Minimally Modified LDL Binds to CD14, Induces Macrophage Spreading via TLR4/MD-2, and Inhibits Phagocytosis of Apoptotic Cells*

Yury I. MillerDagger §, Suganya Viriyakosol, Christoph J. BinderDagger , James R. Feramisco||, Theo N. Kirkland, and Joseph L. WitztumDagger

From the Dagger  Division of Endocrinology and Metabolism, Department of Medicine, the  Veterans Administration San Diego Healthcare System and Department of Pathology and Medicine, and the || Cancer Center, University of California, San Diego, La Jolla, California 92093

Received for publication, September 19, 2002, and in revised form, October 29, 2002

    ABSTRACT
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Minimally modified low density lipoprotein (mmLDL) is a pro-inflammatory and pro-atherogenic lipoprotein that, unlike profoundly oxidized LDL (OxLDL), is not recognized by scavenger receptors and thus does not have enhanced uptake by macrophages. However, here we demonstrate that mmLDL (as well as OxLDL) induces actin polymerization and spreading of macrophages, which results in such pro-atherogenic consequences as inhibition of phagocytosis of apoptotic cells but enhancement of OxLDL uptake. We also demonstrate for the first time that the lipopolysaccharide receptor, CD14, and toll-like receptor-4/MD-2 are involved in these mmLDL effects. Macrophages of the J774 cell line exhibited higher mmLDL binding and F-actin response than its CD14-deficient mutant, LR-9 cells. Similarly, Chinese hamster ovary cells transfected with human CD14 specifically bound mmLDL and responded with higher F-actin compared with control cells. Macrophages from C3H/HeJ mice, which have a point mutation in the Tlr4 gene, responded with lower F-actin to mmLDL and did not spread as well as macrophages from control animals. A significantly higher F-actin response was also observed in Chinese hamster ovary cells transfected with human toll-like receptor-4/MD-2 but not with TLR4 alone or TLR2. Thus, in addition to inhibition of phagocytosis, the recognition of mmLDL by macrophage lipopolysaccharide receptors results in convergence of cellular immune responses to products of microorganisms and to oxidation-specific self-antigens, which could both influence macrophage function and atherogenesis.

    INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Mild oxidation of low density lipoprotein (LDL)1 produces minimally modified LDL (mmLDL), which is still recognized by the LDL receptor and is not recognized by macrophage scavenger receptors (1). These properties differentiate mmLDL from profoundly oxidized LDL (OxLDL), which is not recognized by the LDL receptor but has enhanced uptake by macrophage scavenger receptors, such as CD36 and SR-BI (2). However, mmLDL possesses many pro-atherogenic properties; for example, it induces monocyte adhesion to endothelial cells, production of colony stimulating factors, monocyte chemotactic protein-1, and tissue factor (1, 3, 4). In addition, it activates multiple apoptotic signaling pathways in human coronary cells (5). Indeed, a large number of apoptotic cells accumulate in atherosclerotic lesions (6), despite the presence of many professional phagocytes-macrophages. Phagocytosis of dying cells is crucial for preventing the release of toxic cellular compounds and consequent inflammation. Inhibition of efficient phagocytosis leads to the accumulation of proinflammatory necrotic debris (such as oxidized lipids), plaque instability, and thrombogenesis (7).

We have recently proposed that in mouse peritoneal macrophages phagocytosing apoptotic cells, actin polymerization (which is necessary for filopodia formation) depends on the activity of 12/15-lipoxygenase (12/15-LO), which translocates from the cytosol to the cell membrane at sites where apoptotic cells are bound (8). mmLDL carry hydroperoxides and hydroxides of fatty acids and phospholipids, similar to those produced by 12/15-LO catalysis (9, 10), and thereby could also influence actin polymerization in a manner similar to that of 12/15-LO. Because mmLDL does not bind to scavenger receptors, there are presumably other receptors that mediate mmLDL effects.

In two recent reports, high expression of toll-like receptors (TLRs), and especially of TLR4, has been demonstrated in macrophages in human and mouse atherosclerotic lesions (11, 12). Moreover, the Bruneck (Italy) Study has identified that the D299G TLR4 polymorphism that attenuates receptor signaling is associated with a decreased risk of atherosclerosis (13). TLR4 is a receptor playing an essential role in innate immunity (14). It is also remarkable that the C(-260)T polymorphism in the CD14 promoter, resulting in a significantly higher density of CD14 on monocytes, has been identified as a risk factor for myocardial infarction (15). In another study, a higher density of CD14 was found on monocytes from diabetic patients with cardiovascular disease as compared with nondiabetic patients (16). CD14 does not have its own signaling domain. Thus, upon LPS binding (assisted by LPS-binding protein (LBP)), CD14 associates with TLR4 and triggers TLR4-dependent signaling pathways (17). The TLR4 accessory protein, MD-2, is also capable of binding LPS, and the TLR4/MD-2 complex can serve as an independent LPS receptor (18).

In this paper we report that mmLDL specifically binds to macrophage CD14 and via TLR4/MD-2 activates signaling pathways leading to actin polymerization and cell spreading. Moreover, in macrophages exposed to mmLDL, there is a subsequent inhibition of phagocytosis of apoptotic cells (an actin-dependent process) but, in contrast, an enhancement of OxLDL uptake (an actin-independent pathway). These findings provide a new link between oxidized lipoproteins and macrophage receptors of innate immunity recognizing components of microbial infection and also could explain in part aspects of macrophage dysfunction that lead to the enhanced accumulation of apoptotic cells in atherosclerotic lesions.

    EXPERIMENTAL PROCEDURES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Cells and Reagents-- Resident peritoneal macrophages were harvested without stimulation from 8-10-week-old female mice of C57BL/6 (Charles River, Wilmington, MA), C3H/HeJ, or C3H/HeOuJ (Jackson Lab, Bar Harbor, ME) strains. Murine macrophage-like cell lines J774A.1 and RAW 264.7 were from ATCC (Manassas, VA). An LPS-resistant mutant of J774, LR-9, developed by M. Nishijima, T. Nishihara, and colleagues (19) was obtained from R. Schlegel (Penn State University). These cells have been shown to have a decreased surface expression of CD14 (19). Studies in our lab with an anti-CD14 monoclonal antibody (R & D Systems, Minneapolis, MN) confirmed significantly decreased CD14 expression on the surface of LR-9 cells as compared with J774 cells. Murine fibroblast cell lines stably overexpressing human 15-LO or beta -galactosidase (LacZ) were developed in our lab (9).

