Immune complex processing in patients with systemic lupus erythematosus In vivo imaging and clearance studies

Immune complex processing in patients with

systemic lupus erythematosus. In vivo imaging

and clearance studies.

K A Davies, … , H L Beynon, M J Walport

J Clin Invest.

1992;

90(5)

:2075-2083.

https://doi.org/10.1172/JCI116090

.

Abnormal processing of immune complexes (IC) may be important in the pathogenesis of

systemic lupus erythematosus (SLE). The clearance of large soluble IC (comprising

hepatitis B surface antigen (HBsAg)/anti-HBsAg) radiolabeled with 123I was examined in

12 normal subjects and 10 patients with SLE. IC localization was analyzed by static and

dynamic gamma-scintigraphy. Initial IC clearance from blood was more rapid in patients

(median t1/2 = 2.15 min) than normals (median t1/2 = 5.15 min) due to more rapid uptake in

the liver. However, in the SLE group, up to 12% of complexes were released from the liver

after 30-50 min. Splenic uptake of immune complexes was reduced in the patients and there

was reduced ability to retain IC in this organ. Plasma complement levels and erythrocyte

complement receptor type 1 numbers were reduced in the patients, resulting in defective

opsonization of IC and reduced red cell binding in vivo. These observations support the

hypothesis that IC handling is abnormal in SLE.

Research Article

Immune Complex Processing in Patients with Systemic

Lupus Erythematosus

In Vivo Imaging and

Clearance

Studies

Kevin A. Davies, A. Michael Peters,* Huw L. C. Beynon, and Mark J. Walport

Rheumatology Unit and *Department ofDiagnostic Radiology, Hammersmith Hospital, London W12OHS, United Kingdom

Abstract

Abnormal processing of immunecomplexes (IC)may be

im-portant in the pathogenesis of systemic lupus erythematosus

(SLE).Theclearance of largesoluble IC (comprising hepatitis

B surface antigen (HBsAg)/anti-HBsAg) radiolabeled with

"23I

wasexamined in 12normalsubjects and 10 patientswith

SLE. IC localization was analyzed by static and dynamic

gamma-scintigraphy.Initial ICclearancefrombloodwasmore

rapid in patients (median 11/2 = 2.15min)thannormals

(me-dian t1/2 = 5.15 min) due to morerapid uptakein the liver.

However, in the SLEgroup, upto12%ofcomplexeswere

re-leased from the liver after 30-50 min. Splenic uptake of

im-mune complexes was reduced in the patients and there was

reducedabilitytoretain IC in thisorgan.Plasma complement

levels and erythrocyte complement receptor type 1 numbers

werereducedin the patients, resulting indefective opsonization

ofIC and reduced red cell binding invivo.Theseobservations

supportthehypothesis thatIC handling is abnormal in SLE.

(J. Clin.Invest.1992.90:2075-2083.)Key words: complement

*complementreceptor type1-mononuclearphagocyticsystem

*autoimmune disease * antigenprocessing

Introduction

There are several waysin which abnormal processingof im-munecomplexes (IC)' mayplayaroleinthepathogenesisof

systemiclupuserythematosus(SLE).Immunecomplexesmay

escapeclearance and catabolism bythe mononuclear

phago-cyticsystem(MPS)becauseofabnormaltransportor presenta-tion mechanisms (1). Furthermore, there maybe defects in

MPS functioninpatientswithSLE,andhypothesestoexplain

these have included "saturation" by immunecomplexes (2)

and intrinsic abnormalitiesofphagocyticcell function(3).

Addressreprintrequeststo Dr. KevinA.Davies,RheumatologyUnit,

DepartmentofMedicine,HammersmithHospital,LondonW12OHS, UnitedKingdom.

Receivedfor publication 14 October 1991 and inrevisedform22

April1992.

1.Abbreviations usedinthispaper:A-IgG, aggregatedIgG; CR,

com-plement receptor; HBsAg, hepatitis B surfaceantigen; IC, immune

complex; MPS,mononuclearphagocyticsystem.

Studies in vivo inprimateshave delineatedthemechanism

of clearance from the circulation of soluble immune

com-plexes. Inbaboons, IgG-containingimmunecomplexesfixed

complement,boundtocomplementreceptortype1 (CR1)on

erythrocytes and weretransportedtothe fixed mononuclear

phagocytic system wheretheywere clearedfrom the circula-tion(4).Severalinvestigatorshave used soluble immune

com-plexes or aggregatedIgG to study theclearancepatternof im-munecomplexes from thecirculation of humans (5-9). Ab-normallyfastclearanceofimmunecomplexes inpatientswith SLE was observed and correlated with hypocomplementemia (7). Thesestudieswereperformed using '25I-labeledimmune

complexesand itwas notpossibletoascertainthe site ofrapid

clearance of immune complexes. Imaging studies performed using heat-aggregated IgG showed accelerated uptake in the liver (10).

Wedescribe here theresultsof experimentsusing '231-hepa-titis B surfaceantigen-containingimmunecomplexestotrack the fate of immunecomplexesinnormal humans andpatients withSLE. Severalabnormalities of immune complex kinetics

were identified in patientswith SLE, hypocomplementemia,

and lowCR1 numbers. These were accelerated clearance of

immune complexes in theliver, impaired retention of com-plexes within the liver with release of comcom-plexes back to the

circulation, and impaired immune complex clearance by the

spleen.

Methods

Subjects

26subjects werestudied,18female and 8 males. 12 normal volunteers wererecruited fromthe laboratorystaff, aged from 23 to 59 yr, 6 female and 6 male. Fouradditional normal subjects (threefemale,onemale)

received radiolabeledantigenalone. Thepatientswere10 subjects with SLE, aged from 21to56 yr, 9 females and 1 male, all ofwhom fulfilled the revisedARAcriteria for this disease.Atthetime of thestudy,8of the 10patientshadclinicallyactive diseaseinvolvingmorethanone

organsystem, withanerythrocytesedimentationrate>40mmin the first hour in allcases, andaC4 <50% normal human serumpool,

elevated DNAbinding (>30%byFarrassay),orboth. Twopatients

werestudiedontwooccasions. Fully informed consentwasobtained from all theparticipants for the study, which wasapproved bythe HammersmithHospitalEthics committee and theAdministrationof RadioactiveSubstancesAdvisory Committee (ARSAC).

Allsubjectsreceived 120 mg twicedailyoforalpotassiumiodide for 36 h before each study andafurther 48hafterinjectiontoblock

thy-roid uptake of123I.

