The Interaction of Hageman Factor and Immune Complexes

CHARLESG. COCHRANE,KIRK D. WUEPPER,BARBARAS. AIKEN,SUSAND. REVAK, and HANSL. SPIEGELBERG

From the Department of Experimental Pathology, Scripps Clinic and ResearchFoundation, La Jolla, California 92037

A B S T R A C T The possible interaction of Hageman fac-tor from human or rabbit plasma with a variety of im-munologic reactants was studied. Evidence of an interac-tion was not obtained and neither binding of radiolabeledHageman factor to immune aggregates nor depletion ofthe Hageman factor from the supernate was observed.Cleavage of the labeled Hageman factor molecule intoits 30,000 ;molecular weight-active fragments was notdetectable after incubation with immune complexes.

Isolated Hageman factor was far more sensitive to ac-tivation than Hageman factor in plasma or serum. Therewas no consistent activation of isolated Hageman factorby immunologic reactants as determined by conversionof prekallikrein to its enzymatic form or by shorteningof the clotting time of factor XII-deficient plasma. Avariety of immunologic stimuli were tested: (a) anti-gen-antibody complexes in soluble or precipitated form;(b) particulate antigen-antibody complexes, i.e., zymosan-anti-zymosan in which a surface was presented for ac-tivation; (c) human IgM-IgG and IgG-IgG (rheumatoidfactor) complexes; (d) immune aggregates consisting ofheat or bis-diazotized benzidine-aggregated myeloma pro-teins of all human immunoglobulin classes and subclasses:IgGl,2,34, IgA, IgD, IgM, and IgE. Absorption with im-mune aggregates did not reduce the quantity of Hagemanfactor in solution, nor was the Hageman factor bound tothe precipitates. The presence of plasma or serum withimmune aggregates did not generate activity of the Hage-man factor.

The only preparations of immunoglobulins capable ofactivating Hageman factor were found to be contami-

Dr. Wuepper is the recipient of U. S. Public HealthService Research Career Development Award 1 K4 AM40, 086-01. Dr. Spiegelberg is supported by grants 70 710from the American Heart Association, Special grant 531from the American Cancer Society, California Division,and is a Faculty Research Associate of the American CancerSociety (PRA-38).

Received for publicationz 3 April 1972 and il revised form13 Jutne 1972.

nated writh bacteria. These bacteria, upon isolation, acti-vated Hageman factor.

INTRODUCTION

Little is knowni about the participation of the kinin-form-ing and intrinsic clotting systems in inflammatory injury.Components of these systems may be activated by factorsinducing inflammationi or by plasma and cellular partici-pants brought into the site secondarily. Among the in-citing factors capable of inducing inflammatory reactionsthat have been considered as activators of the kinin-forming and intrinsic clotting systems are complexes ofantigen and antibodv.

Various mlethods have been employed to test for activa-tioni of Hagemain factor and the kinin-forimiing system byimmune reactants. Immune complexes have been addedto fresh plasma to determine if kinin is released orkininogen depleted. Similarly, immune complexes havebeen added to normal plasma not previously exposed toglass to determine the effect upon the recalcification clot-ting time of the plasma. Finally, immune precipitates havebeen exposed to serum that had not contacted glass andthe precipitates, after centrifugation, eluted to search foractive components of the kinin system.

In order to measure activation of Hageman factor di-rectly, it has been necessary to isolate and purify Hage-man factor. In this way, the interaction with immunecomplexes could be measured directly and the effect upoInthe Hagemani factor molecule assessed. Wehave recentlyreported studies describing the purification and character-ization of Hageman factor from human and rabbit plasma(1). The molecule has been labeled with 'I for studies ofits activation (2). When negatively charged particlessuch as kaolin were added, the '"I-labeled Hageman fac-tor became bound to the particles and was activated so asto generate the conversion of prekallikrein to kallikreinand to promote clotting of Hageman-deficient plasma. Bycontrast, when Hageman factor was exposed to an en-

2736 The Journal of Clinical Investigation Volume 51 October 1972

zyme such as trypsin, the molecule was cleaved into threeequal-sized fragments of 30,000 daltons. These fragments,possessing enzymatic activity (3), were termed prekalli-krein activator (PKA) .'

In accordance with these findings, the activation ofpurified Hageman factor by immune reactants was as-sessed by (a) its ability to activate isolated prekallikreinor promote clotting of Hageman-deficient plasma; (b) itsbinding to immune reactants; and (c) its cleavage into30,000 molecular weight subunits after interaction withimmune reactants.

METHODS

Preparation of protein reagents. Hageman factor wasprepared from human and rabbit plasma by methods re-cently described in this laboratory (1). The Hageman fac-tor was employed at a concentration of 50-100 ,g/ml.Briefly, plasma of each species was fractionated with am-monium sulfate. The fraction soluble at 25%o and insolubleat 55% saturation was then applied to a column of DEAESephadex A-50. After elution, the Hageman factor-richfractions were passed successively over a second column ofDEAE Sephadex A-50, G-200 Sephadex and CMSephadexC-50. At no time did the plasma fractions touch glass. TheHageman factor thus obtained was free from apparent con-taminating proteins as judged by electrophoresis in poly-acrylamide gels. The Hageman factor did not contain func-tional activities of other members of the kinin-forming orintrinsic clotting systems (1) and contained undetectablequantities of plasminogen and plasmin. It would not acti-vate prekallikrein or significantly promote clotting of factorXII-deficient plasma without prior activation with eitherkaolin or trypsin. This indicated the Hageman factor wasin precursor form. The purified Hageman factor was labeledwith 'I according to published methods (4). After dialysisthe 'I-labeled Hageman factor was passed over SephadexG-75 to eliminate traces of free 'I. Evidence that the 'Ilabel was bound to Hageman factor was as follows: thelabeled protein eluted from G-200 Sephadex with the samevolume of elution as unlabeled Hageman factor, sedimentedat 4.5S in gradients of sucrose and, upon electrophoresisin polyacrylamide gel, migrated in the identical region whereHageman factor activity was detected.

