INFECTION AND IMMUNITy, June 1975, p. 1235-1243 Vol. 11, No. 6Copyright 0 1975 American Society for Microbiology Printed in U.S.A.

Function of the Classical and Alternate Pathways of HumanComplement in Serum Treated with Ethylene Glycol

Tetraacetic Acid and MgCl2-Ethylene Glycol Tetraacetic AcidR. M. DES PREZ,* C. S. BRYAN, J. HAWIGER, AND D. G. COLLEY

Departments of Medicine and Microbiology, Vanderbilt University School of Medicine and VeteransAdministration Hospital, Nashville, Tennessee 37203

Received for publication 24 January 1975

An immunochemical and functional analysis of the classical and alternatecomplement pathways in human serum was performed in the presence of 10 mMethylene glycol tetraacetic acid (EGTA) and MgCl2-EGTA (MgEGTA), chelat-ing agents which have been recently utilized as a means of distinguishingbetween these two complement pathways. Total hemolytic activity, integrity ofthe Cl complex, hemolytic activity of C2, conversion of factor B (C3 proactiva-tor), and complement-dependent bactericidal activity were studied. The effect ofthese chelators on activation of complement pathways by Escherichia coli, bysensitized erythrocytes as a prototype of activators of the classical pathway, andby zymosan as a prototype of alternate (properdin) pathway activators wasstudied. Human serum containing 10 mM EGTA, which provides almost noionized calcium and considerably less ionized magnesium than unchelatedserum, allowed consumption of complement via the alternate (properdin)pathway, but blocked the classical pathway as judged by disintegration of the C1complex and lack of utilization of C2. However, activity of the alternatecomplement pathway in EGTA serum, as judged by conversion of factor B andbactericidal activity against gram-negative bacteria, was distinctly suboptimal.Addition of magnesium ion in a concentration equimolar to EGTA (MgEGTAserum), while still providing conditions in which the Cl complex dissociated,significantly enhanced alternate complement pathway-mediated bactericidalactivity. However, in MgEGTA serum considerable fluid-phase activation of thealternate pathway, as indicated by decrease in 50% hemolytic complement (CH50)titers and conversion of factor B to its active form in the absence of any activatingchallenge, was observed. Moreover, some fluid-phase consumption of C2 wasobserved in MgEGTA serum, even though, as mentioned, the Cl complex wasshown to be dissociated under these conditions. MgEGTA-related activation ofC2 and of the alternate (properdin) pathway of complement was significantlyenhanced by the presence of zymosan and E. coli. These results indicate that useof the chelating agents EGTA and MgEGTA to differentiate between classicaland alternate pathway activation of human complement is more complex thanhas hitherto been suggested. In EGTA serum, spontaneous activation of eitherpathway does not occur but bactericidal activity, as a measure of biologicfunction of complement, is suboptimal. In MgEGTA serum, bactericidal activityis fully expressed, but there is considerable instability, in terms of fluid-phaseactivation, in Mg2+-dependent components of both pathways. Thus, caution isindicated in the use and interpretation of the effects of these chelating agents onbiologic functions mediated by either pathway of human complement.

The complement system in man participates including immune complex disease (10, 25),in host immune reactions as a result of activa- sickle cell anemia (14), and liver disease (4). Ation of the classical (C142) and the alternate better understanding of these abnormalities(properdin) complement pathways. Acquired requires methods whereby the function of one ordefects in or depletion of the complement sys- anotlier of the two complement pathways cantem are thought to play a role in abnormal host be easily evaluated.defense mechanisms in several disease states We have previously applied the concept of

1235

on February 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

DES PREZ ET AL.

differential chelation of divalent cations usingethylene glycol tetraacetic acid (EGTA) as ameans of distinguishing between the classical(C142 mediated) and alternate (properdin)pathways of complement activation (8). Thisapproach has been frequently employed in sev-eral studies to ascertain the function of comple-ment in patients with sickle cell anemia (14), inparoxysmal nocturnal hemoglobinuria (2), inactivation of B lymphocytes (22), in endotoxin-induced complement changes (7), and in phago-cytosis of cryptococci (5) and of gram-positiveand gram-negative bacteria (9). Despite thesenumerous examples of applying EGTA and/orMgCl2-EGTA (MgEGTA) as chelating agentsfor differentiation of both complement path-ways, a detailed analysis of the effect of thesechelators on such divalent cation-dependentcomplement components as the Cl co'mplex,C2, and the properdin pathway is still lacking.The present study was undertaken to evalu-

ate the functional and immunochemical integ-rity of the human complement system in che-lated serum by measuring total hemolytic activ-ity, the immunochemical pattern of the Clcomplex, hemolytic activity of C2, the immuno-chemical pattern of factor B (C3 proactivator),and complement-dependent bactericidal activ-ity. Different chelating conditions providingvaried concentration of magnesium and calciumions were compared. Three known activators ofcomplement pathways, sensitized erythrocytes,zymosan, and Escherichia coli, were used. Thedata suggest that the use and interpretation ofthe effect of chelating agents on the comple-ment system of man requires caution since it ismore complex than had previously beenthought.

