[CANCER RESEARCH 36, 458-466, February 1976]

induced tumors of rodents are known to shame commontumor-associated antigens (22). “Blocking―factors (eithercirculating tumor antigen on immune complexes) in sena ofmelanoma patients (10, 11, 24—27,31) might be detected bya serological reagent with specificity for common melanoma-associated antigens. If specific antisera to commonmelanoma-associated antigens became available, the isolation and purification of melanoma-associated antigens (4,7,14,28,41)mightbe facilitatedby absorptionontoantibody-coated affinity columns. Purified melanoma antigensand specific antisera might be used as reagents for theidentification and quantitation of serum-blocking factors bymadioimmunoassay.Such testingmight be usefulin thediagnosis of patients with tumors of unknown primaryand/or might aid in the follow-up of patients known to havemelanoma.

Therefore, we attempted to produce, in animals, antiserato malignant melanoma, the specificity of which could bedetermined by a simple, rapid, and sensitive serologicalmicrocytotoxicity test (18, 20). We elected to raise antiserain animals for the following reasons: (a) carcino-embryonicantigen had been detected using xenogeneic antisera (21);(b) uniformly high titers of cytotoxic antibodies can beexpected in xenogeneic antisera; (C) theme is no need todepend on selected human donors, and attempts by Metzgar (33) and ourselves to demonstrate melanoma-specificantibodies in allogeneic semafrom patients with malignantmelanoma have been disappointing.

MATERIALS AND METHODS

Tumor Cell and Fibroblast Cultures. Seven melanomacell lines and 2 adenocamcinoma cell lines were establishedin tissue culture by primary explant from surgical specimens obtained from 7 patients with malignant melanoma, apatient with gastric carcinoma, and a patient with camcinoma of the colon. A 2nd carcinoma of the colon cell line,HT-29M, was kindly provided by Dr. E. Bloom, Departmentof Microbiology and Immunology, UCLA, Los Angeles,Calif. The ovarian carcinoma cell line, 2008, was obtained,in 1972, from the USC-LA County Medical Center, Los Ange

les, Calif. BT-20 breast carcinoma cells were provided byContract E-73-2001-No. 1 within the Special Virus-CancerProgram, NIH, USPHS, through the courtesy of Dr. W. Nelson-Rees.

The tumor cell lines were designated as follows and studiedinthepassagesindicatedinbrackets:(a)melanomas,

SUMMARY

Complement-dependent cytotoxic antibodies to commoncell surface antigens of cultured melanoma cells were produced in guinea pigs. At appropriate dilution, melanomaantisera were cytotoxic only to melanoma target cells. Following absorption with pooled lymphoid cells, additionalabsorption with melanoma cells but not absorption withfibroblasts on carcinoma cells was found to remove all cytotoxic activity from melanoma antisera. Absorption with human fetal skin cells but not with autologous fetal visceralcells was found to remove cytotoxicity from melanoma antisera. Tissue type-specific antigens may be shared by humanmalignant melanomas and fetal skin of black racial origin(at 16 to 18 weeks of gestation). The methods may be useful.in the production of xenogeneic antisera with “operationalmonospecificity―for common melanoma-specific antigens.Sera from 47 patients with malignant melanoma failed toevidence specific cytotoxicity for melanoma target cells.

INTRODUCTION

In malignant melanomas, as well as in many other humancancers, the existence of common tumor-associated antigens, shared by tumors of similar histological type, hasbeen demonstrated by tests for cell-mediated immunity (1,4,7,8,11,13,14,16,26—28,32,37—39),by immunofluomescence (16, 30, 35, 36, 38, 42, 46), and by complementfixation (23, 34). The nature of these common tumom-associated antigens has yet to be elucidated (12), although (e.g.,in human colon carcinomas) common tumor-specific antigens of the embryonal type have been demonstrated (21)and their complex glycoprotein structure well characterized(43).

Further serological studies in malignant melanoma mayprovide a clue as to the genetic origin of the informationcoding for the presence of common melanoma-associatedcell surface antigens, since human melanoma cell lineshave been shown to contain particles with some propertiesin common with RNA tumor viruses (40) and since viral

1Supported, in part, by Grant CA-14846 from the National Institutes ofHealth.

2 Present address: 0-69 Heidelberg, Medizinische Universitãts-Klinik,Bergheimerstr. 58, West Germany.

3 To whom reprint requests are to be addressed, at Harbor General Hospi

tal, 1000 West Carson Street, Torrance, Calif. 90509.Received July 14, 1975; accepted November 3, 1975.

458 CANCER RESEARCH VOL. 36

Production of Antisera with Specificity for Malignant Melanomaand Human Fetal Skin1

Dieter Fritze,2 David H. Kern, Carol R. Drogemuller, and Yosef H. PiIch3

Department of Surgery, UCLA School of Medicine, Harbor General Hospital, 1000 West Carson Street, Torrance, California 90509

on May 18, 2021. © 1976 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Melanoma-specific Antisera

ED-MEL [10—30],@ROB-MEL [7-251, AL-MEL [20—30],CAMER-MEL [5-10], VEG-MEL [7-15], ROBINS-MEL [15-20],DAIR-MEL [25—30];(b) gastric adenocarcinoma, RI-CA [40—70]; (c) ovarian carcinoma, CA-2008[16—25];(d) adenocancinomas of the colon, HT-29M [110—130],WIL-CA [20—30];(e)breast carcinoma, BT-20 [43—50].

Fibroblasts were grown in tissue culture following primary explant from biopsies of normal skin and muscleobtained from melanoma patients ROB, VEG, and lungcancer patient PEARS. Fibmoblast cultures were studied inPassages 5 to 10. All cultures were maintained in RoswellPark Memorial Institute Medium 1640, supplemented with209kfetal calf serum, 0.159k glutamine, 0.25 mg amphotenicm B perml,100lUpenicillinperml,and 100 @gstreptomycm per ml (all from Flow Laboratories, Los Angeles, Calif.).All cell cultures were found to be free of contamination withbacteria and mycoplasma.

