Very Low Density Lipoprotein–Mediated SignalTransduction and Plasminogen Activator Inhibitor Type 1

in Cultured HepG2 CellsCristina Banfi, Luciana Mussoni, Patrizia Rise, Maria Grazia Cattaneo, Lucia Vicentini,

Fiorenzo Battaini, Claudio Galli, Elena Tremoli

Abstract—In normal subjects and in patients with cardiovascular disease, plasma triglycerides are positively correlatedwith plasminogen activator inhibitor type 1 (PAI-1) levels. Moreover, in vitro studies indicate that VLDLs induce PAI-1synthesis in cultured cells, ie, endothelial and HepG2 cells. However, the signaling pathways involved in the effect ofVLDL on PAI-1 synthesis have not yet been investigated. We report that VLDLs induce a signaling cascade that leadsto an enhanced secretion of PAI-1 by HepG2 cells. In myo-[3H]inositol–labeled HepG2 cells, VLDL (100mg/mL)caused a time-dependent increase in [3H]inositol phosphates, the temporal sequence being tris.bis.monophosphate.VLDL brought about a time-dependent stimulation of membrane-associated protein kinase C (PKC) activity andarachidonate release. Finally, VLDL stimulated mitogen-activated protein (MAP) kinase, and this effect was reduced by1-(5-isoquinolinylsulfonyl)-2-methylpiperazine (H7), which suggests that PKC plays a pivotal role in MAP kinasephosphorylation. VLDL-induced PAI-1 secretion was completely prevented by U73122, a specific inhibitor ofphosphatidylinositol-specific phospholipase C, by H7 or by PKC downregulation, and by mepacrine (allP,0.01 versusVLDL-treated cells). 3,4,5-Trimethoxybenzoic acid 8-(diethylamino)-octyl ester, which prevents Ca21 release fromintracellular stores, inhibited VLDL-induced PAI-1 secretion by 60% (P,0.05), and the MAP kinase/extracellularsignal–regulated kinase kinase (MEK) inhibitor PD98059 completely suppressed both basal and VLDL-induced PAI-1secretion. These data demonstrate that VLDL-induced PAI-1 biosynthesis results from a principal signaling pathwayinvolving PKC-mediated MAP kinase activation.(Circ Res. 1999;85:208-217.)

Key Words: plasminogen activator inhibitor type 1n VLDL n fibrinolysis n signalingn hepatoma cell line

A role for plasma triglycerides as a risk factor for cardio-vascular disease has been recently proposed.1 Case-

control studies have shown a positive correlation betweentriglyceride levels and incidence of cardiovascular disease,and most prospective studies have confirmed thisrelationship.1

Hypertriglyceridemia is associated with reduction in HDLlevels, glucose intolerance, and insulin resistance.2,3 In thispathological condition, reduction of the plasma fibrinolyticcapacity has been also documented, and several studies reporta direct relationship between the levels of plasminogenactivator inhibitor type 1 (PAI-1) and plasma triglycerides,the latter being an independent variable determining thelevels of PAI-1 in plasma.4–9 Biological plausibility for thisobservation made in clinical settings has been obtained by invitro studies in cultured cells.

VLDLs have been shown to increase PAI-1 biosynthesis inendothelial cells by inducing transcription of the PAI-1 genepromoter.10,11 Similarly, VLDLs increase PAI-1 synthesis in

HepG2 cells, enhancing steady-state PAI-1 mRNA levels,because of stabilization of the 3.2- and 2.2-kb PAI-1 mRNAtranscripts; the induction of PAI-1 by VLDLs is dependent onthe interaction of the lipoprotein with the apolipoprotein B/Ereceptors and correlates with intracellular triglycerideaccumulation.12

PAI-1 synthesis is regulated by several second-messengersignaling pathways that are cell specific. Protein kinase C(PKC) activation is positively associated with PAI-1 induc-tion, whereas agonist-induced cAMP accumulation has neg-ative effects.13–16A PKC-mediated mechanism is involved inthe induction of PAI-1 synthesis by phorbol esters, tumornecrosis factor-a (TNF-a), and transforming growth factor-bin different cell systems.17–20 Different signaling pathwayshave been proposed for the induction of PAI-1 elevation inbovine aortic endothelial cells by transforming growthfactor-b and TNF-a or lipopolysaccharide.16 A genistein-sensitive phosphorylation step is involved in TNF-a–inducedincrease of PAI-1 gene transcription in human endothelial

Received September 18, 1998; accepted May 17, 1999.From the Institute of Pharmacological Sciences (C.B., L.M., P.R., F.B., C.G., E.T.) and Department of Pharmacology (M.G.C., L.V.), University of

Milan, Italy.Correspondence to Prof Elena Tremoli, Laboratory of Pharmacology of Thrombosis and Atherosclerosis, Institute of Pharmacological Sciences,

University of Milan, Via Balzaretti 9, 20133 Milano, Italy. E-mail elena.tremoli@unimi.it© 1999 American Heart Association, Inc.

Circulation Researchis available at http://www.circresaha.org

208 by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from

cells.21 Finally, calcium-mobilizing agents stimulate PAI-1synthesis in U937 cells.22

LDLs and HDLs have been shown to induce phosphoino-sitide turnover, PKC translocation, and mitogen-activatedprotein (MAP) kinase activation in human and rat smoothmuscle cells, as well as in human skin fibroblasts.23,24

Oxidized LDLs stimulate MAP kinase in smooth muscle cellsand in macrophages.25 In endothelial cells, oxidized LDLsinduce PAI-1 expression through activation of thephosphatidylinositol-phospholipase pathway.26 No informa-tion is available on the signaling pathways evoked byVLDLs. In this study, we demonstrate that VLDLs induceseveral signaling pathways in HepG2 cells and, using variousinhibitors, that the induction of PAI-1 secretion results fromactivating these signaling pathways.

Materials and MethodsMaterialsMEM and FCS were from Life Technologies. The mycoplasmadetection kit and anti-mouse IgG-horseradish peroxidase antibodywere from Boehringer Mannheim. Plastic ware for cell culture wasfrom Costar. Sterile filters were from Millipore Corp. The F1-5ELISA for PAI-1 antigen detection was from Monozyme. [g-32

P]ATP (specific activity, 3000 Ci/mmol), myo-[3H]inositol (specificactivity, 17.7 Ci/mmol), [5,6,8,9,11,12,14,15-3H]arachidonic acid(specific activity, 211 Ci/mmol), [1-14C]palmitic acid (specific ac-tivity, 55 mCi/mmol), and an enhanced chemiluminescence (ECL)detection system were from Amersham. Solvents and silica gel plateswere from Merck. Liquid scintillation Formula 989 was from NEN.Monoclonal antibody (4G10) was from Upstate Biochemicals, Inc.Rabbit polyclonal antibody against phosphospecific p44/p42 MAPkinase was from New England Biolabs. Other reagents werefrom Sigma.

Chemicals1-(6-{[(17b)-3-Methoxyestra-1,3,5(10)-trien-17-yl]amino}hexyl-1H-pyrrole-2,5-dione (U73122) and 1-{6–(-[(17b)-3-methoxyestra-1,3,5(10)-trien-17-yl]amino)hexyl}-2,5-pyrrolidinedione (U73343)were from ICN Biomedicals; tricyclodecan-9-yl-xanthate (D609)was from Calbiochem; butylated hydroxytoluene (BHT), 3,4,5-trimethoxybenzoic acid 8-(diethylamino)-octyl ester (TMB-8), ni-fedipine, EGTA, thapsigargin, nordihydroguaiaretic acid (NDGA),indomethacin,mepacrine,1-(5-isoquinolinylsulfonyl)-2-methylpiper-azine (H7), sphingosine, (1)-a-tocopherol (vitamin E), herbimycinA, phenylmethylsulfonyl fluoride (PMSF), and phorbol 12-myristate13-acetate (PMA) were from Sigma; and (2-[2-amino-3-methoxyphenyl]-4H-1-benzopyran-4-one) (PD98059) was fromNew England Biolabs. EGTA and H7 were dissolved in steriledistilled water, and all of the chemicals were dissolved in DMSOor ethanol.

