yl-MACROGLOBULINS, AND BENCE JONES PROTEINS IDEN

Jowuna of Clinical InvatigatioxVol. 42, No. 6, 1963

TWO TYPES OF y-MYELOMA PROTEINS, P2A-MYELOMA PROTEINS,yl-MACROGLOBULINS, AND BENCE JONES PROTEINS IDEN-

TIFIED BY TWO GROUPS OF COMMON ANTIGENICDETERMINANTS

By JOHN L. FAHEY AND ALAN SOLOMON

(From the Metabolism Service, National Cancer Institute, Bethesda, Md.)

(Submitted for publication December 14, 1962; accepted February 13, 1963)

The serum myeloma proteins are characteristicproducts of malignant plasma cells, and individualmyeloma proteins provide an important source ofinformation on the protein-producing capacities(and limitations) of plasma cell clones. Theglobulins produced by malignant plasma cells areclosely related to the normal v-globulins, but im-munochemical studies of the antigenic determi-nants present on myeloma proteins have uniformlyshown that myeloma proteins are antigenicallydeficient when compared to normal y-globulins(1-8). Similar studies (9-13) have shown thatmacroglobulinemic macroglobulins are antigeni-cally deficient when compared to the normal71-macroglobulins. Antigen deletion has beendemonstrated to occur with malignant transforma-tion (14), and antigenic deficiency of myelomaproteins could be viewed as reflecting a plasmacell change associated with malignancy. Alter-natively, myeloma protein differences may simplyreflect differences between normal y-globulin mole-cules (and between normal plasma cells). Im-munochemical studies have not settled this ques-tion, although marked variability in the antigenicdeterminants on individual myeloma proteins hasbeen noted. A number of studies, however, haveemphasized systematic differences among the pro-teins formed in malignant plasma cells. Suchproteins are now identified as Bence Jones pro-teins, y-myeloma proteins, fl2A-myeloma proteins,or yi-macroglobulins on the basis of specific im-munochemical and physicochemical properties(15-18). Systematic differences of another type,however, may also occur within these four majorgroups.. Two antigenically different groups ofBence Jones proteins have been identified in sev-eral laboratories (19-23). Two types of y-mye-loma proteins were identified by Korngold andLipari (7), and Franklin (24) found two types

of 82A-myeloma proteins. The molecular differ-ences within major y-globulin groups has receivedless attention than the y-, 82A-, yLm-, and BJ-grouping, and the presence of two general sub-divisions has not been established as a consistentfeature of all of the anomalous y-globulin groups.

In the course of immunochemical studies ofy-globulins, differences were observed in the anti-genic composition of S pieces obtained by papaindigestion of individual myeloma proteins (25, 26).In view of previous observations (16, 27, 28)that the antigenic determinants common to ally-globulin groups were present in the S piecesfrom y-globulin molecules and that Bence Jonesproteins were antigenically related to the S pieces,studies were undertaken to determine whether allfour classes of proteins formed in malignantplasma cells could be divided into two groups onthe basis of readily identifiable antigenic differ-ences.Two forms of molecules (types I and II) dif-

fering in antigenic composition were identified ineach anomalous protein group (29). Furtherstudies demonstrated that two sets of antigenicdeterminants (types I and II) were common toall the y-globulin groups, but that the relative fre-quency of types I and II molecules differed in thevarious anomalous y-globulin groups. Moleculeshaving types I and II antigenic determinants werecompared to see whether other molecular param-eters varied in association with two types of anti-genic determinants. These observations form thebasis of the present report.

MATERIALS AND METHODS

Proteins. Twenty-six 'y-myeloma proteins and ten p2A-myeloma proteins were isolated by DEAE-cellulose chro-matography, or by a combination of zone electrophoresisand chromatography of myeloma sera (30). Macro-

811

JOHN L. FAHEY AND ALAN SOLOMON

globulins were isolated from the serum of ten patientswith primary (Waldenstr6m's) macroglobulinemia bypreparation of macroglobulin-rich serum fractions by re-peated precipitation through dialysis against distilled wa-ter, by 35% ammonium sulfate precipitations, or by zoneelectrophoresis and subsequent purification by DEAE-cellulose chromatography or Sephadex G-200 filtration(31). Bence Jones proteins were prepared by DEAE-cellulose chromatography from urine concentrated byultrafiltration, or by addition of 650 g of ammonium sul-fate per L of urine to precipitate the urinary proteins.Bence Jones proteins were obtained from ten patientswith no other anomalous protein, and from eleven pa-tients with y-myeloma protein, three with #2A-myelomaproteins, and five with macroglobulinemia.Each of the proteins was characterized by analytic ul-

tracentrifugation in a Spinco model E ultracentrifuge, bypaper electrophoresis and starch gel electrophoresis, andby immunoelectrophoresis using antiserum reacting withBence Jones proteins, 6.6 S 'y-globulins, 82A-globulins,and 'vi-macroglobulins, as well as antiserum reactingspecifically with each of the last three globulin groups(30, 32). Each of the proteins was classified on thebasis of its immunochemical and ultracentrifugal proper-ties (33). The 'y-myeloma proteins sedimented as 6.6 Scomponents in the ultracentrifuge. The #2A-myelomaproteins frequently contained 9, 11, and 13 S componentsas well as 6.6 S components. The macroglobulins werecomposed of 18, 24, and 32 S components. Bence Jonesproteins had sedimentation coefficients between 2.8 and4.2 S, and had typical heat precipitation properties whentested by the methods of Putnam and associates (34).Hexose analyses were performed by a modification (35)of the thymol-sulfuric acid method of Shetlar and Masters(36).Immunochemical procedures. Antisera prepared by im-

