Interactions of the platelets in paroxysmal nocturnal hemoglobinuria with complement Relationship to defects in the regulation of complement and to platelet survival in vivo

Interactions of the platelets in paroxysmal

nocturnal hemoglobinuria with complement.

Relationship to defects in the regulation of

complement and to platelet survival in vivo.

D V Devine, … , R S Siegel, W F Rosse

J Clin Invest.

1987;

79(1)

:131-137.

https://doi.org/10.1172/JCI112773

.

The blood cells of patients with paroxysmal nocturnal hemoglobinuria (PNH) have

abnormal interactions with complement. The activity of the alternative pathway C3

convertase on the platelets of 9 out of 19 patients with PNH was elevated. 10 patients had

C3 convertase activity within the normal range even though 80-95% of their platelets lacked

the complement regulatory protein decay accelerating factor (DAF) that is absent from the

affected blood cells in PNH. PNH and normal platelets released factor H when C3 was

bound to their surfaces. This may account for the apparent regulation of C3 convertase

activity on platelets that lack DAF. The abnormal uptake of the membrane attack complex of

complement by PNH III erythrocytes was not seen in PNH platelets. 111Indium-labeled

platelet survival times were normal in five of eight patients, which suggests that the lack of

the membrane attack complex defect results in normal platelet survival in PNH.

Research Article

Find the latest version:

http://jci.me/112773/pdf

Interactions of the Platelets

in

Paroxysmal Nocturnal

Hemoglobinuria with Complement

Relationship

toDefects in the

Regulation

of

Complement

andto Platelet

Survival

InVivo

Dana V. Devine,* Robert S. Siegel,t and Wendell F.Rosse**

*Department ofMicrobiology and Immunology and

$Department

ofMedicine, Duke University Medical Center, Durham, North Carolina27710

Abstract

The blood cells of patients with paroxysmal nocturnal hemoglo-binuria (PNH) have abnormal interactions with complement. The activity of the alternative pathway C3 convertase on the platelets of9 out of 19 patients with PNH was elevated. 10 patients had C3 convertase activity within the normal range even though

80-95%

of their platelets lacked the complement regu-latory protein decay accelerating factor (DAF) that is absent from the affected blood cells in PNH. PNH and normal platelets released factor H when C3 was bound to their surfaces. This mayaccount for the apparent regulation of C3 convertaseactivity onplatelets that lack DAF. The abnormal uptake of the mem-brane attack complex of complement by PNH III erythrocytes wasnot seen in PNH platelets.

"'Indium-labeled

platelet survival timeswerenormal in

five

ofeight patients, which suggests that thelack of the membrane attack complex defect results in normal platelet survival in PNH.

Introduction

Paroxysmal nocturnal hemoglobinuria

(PNH)'

is a

hemato-poietic

stemcelldisorder characterized by an increased sensitivity

of

theblood cells to complement-mediated lysis,

hematocyto-penia,

anda

propensity

toward

intravascular

coagulation (1-3). The PNH erythrocyte may be abnormal in one or two of its interactions with complement. PNH II erythrocytes show an

inability

toregulate theC3convertasesformed on the cell surface (4, 5). As a result, these cells are three to five times more sensitive

to

complement

than normal erythrocytes (6). PNH III eryth-rocyteshave

increased

sensitivity to the terminal components of

complement,

themembrane attack complex (7). This defect may actually be

twofold.

PNH IIIerythrocytes are unable to bearasmanymembrane attack complexeson

their surface

as a

Dr.Siegel isnowAssistant Professor ofMedicine,Division of

Hema-tology/OncologyatTheGeorgeWashington UniversityMedicalCenter,

Washington, D.C. Address reprintrequests to Dr. Devine, Box 3934,

DukeUniversity Medical Center, Durham, NC 27710.

Receivedfor publication19May 1986 andinrevisedform11 August 1986.

1.Abbreviations usedinthis paper:CLS, complement

lysis

sensitivity; CoVF,cobravenomfactor; CoVFBb, complexes formed from CoVF andactivated factor B; DAF,decayacceleratingfactor, FITC,fluorescein isothiocyanate; GVB, veronal-buffered salinecontaining0.1%gelatin;

PFA,paraformaldehyde; PNH, paroxysmalnocturnalhemoglobinuria.

normal red cell before a

lytic

poreis created (8). In

addition,

when membrane attackcomplexesaregeneratedinserum,they

readily bindtothe surface ofPNH IIIcellsbut not tothesurface of normal cells (9, 10). Having both the C3convertaseand

ter-minal complex defects makes the PNH IIIredcell 15-25times

moresusceptibletocomplement lysis than normal red cells (6). Recently, the PNH erythrocyte (11, 12)aswellasthe PNH platelet, monocyte, and granulocytes (13), have been foundto

lack decay accelerating factor (DAF), aprotein that regulates both the classical and alternative pathwayC3 convertases on

the surface of erythrocytes (14-16). The deficiency of this mol-eculeseems tobe responsible for the defect in the regulation of the activation of C3 shared by PNHIIand PNH III cells.

How-ever,although PNHIIred cells reconstituted with purified DAF lose their increased susceptibilitytocomplement, PNH IIIcells remain sensitive to the membrane attack complex (17). The molecular basis for the terminal complex defect of PNH III erythrocytes remainstobe elucidated.

The platelets of patientswith PNH are alsoabnormally

sen-sitive to the lyticaction ofcomplement when complement is activated by platelet-specific antibody (18).Thein vivo

signifi-canceof this observation is not clear. Although somepatients withPNHhaveareduced plateletcount,generallyin the context ofpancytopenia, manyhaveplateletcounts well withinnormal

range. It is not known whether PNH platelets also have two defects incomplement regulationsimilar to theerythrocyte. Us-ing antibody-bindUs-ing assays, we have measured DAF in the platelets of20patientswith PNH. We have determined the

rel-ativepercentageof abnormalplateletsin the totalplatelet

pop-ulation in these patients and compared this with the relative percentage of abnormal erythrocytes. We have measured the activityofstabilizedalternativepathwayC3 convertases formed onthe surface of PNHplateletsaswell as theabilityof the PNH

platelettotakeup membrane attackcomplexesunder the con-ditions of reactivelysis.To assess theimpactof defects in

com-plement regulation on the survival ofplatelets,

"'In-labeled

plateletsurvival studies have beenperformedoneight patients with PNH.

