Differential expression of oxidation-specific epitopes and apolipoprotein(a) in progressing and ruptured human coronary and carotid atherosclerotic lesions.

The relationships between oxidation-speciﬁ c epitopes OxLDL. collagen and monitored in a single analysis. In (+)ESI mode, the fragment produced by all protonated PC species, regardless of their parent mass or moiety, is the phosphocholine headgroup with m/z 184. Thus, in the same analysis, 16:0–05:0 (ALDO) PC were monitored employing the MRM pairs 650/184 and 664/184, respectively.

plaque components, independent of dominant plaque type. Those lesions that were prone to lipid accumulation, either intracellular or extracellular, were identifi ed. Foam cell lesions were defi ned as areas of macrophages in the presence or absence of signifi cant extracellular lipid (intimal xanthoma) ( 15 ). Lipid pools within pathologic intimal thickening (PIT) consisted of a proteoglycan-rich matrix with trapped lipid and absence of fibrin and hemorrhage in the deeper intima. These pools were often surrounded by macrophage foam cell or SMC-rich areas toward the lumen. Necrotic core denoted focal areas of necrotic debris with presence of apoptotic macrophage debris, prominent cholesterol crystals, and presence of fi brin with partial or complete loss of proteoglycan matrix ( 15 ). The presence of macrophages within the fi brous cap or shoulder region denoted cases of early and late fi broatheroma, thin fi brous cap atheroma, and ruptured plaques.

Antibodies
Five unique monoclonal antibodies were used in this study to assess the presence of apolipoprotein B-100, OSE, and apo(a). MB47 is an IgG murine monoclonal antibody that binds near the LDL-receptor domain of human apoB-100 ( 18 ). MB47 binds to all apoB-containing lipoproteins equally, and it also binds to fragments of apoB-100 on OxLDL if it is minimally to even extensively modifi ed during oxidation by exposure to copper in vitro. E06 is a natural IgM murine monoclonal antibody cloned from apoE Ϫ / Ϫ mice that binds the PC headgroup of oxidized phospholipids (OxPL) and thus recognizes this whether the OxPL is free or covalently bound to proteins. Covalent binding of OxPL occurs via the reactive oxidized moieties, such as aldehydes, generated on the sn2 side chains when the phospholipids are oxidized. The PC is preserved in this setting and is the moiety recognized by E06. E06 recognizes a variety of OxPL with varying sn2 chain lengths terminated by aldehydes, such as 5, 6, and even 9 carbon lengths. It would not recognize oxovaleryl bound to protein unless it was present as the sn2 side chain of an OxPL ( 19,20 ). MDA2 binds to malondialdehyde (MDA) adducts with lysine residues of proteins, in which MDA acts as a hapten on a protein carrier. It thus recognizes a wide variety of MDA-modifi ed proteins ( 21 ). IK17 is a fully human Fab fragment generated with phage display library technology that also binds to MDA adducts with lysine, but it appears to be more specifi c for MDA-modifi ed LDL, as it does not bind to MDA-modifi ed BSA or polylysine ( 22 ).The chemistry of MDA modifi cation is complex; we believe IK17 detects a more complex MDA adduct, although we have not defi nitely defi ned it. Uniquely, IK17 also binds to OxLDL, whereas MDA2 does not. Thus its epitope appears to be an MDA adduct that is present on both MDA-LDL and OxLDL. It does not bind OxPL. LPA4 is a murine monoclonal IgG antibody binding the sequence TRNYCRNPDAEIRP on apolipoprotein(a), and it does as increased collagen and smooth muscle cell (SMC) expression, and a decrease in reactive oxygen species and macrophages (6)(7)(8).
Despite this wealth of animal data on the relationship of OSE and atherosclerosis, relatively little is known about their relationship to clinically relevant advanced, unstable, or ruptured plaques. Furthermore, a systematic analysis of the presence of OSE in human lesions has not been performed to date. Therefore, the purpose of this study was to determine the presence and relative distribution of well-characterized OSE in various stages of human atherosclerotic lesions, including native coronary lesions, carotid endarterectomy samples, and material from carotid and saphenous vein graft (SVG) embolic protection fi lters. Such knowledge may have signifi cant clinical implications with the emergence in the clinical and translational arenas of oxidative biomarkers, molecular imaging, and therapeutic approaches, including immune modulation and vaccine approaches targeting these moieties (9)(10)(11)(12), broadly characterized as "biotheranostic" (biomarker, therapeutic, diagnostic imaging) applications.

