the isolation and characterization of tumor-specific ...be a separation of protein from nucleic acid...

$: The Isolation and Characterization of Tumor-specific ...be a separation of protein from nucleic acid components (22). The unbound fraction, which contained the TSTA ac tivity, was$
[CANCER RESEARCH 36, 3518-3525, September 1976]

similarities to normal histocompatibility antigens (4, 7, 8,16, 24).

Although the rejection of transplantable tumors amongsyngeneic animals has been used profitably in the study ofanimal tumors, this type of experimental design has beenprecluded in humans by both ethical considerations and thelack of a syngeneic donor-host relationship. Thus, a number of in vitro assays of the host's cellular and humoralimmune responses against putative TSTA have been developed which are applicable to both animal and human tumorsystems (1, 13, 14). Because the TSTA appear to be onlyweakly immunogenic, the antitumor immune responsesevoked on the part of the tumor host are correspondinglyweak (12, 23). Nevertheless, these immune responses areapparently far more tumor specific than can be evoked bythe immunization of xenogeneic animals.

On the basis of the foregoing observations, the expemiments to be described were designed with the followingobjectives: (a) to use the host's antitumor immune responsein order to monitor the purification of the putative TSTA ofrodent and human tumor tissues; (b) to determine whetherany structural and/or immunochemical similarities exist between human and rodent TSTA; and (c) to determinewhether there are any immunochemical and/or structuralrelationships between the TSTA and normal histocompatibility antigens in either the rat or human.

Materials and Methods

Tumor Tissues

In the experiments involving rodents, the methylcholanthrene-induced sarcomas, MC-1 and MC-i 1, 2 TSTA-distinct tumors serially passaged in syngeneic male hoodedrats, were used (31). Early generations of MC-1 and MC-liwere stored in liquid nitrogen and removed at intervals forpassage in syngeneic hooded rats. Human malignant melanoma tissue and breast cancer tissue were obtained eitherat surgery or at autopsy and stored at â€”40Â°until processed.

Purification of the Putative TSTA of Rodent and Human

Tumors

Preparation of Tumor Cell Membranes. The tumor tissuewas thawed and finely minced with stainless steel scissors.The mince was suspended in 4 volumes of 0.25 M sucrosecontaining 0.05 M Tris, 2 mt@iCaCl2, and 2 mM MgCI2, pH7.5, at 4Â°.The tumor tissue was homogenized with an Ultra

Summary

Putative tumor-specific transplantation antigens (TSTA)from both a carcinogen-induced rodent tumor (MC-i ) and 2human tumors were purified. The antigens were solubilizedfrom the tumor cell membranes by limited papain digestionin a manner similar to that described for the isolation ofnormal histocompatibility antigens. The antitumor immuneresponse of the tumor-beaming host was used to monitor thepurification of the putative TSTA in both the rodent andhuman tumor systems. In the case of the rodent tumor, amajor step in the purification of the TSTA involved affinitychromatography on Sephamose beads coupled to autologous antitumom antiserum. A comparable procedure wasutilized in the purification of the TSTA from human tumorsby using affinity chromatography on anti-human @t32-micmoglobulin antiserum coupled to a solid phase. The data obtamed indicate that the TSTA of human tumors contains a@32-micmogIobulinchain that is immunochemically identicalwith, and very similar in size to, that found in normal humanhistocompatibility antigens. A subunit of similar size wasalso identified in the carcinogen-induced rodent tumor.These results suggest that the TSTA in both humans androdents may well be altered histocompatibility antigens.

Introduction

Investigations based on the rejection of experimentallyinduced tumors into previously immunized syngeneic recipients have established the existence of TSTA1 in most animal tumor systems studied (12, 23). Whereas the TSTA oftumors induced by the same virus are common, those induced by chemical carcinogens are most usually unique tothe individual tumor (23). In both cases, however, the TSTAsites of major functional importance, as concerns immunological interaction, would appearto be on the surface membrane of the tumor cell (2). In mumine tumor systems, atleast, a good deal of evidence has accumulated that suggests that the TSTA beamsimmunochemical and structural

I Presented at the Symposium â€œCancer and Chemistry' â€ as part of the

Fourth Conference on Embryonic and Fetal Antigens in Cancer, November 2to 5, 1975, Charleston, S. C. This work was supported in part by the MedicalResearch Council of Canada and the Cancer Society of Montreal.

2 Presenter.

3 The abbreviations used are: TSTA, tumor-specific transplant antigens;

PBS, phosphate-buffered saline containing 0.14 M NaCI-O.OiMsodium phosphate buffer, pH 7.3; $,m, $,-microglobulin; LAI, leukocyte adherence inhibition; NAI, nonadherence index; SDS, sodium dodecyl sulfate; TSA, tumorspecific antigens.

