Atherosclerosis 142 (1999) 47–56

Sublytic C5b-9 induces proliferation of human aortic smooth musclecells

Role of mitogen activated protein kinase and phosphatidylinositol3-kinase

Florin Niculescu a, Tudor Badea a, Horea Rus a,b,*a Department of Pathology, Uni6ersity of Maryland, School of Medicine, 10 South Pine Street, Baltimore, MD 21201, USA

b Medical Clinic no. 1, 3400 Cluj-Napoca, Romania

Received 11 February 1998; received in revised form 9 June 1998; accepted 26 June 1998

Abstract

Proliferation of vascular smooth muscle cells contributes to initimal hyperplasia during atherogenesis, but the factors regulatingtheir proliferation are not well known. In the present study we report that sublytic C5b-9 assembly induced proliferation ofdifferentiated human aortic smooth muscle cells (ASMC) in culture. Cell cycle re-entry occurred through activation of cdk4, cdk2kinase and the reduction of p21 cell cycle inhibitor. We also investigated if C5b-9 cell cycle induction is mediated throughactivation of mitogen activated protein kinase (MAPK) pathways. Extracellular signal regulated kinase (ERK) 1 activity wassignificantly increased, while c-jun NH2-terminal kinase (JNK) 1 and p38 MAPK activity were only transiently increased.Pretreatment with wortmannin inhibits ERK1 activation by C5b-9, suggesting the involvement of phosphatidylinositol 3-kinase(PI 3-kinase). Both PI 3-kinase and p70 S6 kinase were activated by C5b-9 but not by C5b6. C5b-9 induced DNA synthesis wasabolished by pretreatment with inhibitors of ERK1 and PI 3-kinase, but not by p38 MAPK. These data indicated that ERK1 andPI 3-kinase play a major role in C5b-9 induced ASMC proliferation. © 1999 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Aortic smooth muscle cells; Cell cycle; Terminal complement complexes; Mitogen activated protein kinase; Phos-phatidylinositol 3-kinase

1. Introduction

The smooth muscle cells (SMC) are fully differenti-ated cells which proliferate at a low rate [1,2]. Inresponse to vascular injury, aortic SMC (ASMC) dedif-ferentiate and adopt a proliferative phenotype [1,3].The proliferation of SMC plays an important role inatherogenesis, angiogenesis in solid tumors and hyper-tension [3–5]. Proliferation in the atherosclerotic le-sions occurs in the diffuse intimal thickening and in theinnermost part of the underlying media. The proliferat-ing SMC synthesize and degrade the extracellular ma-

trix [3,6].Cell proliferation is accomplished through the activa-

tion of cyclin dependent kinases (cdk), which regulatekey transition check points in the cell cycle. Progressionof cell cycle through G1 into the S-phase requires thecoordinated activation of cyclin D/cdk4,6 and cyclinE/cdk2 complexes [7]. Activation of these complexes isalso regulated by cell cycle kinase inhibitors throughphysically interaction with cdk. The p21 (also known asWAF1, Cip1, Sdi1), an universal cell cycle kinase in-hibitor which inhibits cyclin/kinase catalytic activity,may induce cell cycle arrest in response to activation ofp53 checkpoint pathway [8,9].

Activation of inflammatory process induced by bothhumoral and cellular immune effectors have been impli-

* Corresponding author. Tel.: +1-410-706-7892; fax: +1-410-706-7706.

0021-9150/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved.PII: S 0 0 2 1 -9150 (98 )00185 -3

F. Niculescu et al. / Atherosclerosis 142 (1999) 47–5648

cated in the pathogenesis of atherosclerosis [10]. Amongthe humoral factors involved in atherogenesis, activa-tion of the complement system has been associated withthe prelesional stages as well as with the progression ofatherosclerotic lesions [11–13]. The complement systemcan be activated in vitro by cholesterol containing lipidsderived from atherosclerotic lesions as well as by enzy-matically modified low density lipoproteins [14,15]. TheC5b-9 membrane attack complex of complement hasbeen localized in atherosclerotic lesions in associationwith cell debris, lipid droplets, and cholesterol clefts[16]. In addition, C5b-9 has been found on intactmacrophages and SMC [17].

Activation of complement proteins C5-C9 generatesmembrane-inserted heteropolymeric complexes includ-ing C5b-7, C5b-8 and C5b-9. These complexes arecollectively named terminal complement complexes(TCC) whereas the membrane attack complex repre-sents only the C5b-9 [18]. Sublytic TCC assembly onthe plasma membranes of nucleated cells induces acti-vation of biologically important metabolic events. TCCactivate membrane phospholipases with the subsequentincrease in the mass-level of diacylglycerol and ce-ramide [19] TCC also physically associate and activateheterotrimeric G proteins of Gi/Go family [20]. TheG-protein activation induced by TCC is responsible forthe activation of Ras/Raf-1/extracellular signal regu-lated kinase (ERK) mitogen-activated protein kinase(MAPK) pathway through Gbg subunits [21].

In this study we investigated whether sublytic TCCinduce proliferation of ASMC. In addition, the signaltransduction pathway involved in cell cycle activationby C5b-9 was explored. We show that ASMC prolifera-tion is induced by C5b-9 and this is associated with theactivation of cdk4, cdk2 and down-regulation of thep21 cell cycle inhibitor. The cell cycle activation in-duced by C5b-9 required activation of ERK1 and phos-phatidylinositol 3-kinase (PI 3-kinase) but not p38MAPK or c-jun NH2-terminal kinase (JNK1).