Chinese hamster ovary (CHO-K1) cells stably transfected with CD14 or a CD14 mutant were generated and functionally characterized as previously described (20). The CD14-D5 mutant had a deletion of DPRQY peptide in the CD14 sequence. Surface expression of the protein was determined by FACS analysis with anti-CD14 monoclonal antibody 63D3 (ATCC) (20). Monoclonal antibody 28C5 against human CD14 was a gift from P. Tobias (Scripps Research Institute). TLR2, TLR4, and TLR4/MD-2 stably transfected cells were generated from Pdisplay, a CHO cell line bearing inducible membrane CD25 under the transcriptional control of a human E-selectin promoter containing NF-kappa B-binding sites, as described before (18). TLR2 and TLR4 were engineered to express with an N-terminal hemagglutinin tag, and MD-2 was engineered to express with a C-terminal His tag; their surface expression was analyzed by FACS (18). Escherichia coli LPS was purchased from List Biological Lab (San Jose, CA). Biotinylated Re595 LPS (biotin-LPS) was a gift from R. Dziarski (Indiana University). Murine M-CSF was from R & D Systems.

LDL Isolation and Modification-- LDL was isolated from the plasma of normolipidemic donors by sequential ultracentrifugation. Heavily oxidized LDL (1 mg/ml) was produced by an 18-h incubation in EDTA-free phosphate-buffered saline with 10 µM CuSO4. Minimally modified LDL was prepared by incubating LDL (50 µg/ml) in serum-free Dulbecco's modified Eagle's medium for 18 h with 15-LO fibroblasts (9). As previously characterized in detail, the mmLDL generated by this procedure has normal binding to the native LDL receptor and does not bind to scavenger receptors of macrophages, although it is enriched by oxidized phospholipids detected by monoclonal antibody EO6 and by lipid hydroperoxides as measured by high pressure liquid chromatography (10, 21). Control LDL was prepared for these experiments by an identical incubation with LacZ fibroblasts (9) The extent of LDL oxidation was assessed by measuring its binding to monoclonal autoantibodies EO6, specific to oxidized phospholipid or oxidized phospholipid-protein adducts, and EO14, specific to MDA-lysine epitopes, respectively (22).

All of the native and modified LDL preparations used in these experiments were tested for bacterial LPS with a Limulus Amoebocyte Lysate kit (BioWhittaker, Walkersville, MD). The samples with LPS higher than 2 ng/mg protein were discarded. Because LDL was used at a final concentration of 50 µg/ml or less in all experiments, the LPS contamination in experimental samples was always below 0.1 ng/ml.

F-actin Assay-- Peritoneal macrophages and J774 and RAW cells were plated overnight in culture medium before the experiment. Because quiescent adherent CHO cells have many stress fibers with a high F-actin content, it is difficult to see F-actin changes under these conditions. Thus, we plated the CHO cells for 2 h, the time sufficient for cell attachment but not for spreading, and then stimulated them with LDL samples. The LR-9 cells, which are larger than parental J774, were also plated for 2 h before experiment (in this case J774, analyzed in parallel with LR-9, were also plated for only 2 h). Relative content of F-actin in cells was assessed by flow cytometry (8). The increase in F-actin levels is referred to throughout the text as F-actin response. The data were collected as histograms of distribution of cell fluorescence intensities, and the geometrical means of each histogram were used for further analysis. The experiments were performed in duplicate and repeated at least three times.

CD14 Expression-- J774 and LR-9 cells were thoroughly washed, preincubated for 2 h in serum-free medium, and then gently scraped into bovine serum albumin/phosphate-buffered saline. The cells were stained with biotinylated anti-mouse CD14 (R & D Systems) followed by streptavidin-Alexa Fluor 488 (Molecular Probes, Eugene, OR) and analyzed by flow cytometry.

LDL Binding to Cells-- LDL was biotinylated with EZ-Link Sulfo-NHS-Biotin (Pierce) according to the manufacturer's protocol. Prior to biotinylation, an aliquot of the LDL destined to be the native LDL control, was protected from oxidation by adding 10 µM butylated hydroxytoluene. Nonprotected biotin-LDL was further incubated with the 15-LO cells, as described above, to produce biotin-mmLDL. The cells were scrapped into suspension, incubated with 2-50 µg/ml of biotin-LDL on ice for 2 h, washed, and then stained with streptavidin-Alexa Fluor 488. The cells were also stained with propidium iodide to differentiate live cells and analyzed on a FACSCalibur or a FACScan. Comparison of this FACS assay with a conventional 125I-LDL binding assay (22) showed comparable results (data not shown).

Phagocytosis Assay-- A macrophage phagocytic population was measured according to the procedure published in our lab (23), with some modifications. In short, thymocytes were labeled with 1 µM of 5-chloromethylfluorescein diacetate (CellTracker green CMFDA from Molecular Probes) followed by a treatment with 5 µM dexamethasone to induce apoptosis (23) and incubated with macrophages. The macrophages that have engulfed CellTracker-labeled apoptotic cells differed on FACS from nonphagocytic macrophages by higher fluorescence intensity and were designated as a phagocytic population (23).

OxLDL Uptake by Macrophages-- The cell association and degradation of 125I-OxLDL by resident peritoneal macrophages was determined by the method of Goldstein as performed in our lab (22). To remove aggregates from the 125I-OxLDL prep, it was centrifuged at 14,000 × g for 15 min and filtered through a 0.2-µm filter immediately before addition to the cells. The amount of specific OxLDL associated or degraded per milligram of cell protein was calculated after subtracting nonspecific values determined in presence of excess unlabeled OxLDL.

    RESULTS
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

OxLDL and mmLDL Induce Actin Polymerization and Spreading of Macrophages-- Mouse peritoneal macrophages and the cells of the murine macrophage-like cell line J774 exhibited a rapid and robust increase in polymerized F-actin in response to OxLDL or mmLDL (further referred to as F-actin response) but not to native LDL, as was evident from the increased fluorescence of the cells stained with FITC-phalloidin (Fig. 1, a and b). Note that compared with the content of F-actin in unstimulated peritoneal macrophages or J774 cells, OxLDL, and to an even greater extent mmLDL, increased absolute content of F-actin by up to 70%, whereas native LDL had no such activity. This degree of stimulation was even greater than that observed with M-CSF (5 nM), which also increases F-actin formation (24). This gain in the total cellular F-actin content is substantial, because actin filaments build up the cytoskeleton and are the most abundant protein polymer in the cell. Indeed, as might be anticipated, the increased F-actin response was associated with a major change in cell shape, as noted by the marked degree of cell spreading (Fig. 1, c-e). The F-actin response to mmLDL was blunted in another macrophage line, RAW cells; in fact, the latter did not respond to OxLDL (Fig. 1b).