Prekallikreini. Prekallikrein of rabbit plasma was pre-pared as published previously (5). In general, prekallikreinthat was passed over two columns of DEAE SephadexA-50 was employed. This reagent (12.5 ,ug/ml) was stable,readily activated, and contained no detectable inhibitor ofmembers of the kinin-forming system. Further purificationon CM Sephadex C-50 and by electrophoresis in Pevikon(5) did not change the results.

Immunologic reactantts. Rabbit antibodies to human IgG(HGG), bovine IgG (BGG), and bovine albumin (BSA)were obtained after multiple injections of the proteins insaline. The antisera, containing 500-1000 ,ug antibody N/ml,were fractionated first with ammonium sulfate at 50%saturation, and then by DEAE-cellulose chromatography

1Abbreviations used in this paper: anti-BGG, rabbit anti-body to bovine IgG; anti-BSA, rabbit antibody to bovinealbumin; anti-HGG, rabbit antibody to human IgG; PKA,prekallikrein activator; TBS, tris-buffered saline; BAEe,benzoyl-l-arginine ethyl ester.

to obtain the 'y2-fraction. The quantity of antibody wasmeasured by precipitin reaction. The pools of antibody weredivided into portions and kept frozen. This was essential toavoid contamination with bacteria. To prepare immune com-plexes, antigen was added to the antibody at equivalence.After incubation at 37°C 30 min and at 4°C 30 min, theresulting precipitate was washed three times with cold 0.01M Tris (hydroxymethyl) aminomethane pH 7.4 in 0.15 MNaCl (TBS). It was then resuspended in the same bufferand used for activation of Hageman factor. When testswere performed in an excess of antigen or antibody, or whenthe immune reactants were brought together in the presenceof Hageman factor, the precipitation step was omitted. Zy-mosan was obtained from ICN Nutritional Biochemicals,Cleveland, Ohio. The zymosan was heated at 100°C 30 minand then washed. Antibodies were obtained by injectingrabbits with 5 mg zymosan in incomplete Freund's adj u-vant at monthly intervals. The resulting antibody wasfractionated as noted above for anti-HGG and labeled with'I to allow quantitative measurements of the amountbound to zymosan.

Aggregationt of iinzmnziunoglobuilinis anid mnyeloma proteinis.Myeloma proteins were isolated from the serum of pa-tients with multiple myeloma either by DEAE-cellulosechromatography or Pevikon block electrophoresis. yA- and-yD-myeloma proteins were further purified by SephadexG-200 gel filtration. The -yM-macroglobulins were isolated byeuglobulin precipitation followed by gel filtration on Sepha-dex G-200 columns. The purified myeloma proteins andnormal yG-immunoglobulin were then aggregated by heatat 63°C for 20 min or with bis-diazotized benzidine ac-cording to the method of Ishizaka, Ishizaka, Salmon, andFudenberg (6).

Activation of Hagemant factor. Preparations of anti-bodies and antigens, aggregated myeloma proteins, andkaolin, etc., were added to Hageman factor. The Hagemanfactor employed was most frequently that in the peak tubeseluting from G-200 Sephadex, i.e., from the fourth step inpurification. In the case of human Hageman factor, theconcentration of Hageman factor in these tubes was 240-500 ,ug/ml as determined by quantitative immune diffusionassay. The Hageman factor was readily activated by 25 ,gkaolin or 1-2 ,ug trypsin. Volumes were brought to 0.15 orin some experiments 0.20 ml with TBS. Hageman factorwas incubated with immune reactants for 20 min at either37 or 2?0C. Prekallikrein, 1.2 ,ug or at times 3.5 gtg in0.1 ml, was added and the incubation continued at 37°C 20min. Following this, 3.0 ml 1 mmBAEe was added andthe mixture incubated at 37°C for 30 min. Hydrolysis ofBAEe was then measured by absorption of UV light at253 nm (7). The readings were corrected for the amount ofabsorption at 253 nm given by the protein present as mea-sured at 280 nm. This was generally 70-90%o of the readingat 280 nm. The tubes were then warmed to 37°C for asecond incubation of an additional 1 or 2 hr. The increasein absorption at 253 nm between the two periods was usedas a measure of the degree of activation. Assays of clotpromotion were performed by adding treated or controlHageman factor in 0.1-0.05 ml Hageman-deficient plasmain plastic tubes. After incubation for 6 min at 37°C, thetubes were brought to room temperature and 0.05 ml cepha-lin (Sigma Chemicals, St. Louis, Mo.) and 0.05 ml CaCl2,0.05 M were added to the preparations. The interval inhundredths of minutes between addition of Ca++ and thetime required for a solid clot to form at room temperaturewas recorded.

The Interaction of Hageman Factor and Immune Complexes 2737

TABLE IInhibition of Absorption of Human Hageman Factor to Kaolin

by Plasma*

Hageman-deficient plasma Hagemani Hageman factor

Ilagemani (diluted in factor activity on kaolinfactor 0.1 ml) Kaolin remaining$ (O.D. 2.53 nm)

tnl Mg

0.1 Undiluted 125 100 0.0400.1 1/2 125 94 0.0530.1 1/4 125 84 0.1120.1 1/16 125 50 0.2010.1 - 125 0 0.4270.1 Undiluted 125 0.000

* Isolated Hageman factor was incubated with dilutions ofHageman (XlI)-deficient plasma 37°C 10 min. The kaolinwas removed by sedimenitation and washed before assayingfor bound Hageman factor (PKA) capacity.t Determined by quantitative immune difftision assay on thesuipernate after sedimentationi of kaolill.

Chromnatography anid iltracec trifulgationt were performedas published previously (1).

Antibody to humnan Hageciian factor was prepared by im-munizinig rabbits with Hageman factor obtained after foursteps of purification, i.e, after fractionation by gel filtrationoni G-200 Sephadex (1). Anitibodies to contaminant pro-teins were removed by absorption wNith small quantities ofHageman-deficient plasma. The resulting antibody reactedwith normal plasma to yield a single precipitin band inagar diffusion plates. No precipitate occurred with Hage-man-deficient plasma from three individuals. The antibodyreacted strongly with highly purified human Hagemanfactor obtained from clhromatography oni CM SephadexC-50 (1).