HepatitisB

surface

antigen

(HBsAg)

Thiswas agiftfromDr.J.A.Schifferli(H6pitalCantonale,Geneva,

Switzerland).Theantigenwasin theform ofalargepolymeric protein

(molecularmassaround3,000 kD)andwaspyrogen free.It hasbeen previously characterized both in vitro and in vivo(11).30 ,gantigen

wasused for thepreparationof each batch of IC.

J.Clin.Invest.

©TheAmericanSociety for ClinicalInvestigation,Inc.

0021-9738/92/11/2075/09 $2.00

Anti-HBsAg antibody

Polyclonal antiserum to HBsAg was also a gift from Dr. Schifferli (Croix-Rouge, Bern, Switzerland). One vial of sterile lyophilized im-munoglobulin (> 95%IgGi) wasdissolved in 10 ml sterile water (An-tigen, UK),toproduceafinalconcentration of 50 U/ml, immediately before ICpreparation.

Radiolabeling of

HBsAg

Theantigenwaslabeledwith 123I (Medgenex, Brussels,Belgium),using N-bromosuccinamide ( 12),to ahighspecificactivity(25-30 JLCi/Mg). 301Agofantigenwaslabeledimmediatelybefore each study; free and boundiodinewereseparatedonasterile G25 columnpreblockedwith sterile human albumin. The123Iwas >97%proteinbound in the final preparation in allcases.

IC preparation

HBsAg/anti-HBsAg ICwerepreparedasfollows: 30_,gof radiolabeled HBsAg wasincubatedwith 10mlofantiserum for 60 minat370C.This correspondsto afivefold antibodyexcess.Two5-mlsampleswerethen drawnintosyringes, and the totalradioactivityin each dose measured

using a radioisotope calibrator (ARC 120, Capintec; Amersham,

Amersham,UK).AnICmixturecontaining15g ofantigenwasused for eachsubject,withanactivityof between 300 and 500 MCi. For each batchof labeled ICprepared,onepatientandonenormal controlwere

studied in parallel.

Injection

and

imaging

protocol

Anintravenous cannulawasinsertedinto thesubject'sleft antecubital fossa forinjectionof labeled IC and intoasimilar line in therightarm

for bloodsampling.5 mlof ICwasinjectedas arapidbolusattime0,

with thepatient positioned over a gamma-camera (IGE 400T on-line to anMDS A2computer).Dynamicimagingwasperformed (serial20s

frames) for 50min, followed by 5-min anterior andposteriorstatic imaging at1,4, and 24 h. Foursubjectsunderwentdynamic scanning

for 2 h, and whole-body scanning wasalsoperformed intwocases,

usingaStarcam(IGE)gammacamera.

Blood

samples

Preinjectionserum anderythrocyteswere obtained from allsubjects

for theestimation ofserumcomplementlevels and red cellCR1

num-bers.Postinjection,6-mlvenousbloodsamplesweretakenintoEDTA

pH 7.2at2, 5, 10, 15,20, 25, and 30 min, and thenat10-minintervals upto1-2h.Furthersampleswerethenobtainedatthetimeof

rescan-ning.Thefollowingassayswereperformedoneachbloodsample: Measurementofwhole blood I23.A1-mlaliquotfrom each EDTA sample was counted for 123I activity using an automated gamma

counter(LKB Wallac, Turku,Finland).

Identification

ofprotein-bound 23LA I-ml aliquotof each EDTA

samplewaslysedbymixingwithanequal volume ofwater. Protein-boundactivitywasmeasuredby treatmentof

250-2000-Ml

aliquots of plasmaorlysed whole blood with trichloroacetic acid (BDH Chemicals Ltd., Dagenham, Essex, England)at a10%vol/volfinal concentration, andcountingfor123Iin the washedprecipitate.

Measurementoferythrocyte-bound123I.2ml of each freshly drawn

EDTAsample wasimmediately diluted 1:4 in ice-cold PBS pH 7.4/1%

BSA, and red blood cells were rapidly separated from plasma by centrif-ugationat1,500 g. Cells were subsequently washed three times in the

samebuffer. Therewas nosignificant release of erythrocyte-bound ICs duringthewashing procedure. Washed cells were counted for1231 activ-ity, with correction for isotopic decay in the later samples.

Characterization ofICs. (a) Coprecipitation of'23I-HBsAgusing

StaphylococcalproteinAwas asfollows: precipitation of

'23I-HBsAg-containingICswasperformed by addition of 50,lof a suspension of

Staphylococcalprotein A-coated beads (Calbiochem, Cambridge

Bio-science, Cambridge,UK) to 100 Mi plasma samples, followed by incu-bationfor 10 min at room temperature, and two washes in PBS/ 1% BSAperformed by centrifugationat1,000g. withresuspension ofthe

pellet eachtime. The 1231 activity in pellet and supernate was then measured.

(b) Sucrose density gradient centrifugation was as follows: to ascer-tain complex size, 100-MIA aliquots of fresh plasma were layered ontoa

10-50% sucrose gradient (13), and 1231activity was measured in 16 fractions of 300Mlafter centrifugation for 4 hat30,000 g. Radiolabeled

IgM,IgG,andBSAwereusedassize markers.

Complement

assays

Plasma C3 and C4 levels were measured using single radial immunodif-fusionin 1.2% complement fixation diluent-agarose gels containing antibody (polyclonal goat anti-human C3, or sheep anti-human C4

[Serotec Ltd., Oxford,UK]). CH50wasmeasuredusingaplate

hemo-lyticassay(14).

Assay

for

enumeration

of erythrocyte CRI

Meannumbers of antigen sitesonerythrocytesweremeasuredusinga

radioligand-bindingassay,aspreviouslydescribed (15). Monoclonal antibody to CRIwasEI I( 16) (donated by Dr. Nancy Hogg,Imperial

Cancer Research Fund, London, England). Radiolabelingwas per-formed with 125I(Amersham) using Iodogen ( 17),toaspecificactivity of 1-2 MCi/Mg.

Image analysis

Quantificationofuptakein the liver andspleenwasperformedfrom

radioactivitycount ratesinspecific "regionsof interest" drawn

respec-tivelyaround the liver andspleenonanterior andposteriorimages. Hepatic andsplenicuptakewasquantified bycomparingthe injected

countswith thegeometricmeanofanteriorandposteriorcountsin the

specific regionsof interest. Correctionsweremade forphysical decayof the radionuclide andphotonattenuationaspreviouslydescribed (18).