View larger version (43K):
[in this window]
[in a new window]
 
Fig. 1.   Macrophage F-actin and spreading in response to mmLDL. Adherent cells were incubated with 50 µg/ml of indicated LDL species, 5 nM M-CSF, or 10 ng/ml LPS for 1 h, stained for F-actin and analyzed by flow cytometry. a, representative FACS histograms showing populations of J774 cells incubated with native LDL (gray shading) or mmLDL (bold line). b, geometrical means of the FACS histograms were normalized to medium samples for each cell type. Polymyxin B (PmB) was added at 0.5 µg/ml. *, p < 0.01 versus media. c-e, peritoneal macrophages were incubated with the same concentrations of nLDL for 1 h (c), OxLDL for 1 h (d), and OxLDL for 4 h (e), then fixed, and stained with FITC-phalloidin, and the images were acquired. Scale bar, 15 µm.

Concentrations of OxLDL or mmLDL between 10 and 50 µg/ml appear to saturate the ability of OxLDL to stimulate F-actin formation (not shown). OxLDL and mmLDL caused about a 20% increase in F-actin as early as in 10 min (not shown), and as much as 50-70% F-actin was attained at 1 h (Fig. 1b).

These results were not due to LPS contamination. As noted under "Experimental Procedures," the maximum concentration of LPS in any preparations used was below 0.1 ng/ml. Furthermore, polymyxin B did not inhibit the mmLDL effect (Fig. 1b). When LPS was directly added at 10 ng/ml, which is at least 100-fold higher than the trace amounts of LPS present in any of the LDL samples used, F-actin formation was increased by only 20% in J744, and in this case, polymyxin B abolished this response (Fig. 1b). Clearly, the mmLDL effect was not produced by LPS.

Because mmLDL was produced by incubating LDL with 15-LO-overexpressing fibroblasts (10, 21, 25), we purified it by centrifugation, 0.2-µm filtration, and concentration on a filter with a 100-kDa cut-off. The eluent fraction of molecular mass below 100 kDa as well as LDL incubated with control LacZ fibroblasts (9) had a significantly lower activity. mmLDL produced by a short period of incubation with copper (1 h), replacing 15-LO cells, was also active in eliciting the F-actin response (data not shown).

Role of CD14 in F-actin Response-- Because the response to mmLDL was so pronounced and mmLDL could not bind to scavenger receptors, we focused further studies on the mechanism by which mmLDL induced F-actin formation. We hypothesized that the differences in amplitudes of the F-actin increase between J774 and RAW cells in response to mmLDL and OxLDL (Fig. 1b) might be due to a differential expression of particular receptors that bound the modified LDL and signaled actin polymerization. We screened several receptors, including CD14, which is known to participate in phagocytosis (26), which binds ceramide (27), and which might be involved in LPS-induced macrophage spreading (28). The J774 cells showed a much stronger binding of anti-mouse CD14 and anti-mouse TLR4/MD-2 antibodies than the RAW cells (data not shown). Further, a mutant of J774, LR-9 cells, selected for their resistance to LPS because of the decreased surface expression of CD14 (19) (a finding confirmed in our lab), responded to mmLDL with a significantly smaller increase in the F-actin levels compared with the parental J774 cells (Fig. 2a). The difference in F-actin response by J774 and LR-9 cells correlated with nearly 6-fold increased specific binding of mmLDL to J774 but not to LR-9 cells (Fig. 2b). (Note that the binding of biotinylated mmLDL to J774 cells was abolished by a 25-fold excess of unlabeled mmLDL). Again, note that polymyxin B had no effect on the binding of mmLDL to J774 cells (Fig. 2b).


View larger version (24K):
[in this window]
[in a new window]
 
Fig. 2.   F-actin response and mmLDL binding to J774 and LR-9 cells. a, flow cytometry analysis of F-actin levels in J774 and CD14-deficient LR-9 cells exposed for 1 h to 50 µg/ml of nLDL or mmLDL. The data are shown as percentages of F-actin change compared with the cells in media without any LDL addition. b, flow cytometry analysis of biotinylated LDL binding to J774 and LR-9. The geometrical means of FACS histograms (minus "no LDL" controls) are shown. Some samples contained a 25-fold excess of nonlabeled LDL or 0.5 µg/ml polymyxin B (PmB).

The experiments above provided substantial support for specific binding of mmLDL to CD14. To provide direct evidence, we next tested whether overexpression of CD14 would lead to an increased F-actin response to mmLDL. A CHO cell line stably transfected with human CD14 (CD14 cells) responded to mmLDL with significantly higher levels of F-actin than the wild type CHO-K1 cells (Fig. 3a). Although changes in F-actin levels in CD14 cells were not as pronounced as in macrophages, the 25% increase is still a substantial change in F-actin formation for these large ovary cells, which are not professional phagocytes and do not rapidly form filopodia, as do macrophages. Accordingly, binding of biotinylated mmLDL to the CD14 cells was significantly higher than to the wild type CHO cells (Fig. 3, b and c). To demonstrate specific binding to the CD14 cells, biotinylated mmLDL binding curves were determined in the absence or presence of unlabeled mmLDL. As shown in Fig. 3b, there was specific and saturable binding of mmLDL with a calculated Kd of 15 ± 3 µM. The specific binding of mmLDL to CD14 cells was 10-fold higher than to control CHO cells (Fig. 3c). The mmLDL binding correlated with increased binding of LPS to the CD14 cells (Fig. 3e). This increased LPS binding was specific for CD14, because monoclonal antibody 28C5, directed against the LPS binding site of CD14, reduced LPS binding to levels seen in nontransfected CHO cells (Fig. 3e and Ref 20). However, antibody 28C5 did not block either the F-actin response to mmLDL or mmLDL binding to the CD14 cells (Fig. 3, a and d). These data suggest that although mmLDL has specific binding to CD14, it binds at a site distinct from the LPS binding site (at least that domain to which 28C5 binds). Further support for this comes from studies of binding to the CD14 mutant CD14-D5, in which the peptide DPRQY is deleted, rendering CD14 unable to support binding of LPS (20). Transfection of CHO cells with CD14-D5 resulted in CHO cells in which no specific LPS binding was observed (Fig. 3e). However, binding of mmLDL to CD14-D5 was inhibited by only 25% compared with CD14 cells (Fig. 3d), and the ability to stimulate F-actin formation was not significantly decreased (Fig. 3a). These data confirm that mmLDL binds to CD14 but to a region distinct from the LPS-binding site.