CANCER RESEARCH VOL. 363518

The Isolation and Characterization of Tumor-specificAntigens of Rodent and Human Tumors'

D. M. P. Thomson,2 P. Gold, S. 0. Freedman, and J. Shuster

Montreal General Hospital Research Institute, McGill University, Montreal, Quebec, Canada H3G 1A4

Research. on February 28, 2020. © 1976 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

Characterization of Tumor Antigens

turrax homogenizer as previously described (30, 31). Theprocess was monitored by phase-contrast microscopy, afterstaining with 0.2% trypan blue, to achieve the release ofmore than 90% of the nuclei, but with minimal disruption ofthese structures. The homogenate was then centrifuged for10 mm at 980 x g in 250-mI plastic bottles in a refrigeratedIEC PR 6000 centrifuge. The supemnatant fluid, which constituted two-thirds of the volume in each bottle, was decanted, and the pellet was resuspended in an equal volumeof the same buffer. Both the homogenization and centrif ugation steps were repeated as described above, and thedecanted supemnatants were pooled. An aliquot of 3 M lithium chloride in distilled water was gradually added to thetotal supemnatant to a final concentration of 0.1 M. Themembranes in the supemnatant were sedimented by centrifugation for 1 hr at 75,000 x g in a fixed-angle rotor in a MSESuperspeed-65 centrifuge (19). The supemnatant fluid wasdiscarded, and the crude membrane preparation was suspended in 0.5 MTris buffer, pH 7.5, containing 0.1 M lithiumchloride, to a volume equal to about one-half of the originalpooled supemnatant. The pellets were then homogenized toa smooth suspension with the Ultraturrax homogenizer. Theresuspended crude membranes were again sedimented at75,000 x g for 1 hr. The supemnatant was discarded and themembrane pellet was suspended in 0.01 M Tris-HCI, pH 7.5,containing 1 unit of penicillin and 1 @gof streptomycin perml. The volume was adjusted to approximately 200 ml andstirred overnight at 4Â°.The membranes were sedimented at120,000 x g for 1 hm,and this fraction was then separated bycentrifugation through a dextman-polyethylene glycolaqueous 2-phase system adapted from that described byBrunette and Till (3). On completion of this centrifugationstep, the membranes were found at the interface of the 2-phase system. The membrane fraction was collected, diluted with approximately 4 times its volume of cold distilledwater, and sedimented by centrifugation at 75,000 x g for 1hr. The membrane pellet was resuspended in cold distilledwater and the centmifugation step was repeated. The pelletswere then resuspended in cold PBS, pH 7.3, and stored atâ€”40Â°.

Papain Solubilization of Membrane Protein. The purifiedmembranes were suspended in PBS, pH 7.3, at a proteinconcentration of approximately 10 to 12 mg/mI, as determined by the procedure of Lowry et al. (20). The suspensionwas incubated at 37Â°in a water bath, and twice-crystallizedpapain (Worthington Biochemical Corp., Freehold, N. J.)preactivated in cysteine was then added to a final concentration of0.5 unit per mg of membrane protein. The incubation was continued for 1 hr at 37Â°,during which period thesuspension was stirred continuously. Proteolysis was amrested by the addition of iodoacetamide to a final concentration of 0.05 M. The suspension was rapidly cooled in anice bath and was then centrifuged at 120,000 x g for 3 hr at4Â°.The supernatant recovered after centrifugation was dialyzed overnight at 4Â°against a buffer containing 0.02 MTris-0.38 M NaCI, pH 8.0, and was then concentrated byultrafiltration in an Amicon chamber fitted with a PM-10membrane, to a final volume of approximately 10 ml.

Chromatographic Procedures in the Purification ofTSTA. The soluble tumor membrane components were sub

jected to chromatography on a DEAE-Sephadex A-SOcolumn equilibrated with the buffer described above. lon-ex

change chromatography on DEAE-Sephadex removes contaminants of highly negative charge, and there appears tobe a separation of protein from nucleic acid components(22). The unbound fraction, which contained the TSTA activity, was then dialyzed against a 0.1 M Tmis-0.3 M glycine

0.2 M NaCi buffer, pH 8.0, containing 0.001 M disodiumEDTA, and was then concentrated by ultrafiltration in anAmicon chamber as described above. The concentrate wascentrifuged at 120,000 x g for 1 hr, and the supernatant wasthen subjected to molecular sieve chromatography on acalibrated Sephadex G-iSO column equilibrated with thesame Tris-glycine-NaCI buffer described above. The fractions obtained were concentrated by ultrafiltration and assayed for tumor antigenic activity.

Affinity Chromatography of the Tumor Antigenic Preparations. The â€˜y-giobulinfraction from the sera of syngeneichooded rats, hyperimmunized to the MC-i sarcoma as previously described (30, 31), was precipitated by 50% saturated ammonium sulfate. This material was covalently coupied to cyanogen bromide-activated Sephamose4B in 0.2 Mcitrate buffer, pH 6.5 (6, 31).

