Journal of Clinical InvestigationVol. 45, No. 4, 1966

Characterization of the Y-Globulin Complexes Present inCertain Sera Having High Titers of Anti-y-globulin

Activity *RALPH E. SCHROHENLOHER t

(From the Department of Medicine, University of Alabama Medical Center,Birmingham, Ala.)

It is well established that the principal anti-y-globulin factors in the sera of patients with rheu-matoid arthritis detected by such commonly em-ployed procedures as the sensitized sheep cellagglutination and latex fixation tests are immuno-globulins of the yM (19 S y-globulin) (2) class(3-6). These anti-y-globulins, often referred toas rheumatoid factors, are not limited in occurrenceto rheumatoid arthritis, but are sometimes foundin a variety of other disease states (7-10). Ex-tensive evidence is now available in support of theconcept that these factors are indeed antibodies tovarious antigenic determinants on the yG (7 Sy-globulin) (2) molecule. This evidence has beenreviewed by Kunkel (11, 12).

Certain sera having high titers of anti-y-globulinactivity contain significant quantities of materialssedimenting at a rate intermediate to the 7 and19 S components (13, 14). Such sera have beenencountered in patients having rheumatoid arthritisor a number of other disorders (13-15). Thesematerials were shown to be complexes of yG thatare readily dissociated in acid buffer or urea (14).It seemed possible that the complexes might repre-sent antigen-antibody complexes of the y-globulin-anti-y-globulin type. This possibility was sup-ported by the subsequent demonstration of a lowmolecular weight anti-y-globulin activity in such

* Submitted for publication August 19, 1965; acceptedDecember 16, 1965.A preliminary report of this work was presented at the

annual meeting of the American Rheumatism Association,Philadelphia, Pa., June 18, 1965, and appeared in abstractform (1).

This work was supported by grants TI AM 5000and AM 03555 from the National Institute of Arthritisand Metabolic Diseases, U. S. Public Health Service,and Merck and Company, Rahway, N. J.t Address requests for reprints to Dr. Ralph E. Schro-

henloher, Dept. of Medicine, University of Alabama Medi-cal Center, 1919 7th Ave., S., Birmingham, Ala. 35233.

sera (16). This activity was dearly distinguishedfrom -yM activity in that it was not sensitive tomercaptoethanol treatment. A cryoglobulin of theyG class that precipitated with aggregated y-globu-lin has also been described (17).The present study was undertaken in an effort

to further characterize the y-globulins responsiblefor complex formation. This followed an initialobservation that fractions containing intermediatecomplexes precipitated aggregated y-globulin.These fractions appeared to be free of yM anti--y-globulins. The evidence obtained clearly supportsthe concept that the active y-globulins of the com-plexes are in fact antibodies to yG. The techniqueof fragmentation with pepsin (18, 19) was par-ticularly useful in this respect and in defining thespecificity of these anti-y-globulins.

MethodsThe intermediate complex preparations were isolated

from the sera of three patients (Do, Od, and Go)."These sera had high titers in the various reactions forthe detection of rheumatoid factors and contained moder-ate or large amounts of intermediate 'y-globulin complexes.All three patients had rheumatoid arthritis. Euglobulinprecipitates were prepared and fractionated on SephadexG-200,2 using pH 4.1 sodium acetate buffer, 0.1 mole perL, as previously described (20). Column fractions con-taining the dissociated intermediate complexes werepooled, dialyzed against pH 7.5 sodium phosphate buffer,0.2 ionic strength, and concentrated by ultrafiltration to aprotein concentration of approximately 10 mg per ml.

Reduction by ethyl mercaptan and alkylation by iodo-acetamide were by the procedures previously described(20). Samples that were reduced by mercaptoethanol

1 Sera for this study were kindly provided by Dr.Henry G. Kunkel, The Rockefeller Institute, New York,N. Y. (patient Do), Dr. Israeli A. Jaffe, New York Med-ical College, New York, N. Y. (patient Od), and Dr.Stephan E. Ritzmann, the University of Texas, Galveston,Texas, and Dr. Richard T. Thurm, U. S. Public HealthService Hospital, Boston, Mass. (patient Go).

2 Pharmacia Fine Chemicals, Inc., New Market, N. J.

501

RALPH E. SCHROHENLOHER

were made 0.1 M concentrations with respect to 2-mer-captoethanol and allowed to stand at room temperaturefor 3 hours or were dialyzed against 0.1 M mercaptoetha-nol in pH 7.5 buffer for 16 hours. At the end of the re-duction period all samples were dialyzed against 0.02 Miodoacetamide in pH 7.5 buffer and then buffer.The preparations containing intermediate complexes

were digested by pepsin at pH 4.1 according to the pro-cedure of Nisonoff, Wissler, Lipman, and Woernley (18).The ratio of substrate to enzyme employed was 100: 1.Digestion' mixtures were incubated for 16 to 18 hoursat 370 C. Digestion was terminated by dialysis againstpH 7.5 sodium phosphate buffer, 0.2 ionic strength, inthe cold. The fragments obtained by this procedure willbe referred to as F(ab')2-fragments.

Ultracentrifugal analysis was carried out in a Spincomodel E ultracentrifuge with double sector 12-mm cells.Plate measurements were made according to the proce-dures described by Schachman (21) or Markham (22).The sedimentation coefficients were corrected to theSaow values by the accepted procedures. The partialspecific volumes were assumed to be 0.73.

Density gradient ultracentrifugation was performed ac-cording to the general methods described by Kunkel (23).Continuous gradients of 10 to 40% sucrose prepared ineither pH 7.5 sodium phosphate buffer, 0.2 ionic strength,or pH 3 glycine-saline buffer, 0.1 ionic strength, wereutilized. Centrifugation was performed by using anSW-39 rotor in the Spinco model E ultracentrifuge at33,450 rpm for 16 hours. After centrifugation serialfractions were collected from the bottom of the tube.The activities of fractions from either the pH 3 or 7.5gradients could be determined without prior dialysis.The buffers employed as diluents in the slide precipitinor latex fixation tests were sufficient to completely neu-tralize the relatively weak pH 3 buffer. Also there wasno indication that the sucrose interfered with the activitymeasurements. In some experiments pooled fractionsfrom the various levels of both neutral and acid gradientswere analyzed after dialysis against pH 7.5 buffer. Inall cases the results corresponded exactly to those ob-tained without prior dialysis.

