Inhibitor of the factor VIIa tissue factor complex is reduced in patients with disseminated intravascular coagulation but not in patients with severe hepatocellular disease

Inhibitor of the factor VIIa-tissue factor complex

is reduced in patients with disseminated

intravascular coagulation but not in patients

with severe hepatocellular disease.

M S Bajaj, … , R B Wysolmerski, S P Bajaj

J Clin Invest.

1987;

79(6)

:1874-1878.

https://doi.org/10.1172/JCI113030

.

Inhibition of Factor VIIa-tissue factor activity by a plasma component(s) that requires factor

Xa has been described recently. In this communication, we have developed a specific

radiometric assay (which utilizes 3H-Factor IX and is sensitive to less than 1% of plasma

level) for this inhibitor and have measured its activity in various disease states. Strikingly,

the levels of this inhibitor were found to be normal in patients with advanced chronic

hepatocellular disease but low in patients with disseminated intravascular coagulation

(DIC). When endotoxin was used to induce DIC in rabbits, the levels of this inhibitor fell by

25-90%. Human umbilical vein endothelial cells (HUVE), bovine pulmonary artery

endothelial cells, and a human hepatoma cell line (HepG2) all synthesized and secreted

this inhibitor, whereas a promyelocytic cell line (HL-60) did not and a monocytic cell line

(U937) appears to synthesize only small amounts. When ammonium sulfate-fractionated

human plasma and serum-free conditioned media from both HUVE and HepG2 cells were

electrophoresed on sodium dodecyl sulfate acrylamide gels, two activity peaks

corresponding to Mr approximately 45,000 and Mr approximately 33,000 were eluted in

each case. These observations suggest that (a) the inhibitor is consumed in DIC and that (b)

endothelial cells (or other cells) synthesize sufficient amounts of this inhibitor in vivo to

compensate for any decreased production by liver cells.

[…]

Research Article

Find the latest version:

Rapid Publication

Inhibitor of the Factor Vila-Tissue Factor Complex Is Reduced

in

Patients

with

Disseminated Intravascular Coagulation but Not in Patients

with Severe

Hepatocellular

Disease

Madhu S. Bajaj,Satya V. Rana, RobertB.Wysolmerski, andS.PaulBajaj

SectionofMedicine, Biochemistry, and Pathology, St. LouisUniversitySchoolofMedicine,St. Louis,Missouri63104

Abstract

Inhibition ofFactorVIIa-tissuefactoractivity byaplasma

com-ponent(s) thatrequires factor Xa has been described recently.

In thiscommunication,wehavedevelopedaspecificradiometric

assay(which utilizes3H-Factor IX and is sensitive to<1% of

plasma level)forthis inhibitorandhave measured itsactivityin

various diseasestates.Sikingy,thelevels of this inhibitorwere

foundtobe normal inpatients with advanced chronic

hepato-cellular diseasebutlow inpatientswithdisseminated intravas-cular coagulation (DIC). When endotoxinwas used to induce

DICinrabbits, the levelsofthis inhibitor fellby25-90%. Human umbilical veinendothelialcells(HUVE),bovinepulmonary

ar-tery endothelialcells,andahumanhepatomacell line(HepG2)

allsynthesizedand secreted thisinhibitor,whereasa promyelo-cyticcell line(HL-60)did notandamonocyticcellline(U937)

appears to synthesize only small amounts. When ammonium

sulfate-fractionated humanplasmaand serum-free conditioned

media frombothHUVE andHepG2cellswereelectrophoresed

onsodiumdodecyl sulfateacrylamide gels, twoactivity peaks

correspondingtoMr - 45,000andMr - 33,000 _wereeluted

ineachcase.These observations suggest that(a)the inhibitor

isconsumedinDIC andthat(b) endothelialcells(orothercells) synthesizesufficient amounts of this inhibitor in vivo to

com-pensate foranydecreasedproduction byliver cells.

Introduction

The tissue factorpathwayofbloodcoagulationmayplaya

pri-maryrole inhemostasis, becauseFactorVIla-tissuefactor

(VIIa/

TF)' complex notonly activates FactorXbutalsoFactorIX,

thusbypassing the contactactivation reactions ofthe classical

intrinsic pathway (1, 2). Furthermore, substantial evidence exists

AddresscorrespondencetoDr. S. Paul Bajaj,Hematology Division, St.

Louis UniversityMedicalCenter, 1325 South Grand Boulevard, St. Louis, MO63104.

Receivedfor publication21 January1987.

1.Abbreviations usedinthispaper:VIIa/TF, complex of FactorVIa and tissue factor, BPAE, bovine pulmonary artery endothelial cell;

DEGR-CK, dansyl-glu-gly-arg-chloromethylketone;DIC, disseminated intravascularcoagulation; HPP,human pooled plasma; HUVE, human umbilicalvein endothelialcell; RPP,rabbitpooled plasma.

that there is an inhibitor inplasmathat turnsoffthecoagulation initiated by the VIIa/TF complex (3-5).This inhibitor couldbe aphysiologically important variableinfluencingnormal

coag-ulationaswellaspathologicalthrombus formation. The present

studywas undertaken toidentifythecellularsite(s)ofsynthesis of this inhibitor andto delineatepathological conditions that

influence its plasmalevel. Data arepresentedthat suggest that

the inhibitor issynthesizedinliver and endothelial cells and is

consumed

during

disseminated intravascularcoagulation

(DIC).