Antiserum to human f32m was raised in horses, usingpurified f32m kindly provided by Dr. M. D. Poulik of theWilliam Beaumont Hospital, Royal Oak, Mich. Serial bleedings were obtained from these animals, and their serumantibody activity against @32mwas measured by radioimmunoassay as previously described (28). At the earliest evidence that these animals were producing anti-/32m antibody, presumably of relatively low affinity, the y-globulinfraction of the horse serum was obtained by batch chromatogmaphy on Whatman DE-52 DEAE-ceiiulose equilibratedwith 0.01 M phosphate buffer at pH 7.6. This material wasthen coupled to Sephamose4B using cyanogen bromide asdescribed above (6, 31).

The TSTA-containing fractions obtained from the Sephadex G-1 50 columns described above were subjected to affinity chromatography on an appropriate antibody-boundcolumn. After the column effluent had been thoroughlywashed with 0.01 M Tris-0.14 M NaCI buffer, pH 7.4, theantibody-bound fraction was then eluted with 2.5 M MgCI2,pH 6.8, and 3 M KCNS for the rodent tumor antigens and thehuman tumor antigens, respectively. The MC-1 TSTA wastrace-labeled with 1251before application to the affinity column. The material that did not adhere to the affinity columnwas designated as the unbound fraction, while the materialthat was bound to the column and then eluted with thechaotmopic agent was designated as the bound fraction.

The dialyzed eluate was centrifuged for 1 hr at 75,000 x gand concentrated by ultrafiltration in an Amicon chamberfitted with PM-b membrane. The concentrated eluate wasagain centrifuged at 100,000 x g for 1 hr. This final centrifugation step removed any Sepharose particles that may haveescaped from the column and/or any denatured and aggregated proteins that had been eluted, both of which maynonspecifically interfere with the LAI assay for tumor antigenicity to be described below. Table 1 summarizes thesteps in the isolation of the putative rodent and humantumor antigens.

3@19SEPTEMBER1976



Extract used for inhibition of syngeneicanti-MC-1 immune serumâ€•Fluorescence

indexâ€•with MC-1 sarcoma

cellsPBS9.6Normal

tissueextract9.0Purified

membranesMC-i2.0MC-118.4Papain-soluble

MC-1'SephadexG-150Fraction18.0Fraction26.6Fraction34.0Fraction

47.1Papain-solubie

MC-ii'SephadexG-150Fraction18.5Fraction28.9Fraction38.3Fraction

48.8Affinity

chromatographyMC-1TSAEluate,bound4.6Effluentâ€•,

unbound8.9MC-11TSAEluate,bound9.1Effluent,

unbound9.0

D. M. P. Thomson et a!.

Assays for TSTA Activity. Indirect Membrane Immunofluorescence. Ind imect membrane immunofluomescence wasperformed on washed viable single tumor cell suspensionsprepared from finely minced solid tumor digested in 0.04%trypsin and 0.05% collagenase in the presence of a smallamount of DNase (30). The presence of antigenic activity inthe papain-soluble tumor digest of MC-1 membranes wasdetected and quantitated by the capacity to inhibit MC-itumor-immune antiserum from binding to cell surface antigens on viable MC-1 tumor cells. The fluorescence indexwas computed and expressed as previously described (30).

Tube LA! Assay. Isolated membranes, cell sap, and papain-solubilized membrane fractions obtained upon gelchromatography were assayed for human tumor antigenactivity by determining their capacity to inhibit the adherence of leukocytes in test tubes. The assay was performed,and the NAI was calculated and expressed as previouslydescribed (10, 21). The NAI represents the difference inreactivity to the specific and nonspecific tumor extracts.The higher the NAI, the greater is the differential in reactivitybetween the tumor and control tissue extracts, whereas alow NAI indicates minimal or no difference in reactivity. ANAI of 30 or more is indicative of a positive response, i.e. , asignificant inhibition of leukocyte adherence to glass basedon a large series of breast cancer, melanoma, and controlsubjects (10, 21). Leukocytes from patients with limitedcancer (Stage I or II) that reacted in the tube LAI assay to theappropriate PBS extracts of either breast cancer or melanoma were used to determine whether the isolated tumormaterials contained tumor-specific activity. Control subjects had either benign surgical diseases or unrelated cancers.

Electrophoretlc Analysis of Tumor Antigens

The papain-solubilized membrane preparations, obtainedas described above, and those isolated at each step ofpurification were subjected to assays for tumor antigen. Inaddition, those fractions containing tumor antigenic activitywere madiolabeled with 1251by the chloramine-T method (9)

Table 1Isolation of papain-solubilized tumor-specific antigens

1. Solid tumor, Ultraturrax homogenization2. Differential centrifugation

a. 980 x g for 10 mmb. 75,000 x g for1 hr

3. Membrane purification by aqueous 2-phase polymer system4. Papain digestion of Membranes, 0.5 unit papain activity per mg

membrane protein at 37Â°for 1 hr5. Icebath cooling6. Centrifugation, 120,000 x g for 3 hr7. Supernatantdialysis, 0.38 M NaCi-0.02MTris, pH 88. DEAE-Sephadex-A-50 ion-exchange chromatography, unre

tamed fraction9. SephadexG-150chromatography10. Affinitychromatography

a. Syngeneictumor-immune antiserumb. Xenogeneichorse anti-human f32m

11. Tumor antigenic activity monitored by in vitro assaysof antitumomimmunity

a. Inhibition of indirect membrane immunofluorescenceb. Tube LAI assay

and analyzed by SDS-poiyacrylamide gel electrophoresis in7.5 and 10% gels (32), and in 12.5% 8.0 M urea gels (29). Thegel was then cut into individual i-mm sections and countedin a Nuclear-Chicago y-scintillation spectrometer. Theseprocedures provided information concerning both the degmeeof purity of the preparations in question and the molecular weights of the molecules, or their subunits, involved.