Statistical methods

Elimination half-times,time of 90% hepatic uptake, splenicuptake, and other variableswerecomparedbetween thepatientsand control groupusing theMann-Whitney Utest.Correlations between C4 and C3 levels and IC clearance rate, red cell CR1 numbers and IC binding, and CR1 numbers andhepaticclearancewereanalyzed usingthe

non-parametric Spearmanrankcorrelationtest.

Results

Studies with HBsAg alone

Control experimentswereperformed tocharacterizethe

clear-ancepattern oftheHBsAg when injected alone. 15 Mg of

123I-la-beledantigenwasinjectedinto twoimmune andtwo

nonim-mune normal volunteers.Localizationof antigen to the liver

andspleen occurredinall cases,maximumuptakeinthe liver occurring after 22 and 24

min,

and28 and 35 min in the

im-muneand nonimmune subjects, respectively.Splenic uptake

ofradioactivityat 1 hwasbetween 4 and 7.5% of the injected

dose inthefoursubjects studied,and there was no difference between immune and nonimmune individuals. However,

morerapid splenic localizationoccurred in the immune (mean

t1/2=

9min)than inthenonimmune subjects(mean

_t,/2

=21

min). Less than 2% ofcirculatingactivity was measured on

erythrocytesinthenonimmunesubjectswhereas 12 and 14%

bindingwasobservedintheimmune controls. Maximum

bind-ingin the immunesubjects occurredbetween 5 and 10

min,

much less rapidly than in those patients injected with pre-formed IC(see below).

IC

imaging studies

Initial IC

clearance

and

localization. In SLE patients and

con-trolstherewas

uptake

of IC in theliverand spleen, withrapid

Normal subject

-*-- liver

-0-- spleen

-0-bloodpool

I 11I

0 10 20 30 40 50

0 10 20 30 40 50

time afterinjection(min)

Figure 1.(a) IC clearance kinetics inatypicalnormal controlsubject andapatient with SLE. ICwererapidlyclearedfromthe blood and taken

upbythe liver andspleen. Complexclearance fromthe circulationwas morerapid inthe SLE patient, and therewasmorerapid IC uptake in

the liver.(b) Whole-bodyscanperformedat2 h inanormalsubject(anterior,left;posterior,right).Traceruptakeisseenin the liver and spleen andfree iodineproduced byIC catabolism isvisibleinthe bladder.

kidneys, and other organs revealed no specific uptake above bloodpoollevel in eithergroup(Fig. 1b).The initialrateof IC

clearance from bloodwassignificantlyfasterinthe SLEgroup

(mediant112 =2.15min [range 1.3-6.6

min])

than in the

nor-mal controls (median t112 = 5.15 min [range 3.6-14]) (U = 12.5,P<0.002) (Fig. 2).As is showninFig. 1 _a,therapid

initial removal ofcomplexes from the blood in the patients with SLE correspondedwith the timecourseof localization of immunecomplexesinthe liver. The median timeatwhich 90% of maximumhepatic uptakewasachievedinthepatientswith SLEwas9.0min (range4.3-18min)comparedwith 16.0min

(13-23min)inthe controlsubjects (U= 8.0,P<0.002).At

10min,between 27.3 and 67.5% (median 40.7%) of injected

ICswere detectable in the liver in the normal subjects

com-paredwith 43.0to79.6% ofinjectedIC(median 56.3%)inthe

patientgroup(U = 27,P<0.05).The immunestatusof the

subjecttoHBsAg (measured byastandard ELISAinthe

Virol-ogy Department, Royal Postgraduate Medical School,

Ham-mersmith Hospital)hadnobearingonclearance rate, hepatic

uptake,orsubsequent handlinginthestudiesperformedwith

immunecomplexes.

ICcatabolism and release.Although therewasrapidinitial

hepatic localization ofcomplexes in both normals and

pa-tients, in bothgroupstherewas a fall inhepatic activity ob-served between 30minand 2 h. Thiswas more noticeable in

thepatientgroup, asdemonstratedby comparisonof the ratio

ofhepaticcountsmeasuredat40and 60mininthetwogroups,

median ratio 1.24(range 0.88-1.44)in the studiesperformed

in patientswith SLEcomparedwith 0.75 (0.48-0.95)in the

controls(U= 2,P<0.02) (Fig.3B).

a

100 1

b

a) O *0 aD 60

-0

ai)

c - 40

-20

-01

100

80-CD a)

0 D

60-*0

a)

0

.- 40

20

0-4;

An-0 0 0 0 0₀₀ 0 0 0 goo Il

Controls SLE patients

Figure2.Half-time of initial IC

clearance in normals and

pa-tients. Clearance ofcomplexes

from the circulationwasmore

rapidin the SLE patients than in normals. The subjects whose clearancecurves areshown in Fig. 1 aareindicated byopen

circles. 100 -D CD 90-0

-80-2u 70-D .C 0~

0)H,

60-0 50 A _B . 0 0 0 00 100 0 1.6 -m '0 *0 co=

1.2-E

:

2,C

- _0.8

-X.'

.2E

Cu 0)

0 0.4

-0 co

0.0

Controls SLE_patients

0: 00 0 0.0 40 000 Il

Controls SLEpatients

Measurement ofsequential whole _{blood, protein-bound,}

and red cell-boundactivityafterinjectionof IC indicatedthat significant releaseofprotein-boundradioactivityfrom theliver

was occurringinthe patientgroup after 30-40min. Typical dataareshowninFig. 4,in which thepatternsofIChandlingin

anormal subjectandanSLEpatientarecompared.TCA

pre-cipitation data (Fig.3A)indicated thatagreater proportionof

theactivitydetectedin the bloodat 1 hwasproteinboundin

the SLEgroup(median 86.7%, range 77.0-96.7)than inthe

controls (median69.7%,range52.0-79.1) (U=2.0,P<0.02).

Precipitation of thismaterialusingSepharose-Staphylococcal

protein A showed that the '23I-HBsAg remained complexed withIgG(Fig. 5). Amaximum of 79% ofprotein-bound activ-ity could be precipitated with Sepharose-Staphylococcal

pro-tein Aat100minin thepatientsstudied. The sizeofthis

mate-rial was assessed by sucrose density gradient centrifugation,

which showed itto be composedprimarilyof material of

be-tween 35 and 50

S,

intermediatebetween antigenand the in-jected ICs (Fig. 6,aandc).