Reactivity with aggregated '-globulin was determinedby precipitin formation in'capillary tubes (24) or by asimple glass slide technique. Aggregated y-globulin wasprepared by heating human Cohn Fraction II 3 at 630 Cfor 10 minutes. The aggregates were isolated by prepara-tive! ultracentrifugation and resuspended in saline or pH7.5 buffer. A concentration of about 2.5 mg per mlwas used for capillary tests. A concentration of 1 mgper ml was found satisfactory for the slide test. Theslide test was performed by placing one drop eachof sample, pH 7.5 buffer, and aggregated y-globulin solu-tion in the well of a serological slide and mixing on arotary agitator for 15 minutes. Precipitation was evalu-ated with the aid of low magnification. This test was

3 The Cohn Fraction II preparations utilized in thisstudy were the products of Mann Research Laboratories,New York, N. Y., or Lederle Laboratories, Pearl River,N. Y.

particularly useful for gradient fractions. The resultsof either test were graded on a zero to four plus scaleand were found to correspond closely.The optimal concentrations of samples and aggregated

'y-globulin for preparative scale precipitates were esti-mated by using the slide test, and they varied somewhatfrom one sample to another. Precipitates of the inter-mediate complex preparations from patients Do and Odwere prepared by mixing equal volumes of sample (3 mgper ml in pH 7.5 buffer), aggregated '-globulin'(6 mgper ml in saline), and pH 7.5 buffer. The precipitate ofthe complex preparation from patient Go was preparedby mixing equal volumes of sample (5 mg per ml in pH7.5 buffer), aggregated '-globulin (2.5 mg per ml in sa-line), and pH 7.5 buffer. Precipitates of the F (ab') 2-fragments of the complex preparations were preparedby mixing equal volumes of sample (approximately 6mg per ml in pH 7.5 buffer), aggregrated 'y-globulin(7.5 to 10 mg per ml in saline), and pH 7.5 buffer. Themixtures were incubated at room temperature with oc-casional agitation for 1 hour then in the cold overnight.In all instances the precipitate began to form within afew minutes after mixing. No precipitate formed incontrols prepared with buffer in place of the sample.The precipitates were washed twice by large volumes ofbuffered saline. The precipitates were finally suspendedin a small volume pH 3 glycine-saline buffer, 0.1 ionicstrength, and then dialyzed against the same buffer.The dialysis was necessary to effect complete solution,since pH 7.5 buffer trapped in the precipitate partiallyneutralized the small volume of relatively weak pH 3buffer.The latex fixation test was performed according to the

procedure of Singer and Plotz (25). Agglutination testswith sensitized sheep cells or human group 0 Rh-positivecells coated with Ripley incomplete anti-Rh antibodies 4were performed by the usual procedures (26, 19). Pre-cipitating antisera specific for yG were produced by hy-perimmunizing rabbits with human Cohn Fraction II incomplete Freund's adjuvant and subsequently absorbingthe antisera with F (ab') 2-fragments of Cohn FractionII. Precipitating antisera specific for 'yM were producedby hyperimmunizing rabbits with electrophoretically iso-lated Waldenstrom's macroglobulins in complete Freund'sadjuvant and subsequently absorbing the antisera withhuman Cohn Fraction II. Precipitating antisera specificfor 'yA prepared in the goat were commercial products.5Ouchterlony agar diffusion studies (27) were performedin 2.0%o agar in pH 8.6 barbital buffer, 0.05 ionic strength.Protein concentrations were estimated by the Folin-Ciocalteu method (28) with Cohn Fraction II as thestandard.

Results

Isolation and properties. The euglobulin pre-cipitates from the various sera contained moderate

4 Serum from patient Ripley was kindly provided byDr. Marion Waller, Medical College of Virginia, Rich-mond, Va.

5 Mann Research Laboratories, New York, N. Y.

502

CHARACTERIZATION OF y-GLOBULIN COMPLEXES

or large amounts of intermediate complexes in ad-dition to the usual 7 and 19 S materials. Separa-tion was accomplished by gel filtration on Sepha-dex G-200 with pH 4.1 sodium acetate buffer.The fractionation of a euglobulin preparation frompatient Go is shown in Figure 1. Under theseconditions, the intermediate complexes were dis-sociated into their 7 S components and recoveredwith the 7 S materials in the second peak from thecolumn. The materials contained in this peak,neutralized and concentrated, were utilized in thisstudy without further purification. The 19 S ma-terials, including the yM anti-y-globulins, wererecovered in the first peak. The elution patternobtained with the euglobulin preparations from Pa-tient Od were very similar to those from Go. Theeuglobulins from Patient Do differed in that asmaller quantity of 19 S materials was recovered.The elution patterns did not vary in an appreciablemanner for different euglobulin preparations fromthe same patient or serum sample.Examples of the ultracentrifugal patterns of the

intermediate complex preparations, in pH 7.5 buf-fer, from the three patients are shown in Figure 2.Three preparations from patient Do were exam-ined and gave essentially identical patterns. Alldemonstrated a small 7 S peak, a large broad 16 Speak, and a small amount of faster sedimentingmaterials separated poorly from the 16 S peak.

0.30 AZO;

Z 0.20 3w0

-J

00.10

20 30 40 50

TUBE NUMBER

FIG. 1. FRACTIONATION OF SERUM EUGLOBULINS FROMPATIENT GO BY GEL FILTRATION THROUGH SEPHADEXG-200. The column was eluted with 0.1 M sodium acetatebuffer, pH 4.1. The intermediate complexes were re-covered from the second peak. The 19 S components,including the yM anti-y-globulins, were recovered fromthe first.

FIG. 2. ULTRACENTRIFUGAL PATTERNS OF INTERMEDIATECOMPLEX PREPARATIONS FROM PATIENTS Do (TOP), OD(MIDDLE), AND Go (BOTTOM) AS ISOLATED FROM THE SE-RUM EUGLOBULINS BY GEL FILTRATION AT PH 4.1. Theanalyses were performed in pH 7.5 sodium phosphatebuffer, 0.2 ionic strength. The patterns were recordedafter 32 minutes at 56,100 rpm (top and middle) andafter 40 minutes at 52,640 rpm (bottom). Sedimentationproceeds from left to right. The patterns show large(Do) or moderate (Od and Go) amounts of materialssedimenting faster than 7 S.