Methods

Materials. Human vein endothelial cells(HUVE)wereharvestedfrom humanumbilical cords by the method ofJaffeetal. (6). Bovine pulmonary arteryendothelial cell line (BPAE)describedby Del Vecchio and Smith (7) and the HL-60 cell line established from the peripheral blood ofa

patientwithacutepromyelocytic leukemia (8)wereobtained from the American Type Culture Collection, Rockville, MD. The U937 cell line established from thepleuraleffusion ofapatientwithhistiocytic lym-phoma(9) wasoriginally provided by Dr. Hillel S. Koren, Duke Uni-versity, Durham, NC. The HepG2 cell line derived from a human

he-patoblastoma(10)wasprovided byDr.Jung SanHuang of St. Louis

University,St.Louis,MO. Human brain tissue factorwaspreparedas

describedpreviously (I1); thesuspensionwascentrifugedasearlier (11)

andthe supernatant that clotted recalcified plasma in 24s wasused throughout. Chemicals forelectrophoresiswerepurchased from Bio-Rad Laboratories, Richmond, CA. Tissue culture media and supplements wereobtainedfrom K. C. Biological, Inc., Lenexa, KS, and the hereditary

clotting factor-deficient plasmas were obtained from George King Biomedical, Overland Park, KS. Endotoxin (Bacto lipopolysaccharide,

E. coli 0127, B8)wasobtained fromDifcoLaboratories Inc., Detroit,

MI.Heparin andcycloheximide were received from Sigma Chemical Company, St. Louis, MO. Dansyl-glu-gly-arg-chloromethyl ketone (DEGR-CK) was obtained from Calbiochem-Behring Corp.,LaJolla, CA, andNa(3H)borohydridefromAmersham Corp., Arlington Heights, IL. Bovine plasma was obtained from a local slaughterhouse. Other chemicalswereof the bestcommercially available grade.

Proteins.HumanFactorVII(12),Factor IX(13),and Factor X(13)

werepurified to homogeneity as described previously. Sialyl3H-Factor

IXwas prepared bythe general technique of Van Lenten and Ashwell ( 14). The procedure was slightly modified in that the reagent NaIO4 was usedat atwofold molar concentrationtothesialic acid content of Factor

IX.Specificactivity of the preparation was 2 X

10'

cpm/mg and the radiolabeled protein retained 91% oftheclottingactivity ofthe nonlabeled

control. FactorXa wasprepared asdescribed (13); it was coupled to

Sepharose4Bbyaprocedureasoutlined for Russellsviper venom ( 13).

FactorVIIawasobtained as earlier except insoluble factor Xa (Sepharose-Xa)wasusedinstead of the soluble factor Xa as the activator ( 15). The resin wasremoved bycentrifugation and the supernatant was passed

over asmallChelex-100 column. Factor

VIla

obtained in this manner hadaspecific activity of 35

U/pg;

it was kept at4VCand used within

24 hof preparation. DEGR-Xa was prepared by incubating Factor Xa

1874 M.S.Bajaj,S. V.Rana,R.B. Wysolmerski,and S. P.Bajaj

J.Clin. Invest.

© TheAmericanSociety for Clinical Investigation, Inc.

0021-9738/87/06/1874/05

$1.00

(- 2AM)with30-fold molar excess of DEGR-CK. The conditions em-ployedincluding the removal of free label were exactly the same as

out-lined for Factor XIa (16). DEGR-Xa had<1%of the clotting activity of unmodified Factor Xa. Anti-FactorXand anti-protein C antibody wereprepared as previously described (17).

Gel electrophoresis. Sodiumdodecyl sulfate (SDS) gel electrophoresis wasperformed according to the method ofLaemmli, utilizing 12% acryl-amide gels (18).

Cellculture.HUVE and BPAE cells were grown as previously de-scribed(19).HUVEcellsobtained from three umbilical cords were rou-tinely used for seeding one T-25 cm2 flask. The cells grew to confluence within 4-5 d inM199 supplemented with 20% heat-inactivated fetal calf serum(FCS), 90