Splenic IC uptake.Therewas

significantly

reducedsplenic

uptake of immune complexesduringthe first houramongthe

patients (median, 9.03% ofinjected ICs; range, 4.05-23.7%)

compared with the normal controls (median, 23.9%; range,

17.9-30.7%)(U= 4, P<_{0.02) (Fig.} 7a). The abilityofthe

spleentoretain ICs after initial uptake intotheorganwasalso

abnormal amongtheSLEpatients.The_{activity remaining}in

the spleenat24 hafter injectionwasexpressedas apercentage

of the maximumorganuptakemeasuredduringthe first hour

after injection of the ICs. Among the SLE patients median

uptake was39% maximum _{(range: 24-52%)}compared with

normal subjects (median, 65.5%; range, 58-73%)(U= 0, P <0.002)(Fig.7b). Itwasnotpossibletotrack thelocalization ofIC for>24 hbecause_{of the short half-life of 1231.}

Factors

influencing clearance

pattern in

normals and

patients

The maindifferencesbetween the controlsubjectsandpatients with SLEwereas follows: (a)initial IC clearance was more

rapid in patients than in normals,ascomplexes localized in the

liver, (b) IC release from the liver occurred subsequently in

patients;and(c) therewassignificantlyreduced splenic

com-plex uptake in the patientgroup.The possiblefactors contribut-ingtothese differenceswereexplored:

Hypocomplementemiaand IC size. Among the 10subjects

with SLE therewas aclose correlation between the level of C4

in the plasma and thet112of IC clearance,rsp=0.997 (Fig. 8).

Therewasalsoacorrelation between plasma C3andthe

_t,2

of

Figure3.(A) TCA-precipitable activitymeasured in bloodat I hin

normals andpatients.Agreaterproportionof radioactivitywasstill

proteinboundat 1 hin patients than in normals. (B) Ratio ofhepatic

activitymeasuredat_{40 and 60 min in normals and patients. There}

was afall inhepatic activitybetween 40 and 60 min in the SLEgroup,

correspondingtorelease of ICs.

IC

clearance,

rsp =0.81 1.These observationssuggestthat

hy-pocomplementemia

might

playanimportantrole in

determin-ing

theabnormalICkineticsamongthe

patients

withSLE. One

possible explanation

for this link is that the ICsremained

_larger

in thepatients after injectionthanamongthenormal controls

because ofinefficient solubilization in the presenceof

hypo-complementemia.

Assessment ofIC size

by

sucrose

density

gradient

centrifugation performed

on samplesofplasma ob-tainedimmediatelyaftercomplex injectionwasinsupportof this

hypothesis. Typical gradient

dataareshown inFig.6b. A shift in

peak complex

sizewas seenin the 3- and5-min

samples

drawn froma normalsubject,whichwasnotobservedin the

samples

from anSLE

patient.

Erythrocyte

CR1

number andIC

binding.

The median

erythrocyte CR1

number measured in the normal control

groupwas 950(range 467-1,218). The median value for the SLE

patients

was

significantly

lower,482 (range 78-969)

(U

=

19,

P <

0.02).

Among all subjects there was avery close

linearcorrelation between the maximumbinding in vivo ofIC

to

erythrocytes

and theCR1 number

_(rsp

=0.96). CR1

num-bersonerythrocyteswerealsocorrelatedwiththerateof

clear-anceofICsbytheliver

_(rsp

=0.62, P=0.002).

Patients

studied

ontwooccasions

Twopatientswith SLEwerestudied twice,atintervalsof 8 and 9mo,respectively.Bothhadbeentreatedwithcorticosteroids

andazathioprine duringtheinterim period,with clinicaland serologic improvementmanifestbyarisein C4 anda fallin

DNAbinding (seeTableI).CR1 numbersonerythrocytesrose

by only

14 and 12 moleculespercell, respectively. The

mea-sured

t112,

time for 90% liveruptake, andpercentagehepatic

andsplenic uptake dataareshown in TableI.Inbothsubjectsa

rise in C4 correlated withanincreasedclearance timeand

re-ducedhepatic uptake,with correspondinglygreatersplenic

lo-calizationofcomplexes.ICrelease intothecirculationbetween

30 and 60minoccurredbutwasreducedin thesecondsetof

studies.

Discussion

Theuse of'23I-labeled soluble immunecomplexes offers the opportunity for the analysis of the physiological mechanisms

30000

-a)

C 20000

-CD

U,

0

c

°

1000

0-Normal

₃₀₀₀₀

-20000

-10000

-0 1

I lI

0 10 20 30 40 50 60 70

Time after injection(min)

SLE patient

-0-- _TCA

* Wholeblood

-0--D RBC

\1

I I

0 10 20 30 40 50

Time after injection(min)

Figure 4. Radioactivity measured

sequentiallyin blood after IC injec-tion, whole blood, TCA-precipita-ble, and RBC-bound activity, com-parison between a typical normal I andapatient. There was more rapid 60 70 _{clearance of IC in patients, with}

re-duced binding to erythrocyte CR 1.

ofimmunecomplexclearance fromthecirculationin humans and how these may be perturbed indisease. We found three

main differences between the clearance pattern of immune

complexes in normal subjects and patients with SLE. Clear-anceof immune complexesby theliverwasfaster in patients with SLE than incontrols (the opposite result to that

previ-ously observed for clearance of IgG-coated erythrocytesbythe spleen

[3]),

andtheincreased clearance ratecorrelated with

hypocomplementemia

andreducederythrocyteCR1 numbers. Thesecond differencewasthatuptake andinternalization of immune complexesbythe MPSoftheliverwas abnormal in

patients with SLE;up to 12%oftheinjectedHBsAg,still

com-100

-80

- 6040 -20

-

O--* normal(TV)

-0-- _{patient (CP)}

0 20 40 60 80 100 120 * - normal(HB)

-0--- _{patient (RB)}

I

0 20 40 60 80 100 120

timeafterIC injection (min)

Figure5. Precipitation of IC by StaphylococcusAbeads in normals andpatients.Inthe SLEpatientsbetween 40 min and 2 h therewas

anincrease in theproportionofprotein-bound activityinblood whichwasprecipitable by StaphylococcusA.

plexed with antibody, was released back into thecirculation fromtheliver,aphenomenonnotseenin normal control

sub-jects.Thethirddifferencewasthat therewasmuch reduced IC uptake in the spleen in the

patients compared

with controls.

Weperformed preliminary experimentstostudy the clearance

of

labeled

antigen

alone in immune and nonimmune

subjects

and the results weresimilar to those obtained

previously by

Madietal.(

11).

Antigen cleared in both liver and spleenbut

with kinetics much slower than those of IC. Maximum in vivo binding ofnascentICtoerythrocyte CR1 inimmune

subjects

was 14%,similarboth tothatmeasuredby Madiet al.(11)and

that observed inour

previous

study ofinvivo ICformation and clearancein

patients

receiving radioimmunotherapy ( 19).