Human atherosclerotic lesions
Hearts of patients who had died suddenly with coronary artery disease (CAD) were obtained as previously described ( 13 ). Cases were identifi ed prospectively by the presence and type of CAD and included nonatherosclerotic intimal lesions, pathologic intimal thickening, early and late fi broatheroma, thin cap fi broatheroma (TCFA) and plaque rupture. Eightynine representative lesions from 25 consecutive patients (22 men and 3 women, age at death 47 ± 13) were selected prior to staining.
To rule out postmortem oxidation occurring prior to heart harvesting, we also evaluated carotid endarterectomy specimens (n = 13) from symptomatic patients undergoing clinically indicated procedures. The specimens were removed en bloc and immediately fi xed in formalin. Additionally, we evaluated material derived from distal protection devices (n = 10) obtained during percutaneous intervention of stenotic internal carotid arteries and coronary SVGs. The entire fi lter material was immediately placed in ice-cold phosphate buffered solution of EDTA/BHT (100 µM / 20 µM), and then rapidly lipid extracted and stored at Ϫ 80°C for analyses as described below.
To further rule out postmortem effects, we additionally evaluated fi ve carotid endarterectomy specimens under various handling conditions as follows: each specimen was manually cut into three equal sections and stored at room temperature for 24 h in PBS (phosphate buffered saline), on ice for 24 h in PBS, or on ice in EDTA/BHT for 24 h, respectively. Each of the specimens was then paraffi n embedded, and sections were placed on glass slides and immunostained as above.

Histological preparation
Formalin-fi xed, paraffi n-embedded coronary segments were cut into 5 µm thick sections, mounted on charged slides, and stained with hematoxylin and eosin (H and E) and the modifi ed Movat pentachrome method as previously described ( 14 ).

2775
Total lipid extraction and LC-MS/MS of material from distal protection devices All distal protection devices were of the fi lter variety (Filter-Wire EZ, 110 µm pores, Boston Scientifi c; Accunet Rx, 100 µm pores, Abbott). Filter material was subjected to a Folch lipid extraction with chloroform/methanol (2:1). For total lipid extraction, 500 l of fi lter material homogenates was transferred into a glass tube, 2.5 ml of ice-cold chloroform/methanol and 17:1/17:1 PC were added as internal standards, and the tubes were vortexed at a maximum speed for 30 s. After centrifugation, the lower organic phase was transferred into a fresh glass tube using a Pasteur pipette, and the organic phase was dried under argon to ‫ف‬ 200 l and stored at Ϫ 80°C.
Isocratic high performance liquid chromatography (HLPC) was carried out using a Shimadzu (Columbia, MD) LC-10AD high-performance pump interfaced with a Shimadzu SCL-10A controller. Sample was injected onto a 2.1 mm × 250 mm Vydac (Hysperia, CA) C18 column (Vydac catalog number 201TP52) held at 40°C using a Leap Technologies (Carrboro, NC) PAL autosampler. A buffer of isopropyl alcohol/water/tetrahydrofuran (40/40/20, v/v/v) with 0.2% formic acid at a fl ow rate of 300 l/ min was used for sample elution. The eluate was coupled to a mass spectrometer for further analysis. Separation optimization and verifi cation of HPLC retention times were achieved using 16:0-05:0 (ALDO) PC standard.
All of the mass spectral analyses were performed using an Applied Biosystems (Foster City, CA) 4000 QTrap hybrid quadrupole linear ion trap mass spectrometer equipped with a Turbo V ion source, as previously described and validated ( 24 ). Protonated adducts of the 1-palmitoyl-2-(5 ′ -oxo-valeroyl)-sn -glycero-3phosphocholine (POVPC) were formed using the following settings: CUR, 10 psi; GS1, 40 psi; GS2, 0 psi; IS, 5500V; CAD, high; temperature, 500°C; ihe, ON; DP, 70V; CE, 35V; EP, 15V; and CXP, 15V. The 4000 QTrap is capable of carrying out tandem mass spectrometry, where a specifi ed precursor ion (denoted by its mass-to-charge ratio, m/z ) can be isolated in the fi rst sector of the instrument, fragmented in a second sector collision cell, and the fragments produced then identifi ed by their m/z in a third sector. A specialized form of tandem mass spectrometry is multiple reaction monitoring (MRM), in which multiple MRM not cross-react with plasminogen ( 23 ). All preparations were greater than 99% pure.