Results

Purification of the Papain-solubilized Tumor Antigen ofthe MC-1 Rat Sarcoma. Table 2 reveals that the TSTA activity of the MC-1 tumor is present on the tumor cell membranes rather than in the cell sap. Moreover, this surfaceassociated tumor antigen was solubilized by limited papaindigestion and was partially purified by ion-exchange andmolecular sieve chromatography. From the data obtainedfrom molecular sieve chromatography on Sephadex G-i50,the maximum TSTA activity appeared to reside in the molecuiamweight range of 40,000 to 60,000 daltons (Table 2; Chart1). After affinity chromatography on the anti-MC-i affinity

Table 2

Isolation of MC-1 TSA monitored by inhibition of membraneimmuno fluorescence

aAnti-MC-1,immuneserumwasusedata finaldilutionof 1:10with tissue extracts.

b Number of cells stained with test serum divided by number of

cells stained with normal serum. All samples are read and scored asunknowns.

C Protein contents of fractions were approximately 8 mg/mI.

d MC-1 TSA-bound fraction contained a protein content of ap

proximately500pg/mI. Theeffluent contained a protein content of8 mg/mi.

3520 CANCERRESEARCHVOL. 36



to 20 30 40 50 60


5' 1 II++ + ++

20 41@1@1@1@2OO

FRACTION NUMIER

Chart 1. Sephadex G-150 chromatography. Papain-soluble membrane antigen from DEAE-Sephadex chromatography was applied to the column ofSephadex G-150 (5 x 87 cm) equilibrated with 0.3 M glycine-0.2 M NaCl-0.00iM disodium EDTA and 0.1 M Tris-HCI, pH 8.0; and 10-mI fractions werecollected and assayed for MC-i TSA activity by inhibition of membraneimmunofluorescence. Maximum MC-i TSA activity was in the area shown bythebar.

column, it was found that only the bound fraction containedtumor antigenic activity (Table 2; Chart 2). Hence, a gooddeal of the papain-solubilized membrane preparation thathad copumified with the TSTA to the point of molecular sievechromatography failed to bind to the specific antitumomantibody, appeared in the column effluent, and lacked tumomantigenic activity.

Papain-solubie membrane material from the unrelatedand non-cross-reacting MC-1i tumor applied to the MC-1affinity column showed no specific binding, since the matemialin the eluate did not inhibit in the membrane immunofiuomescenceassay (Table 2).

SDS-GeI Electrophoresis of Purified MC-1 TSA. The 125llabeled MC-i TSA eluted from the anti-MC-i affinity columnwas applied to a Sephadex G-200 column and eluted with asingle major peak with molecular weights of approximately45,000 to 60,000. However, an analysis of the 125l-IabeledMC-i TSA by SDS-polyacmyiamide gel electrophomesis mevealed that, in the presence of a reducing agent, 1251-IabeiedMC-i ISA appeared to be composed of 3 poiypeptidechains with apparent molecular weights of 12,000, 18,000,and 25,000 daltons (Chart 3). The major peak had an apparent molecular weight of 12,000 daltons, and this peak had amobility identical to that of â€˜25i-labeledhuman f32m.

isolation of Tumor Antigenic Material from HumanBreast Cancer and Malignant Melanoma. The development of the LAI assay in test tubes for the detection ofantitumom immunity in patients with breast cancer (10) andmalignant melanoma (21) has provided an immunologicaltechnique for monitoring the purification of the tumor antigens from melanoma and breast cancer preparations. Theresults shown in Tables 3 and 4 reveal that the tumor antigenic activity of these tumors, as monitored by LAI, wasfound in the purified tumor cell membrane preparation andnot in the cell sap. The material containing this activityeluted maximally in the early fractions from the SephadexG-i50 column, in the apparent molecular weight mangeof70,000 to iSO,000 daitons (Fraction 2) (Tables 3 and 4).However, when the LA1-active fractions were chromatographed on anti-human /32m affinity columns, only thebound fraction retained this activity (Tables 5 and 6). NoLAI-active material was found in the unbound fractions at

any time. Moreover, themewas no evidence of cmoss-meactivity of the tumor Ã¤ntigenic activity between the breast andmelanoma systems.

