Potential Role for Ceramide in Mitogen-activated Protein Kinase Activation and Proliferation of Vascular Smooth Muscle Cells Induced by Oxidized Low Density Lipoprotein*

Proliferation of vascular smooth muscle cells (SMC) is a hallmark in the pathogenesis of atherosclerotic lesions. Mildly oxidized low density lipoproteins (UV-oxLDL), which are mitogenic to cultured AG-08133A SMC, activate the sphingomyelin (SM)-ceramide pathway. We report here the following. (i) UV-oxLDL elicited a biphasic and sustained activation of MBP kinase activity, phosphorylation and nuclear translocation of p44/42 mitogen-activated protein kinase (MAPK), and [3H]thymidine incorporation, which were inhibited by PD-098059, a MAPK kinase inhibitor. (ii) The use of preconditioned media (from SMC pre-activated by UV-oxLDL) transferred to native SMC and blocking antibodies against growth factors suggest that UV-oxLDL-induced activation of MAPK and [3H]thymidine incorporation seem to be independent of any autocrine secretion of growth factors. (iii) UV-oxLDL-induced activation of a neutral sphingomyelinase, SM hydrolysis, ceramide production, and [3H]thymidine incorporation were inhibited by two serine-protease inhibitors (serpins), suggesting that a serpin-sensitive proteolytic pathway is involved in the activation of the SM-ceramide signaling pathway. (iv) UV-oxLDL-induced MAPK activation and [3H]thymidine incorporation were mimicked by ceramide generated in the plasma membrane by bacterial sphingomyelinase treatment or by addition of the permeant C2-ceramide. Serpins did not inhibit the MAPK activation and [3H]thymidine incorporation induced by C2-ceramide, indicating that activation of the MAPK and [3H]thymidine incorporation is subsequent to the stimulation of the SM-ceramide pathway. Taken together, these data suggest that mitogenic concentrations of UV-oxLDL are able to stimulate the SM-ceramide pathway through a protease-dependent mechanism and activate p44/42 MAPK, leading to proliferation of vascular SMC.

Atherosclerosis, and its complications, namely myocardial infarction, stroke, and peripheral vascular diseases, is one of the most prevalent cause of morbidity and mortality in Western countries. During atherogenesis, focal lesions spread out progressively and lead to the formation of fibro-atheroma plaques, in which smooth muscle cell (SMC) 1 proliferation plays a critical role (1,2). Among the risk factors identified, low density lipoprotein (LDL) cholesterol level is strongly predictive of coronary heart disease. LDL are believed to have an important role in atherogenesis (3), following oxidative modifications (4 -6), because oxidized LDL are present in atherosclerotic lesions (7) and possess a wide range of biological properties potentially occurring during atherogenesis in vivo (8). Oxidized LDL have recently been shown to be mitogenic to vascular SMC (9 -11). These studies suggest that oxidized LDL may be considered as an additional mitogenic factor, alongside the classical growth factors implicated in SMC proliferation during atherogenesis (6). To date, the mechanism of the oxidized LDL proliferative effect is poorly elucidated and may result from the triggering of a mitogenic intracellular signal either directly by oxidized LDL or indirectly through an autocrine effect involving growth factor secretion and/or growth factor receptor over-expression.
Recently, sphingolipids have emerged as key signaling molecules involved in the regulation of cell growth and differentiation (for reviews, see Refs. [12][13][14][15]. In particular, the sphingomyelin (SM; ceramide phosphocholine)-ceramide pathway appears as a prototypic sphingolipid signaling pathway implicated in the positive or negative regulation of cell growth. Activation of this pathway leads to SM hydrolysis and subsequent generation of ceramide, the backbone of all sphingolipids, which serves as an intracellular second messenger. To date, several agents have been described to stimulate the SMceramide pathway (reviewed in Refs. 12 and 14 -17), including cytokines such as TNF␣, interleukin-1␤, interferon ␥, nerve growth factor, anti-CD28, anti-Fas antibodies, anticancer drugs, and ionizing radiations (18 -21). Cell-permeant ceram-ides, or ceramide produced by treatment of intact cells with exogenous sphingomyelinase, can mimic the effects of various inducers of the SM-ceramide pathway. Thus, various cellular responses including cell proliferation (18), differentiation (22), or apoptosis (19,20,23) seem to be transduced by SM hydrolysis through ceramide generation. We have recently reported that the SM-ceramide pathway is involved in the mitogenic signaling triggered by mildly oxidized LDL (UV-oxLDL) in SMC (11).