munization of rabbits, goats, horses, or sheep with wholenormal 'y-globulin, normal 6.6 S 'y-globulin, normal 18 S'ylM-globulins, 'v-myeloma proteins, 82A-myeloma proteins,'vim-macroglobulins, and Bence Jones proteins weretested for their capacity to detect type I and type II anti-genic determinants. Most of the studies to be describedutilized antisera prepared in rabbits by immunizationwith 'y-globulins prepared by zone (block) electrophore-sis of normal serum, with normal 6.6 S 'y-globulins pre-pared by DEAE-cellulose chromatography, or with puri-fied urinary Bence Jones proteins. Rabbit antiserumagainst type II Bence Jones proteins appeared to bespecific for this type of protein, but the available rabbitanti-type I Bence Jones protein antisera were absorbedwith type II Bence Jones protein or type II myelomaproteins to insure specificity. Specific anti-I and anti-IIantisera also were prepared by absorption of rabbit anti-normal 6.6 S 'v-globulin antiserum with type I or typeII -y-myeloma proteins.

Purified proteins were compared by Ouchterlony testsat protein concentrations of 0.1 to 0.2 mg per ml (BenceJones protein), 0.2 to 0.3 mg per ml (myeloma proteins),and 0.4 to 0.5 mg per ml (-y,-macroglobulins) in 3' X 4-

inch plates of agar in barbital buffer of 0.045 ionic strengthand pH 8.6 containing 0.01 M sodium azide. Immuno-electrophoresis was carried out in similar plates. Dark-field photographs were made at 24, or 48 hours, or both.

RESULTS

Two types of Bence Jones proteins, myelomaproteins, and macroglobulins. Two types of BenceJones proteins, differing in antigenic determi-nants, were readily seen on Ouchterlony agardouble-diffusion tests (Figure 1 A). The precipi-tin lines of Bence Jones proteins a and d intersected,indicating that the two proteins differ in antigenicdeterminants. A similar intersection is seen withthe precipitin lines formed by Bence Jones pro-teins c and f (Figure 1A). Bence Jones pro-teins a, b, c, and others with similar antigenic de-terminants are termed type I. Bence Jones pro-teins d, e, f, and proteins with similar antigenicdeterminants are termed type II. The basis forthis nomenclature is discussed below, but it maybe noted here that type I corresponds to type Bof Korngold and Lipari (22) and type II corre-sponds to their type A. The antigenic differ-ences between types I and II Bence Jones proteinscan also be demonstrated with specific antisera,as shown in Figure 1B, where precipitin lines areformed only between a Bence Jones protein andits corresponding specific antiserum, i.e., betweentype I Bence Jones protein and anti-I antiserum.The antigenic differences shown in Figure 1

are type-specific and are not protein-specific. In-dividual proteins in each Bence Jones proteingroup are closely related antigenically, as shownby the precipitin lines formed by individual BenceJones proteins that fuse (evidence for antigenicsimilarity) when tested against rabbit antinor-mal y-globulin antiserum. This is seen in Figure1 for three Bence Jones proteins of each type.y-Myeloma proteins were divided into two

groups on the basis of antigenic differences. Asshown in Figure 2A, precipitin lines of two y-mye-loma proteins intersected, indicating that eachhad antigenic determinants not present on theother protein. These differences were revealedby a mixture of two antisera prepared againsttypes I and II Bence Jones proteins, respectively.The distinctive antigenic determinants of eachtype of -y-myeloma protein were also identified bymeans of antisera specific for the type I or II anti-

812

TWO TYPES OF MYELOMA PROTEINS, MACROGLOBULINS, AND BENCE JONES PROTEINS 813

TYPE IA.

TYPE IIB.

FIG. 1. A. Two TYPES OF BENCE JONES PROTEINS. Purified Bence Jones proteins a, b, and cfrom patients (E.B., M.T., and G.Y.) have similar antigenic properties, as revealed by fusionof the precipitin lines formed in an Ouchterlony plate using rabbit antinormal human 6.6 S-y-globulin antiserum (AS). These were termed type I Bence Jones proteins. Anothergroup of Bence Jones proteins d, e, and f from patients (Z.O., S.E., and R.F.) shared anotherset of antigenic determinants, and were designated type II Bence Jones proteins. The inter-section of the precipitin lines formed by adjacent types I and II Bence Jones proteins (as be-tween a and d) served to distinguish these two types of proteins. B. TWO TYPES OF BENCEJONES PROTEINS REVEALED BY TESTS WITH SPECIFIC ANTISERA. Type I Bence Jones protein J.H.(BJ I) and type II protein D.R. (BJ II) were tested against specific antisera prepared byabsorption of rabbit antinormal -y-globulin antiserum with appropriate type I or II myelomaproteins. Similar results were obtained with antiserum prepared by immunization of rabbitswith type I or II Bence Jones proteins and absorption of the rabbit serum, where necessary,rith appropriate Bence Jones protein.

genic determinants. Such antisera were preparedby immunization of rabbits with type I or IIBence Jones proteins, or by absorption of rabbitantinormal 6.6 S y-globulin antiserum with typeI or II y-myeloma proteins. As shown in Figure2B, each y-myeloma protein reacts only with thecorresponding specific antiserum.The two types of y-myeloma proteins were com-

pared with the two types of Bence Jones proteins(Figure 2C). Type I y-myeloma proteins andtype I Bence Jones proteins shared antigenic de-terminants (see fusion of precipitin lines in Fig-ure 2C). Similar observations indicated thattype II Bence Jones proteins and type II y-mye-loma proteins shared another set of antigenic de-terminants. These findings demonstrated thatthe two types of y-myeloma proteins were anti-genically related to the two types of Bence Jonesproteins. Therefore, additional studies were un-dertaken to determine whether similar moleculartypes occurred in other y-globulin classes.