Human Sera. Semafrom 47 malignant melanoma patientsat various stages of disease, sena from 7 patients with cancemsother than melanoma, semafrom 5 melanoma patientswho had been treated for 8 to 12 months by alternate weeklyimmunotherapy with BCG and a melanoma cell vaccine(derived from a single melanoma of patient ED), and semafrom patient ED (with progressively growing metastaticmelanoma) were obtained from the Division of Oncology,Department of Surgery, UCLA School of Medicine.

Xenogeneic Antisera. Antisera were prepared in a sepamategroup of 3 to 5 adultfemaleHartleyguineapigsbyimmunization with tumor cells from confluent tissue culturemonolayers. Single cell suspensions of tumor cells wereprepared by 5- to 10-mm exposure to 0.25% trypsin followedby 3 washings in Roswell Park Memorial Institute Medium1640. Tumor cells (10@)suspended in 1 ml of Roswell ParkMemorial Institute Medium 1640 were mixed with an equalvolume of complete Freund's adjuvant and injected intoeach of the4 footpads.Controlguineapigswere immunized with complete Freund's adjuvant serum only. Senawere collected from each group of guinea pigs 2 weeks afterimmunization, heat inactivated for 30 mm at 56°,and frozenin aliquots at —76°.In early experiments, antisera werepooled within each experimental group. In later expeniments, selected antisera from individual guinea pigs wereused. Antisera were designated according to the tumor celltype used for immunization, e.g., ED-MEL, ROB-MEL, RICA, and CA-2008 serum.

Absorption of Xenogeneic Antisera. In the initial seriesofexperiments, 0.3 ml of pooled guinea pig serum (diluted1:20 in Roswell Park Memorial Institute Medium 1640) wasincubated with counted numbers of packed viable cells in agynatory shaking water bath at 37°for 60 mm. Packed cellswere obtained from tissue cultures of the various tumorsand fibroblasts as described above. In addition, lymphoidcells from 12 healthy human donors were isolated on FicollIsopaque gradients (6), pooled, and used for absorption.

4 The abbreviations used are: MEL, melanoma; ED-MEL, ROB-MEL, AL

MEL. CAMER-MEL, VEG-MEL, ROBINS-MEL, DAIR-MEL, cell culturesfrom 7malignant melanomas; RI-CA, CA-2008, HT-29M, WIL-CA, BT-20, cell cultures from carcinomas of the stomach , ovary, colon •and breast; BCG,Bacillus Calmette-Guérin;HL-A, histocompatibility antigen.

A black male fetus was obtained at 16 to 18 weeks ofgestation immediately after 0.9% NaCI solution-inducedabortion. Under sterile conditions the fetus was freed fromadherent placental tissue. No gross abnormalities could bedetected. Skin and adherent muscles were dissected fromthe skeleton, cut in fine pieces, pressed gently through 40and then 80 mesh stainless steel sieves, and washed 5 timesin Hanks' balanced salt solution. The resulting single cellsuspension was designated as “fetalskin.―Separately, asingle cell suspension was prepared from a pool of all fetalinternal organs by the same procedure, and designated as“fetalviscera.―Fresh fetal cells were used for absorption.

In a 2nd series of experiments, 0.5 ml of a melanomaantiserum (ROB-MEL serum) and 0.5 ml of an ovarian cancinoma antiserum (CA-2008 serum) were absorbed separatelywith 0.5 ml of packed human lymphoid cells derived fromthe pool of human donors described above. Following thisabsorption, 0.3 ml of ROB-MEL serum and 0.3 ml of CA2008 serum (each diluted 1:25) were additionally absorbedwith counted numbers of packed cells derived from culturesof tumors on fibmoblasts.

Microcytotoxicity Test for Cytotoxic Antitumor Antibody.The method used was a modification of the procedure oniginally reported by Cohen et a!. (9) and has been described indetail previously (18, 20). Briefly, 3 to 5 x 10' viable tumorcells or fibroblasts prelabeled with [‘25l]iododeoxyumidinewere placed in each well of a Falcon Microtest II plate(Falcon Plastics, Oxnard, Calif.) and incubated overnight toallow the cells to adhere to the well bottoms. The nextmorning, medium was aspirated from the wells and 40 @Iofserum were added to each well followed by incubation for 1hr at 37°.The serum was gently removed by aspiration, and40 @Iof pooled lyophilized rabbit complement (diluted 1:13)(Hyland Laboratories, Costa Mesa, Calif.) were added toeach well. Complement alone was not cytotoxic at thisdilution (17). After 1 hr of incubation with complement, theplates were washed twice, aspirated dry, and sprayed with aplasticfilm.Each serum dilutionwas testedin4 or6 replicate wells. The individual wells of each plate were separatedwith a band saw, and residual radioactivity in each well (ameasure of the number of surviving cells) was measuredwith a y counter. Means and S.D.'s for each set of replicatewells were calculated. Results are expressed in terms ofpercentage of cytotoxicity calculated according to the formula:

% cytotoxicity

= cpm control serum— cpm test serum@

cpm control serum

Controls in all experiments included diluted rabbit complement. In calculating the percentage of cytotoxicity, wechose to compare the cpm recorded following incubationwith test sera to the cpm obtained following incubation withdiluted rabbit complement alone in order to permit thedetermination of cytotoxic activity in control (i.e. , normal)sena. Experimental groups were compared for significanceusing Student's t-test for unpaired data.

FEBRUARY 1976 459

on May 18, 2021. © 1976 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

D. Fritze et a!.

noma antiserum (RI-CA) was found again to be cytotoxiconly to RI-CA target cells (19).