Cell proliferation promoted by extracellular growth stimuli involves various intracellular signaling pathways leading to activation of gene transcription, DNA synthesis, and cell division (24,25). Protein phosphorylation, mediated by a complex regulatory network of protein kinases, plays an essential role in the signal transduction between cell surface and nucleus (26). Despite the broad diversity of mitogens and receptors, these signaling pathways often converge toward the mitogenactivated protein (MAP) kinases, a group of serine/threonine protein kinases also referred to as extracellular-regulated kinases (ERKs; p44/ERK1 and p42/ERK2, p44/42 MAPK) in mammalian cells (26). MAPKs are rapidly activated in response to stimulation of receptors for growth factors, hormones, or cytokines, G protein-coupled receptors, or in response to stress (25,(27)(28)(29). They are activated by phosphorylation on both Tyr and Thr residues (30) by the dual specificity MAPK kinases, which are activated by Raf (29). The activation of MAPKs, associated with their nuclear translocation, is essential to trigger entry into the S phase of cell cycle (31,32).
Cell Culture-Bovine aortic SMC (AG-08133A, from the NIA Aging Cells Repository, Camden, NJ) were grown in RPMI 1640 medium supplemented with 10% FCS, penicillin (100 units/ml), and streptomycin (100 g/ml), at 37°C in humidified 5% CO 2 . After trypsinization, cells were grown in medium containing 10% FCS for 24 h. Then, the medium was removed, and cells were washed once with RPMI and grown for 48 h in RPMI medium containing 1% FCS prior to the addition of native or oxidized LDL. In some experiments, SV40-transfected SMC cells were used. For this purpose, SMC were electroporated (240 V, 960 F) with a plasmid (pAS) carrying the large T antigen (33); these transfected SMC exhibited a two to three times longer life span than the untransfected cells. For the experiments presented here, similar results were obtained with either SMC type.
Lipoprotein Isolation and Oxidation by UV-C-Human LDL (d 1.019 -1.063) were isolated from pooled fresh sera by sequential ultracentrifugation, dialyzed, sterilized by filtration, and stored at 4°C under nitrogen until use, as described previously (10). Mildly oxidized LDL (UV-oxLDL) were prepared by UV-C irradiation under the previously defined standard conditions (254 nm, 0.5 mW/cm 2 for 2 h) (34).
Labeling of LDL with DiIC18 and Determination of LDL Uptake by SMC-Purified LDL were fluorescently labeled with DiIC18 according to Via and Smith (35), re-isolated by ultracentrifugation, and dialyzed and sterilized again as indicated above. DiIC18-labeled LDL were added to the culture medium (100 g of apoB/ml) and incubated with cells for the indicated times. Then, cells were washed once with PBS containing 5 g/liter bovine serum albumin and twice with PBS only. Afterward, cells were homogenized by sonication in 1 ml of distilled water, and an aliquot was used to extract DiIC18 by the procedure of Folch et al. (36) and to read the cell-associated fluorescence extracted in the chloroformic phase (using a Jobin-Yvon JY-3C spectrofluorometer; excitation 545 nm, emission 568 nm).
[ 3 H]Thymidine Incorporation Assays-DNA synthesis was evaluated by [ 3 H]thymidine incorporation (at time 24 h, under standard conditions, or at the time indicated in the figures). Cells (2 ϫ 10 5 /dish) were labeled for 12 h with [ 3 H]thymidine (0.5 Ci/ml) before harvesting and then washed three times with PBS; after addition of 3% perchloric acid, the acid-precipitable material was dissolved overnight in 1 N NaOH, 1% SDS and counted by liquid scintillation (Packard Tricarb 4530 counter).