(Part of this work was reported at the 2ndInternational Congress of Immunology, Brigh-ton, England, 1974.)

MATERIALS AND METHODSHuman blood. Blood was obtained by venipunc-

ture from healthy volunteers and was allowed to clotfor 1 h at room temperature, and the serum was usedfresh or, in most experiments, after storage at -70 Cfor 1 week or less.

Chelators. The preparation of 100 mM stock salinesolutions of ethylenediaminetetraacetate (EDTA).EGTA, EGTA supplemented with an equimolar con-

centration of MgCl2 (MgEGTA), and the source ofthese reagents have been described previously (8).Stock solutions were added to serum or buffer toprovide a final concentration of 10 mM.Complement activation (fixation) experiments.

Serum with or without chelators was incubated at37 C for 30 min in the presence or absence of' threeactivating substances. These included E. coli "E," a

clinical isolate from the bacteriology laboratorv ofVanderbilt Hospital, zymosan (Z), and sheep erythro-cytes sensitized with anti-sheep hemolysin (EA). E.coli "E" was maintained on Trypticase soy agarslants, grown overnight at 37 C in Trypticase soybroth, centrituged, washed three times with saline,and standardized after resuspension in saline byoptical density determination at 500 nm based onpreviously determined curves relating optical densitvto viable colony counts by standard bacterial enumer-ation techniques. Z (Nutritional Biochemicals Corp.,Cleveland, Ohio) was boiled for 1 h, washed threetimes in saline, and resuspended in serum at a f'inalconcentration of 2 mg/ml (26). EA were preparedaccording to the method of Rapp and Borsos (28) fromsheep erythrocytes (Baltimore Biological Laboratory,Cockeysville, Md.) and anti-sheep hemolvsin (DifcoLaboratories, Detroit, Mich.). and concentrated asdescribed previously (8). Normal saline was substi-tuted for either the activating substances or thechelator in control experiments. After incubation, theactivating substances were separated by centrifuga-tion (2,500 x g for 10 min at 4 C) and the supernatantserum was analyzed for complement activitv afteraddition of 0.1 ml of 100 mM CaClJ/ml of' serum tothose samples initially containing either EGTA orMgEGTA.

Determination of hemolytic complement activity(CH50 titers). Serum samples were analvzed bv themethod of Rapp and Borsos (28).

Determination of C2 hemolytic activity. Serumsamples were analyzed according to the method of'Rapp and Borsos (28), using purif'ied guinea pig Cl(Cordis Laboratories, Miami, Fla.). Purit'ied humanC2 from the same source was used as a standard.Immunoelectrophoresis. This was performed in

1% agarose on microscope slides in veronal buffer ofionic strength 0.05 and pH 8.0 according to the methodof Scheidegger (32). In studies utilizing chelated se-rum, it was necessary to include the chelator in a finalconcentration of 10 mM in both the agarose and bufferby the method of Nagaki and Stroud (24).

Analysis of Cl complex. Serum samples with andwithout chelators were analyzed by immunoelectro-phoresis using monospecific rabbit anti-human Cisantiserum kindly provided by Robert Stroud, Univer-sity of Alabama Medical School, Birmingham (24).

Assay of conversion of factor B (C3 proactivator,C3PA) to factor B (C3 activator, C3A). Nonchelatedand chelated serum samples incubated with andwithout the challenge particles for 30 min at 37 Cwere analyzed by immunoelectrophoresis using mono-specific anti-C3A antiserum (Behring Diagnostics,Summerville, N.J.) by the methods of Gotze andMiiller-Eberhard (12). The degree of conversion wasestimated visually and graded from 0 to + + + t

Bactericidal assays. The following bacterialstrains were employed in these assays. E. coli "E" andSalmonella typhi were clinical isolates obtained fromthe bacteriology laboratory, Vanderbilt UniversityHospital. E. coli "C" and Proteus rettgeri were kindlyprovided by Charles E. McCall, Bowman Gray Schoolof Medicine. All four strains were selected for theirexquisite sensitivity to the complement-mediated

INFECT. IMMUN.1236

on February 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

VOL. 11, 1975 HUMAN COMPLEMENT PATHWAYS IN CHELATED SERUM

serum bactericidal system. Bacteria were maintainedon Trypticase sov agar slants and were grown over-night at 37 C prior to each experiment. To obtainlog-phase organisms for bactericidal assays, bacteriawere transferred with a standardized inoculum loopfrom an overnight culture to fresh Trypticase soybroth and allowed to grow at 37 C from 150 to 250 min,depending on the bacterial strain.