The relative amounts of materials sedimenting at16 S or faster were 83, 85, and 86%. Two prepa-rations from Go and one from Od were examined.All demonstrated larger 7 S peaks than the Dopreparations. The relative amounts of complexesin the Go preparations were 41 and 62%. About50%o of the Od preparation sedimented as com-plexes. The various preparations dissociatedcompletely in pH 4 sodium acetate buffer, 0.1mole per L, and sedimented as a single symmetri-cal peak at 6 S. The sedimentation rate for human-yG (Cohn Fraction II) in the pH 4 buffer wasthe same as that observed for the dissociated com-plexes. The absence of fast sedimenting compo-nents after acid dissociation demonstrated that thepreparations were relatively free of yM, which doesnot dissociate at low pH and continues to sedimentrapidly (14).

Ouchterlony agar diffusion indicated that, inaddition to yG, the intermediate complex prepara-tions contained at most trace amounts of -yM and

503

RALPH E. SCHROHENLOHER

TABLE I

Activities of preparations containing intermediatecomplexes isolated from euglobulins by

gel filtration

Precipita-tion with Latex

Prepa- aggregated fixationPatient ration y-globulin* titert

Do I ++ 1:640II +++ 1:320

III +++ 1:320IV +++ 1:1,280

Od I +++ 1:320II ++++ 1:640

III --+++ 1:320

Go I ++ 1:160II ++++ 1:640

III ++++ 1:160IV +++ 1:160

* The strengths of precipitation with aggregated -y-globu-lin by the samples at protein concentrations of 10 mg perml in the capillary tube or slide tests are listed. Multipledeterminations were performed on most preparations andwere within 1 U of the listed values.

t The titers for samples having protein concentrationsof 10 mg per ml are listed. Multiple determinations wereperformed on many preparations and demonstrated amaximal variation of two tubes of a serial twofold dilution.

small amounts of -yA. Two preparations frompatient Do, two from Go, and one from Od wereeach tested with one anti--yA, two anti-yM, andtwo anti-yG antisera. Most tests were performedat least twice, with complete agreement in all in-stances. When tested at 10 mg per ml, each prepa-ration formed strong precipitin lines with only theanti-yG antisera. No lines were formed by theDo preparations with the other antisera. The anti-

TABLE II

Effect of ethyl mercaptan reduction on the activities of inter-mediate complex preparations (Do and Go) and an

isolated yM anti-y-globulin (Jo)

Precipita-tion withaggregated

Sample Description Latex fixation* -y-globulint

Ag/mlDo Untreated 31 +++

Reduced 31 +++Go Untreated 16 ++++

Reduced 16 +++

Jo Untreated 0.9 ++++Reduced Negativeat 115 0

* The protein concentrations of the maximal dilutionsgiving a one plus agglutination are recorded.

t The strengths of precipitation of the undiluted sampleswith aggregated -y-globulin in the capillary tube test arelisted. The protein concentrations were 10 mg per mleach for Do and Go and 2.3 mg per ml for Jo.

yA antiserum formed weak lines and the anti-yMantisera formed very weak lines with the Go andOd preparations. No lines were detected withthese antisera when the preparations were dilutedto 1 mg per ml. The anti-yG antisera continuedto give strong precipitin lines with the dilutedpreparations.The activities of preparations from the three

patients are listed in Table I. The various prepa-rations from all three patients gave moderatelystrong or strong precipitates with aggregatedy-globulin when tested at 10 mg per ml concentra-tion. This precipitin activity could be detectedat concentrations of 1 mg per ml or less. Substi-tution of saline for either the sample or the aggre-

TABLE III

Effect of mercaptoethanol reduction on the activities of inter-mediate complex preparations (Do and Od) and a typical

rheumatoid arthritis (R.A.) serum

Precipita-tion withaggregated

Sample Description Latex fixation* -y-globulint

pg/mlDo Untreated 8 +

Reduced, 3 hours 6 +

Untreated 16 +Reduced, 16 hours 31 +

Od Untreated 13 +Reduced, 3 hours 10 +

Untreated 13 ++Reduced, 16 hours 10 ++

R.A. serum Untreated 1:2,560 ++Reduced, 3 hours Negative at 1:20 0

* The protein concentrations of the maximal dilutions giving a oneplus agglutination are listed for the complex preparations.

t The strengths of precipitation with aggregated y-globulin in theslide test of the Do and Od samples at protein concentrations of 1 mgper ml and the R.A. serum samples at a dilution of 1:10 are listed.

gated y-globulin did not result in precipitation.The preparations also agglutinated latex particlescoated with Fraction II. The agglutinations wereusually strong, being comparable to moderatelyreactive rheumatoid sera. Multiple determinationswere performed on many of the preparations andvaried by no more than two tubes of a serial two-fold dilution. Two preparations from each of thethree patients were tested and found not to agglu-tinate sensitized sheep cells. Several preparationsfrom Do and Go were tested with human cellscoated with incomplete anti-Rh antibody and foundinactive towards this detection system, even athigh protein concentrations. The single prepara-tion from Od tested with the coated human cellswas weakly postive at 10 mg per ml but negative

504

CHARACTERIZATION OF rGLOBULIN COMPLEXES

0.

0.

U)

zw

a

-J

0

O.