Purified ,62A-myeloma proteins were found to

differ in antigenic determinants, since intersectingprecipitin lines developed when appropriate pro-teins were compared (Figure 3A). The two typesof P2A-myeloma proteins were identified as type Ior II molecules by means of Ouchterlony testsusing specific anti-I or anti-II antiserum (Figure3B).Two types of macroglobulins were identified

with similar reagents. Precipitin lines formed bythe two types of y1-macroglobulins intersected(Figure 4A), indicating that each possessed anti-genic determinants not present on the other pro-tein. The yl-macroglobulins were identified astype I or II molecules, i.e., as proteins havingeither type I or II antigenic determinants, byOuchterlony tests employing specific antisera(Figure 4B).Antigenic determinants of indisridual molecules.

The tests with specific antisera (Figures 1 to 4) re-vealed that individual anomalous proteins reactedwith either specific anti-I or anti-II antiserum, butnot with both. This was confirmed in studies of


A. B.TYPE I

C.TYPE II

FIG. 2. A. TWO TYPES OF -y-MYELOMA PROTEINS. Intersection of the precipitin arcs occurred when 'y-myelomaproteins from patients N.S. and T.N. were tested against pooled rabbit antisera (AS) prepared by immunizationwith types I and II Bence Jones proteins. B. Two TYPES OF -y-MYELOMA PROTEINS REVEALED BY SPECIFIC ANTISERAeN.S. and T.N. -y-myeloma proteins tested against specific anti-I and anti-II antisera prepared as for Figure 1B.C. RELATION OF TWO TYPES OF 'y-MYELOMA PROTEINS TO TWO TYPES OF BENCE JONES PROTEINS. A type I BenceJones protein J.H. (BJ I) was seen to share common antigenic determinants with -y-myeloma protein N.S.(-ymp I), as revealed by fusion of the BJ I precipitin line with the precipitin line formed by the 'y-myeloma proteinwhen tested against rabbit antinormal human y-globulin S fragment antisera (AS). Similarly, the precipitin lineformed by type II Bence Jones protein L.E. fused with that formed by y-myeloma protein S.J. (,ymp II), indicat-ing a close antigenic relationship between these types of Bence Jones proteins and y-myeloma proteins.

74 purified proteins, including 28 Bence Jonesproteins, 26 y-myeloma proteins, 10 82A-myelomaproteins, and 10 yl-macroglobulins. Evidently,each anomalous protein has only type I or II anti-genic determinants, but not both types.

Antigenic determinants common to all y-globu-lin classes. The observations made with specificantiserum indicated that types I and II antigenicdeterminants were represented in each of theanomalous y-globulin classes, i.e., among theBence Jones proteins, y-myeloma proteins, /32A-

myeloma proteins, and y1-macroglobulins. In amore direct comparison, representative proteinsfrom each class were tested by the Ouchterlonydouble-diffusion techniques (Figure 5). Whentested with specific antisera (Figure 5), the pre-cipitin line formed by each type I protein fusedwith that of the adjacent protein. This evidenceindicates that type I proteins from all y-globulinclasses share common antigenic determinants.Similarly, type II proteins of all four y-globulinclasses share another set of antigenic determinants

A. B.

FIG. 3. A. TWO TYPES OF fi2A-MYELOMA PROTEINS. Purified ,82A-myeloma proteins from patients J.B. (82A I)and W.H. (fi2A II) were compared using rabbit antinormal y-globulin antiserum (AS). B. Two TYPES OF.82A-MYELOMA PROTEINS REVEALED BY TESTS WITH SPECIFIC ANTISERA. Purified proteins from W.J. and G.F.were tested against specific anti-I and specific anti-II antisera as in Figure 1B.

814


B.

FIG. 4. A. TWO TYPES OF y,-MACROGLOBULINS. Purified y,-macroglobulins from patients with Waldenstr6m'smacroglobulinemia, J.B. (yvim I) and F.R. ('yiM II), were compared using rabbit antinormal 'y-globulin antiserum(AS). B. ANTIGENIC DIFFERENCES IN TYPE I AND TYPE II MACROGLOBULINS REVEALED BY SPECIFIC ANTISERA.

Purified proteins from J.W. (I) and F.R. (II) were used in the test.

(see fusion of adjacent precipitin lines whentested with specific anti-II antiserum, Figure 5).

Properties of type I and II molecules. Theelectrophoretic mobility of the 122 types I and IIproteins is graphically illustrated in Figure 6.Types I and II molecules are found to extendthroughout the entire electrophoretic range fromthe slow y-globulin region to the a-globulin region.Although the anomalous y-globulin groups dif-fered in range of mobility-i.e., /82A-globulins gen-

erally migrated faster than y-myeloma proteins-,the types I and II molecules of each group were

found to have approximately the same electro-phoretic distribution.

Ultracentrifugal studies revealed that the typesI and II Bence Jones proteins had approximatelythe same sedimentation coefficients (Table I).Types I and II y-myeloma proteins sedimented atabout 7 S. Both types I and II 82A-myelomaproteins usually were composed of 6.6, 9, 11, and13 S components. 12A-Myeloma proteins havingprimarily 9 S components were represented amongthe proteins of each type.

