Oxidized Phospholipids in Minimally Modified Low Density Lipoprotein Induce Apoptotic Signaling via Activation of Acid Sphingomyelinase in Arterial Smooth Muscle Cells*

Oxidized phospholipids, including 1-palmitoyl-2-glutaroyl-sn-glycero-3-phosphocholine (PGPC) and 1-palmitoyl-2-(5-oxovaleroyl)-sn-glycero-3-phosphocholine (POVPC), typically present in minimally modified low density lipoprotein, have been found in atherosclerotic lesions. These compounds are gaining increasing importance as inducers of different cellular responses (inflammation, proliferation, or cell death). It was the aim of this study to understand their impact on intracellular signal transduction pathways that are responsible for these biological effects. We found that in arterial smooth muscle cells, PGPC and POVPC activated sphingomyelinases, in particular the acid isoform, which is known to participate in the very early phase of apoptotic stress responses. In addition, mitogen-activated protein kinases, which are involved in induction of stress response and apoptosis were phosphorylated (activated). Finally, activation of caspase 3 was observed, showing that stimulation of smooth muscle cells with POVPC and PGPC is associated with apoptosis. Stimulation of all these enzymes by the oxidized phospholipids almost perfectly matched their activation by minimally modified LDL. Consequently, these phospholipids seem to be responsible for the effect of this particle on cell signaling. Survival and proliferation pathways including NF-κB or AKT kinase were not induced by POVPC and PGPC. Experiments with a specific inhibitor of acid sphingomyelinase named NB6 showed that this enzyme plays a central role in mediating the apoptotic effects of the oxidized lipids. Thus, we conclude that modified phospholipids induce signaling cascades via activation of acid sphingomyelinase finally leading to apoptosis of smooth muscle cells, which is a detrimental process in the development of atherosclerosis.

Oxidation of LDL 1 is a key process in the development of atherosclerosis (1,2). Oxidized LDL exerts multiple effects on the cells of the arterial wall, including inflammation, proliferation, and programmed cell death (3). Especially apoptosis of smooth muscle cells (SMCs) is a prominent feature of the late phase of atherosclerosis (4) and can be triggered by oxidative stress (5). Interaction of SMC with oxidized LDL and the biological consequences thereof are an important issue and much is known about the effects of highly oxidized LDL as a whole particle (6,7). LDL can be oxidized to different degrees. Whereas highly oxidized LDL is modified at both its lipid and protein part, minimally modified LDL (mmLDL) is mainly oxidized at its lipid part, and thus exhibits unique properties, which are different from those of oxidized LDL (8). These features might be due to its content in biologically active lipids mainly originating from oxidation of arachidonic acid of the surface phospholipids (9). Among a large number of degradation products oxidized phospholipids including 1-palmitoyl-2-(5-oxovaleroyl)-sn-glycero-3-phosphocholine (POVPC) and 1-palmitoyl-2-glutaroyl-sn-glycero-3-phosphocholine (PGPC) are present in mmLDL. These compounds are characterized by short polar fatty acyl chains in position 2 and a single hydrophobic fatty acid in position 1 of glycerol (10). As a consequence they are highly exchangeable between cells, tissues, and lipoproteins. Oxidized phospholipids have been found in atheromas and are responsible for multiple phenomena in endothelial cells leading to the progression of atherosclerosis including monocyte adhesion, integrin activation, and tissue factor expression (11)(12)(13). The effects of oxidized phospholipids present in mmLDL on signaling in vascular smooth muscle cells have not been investigated to date.
Sphingomyelinases are very early upstream elements of stress-induced cell signaling (14). They generate ceramide from sphingomyelin affecting several downstream targets, including the family of mitogen-activated protein kinases (MAPK), which are activated by phosphorylation (15). Especially, c-Jun Nterminal kinase (JNK) and p38 MAPK have been reported to be involved in apoptotic stress responses (16,17). At the very end of the apoptotic signaling cascade, caspase 3 is activated, this being responsible for the execution of apoptosis (18).