2. Materials and methods

2.1. Chemicals and antibodies

PD 098,059 inhibitor was from Research Biochemi-cals International (Natick, MA) and SB 202190 and SB203580 were from Calbiochem (La Jola, CA). Wort-mannin was from Biomol (Plymouth Meeting, PA) andPertussis toxin (PTX) was from List Biologic Laborato-ries (Campbell, CA). Nitrocellulose membranes werefrom Millipore (Bedford, MA), P81 phosphocellulosepaper and GF/A glass fiber filters were from Whatman(Maidstone, UK). BCA reagents were from Pierce(Rockford, IL) and protein A/G-agarose was fromOncogene Science. (Cambridge, MA). TLC silicagel 60

plates were from Merck (EM Industries, Gibson, NJ).[3H]thymidine, [a-32P]dCTP and [g-32P]ATP were fromDuPont-NEN (Boston, MA). Rabbit polyclonal IgG tohuman Raf-1, ERK1, JNK1, p38 MAPK, p85 PI 3-ki-nase and cdk4 were from Santa Cruz Biotechnology(Santa Cruz, CA). Rabbit polyclonal IgG to humanp70 S6 kinase, cdk2, p70 S6 kinase specific substrate,myelin basic protein 95–98 peptide were from UpstateBiotechnology (Lake Placid, NY) while Syntide 2 andATF2 were from Santa Cruz Biotechnol. RPMI 1640and fetal bovine serum were from Gibco BRL(Gaithersberg, MD). Smooth muscle cell basal mediumand the growth supplements was from Clonetics (Walk-ersville, MD).

2.2. Culture of primary ASMC in 6itro

Human ASMC from Clonetics were cultured insmooth muscle cell basal medium containing 5% fetalbovine serum, 10 ng/ml human epidermal growth fac-tor, 2 ng/ml human fibroblast growth factor and 5mg/ml insulin for 3 to 4 days. After three to fivepassages, cells were starved for 24 h in medium free ofserum and growth supplements to obtain cells in G0/G1 phase. Over 95% of cells were in G0/G1 phase, asdetermined by the relative DNA content analyzed bypropidium iodide staining and FACS analysis [25]. Allcells stained positive for muscle actin when usingHHF35 monoclonal antibody (Enzo Diagnostics,Farmingdale, NY) and the peroxidase conjugated goatanti-mouse IgG (Jackson Immunoresearch Laboratory,West Grove, PA).

2.3. Acti6ation of serum complement and TCCassembly

Pooled normal human serum (NHS) from healthyadult donors was used as a source of serum comple-ment. The hemolytic activity of complement was inacti-vated by heating NHS at 56°C for 45 min (HI-NHS) orby treating the NHS with K76 (K76-NHS), a comple-ment C5 inhibitor [22]. NHS immunochemically de-pleted of C7 (C7D), C8 (C8D) or C9 (C9D) as well asthe purified human C7, C8 and C9 proteins were fromQuidel (San Diego, CA). C5b6 complexes were assem-bled and purified as described [19,23]. To study thesublytic complement effects and the role of TCC,ASMC were sensitized with anti-HLA class A,B,Cmonoclonal antibody (Ab) (Dako A/S, Glostrup, Den-mark) then incubated for various time periods withNHS (1/10 dilution), HI-NHS (1/10) or C7D (1/5)9C7(10 mg/ml). Sublytic concentrations of anti-HLA Abwere previously determined by dose response experi-ments using NHS, and C7D reconstituted with C7 bymeasuring the released cytoplasmic lactate dehydroge-nase as an indication of cell lysis [24].

F. Niculescu et al. / Atherosclerosis 142 (1999) 47–56 49

2.4. Cell proliferation assay

The number of cells was determined using Cell Titer96 Non-radioactive cell proliferation assay kit(Promega, Madison, WI). ASMC deprived of growthsupplements for 24 h were exposed to the variousstimuli for 48 h. The cells were then incubated for 4 hat 37°C with MTS/PMS solution and the absorbance at490 nm was recorded using an ELISA plate reader. Thecell number was determined using a standard curve ofthe colorimetric reaction.

2.5. Immunoprecipitation and cyclin-dependent kinaseassay

Ab-sensitized ASMC were treated with NHS, orK76-NHS, then lysed with RIPA buffer [25]. Identicalamounts of protein from each sample were incubatedwith 10 mg of rabbit IgG anti-cdk2 or anti-cdk4 and 10ml protein-A/G agarose at 4°C overnight. For the assayof cdk2 kinase activity the immunoprecipitated proteinswere incubated for 30 min at 37°C with 2 mg histone H1(Boehringer Mannheim, Indianapolis, IN) and 1 mCi[g32P]-ATP/sample in an equal volume of 2× reactionbuffer (25 mM Tris–HCl, pH 7.4, 10 mM MgCl2).Cdk4 activity was determined by using a recombinanttruncated form of Rb protein (p56 Rb) as a substrate(QED, San Diego, CA). The reaction was carried outfor 30 min in the reaction buffer containing 400 ng ofthe substrate p56Rb/sample and 10 mCi [g32P]-ATP[25]. Phosphorylated histone H1 and Rb protein bandswere analyzed by 12% SDS-PAGE andautoradiography.

2.6. Determination of p21 protein

p21 Protein in ASMC cell lysates was determinedusing WAF1 (also named p21 or CIP 1) ELISA kitfrom Oncogene (Cambridge, MA). The cell lysates ob-tained from unstimulated cells or cells exposed to Ab+NHS or K76-NHS were diluted 1/4 in the provideddilution buffer. The ELISA was performed accordingto manufacturer’s instruction and a standard curve wasobtained using six standard dilutions. The results wereexpressed in U/mg protein.