Cell CulturesHepG2 cells were cultured as previously described12 in MEMsupplemented with 10% heat-inactivated FCS containing 2 mmol/LL-glutamine, 100 IU/mL penicillin, 100 mg/mL streptomycin, 2.2mg/L sodium bicarbonate, and 1 mmol/L sodium pyruvate under ahumidified atmosphere of 95% air/5% CO2 at 37°C. The cell line wasfound free of mycoplasm or infection. For experiments, cells wereplated at 1.53104 in 12-well plates and used at subconfluence after24-hour preincubation in serum-free medium. Cells were incubatedin the presence or absence of VLDL with appropriate chemicals orvehicle additions (DMSO or ethanol, 0.1% vol/vol). Neither agent,DMSO or ethanol, influenced biochemical response or PAI-1 secre-tion by HepG2 cells; nor did either induce cytotoxicity, as judged bymorphology and trypan blue exclusion.

Lipoprotein PreparationBlood was obtained from normolipidemic subjects among themedical staff attending the E. Grossi Paoletti Center at NiguardaHospital (Milan, Italy). Blood drawn from the antecubital vein afterovernight fasting was anticoagulated with Na2-EDTA (1 mg/mL)containing 10 kallikrein units/mL aprotinin and kept on ice. Plasmawas separated by centrifugation (600g) at 4°C, and VLDLs wereisolated as previously described.12 VLDL particles were sterilizedthrough a 0.45mm filter, stored in sterile tubes at 4°C, and usedwithin 2 weeks. Total protein content in VLDLs was measured bythe Lowry et al27 method. The lipoproteins were essentially free fromcontamination by other lipoproteins as determined by nondenaturinggradient gel electrophoresis. In the study, a total of 20 individualVLDL preparations were used. Average VLDL lipid composition,expressed as percentage of total mass (sum of triglycerides, choles-terol, cholesterol ester, phospholipids, and proteins), was as follows:triglycerides, 59.3665.9%; free cholesterol, 4.7560.6%; cholesterolester, 11.1561.1%; phospholipids, 16.7562.4%; and proteins,860.8%. The composition is given as the mean value of 6preparations.

Quantification of PAI-1 AntigenThe concentration of PAI-1 in conditioned medium of HepG2 cellswas assayed with specific ELISA.12 The possible interference ofVLDL with the assay was excluded by experiments in which PAI-1antigen was determined in medium containing different concentra-tions of the lipoprotein.

Measurement of Inositol Phospholipid FormationSubconfluent HepG2 cells were incubated with MEM containing 2%FCS with 10 mCi/well of myo-[3H]inositol. After 24 hours oflabeling, the cells were incubated for 20 minutes with MEMcontaining 20 mmol/L HEPES (pH 7.3), 0.1% BSA, and 20 mmol/LLiCl. Cells were incubated with or without VLDL for different timeperiods at 37°C. The reaction was terminated by rapid removal of themedium by aspiration, and the cells were rinsed twice with ice-coldPBS and scraped off into 250mL of trichloroacetic acid 5%. Cellswere then pelleted by centrifugation at 15 000g for 5 minutes at 4°C.The supernatants were extracted with diethyl ether saturated withwater, neutralized with 1 mol/L NaOH, evaporated under N2, andresuspended in 250mL of H2O. Inositol phosphates were separatedby HPLC connected to a radiodetector (Radiomatic Flo-One Beta,Canberra Packard).

PKC ActivityCells were incubated for different times in serum-free mediumcontaining 100mg protein/mL VLDL and 100 nmol/L PMA. Atthe end of incubation, cells were washed twice with ice-cold PBS,scraped off, and homogenized with a polytetrafluoroethyleneglass homogenizer in 0.32 mol/L sucrose buffered with20 mmol/L Tris-HCl (pH 7.4) containing (in mmol/L) EDTA 2,EGTA 10, b-mercaptoethanol 50, and PMSF 0.3, and 20mg/mLleupeptin (homogenization buffer). The homogenate was centri-fuged at 100 000g for 30 minutes at 4°C, and the supernatant wascollected for PKC determination (cytosolic fraction). The pelletwas resuspended in homogenization buffer and centrifuged again.The remaining pellet was sonicated in the same buffer (exceptcontaining 0.2% Triton X-100), incubated at 4°C for 45 minutes,and centrifuged at 100 000g for 30 minutes to obtain the partic-ulate fraction. To examine PKC activity, cytosolic and particulatefractions were incubated at 37°C in buffer containing 3mmol/LTris-HCl (pH 7.5), 0.8mmol/L magnesium acetate, and 1 mg Pepa [Ser25]-PKC19 –31 as kinase-specific substrate. The reaction wasstarted by adding 50mmol/L [g-32P]ATP (0.45mCi per sample).Basal activity was measured in the presence of 0.1mmol/L ofEGTA, whereas stimulated activity was evaluated in the presenceof 10 mg of phosphatidylserine and 1 mg of diolein.28 Thereaction was stopped after 5 minutes by spotting 25mL of thesample onto P-81 phosphocellulose paper, adding 25mL of 0.6%H3PO4 to the spot, and washing the paper with tap water.

Banfi et al VLDL-Mediated PAI-1 Signal Transduction in HepG2 209

by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from

Radioactivity retained by the phosphocellulose was determinedby liquid scintillation counting using Formula 989. Proteincontent was measured by Bradford method.29

Release of [3H]Arachidonic Acid From PrelabeledHepG2 CellsCells were incubated for 16 hours in medium containing 5% FCS and0.5 mCi/mL of [3H]arachidonic acid.30 After removal of medium,cells were rinsed 3 times with PBS and incubated with or withoutVLDL for the indicated time. The medium was rapidly removed, andthe radioactivity was quantified by scintillation counting. Lipidswere extracted from the medium by the method of Folch et al.31

HPLC analysis performed on lipid extracts confirmed that theradioactivity was associated almost exclusively with free arachidonicacid.

Analysis of Phospholipase D (PLD) ActivityHepG2 cells were labeled for 16 hours with 1mCi of [14C]palmitatein 1 mL of MEM containing 5% FCS.32 After 3 washings with MEMcontaining 10% serum, cells were incubated for 5 minutes in MEMcontaining 1% ethanol and then stimulated with VLDL or 1mmol/LA23187, for various times. The medium was rapidly aspirated, and,after addition of ice-cold methanol, cells were scraped off and lipidsextracted by the method of Folch et al.31 [14C]PEth was separatedfrom the other phospholipids by thin-layer chromatography on silicagel 60 plates. The solvent system was the organic phase of ethylacetate/isooctane/acetic acid/water (13:2:3:10, vol/vol/vol/vol). Thespots corresponding to phosphatidylethanol (PEth), total phospho-lipids, and phosphatidic acid were identified by iodine vapor,scraped, and then counted for radioactivity in scintillation liquid in abeta counter.