Quantitation of Sphingolipids and Sphingomyelinase Assays-SMC were metabolically labeled to equilibrium with [methyl-3 H]choline (0.5 Ci/ml) or [ 3 H]palmitic acid (0.5 Ci/ml) in RPMI medium containing 1% FCS. After 48 h of incubation, cells (about 300,000/assay) were washed once with PBS and chased for 2 h in fresh RPMI containing 1% FCS. Then, the medium was replaced by fresh 1% FCS-containing medium with or without UV-oxLDL (100 g of apoB/ml). At the indicated times, cells were washed with ice-cold PBS, harvested using a rubber policeman, and sedimented by centrifugation (300 ϫ g for 5 min). Cell pellets were immediately frozen at Ϫ20°C. Cell pellets were suspended in 0.6 ml of distilled water and homogenized by sonication (2 ϫ 10 s, using an MSE probe sonicator). An aliquot was saved for protein determination (37). Lipids from 0.5 ml of the cell lysate were extracted by 2.5 ml of chloroform/methanol (36), and [ 3 H]choline-labeled SM and [ 3 H]palmitoyl-labeled ceramide were quantified as described previously (11,38).
Measurement of MAP Kinase Activity-MAP kinase activity was determined on immunoprecipitates according to a procedure derived from Ikeda et al. (40). Briefly, after incubation with (or without) LDL, cells (2 ϫ 10 7 ) were placed on dry ice, washed twice with cold PBS, and scraped off with a rubber policeman in buffer A (20 mM Tris-HCl buffer, pH 7.4, containing 50 mM NaCl, 50 mM NaF, 5 mM EDTA, 20 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , 100 g/ml PMSF, 1 g/ml leupeptin, and 1 g/ml aprotinin). After centrifugation (300 ϫ g for 10 min), the cell pellets were lysed for 30 min at 4°C in 500 l of lysis buffer B (20 mM Tris-HCl, pH 7.4, containing 10 g/liter Triton X-100, 1 g/liter SDS, 50 mM NaCl, 50 mM NaF, 5 mM EDTA, 20 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , 100 g/ml PMSF, 1 g/ml leupeptin, and 1 g/ml aprotinin). Samples were centrifuged (12,500 ϫ g for 10 min, 4°C). A 5-l aliquot of the supernatant was used for protein determination (37). The supernatant was precleared for 1 h at 4°C with 5 mg of protein A-Sepharose. After centrifugation (12,500 ϫ g for 5 min), the supernatant was incubated for 4 h at 4°C with 5 l of anti-ERK1 antibody (100 g/ml), and 5 mg of protein A-Sepharose was added for 1 h at 4°C. After centrifugation (12,500 ϫ g for 5 min), the pellet was washed twice with buffer A, solubilized in 200 l of extraction buffer (20 mM Tris-HCl, pH 7.4, 60 mM glycerophosphate, 10 mM EGTA, 10 mM MgCl 2 , 0.1 mM NaF, 2 mM dithiothreitol, 1 mM Na 3 VO 4 , 20 mg/ml leupeptin, and 1 mM PMSF) and used for determining the MAP kinase activity by phosphorylation of MBP in the presence of [␥-33 P]ATP (300,000 dpm/assay) as described (40,41). The reaction mixtures were spotted on phosphocellulose discs, which were washed with 10% cold trichloroacetic acid containing 10 mM sodium pyrophosphate. After drying, the discs were counted by liquid scintillation. Alternatively, proteins were separated by electrophoresis in a 10% polyacrylamide gel, and the radioactivity was visualized by autoradiography (Biomax-MR, Eastman Kodak).
Western Blotting-Phosphotyrosine proteins were detected in SMC by Western blotting (42). Cells (2 ϫ 10 6 /assay) were solubilized in 100 l in lysis buffer B for 30 min at 4°C, centrifuged (12,500 ϫ g for 10 min at 4°C), and 100 g of protein of the supernatant were subjected to SDS-polyacrylamide gel electrophoresis (100 V, 9 mA for 16 h). Proteins were transferred to nitrocellulose membranes (300 V, 200 mA for 5 h), blocked with 5% nonfat milk in Tris-buffered saline-Tween 20 (0.1%) at 4°C for 2 h and incubated overnight at 4°C with anti-phosphotyrosine 4G10 monoclonal antibody (dilution 1:500). The membranes were washed four times in Tris-buffered saline-Tween 20, and antibody reactions were detected using horseradish peroxidase-conjugated sheep anti-mouse IgG (dilution 1:3000) and the ECL chemiluminescent detection reagents (Amersham) according to the manufacturer's instructions. Alternatively, the cell lysate was precleared on protein A-Sepharose and immunoprecipitated with an anti-phosphotyrosine (PY20) monoclonal antibody exactly under the conditions described above for immunoprecipitating MAPKs with anti-ERK. Then, tyrosine-phosphorylated proteins, heated (90°C for 10 min) and dissolved in the electrophoresis buffer, were resolved by SDS-polyacrylamide gel electrophoresis, transferred onto nitrocellulose membranes, and probed with anti-ERK1 or anti-ERK2 antibody, as indicated above.
