Enzymatic Modification of Low-Density Lipoprotein in the Arterial Wall

From the Institutes of Clinical Chemistry and Laboratory Medicine (M.T., V.O., A.S., K.J.L.) and Medical Microbiology and Hygiene (P.S., K.P., L.S., S.-R.H., M.H., S.B.), and the Department of Ophthalmology (V.B.G.), University of Mainz, Germany.

ABSTRACT

Objective— Functionally interactive proteases of the plasminogen/plasmin and the matrix metalloproteinase (MMP) system degrade and reorganize the extracellular matrix of the vessel wall in atherosclerosis. Here we investigated whether such proteases are able to confer atherogenic properties onto low density lipoprotein by nonoxidative modification.

Methods and Results— Similar to the recently described enzymatically-modified low-density lipoprotein (E-LDL), native LDL exposed to plasmin or matrix MMP-2 or MMP-9 and cholesterylester-hydrolase (CEH) showed extensive deesterification, with ratios of free cholesterol to total cholesterol rising to 0.8 compared with 0.2 in native LDL. When the ratio exceeded 0.6, both plasmin/CEH-LDL and MMP/CEH-LDL fused into larger particles. In parallel, they gained C-reactive protein–dependent complement-activating capacity. E-LDL produced with any protease/CEH combination was efficiently taken up by human macrophages, whereby marked induction of MMP-2 expression by E-LDL was observed. These in vitro findings had their in vivo correlates: urokinase-type plasminogen activator, MMP-2, and MMP-9 were detectable in both early and advanced human atherosclerotic lesions in colocalization with E-LDL.

Conclusions— Plasmin and MMP-2/MMP-9 may not only be involved in remodeling of the extracellular matrix in progressing plaques, but they may also be involved in lipoprotein modification during genesis and progression of atherosclerotic lesions.

Plasmin and matrix metalloproteinases are present in early human atherosclerotic lesions. Both can synergize with cholesterylesterase to transform low-density lipoprotein into a molecule that binds C-reactive protein, activates complement, and induces macrophage foam cell formation. These processes may serve to remove stranded low-density lipoprotein from tissues, but excessive immune activation could trigger atherogenesis.

Key Words: atherosclerosis ? lipoproteins ? macrophages ? metalloproteinases ? plasminogen activators

Introduction

It is widely held that atherogenesis is triggered by enhanced entrapment of low-density lipoprotein (LDL) in the intima, which is followed by its uptake by macrophages. Both oxidative and nonoxidative processes can generate potentially atherogenic LDL derivatives.1 We are pursuing the concept that enzymatic remodeling of the lipoprotein is a key modification, because proteolytic cleavage of apolipoprotein B (apoB) in conjunction with hydrolysis of cholesteryl esters generates lipoprotein particles that are similar to lesion-derived LDL in structure, biological properties, and composition.2,3 Enzymatically-remodeled LDL (E-LDL) binds C-reactive protein (CRP) and activates complement.4 E-LDL induces foam cell formation in monocytes,5 macrophages,3 and smooth muscle cells,6 stimulates MCP-1 production,7 and directly promotes adhesion and transmigration of monocytes through endothelial cell monolayers.8 These in vitro findings have their in vivo correlates: immunohistological analyses with specific monoclonal antibodies (mAbs) have revealed extensive extracellular deposits of E-LDL at the early stages of atherosclerotic lesion formation.9 CRP and activated complement components are also present in colocalization with E-LDL.4,9 Like in vitro–generated E-LDL, lesioned LDL has a high content of free cholesterol;10–12 therefore, it is apparent that extensive deesterification of cholesteryl esters must indeed occur in the lesions.

We used trypsin in combination with cholesterylester-hydrolase to produce E-LDL in vitro.3 Although both neutral and acid cholesterylester-hydrolase have been demonstrated to be present in the arterial vessel wall of humans and rabbits13,14 and the concentration of the hydrolase is so high in atherosclerotic lesions that it can be directly detected by immunohistochemistry,15 it remains unclear which protease might be acting in synergy with this enzyme to generate E-LDL in vivo. Most recently, we found that macrophages are stimulated by E-LDL to produce cathepsin H, that cathepsin H can replace trypsin as the modifying protease, and that cathepsin H colocalizes with E-LDL in atherosclerotic lesions.16 Thus, the idea is emerging that the specificity of a modifying protease bears little relevance for the creation of atherogenic E-LDL. Subsequent generation of free cholesterol and free fatty acids leads to fusion of E-LDL, which is endowed with the characteristic atherogenic properties. Should this hypothesis be correct, several proteases might be involved in generating E-LDL in situ. Establishing the redundancy of a basic process in atherogenesis would obviously be of interest.

