Formation of soluble immune complexes by complement in sera of patients with various hypocomplementemic states Difference between inhibition of immune precipitation and solubilization

Formation of soluble immune complexes by

complement in sera of patients with various

hypocomplementemic states. Difference

between inhibition of immune precipitation and

solubilization.

J A Schifferli, … , P J Spaeth, A G Sjöholm

J Clin Invest.

1985;

76(6)

:2127-2133.

https://doi.org/10.1172/JCI112217

.

To examine whether the ability of complement to form soluble immune complexes plays a

role in preventing immune complex-mediated diseases, we analyzed the capacity of

complement to inhibit immune precipitation (IIP) and to solubilize preformed immune

aggregates (SOL) in 23 sera of patients with various hypocomplementemic states, and we

correlated the results of these studies with the clinical syndromes found in the various

patients. In sera with deficiency or depletion of early classical pathway components, IIP was

profoundly altered, whereas SOL was delayed but in the normal range. In contrast, in sera

with C3 depletion but intact classical pathway and in properdin-deficient serum, IIP was

initially preserved, whereas SOL was abolished. Since the incidence of immune complex

diseases in various hypocomplementemic states correlates with the severity of IIP defects,

but not with reduced SOL, it is suggested that IIP is an essential biological function of

complement that prevents the rapid formation of insoluble immune complexes in vivo.

Research Article

Find the latest version:

Formation

of

Soluble

Immune

Complexes

by

Complement

in

Sera

of Patients

with Various Hypocomplementemic

States

Difference between Inhibition of Immune Precipitation and Solubilization

Jurg A. Schifferli, GertraudSteiger, GeorgesHauptmann, PeterJ.Spaeth, and Anders G. Sjoholm

Clinique Medicale,DepartmentofMedicine, HopitalCantonalUniversitaire, Geneva,Switzerland;Institutd'HematologieetCentrede

TransfusionSanguine, Strasbourg, France;CentralLaboratory of theSwissRed Cross Blood TransfusionService,Bern, Switzerland; and

Department ofImmunology, InstituteofMedical Microbiology, UniversityofLund, Sweden

Abstract

To examine whether the ability of complement to form soluble immune complexes plays a role in preventing immune complex-mediated diseases, we analyzed the capacity of complement to inhibit immune precipitation (1IP)and tosolubilize preformed immune aggregates (SOL) in 23 sera of patients with various hypocomplementemic states, and we correlated the results of

these studies with the clinical syndromes found in the various patients.

Insera with deficiency or depletion of early classical pathway components, IIP was profoundly altered, whereas SOL was layed but in the normal range. In contrast, in sera with C3 de-pletion but intact classical pathway and in properdin-deficient serum,IIPwasinitially preserved, whereas SOL was abolished. Since the incidence of immune complex diseases in various hypocomplementemic states correlates with the severity of IIP defects, but not with reduced SOL, it is suggested that UIP is an essential biological function of complement that prevents the rapid formation of insoluble immune complexes in vivo.

Introduction

Theassociation between complementdeficiency and immune complex diseases has become evidentoverrecentyears(1). It is therefore possible that complement has aphysiological role in preventing tissue formationordeposition of immune complexes.

Invitro,complement can bring about the formationofsoluble complexes either by inhibitingtheiraggregation atthe time of theantigen-antibody reactionorby solubilizing preformed im-mune precipitates (2, 3). Inhibition of immune precipitation (IIP)' relies predominantly on an intact classical pathway of complement activation,whereas solubilization(SOL)is

depen-dent on the presence of normal alternative pathway function

Dr.Schifferli istherecipient ofaMaxCloettaCareerDevelopmentAward.

Addresscorrespondence to Dr. Schifferli, Clinique MWdicale, H6pital

CantonalUniversitaire, 1211Geneva4, Switzerland.

Receivedforpublication 20April1984andinrevisedform16August 1985.

1.Abbreviations usedinthis paper: AP, alternative pathway;Cl-In,Cl

inactivator; CFD,complement-fixing diluent;

deffs),

deficiency;HSA, human serumalbumin;UIP,inhibition of immuneprecipitation; NHS, normal human serum; P,properdin; SDGU, sucrosedensitygradient

ultracentrifugation; SOL, solubilization.

(4-6). Both processes result in the formation of soluble complexes by the covalent binding of C3 fragments (7, 8).

Assayshave been devised to measure SOL and1IP in vitro in sera of patients with various immune complex-mediated dis-eases(9-22).Abnormalities have been found in different diseases, butnoattempts have been made to compare these two functions indefined situations in order to determine their biological rel-evance.

Inthisstudy, UIP and SOL were analyzed and compared in various hypocomplementemic sera. The results indicated that alterations in TIP and SOL differ depending on the type of hy-pocomplementemia. Inparticular, assays for 1IP revealed rapid andimmediateaggregationof complexes in sera with early clas-sical pathway defects, suggesting that such hypocomplementemic

states are afavorableground for the formation of insoluble com-plexes.

Methods

Immune

complex

model

A '25l-bovineserumalbumin (BSA;ArmourLtd., Eastbourne,United

Kingdom)-rabbitanti-BSAantibody modelwasusedforbothSOLand IIP aspreviously described (3, 6).Briefly,theIgGfraction ofrabbit BSA antiserumwasprepared by Na2SO4 saltprecipitation.Antibodies cross-reacting with human serum albumin(HSA)wereremovedby immu-noadsorption on insolubilized HSA (Sigma ChemicalCo., St. Louis,

MO).Thereafter,therabbit anti-BSAantibodydidnotreactwith

chlo-ramineT '251-radiolabeledHSA(23), i.e.,didnotform solubleor

pre-cipitatingimmunecomplexesasdeterminedbysucrosedensity gradient ultracentrifugation (SDGU).