SDS-Polyacrylamide Gel Electrophoresis of IsolatedMelanoma and Breast Tumor Antigens. The papain-soiuble breast cancer and melanoma antigens in the 70,000- toi50,000-moleculam-weight fraction of a Sephadex G-150column were labeled with 1251by the chloramine-T methodand analyzed by SDS-polyacmylamide gel electrophomesis inthe presence of reducing agents. Both tumor antigens had

+

1.0

E

@ 0.5

0.1

0 0@

8 Sg

+ ++

00

U

SS

+

1.0

EC

0@ 0.5

0.1

FRACTION NUMBER

Chart 2. Affinity chromatography profile of the MC-1 TSA. The 1st largepeak is the unbound or effluent fraction. The column was then washed withapproximately 500 ml (10 times bed volume) of column buffer 0.14 M NaCI0.01 M Tris, pH 7.4. The bound material was eluted with 100 ml (2 times bedvolume) 2.5 M MgCI,-0.01 MTris-HCI, pH 6.8. A small peak was observed. TheMC-i TSA was always trace-labeled with 1251before application, and theradioactivity was measured in each 10-mI fraction. The profile of 1251radioactivity was identical to the A,,,, tracing.

0

U

S

+

+

15 @O 456075Distonc. Migrat.d(mm)

Chart 3. SDS gel electrophoresis of â€œI-labeledMC-i TSA eluted fromanti-MC-1 affinity column. Top, profile of the â€˜251-labeledMC-1 TSA rununreduced on 12.5% 8 PA urea gels; bottom, â€˜25l-labeledMC-1 TSA runreduced on 12.5% 8 M urea gels. Mobility of calibration proteins is indicated.The fastest moving peak of the â€˜251-labeledMC-1 TSA had a mobility identicalwith that of human $@mand was a broad peak. The control MC-11 TSA elutedfrom the MC-i affinity column had too few 1251counts to analyze. /3MG, 1@-microglobulin; BPB, bromphenol blue.

SEPTEMBER1976 3521

6

94

Ia-U,



Antigenic activity determined by tube LAI assay of papain-solubilized malignant melanomaantigenNAIâ€•

of malignant melanomaNAr' SephadexG-150chromatographicrfractionsPurifiedPBS

ex- momLeukocytesfromtractâ€• branes Cell sap1 234Melanoma

Patient 188 12585 115306MelanomaPatient 260 65 1875 60282Control

subject12 3 1515 1234Controlsubject1 16 10 28

Antigenic activity determined by tube LAI assay of papain-solubilized breast cancerantigenLeukocytes

fromNAIâ€•

of breast cancerNAIâ€• SephadexG-i 50chromatographiccfractionsPurified

PBS ex- momtractâ€• branes Cell sap1 234BreastcancerPatientl

BreastcancerPatient2Control subjectControlsubject92

58 870 41 104 0 21 3 i248

110 3178 115 25

0 13 815 22 126

91014

Antigenic activity of TSAof malignantmelanomaisolatedbyaffinitychromatography with anti-humanf3@mNA1â€•

to NAla affinity chromaPBSex- tographytract of

malignantLeukocytesfrom melanoma Unboundâ€•Boundâ€•Melanoma

Patient 3 89 11 69MelanomaPatient 4 84 17 112Control subject 5 â€”i6 â€”6

Antigenic activity of TSA of breast cancer isolated byaffinitychromatographywith anti-humanf3@mNAI'

to NAIâ€•affinity chromaPBSox- tographytract of

breast canLeukocytes from cer Unboundâ€•Boundâ€•BreastcancerPatient3

54 1763ControlSubject â€”9 â€”22 â€”28

D. M. P. Thomson et a!.

Table 3

No. of nonadherentcells in the presenceof specific tumor extractaNAI= â€”No of nonadherentcellsin the presenceof nonspecifictumorextract <

No. of nonadherentcells in the presenceof nonspecific tumor extract

A NAI of 30 or greater indicates significant inhibition of adherence.b A PBS extract of breast cancer was the nonspecific extract.

r Chromatographic fractions were concentrated to approximately 6 mg/mI and tested at a 1 :4 dilution in Medium

199.

Table 4

aA NAIof 30or greateris positive.b A PBS extract of the melanoma antigen was used as the nonspecific control extract.

C Chromatogmaphic fractions were concentrated to approximately 5 to 8 mg/mI protein and tested in tube LAI assay

at a 1:4 dilution in Medium 199.

Table 5 Table 6

a NAI was calculated with PBS extract of breast or with isolates of

soluble breastcancer as specific antigen and PBSextract of ovarian tumor as nonspecific control antigen. Human f32misolatedidentically from a hepatoma was also used as a nonspecific control,and the results remained similar. A NAI of 30 or greater is positive.

b Unbound and bound proteins were concentrated to approxi

mately 4 mg/mI and tested in the tube LAI assay at a proteinconcentration of 75 to 150 j.@g.

by SDS gel electrophoresis, showed the same 3 peaks of12,000, 25,000, and 40,000 daltons (Chart 6).

