Thrombus formation and platelet vessel wall interaction in the nephrotic syndrome under flow conditions

Thrombus formation and platelet-vessel wall

interaction in the nephrotic syndrome under

flow conditions.

J J Zwaginga, … , J J Sixma, T J Rabelink

J Clin Invest.

1994;

93(1)

:204-211.

https://doi.org/10.1172/JCI116947

.

Increased in vitro platelet aggregability and hypercoagulability are generally held to be main

determinants in the prethrombotic state in nephrosis. In vivo, however, thrombotic events

depend on the dynamic interaction of flowing blood with the vessel wall. The present study

confirms that aggregability of platelets of nephrotic patients is significantly increased by

mere stirring or by exogenous stimuli as adenosine diphosphate and arachidonic acid.

Moreover, the nephrotic patients have high von Willebrand factor and decreased red blood

cell deformability, which normally increase platelet-vessel wall interaction. However,

perfusion studies under well-defined flow conditions, in which anticoagulated nephrotic

blood was exposed to deendothelialized human umbilical artery segments and sprayed

collagen, showed normal platelet adhesion and only a modest increase in the deposition of

platelet aggregates. This suggests that some factor counteracts platelet-vessel wall

interaction under flow conditions in the nephrotic syndrome. When tissue factor associated

with endothelial extracellular matrix (ECM) was allowed to generate thrombin, perfusions

with nephrotic blood over this ECM resulted in a strong increase in fibrin generation. The

capacity of patient blood to form increased amounts of fibrin appeared strongly correlated

with the level of hyperfibrinogenemia. Platelet adhesion as well as aggregation in these

experiments was even decreased below control values. This suggests that fibrin coverage

may block the direct contact between blood platelets and matrix. We therefore also […]

Research Article

Find the latest version:

Thrombus Formation and Platelet-Vessel

Wall

Interaction

in

the Nephrotic

Syndrome

under

Flow

Conditions

Jaap Jan Zwaginga,**_{Hein A. Koomans,* Jan J.Sixma,t and Ton J. Rabelink*}

Departments of *Nephrology and $Hematology, UniversityHospital Utrecht, 3508GAUtrecht, The Netherlands

Abstract

Increased invitro plateletaggregabilityandhypercoagulability aregenerallyheldtobemaindeterminantsin the

prethrombo-ticstateinnephrosis. Invivo, however, thromboticevents de-pendonthedynamic interaction offlowingbloodwiththe

ves-sel wall. The presentstudy confirms thataggregabilityof plate-lets ofnephrotic patients is significantly increased by mere

stirringorby exogenous stimuliasadenosinediphosphateand arachidonic acid. Moreover,thenephrotic patients have high

vonWillebrand factor anddecreased red blood cell deformabi-lity, which normally increase platelet-vessel wallinteraction.

However,perfusion studiesunderwell-defined flow conditions,

inwhichanticoagulated nephroticbloodwasexposedto deen-dothelialized human umbilical artery segments and sprayed

collagen, showed normal platelet adhesion andonlya modest increase in thedeposition ofplatelet aggregates. This suggests that some factor counteracts platelet-vessel wall interaction

underflowconditions in thenephrotic syndrome.Whentissue

factor associated withendothelial extracellular matrix(ECM)

was allowed to generate thrombin, perfusions with nephrotic blood over this ECM resulted in a strong increase in fibrin

generation. The capacity ofpatient blood to form increased

amounts offibrin appeared stronglycorrelatedwith the level of

hyperfibrinogenemia.Plateletadhesionaswellasaggregation in theseexperimentswas evendecreased below control values. This suggests thatfibrincoverage may block the direct contact betweenbloodplateletsandmatrix.Wethereforealsostudied

the isolated effect of high

fibrinogen

onplatelet-vessel wall interaction byincreasing fibrinogen concentrations in normal blood.Modulationoffibrinogenconcentrations in normal blood could mimic all the observations in nephrotic blood: platelet aggregationinsuspension increasedwithincreasing

concentra-tionsof

fibrinogen,

whileplatelet adhesion and aggregate for-mation underflow conditions decreased. In perfusions over

tis-suefactor-richmatrix, fibrindeposition increased.Therefore,

ourobservations indicate thatnephrotic hyperaggregabilityin

suspension is not associated with increased platelet vessel wall-interaction underflowconditions. The latter is probably counteractedbyhighlevelsoffibrinogen.Our observations fur-ther suggest thathyperfibrinogenemiamaybe amajor throm-botic risk factor in nephrosis by inducing more fibrin deposi-tions.(J. Clin. Invest. 1994. 93:204-211.) Key words:

fibrino-Addresscorrespondence to Dr. A. J. Rabelink, Department of Nephrol-ogy and Hypertension (F03.226), UniversityHospital Utrecht, P.O. Box85500,3508GA Utrecht, The Netherlands.

Receivedfor publication 12 May 1993 and in revisedform 16 Au-gust 1993.

gen * hypercoaguability * nephrotic syndrome - platelets subendothelium

Introduction

Therisk for acute thromboembolic complications in patients with anephrotic syndrome is unequaled in clinical practice and

generallyattributedtocombined abnormalities inplateletand

coagulation-dependent hemostasis ( 1, 2). Hypercoaguability mayalsocontributeto thedevelopmentofatherosclerosis (3) andglomerular injury (4),other well-knowncomplicationsof thenephrotic syndrome.Variouslaboratoryabnormalities in

nephrosismay play a role in theprethromboticstate:

hyperfi-brinogenemia (5), increased levels of clotting factors (5, 6), decreased levels of antithrombin III (7), decreased fibrinolytic potential (5), and enhanced aggregability of platelets

sus-pended in plasma (8-1 1).

However, theselaboratoryabnormalitiesarehardto extrap-olate to in vivo hemostasis. Studies with perfusion systems,

which allow the dynamic interaction of flowing blood with (injured) vessel wall, have shown that the hemostatic

mecha-nismsare fora largepart flow regulated ( 12). Higherblood flow(shear rate) promotes the transport of platelets towards

the vessel wallbyred blood cells and therefore

platelet-depen-dent hemostasis. Higher platelet count, hematocrit, and

in-creased red bloodcellrigidityaddtotheefficiencyof this

plate-let transport(13, 14).Shear ratealso, modulatessubsequent plateletbindingtoadhesivemoleculesdepositedorpresentat

theinjuredvesselwall,e.g.,vonWillebrand factor(vWF) (15). Coagulation-dependenthemostasis is initiatedbyvessel wall-associated tissue factor and involves thrombin-induced fibrin depositionandplateletaggregation. Again, shearrateregulates these processes ( 16).

Thus far, the influence of flow on hemostasis in the

ne-phroticsyndrome hasnotbeen taken intoaccount.Therefore, whole bloodofnephroticpatientsandnormal controldonors

wasperfusedovermodelsof injuredvesselwall in

well-charac-terizedperfusionchambers.First,citrate-anticoagulatedwhole blood wasperfused overdeendothelialized human umbilical artery segmentsand over sprayed collagen (fibrillar type I).