Incutbationi of immtunte precipitates with serum or plasma.Assays of active componlen1ts of the kimin-formning system intelhiates of the precipitates. Rabbit anti-BGG, anti-HGG,and anti-BSA, 6 or 12 mg antibody protein were mixedwith equivalent amounts of antigen in a plastic 40 ml tube,incubated at 37°C, 30 min, cooled to 4°C for 30 min, andcentrifuged. The precipitates were washed three times inTBS, and mixed thoroughly with 10 ml fresh rabbit serum/6 mg antibody, or, in other experiments, guinea pig serum.The sera were obtained using siliconized needles and plasticcontainers. In two experiments, plasma obtained in acidcitrate dextrose anticoagulant was substituted for serum.Incubation of precipitates and serum (or plasma) followedwith agitation 10 min at 37°C. The precipitate was thenremoved by sedimentation, and treated with 5 ml of 0.5 MNaCl in 0.1 M Tris pH 8.0. The elution with 0.5 M NaClwas repeated twice, the first two eluates combined and di-alyzed against an excess of TBS 18 hr. The dialyzed eluatewas then assayed for its capacity to activate prekallikreinand reduce the clotting time of plasma. Plasma deficientin Hageman factor was generously contributed by Mrs.Gunda Hiatt and plasmas deficient in factors XI and IXby Dr. Samuel Rappaport. The precipitates, after elution,were washed with TBS. Control tubes included (a) im-mune precipitate with TBS instead of serum or plasma,(b) serum or plasma without immune precipitates (i.e.,exposed only to the plastic tube), (c) serum or plasma

mixed with 5 mg kaolin, and (d) precipitates exposed toserum that was washed once with 20 ml TBS before elutionwith 0.5 M NaCl. In some experiments, "2I-labeled Hage-man factor was added to the serum or plasma to follow themolecule through various procedures.

RESULTS

Comparison of the capacity of sculipurificd anid nioni-purified Hagemtan, factor to be activated. It was essentialto determiine the comparative sensitivity of activation ofHageman factor in whole plasma and after isolation.Studies of the activation of Hageman factor should beperformed on the most readily activated preparation.Accordingly, isolated human Hageman factor was ob-tained following the first four steps in purification, i.e.,after elution from Sephadex G-200 (1). The tube coni-taining the greatest concentration of Hageman factor andthe least number of contaminating proteins was selecte(l.The Hageman factor so prepared was then tested foractivation alone, or niixed with various concentrationsof plasma obtained from an individual genetically tlefi-cient in Hageman factor. In order to activate and absorbthe Hageman factor. 125 iAg kaolin was added to eachpreparation in a constant final volume. After incubationi,the kaolin Nas sedimented, washed, and assayed for pre-kallikrein-activating (PKA) capacity. The stllernatewas teste(d by quantitative immtlne diffusion for residualHageman factor protein. The results are slhown in TableI. They reveal that increasing concentrations of Hage-man-deficient plasma interfere with the capacitv of Hage-man factor to become bound to and activated by thekaolin. Similar inhibition of binding of 'I-labeled Hage-man factor to kaolin by normal human plasma was ob-served in other experiments. The possib-ility was con-sidered that inhibitors of PKA in plasma coul(d accouint

TABLE I ILack of A4ctivation of Isolated Hageman Factor

by Immune Precipitate*

PrekallikreinHageman activation

Anti-HGG HGG factor (O.D. 253 nm)

Ag P ?1

625 38 0.1 0.(19625 250 0.1 0.045625 500 0.1 0.025625 250 - 0.023625 250 + kaolin 0.1 0.205

(25 Mg)* Precipitates of human IgG and rabbit antibody exposed toisolated rabbit Hageman factor. After incubation at 22 °C15 min, isolated rabbit prekallikrein (1.25 MAg) added, andincubation continued 30 min at 37°C. BAEe (3 ml) thenadded and hydrolysis meastured after 1 hr at 370C at o.n.253 nm.

2738 Cochrane, Wuepper, Aikin, Revak, and Spiegelberg

for the decreasing activity of Hageman factor on thekaolin. To test this possibility, kaolin was first exposedto isolated Hageman factor and then washed and mixedwith various dilutions of Hageman factor-deficient plasmaat 370 C 15 min. Partial (30%) inhibition of PKA ac-tivity was noted in preparations exposed to undilutedplasma, but considerably less occurred with plasma di-luted one-half. It should be noted, however, that completeinhibition of absorption onto kaolin was not observedwhen larger quantities of kaolin were used (as will benoted in Table V).

In the subsequent studies, isolated Hageman factor wasemployed that was obtained from Sephadex G-200, i.e.,the fourth step of purification. This course was preferableto that of the final step in purification (CM SephadexC-50) because of greater concentrations of the Hagemanfactor and because of a lessened tendency to activatespontaneously.

Tests for activation of isolkted Hageman factor by im-mitune precipitates. Immune precipitates consisting ofrabbit antibodies to BSA, BGG, and HGGand their re-spective antigens were prepared at equivalence. Afterwashing in TBS they were exposed to rabbit Hagemanfactor and assays were performed of possible activation.Little or no specific activation of Hageman factor wasobserved in over 25 such tests usinlg quantities of anti-body ranging from 0.03 to 6 mg. A typical experiment isshown in Table II.

The possibility was considered that activation mightoccur if antibody and antigen were combined in the pres-ence of Hageman factor. To test this, anti-HGG (200 iAg)was added to Hageman factor and the equivalent quan-tity of antigen added. Again, activation failed to occur.In addition, immune complexes prepared with 625 Agantibody and different quantities of antigen, varying fromantibody excess through equivalence into antigen excess(fivefold excess of antigen above that needed at equiva-lence) were also tested for their capacity to activateHageman factor. In all cases, using anti-BSA or anti-human IgG, activation of Hageman factor did not takeplace. In each experiment, addition of kaolin to the im-mune precipitates either during or after incubation withHageman factor brought about activation of the Hage-man factor. Both promotion of clotting and activation ofprekallikrein were observed.

When the immune precipitates or heat aggregateday-globulin and Hageman factor were incubated andadded to factor XII-deficient plasma, specific shorteningof the clotting time upon recalcification was not ob-served. Addition of kaolin to the immune precipitatesand Hageman factor induced marked shortening of theclotting tinie (Table III).