View larger version (31K):
[in this window]
[in a new window]
 
Fig. 3.   F-actin response and mmLDL binding to CD14-transfected CHO cells. a, flow cytometry analysis of F-actin levels in wild type or CD14-transfected CHO cell lines exposed for 1 h to 50 µg/ml of nLDL or mmLDL. The data are shown as percentages of F-actin change compared with the cells in medium without any LDL addition. b and c, binding curve of biotinylated LDL binding to CD14 (b) and nontransfected CHO (c) cells. Geometrical means of FACS histograms (minus "no LDL" controls) are shown. Total binding designates biotin-LDL binding in the absence of a competitor. Nonspecific binding is binding in the presence of a 20-fold excess of nonlabeled LDL. Specific binding is a result of subtraction of nonspecific from total binding. d, specific binding of nLDL and mmLDL to CD14, CD14-D5, and CHO cells in the presence and absence of anti-CD14 antibody 28C5 (6 µg/106 cells), added 1 h prior to addition of LDL. *, p < 0.05 versus CHO-K1. e, biotinylated LPS (50 ng/106 cells) binding to CD14, CD14-D5, and CHO cells in the presence and absence of anti-CD14 antibody 28C5 (6 µg/106 cells).

Role of TLR4/MD-2 in F-actin Response-- Because most CD14-dependent LPS effects are believed to require the TLR4 signaling pathways (29), we studied whether the F-actin response to mmLDL depended on TLR4. Peritoneal macrophages from C3H/HeJ mice have normal CD14 expression, but they are resistant to LPS because of a mutation in the TLR4 gene, Tlr4Lps-d (30). In contrast, macrophages from C3H/HeOuJ mice carry a wild type Tlr4Lps-n and are sensitive to endotoxin. The F-actin response to mmLDL was significantly blunted in Tlr4Lps-d macrophages compared with Tlr4Lps-n macrophages (Fig. 4a). Moreover, in agreement with our experiments with C57BL/6 macrophages (Fig. 1, c-e), Tlr4Lps-n macrophages from C3H/HeOuJ mice readily spread following a 1-h exposure to mmLDL (Fig. 4, b and c). In contrast, the morphology of the Tlr4Lps-d macrophages from C3H/HeJ mice exposed for the same time to mmLDL did not significantly differ from those exposed to nLDL or medium alone (Fig. 4, d and e).


View larger version (20K):
[in this window]
[in a new window]
 
Fig. 4.   Role of TLR4/MD-2 in F-actin response to mmLDL. Peritoneal macrophages were harvested from C3H/HeOuJ (TlrLps-n) and C3H/HeJ (TlrLps-d) mice, plated, and then incubated for 1 h with 50 µg/ml of nLDL or mmLDL. a, F-actin levels were measured by flow cytometry. b-e, phase contrast microphotographs of the macrophages at the end of incubation: TlrLps-n + nLDL (b), TlrLps-n + mmLDL (c), TlrLps-d + nLDL (d), and TlrLps-d + mmLDL (e). f, flow cytometry analysis of F-actin levels in indicated CHO cell lines. TLR2, TLR4, and TLR4/MD-2 stably transfected cells were generated from the CHO cell line Pdisplay (18).

To support the hypothesis of TLR4 involvement in mediating the mmLDL induction of actin polymerization, we next compared the F-actin response of CHO cells stably transfected with either TLR2, TLR4, or TLR4/MD-2. Of these, only the TLR4/MD-2 cells responded to mmLDL with significantly increased levels of F-actin as compared with the control CHO cells (Fig. 4f). Thus, mmLDL can directly activate TLR4/MD-2 in these cells when these molecules are overexpressed.

Next, we tested whether inhibition of several key components of known TLR signaling pathways (14) will influence the F-actin response. The phosphoinositide 3-kinase (PI3K)-specific inhibitors wortmannin and LY294002 (31) efficiently inhibited actin polymerization in J774 cells (Fig. 5). The effect of wortmannin was dose-dependent in the range of 10-100 nM (not shown). In contrast to the PI3K inhibitors, the inhibitor of NF-kappa B translocation, SN50, and the p38 MAPK inhibitor, SB202190, did not have any significant effect (Fig. 5). Wortmannin is among the most selective kinase inhibitors available (32), and it inhibited mmLDL-induced F-actin formation (Fig. 5) at its reported IC50 concentrations. SB202190 was inactive even at a concentration that was 100-fold higher than its reported IC50 (32). SN50 was used at the concentration shown to be effective for NF-kappa B inhibition in many macrophage cell types (33, 34). We also tested the effects of PI3K inhibitors in the CHO cells transfected with CD14 and TLR4/MD-2. Wortmannin and LY294002 inhibited spreading of CD14 and TLR4/MD-2 CHO cells, but the absolute reduction in F-actin formation was not statistically significant (data not shown), presumably because of the low amplitudes of F-actin response in transfected CHO cells (as discussed above).


View larger version (12K):
[in this window]
[in a new window]
 
Fig. 5.   Effect of specific signaling inhibitors on F-actin response. J774 cells were preincubated for 30 min with either medium alone (control), 50 nM wortmannin (wortm), 10 µM LY294002, 50 µg/ml SN50, or 5 µM SB202190 and then incubated for 1 h with medium or 50 µg/ml mmLDL. F-actin levels were measured by FACS. The numbers were normalized to control (no inhibitor treatment) samples. *, p < 0.001 versus control. FITC, fluorescein isothiocyanate

Effect of mmLDL-induced Actin Polymerization on Phagocytosis of Apoptotic Cells and Uptake of OxLDL by Macrophages-- We originally began these studies because of our interest in the mechanism by which OxLDL affected macrophage uptake of apoptotic cells (8, 23). Therefore we examined what impact the mmLDL-induced F-actin formation in macrophages would have on the subsequent ability of the macrophages to phagocytose apoptotic cells. The use of mmLDL in this case helps avoid any possible competition between the OxLDL added during the preincubation and the apoptotic cells used to test phagocytosis, because both bind to scavenger receptors (23). Preincubation with mmLDL resulted in a markedly decreased ability of macrophages to subsequently phagocytose apoptotic cells. In four separate experiments, mmLDL decreased the uptake of apoptotic cells by 35-60% (Fig. 6a). To explain this, we hypothesized that the mmLDL-induced actin polymerization caused extensive cell spreading, which in turn led to a relative inability to mobilize actin polymerization sufficient for phagocytosis. To test this hypothesis, we pretreated the cells with M-CSF, which also caused a comparable increase in F-actin (Fig. 1b) but in an mmLDL-independent manner. M-CSF pretreatment also resulted in subsequent inhibition of phagocytosis of apoptotic cells similar to that seen with mmLDL (Fig. 6a).