Immunohistochemistry
Formalin-fi xed paraffi n sections (5 µm) were incubated overnight at 4°C with primary antibodies MDA2, MB47, E06, and LPA4 at respective dilutions of 1:400, 1:50, 1:1,200, and 1:400. The detection of primary antibodies bound to their respective antigen was achieved using the biotinylated link antibody LSAB2 System-HRP DAB kit (Dako, Carpenteria, CA) with appropriate secondary antibodies directed to mouse IgG or IgM. Histologic sections for antibody staining against IK17 were initially incubated overnight with nonimmune goat anti-human IgG (GAH, Vector, BA-3000) at a dilution of 1:100 in 2% goat serum to reduce nonspecifi c background staining. For IK17 immunostaining, IK17 was diluted 1:600 in 2% goat serum and incubated for 1 h at room temperature (RT). Primary labeling was then visualized using an alkaline phosphatase-labeled goat anti-human secondary antibody (dilution 1:200, Sigma A3813) for 1 h at RT and visualized with Vector Red (Vector SK-5100).

Assessment of immunolocalization of OSE, apo(a), and macrophage markers
The degree of MB47, MDA2, E06, LPA4, and IK17 and macrophage marker positivity was assessed qualitatively and quantitatively. Qualitative assessment within plaque components was performed on a scale of 0-3+: 0 (absent); + <10% of component area; ++ = 10-50% of component area; +++ > 51% of component area. Morphometric measurements of coronary sections were performed using image-processing software (IPLabs, Scanalytics, Rockville, MD) on slides stained with Movat Pentachrome. Quantitative planimetry with computer-assisted color image analysis segmentation with background correction quantifi ed immunohistochemical stains of OSE for each antibody within regions of interest. foamy macrophages, but they were almost absent in the necrotic core and shoulder regions. OxPL, IK17, and apo(a) epitopes were nearly uniformly expressed in SMC and macrophage-rich areas. IK17 epitopes were more prevalent in the necrotic core compared with the other epitopes. The fi brous cap expressed minimal apoB-100 and MDA epitopes, but OxPL, IK17, and apo(a) epitopes were present in 13-22% of the fi brous cap area.
Quantitative immunostaining patterns across plaque types and epitopes. ACROSS PLAQUE TYPES. The extent of macrophage staining was most prevalent in TCFA and plaque rupture, less so in IX, EFA, and LFA, and minimally in AIT and PIT ( Table 3 ). Macrophages were primarily present in the necrotic core and within the fi brous cap and increased with advancing plaque types. ApoB-100 and MDA staining was minimal in most lesion types, representing only 3-4% at most of total lesion area. Staining for apoB-100 was generally within the extracellular space, most prominent within areas of lipid pools and SMCs. OxPL and IK17 epitopes were strongly present in plaque rupture, TCFA, and IX, less so in PIT, EFA, and LFA, and virtually absent in AIT. The predominant site of OxPL staining was in the necrotic core and in foam cell macrophages. IK17 staining was most prominent in the necrotic core with strong positive staining also seen in macrophages. Apo(a) epitopes were strongly present in PR and TCFA, and they were modestly present in IX, PIT, EFA, and LFA, suggesting more nonspecifi c accumulation. Apo(a) staining was intense in the necrotic core, as well as in foam cell and nonfoamy macrophages. There was strong staining for OxPL, apo(a), and IK17 epitopes within the necrotic core of ruptured plaques (mean of total percentage OxPL, apo(a), and IK17 stained area, respectively, 60%, 57%, and 72%). OxPL and apo(a) were strongly expressed in foamy cell macrophages within the fi brous cap, whereas IK17 epitopes were mostly absent within viable macrophages.
ACROSS EPITOPE TYPES. In normal coronary arteries with AIT, staining of macrophages and all OSE and apo(a) was negligible (data not shown). In contrast, IX contained few apoB-100 and MDA epitopes, signifi cantly more macrophage, OxPL, and apo(a) epitopes, and fewer IK17 pairs can be monitored in a single analysis. In (+)ESI mode, the fragment produced by all protonated PC species, regardless of their parent mass or moiety, is the phosphocholine headgroup with m/z 184. Thus, in the same analysis, 16:0-05:0 (ALDO) PC were monitored employing the MRM pairs 650/184 and 664/184, respectively.

Statistical analysis
Mean variables between the various lesions were compared with one-way ANOVA (ANOVA; SPSS) followed by Student t -test for all differences among means. Spearman's correlation was used to demonstrate the relationship between macrophages and OSE. A value of P р 0.05 was considered statistically signifi cant.