Methods

Buffers, complement components, andantisera. Allexperimentswere carriedoutin either veronal-buffered saline(5.0mM sodium barbital,

0.15M NaCl,and0.1% gelatin [pH 7.5]) containing0.0 15 M EDTA

(GVB-EDTA)orveronal-buffered salinecontaining0.15 mM calcium

chlorideand 0.5 mM magnesium chloride (GVB++)or2.5 mM mag-nesium alone(GVB+).Cobravenom factor (CoVF;reference 19),C3

(20), factorB(4),factor D(21),C3nephriticfactor(22),and DAF(15)

were purified usingpreviously described methods. Purified thrombin wasthegift ofDr. John W.Fenton,III(NewYorkDepartment of Health, Albany, NY). AntiserumtohumanerythrocyteDAFwasmisedinrabbits inthis laboratory. This antiserumwasmonospecificonplateletsas

de-ComplementInteractionswithParoxysmalNocturnalHemoglobinuriaPlatelets 131

J.Clin. Invest.

© TheAmerican Societyfor Clinical Investigation, Inc.

0021-9738/87/01/0131/07 $1.00

termined byWesternblotanalysis. Goat anti-human factorH andrabbit anti-human C3c was purchased from Calbiochem-Behring Corp.,

(LaJolla, CA). Monoclonalanti-human C3cwasthegift ofDr. J.Donald Capra(University of Texas, Dallas,TX) and Dr. Gordon Ross(University

of NorthCarolina,ChapelHill,NC).Monoclonalanti-polyC9,which recognizesaneoantigen expressedbypolymerizedC9orMC5b-9(23),

wasthegift ofDr. Ronald J.Falk(UniversityofNorthCarolinaatChapel Hill). Fluorescein isothiocyanate(FITC)-labeled affinity-purified F(ab')2

goatanti-rabbit IgGwaspreparedby Jackson Immunoresearch Labo-ratories (Avondale, PA) andperoxidase-conjugatedrabbitanti-goatwas purchasedfrom BioRadLaboratories, (Richmond, CA). Staphylococcal

ProteinA wasobtained from Sigma Chemical Co., (St.Louis, MO). Whereapplicable, proteins were labeled with 1251 using the Iodogen

method (24).

Plateletpreparation. Bloodwas drawn from normal volunteersor

consenting patients andanticoagulatedinEDTA.Aftercentrifugation

at 1 10g,theplatelet-rich plasmawasremoved andtheplatelets washed threetimes in GVB-EDTA.

Complementlysissensitivityofpatienterythrocytes. Thepercentage

ofabnormalerythrocytes in thebloodofpatientswithPNHwasassessed

using the complementlysissensitivity (CLS)test(6).Thisassaymeasures

thesensitivity of erythrocytestolysisbyantibodyandcomplement.PNH IIerythrocyteshaveCLSvalues between 4 and10.PNH IIIerythrocytes have CLS values>20.

DeterminationofDAF-negativeplateletpopulations

byflow

cytometry. Washed PNH ornormal donorplateletswerereactedwith rabbit

anti-DAF ornonimmune rabbitserum, washedandthen incubated with FITCaffinity-purified F(ab')2goatanti-rabbitIgG. Stained plateletswere fixed in 1%formalin beforefluorescenceanalysis.Plateletsfrom normal donorswere runwith eachgroupofPNHplatelets.

Cell-associated fluorescencewasmeasuredusingacytofluorograph

(model 50H, OrthoDiagnostic Systems Inc.,Raritan,NJ). Plateletswere illuminated with the 488-nmline ofanargonion laser and the

right-anglelight scatterfocusedontoappropriate fiberoptics usinga 32X/

0.40NAobjective(E.Leitz,Inc.,Rockleigh, NJ)mountedinacarrier of localdesign.Fluorescencewasmeasuredbehinda515-530-nm band

passfilter.Analysiswasrestrictedtosignalsfromsingle plateletsbygating

the fluorescencesignalsonthebasis ofaright angleversusforwardscatter

cytogram. The amplifierswere setin linearmode andthedata were analyzedusing the software inanOrtho 2140datahandlingsystem.For statisticalanalysisof thehistograms,tworegionsweredefined. Thefirst,

region 1, representschannels1-999,toexcludeunweighted off-scale data in channel 1000 from statistical analysis. The lower limits of region 2 aredefined by therightshoulder ofthehistogramofthe platelets incubated with nonimmune rabbitserumandFITC-conjugated secondantibody

(negative control).

MeasurementoftotalplateletDAF. To determine whetherplatelets

contain internal pools ofDAF thatmight affectconvertase stability, Western blotanalysiswasperformedonthe plateletsofthe PNHpatients in the study. Washed plateletswere solubilized in nonreducing SDS Laemmlipolyacrylamide gel electrophoresis sample buffer(0.125MTris,

4%SDS, and 20% glycerol [pH6.8]) and 50 ,ugtotalproteinwasloaded

perlane.Afterelectrophoresisinto7.5%polyacrylamide gels, the proteins

weretransferredontonitrocellulosepaperusingthe method ofTowbin

etal.(25).Thenitrocellulosewasprobedfirst with rabbitanti-DAF or

nonimmune rabbitserum andsecond with['25I]staphylococcalprotein A.The paper was dried and autoradiographed.