At the start of each experiment, 0.2-ml aliquots of a10-3 dilution of the log-phase culture were added toplastic tubes (75 by 12 mm) containing 1.8 ml ofserum and 0.2 ml of either stock chelator solution ornormal saline. All tubes were warmed to 37 C imme-diatelv prior to inoculation. The final concentration ofbacteria in each tube was approximately 3 x 103 perml. The tubes were incubated at 37 C with gentleend-over-end tumbling on a tube rotator. Trypticasesoy agar pour plates were made from 0.1-ml aliquotsremoved from each tube at timed intervals; plateswere incubated for at least 24 h before colony enumer-ation. "Complete killing" was defined as the absenceof viable bacteria in aliquots removed from thebacteria-serum mixture after 1 h of incubation. "Sig-nificant killing" was defined as a reduction in thenumber of viable bacteria to less than 10% of theoriginal inoculum by the end of 1 h (6).

RESULTSImmunochemical and functional charac-

terization of the classical and alternate com-plement pathways in human serum chelatedwith EGTA. Three steps in complement activa-tion require divalent cations: Ca2'--dependentassembly of Cl complex (15), Mg2+-dependentC2 activation (16), and Mg2+-dependent activa-tion of the properdin pathway (12, 26. 31).Although normal human serum contains ap-proximately 10-3 M ionized calcium and mag-nesium, serum chelated with 1 x 10-3 MEGTA contains 10- l M Ca2` and 1.6 x 10-6 MMg2+ (2, 13). Under conditions of almost com-plete deprivation of Ca ion and borderlinedepletion of Mg ion by EGTA, some steps ofcomplement activation will be inhibited. Thefollowing experiments were performed to docu-ment in detail the pattern of the inhibitoryeffect of EGTA.

(i) Immunochemical demonstration of theCl complex disintegration in the presence ofEGTA. Nagaki and Stroud, using antiserummonospecific for purified Cls, demonstratedthat the immunoelectrophoretic mobility of C1was altered in the presence of EDTA andbecame similar to the mobilitv of purified Clseither in the presence or absence of the chelator(24). This was taken as evidence that dissocia-tion of Cl into its subcomponents (Clq, Clr,and Cis) could be demonstrated by immuno-electrophoresis. As illustrated in Fig. 1, dis-sociation of the Cl complex took place also in

EGTA serum in a manner quite similar to thatobserved in EDTA serum used as a positivecontrol. Cls migrated toward the anode, indi-cating it dissociation from the Cl complex. Incontrast, in nonchelated serum (negative con-trol), material immunoreactive with anti-Clsremained at the origin in the macromolecularCl complex. Thus, the trimolecular Cl compo-nent of complement dissociated in EGTA se-rum.

(ii) Consumption of total hemolytic com-plement activity and C2 activity in EGTAserum. The effect of EGTA on the functionalintegrity of the complement system and of theC2 component was assessed. For this purposethe consumption of total hemolytic activity andof C2 hemolytic activity by the activatingsubstances was measured. Three activatorswere used: E. coli and Z, known to activate thealternate pathway, and EA as a prototype ofclassical pathway activating agents. In thecontrol. unchelated human serum, E. colicaused almost complete consumption of CH50and C2 (Fig. 2). Z and EA also caused completeconsumption of total hemolytic complementactivity but the pattern of C2 consumptiondiffered: EA challenge produced 100% and Zchallenge 45% consumption of C2, respectively.These patterns of consumption in unchelatedserum suggest involvement of both the classical(C142) and alternate (properdin) pathways incomplement activation by E. coli and Z.Human serum chelated with EGTA showed a

different pattern of decomplementation. E. coliand Z still caused consumption of total comple-ment (CHso), 70% and 98%, respectively. How-ever, consumption of C2 was blocked. Bothwhole complement fixation (CH50) and C2 fixa-tion by EA were almost completely blocked inEGTA serum. These results indicate that chela-tion of serum by EGTA prevented activation ofthe classical pathway by EA and also by E. coliand Z. Hence, consumption of total comple-ment activity by E. coli and Z observed inEGTA serum represented activation of thealternate (properdin) pathway.

(iii) Activation of the alternate (proper-din) pathway as measured by conversion offactor B (C3 proactivator) in EGTA serum.Consumption of whole complement (CH50) butnot C2 in EGTA serum incubated with E. coliand Z provided indirect evidence that thealternate complement pathway was activatedunder these conditions. Direct evidence foralternate pathway activation was obtained byimmunoelectrophoretic analysis of factor B (C3proactivator) conversion to factor B (C3 activa-tor). Using antiserum monospecific to C3 ac-

1237

on February 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

DES PREZ ET AL.

UNCH ELATEDSERUM

EDTA

EGTA

M g EGTA

FIG. 1. Dissociation of Cl 'complex in chelated serum. Immunoelectrophoresis of nonchelated serum orserum containing the designated chelators was carried out as described. Experiments using chelated serum werecarried out with the same concentration of chelator in the agarose gel. Anti-Cls antibody is in the center trough.The movement of material reactive with anti-CIs antibody toward the anode took place in the presence of allthree chelators. In nonchelated serum this material remained at the cathodal side of the well. Anode is at left.

