[CANCER RESEARCH 49. 6840-68-«. December I. 1989]

Flow Cytometric Determination of the Frequency and Heterogeneity of Expressionof Human Melanoma-associated Antigens1

David Herd,' Meenhard Herlyn, Hilary Koprowski, and Michael J. Mastrangelo

Division of Oncology, Department of Medicine, Thomas Jefferson University, Philadelphia, Pennsylvania 19107 fD. B., M. J. M.¡,and H'istar Institute of Anatomy andBiology. Philadelphia. Pennsylvania 19104 ¡M.//.. //. A'./

ABSTRACT

We used flow cytometry to measure the expression of human melanoma antigens on cell suspensions dissociated from metastatic masses.The objective was to study the heterogeneity between tumor samplesfrom different patients and between different tumors excised from a singlepatient. Fifty-three métastasesexcised from 34 melanoma patients »ereanalyzed with a panel of nine murine monoclonal antibodies (MOABs).Melanoma cells were stained by an indirect fluorescent method andanalyzed on a Coulter EPICS C flow cytometer after gating to excludetumor-infiltrating leukocytes and dead cells. The most consistently andmost strongly expressed antigen was the high-molecular-weight proteo-glycan (detected by the MOAB 9.2.27), which was expressed on 95% ofthe melanoma specimens and by a high proportion of cells within eachspecimen (mean ±SE, 79.2 ±5.5). However, strong expression of thisantigen was limited to melanoma cells that had been dissociated mechanically and was markedly diminished by exposure to collagenase. Cultureof collagenase-dissociated tumor cells for 24 to 48 h resulted in reexpres-sion of the antigen. The expression of other melanoma-associated antigens was not affected by collagenase treatment, but for these antigensthere was more variability between cells from an individual tumor andbetween tumors from different patients. The percentage of enzyme-dissociated tumors considered positive for MOAB binding (defined as atleast 10% of cells positive) and the mean ±SE of the percentage ofpositive cells within a tumor were as follows: MOAB ME-9-61 (antigen,p97) = 84% + (41.2 ±5.4%); MOAB ME-20.4 (antigen, nerve growthfactor receptor) = 40% + (18.7 ±5.1%); MOAB ME-24 (antigen,ganglioside G„,)= 84% + (50.8 ±4.8%); MOAB ME-311 (antigen,ganglioside 9-O-acetyl-GD1) = 76% + (42.5 ±5.1%); MOAB ME-361(antigen, mainly ganglioside GD2)= 3% + (1.9 ±0.8%); MOAB 3F8(antigen, ganglioside GD2)= 36% (10.5 ±3.8%); MOAB l4G2a (antigen,ganglioside GD2)= 86% + (46.0 ±6.7%); MOAB L243 (antigen, HLA-DR) = 56% + (22.5 ±5.5%). In 19 cases, we were able to compare theantigenic profiles of two tumors excised from the same patient at differenttimes. Analysis by nonindependent t test showed no significant differencesin MOAB binding between the paired tumors. Moreover, linear regression analysis indicated that there was a linear relationship, with a slopeapproximately = 1, between the percentage of positive cells in Tumor 1versus Tumor 2. Thus, flow cytometric analysis of cells dissociated frommelanoma métastasesgives a different perspective than previous studiesperformed with melanoma cell lines or with tissue sections. In particular,this study suggests that intrapatient heterogeneity of expression ofMOAB-defmed melanoma antigens may not be as important as previouslyestimated.

INTRODUCTION

A number of murine MOABs' have been produced that

recognize antigens strongly associated with human melanomacells (1-4). For many of these antibodies, the antigens havebeen characterized and include such diverse chemical entitiesas high-molecular-weight proteoglycan (5), gangliosides (3),

Received 1/18/89; revised 5/17/89; accepted 8/28/89.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1Supported by Grant CA 39248 from the National Cancer Institute and byfunds from the Nat Pincus Endowment and from the Rae S. ÃœberTrust.

2To whom requests for reprints should be addressed.3The abbreviations used are: MOAB, monoclonal antibody: FITC. fluorescein

isothiocyanate.

and transferrin-like proteins (6).The degree to which melanoma cells within a tumor express

these antigens is not only of considerable theoretical interest,but also of practical importance. Heterogeneous expression ofmelanoma-associated antigens could be a limiting factor in theusefulness of the antibodies as diagnostic or therapeutic agents.Moreover, if those antigens could be made immunogenic, consistency of their expression would be crucial for their inclusionin vaccines. Finally, even if this is not possible, the heterogeneityof their expression might be indicative of similar heterogeneityof other melanoma-associated antigens that are capable ofinducing an antitumor immune response.