Threeinterrelated factors appearedtobeimportant in

de-termining the faster clearance and

localization

of immune complexes intheliver in

patients compared

withnormal

sub-jects. Thesewere

hypocomplementemia,

large IC

size,

and

re-duced binding to erythrocyte CR1, each of which was

asso-ciated with increasedrateof uptake of immune

complexes

by

theliver.Aclosecorrelationwasobservedbetween the levelof

theclassical pathwaycomponentC4 and the initialrateofIC

clearance. Theresultsof

previous

workinvitro

(20-22)

and in

vivo(7)implied that low levels ofcomplement mighthave two

maineffects oncomplex

processing. Firstly,

ICsize wouldbe

expectedtobe greater duetodefective complex solubilization (21, 22). Thiswasconfirmed in thepresentstudyonsix

occa-sions

using

sucrose

density

gradient analysis. Secondly,

defec-tive

coating

of immune

complexes

with

complement

would

reduce

binding

toCR1 onerythrocytes. CR1 numbersonthe

erythrocytes of

patients

withactive SLEare

significantly

lower

than in normal controls

( 15, 23-25).

Thiswasalsothecasein

the present

study,

and low levels of

erythrocyte

CR1 correlated closely with reduced IC

binding

to

erythrocytes.

Thescanning results showed release of

1231I

counts from the

liverduringthe first 60 min among the SLE patients, which was notseen innormal subjects. Two possibleexplanationsforthis

observation are fasterinternalization and catabolism of com-plexes in the SLEpatients followedby early releaseoffree

'23I

orimpairedcellularinternalizationofimmunecomplexeswith releaseof immunecomplexes back into thecirculation.Several linesof evidenceare infavor ofthesecondhypothesis:(a) the

1231counts wereproteinbound asjudgedbytheir coprecipita-tion with

protein

after trichloraceticacid; (b)they werestillin

Un

g

cu

a) .0

0_Ci.).cL

Q°%

Q(I)

on0

.0 G)_c

.0

-0 100

-co

(D

m

80-co}. 60-cn C >

40-0.0

.0 CD)

c-0Q

20-w

..-0

* IC

-0-- HBsAg

I

I I I 1

2 4 6 8 10 12

~~

~Normal

-*11-t=0

--- =3min

-/ -_{t=5 mins}

I

2 4 6 8 10 12

Normal

I

X -0- t-3min

-0-_t=50min

1

l

14 16

100 80

-60

-40

-20

-I

14 16

0

-100

-80

-60

-40

-20

-0

-I I lI

0 2 4 6 8 10 12 14 16

Fraction

Patient

I I i I I I I

0 2 4 6 8 10 12 14 16

*Ag

a5floneak|

Patient

I

_

*Ag

alone

I I

0 2 4 6 8 10

Fraction

12

Figure 6.(a) Sucrose density gra-dientprofilesgenerated by radiola-beledantigen alone and by IC prep-arationimmediatelybefore

injec-tion. Theantigenis - 30 S insize.

(b)Sucrosedensity gradient analysis of plasma samples obtained from

apatientandanormalsubjectatt =0, 3,and 5 min.(c) Sucrose

den-sitygradient analysisof plasma

samplesobtained fromapatientand

anormalsubjectatt=0, 3, and 50 min.At50 min material was present I

--I

in the circulationintermediatein 14 16 size betweenantigen andinjected

ICs.

the form of immunecomplexesbecausethey boundto Staphy-lococcal protein A, which only bound the '23I-HBsAg when complexedwithantibody;and(c)the sizeofthereleased

mate-rialwas - 30 S when analyzed by_sucrosedensity

centrifuga-tion. In contrast, therewasevidencethatuptake andprocessing

of immunecomplexes in both liver and spleen, with

conse-quentrelease offree i23I,wasmoreefficientamongthe normal

controlsubjects(Fig. 3).

There isaprecedent for theseresultsfromexperimentson

clearance mechanisms of immune complexes in mice with SLE-like disease. The clearance of cross-linked mouse

anti-DNPoligomerswasstudied inNZB/W and control mice,both

invivo (26) and in isolated perfused livers(27). Immune

com-plexes were cleared in vivo mainly in the liver, 76% of the

injectedlabellocalizedtothe liverat1h.Initialclearancefrom

the circulationwas morerapidinthe NZB/WF1 lupusmice

than in control mice (26), analogoustotheobservations

re-portedinthepresentstudy. The earlystagesof ICuptakewere

examined by perfusing isolated murine livers viathe portal

veinwith labeled immunecomplexes.Therewasinitially

en-hanced hepatic uptake in the autoimmune mice compared with controls. However, the complexesintheNZBand NZB/

WF1 lupusmice couldbereadily displaced by saturating doses

of heat-aggregated human y-globulin, suggesting impaired bindingtoKupffercells anddefectivephagocytosis (27), simi-lartothatpostulated inthepresentstudy.

The pathway of ICclearance in nonhuman primates has

been established in baboonsandchimpanzees (4, 6, 28). Large

immune complexes injected into the circulation bound effi-cientlytoerythrocytesandwerecleared in the liver.Blocking experiments using monoclonal anti-FcyIII in chimpanzees

showed thatthesereceptorswereinvolved inhepaticclearance

of soluble anti-dsDNA/dsDNA ICs(28). Maximal blockade

ofICuptake bythe liverwasobserved with ICfractions that did

notbindtoerythrocytesandhadpresumably boundless C3b.

Thesedataoffer ahypothesis forthe release of ICs from the

liver, observed in the present studies. Kupffer cells bear the

followingFc andcomplementreceptors:Fc'yI, II,andIII,and

a

c Co

0 0 E

E3

E

b

100

-80

-60

40

-20

-0

-0

100-80

-8 60

E

E

40-E

,e

20

0 0 100 -~

80 -Cn

8

60-E

.E 40 -E

8 20

-0

-6

35-a

30-0

25-o

0

'a

:-(D 10

-5

5-

o-b

80 -70

60 -50

-40 -30

-20 5 10

0

0*::

0

000

TableI. Comparison ofImmune Complex Processing in TwoSLEPatientsStudied onTwoSeparate Occasions

.

0

*:g

S

I I

Controls SLEpatients

Figure7.(a)SplenicICuptake

at 1 h(percentinjected dose)

%.l

in normals and patients.There was reduced uptake in the pa-tientgroup.(b) Retention of IC in the spleen at 24 h in nor-mals and patients. There was areduced capacity of the

spleen

toretain IC in the SLE Controls SLEpatients patients.