Mg/ml

heparin,and 35

jg/ml

endothelial cell growth supplement. Cells were determined to be of endothelial cell origin as described earlier by Wysolmerski and Lagunoff (19). Cells (2-4 d post-confluent) were washed five times with the culture medium (without heparin and FCS) and then cultured overnight (20-24 h) in serum- and heparin-free medium. Conditioned medium was collected, centrifuged

to removeanydebris, and frozen at -20'C until used.BPAEcells were grown in Eagle's minimal essential medium supplemented with 2 mM glutamine, 10% FCS, 50 U/ml penicillin, and 50

jg/ml

streptomycin. HepG2 cells were cultured in Dulbecco's modified Eagle's medium sup-plemented with 10%FCS and 1%nonessential amino acids.Serum-free

conditioned media fromBPAEcells(6 dpostconfluent)andHepG2 cells (2 dpostconfluent) were obtained in a similar manner as for HUVE cells. The HL-60 and U937 cells weremaintained in RPMI-1640 medium supplemented with10%FCS as previously outlined (20). To collect con-ditioned media, the cells werewashed three times, resuspended to a density of - 2X106cells/ml inserum-freeRPMImedia, and cultured

for - 22 h.Allcellsweremaintained at370Cin ahumidified atmosphere of5%CO2in air.Inexperiments where cycloheximide was used, its final concentration was I

Ag/ml

in the medium.

InductionofDIC in rabbits. Our experimental protocol for inducing DIC inrabbitswasthesameasoutlined in detail by Rapaportand as-sociates (21, 22) except thatweuseddexamethasone sodium phosphate (Merck Sharp & Dohme, West Point, PA) instead of cortisone acetate. Six New Zealand rabbits (- 2.5 kg)weredivided intotwogroups. All threerabbits in GroupAreceived 1mlofisotonic saline intramuscularly

dailyfor4d.The threerabbits in GroupBreceived4mg of

dexameth-asone(insteadofsaline) intramuscularlyfor4d.Onday 5 and 24 hr later(day 6), one rabbitfrom each group received intravenous injections ofsaline,and theremainingtwofrom each group received intravenous

injectionsofendotoxin (75

Ag/kg).

Blood samples from theearveins were drawn from all animals 10 min before the first injection of saline

ordexamethasonetoestablish base line levelsoffibrinogen, platelets, FactorV, andFactorVIII.Bloodsampleswereagain collected (starting

on

day

5) 30 min before and 4 and 24 h after the

injections

of saline

orendotoxin. Additionalsampleswereobtained from those animals that survived after the finalinjectionof salineorendotoxin.

Plasma specimens. Patientspecimenswereobtainedatthe St. Louis

UniversityMedical Center. Use of volunteer blood donorswasapproved

by theHumanSubjectsCommittee of St. LouisUniversity.Allsamples

obtainedwereensuredtobe free ofheparin.Patients with chronic he-patocellular disease hadoneof the following: historyoflongstanding

alcohol abuse, chronic activehepatitis, hemochromatosis,or

postnecrotic

cirrhosis. The diagnosticcriteria for thesepatientswere(a)

history

of predisposing cause, (b)stigmata of chronichepatocellulardiseaseon

physicalexamination, (c) abnormal laboratorytests suchasmarkedly decreasedserumalbumin levels, elevatedserumbilirubinlevels,prolonged

prothrombin time, anddecreased proteinCantigen(<47%of normal

asmeasuredbyanelectroimmunoassay),and(d)histopathological

ev-idence of chronic liver diseasebybiopsyorautopsy in five of thepatients.

Patients with DIC hadoneof the following:(a)gram-negative sepsis

(seven patients), (b) gram-positive sepsis (two patients), or(c)severe trauma(onepatient).The laboratorydiagnosticcriteria for these

patients

wereprolonged prothrombintime,prolonged activatedpartial

throm-boplastin time,decreasedplateletcount, and elevated

fibrin/fibrinogen

degradation products (> 10

Ag/ml,

asmeasuredby

Thrombo-Webco

test). Patientsonlong-term warfarintherapyduetoatrialfibrillation, prosthetic heart valves, arterialorvenous thrombosis, orpulmonary thromboembolism whowere onstablemaintenance dosewereselected fortheinhibitoranalysis.TheproteinCantigenin thesepatientswas <53% of normal.Lupusanticoagulant patientswereselected based upon the criteriadescribed elsewhere(23).Two of thepatientshadaclinical history of thrombotic disease. Plasma samples ofpatientswith deep ve-nousthrombosisasdiagnosed byeithervenographyorimpedance pleth-ysmographywerecollectedbeforereceivinganytreatment.

VIIa/TFinhibitor assay.Theassaywasconducted intwosteps and all reagentswerediluted in 0.05 MTris,0.15 MNaCl,pH 7.5

(Tris/

NaCl)containing 1 mg/ml of bovineserumalbumin(Tris/NaCl/Alb).

