Interferon-g Induces Secretory Group IIA Phospholipase A2 in Human Arterial Smooth Muscle Cells INVOLVEMENT OF CELL DIFFERENTIATION, STAT-3 ACTIVATION, AND MODULATION BY OTHER CYTOKINES*

Increased expression of secretory non-pancreatic phospholipase A2 (sPLA2-IIA) could be part of the inflammatory reaction in atherosclerosis. However, the factors controlling sPLA2-IIA production in human vascular cells are unknown. We investigated regulation of sPLA2-IIA expression and secretion by human arterial smooth muscle cells in culture (HASMC). SPLA2-IIA was induced after 3–14 days of culture in non-proliferating conditions. SPLA2-IIA was co-expressed with heavy caldesmon, a cytoskeleton protein, and p27, a G1 cyclin inhibitor, proteins characteristically expressed by differentiated cells. Further incubation with 50–500 units/ml of interferon (IFN)-g significantly increased sPLA2-IIA mRNA and secretion. IFN-g-induced sPLA2IIA was found to be active in cell media and associated with cell membrane proteoglycans. IFN-g induced sPLA2-IIA expression was antagonized by tumor necrosis factor (TNF)-a and interleukin (IL)-10. TNF-a added individually induced a significant but transient (4 h) increase in sPLA2-IIA secretion. IL-10 by itself did not affect sPLA2-IIA expression and secretion. IFN-g-stimulated sPLA2-IIA transcription involved STAT-3 protein. Interestingly, IL-6 but not IFN-g up-regulated the sPLA2-IIA expression in HepG2 cells, thus sPLA2-IIA induction by IFN-g response appears to be cell specific. In summary, conditions leading to cell differentiation induced sPLA2-IIA expression in HASMC and further exposure to IFN-g can up-regulate sPLA2-IIA transcription and secretion. This IFN-g stimulatory effect can be modulated by other cytokines.

position of glycerophospholipids yielding nonesterified free fatty acids and lysophospholipids (1). These products may either act as intracellular second messengers or can be further metabolized into proinflammatory and mitogenic lipid mediators including eicosanoids, platelet activating factor, and lysophosphatidic acid (1). Lysophosphatidylcholine is a mediator of a broad range of cellular processes on vascular and inflammatory cells (2)(3)(4). Many of these lipid mediators accumulate during atherosclerotic lesion development (5,6). SPLA 2 -IIA appears to be involved in several physiological and pathological processes. It contributes to membrane remodeling and removal of oxidized phospholipids (7). In addition, sPLA 2 -IIA has bactericidal and anti-tumorigenic properties, participates in TNF␣-induced activation of nuclear transcription factor and expression of cell-adhesion molecules (8 -10). Regarding pathological situations, sPLA2-IIA can induce acute inflammatory changes when injected in vivo (11). Furthermore, there is a correlation between elevated levels of sPLA 2 -IIA and inflammatory conditions, such as rheumatoid arthritis, septic shock, acute respiratory distress syndrome, and coronary artery disease (12,13). Atherosclerosis has characteristics of an inflammatory process (14). Furthermore, a high incidence of atherosclerosis and high mortality from cardiovascular diseases have been reported in patients with chronic inflammatory diseases that have prolonged periods of high extracellular sPLA 2 -IIA activity (15). Additionally, transgenic mice overexpressing human sPLA 2 -IIA with high levels of the enzyme in plasma and in the aortic intima/media develop more aortic atherosclerosis than non-transgenic littermates (16). In humans it was recently reported that in patients with coronary artery disease the plasma level of sPLA 2 -IIA was an independent risk factor. Furthermore, a high level of sPLA 2 -IIA was also a significant predictor of new coronary events during a 2-year follow up period (13).
Earlier immunohistochemistry studies from our group and others indicate that smooth muscle cells are the main source of sPLA 2 -IIA in human arteries (17)(18)(19). Our data from electron microscopy-immunogold examination revealed that the majority of sPLA 2 -IIA in human atherosclerotic lesions is localized extracellularly associated with collagen fibers and in close contact with extracellular lipid droplets. Intracellular sPLA 2 -IIA was observed in electron-dense vesicles in the cytosol (20). These observations suggest that sPLA 2 -IIA may be involved in the pathogenesis of atherosclerosis. One mechanism by which sPLA 2 -IIA may be atherogenic is by modifying lipoproteins and releasing inflammatory lipid mediators at places of lipoprotein retention in the arterial wall (21). To evaluate the possible biological and pathological function(s) of sPLA 2 -IIA in the arterial wall it is necessary to clarify the mechanisms regulating its expression and secretion by vascular cells. Inflammatory * This work was supported by Swedish Medical Research Council Grants MFR Project 4531 and 13129-01, the Heart and Lung Foundation in Sweden HLF, Project 41346 King Gustaf V:s 80 th year Foundation, Swedish Society of Medicine Project 02-0424, Sahlgrenska University Hospital Foundation, and AstraZeneca, Mölndal, Sweden. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Arterial vascular smooth muscle cells can change their phenotype between contractile/non-proliferating and synthetic/ proliferating forms. These phenotypic changes are accompanied by changes in gene pattern expression and are associated with the pathological arteriosclerotic process (26). The aims of the present study were two. First, to study the expression of sPLA 2 -IIA at mRNA and protein level in an in vitro model of phenotypic modulation of HASMC. Second, to investigate the effect of IFN-␥ and other pro-and anti-inflammatory cytokines in the transcription and secretion of sPLA 2 -IIA by HASMC.