(a) IIP assay.50mloftestserum,0.5 gg

'251I-BSA,

and whenrequired,

theappropriateadditionalreagent,weremixed andadjustedto afinal volumeof 120

,l

withbuffer(complement-fixingdiluent(CFD): barbital-buffered saline,pH7.4,conductivity14mmHo;CFD,OxoidLtd.,

Bas-ingstoke, UnitedKingdom).Afterapreincubationof5 minat_37°C,9

zg of rabbit anti-BSAantibodywasadded and the incubation carried outat37°C. 25-mlaliquotswereremoved after various timeintervals,

mixed withI mlphosphate-bufferedsaline(PBS) (pH 7.4,conductivity 14mmHo;PBS, OxoidLtd.)andcentrifugedat3,000gfor 10minat

4°C. Supernatants andpelletswereseparatedandtheradioactivity

mea-sured.

(b) SOL assay. Immuneprecipitationof'251I-BSA-rabbitanti-BSA antibodywascarriedout for 1 hat_37°Cand 4 hat4°Cin CFD. By thattime, 95%ofthecomplexes hadprecipitated whencentrifugedat 3,000gfor 10min. Theseimmuneaggregateswereuseddirectlywithout

centrifugationinthe assayat

A/5

oftheconcentration used inIIP, i.e.,

0.1

Mg

'25I-BSA plus1.8 anti-BSAantibody.Themixturescontaining50 Ml of test serum and the immuneaggregateswere adjusted to afinal volumeof120MlwithCFD. The assaywascarriedoutsimilarlytoIIP.

Theimmunecomplexloadwashalved inIIP and SOL

experiments

usingproperdin (P)andC4-deficientsera.The variousserawerestored at-70°Cbeforeperformingtheassaysexcept for the fiveseracontaining FormationofSolubleImmuneComplexes by Complement 2127

J.Clin. Invest.

©TheAmericanSociety for Clinical Investigation,Inc.

0021-9738/85/12/2127/07 $1.00

cryoglobulins.These sera werestudied fresh: bloodwasallowedtoclot for 15 minat370C;theserumseparated by centrifugationat370Cwas immediately used in the assays. 10 mM etylenediamine tetraacetate

(EDTA) was usedtoblock allcomplementfunction,and 10 mMetylene

glycol tetraacetate (EGTA) plus 2 mM Mg was used to block classical pathway function. All assaysweredone induplicate.

SDGU

50mlofthemixtures(IIP and SOLassays)wereincubated for 2 h and overlayeredonto a 10-50%(wt/wt)sucrosegradientin PBS and ultra-centrifugedat 80,000 g for 12.5h at40C. Thepolystyrenecentrifuge

tubes and the 24 fractions collectedwerecounted and the percentage of thetotalradioactivity determined for each fraction. RadiolabeledIgM

servedas a19-S(Svedberg unit) marker.

Purified complement proteins

Clq and C4 were purified according topublished methods (24, 25).

Properdinwaspurified asdescribed by Medicuset al.(26).Thepurity

of these threeproteinswasshownbysodiumdodecyl

sulfate-polyacryl-amide gelelectrophoresis(SDS-PAGE)in the presence and in the absence ofreducing agents (27). The P concentrationwas measured bysingle

radial immunodiffusionusinga monospecificantiserum(28), and the result expressed in percentage ofanormal humanserumpoolobtained from 25 donors. The Clq and C4 concentrations and functional activities were determined, respectively, by the method of Lowry (29)usingBSA

asastandard, andbyhemolyticfunctional assays(30,31).Thespecific

activitywas 100% forClq and75% for C4 when comparedwithClq

and C4 of theplasma used for theirpurification.

Complement,

immune

complex,

and

cryoglobulin measurements

Allsera were storedat-70°C until tested, except the fiveseracontaining cryoglobulins,whichweretested fresh. Cl inactivator(Cl-In), C4, C3,

andC2weremeasuredbysingleradial immunodiffusionusing

mono-specificantisera(Miles Laboratories,Inc., Elkhart,IN) accordingtothe manufacturer's notice. Clq was measured accordingtoZiccardi and Cooper(32) usinganti-Clqantiserum raised inagoatbythreeinjections

at2-wkinterval ofpurifiedClqinFreund'sadjuvant(Difco Laboratories, Detroit,MI)(complete adjuvantfor thefirstinjection, incomplete ad-juvant for thetwo followinginjections). C4-binding-protein measure-mentshavebeendescribed (33).CH50,factor B, and alternativepathway (AP)activityweredetermined byhemolyticplate assays(34).Functional C4wasmeasured(31) andcomparedwithantigenicconcentrations. All resultswereexpressed in percentage ofanormalhumanserum(NHS) pool collected from 25 blood donors. Clq bindingassays for immune complexeswerecarriedout asdescribed by Zubleretal.(35).

Cryoglob-ulinsweredetectedby incubation ofserum for 96 h at4°Candtheir

contentthenanalyzed (36).

Results

IIPandSOL assays were standardized. Using the'25I-BSA-rabbit anti-BSA model described, normal serum inhibited immune precipitation completely over the period of time studied (2 h)-i.e., >94% soluble complexes in all 25 normal sera tested (Fig.