Immunoblotting With Anti-PhosphotyrosineMonoclonal AntibodyHepG2 cells were incubated with VLDL and 100 nmol/L PMA forthe indicated time. The cells were then rinsed with calcium- andmagnesium-free PBS and then lysed on ice with 100mL of50 mmol/L b-glycerophosphate, pH 7.2, containing (in mmol/L)sodium orthovanadate 100, MgCl2 2, EGTA 1, and DTT 1, and0.5% Triton X-100, 10mg/mL leupeptin, and 2mg/mL aprotinin.Cell lysates, after protein determination by the Lowry et al27

method, were boiled for 5 minutes in Laemmli buffer,33 resolvedon a 10% SDS-polyacrylamide gel, and transferred to a nitrocel-lulose membrane. Filters were incubated overnight at 4°C with3% BSA in TTBS (20 mmol/L Tris, 134 mmol/L NaCl, and 0.05%Tween 20 [pH 7.6]) and incubated for 1 hour in TTBS containing3% BSA and a 1:2000 dilution of the 4G10 anti-phosphotyrosinemonoclonal antibody. After washing with TTBS, the membraneswere incubated with sheep anti-mouse IgG-horseradish peroxi-dase antibody (1:5000 dilution in TTBS) for 1 hour at roomtemperature, and phosphorylated proteins were detected using theAmersham ECL system.

MAP Kinase ActivityHepG2 cells were first incubated for 48 hours in MEM containing0.1% FCS and for 24 hours with medium alone. The cells were thenincubated with 100mg/mL VLDL for 5 and 10 minutes. Aftertreatment, the medium was removed, and the cells were washed andscraped in 0.5 mL of homogenization buffer, as described by Segeret al.34 After sonication (2 times for 7 seconds) and centrifugation at15 000gfor 10 minutes at 4°C, the supernatants were fractionated onDEAE-cellulose minicolumns. MAP kinase activity was determinedby phosphate incorporation into myelin basic protein (MBP) in thepresence of [g-32P]ATP (2 mCi/sample).34

MAP Kinase ImmunoblottingHepG2 cells were treated with VLDL and FCS for the indicatedtimes. Cells were washed with PBS and lysed in 100mL ofLaemmli buffer. Equal amounts of protein (10mg) were separated

on a 12% SDS-polyacrylamide gel and transferred to a nitrocel-lulose membrane. Western blot analysis was performed with anantibody (1:1000 dilution in TTBS containing 5% milk) againstphosphospecific p44/p42 MAP kinase, which detects p42 and p44MAP kinase (extracellular signal–regulated kinase ERK1 andERK2) only when activated by phosphorylation at tyrosine 204and threonine 202. After incubation with horseradish peroxidase-conjugated secondary antibody, the blot was developed using theAmersham ECL system.

Statistical AnalysisAll experiments were conducted in duplicate with independentseparate cultures (n5numbers of independent experiments). Data areexpressed as mean6SEM. Statistical comparison of control withtreated groups was carried by ANOVA repeated measures followedby the Tukey test. The accepted level of significance wasP,0.05.

ResultsEffect of VLDL on PAI-1 BiosynthesisPrevious studies by our group12 have shown that VLDLsincubated for 16 hours with HepG2 cells concentration-dependently increase PAI-1 mRNA and protein synthesis inthese cells. This effect is prevented by prior incubation ofcells with 15mg/mL antibody against the apolipoprotein B/Ereceptor (C7). In the present study, VLDLs (100mg/mL)were incubated with cells for 16 hours, and PAI-1 wasdetermined in conditioned medium. VLDLs significantlyincreased PAI-1 secretion from the cells (5.0860.18 versus10.0560.42 ng/mL PAI-1 [P,0.01, n514]). This lipoproteinconcentration and incubation time were used, unless speci-fied, throughout the study. In experiments performed with 7individual VLDL preparations, the percentage CV in PAI-1secretion was 6.4.

VLDL-Induced PAI-1 Biosynthesis Is Dependenton Inositol Phospholipid CatabolismWe first examined whether VLDL incubated with HepG2cells induced generation of [3H]inositol phosphates. VLDLinduced generation of [3H]inositol trisphosphate (InsP3)and to a lesser extent of [3H]inositol bisphosphate (InsP2),with the maximal increase occurring between 2 and 5minutes, after which [3H]InsP3 declined, whereas at 10minutes a 3H content was sustained in InsP2, and itincreased further in [3H]inositol monophosphate (InsP)(Figure 1). These data indicate that VLDLs quickly inducephosphatidylinositol breakdown via the activation of phos-pholipase C (PLC). We therefore investigated the effectsof U73122, at concentrations completely inhibiting PLC-dependent processes,35 on PAI-1 secretion. Cells were firstincubated for 1 hour with vehicle or with 5 or 10mmol/LU73122 or its inactive analog U73343 and then exposed toVLDL. Neither U73122 nor U73343 affected basal PAI-1secretion, whereas 10mmol/L U73122 completely pre-vented the effect of VLDL (Figure 2), and U73343 had noeffect. D609 (10mmol/L), a phosphatidylcholine-specificPLC inhibitor,36 did not reduce VLDL-induced PAI-1secretion, but it augmented it by almost 2-fold (P,0.01versus VLDL-treated cells, n54).

Effect of Calcium on VLDL-InducedPAI-1 ReleaseThe importance of Ca21 influx from the extracellular envi-ronment and Ca21 mobilization from intracellular stores in the

210 Circulation Research July 23, 1999

by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from

induction of PAI-1 biosynthesis by VLDL were then evalu-ated. Measurement of VLDL-induced intracellular Ca21

changes, using Fluo 3-AM–loaded cells, was not possiblebecause of turbidity produced by the lipoprotein. We there-fore analyzed the effects of various compounds interferingwith calcium37 on PAI-1 secretion. Neither nifedipine (10 to50 mmol/L), a calcium channel blocker, nor EGTA (0.1 to1 mmol/L) incubated with cells for 1 hour affected subse-quent basal (5.360.2 and 4.9860.51 ng/mL for 50mmol/Lnifedipine and 1 mmol/L EGTA versus 5.1660.54 ng/mL foruntreated cells [P5NS, n54]) or VLDL-stimulated PAI-1secretion (11.8160.51 and 10.6260.86 ng/mL, respectively,versus 11.4560.92 ng/mL for VLDL-treated cells), whichindicates that extracellular calcium is not required for basal orVLDL-induced PAI-1 secretion. TMB-8 (50mmol/L), whichprevents Ca21 release from intracellular stores, significantlyreduced basal and VLDL-stimulated PAI-1 by 30% and64.6%, respectively (Figure 3, top). We also tested whetherCa21 mobilization from intracellular stores could affect PAI-1

biosynthesis. To this end, cells were incubated for 24 hourswith 1 mmol/L thapsigargin before VLDL addition. A markedincrease in PAI-1 secretion in cells incubated with or withoutVLDL was observed (P,0.01) (Figure 3, bottom).

Figure 1. Time-dependent inositol phosphate generation inresponse to VLDL. HepG2 cells were kept in serum-free mediumfor 24 hours and then incubated with 10 mCi/well of myo-[3H]inositol. HepG2 cells labeled with myo-[3H]inositol were treatedat 37°C with or without VLDL (100 mg/mL) in the presence of20 mmol/L LiCl for the indicated times. Inositol phosphates wereseparated by HPLC with radioactivity detector (see Materials andMethods). Data are given as mean6SEM (n53) and express per-centages of initial 3H content (100%) in each inositol phosphate(control). Actual initial values (time 0, n53) were the following: InsP(F), 37756541 cpm; InsP2 ( M ), 5606140 cpm; and InsP3 (E),3046139 cpm. *P,0.01 vs corresponding time 0.

Figure 2. Effect of U73122 on PAI-1 secretion. HepG2 cells were keptin serum-free medium for 24 hours, incubated for 1 hour with the indi-cated concentrations of U73122 or vehicle, and then stimulated for 16hours in the absence (open bars) or presence (hatched bars) of 100mg/mL VLDL in the continuous presence of U73122. PAI-1 antigenlevels were determined in cell supernatants. Each value represents themean6SEM of 4 individual experiments performed in duplicate.Results are presented as percentage of PAI-1 antigen levels in super-natants of HepG2 cells incubated with medium alone for 16 hours.*P,0.01 vs unstimulated cells; §P,0.01 vs cells stimulated by VLDLwithout U73122. – indicates absence of U73122.

