humanmonoclonal yg-cryoglobulins with anti-y...

Human Monoclonal YG-Cryoglobulins

with Anti-y-Globulin Activity

HowARDM. GREY, PETERF. KOHLER, WILLIAM D. TERRY, andEDWARDC. FRANKLIN

From the Department of Experimental Pathology, Division of Allergy,Immunology and Rheumatology, Scripps Clinic and Research Foundation,La Jolla, California 92037; the Immunology Branch of the. National CancerInstitute, Bethesda, Maryland, 20014; and the Department of Medicine,NewYork University, School of Medicine, NewYork 10016

A B S T R A C T Seven human yG-myeloma pro-teins which were also cryoglobulins were studiedwith respect to their reactivity with yG-globulinsas well as with regard to their antigenic classifica-tion within the yG-heavy chain subclasses. Fiveof the seven cryoglobulins studied were positivein at least two of the three tests used to assay foranti-y-globulin activity. One protein was onlyweakly positive in one test system and another wasnegative in all test systems. The structures whichwere recognized by the cryoglobulins were local-ized to the Fc-fragment. Only primate yG-globu-lins contained these antigenic determinants and insome cases the cryoglobulin appeared to showspecificity for one human heavy chain subclass overthe others. Antigenic analysis revealed that fourof the five cryoglobulins with definite antibodyactivity belonged to the yG3-subclass, the fifthbelonged to the yGl-subclass. The two cryoglobu-lins which reacted only weakly or failed to combinewith yG-globulins were both of the yGl-subclass.These findings taken together with the localizationof the combining site to the Fab-fragment suggeststhat many of these cryoglobulins may representantibodies to yG-globulin, and that the cryoprecipi-tate in these cases represents antigen-antibodycomplexes of such a nature that they precipitateonly in the cold.

Address requests for reprints to Dr. Howard M. Grey,Scripps Clinic and Research Foundation, 476 ProspectStreet, La Jolla, Calif. 90237.

Received for publication 11 March 1968.

INTRODUCTION

Immunoglobulins are a physically and chemicallyheterogeneous group of proteins. In some in-stances, they precipitate or gelify on cooling andsuch proteins are known as cryoglobulins. Whilecryoglobulinemia is occasionally seen as a primaryor idiopathic disease, it is most often associatedwith multiple myeloma or Waldenstrom's macro-globulinemia. In these subjects a major or minorfraction of the abnormal homogeneous proteinbehaves as a cryoglobulin. Recently it has becomeapparent that small amounts of cryoglobulins canalso occur in other diseases, most often in those in-cluded in the group of collagen diseases such assystemic lupus erythematosus or related symptomcomplexes resembling this disorder (1-3). Thefrequency of these cryoglobulins is born out bya recent study of 29 unselected cases of cryoglobu-linemia which indicated that about 40% of theproteins studied possessed rheumatoid factor activ-ity, in which 11 of 12 instances was shown to bea yM-rheumatoid factor (2, 3). In one case theanti-y-globulin activity was associated with ayG-protein. The biochemical basis for the phe-nomenon of cold precipitation of these proteins islargely unknown.

The present study was performed first, to studya group of yG-myeloma proteins which were alsocryoglobulins for the purpose of determiningwhether these yG-cryoglobulins possessed antibodyactivity to normal y-globulin. Second, since previ-ous studies on the yG-heavy chain subclasses indi-

The Journal of Clinical Investigation Volume 47 1968 1875

cated that they differ considerably in certain bio-logical and biochemical properties (4-6 and foot-note 1), it was of considerable interest to classifythe yG-cryoglobulins with respect to heavy chainsubclasses.

METHODS

Source anid preparation of proteins. yG-myelomacryoglobulins as well as other yG- and yA-myelomaproteins were obtained from sera of patients with adiagnosis of multiple myeloma or benign monoclonalgammopathy. The homogeneous nature of the cryoglobu-lins was established by starch-gel electrophoresis of thepepsin-digested proteins using a modification of Smithie'svertical gel method (7); -yM-globulin was obtained fromthe serum of a patient with Waldenstrdm's macroglobu-linemia. Human serum albumin (HSA) was obtainedfrom Cutter Laboratories, Berkeley, Calif. Heterologous-y-globulins were a gift from Dr. William Weigle. Thesewere prepared by O-diethylaminoethyl (DEAE) cellu-lose chromatography of whole serum or commerciallyobtained fraction II.

Myeloma proteins were isolated by starch-block elec-trophoresis (8) and in the case of the yA- and -yM-pro-teins, further purified by gel filtration on SephadexG-200 (9).

Myeloma cryoglobulins were prepared by cold precipi-tation of the cryoglobulins followed by five washes inlarge volumes of phosphate-buffered saline, pH 7.2.Immunoelectrophoresis (10) of the washed cryoglobulinsrevealed only yG-globulin when tested with an antiserumprepared against whole human serum.

Antisera. Antisera to heavy chain subclasses of yGwere obtained from rhesus monkeys immunized with yG1,yG2, or -yG4-myeloma proteins and rendered subclassspecific by appropriate absorption. Antiserum to yG3 wasprepared by immunizing rhesus monkeys with pooledhuman yG and absorbing the antiserum with -yG1- and-yG2-myeloma proteins.

