The role of high-density lipoproteins in oxidation and inflammation.

High-density lipoproteins (HDL) in the basal state are anti-inflammatory, capable of destroying oxidized lipids that generate an inflammatory response. However, HDL during acute inflammation are altered and become pro-inflammatory. This "chameleon-like" nature of HDL is considered to be due to the complex composition of HDL. The data reviewed here demonstrate the key role of HDL in modulating inflammation and its implications for atherogenesis.

High-density lipoprotein loses its anti-inflammatory properties during acute influenza a infection.

BACKGROUND: Viruses have been identified as one of a variety of potential agents that are implicated in atherogenesis. METHODS AND RESULTS: C57BL/6J mice were killed before or 2, 3, 5, 7, or 9 days after intranasal infection with 10(5) plaque-forming units (pfu) of Influenza A strain WSN/33. Peak infectivity in lungs was reached by 72 hours, and it returned to baseline by 9 days. No viremia was observed at any time. The activities of paraoxonase and platelet-activating factor acetylhydrolase in HDL decreased after infection and reached their lowest levels 7 days after inoculation. The ability of HDL from infected mice to inhibit LDL oxidation and LDL-induced monocyte chemotactic activity in human artery wall cell cocultures decreased with time after inoculation. Moreover, as the infection progressed, LDL more readily induced monocyte chemotaxis. Peak interleukin-6 and serum amyloid A plasma levels were observed at 2 and 7 days after inoculation. HDL apoA-I levels did not change. ApoJ and ceruloplasmin levels in HDL peaked 3 days after infection. Ceruloplasmin remained elevated throughout the time course, whereas apoJ levels decreased toward baseline after the third day. CONCLUSIONS: We conclude that alterations in the relative levels of paraoxonase, platelet-activating factor acetylhydrolase, ceruloplasmin, and apoJ in HDL occur during acute influenza infection, causing HDL to lose its anti-inflammatory properties.

HDL and the inflammatory response induced by LDL-derived oxidized phospholipids.

Oxidation of low density lipoprotein (LDL) phospholipids containing arachidonic acid at the sn-2 position occurs when a critical concentration of "seeding molecules" derived from the lipoxygenase pathway is reached in LDL. When this critical concentration is reached, the nonenzymatic oxidation of LDL phospholipids produces a series of biologically active, oxidized phospholipids that mediate the cellular events seen in the developing fatty streak. Normal high density lipoprotein (HDL) contains at least 4 enzymes as well as apolipoproteins that can prevent the formation of the LDL-derived oxidized phospholipids or inactivate them after they are formed. In the sense that normal HDL can prevent the formation of or inactivate these inflammatory LDL-derived oxidized phospholipids, normal HDL is anti-inflammatory. HDL from mice that are genetically predisposed to diet-induced atherosclerosis became proinflammatory when the mice are fed an atherogenic diet, injected with LDL-derived oxidized phospholipids, or infected with influenza A virus. Mice that were genetically engineered to be hyperlipidemic on a chow diet and patients with coronary atherosclerosis, despite normal lipid levels, also had proinflammatory HDL. It is proposed that LDL-derived oxidized phospholipids and HDL may be part of a system of nonspecific innate immunity and that the detection of proinflammatory HDL may be a useful marker of susceptibility to atherosclerosis.

Oxidized phospholipids induce changes in hepatic paraoxonase and ApoJ but not monocyte chemoattractant protein-1 via interleukin-6.

In this study, we tested if interleukin-6 (IL-6) plays a role in mediating the effects of oxidized phospholipids (OXPL). Treatment of HepG2 cells with oxidized 1-palmitoyl-2-arachidonyl-sn-glycero-3-phosphoryl choline (OX-PAPC), or biologically active lipids present in mildly oxidized low density lipoprotein, increased apolipoprotein J (apoJ), and decreased paraoxonase (PON) mRNA levels. Antibodies to IL-6 blocked these changes. IL-6 treatment in the absence of OXPL produced the same pattern of mRNA changes observed with OXPL treatment alone. In vivo, OX-PAPC injected into C57BL/6J mice resulted in a marked reduction in PON activity and an increase in apoJ levels in plasma after 16 h. Injection of OX-PAPC into IL-6-deficient C57BL/6J mice (IL-6 -/-) did not alter either PON activity or apoJ levels. We then tested if other mechanisms involved in fatty streak formation depended upon IL-6. Antibody to IL-6 had no effect on OX-PAPC-induced secretion of MCP-1 by endothelial cells nor on MCP-1 mRNA expression in HepG2 cells. C57BL/6J and IL-6 -/- mice fed an atherogenic diet both demonstrated markedly reduced plasma PON activities and the IL-6 -/- mice developed fatty streaks to a greater degree than wild-type mice. We conclude that IL-6 is critical to short term but not long term regulation of PON and that IL-6 is not required for OXPL regulation of MCP-1.