MATERIALS AND METHODS
Reagents-Collagen-I (rat-tail) was purchased from Collaborative Biomedical, Becton Dickinson, Labware, Bedford, MA. Waymouths' cell culture medium and Dulbecco's phosphate-buffered saline with and without calcium and magnesium were purchased from Life Technologies, Inc. Antibiotics, Eagle's modified essential medium, trypsin-EDTA, and non-essential amino acids were from Biowhittaker, Verviers, Belgium. Fetal bovine serum (FBS) was purchased from PAA Laboratories, GmbH, Linz, Austria. Culture vessels were purchased from Life Technologies, Inc., Grand Island NY. Human recombinant interferon-␥ (IFN-␥), interleukin 1␤ (IL-1␤), IL-6, IL-10, and tumor necrosis factor-␣ (TNF-␣) were purchased from IK Immunokontact, Switzerland. Monoclonal mouse antibodies against p27 and p21 proteins were from Transduction Laboratorium, Lexington, Ky; monoclonal mouse anti-sPLA 2 -IIA was from Cayman Chemicals (Ann Arbor, MI); biotin-SP-conjugated affinity pure donkey anti-rabbit IgG (HϩL) and peroxidase-conjugated streptavidin were from Jackson ImmunoResearch, Laboratories, Inc. Bromodeoxyuridine ELISA kit for non-radioactive determination of cell proliferation and protease inhibitors were purchased from Roche Diagnostics Scandinavian, Bromma, Sweden. Chondroitinase ABC and heparitinase I were purchased from Seikagaku Corp., Tokyo, Japan. Salts, buffer substances, and detergents used in this work were of analytical grade and were purchased from Merck, Darmstadt, Germany, and Bio-Rad.
Cell Culture and Differentiation Protocol-HASMC were isolated from inner media of human uterine arteries and cultured in Waymouth's medium, supplemented with 10% FBS, 100 untis/ml penicillin, 100 g/ml streptomycin, 1 mM sodium pyruvate, and 4 mM glutamine at 37°C in a 5% CO 2 , 95% air atmosphere. The medium was changed every 3 days, and cells were detached by treatment with 0.05% trypsin, 0.02% EDTA solutions. Studies were conducted on HASMC between passage 4 and 9. HASMC were seeded at 5 ϫ 10 3 cells/cm 2 in 25-and 80-cm 2 cell culture bottles coated with collagen-I in growing medium (27). After 2 days the medium was removed, cells washed 3 times with Dulbecco's phosphate-buffered saline, and then cultured in Waymouth's medium containing the supplements indicated above but with only 0.5% FBS (serum-poor medium) in order to synchronize the cells by arresting proliferation. After 3 days the medium was removed, cells washed 3 times with Dulbecco's phosphate-buffered saline, and cultured in growing medium for 3-4 days until the cells were confluent. At confluence, cells were stimulated to differentiate in serum-free Waymouth's medium containing the other supplements indicated above (serum-free medium). HASMC were cultured in this medium up to 21 days. Human aortic and pulmonary cells (Clonetics Corp.) were cultured under similar conditions as HASMC. HepG2 liver tumor cells (ATTC, Manassas, VA) were culture in Eagle's modified essential medium containing 10% FBS and supplemented as the growing medium described above. Cells were free of mycoplasma. Endotoxin levels were regularly tested in cell culture medium and cell culture reagents with Coatest/endotoxin, Chromogenix AB (Molndal, Sweden). Levels detected were below 0.01 units/ml.
RNA Preparation and RT-PCR Procedure-Total cellular RNA was isolated from HASMC and CHO-cell line expressing human sPLA 2 -IIA (28) using a single step, acid guanidinium isothiocyanate/chloroform extraction method (17). RT-PCR with 0.1 or 0.5 g of total RNA was performed using a GeneAmp kit from Perkin-Elmer. Oligos and conditions for amplification of cDNA were as described. The oligonucleotides designed from the determined cDNA sequence for human sPLA 2 -IIA (29) amplified a 344-base pair product and the sequences were: sense primer, ATGAAGACCCTCCTACTGTT, and antisense primer, AG-CAGCCTTATCACACTCAC (17). The human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) primer set amplified a 983-base pair product and the sequences were: sense primer, TGAAGGTCGGAGT-CAACGGA and antisense primer, CATGTGGGCCATGAGGTCCA (30). The high molecular weight caldesmon primer set amplified a 530-base pair product and the sequences were: sense primer, AACAACT-GAAAGCCAGGAGG and antisense primer, GCTGCTTGTTACGTT-TCTGC (31). All incubations were done in a Biometra, TRIO-Thermoblock. The PCR products were then separated on a 4% Nusieve GTGagarose gel, FMC Bioproducts Corp., Rockland, ME.