1 A). The immune complex concentrations used in the assay corresponded to half the maximum concentration that could be maintained in solution in NHS; thus, any decrease in IIPwas abnormal. When10mM EDTA was added to block complement function,

TIP

wasabolished and complexes aggregated rapidly so thatprecipitationwas nearly complete in 30 min.

Normal serumsolubilized55% (±25%) preformed immune aggregates in2 h(normal means±2 SDof25 normal sera) (Fig.

1 B).SOL was abolished in the presence of 10 mM EDTA. The formation ofsoluble immune complexes by TIP and

SOLwasstudiedin varioushypocomplementemic sera(Table I). These sera weregrouped accordingtotheir maincomplement defect: (a) deficiencies (def)of early classical pathway compo-nents ( 15 sera); (b) depletion of C3 andproperdindef (5 sera); and (c) def of one of the membrane attack complex pro-teins (3 sera). Immune complexes were measured in all sera, sincetheirpresence could interfere with normal complement function. Clqbinding material wasfound in the fivepatients withcryoglobulinemia(>50%binding),and inClqand C4 de-ficient sera (84 and 12%binding, respectively) (normal: <5%).

IIP andSOL assays in sera with depletion or def of early classicalpathwaycomponents.The

major

observationwas a

se-vere defect in IIP. Initialfastprecipitationwasevident inmost sera, andthiswasfollowed by

partial

redissolutionof the

com-plexes (Fig. 1). Asimilar kineticwasobserved in normal serum chelated with 10 mM EGTA plus 2 mM Mg, indicating that immediate aggregation wasrelated to the absence of classical pathwayfunction. However, IIPremained normalintwo sera:

these twoexceptionsweretheserawith Cl inactivatordeficiency butwithoutsevereclassical pathway depletion.

ThatSOL was possible in these sera was indicated by the partial redissolution of aggregated complexes in the UIP assay, andthiswasconfirmedin theSOLassay: whereasSOLwasin the normal rangeforall 15 sera at2 h,inmost casesthisSOL was preceded by a prolonged lag phasecompared withnormal

serum.

Fig.

1 Bshows the resultsobtainedwith thefiveseraof

patients

with

acquired

depletion

ofclassical

pathway

components (paraproteinemia andbenign tumor). The foursera in which SOLwas notdelayedwere:thetwoCl inactivator-deficientsera

withoutclassical pathwaydepletion, Clq-deficientserum, and

oneserumcontaining cryoglobulins (patient 12).

Toestablish that these defects inUIPand SOL were related

toabsentclassical pathway activity,thetwoassayswererepeated inClqandC4-deficientserarepleted withthemissing compo-nent. UIP andSOL were restored byrepleting C4-deficientserum withpurified C4 (Fig. 2). In separate experiments it was possible

toestablishthatonly 1%ofC4(3 ,ug/ml serum)wasrequiredto

normalize IIP. Inadditionin the IIP assay, earlyprecipitation of complexes wasfaster inthis serum after chelation with 10 mM EGTA plus 2 mM Mg orwith EDTA. Sincethe macro-molecular

Cl

complex delaysimmune aggregation,this obser-vationcould be due to

Cl

dissociationafter chelationofcalcium (37). Furthermore,initialaggregationwasfasterin Clqdeficient than inC4-deficientserum, and therewas nodifferencein the initialaggregation in Clq-deficientserumwhether calcium was chelatedornot(EGTA-MgorEDTA). InClq-deficientserum and in anotherserumwhereClqwasdiminishedpredominantly (patient 6),UIPwasrestoredbypurified Clq: 15

gg/ml

serum (20%of normal) was necessary to normalize IIP in Clq-deficient serum. The immune complexes detected by the Clq binding assay in both sera seem notto have interfered with repletion experiments.

In sera containing cryoglobulins, defective IIP could also have been due to thecryoglobulins rather than hypocomple-mentemia.Weobtained again freshserumof patient 11 (mixed

cryoglobulinemia

IgMk-IgG) in whichhemolyticC4 was 1%of normal,whereas othercomplementproteinswereless depleted. The IIP assay wasperformedin the presenceof absence of pu-rifiedC4(100and200% ofhemolytic C4). Inthis serum, kept

at37°C,

purified

C4wasnotconsumed afteranincubation of

15 min in a control experiment without immunecomplexes, whereas 80% was consumed when complexes were added. Thus

Table I.HyvpocomplementemicSera

Functional complement assays(% NHP)*

Concentrations of reduced complement proteins

Patients and diagnosis CH5O AP5O* (% NHP)§

Earlyclassicalpathway component defects

1 C I qdeficient/SLE-likedisease 2 C4deficient/SLE-likedisease

Acquireddepletion of classical pathway

components/associatedwith: 3 Uterinemyoma andangioedema

4 Paraproteinemia

5 Paraproteinemia 6 Paraproteinemia

7 Paraproteinemia Cryoglobulinemia:

8 TypeIIgG lambda/vasculitis-arthritis

9 Type II IgAlambda-lgG/vasculitis

10 Type II IgMkappa-IgG/vasculitis-nephritis 11 Type II IgMkappa-IgG/vasculitis-nephritis

12 Type IIIIgM-IgG/vasculitis Cl inactivator deficient/angiodema: 13 Treatedwithdanazol

14 Treatedwithstanazol

15 Treatedwith tranexamic acid

C3depletionand Pdeficient

16 Nephritic factor/glomerulonephritis

17 Nephritic factor/glomerulonephritis 18 Nephritic factor/glomerulonephritis 19 Poststreptococcal glomerulonephritis 20 Pdeficient/healthy

