oxidized Eu3+-LDL to Type I Collagen
Since calcium and magnesium are essential components of the interstitial fluid and calcium apatite is assumed to be a component of advanced atherosclerotic lesions,24 we studied the influence of divalent cations on binding of Eu3+-LDL to type I collagen and also varied the concentration of the buffer. Microtitration plates were coated with type I collagen in 10 mmol/L PBS, pH 7.4, for 16 hours at 25°C. The wells were incubated with Eu3+-LDL (0.5 μg per well) in its native form and oxidized to different degrees (REM, 2.0, 4 hours’ oxidation; REM, 3.6, 24 hours’ oxidation) for 90 minutes at room temperature. The buffer (1 mmol/L PBS, pH 7.4) contained 143 mmol/L Na+, 4 mmol/L K+, 1.5 mmol/L Ca2+, and 0.5 mmol/L Mg2+. These concentrations are assumed to be equivalent to those in the interstitial fluid. Dilutions of the buffer were prepared at 1:5, 1:10, 1:50, and 1:100. In addition, 1 mmol/L PBS, pH 7.4, containing only 143 mmol/L Na+, and 4 mmol/L K+, represented the blank buffer. Binding of oxidized Eu3+-LDL was estimated by measuring the fluorescence described under “Methods.” While strongly oxidized Eu3+-LDL (24 hours’ oxidation) reached a maximum binding by increasing the concentration of Ca2+ and Mg2+ (at a buffer dilution of 1:50), moderately oxidized Eu3+-LDL (4 hours’ oxidation) reached its maximal binding at a buffer dilution of 1:10. A further increase of the divalent cations led to a decrease of the binding of both forms of oxidized Eu3+-LDL, whereas native Eu3+-LDL showed a weak but constant increase in its binding to type I collagen. In the nondiluted buffer, native Eu3+-LDL, and Eu3+-LDL oxidized for 4 or 24 hours showed essentially the same binding to type I collagen (Fig 6⇓.)
Competition of a Monoclonal Antibody to Ox-LDL With the Binding of Oxidized Eu3+-LDL to Tissue Proteins
To check for the role of epitopes newly formed by oxidative modification of LDL in binding to tissue proteins, an antibody to ox-LDL was used for competition. In so doing, the mAb OB/04 was subjected to limited proteolysis by pepsin. The F(ab′)2 fragments were used for a competition assay using 10 mmol/L PBS, pH 7.4. As shown in Fig 7A⇓, binding of highly oxidized Eu3+-LDL (REM, 4.1, 24 hours’ oxidation) to laminin and type II collagen was markedly inhibited by the F(ab′)2 of OB/04. Binding of oxidized Eu3+-LDL to type I collagen or poly-d-lysine was also affected, but to a weaker extent. However, increasing the concentration of the antibody fragment to 100 μg/mL led to a reduction of the binding of oxidized Eu3+-LDL to type I collagen by about 40%. In the case of fibronectin, no influence was obtained in the applied range of 0.01 to 10 μg/mL of the competitor in the buffer without divalent cations. Control experiments were performed using the F(ab′)2 of two mAbs from mice with different specificity. Neither anti-chromogranin nor anti-human IgG1 competed with the binding of oxidized Eu3+-LDL to type I or type II collagen, laminin, or poly-d-lysine.
When the experiments were repeated with type I collagen in a buffer containing calcium and magnesium (10 mmol/L PBS, pH 7.4, 143 mmol/L Na+; 4 mmol/L K+; 1.5 mmol/L Ca2+; and 0.5 mmol/L Mg2+) diluted 1:5, the F(ab′)2 fragment also hampered the binding of oxidized Eu3+-LDL by about 40% at the highest concentration of competitor (100 μg/mL). Moreover, in 10 mmol/L PBS, pH 7.4, containing only 143 mmol/L Na+ and 4 mmol/L K+, the competition of the F(ab′)2 fragment was retarded at lower concentrations of the fragment but also reached 40% at 100 μg/mL, as shown in Fig 7B⇑.
Previous diffusion studies indicated that either endothelial cell- or copper-oxidized LDL binds to a greater extent to type I collagen gel than native LDL, using a gel diffusion technique.9 10 11 In the present study we estimated the binding of Eu3+-LDL and 125I-LDL to microtitration plates coated with the tissue proteins (type I, II, III, IV, and V collagen, fibronectin, and laminin) and poly-d-lysine. This method allowed working with small LDL concentrations and short incubation times, keeping lipid peroxidation low, whereas in studies using diffusion techniques, the incubation time was 48 hours.9 11 Iodination of LDL obviously initiated lipid peroxidation, leading to a considerably higher content of lipid hydroperoxides and resulting in a higher negative charge of LDL protein. A higher negative charge was sometimes obtained with LDL labeled with the Eu3+ chelate. This reagent forms a stable covalent thiourea bond with free amino groups of the LDL protein and therefore causes an insignificantly higher EM, but without inducing lipid peroxidation in the lipid phase of LDL. The labeling yield of Eu3+-LDL was about 4 to 25 Eu3+/LDL-protein, giving high sensitivity with low background and low blocking of basic amino acid residues, such as lysines. While the same progression of Cu2+ induced lipid peroxidation of LDL and Eu3+-LDL was monitored, 125I-LDL oxidized at a faster rate.