Analysis and Quantification of the cDNA-Fluorescent labeled PCRamplified products. PCR products were subjected to a semiquantitative analysis following the fluorescent labeling method described in ABI PRISM 377, DNA Sequencer GeneScan TM Chemistry Guide, Perkin-Elmer. PCR products were labeled on one strand using fluorescent 5Ј-FAM-labeled sense primers. The labeled strand from the PCR product was separated and detected by running the gel in the ABI Prism 377 DNA sequencer, Perkin-Elmer essentially as described by the manufacturer. The relative quantity of the amplified fragments was determined by dividing the peak areas corresponding to the fluorescence of the sPLA 2 -IIA fragment against the peak area of the GAPDH fragment. GAPDH and sPLA 2 -IIA fragments were amplified simultaneously in separate PCR reaction tubes. The results presented represent average and standard deviation of three separate GeneScan analyses.
Quantification of sPLA 2 -IIA by ELISA-Cell-associated and -secreted sPLA 2 IIA were measured by enzyme-linked capture antibody immunoassay (32). The antibodies used were: monoclonal antibody against human sPLA 2 -IIA (1 g/l); polyclonal antibody (IgG fraction) against human recombinant sPLA 2 -IIA developed at the lab with demonstrated no cross-reactivity with PLA 2 type V 2 or actin; and biotin-Spconjugated AffiniPure donkey anti-rabbit IgG (HϩL). Microtiter plates (Sero-Wel, Bibby Sterilin Ltd. Stone, Staffs, United Kingdom) were coated with the monoclonal antibody against sPLA 2 -IIA (50 l/well, 5 g/ml) in 15 mM sodium carbonate buffer, pH 9.6, for 18 h in a humid chamber at room temperature. The plates were washed 3 times with 50 mM Tris, 150 mM NaCl pH 7.6 (TBS buffer), containing 0.05% Tween 20 (washing buffer). HASMC in a 25-cm 2 flask were extracted with 0.5 ml of phosphate-buffered saline, pH 7.4, containing 0.5% Triton X-100, 5 M AEBSF, 0.1 mM leupeptin, 5 M aprotinin, and 1 g/ml soybean trypsin inhibitor. For the ELISA, 50 l of cell medium or 50 l of cell extract were added to the wells by quadruplicate and incubated for 1 h at 37°C. This step was repeated once. In total 100 l of cell extract or media were applied in each well. The plates were washed 3 times with washing buffer, followed by the addition of 1:5000 dilution of polyclonal antibody (IgG fraction) against human recombinant sPLA 2 -IIA, and incubated 1 h at 37°C. The plates were then washed 3 times and incubated with 1:5000 dilution of biotin-SP-conjugated donkey antirabbit IgG for 1 h at 37°C. All the antibodies were diluted in TBS buffer containing 1% bovine serum albumin and 0.05% Tween 20 (incubation buffer). The plates were washed 3 times with washing buffer followed by the addition of 50 l/well of the peroxidase-conjugated streptavidin diluted 1:2000 in incubation buffer and incubated 1 h at 37°C. Then the plates were washed again, followed by the addition of 50 l/well of substrate azino diethylbenzthiazoline sulfonate (22 mg/ml) and incubated for 30 min at 37°C after which the plates were read at 405 nm (Spectra MAX Plus Microplate Spectrophotometer System, Molecular Devices, Sunnyvale, CA). Purified human recombinant sPLA 2 -IIA (28) was used to generate a standard curve (62.5-2000 pg/50 l/well).
Preparation of Nuclear Protein Extracts and Electrophoretic Mobility Shift Assay-HASMC were incubated with or without cytokines as described above. After 4 and 24 h incubation cell nuclear extract were prepared. Nuclear extract preparation, annealing, and purification of oligonucleotides and electrophoretic mobility shift assay were performed as described (34). Oligonucleotides were synthesized by Life Technology, Ltd., Paisley, UK. Double-stranded oligonucleotides identical to the sequences from Ϫ563 to Ϫ535 and Ϫ240 to Ϫ208 in the human sPLA 2 -IIA promoter, containing a NF-B and STAT-binding site, respectively, were used. In some experiments for identification of nuclear transcription factors in the EMSA assay, rabbit polyclonal antibodies (Santa Cruz Biotechnology Inc.) against the p65 subunit of NF-B and against STAT 1, 3, 5, and 6 proteins were added to the EMSA binding reactions.
Protein Determination-Protein concentrations were determined by Bradford protein analysis kits (Bio-Rad).
Data Analysis-Data are expressed as mean Ϯ S.D. Each experiment was performed at least 3 times. The results obtained were reproducible between experiments. Variation in the total levels of sPLA 2 -IIA expression, mRNA, secretion, and cell associated, could be observed sometimes between batches of HASMC. Each data point represent n ϭ 3 or 4. Individual statistical comparisons of paired data were evaluated by Student's t test with p Յ 0.05 representing significance. Standard curves and statistical analysis of the results were performed using GraphPad Prism, v2.0 (GraphPad Software, San Diego, CA).