Overexpression of apolipoprotein AII in transgenic mice converts high density lipoproteins to proinflammatory particles.

Previous studies showed that transgenic mice overexpressing either apolipoprotein AI (apoAI) or apolipoprotein AII (apoAII), the major proteins of HDL, exhibited elevated levels of HDL cholesterol, but, whereas the apoAI-transgenic mice were protected against atherosclerosis, the apoAII-transgenic mice had increased lesion development. We now examine the basis for this striking functional heterogeneity. HDL from apoAI transgenics exhibited an enhanced ability to promote cholesterol efflux from macrophages, but HDL from apoAII transgenics and nontransgenics were not discernibly different in efflux studies. In contrast with HDL from nontransgenics and apoAI transgenics, HDL from the apoAII transgenics were unable to protect against LDL oxidation in a coculture model of the artery wall. Furthermore, HDL taken from apoAII-transgenic mice, but not HDL taken from either the apoAI transgenics or nontransgenic littermate controls, by itself stimulated lipid hydroperoxide formation in artery wall cells and induced monocyte transmigration, indicating that the apoAII-transgenic HDL were in fact proinflammatory. This loss in the ability of the apoAII-transgenic HDL to function as an antioxidant/antiinflammatory agent was associated with a decreased content of paraoxonase, an enzyme that protects against LDL oxidation. Reconstitution of the apoAII transgenic HDL with purified paraoxonase restored both paraoxonase activity and the ability to protect against LDL oxidation. We conclude that overexpression of apoAII converts HDL from an anti- to a proinflammatory particle and that paraoxonase plays a role in this transformation.

Reduced aortic lesions and elevated high density lipoprotein levels in transgenic mice overexpressing mouse apolipoprotein A-IV.

Transgenic mouse lines carrying several copies of the mouse apo A-IV gene were produced. Lipoprotein composition and function, and aortic lesion development were examined. Apo A-IV levels in the plasma of transgenic mice were elevated threefold compared with nontransgenic littermates on a chow diet, and sixfold in mice fed an atherogenic diet. Plasma concentrations of total cholesterol, HDL cholesterol, triglycerides, and free fatty acids were similar in transgenic and control mice fed a chow diet. However, with the atherogenic diet, male transgenic mice exhibited significantly higher levels of plasma triglycerides (P < 0.05), total cholesterol (P < 0.01), HDL cholesterol (P < 0.0001), and free fatty acids (P < 0.05), and lower levels of unesterified cholesterol (P < 0.05), than nontransgenic littermates. Expression of the apo A-IV transgene had a protective effect against the formation of diet-induced aortic lesions, with transgenics exhibiting lesion scores of approximately 30% those seen in control mice. HDL-sized lipoproteins isolated from transgenic mice fed the atherogenic diet promoted cholesterol efflux from cholesterol-loaded human monocytes more efficiently than comparable lipoproteins from nontransgenic counterparts. Plasma from transgenics also exhibited higher endogenous cholesterol esterification rates. Taken together, these results suggest that apo A-IV levels influence the metabolism and antiatherogenic properties of HDL.

Mildly oxidized LDL induces an increased apolipoprotein J/paraoxonase ratio.