Figure 3. Effect of TMB-8 and thapsigargin on PAI-1 secretion byHepG2. Top, Cells were kept in serum-free medium for 24 hoursand then incubated for 1 hour with the indicated concentrations ofTMB-8 or vehicle. Bottom, Cells were preincubated in mediumcontaining vehicle or thapsigargin for 24 hours. Cells were thenstimulated for 16 hours in the absence (open bars) or presence(hatched bars) of 100 mg/mL VLDL in the continued presence ofthe calcium modulators. PAI-1 antigen levels were determined incell supernatants. Data are mean6SEM of 4 individual experimentsperformed in duplicate. *P,0.01 vs unstimulated cells; §P,0.05,‡P,0.01 vs VLDL-stimulated cells. – indicates absence of TMB-8in top panel and absence of thapsigargin in bottom panel.

Banfi et al VLDL-Mediated PAI-1 Signal Transduction in HepG2 211

by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from

VLDL-Induced PAI-1 Release Is Dependenton PKCActivation of PLC leads to PKC activation, and PKC hasbeen shown to enhance PAI-1 biosynthesis in several celltypes.14–16 We therefore examined whether PKC is involvedin VLDL-induced PAI-1 release in HepG2 cells. The capacityof VLDL to stimulate PKC activity was assessed by evalu-ating the translocation of this enzyme from cytosol to theparticulate fraction. In unstimulated cells, up to 95% ofcellular PKC activity was recovered in the cytosol fraction.Incubation of cells with VLDL resulted in a maximal increasein membrane-associated PKC at 5 minutes (P,0.01). After10 minutes, it returned toward basal levels, with a concomi-tant and opposite change in cytosol activity (Figure 4). Therelationship between PKC activation and PAI-1 secretion wasthen investigated. VLDL-induced PAI-1 secretion was com-pletely prevented by prior incubation of cells with H7, a PKCinhibitor (35 mmol/L). In additional experiments, HepG2cells were treated with vehicle or 100 nmol/L PMA for 24hours to downregulate PKC.16 The supernatant was thenreplaced with serum-free medium or supplemented with 100mg/mL VLDL or 100 nmol/L PMA. In cells preincubatedwith vehicle, VLDL and PMA increased PAI-1 secretion by217612.2% and 722692% (n54), respectively. After down-regulation of PKC, VLDL- and PMA-induced PAI-1 secre-tion was reduced by 70.6610.1% and 81.469.8%, respec-tively (P,0.01 versus control, n54), further confirming thatPKC activation is involved in VLDL-induced PAI-1secretion.

Effect of VLDL on Arachidonate Releaseby HepG2To assess whether VLDL affected phospholipase-mediatedarachidonate release, HepG2 cells were labeled with [3H]ara-chidonate and then incubated from 0 to 60 minutes withmedium alone or medium containing VLDL. A time-dependent [3H]arachidonate release (0 to 60 minutes) wasdetected in supernatants of VLDL-treated samples; levels offree [3H]arachidonate (confirmed by radio-HPLC) were sig-nificantly greater than in controls, with a peak increment at 45

minutes (Figure 5). Mepacrine, a phospholipase inhibitor, andH7 reduced VLDL-induced [3H]arachidonate release by 50%and 40%, respectively. The involvement of phospholipaseactivation in PAI-1 secretion was then assessed. Prior incu-bation of cells with 15mmol/L mepacrine for 1 hour inhibitedthe VLDL-induced enhancement in PAI-1 by 72.569.5%(P,0.01 versus VLDL-treated cells, n56). Indomethacin, atconcentrations completely suppressing cyclooxygenase activ-ity (10 or 20mmol/L), did not affect PAI-1 secretion by eitherunstimulated or VLDL-stimulated cells. NDGA at 20mmol/Lsignificantly increased basal PAI-1 secretion by HepG2 cells(5.260.40 versus 10.661.56 ng/mL [P,0.01] for untreatedand NDGA-treated cells, respectively), whereas this lipoxy-genase inhibitor failed to affect VLDL-induced PAI-1 secre-tion (10.1561.02 and 11.1560.75 ng/mL in cells treated withVLDL or VLDL plus NDGA, respectively).

VLDL-Induced PAI-1 Release InvolvesTyrosine PhosphorylationTyrosine phosphorylation at 5 and 15 minutes after theaddition of vehicle or VLDL was visualized by immunoblot-ting with the monoclonal antibody 4G10, as described inMaterials and Methods. Figure 6 shows that VLDL enhancedtyrosine phosphorylation of several proteins, with apparentmolecular masses of 40, 44, 70, 80, 110, and 200 kDa. Asimilar pattern of phosphorylation was observed in cellstreated with 100 nmol/L PMA. Preincubation of cells with35 mmol/L H7 attenuated the effect of VLDL on tyrosinephosphorylation (not shown). Herbimycin (1mmol/L), atyrosine phosphorylation inhibitor, inhibited the secretion ofPAI-1 by 81.563% in VLDL-treated cells (P,0.01, n54),which suggests that tyrosine phosphorylation is a prerequisitefor VLDL-induced PAI-1 biosynthesis.

Effect of VLDL on MAP Kinase ActivityEnhanced tyrosine phosphorylation of MAP kinase in re-sponse to PKC activation has been described in several celllines.38,39 To assess whether VLDL induced MAP kinase

Figure 4. Effect of VLDL on PKC activity in HepG2 cells. Cellswere incubated for 24 hours with medium alone and then stimu-lated with or without 100 mg/mL VLDL for the indicated times.Cytosol and particulate fractions were isolated, and PKC activitywas determined. Basal PKC-phosphorylating activities in soluble(open bars) and particulate (hatched bars) fractions were1.5460.25 and 0.3160.11 nmol/min z mg–1 of protein, respec-tively. Data are mean6SEM of 3 individual experiments per-formed in duplicate. *P,0.01 vs basal.

Figure 5. Arachidonate (AA) release in HepG2 cells incubatedwith VLDL. Cells were incubated for 16 hours in medium con-taining 0.5 mCi/mL [3H]AA and then with 100 mg/mL VLDL forthe indicated times. 3H was determined in cell supernatants.Data are mean6SEM of 3 individual experiments performed induplicate. *P,0.01 vs control. Inset, HPLC chromatogramshowing that [3H]AA was the major peak identifiable in thesupernatant of HepG2 cells.

212 Circulation Research July 23, 1999

by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from

activation, experiments were carried out by measuring32Plabeling of MBP in cytosol extracts of HepG2 cells preincu-bated for 24 hours with 0.1% FCS and then incubated for 5and 15 minutes with 100mg/mL VLDL. After 5 minutes,VLDL doubled MAP kinase activity, with the level remainingelevated at 15 minutes. Similar induction of MAP kinaseactivity was observed in cells stimulated with 20% FCS(Figure 7A). To confirm that the observed kinase activityresults from the stimulation of the ERK family of MAPkinases, phosphorylated ERK1 and ERK2 proteins wereidentified by Western blotting in cell lysates of HepG2 cellstreated with VLDL or 20% FCS. Serum phosphorylated bothERK1 and ERK2, whereas VLDL induced phosphorylationof the ERK2 isoform only (Figure 7B). Preincubation of cellswith 35 mmol/L H7 completely prevented ERK2 phosphor-ylation (Figure 7C). MAP kinase is phosphorylated andactivated by MAP kinase/extracellular signal–regulated ki-nase kinase (MEK), a dual-specificity kinase that phosphor-ylates serine and tyrosine residues. PD98059 is a highlyselective inhibitor of MEK that is commonly used to blockMAP kinase activation.40 PD98059 at a concentration knownto inhibit MEK by .50%40 inhibited both basal and VLDL-stimulated PAI-1 secretion by.75% (P,0.01), which indi-cates that MAP kinase was necessary not only for VLDL-induced PAI-1 secretion but also for basal PAI-1 biosynthesis(Figure 8).