View larger version (66K):
[in this window]
[in a new window]
 
Fig. 6.   Effect of preincubation with mmLDL on macrophage phagocytic function and OxLDL uptake. Resident macrophages were preincubated for 4 h with medium only, 50 µg/ml nLDL or mmLDL, or 5 nM M-CSF and then exposed to labeled apoptotic thymocytes (a) or 125I-OxLDL (b). Macrophage phagocytic populations (a) and specific 125I-OxLDL association and degradation by macrophages (b) were measured. *, p < 0.01; **, p < 0.05 versus medium. c-e, phase contrast photographs of macrophage cultures incubated for 4 h with nLDL (c), mmLDL (d), or M-CSF (e).

Uptake of monomeric OxLDL requires the presence of functional scavenger receptors but does not involve an actin-dependent phagocytic mechanism (35). In contrast, uptake of apoptotic cells occurs in part via scavenger receptors and is actin-dependent. To demonstrate that nonphagocytic scavenger receptor-mediated uptake was not influenced, we repeated the mmLDL preincubation with macrophages, but instead of examining the uptake of apoptotic cells, we examined the uptake of aggregate-free 125I-OxLDL. In contrast to inhibition of phagocytosis, pretreatment of macrophages with mmLDL or M-CSF significantly increased subsequent specific uptake and degradation of 125I-OxLDL (Fig. 6b). This result proves that the mmLDL mediated inhibition of phagocytosis of apoptotic cells is not due to down-regulation of scavenger receptors and, to the contrary, is consistent with the known ability of OxLDL (and M-CSF) to up-regulate CD36 (36, 37). Morphologically, macrophages treated with both mmLDL and M-CSF, but not with native LDL, were found to be extensively spread (Fig. 6, c-e), consistent with the hypothesis that the shape and/or actin-related changes in macrophages were the cause of inhibited phagocytosis of apoptotic cells.

    DISCUSSION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

This study reports several novel observations. We demonstrate that OxLDL and mmLDL stimulate fast and robust actin polymerization and macrophage spreading (Fig. 1). Because mmLDL does not bind to scavenger receptors (1), it was unclear how the mmLDL affected F-actin formation. Here we show that mmLDL specifically binds to CD14 (Figs. 2 and 3) and initiates F-actin formation through a signaling pathway that depends on TLR4/MD-2 (Fig. 4) and PI3K (Fig. 5). The mmLDL-induced macrophage spreading culminates on the one hand in a reduced ability of macrophages to efficiently phagocytose apoptotic cells but, on the other hand, in the enhanced uptake of OxLDL, two consequences that would be predicted to enhance atherogenesis (Fig. 6).

Our data demonstrate that mmLDL and OxLDL produced comparable increases in macrophage F-actin (Fig. 1b). In accordance with our data, actin polymerization was recently reported to accompany OxLDL-mediated monocyte adhesion and transendothelial migration (38). We used a well characterized technique for producing mmLDL by incubating LDL with 15-LO-overexpressing cells, which seed fatty acid hydroperoxides into LDL (9, 10, 21). As previously demonstrated, this mmLDL still binds to the LDL receptor and, although enriched in hydroperoxides and oxidized phospholipids, nevertheless does not bind to scavenger receptors that bind OxLDL (9). Thus, we asked whether the mmLDL effect on actin cytoskeleton occurs through binding to cell surface receptors other than OxLDL scavenger receptors. Native LDL did not stimulate F-actin formation (Figs. 1-3). Preliminary studies with J774 and LR-9 cells suggested the involvement of CD14 (Fig. 2). Indeed, we found that specific binding of biotin-mmLDL was 10-fold higher to CD14 cells than to control CHO cells, and this corresponded to increased F-actin formation by the CD14 cells in response to mmLDL (Fig. 3). Because we performed our experiments in medium without serum, this implies that LBP is not obligatory for the mmLDL binding to CD14. In the case of LPS, LBP basically catalyzes transfer of monomers from LPS aggregates and presents the LPS monomers to CD14. In the case of mmLDL, oxidation changes phospholipid presentation on the LDL particle (39), and this alone, without LBP catalysis, could be a sufficient transition causing mmLDL to be recognized by CD14. In favor of this hypothesis are data showing binding of isolated phospholipids to both membrane and soluble CD14 (40, 41). On the other hand, oxidation of the fatty acid chains of phospholipids in the LDL particle changes presentation of the phospholipid head groups, making them immunogenic (39). Similar changes occurring in mmLDL may lead to CD14 recognition.

The anti-CD14 antibody 28C5 or the DPRQY deletion in the CD14 sequence blocked the binding of LPS to CD14, but it did not block mmLDL binding (Fig. 3, d and e), suggesting that CD14-binding sites for LPS and mmLDL differ. Interestingly, binding of phosphatidylinositides to CD14 has been shown to be blocked by anti-CD14 antibodies and to compete with LPS binding to CD14 (40), suggesting similar CD14 binding sites for phosphatidylinositides and LPS. Future studies using various CD14 mutants and monoclonal antibodies to different CD14 epitopes will help map the mmLDL-binding site on CD14. In a paper published after completion of this work, Bochkov et al. (42) demonstrated that a model oxidized phospholipid, 1-palmitoyl-2-(5-oxovaleroyl) phosphatidylcholine, competed with the ability of LPS to bind to CD14 and to activate the NF-kappa B pathway. Whether this is occurring via a similar mechanism as described here is uncertain, because our studies suggest that the mmLDL site on CD14 is different from that of LPS.