NO CHELATORI00-

0 EGTA

o-)wz0Q LLLnC2

NO CHALLENGE E.COL ZYMOZAN EA

FIG. 2. Consumption of total hemolytic comple-

ment activity and C2. Unchelated serum, EGTA

serum, and MgEGTA serum were incubated with E.

coli (10'° organisms/mI of serum), Z (2 mg/mI of

serum), and EA (2 x 108/ml of serum). Incubation

was carried out for 30 min at 37 C before sedimenta-

tion of challenges and determination of residual CH.ioand C2 titers (see text). Figures represent percentage

of control (normal human serum) expressed as mean

value plus or minus the standard error of the mean.

tivator, the change in electrophoretic mobilityof material reacting with this antiserum was

observed in nonchelated human serum incu-bated with E. coli or Z but was not seen inserum incubated with EA (Fig. 3). Such a

change in electrophoretic mobility indicatesconversion of factor B (C3 proactivator) tofactor B (C3 activator) (12) due to activation ofthe alternate (properdin) pathway. In EGTAserum, conversion of C3 proactivator to C3activator was observed in the presence of E. coliand Z but not with EA (Fig. 3). However, theextent of C3 proactivator conversion was con-sistently smaller in EGTA serum than in un-chelated serum. This immunochemical analysisof factor B indicates that activation of thealternate (properdin) pathway of complementby E. coli and Z occurs in EGTA serum but to alesser degree than in unchelated serum.

(iv) Complement-mediated bactericidalactivity of human serum in the presence ofEGTA. Gram-negative bacteria are target cellsfor the complement system in human blood,and evaluation of bactericidal activity ofhuman serum represents one of the most sensi-tive indexes of the biologic function of comple-ment. Therefore, complement-mediated bacte-ricidal activity of human serum was evaluatedunder the chelating conditions studied. In un-chelated human serum, bactericidal activityagainst complement-sensitive strains of E. coli,S. typhi, and P. rettgeri was complete within 10

(I)

U,

1. §z

1238 INFECT. IMMUN.

w w 'WNIPRM

on February 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

VOL. 11, 1975 HUMAN COMPLEMENT PATHWAYS IN CHELATED SERUM 1239

CONTROL E. COLI

NO -'i St&KNoCHELATOR -0& CHELATOR . .'

EGTA ..., EGTA j- c

MgEGTA MEGTAcc.

0~~~~~~~~~EA ZYMOZAN>

NO 0 N0 ..*

CHELATORA'~< CHELATOR

z~~~ ~ ~ ~~~~~4-z

Mg EGTA -- MgEGTA

FIG. 3. Conversion of factor B (C3 proactivator) to factor B (C3 activator) in unchelated and chelated humanserum incubated with E. coli, EA, and Z. The concentrations of the challenge material, time of incubation ofserum, and inclusion of chelators in the immunoelectrophoresis system are the same as described in Fig. 1 and2. Anti-factor B (C3A) antibody is in the center troughs. Nonconverted factor B occupied a position anodal fromthe well; conversion to C3A was associated with movement of the immunoreactive material in a cathodal direc-tion. Anode is at right.

to 20 min (curve A in four panels of Fig. 4). In E. COLI 'E" E COLI 'C"serum containing 10 mM EGTA, bactericidal E

activity was inconsistent (curve C in four panels 100 Dof Fig. 4). Thus, these results indicate furtherthat EGTA serum provides suboptimal condi- x

tions for activation of the complement system '0as manifested by bactericidal activity. It must X, > xc

be noted that heat-labile bactericidal activity inEGTA serum must represent alternate pathway Xactivity since, as demonstrated, the classical < A A Bpathway activation is completely blocked. 00, B,Immunochemical and functional charac- SX) SALMONELLA TYPH PROTE US RETTGER

terization of the classical and alternate com - EDplement pathways in human serum chelated w _M D D

with MgEGTA. The results of the experiments i x-xdescribed above indicated that chelation of x--xc

human serum with EGTA causes disintegration 'o X- xof the Cl complex and blocks consumption of xc

C2 preventing activation of complement byclassic pathway activators such as EA. Further, lserum chelated with EGTA permits activationof the alternate (properdin) pathway but at an

0A B

apparently suboptimal rate. Reasoning that 0 5 10 20 40 60 0 5 10 20 40 60this suboptimal rate might be due to a subopti- TIME IN MINUTES

mal concentration of Mg2+ (1.6 x 10-6 M), FIG. 4. Bactericidal activity of human serumEGTA serum was supplemented with equimolar against four complement-sensitive bacterial strains inconcentration of Mg2+ (1 x 10-3 M) to produce unchelated fresh serum (A), MgEGTA serum (B), anda final concentration of 3 x 10-3 M Mg2+ (2). EGTA serum (C). Controls consisted of heat-inac-Under these conditions, the concentration of tivated serum containing MgEGTA (D) or EGTA (E).ionized calcium is 10-1o M. Thus, chelation with All chelators were in 10 mM concentration.