Jimi et al10 showed an increased binding of 125I-ox-LDL versus native 125I-LDL to collagen gels using mixtures of type I collagen with itself or other collagen types. The binding activities for both LDLs followed the order: I+I, I+III>I+V>I+IV collagen. In contrast, we used each type of collagen in its pure form and found only a very weak binding of native LDL in the absence of divalent cations. This binding did not increase with prolongation of the incubation of native LDL. The favorite binding of our ox-LDL to type I, V, and III collagen is, however, comparable with the results reported previously.10 It should be pointed out that type I and V collagen tended to increase with the progression of atherosclerosis, whereas the proportion of type III collagen to total collagen fell slightly with advancing atherosclerosis.1
Different binding affinities of ox-LDL to the proteins were obtained when ox-LDL, revealing a gradual increase in the oxidative modification, was offered. Maximal increase of binding of ox-LDL to type I and type II collagen or poly-d-lysine was obtained when the oxidative modification of LDL corresponded to REM values from 1.8 to 3.4. This is in good agreement with a recent study in which the diffusion in collagen gels of ox-LDL with different REM values was estimated.11
Negatively charged macromolecules such as ox-LDL may bind to regions on collagen carrying a positive charge. A recent study25 showed a variation in the content of arginine between type I and III collagen compared with type IV collagen, with a loss of nearly 30%. This might explain why we obtained a much weaker binding of ox-LDL by type IV collagen than by type I and III collagen. Comparing the amino acid composition of type I collagen with type II collagen of human origin, Miller25 found a loss of lysine by 42% but an increase of 50% hydroxy-lysine residues in the latter. We found a weaker binding of ox-LDL to collagen type II than to type I collagen, being reduced by 50% to 60%. Furthermore, ox-LDL showed nearly the same binding to poly-d-lysine as to type I collagen. It should be pointed out that regions rich in lysines are assumed to play a potential role in the specific binding of negatively charged oxidatively modified LDL, not only to the extracellular matrix but also to the collagenous domain of the scavenger receptors.12 13 This coiled-coil domain contains 24 uninterrupted Gly-X-Y triplets in which 14 of the Y residues are lysines or prolines. At physiological pH, all of the Gly-X-Y triplets in the bovine scavenger receptor are expected to form a classic righthanded collagenous triple helix, positively charged.26 This collagenous domain is assumed to be responsible for the specific binding of complex polyanionic ligands as a kind of molecular flypaper.27
In a recent study, histidine was shown to prevent binding of LDL to collagen.11 On the one hand, this was due to an inhibition of LDL oxidation by complexing divalent cations such as Cu2+ by histidine. On the other hand, a direct interference of histidine with the binding of ox-LDL to collagen was reported. In our study, l-histidine competed only weakly with the binding of ox-LDL to type I collagen, but it considerably reduced the interaction of ox-LDL with laminin and type II collagen.
In the present study we used for the first time an antibody to ox-LDL as competitor for the binding of ox-LDL to the proteins investigated. The F(ab′)2 competed best with binding of ox-LDL to type II collagen, whereas there was no influence using fibronectin. This was probably due to the weak affinity of ox-LDL for fibronectin. Interestingly, the interactions between ox-LDL and laminin were weakened by almost 40%. This effect was specific for OB/04, the mAb against ox-LDL, since the fragments of anti-chromogranin and anti-IgG1, two other antibodies from mice used as controls, did not compete. Although we do not know the nature or structure of the epitope recognized by the mAb OB/04 in detail, these data indicate that interactions of ox-LDL with the proteins selected for this investigation are not dependent solely on the negative net charge of the oxidized lipoprotein but also on certain epitopes created during oxidative modification. At present, the binding mechanisms of ox-LDL to proteins studied are not elucidated sufficiently to allow a better interpretation of these results.