.6

.4 oo0 11fi ;' o ooooo ooo a

.2 . nli011 11--4 %

2/

0.6 1f10.4ooomo*0 ,Do0o~0.2

C ~~ ~~~I"

4+ AGGREGATED2+ y- GLOBULIN

ACTIVITY0

320 LATEXO FIXATION8 TITER

20

4+ AGGREGATEDy-GLOBULIN

2+ ACTIVITY

0

320 LATEX

80 FIXATIONTITER

20

5 10 15 20TUBE NUMBER

FIG. 3. DENSITY GRADIENT ULTRACENTRIFUGAL EXPERI-MENTS SHOWING THE DISTRIBUTION OF ANTI-y-GLOBULINACTIVITY IN AN INTERMEDIATE COMPLEX PREPARATIONFROM PATIENT Do (UPPER SECTION) AND AN ACID-DISSOCI-ATED SAMPLE OF THE SAME PREPARATION (LOWER SECTION).A pH 7.5 gradient was used for the experiment in theupper section and a pH 3 gradient in the lower. Thegradient fractions are numbered from the bottom. Ag-gregated 'y-globulin precipitation and latex fixation areindicated. The activity of the intermediate complex prep-aration was associated with the faster sedimenting ma-terials. After dissociation of the complexes, the activitywas associated with the 7 S components.

at greater dilutions. Since the yM anti-y-globulinsisolated from the same serum samples were highlyactive toward the coated human cells," it was notlikely that the activities detected in the complexpreparations were due to the presence of smallamounts of yM anti-y-globulins. This conclusionwas supported by the other findings of this study.The activities of several samples each of prepara-

tions from Do and Od and one from Go were de-termined after ethyl mercaptan or mercaptoethanoltreatment. The results are listed in Tables II andIII. In contrast to controls containing yM anti-y-globulin activity, no significant losses of activitywere observed for the intermediate complex prep-arations. Numerous studies have shown that theactivity of yM anti-y-globulins, as measured by

sThe yM anti-'y-globulins from Do and Od were thesubject of a previous investigation (20). The titer ofthe yM anti-'y-globulin from Go (2.8 mg per ml) withhuman red cells coated with Ripley anti-Rh antibody was1: 20,480.

most procedures, is lost after dissociation underconditions similar or identical to those used above(3, 6, 16, 20).The distributions of the anti-y-globulin activities

among the components of the preparations wereexamined by density gradient ultracentrifugation.Analysis of an intermediate complex preparationfrom patient Do on a pH 7.5 gradient is illustratedin the upper section of Figure 3. Both the ag-gregated y-globulin precipitation and latex fixa-tion activities were associated principally with the

......,II.

.°:......:-,. k a ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...,..... .:.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..°^~* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.

FIG. 4. ULTRACENTRIFUGAL ANALYSES OF THE MA-TERIALS PRECIPITATED BY AGGREGATED -v-GLOBULIN FROMINTERMEDIATE COMPLEX PREPARATIONS FROM PATIENTS Do(TOP), OD (MIDDLE), AND Go (BOTTOM) SHOWING THEPRESENCE OF 7 S AND THE ABSENCE OF 19 S MATERIALS.The precipitates were dissolved in pH 3 glycine-salinebuffer. The patterns on the left were recorded immedi-ately after acceleration to 52,640 rpm, and each shows abroad fast sedimenting peak due to the aggregates and aslow sedimenting peak near the meniscus. The patternson the right were recorded 32 minutes later and each showsa single 6.5 S peak. Sedimentation proceeds from leftto right.

505

RALPH E. SCHROHENLOHER

z

-J

0.0

0.2

0.11

n n '4 AGGREGATED:2+ -y-GLOBULIN

oooooooooooo U 0 o ACTIVITY

320 LATEX~~FIXATION

n nnnntrn8TITER

A 21

5 10 15TUBE NUMBER

20

FIG. 5. DENSITY GRADIENT ULTRACENTRIFUGAL EXPERI-MENT SHOWING THE DISTRIBUTIONS OF AGGREGATED 8y-GLOB-ULIN PRECIPITATION AND LATEX FIXATION ACTIVITY IN THEMATERIALS PRECIPITATED BY AGGREGATED -y-GLOBULIN FROMAN INTERMEDIATE COMPLEX PREPARATION FROM PATIENTGo. The precipitate was dissolved in pH 3 glycine-sa-line buffer and the gradient prepared in the same buffer.Gradient fractions are numbered from the bottom. Theresults demonstrate that the slow sedimenting materialsin the precipitate were active.

faster sedimenting materials (i.e., the complexes).An acid-dissociated sample of the same preparationwas fractionated simultaneously on a gradientprepared in pH 3 buffer. The results are illus-trated in the lower section of Figure 3. Dissocia-tion of the complexes into their 7 S componentswas accompanied by a corresponding decrease inthe sedimentation of the activity. These observa-tions are representative of three such studies onpreparations from this patient. Comparable re-sults were also obtained with preparations fromthe other two patients. No yM activity was de-tected in any preparation examined by this method.Control experiments, with isolated yM anti-y-globulins or rheumatoid sera, demonstrated thatthis class of anti-y-globulins would have been re-covered from the bottom (tubes 2 through 8) ofeither gradient. These experiments demonstratedclearly that the activities were associated with theacid-dissociable complexes.

Studies on aggregated y-globulin precipitates.The active components contained in the intermedi-ate complex preparations were further character-ized by ultracentrifugal analysis of the precipitatesformed with aggregated y-globulin. Before analy-sis, the precipitates were washed and dissolved inpH 3 buffer. Two preparations from patient Doand one each from Od and Go were examined bythis technic. Examples of the resulting patternsare shown in Figure 4. Each demonstrated arapidly sedimenting broad boundary due to theaggregated y-globulin and a single component

sedimenting at 7 S. No 19 S materials were de-tected in any of the samples analyzed.The 7 S materials contained in dissolved pre-

cipitates were shown to be active by density gradi-ent ultracentrifugal studies with pH 3 gradients.The results with a precipitate of a Go preparationare illustrated in Figure 5. The anti-y-globulinactivity was clearly associated with the 7 S com-ponents. Similar results were obtained with pre-cipitates from the other two patients. The 7 Smaterials were also isolated from dissolved pre-cipitates of intermediate complex preparationsfrom patient Do by either preparatory ultracentri-fugation or by gel filtration on Sephadex G-200with the pH 3 buffer. The isolated materials,after dialysis into pH 7.5 buffer, formed inter-mediate complexes and precipitated with aggre-gated y-globulin. The ultracentrifugal patternwas quite similar to that of the preparation beforeprecipitation.