2.7. Raf-1 kinase assay

Following assembly of TCC, the cells were lysed inlysis buffer as described [21]. From each cell lysate, 100mg of protein Raf-1 was immunoprecipitated by incu-bating at 4°C overnight with 10 mg of rabbit IgG tohuman Raf-1, and 10 ml protein A/G-agarose/sample.Beads were suspended in 50 ml of 20 mM PIPESreaction buffer pH 7.0, containing 10 mM MnCl2,

1ml/sample [g-32P]ATP (10 mCi) and 2 mg/sample of

Syntide 2. Following incubation for 30 min at 30°C,Syntide 2 phosphorylation was determined by loading10 ml each of the reaction supernatant on P81 phospho-cellulose paper then counting the radioactivity [26,27].

2.8. ERK1 kinase assay

Cell lysates were immunoprecipitated with 10 mgrabbit IgG to human ERK-1 and 10 ml each of proteinA/G-agarose/sample, as described for Raf-1. Theagarose beads were suspended in 50 ml reaction buffercontaining 1 ml/sample of 10 mCi [g-32P]ATP and 10ml/sample myelin basic protein 95–98 peptide [26] for30 min at 30°C. Peptide phosphorylation was assessedby loading 10 ml each of the reaction supernatant onP81 phosphocellulose paper, followed by counting theradioactivity.

2.9. PI 3-kinase assay

Cell lysates were immunoprecipitated overnight withanti-p85 PI 3-kinase polyclonal IgG in the presence ofprotein A/G agarose. The beads were washed with lysisand reaction buffer (20 mM Tris–HCl pH 7.5, 100 mMEDTA, 0.5 mM EGTA, 1 mM ATP, 10 mM MgCl2),then the kinase reaction was developed in reactionbuffer containing 40 mg/per sample phosphatidylinosi-tol (PI) and [g-32P]-ATP, 1 mCi per sample for 30 min at37°C. The reaction was stoped by the addition of 4 MHCl and chloroform/methanol 1:1. Samples were spot-ted on TLC plates, developed and the PI phosphoryla-tion was quantitated.

2.10. JNK1 and p38 MAPK assay

Identical amounts of protein from each lysate wereincubated at 4°C with 10 ml each of anti-JNK1 IgG oranti-p38 MAPK IgG and protein A/G-agarose. Afterwashing with the lysis buffer, JNK1 activity was deter-mined by incubation for 20 min at 30°C with 1 mg ofGST-c-jun1-79 (Biomol Research Laboratory, Ply-mouth Meeting, PA) in the reaction buffer containing[g-32P]-ATP, 1 mCi per sample [28]. p38 MAPK activitywas determined by incubation of immunoprecipitatedproteins for 30 min at 30°C with 50 ml reaction buffer(12.5 mM MOPS, pH 7.5, 12.5 mM b-glycerophos-phate, 7.5 mM MgCl2, 0.5 mM EGTA), containing 1mg of ATF2 and 1 mCi [g-32P]-ATP per sample. Sampleswere then analyzed by 12% SDS-PAGE andautoradiography.

2.11. p70 S6 kinase assay

Cell lysates were immunoprecipitated with rabbit IgGto human p70 S6 kinase and protein A/G-agarose. Thekinase activity associated with the precipitated immune

F. Niculescu et al. / Atherosclerosis 142 (1999) 47–5650

complexes was assayed using a specific substrate pep-tide (AKRRRLSSLRA) and [g-32P]-ATP, 1 mCi/sam-ple. A sample containing 5 ml of supernatant wasspotted on P81 phosphocellulose paper and the ra-dioactivity counted.

2.12. Assay for DNA synthesis

Sensitized ASMC exposed to various stimuli wereincubated with 1 mCi/sample [3H]thymidine in basalmedium without growth supplements for 18 h at37°C. Cells were washed in ice-cold Tris buffer saline,lysed in 0.3 M NaOH, then precipitated with 20%trichloracetic acid and filtered through Whatman GF/A glass fiber filters. The radioactivity on dried filterswas determined by liquid scintillation counting. Theeffect of PD098,059, a specific MEK1 inhibitor [29]was examined by preincubation of ASMC for 30 minwith 100 mM of PD098,059 prior to C5b-9 assembly.Similar experiments were performed with cells prein-cubated with wortmannin (100 nM) as PI 3-kinaseinhibitor and with the p38 MAPK inhibitors SB202190 and SB 203580 (each of 25 mM) [30].

2.13. Isolation of RNA and Northern blot analysis

In brief, total RNA was purified from ASMC usingguanidine isothiocyanate and by ultra-centrifugationthrough 5.7 M CsCl2 cushion for 18 h at 35 000 rpmusing a SW 60 Beckman rotor [25,31]. RNA was de-natured and electrophoresed on 0.8% agarose-formaldehyde gel, then transferred to nitrocellulosemembrane. The membrane was hybridized with 32P-labeled specific cDNA probes [32]. The membranewas exposed to X-ray film, then radiographic densi-ties of each mRNA band were scanned on the Com-puting Densitometer (Molecular Dynamics,Sunnyvale, CA) and the integrated volume was calcu-lated using the Imagequant software (Molecular Dy-namics).

2.14. Preparation of cDNA probes

The p21 cDNA was from Dr David Beach, ColdSpring Harbor Laboratory (Cold Spring, NY). A 1.4kb Hind III fragment of b-actin cDNA was used ascontrol. The cDNA fragments were labeled with [a-32P] dCTP using Oligolabeling kit (Pharmacia, Piscat-away, NJ).