Membraneattack complex deficiencies

21 C7deficient/one episode of meningitis (Neisseria meningitis)

22 C7deficient/one episode of meningitis (Neisseria meningitis)

23 C8deficient/recurrentupperrespiratory

tractinfections

o0 oi

0 0 0 0 0

0 0 0 20 50 100 73 0

25 35 25 15 69

0 0 0

85

100

80

100

100 120 100 100

100

85 90

100

100 63 73

0 0 0 0

0 0 0

0Clq 0 C4

7%Clq, 35% Cl-In,<1%C4, 10% C2 10% Clq, 50% C2

15% Clq,11%Cl-In, <1% C4,25% C2

13%Clq, 17%Cl-In,<1%C4,25%C2

24%Clq,21%Cl-In, 5% C4,32% C2

40%oClq,

5%C4

45%Clq,7%C4

35%Clq, 16% C4,46%C2 33% Clq, 5% C4

15%Clq, 10% C4 39%Cl-In

70%Cl-In

15%Cl-In,20%C4**, 29% C2

10%C3 5% C3 10% C3

15%C3 0Properdin

0 C7 0 C7 0C8

*_{NHP,normalhuman serumpool from 25 controls. tAP50}

lytic

_{alternative pathway. § Only values below the normalrange arementioned.}

11Restoredtonormal withrepletion ofthedeficientcomplementcomponent. I Partiallysis. ** But<1% functionalhemolyticC4.

C4function wasrestored, and thisC4 wasactivatedby BSA-anti-BSA complexes.However, IIPwasonly partially improved; afteranincubation of15 min, the percentage of soluble

com-plexes was 96% in NHS, 70% in thecryoglobulinemic serum, and 75 and 76%,respectively, afterrepletionwith thetwo

con-centrations ofC4.SincetheCH50 oftherepletedsera wasonly 50% ofnormal, it is clear that complement function has not

been fully restored byC4. Thus, hypocomplementemia (here C4depletion) is at least in part responsible for detective IIP; however,cryoglobulinsby themselves could haveanadditional detrimental roleonIIP.

IIPandSOL assays in serawithprofoundC3depletion, and

in

P-deficient

serum.SOLwasabolished in all sera. By contrast, UIPwasonly defective: theaggregationof complexes proceeded

at avery reduced rate;

i.e.,

>85%werestillsolubleafter30 min (Fig. 3).P-deficientserumshowed thewidest differencebetween thetwoassays

(Fig.

4): acompletely preservedUIP contrasted

with nosignificantSOLover aperiod of incubation of2h.In

IIP experiments, addition ofEGTAplus Mgsoas toblock clas-sical pathway activity resulted infast and irreversible

precipi-tation;thesecondphase of

partial

dissolutionwasobserved

only

when the P-deficient serum was repleted (100% ofa normal human pool) withpurified P. Inrepleted serum, SOLwas

re-storedaswell andthe presenceofEGTA

plus

Mg only

delayed

the reaction. In separate

experiments,

SOL was shown to be directly

proportional

tothe P

concentration,

thus

contrasting

with the role ofC4 andClq in IIP. In

addition,

there wasno delayin SOL

using

lowPconcentrations

(down

to25% of

nor-mal).Todeterminewhetherby loweringthe immune

complex

load theP-deficientserumwould be

capable

of

producing

some SOL,

'Iho

the usual immune

precipitate

wasused

(0.01 tsg 12511

BSAplus 0.18

gg

anti-BSA

antibody

per

assay):

there wasstill onlya

marginal

8%SOL

compared

with the 5% in the control P-deficientserumincubated with 10 mM EDTA.

0 30 60 120 0 30 60 120

50- -w

30 60

Time (min)

Figure 3.lIP in the four sera of patients with C3 depletion (pa-tients16-19). Patients: 16:

120 0-* ; 17:---*---; 18:

* ;19:---Ow

Time (min)

Figure 1.(A)lIPof'251I-BSA-anti-BSAantibodycomplexes in various

sera.Theinitial reaction mixturecontaining0.5

sg

'251-BSAand50

ml serum wasdilutedto afinal volume of 120MIwith CFDbuffer,

andincubated for5min at370C.9mgof rabbit anti-BSAantibody

wasadded and theincubationwascontinuedat

370C.

Atvarious time intervals,25-MA aliquotswereremoved, mixed with 1mlof cold PBS buffer,andcentrifuged for15min at3,000g.After separation of

pel-letand supernatant, the percentageof soluble complexeswas deter-mined.NHS: (- o -);NHSchelated with 10 mM EDTA

(NHS-EDTA): (---*---).Patients ofgroup Aof thetable:patients 3-7 (uter-inemyoma,paraproteinemia):(-* -) patients 8-12 (cryoglobulin-emia): (--- * ---). (B) SOL of '25I-BSA-anti-BSA aggregatesindifferent sera.Preformedimmuneaggregates(1.9

Mg)

wereincubatedat370C

with50MItest seruminafinal volume of120Ml.SOL±2 SD in nor-malhuman serum areindicated:(-o ). Patients with early classical pathway depletion(patients 3-7):(-* ).

HIP andSOL inmembraneattack complex protein

def lIP

andSOLwerenormal inserawith C7orC8 def.