Activation of extracellular regulated protein kinase (ERK), which also belongs to the MAPK family, leads to proliferation (19). Proliferation and survival pathways may also be activated independently by AKT-kinase/protein kinase B or the transcription factor NF-B (20,21).
It was the aim of this study to identify the effects of POVPC and PGPC on signal transduction pathways involved in proliferative and apoptotic stress responses of vascular SMC and find out how these phenomena induced by chemically defined compounds are related to the parent (minimally oxidized) lipoprotein particle. We found that POVPC and, to a lesser extent, PGPC activated apoptotic signaling pathways finally leading to caspase 3 activation. They did not stimulate survival and proliferation pathways. More specifically acid sphingomyelinase represents a central element in the POVPC-and PGPC-induced apoptotic signaling cascade. Collectively, our studies suggest that PGPC and especially POVPC are the components that are responsible for the stress effects of mmLDL on SMCs as they mimicked mmLDL in activating the same signaling components.

EXPERIMENTAL PROCEDURES
Materials-Chemicals were purchased from Merck (Darmstadt, Germany). Plastic containers for cell culture were obtained from Sarstedt (Nü rmbrecht, Germany). Media and supplements for cell culture were purchased from Paa Laboratories (Linz, Austria). Micro well plates for fluorescence and luminescence assays were from Falcon (Innsbruck, Austria). TNF-␣ was from Calbiochem (La Jolla, CA). NB6 was provided by Hans-Peter Deigner (Clinic for Anaesthesiology and Intensive Care Therapy, FSU Jena, Germany). Chemicals for gel electrophoreses were purchased from Bio-Rad Laboratories (Hercules, CA). Nitrocellulose membranes were from Schleicher and Schuell (Dassel, Germany). Primary polyclonal antibodies against phosphorylated forms of JNK/SAPK, p38 MAPK, ERK1/2, AKT/PKB, and caspase 3 were purchased from New England Biolabs (Beverly, MA). Alkaline phosphatase-conjugated secondary antibodies and BCIP/NBT Liquid Substrate system for alkaline phosphatase were from Sigma.
Synthesis of Oxidized Phospholipids-POVPC and PGPC were synthesized according to an extensively modified version of the procedure by Watson et al. (9) and Subbanagounder et al. (22). 3 Isolation and Modification of Low Density Lipoprotein-Human LDL (d ϭ 1.019Ϫ1.063) was isolated from pooled fresh plasma using a Beckman NVT-rotor (23). LDL was minimally modified as described by Watson et al. (9). According to the latter procedure mildly oxidized LDL was obtained. It was characterized by a moderate content of lipid peroxidation products (30 -60 nmol of peroxide equivalents/mg of apolipoprotein B (24)), and a slightly enhanced relative electrophoretic mobility (REM) (0.34) as compared with native LDL (0.30) in 0.8% agarose-gels and a virtually intact protein moiety. The protein-content of isolated and modified LDL was determined by the method of Lowry (25).
Determination of Sphingomyelinase Activity-Rat aortic smooth muscle cells (A7r5) were grown in six-well plates until 80% of confluence was reached. Growth was arrested with DMEM ϩ 0.1% FCS for 24 h, and subsequently cells were incubated with either POVPC, PGPC (10 M each), native LDL, or minimally modified LDL (60 g apoB/ml) for 15-360 min. For studying the effects of NB6, cells were pretreated with 10 M of this substance for 30 min. Excess NB6 was removed and cells were then exposed to the oxidized lipids or mmLDL as indicated above.
After treatment of the smooth muscle cells with the respective compounds, cells were washed with cold phosphate-buffered saline, scraped, harvested, and lyzed by incubation with either neutral lysis buffer for determination of neutral SMase activity or acid lysis buffer for determination of acid SMase activity as described previously (26). Cell lysates were subjected to protein quantification according to the method by Bradford (27), and aliquots containing 20 g of protein were analyzed for acid and neutral sphingomyelinase activity using fluorescent sphingomyelin as described (26).
