Three lines of evidence are presented that low density lipoproteins gently extracted from human and rabbit atherosclerotic lesions (lesion LDL) greatly resembles LDL that has been oxidatively modified in vitro. First, lesion LDL showed many of the physical and chemical properties of oxidized LDL, proerties that differ from those of plasma LDL: higher electrophoretic mobility, a higher density, higher free cholesterol content, and a higher proportion of sphingomyelin and lysophosphatidylcholine in the phospholipid fraction. A number of lower molecular weight fragments of apo B were found in lesion LDL, similar to in vitro oxidized LDL. Second, both the intact apo B and some of the apo B fragments of lesion LDL reacted in Western blots with antisera that recognize malondialdehyde-conjugated lysine and 4-hydroxynonenal lysine adducts, both of which are found in oxidized LDL; plasma LDL and LDL from normal human intima showed no such reactivity. Third, lesion LDL shared biological properties with oxidized LDL: compared with plasma LDL, lesion LDL produced much greater stimulation of cholesterol esterification and was de--graded more rapidly by macrophages. Degradation of radiolabeled lesion LDL was competitively inhibited by unlabeled lesion LDL, by LDL oxidized with copper, by polyinosinic acid and by malondialdehyde-LDL, but not by native LDL, indicating uptake by the scavenger receptor(s). Finally, lesion LDL (but not normal intimal LDL or plasma LDL) was chemotactic for monocytes, as is oxidized LDL. These studies provide strong evidence that atherosclerotic lesions, both in man and in rabbit, contain oxidatively modified LDL.

Entry of monocytes into the vessel wall is an important event in atherogenesis. Previous studies from our laboratory suggest that oxidized arachidonic acid-containing phospholipids present in mildly oxidized low density lipoproteins (MM-LDL) can activate endothelial cells to bind monocytes. In this study, biologically active oxidized arachidonic acid-containing phospholipids were produced by autoxidation of 1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphocholine (Ox-PAPC) and analyzed by liquid chromatography and electrospray ionization mass spectrometry in conjuction with biochemical derivatization techniques. We have now determined the molecular structure of two of three molecules present in MM-LDL and Ox-PAPC that induce monocyte-endothelial interactions. These lipids were identified as 1-palmitoyl-2-(5-oxovaleryl)-sn-glycero-3-phosphocholine (m/z 594.3) and 1-palmitoyl-2-glutaryl-snglycero-3-phosphocholine (m/z 610.2). These two molecules were produced by unambiguous total synthesis and found to be identical by analytical techniques and bioactivity assays to those present in MM-LDL and Ox-PAPC. Evidence for the importance of all three oxidized phospholipids in vivo was suggested by their presence in fatty streak lesions from cholesterol-fed rabbits and by their immunoreactivity with natural antibodies present in ApoE null mice. Overall, these studies suggest that specific oxidized derivatives of arachidonic acidcontaining phospholipids may be important initiators of atherogenesis.

It has been proposed that low density lipoprotein (LDL) must undergo oxidative modification before it can give rise to foam cells, the key component ofthe fatty streak lesion of atherosclerosis. Oxidation of LDL probably generates a broad spectrum of conjugates between fragments of oxidized fatty acids and apolipoprotein B. We now present three mutually supportive lines of evidence for oxidation of LDL in vivo: (i) Antibodies against oxidized LDL, malondialdehyde-lysine, or 4-hydroxynonenal-lysine recognize materials in the atherosclerotic lesions of LDL receptor-deficient rabbits; (ii) LDL gently extracted from lesions of these rabbits is recognized by an antiserum against malondialdehyde-conijugated LDL; (iii) autoantibodies against malondialdehyde-LDL (titers from 512 to >4096) can be demonstrated in rabbit and human sera.A growing body of evidence suggests that oxidative modification of low density lipoprotein (LDL) enhances its atherogenicity (for review see ref. 1). Monocyte-derived macrophages, the precursor of most foam cells in early atherosclerotic lesions, cannot take up native LDL rapidly enough to cause lipid loading (2). Oxidative modification converts LDL to a form recognized by the macrophage acetyl-LDL receptor (1, 2) and possibly by other receptors as well (3). This is true whether the oxidation is effected by incubation under appropriate conditions with cultured cells or by autooxidation catalyzed by Cu2+ ions in the absence of cells. Oxidative modification of LDL is accompanied by extensive degradation of its polyunsaturated fatty acids, generating a complex array of shorter chain-length fragments (4). During the oxidation, some of these fragments become covalently linked to apolipoprotein B (5), and much ofthis conjugation involves the E-amino groups of lysine residues. This chemically modified form of apolipoprotein is recognized by the acetyl-LDL receptor (6). Thus we can generate models for oxidized LDL by conjugating the apolipoprotein with single compounds generated during oxidation. Fogelman et al. (7) demonstrated that malondialdehyde (MDA)-conjugated LDL is so recognized. If oxidized LDL contains lysine residues conjugated with a variety of fatty acid fragments of different chain lengths, it should react with antibodies against a variety of such lysine derivatives. We previously showed that immunization of animals with autologous LDL modified by conjugation of lysine groups with glucose yields antisera directed specifically against glucitollysine (8) and that antibodies generated by injection of carbamoylated autologous LDL generates antisera that recognize the carbamoyllysine-not only in LDL but in other conjugated proteins as well (9). In other words, the specificity of these antisera is for very narrowly defined "X"-lysine adducts.The present studies, which use immunochemical methods, offer three lines of evidence that oxidation of LDL occurs in vivo.