Effect of VLDL on PLD ActivationThe formation of PEth, which is generated in the presence ofnoncytotoxic ethanol concentrations by the PLD-catalyzedtransphosphorylation of phosphatidylcholine, is a convenientindex of PLD activity.41 A23187 (1mmol/L) caused a rapid (5minutes) ethanol-dependent accumulation of [14C]PEth inHepG2 cells (from 0.06% of total radioactivity to 0.37%),whereas VLDL did not affect [14C]PEth over 45 minutes,

which indicates that the lipoprotein does not influence thissignaling pathway.

Role of Phosphatidylinositol 3-Kinase (PI 3-K) inVLDL-Induced PAI-1 SecretionWortmannin (0.1 to 1mmol/L), a PI 3-K inhibitor42 that wasincubated with cells 1 hour before VLDL addition, did notaffect basal (5.960.44 and 6.1260.62 ng/mL for untreatedand 1 mmol/L wortmannin–treated cells, respectively) orVLDL-stimulated PAI-1 secretion (15.6661.88 and16.4261.6 ng/mL for VLDL and VLDL plus wortmannin,respectively).

Effect of Antioxidants on VLDL-InducedPAI-1 SecretionTo evaluate whether oxidation/modification of VLDL wasimplicated in the effect of the lipoprotein on PAI-1, experi-ments that included BHT or vitamin E were carried out. BHT(25 mmol/L) did not affect basal PAI-1 secretion (4.560.15and 5.360.52 ng/mL for untreated and BHT-treated cells,

Figure 6. Effect of VLDL on tyrosine phosphorylation in HepG2cells. Time-dependent tyrosine phosphorylation in cells treatedwith VLDL. Cells were incubated for 24 hours with mediumalone and then stimulated with or without 100 mg/mL VLDL and100 nmol/L PMA for the indicated times. Cell lysates (100mg/lane) were fractionated on a 10% SDS-polyacrylamide gelfollowed by immunoblotting with anti-phosphotyrosine antibodyas described (see Materials and Methods). Molecular massesare indicated in kDa. Results are representative of 3 indepen-dent experiments.

Figure 7. Effect of VLDL on MAP kinase activation in HepG2cells. A, Cells were kept in medium containing 0.1% FCS for 48hours and then stimulated with 100 mg/mL VLDL and 20% FCSfor the indicated times. MAP kinase activity was determined incell lysates as described (see Materials and Methods). Data aremean6SEM of 3 individual experiments. – indicates absence ofstimulus. B, Cells were incubated for 48 hours in medium con-taining 0.1% FCS and then stimulated with 100 mg/mL VLDLand 20% FCS for the indicated times. C, Cells kept in 0.1%FCS for 48 hours were preincubated with 35 mmol/L H7 andthen stimulated with 100 mg/mL VLDL for the indicated times.Cell lysates (10 mg/lane) were fractionated on a 12% SDS-poly-acrylamide gel followed by immunoblotting with anti-phosphory-lated ERK1 and ERK2 antibody as described (see Materials andMethods). Results are representative of 3 independentexperiments.

Banfi et al VLDL-Mediated PAI-1 Signal Transduction in HepG2 213

by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from

respectively) or VLDL-enhanced PAI-1 release (9.3960.77and 9.860.48 ng/mL for VLDL and for VLDL plus BHT,respectively). Similarly, no change in PAI-1 secretion wasrecorded in the presence of 50mmol/L vitamin E. Thesefindings rule out the hypothesis that active oxidation ofVLDL under our experimental conditions was responsible forthe observed effect of the lipoprotein on PAI-1 biosynthesis.

DiscussionIn this study, we report that VLDLs activate in HepG2 cellsinositol phosphate generation and PKC, which in turn induceMAP kinase phosphorylation. We have shown by pharmaco-logical studies that coordinated activation of this signalingpathway is essential for PAI-1 biosynthesis by these cells.

As previously observed for other lipoprotein classes, eg,LDL and HDL,23,43 VLDLs induce phosphoinositide catabo-lism with the formation of inositol phosphates, and therelevance of this pathway to VLDL-induced PAI-1 secretionis supported by the data obtained with a specific inhibitor ofthis signaling pathway. At variance, inhibition ofphosphatidylcholine-specific activity by D609 was associatedwith VLDL-induced increases in PAI-1 secretion, whichsuggests that this pathway too is involved in the regulation ofPAI-1 by mechanisms not yet understood. In addition, aninvolvement of PLD or PI 3-K activation in the VLDL-induced PAI-1 secretion was excluded.

Blockage of calcium influx by a calcium channel blockeror by EGTA did not affect the VLDL-induced PAI-1 biosyn-thesis, whereas TMB-8 significantly reduced it, which sug-gests that the effect of VLDL requires release of sequesteredCa21. Thapsigargin, which inhibits the Ca21-ATPase of intra-cellular Ca21 stores and, consequently, activates Ca21 influxfrom the extracellular space,44 strongly induced PAI-1 secre-tion both in basal and in VLDL-stimulated cells; this findingis in agreement with a previous report on U937 cells exposedto the calcium ionophore A23187.22

Cytosolic PKC is normally functionally inactive, whereasafter agonist stimulation and generation of inositol phos-phates and diacylglycerol, the enzyme translocates to themembrane, where the lipid-rich environment and Ca21 serve

as cofactors for enzymatic kinase activity.45 The observationthat VLDLs induce PKC translocation suggests that activa-tion of this enzyme may be crucial to the signal transductionpathways triggered by this lipoprotein. In HepG2 cells, 2calcium-dependent (a andbII) and 3 calcium-independent (d,e, andz) isoforms of PKC have been described.46 Althoughwe have not explored the role of each PKC isoform, involve-ment of the calcium-dependent PKCa can be hypothesized.Preliminary experiments have shown that PKCa isoformtranslocates from the cytosol to the membrane fraction, whichsupports this hypothesis at least in part (data not shown).

Activation of PKC, brought about by its putative ligandPMA, induces PAI-1 in HepG2 cells.15,47 Inhibition of PKCactivation by H7 and PKC downregulation prevent VLDL-induced PAI-1 secretion, which indicates that, similarly toother agonists, this lipoprotein fraction increases PAI-1 se-cretion via this signaling pathway.

We also show that VLDLs induce release of arachidonatefrom membrane lipids. This increase is time dependent andpartially reduced by the PKC inhibitor, H7. Unesterifiedarachidonate may originate from the activation of severalenzymes, among which phosphatidylinositol-specific PLC(PI-PLC) and phospholipase A2 are the most likely candi-dates.48 The reduction of arachidonate release by H7 indicatesthat at least 50% of the release of this fatty acid is related toa PKC-dependent mechanism. PKC has been implicated inphospholipase A2 phosphorylation, and this process is sec-ondary to MAP kinase activation.49 In addition, the resultsobtained with inhibitors of arachidonate oxidation rule out thepossibility that products of arachidonate metabolism areimplicated in VLDL-induced enhancement of PAI-1secretion.