Qualitative and quantitative patterns of coronary artery immunostaining
Qualitative immunostaining patterns of macrophage, OSE, and apo(a) by plaque composition. The immunostaining patterns and intensity by plaque composition in all plaque types are displayed in Table 1 . The macrophage marker KP-1 strongly stained foam cells and the necrotic core, as expected. ApoB-100, detected by MB47, was seen primarily in early lesions in lipid pools and in lesser amounts in SMC-rich areas and the necrotic core. Foamy macrophages did not stain for apoB-100 with MB47. Malondialdehyde (MDA) epitopes, detected by MDA2, demonstrated a similar pattern of staining as apoB-100 but showed some positivity in foamy macrophages. OxPL epitopes, detected by E06, were strongest in macrophage-rich areas, lipid pools, and the necrotic core. IK17 epitopes were strongest within necrotic cores and were mainly present in areas rich in foamy macrophages. Apo(a) epitopes, detected by LPA4, were consistently present throughout early and late lesions, especially in macrophages and the necrotic core.
Quantitative immunostaining patterns of macrophages and oxidation markers by plaque composition and cell type in all plaque types. Macrophage expression was primarily present in fi brous caps, shoulder areas, and the necrotic core ( Table 2 ). ApoB-100 and MDA epitopes were primarily present in macrophage-rich areas in the fi brous cap and differences were noted in expression of epitopes in all six lesion types by ANOVA (all P < 0.0001). In posttest Bonferonni analysis among paired comparison of epitope types, in general, signifi cant differences were not present between apoB and MDA2 epitopes, but they were present epitopes ( Fig. 1 ). PIT lesions were enriched in apo(a) epitopes. EFA and LFA were strongly positive for macrophages, OxPL, IK17, and apo(a) epitopes, whereas TCFA and plaque ruptures showed enhanced staining for macrophages, OxPL, apo(a), and IK17 epitopes. Signifi cant  Thin cap fi broatheroma. Immunostaining of a thin cap fi broatheroma ( Fig. 6 ) shows an intact thin fi brous cap overlying a late necrotic core. SMCs are conspicuously absent in the fi brous cap but present in the media and intimal/medial interface. There is marked infi ltration of macrophages in the thin fi brous cap and underlying necrotic core. ApoB-100 is predominantly localized to the necrotic core and surrounding acellular areas, and MDA epitopes are seen in similar areas, except the intensity of staining is markedly weaker. OxPL are primarily present in the fi brous cap rich in macrophages and in the extracellular matrix, including in the periphery of the necrotic core. There is intense staining for apo(a) in the fi brous cap and regions of the necrotic core, mostly in the extracellular matrix. IK17 is strongly positive in the late necrotic core.
Plaque rupture. Immunostaining of plaque rupture ( Fig. 7 ) shows a coronary lesion with fi brous cap disruption and superimposed luminal thrombus. SMCs are present in the medial layer but not in the plaque. In contrast, there is marked infi ltration of macrophages in the fi brous cap and underlying necrotic core. Staining for apoB-100 is essentially negative. MDA epitopes are moderately expressed in the fi brous cap within cells and extracellular matrix and necrotic core. OxPL are primarily localized to macrophages in the periphery of the necrotic core. There is intense reaction to apo(a) in the disrupted fi brous cap overlying the necrotic core. IK17 epitopes are present only in the late necrotic core.

Presence of OSE in carotid endarterectomy specimens
We also studied carotid endarterectomy specimens that were surgically removed en bloc and then immediately fi xed in neutral buffered formalin. Remarkably, these carotid lesions displayed fi ndings very similar to those observed with the coronary lesions. An example of immunostaining of a carotid TCFA with a necrotic core and presence of macrophages signifying the presence of OSE is shown in Fig. 8 . Similar to the fi ndings in the coronary arteries, apoB and MDA epitopes were predominantly localized to the less lytic areas of necrotic core. However, OxPL, apo(a) and IK17 epitope staining was strongly positive within macrophages and the necrotic core.

Presence of OxPL in carotid and coronary SVG distal protection devices documented by LC-MS/MS
To obtain physical evidence for the presence of oxidized lipids in these necrotic lesions, we collected plaque material released from such lesions during interventional procedures and trapped by distal protection devices. The fi lters were placed in potent antioxidants, frozen at Ϫ 70°C, and then lipid extracted and examined by LC-MS/MS for the presence of OxPL. Protonated adducts of 1-palmitoyl-2-(5 ′ -oxo-valeroyl)-sn -glycero-3-phosphocholine (POVPC) ( m/z 594), a well validated bioactive, pro-infl ammatory and pro-atherogenic OxPL ( 25,26 ) that reacts with E06 ( 27 ), were detected from both carotid and coronary distal SVG protection devices ( Fig. 9 ). This fi nding demonstrated among OxPL, apo(a), and IK17 epitopes (ranging from P < 0.05 to P < 0.001).