To date, much of the work on this subject has utilizedmelanoma cell lines (4) or tumor tissue sections that have beenfixed and stained by immunoperoxidase methods (7). Expression of antigens by established cell lines may not reflect that ofthe tumor cells from which they originated. Immunoperoxidasestudies are compromised by difficulty in quantitation and byfixation artifacts. In this study, we sought to avoid these pitfallsby analyzing melanoma cells dissociated from metastaticmasses and by using a quantitative assay method, flow cytom-etry. We show that binding of a panel of well-characterizedMOABs is quite heterogeneous between tumors from differentpatients, but that the heterogeneity does not appear to extendto different tumors from the same patient.

MATERIALS AND METHODS

Source of Tumors. Metastatic masses were excised from 34 patientswith melanoma. For 19 patients, a second mass was excised 2 to 25mo later, so that the total number of tumors examined was 53. All ofthe patients were treated with active immunotherapy, consisting of anautologous melanoma vaccine with cyclophosphamide pretreatment (8,9). The sites of the original tumor specimens were as follows: lymphnode, 18; subcutaneous, 11; lung, 2; liver, 1; and bowel, 2.

Preparation of Tumor Cells. We used a modification of the methodof Peters el a!. (10). Freshly excised tumor masses were trimmed ofskin, fat. and necrotic tissue and minced in cold Hanks' balanced salt

solution supplemented with 1% human albumin and 0.1 % EDTA. Cellsthat were released into the medium by mechanical dissociation wereput aside and stored separately. The minced tumor pieces were placedin an enzyme solution, consisting of collagenase (140 mg) and DNase(30 mg) in 100 ml of the same solution. The collagenase was type I(Sigma Chemical Co., St. Louis, MO) from Clostridium histolyticum;the DNase was type I (Sigma) extracted from bovine pancreas. Thedissociation process was carried out in baffled, trypsinizing flasks in a37°Cwater bath with constant stirring using magnetic stir bars and an

immersible stirrer. After 30 min of dissociation, the enzyme solutioncontaining the cell suspension was removed, and fresh enzyme solutionwas added. The dissociation process was continued until no visibletumor tissue remained. The total time that a tumor sample was exposedto collagenase/DNase varied from 60 to 90 min. The tumor cells werewashed twice in the Hanks' solution, resuspended in freezing medium

(RPMI 1640 plus 1% human albumin plus 10% dimethyl sulfoxide),and frozen in a controlled rate freezer (Union Carbide. Indianapolis,IN) at PC/min. They were stored in the liquid phase of liquid nitrogenuntil needed.

Monoclonal Antibodies. The source and specificity of the murineMOABs used are shown in Table 1.

6840

Research. on January 8, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEROGENEITY OF MELANOMA ANTIGENS

Table I Sources and specificities of monoclonal antibodies

MOABME-9-6!ME-20.49.2.27ME-24ME-3I1ME-361l4G2a3F8L243IsotypclgG2bIgGlIgG2aIgMIgG3IgGZa*IgG2aIgG3lgG2aAntigenp97

(mclanotransfcrrin)Nervegrowth factorreceptorHigh-molecular-weight

proteoglvcanGangliosideGD3rGanglioside9-O-acetyl-GD3GangliosideGD/GangliosideGD¡GangliosideGD2HLA-DRSourceWistar0WistarNCI*WistarWistarWistarScripp/Sloan-Kettering*Becton

Dickinson*ReferenceHerlyn

et al.(29)Rossera/.(13)Bumol

and Reisfeld(2)Thurinet al.(30)Thurinel al.(24)Thurinet al.(21)Mujooet at.(22)Cheunget al. (20)

" M. Herlyn, Wistar Institute of Anatomy and Biology, Philadelphia, PA.' I. Green, Biological Response Modifiers Program, National Cancer Institute.' Minor cross-reactivity with ganglioside GD2.d IgG3 variant also tested and gave identical results.' Minor cross-reactivity with ganglioside Go3-^R. Reisfeld, Scripps Clinic and Research Foundation, LaJolla, CA.* N-K. V. Cheung, Memorial-Sloan Kettering Cancer Center, New York, NY.* Becton-Dickinson Immunocytometry Systems, Mountainview, CA.