Bothsurfacesinduceplateletadhesionaswell as platelet

aggre-gation; collagenasoneof theconstituents ofthedeepervessel wall,however, isamore potentstimulus ofplateletaggregation ( 17).Inthis waywedeterminedthatincreasedvWFlevels and red blood cell rigidity and the in vitro hyperaggregability in

nephrosis do not coincide with an increased platelet-vessel wall interaction under physiologic flow conditions. Second, low molecularweightheparin(LMWH)I_{was used as}

anticoag-ulant in additionalperfusions.Where LMWHinhibits

coagula-1. Abbreviations used in thispaper:ATIII, antithrombin III; ECM, extracellularmatrix; FPA, fibrinopeptide A; LMWH,low molecular weightheparin; PRP, platelet-rich plasma.

204 Zwagingaetal.

J.Clin. Invest.

© The American Society for ClinicalInvestigation, Inc.

tion during in vitro handling of the blood, it does allow throm-bin generation by tissue factor associated with extracellular matrix of cultured human umbilical vein endothelial cells (ECM). Theprothrombin-thrombin transition during perfu-sions can be monitored by measuring the release of peptide fragment1 +2 from prothrombin in the perfusate. Hypercoag-ulability in nephrosiswasillustratedby increased fibrin genera-tion (monitored by thrombin-dependent fibrinopeptide-A [FPA] release from fibrinogen) and fibrin deposition on the perfused matrix. Fibrinogen as platelet-binding and -interlink-ing protein and as substrate of thrombin-induced fibrin forma-tionis almost invariably increased in nephrotic patients ( 1, 2).

Inperfusions with control blood in which fibrinogen was artifi-cially raisedwecould furthersubstantiate thatnephrotic hyper-fibrinogenemia may play a major regulatory role in hemostasis andthrombogenesis.

Methods

Patients. Studieswere performed in two groups of 10 nephrotic pa-tients(characteristics in TableI)withoutimportant loss of renal func-tionasestimated from creatinine clearance (92±14 ml/min in group 1 and91±13ml/minin group 2). Blood of patients in group 1 was used in plateletaggregation studies, in red blood celldeformability measure-ments, and in citrateperfusions to substantiate within the same patient theeffect of in vitro abnormalitiesonactual platelet-dependent hemo-stasis. The blood ofpatientsin group 2wasusedformeasurementsof

clottingfactors andforLMWHperfusions(vide infra)torelate coagula-tionabnormalitiestocoagulation-dependentthrombusformation dur-ing perfusions. Parallel to every experiment with patient blood, control

experiments were performed with blood collected from healthy membersfrom the medicalstaff.Daily variations in experimental

con-ditions (e.g.,reactivityofperfusion surface)wereinthis way accounted for.Allpatientsand controlsubjectswerewithout medication forat

least2 wkbefore the study; they gave informed consent and the study was approvedbytheUniversity Hospital Ethical Committee for Stud-ies in Humans.

Red blood celldeformability. Red blood cell deformability was measured by alaser-diffractometer set-up, essentially as described in detailpreviously ( 18, 19).Inthis method laser light is diffracted bya

red blood cellsuspension. When shear stress (in the present study 30

N/iM2)isappliedtothesuspensionandtothe red bloodcells, it will

causeelongationof the diffraction pattern. Thiselongationisa mea-surefor the actualshear-induced red blood cell deformation.Higher

valuesindicatehigherdeformabilityof red blood cells. To study intrin-sic red blood cell qualities, 15 q1ofEDTA(2mM)-anticoagulated

whole blood was suspended in 2.25 ml of7% polyvinylpyrrolidone

TableI. PatientGroup Characteristics

GroupI Group 2

Histological diagnosis(biopsy proved, n):

Membraneousglomerulopathy 5 6

Membranoproliferativeglomerulonephritis 1 I

Mesangioproliferativeglomerulonephritis 2 1 Focalglomerulosclerosisandhyalinosis 2 1

Minimal change glomerulopathy I

Male/female(n) 7/3 8/2

Age range(yr) 18-56 18-48 Proteinuria(g/24 h) 7.6±3.9 5.7±2.1

Plasmaalbumin

(g/liter)

19±4 18±4

(PVP;360,000mol wt, Sigma ChemicalCo.,St.Louis, MO)in PBS

(1:10 vol/vol0.1 Msodiumphosphatebuffer, pH 7.4,in 150 mMol

NaCl).Tostudyinteraction ofthenephroticred cell withplasma

com-ponents the PVPwasdiluted inautologous plasma (50 mg/ml).PVP presenceresultinginahigh viscosityof the mediumwasnecessaryto

obtain shearstressvalues that willcausedeformation of red cells. Platelet aggregationinsuspension. Blood frompatientsand

con-trolswascollected into 1:10(vol/vol) 110 mM trisodium citrateby

clean venepuncture. Platelet-richplasma (PRP)wasprepared (10-min

centrifugationat150 g and20'C)and theplateletcount wascorrected

to

200,000/,gl

withautologousplateletpoorplasma(10-min

centrifuga-tionat3,000g and20'C).Within 2hafter bloodcollection, aggrega-tionstudies with thisPRPwereperformedat370C inaPAP-4

aggrego-meter(Bio/Data,Hatboro,PA)at900 rpm.AA(Bio/Data)in final concentrations of 1.5-, 0.75-, 0.375-, 0.250 mM and ADP(Dade,

Aquada, PA)inconcentrations of5-,2.5-, 1.67-,0.84

,M

wereusedas

stimuli.Inaddition,spontaneousplateletaggregationwasdetermined

by themerestirring of the PRP. To avoid interferencebyturbidity

differences duetohigh plasmalipids,wedeterminedplatelet aggrega-tionnotbythechangeinlighttransmission butbydirectlycounting

single (not aggregated) plateletsremainingin thePRPafter 3 min of

AAandADPstimulation andafter 15 min ofmerestirring. Single

plateletcounts weredeterminedbyaplatelet analyzer(model 810,

BakerInstruments,Allentown, PA)withsize aperturessetbetween3.2 and 16 rm3asdescribedbyVerhoeven etal. (20)andexpressedas

percentage of initialplateletcount.

Perfusionstudies with citratedblood.Blood frompatientsand

con-trolswascollectedinto 1:10(vol/vol) 110mMtrisodium citrate. To avoid that initial differences in hematocrit andplateletcountbetween

patientand control blood would influenceadhesion,perfusateswere

standardizedto aplateletcountof200,000platelets/glanda hemato-crit of 0.40. Standardizedperfusateswereobtainedbyadding autolo-gous PRP, plateletpoorplasma (PPP),and red blood cells(RBC) togetherin properquantities. PRP, PPP,andRBCwereobtainedby centrifugationof whole blood.