3. Results

3.1. Effect of C5b-9 on cell proliferation

To test if C5b-9 is mitogenic for ASMC the cell

Fig. 1. Effect of complement activation on ASMC proliferation.ASMC exposed to various stimuli were allowed to grow for 48 hbefore cell counting using a non-radioactive cell proliferation kit. Theincrease attributed to C5b-7, C5b-8 and C5b-9 was tested usingASMC sensitized with Ab and C8D (C5b-8), C9D(C5b-8) or C9D+C9 (C5b-9) and compared with Ab+C7D (C5b6) and serum freemedium (SFM). The increase in cell number induced by C5b-9 was2-fold over SFM level (!PB0.001). C5b-8 induced a small increase(1.2-fold) while C5b6 and C5b-7 had no effect. Data are mean9S.E.M. from three separate experiments.

number was determined after stimulation. ASMCwere allowed to grow for 48 h in the presence ofvarious stimuli before counting the viable cells. Theincrease attributed to C5b-7, C5b-8 and C5b-9 wasdetermined using sensitized ASMC exposed to C8D,C9D or C9D+C9, respectively. Cells treated withC5b6 (Ab+C7D) and in basal medium withoutgrowth supplements (SFM) were used as negativecontrols. C5b-9 induced 2-fold increase in cell numberover the level of cells in SFM or C5b6 treated (Fig.1). C5b6 and C5b-7 had no effect while C5b-8 in-duced only an increase of 1.2-fold (Fig. 1).

3.2. Effect of complement acti6ation on cell cycledependent kinases

To elucidate the mechanisms involved in cell cyclere-entry induced by TCC, the activity of cdk2 and

Fig. 2. Activation of cdk4 and cdk2 is induced by TCC. Unstimulated(lane 1) or Ab-sensitized ASMC treated with NHS (lanes, 2,4,6) orK76-NHS (lanes 3,5,7) for 3, 6 and 18 h were lysed then immunopre-cipitated with anti-cdk4 IgG. The cdk4 kinase activity was testedusing the recombinant truncated form of Rb protein (p56Rb) as asubstrate (cdk4). From the same lysate cdk2 was immunoprecipitatedwith anti-cdk2 IgG. The cdk2 activity of the IP was determined usinghistone H1 as a substrate (cdk2).

F. Niculescu et al. / Atherosclerosis 142 (1999) 47–56 51

Fig. 3. Effect of complement on p21 mRNA expression and protein. (A) Sensitized ASMC were treated with NHS up to 18 h. Total RNA wasextracted as described in Section 2. p21 and b-Actin mRNA expression was determined by Northern blot analysis (10 mg total RNA/lane, upperpanel). The radiographic densities of p21 mRNA were quantified and normalized by the densities of b-actin (lower panel). The p21 mRNAexpression was increased within 60 min after complement treatment then decreased at 3, 6 and 18 h to levels lower than that of quiescent cells.Unstim=unstimulated cells. (B) p21 Protein levels were determined by ELISA in ASMC stimulated with Ab and NHS(square containing centredot) and compared with unstimulated cells ("). The p21 levels increased during the first 3 h of treatment with complement then decreased at 18h to levels lower than the unstimulated one. p21 Protein in cells treated with Ab+K76-NHS induced a sustained protein increase at 3 and 6 h.At 18 h p21 level in K76-NHS treated cells was similar to that in quiescent cells.

cdk4 were determined. Fig. 2 shows that cdk4 activitywas low in quiescent cells. Activation of complementincreased the kinase activity by 2-fold as early as 3 h.The activity reached a maximum of 3-fold increase at 6h and returned to the basal level at 18 h. A 3-foldincrease in the cdk2 activity induced by antibody andNHS was noted only at 18 h (Fig. 2) when 21% ofASMC were in S-phase (data not shown). The numberof cells in S-phase was determined by the analysis ofDNA content using the CycleTEST Plus kit from Bec-ton Dickinson (San Jose, CA) as previously described[25]. To determine whether NHS mediated increase incdk4 and cdk2 activity was due to complement activa-tion and required activation of C5, K76-NHS which isdevoid of C5 activity was used in similar experiments.Fig. 2 revealed that K76 abolished NHS effect on cdk2activity and induced a modest and transient increase ofcdk4 at 3 h. These results suggest that C5b-9 assemblyis required for both cdk4 and cdk2 activation.

3.3. Effect of complement acti6ation on p21 inhibitor

Since p21 regulates cell cycle progression in G1 phasewe investigated whether cdk activity induced by C5b-9may be associated with changes in p21 mRNA andprotein. p21 mRNA was present in quiescent cells andincreased up to 1 h following complement activation.p21 mRNA expression progressively decreased, fallingat 18 h during S-phase, under the level seen in quiescentcells (Fig. 3A). The p21 protein level showed a similarpattern as the mRNA expression (Fig. 3B). As shownin Fig. 3B, K76 abolished the NHS induced p21 down-regulation, but induced a sustained increase at both 3and 6 h. These data indicate that down-regulation ofp21 is due to complement activation and also requiredthe activation of C5 the first component of TCC. Onthe other hand, this suggests that C5b-9 induced p21reduction is required for activation of cdk4 and cdk2and S-phase entry.

F. Niculescu et al. / Atherosclerosis 142 (1999) 47–5652

3.4. Acti6ation of Raf-1 by C5b-9

ASMC exposed to C5b-9, but not to C5b6, showedan increase in Raf-1 kinase activity, as assessed byphosphorylation of Syntide 2 (Fig. 4). This increasereached the maximum of 3.6-fold over the C5b6 andbasal level at 10 min (PB0.01). At 30 min, the kinaseactivity returned to the basal levels.

3.5. Acti6ation of ERK1 kinase by C5b-7, C5b-8 andC5b-9

As shown in Fig. 5A, ERK1 activity was increased5.3-fold at 10 min by C5b-9, over the control and C5b6level, respectively (PB0.001). ERK1 activity was alsoincreased by C5b-7 about 2.8-fold and by C5b-8 about3-fold over the C5b6 level. ERK1 activity inducible byC5b-9 was completely abolished by PD098,059, theMEK1 specific inhibitor (Fig. 5B). The involvement ofPI-3 kinase in ERK1 activation was also tested bypretreating the cells with the specific inhibitor, wort-mannin. The increase in ERK1 activity was abolishedby wortmannin (Fig. 5B).