Here, we report the analysis of 3 enzymes known to be present in atherosclerotic lesions, namely plasmin, metalloproteinase-2 (MMP-2), and MMP-9, for their potential to modify LDL nonoxidatively and confer potential proatherogenic properties to the lipoprotein particle.

Methods

The Methods section can be found in an online supplement available at http://atvb.ahajournals.org.

Isolation and Modification of LDL

LDL from healthy subjects, aged 18 to 65 years, was isolated by preparative ultracentrifugation (d=1.020 to 1.062 g/mL). Concentrations refer to the total cholesterol concentration in the lipoprotein samples.

For enzymatic modification, LDL (5 to 7 mg/mL) was incubated in veronal buffered saline (VBS) with a protease at 37°C for 18 hours, followed by further incubation with cholesterylester-hydrolase (CEH; Sigma). Reactions were monitored by using the ratio of free to total cholesterol17 and measurements of turbidity at 560 nm. The following proteases were used: trypsin (10 μg/mL; Sigma), plasmin (0 to 20 U/mL; Calbiochem), MMP-2 (1 μg/mL; Sigma), and MMP-9 (1 μg/mL; Sigma).

Complement Consumption Assay

A human serum pool from 10 healthy donors containing <1 μg/mL CRP was diluted 10-fold with VBS. Lipoproteins and CRP were added to the given final concentrations, and samples were incubated for 60 minutes at 37°C. A 5% suspension of antibody-coated sheep erythrocytes was added to each tube, and hemolysis was read after 60 minutes at 37°C by measuring the absorbance of supernatants at 412 nm. Curves obtained from the hemolysis readings were inverted to depict complement consumption.

C3-Conversion

C3-conversion in human serum was assessed by 2D quantitative immunoelectrophoresis as described.18 Briefly, lipoproteins were added to human serum (final concentration: 400 μg/mL cholesterol±10 μg/mL recombinant CRP). After 4 hours at 37°C, samples were diluted with 300 μL saline/10 mmol/L EDTA (pH 7.5), and 10 μL samples were subjected to 2D immunoelectrophoresis. C3-turnover was assessed by planimetry of the areas delimited by the C3 and C3b/C3c arcs.18

Monocyte Isolation and Cell Culture

Monocytes were isolated from buffy coats of healthy donors as described7 and cultured for 7 days in medium supplemented with 10% AB serum. Cells were then incubated for 16 hours in the presence of lipoproteins at a concentration of 100 μg cholesterol/mL, washed twice, and subsequently fixed in 10% formaldehyde. Intracellular lipid-droplets were stained with oil red O (Sigma). Total cholesterol was determined by using an enzymatic assay (Roche).17 The significance of differences was determined by the Student t test for paired samples.

Immunohistochemistry

Specimens of early (initial lesions and fatty streaks) and advanced human coronary atherosclerotic lesions19 were used for immunohistochemical staining as described.9 Primary antibodies were used for E-LDL,9 urokinase-type plasminogen activator (uPA), MMP-2, and MMP-9. Negative controls included replacement of the primary antibody by irrelevant isotype-matched monoclonal mouse antibodies.

Chemiluminescence Immunoassay for Antibody Binding

Reactivity of native and modified LDL with the mAbAIL-3 raised against E-LDL was assessed in a microtitre plate chemiluminescent assay. AIL-3 was used diluted 1:5000, and binding was detected by chemiluminescence with a goat anti-mouse IgG conjugate (Sigma).

Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis and Western Blot Analyses

Monocytes from 4 healthy donors were isolated and incubated with native or modified LDL for 24 hours. Thereafter, cells were harvested and lysed in sodium dodecyl sulfate (SDS)-loading buffer. Proteins (from 106 cells; Figure 3B) or lipoprotein samples, respectively, were separated in 10% SDS polyacrylamide gels, transferred to nitrocellulose membranes, and probed with primary antibodies for MMP-2, MMP-9, or E-LDL. Blots were developed by enhanced chemiluminescence.