Relationship betweenimmunecomplexsizeand abnormal-ities in HIPandSOL. To see whether thedifferencesin UIP and SOL obtainedin thevariousseracould be relatedtothesize of thesoluble immune complexes

formed,

the experiments were

repeatedand thesize ofthecomplexes determinedbySDGU afteran incubation of2 h. In normal human serum, UIP and SOL produced complexes of similar

size;

however, since SOL

wasnotcomplete(normal:55%),afraction oftheinitialinsoluble complexes wasrecovered in the pellets ofthe

ultracentrifuged

tubes(Fig.5).

100- NHS NHS

C4def+C

C'idef Cdf

50 \- .. j 4Cdef

C4def+MgEGTA

C4def+EDTA

.-_._{aA_...}.4 _ CC4def+EDTA

0~

0 30 60 120 0 30 60 120

Time(min)

Figure2.UIP(A) and SOL (B) inC4-deficient(def)serum. C4 def:

C4-defserum; +C4: repleted with 300

Mg/ml

serumofC4; +EDTA:

che-latedwithEDTA to block all complement function; +Mg-EGTA:

che-lated with 2mM Mg + 10 mM EGTA to block only classical pathway

activity.

In patients with early classical pathway depletion or def (group a), the size of the soluble complexes formedwasnormal

in both assays. However, in UIP experimentsa fraction of the

complexes was recovered in the pellets. This fraction

corre-spondedwell to the fraction of insolublecomplexesdetermined

by centrifugation at3,000 gin theUIPassay.Afterrepletionof C4 andClqin thecorrespondingdeficientsera,nopelleted in-solublecomplexeswere formed. Thusclassicalpathwaydefects

affected only the proportionofthe soluble immune complexes formed butnot their size.

Incontrast, inthe fourC3-depletedsera notsustaining any SOL (group b), thecomplexes formed duringIIP were larger

than in normal human serum intwo,did not forma

homoge-neous population in one, and aggregatedinone(Fig. 6). The solublecomplexes formedinP-deficientserum werelargerthan those formed in normal serum; however, their size was

nor-malizedbyrepletion withP(Fig.7).Thus,alternativepathway dysfunction abolished the formation ofnormal size soluble

complexes. In membrane attack complex def(group c), the

size of the solublecomplexeswas notmodified.

Discussion

Several lines ofevidence suggestthatcomplement is involved

inthe elimination of immunecomplexes. Firstly, complement-deficient states are associated with various types of immune complex diseases(1).Secondly, complement affects directlythe

structureofimmunecomplexeseitherbybreakingupinsoluble immuneaggregatesorbymaintaining immunecomplexes sol-ubleatthetime of theantigen-antibody reaction (2, 3). Inthis studywecomparedUIP and SOL-thetwocomplement func-tionsthatbringabouttheformation of solublecomplexes-in varioushypocomplementemic states to see whether defects in

A B

1uu-Pdef+P

Pdef

Pdef +P+EGTA-Mg

Pdef+EGTA-Mg Pdef Pdef+

0 30 60 120 0 30 60 120

Time(min)

Figure 4.lIP(A)andSOL (B)inP-deficient(def) serum. 2130 J.A.

Schifferli,

G. Steiger, G. Hauptmann, Spaeth, Sjdholm

y

.!. 50

m :3 0

% cm

15

-Fraction number

Figure 5. SDGUof soluble immune complexes formed byUIPand

SOLafteranincubation of 2 h in NHS. 24 fractions were collected, andresults wereexpressed in percentage of total cpm recovered. Bot-tomofthegradient is on the left. IgM served as a 19-S (Svedberg unit) marker. The percentage of complexes recovered in the pellets were 4%

forUIPand 25%for SOL.

these twofunctionscouldexplain the association of complement deficiencyand immune complexdisease.HIPandSOL were not affected similarly in the various sera tested. Abnormalities in HIP were related to early classical pathway defects, whereas SOL wasabolished in the absence of AP function.

In Clq and in C4 deficiency, HIP was severely defective mainlyin itsinitial phase; however, immune aggregation was followed by partial redissolution of the complexes, suggesting that theSOL process wasintact. Thiswasconfirmedin the SOL assaywhich wasonly preceded byaprolonged lag phase in C4-deficientserum. That defective classicalpathway function was

responsiblefor the incapacity to prevent rapid immune aggre-gation was demonstrated byrepletingthetwodeficientserawith theirrespective missingcomponent.

It has been shown that C4 fragments, despite binding to complexes, have littleor no capacity to delay immune aggre-gation (37, 38). The findingthat only 1%hemolytical C4 was sufficientto restoreIIP in theC4-deficient serum suggests that themain role of C4 istoprovideaninitial C3 convertase attached

tothecomplex so as toobtain immediateAPactivation. The relative roleof C4 in HIP was also evidentin the three patients

1 5 10 15 20 25 Fraction number

Figure 6. SDGU of complexes formed during IIP after an incubation

of2 hin thefourserawith C3depletion(patients 16-19).The per-centageof complexes recoveredin thepelletswere:NHS:0.5%; pa-tient 16: 34% (-); 17: 68% (-- -_{-); 18: 29%(-}-_{-);19: 11%}

(----). s,Svedberg unit.

Fraction number

Figure 7. SDGU of complexes formed in P-deficient serum repleted ornotrepleted with P. The percentage of complexes recovered in thepelletswere: Pdeficient: 13% (- * -); P deficient + P: 8%

(-- A-v- -); normal serum: 8% (- o -). s, Svedberg unit.

withCl inactivator deficiency, since only the patient with pro-found C4 depletion presented an initial defect in IlP.