Phenotypic Changes during Differentiation of Cultured Human Arterial Smooth
Muscle Cells-To investigate the regulation of sPLA 2 -IIA expression in HASMC, we used an in vitro model of phenotypic modulation or differentiation in culture shown in Fig. 1 and described under "Experimental Procedures." We studied the following parameters: 1) growth of the cells by determination of cell number and DNA synthesis; 2) cell morphology; 3) expression of the cell differentiation markers, cyclin-dependent kinase inhibitors of mammalian cell cycle p27 and p21 and heavy caldesmon, a marker of human smooth muscle cell differentiation. As shown in Fig. 2A after 3 days of cell synchronization in 0.5% FBS (serum-poor medium) there was a marked reduction in DNA synthesis with little change in cells numbers. These results reflect the expected reduction in proliferation rate of confluent cells cultured in the absence of serum. Phase-contrast microscopy showed no morphological changes between HASMC maintained 1, 3, 7, and 14 days in serum-free medium (Fig. 2B). These HASMC cultures were judged as normal dense cultures of fully confluent spindle shaped cells.
The expression of p27 and p21 cyclin-dependent kinase (Cdk2) inhibitors, as markers of cell differentiation, was evaluated by Western blot analysis of cell extracts. As shown in Fig.  3A p27 protein was not detected in HASMC after 2 days in 10% FBS when proliferation was as highest, whereas the same cells after 1 or 14 days in defined serum-free medium (non-proliferating conditions) showed as expected a markedly increase in the expression of p27. The p21 protein was detected under proliferating and non-proliferating conditions, however, its expression was increased after 1-day culture in serum-free medium. We also analyzed the mRNA level for heavy caldesmon, a marker of smooth muscle cell differentiation found in the vascular wall, in parallel to analysis of mRNA for sPLA 2 -IIA and GAPDH. PCR-amplified fragments showed the corresponding size of the different fragments and only one well defined band was amplified (Fig. 3B). These results shown that post-confluent HASMC expression of sPLA 2 -IIA and heavy caldesmon mRNA increased significantly after 7 and 14 days in culture in serum-free medium. In contrast, the levels of GAPDH mRNA remained constant during culture. These observations indicate that the in vitro protocol used for culture of HASMC lead the cells to a quiescent, differentiated state that promotes expression of sPLA 2 -IIA. Similar results were observed with human smooth muscle cells from aorta and pulmonary artery (data not shown).
Modulation of sPLA 2 -IIA mRNA, Cell-secreted and Cell-associated sPLA 2 -IIA during in Vitro Differentiation of HASMC-Analysis of mRNA content by semiquantitative RT-PCR showed that subconfluent proliferating cells cultured in growing medium (3-day 10% FBS) and confluent cells after 1 day in defined serum-free medium (1-day 0% FBS), exhibited a low constitutive expression of sPLA 2 -IIA (Fig. 4A). When these cells were switched to defined serum-free medium, we observed a time-dependent increase in the mRNA content of sPLA 2 -IIA after 3 days that remained constant up to 14 days in culture. The results of sPLA 2 -IIA protein levels presented in Fig. 4, B and C, indicate that similar to the mRNA levels, culture of confluent cells in defined serum-free medium induced a timedependent increase in the cell-secreted and cell-associated protein content of sPLA 2 -IIA. Despite an increase in sPLA 2 -IIA mRNA levels no sPLA 2 -IIA protein could be detected in cell media or associated to the cells after 3 days in serum-free media (3-day 0%). A significant increase in sPLA 2 -IIA cellassociated protein was observed after 7 days (2-2.5-fold) that continued up to 14 days of culture in serum-free medium, where after, sPLA 2 -IIA protein levels started to decrease. The amount and period of sPLA 2 -IIA secretion varied between different batches of HASMC. In some cultures, secretion was highest after 7 days in serum-free medium. To further investigate if this increase in sPLA 2 -IIA mRNA and cell-associated protein levels in culture was due to the absence of serum, cells expressing sPLA 2 -IIA after 5 or 7 days in defined serum-free medium were switched to 10% FBS growing medium. In these cells sPLA 2 -IIA mRNA, secreted and cell-associated sPLA 2 -IIA protein levels decreased (Fig. 4, A, B, and C). In contrast, an increase in the sPLA 2 -IIA mRNA and cell-associated protein levels were observed when the same cells were switched back to defined serum-free medium (data not shown). These results together with the results presented in Figs. 2 and 3 indicate that sPLA 2 -IIA mRNA and sPLA 2 -IIA protein levels correlate with the degree of cell proliferation in cultures of HASMC. To search for the mechanisms of sPLA 2 -IIA induction, proliferating HASMC nonexpressing sPLA 2 -IIA and 1-day post-confluent HASMC were incubated for 1 or 3 days with conditioned media from 3 and 7 days post-confluent HASMC. However, these conditioned media from HASMC did not induce or enhance the expression of sPLA 2 -IIA at mRNA or protein levels (data not shown).