Figure 3. A, Combined treatment with plasmin and CEH, MMP-2 and CEH, or MMP-9 and CEH transforms LDL into a foam cell inducer. Monocyte-derived macrophages were incubated for 16 hours in the absence or presence of 100 μg/mL native LDL, with LDL modified by combined treatment with either plasmin plus CEH (plasmin/CEH-LDL, left), MMP-2 plus CEH (MMP-2/CEH-LDL, middle), or MMP-9 plus CEH (MMP-9/CEH-LDL, right). Total cholesterol was determined by cholesterol measurement assays from Roche and referred to total cellular protein. Means and error bars (SD) are calculated from 3 different experiments. n.s. indicates not significant; *P<0.05. B, MMP-2 protein expression in monocytes is increased by MMP-2/CEH-LDL. Monocytes were incubated with LDL and MMP-2/CEH-LDL or MMP-9/CEH-LDL (25 and 50 μg/mL) for 24 hours and analyzed by Western blotting. The mature 72-kDa and 95-kDa bands of MMP-2 (top) and MMP-9 (bottom), respectively, are shown. The 100-kDa band of the molecular weight marker is indicated by arrows.

Results

LDL Modified With Plasmin, MMP-2 or MMP-9, and CEH is Similar to LDL Modified with Trypsin

Enzymatic transformation of LDL to potentially atherogenic E-LDL has previously been achieved by combined treatment with trypsin and CEH. It was then of interest to discern whether plasmin or MMP could replace trypsin. Both plasmin/CEH-LDL and MMP/CEH-LDL displayed marked electronegativity in agarose gel electrophoresis, similar to trypsin-generated E-LDL (not shown). E-LDL preparations were analyzed for their content of free and esterified cholesterol. In all cases, the molar ratios of free to total cholesterol rose from 0.2 to reach 0.8, as previously found for both trypsin- and cathepsin H–modified LDL16,17 and for lesion-derived LDL.2 Deesterification was not noted when 1 of the 2 enzymes was omitted. Furthermore, we performed 1 experiment with tissue inhibitor of metalloproteinase-1 (TIMP-1) and found that TIMP-1 inhibited MMP-9–mediated enzymatic modification of LDL (data not shown). Taken together, these results confirmed the report that proteolytic nicking of apoB renders cholesterylesters accessible to the action of CEH.20

LDL Modified With Plasmin, MMP-2 or MMP-9, and CEH has CRP-Dependent Complement-Activating Capacity

The ability of different LDL-preparations to activate complement was analyzed by 2D immunoelectrophoresis of C3. Figure 1 depicts results obtained with 50 μg/mL native LDL, plasmin/CEH-LDL, MMP-2/CEH-LDL, or MMP-9/CEH-LDL. Low basal C3 turnover of 10% to 20% occurred in samples spiked with native LDL in the absence or presence of CRP. Plasmin/CEH-LDL, MMP-2/CEH-LDL, or MMP-9/CEH-LDL alone slightly augmented C3 cleavage. However, when CRP was added to the modified LDL preparations, C3 consumption was markedly enhanced to 50% to 70%. No enhancement of complement activation was observed with control serum samples incubated in buffer alone, in serum spiked with the enzyme mix, in CRP without lipoproteins, and when either protease or CEH was omitted.

Figure 1. 2D immunoelectrophoresis of C3 in human serum. Serum samples were incubated with 50 μg/mL native LDL, plasmin/CEH-LDL, MMP-2/CEH-LDL, and MMP-9/CEH-LDL in the absence (left) or presence (right) of CRP. Note substantial C3 conversion provoked by plasmin/CEH-LDL, MMP-2/CEH-LDL, and MMP-9/CEH-LDL, but not native LDL in the presence of CRP. Results are representative of 3 separate experiments.

Dose-response and kinetic experiments were undertaken using the combination of plasmin and CEH or plasmin alone. First, the protease concentration was varied from 0 to 2 U/mL. After overnight incubation, CEH was applied at 20 μg/mL for another 10 to 48 hours. Formation of E-LDL was monitored by assessment of cholesterol deesterification, turbidity measurements, and complement consumption tests. It was found that as little as 0.0008 U/mL plasmin sufficed to promote formation of functional E-LDL. At this lowest concentration, SDS-PAGE revealed that very limited nicking of apoB had occurred, leading to appearance of several high molecular weight products (>200 kDa) and one major product of lower molecular weight (75 kDa). Plasmin alone was used at 0.2 U/mL. The respective SDS-PAGE did not differ from the one using plasmin/CEH-LDL at 0.2 U/mL. (Figure I, available online at http://atvb.ahajournals.org).