In the response-to-retention hypothesis of early atherogenesis, Williams and Tabas28 consider LDL as an absolute requirement for lesion development, provoking normal cellular and matrix elements to participate in a cascade leading to atherosclerosis. Ox-LDL was shown to stimulate the collagen production in the arterial smooth muscle cells by growth factors such as platelet-derived growth factor and transforming growth factor β1.29 Furthermore, in a recent immunohistochemical study on human atherosclerotic plaques, epitopes derived from lipid peroxidation and epitopes specific for apoB were shown to colocalize predominantly extracellularly along laminar and fibrous structures.30 Thus, subendothelial oxidation of LDL might not only be important in creating epitopes recognized by the scavenger receptors of macrophages but, as our results suggest, also be involved in a process of retention of this lipoprotein by matrix proteins, predominantly various types of collagen and to a lesser extent laminin.
Selected Abbreviations and Acronyms
|SDS-PAGE||=||SDS-polyacrylamide gel electrophoresis|
This article is being published in remembrance of Prof Dr Esterbauer. This work was supported by the Austrian Research Council, joint project program “Biomembranes and Atherosclerosis,” project F00710 (to Dr Greilberger and Dr Jürgens). We thank Gerhard Ledinski and Alexandra Tieber for their skillful technical assistance.
Bihari-Varga M, Gruber E, Rotheneder M, Zechner R, and Kostner GM. Interaction of lipoprotein Lp(a) and low-density lipoprotein with glycosaminoglycans from human aorta. Arteriosclerosis.. 1988;8:851-857.
Quinn M, Parthasarathy S, Fong LG, Steinberg D. Oxidatively modified low-density lipoproteins: a potential role in recruitment and retention of monocyte/macrophages during atherogenesis. Proc Natl Acad Sci U S A.. 1987;84:2995-2998.
Parthasarathy S, Printz DJ, Boyd D, Joy L, Steinberg D. Macrophage oxidation of low-density lipoprotein generates a modified form recognized by the scavenger receptor. Arteriosclerosis.. 1986;6:505-510.
Sparrow CP, Parthasarathy S, Steinberg D. A macrophage receptor that recognizes oxidized low density lipoprotein but not acetylated low density lipoprotein. J Biol Chem.. 1989;264:2599-2604.
Hoover GA, McCormick S, Kalant N. Interaction of native and cell-modified low density lipoprotein with collagen gel. Arteriosclerosis.. 1988;8:525-534.
Kalant N, McCormick S, Parniak MA. Effects of copper and histidine on oxidative modification of low density lipoprotein and its subsequent binding to collagen. Arterioscler Thromb.. 1991;5:1322-1329.
Krieger M, Acton S, Ashkenas J, Pearson A, Penman M, Resnik D. Molecular flypaper, host defense, and atherosclerosis. J Biol Chem.. 1993;268:4569-4572.
Schmut O. The organization of tissue of the eye by different collagen type. Arch Klin Exp Ophthalmol.. 1978;207:189-199.
Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with the Folin phenol reagent. J Biol Chem.. 1951;193:265-275.
El-Saadani M, Esterbauer H, El-Sayer M, Goher M, Nassar AY, Jürgens G. A spectrophotometric assay for lipid peroxides in serum lipoproteins using a commercially available reagent. J Lipid Res.. 1989;30:627-629.
Hammer A, Kager G, Dohr G, Rabl H, Ghassempur I, Jürgens G. Generation, characterization, and histochemical application of monoclonal antibodies selectively recognizing oxidatively modified apoB-containing serum lipoproteins. Arterioscler Thromb Vasc Biol.. 1995;15:704-713.
Chen Q, Esterbauer H, Jürgens G. Studies on epitopes on low density lipoprotein modified by 4-hydroxy-nonenal: biochemical characterisation and determination. Biochem J.. 1992;288:249-254.
Stary HC, Chandler AB, Dinsmore RE, Fuster V, Glagoy S, Insull W, Rosenfeld ME, Schwartz CJ, Wagner WD, Wissler RW. A definition of advanced types of atherosclerotic lesions and a histological classification of atherosclerosis. Arterioscler Thromb Vasc Biol.. 1995;15:1512-1531.
Miller EJ. Chemistry of the collagens and their distribution. In: Piez KA, Reddi AH, eds. Extracellular Matrix Biochemistry. New York: NY: Elsevier Science Publishing Co; 1984:41-78.
Williams KJ, Tabas I. The response-to-retention hypothesis of early atherogenesis. Arterioscler Thromb Vasc Biol.. 1995;15:551-561.
Liu K, Massaeli H, Pierce GN. The action of oxidized low density lipoprotein on calcium transients in isolated rabbit cardiomyocytes. J Biol Chem.. 1993;6:4145-4151.
Jürgens G, Chen Q, Esterbauer H, Mair S, Ledinski G, Dinges HP. Immunostaining of human autopsy aortas with antibodies to modified apolipoprotein B and apoprotein (a) Arterioscler Thromb.. 1993;13:1689-1699.