In order to find if fresh serumi was important in theactivation of Hageman factor, fresh normal rabbit or

TABLE II IThe Effect of Aggregated -y-Globulin and Kaolin on the Clot-

Promoting Activity of HumanHageman Factor

Hageman Clotting BAEeActivator* factor timet hydrolysis§

pg 0.3 ml minKaolin

25 + 5.98 0.13150 + 4.80 0.272

100 + 2.86 max

Heat aggregated HGG250 + 8.72 0.089500 + 9.20 0.088

1000 + 9.10 0.102+ 8.05 0.078

* Potential activator was added to purified human Hagemanfactor and sufficient tris-buffered saline (pH 7.5) to yield atotal volume of 0.5 ml. Incubation was carried out at 22°C20 min. Similar results were obtained with incubation at 37°C.I0.1 ml incubated mixture added to 0.05 ml cephalin and0.05 ml factor XII-deficient plasma. After 6 min at 37°C0.05 ml, 0.05 MCaCI2 added at room temperature. The clottingtime was recorded when a firm clot formed. The average of twotimes is given.§ 0.1 ml incubated mixture added to 0.1 ml prekallikrein(3.7 ,ug) and incubated 37°C 30 min. BAEe then added andthe increase in BAgenerated between 60 and 120 mim recorded.max, maximal hydrolysis at 60 min.

human serum was added to the Hageman factor andimmune precipitates. After incubation at 370C 15 minand washing of the precipitates, prekallikrein was addedas before. Again, there was no activation of the Hage-man factor as determined by conversion of prekallikrein.Washing of the precipitates after exposure to serum orplasma was necessary to eliminate inhibitors of kalli-krein.

A particulate antigen was also tested. Zymosan (500mug), zymosan with antibody (14.4-20.9 Ag antibody/500 gg zymosan), or zymosan with antibody exposedto fresh rabbit serum (one-fifth dilution) was employed.The particles exposed to serum were washed in TBSbefore adding Hageman factor. Evidence that the zymo-san exposed to antibody and then fresh rabbit serumhad bound C3 was demonstrated by immune adherenceof rabbit platelets to these particles. Platelets did notadhere to zymosan, or zymosan with antibody alone.Exposure of each preparation of zymosan to rabbitHageman factor revealed no specific activation of theHageman factor. There was no shortening of the clot-ting time of factor XII-deficient plasma and no con-version of prekallikrein to its active form. Again, 50Ag kaolin incorporated in the mixture, induced activa-

The Interaction of Hageman Factor and Immune Complexes 2739

10

5 0.5000.200

o.ioQC

Of

IgGI

-l1gG2 1 1g63

Clot PrometioI - 1-I

-

Prokalikroio Activatiola A A A A A A A - A A

AM=Myeloma Protein Plus Kaolin, 25igFIGURE 1 Tests for activation of humain Hageman factorby aggregated myeloma proteins. The first two columns ineach class represent different myeloma proteins aggregatedby bis-diazotized benzidine. The third column shows proteinsaggregated by heat (630C, 20 min). 250 ,ug of aggregatedmyeloma protein was used. Activation of prekallikrein wasmeasured by the ability of the activated enzyme to cleaveBAEe. 25 Ag kaolin (triangles) was mixed with the aggre-gated myeloma protein to test possible inhibition of activa-tion by the proteins.

tion of the Hageman factor to the same extent as whenthe kaolin was added alone to the Hageman factor.

Tests of the activationl of Hagemiian factor zwith, aggre-gated putrified humiiian mjiycloi ma proteinis. 'Myeloma pro-teins, aggregated by either bis-diazotized benzidine orheat, were added to isolated human Hageman factor.The formation of PKA and clot-promoting propertieswere tested. The results are given in Figs. 1 and 2.They demonstrate that neither PKA nor clot-promotingactivity was generated by aggregates of immnnunoglobu-lins IgGl,2,3,4, IgA, IgD, and IgMI. A single-aggregatedIgE myeloma protein also failed to activate Hagemanfactor. The results in Figs. 1 and 2 represent duplicateand, in most cases, quadruplicate determinations. Lowerdoses of aggregated proteins, i.e., 125 yg, also failed toactivate Hageman factor. Each determination shownrepresents an experiment with a different myelomaprotein of that class or subclass. The proteins in thefirst column were heavily aggregated wvith bis-diazo-tized benzidine so that a fine precipitate developed,while those in the second column were lightly aggre-gated. Ultracentrifugation of the latter proteins in gra-dients of sucrose revealed the presence of aggregates.In order to find if inhibitors existed in the prepara-tions of myeloma proteins that would prevent detectionof activation, kaolin was added in minimal quantity(25 ,ug) to the combination of Hageman factor andeach myeloma protein. As noted in Figs. 1 and 2, clot-promoting and PKA activity were readily detected. Theclotting times and rate of activation of prekallikrein

were the same as those when kaolin was added to theHageman factor alone.

Some clot-promoting activity was demonstrated withtwo IgG4 and one of three IgA proteins when they wereadded to Hageman factor although no PKA activitywas detected with these preparations. All three proteinshad a fast 8-electrophoretic mobility and could not bepurified free of other ,-globulins. Therefore, tests wereperformed in order to determine if these proteins pro-mote clot formation in the absence of Hageman factor.The results showed indeed spontaneous clot-promotingactivitv in both IgG4 and the single positive IgA pro-tein. It wvas therefore not possible to attribute clot-promoting activity to activation of Hageman factor.

In order to find if the aggregated myeloma proteinsexhibited other biologic activities, the proteins weretested for their ability to utilize complement. 62 Ag oftwo different heat-aggregated myeloma proteins of eachclass and subclass was added to 0.25 ml fresh humanserum. After incubation at 4°C 16 hr, the residualhemolvtic-complement activity was determined. IgGi,IgG3, aind IgM reduced hemolvtic levels 28, 54, and26%' , respectively. The other aggregated myeloma pro-teins failed to reduce hemolytic activity.