Starch gel electrophoresis showed that the BenceJones proteins were heterogeneous, but the rela-tive number of components did not correlate withthe Bence Jones protein type. y-Myeloma pro-

teins may appear homogeneous, or show electro-

TYPE I TYPE II

FIG. 5. COMPARISON OF TYPE I AND TYPE II PROTEINS FROM SEVERAL 'Y-GLOBULIN CLASSES. Bence Jones pro-

teins, 'y-myeloma proteins, iAu-myeloma proteins, and 'yv-macroglobulins from J.H., R.P., W.J., and J.B., respec-

tively, were used as type I proteins; those from E.L., T.N., G.F., and F.R. were used as type II proteins. Spe-cific anti-I and anti-II antisera prepared from rabbit antinormal 'y-globulin antiserum were used (AS).

A.


-30 20 10 T+10 20 30 -30 20 10 1+ 10 20 30

FIG. 6. ELECTROPHORETIC DISTRIBUTION OF TYPE I AND TYPE II MYELOMA PROTEINS,

MACROGLOBULINS, AND BENCE JONES PROTEINS. Electrophoretic mobility was deter-mined by zone electrophoresis of serum, urine, or purified proteins on filter paper

strips with the Durrum-Spinco apparatus and barbital buffer at pH 8.6, ionicstrength 0.09. Mobility values are plotted in relation to the point of sample applica-tion, and a normal serum electrophoretic pattern is included for orientation.

TABLE I

Some physicochemical properties of type I and type II molecules*

Ultracentrifugationt Starch gel electrophoreticType of proteint S(20.w) Electrophoretic mobility pattern Hexose

S mm from origin on paper % of proteinBence Jones proteins

II3.71El3I - 3 13 g00%}HeterogeneityII ~~3.7.1 CI10]1 1 0

'y-Myeloma proteinsI 6.96 [8] -10 [36] 63% [43] Electrophoretic 1.5 [8]

II 7.04 [2] - 5 [21] 68% [19]f heterogeneity: 1.6 [7]

162A-Myeloma proteinsI 6.6, 9, 11, 13 [1O] +12 [10] 70% [1CO Polymeric 3.7 [4]

II 6.6, 9, 11, 13 [11] +11 [10] 82% [ll]heterogeneityt 3.0 [2]

yi-Macroglobulins

I 18, 24, 32 [17] -10 [16] 25% [ 4] Anomalous protein 6.1 [4]

II 18, 24, 32 [ 4] -10[ 4] 12% [17]fband in gel 6.0 [1]

* On naming of types, see Discussion. Numbers in brackets indicate number of samples tested.t Analysis of isolated Bence Jones proteins and y-myeloma proteins and of whole serum for 132A-myeloma proteins

and y1-macroglobulins. Mean values are recorded.$ Starch gel electrophoretic patterns indicating heterogeneity on the basis of multiple components differing in net

electrical charge (electrophoretic heterogeneity) or in size (polymer-type heterogeneity) have been described in detail (33).

NORMAL HUMAN SERUM

b9i0000I0 MYELOMAgo 'L| * i PROTEINS ANOMALL.111.1:.i :..:: :z~ti.±.. SERUM

GAMMAI |2A - GLOBULINS

*,~~~~~~~~~~~~~~~~~~~~~~~Y*.A*,,-,.IM,.

tA. .V A..PROTEINS

:.el.0 el MACR el: . BENCE JONES-~~~~ J ~~~GLOBULINS PROTEINS

816


ANTISERA

ANTI-I

ANTI-6.6S --

ANTI-II ->

E.S. SERUM (TYPE I ymp) E.McC. SERUM (TYPE II ),mp)

(+) (-) (+)y ^ a A

FIG. 7. IDENTIFICATION OF TYPES I AND II MYELOMA PROTEINS BY IMMUNOELECTROPHORESISOF WHOLE SERUM. Serum from patient E.S. was used in the first set of immunoelectropho-retic analyses, and serum from E.McC. in the second. After electrophoresis in agar, specificanti-I antiserum was added to the upper trough, specific anti-6.6 S -y-globulin antiserum to themiddle trough, and specific anti-II antiserum to the lower trough. The arrows indicate theanomalous protein revealed by specific anti-I and anti-II antisera. As may be seen in themyeloma sera, the normal -y-globulins react with both antisera, but the characteristic bowingof the myeloma globulin is found only with homologous antiserum.

phoretic heterogeneity on starch gel electrophore-sis (33). Both homogeneous and heterogeneousy-myeloma proteins were represented within eachtype. Sixty-three per cent of type I y-myelomaproteins showed electrophoretic heterogeneity, as

did 68% of the type II y-myeloma proteins.The hexose content of purified protein was

measured (Table I). Types I and II y-myelomaproteins had similar hexose contents, i.e., about1.5%o. A wider range of hexose values (2 to4.5%) was found with /2A-myeloma proteins, butno significant difference was noted between thetwo types of f,2A-globulins. Both types I and II-y1-macroglobulin contained approximately 6%ohexose.