Perfusions withsteadyflowwerecarriedoutinamodified annular

perfusionchamber(21 )adaptedfrom theoriginaldescribedby Baum-gartner(12)orwith the doublerectangularperfusionchamber

devel-oped bySakkariassenetal. (22).In the annularperfusionchamber inverteddeendothelialized artery segments isolated from human umbil-ical cord wereusedas amodelofinjuredvessel wall. Arterieswere

mounted on two separatecentral rods, whichfitted in serially

con-nected annular perfusion chambers. The rectangular perfusion

chamber containedtwopairsof Thermanoxcoverslips (Lab-Tek

Divi-sion,MilesLaboratories,Naperville,IL)withsprayedfibrillarequine

collagen (18.5 ,ug/cm2; Collagen Reagent Horm, Hormon Chemie, Munich,FRG)asmodelofinjuredvessel wall(17).

Perfusionswerecarriedout ashasbeendescribedpreviously(16).

Inshort: Theperfusionchamberswererinsedwithprewarmed

(370C)

Hepes-bufferedsaline(HBS:Hepes 10mM, NaCl 0.15M,pH7.4).

Perfusates(volume 14ml)prewarmedfor 15 minat

_370C,

werethen recirculatedfor 5 minatvarious shearrates(300/s,toapproachlower "venous" shear rates,or1,300/s,as ashearratetypicalfor the

"arte-rial" vascularbed).After eachperfusionrunthearteries and

coverslips

wereremoved and the system rinsedthoroughlywith HBS. The

cover-slipsandvesselwallsegments werefixed with

respectively

0.5% and 2.5%glutardialdehydein PBS(1:10vol/vol0.1 M

sodiumphosphate

buffer,pH7.4in 150mM_NaCl),followedbyosmium tetroxide

(2%)

postfixation,ethanoldehydration,andsubsequentembeddinginepon (12).Thermanoxcoverslipswereremovedfrom the embedded colla-gen surfacebythermoshock.

Perfusionstudies with

LMWH-anticoagulated

blood. Exceptfor

the method of

anticoagulation

and the

_perfused

surface,

perfusions

werecarriedout asdescribed above. Blood forperfusion

experiments

wascollected into 1 / 10 vol ofLMWH(Fragmin,KabiVitrum,

Stock-holm,Sweden)200anti-XaU/mlsaline. TheLMWHconcentration

amount ofLMWHwere added topatientor tocontrol plasma (both anticoagulated with 20 mM trisodium citrate) showed comparable anti-Xaactivity in both plasmas. Perfusateswerestandardizedfor he-matocrit and platelets as described in the citrate anticoagulated perfu-sions.

As aperfusionsurface we used matrix from human umbilical vein endothelial cells (HUVECs). HUVECs were isolated from umbilical cordsaccordingtothe methodoriginallydescribed byJaffeetal.(23) with some minor modifications (24). HUVECs of the second passage were identified by their typical characteristics and subcultured on Thermanoxcoverslips that were previously coated and glutardialde-hyde fixed with gelatin (24). Cell monolayers were grown to con-fluence (50,000cells/cm2)in 5-7 d. HUVECs' tissuefactor deposition

intheir ECM,wasstimulated by addition ofPMA(SigmaChemical Co.)tothe culture medium(16).PMA wasdissolved in DMSO and subsequently diluted 1:1,000 vol/vol in the culture medium: 20 ng PMA/ ml. Additionof DMSO alone in the same concentration had no influence on the cells. After 10 h, matrix ofthe stimulated HUVEC was isolated by 0.1MNH40H.Coverslips with isolated ECM were washed three times with PBS and used forperfusion studies. The perfused coverslips were fixed, dehydrated, epon embedded, and "thermo-shocked"asdescribed above. Formicrophotographsofenface

mor-phology, coverslips were stained after perfusion and fixation, with May-Grunwald/Giemsa.

Evaluationofplatelet depositionandthrombusformation.Sections of the epon-embedded collagen, 1,gmthick, perfused matrix,and

ves-sel wall segmentswereprepared and stained with methylene blue and basic fuchsin for evaluation of plateletdeposition andaggregation.

Plateletdeposition andaggregationwereevaluatedbylightmicroscopy

(Dialux 20 EB, Leitz GmbH, Wetzlar, FRG)at 1,000magnification, usinga speciallyconstructedeyepiece micrometer in the ocular. At least 1,000intersection pointsat IO-Mm intervals, evenly distributed

overthe totalof theperfusedsurfaceswereevaluatedfor each section. Platelet adherencewasdetermined from the percentage ofartery, ECM of collagen surface directly covered by contact platelets, platelets

spreadout onthesurface andplateletaggregates(between2and 10,um

inheight), accordingtothecriteria described byBaumgartneretal. (25). Aggregate formationwasdefined and expressedasthe percentage ofspread platelets covered with aggregatedplateletsandwasthereforea

measure of platelet-platelet interaction independent of total surface coverage. Microscopically visible fibrindeposited onperfused ECM and above 2tminheightwasexpressedaspercentage surface coverage. Studies with reconstitutedperfusates. Additionalperfusionswith reconstituted bloodwereperformed toexclude thepossibility ofan

adhesion defect innephroticplatelets that would conceal adhesion pro-moting qualities of nephrotic plasma. Plasmaorwashedplatelets of fourpatientsofgroup 2wererespectivelymixed withplasmaorwashed platelets and red blood cellsfrom control subjects. The patients had high levels of vWF-Ag (2.11±0.43 U/ml) andfibrinogen(6.7± 1.3 g/ liter).Perfusions were performed induplicatefor 5 min at a shear of 1,300s-'overinverted umbilical artery segments. The artery segments wereevaluated asdescribed above.

Studieswith varyingfibrinogen levels. Theeffect of increased

fi-brinogenlevels, suchas occurin thenephroticsyndrome,on platelet-vesselwallinteraction and in vitro platelet aggregation was evaluated byaddingfibrinogentobloodofnormaldonors.This approachwas

technically more feasible than defibrinating blood of nephrotic pa-tients.Inthe later procedure it is hard to avoid substantial loss ofvWF

with the defibrination. Purified fibrinogen (Kabi Vitrum; dissolved to 400 mg/ml in and dialyzed for 6hagainst 0.15MNaClwas added to bloodof control subjects until concentrations of 2, 3, and 4 or 4.5 g/literwereobtained. Because purified fibrinogen contained high lev-elsof freeFPA, we could not measure the matrix-induced increase in

FPAinperfusates with added fibrinogen. Human albumin in 0.15 M

NaClin the same concentrationswasadded to perfusates to serve as control. Platelet aggregation studies after stimulation with ADP, AA, and spontaneous aggregationwerealsoperformed with control PRP to which

_fibrinogen

wasadded insimilarfashion.

Analytical methods. EDTA(2 mM) anticoagulatedwhole blood wascollected for determination ofhematocrit,meancell volume, and platelet count(model-S. Coulter, Harpenden, UK). Fibrinogen, fac-torsII, V,VII, and X, antithrombin III (AT III), and vWF-related properties were determined in blood collected into 0.1 vol 130 mM trisodium citrate. Reference plasmawas afreshly frozen plasma pool from 40 normal subjects, storedat-70'C. Plasma fibrinogen levelwas

determined accordingtoClauss (26). Factor II, V, VII, and X

coagu-lantactivitywasmeasured inanone-stageclotting assay with

commer-cially available deficient plasmas (George King Bio-Medical Inc., Overland Park,KS)assubstrate. AT IIIwasassayed with an ATIII-kit (Kabi-Vitrum) accordingtothe manufacturers instructions. vWF-re-latedantigen was determined byanELISA (27), the ristocetin cofactor activity was measured with formalin-fixed platelets and 1 mg/ml risto-cetin sulfate (H. Lundbeck & Co., Copenhagen, Denmark) (28). Albu-min,creatinine, and proteinuria were determined by standard labora-torytechniques.