Anti-K and X antisera were obtained from rabbits im-munized with K or. X Bence Jones proteins.

Diffusion in gel. Proteins were studied by doublediffusion in 1%7 agarose gel. Several of the cryoglobulinpreparations precipitated in the gel at room temperature,and it was therefore necessary to modify the usual pro-cedure. Antiserum was added to appropriate wells andallowed to completely diffuse into the agar while theplates were heated to 37'C in a humid chamber. All anti-gen solutions were heated at 50'C, the temperature re-quired to totally solubilize the least soluble cryoprotein.Antigens were rapidly added to wells, and the plate in-cubated at 37°C until all solutions had completely dif-fused into the agar. The plate was then submerged inheated mineral oil and incubated at 500 C. Under these

1Muller-Eberhard, H. J., M. A. Calcott, and H. M.Grey. Interaction between C'lq and 'yG-globulins of dif-ferent heavy chain subtypes. In preparation.

conditions proteins did not spontaneously precipitate inthe gel. Plates were evaluated at 6, 24, and 48 hr.

Estimates of the quantity of each subclass present inthe washed cryoprecipitates were obtained by finding thehighest dilution of each preparation which would give aprecipitin line with the subclass specific antisera. Dilu-tions of known concentrations of myeloma proteins repre-senting each of the subclasses were similarly tested withthe same antisera on the same plate. It was assumed thatthe highest dilution of the cryoglobulin that gave a pre-cipitin line contained the same quantity of a subclass asthat present in the highest dilution of the myeloma pro-tein that gave a precipitin line.

Rheumatoid factor test. The test for rheumatoid fac-tor was performed by the latex agglutination test using acommercial source (Hyland Laboratory, Los Angeles,Calif.) of gamma globulin coated latex particles. Serialdilutions of purified cryoglobulins were tested for theircapacity to agglutinate the latex particles. Results areexpressed as the lowest concentration of cryoglobulinthat gave definite agglutination.

Coprecipitation of radioiodinated proteins with cryo-globulins. Normal -y-globulin from a variety of species,human serum albumin, as well as human yA- and yM-paraproteins, were labeled with "1I (I*) by means of thechloramine T method (11). Radioisotope detection wasdone with a gamma well-type scintillation counter. Totest for binding of the labeled proteins to the cryoglobulinprecipitate, 1-ml aliquots containing 10 mg of the purifiedcryoglobulin in phosphate buffered saline, pH 7.2, weremixed with 0.1 ml of the buffer containing 0.1-1 ,ug ofthe labeled protein to be tested. The mixture was incu-bated at 37'C for 1 hr and then for 16 hr at 4°C. Thecryoglobulin-I* myeloma protein mixture was centri-fuged at 2000 rpm for 30 min in the cold; the supernatantwas decanted and the precipitate washed and resuspendedin 1 ml of cold, phosphate-buffered saline. The centrifu-gation was repeated, the supernatant discarded, and theprecipitates assayed for radioactivity. It was found thatthe percentage of radioactivity precipitated did not varyover the range of concentrations of I* protein used.

Enzymatic digestion of 'y-globulin. Fragments ofcryoglobulins and normal human 'y-globulin were pro-duced by papain digestion (12) using a protein: enzymeratio of 100:1. Digestion was carried out in the presenceof 0.01 M cysteine and 0.002 M ethylenediaminetetraacetate(EDTA) at pH 7.5. Digestion was allowed to proceedfor 45 min at 37°C and was stopped by the addition of afive-fold molar excess of iodoacetamide. After 1 hr at40C, the digestion products were dialized against 0.1 Mphosphate buffer, pH 7.5. Pepsin digestion (13) was per-formed using 100: 1, protein: enzyme ratio at pH 4.0.Digestion proceeded for 16 hr at 37°C and was stoppedby dialysis against a large volume of 0.1 M phosphatebuffer, pH 7.5.

Ultracentrifugation. Analytic ultracentrifugation wascarried out in a Spinco model E ultracentrifuge equippedwith schlieren optics. Centrifugation was performed at20°C, at a speed of 52,640 rpm. Sedimentation coefficientsand areas were calculated with the aid of a Gaertner

1876 H. M. Grey, P. F. Kohler, W. D. Terry, and E. C. Franklin

comparator (Gaertner Scientific Corp., Chicago, Ill.).The sedimentation coefficients were corrected to the Sso, w

values by the accepted procedures (14). The partialspecific volumes were assumed to be 0.73.

RESULTS

Electrophoretic and antigenic analysis. Seveny/G-cryoglobulins were available in sufficient quan-tity to perform detailed studies. It was difficult toperform agar or starch-gel electrophoresis on theseproteins to establish their monoclonal nature be-cause of their tendency to aggregate, even at roomtemperature. In order to circumvent this problemwe treated the proteins with pepsin which de-stroyed the property of cryoprecipitation andthereby made it possible to evaluate the electro-phoretic homogeneity of these proteins. The resultsof starch-gel electrophoresis of the pepsin-treatedcryoglobulins is shown in Fig. 1 and illustrates thediscrete bands that are the typical patterns ob-tained with the pepsin-treated myeloma proteinscompared with the diffuse smear obtained withnormal yG-globulin (HGG, Fig. 1, slot 7).