Indirect Immunofluorescence-SMC, grown on uncoated glass coverslips, were fixed in PBS and 3% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100, and incubated with anti-ERK1 antibody (1:40) for 30 min. After washing three times with PBS, incubation with a secondary fluorescein isothiocyanate-conjugated anti-rabbit IgG for 30 min, and washing again with PBS, slides were mounted in Fluoprep and examined by fluorescence microscopy using a Diaplan (Leitz) or a confocal laser scanning microscope (Zeiss, LSM model).

MAPK Activation in AG-08133A SMC Treated by UV-ox-LDL-
The UV-oxLDL used here were defined by moderate lipid peroxidation (4 Ϯ 1 nmol of thio-barbituric acid-reactive substances/mg of apoB) and only minor modifications of apoB (as shown by minor alterations of the electrophoretic mobility, no significant loss of surface amino groups and cellular uptake through the apoB/E receptor) (34,43). We utilized here a concentration of UV-oxLDL (100 g of apoB/ml) that gave maximal mitogenic effect to cultured AG-08133A SMC (as shown by [ 3 H]thymidine incorporation) and only a low level of cytotoxicity, as assessed by LDH release (10,11).
Mitogenic concentrations of UV-oxLDL induced a biphasic increase of MBP kinase activity in SMC (Fig. 1, A and B). A first, transient peak of MAPK activity, maximal at 1 h, was followed by a second rise, sustained for at least 3 h. Native LDL induced a weak transient peak but not the second sustained peak of MBP kinase (Fig. 1, A and B). Similar results were obtained when using whole cell extracts (Fig. 1A) and anti-ERK immunoprecipitates (data not shown). To provide more evidence for the MAPK activation in SMC stimulated by mitogenic doses of UV-oxLDL, tyrosine phosphorylation of MAPK was examined by immunoblotting. As reported in Fig. 1C, UV-oxLDL induced an increase of tyrosine phosphorylation of p44 and p42 that was evident after 1 h of incubation and plateaued at 3-5 h. As shown in Table I, both the MAPK activation (peaks at 1 and 3 h) and the proliferative effect induced by UV-oxLDL were inhibited by 10 M PD-098059, a MAPK kinase inhibitor (44). This suggests the involvement of the MAPK cascade in the mitogenic signaling triggered by UV-oxLDL.
Subsequently to mitogenic stimulation, activated MAPKs have been shown to be translocated into the nucleus (31,45,46). Therefore, we investigated, by immunofluorescence microscopy, the subcellular localization of MAPKs in SMC incubated with mitogenic doses of UV-oxLDL. As illustrated in Fig.  2, MAPKs underwent a nuclear translocation, which was maximal between 3 and 5 h after the beginning of the incubation of SMC with UV-oxLDL.
MAPK Activation and Proliferation Triggered by UV-oxLDL in AG-08133A SMC Do Not Result from an Autocrine Secretion of Growth Factors-As oxidized LDL have been shown to enhance the expression of various growth factors (6), we investigated whether an autocrine mechanism (secretion of growth SMC were treated under the conditions of A with 100 g of apoB/ml of UV-oxLDL or native LDL (upper and lower panels, respectively), and assays for MBP phosphorylation were performed under the conditions described under "Experimental Procedures." C, tyrosine phosphorylation of MAPK. SMC were incubated with UV-oxLDL or native LDL (100 g of apoB/ml) for the indicated times. SMC lysates were immunoprecipitated with anti-phosphotyrosine protein antibody, resolved on SDSpolyacrylamide gel electrophoresis, and revealed with anti-ERK1/ERK2 antibody. factors by AG-08133A SMC treated by UV-oxLDL) may be involved in the mitogenic signaling and the proliferative effect induced by oxidized LDL.