Relative frequency of types I and II molecules.Identification of anomalous types I and II pro-

teins in serum and urine was facilitated by im-munoelectrophoretic procedures using specificanti-I or anti-II antiserum in the antibody troughs.Typical results for two myeloma sera are shownin Figure 7. The E.S. serum contains a 6.6 Sy-myeloma protein (intense precipitin arc withspecific anti-y-globulin antiserum in center trough)that reacts with specific anti-I antiserum (uppertrough), but not with anti-II antiserum (lowertrough), identifying the protein as a type I y-mye-

loma protein. The E.McC. serum also contains a

-y-myeloma protein which, in contrast, fails to re-

act with anti-I antiserum, but does react with spe-

cific anti-II antiserum, i.e., a type II y-myelomaprotein.Serum and urine proteins from 120 patients

with multiple myeloma or macroglobulinemiawere examined to determine the relative frequencyof molecules having the types I and II antigenicdeterminants (Table II). Seventy-five purifiedproteins, representing all four classes of y-globu-lins, were tested by Ouchterlony analysis (Figures1 to 4), and additional sera or urine were tested byimmunoelectrophoresis. The frequency of typeI molecules was 82%o among the yi-macroglobu-lins, 67% among 6.6 S y-myeloma proteins, 53%ofor Bence Jones proteins, and 50%7o for the /82A-myeloma proteins. The difference between thefrequency of type I protein occurrence in the 6.6S y-myeloma protein group and in other anoma-lous protein groups was not statistically significant(p > 0.05 in each case).

TABLE II

Frequency of type I and type II molecules inmultiple myeloma and macroglobulinemia*

Purified Whole Totalproteins serum percentage

Proteins I II I II I II

6.6S 'y-Myeloma proteins [66] 19 7 25 15 67 3312A-Myeloma proteins [22] 4 6 7 5 50 50-yi-Macroglobulins [22] 9 1 9 3 82 18Bence Jones proteins [30] 16 12 2 53 47

* Numbers in brackets indicate the number of samples tested. Onnaming of types, see Discussion.


A. v-MYELOMA PROTEIN

TYPEI

TYPEII

B. 12A-MYELOMA PROTEIN

TYPEII

TYPEI

C. y-MACROGLOBULINS

FIG. 8. COMPARISON OF BENCE JONES PROTEIN AND ANOMALOUS SERUM GLOBULIN OF THE

SAME PATIENT. Ouchterlony tests were conducted with specific antisera, and only positiveresults are shown, as with R.D.'s Bence Jones protein and y-myeloma protein, which reactedonly with specific anti-I antiserum (AS) and not with anti-II antiserum.

Bence Jones protein and anomalous globulin inthe same patient. Both urinary Bence Jones pro-tein and an anomalous serum protein were avail-able from 20 patients. Ouchterlony tests, asshown in Figure 8, showed fusion of the precipi-tin lines formed by the two anomalous proteins ineach case, indicating that the two anomalous pro-teins from each patient were of the same type (I

TABLE III

Relationship of Bence Jones protein to anomalousglobulin in the same patient

Urinary Bence Jonesprotein type

Anomalous serumprotein type* I II

-y-Myeloma proteinsI 6 0

II 0 5

32A-Myeloma proteinsI 2 0

II 0 1

-yi-MacroglobulinsI 5 0

II 0 1

* On naming of types, see Discussion.

or II). The anomalous serum proteins, testedagainst Bence Jones proteins, include 6.6 S y-mye-loma proteins, 182A-myeloma proteins, and yl-mac-roglobulins of types I and II, as shown in TableIII.a

DISCUSSION

The demonstration of two separate groups ofantigenic determinants, common to all a-globulingroups-i.e., in the y-myeloma proteins, 82A-mye-loma proteins, y-macroglobulins, and Bence Jonesproteins-, which permits a division of each groupinto two types, confirms and extends the observa-tions of Korngold and Lipari (7), who identifiedthree types of myeloma proteins on the basis ofdifferences in antigenic determinants and desig-nated them as types I, II, and III. Subsequently,myeloma proteins of type III were identified asft2A-myeloma proteins (18). Both types I and IImyeloma proteins of Korngold and Lipari (7),however, were closely related antigenically to nor-mal y2-globulins. Both types shared some commonantigenic determinants, but each type had anti-genic determinants not present in the other group.Our own observations on the two groups of

818


y-myeloma proteins agree with those of Korngoldand Lipari (7), and we have used the same termi-nology, i.e., types I and II, for the specific anti-genic determinants distinguishing the two types ofy-myeloma proteins.

Identifications of two antigenically distinct typesof Bence Jones protein also agree with those ofKorngold and Lipari (22), whose type B is equiv-alent to our type I and whose type A is the same asour type II. The present studies and the recentstudies of Mannik and Kunkel (37), however, ex-tend the earlier work of Korngold and Lipari(22) and Franklin (24) to show that both groupsof distinctive antigenic determinants are repre-sented in all classes of anomalous proteins. Mi-gita and Putnam (38) also identified types I andII y-myeloma proteins on the basis of antigenicdeterminants common to one of the two types ofBence Jones proteins. These authors (38) rec-ognized that 182A-myeloma proteins and y-macro-globulins also had limited antigenic composition,but the proteins in these latter two classes avail-able to them were all type I globulins.We have referred to the two groups of anti-

genic determinants common to all y-globulingroups as types I and II antigenic determinants,in accord with Korngold and Lipari's terminologyfor y-myeloma proteins (7), and (because mole-cules with type I antigenic determinants occurmore frequently) have used the same designationfor all y-globulin groups, i.e., in the Bence Jonesproteins, y-myeloma proteins, 182A-myeloma pro-teins, and y1-macroglobulins.Many of the observations reported here empha-

size the antigenic differences between types I andII y-myeloma proteins. These two types of y-mye-loma proteins also share common antigenic deter-minants, i.e., those characteristic of 6.6 S y-globu-lins. Similarly, types I and II fB2A-myeloma pro-teins share antigenic determinants characteristicOf 82A-globulins, and types I and II y1-macro-globulins share the specific determinants of ylM-globulins.The presence of 6.6 S y antigenic determinants