Morphometric patterns of immunostaining
Intimal xanthoma. Immunostaining of normal segments of coronary arteries revealed no staining with any of the oxidation-specifi c monoclonal antibodies, but there was occasional minimal endothelial cell staining noted with monoclonal antibody LPA4 for apo(a) epitopes. Immunostaining of an intimal xanthoma ( Fig. 2 ) shows an early superfi cial lesion rich in macrophages without evidence of necrosis. There is a relative absence of staining for apoB-100 and MDA epitopes but strong positive staining for OxPL, which are present in areas rich in macrophages and SMCs. IK17 and apo(a) epitopes are expressed in select populations of macrophages in a distribution different from E06.
Pathologic intimal thickening. Immunostaining of pathologic intimal thickening ( Fig. 3 ) shows a superfi cial plaque with an acellular lipid pool in a proteoglycan-rich matrix. SMCs are mostly present in the medial wall and superfi cial layers outside the lipid pool, whereas macrophages are present above the regions of the lipid pool. There is moderate staining for apoB-100 but intense staining for apo(a) in areas of the lipid pool. Staining for MDA epitopes is essentially negative. OxPL staining is primarily present in areas of superfi cial macrophages and SMCs above the area of the lipid pool, whereas weak to moderate staining for IK17 epitopes is seen in the lipid pool.
Early fi broatheroma. Immunostaining of an early fi broatheroma ( Fig. 4 ) shows a superfi cial plaque with an early necrotic core characterized by macrophages, free cholesterol, and extracellular matrix. SMCs are mostly present in the medial wall and fi brous cap, whereas macrophages stain intensely in the fi brous cap and necrotic core. Staining for apoB-100 and MDA epitopes is essentially negative. OxPL are primarily present in areas rich in macrophages in the fi brous cap, whereas weak to moderate staining for IK17 epitopes is seen in the necrotic core. Expression of apo(a) is primarily localized to the extracellular matrix of the fi brous cap.
Late fi broatheroma. Immunostaining of a late fi broatheroma ( Fig. 5 ) shows an intact thick fi brous cap overlying a late necrotic core characterized by cell debris, free cholesterol, and relative absence of extracellular matrix. SMCs are present in the media and superfi cial area of the fi brous cap. Macrophages are mainly present in the necrotic core and pericore regions, with some infi ltration in the fi brous cap. There is relatively weak staining for apoB-100 in the region of the necrotic core, and MDA epitopes are essentially absent. OxPL and apo(a) epitopes show a similar distribution and are found primarily in the extracellular matrix of the fi brous cap and necrotic core. IK17 epitopes are strongly positive in the necrotic core and absent elsewhere. that the OxPL detected in both the coronary and carotid lesions by E06 immunostaining represented epitopes present in vivo and not products of artifactual oxidation that occurred during tissue processing.

Effect of time to harvest and expression of OSE
Immunostaining of freshly procured carotid endarterectomy specimens stored at room temperature for 24 h, on ice for 24 h, or in antioxidants for 24 h demonstrated no appreciable qualitative differences in immunostaining patterns among antibodies with different methods of storage prior to analysis ( Fig. 10 ).