Then we studied whether absorption with melanoma cellswould remove more cytotoxic activity from melanoma antisera than would absorption with pooled lymphocytes orgastric carcinoma cells. These results are summarized inChart 2. Prior to absorption, both melanoma antisera (EDMEL, ROB-MEL) were cytotoxic to target cells from 3 melanomas. Whereas gastric carcinoma cells or pooled lymphocytes failed to remove significant cytotoxicity, comparablenumbers of cells from 5 different melanoma cultures uniformly reduced the cytotoxicity of both melanoma antiserasignificantly. The melanoma cells used in these absorptionexperiments could have removed cytotoxicity simply by expressing stronger species and/or HL-A antigens than theother tissue cells. Therefore, we studied whether more cytotoxicity was removed from both melanoma antisera by melanoma cells than by autologous skin fibroblasts. The resultsof this experiment are illustrated in Chart 3. Whereas absorption with gastric carcinoma cells or pooled Iymphocytes failed to remove more than 20% cytotoxicity, in mostcytotoxicity tests (5 of 6), absorptions with autologous melanoma cells were found to remove significantly more cytotoxicity from both melanoma antisera than did identical numbers of autologous fibroblasts. In 1 experiment, fibroblastsand autologous melanoma cells removed similar quantitiesofcytotoxicity.Inthisinstance,theantiserummighthavebeen too dilute to demonstrate different absorption between melanoma cells and autologous fibroblasts. Mostimportant, significantly more cytotoxicity was removedfrom ROB-MEL antiserum by absorption with ROB-MELmelanoma cells than by ROB fibmoblasts in tests on ROBMEL melanoma target cells.

Since melanoma cells might share antigens with humanfetal tissue, both melanoma antisera were absorbed withincreasing numbers of cells derived from either skin andmuscle or internal organs of a fetus of black racial origin.The results obtained are illustrated in Chart 4. Absorptionwith fetal skin cells as well as melanoma cells was found toremove all cytotoxicity, whereas absorption with fetal visceral cells or pooled lymphocytes removed significantlyless cytotoxicity from both melanoma antisera.

We attempted to obtain similar results by absorbing antisena first with pooled lymphoid cells to remove antispeciesand/ or anti-HL-A reactivity, followed by additional absorption with cells from cultures of fibroblasts, carcinomas, ormelanomas. The results of this 2nd series of experimentsareillustratedinCharts5 to9.

Xenogeneic antisera were collected from individualguinea pigs following immunization with cells from each of3 melanomas or carcinomas. These sera were selected forhigh cytotoxicity to ED-MEL melanoma target cells. ThedatadepictedinChart5 illustratethat,atdilutionsof1:8and 1:32, all antisera, except the RI-CA antiserum, werefound to be cytotoxic to ED-MEL target cells. Cytotoxicitywas complement dependent, since heat-inactivated rabbitcomplement abrogated cytotoxicity completely. One melanoma antiserum (ROB-MEL) and 1 ovarian carcinoma antiserum (CA-2008) were each absorbed with packed pooledlymphoid cells at a cell to serum ratio of 1:1. Prior to andfollowing this absorption, both antisera were tested on tar

460 CANCERRESEARCHVOL. 36

RESULTS

Human Sera. Initially, semafrom 47 melanoma patients atvarious stages of the disease were screened for cytotoxicityto target cells from 3 different melanoma cultures. All serawere allogeneic with respect to the target cells. Most semafrom patients with melanoma, as well as semafrom 7 patients with various cancers other than melanoma, showedless than 20% cytotoxicity for all 3 melanomas. The serafrom only 4 melanoma patients were significantly cytotoxicfor all 3 melanoma cell lines. These senawere tested by themicrolymphocytotoxicity test on a panel of lymphocytesobtained from 33 healthy Caucasian donors and were foundto contain a broad spectrum of anti-HL-A neactivities (E.Zellem and Dr. A. L. Waiford, Department of Pathology,UCLA, personal communication).

Sera from 5 melanoma patients (including Stages I, II , andIII) were screened for cytotoxic activity on target cells from 3melanomas and 3 carcinomas prior to and during thecourse of immunotherapy with BCG and allogeneic melanoma cell vaccine (provided by Dr. Wendell Winters, Division of Oncology, Department of Surgery, UCLA). In 3 of the5 sera, cytotoxicity to ED-MEL target cells increased significantly during treatment with ED-MEL cell vaccine. However,autologous sera from patient ED, who was not vaccinatedwith his own melanoma cells, evidenced no change in cytotoxicity to target cells from his own tumor. Sera from the 2vaccinated patients who showed the most impressive increase in cytotoxicity to ED-MEL target cells evidenced nochange in cytotoxicity to target cells from 2 other melanomas. Therefore, the increase in cytotoxicity during immunotherapy With BCG and melanoma cell vaccine wasfound to be confined to target cells derived from the immunizing melanoma cell line. Since these results suggestedthat specific cytotoxic activity to common melanoma-associated antigens was not detected in sera of melanoma patients even during immunization with allogeneic melanomacell vaccine, these data are not presented in detail. We thenproceeded to study xenogeneic antisera.

Xenogeneic Antisera The characterization of the xenogeneic antisera was approached in 2 ways. Melanoma antiserawere diluted until they appeared to lyse melanoma targetcells selectively. At these and higher dilutions, the sera weretested prior to and following absorption with cells fromvarious tissues. In the second approach, antisera were absorbed with packed pooled human lymphoid cells to remove antispecies and anti-HL-A reactivity. Then, these serawere tested prior to and following additional absorptionwith cells from various tissues.

Initially, the cytotoxicity of antisera to 2 melanomas (EDMEL, ROB-MEL) and a gastric carcinoma (RI-CA) was testedon target cells from 3 melanomas and 3 carcinomas. Theresults obtained are illustrated in Chart I . Both melanomaantisera were cytotoxic to all 3 melanomas over a widerange of serum dilutions, whereas complete Freund's adjuvant serum and gastric carcinoma antiserum were completely noncytotoxic. At dilutions of 1:8 to 1:64, both melanoma antisera were also cytotoxic to the carcinoma targetcells. However, at 1:128 and higher dilutions, both melanoma antisera showed less than 20% cytotoxicity to all 3carcinoma target cells (Chart 1, arrows). The gastric carci

on May 18, 2021. © 1976 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

I I I I 5x)03OCpm(:200

I@@ I IOcpm 2x)O3

I00 Ocpm

TARGET CELLS

AL-MELED-MEL

:8 :32 1:128 1:51212,048

ROB-MEL

>.I-.