Inthefirst step, 50 ul of FactorX(70

ug/ml)

wasincubatedwith 100

til

of Factor VII(25

gg/ml),

50Mlof diluted tissuefactor, and 31Mlof

75 mMCaCI2at

370C

for 10 min.Duringthis incubationtime,Factor

VIIwascompletely convertedtoFactorVIa,and>90%of FactorX wasconverted to Factor Xa. The reactionwasquenched by the addition of 36Mlof 75 mM Na2EDTA. In the second step,

15-Ml

aliquotsof the aboveincubationmixturewereaddedtovarious tubescontaining50Ml

of 3H-Factor IX (5

jg/ml)

and 20

M1

of different dilutions of plasmaor testmaterial. The reactionwasinitiatedbytheaddition of 15Mlof 70

mMCaC12. The tubeswereincubatedfor the desired length of timeat

370Catwhichpoint100

Al

of coldstoppingbuffer

(Tris/NaCI

containing 50mMbenzamidine, 50 mMEDTA, and 5 mg/ml bovineserum

al-bumin)wasadded. Trichloroacetic acid(TCA)solublecountsinthese tubeswerethendeterminedasdescribed earlier (24, 25). Background

counts

('

2% oftotal) weresubtracted from thesamplecounts to

de-termine the percent TCA-solublecounts.Completeactivation of Factor IXyielded - 35% TCA-solublecounts.Humanpooledplasma from 20

donorswasarbitrarily assignedtocontain 100U/mlof

VIIa/TF

inhibitor

activity.Thus,a testsample containing10 U/ml of inhibitoractivityis

equalto 10%ofplasmalevelsof inhibitor.

Results

Quantitative

assay

for

VIIa/TF

inhibitor activity.

The present

assay was

developed

based upon the

rationale

that TF should be the

only

limiting

reagent, and that the smallamounts

of

Factor IX

contributed by plasma in

the assay shouldnot

influence

the

fraction

of

3H-Factor

IX

activated

at

early

time

points.

Both

conditions

were metwhena

diluted

TF

preparation

anda

_high

Factor VII

concentration

were

employed,

which in the absence of the inhibitor

yielded

10% TCA-soluble counts

(after

30

min

incubation; Fig.

1, inset

[-])

at a3H-Factor IX

concentration

of

44 nM

(-

5

times

below the Km

value,

reference

2). Thus,

when the

concentration of tissue factor

was

doubled,

the

initial

rate

of

3H-Factor

IX

activation

was

correspondingly

increased

(Fig.

1, inset

[X]).

Similarly,

when nonlabeled Factor IX ata

concentration of

17 nM

(the

amount

of

Factor IX

contributed

by

10%

plasma

in the assay is 7

nM)

wasalso present in the

assay,

it

hadno

detectable influence

onthe

percent activation

of 3H-Factor

IX

(Fig.

1,

inset

_[01]).

Also,

the

inhibitory

activity

was not

increased

whenFactor X

concentration

wasincreased

by

a

factor

of

3/2. Furthermore,

inhibition

was notobserved whenFactor Xa

formed in

the

first

step and Factor X

present

in normal

plasma

wereneutralized

by

anti-FactorXantibodies

before

conducting

the second step

of

the assay. Inhibitionwas

alsonotobserved when

preformed

FactorVIa and active site blocked Factor Xa

(DEGR-Xa)

wereused in the first

step

and

congenitally

Factor

X-deficient

plasma

wasusedin thesecond

step of the assay.Fromthese

observations,

weconclude that the

inhibitory

activity

observed

in

ourassay is due to a

VIIa/TF

inhibitor

that

specifically requires

FactorXa forittofunction.

Because

inhibition

was more

pronounced

whenthe

incubation

was carried out for 30 mn (Fig. 1, inset), we selected

this

%HLIMAN

FOOLED

PLASMA"-) Figure 1.

VIIa/TF

inhibitor assay

2 4 6 8 10

standard curve. The incubation

8 7 mixturescontained 3H-FactorIX

6 'i"' (2.5

gg/ml

or 44 nM), Factor

VIIa

(1.4

Ag/ml

or28 nM), Factor Xa

4 ₁15

(2,

g/ml

or34nM),diluted TF

d

_-,

, (2.8%

vol/vol),

CaCl2

(10

mM),

< a

O-,>/<and

_{variousdilutions of HPP}

(-)

2-0 2 - <05 ; j

orrabbit

pooled plasma (RPP)

(o).

o 0 2 3( Eachreactionvolume was 100

Al

10 _(M20 30

TIME(MIN) and the diluent was Tris/NaCl/ Alb. The reactions were terminated

at30 min and percent TCA-soluble cpm were determined as described (24, 25). The logarithmof percent TCA-soluble CPM is plotted against the percent HPP or the percent RPP. Human plasmawas

pooled from 20 donors and rabbit plasmawaspooled from four ani-mals. Inset, kinetics ofactivationof 3H-FactorIXunder various

con-ditions. Percent TCA-soluble cpmareplotted against time. Each

reac-tionmixture contained all of the reagents listed above with the follow-ing noted changes. (m) 9% HPP, (a) 6% HPP, (A) 3% HPP, (A) 1%

HPP,

(0)

0%HPP(buffer control), (o) 0%HPPand17nMnonlabeled FactorIXinaddition to44nM3H-FactorIX, and (X) 0%HPPand 5.6%(instead of 2.8%) tissue factor.

cubation

time for

routine analysis.

Wealso

plotted

the data in several ways and

found

thatthe best fit

(r

=

0.98)

wasobtained when thelogarithm

of

percent

TCA-soluble

counts was

plotted

against

the

inhibitor units (Fig.