Kinetic experiments were then performed using 0.02 U/mL plasmin. In these experiments, LDL preparations were reequilibrated in VBS before enzyme treatment so that complement activation tests could be performed on aliquots taken at different times. Deesterification was followed after addition of CEH, and it was found that the ratio of free to total cholesterol rapidly rose from 0.2 to reach 0.5 to 0.6 within 15 minutes. Thereafter, deesterification proceeded slowly and the lipoprotein solutions remained clear until the ratio of free to total cholesterol exceeded 0.6. At varying times thereafter, solutions became turbid because of LDL fusion. Once initiated, this process proceeded rapidly, and maximal turbidity was reached within 2 hours (Figure 2A). The total time required for LDL solutions to turn turbid ranged from 6 to 55 hours with 6 different LDL preparations.

Figure 2. A, Plasmin-treated LDL was incubated with CEH; the content of free cholesterol and turbidity of the solution was followed over time. B, No complement activation was observed either in the presence () or absence () of CRP as long as the lipoprotein solutions remained clear, despite the fact that cleavage of >60% of the cholesteryl esters had occurred. C, Complement activation was noted when solutions turned turbid because of lipoprotein fusion. In the absence of CRP (), complement activation occurred at high E-LDL concentrations. In the presence of CRP (), complement consumption was noted already at low E-LDL concentrations. In A through C, results are representative of 3 separate experiments.

Functional tests revealed that complement-activating function correlated with the fusion step. Thus, when a sample from the foregoing experiment was tested at 1 hour, during which extensive deesterification had already occurred, no complement consumption was discerned with or without CRP (Figure 2B). In contrast, once an LDL preparation had become turbid, spontaneous complement consumption (in the absence of added CRP) was observed at high E-LDL concentrations of >100 μg/mL. However, complement consumption could be triggered at lower E-LDL concentrations by addition of 5 μg/mL CRP (Figure 2C). This activation pattern was found for all E-LDL preparations (plasmin/CEH-LDL, MMP-2/CEH-LDL, and MMP-9/CEH-LDL) without exception.

Enhanced Macrophage Uptake of LDL Modified With Plasmin, MMP-2 or MMP-9, and CEH

Plasmin/CEH-LDL, MMP-2/CEH-LDL, and MMP-9/CEH-LDL induced foam cell formation as revealed by oil red staining (not shown). Results of quantification of cellular cholesterol are shown in Figure 3A. Incubation of cells with 100 μg/mL (cholesterol) native LDL for 16 hours caused a slight but statistically not significant increase of total cellular cholesterol (10%). In contrast, incubation with either plasmin/CEH-LDL, MMP-2/CEH-LDL, or MMP-9/CEH-LDL led to massive cholesterol accumulation. Total cellular cholesterol content increased 2- to 4-fold, corresponding to a >10-fold higher cholesterol uptake compared with that found with native LDL. Foam cell formation was not noted if 1 of the enzymes was omitted.

MMP-2 Protein Expression in Monocytes Is Increased by E-LDL

To investigate whether MMP-2/CEH-LDL or MMP-9/CEH-LDL induce MMP-2 or MMP-9, monocytes were treated with lipoprotein preparations and analyzed by Western blot using specific antibodies directed against MMP-2 and MMP-9. Expression of MMP-2 was enhanced in MMP-2/CEH-LDL–treated cells compared with LDL-treated cells, a result that was confirmed in 4 donors. Figure 3B (top) depicts a representative blot; control incubations without lipoprotein were additionally used for comparison. In contrast, expression of MMP-9 (bottom) was not different in MMP-9/CEH-LDL–treated cells compared with LDL-treated cells.

Plasmin, MMP-2, and MMP-9 Colocalize With E-LDL in Human Atherosclerotic Lesions

We then sought evidence for the presence of MMP-2 and MMP-9 using a specific mouse mAb against these enzymes and the plasmin cascade using a specific mouse mAb against the uPA in human coronary atherosclerotic lesions. Tissue sections were stained for uPA, MMP-2, and MMP-9 and compared with E-LDL staining. The predominant manifestation of E-LDL deposits in an early lesion (Figure 4D) was a diffuse deposition in the deep fibroelastic and fibromuscular layers of the intima adjacent to the media as described previously.4,9 There was a close association and overlapping of E-LDL and uPA (Figure 4A) with MMP-2 (Figure 4B) and MMP-9 (Figure 4C) epitopes within the deeper portion of the intima. In a more advanced atherosclerotic lesion (Figure 4F through 4J), uPA (Figure 4F), MMP-2 (Figure 4G), and MMP-9 (Figure 4H) also colocalized with the E-LDL epitopes (Figure 4I). The proteases and E-LDL were diffusely scattered among the extracellular components of the lipid cores of lesions, predominantly in the central part. In addition, intracellular staining for uPA, MMP-2, and MMP-9 of foam cells was evident. As reported previously, these cells did not stain positively for E-LDL, which was likely because of destruction of the neoepitope in the course of extensive proteolysis, as has been demonstrated in vitro.9 Control staining performed with the irrelevant isotype-matched antibodies yielded negative results with all tissue specimens (Figure 4E and 4J).