Macromolecular CIhas adirect inhibitory effect on immune aggregation at the time of the antigen-antibody reaction, and this effect is proportional to itsconcentration(37). Thus, Clq incorporated in the Cl complex does not only initiate classical pathway activation but directly prevents the immediate aggre-gation reaction. ThisCl effectwas probably present in C4-de-ficient serum,sinceinitial immuneaggregationwasnot asfast

asinClq-deficientserum.

InP-deficientserum, 1IP andSOL were affected very differ-ently. The essential roleofProperdinin SOL is wellestablished; however, we wanted to know if a serum deficient in only this component, and in the face ofareduced immunecomplex load, whether some SOL would be possible. The results were clear: no significant SOL occurred in any of the experiments done usingourimmunecomplex model. ThisSOLwasrestored by purifiedPandwasdirectly proportionalto Pconcentration. The absentSOL contrasted with thecapacityof thisserum toprevent immune precipitation,confirmingresultsobtained with a serum depletedofProperdin (6). Complementmediated UIP, likeSOL, didnotrequiretheassembly of the membrane attackcomplex (C5b to C9), since IIP and SOL were normal in C7 and C8 deficientsera.

In seraofpatients with complement-depleted states, alter-ationsin UIP and SOL could bedirectlyrelatedtomissing

com-plement functionorproteins:early classicalpathwaydepletion correlatedwith initial defects in UIP and C3 depletionwith ab-olition of SOL.Inthelater group, UIPwas not asmuchpreserved

asin Pdeficiency: aggregationofcomplexescouldnotbe pre-vented completely, but did proceedat a very reduced rate. A

similarslowaggregation has beenreported previouslyin serum

depleted ofC3 with cobra venom factor and in C3-deficient

serum(3). Thus,

despite

the absence of C3 therewasstilla

con-siderable delay in immuneaggregationinsuch sera,

contrasting

with the rapid aggregation observed in early classical pathway componentdeficienciesor

depletion.

Polyclonal and monoclonal IgM rheumatoid factors have been reportedto block normal IIP

(21, 22);

it is

possible

that

this effect waspresent in some

patients

with classical

pathway

depletion and

cryoglobulinemia

typesII andIII.

However,

hy-pocomplementemia

wasinvolved in the defective I1Pas

_well,

since

repletion

of C4 inonesuchserum restored IIP in part.

Immuneeliminationofcomplexesinvivohas been shown

to be proportional totheirsize, large complexes disappearing first (39). To see whether the solublecomplexes formed byHIP andSOL invarioussera wereof similarsize,weanalyzedthem by SDGU.Innormal serum,HIPandSOLproduced largesoluble complexes (>19 S) of identical size. In pathological sera, the proportion of soluble complexesbecomingorremainingsoluble wasreduced, asalready shown in the HIP andSOLassays; how-ever, the size ofthe soluble complexes formed remained

un-changedaslongasalternative pathwayfunctionwaspreserved. OnlyserawithC3

depletion

and P-deficient serum, those which didnotsustain anySOL, producedsolublecomplexesoflarger sizeduringIIP.These results suggest that under normal

circum-stancesall solublecomplexes formedeither

by

HIPand

by

SOL have beenprocessed by the AP.Thus, SOL of

preformed

aggre-gatesproducesonlyonetypeof complexes;

i.e.,

APprocessed.

Incontrast, at thetimeoftheantigen-antibody reaction,

com-plement canmaintaincomplexes solublebydifferentsuccessive mechanisms:(a) Cl itselfprotectscomplexesfrom

aggregation

forashortperiodoftime (6);(b)classicalpathwayactivation is capableofproducing large complexes ofreduced

precipitability;

and (c) APaction(including P) reduces the size of such

com-plexes. No invivo experimenthave been done yetto seewhether such differences in size of large complement reacted soluble complexes (>19S) influence their elimination.

Invivo it islikelythatcomplementiscontinuously involved in the formation ofsolublecomplexes sothat thereaction

be-tweenantigen andantibodydoes notleadtothe formation of insoluble material.Prevention of immune aggregation could al-lowcomplexesformed inthe extravascular spacetodiffuse into thelymphaticandbloodvessels,and thoseformedin the blood

stream not todepositin tissues. Furthermore, ithas been sug-gestedthatC3

binding

tothecomplexes is of

major importance

for thesafe disposal oflarge soluble complexes(40). Indeed,in primates,whoseerythrocyteshave C3b receptors, suchcomplexes have been showntobedeliveredtothefixed macrophagesystem

onthesurface oferythrocytes.

Clinical indications on the relevance of complement-me-diated lIPcomesfromthe strongassociationbetweendifferent early classical pathway component deficiencies and immune complex diseases (1).TheClq-andC4-deficient patients

inves-tigated

hereareno exceptions; mostsuchpatients suffer from SLE-like diseases. Interestingly, C3 deficiency isless often

as-sociatedwith immune complex disease than are Clqand C4 deficiencies.Thepartial delayinimmuneaggregationcould still beprotectiveto someextent.Of particular interest isthe obser-vation that

lIP

is preserved in P-deficient serum: none of the

sevenpatients describedtodate hasanimmunecomplexdisease (29, andSjoholm,A., and P.Spaeth,unpublished observation). Thiscastssomedoubts on thepossible protective role of SOL, since SOL is abolished in the absence of P. It is therefore likely that in many situations, normal classical pathway function is essential andsufficienttoprotect theorganismagainst the harm-fuleffectsof immune complex formation and deposition in var-ioustissues.