Approximately 70% of the material applied with TSTAactivity was bound to the anti-human f32m column. Afterapplication of the chaotropic agent, 70 to 90% of the boundfraction was recovered as determined by measurements ofhuman f32m or total protein. The unbound fraction contamed no, or minimal amounts, of human f32mas determined by radioimmunoassay for human f32m. In contrast,the bound fraction was considerably enriched in /32m.

a NAI was calculated with PBS extract of ovarian tumor as a

nonspecific control antigen. Humanf.12misolated identically from ahepatoma was also used as a nonspecific control, and the resultswere unchanged.

b Unbound and bound fractions were concentrated to approxi

mately 2 mg/mI and tested in the tube LAI assay at a proteinconcentration of 75 to iSO @g.

similar profiles on SOS electrophomesis (Charts 4 and 5).There were 2 major peaks; the peak with the fastest mobilitymoved identically with that for 1251-iabeled human f32m. Theminor intermediate peak had an electrophoretic mobility ofapproximately 25,000 daitons. The slowest component hadan approximate molecular weight of 40,000 (Charts 4 and5).

The material that bound to the anti-human /32m affinitycolumn and that had demonstrated LAI activity was alsolabeled with 1251The madiolabeled material, when analyzed

CANCER RESEARCH VOL. 363522



v-@ â€˜, â€˜@----â€”-@â€”1102030 40506014..Ai@/@\@

jf%@@

10@ f@ +

@ lb @[email protected]%8MUrea

13 30 45 60 75


Since the papain-solubilized TSA of the MC-1 tumorretains its antigenicity, as determined by its ability to neutralize MC-i tumor-immune serum in the membrane immunofluomescence assay, the technique of affinity chromatogmaphywas chosen to achieve selective isolation and pumification of the MC-i tumor-specific antigen. The Sephamoseanti-MC-i antibody column, unlike conventional techniques, could selectively isolate the tumor-specific moietyfrom a variety of other membrane components. The isolatedTSA of the MC-i rat sarcoma was approximately 50,000 to

0.U

2000

@ 15000.U

l000

500

2000

@ 1500

a-U

,,:;- 1000I,'

500

0 0

80

:i@@3

2 4 BPS

0

Ia-

U

7.5%

1.

3

1 2 4 BPB

(@ (

DISTANCE MIGRATED (mm)

Chart 5. SDS gel electrophoresis of the â€˜251-labeledpapain-solubilizedmelanoma antigen from Fraction 2 of Sephadex G-i50 chromatography. A,7,5% gel; B, 10% gel; C, 12.5% 8 M urea gel. The â€˜Â°â€˜I-Iabeledmelanomaantigen was run reduced. Mobility of calibration proteins is indicated: 1,albumin (68,000); 2, aldolase (40,000); 3, chymotrypsinogen (25,000); 4,cytochrome c (12,400). BPB, bromphenol blue.

4

600A

100

0

a.U

B

20 30 40 50 60Distance Migrated (mm)

Chart 4. SOS gel electrophoresis of the â€œ'51-labeledpapain-solubilizedbreast cancer antigen from Sephadex G-i50 chromatography (Fraction 2).Fraction 2 from the column with a molecular weight range of 70,000 to150,000 had maximum LAI activity. The 1251breast cancer antigen was runreduced on 7.5, 10, and 12.5% 8 M urea gels. The peak with the fastestmobility moved in a manner identical with the human f32mpeak. Mobility ofcalibration proteins is indicated. BPB, bromphenol blue.

Discussion

BPB

C

The results of the present investigation indicate that thetumor antigens of both a carcinogen-induced rodent tumorand those human tumors that were studied are present inthe tumor cell membrane and can be soiubilized by limitedpapain digestion in a manner similar to that used for thepurification of histocompatibility antigens from various species. Moreover, it was possible to use the host's antitumorimmune response as a method of monitoring the pumification of the tumor antigenic preparation. The data obtainedsuggest that, in the case of the human tumor studied, atleast, the surface tumor antigens involved contain a /32mchain that is immunochemically identical with and is verysimilar in size to that found in normal human histocompatibility antigens. A subunit of similar size was also identifiedin the carcinogen-induced rodent cancer.

15 30 45 60 75 90 105

SEPTEMBER 1976 3523

g,@ Â§.



D. M. P. Thomson et al.

4

3

0

Distance Migrated(mm)

Chart 6. SDS gel electrophoresis of breast cancer antigen eluted fromanti-human f32maffinity column. The eluted breast antigen was labeled with1251 and run reduced on 10% gel. Top, profile of the â€˜25l-labeled breast

antigen; bottom, profile of â€˜25I-labeledhuman [email protected] of calibrationproteins is indicated.

60,000 daitons as determined by Sephadex G-200 chromatogmaphy. On SDS-geis, when reduced, this material consisted of 3 subunits with molecular weights of approximately 11,000, 18,000, and 25,000 daltons. The 11,000-moleculamweight subunit had a mobility identical with that ofhuman f32m.

The isolated MC-i TSA of the chemically induced tumorthus appears to be composed of subunits with molecularweights similar to those described for the rat Ag-B histocompatibility antigens (17, 18).