One downstream event of PKC activation is stimulation ofMAP kinases.50,51 MAP kinases form a family of serine/threonine kinases uniquely activated by dual phosphorylationof threonine and tyrosine residues. This pathway may origi-nate from several distinct classes of cell-surface receptorssuch as tyrosine kinases and G protein–coupled recep-tors.52–54 This cascade of events involves activation of Ras,which in turn activates Raf and leads to MAP kinaseactivation. PKC regulates the MAP kinase cascade in anumber of ways. In some cells, PKC regulation of MAPkinase is Ras dependent, and in others it is Ras indepen-dent.55,56 We show here that both tyrosine phosphorylationand MAP kinase activation are induced by VLDL in HepG2cells. MAP kinase activation evoked by VLDL, as evaluatedby MBP assay, required shorter time than that of ERK, whichindicates that other members of the MAP kinase family mightbe phosphorylated by VLDL.57 Interestingly, VLDL inducedphosphorylation of ERK2 isoform only, a phenomenon thatrequires investigation. Inhibition of tyrosine and MAP kinasepathways results in the impairment of the VLDL-inducedincrease in PAI-1. Indeed, the inhibition of MAP kinaseresults in almost complete suppression of basal and VLDL-stimulated PAI-1 secretion, which suggests that MAP kinaseactivation is the final essential step in PAI-1 biosynthesis inHepG2 cells. Interestingly, inhibition of PKC activation byH7 prevented ERK2 phosphorylation induced by VLDL,

Figure 8. Effect of PD98059 on VLDL-induced PAI-1 release.Cells, kept in serum-free medium for 24 hours, were treatedwith vehicle or with 30 mmol/L PD98059 for 1 hour and thenwith 100 mg/mL VLDL in the continued presence of the inhibitor.PAI-1 antigen levels were determined in cell supernatants. Dataare mean6SEM of 4 individual experiments performed in dupli-cate. *P,0.01 vs supernatant of unstimulated cells; §P,0.01 vsVLDL-stimulated cells. – indicates absence of PD98059.

214 Circulation Research July 23, 1999

by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from

which suggests that PKC plays a pivotal role in VLDL-induced MAP kinase phosphorylation.

On the basis of the data discussed above, we conclude thatVLDL or its components, through activation ofphosphatidylinositol-specific PLC, induce InsP3 and diacyl-glycerol formation and intracellular Ca21 increases, leading toactivation of PKC and its translocation from the cytosol to themembrane fraction. PKC activates MAP kinases, which inturn result in PAI-1 upregulation. This signaling pathwaymay also induce, in addition to PAI-1, a variety of othercellular functions as well as the synthesis of biologicallyactive proteins.

The mechanisms by which VLDLs induce the signalingcascade discussed above have not been addressed in thisstudy.

The VLDL fraction contains VLDL and VLDL remnants,resulting from lipoprotein lipase-mediated hydrolysis of tri-glycerides.58 Both VLDL and VLDL remnants bind and aretaken up by a variety of receptors present on liver cells, eg,the LDL receptor and the LDL receptor–related protein.59 Inaddition, VLDL may be catabolized by a specific VLDLreceptor present on several cell systems, including HepG2cells.60 Previous studies by our group have shown thatVLDL-induced PAI-1 is fully prevented by the C7 antibody,which specifically recognizes LDL receptor, suggesting amechanism mediated by the ligand activation of the apoli-poprotein B/E receptor.12 The possibility, however, that inaddition to LDL receptor, VLDL receptor is also involved inPAI-1 induction cannot be ruled out.

Interestingly, the C7 antibody reduces S6 kinase activationand phosphoinositide turnover induced by LDL in humanvascular smooth muscle cells23 and Ca21-mediated VCAM-1and E-selectin expression in human cultured endothelial cellsexposed to native LDL.61 LDL- and HDL-induced surfactantsecretion by alveolar type II cells involves activation ofphosphoinositide hydrolysis and increases in intracellularcalcium concentration and PKC activity, and these effects areinhibited by pertussis toxin, which suggests that the LDLreceptor also interacts with a heterotrimeric G protein.62 Themechanisms, however, by which lipoprotein-receptor interac-tions initiate signal transduction are unknown.

Thus, ligand activation of the apolipoprotein B/E receptoror lipid components formed after uptake and degradation ofVLDL or VLDL remnants may be responsible for theactivation of the signaling cascade leading to PAI-1biosynthesis.

We have previously shown that the exposure of HepG2cells to VLDL results in intracellular triglyceride accumula-tion, which in turn correlates with PAI-1 secretion.12 Incuba-tion of HepG2 cells with linoleic acid, either as triglyceride orcomplexed to albumin, increases PAI-1 secretion,63 and adirect effect of unsaturated fatty acids on PAI-1 gene tran-scription has been recently demonstrated by Nilsson et al64 inendothelial cells. Interestingly, arachidonic and linoleic acidsactivate MAP kinase in vascular smooth muscle cells.65,66

Finally, experiments with antioxidants indicate that activeoxidation of the lipoprotein under in vitro experimentalconditions was not likely to mediate the event.

PAI-1 levels have been shown to be elevated in patientswith risk factors for cardiovascular disease, eg, hypertension,type 2 diabetes, insulin resistance, and hypertriglyceridemia.4

Elevated levels of this protein may reduce fibrinolytic activityresulting, in vivo, in a reduced fibrin dissolution and throm-bus removal. The interaction of VLDL with HepG2 cellsinduces increases in PAI-1 antigen in its active form.12 Thiseffect is obtained at concentrations of the lipoprotein similarto those present in vivo in normolipidemic subjects. In thiscontext, it is worth mentioning that a correlation betweenPAI-1 levels and triglycerides has been observed not only inpatients at high risk of cardiovascular disease but also inhealthy subjects with normal lipid profile.5,7 PAI-1 levels arenot constant during the day but show diurnal variation withremarkable postprandial increases, with particular incrementsin patients with hypertriglyceridemia. These changes duringthe day may represent a predisposing condition toward aneffect of VLDL in competent cells, leading to increase inPAI-1, which in turn results in reduced fibrinolysis.

On the basis of the present in vitro findings, we proposethat VLDLs induce biosynthesis of PAI-1 as the result of aprincipal signaling pathway involving PKC-mediated MAPkinase activation. However, the role of other potential relatedsignaling pathways in this process cannot be excluded. Thefinding that selective inhibition of the proposed signalingcascade results in almost complete suppression of the secre-tion of this antifibrinolytic protein may provide a series oftarget candidates for pharmacological strategies aimed atreducing PAI-1 biosynthesis.

AcknowledgmentsThis work was supported by a HIFMECH contract (BMH4-CT96-0272) from the European Community and by MURST 60%. Wethank Giovanna Giandomenico, Franca Meloni, and Laura Lucchifor excellent technical assistance. We also thank the E. GrossiPaoletti Center at Niguarda Hospital (Milan, Italy) for providingblood samples for VLDL preparation.

References1. Hokanson JE, Austin MA. Plasma triglyceride level is a risk factor for

cardiovascular disease independent of high-density lipoprotein cholester-ol level: a meta-analysis of population-based prospective studies.J Car-diovasc Risk. 1996;3:213–219.

2. Fuller J, Shimpley MJ, Rose G, Jarrett RJ, Keen H. Coronary heartdisease risk and impaired glucose tolerance: The Whitehall Study.Lancet.1980;i:1373–1376.

3. Frayn KN. Insulin resistance and lipid metabolism.Curr Opin Lipidol.1993;4:197–204.

4. Hamsten A, Eriksson P. Fibrinolysis and atherosclerosis.Baillieres ClinHaematol. 1995;8:345–363.

5. Hamsten A, Wiman B, de Faire U, Blomback M. Increased plasma levelsof a rapid inhibitor of tissue plasminogen activator in young survivors ofmyocardial infarction.N Engl J Med. 1985;313:1557–1563.

6. Hamsten A, de Faire U, Walldius G, Dahlen G, Szamosi A, Landou C,Blomback M, Wiman B. Plasminogen activator inhibitor in plasma: riskfactor for recurrent myocardial infarction.Lancet. 1987;2:3–9.