Measurements in theperfusates. Aradioimmuno-assaykit (Mal-linckrodt Inc., St. Louis, MO)was used forFPA measurements. A enzymeimmunoassay (Enzygnost Fl + 2, Behring, Marburg, FRG) was used to determine the activation peptide ofprothrombin when thrombin is formed. Before, during, and after perfusion, samples (900

Ml)

from the

perfusate

werecollected and addedto100Ml of

anticoagu-lantmixture provided by the FPA-kit. FPA and fragment 1 +2 genera-tion wascalculated in these samples by subtracting baseline

concentra-tions fromconcentrations in the perfusateatthe end of the perfusions. FPA generation was expressed in nanograms per milliliter plasma, frag-ment 1 +2in nanomolar.

Statistical analysis. Data are expressed as mean±SD. The

signifi-cance ofdifferences between patient and control data obtained in various experiments were compared with the Student's t test for paired data. Linear regression wasapplied to test for correlations between fibrinogen levels and platelet-vessel wall interaction during the various experiments. Differences in the fibrinogen studiesweretested with AN-OVA with shear rate, stimulusconcentration andfibrinogen

concen-tration asindependent variables. Thedifferencesbetween themeans

were analyzed at1%and5%significance level by the leastsignificance

difference test.

Results

Aggregation studies within suspension. Non aggregated

(sin-gle)plateletsremaining in PRP after stimulation with ADPas

well as after stimulation with AA (TableII)were significantly

decreased in PRP ofnephrotic subjects (group 1). This in-creasedaggregabilityinnephroticPRP wasparticularly appar-ent at lowconcentrations of thestimulatingagent, when con-trolsubjectsdisplayedalmost noplateletaggregation. In

addi-tion, we found significant spontaneous aggregation in the nephrotic PRP after mere stirring.

Plateletadhesion and aggregate formation with

flowing

ci-trated blood. Except for increased levels of fibrinogen,

ne-phrotic subjects of group 1 (n = 10) also hadsignificantly in-creasedlevels ofvWF-antigenandvWF-ristocetin cofactor ac-tivity. In addition, red blood cell deformability was reduced

(Table IIIB). These abnormalities-known as platelet

adhe-sionpromotingfactors-were presentin all thepatients. Plate-let adhesion (expressed aspercentagesurface area covered with platelets) to deendothelialized umbilical artery, however, did

not differ between nephrotic patients (group 1 ) and control subjects (Table IIIA). Spread platelets were subdivided in spread platelets without and with aggregates > 2 ,m on top. Platelet aggregateformation,thusdefinedand expressed as per-centage oftotalplatelet coverage wasslightly butsignificantly

higher in the nephrotic subjects (Table III). Perfusions over

TableII.PlateletAggregationStudiesin Suspension

afterStimulationwithADP-andAAandSpontaneous Aggregation(PatientGroup1)

Patients (n=10) Controls (n= 10)

ADP-stimulated aggregation(uM)

5 19±8 24±16

2.5 35+14* 52±12 1.67 41±9* 60±14

0.84 52±9t 79±9

AA(mM)

1.5 14±12 36±26 0.75 26± 19 55±33 0.375 27±18* 82±6 0.250 39±23* 89±8 Spontaneousaggregation 69.1±20.9* 86.9±10.2 Plateletaggregation is expressed as percentage single (nonaggregated) plateletsof initial platelet counts in PRP after 3-min stimulation with ADP andAAaregiven at 3-min stimulation or after 15-min merestirring. All values expressed as means±SD. *P < 0.05,t_P

<0.01 comparedtocontrolsubjects.

umbilical arteries performed atlower (300/s) shear rates as modelforvenousflowresulted in anexpected lowered platelet

adherence.In our study theseperfusionsat lower shear did not enhance the difference in aggregate formation between

ne-phrotics andcontrolsubjects.Platelet adhesion to sprayed col-lagenat 1,300/s shear was less in both patients and control subjects than to umbilical arteries (Table IIIA). However,

ag-TableIII.Platelet Adhesion andAggregate Formation at UmbilicalArteriesand Sprayed Collagen (A) Related

toFactors Determining Platelet-Vessel WallInteraction

(B,PatientGroup1)

A Patients (n=10) Controls (n= 10)

Umbilical arteries (shear 1,300/s)

Adhesion (%) 51.9±16.1 55.3±14.6 Aggregate(%) 23.7±6.1* 20.8±6.6 Umbilical arteries (shear 300/s)

Adhesion (%) 38.7±5.4 44.5±5.5 Aggregate(%) 22.9±3.2* 18.8±6.2 Collagen(shear1300/s)

Adhesion (%) 33±7.5 35.9±9.7 Aggregate (%) 64.0±11.6* 58.5±12.9 B

Deformabilityindex 0.21±0.08* 0.37±0.02

Fibrinogen

(g/liter)

7.2±2.0V 2.8±0.5

vWF-ag(U/ml) 2.18±0.55* 0.73±0.08 vWF-RCof(U/ml) 1.59±0.39t 0.65±0.05 RBCdeformabilityindex is determined at 30N/m2.Abbreviations: vonWillebrand antigen Ag) and ristocetin cofactor

(vWF-RCof).Aggregateformation is expressed as the percentage of total platelet adhesionconsisting of aggregateformation(> 2

gm).

All

val-ues aregivenasmean±SD. *P <0.05,*P <0.01 compared to controlsubjects.

200

160

120

80

40

10

5

ATIII 11 V Vll X I

Figure 1.Means±SD of clotting factors (II, V, VII, X, I) and AT III of patient group 2. Values are expressed as percentage (ATIII, II, V, VII,X), or asg/liter (factor I(=fibrinogen). The solid lines refer to the reference range(mean±2SD).

gregateformation wasincreased in both patients and control subjects. Again, platelet adhesion on collagen between

ne-phrotics and control subjects did not differ, whereas aggregate formation was slightly increased in the nephrotic patients. Platelet counts at the end of the perfusions were always >

_I00,000/jAl.

Above thisvalue,platelet adhesion isnot deter-mined by varying platelet counts ( 16).