The TLR4/MD-2-CHO cells also responded to mmLDL with significantly elevated F-actin (Fig. 4). Although TLR4 alone does not bind LPS, MD-2 has been shown to bind LPS without assistance from either LBP or CD14, and the LPS-MD-2 binding constant is comparable with that of LPS-CD14 (18). It could be also that mmLDL binds to MD-2, although it was not a focus of this study. In several studies, TLR4 activation by various agonists has been reported, with indirect data suggesting potential binding of these agonists (43, 44). Studies with isolated MD-2 and CD14 molecules will help elucidate their binding to mmLDL. Our other experiments also support the role of TLR4 in mediating the F-actin response to mmLDL. C3H/HeJ macrophages with the mutated Tlr4 gene lost their ability to respond to mmLDL by either enhanced actin polymerization or spreading (Fig. 4). The C3H/HeJ mice are known not only for their resistance to LPS but also for resistance to cholesterol-induced atherosclerosis (45, 46). Because we show that the mmLDL-induced macrophage spreading inhibits phagocytosis of apoptotic cells (Fig. 6) and could be therefore considered pro-atherogenic, it is tempting to propose that among other factors (46), the TLR4Lps-d mutation and mmLDL resistance make the C3H/HeJ mice less susceptible to atherosclerosis. However, C57BL/10ScN mice, in which the Tlr4 gene is deleted (Tlr4lps-del), cross-bred with ApoE-/- mice, did not show any changes in atherosclerosis progression compared with control ApoE-/- mice when these mice were fed a Western diet (47). Such a diet, which induces extreme hypercholesterolemia and excessive atherogenic pressure, may have obscured an ability of TLR4 to modulate atherogenesis. Similarly, Rag (1 or 2)-deficient mice bred into the ApoE-/- background do not show any attenuation in lesion formation when such mice are fed a Western diet, but, if chow-fed, reductions of atherosclerosis of 40-80% are found (48, 49). In addition, importance of this pathway in humans was strongly suggested by the recent Brunek (Italy) study, which identified that the D299G TLR4 polymorphism, attenuating receptor signaling, was associated with a decreased risk of atherosclerosis (13).

TLR signaling in most cases is associated with pro-inflammatory gene expression (14). What is known about the role of TLR in the cell cytoskeleton regulation? TLR2 activates NF-kappa B via PI3K and a Rho family small GTPase Rac1 (50). In addition to transcriptional regulation, PI3K and Rac are well known regulators of actin polymerization and macrophage phagocytosis (51, 52). Our data show that PI3K might also be involved in mmLDL-induced actin polymerization (Fig. 5), although it is not yet clear whether the PI3K activation depends on TLR4 signaling in this case. Another study reports that a high dose of LPS stimulates spreading of J774 cells via activation of beta 2-integrin (28). The authors of this elegant work show that the LPS-induced spreading (occurring in 10 min) is linearly dependent on the signaling chain, which is common for TLRs: from MyD88 to IRAK to p38 MAPK. Subsequently, p38 controls spreading independently of its role in transcription but rather through activation of another small GTPase of the Ras family, Rap1, which in turn activates beta 2-integrin (28). We also found that an inhibitor of NF-kappa B nuclear translocation did not affect the F-actin response to mmLDL (Fig. 5). However, unlike in the studies of Schmidt et al. (28), p38 inhibition did not significantly influence the mmLDL effect (Fig. 5).

Our finding that the biological effect of mmLDL is mediated by the LPS receptors agrees with a general concept of these receptors' function. Both CD14 and TLR family receptors are "pattern recognition receptors" activated by a variety of chemically different ligands. Besides LPS, these receptors recognize conserved products of microbial metabolism, such as peptidoglycan, lipoteichoic acids, and other components of bacterial cell walls (53, 54), as well as nonmicrobial substances. For instance, such different stimuli as the plant product Taxol (43), heat shock protein 60 (55), and the respiratory syncyntial virus coat protein F (44) have all been reported as TLR4 agonists. CD14 binds ceramide, which can be found in modified LDL (27, 56). The results of our study suggest that structural changes in the LDL particle upon its mild oxidation make it also recognizable by immune competent cells and lead to TLR4 activation. Further studies will be needed to define the ligand(s) on mmLDL mediating the binding and biological effects.

These data and other results from our group support the hypothesis of convergence of immune responses to oxidation-specific self-antigens and to microbial infection. This hypothesis was first substantiated by data from our group showing that a natural autoantibody to mmLDL, EO6, is genetically identical to the classic germline anti-phosphorylcholine antibody T15, which binds avidly to Streptococcus pneumoniae. More specifically, EO6/T15 recognize bacterial cell wall polysaccharides that are rich in phosphorylcholine (57), as well as phosphorylcholine-containing oxidized phospholipids present in oxidized LDL (22) and cells undergoing apoptosis (23). The present study demonstrates correspondence not only in humoral (EO6-T15 antibody) but also in cellular (CD14 and TLR4 receptors) immune responses to mmLDL and to microbial infection.

Finally, this study shows that the CD14/TLR4/MD-2-mediated macrophage spreading in response to mmLDL inhibits phagocytosis of apoptotic cells (Fig. 6a). This result cannot be explained by the competition between mmLDL and apoptotic cells for the same macrophage receptors because mmLDL does not bind to scavenger receptors (1, 21). Moreover, mmLDL pretreatment actually increases the subsequent uptake of OxLDL by macrophages (Fig. 6b), showing that scavenger receptor expression was not suppressed and in fact was increased. In a similar manner, macrophage spreading caused by M-CSF also resulted in decreased phagocytosis, but increased OxLDL uptake (Fig. 6, a and b). The exact mechanisms by which macrophages exposed to mmLDL (or M-CSF) become less efficient phagocytes but have a higher rate of OxLDL uptake are not yet known. Presumably, the higher rate of OxLDL uptake induced by mmLDL exposure is mediated by activation of peroxisome proliferator-activated receptor gamma  and enhanced CD36 expression (36). Because mmLDL, as well as OxLDL, is prevalent in atherosclerotic lesions, our observations could have important implications in understanding macrophage function in atherosclerotic lesions. Considering phagocytosis of apoptotic cells as an anti-inflammatory process and extensive uptake of OxLDL as a pro-atherogenic event, one would predict a net pro-atherogenic role of mmLDL as a consequence of these effects on macrophages. Thus, this study defines novel pathways by which both mmLDL and microbial infection could influence the progress of atherogenesis.

    ACKNOWLEDGEMENTS

We thank E. Raz for critical reading the manuscript, D. J. Young in the UCSD Cancer Center Flow Cytometry Shared Resource for flow cytometer expertise, and the UCSD Cancer Center Digital Imaging Shared Resource.

    FOOTNOTES

* This work was supported by American Heart Association Grant AHA WSA 0160111Y (to Y. I. M.), National Institutes of Health Grants HL56989 (to La Jolla SCOR in Molecular Medicine and Atherosclerosis), 1 P01 HL66941 (to J. R. F.), and PO1GM37696 (to S. V. and T. N. K.), and funds from the Medical Research Service of the Department of Veterans Affairs (to S. V. and T. N. K.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§ To whom correspondence should be addressed: University of California, San Diego, 1080 Basic Science Bldg., 9500 Gilman Dr., La Jolla, CA 92093-0682. Tel.: 858-822-5771; Fax: 858-534-2005; E-mail: yumiller@ucsd.edu.