DISCUSSION
This is the fi rst comprehensive analysis evaluating several well-characterized OSE and apo(a) in a wide range of human coronary, carotid, and SVG atherosclerotic lesions. The study demonstrates a differential expression of apoB-100, OSE, and apo(a) in the progression of atherosclerosis from early lesions to plaque rupture. Plaques that develop during early atherosclerosis are enriched in apoB-100, apo(a), and all the OSE measured in this study. However, as lesions progress, apo(a), OxPL, and IK17 epitopes become progressively enriched. Specifi cally, apo(a) epitopes were present in most lesions, whereas OxPL and IK17 epitopes were mainly associated with foamy macrophages of the fi brous cap and the necrotic core. IK17 epitopes were most specifi cally associated with necrotic cores and plaque rupture. Additionally, a similar distribution was noted in TCFA of freshly procured carotid endarterectomy specimens. Finally, LC-MS/MS documented the presence of OxPL in material trapped by distal protection devices, documenting their presence in vivo in clinically symptomatic plaques. These observations provide a framework for understanding the relationship between OSE and Lp(a) and the progression and destabilization of human coronary and carotid atherosclerosis.
This study used unique monoclonal antibodies targeting well-characterized lipoprotein and OSE to study their relationship to human atherosclerosis, as described in Methods. For example, MB47 detects not only the native apoB-100 moiety of unoxidized LDL but also apoB of minimally oxidized LDL and even apoB fragments of extensively oxidized LDL. MDA2 can recognize both MDA-LDL as well as MDA-lysine epitopes on other proteins in the vessel wall, such as apoAI ( 28,29 ), and it has also been utilized for noninvasive imaging of experimental atherosclerotic lesions (3)(4)(5). E06 is a well-characterized murine IgM natural antibody cloned from apoE Ϫ / Ϫ mice ( 20 ) that has previously been used to stain mouse and rabbit atherosclerotic lesions and to image atherosclerotic lesions in apoE Ϫ / Ϫ mice using magnetic resonance techniques (3)(4)(5). E06 is also used in plasma immunoassays to detect OxPL on circulating apoB-100 particles (OxPL/ apoB) in humans ( 10 ). In support of these pathologic fi ndings, plasma levels of the OxPL epitopes on apoB-100 particles (OxPL/apoB) measured by E06 were recently shown to be strongly associated with the presence of angiographically defi ned coronary artery disease, to predict the presence and progression of carotid and femoral atherosclerosis, and to predict cardiovascular death, myocardial infarction, and stroke in unselected epidemiological populations ( 30,31 ). LPA4 binds to the apolipoprotein(a) portion of Lp(a) and is used in immunoassays to measure plasma Lp(a) levels, but it has not been previously used to immunostain tissues ( 23 ). Apo(a) itself is not an oxidationspecifi c epitope, but apo(a) and Lp(a) bind OxPL, which may refl ect the key atherogenic component of Lp(a) ( 32 ). IK17 is a human Fab fragment derived from a phage display library, and it binds to a unique MDA-like epitope present on both MDA-LDL and copper-oxidized LDL. IK17 has also been used in detecting and imaging atherosclerosis in LDLR Ϫ / Ϫ mice and apoE Ϫ / Ϫ mice (3)(4)(5)22 ).
The differences in immunoreactivity patterns present among the different antibodies suggest that specifi c epitopes are generated and/or enriched during different pathophysiological stages of lesion development, progression, and destabilization. Such epitopes can be generated through peroxidation of unsaturated fatty acids present on lipoproteins, phospholipids, and cell membranes in the vessel wall, as well when cells, such as macrophages, undergo apoptosis. Despite the fact that these epitopes are generated through oxidative modifi cation, expression of specifi c epitopes appears dependent on the stage of lesion progression. For example, MDA epitopes were more common in early lesions, such as in PIT and early fi broatheromas, whereas OxPL, apo(a), and IK17 epitopes were more prevalent in advanced lesions. Most of the prior data with antibody MDA2 were generated in mouse and rabbit models, with early to intermediate lesions showing diffuse staining patterns both intra-and extracellularly ( 33 ). MDA2 has been used to stain early lesions in teenage subjects in the Bogalusa Heart Study ( 34 ). Surprisingly, in the current study, MDA epitopes did not become more prevalent as lesions advanced. It may be postulated that MDA epitopes are generated early in the development of atherosclerosis and may more closely refl ect plaque initiation and plaque accumulation rather than plaque destabilization. It is also possible that there is decomposition of these epitopes in more aged tissues. We have recently shown that complement factor H also binds MDA epitopes, both OxPL epitopes as lesions progress. Alternatively, changes in extracellular matrix components in later stage lesions may allow greater accumulation of noncell-associated epitopes. Similarly, although both the carotid and coronary plaque are considered TCFA, the extent of macrophages, particularly in the cap region, appears much greater for the carotid, and E06 staining is present both in the macrophages of the cap and cap ECM, whereas relatively less in the necrotic core. These data in human lesions that span a broad array from early lesions to plaque rupture can only be tangentially compared with animal data in which most lesions studied are at the fattystreak level. Nonetheless, the staining patterns for equivalent lesions seem similar, and these epitopes are equally expressed in animals and humans and are not species specifi c, but specifi c to the oxidative pathways generating these epitopes.