Measurementof nephritic factor-stabilized C3 convertaseactivity. To measure theactivity of an alternative pathway C3 nephritic

factor-stabilized C3 convertase on the platelet surface, 5 X 108 platelets in

GVB-EDTAwere first incubatedwith 1 mg C3 and CoVF-factor Bb

complexes (CoVFBb) for30minat 37°C. The CoVFBbcomplexes were

prepared byincubating 50

Mg

CoVF in 5mM

MgCl2

with 50

Mg

factor BandsufficientcrudefactorD to convert all the factor B to Bb. The

magnesiumwaschelated withEDTA before adding the CoVF complexes

totheplatelets. Aftera30-minincubation,these platelets(PC3b) were

washedin GVB-EDTAthreetimes and a sample removed to measure

platelet-bound

C3b. The PC3bwerethen washedinto

_GVB+

and

exposed

tofactor B, factorD,and C3

nephritic

factor for 10minat370C.After

onewashwith

GVB+,

the

platelets

wereincubatedwith 1 mgC3 for 30 minat

370C.

After three washes with

GVB-EDTA,

these

_platelets

(PC3bBbC3NeF)were

assayed

for

platelet-bound

C3b.

A radiolabeled monoclonal

antibody-binding

assay was used to

measuretheamountof C3b

deposited

onthe

platelet

surfaceunderthe aboveconditions.

Equal

volumes of '25I-labeledanti-C3c and washed

plateletswereincubated for 30minat roomtemperature.

Aliquots

of the

platelet

and

antibody

suspension

were

_{centrifuged through phthalate}

esters

(Fisher

Scientific

Co.,

Pittsburgh, PA) (1.5

parts

butyl phthalate/

1 partbis

_{[2-ethylhexyll}

phthalate)

toseparate bound from unbound

antibody.

Nonspecific

antibody

binding

wasdefinedasthat

antibody

binding

to

platelets

exposed only

to

_C3,

butnotCoVF

complexes

inthe first step

(PC3).

The determination of molecules of

_antibody

boundper

plateletwasmadefroma

knowledge

ofthe

specific

activity

ofthe

_antibody

and theplateletcount.Aconversion ratiowascalculatedasfollows: moleculesbound/PC3bBbC3NeF-_{moleculesbound/PC3b}

moleculesbound/PC3b-_{moleculesbound/PC3}

Measurement

ofplatelet

factor

Hrelease. 5X 108washed

platelets

in

GVB+

wereincubated with 1 mg

C3,

50

Mg

factorB,and crude factor

Dfor 30minat

37°C.

Immediately

after

_washing

threetimes in

PBS,

the

platelets

were

resuspended

in I mlPBS

containing

1%

paraformal-dehyde

(PFA)

andincubatedatroomtemperature for1h. After

washing

threetimes in GVB-EDTA toblock unreacted

aldehyde

groups, the

plateletswerestained for flowcytometryasdescribed above

using

goat

anti-factorHorrabbit anti-C3c

(both

from

_{Calbiochem-Behring Corp.)}

andthe

appropriate

FITC-conjugated

second antibodies. Control

platelets

were

prepared

using

PFA-fixed

platelets

that had been

exposed

to

com-plementcomponents. Controls for

antibody

binding

were

performed

using

nonimmunesera.

Thrombin-stimulated factorHreleasewasassessed

by treating

plate-letsat5X

_108/ml

in GVB-EDTA with 100

U/ml

thrombin for 20min

at

37°C.

The

platelets

werethenfixed inPFAandstainedasdescribed

above.

Quantitation

ofplatelet

factor

H.Theamountof factorHreleased

by

thrombin-stimulatedor

detergent-solubilized

platelets

wasmeasured

by

a

radioimmunoassay

described in detail elsewhere

(Devine,

D.

V.,

and W. F.Rosse,

manuscript

in

_{preparation). Briefly,}

microtiter

plate

wellswerecoated withamonoclonal

_antibody

tohumanfactorH

(Gen-zyme

Corp.,

Boston,

MA),

supernatesfrom thrombin-stimulated

_platelets

orTriton

X-100-solubilized platelets

were added to the wells. After

washing

the plate,

'25l-labeled

polyclonal

anti-factorHwasaddedto

each well. The boundcpmweremeasured

using

ascintillation

well-type

gammacounter.Theamountof factorH per 108

platelets

wasdetermined

by

comparison

with a standard curve

prepared

using purified

serum

factor H.

Uptake

of

membraneattack

_complexes

_from

serum.Wewere

inter-ested in

determining

ifthe

platelets

from

_patients

with PNHtakeup membrane attack

complexes

undertheconditions of reactive

_lysis

asdo PNHIII

erythrocytes.

CoVFBb

complexes

were

prepared

exactly

as

de-scribed above. 5 X 108

platelets

inGVB-EDTAwere incubated with CoVF

complexes

and EDTA-chelated normal humanserum

₍₉

_parts

serum/l

part 0.2M EDTA

_{[pH 7.2])}

for30 minat

_37°C.

After three

washes in

GVB-EDTA,

platelet-bound

C5b-9wasassessed

by

'25I-labeled

monoclonal

anti-poly

C9

binding

asdescribed above.The ttestwas

usedfor statistical

analysis

of

antibody-binding

studies

(26).

"'In-labeled

platelet

survivalstudies. Platelet survival studieswere

performed

on

_patients

after

obtaining

written informed consent. The

protocol

was

approved

by

the Duke

University

Medical CenterHuman ResearchCommittee.

Briefly,

platelets

wereisolatedandlabeled

_according

tothemethodofThakuretal.

(27),

asmodified

by

Heatonetal.

(28).

Therewas nolossof

platelets

during

the

_{labeling procedure.}

200MCiof

"'In-labeled

autologous

platelets

were

injected intravenously

and serial

blood

_samples

wereobtained 15 min and 2 hafter

injection,

then every 12 hthereafter for4d. Toestimatemeansurvival

time,

datawere

_analyzed

using

themaximum likelihood estimate ofthe

_{integer-ordered}

gamma

function as recommended by the International Committee for

dardization in Hematology (29). Platelet turnover was calculated

ac-cordingtotheformula:

platelet

count/Asl

90%

platelets/,Al

per h=

(30).tCun/lX

0

lifespan (h) initial recovery

Images of the anterior and posterior chest and anterior and posterior

abdomenwere obtained 24 h after injection using a large crystal

scin-tillationcamerafittedwith a medium energy collimator. Both gamma photon peaks of"'In (at 173 and 247 keV) were summed and each

imagewasacquired for 300 s.