3.6. Acti6ation of PI 3-kinase by C5b-9

Since wortmannin abolished ERK1 the activation ofPI 3-kinase induced by C5b-9 was examined. PI 3-ki-nase was immunoprecipitated from cell lysate afterexposure to C5b-9 then the kinase assay was performedin the presence of PI. C5b-9 induced an increase of PI3-kinase activity of about 4.8-fold over the C5b6 levelat 10 min and this activity was blocked by wortmannin(Fig. 6A). Since TCC activate Raf-1/ERK1 pathway

Fig. 5. ERK1 activation by TCC. ERK1 was determined in anti-ERK1 IP at the indicated time points following C5b-7, C5b-8 andC5b-9 assembly using MBP 95-98 peptide as substrate. (A) ERK1activity in response to C5b-9 assembly increased 5.3-fold at 10 min(! PB0.001) compared with unstimulated cells (CTR) or C5b6.C5b-7 and C5b-8 also increased significantly ERK1 activity (PB0.01). Data represent mean9S.E.M. from four experiments per-formed in quadruplicate. (B) ASMC were pretreated with 100 mMPD098,059 (") or 100 nM wortmannin ( ) for 30 min prior toC5b-9 assembly. Exposure to C5b-9 () generated an increase inERK1 activity similar to that shown in Fig. 2A. Pretreatment withPD098,059 and wortmannin abolished the induction of ERK1 byC5b-9. Unst.=unstimulated cells.

Fig. 4. Raf-1 activation by C5b-9. Raf-1 was immunoprecipitatedfrom ASMC lysate following C5b-9 assembly using purified C5b6,C7, C8, and C9. Raf-1 kinase activity was determined using a specificpeptide substrate. The Raf-1 activity induced by C5b-9 ( ) reachedmaximum 2.3-fold over the C5b6 level (!PB0.01) at 10 min. C5b6() had no effect. The data represent mean9S.E.M. from threedifferent experiments.

through Gi/o at the plasma membrane [20], the effect ofPTX, a Gi-protein inhibitor, on PI 3-kinase activationwas tested. PTX reduced the C5b-9 induced PI 3-kinaseactivity by 70% (Fig. 6B).

3.7. Acti6ation of JNK1 and p38 MAPK by C5b-9

JNK/SAPK (stress activated protein kinases) are agroup of serine/threonine kinases structurally relatedbut distinct from ERK or p38 MAPK. Although JNK/SAPK and p38 induced transcriptional activation ofAP-1 proteins, c-Jun is specifically activated by JNK[33]. JNK1 activity was increased in response to C5b-9but not to C5b6, as determined by phosphorylation ofGST-c-Jun 1-79. This activation was transient andrapidly declined to the basal level at 15 min (Fig. 7).

F. Niculescu et al. / Atherosclerosis 142 (1999) 47–56 53

C5b-9 induced a 1.8-fold increase in p38 MAPK activ-ity within 5 min, which declined to the C5b6 level at 10min (Fig. 7).

3.8. Acti6ation of p70 S6 kinase

Since activation of both ERK1 and p38 MAPKrequired PI 3-kinase activity, we examined the p70 S6kinase activity in response to C5b-9 as a target kinasesituated downstream to PI 3-kinase. C5b-9 induced3.5-fold increase in p70 S6 kinase activity at 5 min overthe C5b6 level and this increase was abolished by both

Fig. 7. Activation of JNK1 and p38 MAPK by C5b-9. ASMC weretreated with C5b6 or C5b-9 assembled from purified components forup to 15 min. Cells were lysed and JNK1 or p38 MAPK activity wasassayed in anti-JNK1 or anti-p38 MAPK IP. GST-NH2-terminalc-Jun protein (GST-c-Jun 79) and ATF2 were used for JNK1 and p38MAPK, respectively. C5b-9 but not C5b6 induced an increase ofJNK1 and p38 MAPK activity. Both kinases had a maximum at 5min. Unstim.=unstimulated cells.

Fig. 6. Activation of PI 3-kinase by C5b-9. (A) ASMC were treatedwith C5b6 or C5b-9 using purified components for up to 15 min.Cells were lysed and PI 3-kinase activity was assayed in anti-PI3-kinase immunoprecipitates by incubation with [g-32P]-ATP andphosphatidylinositol as described in Section 2. The phosphorylationproducts were analysed by thin layer chromatography and quanti-tated by liquid scintillation. A representative experiment from threedifferent experiments is shown as a diagram. C5b-9 (square contain-ing centre dot) increased PI 3-kinase activity with a maximum at 10min. This increase was abolished by wortmannin (WO) ( ). C5b6had no effect on PI 3-kinase activity (). (B) ASMC exposed toC5b-9 with or without PTX (500 ng/ml for 18 h) were lysed and PI3-kinase activity was determined. An autoradiograph of a representa-tive experiment for three different experiments indicating the positionof PI3P is shown. PTX produced a nearly complete inhibition ofC5b-9 induced PI 3-kinase activation. PI3P=phosphatidylinositol-3-phosphate.

wortmannin and PD098,059 (Fig. 8). These data sug-gest that both PI 3-kinase and MEK1 are involved inp70 S6 kinase activation.