The origin and generation of the OSE merits some discussion. ApoB-100 is present on VLDL, VLDL remnants (IDL), LDL, and Lp(a). Once such apoB-containing lipoproteins enter the arterial intima, they bind to extracellular matrix ( 37 ) and then undergo oxidation mediated by a variety of free radical-mediated mechanisms. This likely in the eyes of patients with macular degeneration and in human coronary atherectomy specimens ( 35 ). In fact, prior side-by-side staining of such tissues with anti-CFH antibodies and MDA2 has shown partial colocalization but also independent staining. It is possible that CFH may bind and mask some of these epitopes. It is also possible that these epitopes may be bound by autoantibodies to MDA-LDL that have been well characterized ( 36 ). Additionally, some of these epitopes may be transported out of the lesion wall on migrating macrophages or lipoproteins.
In contrast, apo(a), OxPL, and IK17 epitopes were highly prevalent in foamy macrophages in thin fi brous caps, necrotic cores, and ruptured plaques, suggesting that they more closely refl ect advancing and unstable plaques. In conjunction with these fi ndings in late lesions, there was also a similarly increased expression of macrophage markers and a lack of SMC, consistent with the pathophysiological fi ndings of plaque vulnerability. Thus, the expression of apo(a), OxPL, and IK17 epitopes may more closely refl ect plaque infl ammation, destabilization, and rupture. Some variability was noted in the earlier stage coronary lesions in cell-associated OxPL staining, and some of this may be due to destruction of IK17 epitopes. IK17 epitopes, which appear primarily in the necrotic core where macrophage foam cells and necrotic debris have accumulated, may represent a byproduct of very advanced oxidation of MDA-related adducts. Interestingly, prior attempts in our laboratory to detect explains the presence of apoB and MDA epitopes in relatively early lesions. As plaques progress, macrophage number and activation increase, which among other properties, leads to further oxidation of LDL, resulting in progressive lipid oxidation and accumulation of OxPL and IK17 epitopes on circulating lipoproteins, which appear to refl ect advanced MDA-related epitopes, have been unsuccessful (S. Tsimikas, unpublished observations) , suggesting that they may only be present in highly oxidized lipids and modifi ed proteins within advanced lesions. It is important to note that both E06 and IK17 bind to apoptotic cells and apoptotic debris ( 2,22 ), which are clearly enriched in late lesions and a major contributor to the unstable plaque. Indeed, Lp(a) and the OxPL it contains are likely to be a major stimulus leading to macrophage apoptosis and cell death ( 24 ). Further studies are needed to defi ne the pathways through which these various OSE are generated in vivo.
In the current study, it was clearly demonstrated that Lp(a) is quite ubiquitous in atherosclerotic plaques, even in early lesions, and that it progressively increases as lesions progress to plaque rupture. The fact that apoB, apo(a), and OxPL did not necessarily uniformly colocalize suggests that the apoB and OxPL components of Lp(a) may be degraded and removed but that the apo(a) component continues to be bound to the plaque and have a longer residence time ( 38 ). It also suggests that additional and quantitatively signifi cant amounts of OxPL are formed in the vessel wall in situ and independently of those OxPL potentially carried by Lp(a). Consistent with the current fi ndings of the enhanced presence of Lp(a) in vulnerable plaques, a correlation was previously noted with plasma Lp(a) levels, apo(a) immunopositivity in atherectomy specimens, and the severity of the clinical presentation in patients with acute coronary syndromes ( 39 ). Indeed, as noted above, Lp(a) may well promote macrophage cell death via a CD36/TLR2-dependent process, thus contributing to plaque rupture ( 24 ). With the confl uence of data showing that Lp(a) is an independent and potentially causal risk factor for cardiovascular disease ( 40,41 ) and with emerging therapeutic interventions to lower Lp(a) ( 42-44 and B. C. Kolski and S. Tsimikas, unpublished observations), these data provide a scientifi c rationale why Lp(a) may contribute to both plaque progression and plaque destabilization, and they validate potential therapeutic targeting approaches.  6. Immunostaining of OSE in human coronary TCFA or high-risk plaque for rupture. A: Human coronary lesion with an thin intact fi brous cap (arrow) overlying a late necrotic core (NC). Movat pentachrome; I, intima; M, media. B: ␣ -Actin staining for SMCs is positive in the media and intimal/medial interface, whereas the majority of plaque, in particular the fi brous cap, is negative. C: Localization of CD68 staining shows marked infi ltration of macrophages in the thin fi brous cap and underlying necrotic core. D: Staining for MD47 is predominantly localized to the necrotic core and surrounding acellular areas. E: MDA2 expression is seen in similar areas as in (D), except the intensity of staining is markedly weaker. F: E06 is found primarily localized to fi brous cap macrophages and extracellular matrix including in the periphery of the necrotic core. G: Intense reaction against LPA4 is seen in the fi brous cap and regions of the necrotic core, mostly in the extracellular matrix. H: IK17 is strongly positive in the late NC. B-G: Immunoperoxidase (brown reaction product). H: Alkaline phosphatase (red reaction product). Magnifi cation ×200.
fi ndings in the later stage lesions. However, this cannot explain the fi ndings with the carotid or SVG, as they were processed immediately in EDTA/BHT and handled optimally subsequently to maximally limit artifactual oxidation.