Theconsequences ofacquired defectsof complement com-ponentsarelikelyto besimilar to inheriteddeficiencies, since

comparable

defects in

lIP

are observed. Nephritic factorand profoundC3depletion areassociatedwith membranoprolifer-ative

glomerulonephritis

type II; however, such

patients

have

no vasculitis. Furthermore, deposits in the glomerulicontain

mostoften C3 fragmentsbutlittleor noimmunoglobulin(41). Thus, there isnoevidencethatthisdiseaseis mediated

by

im-mune

complex deposition.

Incontrast,

generalized

vasculitis and arthritisare

prominent

in

cryoglobulinemia

and

suggest

adisease process similartoacuteserumsickness in which vasculitic lesions areduetoimmune

complex deposition

from the bloodstream.

Thus,

it could well be that in many

pathological

situations

(cryoglobulinemia,

SLE,

bacterial

endocarditis),

thedepletion oftheclassical

_pathway

of

complement

contributestothe for-mationof insoluble

_complexes.

Acknowledaments

WethankDr. L.Perrin forperformingtheClq bindingtests.Weare indebtedtoMrs.C.Todisco fortypingthemanuscript.

The workwassupported bygrantnumber3.896-0.83from the Fonds National Suisse delaRechercheScientifique.

References

1.Ross, S.C., andP. Densen. 1984.Complementdeficiencystates andinfection:epidemiology,pathogenesisandconsequencesofNeisserial and otherinfections inanimmunedeficiency.Medicine(Balt.). 63:243-273.

2. Miller, G. W., andV. Nussenzweig. 1975. A newcomplement

function:solubilization ofantibody-antigenaggregates.Proc.Nat!. Acad. Sci. USA. 72:418-422.

3.Schifferli,J.A., S.R.Bartolotti,andD. K.Peters.1980. Inhibition of immuneprecipitation by complement. Clin.Exp.Immunol. 42:387-394.

4.Takahashi, M., S.Takahashi, V. Brade, andV.Nussenzweig. 1978. Requirements for the solubilization of immuneaggregateby complement.

J.Clin.Invest.62:349-358.

5. Fujita, T., Y. Takata, and N. Tamura. 1981. Solubilization of immuneprecipitates by the 6 isolated alternative pathwayproteins.J. Exp. Med. 154:1743-1751.

6. Schifferli,J. A., P.Woo, andD. K. Peters. 1982. Complement mediated inhibitionof immuneprecipitation.I.Roleof theclassical and alternative pathways. Clin. Exp. Immunol. 47:563-569.

7.Takata, Y., N. Tamura,and T. Fujita. 1984. Interaction of C3 withantigen-antibody complexes inthe processofsolubilization of im-muneprecipitates.J. Immunol. 132:2531-2537.

8. Hong, K.,Y.Takata,K.Sayama,H.Kozono, J. Takeda,Y. Nak-ano, T.Kinoshita,andK. Inoue. 1984. Inhibition of immune precipi-tationbycomplement.J.Immunol. 133:1464-1470.

9. Bartolotti, S. R., B. Pussell,A. Dash, and D. K. Peters. 1979. Complex dissolution:anassayfor complementfunction and immune complex load.KidneyInt. 16:92.(Abstr.)

10.Naylor,J.F., S.A.Ward, S.E.Moore, and J.D.Smiley. 1979. Decreased complement solubilization of immune complexes in sera

containing hightiters of rheumatoid factor. Arthritis Rheum. 22:642. (Abstr.)

11.Zeitz,H.J., G.W.Miller,T.F.Lint,M. A.Ali, andH.Gewurz. 1981.Deficiency of C7 with systemic lupus erythematosus.Solubilization

ofimmune complexesincomplement-deficientsera.ArthritisRheum. 24:87-93.

12.Aguado,M.T.,L. H.Perrin,P. A.Miescher,and P. H. Lambert.

1981. Decreasedcapacitytosolubilizeimmune complexesin serafrom

patientswithsystemiclupus erythematosus. Arthritis Rheum. 24:1225-1229.

13.Schifferli,J. A.,S.M.Morris,A.Dash,and D. K. Peters. 1981.

Complement mediated solubilization in patients with systemiclupus

erythematosus,nephritisorvasculitis. Clin. Exp. Immunol. 46:557-564.

14.Sakurai, T.,T.Fujita,I.Kono, T. Kabashima,K.Yamane, N.

Tamura,and H.Kashiwagi.1982.Complement mediatedsolubilization

ofimmune complexesinsystemic lupus erythematosus.Clin.Exp.

Im-munol. 48:37-42.

and the presence of incompletely solubilized immune complexes in SLE sera.Clin. Exp.Immunol. 54:439-447.

16. Baatrup, G., I. Petersen, S. E. Svehag, and I. Brandslund. 1983. A standardized method for quantitating the complement mediated im-mune complex solubilizing capacity of human serum. J. Immunol. Methods. 59:369-380.

17. Baatrup, G., I. Petersen, E. Kappelgaard, H. H. Jepsen, and S.E.Svehag. 1984. Complement mediated solubilization inhibition and complementfactor levels in SLE patients. Clin. Exp. Immunol. 55:313-318.

