Cholera toxin, pertussis toxin, heparin, and genistein were purchased from Calbiochem. All other reagents, such as H₂O₂ and polyethylene glycol conjugated-catalase (PEG-Cat), were purchased from Sigma unless otherwise stated.

hAoSMCs were purchased from Clonetics and cultured in SmGM-2 Bullet kit medium (Clonetics) in 5% CO₂ at 37℃. Twenty-four hours before experiments, growth medium was removed and replaced with DMEM medium (Gibco) containing 0.1% FBS (Gibco) for 24 hours (maintenance medium).

n-LDL (density 1.019-1.063 g/mL) was isolated from the plasma of normocholesterolemic subjects (serum cholesterol <6.2 mmol/L) by differential ultracentrifugation as previously described.16 During LDL manipulation and storage, particular precautions were taken in order to maintain nLDL integrity and prevent oxidation. The content of endotoxin was below the detection limit (<1 ng/mL), as measured using an endotoxin assay kit (Pharmingen). Prepared nLDL was confirmed by thiobarbituric acid reactive substances (TBARS) assay measurement using malondialdehyde as a standard and compared its oxidation with commercial nLDL and Ox-LDL (Intracel Co., Frederick, MD, USA).

Pretreated cells grown on gelatin-coated glass were stained with DCF-DA (2',7'-dichlorofluoroscein diacetate, 10 µmol/L), followed by incubation at 37℃ for 30 minutes. Cells were washed three times with phosphate-buffered saline (PBS) and imaged at 488/535 nm using an epifluorescence microscope (Olympus BX51).

The amounts of IL-8 released into the media were assayed using a human IL-8 ELISA Kit II (OptEIA, BD Biosciences, NJ, USA) according to the manufacturer's instructions.

Cells treated with different chemicals were lysed in SDS sample buffer (62.5 mmol/L Tris pH6.8, 2% SDS, and 10% glycerol). Each sample was resolved on a 10% SDS-PAGE, transferred onto PVDF membranes (Pall), analyzed with antibodies according to the supplier's instructions, and then visualized with peroxidase and an enhanced-chemiluminescence system (ECL kit, Amersham, NJ, USA). Polyclonal phospho-p38 MAPK was purchased from Cell Signaling Technology and normalization was performed with the smooth muscle cell (SMC)-specific α-actin antibody (BD Biosciences, NJ, USA).

Transwell plate (13 mm gelatin-coated polycarbonate filter, 8 µm pore size, Costar Corp.) was used for the migration assay. Briefly, suspended cells (1×10⁵) were added to the upper chamber of the Transwell insert, and IL-8 or neutralizing antiserum against IL-8 (R&D system, MN, USA) was mixed with the medium [DMEM+1% FBS+0.1% bovine serum albumin (BSA)] in the lower chamber. After the transwell was incubated for 8 hours, the filters were fixed with methanol (10 minutes at 4℃) and stained with hematoxylin. The number of migrated cells was determined by microscopy (Olympus BX51, 400X high power field).

All data represent mean±S.D., and unpaired Student's t-test was used to assess significant differences between the two groups. A value of p<0.05 was accepted as significant.

First, we tested whether prepared LDL underwent any change such as oxidative modification. However, LDLs prepared in 3 independent experiments showed no oxidation, because the value was -0.60±0.06 µmol/L in TBARS assay. This value was comparable to commercially available nLDL, which was -0.55±0.04 µmol/L, whereas, oxidized LDL was 7.95±0.60 µmol/L.