1).

Inthis assay, the within-run

coefficient

of variationwas5-10% and thebetween-run

coeffi-cient

of variation

was< 10%.

Sensitivity

of the assaywasabout 0.5%

of

plasma level.

VIIa/TF

inhibitor

activity in various diseasestates. VIIa/ TF

inhibitor

was

assayed

in

plasma

from

25

healthy

volunteers

using

human

pooled plasma (HPP)

asareference. The values

ranged

from

72to 142 U/ml

with

ameanof 101

U/ml

(Fig.

2).

When

plasma inhibitor

levels

of

10

patients

with

DIC

were

as-sayed,

the

inhibitor

wasfoundtobe

significantly

reduced

(mean

57±30

U/ml;

P<

0.001). However,

when

plasma

inhibitor levels of 12

patients

with hepatocellulardisease (mean, 107±33), 12

patients

on

chronic warfarin

therapy (mean, 105±35),

11

patients

with the lupus

anticoagulant

(mean, 10

1±39),

and 11

patients

with

deepvein

thrombosis (mean, 107±30)

were

measured,

the

inhibitor activity

was

found

tobe

within

normal

limits

(Fig. 2).

VIIa/TF

inhibitor

activity in

plasma

from

rabbits

after

in-duction of

DIC.

Fibrinogen,

platelets, Factor V, and FactorVIII

levelsremained

within

normal limits (±

10%)

in the rabbit treated with

saline

only

(rabbit

1

of Fig.

3 A) and in the rabbit treated with dexamethasone followed by saline (rabbit 1 of Fig. 3 B).

140 ° 8 0

0~~~~~~~~~~~~~

z120 oo00 _x

180

L K 8~_8o0 0

>20

NOR 01 LIV WAR LUP OVT

Figure2.Distribution of VIIa/TF inhibitor activity

inplasma from25 normal

subjects (NOR),10patients withDIC (DIC), 12

pa-tients withhepatocellular

disease(LIV), 12patients onchronicwarfarintherapy (WAR), 11patientswith the

lupusanticoagulant (LUP),

and11patientswithdeep

venousthrombosis (DVT).

Arrowrepresentsmeanof

datapointsforeach patient

group.

z

W100,

wIO

X~80 cr p60

m M40

We 20 5

0 20 40 0 20 40

HOURSAFTERENDOTOXIN

Figure 3.VIIa/TF inhibitor activity in plasma ofrabbits after endotoxin treatment.The standard curve labeled RPPin Fig. 1 was used to measure the inhibitory activity. Rabbits in Group A were givensaline daily for 4 d and rabbits inGroup B were given dexa-methasonedaily for4d.On day 5 (zero houronabscissa) and 24hlater (indicated by arrow), one rabbit in each group(labeledIinAand B) re-ceived saline,andtworabbits ineachgroup(labeled2 and3inboth A

andB)receivedendotoxin. For details,seeMethods.

Fibrinogen, platelets, Factor V, and Factor VIII levels fellto

between 60 and 70% of the basal levelsinthetworabbits(labeled 2 and 3 in Fig. 3 A) of GroupA 4 hafterreceivingendotoxin

andremained depressed tothe samedegree after 24 h except for fibrinogen, which rose in bothcases to between 130 and 140%. Onerabbit(labeled 3 inFig. 3A)died within 2 h after

receiving the second injection of endotoxin. The otherrabbit (labeled 2 inFig. 3 A) did notshow furtherdepressionofany

of the aboveclottingfactors. Thetworabbits thatreceived

en-dotoxin after dexamethasone treatment showed excessive

re-duction in the clottingfactorsascomparedwith thosethat

re-ceived salineinsteadofdexamethasone. After 4 h of endotoxin

injection, the levels of various factorsinthetworabbits(labeled

2and 3inFig. 3 B;firstnumber isforrabbit 2 and thesecond number is forrabbit 3) treated with dexamethasone were:

fi-brinogen,43and34%;platelets,41 and47%; Factor V,73 and

68%; and Factor VIII, 52 and 28%. After 24 h of endotoxin

injection the levels were: fibrinogen, 129 and 118%; platelets,

37and43%; Factor V,45 and 32%;andFactorVIII, 23 and

28%. Both rabbitsdiedwithin2 hofreceiving the second

injec-tion ofendotoxin. The abovemeasurementsestablish thatwe

did induce DIC in all four rabbits that received endotoxin. Moreover,aspreviously reported (21, 22), steroid

(dexameth-asoneinourstudy)treatmentenhanced intravascular clotting. VIIa/TF inhibitor levelsfell in all fourrabbits thatreceived endotoxin(Fig. 3).Thelevel of the inhibitor in thetworabbits that didnotreceivedexamethasone(labeled2 and 3 inFig. 3

A) fell by - 25% after 24 hfrom the first endotoxin injection.