We have examined the effects of mildly oxidized LDL and atherosclerosis on the levels of two proteins associated with HDL; apolipoprotein J (apoJ), and paraoxonase (PON). On an atherogenic diet, PON activity decreased by 52%, and apoJ levels increased 2.8-fold in fatty streak susceptible mice, C57BL/6J (BL/6), but not in fatty streak resistant mice, C3H/HeJ (C3H). Plasma PON activity was also significantly decreased, and apoJ levels were markedly increased in apolipoprotein E knockout mice on the chow diet, resulting in a 9.2-fold increase in the apoJ/PON ratio as compared to controls. Furthermore, a dramatic increase in the apoJ/PON ratio (over 100-fold) was observed in LDL receptor knockout mice when they were fed a 0.15%-cholesterol-enriched diet. Injection of mildly oxidized LDL (but not native LDL) into BL/6 mice (but not in C3H mice) on a chow diet resulted in a 59% decrease in PON activity (P < 0.01) and a 3.6-fold increase in apoJ levels (P < 0.01). When an acute phase reaction was induced in rabbits, or the rabbits were placed on an atherogenic diet, hepatic mRNA for apoJ was increased by 2.7-fold and 2.8-fold, respectively. Treatment of HepG2 cells in culture with mildly oxidized LDL (but not native LDL) resulted in reduced mRNA levels for PON (3.0-fold decrease) and increased mRNA levels for apoJ (2.0-fold increase). In normolipidemic patients with angiographically documented coronary artery disease who did not have diabetes and were not on lipid-lowering medication (n = 14), the total cholesterol/HDL cholesterol ratio was 3.1+/-0.9 as compared to 2.9+/-0.4 in the controls (n = 19). This difference was not statistically significant. In contrast, the apoJ/PON ratio was 3.0+/-0.4 in the patients compared to 0.72+/-0.2 in the controls (P < 0.009). In a subset of these normolipidemic patients (n = 5), the PON activity was low (48+/-6.6 versus 98+/-17 U/ml for controls; P < 0.009), despite similar normal HDL levels, and the HDL from these patients failed to protect against LDL oxidation in co-cultures of human artery wall cells. We conclude that: (a) mildly oxidized LDL can induce an increased apoJ/PON ratio, and (b) the apoJ/PON ratio may prove to be a better predictor of atherosclerosis than the total cholesterol/HDL cholesterol ratio.

The Yin and Yang of oxidation in the development of the fatty streak. A review based on the 1994 George Lyman Duff Memorial Lecture.

Recent data support the hypothesis that the fatty streak develops in response to specific phospholipids contained in LDL that become trapped in the artery wall and become oxidized as a result of exposure to the oxidative waste of the artery wall cells. The antioxidants present within both LDL and the microenvironments in which LDL is trapped function to prevent the formation of these biologically active, oxidized lipids. Enzymes associated with LDL and HDL (eg, platelet activating factor acetylhydrolase) or with HDL alone (eg, paraoxonase) destroy these biologically active lipids. The regulation and expression of these enzymes are determined genetically and are also significantly modified by environmental influences, including the acute-phase response or an atherogenic diet. The balance of these multiple factors leads to an induction or suppression of the inflammatory response in the artery wall and determines the clinical course.

Anti-inflammatory HDL becomes pro-inflammatory during the acute phase response. Loss of protective effect of HDL against LDL oxidation in aortic wall cell cocultures.

We previously reported that high density lipoprotein (HDL) protects against the oxidative modification of low density lipoprotein (LDL) induced by artery wall cells causing these cells to produce pro-inflammatory molecules. We also reported that enzyme systems associated with HDL were responsible for this anti-inflammatory property of HDL. We now report studies comparing HDL before and during an acute phase response (APR) in both humans and a croton oil rabbit model. In rabbits, from the onset of APR the protective effect of HDL progressively decreased and was completely lost by day three. As serum amyloid A (SAA) levels in acute phase HDL (AP-HDL) increased, apo A-I levels decreased 73%. Concomitantly, paraoxonase (PON) and platelet activating factor acetylhydrolase (PAF-AH) levels in HDL declined 71 and 90%, respectively, from days one to three. After day three, there was some recovery of the protective effect of HDL. AP-HDL from human patients and rabbits but not normal or control HDL (C-HDL) exhibited increases in ceruloplasmin (CP). This increase in CP was not seen in acute phase VLDL or LDL. C-HDL incubated with purified CP and re-isolated (CP-HDL), lost its ability to inhibit LDL oxidation. Northern blot analyses demonstrated enhanced expression of MCP-1 in coculture cells treated with AP-HDL and CP-HDL compared to C-HDL. Enrichment of human AP-HDL with purified PON or PAF-AH rendered AP-HDL protective against LDL modification. We conclude that under basal conditions HDL serves an anti-inflammatory role but during APR displacement and/or exchange of proteins associated with HDL results in a pro-inflammatory molecule.