Quantification of Caspase 3 Activity-A7r5 cells were seeded into 96 micro well plates employing a cell number of 15,000 cells/well and grown for 24 h in 100 l DMEM ϩ 10% FCS. Subsequently, the medium was removed and cells were incubated for 2-8 h with a 100 l suspension of either POVPC or PGPC corresponding to final lipid concentrations of 1, 10, or 50 M. For inhibition experiments, cells were preincubated with 10 M NB6 for 30 min prior to addition of oxidized lipid. A 100 l solution of (Z-DEVD) 2 -rhodamine 110 in cell lysis/activity buffer was added to each well followed by shaking the plate for 30 min on an Eppendorf thermomixer. After incubation at room temperature for an additional 4 h, fluorescence in the wells was measured (excitation wavelength 485 nm, emission wavelength 530 nm) using a PerkinElmer Life Sciences LS-50 B fluorometer with plate reader equipment.
Measurement of NF-B Activity-For transient transfection, A7r5 cells were seeded into 6-well plates (9.6 cm 2 ), grown to 50 -70% confluence, and incubated with the transfection mixture containing 0.3 g of pNF-B Luc plasmid, 0.5 g of pSV␤-galactosidase, and and 4.8 l of Tfx TM -50 transfection reagent per ml of serum-free DMEM (supplemented with HEPES and L-glutamine) for 1 h. Afterward, cells were supplied with fresh medium (DMEM ϩ 10% FCS) and incubated for an additional 48 h followed by stimulation with POVPC, PGPC (10 M each), or mmLDL (60 g/ml apoB) for 60 -240 min.
For termination, stimulated cells were washed twice with ice-cold phosphate-buffered saline, lyzed, and scraped. Aliquots of the cell lysates were used to measure luciferase activity. 100 l of luciferase substrate reagent were mixed gently with 40 l of the cell lysate, and after a 20 s delay, light emission was measured in a BIOLUMAT LB9500 luminometer from Berthold (Vienna, Austria) using an integration time of 10 s. The background signal obtained by measurement of untransfected cells was substracted. For the ␤-galactosidase assay 150 l of cell lysate were mixed with 150 l of ortho-nitrophenylgalactopyranoside-reagent by brief vortexing. After 3 h of incubation at 37°C, the reaction was terminated by adding 500 l of 1 M Tris. The mixture was vortexed again, and the absorbance was read at 420 nm. Luciferase activity was standardized with reference to ␤-galactosidase activity by dividing luciferase by ␤-galactosidase activity, and this is reported as a percentage activity compared with control cells.

Activation of Sphingomyelinases by Oxidized Phospholipids-Sphingomyelinases are central elements in stress-induced signal transduction leading to apoptosis or proliferation (29).
We investigated the effects of the pure oxidized phospholipids PGPC and POVPC, which are typically present in minimally oxidized LDL, on acid and neutral sphingomyelinases in smooth muscle cells using fluorescent sphingomyelin as a substrate (26). Minimally modified LDL and native LDL were applied as control stimuli. In addition, experiments were performed with NB6, which has been shown to specifically inhibit acid sphingomyelinase on the transcriptional level (30). As shown in Fig. 1A, 10 M POVPC stimulated activity of acid sphingomyelinase to the same extent as mmLDL (60 g/ml) after 30 min (180% of control). In contrast, PGPC and native LDL showed only minor effects. The maximum effect of POVPC was reached after 30 min. However, significant activation was already detectable after 5 min (data not shown). After 360 min of incubation with POVPC, aSMase activity returned to control levels. Incubation of smooth muscle cells with 10 M NB6 for 30 min prior to stimulation with either POVPC, PGPC, or mmLDL led to a decrease of acid sphingomyelinase activity to control levels.
The neutral isoform of this enzyme was activated by POVPC and PGPC to the same extent (Fig. 1B). Nonetheless, in contrast to the acid isoform, neutral SMase remained activated after 360 min. MmLDL caused substantially higher stimulation as compared with the oxidized phospholipids. In this context, it has to be emphasized that native LDL also strongly activated neutral SMase. Thus, the lipoprotein effect seems not to be specific for the oxidized particle. Treatment of SMC with the acid SMase inhibitor NB6 prior to stimulation with either the oxidized phospholipids or mmLDL did not affect neutral sphingomyelinase activation at all.