The effect of RNA and protein synthesis inhibitors on MAPK activation elicited by UV-oxLDL was investigated by preincubating the cells for 2 h with 50 nM actinomycin D or 10 M cycloheximide just before addition of 100 g of apoB/ml of UV-oxLDL. This treatment did not block the MAPK activation (neither the first transient nor the sustained peaks) induced by UV-oxLDL (Table I). This suggests that the MAPK activation (both peaks) triggered by UV-oxLDL does not require RNA and protein synthesis (e.g. synthesis of growth factors).
Then, to investigate the potential involvement of growth factors in the mitogenic effect of UV-oxLDL, two additional sets of experiments were performed using antibodies against growth factors and transfer of preconditioned medium (Tables  2 and 3). As shown in Table II, under the experimental conditions used here, PDGF exhibited a mitogenic effect, whereas bFGF and EGF were not significantly mitogenic to AG-08133A SMC. Moreover, neutralizing antibodies (anti-PDGF, anti-bFGF, and anti-EGF antibodies able to block the mitogenic effect of human and bovine growth factors) did not inhibit the mitogenic effect of UV-oxLDL (Table II).
In a separate set of experiments, preconditioned medium (from SMC treated with UV-oxLDL) was transferred to untreated (reporter) SMC to test its potential mitogenic effect. The preconditioned medium was prepared as follows. AG-08133A SMC were "pulsed" for variable periods of time with a mitogenic concentration of UV-oxLDL. Then, this pulse medium was discarded; after washing, the SMC were "chased" for 2 or 4 h in UV-oxLDL-free medium that was used as preconditioned medium. These preconditioned media were immediately transferred to reporter AG-08133A SMC. Preconditioned media (UV-oxLDL-free) exhibited no or only poorly mitogenic activity (Table III).
Together, these data suggest that, in our model system, the mitogenic effect of UV-oxLDL is probably not mediated by secreted (autocrine) mediators, such as growth factors.
Potential Role of Ceramide in MAPK Activation Evoked by UV-oxLDL-Proliferation of vascular SMC induced by mitogenic doses of UV-oxLDL has been shown to be associated with the activation of the SM-ceramide pathway and to be mimicked by short chain permeant ceramides (11). Here, we show that SM hydrolysis was associated with the activation of a neutral SMase, whereas SMase activities at acidic and alkaline pH remained unchanged (Fig. 3A). The neutral SMase activated by UV-oxLDL was not affected by the presence of EDTA in the reaction buffer (data not shown), thus suggesting that this neutral SMase is magnesium-independent.
In an attempt to understand whether the activation of the SM-ceramide and MAPK pathways triggered by UV-oxLDL were independent or causally related, two sets of experiments were designed using SM hydrolysis inhibitors and exogenous ceramide.
Serine protease inhibitors (serpins) have recently been shown to be able to block the activation of the SM-ceramide pathway involved in daunorubicin-induced apoptosis (47). In our experimental model system, effective concentrations of the two serpins used here, DCIC and TPCK, inhibited concomitantly SMase stimulation (Fig. 3B), SM hydrolysis (Fig. 3C), ceramide production (Fig. 3D), MAPK activation (Fig. 3E), and [ 3 H]thymidine incorporation (Table IV) induced by mitogenic concentrations of UV-oxLDL (100 g of apoB/ml). Finally, we examined whether the inhibitory effect of serpins on [ 3 H]thymidine incorporation may result from a broad toxic effect. Over the time of experiment (5 h), we did not detect any significant toxicity of the two serpins (at the used concentrations, 10 M) (Fig. 3F). In longer time-course experiments (24 or 48 h), DCIC was not toxic to SMC, and TPCK was only slightly toxic, as evaluated by trypan blue staining. Therefore, the inhibitory effect of TPCK or DCIC on [ 3 H]thymidine incorporation in SMC treated by UV-oxLDL (Table IV) does not seem to result from a toxic effect but rather from an inhibition of [ 3 H]thymidine  incorporation. To investigate whether this inhibitory effect was a general effect of protease inhibitors or was relatively specific to TPCK or DCIC, we examined the effect of other protease inhibitors (which inhibit necrosis or apoptosis induced by UV-oxLDL) (48). As reported in Table IV, leupeptin, antipain, and pepstatin A were ineffective in contrast to TPCK or DCIC. As C 2 -ceramide and ceramide generated at the plasma membrane by bacterial SMase treatment are mitogenic to SMC (11), we investigated the potential role of ceramides in the activation of MAPK. SMase treatment, used under non-lytic conditions leading to SM hydrolysis and ceramide generation in the plasma membrane (11), elicited a biphasic activation of MBP kinase and enhanced [ 3 H]thymidine incorporation (Fig. 4). In contrast, treatment of cells by bacterial phospholipase C (PLC, used under non-lytic conditions leading to about 40% hydrolysis of cellular phosphatidylcholine without significant hydrolysis of SM) (11) induced no significant activation of MBP kinase nor [ 3 H]thymidine incorporation (Fig. 4).