Lipid-induced changes in intracellular iron homeostasis in vitro and in vivo.

Iron promotes cellular damage via its capacity to catalyze hydroxyl radical formation and by peroxidation of unsaturated lipids. The major cellular iron storage depot, ferritin, acts as a critical antioxidant defense by sequestering unbound or "free" iron, limiting its participation in damaging oxidative reactions. In this study, we investigated the relationship between LDL modified by artery wall cells and the regulation of intracellular free iron levels in the mouse model and in a human aortic endothelial and smooth muscle cell coculture system. We found in response to an atherogenic diet, fatty streak-resistant C3H/HeJ mice exhibited higher levels of liver apoferritin and lower intracellular concentrations of free iron than did fatty streak-susceptible C57 BL/6J mice. Also, ferritin repressor protein mRNA was not significantly suppressed after 15 wk on the atherogenic diet in female C57BL/6J mice, which exhibit the most extensive fatty streak formation, but was significantly reduced in C3H/HeJ mice. Iron loading of coculture cells resulted in elevations of cellular free iron and enhanced LDL-induced monocyte transmigration. Pretreatment of cells with apoferritin completely abolished iron-induced LDL modification. Addition of LDL to cocultures resulted in elevations in lipid peroxidation products, intracellular free iron, apoferritin mRNA expression, and apoferritin synthesis, suggesting a possible relationship between the oxidative modification of LDL and iron metabolism.

A new antiinflammatory compound, leumedin, inhibits modification of low density lipoprotein and the resulting monocyte transmigration into the subendothelial space of cocultures of human aortic wall cells.

Addition of leumedin, N-[9H-(2,7-dimethylfluorenyl-9-methoxy) carbon]-L-leucine at 30-60 microM together with LDL almost completely prevented the induction of monocyte chemotactic protein mRNA, reduced monocyte chemotactic protein 1 levels by 84%, and inhibited monocyte migration into the subendothelial space of cocultures of human aortic wall cells by < or = 98%. LDL incubated with leumedin formed a stable complex that remained intact even after refloating in an ultracentrifuge. Leumedin at 50 microM did not change conjugated diene formation during coculture modification of LDL or Cu++ catalyzed oxidation of LDL. Unlike LDL from control rabbits, LDL isolated from rabbits that were injected with 20 mg/kg leumedin was remarkably resistant to modification by the coculture and did not induce monocyte migration to a significant degree. Moreover, HDL isolated from rabbits injected with leumedin was far more effective in protecting against LDL modification by the artery wall cocultures than HDL from control rabbits. We conclude that leumedins can associate with lipoproteins in vivo, rendering LDL resistant to biological modification and markedly amplifying the protective capacity of HDL against in vitro LDL oxidation by artery wall cells.

Lipopolysaccharide-induced inhibition of scavenger receptor expression in human monocyte-macrophages is mediated through tumor necrosis factor-alpha.

The exposure of mononuclear cells to LPS results in a variety of cellular alterations including changes in the expression of various membrane receptors. In human monocyte-macrophages the development of the scavenger receptor, which mediates the uptake and internalization of chemically modified proteins, was suppressed by 100 ng/ml of LPS, concomitant with a reduction in receptor mRNA. Removal of LPS from the media resulted in a rapid increase in scavenger-receptor activity and mRNA that was further enhanced by macrophage-CSF and granulocyte/macrophage-CSF. However, neither macrophage-CSF nor granulocyte/macrophage-CSF could overcome the suppression of scavenger-receptor activity in the presence of LPS. The LPS-induced suppression of the scavenger receptor could be overcome by the co-addition of neutralizing antibody to TNF-alpha. TNF-alpha added to human monocyte-macrophages in the absence of LPS suppressed scavenger receptor activity to the same extent as did LPS. In contrast, the co-addition of LPS and neutralizing antibodies to either IFN-gamma or to IL-1 beta did not overcome the inhibitory effects of LPS on scavenger receptor activity. We conclude that the LPS-induced suppression of the scavenger receptor is mediated primarily through TNF-alpha.

Interaction of monocytes with cocultures of human aortic wall cells involves interleukins 1 and 6 with marked increases in connexin43 message.