CR1, CR3, and CR4 (29-31). Myoneset al. (32) demon-strated by radioligand bindingassaysusing specific monoclo-nal antibodies that CR1 wasexpressedon monocyte-derived

macrophagesat - 25% ofthe level ofCR3 and 50% ofthe level

of CR4 expression, but only showed weak staining of tissue macrophages byimmunoperoxidase-labeled anti-CR1

antibod-ies. More recently, immunohistochemical analysis of CR

ex-pression has been performed on normal human liver tissue

obtainedbypercutaneousbiopsy(33).Kupffercellswere

iden-tified using monoclonal antibody EBM 1. These cells

cos-tained strongly in allcaseswith monoclonalantibodiestoboth CR1 and CR3 whereas weakerstainingwasdemonstrated with anti-LeuM5 (anti-p 150,95 ), suggestingalower levelof

recep-torexpression. Itmaybe thatligation of both Fc and CRs is

requiredtomediate efficiently endocytosis and phagocytosis of ICs. In hypocomplementemic subjects, immune complexes bearing low numbers of C3 and C4maybindtransientlytoFc

receptorsontheKupffer cell surface and thenaproportionare

released back into thecirculation.

We previously studied the clearance mechanisms in

hu-mans of 251I-soluble tetanus toxoid (TT)/antitetanus toxoid

complexes (7, 8) and found that complexeswerecleared from

the circulationmorerapidly in patients with SLE and/or

hypo-complementemia than in normal subjects. The site of

clear-r =0.997 6

5-o

__°_ Figure 8. Correlation between

0 20 40 60 80 10 120 plasma C4 and initial IC

clear-plasma C4(%NHS) ance rate in patients with SLE.

Patient DNA C4 T,,2 T90 MaxSp. Max Liv E-CRI E-IC

% % min %injected mol/cell %bound

R.B. 1 66 18 1.33 6.0 12.27 46.45 490 26 R.B. 2 33 28 1.90 9.0 16.1 41.05 504 28 C.P. 1 64 15 1.30 9.0 21.6 58.02 570 32 C.P. 2 20 40 2.3 12.5 28.0 53.10 582 34

DNA, percent DNAbinding measured by Farr assay;

_T,/2,

half-time forinitial IC clearance; T90, time for 90% hepatic IC uptake; Max

Sp/Liv,maximum percent splenic and hepatic IC uptake; E-IC, max-imum percentbinding of IC to erythrocytes. C4 levels are measured by radialimmunodiffusionassay, and reported as a percentage of values obtained using poolednormal human serum.

anceof these complexeswasnotascertained andweinterpreted theveryfast clearancerateof complexestobedue totrapping of ICs outside the MPS, possibly in the pulmonary vasculature. Thisinterpretation wasprobably incorrect inthelight ofthe present data.Other studieshave also shown IC clearance to be

abnormally fast inpatients with SLE, notablya recentstudy by Madietal. (9), using similar HBsAg/anti-HBsAgcomplexes tothose inthepresentstudy, in whichthe clearancesites ofthe immunecomplexes were not visualized.

Another model forthe studyofIC clearanceisthe useof aggregated IgG (A-IgG) (6, 10). Halma et al. (10) reported

abnormally rapid initial A-IgG clearance by the liverin

pa-tients with SLEand decreasedmaximumsplenic uptake, but

therewere nocorrelations observed betweenE-CR 1 andA-IgG

binding

to erythrocytes or between complement levels and elimination kinetics.

Itis ofsomeinterest to contrast the presentfindings with the resultsof analyses of MPS function using IgG-or comple-ment-coatederythrocytesasprobes. Erythrocytescoatedwith anti-rhesus IgG, which did notfix complement, weremainly

removedfromthecirculationin the spleen, clearancemediated byFc receptors(34, 35).In contrast,erythrocytescoatedwith IgM and C3areretainedtransiently in the liver by

ligation

of C3receptors(34, 36, 37). Catabolism oftheC3onthe erythro-cytes to C3dg was accompanied by theirrelease back to the

circulation.Wedonothave

precise

data on theratio of

immu-noglobulintoC3on ourimmune complexes,asthese parame-ters

changed

quickly

in vivoafter IC

injection

(illustrated

in

Fig. 6

b,

which shows thechangesin IC size after

injection).

However, up to 79%oftheimmune complexes boundto

eryth-rocyte CR1 in

vivo, showing

thepresenceofsufficient C3bon

the

complexes

to

ligate

these receptors

efficiently.

Further-more, themedian

t1/2

of IC

binding

to

erythrocyte

CRIwas7.8

min,longer than the median

_t,12

ofcomplexclearance,

show-ingthat the ICstillbore C3boriC3batthetimeofclearance. Clearance of IgG-coated

erythrocytes

by

the

spleen

was

foundto be delayed in

patients

with SLE

(3,

38) compared

with normal subjects, andthese datawere

interpreted

to sup-port the

hypothesis

that"reticulo-endothelial" blockade

(2),

possibly by

high

levelsof

endogenous

immune

complexes,

may

occurin

patients

with SLE. The present data and other studies

onthe clearanceof soluble immune

complexes

donot

support

the

hypothesis

thatthere is blockade of the MPS in

patients

a

a_

ME

.2 E c 0 co

0ME

withSLE. Itispossiblethat there is an isolated defect ofsplenic

MPSfunction, but abnormal splenic blood flow has been dem-onstrated in patients with SLE and this may be a factor

in determining the delayed uptake of IgG-coated

erythro-cytes(39).

Thethird major difference inimmunecomplex

processing

between normal subjects andpatients with SLE observed in

this study relatedtotheuptakeof IC in thespleen.At 1 h after

injection only - 10% ofinjected ICwerepresent in thespleen

in thepatients compared withonequarter inthe normals,

sug-gesting preferential uptake. in the liver and reduced initial splenic localization. Wealso found thatantigen wasretained

much moreefficiently in the spleen in thenormalsthan inthe

hypocomplementemic subjects: 15-20% injected activitywas

still measurableinthespleenat 24 h in normalscomparedwith 3-5%inpatients. The splenic clearance of IgG-coated

erythro-cyteshasalsobeen shown to be abnormal inpatientswithSLE,

asdiscussed above, although early work suggestedacorrelation between

_tl

2ofE-IgGclearance and disease activity (3, 38),

laterstudieshavefailedtoconfirm these observations ( 35, 40). Theuptake of radiolabeled A-IgG bythespleen has also been

shownpreviouslytobereduced in SLEpatients (5).The

ana-tomicalorganizationof thespleen isideallysuitedtothe

pro-cessing of erythrocytes, particulate antigens,andopsonized im-mune complexesdeliveredonerythrocytes boundtoCR1.The

splenic

redpulp hasacomplex and highly specialized

vascula-tureorganizedin areticular mesh thatfunctions asafiltration bed(41).Plasmaand cellsare

progressively

separatedasblood

passesthroughthe organ,resulting ina

relatively

elevated

he-matocrit in theredpulp,anenvironment suitedtothe interac-tion ofopsonized IC boundtored cell CR1and theperiarterial macrophages. Hypocomplementemia, and reduced

erythro-cyteCR1 result in defective

opsonization

of IC and much

re-duced

binding

to

erythrocytes.