7. Juhan-Vague I, Vague P, Alessi MC, Badier C, Valadier J, Aillaud MF,Atlan C. Relationships between plasma insulin triglyceride, body massindex, and plasminogen activator inhibitor 1.Diabete Metab. 1987;13:331–336.

8. Metha J, Metha P, Lawson D, Saldeen T. Plasma tissue plasminogenactivator inhibitor levels in coronary artery disease: correlation with ageand serum triglyceride concentrations.J Am Coll Cardiol. 1987;9:263–268.

Banfi et al VLDL-Mediated PAI-1 Signal Transduction in HepG2 215

by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from

9. Mussoni L, Mannucci L, Sirtori M, Camera M, Maderna P, Sironi L,Tremoli E. Hypertriglyceridemia and regulation of fibrinolytic activity.Arterioscler Thromb. 1992;12:19–27.

10. Stiko-Rahm A, Wiman B, Hamsten A, Nilsson J. Secretion of plasmin-ogen activator inhibitor-1 from cultured human umbilical vein endothelialcells is induced by very low density lipoprotein.Arteriosclerosis. 1990;10:1067–1073.

11. Eriksson P, Nilsson L, Karpe F, Hamsten A. Very-low-density lipoproteinresponse element in the promoter region of the human plasminogenactivator inhibitor-1 gene implicated in the impaired fibrinolysis of hy-pertriglyceridemia.Arterioscler Thromb Vasc Biol. 1998;18:20–26.

12. Sironi L, Mussoni L, Prati L, Baldassarre D, Camera M, Banfi C, TremoliE. Plasminogen activator inhibitor type-1 synthesis and mRNAexpression in HepG2 cells are regulated by VLDL.Arterioscler ThrombVasc Biol. 1996;16:89–96.

13. Santell L, Levin EG. Cyclic AMP potentiates phorbol ester stimulation oftissue plasminogen activator release and inhibits secretion of plasminogenactivator inhibitor-1 from human endothelial cells.J Biol Chem. 1988;263:16802–16808.

14. Kooistra T, Bosma PJ, Toet K, Cohen LH, Griffioen M, van den Berg E,le Clercq L, van Hinsbergh VW. Role of protein kinase C and cyclicadenosine monophosphate in the regulation of tissue-type plasminogenactivator, plasminogen activator inhibitor-1, and platelet-derived growthfactor mRNA levels in human endothelial cells: possible involvement ofproto-oncogenes c-jun and c-fos.Arterioscler Thromb. 1991;11:1042–1052.

15. Bergonzelli GE, Kruithof EK, Medcalf RL. Transcriptional antagonism ofphorbol ester-mediated induction of plasminogen activator inhibitor types1 and 2 by cyclic adenosine 39,59-monophosphate.Endocrinology. 1992;131:1467–1472.

16. Slivka SR, Loskutoff DJ. Regulation of type 1 plasminogen activatorinhibitor synthesis by protein kinase C and cAMP in bovine aorticendothelial cells.Biochim Biophys Acta. 1991;1094:317–322.

17. Tranque P, Robbins R, Naftolin F, Andrade-Gordon P. Regulation ofplasminogen activators and type-1 plasminogen activator inhibitor bycyclic AMP and phorbol ester in rat astrocytes.Glia. 1992;6:163–171.

18. Le Magueresse-Battistoni B, Pernod G, Kolodie L, Morera AM,Benahmed M. Tumor necrosis factor-alpha regulates plasminogen acti-vator inhibitor-1 in rat testicular peritubular cells.Endocrinology. 1997;138:1097–1105.

19. Halstead J, Kemp K, Ignotz RA. Evidence for involvement ofphosphatidylcholine-phospholipase C and protein kinase C in trans-forming growth factor-beta signaling.J Biol Chem. 1995;270:13600–13603.

20. Rydzewski B, Zelezna B, Tang W, Sumners C, Raizada MK. AngiotensinII stimulation of plasminogen activator inhibitor-1 gene expression inastroglial cells from the brain.Endocrinology. 1992;130:1255–1262.

21. van Hinsbergh VW, Vermeer M, Koolwijk P, Grimbergen J, Kooistra T.Genistein reduces tumor necrosis factor alpha-induced plasminogen ac-tivator inhibitor-1 transcription but not urokinase expression in humanendothelial cells.Blood. 1994;84:2984–2991.

22. Peiretti F, Fossat C, Anfosso F, Alessi MC, Henry M, Juhan-Vague I,Nalbone G. Increase in cytosolic calcium upregulates the synthesis oftype 1 plasminogen activator inhibitor in the human histiocytic cell lineU937.Blood. 1996;87:162–173.

23. Scott-Burden T, Resink TJ, Hahn AW, Baur U, Box RJ, Buhler FR.Induction of growth-related metabolism in human vascular smoothmuscle cells by low density lipoprotein.J Biol Chem. 1989;264:12582–12589.

24. Deeg MA, Bowen RF, Oram JF, Bierman EL. High density lipoproteinsstimulate mitogen-activated protein kinases in human skin fibroblasts.Arterioscler Thromb Vasc Biol. 1997;17:1667–1674.

25. Kusuhara M, Chait A, Cader A, Berk BC. Oxidized LDL stimulatesmitogen-activated protein kinases in smooth muscle cells and macro-phages.Arterioscler Thromb Vasc Biol. 1997;17:141–148.

26. Chautan M, Latron Y, Anfosso F, Alessi MC, Lafont H, Juhan-Vague I,Nalbone G. Phosphatidylinositol turnover during stimulation of plasmin-ogen activator inhibitor-1 secretion induced by oxidized low densitylipoproteins in human endothelial cells.J Lipid Res. 1993;34:101–110.

27. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurementwith the Folin phenol reagent.J Biol Chem. 1951;193:265–275.

28. Pascale A, Fortino I, Govoni S, Trabucchi M, Wetsel WC, Battaini F.Functional impairment in protein kinase C by RACK1 (receptor foractivated C kinase 1) deficiency in aged rat brain cortex.J Neurochem.1996;67:2471–2477.

29. Bradford MM. A rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dyebinding.Anal Biochem. 1976;72:248–254.

30. Rise P, Colombo C, Galli C. Effects of simvastatin on the metabolism ofpolyunsaturated fatty acids and on glycerolipid, cholesterol, and de novolipid synthesis in THP-1 cells.J Lipid Res. 1997;38:1299–1307.

31. Folch J, Lees M, Sloane-Stanley GHS. A simple method for isolation oftotal lipids from animal tissues.J Biol Chem. 1956;226:497–509.

32. Martinson EA, Trilivas I, Brown JH. Rapid protein kinase C-dependentactivation of phospholipase D leads to delayed 1,2-diglyceride accumu-lation. J Biol Chem. 1990;265:22282–22287.

33. Laemmli UK. Cleavage of structural proteins during the assembly of thehead of bacteriophage T4.Nature. 1970;227:680–685.

34. Seger R, Seger D, Reszka AA, Munar ES, Eldar-Finkelman H,Dobrowolska G, Jensen AM, Campbell JS, Fischer EH, Krebs EG. Over-expression of mitogen-activated protein kinase kinase (MAPKK) and itsmutants in NIH 3T3 cells: evidence that MAPKK involvement in cellularproliferation is regulated by phosphorylation of serine residues in itskinase subdomains VII and VIII.J Biol Chem. 1994;269:25699–25709.

35. Tatrai A, Lee SK, Stern PH. U-73122, a phospholipase C antagonist,inhibits effects of endothelin-1 and parathyroid hormone on signal trans-duction in UMR-106 osteoblastic cells.Biochim Biophys Acta. 1994;1224:575–582.