Figure 4. Representative example of the immunohistochemical analysis of uPA, MMP-2, MMP-9, and E-LDL in different stages of human coronary atherosclerotic lesion development. Lumen is to the upper left-hand corner. Demarcation between intima and media is indicated by an arrowhead. Magnification x64 (A through E) and x32 (F through J), respectively. A through E, Sequential sections of an early lesion (fatty streak). F through J, Sequential sections of an advanced lesion. A and F, uPA stain. B and G, MMP-2 stain. C and H, MMP-9 stain. D and I, E-LDL stain. E and J, Control stain.

MAb AIL-3 Recognizes LDL After Modification With Plasmin, MMP-2 or MMP-9, and CEH

The anti–E-LDL antibody AIL-3, which was derived against LDL modified by trypsin and CEH,9 also reacted with E-LDL generated by incubation with plasmin, MMP-2 or MMP-9, and CEH as demonstrated by chemiluminescence immunoassay and Western blotting. Figure 5 shows a representative AIL-3 chemiluminescent binding assay to native LDL, trypsin/CEH-LDL, plasmin/CEH-LDL, MMP-2/CEH-LDL, and MMP-9/CEH-LDL. Specific antibody binding was observed to all LDL preparations treated with a protease and CEH. In contrast, native LDL preparations showed no significant antibody binding. Intact apoB of native LDL was not recognized by mAb AIL-3 in Western blots (Figure II, available online at http://atvb.ahajournals.org). After enzyme treatment with either trypsin or plasmin, mAb AIL-3 reacted mainly with 3 fragments of MW 68, 48, and 30 kDa. Unlike the chemiluminescence immunoassay, a band was not seen with MMP-2/CEH-LDL and MMP-9/CEH-LDL. This probably reflects a loss of the epitope after SDS-polyacrylamide gel electrophoresis.

Figure 5. Immunoassay showing binding of anti–E-LDL antibody AIL-3 to LDL and E-LDL treated with trypsin, plasmin, MMP-2, or MMP-9, respectively. Equal amounts of primary antibody (1:5000 dilution) were added per well. The amount of antibody bound was then measured with alkaline phosphatase-labeled goat anti-mouse IgG using chemiluminescent technique. Luminescence was measured as number of flashes of light (RLU indicates relative light units). Means and error bars (SD) are calculated from 3 different experiments.

Discussion

Ample evidence indicates MMPs can weaken the fibrous cap of arterial plaques and thereby make them prone to rupture.21,22 The plasminogen cascade represents another proteolytic system involved in the genesis and progression of atherosclerotic lesions. Its contribution to the development of experimental neointimal lesions after injury and to aortic medial destruction was demonstrated in uPA and plasminogen activator inhibitor-1–null mice23 and apoE–null mice,24 respectively. Despite this detrimental role, the overall effect of MMPs and the plasminogen cascade in the pathogenesis of atherosclerosis and its sequelae is not entirely clear. New insights obtained from recent studies with MMP inhibitors and genetic manipulation also point to beneficial effects of MMPs in the process.25

Here we report another role for MMPs and plasmin in the arterial vessel wall. Both proteases can degrade LDL and trigger its conversion to a CRP-binding, complement-activating particle. E-LDL with atherogenic properties was originally produced in vitro by the consecutive action of trypsin, CEH, and neuraminidase.3 The latter enzyme was needed to confer maximal complement activating properties to E-LDL. With the discovery that LDL modified by trypsin and CEH could bind CRP, converting it to a very potent complement activator,4 incubation with neuraminidase was omitted. Trypsin-generated E-LDL is similar to lesion derived lipoprotein remnants in having a high content of unesterified cholesterol and a tendency to fuse and form heterogeneously-sized lipidic droplets. Furthermore, in vitro–generated E-LDL has neoepitopes that are detectable in lesion areas rich in extracellular lipids. Like E-LDL, lesion-derived material also activates complement, concurring with the fact that complement activation is one of the first steps in lesion initiation, even preceding monocyte infiltration.26