FACS Analysis. The melanoma samples were tested with the panelof anti-melanoma MOABs by a two-step procedure. Tumor cell suspensions were incubated with MOABs on ice for 30 min, washed twice,and resuspended in FITC-conjugated Fab fragment of affinity-purifiedgoat anti-mouse immunoglobulin (Cappel-Organon Teknika, WestChester, PA) for 30 min. Then, they were washed again and analyzedwith a Coulter EPICS C flow cytometer (Coulter Electronics, Hialeah,FL), equipped with a 488-nm argon laser (used at 500 mW). Onealiquot of cells was incubated with a MOAB that binds to all humanleukocytes (hybridoma HB12; American Type Culture Collection) toallow identification of leukocytes. Ethidium bromide (10 jig/ml) wasadded to all samples just before analysis to allow gating of dead cells.

Flow cytometry analysis was performed by a modification of themethod of Schroff et al. (11). Gates were set on the forward light scatterhistogram to include only cells in the size range of melanoma cells andto exclude debris, erythrocytes, and clumps. A second gate was set onthe red fluorescence histogram to exclude dead cells that had taken upethidium bromide. Next, the tumor cell sample stained with anti-humanleukocyte antibody was analyzed with a two-dimensional histogram offorward light scatter versus log of green fluorescence; the lower limit ofthe forward light scatter window was adjusted upward until it containedfewer than 5% leukocytes. Finally, the samples stained with anti-melanoma antibodies were analyzed on the green fluorescence channelusing logarithmic amplification. In the few cases in which the leukocytescould not be completely excluded from the melanoma cell light scattergate, we monitored a two dimensional plot of log green fluorescenceversus forward light scatter to be certain that any green positive cellsexhibited sufficiently high forward light scatter to preclude their beingleukocytes.

The percentage of cells "positive" for an antibody was determined

by the Immuno program developed by Coulter Electronics. The negativehistogram (tumor cells stained with FITC-goat anti-mouse only) issubtracted channel by channel from the positive histogram (FITC-goatanti-mouse plus MOAB) after the setting of a match range for the twohistograms. This method allows for the counting of dimly stained cells,which could be disregarded by arbitrarily setting a dividing line betweenpositive and negative populations. A tumor was considered "positive"for binding of a MOAB if at least 10% of the cells were "positive" by

the Immuno program.Statistics, t test for nonindependent samples and linear regression

analysis were performed on an IBM microcomputer with a statisticalsoftware package (Systat: Stats and MGLH modules; Systat, Inc.,Evanston, IL).

RESULTS

Enzymatically versus Mechanically Dissociated MelanomaCells. Melanoma cell samples obtained by enzymatic dissociation of metastatic masses from 34 patients were analyzed forbinding of the panel of 9 MOABs. Fig. 1 shows the percentageof cells within each tumor that bound each antibody. A highproportion of cells within most of the tumors expressed the

1DO

SO-80

a~70-S

60-°-

50-

J25

30-20-10-!•

$*'f

-f.

t.MOAB

9-81Antigenp97t-.t•t'20.4N»t

*.''•-.*'9.2.27HUWtti'tj24003|'•.

+±Î

V-311381AG03

GD2*t3F8C02+!..*14G20002i..;»¡L243DR

Fig. 1. Binding of MOABs to enzyme-dissociated melanoma cells. Each "+"

represents a tumor from a different patient. The dotted ( ) line connects themean values for each antibody.

antigens identified by MOABs ME-9-61 (p97), ME-24 (ganglioside GD3),ME-311 (ganglioside 9-0-acetyl-GD3), and 14G2a

(ganglioside GD2>.On the other hand, very few tumors werepositive for the antigens identified by 9.2.27 (high-molecular-weight proteoglycan), ME-361 (ganglioside GD2),and 3F8 (ganglioside GDJ). Intermediate levels of expression were observedfor MOABs against nerve growth factor receptor (ME-20.4)and HLA-DR, consistent with other studies (12, 13).