Properties of the F(ab')2-fragments. Anti-bodies of the yG class have been shown to becleaved by pepsin, in the absence of mercaptans,into bivalent fragments (18, 19). These frag-ments, termed F(ab')9-fragments, sediment in theultrac'entrifuge at 5 S. If the y-globulins responsi-ble for the complexes are antibodies of the yG

FIG. 6. ULTRACENTRIFUGAL EXPERIMENTS ILLUSTRATINGTHE ABSENCE OF COMPLEXES AFTER PEPTIC DIGESTION OF ANINTERMEDIATE COMPLEX PREPARATION FROM PATIENT DoAND THE FORMATION OF COMPLEXES BY THE RESULTINGF (AB') 2-FRAGMENTS WITH HUMAN -yG. Upper pattern:the pepsin-digest; middle pattern: 'yG control; lower pat-tern: after addition of yG to the pepsin-digest. The pat-terns were recorded after 64 minutes at 52,640 rpm.Sedimentation proceeds from left to right. All sampleswere in pH 7.5 sodium phosphate buffer, 0.2 ionicstrength. The concentration of the intermediate complexpreparation before digestion was 10 mg per ml. That ofthe yG in the control and mixture was 6 mg per ml.

506

CHARACTERIZATION OF rGLOBULIN COMPLEXES

class, it would be reasonable to expect theirF(ab')2-fragments to be active. Samples of thevarious intermediate complex preparations weretherefore subjected to this system of degradationand the resulting fragments characterized.Examples of the ultracentrifugal patterns of the

pepsin-digested complex preparations in pH 7.5buffer are shown in Figures 6 and 7. A minimumof two such preparations from each of the threepatients were examined by this technic. The char-acteristic fast sedimenting complexes were ab-sent after peptic digestion. Also, no 7 S materialswere detected. The preparations contained largequantities of F (ab') 2-fragments sedimenting at5.2 to 5.5 S. Small amounts of slower sedimentingcomponents were also present. The effectivenessof the fragmentation procedure was also indicatedby Ouchterlony agar diffusion studies on a repre-sentative preparation from each of the three pa-tients. By using an antiserum to human FractionII, the pepsin split complex preparations gave theexpected reaction of partial identity with the un-treated preparation (19).The activities of the F(ab')2-fragments and the

corresponding untreated intermediate complexpreparations are listed in Table IV. All the pep-sin-digests examined (six from Do, four from Od,and two from Go) continued to precipitatestrongly with aggregated y-globulin. Two of thedigests from each patient were tested for latexfixation activity. Although the titers were com-parable to the original material, the strengths ofagglutination were sometimes less. This was es-pecially apparent for some of the Go and Od frag-ment preparations. Conversion of the F(ab')2-fragments to the univalent fragments by reductionwith 0.01 M cysteine and alkylation with 0.02 Miodoacetamide resulted in the expected loss of ac-tivity and decrease in sedimentation rate to 3.5 S.The F(ab')2-fragments did not agglutinate humanred cells coated with incomplete anti-Rh antibody.

Peptic digestion of the yG molecule under theconditions utilized in this study results in the de-struction of part of the molecule (18, 19). Forhuman -yG, this part is approximately equivalentto the fast- or Fc-fragment produced by papaincleavage. The absence of complexes in the inter-mediate complex preparations after peptic diges-tion, even though the resulting fragments demon-strated anti-y-globulin activity, indicated that the

FIG. 7. ULTRACENTRIFUGAL EXPERIMENTS ILLUSTRATINGTHE ABSENCE OF COMPLEXES AFTER PEPTIC DIGESTION OF ANINTERMEDIATE COMPLEX PREPARATION FROM OD AND THE

FORMATION OF COMPLEXES BY THE RESULTING F(AB' 2-FRAGMENTS WITH HUMAN yG. a = the pepsin-digest;b = yG control; c = after addition of yG to the pepsin-digest; and d = the mixture after acid dissociation. Thepatterns were rec6rded after 48 minutes at 56,100 rpm.Sedimentation proceeds from left to right. Samples a, b,and c were in pH 7.5 sodium phosphate, 0.2 ionic strength.Sample d was in 0.1 M sodium acetate, pH 4.1. The con-centration of the intermediate complex preparation beforedigestion was 10 mg per ml. The concentration of theyG in the control and mixture was 10 mg per ml.

activity of these preparations is specific for groupsin the part of the molecule destroyed by the pepsin.Further evidence for this was obtained by the for-mation of fast sedimenting complexes with addedhuman -yG (Fraction II y-globulin). The resultsof such an experiment with a fragment preparationfrom patient Do are shown in Figure 6. A largeamount of complexes sedimenting at 8.6 S and asmall amount sedimenting at a faster rate werepresent in the mixture. This was accompanied bya corresponding decrease in the quantity of 5 Sfragments. The complexes were dissociated into 5and 6 S components after dialysis into 0.1 M so-dium acetate buffer, pH 4.1. Essentially identical

507

RALPH E. SCHROHENLOHER

TABLE IV

Activity of F(ab')2-fragments and corresponding untreated intermediate complex preparations

Precipitation with aggregatede-globulin* Latex fixation titer

Experi-Patient ment Untreated F (ab') 2-fragments Untreated F (ab') 2-fragments

Do 1 +++ +++ 1:320 1:1602 ++ +++ 1:1,280 1:6403 +++ ++++4 +++ ++5 ++ ++6 ++ +++

Od 1 +++ ++++ 1:320 1:3202 +++ ++ 1:160 1:1603 +++ ++4 ++++ ++++

Go 1 +++ +++ 1:160 1:802 +++ +++ 1:160 1:80

* The strengths of precipitation of the undiluted samples with aggregated y-globulin in the capillary or slide tests arelisted.

results were obtained with a second fragmentpreparation from this patient.The results of a similar experiment with an

F(ab')2-fragment preparation from patient Od areshown in Figure 7. Although the results were notas dramatic as with the fragments from patient Do,the formation of complexes after the addition of yGwas apparent. These sedimented at approximately9 and 11 S and were dissociated after dialysis intothe pH 4.1 buffer. The ultracentrifugal pattern ofthe dissociated mixture resolved into two compo-