Tests for activation of Hageman factor with twodlifferent cryoglobulins were performed using 50, 100,200, 500, and 1000 ,ug of the cryoglobulins. One w-asan IgM\-IgG-mixed cryoglobulin and the other an IgG-IgG cryoglobulin. The cryoglobulins were allowed towarm from 0 to 37°C in the presence of Hagemanfactor over a 30 min period. The proteins were therebysolubilizedl gradually. The single examples in threeexperinments shown in Fig. 2 were typical, showingneither gelneration of PKA nor clot-promoting activi-ties. Two otlher

1-2-3.

5.6-

E-

C._

10-

E 0.5O00.200-

o 0.1000

a. 0.000

IgA

IgAM-IgG-mixed cryoglobulins were

IgoClot

.A A

Pro allikru

I 1gMPromotion

I Ia

n Act vati on

Cryoglobulin Cryoglobulin

IA

J_E-

pwZuiwi.i.3.iipi;I I

-=Myelo,a Protoil Plus Kaolin, 25jig

FIGURE 2 Same as Fig. 1 except IgMI-IgG and IgG-IgGcryoglobulinis were also tested. 250 ,ug of each protein wasassayed. The cryoprecipitate was allowed to warm anddissolve in the presence of Hageman factor at 370C.

2740 Cochrane, Wuepper, Aikin, Revak, and Spiegelberg

1.2-3-4-5-6-7-1.1-

1-1-

1-1-1-I-

q1.

.1L

II14111

c

S.-.2

AA

IAA

0

4*

-c 200±°200t I I I I i

12 Aio ** HGG-Anti4HGG + HF-1251X 10 --BGG-Amti4GG + HF-1251~ 6S~ b -**-Saliee + HF-1251

6^- -_ E4-

:2-

S B 4 8 12 16 20 24 28 32 36Tube Number

FIGURE 3 Sedimentation of 'mI-labeled human Hagemanfactor with soluble immune complexes of HGG-anti-HGGor BGG-anti-BGG after incubation 15 min, 25°C. The com-plexes were solubilized from washed precipitates by addi-tion of antigen equal in amount to 10 times that employedat equivalence (see text). Sedimentation in 5-20% sucrose,65,000 rpm, 5 hr.

tested in amounts of approximately 150 ,ug with similarresults. And finally, an IgM paraprotein, IgMLay, ob-tained from a patient with Waldenstrom's macroglobu-linemia (kindly donated by Dr. John Johnson, Depart-ment of Rheumatology, Scripps Clinic and ResearchFoundation, La Jolla, Calif.) (8) shown to bind butnot to precipitate IgG, was tested for its ability toactivate Hageman factor. 250 and 500 yg in 0.1 mlcombined with 50 and 200 lAg respectively of heat-aggregated human IgG failed to activate either clot-promoting or PKA activity. In contrast, when 50 ugkaolin was added to the mixture of three proteins in aseparate tube as a control, the Hageman factor wasactivated. The results with IgML.y were similar tothose presented in Fig. 2.

On three occasions during the course of these studies,aggregated immunoglobulin preparations were foundthat activated Hageman factor. These preparations hadbeen held at 4°C for 3 wk to several months beforeuse, and were found upon culture to contain Pseudo-monas alcaligenes in two preparations and Escherichiacoli in the other. When cultures of the bacteria weretested by themselves, they were found to be potentactivators of Hageman factor.

Tests of binding of radiolabeled Hagernan factor tosoluble immuine complexes. Hageman factor from rab-bit and human plasma was labeled with 'I and testedfor its ability to bind to immune complexes, both solubleand insoluble. Soluble complexes of rabbit antibodies tohuman IgG and bovine IgG containing 250 lAg antibodywere prepared and incubated with 'I-labeled Hageman

factor. Similarly, isolated rheumatoid factor (latex ag-glutination titer of 512) was mixed at 37°C with 'I-labeled human Hageman factor and 100 Ag heat-aggre-gated human IgG. After 15 min incubation, the mix-ture was ultracentrifuged in a gradient of sucrose (5-20%). Typical patterns of centrifugation are shown inFigs. 3 and 4. As noted, the immune complexes orrheumatoid factor-aggregated IgG complexes sedi-mented more rapidly than the 'I-labeled Hagemanfactor. The pattern of Hageman factor in the presenceof immune complexes was indistinguishable from thatin the absence of immune complexes. In addition, noevidence of breakdown of Hageman factor into its 2.6Ssubunits was observed. The addition of fresh rabbitplasma or human serum to the incubation mixture(diluted one-fifth in the protein solutions) did not alterthe pattern of sedimentation of the Hageman factor. Itshould be emphasized that the 'I-labeled Hageman fac-tor was fully capable of binding to and being activatedby kaolin. When tested before sedimentation, the Hage-man factor was found not to be activated by the im-mune complexes.

Attempts to elute active produtcts of Hageman factorfrom immuine precipitates exposed to normal plasmaor serum. Precipitates of rabbit anti-BSA, BGG, andHGGwere prepared with equivalent amounts of anti-gen, exposed to fresh serum or plasma, and eluted with0.5 M NaCl as described in Methods. The eluates weredialyzed and tested in various quantities for theirability to activate prekallikrein and promote clotting ofplasma. Five experiments were performed, yieldingsimilar data.

The results of one such experiment are shown inTable IV. As noted, active Hageman factor did notappear in the eluate as tested by prekallikrein activa-tion, while eluates of kaolin treated with normal serum

bemstsiatid Fater

x 20--i

-2 'Iem. FactsorAgg IIg15 + NF-1251

1 4 3 12 1'6 20 24Tube Number

FIGURE 4 Sedimentation of rheumatoid factor-aggregatedhuman IgG with 'I-labeled human Hageman factor afterincubation. Conditions of incubation and sedimentation werethe same as those in Fig. 3.