Effect of Interferon-␥ on sPLA 2 -IIA Expression-Incubation of 7 days post-confluent HASMC with IFN-␥ for 4 or 24 h stimulated the transcription of sPLA 2 -IIA inducing a dose-dependent (50 -500 units/ml) increase of sPLA 2 -IIA mRNA and also of cell associated and cell secreted sPLA 2 -IIA (Fig. 5, A, B,  and C). IFN-␥ at 500 units/ml, the highest concentration used, increased the mRNA levels significantly but the sPLA 2 -IIA protein levels decreased. Highest induction of sPLA 2 -IIA protein levels was observed with 100 units/ml. The IFN-␥-sPLA 2 -IIA up-regulation remained constant for 48 h (data not shown). The induction of sPLA 2 -IIA expression by IFN-␥ was also observed in postconfluent HASMC after 1 or 3 days culture in serum-free medium (data not shown).
Effect of IFN-␥ on PLA 2 Activity-SPLA 2 IIA activity was measured in the cell media from HASMC incubated with or without 100 units/ml IFN-␥ during 48 h. IFN-␥ significantly increased the amount of active sPLA 2 -IIA secreted in the cell medium (Fig. 6). To investigate if secreted sPLA 2 -IIA could be found associated to extracellular membrane proteoglycans, HASMC were pretreated with chondroitinase ABC (0.1 units/ ml) and heparitinase I (0.01 units/ml) for 2 h at 37°C before collecting the cell media. As shown in Fig. 6, the degradation of cell surface proteoglycans increased the amount of sPLA 2 -IIA released in the cell medium. This indicated that part of the sPLA 2 -IIA secreted by HASMC is associated with cell membrane proteoglycans.
Modulation of sPLA 2 -IIA Expression by Different Cytokines- Fig. 7 shows the results obtained when studying the effect of different cytokines on the sPLA 2 -IIA mRNA level (I) as measured by semiquantitative RT-PCR and the amount of sPLA 2 -IIA protein secreted into the cell media (II). The RT-PCR results (Fig. 7I)  nizing effect was also observed when TNF-␣ and IL-10 were added simultaneously (P). This effect became stronger at secretion level (Fig. 7II, P). Thus, suggesting a possible synergistic effect between these cytokines.
The amount of cell-secreted and cell-associated sPLA 2 -IIA protein after 4 and 24 h incubation with cytokines were measured in parallel to the mRNA levels. The levels of cell-secreted sPLA 2 -IIA are presented in Fig. 7II. Results from cell-associated sPLA 2 -IIA protein level are not shown. There were similarities and discrepancies between cell-secreted and cell-associated sPLA 2 -IIA protein compared with the mRNA levels. IFN-␥, as shown in Fig. 5, increased cell-secreted levels of sPLA 2 -IIA. This correlates with the results obtained at the mRNA level. IFN-␥ was also the most potent cytokine increasing sPLA 2 -IIA secretion (Fig. 7II, B) and cell-associated levels after 24 h incubation. IL-1␤ added together with IFN-␥ stimulated secretion of sPLA 2 -IIA after 24 h incubation (D). A decreased level of cell-associated sPLA 2 -IIA protein (data not shown) accompanied this. IL-1␤ ϩ IFN-␥ added together decreased the levels of mRNA message (Fig. 7I, D). These results together suggest that IL-1␤ regulate mainly secretion of sPLA 2 -IIA in HASMC. TNF-␣ and IL-10 did not affect IFN-␥induced sPLA 2 -IIA secretion (Fig. 7II, E and F, respectively). IL-6 (H) and TNF-␣ (I) added individually only induced a transient increase in secretion of sPLA 2 -IIA lasting 4 h. The highest secretion of sPLA 2 -IIA was obtained with TNF-␣ after 4 h incubation. After 24 h incubation the levels of TNF-␣induced sPLA 2 -IIA secretion was lower than in the control cells. In addition, no changes in the levels of cell-associated sPLA 2 -IIA were observed (data no shown). These results sug-gest that TNF-␣ appears to regulate sPLA 2 -IIA translation and secretion. This TNF-␣ stimulatory effect on sPLA 2 -IIA secretion was antagonized when TNF-␣ was added together with IFN-␥ (E), IL1-␤ (M), or IL-6 (N). The strongest antagonizing effect was observed with IL1-␤ (M).

IL-6 but Not IFN-␥ Induces snpPLA2 Expression and Secretion in HepG2
Cells-We compared the IFN-␥ induced sPLA 2 -IIA expression in HASMC with HepG2 liver cells. IL-6 was reported previously to increase sPLA 2 -IIA mRNA accumulation and sPLA 2 -IIA secretion in HepG2 liver cells (51). Therefore, we incubated HepG2 cells with IL-6 and IFN-␥ individually or combined and after 24 h incubation measured sPLA 2 -IIA mRNA accumulation and sPLA 2 -IIA secretion. In Fig. 8 it can be observed that contrary to the results obtained with

FIG. 3. Expression of cell differentiation markers in HASMC.