Since Schroffer al. (11) have shown that the high-molecular-weight proteoglycan is sensitive to collagenase, we consideredthe possibility that the enzymatic dissociation process may havecaused a loss of other antigens as well. For 22 of the melanomasamples, we were able to obtain sufficient quantities of cellsdissociated mechanically without exposure to collagenase/DNase, and the results of analyses of these cells are shown inFig. 2. As expected, the high-molecular-weight proteoglycanantigen identified by MOAB 9.2.27 was strongly expressed bymechanically dissociated cells, but the expression of the otherantigens was similar to that observed with enzymatically dissociated cells. Thus, the minimal binding of MOABs ME-361and 3F8 was unlikely to have been an artifact of the dissociationprocess. Fig. 3 is a summary graph showing the proportion ofenzymatically versus mechanically dissociated tumors considered "positive" for each MOAB, i.e., at least 10% of the cells

binding the MOAB as defined by the Immuno analysis.As previously shown by Schroff et al. (11), incubation of

enzymatically dissociated melanoma cells for 24 to 48 h intissue culture medium resulted in strong reexpression of theproteoglycan antigen, but not of the GD2-associated antigensidentified by MOABs ME-361 and 3F8 (data not shown).

6841

Research. on January 8, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEROGENEITY OF MELANOMA ANTIGENS

10090801

708

60°-50inÕ

+°-"3°-''

20-10-

0-•

|++\+•

$tUOAB

9-61Antlg«!p97+frt/.

+.'''•t'+20.4NGFg

+|+t&

1/*'•••+

".^Î'tî,9.2.27

24HUWGD3frt*+

+ftt

\+î

V311361AGD3

G024.j

+T•'

+'•t/

,'\+;+3F8

14G20L243002GD2 DR

Fig. 2. Binding of MOABs to mechanically dissociated melanoma cells. Each"+" represents a tumor from a different patient. The doited ( ) line connects

the mean values for each antibody.

•ICEQ TCM

20.4NGF

9.2.27 24 311 361 3F8 14G2oHUW GD3 AGD3 GD2 GD2 GD2

L243DR

Fig. 3. Comparison of MOAB binding to melanoma cells dissociated enzy-matically (TCE) versus mechanically (TCM). Tumors were considered "positive"

if at least 10% of the cells were positive for the antibody, as defined by Immunoanalysis.

Heterogeneity of Expression of Melanoma-associated Antigens. For 19 patients we were able to obtain and analyze asecond tumor specimen (Tumor 2) which arose 2 to 25 mo(median 4 mo) after excision of the original tumor (Tumor 1).In 17 cases. Tumor 2 developed during or shortly after a timewhen the patients were receiving immunotherapy with an autol-ogous melanoma vaccine that had been prepared with cells fromTumor 1 (8, 9). Thus, differences in expression of MOAB-identified antigens between Tumor 1 and Tumor 2 could havebeen due to "natural" heterogeneity or to variations induced by

successful immunization.Surprisingly, we observed little variation between paired tu

mors from the same patient. Analysis by a t test for noninde-pendent samples showed no significant differences for any ofthe MOABs tested (P > 0.10 for all). Linear regression analyseswere performed for each antibody by the following model

% (+) of cells in Tumor 2 = K + C x [% (+) of cells in Tumor 1]

where C is the regression coefficient, representing the slope ofthe regression curve, and K is a constant representing the Yaxis intercept.

As indicated above, the initial tumors were uniformly positivefor the antiproteoglycan MOAB 9.2.27 and uniformly negativefor anti-GD2 MOABs, ME-361 and 3F8. This pattern wasidentical for the second tumors, and therefore a close correspondence was obvious without regression analysis. The linearregression data for the other six MOABs are shown in Table 2,and the regressions curve for four of the MOABs are shown in

Table 2 Linear regression analysis data for binding of MOABs by cells frompaired tumors from the same patient

For each MOAB. the percentage of "+" cells in Tumor I was plotted againstthe percentage of "+" cells in Tumor 2 (see text).

AntibodyAntigenME-9-61

p97ME-20.4 Nerve growth factorME-24 GD3ME-311 O-Acetyl-Goj14G2a GD2L243 HLA-DRRegression

coefficient(slope)0.667

±0.17 10-*0.877 ±0.156*0.639 ±0.286''0.785 ±0.188*0.831 ±0.192'0.744 ±0.059*Constant

(y-intercept)13.959±8.934r

4.870 ±5.30916.001 ±15.7792.936 ±9.7003.177 ±11.802

-0.446+1.456"Mean ±SE.

*/>< 0.001.c For all. P > 0.20.d P < 0.05.