We

hypothesize

thatit is

abnor-maldelivery of ICtothe mononuclear

phagocytic

systemthat

resultsinreducedinitialuptake and impaired complex

reten-tion inthespleen.Aparallelmay bedrawnbetweenthese

re-sults and

experiments demonstrating

the

importance

of the complement system in determining the

binding

of immune complexesto follicular dendritic cells (42-44),aprocess

im-portantin thegeneration ofmemory Blymphocytes.We specu-late that hypocomplementemia and abnormal processing of immune complexes may be associated with abnormal path-waysof antigen processingandpresentationand that this may be afactor inpromoting the autoimmuneresponseinSLE.

Inconclusion, wehavepresented evidencethat the clear-ance pathway oflarge soluble immune complexes from the

circulation in patients withSLEisabnormal and,in particular,

that hepatic processing of immune complexes is impaired. There was reduced uptake of immune complexes in the spleen. Theseabnormalities may play a role in the

immunopathogene-sis ofthedisease byallowing immune complex deposition in

tissues outsidethe MPS and, speculatively, by promoting ab-normal pathways of antigen presentation.

Acknowledgments

Wearegrateful to Ms. B. Henderson in Nuclear Medicine at Ham-mersmith Hospital for her assistance with the scanning, and Dr. Jurg Schifferli and Professor P. J. Lavender for their helpful advice.

The work described in this study was funded by the Arthritis and

RheumatismCouncil of GreatBritain.

References

1.Lachmann,P.J.,andM.J.Walport.1987.Deficiencyofthe effector

mecha-nisms of the immune response andautoimmunity.InAutoimmunityand

Au-toimmuneDisease,Ciba FoundationSymposium129. J.Whelan,editor.Wiley

Ltd.,Chichester,England. 149-171.

2.Haakenstad,A.O., andM.Mannik. 1974. Saturation of the

reticuloen-dothelialsystem with soluble immunecomplexes.J.Immunol.112:1939-1947.

3.Hamburger,M. I.,T. J.Lawley,R. P.Kimberly,P. H.Plotz,and M. M.

Frank. 1982. A serialstudy ofsplenicreticuloendothelial systemFc-receptor

function insystemic lupus erythematosus. ArthritisRheum.25:48.

4.Cornacoff,J.B.,L.A.Hebert,W. L.Smead,M. E.VanAman,D.J.

Bir-mingham, and F. J. Waxman. 1983. Primateerythrocyte-immune complex-clearingmechanism. J.Clin.Invest.71:236-247.

5. Lobatto, S., M. R.Daha, F. C.Breedveld, E. K.Pauwels,J.H. Evers

Schouten,A. A.Voetman,A.Cats,andL. A.vanEs.1988. Abnormal clearance of soluble aggregatesofhumanimmunoglobulinG inpatientswithsystemic

lupuserythematosus.Clin. Exp.Immunol.72:55-59.

6.Lobatto,S.,M. R.Daha,A.A.Voetman,J.H.Evers-Schouten,A. A. Van

Es,E. K.J.Pauwels,andL. A.vanEs. 1987. Clearance of soluble aggregates of

immunoglobulinG inhealthyvolunteers andchimpanzees.Clin.Exp.Immunol.

68:133-141.

7.Schifferli,J.A.,Y.C.Ng,J. P.Paccaud,and M. J.Walport.1989. The role ofhypocomplementemiaand lowerythrocyte complementreceptortype I

num-bers indeterminingabnormal immunecomplexclearanceinhumans. Clin.Exp.

Immunol. 75:329-335.

8.Schifferli, J.A.,Y.C.Ng, J. Estreicher,and M. J. Walport. 1988. The clearance oftetanustoxoid/anti-tetanustoxoid immunecomplexesfrom the

circulation ofhumans. J. Immunol. 140:899-908.

9.Madi, N.,G. Steiger,J.Estreicher,and J.A.Schifferli. 1991.Immune

adherence and clearanceofhepatitisBsurfaceAg/Abcomplexesis abnormal in

patientswithsystemiclupuserythematosus.Clin.Exp.Immunol. 85:373-378. 10.Halma,C.,F.C.Breedveld,M. R.Daha,D.Blok,J. H.Evers-Schouten,J.

Hermans, E. K.J.Pauwels, and L. A.van Es. 1991. Elimination of soluble 1231-labeled aggregates ofIgGinpatientswithsystemic lupuserythematosus.

ArthritisRheum. 34:442-452.

11.Madi, N.,J.-P.Paccaud,G.Steiger,and J.A.Schifferli. 1989. Immune adherenceofnascenthepatitisBsurfaceantigen-antibodycomplexesinvivo in humans. Clin.Exp. Immunol.78:201-206.

12. Reay, P. 1982. Use ofN-bromosuccinamidefor the iodinationofproteins

forradioimmunoassay.Ann.Clin. Biochem. 19:129-133.

13.Johns, P.,and D. R. Stanworth. 1975.Asimplenumerical methodforthe

constructionof isokineticsucrosedensity gradients,andtheirapplicationtothe

characterisation ofimmunoglobulin complexes.J. Immunol.Methods. 10:231-239.

14.Weir,D.M. 1986. HandbookofExperimentalImmunology. Blackwell

ScientificPublications, Oxford, England.

15.Walport,M.J.,G. D.Ross,C. MackworthYoung,J. V.Watson,N.Hogg,

andP.J. Lachmann. 1985.Familystudiesoferythrocyte complementreceptor type 1 levels: reduced levels inpatientswith SLEareacquired,notinherited. Clin.

Exp.Immunol. 59:547-554.

16.Hogg, N.,G. D.Ross,D. B.Jones,M.Slusarenko,M. J.Walport,and P. J. Lachmann.1984. Identificationofananti-monocytemonoclonalantibodythat isspecificfor membranecomplementreceptor typeone(CR1 ).Eur. J.Immunol. 14:236-243.