Inrabbit2, the level of the inhibitor began rising 48hafter the

secondendotoxininjectionandwasnormalafter 72 h. Thelevel oftheinhibitorinthetworabbits that received dexamethasone (labeled2and 3inFig. 3 B) fell by - 70% (average of 50 and

90%)after 24 h from the firstendotoxin injection. Thus, although

werecognize thatwehave obtained data withalimited number

ofanimals,webelieve it isvalidtoconclude that theinhibitor isreducedinDIC.

VIIa/TF inhibitor activityin conditioned media ofvarious cell cultures. Studies inaprevioussection established that the

inhibitor isnot reduced inpatients with severe hepatocellular

disease. Toaccount forthisobservationandidentify site(s) of synthesisofthisinhibitor, weassayed activity of the inhibitor

insupernatantsof several cellcultures.These dataarepresented

inTableI.HepG2, HUVE, and BPAE cells all synthesized and

secreted this inhibitor. Furthermore, cycloheximide inhibited

thesynthesis. U937 cellsupernatantscontainedsmallamounts

of thisinhibitor,whereasHL-60 cell supernatants containedno

detectableactivity.Ineachcelltype,inhibitionwasnotobserved

ifFactor Xa formed in thefirststepwasneutralized by

Anti-1876 M.S.

Bajaj,

S. V. Rana,R. B. Wysolmerski, andS. P. Baja]

4 A + B

~-

1a- 1 *

Table I. VIIa/TF Inhibitor Activity in Supernatantsof Various Cell Cultures

Activityin

conditionedmedium

adjustedto

Cell type Origin Treatment 106cells/ml

HepG2 Human None 4.9±0.5

hepatoblastoma

Cycloheximide -0.3

HUVE Humanumbilical None 3.9±0.3

vein

Cycloheximide -0.3

BPAE Bovine None 3.3±0.3

pulmonary artery

U937 Human None -0.3

histiocytic

lymphoma

HL-60 Human None Undetectable

promyelocytic leukemia

HepG2(- 5 X106),HUVE( 3 X106),and BPAE (- 3X106)cells were grown toconfluentmonolayersinT-25cm2cultureflasks

con-taining5 mlofappropriate medium supplemented with 10%

inacti-vatedFCS. Conditioned mediawereobtainedbyincubatingcultures

for 20-24hin the absenceof FCS. Whencycloheximidewasused, its concentrationwas1

_Ag/ml.

U937 and HL-60 cellswere grownin T-75 cm2 flasks. Cellswerespun,washed,andresuspended inmedia

with-outFCS and culturedfor - 22 h.Afterculturing, the number ofcells

were - 2 X106/ml forboth U937 andHL-60 celllines.Cells were

spunandsupernatants wereassayed before and after20-fold

concen-tration.Activity in conditioned medium isthe meanoftwo

experi-mentsperformedinduplicate. Complete growth media when tested fortheinhibitorwerefoundtobenegative.ForHepG2,HUVE, and

U937cells,activityis presentedaspercentageofHPPandforBPAEit is presentedaspercentageofbovineplasmafromasingle animal(a

standardcurveusing bovine plasmawasconstructedfor this purpose).

FactorXantibodies. Inhibition was also not observed if the test

samples

wereincubated at 100°C for 5 min before assay.

In

further

experiments, we concentrated HPP and the

con-ditioned

media from HepG2 and HUVE cells. These samples were then run on SDS gels. The proteins were eluted and assayed for the

inhibitor activity.

Ineach case, two peaks

ofactivity

were identified-one corresponding to 45,000±2,000 and the second

corresponding

to33,000±2,000 mol wt (Fig. 4).

Discussion

Inthis

communication,

wehave developed aradioassay for

VIIa/

TF

inhibitor

that

specifically

requires Factor Xa to

function.

Boththe sensitivity and precision of the assay are appropriate

for

measurements of the inhibitor levels in clinical

specimens.

Using

this assay, we show that the

inhibitor

levels are

significantly

depressed

in

patients

with DIC.

Of

the 10

patients

investigated,

onepatient withtraumahadalevelof 35%, two

patients

with

gram-positive sepsis

had

levels

of

- 70%, and seven

patients

with

gram-negative

sepsis

had levels of 27-87% (Fig. 2). Our

animal studies using

a

rabbit

model system support our mea-surements

of

the

inhibitor

level inDIC

patients.

Furthermore, in

earlier studies

(21, 22), enough evidence has been presented

.D 926645 31

-+ t 4

2.0. _i_

1.0

AS

>

20

-I

I.-> 1nmi

VA.