Platelet thrombi andfibrin deposition bymatrix-induced

thrombinformation. Levels of clottingfactors and

antithrom-bin-III were measured in the nephrotic patients of group 2. Except for a large increase in fibrinogen all were within or

moderately above (factor VII) the normal range (Fig. 1). In

perfusionswithLMWH-anticoagulatedblood ofthesepatients

overendothelial cellECM, platelet adhesionwasdecreasedas

compared tonormalcontrols(Table IVA). Aggregate forma-tiononadherentplateletswasalso reducedcomparedto

con-trols.Inthepatient perfusions,however,there was a morethan twofoldincrease in fibrindeposition directlyonthe ECM (Ta-ble IVA, Fig. 2). Additionally, FPA generation,as more

quan-TableIV.(A) PlateletAdhesion,AggregateFormation, andFibrin

DepositiononTissueFactor-rich Endothelial CellMatrixECM; (B) FragmentI+2andFPALevelsbeforeand Generation

during PerfusionswithLMWH-AnticoagulatedBlood

fromNephrotic Subjects (Group 2)and Control

ECM (shear 1300/s) Patients (n=10) Controls(n=10)

A Adhesion (%) 24.1±5.6* 41.8±5.5 Aggregate (%) 48.3±5.1* 57.5±2.2 Fibrindeposition(%) 60.0±17.3* 25.5±1.8

B Fl +2preperfusion(ng/ml) 3.6±1.7 1.8±0.9

Fl +2generation(ng/ml) 5.4±2.9 3.4±4.0 FPApreperfusion(ng/ml) 24.9±11.3t 7.8±2.8

FPAgeneration(ng/ml) 303±187* 107±76

Fibrinogen(g/liter) 6.2±1.9* 2.9±0.6

Adhesion is expressedaspercentage surface coverage.Aggregate for-mation is expressedasthe percentage of theplateletadhesion

con-sistingof aggregate formation(>2

Am).

Fibrindeposition(>2Mm)

expressed aspercentage of total surface. All valuesaregivenas

Figure

2.

Scanning

elec-tron

_micrograph

of fi-brindepositionson

tis-sue-factor rich endothe-lial cell matrix ina

nephrotic subject,

after

5-min

perfusion

with wholeblood

anticoagu-lated with LMWH.

titative measure ofthe

fibrinogen

conversion to

fibrin,

was

threefold higher thanincontrols (Table IVB). To correlatethe levelof hyperfibrinogenemia in patients (Table IVB) withFPA

generation and fibrin deposition we expressed patient data from allexperimentsaspercentage of values obtained in paral-lelperformed experiments withcontrolblood. Thus expressed, thrombin-induced FPA generation in patients correlated

stronglywiththesubstrate

(fibrinogen)

concentration (Fig. 3;r

=0.81,P <0.01). Initial

fibrinogen

levelsalsocorrelated with fibrindeposition (r=0.75,P<0.05). The

increased

levels of FPAinpatientbloodbeforeperfusions (baseline) possibly

re-flect increased fibrin generationinvivo oratbloodcollection

(Table IVB). Matrix-induced thrombin generation in

per-fusates, as reflected by the fragment 1 + 2 generation in the

perfusate, was not different between patients and controls (Table IVB).

Studies with reconstituted

perfusates.

Perfusion ofplatelets

ofcontrol subjects reconstitutedinnephrotic plasmaover in-vertedumbilical arteriesdid notdemonstrate increased platelet adhesion when compared to perfusions with control platelets reconstituted incontrol plasma: 58.2±5.4% vs. 54.1±4.6% of

surfacecoverage. Nor didcontrol plateletsinnephrotic plasma

show increased aggregate formation (> 2 ,um) compared to controlplateletsincontrol plasma: 47.4±5.8%vs. 49.5±5.5% oftotal plateletadhesion. Similarly, there was no significant difference in adhesion and aggregate formation whether ne-phroticplateletswerereconstitutedin their ownplasmaor con-trolplasma;plateletadhesionwasrespectively 59.1±6.6%

(ne-phrotic plasma)vs.55.8±5.1% (control plasma)andaggregate

formation

respectively 36.9±7.8% vs. 34.8±6.3%. However,

aggregateformationwas lower (P < 0.05) in thereconstituted perfusion studies with nephrotic platelets compared to those

withcontrolplatelets.

Fibrinogen

studies. As main

interlinking protein

of

acti-vated

platelets

andassubstrateofthrombin-inducedfibrin

gen-eration,

we

_investigated

whether

high

levels of

fibrinogen

in normal blood could would have similar effects as were ob-tained in theexperimentswithnephrotic patients.For this pur-pose, the

fibrinogen

concentration in control blood was in-creased. Hyperfibrinogenemia stimulated in vitroplatelet

ag-gregation (Table V).

Addition offibrinogento the

perfusate

resultedina

progressive

decreasein

platelet

adhesionto

umbili-FPA generation, %

300 r

200

100

0

0@0

0

0

0 0

0

100 200 300 400 500

Fibrinogen, %

Figure 3. FPA generation versus baseline fibrinogen in patientgroup 2,both expressed as percentageofparallelperformed control

experi-ment.Regression line,r=0.81, P<0.01.

Table V Nonaggregated Platelets Remaining in PRP of Control

Subjects after Stimulation withADP, AA, and Mere

Stirring;

AggregabilityasFunctionof

Different

Fibrinogen Concentrations in PRP

Fibrinogen concentration 2.0 3.0 4.5 (gliter)

ADP-stimulated aggregation

(,sM)

5 34±12 27±14 22±11*

2.5 62±9 47±10* 36±14*

AA(mM)

1.5 8±2 7±3 5±3 0.75 93±6 90±7 73±9*

Spontaneous aggregation 90±8 90±7 77±7* Higher platelet aggregability is expressed in a lower percentage (of initial platelet counts in PRP) single (nonaggregated) platelets re-maining in PRP after 3-min stimulation with ADP and AA or after

15-min mere stirring. Values are mean±SD of duplicate experi-ments. *P<0.05, t P < 0.01, compared to values obtained with PRPwith 2.0 g/liter fibrinogen.

calarteries (Table VI). Thiswasparticularlyapparent at high

shearrates at which platelet adhesion is normally increased.

Aggregateformationalsodecreased with increasing fibrinogen

concentrations both at 1,300/s and 300/s shear. Similarly, platelet adhesion andaggregateformation decreasedat increas-ing fibrinogen concentrations in LMWH perfusions over

ECM. Most of the perfusion surface in theseperfusions was coveredwithfibrin bothathighand lowfibrinogen

concentra-tions. However,atthehigh fibrinogen concentration larger fi-brindepositions had formed (Table VII).

Discussion

Increasedplatelet adhesion,aggregateformation,and coagula-tion at injured vessel wall have been proposed as important

Table VI.PlateletAdhesion and AggregateFormation onUmbilicalArteries inControlPerfusates

atDifferentConcentrations

of

Fibrinogen

Fibrinogenconcn. Adhesion Aggregate

(glliter) %

Shearrate 1,300/s

2.0 59.3±7.4 35.5±10.8

3.0 43.3+5.7* 22.2±7.0*

4.5 35.6±5.6* 19.3±6.3* Shearrate300/s

2.0 44.0±6.7 25.4±6.5 3.0 35.9±6.5* 19.7±8.9* 4.5 34.9±2.9* 12.6±3.9t

Adhesion isexpressedaspercentage surface coverage. Aggregate for-mation is expressedasthe percentage of theplateletadhesion

con-sistingof aggregate formation(>2

gm).

Valuesaregivenas

mean±SD ofduplicateexperiments. *P<0.05;*_P<0.01

com-paredtolowfibrinogen (2.0g/liter)concentration.