Medium from cocultures of human aortic endothelial cells (HAEC) and smooth muscle cells (HASMC) taken from the same donor contained approximately two- to fourfold more macrophage colony-stimulating factor, granulocyte/macrophage colony-stimulating factor, and up to 5.1-fold more transforming growth factor beta than could be accounted for by the sum of the activities of media from equivalent numbers of HAEC and HASMC cultured separately. After pulse labeling, immunoprecipitated [35S]fibronectin and [14C]collagen were also found to be substantially increased in the coculture compared to the sum of HAEC and HASMC cultured separately. The cocultivation of HAEC and HASMC resulted in a 2.7-fold increase in connexin43 messenger RNA. When direct physical contact between HAEC and HASMC was prevented by a membrane that was permeable to medium, the levels of [35S]fibronectin and [14C]collagen in the coculture were significantly reduced. Monocytes cultured alone contained low levels of [35S]fibronectin and [14C]collagen but when added to the coculture there was up to a 22-fold increase in [35S]fibronectin and a 1.9-fold increase in [14C]collagen compared to the coculture alone. The increase in fibronectin was prevented in the presence of neutralizing antibody to interleukin 1 and antibody to interleukin 6 by 45% and 67%, respectively. Addition of monocytes to cocultures also induced the levels of mRNA for connexin43 by 2.8-fold. We conclude that the interaction of HAEC, HASMC, and monocytes in coculture can result in marked increases in the levels of several biologically important molecules and that increased gap junction formation between the cells and interleukins 1 and 6 may be partially responsible for these changes.

Processing of lipoproteins in human monocyte-macrophages.

Subcellular fractionation of human monocyte-macrophages (HMM) yielded a fraction rich in endosomes, lysosomes, and mitochondria. This pellet was further fractionated in a metrizamide gradient and the subcellular organelles were distributed among seven distinct bands. All of the bands contained lysosomal enzymes in similar amounts. However one band, poor in mitochondria, was markedly enriched in cathepsin D and cholesteryl ester hydrolase activities. A number of different ligands (low density lipoproteins (LDL), malondialdehyde-altered LDL, beta-migrating very low density lipoprotein, high density lipoprotein, reductively methylated LDL, mannose-bovine serum albumin, and transferrin) were presented to HMM at a concentration of 20 micrograms/ml at 4 degrees C. Three minutes after warming the cells at 37 degrees C all ligands except two were found predominantly in the cathepsin D- and cholesteryl ester hydrolase-rich fraction. Unlike the other ligands, LDL had distributed to other more dense fractions and reductively methylated LDL was found mainly in less dense fractions. At a lower concentration, 2 micrograms/ml, the distribution of LDL was identical to the other ligands. In vitro incubation of the fractions obtained from the gradient suggested that cathepsin D was largely responsible for the hydrolysis of the lipoproteins. We conclude that studies of LDL metabolism in HMM must take into account the different processing of this ligand at commonly used concentrations.

Modification of the Recalde method for the isolation of human monocytes.

A modification of the method for monocyte isolation reported by Recalde (1984. J. Immunol. Methods. 69: 71-77) is described. Application of the modified method to 36 consecutive healthy adult donors gave a monocyte purity of 73 +/- 8% monocytes with less than 1% polymorphonuclear leukocytes and a yield of 3.44 +/- 0.93 x 10(5) monocytes/ml blood. While the monocyte purity of the modified Recalde method was lower than that obtained by elutriation (method BB in Fogelman et al. 1981. J. Lipid Res. 22: 1131-1141) in 15 donors (71 +/- 10% vs. 83 +/- 6%) the monocyte yield and the viability of the cells before and after culture were similar in both methods. The modified Recalde method does not require the expensive or complicated equipment required for elutriation and permits the isolation of human monocytes for culture in autologous serum from multiple donors in a single day.

Lipoprotein receptors and endothelial cells.

The interaction of lipoproteins with endothelial receptors can result in alterations in macromolecular transport, in changes in monocyte adherence to the endothelium, and in the production of monocyte chemotactic factor by the endothelium. Monocyte migration in response to such factors can further alter lipoprotein transport into the subendothelial space, and all of these factors may play a role in the development of the atherosclerotic lesion.

Low density lipoproteins transfer bacterial lipopolysaccharides across endothelial monolayers in a biologically active form.