* .v

0.51-0

10 30 50

Slice Numt

21 14 Figure 4. Elution ofVIla/

_______ TFinhibitor activity from

SDSgels. Proteins present

inthe 40-70% ammonium

sulfateprecipitateofHPP,

HPP

HepG2,

andHUVEwere

dialyzed extensively against

~~l

~* iTris/NaClbuffer. - 150

,d

ofsamplecontaining- 12 U/ml of VIIa/TF inhibitor activity was applied to each HepG2 gel in the absence of a

re-ducing agent. Gelswere ' sliced into 1-mmsegments

and eachslicewas soaked

in0.1mlofTris/NaCl/Alb

buffer in plastic tubes

over-HtUVE night. Each sample was thendialyzed extensively -- against

Tris/NaCl

bufferto

70 90

removeSDS.TheVIIa/TF ber inhibitoractivity was

deter-mined asdescribed in Methodsexcept 60z1 (instead of20

_,d)

ofsample and 10Ml (instead of 50 ul) of fivefoldconcentrated3H-FactorIXwas used.This en-hanced thesensitivity ofthe assaythreefold. Duplicate stained gels (20

Ml

sample)

in eachcase arealso

depicted.

Molecular

weight

markers obtainedfrom Bio-RadLaboratoriesare phosphorylase B (92,500),

bovineserumalbumin (66,000), Ovalbumin (45,000), carbonic

anhy-drase(31,000), soybean trypsin inhibitor (21,500), and lysozyme (14,000).

tosuggest that the endotoxin-induced

clotting

is initiatedby the tissue factor pathway. Thus,ineightof ourpatients (sevenwith

gram-negative sepsisand onewithtrauma),theclottingis most

likely initiated by tissuefactor. Ifso,ourfindings supporta

hy-pothesis

that reduced levels of the inhibitorinpatientswith DIC

representdepletion through consumption.

The lupus

anticoagulant

is associated withanincreased

in-cidence of thrombosis (23).Decreased levels(antibody-mediated

or

otherwise)

of

VIIa/TF

inhibitor could contribute to the

pathogenesis

ofthrombosis inthese

patients.

For thisreason, we measured thelevels

of

this

inhibitor

in 11

patients

withthe

lupus

anticoagulant.

None of thepatients appearedtohave

ab-normal levels

of this inhibitor.

Thus, thecause ofthrombosis

in the

majority

of

patients

with the

lupus

anticoagulant

does

notappear to be due todecreased levels of this inhibitor. We

alsomeasured the levels of this inhibitorinpatientswithdeep veinthrombosis in a search

for evidence

thata

deficiency

ofthis

inhibitor could

predispose

suchindividualstothrombosis. None

of the

patients

examined thus far showed reducedlevels of this

inhibitor.

The

inhibitor

levelswerealsonotreduced in patients

undergoing

chronicwarfarin

therapy, suggesting

that the

inhib-itorisnot avitamin K-dependent protein.

Wealsomeasured the VIIa/TF

inhibitor

activity

in 12 pa-tients with advanced

hepatocellular

disease. The inhibitorlevels

were

found

tobe normal inthese

patients.

Thiswas an

unex-pected

finding,

because ina

preliminary

report, Brozeand

Mil-etich (26)

demonstrated thataliver cell ine

(HepG2)

produces this inhibitor.Inthis report,wehave foundthat in addition to

the

HepG2 cells, endothelial

cellsalso

produce

this inhibitorin

significant

amounts. In

fact,

our

patient

data would indicate

that

endothelial

cells may be the

major

sourceof thisinhibitor

invivo.

Toexamineifthe inhibitor foundin humanplasmaissimilar tothe inhibitor produced by HepG2andHUVE cells inculture, we performed SDS gel electrophoresisonprotein samplesfrom

each of these threesources. In each case, theinhibitoryactivity

wasfoundat45,000±2,000 D andat 33,000±2,000 D.

Warn-Cramerandassociateshave also found that theinhibitoryactivity in human plasma ispresent atmolwtof - 43,000and - 34,000

(27 and Dr. B. Warn-Cramer, personal communication). Whether the 34,000 mol wt form is a degradation productof

the43,000 mol wt form is not known. However, both forms require Factor Xa to function (Dr. B. Warn-Cramer, personal communication).

Acknowledgments

We are indebted toDr.J. Heinrich Joist(Director,Clinical Hemostasis Laboratory) for his generous support andconstantencouragement. We

thank Dr.Coy Fitch(Chairman, Department of Medicine) for allowing

Dr.M. S. Bajaj to carry out these studies during her residency (Internal Medicine) training. We thank Drs. Joachim Reimers and Steve Janney for their interest and helpful discussions.

This work was supported by National Institutes of Health Research GrantHL-36365 and American Heart Association Grant-In-Aid 86-1182.

Dr.S. P.Bajaj is supported in part byanAmerican HeartAssociation EstablishedInvestigatorAward(No. 83-176). Dr.Wysolmerskiis sup-ported byatraining grant(HL-07050).

References

1.0sterud, B., and S. I. Rapaport. 1977. Activation of factorIXby the reaction product of tissue factor and factor VII. Proc. Natl. Acad. Sci. USA. 74:5260-5264.

2. Warn-Cramer, B. J., and S. P. Bajaj. 1986. Intrinsic versus extrinsic coagulation. Biochem. J. 239:757-762.

3. Lanchantin, G. F., and A. G. Ware. 1953. Identification of a thromboplastin inhibitor in serum and in plasma. J. Clin. Invest. 32: 381-389.