The Interaction of Hageman Factor and Immune Complexes 2741

TABLE IV Human or rabbit Hageman factor labeled with 12ILack of Ac/tivation of Hageman Factor in Normal Serum by was not bound by the immune precipitates exposed to

Immune Precipitates each respective serum (Table V) while up to 10% ofthe "2I-labeled Hageman factor bound to kaolin under

Conversion of prekallikrein to kallikrein identical conditions. There are several factors of note

Hydrolysis of BAEe in the experiment summarized in Table V. (a) TheEluate* Prekallikrein (O.D. 253 nm)T amount of Hageman factor in the washed precipitate

ml lAg was considerably reduced after a second wash, thereElucate of normal rabbit seruLm immune precipitate being less than 0.5% of the total Hageman factor then

0.25 1.25 0.036 remainiing. The amount bound to kaolin was not reduced0.5( 1.25 0.008 more than 1% of the total by the second wash. (b)(.50 - 0.054 After initial sedimentation and(I careful decanting of the

Elua.lte of TI3S ± immunlltlle precipitalte supernatanit seruim, 12-15%', of the '"I-labeled Ilagemiialn0.2 5 1.25 0.032 factor was left in the )lastic tuhe containing the pre-0.50) 1.25 0).0)02 cipitate. Over half of tlis was bound to the tube, for

0.50 - 0.052 wvhen the precipitates were waslhedl and decanted intoa fresh plastic tube, 6-8% of the original Hageman

Eluate of normal rabbit seruLm + immtune precipitate after factor wastil detected in the tube.nTh resalwashin factor was still detected in the tube. The residualwashing Hageman factor resisted removal bvwashing in buf-

0.25 1.25 0.064 fered saline, but was niot active in tllat it failed to con-

0.50 0.096 Vert prekallikreini to kallikrein (although it would aftertreatment with typsiin). (c) The immune precipitates,

Vitialte o)f nlormaltll rabhi serumlzl ± ka<lolin wvith 2-3%c of the Hageman factor remainiing, as show0.25 1.25 0.208 in Table V, did not activate prekallikrein. (d) This0.25 0.042 experiment was also performiied at 25 and 4°C. Less

Hageman factor was bound to kaolin at 4°C, but whatPromotion of clotting in plassma did bind was inhibited less by the serum than that at

Clotting time (min)§ 25 or 370C. Maximal binding occurred at 25-37°C witlhless inhibition of the active Hageman factor being

uXII-deficienit XI-deficient IX-deficient observed at 250C. The amount of Hageman factor inEluateofplasma plasma plasma the immune precipitates was constant at each tempera-

NRS+ immunie ture. (e) As noted in Table I, the presence of plasmaprecipitate 16.3 3.13 7.6 proteins inhibits uptake of Hageman factor by activator.

TBS + immune This finding was confirmed in the studies recorded inprecipitate 26.6 5.24 27 Table V in whiclh significantly more Hageman factor

NRS+ immunie was bound to kaolin when the serum vwas diluted. Asprecipitate 4.22 25 noted, this was reflected in greater loss of Hageman

-wash (X 1 3>35 6.36 >25 factor fronm the serum as determined by quantitativeimmune assay. To test if the "I-labeled Hageman factor

* Normal rabbit serum (10 ml) incubated with BGG-anti- in the supernatant serum consisted at least in part ofBGG precipitate or kaolin. Centrifuged precipitate then the 30,000 mol. wt. cleavage product, i.e., PKA, theeluted with 0.5 MNaCl in tris buffer, pH 8.0. Eluate dialyzed supernatant serum was subjected to ultracentrifugationagainst TBS before use. in a gradient of sucrose. The "«I activity sedimenteclI Prekallikrein incubated 1 hr with BAEe. No significant dif- in a sharp peak at 4.5S, i.e., in the same position asference in results noted after 2 additional hr incubation. native Hageman factor. This indicated that Hageman§ See Methods. Hfactorh by incuato wtha immun

factor lhad not beeni cleaved bv ilncubation wvith iimmnuiie

exhibited active Haffeman factor. Although there was p)recip'itates.clot-promoting activity in the dialyzed eluates of un-washed precipitates, this was not eluted from the pre-cipitates if they were washed once in TBS before elu-tion with 0.5 M NaCl. In addition, the clot-promotingmaterial shortened the clotting times of not only factorXII-deficient plasma, but factor XI- and IX-deficientplasmas as well.

DISCUSSION

The experiments presented offer no evidlence that Hage-man factor interacts with immune complexes and failto support the possibility that immune aggregates are

capable of activating Hageman factor: (a) radiolabeledHageman factor was not bound by immune aggregates

2742 Cochrane, Wuepper, Aikin, Revak, and Spiegelberg

while it was by a known activator, kaolin; (b) Hage-man factor was not cleaved into its 30.000 mol. wt. sub-units, the prekallikrein activators as occurs upon acti-vation with enzymes; (c) complexes of antigen andantibody either in precipitate or soluble form failed toactivate isolated Hageman factor so that it would acti-vate prekallikrein or promote clotting of factor XII-deficient plasma; (d) a particulate antigen, zymosan,together with antibody, failed to induce activation; (e)aggregated human myeloma proteins of all knownclasses and subclasses failed to activate Hageman fac-tor. This eliminated the possibility that a single classor subclass of immunoglobulin could be effective thatmay have been deficient in the antibodies used for test-ing. To test whether the physical form of the aggregatewas important in activating Hageman factor, the im-munoglobulins were aggregated by both heat (63°C for20 min) and bis-diazotized benzidine so as to obtainsmall and large (visible) aggregates. No effect wasnoted with each type of aggregated immunoglobulin.

The addition of plasma or serum to immune aggre-gates did not lead to activation of Hageman factor. Infact, serum or plasma inhibited the uptake of Hage-man factor into a known activator, kaolin (Table I),and in separate studies was found to block PKA ac-tivity when the Hageman factor was activated by kaolin.Immune precipitates or soluble complexes did not bindHageman factor in the presence of serum. Zymosan,used as antigen in order to provide a surface for acti-vation, also did not activate Hageman factor whencombined with antibody and fresh normal serum. Inaddition, serum after absorption with immune precipi-tates did not shown a diminution of Hageman factor.

These studies were facilitated by the availability ofMI-labeled human and rabbit Hageman factor and by

antibodies directed to the former. These reagents al-lowed quantitative measurements to be performed onthe amount of Hageman factor participating in reac-tions whether or not it was activated. This is especiallyimportant in the present studies in which activation byimmune complexes was not observed. All of the Hage-man factor could be accounted for at the end of thereaction, which substantiated considerably the results.In addition, it was determined at the outset of theseexperiments that Hageman factor was far more readilyactivated after partial purification. Thus, the ability todetect activation was particularly sensitive.