18. Naama, J. K., W. S. Mitchell, and K Whaley. 1983. Inhibition of complement mediated solubilization of antigen-antibody complexes by sera frompatients with rhumatoid arthritis. Clin. Exp. Immunol. 54: 429-438.

19. Corvetta, A., P. J. Spaeth, P. A. Ghirelli, F. Orecchioni,R. Buetler, M. Montroni, and U. E. Nydegger. 1983. Complement activation and impaired capacity to solubilize immune complexes or to prevent their formation in essential mixed cryoglobulinemia. Diagn. Immunol. 1:315-323.

20. Naama, J. K., W. S. Mitchell, A. Zoma, J. Veitch, and K. Whaley. 1983. Complement mediated inhibition of immune precipitation in pa-tients with immune complex diseases. Clin. Exp. Immunol. 51:292-298. 21.Balestrieri, G.,A.Tincani,P.Migliorini, C. Ferri, R. Cattaneo, and S. Bombardieri. 1984. Inhibitory effect of IgM rheumatoid factor

onimmune complex solubilization capacity and inhibition of immune precipitation.Arthritis Rheum. 27:1130-1136.

22.Mitchell, W. S., J. K. Naama, J. Veitch, and K. Whaley. 1984. IgM-RF prevents complement-mediated inhibition of immune precip-itation.Immunology.52:445-448.

23.McConahey, P.J., andF.J.Dixon. 1966. A method for trace iodination ofproteins for immunological studies. Int. Arch. AllergyAppl. Immunol. 29:185-189.

24. Tenner, A. J., P. H. Lesarve, and N. R. Cooper. 1981. Purification andradiolabeling of human Clq.J.Immunol. 127:648-653.

25. Hammer, C. H., G.H.Wirtz, L. Renfer, H. T. Gresham, and

B.F.Tack.1981. Large scale isolation offunctionally active components of the human complement system. J.Biol. Chem. 256:3995-4006.

26.Medicus,R.G.,A.F.Esser, H.N.Fernandez, and H.J. Muller-Eberhard. 1980.Native and activatedProperdin: interconvertability and identity of amino- and carboxyterminal sequences. J. Immunol. 124: 602-606.

27.Laemmli,U.K. 1970.Cleavage ofstructureproteins duringthe assembly ofthe head ofbacteriophageT4. Nature(Lond.).227:680-685.

28.Sjoholm, A. G., J.A.Braconier, and C. Soderstrom. 1982.

Pro-perdin deficiency inafamily with fulminant meningococcal infections. Clin. Exp. Immunol. 50:291-297.

29. Lowry,B.H.,N. J.Rosebrought,A.L.Farr, andR. J.Randall. 1951.Protein measurement with the folin phenol reagent. J. Biol. Chem. 193:265-275.

30. Praz, F.,R. A. Barreira, and P. Lesavre. 1982. A one-step pro-cedure forpreparation of classical pathway (Clq) and alternative pathway (Factor D)depleted human serum. J. Immunol. Methods. 50:227-231. 31.Gaither, T. A., D. W.Alling, and M.M. Frank. 1974.Anew

one-step method for the functional assay of the fourth component (C4) ofhumanandguineapig complement. J. Immunol. 113:574-583.

32.Ziccardi, R. J., and N.R.Cooper. 1977. The subunit composition andsedimentationproperties of human Cl.J.Immunol. 118:2047-2052. 33.Schifferli,J.A.,A.Bakkaloglu, N. Amos, andD. K. Peters.1984. C4binding protein in sera ofpatients with systemic lupus erythematosus

andmixedessential cryoglobulinemia. Complement. 1:85-89. 34. Lachmann, P. J., and M. J. Hobart. 1976. Complement

tech-nology. InHandbook of Experimental Immunology.D. M.Weir, editor. BlackwellScientific Publications, Oxford. Thirded.SA: 1-23.

35. Zubler, R. H., G. Lange,P. H. Lambert, and P.A. Miescher. 1976.Detection of immune complexes in unheatedserabyamodified 125-1-Clqbindingtest.J.Immunol. 116:232-235.

36. Brouet,J.C., J. P. Claudel, F. Danon, M. Klein, and M. Selig-mann. 1974.Biological and clinical significance of cryoglobulins:areport of 86cases. Am.J.Med. 57:775-788.

37.Schifferli,J. A., G.Steiger,and M.Schapira. 1985. The role of Cl, Cl inactivator and C4 in modulating immuneprecipitation.Clin. Exp. Immunol. 60:605-612.

38. Naama,J. K.,A. 0. Hamilton,A. C. Yeung-Laiwah, andK.

Whaley. 1984. Prevention of immune precipitation by purified classical pathway complement components. Clin. Exp. Immunol. 58:486-492.

39.Wilson, C. B., and F. J. Dixon. Quantitation of acute and chronic

serumsickness in the rabbit. J. Exp. Med. 134:7S-18S.

40. Waxman, F.J., L.A.Hebert,J.B.Cornacoff,M. E. VanAman, W. L.Smead, E. H. Kraut, D. J. Firmingham, andJ.M.Taguiam. 1984. Complement depletion accelerates the clearance of immune complexes from thecirculation in primates. J. Clin.Invest.74:1329-1340.

41.Cameron, J.S., D.R.Turner, J. Heaton,D.G.Williams, C. S. Ogg, C. Chantler, G. B. Haycock, andJ.Hicks. 1983.Idiopathic

mes-angiocapillaryglomerulonephritis. Comparisonof typesIand II in chil-dren and adults andlong-termprognosis.Am.J. Med. 74:175-192.