Results

Wehave studied the platelets of 20 patients with PNH of whom 8 have been examined multiple times. The binding of anti-DAF

antibody

to

platelets

wasused to measure the percentage

ofDAF-deficient

platelets in each

patient.

This percentage was then compared with the percentage of complement-sensitive eryth-rocytes as measured by the CLS test (Fig. 1). For each patient studied, therewas agreaterpopulation of abnormal (DAF-de-ficient) platelets than abnormal erythrocytes. The binding of anti-DAF to normal donor platelets was highly heterogeneous (Fig. 2 A). Most patients had platelets which bound little if any anti-DAF

(Fig.

2 C). Even inthose fourPNH patients whose platelets showed

significant

anti-DAF binding there was never the

intensity

of staining seen in normal platelets (Fig. 2 B).

By Western blot

analysis,

the platelets of patients with >80% DAF-negative platelets by fluorescence analysis did not contain

cryptic pools

of DAF

(Fig. 3). Using

this method, DAF was

readily detected in normal platelets

(Fig.

3, lane 1).

The

activity of

the alternative pathway C3 convertase was measuredonthe

surface of

normal and PNH platelets to

deter-mine

theeffect

of

DAF

deficiency

on theregulation of this en-zymecomplex. PNH platelets

did

not bind fluid phase C3b more readily than normal platelets because similar amounts of mono-clonal anti-C3c boundtonormal donoror PNH

platelets,

which werereacted with

fluid

phase CoVF-factor Bb complexes and C3

(Table I).

C3

nephritic

factor-stabilized convertases were

established

atthese C3b sites and the functional

activity

of these

100

90+

80+

70±

z

60-E c

50-4.n

-:

w

40-cL

30-

20-

10--A

//////

Figure1.Percentageof abnormal

erythrocytesandabnormalplatelets foragivenpatientwithPNH.The percentageof abnormalerythrocytes

wasdetermined bytheCLStestof Rosse and Dacie(6).The percentage ofabnormalplateletswasdetermined by measuringtheamountof platelets that failedtostain withantibodyto ________________lDAF.Thiswasachievedby

fluores-RBC Platelets cenceflowcytometry.

200-0

w

E z

2001

A

8

C

200 400 600 800

Goted Green Fluorescence

1000

Figure 2.Cytograms of stainingpatternsofnormal and PNH platelets

withrabbit antibodyto DAF. Gated greenfluorescenceis shown on a

linear scale. Anti-DAF staining is shownin thestippledcurveswhile nonimmune rabbitserum controls are notstippled.A,stainingpattern

of normalplatelets.Note that thestaining is highly heterogeneous.B,

stainingpattern seenin4of 20patients. Note that thereareno plate-letsfound in the higher intensity channelsontheright.C,staining

pattern seeninmostofthe PNHpatients examined.There waslittle, ifany,significant staining withtheanti-DAF antibody.

convertases wasmeasured. The

activity

of these C3 convertases

onthe

platelets of

13 normal donorsas

defined

by the conversion ratio was 9.1 (SD = _{3.3). Values}> 2 SDs above the mean of the

activity

onnormal donorswereconsidered elevated. Notall

patients

with PNH had elevated

activity of

theC3 convertases formed onthe platelet

surface

(Fig.

4). 10

of

19

patients

had normalC3 convertase activity, while 9 had elevated activity.

C3convertase

activity

wasalsomeasuredontheplatelets of

four

patients without

C3

nephritic

factor

stabilization.

Although the

conversion

ratiosfor both normal andPNHplateletswere

lower than those

obtained

with stabilized convertases, the pattern

of

elevatedconvertase

activity

on the

platelets

of onlysome

of

thePNH

patients

wasconsistent with the results of

experiments

with stabilized C3 convertases. Three of the

four

patients

had elevatedconvertase

activity

withastabilizedconvertase.Without the

addition

of C3

nephritic factor,

theconvertase

activity

was

2.9, 3.6, and 4.7 times that seen on normal platelets run in

parallel experiments.

The other

patient

had

platelets

onwhich

astabilized C3 convertase had normal

activity.

When the

ex-periment

was

performed

withthis

person's platelets

without C3

nephritic

factor, theconvertase

activity

was0.9 timesthatof the normal platelet controls.

Becausethe presence ofDAF onthe

platelets

couldaccount

for the apparent

regulation

of

the C3convertase

activity

onthe platelets ofsome PNH

patients,

the C3convertase

activity

was

compared

with the percentage of

DAF-negative platelets (Fig.

NRS

kD

Anti-

DAF

1 1

23

4

92.5

6

9

--46

30

Figure 3.Westernblotanalysis of plateletDAF.Washedplatelets fromanormal donor orPNHpatients with>80% DAF-negative platelets by flow cytometry were electrophoresed on SDS-polyacryl-amide gel electrophoresis and transferredtonitrocellulose paper. The nitrocellulose was incubated with nonimmune rabbit serum or rabbit anti-DAF,thenwith1251I-labeledstaphylococcal proteinA. An autora-diograph of the nitrocellulose paper is shown here.Normaldonor

platelets were electrophoresedin Lane 1.Lanes2-4contain platelets from three differentPNHpatients.

5). 8

of

the 10

patients

with normal convertase

activity

had >70%

DAF-negative platelets by

flow cytometry

analysis.

Whenplateletswereincubated

with

C3, factor B,

factor

D, CoVF, and

magnesium,

C3bwasboundtothe

platelet surface.

Some

of

the C3b binds factorB toformthe alternative

pathway

C3 convertase onthe

platelet

surface. The

deposition

of C3

by

this

technique triggered

the release of factorHfrom the

platelet.

Factor H then became boundto the platelet surface and was

detectedby

flow

cytometry.

Fig.

6 shows factorHrelease

from

normal

platelets

in response to C3

deposition.

Very

similar

staining

patternswereseenwhenPNH

platelets

contained

plate-let-boundC3

(Fig.

7). This

finding

wasobserved in theplatelets

ofpatients with

eitherPNH II or PNH IIIerythrocytes.