I

on February 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

DES PREZ ET AL.

MgEGTA provides near normal concentrationsof ionized magnesium but almost total chela-tion of ionized calcium. MgEGTA serum wasstudied in a parallel series of' experiments as-sessing the immunochemical and functionalproperties of' both complement pathways.

(i) Immunochemical demonstration of theCl complex disintegration in the presence ofMgEGTA. Addition of MgEGTA to humanserum caused the material immunoreactivewith anti-Cls antibody to migrate toward theanode in a manner similar to EDTA and EGTA(Fig. 1). This was taken as evidence that the Clcomplex was dissociated in MgEGTA serum.

(ii) Consumption of total hemolytic com-plement activity and C2 activity in MgEGTAserum. The pattern of decomplementation inhuman MgEGTA serum was different from thatobserved in either nonchelated or EGTA serum(Fig. 2). Significant decrease in whole comple-ment (69%) and C2 (38'%) titers was observedaf'ter incubation of' MgEGTA serum for 30 minat 37 C in the absence of' any challenge material(fluid-phase activation). Both E. coli and Zproduced complete consumption of' whole com-plement and a high degree of' consumption ofC2. The fluid-phase activation manifested by thespontaneous and significant decrease in wholecomplement titers (in the absence of' decomple-menting challenge materials) and to a lesserdegree in C2 titers was unanticipated andindicated that C2 and perhaps other compo-nents of' complement pathways were modifiedor unstable in MgEGTA serum.

(iii) Activation of the alternate (proper-din) pathway as measured by conversion offactor B (C3 proactivator) in MgEGTAserum. Activation of factor B as measured by achange in electrophoretic mobility was observedin MgEGTA serum in the absence of' any chal-lenge material whatsoever (control), or with EAwhich, as described above, did not activate thealternate pathway in nonchelated or EGTAserum (Fig. 3). Furthermore, conversion of' C3proactivator (factor B) to C3 activator was morecomplete than in EGTA serum after incubationwith E. coli or Z. These results provide evidencethat conditions for activation of the alternate(properdin) pathway by various activating sub-stances (E. coli, Z) are more favorable inMgEGTA serum than in EGTA serum. How-ever, spontaneous and partial fluid-phase alter-nate pathway activation in MgEGTA serum inthe absence of' activating substances considera-bly complicates interpretation of the data withactivating substances.

(iv) Complement-mediated bactericidalactivity of human serum in the presence of

MgEGTA. When human serum chelated withMgEGTA was tested for complement-depend-ent bactericidal activity, complete killing of allfour tested gram-negative bacterial strains wasregularly observed (curve B in four panels of Fig.4). There was a characteristic delay in the onsetof bactericidal activity in MgEGTA serum com-pared to unchelated serum. Nevertheless, com-plete killing of the organisms tested occurredduring first 40 min of incubation in MgEGTAserum.

DISCUSSIONThe results of the present study will be

discussed in relation to: (i) the role of' divalentcations in the function of the classical andalternate complement pathways; (ii) the natureof complement activation in the presence ofEGTA and MgEGTA; and (iii) the complexityof the action of the chelators due to the biologiceffects of' different absolute and relative concen-trations of' ionized calcium and magnesium.The importance of' divalent cations in the

complement system was initially reported in1906 by Cernovodeanu and Henri (3). Mayer etal. found in their classic studies that Mg2-markedly enhanced the hemolytic activity of'complement (21) and Levine et al. indicatedthat Ca2t was required for the function of' Clwhereas Mg2+ was necessary for the action of C2(16, 17). Subsequent studies of Pillemer et al.postulated that magnesium ion was an essentialcomponent of' the properdin system (26). Sand-berg and Osler confirmed that Mg2- was re-quired for alternate (properdin) pathway activ-ity (31), and Miiller-Eberhard and Gotze havesuggested that magnesium ion is required forthe function of' factor D (C3PA convertase) (23).Thus, calcium ion is required during the earlyCl-dependent step of' the classical pathwayactivation and magnesium ion is necessarv forthe C2-dependent step in the classical pathwayand for the function of' the properdin system.These different divalent ion requirements in thetwo pathways prompted the idea to utilizechelators not only to block calcium-dependentand magnesium-dependent complement steps(15-17) but also to differentiate one pathwayfrom other (8). This approach has been em-ployed in several investigations of the functionof complement in several different systems(2, 5, 7, 9, 10, 22).Our data support these findings but with

some qualit'ications. First, immunochemicalanalysis of the Cl complex using anti-C isantibody indicated disintegration of Cl inEGTA serum in a manner analogous to theeffect of EDTA as previously shown by Nagaki