The now 8-year-old concept that we have been pursuing proposes extracellular enzymatic degradation of LDL to represent a key step leading to generation of an atherogenic lipoprotein.3,27,28 One of the most prevalent criticisms of this concept has been the fact that trypsin is an intestinal enzyme with no connection to atherosclerotic lesion formation. After the demonstration that cathepsin H can substitute for trypsin in generating E-LDL, the present observation that MMPs and plasmin are similarly effective may resolve an issue of basic importance. LDL treated with any protease followed by incubation with CEH shows the same properties as trypsin/CEH-treated LDL. E-LDL is highly enriched in unesterified cholesterol, binds CRP to activate complement, and efficiently drives foam cell formation. Finally, as demonstrated by chemiluminescence immunoassay or Western blotting, it shares the epitope for the E-LDL specific mAb AIL-3, which has previously been used to stain nonoxidized modified LDL in early human atherosclerotic lesions.9 Thus, E-LDL generated by all tested proteases are endowed with the same atherogenic properties as the originally described trypsin-generated E-LDL.

The present study provides new information on the in vitro formation of functional E-LDL, and a number of points merit comment. It has become clear that very limited nicking of the apoB "cage," unrelated to any specific apoB breakdown patterns, suffices to render the underlying cholesterylesters accessible to CEH. This accords with the finding that any one of the tested proteases can assume the modifying function; that is, that specificity plays no recognizable role. Second, the kinetic experiments reveal that CRP binding and complement activating function correlate not with the onset of deesterification but with the fusion step that occurs after extensive scission of the cholesterylesters has occurred. This finding is in line with the fact that neither native nor oxidized LDL possess the complement-activating properties. We have proposed that complement activators are present but masked in native LDL molecules, and that they become exposed when the lipoprotein is entrapped in tissues.28,29 This is thought to primarily subserve a physiological function, because innate immune mechanisms can thus be recruited to remove the lipoprotein with its insoluble cargo. Exposure of the activators requires fusion of E-LDL particles, which occurs only by the combined action of proteases (like plasmin, MMP-2, or MMP-9) and CEH. Thus, the action of plasmin and MMPs is not thought to be primarily atherogenic. Fusion of E-LDL particles is not equivalent to lipoprotein aggregation, which can occur after massive oxidation. The time required for fusion to occur was found to vary considerably among the LDL preparations. The reasons for this are presently unknown, but one possibility is that differences in LDL subfractions play a role.30,31

Evidence for the presence of active MMPs and plasmin in atherosclerotic lesions is abundant,32–34 and MMPs and plasmin interact functionally and cooperate in extracellular matrix degradation.23,24,35 Plasmin plays a role in media destruction and aneurysm formation through activation of pro–MMP-3 (stromelysin-1), pro–MMP-9, pro–MMP-12 (metalloelastase), and pro–MMP-13 (collagenase-3) in atherosclerotic lesions.24 However, none of the data have hitherto pointed to the possibility that MMPs and plasmin might also have a role in early atherogenesis by their remodelling action on tissue-stranded LDL. The present demonstration of colocalization between the proteases and E-LDL within early atherosclerotic lesions is of distinct interest. It is conceivable that MMPs and plasmin serve a dual role in the early lesion: (1) they provide access for macrophages to the region, and (2) they convert the trapped lipoprotein into a molecule that can be efficiently removed.

What are the origins of MMPs and plasmin in the atherosclerotic lesions? As shown in this article for MMP-2 and previously for cathepsin H,16 at least one possibility is the induction of such proteases by E-LDL. Of course, we cannot speculate which of the 4 detected enzymes (cathepsin H, MMP-2, MMP-9, or plasmin) assumes the decisive role of LDL modification in vivo. The MMP system itself is highly redundant, and proteolytic activity on a given substrate is not dependent on the presence of a single enzyme.36 However, given the probable redundancy of the proteolytic modification step, it appears unlikely that specific inhibitors of any protease will alone be able to efficiently halt the formation of E-LDL. Thus, reducing the E-LDL tissue-load through diminishing LDL insudation into the vessel wall will remain the mainstay of therapeutic strategies in atherosclerosis.

Acknowledgments

This work received partial support from the Deutsche Forschungsgemeinschaft (Bh2/3-1) and the Boehringer Ingelheim Foundation (to V.B.G.). We thank Rosemarie Schweigert, Antje Canisius, and Klaus Adler for expert technical assistance.

References

Williams KJ, Tabas I. The response-to-retention hypothesis of atherogenesis reinforced. Curr Opin Lipidol. 1998; 9: 471–474.

Seifert PS, Hugo F, Tranum-Jensen J, Z?hringer U, Muhly M, Bhakdi S. Isolation and characterization of a complement-activating lipid extracted from human atherosclerotic lesions. J Exp Med. 1990; 172: 547–557.