on y-globulins complicated the demonstration oftypes I and II molecule differences. Antiserareacting with 6.6 S y antigenic determinantsand with types I and II determinants oftenshowed spur formation of the precipitin linesformed by types I and II myeloma proteins

in adjacent wells on Ouchterlony plates, butdid not readily distinguish the complete sepa-rateness of the types I and II antigenic determi-nants. Antisera reacting specifically with type Ior II antigenic determinants were obtained by im-munization of rabbits with purified type I or IIBence Jones proteins. For example, rabbit anti-sera prepared against type II Bence Jones pro-teins reacted with the type II antigenic determi-nants in all y-globulin groups, but did not react(to form precipitin lines) with type I proteins.Specific antisera could also be prepared by usingtype I or II y-myeloma proteins to absorb apotent rabbit antinormal y-globulin antiserum.These absorbed antisera reacted with type I orII antigenic determinants in all y groups (6.6 Sy-, /2A-, and yiM-globulins, and Bence Jones pro-teins). The number of antigenic sites detected byspecific anti-I or anti-II antisera has not beendetermined. Precipitin reactions formed readily,however, as is consistent with the presence of twoor more determinants on the antigenic molecule,and the types I and II antigenic determinants mayrepresent two groups of individual antigenicdeterminants.The type I antigenic determinants reflect one

structural configuration, and the type II determi-nants, a different configuration of these y-globu-lins. Further studies (37-39) have shown thatthe types I and II antigenic determinants arepresent on S pieces (obtained by papain digestion)and "L" polypeptide chains (produced by reduc-tion and alkylation) of the y-globulin molecules.The F pieces and "H" chains do not have the typesI and II antigenic determinants, but appear toaccount for the y-, 82A-, or yiM-specific propertiesof the anomalous proteins. Therefore, the mye-loma protein (or other anomalous protein) maybe classified on the basis of immunochemical teststhat detect specific properties in two parts of themolecule. For example, a type I -y-myeloma pro-tein contains the structural configurations (andpolypeptide chain) characteristic of 6.6 S y-globu-lins in one- part of the molecule, while anothermolecular part contains a type I polypeptide chain,i.e., a polypeptide with type I antigenic determi-nants. The type II y-myeloma protein differs inhaving a type II (instead of a type I) polypeptidechain, but is similar in the y-globulin-specific partof the molecule. A type II 132A-myeloma protein


is similar to the type II y-myeloma protein in thecommon type of polypeptide chain, but differsin having a 12A-globulin-specific polypeptide inanother part of the molecule (39).The wide range of electrophoretic mobilities

encountered for both types I and II moleculesindicates that the molecular configurations re-sponsible for the two sets of antigenic determi-nants probably have no relation to the moleculardifferences responsible for electrophoretic mobility.Similarly, the hexose content, ultracentrifugal be-havior, and presence or absence of electrophoreticheterogeneity (as revealed by starch gel electro-phoresis) or polymer-type heterogeneity (as seenon ultracentrifugation or starch gel electrophore-sis) (33) were not related to the type I or II anti-genic determinants, nor, presumably, to molecu-lar pieces represented by these two sets of anti-genic determinants. These observations on thephysicochemical properties of types I and II pro-teins support the view that the types I and II anti-genic determinants are segregated in a similarpart of the molecule.

Because all the anomalous proteins formed inmalignant plasma cells had type I or II antigenicdeterminants, we may wonder whether these de-terminants are present as well in the normaly-globulins, or whether they reflect abnormalitiesof malignant plasma cell metabolism. Two obser-vations indicate that types I and II antigenic de-terminants are present among normal y-globulins:antiserum prepared against normal 6.6 S y-globu-lins contains antibodies against types I and IIantigenic determinants (see above); and twotypes of molecules have been identified amongnormal 6.6 S y-globulins, 18 S y1-macroglobu-lins (40) and ,82A-globulins (24). Thus, neithertype I nor type II antigenic determinants repre-sent abnormalities peculiar to products of malig-nant plasma cell metabolism, but rather, reflecta selection from among the antigenic determinantspresent in the heterogeneous normal y-globulinpopulation.That all of 140 anomalous proteins contained

either type I or type II antigenic determinantsindicates that individual clones of plasma cellssynthesize either type I or II antigenic determi-nants. Observations in 20 patients with bothBence Jones proteins and an anomalous serumglobulin, showing that both proteins were of the

same antigenic type (I or II) in each case, sup-port the view that malignant plasma cell tumors(clones) are limited to only one antigenic formof -y-globulin.The possibility that the Bence Jones protein

represents a y-globulin subunit, identical with oneof the subunits composing the larger myelomaprotein molecule, is emphasized by the demonstra-tion that the type I or II antigenic determinantson Bence Jones proteins are identical with the de-terminants present in the corresponding mye-loma protein. This possibility, indicated by im-munochemical studies (8, 22, 23, 27, 28, 37, 38,41-44) and strengthened by the present obser-vations, also has been supported by the physico-chemical studies of Poulik and Edelman (45)and Edelman and Gally (46) demonstrating elec-trophoretic and other similarities between theBence Jones protein and the "L" chain of mye-loma protein from the same patients.Tumors forming y-myeloma proteins produce

type I molecules more frequently than type IIproteins. A possible explanation for the 2: 1 ra-tio of type I to type II y-myeloma tumors wouldbe random malignant transformation in a heter-ogeneous normal plasma cell population in whichtype I y-globulin-producing plasma cells outnum-bered type II cells by a ratio of 2: 1. Alterna-tively, type I protein-producing cells might bemore susceptible to malignant transformation, orindividual normal plasma cells might have the dualpotentiality of forming types I and II molecules,and might lose their type II synthetic capacitymore frequently on malignant transformation. Itis noteworthy, moreover, that the anomalous pro-tein in other y-globulin systems (e.g., yi-macro-globulins, 182A-myeloma proteins, and Bence Jonesproteins) also occur in type 1: 11 ratios not sig-nificantly different from the 2: 1 ratio amongy-myeloma proteins. Type I molecules also pre-dominate among the normal serum y-globulins(40), indicating that the predominance of typeI y-myeloma proteins is not a peculiarity of malig-nant plasma cells. This finding is also compatiblewith the view that malignant transformation oc-curs in a heterogeneous normal plasma cell popu-lation.The selective increase of type I or II y-globulin-

molecules is characteristic of multiple myeloma.Studies are under way to determine the diagnostic