Table VII.Platelet Adhesion,Aggregate Formation, and Fibrin Depositions on ECM at

Different

Fibrinogen Concentrations inControl Perfusate

Fibrinogen Adhesion Aggregate Fibrin (> 2, > 5jim) Fibrin (> 5jim)

g/liter %

2.0 26±5 72±19 63±38 10±6

4.0

_21±5

_57±13*

_31±20*

_39±21*

Adhesion is expressed as percentage surface coverage. AG/TS, aggre-gateformation(> 2

gm)

(AG) expressed as percentage of total surface coverage(TS). Fibrin deposition (> 2

_gm)

expressed as percentage of total surface. Values are given as mean±SD of duplicate experi-ments. *P<0.05;compared to low fibrinogen concentration.

mechanisms inthedevelopmentof thrombosis(29) and

ath-erosclerosis (3).It hasbeenpostulated that increased platelet-vessel wall, platelet-platelet interaction, and hypercoagulabi-lity are also important features of the prethromboticstate in

patients

with

nephrotic

syndrome ( 1, 2). This notionhas been

extrapolated frominvitroobservations, showingabnormalities inplateletfunctionandcoagulation factors(5-11). Although flow isacrucial determinant of bothplatelet-and

coagulation-dependent

hemostasis,

noinformation is availableonthe

signif-icanceof these abnormalities under flow conditions in the

ne-phrotic syndrome.

The present study confirms that platelet aggregability in

suspension

is increased after stimulation with ADP, arachi-donicacidandbymerestirring (Table II). Platelet hyperaggre-gability inthenephrotic syndrome probablyhas a multifacto-rialgenesisandhas been associated with lowserum

albumin,

hypercholesterolemia,

and

hyperfibrinogenemia

(8, 9).

Only

fewstudies deal with the mechanism and relevance of these associatedchanges.Plateletfunctionafter stimulationwithAA

normalizes after correction of

hypoalbuminemia

in vitro and in vivo (9, 10). Hypoalbuminemia might increase cyclooxy-genase's

efficiency

toform theproaggregatorythromboxane A2

in the presence ofahigher concentration of free arachidonic acid that is normally boundtoalbumin (8, 9). However, plate-lethyperaggregabilitytootherstimuli suchas ADP,collagen, and ristocetin could not be relatedto

hypoalbuminemia (8,

11 ).Fibrinogen links GPIIbIIIareceptors onactivated

platelets

(30). Since

hyperfibrinogenemia

is characteristic of the

ne-phrotic syndrome, itseems

logical

to

postulate

arole for

fibrin-ogenin the platelet

dysfunction

in

nephrosis.

The present

ob-servationthatplatelet

aggregation

in

suspension

increases with

increasing concentrations of

fibrinogen

supportsthis

hypothe-sis.Meadeetal.

(31 )

reported

increased

aggregation velocity

at

higher

fibrinogen concentrations;

the maximum

aggregation

response, however, decreased with

increasing fibrinogen

con-centrations. The latter difference

might

be

explained by

the fact

that in ourstudy

single

(nonaggregated) platelets

were

mea-sured instead of

light

transmission which is

dependent

onthe

finalsize ofthe aggregates.

Increased platelet aggregation tests in suspension are

fre-quentlyusedasindicator for increased

platelet-vessel

wall

in-teractioninvivo

(i.e.,

underflow

conditions) ( 1, 2). Moreover,

vWFwasincreasedand red blood cell

deformability

decreased

in ourpatients (Table IIIB).Theseareconditions

respectively

associated with increased

platelet adhesion,

andmoreeffective

perfusions

with citrated blood wefoundanormal

platelet

ad-hesionand

only

a

modestly

increased

platelet

aggregation

(Ta-ble

IIIA).

The

perfusion

system with umbilical arteries is well

capable

of

platelet

adhesion above the 50% found inour

study

(22).

Therefore,

the absenceof increased

platelet

adhesion in

nephrotic

subjects

couldnotbe

explained by

a nearmaximal

platelet

adhesion stimulus.

Moreover,

perfusion

studies over

tissue factor rich endothelial cell

matrix,

in which

blood-anti-coagulation

with LMWH allowed matrix initiated thrombin

formation,

showed even less

platelet

adhesion and aggregate

formationin

nephrotic subjects

thanin controls. The decrease in

platelet

adherenceinthismodelwasassociated with

exten-sive fibrin

depositions

on the matrix

(ECM) (Table IVA).

Therefore,

the

expected

increasein

platelet-vessel

wall

interac-tionin

flowing nephrotic

bloodwasinsomeway counteracted

while fibrin

deposition

wasincreased.

Toexcludethe

possibility

thata

platelet

defect in the

ne-phrotic

syndrome

accounted fortheabsenceof increased

plate-let-vesselwall

interaction,

wealso

perfused

umbilical arteries with reconstituted

blood,

i.e.,

washedplatelets of control

sub-jects

incontrol

plasma

and

nephrotic

plasma,

and washed

plate-lets of

nephrotic subjects

in

nephrotic plasma

and control

plasma.

Similartothe whole blood

perfusion

studies

nephrotic

plasma

didnotincrease

platelet

adhesion

compared

tocontrol

plasma,

neither with control

platelets

norwith

nephrotic

plate-lets.Therewas adecreased aggregateformation inthe reconsti-tution

perfusions

with

nephrotic

platelets. However,

thiswas

thecasebothin

nephrotic plasma

aswellasin control

plasma.

Since

nephrotic platelets

are

_{hyperaggregable}

in

nephrotic

plasma

(in

aggregometerstudiesand inlesser degree inwhole blood

perfusion studies),

this

_phenomenon

is

likely

tobe due

tolossof

hyperaggregable platelets

(and

thusselection of

rela-tively refractory

platelets)

during

the

centrifugation

and

wash-ing

procedures

whicharenecessaryfortheisolation ofplatelets.

The presenceofa

platelet

adhesion defectthereforeseemsnot a

satisfactory explanation

forthe absenceofanincreased

plate-let-vessel wallinteraction.

Toevaluatetheisolatedroleof

hyperfibrinogenemia

apart

from other

factors,

we

performed

additional

experiments

with control blood in which the

fibrinogen

concentration was in-creased. In these

studies,

addition of

fibrinogen

to control blood couldmimicmostoftheobservations in

nephrotic

sub-jects. Fibrinogen again

promoted platelet aggregation

in

suspen-sion,

but it did not stimulate

platelet-vessel

wall interaction

underflow conditions. In

fact, perfusion

studies withcontrol

blood at

increasing fibrinogen

concentrations revealed a

de-crease in both

platelet

adhesion and aggregate

formation;

in

LMWH

perfusions

decreased platelet adherencewasreplaced

by

more

_pronounced

fibrin

deposition.

In asfarasthis

observa-tioncanbe

extrapolated

toour

nephrotic patients,

hyperfibrin-ogenemia

in

patients

may thus counteract vWF and red blood

cellenhanced

_platelet

adhesion.