A, Western blot analysis of endogenous cell cycle inhibitor p27 and p21 expression by HASMC cultured as schematized in Fig. 1. Proliferating HASMC after 2 days in fetal bovine serum-supplemented medium (FBS 10%, 2d) and non-proliferating HASMC after 1 and 14 days in serumfree medium (0%, 1d and 14d) to stimulate differentiation were used. p21 and p27 were detected with monoclonal antibodies. Recombinant p21 and p27 were used as positive standard (st). B, increase in mRNA for sPLA 2 -IIA, high molecular weight caldesmon and GAPDH during culture of HASMC. Cells were harvested at different times during culture in serum-free medium as described in the legend to HASMC, HepG2 cells did not respond to IFN-␥. In addition, IL-6 was more potent stimulating sPLA 2 -IIA mRNA accumulation (Fig. 8) in HepG2 liver cells than in HASMC.

SPLA 2 -IIA Up-regulation and NF-B and STAT Activity in Human Arterial Smooth Muscle Cells Induced by Cytokines-
EMSA were performed to explore if the nuclear transcription factor NF-B and the family of STAT transcription factors were involved in the regulation of sPLA 2 -IIA transcription. Nuclear extracts were prepared from HASMC incubated with or without cytokines incubated for 4 and 24 h as described above in the experiments shown in Fig. 7. Two different sequences, homologous to STAT and NF-B-binding sites, were chosen from the sPLA 2 -IIA promoter. Nuclear extracts from HASMC stimulated with IL-1␤ and TNF-␣ for 4 or 24 h resulted as expected in one major complex with NF-B-labeled oligonucleotide (Fig.  9). NF-B activation was observed whether cells were incubated with IL-1␤ or TNF-␣ individually or in combination with other cytokines. For identification of the protein-DNA complex we also used an antibody against the p65 subunit of NF-B (data not shown). Stimulation of HASMC with IFN-␥ for 4 or 24 h resulted in increased binding of STAT proteins to the DNA oligo. STAT activation was observed in all incubations whether IFN-␥ was added individually or in combination with other cytokines (Fig. 9). Unstimulated (controls) HASMC showed no NF-B or STAT complex formation with labeled DNA oligonucleotides. Thus indicating the absence of activation of NF-B or STAT by lipopolysaccharide contaminants and that the cytokines used were functionally active.
There are several known mammalian STAT proteins. In order to identify the STAT proteins involved in IFN-␥ stimulation of sPLA 2 -IIA expression we used antibodies against STAT1, STAT2, STAT3, and STAT5. Fig. 10A shows that only the antibody against STAT3 protein blocked the binding of STAT3 to the labeled oligo. Non-immune rabbit IgG did not affect STAT-DNA complex formation. As phosphorylation of STAT3 at Tyr 705 is essential for dimerization and DNA binding, phosphorylation at this site is a marker of STAT3 activity. Fig. 10B show the presence of active phosphorylated STAT 3 protein in an immunoblot analysis of cell extracts from HASMC after incubation with IFN-␥ for 15 and 35 min. In control cells the STAT3 band is unphosphorylated. These results together indicate that in HASMC the up-regulation of sPLA 2 -IIA expression by IFN-␥ involved activation of STAT3 nuclear transcription factor but not NF-B activation. DISCUSSION We believe that this is the first study investigating the control of the sPLA 2 -IIA expression at mRNA and protein level by human arterial smooth muscle cells in culture. We found that the expression of sPLA 2 -IIA by HASMC is induced once cells are exposed to non-proliferating culture conditions that promote cell differentiation. Taking advantage of continual culture of HASMC in serum-free medium, we demonstrated that sPLA 2 -IIA mRNA, cell-associated and cell-secreted protein levels are found in non-proliferating, quiescent HASMC, expressing cyclin-dependent kinase inhibitor p27 (Fig. 3). This is a marker of cell differentiation found in vascular smooth  6. SPLA 2 -IIA enzymatic activity in cell medium and associated to cell surface proteoglycans after 24 h incubation with IFN-␥. HASMC were cultured 7 days in serum-free medium following the in vitro protocol to induce sPLA 2 -IIA expression shown in Fig. 1. Medium was removed and cells were then incubated for 48 h with or without (Control) IFN-␥ in serum-free medium in duplicate sets. Before collecting the cell medium, one set of Control and IFN-␥treated cells was treated with chondroitinase ABC plus heparitinase I. Cell media were collected and assayed for PLA 2 activity using [ 14 C]phosphatidylcholine-liposomes. muscle cells in vitro and in vivo (35,36). This was supported by the presence of heavy caldesmon mRNA a specific marker for HASMC differentiation (37). No large differences in the total levels of p21 were found (Fig. 3), probably because its inhibitory action is achieved by changes in the ratio of p21 free to cyclin-Cdk bound complexes without affecting its total intracellular levels (36). The presence of serum, a source of growth factors, induces cell proliferation and