Bhakdi S, Dorweiler B, Kirchmann R, Torzewski J, Weise E, Tranum-Jensen J, Walev I, Wieland E. On the pathogenesis of atherosclerosis: enzymatic transformation of human low density lipoprotein to an atherogenic moiety. J Exp Med. 1995; 182: 1959–1971.

Bhakdi S, Torzewski M, Klouche M, Hemmes M. Complement and atherogenesis: binding of CRP to degraded, nonoxidized LDL enhances complement activation. Arterioscler Thromb Vasc Biol. 1999; 19: 2348–2354.

Kapinsky M, Torzewski M, Büchler C, Duong CQ, Rothe G, Schmitz G. Enzymatically degraded LDL preferentially binds to CD14(high) CD16(+) monocytes and induces foam cell formation mediated only in part by the class B scavenger-receptor CD36. Arterioscler Thromb Vasc Biol. 2001; 21: 1004–1010.

Klouche M, Rose-John S, Schmiedt W, Bhakdi S. Enzymatically degraded, nonoxidized LDL induces human vascular smooth muscle cell activation, foam cell transformation, and proliferation. Circulation. 2000; 101: 1799–1805.

Klouche M, Gottschling S, Gerl V, Hell W, Husmann M, Dorweiler B, Messner M, Bhakdi S. Atherogenic properties of enzymatically degraded LDL: selective induction of MCP-1 and cytotoxic effects on human macrophages. Arterioscler Thromb Vasc Biol. 1998; 18: 1376–1385.

Klouche M, May AE, Hemmes M, Messner M, Kanse SM, Preissner KT, Bhakdi S. Enzymatically modified, nonoxidized LDL induces selective adhesion and transmigration of monocytes and T-lymphocytes through human endothelial cell monolayers. Arterioscler Thromb Vasc Biol. 1999; 19: 784–793.

Torzewski M, Klouche M, Hock J, Messner M, Dorweiler B, Torzewski J, Gabbert H E, Bhakdi S. Immunohistochemical demonstration of enzymatically modified human LDL and its colocalization with the terminal complement complex in the early atherosclerotic lesion. Arterioscler Thromb Vasc Biol. 1998; 18: 369–378.

Kruth HS. Localization of unesterified cholesterol in human atherosclerotic lesions. Demonstration of filipin-positive, oil-red-O-negative particles. Am J Pathol. 1984; 114: 201–208.

Simionescu N, Vasile E, Lupu F, Popescu G, Simionescu M. Prelesional events in atherogenesis. Accumulation of extracellular cholesterol-rich liposomes in the arterial intima and cardiac valves of the hyperlipidemic rabbit. Am J Pathol. 1986; 123: 109–125.

Chao FF, Amende LM, Blanchette-Mackie EJ, Skarlatos SI, Gamble W, Resau JH, Mergner WT, Kruth HS. Unesterified cholesterol-rich lipid particles in atherosclerotic lesions of human and rabbit aortas. Am J Pathol. 1988; 131: 73–83.

Sakurada T, Orimo H, Okabe H, Noma A, Murakami M. Purification and properties of cholesterol ester hydrolase from human aortic intima and media. Biochim Biophys Acta. 1976; 424: 204–212.

Hajjar DP, Weksler BB, Falcone DJ, Hefton JM, Tack-Goldman K, Minick CR. Prostacyclin modulates cholesteryl ester hydrolytic activity by its effect on cyclic adenosine monophosphate in rabbit aortic smooth muscle cells. J Clin Invest. 1982; 70: 479–488.

Hakala JK, Oksjoki R, Laine P, Du H, Grabowski GA, Kovanen PT, Pentikainen MO. lysosomal enzymes are released from cultured human macrophages hydrolyze LDL in vitro and are present extracellularly in human atherosclerotic lesions. Arterioscler Thromb Vasc Biol. 2003; 23: 1430–1436.

Han SR, Momeni A, Strach K, Suriyaphol P, Fenske D, Paprotka K, Hashimoto SI, Torzewski M, Bhakdi S, Husmann M. Enzymatically modified LDL induces cathepsin H in human monocytes: potential relevance in early atherogenesis. Arterioscler Thromb Vasc Biol. 2003; 23: 661–667.

Suriyaphol P, Fenske D, Z?hringer U, Han SR, Bhakdi S, Husmann M. Enzymatically modified nonoxidized low-density lipoprotein induces interleukin (IL)-8 in human endothelial cells: role of free fatty acids. Circulation. 2002; 106: 2581–2587.

Bhakdi S, Roth M, Nürnberger W. A simple method for quantitative measurement of C3d in human plasma. J Immunol Methods. 1984; 74: 79–86.