Themechanism ofthesephenomena isnot clear yet. One

possible

explanation might

be that at low shear stress, as in the aggregometer,

fibrinogen-induced

platelet aggregates are not

exposed

to strong

dissolving

forces. Instead, in the perfusion system where much

higher

shear stressexistthese aggregates

might

bedissolved. Alternatively, aggregationin theperfusate

was

increased,

as inthe in vitroaggregation tests, so that less

platelets

were available for adhesion. Inagreement with this

notionseveralstudieshavereportedinvivoincreased levels of

the

_platelet

release

_{product #-thromboglobulin (32, 33).}

How-ever, we previously demonstrated thatplatelet availability is notdetermining platelet adhesion in theperfusion systemas

long as the platelet count in the perfusate is more than

100,000/I,as wasthe case in the presentstudy(16).Fibrin depositions bindto subendothelium andmightthus interfere with ligand systems in the subendothelium, therebydecreasing

platelet adhesion. The present observation thatplatelet adhe-sion decreases withincreasing fibrin depositions supports this notion. Finally, it is tempting tospeculate thatfibrinogenor

fibrin interfered with vWF-GPIIbIIIa mediated adhesive pro-cesses.Supportfor thishypothesis was derived from results in perfusions over laminin surfaces on whichplateletsadhere via the VLA-6 integrin receptorsupported by vWF-GPIb interac-tion and independent ofvWF-GPIIbIIIa interactions. Con-trary to the present perfusions, adhesion to this laminin was

not decreased byaddingextrafibrinogentotheperfusate

(un-publishedobservations, S. Endenburg, Utrecht).

Theincreased fibrindepositionattissue factor richmatrix wasaccompaniedby athreefold increasein FPAgenerationin

nephrotic perfusates above control values. These findings re-flect thrombin-induced fibringeneration and might be caused by increased thrombin formation, by less effective thrombin

inhibition or by more effective fibrinogen-fibrin transition. Generation offragment 1 + 2, as a measure of activation of

prothrombin into thrombin, was not different from control.

Thissuggeststhat theenhanced fibrin formation isnot deter-minedbyincreasedthrombin formation. This mightimplicate that thrombin is longer active at matrix perfused with

ne-phrotic blood. One cause of prolonged action of thrombin

might bea reductioninlevels of AT III. Although depressed levelsofAT IIIhave beenreported frequentlyinpatientswith heavyproteinuria(7), plasmalevelswerenormal in the pres-entstudy(Fig. 1).Inaddition, depressed levels ofAT IIIwould beassociatedwith increased formation ofthrombin, which was notapparentfromthe(normal) fragment 1 + 2formationin our study. Alternatively, one might postulate that up to the

supraphysiological plasmalevelsfoundin ourpatients, fibrino-genavailabilityas asubstrateisratelimiting for tissue

factor-dependent thrombinininducing fibrin formation. This

possi-bility is suggested bythe strongcorrelationbetweeninitial

fi-brinogen level and FPAgeneration inthe nephrotic subjects

(Fig. 3). Further support comesfrom in vivo observationsin

rabbits where intravascular fibrin deposition increased with

elevation oftheplasmafibrinogenlevel (34).Finally,although

perfusion timewas only 5

min,

we cannot exclude that de-creased or lesseffectivefibrinolysisinnephrosisalso may have

contributedsomewhat totheextensiveness ofthedepositions. Extrapolated from our experiments, the high risk for

thromboembolic complications in nephrotic subjectsmay

de-pendonfibrindepositions,even at relatively highshearrates. Invivofibrin depositioninnephroticsubjects is also suggested

byelevated baseline FPA levels in the present study (Table IVB), as well as in studiesbyothers (35). Fibrindeposition

normallyincreases withlowershear rates and promotes inclu-sionofRBCs and other cellelements( 12). Anincreased

ten-dencytoform these"mixed-thrombi" will therefore be most

obviousatlowershears in the venous vasculature or at sites of

turbulence,such asatheroscleroticlesions.Furthermore, fibrin

depositions

arealsopresentin atheroscleroticlesions, and are proposed tocontributetoatherosclerosis by causing endothe-lial celldisorganisation,bystimulating the proliferation of vas-cular smooth muscle cells and by providing an adsorptive

face for LDL accumulation(36, 37).Finally,innephrotic pa-tients elevated fibrinogenin renal veinblood correlated with

intraglomerular fibrin depositions,and maycontributetothe development ofglomerularinjury(38).

Insummary, the present studies challenge the concept of

increased in vivo platelet-vessel wall interaction in the ne-phrotic syndrome. We suggest that a plasma factor,

presum-ablyfibrinogen, counteractsthisinteraction. Therefore, in vi-troaggregationtestsdonotpredictplateletfunctionunderflow conditions in thenephrotic syndrome. Ourobservationsalso suggest that thethrombotic riskinnephrosis is determined by coagulation abnormalities rather thanby plateletdysfunction. Hyperfibrinogenemiamaybe themajorfactor determiningthe

thrombotic risk in nephrosis by inducing fibrin depositions proportionaltothefibrinogenlevel.Thisalsoimpliesthat ther-apy for hypercoagulability inthe nephrotic syndrome should focusonthisabnormality.

Acknowledgments

This study wassupportedby the DutchKidney Foundation.A.J.

Rabe-link isafellow of the Royal Dutch Academy ofSciences(KNAW).

References

1.Cameron,J.S. 1984. Coagulationandthromboemboliccomplications in thenephrotic syndrome. Adv. Nephrol.75-114.

2. Llach,F.1985. Hypercoaguability, renal vein thrombosisandother throm-botic complications of nephrotic syndrome. Kidney Int. 28:429-439.

3. Ross,R.1986. The pathogenesis of atherosclerosis:anupdate.N.Engl.J. Med. 314:488-500.

4.Barnes,J.L., andM. A.Ventakatachalam. 1985. The role of platelets and polycationic mediators in glomerular vascular injury. Semin. Nephrol. 5:57-68. 5.Thompson,C., C.D.Forbes, C.R. M.Prentice, andA.C. Kennedy. 1974. Changes in blood coagulation and fibrinolysis in the nephrotic syndrome. Q. J. Med. 43:399-407.

6.Kendall,A.G., R.E.Lohmann, andJ. B.Dossetor. 1971.Nephrotic syn-drome:ahypercoaguablestate.Arch.Intern.Med. 127:1021-1027.

7.Kauffman,R.H.,J.J.Veltkamp,N. H. vanTilburg,and L. A. van Es. 1978. Acquired antithrombin III deficiency and thrombosis in the nephrotic syndrome.Am.J.Med. 65:607-613.

8.Machleidt, C.,T.Mettang,E.Starz, J. Weber,T.Risler,andU. Kuhlmann. 1989.Multifactorialgenesis of enhanced platelet aggregability in patients with nephrotic syndrome. Kidney Int. 36:1119-1124.

9.Remuzzi, G., G. Mecca,D.Marchesi,M.Livio, G. De Gaetano,M. B. Donati,and M.J.Siver. 1979. Platelethyperaggregabilityandthenephrotic syn-drome.Thromb.Res. 16:345-354.