820


value of immunochemical tests for discrete type Ior II globulin increases in the recognition of mul-tiple myeloma and macroglobulinemia.

SUMMARY

Two groups of 6.6 S y-myeloma proteins, dif-fering in antigenic composition, are described.About 67% of the y-myeloma proteins have typeI antigenic determinants; the remaining 33%have type II. The electrophoretic distribution,ultracentrifugal properties, hexose content, andproportion of homogeneous and heterogeneousmyeloma proteins are similar for types I and IIy-myeloma proteins.Two antigenically different types of 62A-mye-

loma proteins, y1-macroglobulins, and BenceJones proteins are also described.The antigenic determinants forming the basis

for identification of type I molecules appear to bethe same in all y-globulin classes, i.e., among theBence Jones proteins, y-myeloma proteins, /B2A-myeloma proteins, and yi-macroglobulins. Simi-larly, the type II proteins from each y-globulinclass share another set of common antigenic de-terminants.One hundred forty myeloma proteins, macro-

globulins, and Bence Jones proteins were exam-ined and found to have either type I or type IIantigenic determinants; no molecules had bothtypes. When the urinary Bence Jones proteinand the anomalous serum protein were comparedin 19 cases, the two proteins were found to beof the same antigenic type. This evidence indi-cates that clones of malignant plasma cells pro-duce either type I or type II molecules, but areunable to form both.

ACKNOWLEDGMENTS

The authors wish to thank Dr. L. Korngold, Hospitalfor Special Surgery, New York, N. Y., for testing repre-sentative Bence Jones proteins. Miss Carla McLaughlinprovided skillful assistance throughout the study.

REFERENCES

1. Kunkel, H. G., R. J. Slater, and R. A. Good. Re-lation between certain myeloma proteins andnormal gamma globulin. Proc. Soc. exp. Biol.(N. Y.) 1951, 76, 190.

2. Lohss, F., E. Weiler, and G. Hillmann. Myelom-Plasma-Proteine. III Mitteilung. Zur Immuno-

chemie der 'v-Myelom-Proteine. Z. Naturforsch.1953, 86, 625.

3. Lohss, F., and G. Hillmann. Myelom-Plasma-Pro-teine. IV Mitteilung. Zur Immunochemie dera- and fi-Myelomproteine. Z. Naturforsch. 1953,86, 706.

4. Slater, R. J., S. M. Ward, and H. G. Kunkel.Immunological relationships among the myelomaproteins. J. exp. Med. 1955, 101, 85.

5. Smith, E. L., D. M. Brown, M. L. McFadden, V.Buettner-Janusch, and B. V. Jager. Physical,chemical, and immunological studies of globulinsfrom multiple myeloma. J. biol. Chem. 1955, 216,601.

6. Deutsch, H. F., J. I. Morton, and C. H. Kratochvil.Antigenic identity of hyperglobulinemic serumcomponents with proteins of normal serum. J.biol. Chem.. 1956, 222, 39.

7. Korngold, L., and R. Lipari. Multiple-myeloma pro-teins. I. Immunological studies. Cancer 1956, 9,183.

8. Thije, 0. J. ten. Urinary and serum proteins inmyelomatosis. Acta med. scand. 1956, 153, 253.

9. Habich, H. Zur Antigenanalyse der Paraproteinebei Makroglobulinamien. Schweiz. med. Wschr.1953, 83, 1253.

10. Kanzow, U., W. Scholtan, and A. Miiting. Sero-logische Differenzierung von Makroglobulinamien.Klin. Wschr. 1955, 33, 1043.

11. Korngold, L., and G. Van Leeuwen. Macroglobu-linemia I. The antigenic relationship of patho-logical macroglobulins to normal 'y-globulins. J.exp. Med. 1957, 106, 467.

12. Morton, J. I., and H. F. Deutsch. Human serummacroglobulins and dissociation units. II. Im-munochemical properties. J. biol. Chem. 1958, 231,1119.

13. Reisner, C. A., and E. C. Franklin. Studies of mer-captoethanol-dissociated normal human 19 S'y-globulin and pathologic macroglobulins frompatients with macroglobulinemia. J. Immunol.1961, 87, 654.

14. Hauschka, T. S., B. J. Kvedar, S. T. Grinnell, andD. B. Amos. Immuno-selection of polyploids frompredominantly diploid cell populations. Ann. N. Y.Acad. Sci. 1956, 63, 683.

15. Burtin, P. etude Immuno-electrophoretique desproteines du myelome in Analyse Immuno-elec-trophoretique by P. Grabar and P. Burtin. Paris,Masson, 1960, p. 157.

16. Heremans, J. F., and M-Th. Heremans. Immuno-electrophoresis. Acta med. scand 1961, suppl. 367,27.

17. Osserman, E. F., and D. Lawlor. Immunoelectro-phoretic characterization of the serum and urinaryproteins in plasma cell myeloma and Walden-strom's macroglobulinemia. Ann. N. Y. Acad.Sci. 1961, 94, 93.