Hailiday and Miller (11) have described a simple in vitrotechnique of detecting host antitumom immunity called LA1.Holan et a!. (15) described a modified version of this assayin matsthat was performed in test tubes. In our laboratorythis technique was adapted and modified for the detectionof sensitization to human breast cancer and malignant melanoma tumor antigens (10, 21). With this assay, leukocytessensitized to the tumor antigen respond by not adhering toglass. It appeared feasible, therefore, to use the same mesponding cells to monitor the purification of the sensitizingantigen, since, as long as some form of the sensitizingantigen was added to the test tubes, the sensitized cellsshould demonstrate the response of nonadhemence toglass.

The results of the present study indicate that the tube LAIassay can be used to monitor the isolation of tumor antigens and that tumor antigens isolated from human breastcancer and melanoma tissue have similar physicochemicalproperties. Both tumor antigens appear to be composed ofsubunits of similar size. SDS-polyacmylamide gel electrophometic studies revealed that both human breast and melanoma tumor antigens consisted of 3 molecular speciesweighing about 11,000, 25,000, and 40,000 daltons. Although the HLA molecules are now thought to be composedof 2 distinct subunits (f32m and a heavier ailoantigenicchain), similar structural components have been obtainedfor the HLA antigens isolated from human lymphoid cells

after papain digestion (5, 25, 26). Furthermore, @32m,the11 000-dalton polypeptide chain found in all histocompatibility antigens, appears to be an integral part of the breastand melanoma tumor antigens.

It appears that rodent and human TSTA have structuralsimilarities and, at least in man, the 11 000-dalton subunit isapparently f32m.Thus, the TSTA in rodents and humans isvery similar to the AG-B and HLA species-specific histocompatibility antigens. The tumor-specific polypeptide chainmay, in fact, represent a modified histocompatibility antigen. By analogy to the histocompatibility antigens, thelarger subunits therefore appear to carry the distinctivetumor antigenic determinant. The biological significance ofthese observations is compatible with the hypothesis (27)that histocompatibility molecules serve as adapters thatcombine with non-self-components that are recognized bycytotoxic T-lymphocytes. Current work in our laboratory isdesigned to establish a similarity, at the level of primarystructure, between the histocompatibility antigens and tumomantigens.

Acknowledgments

We thank Dr. J. Marti, Dr. N. Grosser, and Dr. Maximo Flores for testingthe fractions. We are grateful to J. Weatherhead, P. Friedlander, and B.Chaland for able technical assistance.

References

1. Baldwin, R. W. In Vitro Assays of Cell-Mediated Immunity to Human SolidTumors: Problems of Quantitation, Specificity, and Interpretation. J.NatI. Cancer Inst., 55: 745-748, 1975.

2. Baldwin, R. W., and Embleton, M. J. Detection and Isolation of TumourSpecific Antigen Associated with a Spontaneously Arising Rat MammaryCarcinoma. Intern. J. Cancer, 6: 373â€”382,1970.

3. Brunette, D. M., and Till, J. E. Rapid Method for Isolation of L CellSurface Membranes Using an Aqueous 2-Phase Polymer System. J.Membrane Biol., 5: 215-224, 1971.

4. Comoglio, P. M., Bestini, M., and Forni, G. Evidence for a MembraneCarrier Molecule Common to Embryonal and Tumour-Specific AntigenicDeterminants Expressed by a Mouse Transplantable Tumour. Immunology, 29: 353-364, 1975.

5, Cresswell, P., Springer, T., Strominger, J. L., Turner, M. J. , Grey, H. M.,and Kubo, R. T. Immunological Identity of the Small Subunit of HL-AAntigens and @32-Microglobulinand Its Turnover on the Cell Membrane.Proc.NatI.Acad.Sci.U.S.,71:2123-2127,1974.

6. Cuatrecasas, P. Protein Purification by Affinity Chromatography. Derivatizations of Agarose and Polyacrylamide Beads. J. Biol. Chem. , 245:3059-3065, 1970.

7. Editorial: Viruses and Histocompatibility Antigens: An Unexpected Interaction (from E. Lennox). Nature, 256: 7-8, 1975.

8. Gooding, L. R., and Edidin, M. Cell Surface Antigens of a Mouse Testicular Teratoma. Identification of an Antigen Physically Associated with H-2Antigens on Tumour Cells. J. Exptl. Med., 140: 61-78, 1974.

9. Greenwood, F. C., Hunter, W. M., and Glover, J. S. The Preparation ofâ€˜31l-LabelledHuman Growth Hormone of High Specific Radioactivity.Biochem.J., 89: 114-123, 1963.

10. Grosser, N., and Thomson, D. M. P. Cell-mediated Antitumor Immunityin Breast Cancer Patients Evaluated by Antigen-induced Leukocyte Adherence Inhibition in Test Tubes. Cancer Res., 35: 2571-2579, 1975.

11. Halliday, W. J. , and Miller, S. Leukocyte Adherence Inhibition: A SimpleTest for Cell-Mediated Tumour Immunity and Serum Blocking Factors.Intern. J. Cancer, 9: 477-483, 1972.