0

>(0

0I->.0

Chart 1. Percentage of cytotoxicity obtained with 2xenogeneic melanoma antisera (ED-MEL and ROBMEL), a gastric carcinoma antiserum (RI-CA), and anantiserum collected after immunization with completeFreund's adjuvant (CFAS). Target cells are derived from3 melanomas (top) or 3 carcinomas (bottom).

RI-CA (GASTRIC)HT-29M (COLON-CA) CA-2008(OVARY)

I-zw0

wa.

Ocpm

.---.CFASa—cRI-CA

o—oROB@MEL“—“ED-MEL

t@@i \

00 Ocpm

80

60

40

20

0

-20 510―

PERCENTCYTOTOXICITY

0 20 40 60 80 00I@ - I @_—H

4xIO3cpm 0

PERCENT CYTOTOXICITY

too 80 60 40F- ‘0 cpm

20 0

I H4x103

Absorbed wIIh@

I.@ RI-CA2 I0@POOLEDLYMPHOS

3 5x106ED-MEL4_ iO@ROBINS MEL5-5x106DAIRMEL6@ AL-MEL

7. 3xIO@ ROB-MEL

ED-MELSERUM :200ROB-MEL SERUM -200

5 x)O3cpm 0

-J_J _J

(@0

2

3

4567

a2xIO3cpm

:200

@-1-@_____ I@ NT. 2

:@ (:60@@ 160

Chart 2. Percentage of cytotoxicity obtained with 2 xenogeneic melanoma antisera, ED-MEL (right), and ROB-MEL (left), prior to and following absorptionwith cells from: (1) gastric carcinoma; (2) a pool of lymphocytes; (3 to 7) 5 different melanomas. Each serum was tested on melanoma target cell ED-MELROB-MEL and AL-MEL NT., not tested.

461

Melanoma-specific Antisera

:8 32 1:128 1512 :2048 8 :32 1:128 1:512 .2048

FEBRUARY 1976

on May 18, 2021. © 1976 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

‘I I

get cells derived from cultures of 2 melanomas (ED-MEL,ROB-MEL) on fibnoblasts (VEG). At dilution of 1:100 or1:200, the absorption with packed lymphocytes was foundto remove 40 to 60% of the cytotoxicity for all 3 target cells.The ovarian carcinoma antiserum became completely non

p<O.005 cytotoxic to all 3 target cells, whereas the melanoma antise

mumwas still significantly cytotoxic to ED-MEL melanomacells (Chart 6). When the ROB-MEL melanoma antiserumwas additionally absorbed with ROB-MEL cells, the cytotoxicity to ED-MEL target cells was removed completely. Bycontrast, absorption with 3 types of carcinoma cells failed

p<005 to reduce cytotoxicity further. Since the ovarian carcinoma

. antiserum already became noncytotoxic, it remained noncy

totoxic following additional absorption with ovarian cancinoma cells (Chart 6).

We then studied whether additional absorption of ROBp<O.O) MEL melanoma antiserum with melanoma cells would me

move more cytotoxicity than would additional absorptionwith fibroblasts. The results obtained are illustrated in Chart

p<O.02@ Additional absorption of ROB-MEL antiserum with cells

from 3 different melanomas was found to remove all cytotoxicity, whereas absorption with cells from 3 fibmobiast cul

p<üOO) tunes removed significantly less cytotoxicity.

Finally, we absorbed ROB-MEL antiserum additionalywith carcinoma cells, fibmoblasts, or melanoma cells. Priorto and following absorption, this antiserum (diluted 1:50)was tested on target cells from 2 recently established melanoma cell lines (CAMER-MEL, Passage 5; VEG-MEL, Passage 6). The results depicted in Charts 8 and 9 show thatfibrobiasts and carcinoma cells absorbed significantly lessadditional cytotoxic activity than did melanoma cells. Inthese absomptions, VEG fibmoblasts were autologous to theVEG-MEL melanoma target cells (Chart 9), and ROB fibmo

NUMBER OF CELLS USED FOR ABSORPTION

0 IO@ 06 IO@ o8

5.106

>-

0

0

0

>- __________________ ______________0zw0a.Lia.

D. Fritze et a!.

PERCENT ABSORPTION

-Ic? 9 19 29 39 i@O 59 60

-I I ED-MEL

@ 1.200

I@@@ /200

:@::@ â€N.S.

1:60

ROB-MEL

1:100

/:/OO

/:6O

I j I

I I I

SERA ABSORBEDWITH CELLS FROM:

— /0@RI-CAIO@POOLEDLYMPHOS

@-::::@--3x106ROB-SKIN@::@3x106ROB-MEL

Chart 3. Percentage of absorption obtained with each of 2 xenogeneicmelanoma antisera, ED-MEL (top), and ROB-MEL (bottom), using cells fromgastric carcinoma (RI-CA), pooled lymphocytes, autologous fibroblasts, orautologous melanoma cells (both from patient ROB). Each serum was testedon 3 melanoma target cells.

ED-MEL SERUM I200

0@ l0@ I0@ 08@ -@-@---@ -

5.106

‘00T°@P@ T0

Chart 4. Percentage of cytotoxicity obtained with 2 xenogeneic melanoma antisera, ED-MEL (top), and ROB-MEL (bettom), prior to and following absorption with increasing numbers of cells derived from human fetal skin and muscle (FetalSkin), fetal internal organs (Fetal Viscera), pooled lymphocytes (PooledLymphos.),or ED-MELmelanomacells. Eachserum was tested on target cells from 3 melanomas: ED-MEL(left), ROB-MEL (center), and AL-MEL (right).