Published, JBC Papers in Press, November 6, 2002, DOI 10.1074/jbc.M209634200

    ABBREVIATIONS

The abbreviations used are: LDL, low density lipoprotein; OxLDL, oxidized LDL; mmLDL, minimally modified LDL; nLDL, native LDL; LO, lipoxygenase; TLR, toll-like receptor; LPS, lipopolysaccharide; LBP, LPS-binding protein; M-CSF, macrophage colony-stimulating factor; PI3K, phosphoinositide 3-kinase; CHO, Chinese hamster ovary; FACS, fluorescence-activated cell sorter.

    REFERENCES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

1. Berliner, J. A., Territo, M. C., Sevanian, A., Ramin, S., Kim, J. A., Bamshad, B., Esterson, M., and Fogelman, A. M. (1990) J. Clin. Invest. 85, 1260-1266[Medline] [Order article via Infotrieve]
2. Glass, C. K., and Witztum, J. L. (2001) Cell 104, 503-516[Medline] [Order article via Infotrieve]
3. Berliner, J. A., Subbanagounder, G., Leitinger, N., Watson, A. D., and Vora, D. (2001) Trends Cardiovasc. Med. 11, 142-147[CrossRef][Medline] [Order article via Infotrieve]
4. Liao, F., Berliner, J. A., Mehrabian, M., Navab, M., Demer, L. L., Lusis, A. J., and Fogelman, A. M. (1991) J. Clin. Invest. 87, 2253-2257[Medline] [Order article via Infotrieve]
5. Napoli, C., Quehenberger, O., De, Nigris, F., Abete, P., Glass, C. K., and Palinski, W. (2000) FASEB J. 14, 1996-2007[Abstract/Free Full Text]
6. Martinet, W., and Kockx, M. M. (2001) Curr. Opin. Lipidol. 12, 535-541[CrossRef][Medline] [Order article via Infotrieve]
7. Libby, P. (2001) Am. J. Cardiol. 88, J3-J6[Medline] [Order article via Infotrieve]
8. Miller, Y. I., Chang, M. K., Funk, C. D., Feramisco, J. R., and Witztum, J. L. (2001) J. Biol. Chem. 276, 19431-19439[Abstract/Free Full Text]
9. Benz, D. J., Mol, M., Ezaki, M., Mori-Ito, N., Zelaan, I., Miyanohara, A., Friedmann, T., Parthasarathy, S., Steinberg, D., and Witztum, J. L. (1995) J. Biol. Chem. 270, 5191-5197[Abstract/Free Full Text]
10. Ezaki, M., Witztum, J. L., and Steinberg, D. (1995) J. Lipid Res. 36, 1996-2004[Abstract]
11. Xu, X. H., Shah, P. K., Faure, E., Equils, O., Thomas, L., Fishbein, M. C., Luthringer, D., Xu, X. P., Rajavashisth, T. B., Yano, J., Kaul, S., and Arditi, M. (2001) Circulation 104, 3103-3108[Abstract/Free Full Text]
12. Edfeldt, K., Swedenborg, J., Hansson, G. K., and Yan, Z. Q. (2002) Circulation 105, 1158-1161[Abstract/Free Full Text]
13. Kiechl, S., Lorenz, E., Reindl, M., Wiedermann, C. J., Oberhollenzer, F., Bonora, E., Willeit, J., and Schwartz, D. A. (2002) N. Engl. J. Med. 347, 185-192[Abstract/Free Full Text]
14. Medzhitov, R. (2001) Nat. Rev. Immunol. 1, 135-145[CrossRef][Medline] [Order article via Infotrieve]
15. Hubacek, J. A., Rothe, G., Pit'ha, J., Skodova, Z., Stanek, V., Poledne, R., and Schmitz, G. (1999) Circulation 99, 3218-3220[Abstract/Free Full Text]
16. Patino, R., Ibarra, J., Rodriguez, A., Yague, M. R., Pintor, E., Fernandez-Cruz, A., and Figueredo, A. (2000) Am. J. Cardiol. 85, 1288-1291[CrossRef][Medline] [Order article via Infotrieve]
17. Jiang, Q., Akashi, S., Miyake, K., and Petty, H. R. (2000) J. Immunol. 165, 3541-3544[Abstract/Free Full Text]
18. Viriyakosol, S., Tobias, P. S., Kitchens, R. L., and Kirkland, T. N. (2001) J. Biol. Chem. 276, 38044-38051[Abstract/Free Full Text]
19. Nishijima, M., Hara-Kuge, S., Takasuka, N., Akagawa, K., Setouchi, M., Matsuura, K., Yamamoto, S., and Akamatsu, Y. (1994) J Biochem. (Tokyo) 116, 1082-1087[Abstract]
20. Viriyakosol, S., and Kirkland, T. N. (1995) J. Biol. Chem. 270, 361-368[Abstract/Free Full Text]
21. Sigari, F., Lee, C., Witztum, J. L., and Reaven, P. D. (1997) Arterioscler. Thromb. Vasc. Biol. 17, 3639-3645[Abstract/Free Full Text]
22. Hörkkö, S., Bird, D. A., Miller, E., Itabe, H., Leitinger, N., Subbanagounder, G., Berliner, J. A., Friedman, P., Dennis, E. A., Curtiss, L. K., Palinski, W., and Witztum, J. L. (1999) J. Clin. Invest. 103, 117-128[Abstract/Free Full Text]
23. Chang, M. K., Bergmark, C., Laurila, A., Hörkkö, S., Han, K. H., Friedman, P., Dennis, E. A., and Witztum, J. L. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 6353-6358[Abstract/Free Full Text]
24. Nakamura, I., Lipfert, L., Rodan, G. A., and Le, T. D. (2001) J. Cell Biol. 152, 361-373[Abstract/Free Full Text]
25. Lee, C., Sigari, F., Segrado, T., Hörkkö, S., Hama, S., Subbaiah, P. V., Miwa, M., Navab, M., Witztum, J. L., and Reaven, P. D. (1999) Arterioscler. Thromb. Vasc. Biol. 19, 1437-1446[Abstract/Free Full Text]
26. Schlegel, R. A., Krahling, S., Callahan, M. K., and Williamson, P. (1999) Cell Death Differ. 6, 583-592[CrossRef][Medline] [Order article via Infotrieve]
27. Pfeiffer, A., Bottcher, A., Orso, E., Kapinsky, M., Nagy, P., Bodnar, A., Spreitzer, I., Liebisch, G., Drobnik, W., Gempel, K., Horn, M., Holmer, S., Hartung, T., Multhoff, G., Schutz, G., Schindler, H., Ulmer, A. J., Heine, H., Stelter, F., Schutt, C., Rothe, G., Szollosi, J., Damjanovich, S., and Schmitz, G. (2001) Eur. J. Immunol. 31, 3153-3164[CrossRef][Medline] [Order article via Infotrieve]