Throm-bin-stimulated

platelets

also released factor H into the

sur-NORMAL PNH Figure 4. Alternative path-way C3 convertase activity

40_ onnormal and PNH

plate-lets. After creatingnephritic

factor-stabilized C3

conver-30_ tases on the surface of the

platelet

using purified

com-je . ponents,, the

platelets

were 20 exposedtoadditional C3.

TheC3 conversion ratio __________-- represents theamountof

AA.

C3 boundperC3

conver-10

_.

tasecreated. Platelet-bound

. ° C3wasassessed by

'25I-la-*

r00

beled monoclonal anti-C3c antibodybinding.Lines connect measurements obtained on the same patient. For a given tient, several months separate each measurement. Filled circles, pa-tients with PNH III erythrocytes. Open circles, papa-tients with PNH III erythrocytes. Half-filled circle, a patient with both PNH II and PNH IIIerythrocytes.

rounding medium;

however, factorH

did

not

bind

backtothe platelet surface in the absence

of

platelet-bound

C3b

(Fig. 6).

Wehave

quantitated

theamountof factorHreleased

from

theplatelets

of

five

patients with

PNH

by thrombin

or

detergent

solubilization.

The

platelets of

two

of

these

patients

had

elevated

C3 convertase

activity.

While

thrombin-stimulated normal

platelets released an average of 50 ng

factor

H/108

platelets,

thrombin-stimulated

PNH

platelets

released

markedly lower

amounts

of

factor H.In responseto

thrombin

stimulation,

the platelets from PNH

patients with

normal C3convertase

activity

released 14, 15, and 17 ng factor

H/108 platelets,

whereas the

platelets from

two

patients

withelevated C3convertase

activity

released4 and 6 ng factor H/

108

platelets. However,

radioim-munoassays

of

TritonX-

100-solubilized platelets

from thesame

platelet

preparations

demonstrated that the total

factor

H

con-tained

in the

platelets of

thesetwo

patients

was 19 and 24 ng/

108

platelets. The totalamountof factorHdetected in

detergent-solubilized

normal

platelets is virtually identical

tothatamount

released by

thrombin stimulation

(Devine,

D. V., and W. F. Rosse,

manuscript

in

preparation).

When platelets were reacted with CoVF-treated

EDTA-serum

(reactive lysis),

neither normalor PNHplatelets took up membrane attack complexes from the serum. The amount of monoclonal

anti-poly

C9bindingtoPNHplateletswasnot

sta-tistically different

from the

binding

tonormal

platelets

(t=

0.44,

TableI.Deposition of C3onto PNHand

Normal Platelets from FluidPhase

Molecules of monoclonal anti-C3c Platelets bound/platelet

mean±SD

Normal donor platelets (n=13) 2590±1362

PNHpatient platelets (n=_{14) 2385±1059*}

Reactivelysisconditions were created using CoVFBb complexes and

purifiedC3.Platelet-bound C3 was measured using

'25I-labeled

mono-clonalanti-C3c.

* Notsignificantlydifferent from normal (P> 0.5).

35

'3

30-z

JoO

2

I-Figure5.Comparison of

alterna--- tive pathway C3 convertase

activ-ityandthe percentage of DAF-neg-ative platelets for patients with

PNH.C3 convertase activity was assessedasdescribed in Fig.4.The *. percentage of DAF-negative

plate-20 40 60 80 100 letswasmeasured

using

fluores-PERCENTDAF-NEGATIVE PLATELETS cence flowcytometry.

A

200 B D

1 200 400 600 800 1000 200 400 600 800 1o0o Gated GreenFluorescence Gated Green Fluorescence

Figure 6. Flow cytometric analysis of platelet bound C3 and factor H

onnormal platelets. Thebinding of anti-C3 (A) or anti-factor H (B-D) areindicated by the stippled curves. Nonimmune serum control curves arenotstippled. Platelets were incubated withCoYF,C3,

fac-torB, factor D, and magnesium (A and B), with thrombin (C) or with buffer alone (D). The anti-factor H and nonimmune serum curves are coincidentinC and D.

P>0.5)(Fig. 8). However, the monoclonal

anti-poly

C9 readily detected the increased uptake of membraneattack complexes by PNH III

erythrocytes

under

identical

reactive

lysis

conditions.

Inaddition, PNHplatelets exposedtoreactive lysisconditions

bound

'25I-labeled

anti-C3cin amountssimilartonormal plate-lets (data not shown).

Of the

eight

patients

who

participated

in"'In-labeled platelet survival

studies,

five had normal platelet

survival,

while three

haddecreased survival (Table II). Of the three withdecreased

A

200 400 600 800 1000 GATED GREENFLUORESCENCE

B

200 400 600 800 GATED GREEN FLUORESCENCE

Figure7.Flow cyto-metric analysis of

plate-letbound C3 and factor

H onPNH platelets. 000 The

binding

of anti-C3 000 (A)oranti-factorH(B

andC)areindicated by

C thestippled curves. Nonimmune serum control curves are not

stippled..Platelets treated with C3, factor B, factor D,CoVF, and magnesium are shown in A andB.Untreated

plateletsareshown inC.

M

Figure

8.

Uptake

of

O 5 _{m membrane attack}

com-Z 4 g plexesbynormal and

N, _{gPNHerythrocytesand}

o3 : : platelets under the

con-ditionsofreactive

lysis.

° 2 Washed

platelets

or

erythrocytes

were

ex-- r T > posed to

CoVF-factor

B

complex-treated

EDTA-0

*:.:.:chelated

serum.The

up-PLATELETS R BC takeof membrane

at-tack

complexes

was

measured by'251-labeledmonoclonal anti-poly C9 antibodybinding.

This monoclonal antibody recognizes a neoantigen expressed by poly-merized C9. Normal platelets bound a mean of 766 molecules of

anti-poly C9/platelet (SD = 412; n =_9),while PNH platelets bound 720 molecules anti-poly C9/platelet (SD=768;n = 11). The PNH

eryth-rocytes shown inthis experimentcontained >95%PNHIIIcells.

survival,

twohad splenic enlargement. Thepercentageof

DAF-negative

platelets

weremeasuredinfour ofthese

patients.