In contrast to the starch-gel electrophoresis datawhich indicated the cryoglobulins were essentiallyhomogeneous' immunodiffusion analysis revealedthat in several instances the cryoprecipitates wereantigenically heterogeneous, in that more than onelight chain type or heavy chain subclass waspresent in the washed cryoprecipitate. However,a predominant light and heavy chain type wasalways found. Since the electrophoretic analysisindicated that the major protein moiety in each ofthe cryoprecipitates was "monoclonal" with regardto electrophoretic mobility, it was assumed that

Kup Fin Man Pea Pur Cra HGG Spe1 2 3 4 5 6 7 8 +

FIGURE 1 Starch-gel electrophoresis (glycine buffer,pH 8.8) of the F (ab')2 fragments of cryoglobulins andnormal 'y-globulin, obtained by pepsin digestion. TheF(ab')2 fragments of the cryoglobulins yielded discretebands whereas the F(ab')2 fragment of normal y-globu-lin gave a diffuse smear.

the major antigenic type present in the cryoprecipi-tate represented the antigenic structure of the"monoclonal" protein. Table I gives the results ofthis antigenic analysis. In those cases where signifi-cant amounts of more than one heavy chain sub-class or light chain type were present the minorantigenic type is placed in parenthesis next to themajor type. In some instances the minor "contami-nating" antigenic type made up as much as one-fourth to one-third of the total cryoprecipitate.Five of seven proteins had type K light chains asthe sole or major class present. Four proteinswere of the yG3-heavy chain subclass and threewere of the yGl-subclass.

Determination of anti-y-globulin activity. Thefinding that in some instances the cryoprecipitatecontained the myeloma protein plus other y-globu-lin suggested the possibility that the myelomaprotein cryoglobulin might be interacting with the

TABLE IAntigenic Classification and Anti-y-Globulin Activity of YG-Monodonal Cryoglobulins

%o Radioactivity in cryoprecipitateLatex agglu-

yG-heavy chain Light chain tination of NormalCryoglobulin subclass type cryoprecipitate -yA HSA yG

Man yG3(-yGl,yG2)* K 0.11 3.0 2.2 33.0Kup yG3 K(X) 0.06 4.7 5.9 33.8Pea -yGl X 0.06 3.2 2.1 13.7Cra -yG3(-yG1) K 0.06 5.7 5.7 35.3Spe yG1 X >10 3.4 2.3 15.1Pur yG3(yGl) K(X) >10 5.1 3.1 26.1Fin yG1 K >10 2.5 3.5 6.8

* Figures in parentheses indicate the presence and classification of significant amounts (>5%) of other immunoglobulinantigens.T Lowest concentration of protein (mg/ml) to give positive reaction.

Multiple Myeloma Proteins with Anti--v-Globulin Activity 1877

normal y-globulin present in the patient's serum.The first system used to test this possibility was thelatex agglutination method for measuring rheuma-toid factor activity. The results of this test areshown in Table I. Four of the seven cryoglobulinstested possessed a high titer of rheumatoid factoractivity and three proteins were negative. Testsdone on the serum supernatants obtained aftercryoprecipitation demonstrated only weakly posi-tive or negative reactions.

Another indication of interaction between thecryoglobulins and human yG-globulin was obtainedby determining the extent of coprecipitation ofseveral radioiodinated proteins. The degree ofretention of I*yG, I*yA, and I*HSA in the cryo-globulin precipitate is given in Table I. Only2-6% of the added yA-myeloma protein or HSAwas present in the cryoprecipitate whereas therewas 13-35%o of the added normal -yG in six ofseven cryoprecipitates. When the data obtainedwith the iodinated human y-globulin is comparedwith the rheumatoid factor activity as measured bythe latex fixation method (Table I), good butimperfect correlation is observed. Three proteins,Man, Kup, Cra, that had high I*yG-binding had ahigh rheumatoid factor activity; the one protein,Fin, which did not bind I*yG had a negativerheumatoid factor test and one protein, Spe, wasweakly positive in one and negative in the othertest. However, two proteins showed discordantresults. One protein, Pea, only weakly bound theI*yG but gave a strongly positive rheumatoid fac-tor test, whereas another protein, Pur, that showedgood binding of the I*yG was negative by the latexagglutination technique.

Ultracentrifuge studies. There are obviousdifficulties in evaluating the interaction of a yG-immunoglobulin with another protein of the sameimmunoglobulin class. Since several of the cryo-globulins behaved as rheumatoid factors, i.e. anti-bodies to -yG-globulin, it was considered of particu-lar interest to digest the cryoglobulins with papainand pepsin in order to separate the Fab- and Fc-fragments thus allowing each to be studied inde-pendently. If specific interaction between the Fabor F(ab')2 of the cryoglobulin and -yG-globulincould be demonstrated, this would represent fur-ther evidence that the cryoglobulins may act asclassical antibodies with regard to the localizationof the combining sites to the Fab-fragment. When

/ ~ ~~~~Mon Fab +yG

Man FcayG

FIGURE 2 Analytic ultracentrifuge analysis of the in-teraction between papain fragments of cryoglobulin Manand normal 'yG-globulin. Pictures taken 64 min afterreaching speed of 52,640 rpm. Man Fab formed complex(a) which sedimented more rapidly than yG by itself (b)whereas, the Fc-fragment did not complex with -yG (c).

the seven cryoglobulins, were digested with pepsinor papain, all of them lost the property of cryo-precipitation. When analyzed in the analytic ultra-centrifuge the usual 3.5S and 55 peaks were ob-served with the papain and pepsin digested pro-teins, respectively, and no suggestion of higherpolymers were seen when analyzed at 10 or 200 C.