de-differentiation suppressing both sPLA 2 -IIA mRNA and protein expression (Fig. 4). The results indicate that sPLA 2 -IIA expression by HASMC in culture requires conditions that lead to a phenotypic change from proliferating, de-differentiated cells, toward quiescent (nonproliferating) cells (30). Pulmonary artery smooth muscle cells showed similar behavior. These results agree with previous data reported by Anderson and co-workers (39) showing that the expression of sPLA 2 -IIA by confluent human coronary ar-tery vascular smooth muscle cells is not changed over a 10-day culture period in the presence of serum. Together these results support our previous immunohistochemistry work showing the presence of sPLA 2 -IIA associated with ␣-actin positive, spindleshaped differentiated smooth muscle cells in intima and media of human arteries (17,20). Inflammatory cytokines are reported to up-regulate the expression and secretion of sPLA 2 -IIA in different cell systems (23). Our previous electron microscopy study indicates that human atherosclerotic lesions contain more extracellular sPLA 2 -IIA than adjacent non-atherosclerotic regions in the same coronary artery (20). Although, the exact mechanism responsible for this increase in extracellular sPLA 2 -IIA is not established, one possibility is that proinflammatory cytokines stimulate sPLA 2 -IIA synthesis and secretion by arterial smooth muscle. IFN-␥, IL-1␤, and TNF-␣ are proinflammatory cytokines present in atherosclerotic lesion (40). The results presented here showed that IFN-␥ was the most potent of all cytokines studied in stimulating sPLA 2 -IIA mRNA and protein levels (Fig. 5). IFN-␥ already after 4 h incubation induced a significant increase in mRNA and cellassociated and cell-secreted protein levels of sPLA 2 -IIA in HASMC, this effect was sustained for 48 h. The secreted sPLA 2 -IIA was catalytically active as shown in Fig. 6. Furthermore, treatment of HASMC with enzymes that degrade glycosaminoglycans increased the amount of active sPLA 2 -IIA released into the cell medium supporting our previous suggestions that pericellular glycosaminoglycans are an important compartment of active enzyme. The role of IFN-␥ in stimulation of sPLA 2 -IIA expression was supported by EMSA and immunoblot analysis showing the presence of activated STAT3, a characteristic transcription factor of the IFN-␥ intracellular signaling pathway, in HASMC (41). IFN-␥ is a potent inhibitor of HASMC proliferation (42). Our results indicate that expression of sPLA 2 -IIA in HASMC is stimulated under nonproliferating conditions. Therefore, one may speculate that up-regulation sPLA 2 -IIA expression by IFN-␥ may also be a cellular Arterial smooth muscle cells in vitro and in vivo respond markedly to IFN-␥ by expressing class II major histocompati-bility antigens such as HLA-DR (42,43). These genes are up-regulated in smooth muscle cells in human atherosclerotic plaques, probably induced by secretion of IFN-␥ by a subset of T cells present in the arterial wall (44). In addition, extracellular sPLA 2 -IIA was reported to increase T-lymphocyte response (45). Taken together, these results suggest that IFN-␥ signaling may promote atherogenesis by stimulating sPLA 2 -IIA expression and creating a positive feedback mechanism, sustaining chronic inflammation at places of lipid deposition in the arterial wall.
Several studies previously reported that IL-1␤ and TNF-␣ induce sPLA 2 -IIA expression and secretion over a long period in different animal cell systems, including rat vascular smooth muscle cells (46 -48). Our results show that incubation of HASMC with IL-1␤ induced a moderate increase of sPLA 2 -IIA mRNA and sPLA 2 -IIA protein secretion that was stable during 24 h of incubation. However, TNF-␣ induced a strong and transit sPLA 2 -IIA mRNA accumulation and sPLA 2 -IIA secretion by HASMC after 4 h incubation, decreasing to the levels observed in nonstimulated cells (control) in the case of mRNA, and below those control levels in the case of the amount of secreted protein, after 24 h incubation (Fig. 7, I and II). TNF-␣ was the most potent of all cytokines studied in stimulating secretion of sPLA 2 -IIA with cytokines after 4 h incubation. Unexpectedly and interestingly, incubation of HASMC for 4 h with the combination of IL-1␤ and TNF-␣, decreased the secretion of sPLA 2 -IIA bellow the control levels, indicating that these cytokines antagonized each other when present together. On the other hand, EMSA results showed the presence of active NF-B after 4 and 24 h incubation with these cytokines (Fig. 9, EMSA results after 24 h). Therefore, contrary to what is reported with rat aortic smooth muscle cells, binding of NF-B transcription factors to recognition elements in the promoter region of human sPLA 2 -IIA do not necessarily induce transcription of sPLA 2 -IIA in HASMC (48). However, the mechanism for this difference in response to IL-1␤ between human aortic cells and rat vascular cells remains to be clarified.