Similarly, the synthetic short chain permeant C 2 -ceramide induced a biphasic activation of the MBP kinase (Fig. 5A) and enhanced [ 3 H]thymidine incorporation (Fig. 5B). In contrast, C 2 -dihydroceramide (DHC), an inactive analog of C 2 -ceramide, used under conditions similar to C 2 -ceramide, did not activate MBP kinase nor increase [ 3 H]thymidine incorporation (Fig. 5,  A and C). The data of Figs. 4 and 5 suggest that MAPK activation and [ 3 H]thymidine incorporation induced by SMase treatment or by C 2 -ceramide did not result from a nonspecific effect, as bacterial PLC and DHC were inactive.
Effective concentration of the MAPK kinase inhibitor PD-098059 abrogated both the MAPK activation and [ 3 H]thymidine incorporation elicited by C 2 -ceramide (Fig. 5, B and C), thus suggesting that [ 3 H]thymidine incorporation evoked by C 2 -ceramide is dependent on MAPK activation. In contrast, TPCK and DCIC did not inhibit significantly the MAPK activation and the mitogenic effect elicited by the cell-permeant short chain ceramide (Fig. 5, B and C). Together, these data strongly suggest that, in this experimental model system, serpins inhibit the UV-oxLDL-induced ceramide production (probably by acting upstream from SM hydrolysis) and subsequently [ 3 H]thymidine incorporation. DISCUSSION The present study shows that, in cultured vascular AG-08133A SMC, (i) mitogenic concentrations of UV-oxLDL elicit the activation of the SM-ceramide pathway and the stimulation and nuclear translocation of MAPK and (ii) ceramides (natural ceramide generated in the plasma membrane by SMase treatment or short chain permeant C 2 -ceramide) are effective mitogenic mediators, which are also able to induce activation and nuclear translocation of MAPK and subsequent stimulation of [ 3 H]thymidine incorporation.
Oxidized LDL have been shown to induce a significant increase of [ 3 H]thymidine incorporation and proliferation of cultured aortic SMC (9 -11). This mitogenic effect is stronger than that of native LDL (10,11), which were only poorly mitogenic to SMC, despite their ability to induce some intracellular signaling (49).
Moderate concentrations of UV-oxLDL (100 g of apoB/ml) induced a significant mitogenic response in vascular SMC, in a medium supplemented with either human lipoprotein-depleted serum (10) or 1% FCS (11). In G 0 -arrested cells, initiation of DNA synthesis is a complex process that requires potent growth factors enabling cells to pass the G 1 restriction point controlling S phase entry (50). We report here that, similarly to growth factors, the proliferative effect of UV-oxLDL on SMC was associated with a biphasic, sustained stimulation of MBP kinase activity, tyrosine phosphorylation. and nuclear translocation of p44/p42 MAPKs. In contrast, native LDL induced only a transient activation of MAPK. The role of MAPK activation in the proliferation of SMC induced by UV-oxLDL was supported by the inhibitory effect of PD-098059, a MAPK kinase inhibitor, which blocked both the MAPK activation (45) and the SMC proliferation.