In order to determine whether the Fab-fragmentof the cryoglobulins was responsible for the inter-action observed with normal -y-globulin in theprevious test systems, Fab- and Fc-fragments ofMan cryoglobulin were isolated, mixed with nor-mal -y-globulin, and studied in the analytic ultra-centrifuge (Fig. 2). The papain Fab-fragment ofthe Man protein interacted with normal -yG toform a complex which sedimented with an s rateof 6.9, 0.6 U greater than the -yG-protein by itself(Fig. 2 a and b) ; whereas when the Man Fc-fragment was mixed with -yG-globulin no evidenceof interaction was obtained (Fig. 2 c). The area


under the faster sedimenting peak in Fig. 2 a wasgreater than that observed when yG was centri-fuged separately (Fig. 2 b) and the area underthe 3.5S peak was less than that observed whenthe Fab-fragment was centrifuged separately. Thisalso indicates that complexing had occurred be-tween the Man Fab-fragment and normal yG-globulin.

For further ultracentrifugal analysis of the inter-action between the cryoglobulins and -yG-globulin,pepsin digested F(ab')2 fragments were used.Representative myeloma proteins of the four heavychain subclasses were used as antigens rather thannormal y-globulin so that it could be determinedwhether the cryoglobulins recognized structuresspecific for the heavy chain subclasses or whetherthe reactive regions were common to the foursubclasses. Complexing between the cryoglobulinF (ab') 2 and the yG-myeloma proteins is illustratedin Figs. 3 and 4. The complexing was character-ized, as was the case for the papain Fab-fragments,by a diminution of the 5S F (ab')2 peak and an in-crease in the area and s rate of the faster sediment-ing peak. Fig. 3 illustrates the interaction betweenF (ab')2 fragment of the myeloma protein Manwith a -yGl-myeloma protein. The 5S peak wasalmost absent in the cell which contained bothF (ab')2 and a yGl-myeloma protein (Fig. 3 b).A single, slightly assymetrical peak was observedwhich was larger and sedimented slightly ahead ofthe peak observed when the yGl-protein was cen-trifuged by itself (Fig. 3 c) at the same proteinconcentration. The s20,,0 of the yG-myeloma proteinwas 6.6 and that of the 5S-7S complex 8.2. Thelower two frames show a mixture of Man F (ab')2with a yA-myeloma protein. The yA-protein usedin this case was a polymer. No diminution in the5S peak was observed nor was there a noticeablechange in the sedimentation pattern of the yA-polymer. Fig. 4 illustrates the reactions betweenthe F (ab')2 of the myeloma protein Pur and twoyG- and a yA-myeloma protein. Fig. 3 b and gshow the interaction between the F(ab')2 of thecryoglobulin and the yG-proteins and Fig. 2 ddemonstrates the lack of interaction with a yA-protein. This figure also demonstrates that PurF (ab')2 reacted to a greater degree with the pro-tein of the yG2-subclass than with that of the yG4-subclass. Although the 5S peaks were depleted toalmost the same extent when the schlieren pattern

Man F(ab')2

Man F(ab )2 IyG

YG

Man F(ob') + yA2s~~~~~~~~~~

g ~~~~~~~~....:e.

yA

FIGURE 3 Analytic ultracentrifuge analysis of the inter-action between the F (ab') 2 fragment of cryoglobulinMan and myeloma proteins. Pictures taken 64 min afterreaching speed of 52,640 rpm. Man F(ab'),, complexedwith 'yGl-protein (b) but not with yvA-protein (d).Complexes caused diminution of 5S peak and increase ins rate and area of faster sedimenting peak when comparedto unmixed proteins (a and c).

of Fig. 3 b and g were compared, the complexesformed between Pur F (ab') . and 7yG2 were larger,having a sedimentation rate of 8.4, whereas thecomplexes formed between Pur F(ab'), and -yG4had an s rate of 6.7 only 0.2 Svedberg units greaterthan the s rate of the uncomplexed yG4-proteinshown in Fig. 31.

Table II summarizes the ultracentrifuge resultsobtained with the F (ab') . of the seven cryoglobu-lins studied and indicates the maximum degree ofinteraction observed and gives a comparison of therelative intensity of the reaction obtained with each

Multiple Myeloma Proteins with Anti-v-Globulin Activity 1879

of the four yG-heavy chain subclasses as measuredby the s rates of the complexes formed. Five of theseven proteins showed definite interaction betweentheir pepsin F (ab')2 and one or more of the yG-myeloma proteins used. Three of the reactionsshowed some degree of specificity for one heavychain subclass over another; although wherever areaction was observed, all subclasses interacted tosome extent. Two proteins, Fin and Spe, showedno detectable interaction.

Pur F(ab')2

Pur F(ab')2 + yG2

YG2

y G4

Pur F(ab')2 + yG4: ... -.-..-..