ED-MEL SERUM 1:60

0 06 o@ 08

5' @06

0

•I.5xO@@@

ROB-MEL SERUM :60

Ls,io3

ROB-MEL SERUM1:200-i-O T°

80@

601

40j

20@ \\

01 3,0@

:@

tooOcpm

801

60@

40@@

2O@

013,103

ED-MEL

A—A Fe/cl Visce,'c j@ Fe/cl Skin

.—o Pooled 4vmpdios. 1 =::}Ed-MeI

ROB-MEL AL-MEL

TARGET-CELLS

462 CANCER RESEARCH VOL. 36

U) 0

ii

on May 18, 2021. © 1976 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

0‘C0I-0

>-C.)

I-zLUC)

LU0.

Melanoma-specific Antisera

sera of 47 patients at various stages of malignant melanomaproved unsuccessful. In only 4 of these sera, cytotoxicitywas detected for all 3 melanoma cell lines. Since these 4sera were found to contain a broad spectrum of anti-HL-Aneactivities, their cytotoxicity was most probably directedagainst the HL-A antigens previously identified in the 3melanomas (17). Lewis et a!. (30) and Bodurtha et a!. (5) meported that cytotoxic antibody to allogeneic melanoma cellsappeared infrequently in sena of melanoma patients. SinceMetzgar et a!. (33) also noted that sera from melanoma patients “oftenshow different serologic specificities, areusually low titered, and have not been useful reagents forcharacterizing melanoma-specific antigen(s),' â€we studiedmelanoma antisera produced in guinea pigs.

The characterization of the xenogeneic antisera was appnoached by 2 methods. In the 1st series of experiments,melanoma antisera were diluted until they were found tolyse melanoma cells selectively (Chart 1). At these dilutions,the antisera were tested on melanoma target cells prior toand following absorption with cells from a gastric carcinoma, a pool of normal human lymphocytes or cells fromeach of 5 different melanomas. We found that absorptionwith melanoma cells from all 5 lines reduced the cytotoxicactivity of 2 melanoma antisera significantly more than didabsorption with cells from the other tissues (Chart 2). Similamly,absorption with melanoma cells was more effective inremoving cytotoxicity from 2 melanoma antisera than wasabsorption with autologous fibroblasts (Chart 3). Thesefindings suggested that, following absorption, these antisena reacted preferentially with common melanoma-specificantigens expressed by cultured melanoma cells.

In addition, we noted that fresh cells from human fetalskin (and muscle) shared with cultured melanoma cells theability to remove more cytotoxic activity from melanomaantisera than could autologous fetal visceral cells or pooledlymphocytes(Chart4).Sincefetalskincellsand fetalvisceral cells had been obtained from the same fetus, HL-Aantigens probably do not account for the difference notedin absorption with both tissues. However, species and/onHL-A antigens might be expressed on fetal skin at an earlierstage of development or at higher concentration than oninternal organs. Fossati et a!. (15) reported that lymphocytes from 6 of 10 melanoma patients were cytotoxic toepithelial-like cells cultured from human embryos at 1 to 2months of gestation . However, this cell-mediated cytotoxicity was not specific for lymphocytes from melanoma patients. Fossati et a!. (15) found also that lymphocytes from 5of 5 melanoma patients failed to be cytotoxic to fibroblastlike cultures derived from lung and heart of 2 4- to 5-monthold fetuses. Thus, our observation that absorption With fetalcells from internal organs failed to remove cytotoxicity frommelanoma antisera might gain support by inspection ofearlierdata.

Our data are consistent with the hypothesis that tissuetype-specific antigens may be shared by malignant melanomas and black fetal skin. Fetal tissues of other races mustbe studiedbeforefirmconclusionscan be drawn.

In a 2nd series of experiments, xenogeneic antisera werefirst absorbed with pooled lymphoid cells to remove antispecies and/or anti-HL-A reactivity. Then, these antisera were

(MELANOMA) (CASTRIC (OVARY) (BREAST)

-2xO@

ED-MEL ROB-MEL AL-MEL:8 :32 :8 :32 I:B :32

R)-CA CA-20061:8 :32 :8 :32 I:@ :32

E@—

fliLfiED- MEL TARGET-CELLS

Chart 5. Percentage of cytotoxicity for ED-MEL target cells obtained withxenogeneic antisera to 3 melanomas (left) and 3 carcinomas (right) at dilutions of 1:8 and 1:32, using either active or heat-inactivated rabbit complement.

NO.OF CELLS FOR ABSORPTIONS

o---0CA-2008 SERUM :200

Absorbed with:CA-2008

blasts were autologous to the melanoma cells used to prepane the ROB-MEL antiserum (Charts 7, 8, and 9).

In summary, in all experiments, additional absomptionswith melanoma cells were found to remove significantlymore cytotoxicity from ROB-MEL melanoma antiserum thandid absorption with cells from any of the other tissues.

DISCUSSION

Initial attempts to demonstrate cytotoxic antibodies tocommon melanoma-associated cell surface antigens in

0‘C

0I-.

I-zLU

LU0.

PACKEDLYMPHOS.

9 ((:() (@O@ 5.IO@ (X@O6 5@@6

cpm5OT(xIO@

.—. ROB-MEL SERUM :200

Absorbed wIth:

I WIL-CA (COLON)2 RI- CA (GASTRIC)3 CA-2X8 (OVARY)4 ROB-MEL

40-

30-

20-

0-

0-

ED-MEL TARGET-CELLSChart 6. Percentage of cytotoxicity for ED-MEL target cells obtained with

xenogeneic melanoma antiserum (ROB-MEL) or xenogeneic ovarian carcinoma antiserum (CA-2008), each diluted 1:200, prior to and following absorption with equal volumes of packed pooled lymphoid cells and followingadditional absorption with cells from 3 carcinomas or ROB-MEL melanoma.CA-2008 antiserum was absorbed additionally with CA-2008 cells.

FEBRUARY 1976 463

on May 18, 2021. © 1976 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

PACKEDLYMPHOS.