36. Schutze S, Potthoff K, Machleidt T, Berkovic D, Wiegmann K, KronkeM. TNF activates NF-kappa B by phosphatidylcholine-specific phospho-lipase C-induced “acidic” sphingomyelin breakdown.Cell. 1992;71:765–776.

37. Li Q, Tallant A, Cathcart MK. Dual Ca21 requirement for optimal lipidperoxidation of low density lipoprotein by activated human monocytes.J Clin Invest. 1993;91:1499–1506.

38. Cobb MH, Boulton TG, Robbins DJ. Extracellular signal-regulatedkinases: ERKs in progress.Cell Regul. 1991;2:965–978.

39. Thomas G. MAP kinase by any other name smells just as sweet.Cell.1992;68:3–6.

40. Dudley DT, Pang L, Decker SJ, Bridges AJ, Saltiel AR. A syntheticinhibitor of the mitogen-activated protein kinase cascade.Proc Natl AcadSci U S A. 1995;92:7686–7689.

41. Dawson RM. The formation of phosphatidylglycerol and other phospho-lipids by the transferase activity of phospholipase D.Biochem J. 1967;102:205–210.

42. Yano H, Nakanishi S, Kimura K, Hanai N, Saitoh Y, Fukui Y, NonomuraY, Matsuda Y. Inhibition of histamine secretion by wortmannin throughthe blockade of phosphatidylinositol 3-kinase in RBL-2H3 cells.J BiolChem. 1993;268:25846–25856.

43. Voyno-Yasenetskaya TA, Dobbs LG, Erickson SK, Hamilton RL. Lowdensity lipoprotein- and high density lipoprotein-mediated signal trans-duction and exocytosis in alveolar type II cells.Proc Natl Acad Sci U S A.1993;90:4256–4260.

44. Macarthur H, Hecker M, Busse R, Vane JR. Selective inhibition ofagonist-induced but not shear stress-dependent release of endothelialautacoids by thapsigargin.Br J Pharmacol. 1993;108:100–105.

45. Farrar WL, Thomas TP, Anderson WB. Altered cytosol/membraneenzyme redistribution on interleukin-3 activation of protein kinase C.Nature. 1985;315:235–237.

46. Ducher L, Croquet F, Gil S, Davy J, Feger J, Brehier A. Differentialexpression of five protein kinase C isoenzymes in FAO and HepG2hepatoma cell lines compared with normal rat hepatocytes.BiochimBiophys Res Commun. 1995;217:546–553.

47. Arts J, Grimbergen J, Bosma PJ, Rahmsdorf HJ, Kooistra T. Role of c-Junand proximal phorbol 12-myristate-13-acetate-(PMA)-responsive elementsin the regulation of basal and PMA-stimulated plasminogen-activator inhib-itor-1 gene expression in HepG2.Eur J Biochem. 1996;241:393–402.

48. Dennis EA. Phospholipases. InThe Enzymes.3rd ed. New York, NY:Academic Press, 1983;16:307–353.

49. Lin LL, Wartmann M, Lin AY, Knopf JL, Seth A, Davis RJ. cPLA2 isphosphorylated and activated by MAP kinase.Cell. 1993;72:269–278.

50. Cobb MH, Goldsmith EJ. How MAP kinases are regulated.J Biol Chem.1995;270:14843–14846.

51. Adams PD, Parker PJ. TPA-induced activation of MAP kinase.FEBSLett. 1991;290:77–82.

52. Blenis J. Signal transduction via the MAP kinases: proceed at your ownRSK. Proc Natl Acad Sci U S A. 1993;90:5889–5892.

53. Blumer KJ, Johnson GL. Diversity in function and regulation of MAPkinase pathways.Trends Biochem Sci. 1994;19:236–240.

216 Circulation Research July 23, 1999

by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from

54. Post GR, Brown JH. G protein-coupled receptors and signaling pathwaysregulating growth responses.FASEB J. 1996;10:741–749.

55. Lange-Carter CA, Pleiman CM, Gardner AM, Blumer KJ, Johnson GL. Adivergence in the MAP kinase regulatory network defined by MEKkinase and Raf.Science. 1993;260:315–319.

56. Leevers SJ, Marshall CJ. Activation of extracellular signal-regulatedkinase, ERK2, by p21 ras oncoprotein.EMBO J. 1992;11:569–574.

57. Lee H, Ghose-Dastidar J, Winawer S, Friedman E. Signal transductionthrough extracellular signal-regulated kinase-like pp57 blocked in differ-entiated cells having low protein kinase cb activity. J Biol Chem. 1993;268:5255–5263.

58. Havel RJ, Kane JP. Lipoprotein and lipid metabolism disorders. In:Scriver CR, Beaudet AL, Sly WS, Valle D, eds.The Metabolic Bases ofInherited Disease. New York, NY: McGraw-Hill; 1989:1129–1302.

59. Javitt NB. HepG2 cells as a resource for metabolic studies: lipoprotein,cholesterol, and bile acids.FASEB J. 1990;4:161–168.

60. Webb JC, Patel DD, Jones MD, Knight BL, Soutar AK. Character-ization and tissue-specific expression of the human “very low densitylipoprotein (VLDL) receptor” mRNA. Hum Mol Genet. 1994;3:531–537.

61. Allen S, Khan S, Al-Mohanna F, Batten P, Yacoub M. Native low densitylipoprotein-induced calcium transients trigger VCAM-1 and E-selectin

expression in cultured human vascular endothelial cells.J Clin Invest.1998;101:1064–1075.

62. Pian MS, Dobbs LG. Lipoprotein-stimulated surfactant secretion inalveolar type II cells: mediation by heterotrimeric G proteins.AmJ Physiol. 1997;273:L634–L639.

63. Banfi C, Rise P, Mussoni L, Galli C, Tremoli E. Linoleic acid enhancesthe secretion of plasminogen activator inhibitor type 1 by HepG2 cells.JLipid Res. 1997;38:860–869.

64. Nilsson L, Banfi C, Diczfalusky U, Tremoli E, Hamsten A, Eriksson P.Unsaturated fatty acids increase plasminogen activator inhibitor-1expression in endothelial cells.Arterioscler Thromb Vasc Biol. 1998;18:1679–1685.

65. Rao GN, Baas AS, Glasgow WC, Eling TE, Runge MS, Alexander RW.Activation of mitogen-activated protein kinases by arachidonic acid andits metabolites in vascular smooth muscle cells.J Biol Chem. 1994;269:32586–32591.

66. Rao GN, Alexander RW, Runge MS. Linoleic acid and its metabolites,hydroperoxyoctadecadienoic acids, stimulate c-Fos, c-Jun, and c-MycmRNA expression, mitogen-activated protein kinase activation, andgrowth in rat aortic smooth muscle cells.J Clin Invest. 1995;96:842– 847.

Banfi et al VLDL-Mediated PAI-1 Signal Transduction in HepG2 217

by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from

Fiorenzo Battaini, Claudio Galli and Elena TremoliCristina Banfi, Luciana Mussoni, Patrizia Risé, Maria Grazia Cattaneo, Lucia Vicentini,

Inhibitor Type 1 in Cultured HepG2 CellsMediated Signal Transduction and Plasminogen Activator−Very Low Density Lipoprotein

Print ISSN: 0009-7330. Online ISSN: 1524-4571 Copyright © 1999 American Heart Association, Inc. All rights reserved.is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation Research

doi: 10.1161/01.RES.85.2.2081999;85:208-217Circ Res. 

http://circres.ahajournals.org/content/85/2/208World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://circres.ahajournals.org//subscriptions/

is online at: Circulation Research Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. Further informationEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculation Researchin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

by guest on July 8, 2015http://circres.ahajournals.org/Downloaded from