1240 INFECT. IMMUN.

on February 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

VOL. 11, 1975 HUMAN COMPLEMENT PATHWAYS IN CHELATED SERUM

and Stroud (24). Second, EGTA blocked fixa-tion of complement via the classical pathway byEA. This blocking effect was assessed by mea-surement of total hemolytic activity and C2hemolytic activity. Third, as predicted, EGTAserum allowed consumption of complement byE. coli and Z, both known to activate thealternate pathway (12, 18, 26). Activation of thealternate (properdin) pathway was documentedimmunochemically by conversion of factor B(C3 proactivator) to factor B (C3 activator).However, this conversion was not complete inEGTA serum challenged with E. coli or Z(Fig. 3). Likewise, bactericidal activity, ex-amined as a sensitive biologic end point ofthe complement system, was not fully ex-pressed in EGTA serum. Thus, although in-hibition of the classical (C142) pathway ofcomplement was documented by several meansin EGTA serum, the alternate (properdin) path-way did not function at an optimal rate. Thesuboptimal function of the alternate pathway inEGTA serum is best explained by suboptimalconcentration of ionized magnesium.An additional limitation of the use of EGTA

serum or plasma is that the low magnesium ionconcentration in the presence of EGTA may bedetrimental to a variety of biologic systems,including the structural integrity of target cells.For example, magnesium ion is considered to bethe critical divalent cation for maintaining thestructural integrity of the E. coli cell wall (1, 30).Experiments from this laboratory reported else-where (1) have demonstrated that the relativelycomplement-resistant strain E. coli ATCC25911 was killed more readily in EGTA serumthan in MgEGTA serum. This is exactly oppo-site to the effects of these two chelators onrelatively complement-sensitive strains, as il-lustrated in Fig. 4. This effect of EGTA on thecomplement-resistant strain was attributed tocell wall damage as a consequence of magne-sium ion deprivation, rendering it much moresensitive to the serum bactericidal system.As discussed above, the concentration of

magnesium ion in EGTA serum was distinctlysuboptimal for activation of the alternate (pro-perdin) pathway as measured by incompleteconversion of factor B and by inconsistentcomplement-mediated bactericidal activity.This suboptimal activity in EGTA serum hasled to the addition of equimolar concentrationof magnesium ion to EGTA. MgEGTA serumprepared in this fashion and used in this andother laboratories (R. B. Johnston, Jr., personalcommunication) is deficient in ionized calcium(10-1o M) but the concentration of Mg2+ isincreased to 3.0 x 10-3 M. Under these chelat-

ing conditions, the Cl complex was dissociatedas demonstrated by electrophoretic mobility ofCls, but Mg-dependent steps such as activationof C2 and the properdin pathway became spon-taneously triggered or unstable in the fluid-phase during incubation of serum at 37 C.Despite this fluid-phase activation, comple-ment-dependent bactericidal activity was fullyand consistently expressed in MgEGTA serum.It appears that MgEGTA, added just prior to atest reaction, can be a useful reagent for indicat-ing the presence or absence of alternate path-way-mediated biologic activity of complementbut that caution should be exercised in thequantitation of the phenomena observed underthese conditions.The spontaneous fluid-phase activation or

instability of Mg-dependent reactions (C2 stepand properdin pathway) in MgEGTA serummerits emphasis. This phenomenon was tem-perature dependent since it occurred in MgEG-TA serum incubated alone at 37 C; incubationof MgEGTA serum at 4 C prevented spontane-ous activation of C2 or conversion of C3 proac-tivator. The presence of a particulate matrixsuch as EA enhanced these MgEGTA-relatedprocesses of C2 consumption or of alternatepathway activation causing appreciable loss ofwhole complement activity.The consistent fall in C2 titer in MgEGTA

serum both in the presence and absence ofchallenge particles is unexplained. Since the C1complex was observed to dissociate into itscomponent parts in MgEGTA serum as well asin EGTA serum, it seems unlikely that activa-tion of the classical pathway based on an intactCl complex could account for this observation.Since C2 titers did not decrease in EGTA serumafter the same manipulations, the presence ofthe chelator per se could not account for thisphenomenon. Since magnesium is required forthe reactivity of C2 in immune hemolysis (16),it is reasonable to postulate that some aspect ofthe C2 activating mechanism may be spontane-ously triggered under conditions of relativemagnesium ion excess in a manner analogous tospontaneous activation of factor B. Since the C1complex is dissociated in MgEGTA serum, theobserved decrease in C2 titers in MgEGTAserum cannot be ascribed to the newly de-scribed C1 bypass activator pathway (20). How-ever, there is a possibility that under conditionsof a relative excess of Mg ion, Cls is able to acton C2. Polley and Miiller-Eberhard observedinactivation of C2 during its purification in 2 x10-3M EDTA, pH 6.0, and they attributed thisphenomenon (and C4 inactivation) to Cls (27).Gigli and Austen demonstrated that fluid-phase

1241

on February 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

1242 DES PRE

activation of C2 by Cl involves unmasking theC2 receptor site on Cls which is otherwiseblocked by Clr (11). Such unmasking may beaccelerated by a relative excess of Mg ion inMgEGTA serum.The spontaneous activation of the alternate