0 (1:1) IxlO@5xIO@ 1*106 5x@(6I@ I I I

0. Fritze at a!.

NO.OFCELLS FORABSORPTIONS:

NO. OF CELLS FOR ABSORPTIONS

@pm—I@iO@

ROB-MEL SERUM :200

Chart 7. Percentage of cytotoxicity for ED-MEL target cells obtamed with xenogeneic melanoma antiserum ROB-MEL prior to andfollowing absorption with pooled lymphoid cells and following additional absorption with increasing numbers of cells from 3 culturesof fibroblasts or melanomas.

PACKED0 LYMPIIOS.)@IO@5@)O@ ( @iO658)06

I : ,@ i

50

40

@ 30

@:20C)

zLU

@ 0LU0.

0@

-10

1 ROB2VEG @5KIN3 PEARS SFIBROBLASTS

0—0 ROB1 MELANOMAA—― EDo—o AL @CELLS

Absorbedwith:

@ Absorbedwith:

ED-MEL TARGET-CELLS

LROB-MEL SERUM :50

80-

Chart 8. Percentage of cytotoxicity for CAMER-MEL target cells@ 7O(Passage 5) obtained with xenogeneic melanoma antiserum ROB- @3MEL prior to and following absorption with pooled lymphoid cells(PACKED LYMPHOS.) and following additional absorption with in- e—creasing numbers of cells from cultures of 3 fibroblasts, 2 cardnomas, or ED-MEL melanoma.

z 5OLU

40

30 lxl0@if@cpm

additionally absorbed with cells from cultures of fibroblasts,carcinomas, and melanomas (Charts 5 to 9). Although onecannot be certain that all antispecies and anti-HL-A reactivity was eliminated from these antisera by absorption withpooled lymphoid cells, our results show reductions in cytotoxicity of from 40 to 60%. In fact, at identical dilutions, anovarian carcinoma antiserum became completely noncytotoxic to 2 melanoma cell lines and 1 fibroblast line, whereasa melanoma antiserum continued to exhibit significant cytotoxicity for melanoma target cells (Chart 6). Moreover, additional absorptions with melanoma cells were found to memove significantly more cytotoxic activity from this melanoma antiserum than did absorptions with cells from any ofthe other tissue types (Charts 7 to 9).

Although serological cross-reactivity based on reactions

Absorbedwith:

I ROBf SKIN

V@ F/BRO&ASTS3 PEARS

4 CA-2008i'OVARY)

5 W/L-CA (COLON)6 ED-MEL (MELANOMA)

to membrane antigens related to fetal calf serum (29) cannot be entirely ruled out, we suggest that this explanation isunlikely. All tumor cells and fibroblasts were cultured underidentical conditions in medium containing 20% fetal calfserum, but the cross-reactivity was observed only amongthe melanomas. The cross-reactivity could be due to common HL-A antigens or other non-HL-A normal cell surfaceantigens. However, we have demonstrated previously thatthe 3 melanomas most often used in this study (ED-MEL,ROB-MEL, AL-MEL) have individual HL-A antigens withoutany major cross-reacting specificity being shared by any 2of these cell lines (17).

Although new alloantigenic systems have been describedon human lymphocytes (45), the HL-A system is the strongest histocompatibility system in man (44), and may be identi

ED-MEL

CAMER- MEL TARGET-CELLS

464 CANCER RESEARCH VOL. 36

on May 18, 2021. © 1976 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Melanoma-specific Antisera

NO.OFCELLSFORABSORPT)ONS ficity―for the detection of common melanoma-associatedantigens. Specific cytotoxic antibodies to these antigenswere detected in guinea pig antisera by absorption withmelanoma cells and by failure of absorption with fibroblastsand carcinoma cells. Tissue type-specific antigens may beshared by human malignant melanomas and skin fromblack fetuses (at 16 to 18 weeks of gestation) and mayexplain the serological cross-reactivity among melanomas.ROB-MEL SERUM :50

-I.70

60

>.

@ 50C)‘C

0@:‘0C-)

zLUC)

LU0.

20

(0

0 (.7.(O@cpm

1 ROB , SKIN2 VEG@ F/BRO8LASTS3 PEARS

4 CA-2X18 (0 VANY)5 WIL-CA (COLON)6 ED-MEL (MELAROMA)

VEG-MEL TARGET-CELLS

Chart 9. Percentage of cytotoxicity for VEG-MEL target cells (Passage 6)obtained with xenogeneic melanoma antiserum ROB-MEL prior to and following absorption with pooled lymphoid cells (PACKED LYMPHOS.) and following additionalabsorptionwith increasingnumbersof cellsfrom culturesof 3fibroblasts, 2 carcinomas, or ED-MEL melanoma.

Absorbedwith:

cal with on closely related to the human species antigens(2). Therefore, weakennon-HL-A normal antigens are pnobably not responsible for the cross-reactivity between melanomas observed in this study. Moreover, pooled lymphoidcells were used to absorb xenogeneic melanoma antiseraextensively. Blood Group A-like substances have been detected in some human tumors (3), but it appears unlikelythat the serological cross-reactivity among all melanomacell cultures should be due to the same blood group antigens. We cannot muleout that these serological reactionsmight be due to antigens specified by a virus, since virusparticles with some properties in common with RNA tumorviruses have been detected in some human melanoma celllines (40).

Our results essentially confirm and extend the obsenvations of Metzgan et a!. (33). Following immunization of 2monkeys with 2 human melanoma cell lines, Metzgan et a!.(33) detected (by complement-dependent cytotoxicity andmixed agglutination) specific antibodies to common melanoma-associated membrane antigens in these xenogeneicantisera. Although absorptions with human fetal tissueswere not done, they speculated that common melanomaassociatedantigensmightbe ofthefetaltype.