An early and transient activation of MAPK has been reported to be induced by native LDL (44,51,52) or oxLDL (44,(52)(53)(54), but, in these studies, no information was provided on the late events (over 5 h) and nuclear translocation of MAPK, which are thought to play a key role in the full mitogenic response (55). When monitoring the MAPK activation during a longer period of time, we observed an obvious difference between MAPK activation triggered by native LDL and UV-oxLDL, as only UV-oxLDL induced the biphasic and sustained activation of MAPK. This response is quite similar to that observed with classical mitogens that induce a full mitogenic response (30,31,55).
We therefore considered the possibility that oxidized LDL may promote indirectly cell growth via an auto/paracrine mechanism resulting from enhanced expression of growth factors. Several data obtained suggest that a role of autocrine growth factors is unlikely in our model system (cultured vascular AG-08133A SMC used under the described experimental conditions). (i) The interval of time between cell contact with and MAPK activation by UV-oxLDL (the first peak occurred after a 1-h pulse with UV-oxLDL) is relatively short for an autocrine signaling involving stimulation of growth factor synthesis by UV-oxLDL. For instance, as oxLDL-induced PDGF-A mRNA peaked after 2 h of incubation (49) and protein synthesis and secretion of PDGF need some additional time, PDGF secretion cannot explain the MAPK activation occurring at 1 h. (ii) Similarly, MAPK activation triggered by exogenous SMase (which mimics the mitogenic effect of UV-oxLDL) occurs very rapidly (first peak at 5 min), thereby excluding a role of growth factors (gene induction, protein synthesis, and secretion of growth factors being very unlikely in this very short time). (iii) Inhibition of mRNA or protein synthesis by effective doses of actinomycin D or cycloheximide did not block the MAPK activation induced by UV-oxLDL. (iv) MAPK activation by ceramide (ce- Anti-EGF, 60 161 Ϯ 10 0 Anti-EGF, 60 101 Ϯ 8 100 Anti-FGF, 1 159 Ϯ 11 0 Anti-FGF, 1 102 Ϯ 9 a [ 3 H]Thymidine incorporated into DNA ranged between 346,000 and 512,000 dpm/mg of cell protein. b p Ͻ 0.01% (comparison to the relative controls, i.e. between assays with and without antibody or UV-oxLDL). ramide generated by SMase and permeant C 2 -ceramide) was not abrogated by actinomycin D or cycloheximide (data not shown). (v) Blocking antibodies against PDGF, bFGF, and EGF (used under conditions inhibiting the response induced by the related growth factor) did not affect the MAPK activation nor the mitogenic response triggered by UV-oxLDL. Moreover, in incorporation of reporter AG-08133A SMC Preconditioned media were prepared as follows. SMC were preincubated (or "pulsed") for 1 or 16 h in RPMI containing 1% FCS and 100 g of apoB/ml of UV-oxLDL; then, this medium (containing UV-oxLDL) was discarded, and SMC were incubated ("chased") for an additional 2 or 4 h in RPMI containing 1% FCS (preconditioned culture medium). The preconditioned media ("chase" media) were then transferred to reporter AG-08133A SMC. After addition of the preconditioned medium, the MAPK activities and [ 3 H]thymidine incorporation were determined on the reporter SMC, at 1 and 3 h for MAPK (time of the MAPK peaks evoked by UV-oxLDL) and 16   agreement with Stiko-Rahm et al. (49), when the preconditioned medium (obtained by preincubating SMC with oxLDL) was transferred on reporter SMC, no significant mitogenic effect was observed. Taken together, these data strongly suggest that an autocrine secretion of growth factors triggered by oxLDL plays only a minor role in our model system. Therefore, from our recently reported data (11), we hypothesized that UV-oxLDL may act directly as a mitogen to SMC by triggering the SM-ceramide pathway and subsequently the MAPK cascade. Of course, these data do not exclude the possibility that, in atherosclerotic lesions, growth factors may play a prominent role in the genesis of lesions (see review in Ref. 6).