Three had >90%

DAF-negative platelets

andonehad 77%

DAF-neg-ative

platelets.

Ofthese

patients,

twohad normal platelet survival andturnovertimes. The othertwo werethe

patients

with

splenic

enlargement.

Discussion

The abnormal interactions of complement with erythrocytes in PNHcauseincreased destruction of red cells. This

phenomenon

is readily demonstrated by the increased sensitivity of PNH cells

tolysis by antibody and complement invitro(6). In vivo, the activation ofcomplement by the alternative pathway is themost

likelycauseofred celldestruction(1). AsterandEnright dem-onstratedthat the platelets of patients with PNHaremore

readily

lysed by complement in thepresenceof complement-fixing

anti-TableII "'In-labeled Platelet StudiesinPNHPatients

DAF-Platelet negative Patient count/ul Survival Recovery Turnover platelets

h % pit/t/h %

1* 113,000 106 25 3837 95

2 12,000 157 81 85 N.D.t

3 172,000 176 65 1349 90

4 231,000 >200 74 992 N.D.

5 76,000 >200 55 612 N.D.

6 309,000 124 65 3439 N.D.

7* 375,000 92 27 >4000 90

8 567,000 200 N.D. 1634 77

Normal 200-300,000 180-200 60-70 1000-1500 0

Plateletswereisolatedfrom the patient'speripheralblood and labeled with"'Inoxine. After the labeled plateletsweregiven backto the pa-tient, blood samples were drawnat 15min,2 h, and every 12 h after injection. Survivalwascalculatedusingthemaximum likelihood

esti-mateof theinteger-orderedgamma function(29). *Splenomegalypresent.

Not done.

ComplementInteractions withParoxysmalNocturnalHemoglobinuriaPlatelets 135

2001

200].

I

!

cr.

200 i

L

200

-z

_

m

Ij

Ct.

I:to

Uk

200 400 600 800 GATED GREEN FLUORESCENCE

.r

F--. I

.1

:1

F-:11

11

.i

i

I

I

0

.t 6

t ,1:

t

I.

i:

t

I% ..

plateletantibody ( 18) thanarenormalplatelets. The interactions

of

platelets

with

thealternative

pathway

of

complement

arenot

aswell

understood.

Studies from this laboratory

using

CoVF-treated whole serum and

quantitative

antiglobulin consumption

assays

for

platelet-bound

C3

indicate that PNH

platelets

bind moreC3 than do

normal platelets (31).

To study the

interactions

of normal andPNH

platelets

with

complement

components in the absence of otherserum

proteins,

weused

purified

componentsand monoclonal antibodies

specific

for the C3c

region

of C3. We have demonstrated

that

the C3

convertase

activity

onthe

platelets

of many

patients

with PNH wasnot greater than thaton normal

platelets. This

could not

be

attributed

tothe presence of DAF eitheronthe

platelet

surface

orin

internal pools

that

might

be

expressed

when

complement

bindstotheplatelet. The

platelets

of 16 of 20

patients

examined werenearlyall

DAF-deficient.

Because

platelets

thatwere

lacking

in DAFdid not bind abnormalamounts

of C3,

DAFmaynot

playaspivotalarole in the

regulation

of the alternative

pathway

C3

convertase onthe

platelet

asit doeson the red blood cell

surface (1

1,

12). However,

DAF may be crucialtothe

regulation

of

the

classical pathway C3

convertase,

C4b2a,

onthe

platelet

surface. The absenceof

this

regulatory protein

onPNH

platelets

may

explain

the

increased

susceptibility

of the

PNH

platelet

to

lysis

by

antibody

andcomplement

(18,

31).

In

fact,

preliminary

data

obtained

on some

of the

same

patients reported

in this

study indicate

that

when

complement is activated

by

antibody

in

awhole serum system, PNH

platelets bind

bothmoreC3 and moreC9 than normal

platelets.

The

hypothesis

that DAF

is

not

crucial in

the

regulation

of

the

alternative

pathway

C3

convertase on

platelets is further

supported by

the

observation

that when

C3

was

deposited

onto

the surface of

platelets

via

the

alternative

pathway,

theplatelets

of

both normal donors and

patients

with

PNHreleased

factor

H, a plasma protein that

performs

a

similar

function

toDAF.FactorH

became

boundtothe

platelet surface

only

when C3 was presentontheplatelet. FactorHreleased by

thrombin stimulation did

notbindtotheplatelet

surface

in the absence

of

platelet-bound C3. The platelet,

therefore,

appears

to havea second mechanism to regulate

surface-bound

alter-native pathway C3 convertases. Other complement components in addition to factor H (32) have been described in platelets,

including Cl

inhibitor (33) and factorD(32). Whereas the

in-tracellular localization of

factor

Hhas not yet beendetermined,

the

othertwo

complement

components appeartoreside in the platelet alpha granules.

InPNHplatelets, the total amountof factor H present ap-pearsto belower than that in normal platelets. The factor H released bythrombin-stimulated platelets of two patients with

increased

C3 convertase activity was markedly deficient, although the total amount of

factor

H contained in these platelets was equivalenttothatamountreadily released from PNH platelets with normal C3 convertase activity. In addition, we report else-where that the inhibition offactor H release from normal platelets markedly enhances C3 convertase activity. Therefore, it appears that some PNHplatelets may be unable to release a sufficient amount of

factor

H to regulate the activity of the alternative

pathway

C3 convertase. The reason that this occurs in only some of the PNH patients studied remains to be elucidated.

Previous

studies from both our laboratory (31) and others

(34)

onthe

interaction

of the alternative pathway of complement with the platelets in PNH were all performed using whole serum systems. Thedata reported in this study support the observation

of Zimmerman

andKolb

₍₃₄₎

that the

platelets

inPNHdonot

take upmoreC3 than normal

platelets

when

exposed

to CoVF-treated serum.