Influence of POVPC and PGPC on JNK and p38 MAPK-JNK and p38 MAPK belong to the family of MAPKs and mediate apoptotic stress response (31). They are supposed to be downstream targets of ceramide and, as a consequence, of sphingomyelinase activity. Thus, we investigated the role of JNK and p38 MAPK in cell signaling triggered by oxidized phospholipids and the potential involvement of acid sphingomyelinase as an upstream element. For this purpose, SMC were exposed to 10 M POVPC or PGPC or 60 g/ml mmLDL for 5-30 min with or without pretreatment with NB6 as an inhibitor of aSMase. Fig. 2, A and B show data from the Western blot analysis of cell extracts for phosphorylated (activated) tion. Addition of NB6 to the cells prior to stimulation abolished POVPC-, PGPC-, and mmLDL-induced phosphorylation of JNK and p38 MAPK, indicating that acid sphingomyelinase is essential for SMC signaling initiated by mmLDL and its oxidized phospholipid constituents. Activation of Caspase 3 by Oxidized Phospholids-To provide more direct evidence that the incubation of smooth muscle cells with POVPC or PGPC does, in fact, lead to apoptosis we measured caspase 3 activity after stimulation of the cells with various concentrations (1-50 M) of POVPC and PGPC for different lengths of time. According to Fig. 3, POVPC induced caspase 3 activation in a time-and concentration-dependent manner (Fig. 3A). Maximum caspase 3 activity was reached after incubation with 50 M POVPC for 6 h. In contrast, caspase 3 was only slightly activated by PGPC. Inhibition of acid sphingomyelinase by NB6 led to a substantial decrease of POVPC-induced caspase 3 activity, providing additional evidence that acid sphingomyelinase seems to be a central element in mediating the apoptotic effect of POVPC. Data obtained from analysis of caspase 3 activity were confirmed by Western blot analysis (Fig. 4), showing activated caspase 3 in lysates of cells, which have been exposed to the oxidized phospholipids.
Lack of Activation of Proliferative and/or Salvage Pathways-Further experiments were performed to find out whether or not POVPC and PGPC affect survival and proliferation pathways in addition to apoptotic signaling cascades. In this context, activity of NF-B was measured using a luciferase reporter gene assay. This transcription factor might play a role in atherogenesis because it mediates survival, proliferation and inflammation (21).
To make sure that the luciferase expression system worked properly, activation of NF-B by the cytokine TNF-␣ was examined, this being a potent activator of NF-B in many cell types including vascular smooth muscle cells (32). Fig. 5 shows that TNF-␣ induced NF-B activity in smooth muscle cells in a time-dependent manner.
Stimulation of cells with mmLDL (60 g apoB/ml) for 60 or 240 min did not lead to any NF-B-induced luciferase expression. Under the same conditions, NF-B levels were neither affected by POVPC nor PGPC (10 M each), coinciding with results obtained by stimulation of the cells with mmLDL. Consequently, we may conclude that the NF-B pathway is not involved in POVPC-and PGPC-induced signaling.
In addition to NF-B, there are two other components important for survival and proliferation signaling pathways. AKT-kinase/PKB and ERK are protein kinases that participate in cytokine-and growth factor-induced signal transduction leading to proliferation of most cells (19,20). Exposure of cells to 10 M of the oxidized lipids or mmLDL (60 g apoB/ml) neither stimulated ERK nor AKT/PKB. Western blot analysis (Fig. 6) of stimulated cell extracts for the phosphorylated (activated) proteins did not show any differences compared with the controls. These results again confirm the assumption that mmLDL and its biologically active components POVPC and PGPC do not activate signaling pathways responsible for proliferation and/or survival of vascular smooth muscle cells.