(properdin) pathway in MgEGTA serum asmanifested by conversion of C3 proactivator isprobably due to the appreciable excess of mag-nesium ion in relation to calcium ion. Thealternate pathway of complement activated inthe presence of target cells by the relative excessof Mg2+ in MgEGTA serum will trigger themembrane lesion responsible for complete kill-ing of gram-negative bacteria (Fig. 4) or for lysisof paroxysmal nocturnal hemoglobinuria eryth-rocytes as previously demonstrated by Bryantand Jenkins (2). The critical condition for thisspontaneous activation is not the absolute con-

centration of magnesium but rather the ratio ofionized magnesium to ionized calcium as re-cently suggested by May et al. (20) and Lam-bert et al. (P. H. Lambert, L. Perrin, and J. C.Cerottini, J. Immunol. 111:307).The concept emerging from these and other

studies is summarized in Table 1. In vitroelimination of calcium and magnesium ion bychelation with EDTA inhibits the classical andalternate pathways. Chelation of human serumwith EGTA inhibits the classical pathway dueto elimination of Ca2+ but leaves enough Mg2+to allow the function of the alternate pathway ata suboptimal rate resulting in incomplete bacte-ricidal activity. Human serum containingMgEGTA has an increased total concentrationof Mg2+ compared to nonchelated serum butstill vanishingly small concentration of Ca2.These divalent ion concentrations result in thefluid phase activation of the Mg2+-dependentC2 step and of the properdin pathway. How-ever, complement-dependent bactericidal ac-

tivity is fully expressed. The ratio of ionizedmagnesium to ionized calcium in MgEGTA

:Z ET AL. INFECT. IMMUN.

serum is approximately 30 million; normalserum has approximately equal concentrationsof both ions. Thus, although both EGTA andMgEGTA can be used as reagents permittingalternate pathway activation, their effects areconsiderably more complex than has hithertobeen suggested and caution is needed in inter-pretation of biologic effects studied in thepresence of these chelators.

ACKNOWLEDGMENTSThe expert technical assistance and collaboration of

Aleksandra Hawiger is most gratefully acknowledged. Wewish to thank Elzora Knippers for her excellent technicalassistance. We are indebted to Robert Stroud, Universityof Alabama, for antiserum monospecific for Cls.

This work was supported by Veterans Administrationtraining program TR no. 74 and by grants from the UnitedStates Public Health Service (HL 08399 and HL 16646).

LITERATURE CITED1. Bryan, C. S. 1974. Sensitization of E. coli to the serum

bactericidal system and to lysozyme by ethyleneglycol-tetraccetic acid (EGTA). Proc. Soc. Exp. Biol. Med.145:1431-1433.

2. Bryant, R. E., and D. E. Jenkins, Jr. 1968. Calciumrequirements for complement dependent hemolyticreactions. J. Immunol. 101:664-668.

3. Cernovodeanu, P., and V. Henri. 1906. Activation dupouvoir hemolytique de certains serums por les sels demagnesium. C. R. Soc. Biol. 60:571-573.

4. DeMeo, A. N., and B. R. Andersen. 1972. Defectivechemotaxis associated with a serum inhibitor in cir-rhotic patients. N. Engl. J. Med. 286:735-740.

5. Diamond, R. D., J. E. May, M. A. Kane, M. M. Frank,and J. E. Bennett. 1974. The role of the classical andalternate complement pathway in host defenses againstcryptococcus neoformans infection. J. Immunol.112:2260-2270.

6. Fierer, J., F. Finley, and A. I. Braude. 1972. A plaqueassay on agar for detection of gram-negative bacillisensitive to complement. J. Immunol. 109:1156-1158.

7. Fine, D. P. 1974. Activation of the classic and alternatecomplement pathways by endotoxin. J. Immunol.112:763-769.

8. Fine, D. P., S. R. Marney, Jr., D. G. Colley, J. S. Sergent,and R. M. Des Prez. 1972. C3 shunt activation inhuman serum chelated with EGTA. J. Immunol.109:807-809.

9. Forsgren, A., and P. Quie. 1974. Influence of the alternate

TABLE 1. Summary of data on the function of the classic and alternate complement pathways in human serumunder conditions of divalent cation chelation

Approximate Complement pathwaysDeterminant Mg2+ Ca2+ ratio of

Mg2+ to Ca2+ Classic Altemate

Unchelated 10-3 M 10- 3M 1.0 Functional Functional10 mM EDTA serum 10-9 M 10-11 M 102 Inhibited Inhibited10 mM EGTA serum 1.6 x 10- M 5.0 x 10- I'M 3.1 x 104 Inhibited Suboptimally

functional10 mM MgEGTA serum 3.0 x 10-3 M 10- 10M 3.0 x 107 Partially Functional"

inhibitedaa Cl complex dissociated; C2 activated in a fluid phase.b Factor B (C3 proactivator) activated in a fluid phase.

on February 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

VOL. 11, 1975 HUMAN COMPLEMENT PATHWAYS IN CHELATED SERUM

complement pathway on opsonization of several bacte-rial species. Infect. Immun. 10:402-404.