The identification of SMase(s) involved in ceramide generation has proven to be equivocal. At least three forms of SMase are candidates for SM hydrolysis. Acid SMase has been associated with signaling triggered by TNF␣ (39), Fas (19), and CD28 (18). Activation of a membrane-bound neutral SMase has also been described with TNF␣ (39), anti-Fas antibody (20), and daunorubicin (21). Finally, a cytosolic magnesium-independent neutral SMase has been implicated in vitamin D3induced differentiation (56). In our study, UV-oxLDL induced only activation of a neutral SMase (peaking at 1 h), which may be similar to the enzyme described by Okazaki et al. (56). The activation of the neutral SMase was inhibited by TPCK and DCIC concomitantly with SM hydrolysis and ceramide generation, thus suggesting that the neutral SMase is involved in SM hydrolysis and mitogenic signaling triggered by UV-ox-LDL. These data are consistent with the role of this SMase in the TNF␣-induced mitogenic signaling in fibroblasts (38). The molecular mechanism of the activation of this neutral SMase by oxLDL remains unknown, but its inhibition by serpins (Ref. 47 and present paper) or by caspase inhibitors (57,58) supports the hypothesis for a role of protease(s) in SMase activation.
Furthermore, the two serpins TPCK and DCIC inhibited concomitantly ceramide generation (subsequent to SM hydrolysis), MAPK activation, and [ 3 H]thymidine incorporation triggered by UV-oxLDL but did not inhibit the MAPK activation and mitogenic effect induced by short chain ceramide. Taken together, these data suggest that serpins act upstream from the SMase, and do not interfere, in our model system, between ceramide and downstream signaling, i.e. MAPK activation and DNA synthesis.
Another major question is to know whether the activation of the SM-ceramide pathway is involved in the mitogenic effect or is only a parallel consequence of the response of cells to UV-oxLDL. The former hypothesis is supported by the fact that both MAPK activation and mitogenic effect of UV-oxLDL were mimicked by exogenous short chain permeant ceramide and by ceramide generated at the plasma membrane by exogenous SMase. A nonspecific effect of these mediators (for instance membrane damages triggering the activation of the MAPK pathway) is unlikely because (i) phospholipid hydrolysis by exogenous phospholipase C did not trigger MAPK activation or DNA synthesis, although it induced probably plasma membrane structural changes similar to those elicited by SMase, and (ii) C 2 -dihydroceramide was ineffective, in contrast to C 2 -ceramide.
Together, the above reported data are consistent with the hypothesis that UV-oxLDL may induce the sequential activation of the SM-ceramide pathway, MAPK cascade, and downstream mitogenic signaling. This is consistent with Raf1 phosphorylation (59) and MAPK activation induced by endogenous or cell-permeant ceramides in HL60 leukemic cells (60, 61),  fibroblasts (62,63), and endothelial cells (64). Although MAPK activation by ceramide has been demonstrated in cellular systems undergoing apoptotic or inflammatory responses, our results strongly implicate activation of MAPK by ceramide in a proliferative signaling pathway consistent with the well documented mitogenic role of MAPK (30,31,55).
In various experimental models, other sphingolipid mediators such as lactosylceramide, sphingosine 1-phosphate, or sphingosylphosphocholine have been shown to stimulate the p42/p44 MAPK or exert a proliferative effect (54,(65)(66)(67). The possibility that, in our model system, ceramide may be degraded and converted into sphingosine 1-phosphate or sphingosylphosphocholine, or utilized as a precursor for lactosylceramide biosynthesis, cannot be ruled out. To date, the respective role of the ceramide pathway and other pathways involving other sphingolipid mediators in the oxLDL-induced mitogenic effect remains to be elucidated.
In conclusion, the present paper reports several novel results concerning the mitogenic intracellular signaling triggered by oxLDL in vascular SMC. (i) The mitogenic effect of UV-oxLDL is mediated through serpin-sensitive serine proteases, which activate in turn the SM-ceramide pathway. (ii) The mitogenic effect resulting from the activation of the SM-ceramide pathway is mediated through a biphasic and sustained activation and nuclear translocation of MAPK. Finally, our data suggest that the activation of SM-ceramide pathway induces in turn a sustained activation of the MAPK pathway, associated with nuclear translocation of MAPK, which is a critical event during G 1 progression (55) and seems to be sufficient for inducing cell proliferation (68). As oxLDL are present in atherosclerotic lesions, but not in normal vascular wall (7), this mitogenic stimulus is locally persistent and specific to atherosclerotic areas. It is therefore suggested that, besides the other cytokines and growth factors potentially involved in atherogenesis (6), oxLDL may play a critical role in the proliferation of SMC occurring in atherosclerotic plaque.