Since we previously demonstrated that stimulation of hAoSMCs with nLDL increased IL-8 production via p38 MAPK activation 16, we attempted to determine the nature of the receptor responsible for IL-8 production. As shown in Fig. 1A, increased IL-8 production by nLDL from 4.96±0.36 ng/mL to 14.03±0.56 ng/mL (Fig. 1A, *vs. untreated, p<0.01) was completely blocked by pretreatment of the cells with PTX (Fig 1A, **vs. nLDL=3.23±0.03 ng/mL vs. 14.03±0.56 ng/mL, p<0.01). This effect of pertussis toxin (PTX) was associated with inhibition of p38 MAPK phosphorylation (Fig. 1B). On the other hand, cholera toxin (CTX) treatment had no effect on IL-8 production (Fig. 1A, #vs. nLDL=13.63±0.39 ng/mL vs. 14.03±0.56 ng/mL, not significant) and p38 MAPK phosphorylation (Fig. 1B). Furthermore, inhibitors related to receptor-associated signal transduction, such as heparin for the classical nLDL receptor or genistein for receptor tyrosine kinase, were employed, however, these agents had no inhibitory effect on IL-8 production (Fig. 1A, nLDL+heparin vs. nLDL=13.91±0.84 ng/mL vs. 14.09±1.38 ng/mL, not significant; nLDL+genistein vs. nLDL=13.89±0.40 ng/mL vs. 14.09±1.38 ng/mL, not significant) and p38 MAPK activation (Fig. 1C). In addition to nLDL, sphingosine-1-phosphate (S1P) and lysophosphatidic acid (LPA) have been known as agonists binding to Gi-protein coupled receptors. Therefore, nLDL was compared with S1P and LPA in terms of IL-8 production. Unlike nLDL, however, Fig. 1D shown that S1P and LPA did not induce IL-8 production (*vs untreated=13.57±0.29 ng/mL vs. 5.53±0.59 ng/mL, p<0.01; S1P or LPA vs. untreated=5.56±0.34 ng/mL (S1P) or 5.51±0.50 ng/mL (LPA) vs. 5.53±0.59 ng/mL, not significant).

We next tested whether nLDL induces H₂O₂ generation and also H₂O₂ generation is dependent on PTX-sensitive G-protein coupled receptor. As shown in Fig 2, incubation of hAoSMCs with nLDL resulted in a significant increase of DCF fluorescence (*vs. untreated=7.42±0.72 vs.1.00±0.14, p<0.01, n=3). On the other hand, preincubation of the cells with PTX (#vs. nLDL=1.10±0.10 vs. 7.42±0.72, p<0.01, n=3) or PEG-Cat (##vs. nLDL=0.92±0.19 vs. 7.42±0.72, p<0.01, n=3) significantly prevented the increase of nLDL-induced DCF fluorescence. Furthermore, H₂O₂ markedly increased DCF fluorescence (**vs. untreated=5.36±0.38 vs. 1±0.14, p<0.01, n=3). Fig. 2B shows relative fluorescence values as bar graph from 3 independent experiments.

Since p38 MAPK is a critical component of redox-sensitive signaling pathways in VSMCs,17-19 we tested whether nLDL-induced p38 MAPK activation and IL-8 production were associated with increased intracellular H₂O₂ generation. As shown in Fig. 3A, nLDL-induced p38 MAPK activation was completely blocked with pretreatment of the cells with cell-permeable PEG-Cat. Furthermore, increased IL-8 production by nLDL was also abrogated by incubation with PEG-Cat (Fig. 3B, *vs. untreated=15.12±0.36 ng/mL vs. 4.83±0.29 ng/mL, p<0.01; #vs. nLDL=4.98±0.41 ng/mL vs.15.12±0.36 ng/mL, p<0.01). Interestingly, exogenous H₂O₂ at a low concentration of 10 µmol/L induced the phosphorylation of p38 MAPK as early as 10 minutes (Fig. 3C) and increased IL-8 production in a dose-dependent manner (Fig. 3D, *vs. untreated control=6.42±0.39 ng/mL vs. 4.95±0.36 ng/mL, p<0.05; #vs. untreated=10.23±1.26 ng/mL vs. 4.95±0.36 ng/mL, p<0.01). Since exogenous H₂O₂ at concentrations higher than 20 µmol/L significantly decreased cell viability as early as 1 hour, we failed to completely mimic IL-8 production to the level of nLDL stimulation. Therefore, these data suggest that intracellular localization of H₂O₂ generation could be important in the activation of redox-sensitive protein kinases.

We next examined whether the secreted IL-8 had any effect on the migration of hAoSMCs that is one of critical phenomena of hAoSMC during atherogenesis. Therefore, the effect of IL-8 on hAoSMC migration was determined by Transwell assay as described in the methods section. As shown in Fig. 4, IL-8 treatment at 15 ng/mL significantly increased hAoSMCs migration (*vs. untreated=1.88±0.30 vs. 1.04±0.30, p<0.01). The migration activity of IL-8 was completely blocked with neutralizing antiserum (0.5 µg/mL) against IL-8 (#vs. IL-8=0.98±0.16 vs.1.88±0.30, p<0.01). PDGF-BB as a positive control increased the migration as much as 2.45±0.11 fold.