In previous studies, we demonstrated that common cellsurface antigens might be shared by carcinogen-inducedmunine sarcomas (20). By using similar methods to produceand absorb xenogeneic tumor antisera, we have now detected common tumor-associated antigens on cultured human melanoma cells and provided some evidence for theirpossible relationship to antigens of fetal skin. We suggestthat our methods may be useful to produce and to quantitate xenogeneic melanoma antisera with “operationalspeci

FEBRUARY 1976 465

ACKNOWLEDGMENTS

We wish to thank Hildy Heinkel, Department of Illustration, UCLA, Schoolof Medicine, for her expert assistance In the preparation of the figures.

REFERENCES

1 . Baldwin, R. W., Embleton. M. J., Jones, J. 5. P., and Langman, M. J. 5.Cell-Mediated and Humoral Immune Reactions to Human Tumors. Intern. J. Cancer, 12: 73—83,1973.

2. Batchelor,J. R. . and Sanderson, A. R. HL-A Substances:PropertiesandImmunization of Rabbits. Transplant. Proc., 1: 489—490,1969.

3. Bloom. E. T., Fahey,J. L., Peterson, I. A., Geering, G., Bernhard, M., andTrempe, G. Anti-tumor Activity in Human Serum: Antibodies DetectingBlood-Group-A-Like Antigen on the Surface of Tumor Cells In Culture.Intern. J. Cancer, 12: 21—31,1973.

4. Bluming, A. Z., Vogel, C. L., Ziegler, J. L. , and Kiryabwire, J. W. M.

Delayed Cutaneous Sensitivity Reactions to Extracts of Autologous Malignant Melanoma—A Second Look. J. NatI. Cancer Inst. , 48: 17—24,1972.

5. Bodurtha, A. J., Chee, D. 0., Laucius, J. F., Mastrangelo, N. J., andPrehn, R. T. Clinical and Immunological Significance of Human Melanoma Cytotoxic Antibody. Cancer Res., 35: 189—193,1975.

6. Boyum, A. Separation of Leukocytes from Blood and Bone Marrow.Scand. J. Clin. Lab. Invest. 21(Suppl.): 97, 1968.

7. Char, D. H., Hollinshead, A., Cogan, D. G., Ballintine, E. J., Hogan M. J.,and Herberman, R. B. Cutaneous Delayed Hypersensitivity Reactions toSoluble Melanoma Antigen. New EngI. J. Med., 291: 274—277,1974.

8. Cochran, A. J., Jehn, U. W., and Gothoskar, B. P. Cell-Mediated Immunityin Malignant Melanoma. Lancet, 1: 1340—1341, 1972.

9. Cohen, A. M., Burdick, J. F., and Ketcham, A. S. Cell-Mediated Cytotoxicity: AnAssayUsing“'l-lododeoxyuridine-LabeledTargetCells.J. lmmunol., 107: 895—898,1971.

10. Currie, G. The Role of Circulating Antigen as an Inhibitor of TumourImmunity in Man. Brit. J. Cancer, 28: 153—161,1973.

11. Currie, G. A., and Basham, C. Serum Mediated Inhibition ofthe Immunological Reaction of the Patient to His Own Tumor: A Possible Role ofCirculating Antigen. Brit. J. Cancer, 26: 427—438,1972.

12. Davies, D. A. L. What Are Tumour Specific Antigens? Nature. 254: 653—654, 1975.

13. DeVries, J. E., Rumke, P., and Bernheim, J. L. Cytotoxic Lymphocytes inMelanoma Patients. Intern. J. Cancer, 9: 567—576,1972.

14. Fass, L., Herberman, R. B., Ziegler, J. L., and Kiryabwire, J. W. M.Cutaneous Hypersensitivity Reactions to Autologous Extracts of Malignant Melanoma Cells. Lancet, 1: 116—118,1970.

15. Fossati, G., Canevari, S., and Della Porta, G. Brief Communication:Cellular Immunity in Cancer Patients against Embryonal Cells. J. NatI.Cancer Inst., 51: 667—669,1973.

16. Fossati, G.. Colnaghi, M. I., Della Ports, G., Cascinelli, N., and Veronesi,U. Cellular and Humoral Immunity against Human Malignant Melanoma.Intern. J. Cancer, 8: 344—350,1971.

17. Fritze, D., Kern, D. H., Drogemuller, C. A., and PlIch, V. H. New MicroMethod for the Detection of HL-A Antigens on Cultured Human TumorCells. Transplantation, 20: 211—218,1975.

18. Fritze, D., Kern, D. H., and Pilch, V. H. Brief Communication: Quantitation of Cytotoxic Antitumor Antibody in vitro: A Microassay with “el..Iododeoxyuridine as a Tumor Cell Label. J. NatI. Cancer Inst., 53: 1403—1407,1974.

19. Fritze, D., Kern, D. H., and PlIch, V. H. Serologic Evidence for Crossreacting Tumor-Associated Antigens in Two Chemically Induced MurineSarcomas and in Human Malignant Melanomas. Behring Inst. Mitt. •56:90—97,1975.

20. Fritze, D., Kern, D. H., Waldman, S. R., and Pilch, V. H. SerologicEvidence for Cross-Reacting Antigens in Two Carcinogen-Induced Murine Sarcomas. Intern. J. Cancer, 15: 116—131, 1975.

21 . Gold, P. , and Freeman, S. 0. Demonstration of Tumor Specific Antigensin HumanColonicCarcinomatabyImmunologicalToleranceandAbsorption Techniques. J. Exptl. Med., 121: 439—463,1965.

on May 18, 2021. © 1976 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

22. Gross, L. Oncogenic Viruses, Ed. 2. Oxford: Pergamon Press, 1970.23. Gupta. R. K., and Morton, D. L. SuggestiveEvidence for in vivo Binding

of Specific Antitumor Antibodies in Human Melanomas. Cancer Res., 35:58—62,1975.

24. HellstrOm, I., and HellstrOm, K. E. Cell-Mediated Immune Reactions toTumor Antigens with Particular Emphasis on Immunity to Human Neoplasms. Cancer, 34: 1461—1468,1974.