3.9. In6ol6ement of MAPK and PI 3-kinase pathwaysin DNA synthesis induced by C5b-9

To test the role of MAPK and PI 3-kinase signalingin ASMC proliferation induced by C5b-9, the effect ofkinase specific inhibitors on DNA synthesis was exam-ined. Cells were exposed to C5b6, C5b-7, C5b-8, orC5b-9 assembled from purified proteins, then allowedto incorporate [3H]-thymidine for 18 h. Significant in-crease in thymidine incorporation occurred mostly inresponse to C5b-9 (Fig. 9A). Increased incorporationinduced by sublytic C5b-9 is about 30% of the increaseinduced by Ab and NHS (data not shown). The effectof C5b-7 and C5b-8 was not significant. BothPD098,059 and wortmannin were able to inhibit thecomplement induced DNA synthesis (Fig. 9B). The p38

Fig. 8. Activation of p70 S6 kinase by C5b-9. ASMC were exposed toC5b-9 () or pretreated with PD098,059 ( , PD) or wortmannin(square containing centre dot, WO) prior to C5b-9 exposure. C5b-9increased p70 S6 kinase activity which was abolished by wortmannin.Inhibition by PD098,059 at 10 min was less effective than at 5 min.

F. Niculescu et al. / Atherosclerosis 142 (1999) 47–5654

Fig. 9. Effects of complement activation on DNA synthesis inASMC. (A) ASMC were exposed to C5b6, C5b-7, C5b-8 or C5b-9assembled from purified components and cultured for 18 h in basalmedium free of serum and growth factors, containing 1 mCi/sample[3H]thymidine. Induced DNA synthesis was significantly higher inC5b-9 treated cells than in unstimulated cells (UNSTIM) (!PB0.001). Results are mean9S.E.M. from three separate experiments.(B) Pretreatment with 100 mM PD098,059 (+PD) or 100 nM wor-mannin (+WO) completely abolished the thymidine incorporationinduced by Ab+NHS. Results are expressed as percent of Ab+NHS level which is considered at 100%.

sublytic C5b-9 can be one of the factors responsible forASMC proliferation. Our data showed for the first timethat C5b-9 is mitogenic for human ASMC in culture.This observation may be significant for the pathogene-sis of atherosclerosis since C5b-9 is present in humanaortic atherosclerotic lesions in association with cellmembranes [16] and may stimulate in vivo ASMCactivation and proliferation.

Activation and progression of the cell cycle in ASMCby C5b-9 is achieved through an initial increase of cdk4followed by cdk2 activation concomitant with the in-creased DNA synthesis. Both kinases have been impli-cated in controlling cell cycle passage through the G1checkpoints [7]. The requirement of cell cycle kinases inSMC proliferation has been also shown by others usingantisense oligonucleotides against cdk2, cdc2 and pro-liferating-cell nuclear antigen in an experimental modelof carotid artery injury in rats, in which administrationof antisense oligonucleotide reduced the neointima for-mation and SMC proliferation [34,35]. Our study alsoimplicated p21 as a negative regulator of ASMC prolif-eration since high levels of p21 mRNA and protein areexpressed in growth arrested cells. After exposure toC5b-9 for 18 h the p21 mRNA and protein levels werereduced under the level of unstimulated cells. Thesedata suggest that p21 down-regulation may be an im-portant factor in activation of the cell cycle by C5b-9 inASMC. The p21 induction within 3 h after complementactivation is probably due to its property to respond asan immediate early gene to growth factors [36]. Ourdata indicating a negative regulatory role for p21 inASMC cell cycle are consistent with a recent reportwhich showed the inhibition of intimal hyperplasia ininjured arteries through p21 gene transfer by an aden-oviral vector [37,38].

Sublytic assembly of TCC on nucleated cells is ableto trigger the activation of a complex signal transduc-tion pathway responsible for many of its metaboliceffects [18]. Analysis of MAPK pathways induced byC5b-9 and the role of MAPK in ASMC proliferationrevealed that ERK1, JNK1, and p38 MAPK are acti-vated by C5b-9 and the enhanced DNA synthesis isabolished by inhibitors of MEK1 and PI 3-kinase butnot by inhibitors of p38 MAPK. ERK1 activationinduced by C5b-9 was higher than those of JNK1 andp38 MAPK. JNK1 activity induced by C5b-9 in ASMCwas transitory which contrasted with a higher andsustained activity seen in oligodendrocyte [25] and indi-cate that JNK1 may be less important in mediation ofC5b-9 effects in ASMC. The finding that ERK1 activitywas also inhibited by wortmannin indicated that ERK1pathway required for cell cycle induction is activatedthrough PI 3-kinase. Inhibition of PI 3-kinase activityby PTX suggested that PI 3-kinase is the downstreamkinase activated by C5b-9 through Gi protein. Thus itis reasonable to conclude that C5b-9 signaling is medi-

MAPK specific inhibitors SB 202190 and SB 203580were not effective in inhibiting DNA synthesis (datanot shown).

4. Discussion

After arterial injury, the rapid proliferative responseof SMC in media is followed by a second peak ofproliferation in the neointima, leading to intimal hyper-plasia [3]. Despite the important clinical consequencesof this process, the factors involved in proliferation andits regulatory mechanisms are not yet completely under-stood. Since activation of the complement system hasbeen implicated in both experimental and humanatherosclerosis [11,12] we have investigated whetheractivation of complement and membrane assembly of

F. Niculescu et al. / Atherosclerosis 142 (1999) 47–56 55

ated by interaction with G-protein which induces PI3-kinase activation through Gbg subunits. PI 3-kinaseas a signaling molecule was situated above Ras andRaf-1/MAPK inducing src activation and shc phospho-rylation by interacting with Gbg subunits [39–41]. PI3-kinase activation stimulate several kinases importantfor different signaling pathways since activation of bothp70 S6 kinase (Fig. 8) and p38 MAPK are inhibited bywortmannin. p70 S6 kinase induced through PI 3-ki-nase activation may be important in ASMC prolifera-tion since this kinase plays an essential role inmitogenesis [42].