FIGURE 4 Analytic ultracentrifuge analysis of the in-teraction between F(ab')2 fragments of cryoglobulin Purand myeloma proteins. Pictures taken 64 min after reach-ing speed of 52,640 rpm. Pur F (ab') 2 complexed withyG2- and yG4-proteins (b and g) but not with -yA (d).The complexes formed with 'yG2 had a faster sedimenta-tion rate than those formed with yG4.

TABLE I IUltracentrifuge Analysis of Complex Formation between

F(ab')2 of Cryoglobulins and 7G-Myeloma Proteinsof Different Heavy Chain Subclasses

Degree of interaction

Increase ins rate of

Diminu- "7S" peak Heavy chainCryo- tion of due to SS- subclass specificity

globulin SSpeak* -yG-complex of interaction

Man 4+ 1.6 1 >443 2Kup 1+ 0.2 2 3 4Pea 3+ 0.3 -2a3az4Cra 3+ 1.0 3>1>2>4Spe 0 0Pur 4+ 1.9 2 > 1 3 z4Fin 0 0

* 1+10-25%; 2+25-50%; 3+50-75%; 4+75-100%.

When the three tests (rheumatoid factor activ-ity, I*y-globulin precipitation, and F(ab')2-y-globulin complexing) used for evaluating anti-y-globulin activity of the cryoglobulins werecompared, the correlation was again good but notcomplete. Four proteins showed positive reactionsin all three tests (Man, Kup, Pea, and Cra); oneprotein, Fin, was negative in all three tests; oneprotein, Spe, was weakly positive in one test andnegative in the other two; and one protein, Pur,was strongly positive in two tests but negative inthe rheumatoid factor assay. It is of interest thatthis last protein which showed the most discordantresults was the only protein which demonstratedspecificity for the yG2-subclass.

It has previously been shown that most -yM-rheumatoid factors react with the Fc-fragment ofyG-globulin. The finding that some of the cryo-globulins reacted preferentially with certain of theyG-heavy chain subclasses suggested that thedeterminants that the cryoglobulins reacted withwere also located on the Fc-fragment. This wasdirectly demonstrated by studying the interactionbetween the F(ab')2 of a cryoglobulin with theFc- and Fab-fragments of normal y-globulin. Theultracentrifuge analysis of this interaction is shownin Fig. 5 and demonstrates the complexing ofnormal -yG-Fc-fragment with the F(ab')2 fromMan cryoglobulin to produce a complex which hada sedimentation rate of 6.5S (Fig. 5 a). There wasno evidence of complex formation between 'yG-Fab-fragment with the Man F(ab')2 (Fig. 5 b).


.!m;;.. P u r F (a b') 2

+ y AA..

Mon F(ob')2+ yG Fob

FIGURE 5 Analytic ultracentrifuge analysis of the in-teraction between F (ab')2 fragment of cryoglobulin Manand papain fragments -of normal y-globulin. Picture taken64 mm after reaching speed of 52,640 rpm. Man F (ab')interacted with. the Fc-fragment (a) to form a 6.5S com-plex, whereas no complex formed with the Fab-fragmentof normal 'v-globulin (b).

Reaction. of cryoglobulins with heterologous-y-globulins. In order to further investigate thespecificity of the reaction between the cryoglobu-lins and y-globulin, we made a study of the reactionof two cryoglobulins, Man and Kup, with radio-iodinated y-glo'bulins of several infra-human spe-cies. The y'globulins available for study wereobtained from: gorilla, monkey, horse, cow, dog,guinea pi and turkey. Both the gorilla and

RGGNSA

*'a2.....c-etLi S.: _O

'vgoui. Inrasn aons of hua 'v-globuli(HGG), rnabbiti 'vgoulircnt(rifgG)eo humanysserumth al-bumcion (HSA)weren added tor3agmn of MncryoglobulinMa

6mincuae afte 3eahinfop1hread 1f5264hrpmatn4bforecen

tRifeaction.niiin of cryo precuips itat formtironoc-

curedifonly afte the raddtion ofteeHGG. yolou

lis n yg'ouin e ae tuyofte ectooftwo ~~~crygouis andKp2ihrdoioiae ygouin fsvra nr-hmnsecie. Te Wgouisaalbefrsuywr

° I \ 5iO2mg Prti\ de

FIUE6Ihbto o f'rorcpt bynoml\ua

( G-, rabi y Protbuln Addedrhua srm l

bumin (HIS-A) were added to 3 mg of Man cryoglobulin,incubated at 37° for I hr. and 16 hr at 4°C before cen-trifugation.. -Inhibition' of cryoprecipitate formation oc-curred only after the addition of HGG.

monkey y-globulin coprecipitated with the twocryoglobulins studied to the same extent as humany-globulin (35-40%o); whereas, none of the fivenonprimate y-globulins reacted to a significant ex-tent. A labeled preparation of human yM from acase of Waldenstrdm's macroglobulinemia was alsostudied by the coprecipitation procedure. None ofthe seven cryoglobulins reacted with the yM-preparation.