10. Gewurz, H., R. J. Pickering, and S. E. Mergenhagen.1968. The complement profile in acute glomerulone-phritis, systemic lupus erythematosus and hypocom-plementemic chronic glomerulonephritis: contrastsand experimental correlations. Int. Arch. Allergv Appl.Immunol. 34:556-570.

11. Gigli, I., and K. F. Austen. 1969. Fluid phase destructionof C2bu by Clhu. 11. Unmasking by C4i6u of C1muspecificity for C2"u. J. Exp. Med. 130:833-849.

12. Gotze, O., and H. J. Miiller-Eberhard. 1971. The C3-activator system: an alternate pathway of complementactivation. J. Exp. Med. 134:90S-108S.

13. Hovig, T. 1964. The effect of calcium and magnesium onrabbit blood platelet aggregation in vitro. Thromb.Diath. Haemorrh. 12:179-200.

14. Johnston, R. B., Jr., S. L. Newman, and A. G. Struth.1973. An abnormality of the alternate pathway ofcomplement activation in sickle-cell disease. N. Engl.J. Med. 288:803-808.

15. Lepow, I. H., G. B. Neff, E. W. Todd, J. Pensky, and C.F. Hiers. 1963. Chromatographic resolution of the firstcomponent of human complement into three activities.J. Exp. Med. 117:983-1008.

16. Levine, L., M. M. Mayer, and H. J. Rapp. 1954. Kineticstudies on immune hemolysis. VI. Resolution of the C'reaction step into two successive processes involvingC'2 and C'3. J. Immunol. 73:435-442.

17. Levine, L., A. G. Osler, and M. M. Mayer. 1953. Studieson the role of Ca and Mg in complement fixation andimmune hemolysis. J. Immunol. 71:359-379.

18. Marcus, R. L., H. S. Shin, and M. M. Mayer. 1971. Analternate complement pathway; C-3 cleaving activitynot due to C4, 2a on endotoxic lipopolysaccharide aftertreatment with guinea pig serum; relation to properdin.Proc. Natl. Acad. Sci. U.S.A. 68:1351-1354.

19. May J. E., and M. M. Frank. 1973. Hemolysis of sheeperythrocytes in guinea pig serum deficient in the fourthcomponent complement. II. Evidence for involvementof Cl and components of the alternate complementpathway. J. Immunol. 111:1668-1676.

20. May J. E., W. Rosse, and M. M. Frank. 1973. Paroxyso-mal hemoglobinuria alternate-complement-pathway-

mediated lysis induced by magnesium. N. Engl. J.Med. 289:705-709.

21. Meyer, M. M., A. G. Osler, 0. G. Bier, and M. Heidel-berger. 1946. The activating effect of magnesium andother cations on the hemolytic function of complement.J. Exp. Med. 84:535-548.

22. Miller, G. W., P. H. Saluk, and V. Nussenzweig. 1973.Complement-dependent release of immune complexesfrom the lymphocyte membrane. J. Exp. Med. 138:495-507.

23. MUiller-Eberhard, H. J., and 0. G6tze. 1972. C3 Pro-activator convertase and its mode of action. J. Exp.Med. 135:1003-1008.

24. Nagaki, K., and R M. Stroud. 1969. Specific antiserato Cls: detection of different electrophoretic speciesof Cls. J. Immunol. 103:141-145.

25. Perrin, L. H., P. H. Lambert, U. E. Nydegger, and P. A.Miescher. 1973. Quantitation of C3PA (properdin fac-tor B) and other complement components in diseasesassociated with a low C3 level. Clin. Immunol. Im-munopath. 2:16-27.

26. Pillemer, L., L. Blum, I. H. Lepow, E. W. Todd, andA. C. Wardlaw. 1954. The properdin system and im-munity. I. Demonstration and isolation of a new serumprotein, and its role in immune phenomena. Science120:279-285.

27. Polley, M. J., and H. J. Muller-Eberhard. 1968. Thesecond component of human complement: its isolation,fragmentation by C'l esterase, and incorporation intoC3 convertase. J. Exp. Med. 128:533-551.

28. Rapp, H. J., and T1. Borsos. 1970. Molecular basis ofcomplement action. Appleton-Century-Crofts, NewYork.

29. Ruddy, S., I. Gigli, and K. F. Austen. 1972. The com-plement system of man. N. Engl. J. Med. 287:489-495, 545-549, 592-596, 642-646.

30. Salton, M. R., Jr. 1964. The bacterial cell wall. Elsevier,Amsterdam.

31. Sandberg, A. L., and A. G. Osler. 1971. Dual pathwaysof complement interaction with guinea pig immuno-globulins. J. Immunol. 107:1268-1273.

32. Scheidegger, J. J. 1955. Une micro-methode de l'immunoelectrophorese. Int. Arch. Allergy Appl. Immunol. 7:103-110.

1243

on February 17, 2020 by guest

http://iai.asm.org/

Dow

nloaded from