10. Jackson, C. A., M.Creaves,A.D. Patterson, C. B. Brown,and F. E. Preston. 1982.Relationship between plateletaggregation,thromboxane synthesis andalbumin concentration in nephrotic syndrome.Br.J. Haematol. 52:69-75. 11.Bennet, A., and J. S. Cameron. 1987.Platelethyperaggregability inthe nephrotic syndrome which isnotdependentonarachidonic acid metabolismor onalbuminconcentration. Clin.Nephrol.27:182-188.

12. Baumgartner,H. R. 1973. The role of blood flow inplateletadhesion, fibrindepositionandformation of mural thrombi. Microvasc. Res. 5:167-179.

13.Turitto,V.T.,H.J.Weiss,and H. R.Baumgartner. 1979.Rheological factorsinfluencingplatelet interaction with vessel surfaces. J. Rheol. 23:735-749. 14.Aarts,P. A. M.M.,R. H.Heethaar,and J. J.Sixma. 1984. Red blood cell deformabilityinfluences platelets-vesselwallinteractioninflowingblood.Blood. 64:1228-1233.

15.Sixma,J.J.,K.S.Sakariassen,and P. A.Bolhuis. 1982.The roleof factor VIII-vonWillebrandfactor in adhesion of platelettothe vessel wall. In Pathobiol-ogyoftheEndothelial Cells.H.L. Nossel and K.Vogel,editors.)AcademicPress, Inc.,New York. 139-152.

16.Zwaginga,J.J.,J.J.Sixma, andP.G. de Groot. 1990. Activation of endothelial cells inducesplateletthrombus formationontheirmatrix:studiesofa

newin vitrothrombosis model with low molecularweightheparinas anticoagu-lant.Arteriosclerosis. 10:49-61.

17.Houdijk, W.P.M., P.G.deGroot,P. F. E. M.Nievelstein,K.S. Sakarias-sen,andJ. J.Sixma.1986.VonWillebrand factor and fibronectin butnot throm-bospondinareinvolved in platelet adhesiontothe extracellularmatrix ofhuman vascularendothelialcells.Arteriosclerosis. 6:24-33.

18. Clark, M.R., N. Mohandas, andS.B.Shohet. 1983. Osmotic gradient ektacytometry:comprehensive characterizationofredcellvolumeand surface maintenance.Blood. 61:899-910.

19. Hardeman,M.R., P.Goedhart,and D. Breederveld.1987. Laser diffrac-tionellipsometry of erythrocytes under controlledshear stressusingarotational viscosimeter.Clin. Chim.Acta. 165:227-234.

20. Verhoeven, A.J. M.,M. E.Mommersteeg, J. W. N.Akkerman. 1984. Metabolicenergyis required in human plateletsatany stageduring optical aggre-gationandsecretion.Biochim. Biophys.Acta.800:242-250.

21.Nievelstein,P. F.E. M., Ph. G. de Groot, P. D'Alessio, F.H.G.Heynen, E.Orlando and J. J.Sixma. 1990.Plateletadhesiontovascularcells:the role of exogenous vonWillebrand Factor in platelet adhesion. Arteriosclerosis. 10:462-469.

22. Sakariassen,K.S.,J. D.Banga, Ph.G. de Groot, andJ. J.Sixma. 1984. Comparisonofplatelet interaction with subendothelium of human renal and umbilical arteries and the extracellular matrix produced by humanvenous endo-thelial cells.Thromb. Haemostasis. 52:60-65.

23. Jaffe, E. A., R. L. Nachman, C. G. Becker, and C. R. Minick. 1973. Cultureof human endothelial cells derived from umbilicalcordveins: identifica-tionof morphology and immunologic criteria.J. Clin.Invest. 52:2745-2756.

24. Willems, C., G. C. B. Astaldi, P.G. de Groot, M. C.Janssen, M.D. Consalves,W. P.Zeylemaker, andJ. A. vanMourik.1982.Media conditioned by cultured humanvascularendothelialcellsinhibitthe growth of vascular smooth muscle cells. Exp. Cell. Res. 139:191-197.

25.Baumgartner, H. R., R.Muggli,T. B.Tschopp,and V. T. Turitto. 1976. Platelet adhesion, release and aggregation in flowing blood: effects of surface propertiesand plateletfunction. Thromb.Haemostasis.35:124-138.

26. Clauss, A.1957.Gerinnungsphysiologische schnell Methodezur Bestim-mung des Fibrinogens. Acta Haematol. 17:237-246.

27. Hamer, R. J., J. A. Koedam, N. H.Beeser-Visser, andJ. J.Sixma. 1986. HumanfactorVIIIpurification fromcommercialfactor VIIIconcentrate, charac-terization, identification and radiolabeling. 1986. Biochim. Biophys. Acta. 873:356-366.

28. MacFarlane, D. E., J.Stibbe,E. P.Kirkby, M. B.Zucker,R. A.Grant,and J. McPherson. 1975. Amethod for assayingvon Willebrand factor (ristocetin cofactor). Thromb.Diath.Hemorrh. 34:306-308.

29.Sixma,J.J.,and J. Wester. 1977.Thehemostatic plug. Semin. Hematol. 14:265-299.

30. Niiya, K., E. Hodson, R. Bader, V.Byers-Ward,J.A.Koziol,E. F.Plow, and Z.M.Ruggeri. 1987.Increasedsurface expressionof the membrane glyco-protein lIb/IIIa complexinduced by platelet aggregation: relationship tothe bindingof fibrinogen and platelet aggregation. Blood. 70:475-483.

31.Meade,T. W., M.V. Vickers,S.G. Thompson, andM. J.Seghatchian. 1985. The effectof physiological levels of fibrinogen on platelet aggregation. 1985. Thromb.Res. 38:527-534.

32.Adler, A. J., A. P.Lundin,M. V. Feinroth, E. A.Friedmann,and G. M. Berlyne. 1980. Beta-thromboglobulin levels in nephrotic syndrome.Am.J.Med. 69:551-554.

33.Kuhlmann, U., J.Streurer, K. Rhyner,A.von Felten,J.Brinner,and W. Siegenthaler. 1981. Plateletaggregationandbeta-thromboglobulinlevels in ne-phrotic patients withandwithout thrombosis.Clin. Nephrol. 15:229-235.

34. Gurewich, V., B.Lipinsli,and E.Hyde. 1976. The effectof thefibrinogen concentration and theleucocyte countonintravascularfibrindeposition from soluble monomer complexes. Thromb. Haemostasis. 36:605-614.

35.Sagripanti, A.,A.Cupisti,M.Ferdeghini,E.Pinori, and G. Barsotti. 1989. Molecular markers of hemostasis activation innephrotic syndrome. Nephron. 51:25-28.

36.Thompson,W. D.,andE. B.Smith. 1989. Atherosclerosis andthe coagula-tion system.J. Pathol. 159:97-106.

37.Ishida,T.,andK.Tanaka. 1982.Effects of fibrinand fibrinogen-degrada-tion products on the growth of aortic smooth muscle cells in culture. Atherosclero-sis. 44:161-174.