The present studies are in keeping with those ofRatnoff (9) and Robbins and Stetson (10) in whichimmune complexes failed to promote clotting of platelet-poor plasma. By contrast, several studies have beenreported with results that differ from these.

Davies, Lowe, and Holman (11-13) and Movat, Di-Lorenzo, and Treloar (14-16) have reported that when

TABLE VInability of Immune Precipitates to Bind

in Normal HumanSerum*Hageman Factor

HF activity onBinding of HF to HF remaining Ag-Ab precipitateAg-Ab precipitate NHS in supernate or kaolin

or kaolin (dilution) (% of total)$ (O.D. 253 nm)

Ag-Ab precipitate2.9 undiluted 0.0002.8 1/2 99 0.000

Kaolin6.1 undiluted 0.030

10.0 1/2 91 0.113

* Normal human serum (NHS) containing 1251-labeledHageman factor was mixed with BSA-anti-BSA precipitate(500 ug Ab) or kaolin (500 ,ug) 37°C 15 min. The precipitatesor kaolin were washed once and then transferred to fresh plastictubes for counting 1211. The results recorded are typical ofthree separate experiments.I Measured by quantitative immune diffusion assay. NHSwithout addition of Ag-Ab precipitate control (100%).

serum was incubated with immune precipitates, sub-stances were generated that increased vascular perme-ability. Sera of guinea pig, rabbit, and man were em-ployed although serum of man generated little or noactivity (15). Similar conclusions were reached inexperiments of Eisen and Smith (17). The dataraised the possibility that the kinin system in serumwas activated by the immune precipitates. Supportingthis interpretation were the findings that the permeabil-ity occurred in guinea pigs treated with the antihista-minic mepyraminic maleate (5 mg/kg), the activity wasenhanced if the serum was pretreated with inhibitorsof serum carboxypeptidase known to possess kininaseactivity, generation of the permeability factor was in-hibited by various agents known to block formationof kinin, and serum deficient in Hageman factor wasinactive. Contrary to this interpretation, however, wasthe finding that EDTA (5 x 10-2 M) inhibited genera-tion of the permeability factor (15). The activation ofHageman factor and the kinin system is unaffected bythe presence of EDTA at this concentration.

While several alternative hypotheses might be con-sidered, one that these workers have taken pains toeliminate must still be kept in mind; namely, that themajority of permeability results from the generation ofanaphylatoxin through activation of the complementsystem. Several observations support this conclusion:(a) Inhibitors of the complement system preventedactivation of the permeability factor: chelation of theserum with EDTA, heating or addition of salicylaldox-inme to the serum before addition of immune precipi-

The Interaction of Hageman Factor and Immune Complexes 2743

tates. It was difficult to generate the permeability factorfrom citrated plasma. (b) Permeability was enhancedby treating the serum with agents that block serumcarboxypeptidase. This would prevent degradation ofanaphylatoxin as well as kinins and would thereforeprotect both activities. Indeed, it is known that ana-phylatoxin cannot be measured in human serum afterincubation with immune precipitates which conforms tothe observation of the authors that generation of per-meability factor was not observed in human serum.(c) Adequacy of treatment of the test animals withmepyramine maleate was assessed by intradermial in-jections of histamine (1 /,g) rather than by quantitiesof anaphylatoxin shown to y ield reactions of perme-ability equal to those in the tests. It is known thathistamine released endogenously is more difficult tosuppress with antihistaminics than histamine adminis-tered exogenously.

Further analysis of the serum after treatment withimmune precipitates allowedl separation of several per-meability factors by differing physical properties (18).In addition, elution of unwashed precipitates followedby chromatographic analysis allowed identification ofseveral factors capable of activating members of thekinin system (18, 19). However, it is quite possiblethat activation of the eluates obtained fronm immuneprecipitates resulted from the process of dialvsis andchromatography. In fact, this has been shown to occurwhen proper precautions are not taken (3, 5, 20). Inthe present studies, abundant Hageman factor, i.e.about 12-15% of that added in normal serum, remainedin the plastic tubes after serumiwas decantedI from theimmune precipitates. One-half of this was bound tothe tube and essentially all of the remainder could bewashed from the precipitate with physiologic saline.The Hageman factor bound to the plastic tube andthat removed from precipitate by washing was in-active. Thus elution of the serum treated, unwashedimmune precipitate with high concentrations of NaCl(18) would yield abundant Hageman factor that couldbe activated during further treatment.

Kaplan, Spragg, Austen, and Gigli (21, 22) havereported that IgMLay, obtained from a patient withWaldenstroms macroglobulinemia. when reacted withIgG, activated isolated Hageman factor. In the studiesreported here, several IgM proteins having rheumatoidfactor activity including IgMLay, failed to convert iso-lated Hageman factor to its active form. When smallamounts of kaolin were added with these reactants, nor-mal activation was observed indicating that inhibitorsof the activation were not present in the IgM\ or IgGproteins.

Epstein, Tan, and Melmon (23) have presented well-controlled studies showing that immunoglobulins with

rheumatoid factor activity reacted with IgG to releasekinin from plasma that contracted rat uterus and de-pleted levels of kininogen. The amount of kinin releasedparalleled the titer of rheumatoid factor. It should benoted that these investigators employed chelated plasmawhich did not support generation of the vascular per-meability factor studied by MIovat and DiLorenzo (15).The two factors are therefore probably different. Hage-man-deficient plasma failed to support generation of thekinin. In the present studies, isolated Hageman factorwas neither bound to nor activated by rheumatoid fac-tor-aggregated IgG complexes. Addition of serum orplasma had no effect oIn the results. In addition, rheu-matoi(I factor of high titer (1,000-10,000 reciprocaldilution) isolated chromatographically. was combinedwith heat-aggregated IgG. This was then added toplasma pretreated with 1-10-orthophenanthroline underconditions identical with those of Epstein et al. (23).Nevertheless, with seven different preparations of rheu-matoid factor we failed to obtain either specific kininrelease or kininogen depletion. Addition of 200 Ag kao-lin/nml plasma in the presence of each rheumatoid fac-tor and heat-aggregated IgG generated kinin anddepleted kininogen levels up to 40%ff6. The reason forthis discrepancy is not apparent.