Inhibition of cryoprecipitation by normal yG-globulin. In a further attempt to characterizesome cryoglobulins as a special form of precipi-tating antibody to y-globulin, the effect of theaddition of increasing amounts of y-globulin on theformation of the cryoprecipitate was studied.Fig. 6 shows these results and indicates the man-ner in which increasing amounts of normal humany-globulin added to the cryoglobulin led to a de-crease in the total cryoprecipitate formed, whereasrabbit y-globulin and human serum albumin hadno such effect. This inhibition of precipitation issimilar to that observed in classical immunsys-tems studied by quantitative precipitin analysiswhere addition of antigen in excess of that neededto maximally precipitate the antibody results inthe formation of soluble antigen-antibody com-plexes.

DISCUSSION

The present study of seven monoclonal yG-cryo-globulins has demonstrated that these proteins dif-fer from noncryoglobulin monoclonal yG-proteinswith respect to two important features. First,there is an extremely high incidence (five ofseven) of proteins that interact with .yG-globulinsand may be considered to possess antibody activityto yG-globulin;'second,'there is a similarly highincidence of the yG3-heavy chain subclass in thisgroup of myeloma proteins compared to thatfound in noncryoglobulin yG-myeloma proteinswhere the incidence ranges from 5-10 % (15, 16).It is of interest in this regard that one cryoglobu-lin which was negative by all three tests for anti-y-globulin activity and another protein which wasonly weakly positive to one test were both of theyGl-subclass, whereas, of the other five proteinswith anti-yG-activity, four were of the yG3-sub-class. That the anti-j-globulin activity was as-sociated with the myeloma protein which was themajor component of the cryoprecipitate and not

Multiple Myeloma Proteins with Anti--v-Globulin Activity 1881

due to small amounts of nonmyeloma 19S or 7Srheumatoid factor is indicated by the ultracentri-fuge analysis of the interaction between the pep-sin treated cryoglobulins and yG-myeloma pro-teins. As shown in Figs. 3 and 4 most, if not all,of the F(ab')2 was capable of complexing withthe yG-myeloma proteins as evidenced by the al-most complete disappearance of the 5S peak. Ifthe anti-y-globulin activity was associated with aminor fraction of nonmyeloma rheumatoid factoronly minor alterations in the size of the 5S peakshould have been observed.

The significance of why complexes betweenthe F (ab')2 of cryoglobulins and certain yG-mye-loma proteins sedimented at s rates only 0.1-0.3units greater than the uncomplexed yG-myelomaproteins is unclear but it may reflect a highly dis-sociating complex, whereas those combinationswhich yielded complexes which sedimented withs rates 1-2 units greater than the uncomplexedyG-proteins may have formed complexes of higheraffinity that sedimented with an s rate compatiblewith stable 5S-7S complexes. An alternative ex-planation involving differences in the frictionalcoefficients of the complexes formed is also pos-sible, but appears less likely.

It would seem likely that both of the distinctivefeatures of these cryoglobulins; viz. antibody ac-tivity toward y-globulin and the yG3-heavy chainstructure, are important in the formation of thecryoprecipitate. A possible explanation for theformation of the cryoprecipitate in those proteinsthat possess anti-y-globulin activity would bethat it represents antigen-antibody complexes ofsuch a nature that precipitating complexes areformed only in the cold. The temperature depend-ence of the precipitation could be explained bypostulating the antigen-antibody bond formed isone of very low affinity, since with complexes oflow affinity, the antigen-antibody equilibrium tendsto be shifted toward dissociation at higher tem-peratures.2

Why cryoglobulins with anti-y-globulin activitytend to be of the yG3-subclass is not immediatelyapparent; however, there are at least two possiblereasons for this. First, it is possible that there isa close relationship between the structure of anantibody site and the light or heavy chain class orsubclass to which the y-globulin molecule belongs.

2Grey, H. M. Unpublished observations.

Or, put in more precise terms, the type of varia-tion that can exist in the variable regions of theheavy and light chain is limited by and related tothe structure of the constant region. Such a rela-tionship between antigenic class and antibodyspecificity has been observed in the case of coldagglutinins with I blood group specificity: mostof which have only type K light chains (17), andcertain antibodies produced in the guinea pig tothe dinitrophenyl hapten (18). The second pos-sible explanation for the preponderence of theyG3-subclass in this group of cryoglobulins withanti-yG-activity is that the antigen-antibody com-plexes formed with yG3-antibody molecules areless soluble than those formed with antibody mole-cules of the other subclasses. The complexes madeup of yG3-molecules would therefore have theobserved property of cryoprecipitability, whereascomplexes made up of the other subclasses wouldbe more soluble. Observations on yG3-myelomaproteins without antibody activity give some in-direct evidence for this hypothesis in that manyyG3-proteins have a tendency to form aggregates.8

Another indication that the over-all structureof the protein is important with regard to the phe-nomenon of cryoprecipitability came from obser-vations made with the pepsin digested cryoglobu-lins. Pepsin treated cryoglobulins lost their abilityto precipitate in the cold. Combination with yGstill occurred however, since addition of normalyG-globulin to the F(ab')2 fragments of the cryo-globulins resulted in the formation of soluble com-plexes demonstrable in the ultracentrifuge. How-ever, on the addition of normal yG to the F(ab')2fragment of the cryoglobulin, over a wide range ofyG-concentrations, no cryoprecipitate formed.This would indicate that the phenomenon of coldprecipitation is dependent on the integrity of thewhole molecule and suggests that the structure ofthe Fc-fragment is important in determining thesolubility of the complexes that form.