.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

FIG. 8. ULTRACENTRIFUGAL ANALYSES OF THE MATER-

IALS PRECIPITATED BY AGGREGATED y-GLOBULIN FROM PEP-

SIN-DIGESTED INTERMEDIATE COMPLEX PREPARATIONS FROMPATIENTS Do (LEFT) AND OD (RIGHT) SHOWING THE

PRESENCE OF 5 S MATERIALS. The precipitates were dis-solved in pH 3 glycine-saline buffer. Sedimentation pro-ceeds from left to right. The patterns were recordedafter 32 minutes at 56,100 rpm (left) and 52,640 rpm(right). Only a single peak sedimenting at 5.4 S was de-tected in each. The aggregated y-globulin has completelysedimented from view.

nents about equal in size sedimenting at about 5and 6 S. These observations were confirmed byusing a second fragment preparation from this pa-tient. The results obtained when human yG wasadded to fragment preparations from patient Gowere similar to those shown for the fragment prep-aration from patient Od.The active components in two F(ab')2-fragment

preparations from patient Do and one from Odwere further characterized by ultracentrifugalanalysis of the materials precipitated by aggre-gated y-globulin (Figure 8). Before analysis, theprecipitates were washed and dissolved in pH 3buffer. In addition to the aggregates, only asingle 5 S component was detected. It was notpossible to examine Go fragments by this methoddue to the limited supply of serum from this pa-tient. Density~gradient ultracentrifugal studiesof the dissolved precipitates on pH 3 gradientsprovided additional evidence that the anti-y-globu-lin activity was associated with the 5 S materials.Similar results were obtained when a fragmentpreparation from patient Do was analyzed on apH 7.5 gradient.

Discussion

Preparations that demonstrated large or moder-ate amounts of intermediate complexes by ultra-centrifugal analysis were found to precipitate ag-gregated y-globulin and agglutinate Fraction II-coated latex particles. Density gradient ultra-

508

CHARACTERIZATION OF y-GLOBULIN COMPLEXES

centrifugation demonstrated that the activity wasassociated principally with the fast sedimentingmaterials in the preparations. A variety of evi-dence indicated that the activity was not due tothe presence of yM anti-y-globulins, but was infact due to a yG anti-y-globulin. The preparationswere found to be essentially free ofyM by both ana-

lytical and density gradient ultracentrifugal analy-ses after acid dissociation. Only 7 S materialswere detected. Gamma M would have been de-tected readily since, unlike intermediate complexes,they are not dissociated at low pH and continueto sediment rapidly (14). Ouchterlony agar dif-fusion demonstrated at most trace amounts of -yM.Only yG was detected in significant amounts.The yM anti-y-globulins from the sera studiedwere highly reactive towards sensitized humancells, whereas the complex preparations were not.The presence of sufficient yM anti-y-globulin togive the activities detected would have most cer-

tainly resulted in significant activity towards thesensitized cells. The activity of the complex prep-arations also differed from that of typical yM anti-y-globulins in that it was not sensitive to treatmentby mercaptans.

Direct evidence that the activity was indeed as-

sociated with the 7 S components of the complexeswas obtained by density gradient ultracentrifugalstudies of acid-dissociated preparations. Thesestudies failed to show the presence of yM (19 S)activity. Only 7 S activity was detected. Controlstudies demonstrated that yM anti-y-globulin ac-

tivity would have been easily detected by thistechnique. Similarly, analytical ultracentrifugalanalyses of the materials precipitated by aggre-gated y-globulin, dissolved in pH 3 buffer, revealedonly large quantities of 7 S materials. These ma-

terials, isolated from the dissolved precipitates,were shown to be active and to form complexes.The results of the peptic digestion studies were

consistent with those expected for a typical yG-antibody (18, 19). After digestion, the prepara-

tions sedimented at 5 S and continued to react inthe tests for anti-y-globulin activity. The activitywas shown by a number of studies to be associ-ated with the 5 S fragments and not undigestedmaterial. The absence of detectable amounts ofcomponents sedimenting faster than 5 S indicateda minimum of undigested material. The distri-bution of activity in a pepsin-digest was observed

to correspond to the 5 S materials by densitygradient ultracentrifugation. Analytical ultracen-trifugal analyses of the materials precipitated byaggregated y-globulin from pepsin-digests dem-onstrated only 5 S components. The fragmentseluted from the precipitates were shown to be ac-tive by density gradient ultracentrifugation. Theactivity of the pepsin-digests differed from thatof the untreated preparations in that it was sensi-tive to treatment by mercaptans. This sensitivityis consistent with that demonstrated for the 5 Sfragments produced from antibodies by peptic di-gestion (18). The specificity of these yG anti-y-globulins was also evident from the digestion stud-ies. The absence of complexes in the pepsin-di-gests at neutral pH indicated that the group orgroups that react with the active sites to formcomplexes were destroyed. This was further indi-cated by the formation of complexes with unalteredyG.From the results of these studies it is apparent

that the yG factors responsible for the formationof intermediate complexes are antibodies specificfor one or more groups on the part of the yG mole-cule destroyed by peptic digestion. Intermediatecomplexes are therefore formed by the binding ofyG to the active sites of yG anti-y-globulins.Some evidence that the intermediate complexesrepresent antigen-antibody complexes was alsoobtained by Kunkel, Muller-Eberhard, Fudenberg,and Tomasi (14). It was found that the additionof yG to sera or euglobulin fractions producedmarked effects on the ultracentrifugal distributionof the complexes. Those complexes sedimentingat 10 S were increased at the expense of thosesedimenting at faster rates. The effect of the ex-cess yG suggested the action of excess antigen onantigen-antibody complexes. In contrast to thepresent study, efforts by these workers to demon-strate an anti-y-globulin activity for the complexesby various serological tests were not successful.The reason for this is not apparent. However, ina later study, Chodirker and Tomasi (16) demon-strated an anti-y-globulin activity by the latexfixation test in fractions that contained intermedi-ate complexes, but were free of yM. These frac-tions were isolated from sera by ion exchange chro-matography or density gradient ultracentrifuga-tion. In contrast to control rheumatoid sera, con-siderable activity remained in sera containing