4.Carson, S. D. 1981. Plasma high density lipoproteins inhibit the activation of coagulation factorXbyfactor VIIa and tissue factor. FEBS

(Fed.

Eur.Biochem.Soc.) Lett. 132:37-40.

5.Sanders,N.L., S. P. Bajaj,A.Zivelin, and S. I. Rapaport. 1985. Inhibition of tissue factor/factor VIIa activity in plasma requires factor

Xandanadditional plasma component. Blood. 66:204-212.

6.Jaffe, E. A., R. L.Nachman, G. G. Becher, and C. R. Minick. 1973. Culture of endothelial cells derived from umbilical veins. J. Clin.

Invest.52:2745-2756.

7.Del Vecchio, P. J., and J. R. Smith. 1981. Expression ofangiotensin-converting enzyme activity in cultured pulmonary artery endothelial cell.J.Cell.Physiol. 108:337-345.

8.Collins,S. J.,R.C.Gallo, and R. E. Gallagher. 1977. Continuous

growth and differentiation ofhumanmyeloid leukemic cells insuspension

culture. Nature(Lond.). 270:347-349.

9. Sundstrom, C., andK.Nilsson. 1976. Establishmentand

char-acterizationof ahumanhistiocytic lymphoma cell line (U937).Int. J.

Cancer. 17:565-577.

10. Aden, D.P., A. Fogel, S. Plotkin, I. Damjanov, and B.B.Knowles. 1979. Controlsynthesis of HBsAg inadifferentiated human liver

car-cinoma-derived cell line.Nature(Lond.).282:615-616.

11.Bajaj, S. P., S. I. Rapaport, and W. A. Russell. 1983. Redeter-mination of the rate-limiting step in the activation of factor IX by factor XIa andfactor VIla/tissue factor. Biochemistry. 22:4047-4053.

12. Rao, L. V. M., and S. P. Bajaj. 1984. Purification of human factorVII utilizingO-(diethylaminoethyl)-Sephadex and

Sulfopropyl-Sephadex chromatography. Anal Biochem, 136:357-361.

13.Bajaj, S. P., S. I. Rapaport, and C. Prodanos. 1981.Asimplified procedure for purification of humanprothrombin,factorIXand factor X. Prep.Biochem. 11:397-412.

14. VanLenten, L., and G. Ashwell. 1971. Studiesonthechemical andenzymatic modifications ofglycoproteins.J.Biol. Chem.

246:1889-1894.

15.Bajaj, S. P., S. I. Rapaport, and S.F. Brown. 1981. Isolation and characterization ofhumanfactor VII. J. Biol. Chem. 256:253-259.

16. Warn-Cramer, B. J., and S. P. Bajaj. 1985. Stoichiometry of

bindingofhighmolecularweight kininogentofactorXI/XIa. Biochem.

Biophys.Res.Commun. 133:417-422.

17.Bajaj, S. P., S. I. Rapaport, S.L.Maki,and S.F. Brown. 1983.

Aprocedure for isolation of humanproteinC andproteinSasbyproducts

of thepurificationof factors VII, IX,Xandprothrombin. Prep.Biochem. 13:191-214.

18.Laemmli,V. K. 1970. Cleavage of structural proteins during the assembly ofthe head ofbacteriophageT4. Nature(Lond.).227:680-685. 19.Wysolmerski, R.,and D.Lagunoff. 1985. The effect of ethchlor-vynol on cultured endothelial cells. Am. J. Pathol. 119:505-512.

20. Hudig, D., and S. P. Bajaj. 1982. Tissue factor-likeactivityof the humanmonocytictumorcell lineU937. Thromb. Res. 27:321-332. 21. Lerner,R.G., S. I.Rapaport, and J. M.Spitzer. 1968. Endotoxin inducedintravascular clotting.Thromb.Diath.Haemorrh. 20:430-443. 22.Shen, S. M.-C., and S. I. Rapaport. 1974. Increased fibrinogen consumption following endotoxininjectionincortisone-treated rabbits. Proc. Soc.Exp. Biol. Med. 145:851-854.

23.Meuh,J.R., K. D.Herbst, and S.I.Rapaport. 1980. Thrombosis inpatients with the lupus anticoagulant.Ann.Intern.Med. 92:156-159. 24. Zur, M.,and Y.Nemerson. 1980. Kinetics of factorIXactivation via the extrinsic pathway. J. Biol. Chem. 255:5703-5707.

25.Bajaj,S.P. 1982.Cooperative

Ca2"

binding to human factor IX.

J.Biol. Chem. 257:4127-4132.

26. Broze, G. J., and J. P. Miletich. 1986. Characterization of the inhibition of tissue factor in serum. Circulation.74:11-409.(Abstr.)

27.Warn-Cramer, B.,L. V.M. Rao, S. Maki, A. Zivelin, and S.I.

Rapaport. 1986. Partial purification of a factor X-dependent plasma inhibitorofthefactorVIIa/tissue factor complex. Circulation. 74:II-408. (Abstr.)