Fig. 4. For all of the MOABs there was a statistically significantlinear relationship between the percentage of cells binding theMOAB in Tumor 2 versus Tumor 1. Moreover, in no case wasthe slope significantly different from 1 or the constant significantly different from 0. Thus, for all 9 MOABs the percentageof cells binding Tumor 2 could have been accurately predictedto be close to the percentage that was bound by Tumor 1.

DISCUSSION

Flow cytometric analysis proved to be an efficient and reliablemethod of measuring the binding of a panel of well-characterized MOABs to melanoma cells dissociated from metastaticmasses. The results give a somewhat different perspective ofthe expression of these melanoma-associated antigens thanprovided by previously published studies.

It is apparent that the expression of these antigens is moreheterogeneous than indicated by assays of melanoma cell lines.For example, although the p97- and GDi-related antigens wereeach detected in 84% of tumors by MOABs ME-9-61 and ME-24, respectively, the percentage of positive cells within anindividual tumor varied widely, from 0 to 91.9% for p97 andfrom 0 to 96.9% for GD3-The more uniform expression of thesecell surface antigens in tissue culture (14) could be related toselection of cells that can adhere well to plastic. In fact, Chereshet al. (15) have shown that membrane ganglioside expression isnecessary for attachment of cells in vitro and that anti-ganglio-side antibodies inhibit such attachment. Another major difference between melanoma cells in vivo and in vitro is the muchhigher proliferative rate of the latter. Although the antigenrecognized by MOAB 9.2.27 is not cell cycle dependent (16),this may not be the case for other melanoma-associated antigens(12, 17).

The one antigen for which variability among patients' speci

mens was minimal was the high-molecular-weight proteoglycanantigen recognized by MOAB 9.2.27. All but one tumor waspositive, and the percentage of positive cells within a tumorwas consistently high (75.6 ±5.6%). A similar result wasreported by Morgan et al. (1), who also used flow cytometry tomeasure melanoma antigen expression. Those authors foundthat, in 26 of 30 melanoma specimens, more than 90% of thecells were positive for antibody 9.2.27. Lindmo et al. (18) andSchroffe? al. (11) have shown that the high-molecular-weightproteoglycan antigen may be reversibly removed from the cellsurface by trypsin or collagenase and that the antigen is regenerated after 48 h, an observation confirmed by us in this paper.It is possible that the sensitivity to collagenase was due tocontamination of the relatively crude collagenase preparation

6842

Research. on January 8, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEROGENEITY OF MELANOMA ANTIGENS

Fig. 4. Comparison of paired tumors fromthe same patient, linear regression analysis.Each panel shows a scattergram plot of thepercentageof MOAB-positive cells in the original tumor (Tumor 1) versusthat of a tumor(Tumor 2) excised 2 to 25 mo (median. 4 mo)later. For all four MOABs (ME-9-61. ME-24,ME-311, and 14G2a)the calculated regressioncunes were statistically significant (seeTable2).

-*-*10 20 30 « 30 »0 70 »0 »0 100

Tumor 1: % Cells (+)

x « » go TO to ta tooTumor 1: % Cells (+)

1020304030S070BOKJ100Tumor 1: % Cells (+)

MOAB 14G2a100

- «0-

= 70-1

«U

10 20 M 40 30 00 70 SO SO

Tumor 1: % Cells (+)

with other proteolytic enzymes. However, this finding is important, because collagenase is commonly used in the preparation of tumor cell vaccines for use as immunotherapy (8, 9, 19).It is possible that other, as yet undefined, tumor-associatedantigens are equally susceptible to collagenase and that moregentle treatments would result in more immunogenic vaccines.

We found minimal binding of the antibodies, ME-361 and3F8, to our melanoma cell suspensions. The latter reagent hasbeen shown to be specific for ganglioside GD2 (20), and theformer binds predominantly to GD2and to a much lesser extentto GDJ (21). These data are consistent with published workindicating that the GD2 is a major melanoma ganglioside onlyin cells adapted to tissue culture. For example, Tsuchida et al.(14) chemically assayed the ganglioside content of freshly excised surgical specimens and of tissue culture lines derived fromthose specimens. Although the cells lines were rich in GD2, thefreshly excised tissues contain minimal amounts; in contrast,GDJ levels were high in both surgical specimens and cell lines.The relatively strong binding of MOAB ME-361 to melanomatissue observed by Thurin et al. (21) may have been due to theuse of acetone fixation which allowed detection of smallamounts of non-surface GD2-