Oxidative modification of LDL renders it immunogenic and autoantibodies to epitopes of oxidised LDL, such as malondialdehyde (MDA)-lysine, are found in serum and recognise material in atheromatous tissue. However, there has been no prospective study to assess the importance of oxidised LDL among patients with vascular disease. We compared the titre of autoantibodies to MDA-modified LDL and native LDL in baseline serum samples of 30 eastern Finnish men with accelerated two-year progression of carotid atherosclerosis and 30 age-matched controls without progression. Neither group had specific antibody binding to native LDL. A titre was defined as a ratio of antibody binding to MDA-LDL/binding to native LDL. Cases had a significantly higher titre to MDA-LDL (2.67 vs 2.06, p = 0.003). Cases also had a greater proportion of smokers (37% vs 3%), higher LDL cholesterol (4.2 mmol/l vs 3.6 mmol/l), and higher serum copper concentration (1.14 mg/l vs 1.04 mg/l). Even after adjusting for these variables and the severity of baseline atherosclerosis, the difference in antibody titre remained significant in a multifactorial logistic model (p = 0.031). Thus, the titre of autoantibodies to MDA-LDL was an independent predictor of the progression of carotid atherosclerosis in these Finnish men. Our data provide further support for a role of oxidatively modified LDL in atherogenesis.

To determine whether oxidized LDL enhances atherogenesis by promoting monocyte recruitment into the vascular intima, we investigated whether LDL accumulation and oxidation precede intimal accumulation of monocytes in human fetal aortas (from spontaneous abortions and premature newborns who died within 12 h; fetal age 6.2 Ϯ 1.3 mo). For this purpose, a systematic assessment of fatty streak formation was carried out in fetal aortas from normocholesterolemic mothers ( n ϭ 22), hypercholesterolemic mothers ( n ϭ 33), and mothers who were hypercholesterolemic only during pregnancy ( n ϭ 27). Fetal plasma cholesterol levels showed a strong inverse correlation with fetal age ( R ϭ Ϫ 0.88, P Ͻ 0.0001). In fetuses younger than 6 mo, fetal plasma cholesterol levels correlated with maternal ones ( R ϭ 0.86, P ϭ 0.001), whereas in older fetuses no such correlation existed. Fetal aortas from hypercholesterolemic mothers and mothers with temporary hypercholesterolemia contained significantly more and larger lesions (758,651 Ϯ 87,449 and 451,255 Ϯ 37,448 m 2 per section, respectively; mean Ϯ SD) than aortas from normocholesterolemic mothers (61,862 Ϯ 9,555 m 2 ; P Ͻ 0.00005). Serial sections of the arch, thoracic, and abdominal aortas were immunostained for recognized markers of atherosclerosis: macrophages, apo B, and two different oxidation-specific epitopes (malondialdehyde-and 4-hydroxynonenal-lysine). Of the atherogenic sites that showed positive immunostaining for at least one of these markers, 58.6% were established lesions containing both macrophage/foam cells and oxidized LDL (OxLDL). 17.3% of all sites contained only native LDL, and 13.3% contained only OxLDL without monocyte/ macrophages. In contrast, only 4.3% of sites contained isolated monocytes in the absence of native or oxidized LDL. In addition, 6.3% of sites contained LDL and macrophages but few oxidation-specific epitopes. These results demonstrate that LDL oxidation and formation of fatty streaks occurs already during fetal development, and that both phenomena are greatly enhanced by maternal hypercholesterolemia. The fact that in very early lesions LDL and Ox-LDL are frequently found in the absence of monocyte/macrophages, whereas the opposite is rare, suggests that intimal LDL accumulation and oxidation contributes to monocyte recruitment in vivo. ( J. Clin. Invest. 1997. 100:2680-2690.)