Limitations
Artifactual oxidation may have occurred during procurement of the coronary specimens due to delay in harvesting after sudden death, and it may have impacted the Fig. 7. Immunostaining of OSE in human coronary plaque rupture. A: Human coronary lesion with fi brous cap disruption (arrow) and superimposed luminal thrombus (Th, arrow). Movat pentachrome; I, intima; M, media. B: ␣ -Actin staining for SMCs is positive in the medial layer but negative in the plaque. C: Localization of CD68 staining shows marked infi ltration of macrophages in the fi brous cap and underlying late necrotic core (NC). D: Staining for MD47 is essentially negative. E: MDA2 expression is moderately expressed in the fi brous cap within cells and extracellular matrix and necrotic core. F: E06 is found primarily localized to macrophages in the periphery of the necrotic core. G: Intense reaction against LPA4 is seen in the disrupted fi brous cap overlying the necrotic core. H: Weak to moderate staining of IK17 is seen in the late NC. B-G: Immunoperoxidase (brown reaction product). H: Alkaline phosphatase (red reaction product). Magnifi cation ×200.

Journal of Lipid Research Volume 53, 2012
In these specimens, we saw the same pattern of OSE expression. In addition, the fact that MDA epitopes did not increase with lesion size and that the experimental study of various exposures of carotid endarterectomy specimens showed no major differences in immunostaining suggests that if postmortem oxidation did occur in the coronary specimens, it did not signifi cantly impact the fi ndings.

Clinical implications and conclusions
This comprehensive analysis of a range of human plaque types and stages of atherosclerosis demonstrates that OSE,  (A) showing a necrotic core (NC) with a overlying intact thin fi brous cap, which is partially degraded by infi ltrating macrophages. C: Cluster of macrophages within the fi brous cap (arrow; H and E staining). D: Immunostaining for MD47 is predominantly localized to the less lytic areas of necrotic core (lower left) and extracellular matrix of the fi brous cap; there is a relative absence of staining within fi brous cap macrophages. E: MDA2 expression is diffusely present in the fi brous cap, including macrophages and less lytic areas of the necrotic core; the staining intensity is the weakest of the fi ve OSE epitopes. F: EO6 is strongly positive within the fi brous cap, including macrophages, as well as connective tissue and necrotic core. G: Immunostaining against LPA4 is intense in the less lytic areas of the necrotic core and connective tissue matrix. H: Staining pattern for IK17 mirrors that of LPA4 in which a strong signal is noted in the less lytic areas of the necrotic core and connective tissue matrix. D-G: Immunoperoxidase (brown reaction product). H: Alkaline phosphatase (red reaction product). A: Magnifi cation ×10.25. B-H: Magnifi cation ×100. in particular those detected by antibodies E06, IK17, and LPA4, are highly prevalent and progressively enriched in advancing atherosclerotic lesions and ruptured plaques. Several translational "biotheranostic" (biomarker, therapeutic, and diagnostic imaging) approaches targeting OSE, such as in-vitro assays, in-vivo and noninvasive diagnostic imaging modalities, and therapeutic applications have reached the investigational clinical frontier ( 9,11,12 ). As these studies ultimately make their way to clinical applications, the data generated in this study will provide a foundation for appropriate clinical applications and interpretation of the fi ndings. Furthermore, the presence of vasoactive OxPL within distal protection devices is consistent with our prior data showing increased OxPL in plasma following PCI ( 23 ), and it suggests the design of studies with therapies directed to binding and inactivating OxPL or other OSE released during percutaneous coronary, carotid, SVG, or peripheral interventions. For example, this could be achieved by infusing oxidation-specifi c antibodies or similarly directed therapies at time of presentation of an acute vascular syndrome or prior to intervention, such as with humanized E06-type antibodies or the human antibody IK17, which has been shown to bind OSE, prevent foam cell formation, and reduce atherosclerosis progression when infused in atherosclerotic mice ( 9 ). Because this study is retrospective and cross-sectional, it does not prove causality for the role of OSE in plaque progression and rupture. Future studies using human OSE antibodies therapeutically and showing clinical effi cacy will be needed to prove the role of OSE in mediating CVD events.