28. Schmidt, A., Caron, E., and Hall, A. (2001) Mol. Cell. Biol. 21, 438-448[Abstract/Free Full Text]
29. Guha, M., and Mackman, N. (2001) Cell Signal. 13, 85-94[CrossRef][Medline] [Order article via Infotrieve]
30. Poltorak, A., He, X., Smirnova, I., Liu, M. Y., Huffel, C. V., Du, X., Birdwell, D., Alejos, E., Silva, M., Galanos, C., Freudenberg, M., Ricciardi-Castagnoli, P., Layton, B., and Beutler, B. (1998) Science 282, 2085-2088[Abstract/Free Full Text]
31. Vanhaesebroeck, B., Leevers, S. J., Ahmadi, K., Timms, J., Katso, R., Driscoll, P. C., Woscholski, R., Parker, P. J., and Waterfield, M. D. (2001) Annu. Rev. Biochem. 70, 535-602[CrossRef][Medline] [Order article via Infotrieve]
32. Davies, S. P., Reddy, H., Caivano, M., and Cohen, P. (2000) Biochem. J. 351, 95-105[CrossRef][Medline] [Order article via Infotrieve]
33. Abate, A., and Schroder, H. (1998) Life Sci. 62, 1081-1088[CrossRef][Medline] [Order article via Infotrieve]
34. Ropert, C., Almeida, I. C., Closel, M., Travassos, L. R., Ferguson, M. A., Cohen, P., and Gazzinelli, R. T. (2001) J. Immunol. 166, 3423-3431[Abstract/Free Full Text]
35. Lougheed, M., and Steinbrecher, U. P. (1996) J. Biol. Chem. 271, 11798-11805[Abstract/Free Full Text]
36. Tontonoz, P., Nagy, L., Alvarez, J. G., Thomazy, V. A., and Evans, R. M. (1998) Cell 93, 241-252[Medline] [Order article via Infotrieve]
37. Huh, H. Y., Pearce, S. F., Yesner, L. M., Schindler, J. L., and Silverstein, R. L. (1996) Blood 87, 2020-2028[Abstract/Free Full Text]
38. Mine, S., Tabata, T., Wada, Y., Fujisaki, T., Iida, T., Noguchi, N., Niki, E., Kodama, T., and Tanaka, Y. (2002) Atherosclerosis 160, 281-288[CrossRef][Medline] [Order article via Infotrieve]
39. Friedman, P., Hörkkö, S., Steinberg, D., Witztum, J. L., and Dennis, E. A. (2002) J. Biol. Chem. 277, 7010-7020[Abstract/Free Full Text]
40. Wang, P. Y., Kitchens, R. L., and Munford, R. S. (1998) J. Biol. Chem. 273, 24309-24313[Abstract/Free Full Text]
41. Sugiyama, T., and Wright, S. D. (2001) J. Immunol. 166, 826-831[Abstract/Free Full Text]
42. Bochkov, V. N., Kadl, A., Huber, J., Gruber, F., Binder, B. R., and Leitinger, N. (2002) Nature 419, 77-81[CrossRef][Medline] [Order article via Infotrieve]
43. Kawasaki, K., Akashi, S., Shimazu, R., Yoshida, T., Miyake, K., and Nishijima, M. (2000) J. Biol. Chem. 275, 2251-2254[Abstract/Free Full Text]
44. Kurt-Jones, E. A., Popova, L., Kwinn, L., Haynes, L. M., Jones, L. P., Tripp, R. A., Walsh, E. E., Freeman, M. W., Golenbock, D. T., Anderson, L. J., and Finberg, R. W. (2000) Nat. Immunol. 1, 398-401[CrossRef][Medline] [Order article via Infotrieve]
45. Liao, F., Andalibi, A., deBeer, F. C., Fogelman, A. M., and Lusis, A. J. (1993) J. Clin. Invest. 91, 2572-2579[Medline] [Order article via Infotrieve]
46. Shi, W., Haberland, M. E., Jien, M. L., Shih, D. M., and Lusis, A. J. (2000) Circulation 102, 75-81[Abstract/Free Full Text]
47. Wright, S. D., Burton, C., Hernandez, M., Hassing, H., Montenegro, J., Mundt, S., Patel, S., Card, D. J., Hermanowski-Vosatka, A., Bergstrom, J. D., Sparrow, C. P., Detmers, P. A., and Chao, Y. S. (2000) J. Exp. Med. 191, 1437-1442[Abstract/Free Full Text]
48. Dansky, H. M., Charlton, S. A., Harper, M. M., and Smith, J. D. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 4642-4646[Abstract/Free Full Text]
49. Reardon, C. A., Blachowicz, L., White, T., Cabana, V., Wang, Y., Lukens, J., Bluestone, J., and Getz, G. S. (2001) Arterioscler. Thromb. Vasc. Biol. 21, 1011-1016[Abstract/Free Full Text]
50. Arbibe, L., Mira, J. P., Teusch, N., Kline, L., Guha, M., Mackman, N., Godowski, P. J., Ulevitch, R. J., and Knaus, U. G. (2000) Nat. Immunol. 1, 533-540[CrossRef][Medline] [Order article via Infotrieve]
51. Leverrier, Y., and Ridley, A. J. (2001) Curr. Biol. 11, 195-199[CrossRef][Medline] [Order article via Infotrieve]
52. Cantley, L. C. (2002) Science 296, 1655-1657[Abstract/Free Full Text]
53. Schnare, M., Barton, G. M., Holt, A. C., Takeda, K., Akira, S., and Medzhitov, R. (2001) Nat. Immunol. 2, 947-950[CrossRef][Medline] [Order article via Infotrieve]
54. Kaisho, T., and Akira, S. (2002) Biochim. Biophys. Acta 1589, 1-13[Medline] [Order article via Infotrieve]
55. Ohashi, K., Burkart, V., Flohe, S., and Kolb, H. (2000) J. Immunol. 164, 558-561[Abstract/Free Full Text]
56. Schissel, S. L., Tweedie-Hardman, J., Rapp, J. H., Graham, G., Williams, K. J., and Tabas, I. (1996) J. Clin. Invest. 98, 1455-1464[Abstract/Free Full Text]
57. Shaw, P. X., Hörkkö, S., Chang, M. K., Curtiss, L. K., Palinski, W., Silverman, G. J., and Witztum, J. L. (2000) J. Clin. Invest. 105, 1731-1740[Abstract/Free Full Text]


Copyright © 2003 by The American Society for Biochemistry and Molecular Biology, Inc.