12. HellstrOm, K. E., and HellstrOm, I. Cellular Immunity against TumorAntigens. Advan. Cancer Res., 12: 167-223, 1969.

13. Herberman, R. B. Cellular Immunity to Human Tumor-Associated Antigens. A Review. Israel J. Med. Sci., 9: 300-307, 1973.

14. Herberman, R. B., and Oldham, R. K. Problems Associated with Study ofCell-Mediated Immunity to Human Tumors by Microcytotoxicity Assays.J. NatI. Cancer Inst., 55: 749-753, 1975.

15. Holan, V., Hasek, M., Bubenik, J., and Chutna, J. Antigen-Mediated

15 30 45 60

3524 CANCERRESEARCHVOL. 36




Macrophage Adherence Inhibition. Cellular Immunol., 13: 107-116,1974.

16. Invernizzi, G., and Parmiani, G. Tumour-Associated Transplantation Antigens of Chemically Induced Sarcomata Cross Reacting with AllogeneicHistocompatibility Antigens. Nature, 254: 713-714, 1975.

17. Katagiri, M., Natori, T., Tanigaki, N., Kreiter, V. P., and Pressman, D.Papain-Solubilized Ag-B Antigens. I. Isolation and Characterization ofTwo Components Composing Ag-B Antigens. Transplantation, 19: 230-239, 1975.

18. Katagiri, M., Tanigaki, N., and Pressman, D. Papain-Solubilized Ag-BAntigens. II. Characterization of Small Sized Ag-B Molecules. Transplantation, 20: 135-141 , 1975.

19. Kornfeld, A., and Siemers, C. Large Scale Isolation and Characterizationof Calf Thymocyte Plasma Membranes. J. Biol. Chem., 249: 1295-1301,1974.

20. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. , and Randall, R. J. ProteinMeasurement with the Folin Phenol Reagent. J. Biol. Chem., 193: 265-275, 1951.

21. Marti, J. H., and Thomson, D. M. P. Antitumor Immunity in MalignantMelanoma Assay by Tube Leukocyte-Adherence Inhibition. Brit. J.Cancer, 34: 1-18, 1976.

22. Miyakawa, V., Tanigaki, N., Yagi, Y., and Pressman, D. An EfficientMethod for Isolation of HL-A Antigens from Hematopoietic Cell Lines. J.Immunol., 107: 394-401, 1971.

23. Old, L. J., and Boyse, E. A. Immunology of Experimental Tumors, Ann.Rev. Med., 15: 167-186, 1964.

24. Ostberg, L., Rask, L., Wigzell, H., and Peterson, P. A. Thymus Leukae

mia Antigen Contains f32-Microglobulin. Nature, 253: 735-736, 1975.25. Nakamuro, K., Tanigaki, N., and Pressman, 0. Multiple Common Proper

ties of Human fJ,-Microglobulin and the Common Portion FragmentDerived from HL-A Antigen Molecules. Proc. NatI. Acad. Sci. U.S., 70:2863-2865, 1973.

26. Peterson, P. A. , Rask, L. , and Lindblom, J. B. Highly Purified PapainSolubilized HL-A Antigens Contain f@-Microglobulin. Proc. NatI. Acad.Sci.U.S.,71:35-39,1974.

27. Schrader, J. W., Cunningham, B. A. , and Edelman, G. M. FunctionalInteractions of Viral and Histocompatibility Antigens at Tumor Cell Surfaces. Proc. NatI. Acad. Sci. U. S., 72: 5066-5070, 1975.

28. Smith, M., Gold, P., Freedman, S. 0., and Shuster, J. Studies of theLinkage Relationship of Beta-2-Microglobulin in Man-Mouse SomaticCell Hybrids. Ann. Human Genet., 39: 21-31 , 1975.

29. Swank, A. T. , and Munkres, K. D. Molecular Weight Analysis of Oligopeptides by Electrophoresis in Polyacrylamide Gel with Sodium DodecylSulfate. Anal. Biochem., 39: 462-477, 1971.

30. Thomson, D. M. P., and Alexander, P. A Cross-Reacting EmbryonicAntigen in the Membrane of Rat Sarcoma Cells Which Is Immunogenic inthe Syngeneic Host. Brit. J. Cancer, 27: 27-47, 1973.

31. Thomson, D. M. P., Sellens, V. , Eccles, S. , and Alexander, P. Radioimmunoassay of Tumor Specific Transplantation Antigen of a ChemicallyInduced Rat Sarcoma: Circulating Soluble Tumor Antigen in TumorBearers. Brit. J. Cancer, 28: 377-388, 1973.

32. Weber, K., and Osborn, M. The Reliability of Molecular Weight Determinations by Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis. J.Biol. Chem. ,244: 4406-4412, 1969.

3525SEPTEMBER1976



1976;36:3518-3525. Cancer Res D. M. P. Thomson, P. Gold, S. O. Freedman, et al. of Rodent and Human TumorsThe Isolation and Characterization of Tumor-specific Antigens

Updated version

http://cancerres.aacrjournals.org/content/36/9_Part_2/3518

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/36/9_Part_2/3518To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



the isolation and characterization of tumor-specific ...be a separation of protein from nucleic acid...

Documents