Our findings are consistent with previously reporteddata showing that sustained ERK activation is requiredfor the initiation of SMC division and antisenseoligonucleotides to p44 and p42 ERKs abolished boththe basal and the PDGF-induced thymidine incorpora-tion [43]. Since endothelial cell proliferation is inducedby C5b-9 through release of bFGF and PDGF [44],ERK1 pathway activation may play a role in the mito-genic effect through a paracrine mechanism.

The data reported in this study suggest that activa-tion of complement and generation of sublytic C5b-9 inthe arterial wall exposed to injury may play a signifi-cant role in inducing proliferation of ASMC in vivo.C5b-9 stimulate vascular ASMC to enter cell cyclethrough activation of ERK1 pathway and PI 3-kinaseactivation in a G-protein dependent manner. Identifica-tion of complement activation as an important patho-genetic factor in ASMC proliferation and of specificsignaling pathways involved in this proliferation mayopen a new direction for future successful therapeuticstrategies to suppress intimal hyperplasia and theatherosclerotic process.

Acknowledgements

We want to thank Dr Moon L. Shin for the supportand encouragement to carry out this work. We alsothank Dr Myna Chi for aortic smooth muscle cellculture and Mrs N. Dehghan for the manuscript prepa-ration. This work is supported by the US Public HealthGrants, RO1 AI19006 and RO1 NS 199006.

References

[1] Owens GK. Regulation of differentiation of vascular smoothmuscle cells. Physiol Rev 1995;75:487–517.

[2] Parker TG, Schneider MD. Growth factors, proto-oncogenes,and plasticity of the cardiac phenotype. Ann Rev Physiol1991;53:179–200.

[3] Ross R. The pathogenesis of atherosclerosis: a perspective forthe 1990s. Nature 1993;362:801–9.

[4] Fidler IJ, Ellis LM. The implications of angiogenesis for thebiology and therapy of cancer metastasis. Cell 1994;79:185–8.

[5] Jackson CL, Schwartz SM. Pharmacology of smooth muscle cellreplication. Hypertension 1992;20:713–36.

[6] Koyama H, Raines EW, Bornfeldt KE, Roberts JM, Ross R.Fibrillar collagen inhibits arterial smooth muscle proliferationthrough regulation of Cdk2 inhibitors. Cell 1996;87:1069–78.

[7] Sherr CJ. G1 phase progression: cycling on cue. Cell1994;79:551–5.

[8] Xiong Y, Hannon GJ, Zhang H, Casso D, Kobayashi R, BeachD. p21 is a universal inhibitor of cyclin kinases. Nature1993;366:701–4.

[9] el-Deiry WS, Harper JW, O’Connor PM, Velculescu VE, Can-man CE, Jackman J, Pietenpol JA, Burrell M, Hill DE, Wang Y,et al. WAF1/CIP1 is induced in p53-mediated G1 arrest andapoptosis. Cancer Res 1994;54:1169–74.

[10] Stemme S, Hansson GK. Immune mechanisms in atherogenesis.Ann Med 1994;26:141–6.

[11] Niculescu F, Rus HG, Vlaicu R. Immunohistochemical localiza-tion of C5b-9, S-protein, C3d and apolipoprotein B in humanarterial tissues with atherosclerosis. Atherosclerosis 1987;65:1–11.

[12] Seifert PS, Hugo F, Hansson GK, Bhakdi S. Prelesional comple-ment activation in experimental atherosclerosis. Terminal C5b-9complement deposition coincides with cholesterol accumulationin the aortic intima of hypercholesterolemic rabbits. Lab Invest1989;60:747–54.

[13] Guettier C, Hinglais N, Bruneval P, Kazatchkine M, Bariety J,Camilleri JP. Immunohistochemical localization of S protein/vit-ronectin in human atherosclerotic versus arteriosclerotic arteries.Virchows Archiv A Pathol Anat Histopathol 1989;414:309–13.

[14] Seifert PS, Hugo F, Tranum-Jensen J, Zahringer U, Muhly M,Bhakdi S. Isolation and characterization of a complement-acti-vating lipid extracted from human atherosclerotic lesions. J ExpMed 1990;172:547–57.

[15] Bhakdi S, Dorweiler B, Kirchmann R, Torzewski J, Weise E,Tranum-Jensen J, Walev I, Wieland E. On the pathogenesis ofatherosclerosis: enzymatic transformation of human low densitylipoprotein to an atherogenic moiety. J Exp Med 1995;182:1959–71.

[16] Rus HG, Niculescu F, Constantinescu E, Cristea A, Vlaicu R.Immunoelectron-microscopic localization of the terminal C5b-9complement complex in human atherosclerotic fibrous plaque.Atherosclerosis 1986;61:35–42.

[17] Rus HG, Niculescu F, Porutiu D, Ghiurca V, Vlaicu R. Cellscarrying C5b-9 complement complexes in human atheroscleroticwall. Immunol Lett 1989;20:305–10.

[18] Shin ML, Rus HG, Niculescu FI. Membrane attack by comple-ment: assembly and biology of terminal complement complexes.In: Lee AG, editor. Biomembranes, vol. 4. JAI Press, 1996:123–49.

[19] Niculescu F, Rus H, Shin S, Lang T, Shin ML. Generation ofdiacylglycerol and ceramide during homologous complementactivation. J Immunol 1993;150:214–24.

[20] Niculescu F, Rus H, Shin ML. Receptor-independent activationof guanine nucleotide-binding regulatory proteins by terminal-complement complexes. J Biol Chem 1994;269:4417–23.

[21] Niculescu F, Rus H, Van Biesen T, Shin ML. Activation of Rasand mitogen-activated protein kinase pathway by terminal com-plement complexes is G protein dependent. J Immunol1997;158:4405–12.