However,

these workers

reported

an increased

uptake of

C8,

presumably

in the form of membrane attack

com-plexby the

platelets

ofonePNH

patient

under thesame

con-ditions. We

did

not seeany increase in the

uptake

of membrane attack

complexes

underour

experimental

conditions,

which

dif-fered somewhat

from those of Zimmerman and Kolb

(34),

par-ticularly with regard

to the presence of EDTA in the reaction

mixture.

Inwholeserumsystems, the

interactions

of the

platelets

with

activated

coagulation

factors,

especially thrombin,

aswell

aswith

complement

componentsmustbe taken into consider-ation. The presence of thrombin has been shownto

markedly

affect the interaction of

platelets

with

complement (35, 36).

Plateletscan

actively

rid themselves of

surface-bound

mem-brane attack

complexes

bya

calcium-dependent shedding

process

similar

to

the

patching

and

capping phenomena

seenin nucleated cells

(37).

In this

investigation,

this process was inhibited

by

performing

the

experiments

in thepresence of EDTA. Under these

conditions,

wehave

demonstrated

that the

platelets

from

patients

with PNH III

erythrocytes

do nottake up membrane attack

complexes

thatare

generated

inthe fluid

phase. However,

the PNHIII

_erythrocyte

willtake up

large

amountsof this

com-plex

resulting

in cell

lysis

(9,

10).

Because this ismost

likely

the

mechanism

that

contributes

tothe

severely

shortened in vivo survival ofthe PNH III

erythrocyte,

it follows thatweand others

(18)

have

found

in vivo

platelet

survivalwas normal inmost

PNH

patients.

This is

reflected

in the observed percentages of

abnormal

erythrocytes

and

abnormal

platelets

inthe

peripheral

blood

of

patients

with PNH. Becausethe

erythrocyte

lifespan

is shortened and the

platelet

lifespan

is

normal or

_nearly

so, the percentage

of

abnormal

platelets

may more

accurately

reflect theextentof

proliferation

ofthe abnormal PNH clone. Shortened

platelet

survival timewas

only

seenin those

patients

who had

splenomegaly

orin theone

patient

in the

study

whowas

entering

an

aplastic

crisis.

In

addition,

thePNH III

erythrocyte

is

markedly

sensitive

to

lysis

by

complement

because thecells will bear

fewer of

the membrane attack

complexes of

complement

than normal

erythrocytes

before

they

lyse. Although

we do nothave

direct

evidence

that

this

is also the casefor PNH

platelets,

it would

explain

the

observation

that PNH

platelets

are more

_susceptible

to

complement

lysis

when

complement

is

activated

atthe

platelet

surface,

e.g.,

by

anti-platelet antibody (18).

Acknowledgments

Weacknowledge the assistance ofDr. R. Edward Coleman with the

platelet survival studies. Our thankstoFrankCornew for performing theflowcytometryanalyses.

This workwassupportedbyNational Institutes ofHealth grant AM-31379 awarded to Dr. Rosse.

References

1.Rosse,W. F.,andC.J. Parker. 1985. Paroxysmal nocturnal he-moglobinuria. Clin. Haematol. 14:105-125.

2.Ham,T.H., andJ. H.Dingle. 1939. Studiesondestruction ofred blood cells. II. Chronichemolyticanemiawith paroxysmal nocturnal hemoglobinuria: certainimmunologicalaspectsofthehemolytic

mech-anism withspecial referencetosernm complement.J. Clin. Invest. 18:

657-672.

3. Peytremann, R., R. S. Rhodes, and R. C. Hartmann. 1972. Thrombosis in paroxysmal nocturnal hemoglobinuria (PNH) with

par-ticular referencetoprogressive,diffuse hepaticvenousthrombosis. Semin. Hematol. 5:115-136.

4.Parker, C. J., P. J. Baker, and W. F. Rosse. 1982. Increased

en-zymatic activity ofthealternativepathway convertase when bound to

the erythrocytes ofparoxysmal nocturnal hemoglobinuria.J.Clin. Invest.

69:337-346.

5. Pangburn, M. K., R. D. Schreiber, J. S. Trombold, and H. J. Muller-Eberhard. 1983. Paroxysmal nocturnal hemoglobinuria: defect

inthe factor H-like functions of the abnormal erythrocytes.J.Exp. Med. 157:1971-1980.

6. Rosse, W. F., and J. V. Dacie. 1966. Immune lysis of normal human and paroxysmal nocturnal hemoglobinuria(PNH)redblood cells. I. The sensitivity of PNH red cells to lysis by complement and specific

antibody. J. Clin.Invest. 45:736-748.

7. Packman, C. H., S.I.Rosenfield,D. E.Jenkins,Jr., P. A.Thiem, andJ. P.Leddy. 1979. Complement lysis ofhumanerythrocytes. Differing susceptibility oftwotypesof paroxysmal nocturnal hemoglobinuria cells

toCSb-9.J. Clin.Invest. 64:428-433.

8. Roualt,T. A., W. F.Rosse,S.Bell,and J.Shelburne. 1978. Dif-ferences in the terminalstepsofcomplementlysisof normal and

par-oxysmal nocturnalhemoglobinuriared cells. Blood. 51:325-330. 9.Gotze, O.,and H. J.Muller-Eberhard. 1970.Lysis of erythrocytes by complement intheabsence ofantibody.J. Exp. Med. 132:898-915. 10.Parker, C. J., T. Wiedmer,P.J.Sims,and W. F. Rosse. 1985. Characterization of the complement sensitivity of paroxysmal nocturnal hemoglobinuriaerythrocytes. J.Clin.Invest.75:2074-2084.

11.Nicholson-Weller, A., J. P. March, S. I.Rosenfield,and K. F. Austen.1983. Affected erythrocytes ofpatientswithparoxysmalnocturnal

hemoglobinuriaaredeficientinthe complementregulatoryprotein,decay

accelerating factor.Proc.NatL Acad. Sci. USA. 80:5066-5070. 12. Pangburn,M. K., R. D. Schreiber,andH.J.Muller-Eberhard.