Stary HC. Natural history and histological classification of atherosclerotic lesions: an update. Arterioscler Thromb Vasc Biol. 2000; 20: 1177–1178.

Chao FF, Blanchette-Mackie EJ, Tertov VV, Skarlatos SI, Chen YJ, Kruth HS. Hydrolysis of cholesteryl ester in low density lipoprotein converts this lipoprotein to a liposome. J Biol Chem. 1992; 267: 4992–4998.

Galis ZS, Sukhova GK, Lark MW, Libby P. Increased expression of matrix metalloproteinases and matrix degrading activity in vulnerable regions of human atherosclerotic plaques. J Clin Invest. 1994; 94: 2493–2503.

Lee RT, Schoen FJ, Loree HM, Lark MW, Libby P. Circumferential stress and matrix metalloproteinase 1 in human coronary atherosclerosis. Implications for plaque rupture. Arterioscler Thromb Vasc Biol. 1996; 16: 1070–1073.

Carmeliet P, Moons L, Lijnen R, Janssens S, Lupu F, Collen D, Gerard RD. Inhibitory role of plasminogen activator inhibitor-1 in arterial wound healing and neointima formation: a gene targeting and gene transfer study in mice. Circulation. 1997; 96: 3180–3191.

Carmeliet P, Moons L, Lijnen R, Baes M, Lemaitre V, Tipping P, Drew A, Eeckhout Y, Shapiro S, Lupu F, Collen D. Urokinase-generated plasmin activates matrix metalloproteinases during aneurysm formation. Nat Genet. 1997; 17: 439–444.

Gnasso A, Motti C, Irace C, Carallo C, Liberatoscioli L, Bernardini S, Massoud R, Mattioli PL, Federici G, Cortese C. Genetic variation in human stromelysin gene promoter and common carotid geometry in healthy male subjects. Arterioscler Thromb Vasc Biol. 2000; 20: 1600–1605.

Seifert PS, Hugo F, Hansson GK, Bhakdi S. Prelesional complement activation in experimental atherosclerosis. Terminal C5b-9 complement deposition coincides with cholesterol accumulation in the aortic intima of hypercholesterolemic rabbits. Lab Invest. 1989; 60: 747–754.

Bhakdi S. Complement and atherogenesis: the unknown connection. Ann Med. 1998; 30: 503–507.

Bhakdi S, Lackner KJ, Han S-R, Torzewski M, Husmann M. Beyond cholesterol: the enigma of atherosclerosis revisited. Thromb Haemos. 2004; 91: 639–645.

Bhakdi S, Torzewski M, Paprotka K, Schmitt S, Barsoom H, Suriyaphol P, Han S-R, Lackner KJ, Husmann M. Possible protective role for CRP in atherogenesis: complement activation by modified lipoproteins halts before the detrimental terminal sequence. Circulation. 2004; 109: 1870–1876.

Baumstark MW, Kreutz W, Berg A, Frey I, Keul J. Structure of human low-density lipoprotein subfractions, determined by X-ray small-angle scattering. Biochim Biophys Acta. 1990; 1037: 48–57.

Winkler K, Konrad T, Fullert S, Friedrich I, Destani R, Baumstark MW, Krebs K, Wieland H, Marz W. Pioglitazone reduces atherogenic dense LDL particles in nondiabetic patients with arterial hypertension: a double-blind, placebo-controlled study. Diabetes Care. 2003; 26: 2588–2594.

Loftus IM, Naylor AR, Goodall S, Crowther M, Jones L, Bell PR, Thompson MM. Increased matrix metalloproteinase-9 activity in unstable carotid plaques. A potential role in acute plaque disruption. Stroke. 2000; 31: 40–47.

Pasterkamp G, Schoneveld AH, Hijnen DJ, de Kleijn DP, Teepen H, van der Wal AC, Borst C. Atherosclerotic arterial remodeling and the localization of macrophages and matrix metalloproteases 1, 2 and 9 in the human coronary artery. Atherosclerosis. 2000; 150: 245–253.

Silence J, Lupu F, Collen D, Lijnen HR. Persistence of atherosclerotic plaque but reduced aneurysm formation in mice with stromelysin-1 (MMP-3) gene inactivation. Arterioscler Thromb Vasc Biol. 2001; 21: 1440–1445.

Galis ZS, Khatri JJ. Matrix metalloproteinases in vascular remodeling and atherogenesis: the good, the bad, and the ugly. Circ Res. 2002; 90: 251–262.

Libby P. Inflammation in atherosclerosis. Nature. 2002; 420: 868–874.