25. HellstrOm, K. E., and HellstrOm, I. Lymphocyte Mediated Cytotoxicity andBlocking Serum Activity to Tumor Antigens. Advan. Immunol., 18: 209—277,1974.

26. HellstrOm, I., Warner, G. A., HellstrOm, K. E., and SjOgren, H. 0. Sequential Studies on Cell-Mediated Tumor Immunity and Blocking SerumActivity in Ten Patients with Malignant Melanoma. Intern. J. Cancer, 11:280—292,1973.

27. Heppner, G. H., Stolbach, L., Byrne, M., Cummings, F. J., McDonough,E., and Calabresi, P. Cell-Mediated and Serum-Blocking Activity to Tumor Antigens in Patients with Malignant Melanoma. Intern. J. Cancer,11: 245—260,1973.

28. Hollinshead,A. C., Herberman, R. B., Jaffurs, W. J., Alpert, L. K., Minter,J. P., and Harris, J. E. Soluble Membrane Antigens of Human MalignantMelanoma Cells. Cancer, 34: 1235-1243, 1974.

29. Irie, R. F., Irie, K., and Morton, D. L. Natural Antibodyin Human Serumtoa Neoantigen in Human Cuftured Cells Grown in Fetal Bovine Serum. J.NatI. Cancer Inst., 52: 1051—1058,1974.

30. Lewis, M. G., Ikonopisov, R. L., Nairn, R. C., Phillips, T. M., HamiltonFairley, G., Bodenham, D. C., and Alexander, P. Tumor-SpecifIc Antibodies in Human Malignant Melanoma and Their Relationship to the Extentof the Disease. Brit. Med. J., 3: 547—552,1969.

31. Lewis, M. G., Phillips, T. M., Cook, K. B., and Blake, J. Possible Explanation of Loss of Detectable Antibody in Patients with Disseminated Malignant Melanoma. Nature, 232: 52—53,1971.

32. Lieberman, A., Epstein, W., and Fudenberg. H. H. ImmunopathologicalChanges in Patients with Cutaneous Malignant Melanoma followingIntratumoral Inoculation of BCG: Correlation with Cell-Mediated Immunity. Intern. J. Cancer, 14: 401—416,1974.

33. Metzgar, R. 5., Bergoc, P. M., Moreno, M. A., and Seigler, H. F. BriefCommunication: Melanoma-Specific Antibodies Produced in Monkeysby Immunization with Human Melanoma Cell Lines. J. NatI. Cancer Inst.,50: 1065—1068,1973.

34. Morton, D. L., Eilber, F. R., and Malmgren, R. A. Immune Factors inHuman Cancer: Malignant Melanomas, Skeletal and Soft Tissue Sarcomas. Progr. Exptl. Tumor Res., 14: 25—42,1970.

35. Morton, D. L., Malmgren, R. A., Holmes, E. C., and Ketcham, A. S.Demonstration of Antibodies against Human Malignant Melanoma byImmunofluorescence. Surgery, 65: 233—240,1968.

36. Muna, N. M. , Marcus, S. , and Smart, C. Detection by Immunofluorescence of Antibodies Specific for Human Malignant Melanoma Cells.Cancer, 23: 88—93,1969.

37. Nagel, G. A., Piessens, W. F., Stilmant, M. M., and Lejeune, F. Evidenceof Tumor-Specific Immunity in Human Melanoma. European J. Cancer,7: 41—47, 1971.

38. Nairn, R. C., Nind, A. P. P., Guli, E. P. G., Davies, D. J., Little, J. H.,Davis, N. C., and Whitehead, R. H. Antitumor Immunoreactivity in Patients with Malignant Melanoma. Med. J. Australia, 1: 397—403,1972.

39. Oettgen, H. F., Aoki, T. Old, L. J., Boyse, E. A., Dc Harven, E., and Mills,G. M. Suspension Culture of a Pigment-producing Cell Line Derived froma Human Malignant Melanoma. J. NatI. Cancer Inst., 41: 827—831, 1968.

40. Parsons, P. G., Gross, P. , and Pope, J. H. Detection in Human MelanomaCell Lines of Particles with Some Properties in Common with RNATumorViruses. Intern. J. Cancer, 13: 606—618,1974.

41. Relsfeld, R. A., Pelligrino, M. A., and Kahan, B. D. Salt Extraction ofSoluble HL-A Antigens. Science, 172: 1134—1136,1971.

42. Rhomsdale, M. D., and Cox, I. 5. Human Malignant Melanoma Antibodies Demonstrated by Immunofluorescence. Arch. Surg., 100: 491—497,1970.

43. Rule, A. H., and Goleski-Reilly, F. Phase-specific Oncocolon Antigens: ATheoretical Framework for â€Carcinoembryonic Antigen―Specificities.Cancer Res., 34: 2083—2087,1974.

44. Walford, R. L. The Isoantigenic Systems in Human Leukocytes: Medicaland Biological Significance. Series Haematol. 2: 1—96,1969.

45. Walford, R. L., Gosset, T., Smith, G. S., Zeller, E., and Wilkinson, J. ANew Alloantigenic System on Human Lymphocytes. Tissue Antigens, 5:196—204,1975.

46. Wood, G. W., and Barth, R. F. Immunofluorescent Studies of the Serologic Reactivity of Patients with Malignant Melanoma against Tumorassociated Cytoplasmic Antigens. J. NatI. Cancer Inst., 53: 309—316,1974.

466 CANCER RESEARCH VOL. 36

D. Fritze et a!.

on May 18, 2021. © 1976 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1976;36:458-466. Cancer Res   Dieter Fritze, David H. Kern, Carol R. Drogemuller, et al.   and Human Fetal SkinProduction of Antisera with Specificity for Malignant Melanoma

  Updated version

  http://cancerres.aacrjournals.org/content/36/2_Part_1/458

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

  .pubs@aacr.orgDepartment 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/36/2_Part_1/458To request permission to re-use all or part of this article, use this link

on May 18, 2021. © 1976 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from