1983.Deficiencyofanerythrocyte membrane protein withcomplement

regulatory activity in paroxysmal nocturnal hemoglobinuria.Proc.

Nat/.

Acad.Sci. USA.80:5430-5434.

13. Nicholson-Weller, A., D. B.Spicer, and K. F. Austen. 1985. Deficiency ofthecomplement regulatory protein, "decay accelerating factor",on membranesof granulocytes,monocytes, andplateletsin

par-oxysmal nocturnal hemoglobinuria.N.Engl. J. Med. 312:1091-1097.

14.Hoffman,E.M. 1969.Inhibition of complement byasubstance isolated from human erythrocytes.II.Studiesonthe siteandmechanism of action.Immunochemistry.6:405-419.

15.Nicholson-Weller, A., J. Burge, D. T. Fearon, P. F. Weller, and K. F. Austen. 1982. Isolation ofa human erythrocyte membrane

gly-coproteinwithdecay-accelerating activityfor the C3convertasesof the complementsystem.J. Immunol. 129:184-189.

16.Medof, M. E., T. Kinoshita, andV.Nussenzweig. 1984. Inhibition of complement activationonthesurface of cells after incorporation of

decay-acceleratingfactor (DAF) into their membranes. J. Exp. Med. 160:1558-1578.

17.Medof,M.E.,W. F.Rosse,T.Kinoshita,R.Silber,and V.

Nus-senzweig. 1985. Partial correction of theparoxysmal nocturnal hemo-globinuria (PNH)hemolyticdefects with decay accelerating factor (DAF). Clin.Res.33:554.(Abstr.)

18.Aster, R.H., and S.E.Enright. 1969.Aplatelet and granulocyte membrane defect inparoxysmal nocturnalhemoglobinuria.Usefulness forthedetection ofplateletantibodies.J.Clin.Invest. 48:1199-1210.

19.Ballow, M.,andC. G. Cochrane. 1969.Twoanticomplementary factorsincobravenom:hemolysis ofguineapig erythrocytes byoneof

them. J.Immunol. 103:944-952.

20. Tack, B. F., and J. W. Prahl. 1976. Third component of com-plement:purificationfrom plasma andphysicochemical characterization. Biochemistry. 15:4513-4521.

21. Lambris,J. D., N. J. Dobson, and G. D. Ross. 1980. Releaseof

endogenousC3b inactivator from lymphocytes inresponse totriggering

membranereceptors for

#IlH

globulin.J.Exp.Med. 152:1625-1644.

22. Schreiber, R. D., 0. Gotze, and H. J. Muller-Eberhard. 1976.

Nephritic factor itsstructure and function and itsrelationshiptoinitiating factor ofthealternative pathway.Scand.J.Immunol. 5:703-713.

23. Falk, R. J., A. P. Dalmasso, Y. Kim, C. H.Tsai,J.I.Scheinman,

H.Gewurz, and A. F. Michael. 1983. Neoantigenof thepolymerized ninthcomponentof complement. Characterization ofamonoclonal

an-tibodyandimmunohistochemical localizationin renaldisease.J. Clin.

Invest.72:560-573.

24.Fraker, P. J., and J. C. Speck. 1978. Protein and cell iodinations with a sparingly soluble chloramide 1,3,4,5-tetrachloro-3,6-diphenyl-glycuril.Biochem. Biophys.Res. Commun. 80:849-857.

25. Towbin, H., T.Staehlin, and J. Gordon. 1979.Electrophoretic transfer of proteins with polyacrylamide gelsto nitrocellulose sheets: procedure and someapplications.Proc.

Nail.

Acad.Sci. USA.

76:4350-4354.

26.Sokal, R. R., and F. J.Rohlf. 1981.InBiometry.W.H.Freeman &Co.Publishers, New York. 1-859.

27. Thakur, M. L., M. J. Welch, J. H. Joist,and R. E. Coleman.

1976.Indium-I1I1 labeledplatelets:studiesonpreparationandevaluation of in vitroandin vivofunctions. Thromb.Res.9:345-357.

28.Heaton, W. A., H. H. Davis,and M. J.Welch. 1979.

Indium-I11: a newradionuclidelabelfor studyinghumanplatelet kinetics.Brit.

J.Haematol. 42:613-622.

29.Murphy,E.A.,andM. E.Francis. 1971.Theestimation ofblood

platelet survivalII. Themultiple hitmodel.Thromb. Diath. Haemorrh. 25:53-80.

30.Harker,L. A. 1970.Thrombokinetics inidiopathic thrombocy-topenicpurpura. Brit. J.Haematol. 19:95-104.

31. Dixon,R.H.,andW. F. Rosse. 1977.Mechanism of complement mediated activation ofhumanbloodplatelets in vitro.Comparisonof normalandparoxysmal nocturnal hemoglobinuria platelets.J. Clin. In-vest. 59:360-368.

32.Kenney, D. M.,andA.E.DavisIII. 1981. Association of

alter-native complement pathwaycomponentswith human bloodplatelets: secretion and localization of factorD and

#I1H

globulin.Clin. Immunol.

Immunopathol. 21:351-363.

33. Schmaier,A.H.,P. M.Smith,and R. W.Colman. 1985. Platelet

Cl inhibitor.Asecretedalpha-granule protein.J. Clin.Invest. 75:242-250.

34.Zimmerman,T.S.,and W. P. Kolb. 1976. Human

platelet-ini-tiatedformation and uptake of the C5-9complexofhumancomplement.

J.Clin.Invest.57:203-211.

35. Polley,M.J., andR.Nachman. 1978. The humancomplement

systeminthrombin-mediated plateletfunction.J.Exp.Med. 147:1713-1726.

36.Polley, M.J.,and R. Nachman. 1979.Humancomplementin thrombin-mediatedplateletfunction:uptakeof the CSb-9complex.J. Exp.Med. 150:633-645.

37. Wiedmer, T., and P. J. Sims. 1986.

Repolarization

of the

mem-branepotentialof blood

platelets

after

complement damage:

evidence fora

Cal'-dependent exocytotic

elimination of

C5Sb-9

pores.Blood. 68:

556-561.