509

RALPH E. SCHROHENLOHER

complexes after mercaptoethanol treatment. Theactive components and the specificity were notcharacterized further. A cryoglobulin of the yGclass that had the properties of the intermediatecomplexes and precipitated with aggregated y-glob-ulin has also been reported (17). The activitywas not sensitive to mercaptoethanol treatment andwas shown to be associated with the 7 S com-

ponents.The classical yM anti-y-globulins also react with

antigenic determinants in the part of the yG mole-cule destroyed by peptic digestion (19, 29). Sincethis part of the molecule appears to contain sev-

eral such determinants (12), the exact relation-ships of the specificities of the two classes of anti-y-globulins will require additional studies. Thedifference in the capacity of the yM and yG anti-y-globulins from the same patient to agglutinate sen-

sitized red cells suggests that the specificities are

not identical. Since the yM anti-y-globulins inindividual sera can be heterogeneous with respectto reactivity in various test systems (30), it is pos-

sible that some molecules of corresponding speci-ficity exist. The possibility also exists that thismay reflect differences in avidity or agglutinatingefficiency. Such differences have been demon-strated for other yM and yG antibodies (31). In-hibition as a factor in the failure of the yG anti-y-globulins to agglutinate sensitized cells seems un-

likely since the F(ab')2-fragments of the prepara-

tions also failed to agglutinate sensitized humancells. The potentially inhibitory part of the yGmolecule was destroyed by the digestion procedure(19).The accumulated evidence from this and other

laboratories indicates that the anti-y-globulins re-

sponsible for intermediate complex formation are

yG counterparts to the classical yM anti-y-globulins.The existence of such counterparts constitutes ad-ditional evidence that the rheumatoid factors rep-

resent an immune response to some form of yG.The existence of both yG and yM anti-y-globulinsof similar specificities is consistent with the knownmolecular heterogeneity of antibodies. The simul-taneous presence of activity in all three majorclasses of immunoglobulins has been demonstratedfor a variety of human antibodies, including anti-nuclear (32) and antithyroglobulin (33) factors.Also pertinent in this respect is the reported iden-tification of /A anti-y-globulin factors that are

reactive with human and rabbit y-globulin (34).The existence of human -yG anti-y-globulins is alsoconsistent with the production of both 7 S and 19S anti-y-globulin factors in rabbits immunizedwith altered autologous y-globulin (35). Theproduction of both -yM and yG antibodies in re-sponse to a single antigen commonly occurs in avariety of species, including humans (36-39).The -yM antibodies have been shown to be the firstantibodies produced in response to an antigenicstimulus. The yG antibodies usually appear laterand frequently persist over a longer period of time.Often repeated or larger doses of antigen are re-quired to produce yG antibodies. The incidenceand distribution of the intermediate complexes ap-pears compatible with these experimental data con-cerning the relationships of the yG and yM anti-bodies. Although exact information is not avail-able, it is known that complexes are found in largequantities in only a limited number of patients andare usually associated with high titers of yM anti-y-globulin activity (14).A group of -yG anti-y-globulins that are specific

for antigenic determinants in human -yG revealedby peptic digestion have been described (19).These factors do not form complexes and areclearly distinguished from the anti-y-globulinsresponsible for intermediate complex formationin their specificity and serological activity. Thesefactors agglutinate red cells coated with the Fab-or F(ab')2-fragments of an incomplete anti-Rhantibody but do not react in the common tests forrheumatoid factor activity.

Summary

Preparations containing significant amounts ofy-globulin complexes of the intermediate type wereisolated from the sera of three patients with rheu-matoid arthritis. The preparations precipitatedheat-aggregated y-globulin and agglutinated latexparticles coated with Fraction II y-globulin. Evi-dence from a variety of studies demonstrated thatthe anti-y-globulin activities were not due to thepresence of small amounts of yM (19 S) anti-y-globulins, but were associated with the yG (7 S)components of the complexes.The complex preparations were degraded by

peptic digestion to fragments that sedimented at5 S. Complexes and 7 S materials were not pres-

510

CHARACTERIZATION OF y-.GLOBULIN COMPLEXES

ent after digestion. The fragments demonstratedanti-y-globulin activities similar to those of theoriginal preparations and were found to form com-plexes with unaltered human -yG.The results provide evidence that the yG factors

responsible for the formation of the intermediatecomplexes are antibodies that are specific forgroups on the part of the yG molecule destroyedby peptic digestion. This group of anti-y-globulinsappears to represent -yG counterparts to the clas-sical /M anti-y-globulins (19 S rheumatoidfactors).

Acknowledgments

The author wishes to thank Dr. Henry G. Kunkel forhis helpful advice and Mrs. Mary Nell Beard for hercapable technical assistance.

References

1. Schrohenloher, R. E. Characterization of the 'yG-globulin complexes present in certain sera havinghigh titers of anti-vy-globulin activity (abstract).Arthr. and Rheum. 1965, 8, 466.

2. Nomenclature for human immunoglobulins. Bull.Wld. Hlth. Org. 1964, 30, 447.

3. Franklin, E. C., H. G. Kunkel, H. J. Mfiller-Eber-hard, and H. R. Holman. Relation of highmolecular weight proteins to the serological reac-tions in rheumatoid arthritis. Ann. rheum. Dis.1957, 16, 315.

4. Heimer, R., 0. M. Federico, and R. H. Freyberg.Purification of a rheumatoid factor. Proc. Soc.exp. Biol. (N. Y.) 1958, 99, 381.

5. Lospalluto, J., and M. Ziff. Chromatographic stud-ies of the rheumatoid factor. J. exp. Med. 1959,110, 169.

6. Kunkel, H. G., E. C. Franklin, and H. J. MullerEberhard. Studies on the isolation and characteri-zation of the "rheumatoid factor." J. clin. Invest.1959, 38, 424.

7. Kunkel, H. G., H. J. Simon, and H. Fudenberg. Ob-servations concerning positive serologic reactionsfor rheumatoid factor in certain patients withsarcoidosis and other hyperglobulinemic states.Arthr. and Rheum. 1958, 1, 289.

8. Dresner, E., and P. Trombly. The latex-fixationreaction in nonrheumatic diseases. New Engl. J.Med. 1959, 261, 981.

9. Cathcart, E. S., R. C. Williams, Jr., H. Ross, andE. Calkins. The relationship of the latex fixationtest to the clinical and serological manifestationsof leprosy. Amer. J. Med. 1961, 31, 758.