Actiz'ation by bactcria. In the only instances inwvhich aggregated immunoglobulins, immune complexes,or isolated rheumatoid factors were capable of generat-ing activity fronm the precursor Hageman factor, bac-terial contamination wvas found in the preparation ofimimiuinoglobulin. The bacteria, upon culture and isola-tion, were capable of activating the Hageman factor.In two cases, Pscudomona.s alcaligenes and in a third,E. coli were identified. This finding may illustrate amechanism of activation of the kinin-forming and in-trinsic clotting systems of considerable importance inthe host's defense against invading organisms. In addi-tion, it represents a pitfall to investigations of the acti-vationi of the kinin-forming system than can misleadproper interpretation.

ACKNOWLEDGMENTS

This study is publication No. 580 from the Department ofExperimental Pathology, Scripps Clinic and ResearchFoundation, La Jolla, Calif.

The work was supported by U. S. Public Health ServiceGrant A1-07007, the National Multiple Sclerosis SocietyGrant 459, the Council for Tobacco Research, U. S. A.,and the American Heart Association.

REFERENCES1. Cochrane, C. G., and K. D. Wuepper. 1971. The first

component of the kinin-forming system in human andrabbit plasma. It relationship to clotting factor XIT(Hageman Factor). J. Exp. Med. 134: 986.

2 Generously contributed by Dr. Nathan Zvaifler.

2744 Cochrane, Wuiepper, Aikin, Revak, and Spiegelberg

2. Cochrane, C. G., S. S. Sitzer, and B. S. Aikin. 1972.Hageman Factor: its structure and activation. Fed.Proc. 31: 623.

3. Wuepper, K. D., E. S. Tucker III, and C. G. Coch-rane. 1970. Plasma kinin system: proenzyme components.J. Immunol. 105: 1307.

4. McConahey, P. J., and F. J. Dixon. 1966. A methodof trace iodination of proteins for immunologic studies.Int. Arch. Allergy Appi. Immunol. 29: 185.

5. Wuepper, K. D., and C. G. Cochrane. 1972. Plasma pro-kallikrein: isolation, characterization, and mechanism ofactivation. J. Exp. Med. 135: 1.

6. Ishizaka, T., K. Ishizaka, S. Salmon, and H. Fuden-berg. 1967. Biologic activities of aggregated y-globulin.VIII. Aggregated immunoglobulins of different classes.J. Immunlol. 99: 82.

7. Schwert, G. W., and Y. Takenaka. 1955. A spectro-photometric determination of trypsin and chymotrypsin.Biochim. Biophys. Acta. 16: 570.

8. Stone, M. J., and H. Metzger. 1968. The specificity of amonoclonal macroglobulin (-yM) antibody: reactivitywith primate yG immunoglobulins. J. Immunzol. 102: 222.

9. Ratnoff, 0. D. 1969. Some relationships among hemo-stasis, fibrinolytic phenomena, immunity, and the in-flammatory response. Adv. Immunol. 10: 188.

10. Robbins, J., and C. A. Stetson, Jr. 1959. An effect ofantigen-antibody interaction on blood coagulation. J.Exp. Med. 109: 1.

11. Davies, G. E., and J. S. Lowe. 1960. A permeabilityfactor released from guinea-pig serum by antigen-anti-body precipitates. Br. J. Exp. Pathol. 41: 335.

12. Davies, G. E., and J. S. Lowe. 1962. Further studieson a permeability factor released from guinea-pig serumby antigen-antibody precipitates: relationship to serumcomplement. Int. Arch. Allergy Appl. Immunol. 20: 235.

13. Davies, G. E., G. Holman, and J. S. Lowe. 1967. Roleof Hageman factor in the activation of guinea-pig pre-kallikrein. Br. J. Pharmacol. 29: 55.

14. Movat, H. Z. 1967. Activation of the kinin system byantigen-antibody complexes. In International symposium

on Vaso-active Polypeptides: Bradykinin and RelatedKinins. M. Rocha e Silva and H. A. Rothschild, editors.Edart Livraria Editora, Ltda., Sao Paulo, Brazil. 177.

15. Movat, H. Z., and N. L. DiLorenzo. 1968. Activationof the plasma kinin system by antigen-antibody aggre-gates. I. Generation of permeability factor in guineapig serum. Lab. Invest. 19: 187.

16. Movat, H. Z., N. L. DiLorenzo, and M. P. Treloar.1968. Activation of the plasma kinin system by antigen-antibody aggregates. II. Isolation of permeability-en-hancing and kinin-releasing fractions from activatedguinea pig serum. Lab. Invest. 19: 201.

17. Eisen, V., and H. G. Smith. 1970. Plasma kinin forma-tion by complexes of aggregated 'y-globulin and serumproteins. Br. J. Exp. Pathol. 51: 328.

18. Movat, H. Z., M. P. Treloar, and Y. Takeuchi. 1969. Asmall molecular weight permeability factor in guineapig serum: adsorption to antigen-antibody aggregates.J. Immunol. 103: 875.

19. Treloar, M. P., and H. Z. Movat. 1970. Isolation oftwo small molecular weight activators of plasma kinin-system in the guinea pig. Fed. Proc. 29: 576a.

20. Wuepper, K. D., and C. G. Cochrane. 1972. Isolationand mechanism of activation of components of theplasma kinin-forming system. Int Biochemistry of theAcute Allergic Reaction. K. F. Austen and E. L.Becker, editors. Blackwell Scientific Publications Ltd.,Oxford. In press.

21. Kaplan, A. P., J. Spragg, and K. F. Austen. 1972. Theimmunologic activation of the bradykinin-forming systemin man. In Biochemistry of the Acute Allergic Response.K. F. Austen and E. L. Becker, editors. Blackwell Sci-entific Publications Ltd., Oxford. In press.

22. Kaplan, A. P., I. Gigli, and K. F. Austen. 1971. Im-munologic activation of Hageman factor and its rela-tionship to fibrinolysis, bradykinin generation and com-plement. J. Clint. Invest. 50: 51a. (Abstr.)

23. Epstein, W. V., M. Tan, and K. L. Melmon. 1969.Rheumatoid factor and kinin generation. Ann. N. Y.Acad. Sci. 168: 173.

The Interaction of Hageman Factor and Immune Complexes 2745