Another interesting cytokine that we studied is IL-6. This is a multifunctional cytokine produced at local tissue sites in almost all situations of homeostatic perturbation (49). Cells present in vascular wall such as monocyte-derived macrophages, T cells, endothelial cells, fibroblast and smooth muscle cells, produce IL-6, usually as a response to IL-1␤, TNF-␣, or lipopolysaccharide stimulation. In addition, IL-6 is required for the induction of acute phase reaction proteins such as C-reactive protein (50). SPLA 2 -IIA is classified as an acute phase protein and therefore IL-6 may also be responsible for its synthesis by the liver (51). However, whether IL-6 has a pro-or anti-inflammatory function in local or systemic responses remains to be established (50,52). Previously, IL-6 has been shown to stimulate gene expression and synthesis of sPLA 2 -IIA in HepG2 liver hepatoma cells (51). In the present work we compared the effect of IL-6 in the expression and secretion of sPLA 2 -IIA by HepG2 cells and HASMC. Our results showed that incubation of HASMC for 4 h with IL-6 induced a significant increase in secretion of sPLA 2 -IIA accompanied by an increase in mRNA accumulation after 4 h incubation (Fig. 7). This stimulation was transient and it was also lower than the secretion induced by TNF-␣. In combination with other cytokines, IL-6 antagonized sPLA 2 -IIA secretion induced by IFN-␥, IL-␤, and TNF-␣ in HASMC. Interestingly, experiments with HepG2 cells (Fig. 8) indicated that these cells responded to IL-6 but not to IFN-␥ by increasing both sPLA 2 -IIA mRNA accumulation and sPLA 2 -IIA secretion after 24 h of incubation. These results suggest that sPLA 2 -IIA induction by IFN-␥ and IL-6 cytokines are cell specific. In addition, these experiments suggest that the pro-or anti-inflammatory properties of IL-6 may be modulated by the presence of other proinflammatory cytokines. Our results support previous data by Crowl and coworkers (51) suggesting that the main source of sPLA 2 -IIA in plasma may be hepatic cells stimulated by IL-6. However, IL-6 or IFN-␥ stimulated vascular smooth muscle cells may also contribute to increase plasma sPLA 2 -IIA concentration by transit leakage from the vascular wall that may take place during acute inflammatory conditions. Circulating levels of IL-6, C-reactive protein, and sPLA 2 -IIA are increased in patients with cardiovascular disease and are associated with poor prognostic outcome (13,53). However, the mechanism for this association is unknown. Increased levels of circulating IL-6 is reported to exacerbate early atherosclerosis (54). Furthermore, low density lipoprotein modification by PLA 2 activity in total plasma induces atherogenic small, dense low density lipoprotein particles with high affinity for arterial proteoglycans (55). Taken together, these results suggest that IL-6 may contribute to atherosclerosis by stimulating hepatic cells and arterial smooth muscle cells to secrete sPLA 2 -IIA and as a consequence increase the extracellular activity of sPLA 2 -IIA in plasma and arterial wall.
There is ample support for the participation of proinflammatory cytokines in atherosclerosis. However, the potential contribution of anti-inflammatory cytokines in the modulation of atherogenesis is less documented. IL-10 is an anti-inflammatory cytokine, expressed in atherosclerotic lesions and is reported to be an anti-atherosclerotic cytokine (56). Our results shown that IL-10 by itself did not affect the expression or secretion of sPLA 2 -IIA. However, IL-10 was able to block IFN-␥-induced sPLA 2 -IIA expression and secretion (Fig. 7). IL-10 markedly inhibited IFN induced sPLA 2 -IIA transcription despite a strong STAT3 activation observed with EMSA (Fig. 9). Inhibitor effect of IL-10 on IFN-␥ and TNF-␣ induced gene expression are reported previously in the literature (57). Our results suggest that the anti-inflammatory or protective function of IL-10 in atherosclerosis may be partially due to its capacity to decrease sPLA 2 -IIA expression in the presence of proinflammatory cytokines.
Our results showed that sPLA 2 -IIA binds to cell surface proteoglycans (Fig. 6). It has been proposed that this binding is important in the hydrolysis of cell phospholipids for the gener-ation of arachidonic acid (58). While this hypothesis has been challenged (59), results from our group indicate that sPLA 2 -IIA activity can be modulated by its interaction with different types of glycosaminoglycans (60). Cytokines are reported to differentially influence synthesis and composition of proteoglycans in different cell systems (38). Therefore, one may expect that the levels of sPLA 2 -IIA secretion after cytokine stimulation may be affected by the direct effect on sPLA 2 -IIA secretion and changes in cell membrane proteoglycan composition influencing binding of sPLA 2 -IIA.
In summary, the results presented here showed that differentiated HASMC constitutively express sPLA 2 -IIA. SPLA 2 is mainly cell associated in unstimulated cells. Exposure to proinflammatory cytokines can stimulate sPLA 2 -IIA transcription and secretion of intracellular storage or de novo synthesized sPLA 2 -IIA in a transient or sustained form. This may have physiological relevance for the triggering of an acute or chronic inflammatory response in the arterial wall. Interestingly, a recent report shows the simultaneous presence of mRNA transcripts for sPLA2-IIA, IFN-␥, IL-1␤, and TNF-␣ in human atherosclerotic lesion (25). However, further work is required to elucidate the molecular regulatory mechanisms controlling transcription, translation, and secretion of sPLA 2 -IIA by HASMC. These results suggest that changes in the balance of different pro-and anti-inflammatory cytokines present in the arterial wall may regulate the levels of extracellular sPLA 2 -IIA and its potential contribution to atherogenesis.