[22] Shirazi Y, Rus HG, Macklin WB, Shin ML. Enhanced degrada-tion of messenger RNA encoding myelin proteins by terminalcomplement complexes in oligodendrocytes. J Immunol1993;150:4581–90.

[23] DiScipio RG, Smith CA, Muller-Eberhard HJ, Hugli TE. Theactivation of human complement component C5 by a fluid phaseC5 convertase. J Biol Chem 1983;258:10629–36.

F. Niculescu et al. / Atherosclerosis 142 (1999) 47–5656

[24] Carney DF, Lang TJ, Shin ML. Multiple signal messengersgenerated by terminal complement complexes and their role interminal complement complex elimination. J Immunol1990;145:623–9.

[25] Rus HG, Niculescu F, Shin ML. Sublytic complement attackinduces cell cycle in oligodendrocytes. S phase induction isdependent on c-jun activation. J Immunol 1996;156:4892–900.

[26] Gardner AM, Lange-Carter CA, Vaillancourt RR, Johnson GL.Measuring activation of kinases in mitogen-activated proteinkinase regulatory network. Methods Enzymol 1994;238:258–70.

[27] Vaillancourt RR, Harwood AE, Winitz S. Analysis of guaninenucleotides associated with protooncogene ras. Methods Enzy-mol 1994;238:255–8.

[28] Coso OA, Chiariello M, Kalinec G, Kyriakis JM, Woodgett J,Gutkind JS. Transforming G protein-coupled receptors potentlyactivate JNK (SAPK). Evidence for a divergence from thetyrosine kinase signaling pathway. J Biol Chem 1995;270:5620–4.

[29] Lazar DF, Wiese RJ, Brady MJ, Mastick CC, Waters SB,Yamauchi K, Pessin JE, Cuatrecasas P, Saltiel AR. Mitogen-ac-tivated protein kinase kinase inhibition does not block thestimulation of glucose utilization by insulin. J Biol Chem1995;270:20801–7.

[30] Jiang Y, Chen C, Li Z, Guo W, Gegner JA, Lin S, Han J.Characterization of the structure and function of a new mitogen-activated protein kinase (p38b). J Biol Chem 1996;271:17920–6.

[31] Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ. Isola-tion of biologically active ribonucleic acid from sources enrichedin ribonuclease. Biochemistry 1979;18:5294–9.

[32] Thomas PS. Hybridization of denatured RNA and small DNAfragments transferred to nitrocellulose. Proc Natl Acad Sci USA1980;77:5201–5.

[33] Gutkind SJ. The pathways connecting G protein-coupled recep-tors to nucleus through divergent mitogen-activated protein ki-nase cascades. J Biol Chem 1998;273:1839–42.

[34] Morishita R, Gibbons GH, Ellison KE, Nakajima M, von derLeyen H, Zhang L, Kaneda Y, Ogihara T, Dzau VJ. Intimalhyperplasia after vascular injury is inhibited by antisense cdk 2kinase oligonucleotides. J Clin Invest 1994;93:1458–64.

[35] Morishita R, Gibbons GH, Ellison KE, Nakajima M, Zhang L,Kaneda Y, Ogihara T, Dzau VJ. Single intraluminal delivery ofantisense cdc2 kinase and proliferating-cell nuclear antigenoligonucleotides results in chronic inhibition of neointimal hy-perplasia. Proc Natl Acad Sci USA 1993;90:8474–8.

[36] Michieli P, Chedid M, Lin D, Pierce JH, Mercer WE, Givol D.Induction of WAF1/CIP1 by a p53-independent pathway. Can-cer Res 1994;54:3391–5.

[37] Chang MW, Barr E, Lu MM, Barton K, Leiden JM. Aden-ovirus-mediated over-expression of the cyclin/cyclin-dependentkinase inhibitor, p21 inhibits vascular smooth muscle cell prolif-eration and neointima formation in the rat carotid artery modelof balloon angioplasty. J Clin Invest 1995;96:2260–8.

[38] Yang ZY, Simari RD, Perkins ND, San H, Gordon D, NabelGJ, Nabel EG. Role of the p21 cyclin-dependent kinase inhibitorin limiting intimal cell proliferation in response to arterial injury.Proc Natl Acad Sci USA 1996;93:7905–10.

[39] Yamauchi K, Holt K, Pessin JE. Phosphatidylinositol 3-kinasefunctions upstream of Ras and Raf in mediating insulin stimula-tion of c-fos transcription. J Biol Chem 1993;268:14597–600.

[40] Hawes BE, Luttrell LM, van Biesen T, Lefkowitz RJ. Phos-phatidylinositol 3-kinase is an early intermediate in the G betagamma-mediated mitogen-activated protein kinase signalingpathway. J Biol Chem 1996;271:12133–6.

[41] Lopez-Ilasaca M, Crespo P, Pellici PG, Gutkind JS, Wetzker R.Linkage of G protein-coupled receptors to the MAPK signalingpathway through PI 3-kinase gamma. Science 1997;275:394–7.

[42] Edelmann HML, Kuhne C, Petritsch C, Ballou LM. Cell cycleregulation of p70 S6 kinase and p42/p44 mitogen-activatedprotein kinases in swiss mouse 3T3 fibroblasts. J Biol Chem1996;271:963–71.

[43] Robinson CJ, Scott PH, Allan AB, Jess T, Gould GW, Plevin R.Treatment of vascular smooth muscle cells with antisense phos-phorothioate oligodeoxynucleotides directed against p42 and p44mitogen-activated protein kinases abolishes DNA synthesis inresponse to platelet-derived growth factor. Biochem J1996;320:123–7.

[44] Benzaquen LR, Nicholson-Weller A, Halperin JA. Terminalcomplement proteins C5b-9 release basic fibroblast growth factorand platelet-derived growth factor from endothelial cells. J ExpMed 1994;179:985–92.

.