The finding in the present study, that there ap-pears to be a high incidence of antibody activityto y-globulin among yG-cryoglobulin myelomaproteins, confirms and extends previous reportsin which studies of unselected cases of cryoglobu-linemia were studied (2, 3, 19). In a previousseries (2, 3), 12 of the 29 cases of cryoglobu-

8 Kunkel, H. G., and H. M. Grey. Unpublished ob-servations.


linemia studied had a positive test for rheumatoidfactor. In one case the rheumatoid factor was as-sociated with a yG-myeloma protein, whereas, inthe remaining 11 cases the cryoglobulin was ofthe mixed type (19) in which the yM-componentacted as rheumatoid factor and the yG-globulinserved as antigen for the rheumatoid factor.

Most anti-y-globulin antibodies studied comefrom the sera of patients with connective tissuedisorders and have been yM-immunoglobulins.However, yG-antibodies directed against 'yGglobulin antigen are also quite common but lessamenable to study because of the difficulty in iso-lating the antibody from the antigen since bothare 7S yG-globulins. The technique used in thepresent study was one employed by Schrohenloherin studying -yG-rheumatoid factor; that is, pepsindigestion of the antigen-antibody complex, therebydestroying the antigenic determinants which re-side in the Fc-fragment (20). Using this tech-nique, Schrohenloher was able to demonstratecomplexing between the F(ab')2 of the yG-rheu-matoid factor with normal 'yG-globulin by ultra-centrifuge analysis. The complexing observed wassimilar in degree and size of complex formed tothose observed in the present study using mye-loma proteins.

The current study adds to the recently expand-ing list of paraproteins which may be consideredto have antibody activity. Several cases of mono-clonal yM-globulins with anti-y-globulin activityhave been described (21-23). Recently, one suchcase which also was a "mixed" cryoglobulin in apatient with Waldenstr6m's macroglobulinemiawas studied in detail and was shown to combinespecifically with yG-globulin from human or otherprimate sera but, as in the present study, not withother mammalian y-globulins (23). It was alsopossible to demonstrate that the binding site waslocated in the Fab-fragment of the yM-molecule.Other antibody specificities that have been as-cribed in the past to monoclonal immunoglobulinsare: cold agglutinins with specificity for the Ired cell antigen (24); an antigen present on agederythrocytes, but not on fresh ones (25); streptolysin 0 and staphlysin T (26); a- and /8-lipopro-teins (27); and dinitrophenyl hapten (28). It isof considerable interest that most of the mono-clonal proteins studied with antibody activityhave specificity for autoantigens. The significance

of this at present is unclear but it may be an im-portant factor in the pathogenesis of certain dys-proteinemic states.

The question is often raised whether all or in-deed any of the myeloma proteins mentionedabove which exhibit affinity for a variety of sub-stances are truly antibodies. In the present studytwo of the most important criteria for antibodieshave been met by the cryoglobulins. First, thedegree of specificity with which the cryoglobu-lins reacted with primate yG-globulin and notwith other primate immunoglobulins or mam-malian yG-globulins is in keeping with, and isconsidered strong evidence for the antibody na-ture of the cryoglobulins. Second, the reactive re-gion on the cryoglobulin molecule has been shownto be the Fab-fragment which also points to theantibody nature of these proteins. A third propertywhich classical antibodies share but which is notshared by these paraproteins because of the verynature of the disease from which they arise, is theinducibility of antibody formation. It is perhapsthis last point which has led to much scepticismand caution in designating some paraproteins asantibodies. However, the great similarities theseproteins share with antibodies with respect to thenature of their reactivity with a specific ligandand the localization of the reactive site to thesame portion of the molecule in which it occurs ininduced antibodies, make it appear likely that theyare indeed antibodies. As more information be-comes available on the nature and further locali-zation of the antibody combining site within theimmunoglobulin molecule, it will be possible totest the myeloma "antibodies" by these more re-fined criteria as well.

ACKNOWLEDGMENTSWe wish to acknowledge the generous gift of serum(Man) we received from Dr. Stanley J. Altman. It isalso a pleasure to acknowledge the fine technical as-sistance of Miss Sharon Ford.

This is Publication No. 264 of the Department of Ex-perimental Pathology, Scripps Clinic and ResearchFoundation. This investigation was supported by Re-search Grant AI-07007 and AI-06894 of the U. S. PublicHealth Service, and in part, by Atomic Energy Com-mission Contract AT(04-3)-410, and the American HeartAssociation grant No. 67-795. This study was done duringthe tenure of an Established Investigatorship of theAmerican Heart Association to one of the authors(H.M.G.).

Multiple Myeloma Proteins with Anti---Globulin Activity 1883

REFERENCES

1. Christian, C. L., W. B. Hatfield, and P. H. Chase.1963. Systemic lupus erythematosus. Cryoprecipitationof sera. J. Clin. Invest. 42: 823.

2. Meltzer, M., and E. C. Franklin. 1966. Cryoglobu-linemia-a study of twenty-nine patients. I. IgG andIgM cryoglobulins and factors affecting cryoprecipita-bility. Am. J. Med. 40: 828.