10. Williams, R. C., Jr., and H. G. Kunkel. Rheumatoidfactor, complement, and conglutinin aberrationsin patients with subacute bacterial endocarditis.J. clin. Invest. 1962, 41, 666.

11. Kunkel, H. G. Immunological aspects of connectivetissue disorders. Fed. Proc. 1964, 23, 623.

12. Kunkel, H. G., and E. M. Tan. Autoantibodies anddisease in Advances in Immunology, F. J. Dixon,Jr., and J. H. Humphrey, Eds. New York,Academic Press, 1964, vol. 4, p. 351.

13. Kunkel, H. G., H. H. Fudenberg, and H. J. Mueller-Eberhard. Gamma globulin complexes in the seraof certain patients with rheumatoid arthritis (ab-stract). Arthr. and Rheum. 1960, 3, 264.

14. Kunkel, H. G., H. J. Muller-Eberhard, H. H. Fu-denberg, and T. B. Tomasi. Gamma globulincomplexes in rheumatoid arthritis and certainother conditions. J. clin. Invest. 1961, 40, 117.

15. Tomasi, T. B., Jr., H. H. Fudenberg, and N. Finby.Possible relationship of rheumatoid factors andpulmonary disease. Amer. J. Med. 1962, 33, 243.

16. Chodirker, W. B., and T. B. Tomasi, Jr. Low-mo-lecular-weight rheumatoid factor. J. clin. Invest.1963, 42, 876.

17. Meltzer, M., and E. C. Franklin. Interaction ofy globulins. A study of cryoglobulins with andwithout rheumatoid factor activity (abstract).Arthr. and Rheum. 1962, 5, 117.

18. Nisonoff, A., F. C. Wissler, L. N. Lipman, and D. L.Woernley. Separation of univalent fragmentsfrom the bivalent rabbit antibody molecule by re-duction of disulfide bonds. Arch. Biochem. 1960,89, 230.

19. Osterland, C. K., M. Harboe, and H. G. Kunkel.Anti-y-globulin factors in human sera revealedby enzymatic splitting of anti-Rh antibodies. VoxSang. (Basel) 1963, 80, 133.

20. Schrohenloher, R. E., H. G. Kunkel, and T. B. To-masi. Activity of dissociated and reassociated19S anti-'y-globulins. J. exp. Med. 1964, 120,1215.

21. Schachman, H. K. Ultracentrifugation in Biochem-istry. New York, Academic Press, 1959.

22. Markham, R. A graphical method for the rapiddetermination of sedimentataion coefficients. Bio-chem. J. 1960, 77, 516.

23. Kunkel, H. G. Macroglobulins and high molecularweight antibodies in The Plasma Proteins, F. W.Putnam, Ed. New York, Academic Press, 1960,vol. 1, p. 279.

24. Franklin, E. C., H. G. Kunkel, and J. R. Ward.Clinical studies of seven patients with rheumatoidarthritis and uniquely large amounts of rheuma-toid factor. Arthr. and Rheum. 1958, 1, 400.

25. Singer, J. M., and C. M. Plotz. The latex fixationtest. I. Application to the serologic diagnosis ofrheumatoid arthritis. Amer. J. Med. 1956, 21, 888.

26. Rose, H. M., C. Ragan, E. Pearce, and M. 0. Lip-man. Differential agglutination of normal andsensitized sheep erythrocytes by sera of patientswith rheumatoid arthritis. Proc. Soc. exp. Biol.(N. Y.) 1948, 68, 1.

27. Ouchterlony, 0. Antigen-antibody reactions in gels.IV. Types of reactions in coordinated systems of

511

RALPH E. SCHROHENLOHER

diffusion. Acta path. microbial. scand. 1953, 32,231.

28. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, andR. J. Randall. Protein measurement with theFolin phenol reagent. J. biol. Chem. 1951, 193, 265.

29. Franklin, E. C. Interaction of rheumatoid factorand gamma globulin in Proceedings of the 10thInternational Congress of Rheumatology, Rome1961. Minerva med. 1961, 2, 804.

30. Williams, R. C., Jr., and H. G. Kunkel. Separationof rheumatoid factors of different specificities us-ing columns conjugated with gamma-globulin.Arthr. and Rheum. 1963, 6, 665.

31. Greenbury, C. L., D. H. Moore, and L. A. C. Nunn.Reaction of 7S and 19S components of immunerabbit antisera with human group A and AB redcells. Immunology 1963, 6, 421.

32. Barnett, E. V., R. F. Bakemeier, J. P. Leddy, andJ. H. Vaughn. Heterogeneity of antinuclear fac-tors in lupus erythematosus and rheumatoid ar-thritis. Proc. Soc. exp. Biol. (N. Y.) 1965, 118,803.

33. Fahey, J. L., and H. Goodman. Antibody activityin six classes of human immunoglobulins. Science1964, 143, 588.

34. Heimer, R., and F. M. Levin. GammalA and Gamma2rheumatoid factors (abstract). Arthr. and Rheum.1964, 7, 738.

35. Williams, R. C., Jr., and H. G. Kunkel. Antibodiesto rabbit y-globulin after immunizing with variouspreparations of autologous -y-globulin. Proc. Soc.exp. Biol. (N. Y.) 1963, 112, 554.

36. Bauer, D. C., M. J. Mathies, and A. B. Stavitsky.Sequences of synthesis of -y-1 macroglobulin and'y-2 globulin antibodies during primary and sec-ondary responses to proteins, salmonella antigens,and phage. J. exp. Med. 1963, 117, 889.

37. Uhr, J. W., and M. S. Finkelstein. Antibody forma-tion IV. Formation of rapidly and slowly sedi-menting antibodies and immunological memory tobacteriophage 0X 174. J. exp. Med. 1963, 117,457.

38. LoSpalluto, J., W. E. Miller, Jr., B. Dorward, andC. W. Fink. The formation of macroglobulinantibodies. I. Studies on adult humans. J. clin.Invest. 1962, 41, 1415.

39. Fink, C. W., W. E. Miller, Jr., B. Dorward, andJ. LoSpalluto. The formation of macroglobulinantibodies. II. Studies on neonatal infants andolder children. J. clin. Invest. 1962, 41, 1422.

512