Rabbit aortic endothelial cells (RAECs) were grown on micropore filters in a device that allowed in situ determination of transendothelial electrical resistance (TEER). Incubation of confluent RAEC monolayers with 2 ng.ml-1 of bacterial LPS for 3 h did not change the protein content or the number of cells on the filters, but resulted in a marked decline in TEER (from 14.1 +/- 0.9 to 5.1 +/- 0.6 omega.cm2) and a significant increase in LDL transport across the monolayers (from 154 +/- 13 to 456 +/- 41 ng. h-1 per cm2). In contrast, exposure of RAEC monolayers for 3 d to as much as 5 micrograms.ml-1 of LPS complexed to LDL (LPS-LDL) did not alter the TEER or LDL transport. LPS-LDL was transported across the monolayers at the same rate as LDL. While microgram quantities of LPS complexed to LDL did not disrupt the integrity of the endothelial monolayer, incubation of RAECs with transported LPS-LDL at concentrations of 25-100 ng LPS.ml-1 resulted in a two- to ninefold increase in the secretion of monocyte chemotactic activity by these cells. Incubation of rabbit aortic smooth muscle cells with transported LPS-LDL at concentrations of 25-100 ng LPS.ml-1 resulted in a two- to threefold increase in the secretion of monocyte chemotactic activity. We propose that LDL protects endothelial cells from the acute toxicity of LPS but the resulting complexes are transported across the endothelium in a biologically active form that can initiate an inflammatory response.

Macrophage lipoprotein receptors.

Macrophages possess a number of surface receptors that are capable of mediating the internalization of lipoproteins. The low-density lipoprotein (LDL) receptor of human monocyte macrophages recognizes apolipoprotein B-100 and apolipoprotein E and is rapidly regulated in response to changes in intracellular cholesterol levels. In contrast, in J774 macrophages LDL receptor regulation is defective and LDL can cause massive cholesterol accumulation. The beta migrating very low density lipoprotein (beta-VLDL) receptor is poorly regulated by cellular cholesterol concentrations, readily recognizes apolipoprotein E, poorly recognizes apolipoprotein B-100, and is immunologically related to the LDL receptor. The scavenger receptor (acetyl-LDL receptor) appears to have a molecular weight of 250,000 and is not regulated by cellular cholesterol levels. This receptor recognizes LDL that has been chemically or biologically altered. LDL complexes can also enter macrophages and cause cholesterol accumulation. Examples of such complexes are LDL-dextran sulphate complexes, LDL-proteoglycan aggregates, LDL-mast cell granule complexes, LDL-heparin-fibronectin-denatured collagen complexes, and LDL-antibody complexes. The entry of lipoprotein into macrophages by a pathway that is poorly regulated or is not regulated by cellular cholesterol concentrations appears to be a prerequisite for the formation of arterial foam cells.

The role of lipoproteins and receptor-mediated endocytosis in the transport of bacterial lipopolysaccharide.

The addition of bacterial lipopolysaccharide (LPS) from Escherichia coli 0111:B4 to human monocyte-macrophages cultured in serum results in suppression of scavenger receptor activity. The present studies were performed to examine if the effect on scavenger receptor activity was mediated by LPS alone or by LPS in association with lipoproteins. Radioiodinated LPS (125I-LPS) was added to human plasma in vitro and to normal and hyperlipidemic rabbit plasma in vitro and in vivo to determine the distribution of 125I-LPS among the lipoprotein classes. It was found that all lipoprotein classes bound LPS in direct proportion to their plasma cholesterol concentration. LPS alone was compared to LPS bound to low density lipoprotein (LDL), high density lipoprotein, or reductively-methylated LDL for their abilities to suppress scavenger receptor activity in monocyte-macrophages in lipoprotein-free serum. Only LPS bound to LDL (LPS-LDL) demonstrated an effect similar to that observed when LPS was added to cells in serum. Either unlabeled LDL or unlabeled LPS-LDL complexes competed with the uptake of 125I-LPS-LDL complexes, which appeared to proceed by receptor-mediated endocytosis. In contrast to the uptake of 125I-LDL, the uptake of 125I-LPS-LDL by cultured monocyte-macrophages was not followed by its hydrolysis and the release of its radioactive degradation products into the medium. The association of LPS with lipoproteins was very stable and appeared to be mediated by a lipid-lipid interaction. We hypothesize that LPS bound to lipoproteins may be transported into the artery wall and may initiate the atherosclerotic reaction.