3. Meltzer, M., E. C. Franklin, K. Elias, R. T. Mc-Cluskey, and N. Cooper. 1966. Cryoglobulinemia-aclinical and laboratory study. II. Cryoglobulins withrheumatoid factor activity. Am. J. Med. 40: 837.

4. Terry, W. D. 1965. Skin-sensitizing activity relatedto 'y-polypeptide chain characteristics of human IgG.J. Immunol. 95: 1041.

5. Ishizaka, T., K. Ishizaka, S. Salmon, and H. Fuden-berg. 1967. Biologic activities of aggregated y-globu-lin. VIII. Aggregated immunoglobulins of differentclasses. J. Immunol. 99: 82.

6. Takatsuki, K., and E. F. Osserman. 1964. Structuraldifferences between two types of "heavy chain" dis-ease proteins and myeloma globulins of correspondingtype. Science. 145: 499.

7. Fahey, J. L. 1963. Heterogeneity of myeloma proteins.J. Clin. Invest. 42: 111.

8. Kunkel, H. G. 1954. Zone electrophoresis. MethodsBiochem. Analy. 1: 141.

9. Flodin, P., and J. Killander. 1962. Fractionation ofhuman-serum proteins by gel filtration. Biochim.Biophys. Acta. 63: 403.

10. Scheidegger, J. J. 1955. Une micro-methode d'immuno-electrophorese. Intern. Arch. Allergy Appl. Immunol.7:103.

11. McConahey, P. J., and F. J. Dixon. 1966. A methodof trace iodination of proteins for immunologicstudies. Intern. Arch. Allergy Appl. Immunol. 29: 185.

12. Porter, R. R. 1959. The hydrolysis of rabbit -y-globu-lin and antibodies with crystalline papain. Biochem. J.73: 119.

13. Nisonoff, A., F. C. Wissler, L. N. Lipman, and D. L.Woernley. 1960. Separation of univalent fragmentsfrom the bivalent rabbit antibody molecule by reduc-tion of disulfide bonds. Arch. Biochem. Biophys. 89:230.

14. Schachman, H. K. 1959. Ultracentrifugation inBiochemistry. Academic Press, Inc., New York.

15. Grey, H. M., and H G. Kunkel. 1964. H-chain sub-groups of myeloma proteins and normal 7S y-globu-lin. J. Exptl. Med. 120: 253.

16. Terry, W. D., and J. L. Fahey. 1964. Subclasses ofhuman -y2-globulin based on differences in the heavypolypeptide chains. Science. 146: 400.

17. Harboe, M., R. Van Furth, H. Schubothe, K. Lind,and R. S. Evans. 1965. Exclusive occurrence of Kchains in isolated cold haemagglutinins. Scand. J.Haemat. 2: 259.

18. Nussenzweig, V., and B. Benacerraf. 1967. Antihaptenantibody specificity and L-chain type. J. Exptl. Med.126: 727.

19. LoSpalluto, J. B. Dorward, W. Miller, Jr., and M.Ziff. 1962. Cryoglobulinemia based on interaction be-tween a gammamacroglobulin and 7S gamma globu-lin. Am. J. Med. 32: 142.

20. Schrohenloher, R. E. 1966. Characterization of they-globulins complexes present in certain sera havinghigh titers of anti-vy-globulin activity. J. Clin. Invest.45: 501.

21. Kritzman, J., H. G. Kunkel, J. McCarthy, and R. C.Mellors. 1961. Studies of a Waldenstrbm-type macro-globulin with rheumatoid factor properties. J. Lab.Clin. Med. 57: 905.

22. Curtain, C. C., A. Baumgarten, and J. Pye. 1965.Coprecipitation of some cryomacroglobulins with im-munoglobulins and their fragments. Arch. Biochem.Biophys. 112: 37.

23. Metzger, H. 1967. Characterization of a humanmacroglobulin. V. A Waldenstrbm macroglobulin withantibody activity. Proc. Natl. Acad. Sci. U. S. 57:1490.

24. Fudenberg, H. H., and H. G. Kunkel. 1957. Physicalproperties of the red cell agglutinins in acquiredhemolytic anemia. J. Exptl. Med. 106: 689.

25. Ozer, F. L., and H. Chaplin, Jr. 1963. Agglutinationof stored erythrocytes by a human serum. Characteri-zation of the serum factor and erythrocyte changes.J. Clin. Invest. 42: 1735.

26. Zettervall, 0. 1967. Some properties of human Gmyeloma globulins with antibody activity. In GammaGlobulins-Structure and Control of Biosynthesis.Interscience Publishers, Inc., New York.

27. Beaumont, J. L., B. Jocotot, C. Vilain, and V. Beau-mont. 1965. Presence d'un auto-anticorps anti-a-lipiproteine dans un serum de myelome. C. R.' Acad.Sci. 260: 5960.

28. Eisen, H. N., J. R. Little, K. Osterland, and E. S.Simms. 1967. A myeloma G-protein with antibody(anti-2, 4-dinitrophenyl) activity. Cold Spring HarborSymp. Quant. Biol. 32: 75.


humanmonoclonal yg-cryoglobulins with anti-y...

Documents