17.Fraker,P.J.,and J. C.Speck. 1978. Protein and cell membrane iodin-ations withasparinglysolublechloroamide,

1,3,4,6,-tetrachloro-3a,6a-diphenyl-glycoluril. Biochem.Biophys.Res.Commun. 80:849-857.

18.Myers,M.J.,J. P.Lavender,J. B.deOlivera,andA.Maseri. 1981.A

simplifiedmethodofquantitatingorganuptake usingagammacamera.Br. J.

Radiol.54:1062-1067.

19.Davies,K.A.,S.Stewart,V.Hird,G.Sivalopenko,A.E.Epenetos,and M. J.Walport. 1990. Invivo formation and clearance of immune complexes in

humans.J. Immunol. 144:4613-4620.

20.Miller,G.W.,and V.Nussenzweig.1975.Anewcomplement function: solubilization ofantigen-antibody aggregates. Proc. Natl. Acad. Sci. USA. 72:418-423.

21.Schifferli,J.A.,P.Woo,andD. K.Peters. 1982.Complement-mediated

inhibition of immuneprecipitation.Roleof the classical and alternative

path-ways. Clin. Exp.Immunol.47:555-562.

22.Hong, K.,Y.Takata,K.Sayama,H.Konozo,J.Takeda,Y.Nakata,T.

Kinoshita,andK.Inoue. 1984. Inhibitionof immuneprecipitationby

comple-ment.J. Immunol. 133:1464-1469.

23.lida,K.,R.Mornaghi,andV.Nussenzweig.1982.Complement receptor

(CR1) deficiencyinerythrocytesfrompatientswithsystemiclupus

erythemato-sus.J.Exp. Med. 155:1427-1438.

systemic lupus erythematosus detected by immune adherencehemagglutination

andradioimmunoassayswith monoclonalantibody.Arthritis Rheum. 27:1329-1335.

25. Holme, E., A. Fyfe, A. Zoma, J. Veitch, J. Hunter, and K. Whaley. 1986. Decreased C3breceptors(CR1)onerythrocytes from patientswith systemic

lupuserythematosus.Clin.Exp.Immunol.63:41-48.

26. Finbloom, D. S., and P. H. Plotz. 1979.Studies of reticuloendothelial

function in the mouse with modelimmunecomplexes. II. Serum clearance, tissue uptake, and reticuloendothelial saturation in NZB/W mice. J. Immunol. 123:1600-1603.

27. Magilavy, D. B., A. Rifai,and P. H. Plotz. 1981. Anabnormality of

immune complexkineticsinmurine lupus. J. Immunol. 126:770-774. 28. Kimberly, R. P., J. C. Edberg, L. T. Merriam, S. B. Clarkson, J. C. Unke-less, and R. P. Taylor. 1989. In vivo handling of soluble complement fixing Ab/dsDNA immune complexes in chimpanzees. J. Clin. Invest. 84:962-970.

29. Ross, G. D., and M. E. Medof. 1985. Membrane complement receptors

specific forboundfragments ofC3. Adv. Immunol. 37:217-267.

30.Smedsr0d,B., H.Pertoft,G. Eggertsen, and C. Sundstrom. 1985. Func-tional and morphological characterization of cultures ofKupifercells and liver endothelial cells prepared by means ofdensity separation in Percoll, and selective substrate adherence.CellTissue Res. 241:639-649.

31. Unkeless, J. C. 1991. Function and heterogeneity of human Fc receptors

forimmunoglobulinG. J.Clin.Invest. 83:355-361.

32. Myones, B. L., J. G. Dalzell, N. Hogg, and G. D. Ross. 1988. Neutrophil andmonocyte cell surfacep150,95 has iC3b-receptor (CR4) activity resembling CR3. J.Clin.Invest.82:640-651.

33. Hinglais, N., M. D. Kazatchkine, C. Mandet, M. D. Appay, and J. Bariety. 1989. Human liver Kupffer cells expressCR1, CR3, and CR4 complement recep-torantigens.Animmunohistochemicalstudy.Lab. Invest. 61:509-514.

34. Frank, M. M., T. J. Lawley, M.I.Hamburger, and E. J. Brown. 1983. Immunoglobulin G Fc receptor-mediated clearance in autoimmune diseases. Ann.Intern.Med.98:206-218.

35.Kimberly, R. P., and P. Ralph. 1983.Endocytosis bythemononuclear

phagocytesystem andautoimmune disease.Am.J.Med. 74:481-493.

36. Brown, D.L.,P. J.Lachmann, andJ. V. Dacie. 1970. The in vivo behavior

ofcomplement-coatedredcells:studies inC6-deficient, C3-depleted,and normal rabbits.Clin.Exp.Immunol.7:401-422.

37.Atkinson, J. P., and M. M. Frank. 1974.Studiesonthe invivo effects of

antibody.Interaction ofIgMantibodyandcomplementin the immune clearance and destruction of erythrocytes in man. J.Clin.Invest.54:339-348.

38. Frank, M. M., M. I. Hamburger, T. J.Lawley,R.P.Kimberly,and P. H. Plotz. 1979. Defective reticuloendothelial system Fc-receptor function in sys-temiclupus erythematosus. N. Engl. J.Med. 300:518-523.

39.Walport, M. J., A. M. Peters, K. B. Elkon, C. Pusey, J. P. Lavender, and

G.R. V.Hughes.1985. Thesplenic extractionratio ofantibody-coated

erythro-cytes anditsresponse toplasma exchange andpulse methylprednisolone. Clin.

Exp.Immunol. 60:465-473.

40.VanDerWoude, F. J., M. Van DerGiessen,C. G. M.Kallenberg, W. Ouwehand, H.Beekhuis,J. M. Beelen, W. J. VanSon,P.H. J.Hoedemaker,

G. K. Van Der Hem, and T. H. The.1984. ReticuloendothelialFc receptor

func-tion in SLEpatients.I.PrimaryHLAlinkeddefectoracquired dysfunction sec-ondary todisease activity?Clin.Exp.Immunol. 55:473-480.

41.Weiss,L. 1988. TheSpleen, Cell andTissue Biology.L.Weiss,editor. UrbanandSchwarzenberg, Baltimore,MD. 517 pp.

42. Pepys, M. B. 1974. Roleofcomplement ininduction ofantibody produc-tion invivo. Effect ofcobra venomfactorand otherC3-reactiveagents on

thy-mus-dependent and thymus-independent antibody responses. J. Exp. Med.

140:126-145.

43.Pryjma,J., and J. H. Humphrey. 1975. Prolonged C3depletion bycobra venomfactor in thymus-deprived mice and its implication forthe roleofC3 as an