18. Korngold, L. Abnormal plasma components andtheir significance in disease. Ann. N. Y. Acad.Sci. 1961, 94, 110.

19. Bayne-Jones, S., and D. W. Wilson. Immunologi-cal reactions of Bence-Jones proteins. II. Dif-ferences between Bence-Jones proteins from vari-ous sources. Bull. Johns Hopk. Hosp. 1922, 33,119.

20. Robinson, S. H. G. An investigation of the anti-genic properties of four specimens of Bence-Jonesprotein obtained from cases of myelomatosis.Brit. J. exp. Path. 1927, 8, 454.

21. Hektoen, L., and W. H. Welker. Immunologicaldifferences of crystalline Bence-Jones proteins.Biochem. J. 1940, 34, 487.

22. Korngold, L., and R. Lipari. Multiple-myelomaproteins. III. The antigenic relationship of BenceJones proteins to normal 'y-globulin and multiple-myeloma serum proteins. Cancer 1956, 9, 262.

23. Burtin, P., L. Hartmann, R. Fauvert, and P. Grabar.etudes sur les proteines du myelome. I. Ptudecritique des techniques d'identification de la pro-teine de Bence-Jones et de leur valeur diagnostique.Rev. franc. Atud. clin. biol. 1956, 1, 17.-

24. Franklin, E. C. Two types of 'yA-globulin in serafrom normals and patients with multiple myeloma.Nature (Lond.) 1962, 195, 392.

25. Fahey, J. L. Studies of 'y and 8,2-globulins: com-parison of S and F (papain) fragments of mye-loma proteins from inbred mice. J. Immunol. Inpress.

26. Fahey, J. L. Electrophoretic properties of gammaglobulins. Studies of subunits of human y-mye-loma proteins. In preparation.

27. Franklin, E. C., and D. R. Stanworth. Antigenicrelationships between immune globulins and cer-tain related paraproteins in man. J. exp. Med.1961, 114, 521.

28. Askonas, B. A., and J. L. Fahey. Enzymaticallyproduced subunits of proteins formed by plasmacells in mice. II. j32A-myeloma protein and BenceJones protein. J. exp. Med. 1962, 115, 641.

29. Fahey, J. L. Heterogeneity of 'v-globulins. Advanc.Immunol. 1962, 2, 41.

30. Fahey, J. L. Chromatographic studies of anoma-lous 'y-, 82A- and macroglobulins and normal'y-globulins in myeloma and macroglobulinemicsera. J. biol. Chem. 1962, 237, 440.

31. Flodin, P., and J. Killander. Fractionation of hu-man-serum proteins by gel filtration. Biochim.biophys. Acta (Amst.) 1962, 63, 403.

32. Fahey, J. L., and C. L. McLaughlin. In preparation.33. Fahey, J. L. Heterogeneity of myeloma proteins.

J. clin. Invest. 1963, 42, 111.34. Putnam, F. W., C. W. Easley, L. T. Lynn, A. E.

Ritchie, and R. A. Phelps. The heat precipita-tion of Bence-Jones proteins. I. Optimum condi-tions. Arch. Biochem. 1959, 83, 115.

35. Goulian, M., and J. L. Fahey. Abnormalities in se-rum proteins and protein-bound hexose in Hodg-kin's disease. J. Lab. clin. Med. 1961, 57, 408.

36. Shetlar, M. R., and Y. F. Masters. Use of thymol-sulfuric acid reaction for determination of carbo-hydrates in biological material. Analyt. Chem.1957, 29, 402.

37. Mannik, M., and H. G. Kunkel. Classification ofmyeloma proteins, Bence Jones proteins, andmacroglobulins into two groups on the basis ofcommon antigenic characters. J. exp. Med. 1962,116, 859.

38. Migita, S., and F. W. Putnam. Antigenic relation-ships of Bence Jones proteins, myeloma globulins,and normal human 'y-globulin. J. exp. Med. 1963,117, 81.

39. Fahey, J. L. Structural basis for the differencesbetween type I and II human gamma globulinmolecules. J. Immunol. In press.

40. Fahey, J. L. Two types of 6.6S 'y-globulins and 18S'v1-macroglobulins in normal serum and 'y-micro-globulins in normal urine. J. Immunol. In press.

41. Lohss, F., and J. Wollensak. Zur Natur der Bence-Jones Proteine. Z. Naturforsch. 1959, 14B, 328.

42. Deutsch, H. F., C. H. Kratochvil, and A. E. Reif.Immunochemical relation of Bence-Jones proteinsto normal serum proteins. J. biol. Chem. 1955,216, 103.

43. Stiehm, E. R., J. I. Morton, and H. F. Deutsch.Some immunologic properties of papain digestfragments of multiple myeloma proteins. J. Im-munol. 1960, 85, 337.

44. Olins, D. E., and G. M. Edelman. The antigenicstructure of the polypeptide chains of human'y-globulin. J. exp. Med. 1962, 116, 635.

45. Poulik, M. D., and G. M. Edelman. Comparison ofreduced alkylated derivates of some myelomaglobulins and Bence-Jones proteins. Nature(Lond.) 1961, 191, 1274.

-46. Edelman, G. M., and J. A. Galley. The nature ofBence-Jones proteins. Chemical similarities topolypeptide chains of myeloma globulins and nor-mal 'y-globulins. J. exp. Med. 1962, 116, 207.

822

Documents

yl-MACROGLOBULINS, AND BENCE JONES PROTEINS IDEN