DISCUSSION
This work reports on the role of POVPC and PGPC in mmLDL-induced signaling of vascular SMC. We demonstrate that both phospholipids under consideration evoked apoptotic cell death by time-and concentration-dependent induction of caspase 3 activity, albeit to a different extent. POVPC was the more potent apoptotic lipid. Both POVPC and PGPC activated the sphingomyelinase pathway, particularly acid sphingomyelinase, to which an important role in stress-induced apoptosis has been assigned (33). We found a maximum of SMase activity in SMCs after 30 min of incubation with POVPC and PGPC. A significant increase of aSMase activity (150% of control) was already found after 5-10 min (data not shown). POVPC and PGPC also activated p38 MAPK and JNK, both playing an eminent part in apoptotic signaling. Phosphorylation of JNK and p38 MAPK due to cell stimulation with POVPC and PGPC was again very fast (within 5-30 min), POVPC being more potent than PGPC.
POVPC-and PGPC-induced activation of acid SMase could be abrogated by a specific inhibitor of acid sphingomyelinase (NB6). NB6 also prevented activation of the downstream targets p38 MAPK and JNK. Finally, the execution of apoptosis by caspase 3 could be inhibited by NB6 as well. These results indicate that acid sphingomyelinase is a central mediator in oxidized phospholipid-triggered apoptotic signaling.
Survival pathways, based on activation of ERK, NF-B and AKT-kinase/PKB, were neither activated by POVPC nor PGPC. Obviously, the SMCs are bound to death if these compounds are present in sufficient amounts, e.g. in an atheroma.
Remarkably, POVPC and, to a minor extent, PGPC closely mimic the biological activity of mmLDL. They activate the same signaling pathways leading to apoptosis. POVPC might be a component in the oxidized particle, which determines its properties to a major extent.
Activation of aSMase and its downstream elements occurs extremely fast (within a few minutes). This might be due to the rapid transfer of the respective lipids into the plasma membrane of the cell. The oxidized phospholipids have only one long hydrophobic side chain in position 1 of glycerol. The fatty acyl chain at position 2 is short and contains either a polar (C ϭ O) or a charged polar (COO Ϫ ) group. Hence, this chain becomes a part of a tremendously large head group. Consequently, these lipids may be considered highly surface active substances, similar to lysophophatidylcholine, which is also known for its potency to activate signaling and apoptosis (34,35) Activation of sphingomyelinase and its downstream components after 10 min is too fast to be explained in terms of an increased protein expression level. It might rather be due to a direct interaction of the lipids with the plasma membrane or the enzyme itself. On the one hand, accumulation of the oxidized lipids could cause unspecific alterations of membrane composition and, as a consequence, lead to activation of SMase. On the other hand, a direct contact of the lipids with SMase could initiate enzyme activation as well.
In this context it has to be emphasized that POVPC is a more efficient stimulant than PGPC. POVPC is able to modify proteins covalently due to its aldehydic group (36). Similar protein modification by aldehydic groups has already been reported for 4-hydroxynonenal in endothelial cells. This lipid was shown to react with the free aminogroups of the epidermal growth factor receptor leading to activation of the receptor and its associated signaling pathways (37,38).
In summary, our data provide strong evidence that POVPC is one of the main biologically active lipids rendering mmLDL apoptotic for vascular SMC. Acid sphingomyelinase has been identified as a key mediator of the involved signaling pathways finally leading to apoptosis of SMC, which is a hallmark in the development of atherosclerosis. FIG. 6. Western blot analysis of phospho-ERK and phospho-AKT/PKB. Effects of POVPC and PGPC as compared with minimally modified LDL in SMCs. A7r5 cells were incubated with either POVPC, PGPC (10 M each), or mmLDL (60 g apoB/ml) for 30 or 60 min after treatment with or without 10 M NB6 for 30 min. Cells were harvested, lyzed, and Western blots of sample aliquots containing 20 g of protein were performed with anti phospho-ERK1/2 or anti phospho-AKT/PKB antibodies as described under "Experimental Procedures." The same results were obtained in 3 independent experiments.