In light of these data, our observation that melanoma cellsstrongly and consistently bound the MOAB 14G2a is puzzling,since this reagent also seems to bind to GD2(22). Furthermore,we observed strong and consistent binding of MOAB ME-311which recognizes an alkali-labile ganglioside which is thoughtto represent an 9-0-acetyIated form of GD3 (23, 24). Chemicalassay of melanoma tissue by Tsuchida et al. (25) showed thatamounts of GDI and alkali-labile ganglioside were similarly low(GD2= 2.0 ¿/g/g,alkali-labile ganglioside = 3.7 /¿g/g;cf. GOÌ=47.7 Mg/g)- Thus, the high reactivity of these two MOABs isunexplained, but could be due to a concentration of 9-0-acetyl-GD3 on the cell surface or to previously undetected, low-levelcross-reactivity to the predominant ganglioside, GD3.

Although the expression of most of the melanoma-associatedantigens exhibited considerable variability when specimensfrom different patients were compared, we were surprised tofind that the melanoma phenotype of an individual patient was

quite stable over time. Regression analysis indicated that thedegree of binding of each MOAB by a "new" tumor was similar

to that of a tumor excised 2 to 25 mo previously. This observation is remarkable in view of: (a) the relatively small samplesize; (b) the technical variability inherent in surface markeranalysis; and (c) the intervening administration of a melanomacell vaccine.

The subject of phenotypic variability of melanoma-associatedantigens among tumors removed from the same patient has notbeen extensively addressed in the literature. Albino et al. (26)reported that cell lines derived from three metastatic massesfrom the same patient varied in growth rate and reactivity witha battery of MOABs. However, it is not possible to determinewhether these differences reflected distinct populations presentin vivo or were due to variations related to adaptation to tissueculture. Natali et al. (27) studied freshly harvested metastaticmasses from seven melanoma patients by immunofluorescentstaining of cryostat tissue sections. Although these investigatorsinterpreted their findings as indicating heterogeneity in theexpression of the high-molecular-weight proteoglycan antigenand of HLA-DR, their assay was semiquantitative and thus didnot allow for statistical analysis. The only published studycomparable to ours was that of Morgan et al. (28) who analyzedby flow cytometry 30 tumor specimens from 10 patients; theseauthors reported that multiple tumors from the same patientexhibited comparable, i.e., high, levels of MOAB 9.2.27 binding, similar to what we have observed.

In summary, the result of previously published studies do notcontradict our finding that the melanoma phenotype exhibitslittle variability among the metastatic lesions of a given patient,at least in regard to the antigens recognized by the limited panelof MOABs that we studied. The critical question of whetherthis finding applies to melanoma antigens relevant to immu-nologically mediated tumor regression remains to be answered.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the assistance of the followingindividuals: Cannella Clark and Carlotta Green, who processed and

6843

Research. on January 8, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEROGENEITY OF MELANOMA ANTIGENS

analyzed the tumor specimens; Marsha Golden, who performed theflow cytometry analysis; and Ellen Hart, RN, who monitored thepatients and arranged for timely specimen delivery.

REFERENCES

1. Morgan, A. C., Jr., Galloway, D. R., and Reisfeld. R. A. Production andcharacterization of monoclonal antibody to a melanoma specific glycopro-tein. Hybridoma, /: 27-36. 1981.

2. Bumol, T. F., and Reisfeld, R. A. Unique glycoprotein-proteoglyean complexdefined by monoclonal antibody on human melanoma cells. Proc. Nati. Acad.Sci. USA, 79: 1245-1249, 1982.

3. Cheresh, D. A., Varki, A. P., Varki. N. M.. Stallcup. W. B.. Levine, J., andReisfeld, R. A. A monoclonal antibody recognizes an O-acetylated sialic acidin human melanoma associated ganglioside. J. Biol. Chem.. 259:7453-7459.1984.

4. Herlyn, M., Clark, W. H., Mastrangelo, M. J., Guerry, D., Elder, D. E.,LaRossa, D., Hamilton, R., Bondi, E.. Tuthill. R.. Steplewski. Z., andKoprowski, H. Specific immunoreactivity of hybridoma-secreted monoclonalanti-melanoma antibodies to cultured cells and freshly derived human cells.Cancer Res., 40: 3602-3609, 1980.

5. Harper, J. R., Bumol, T. F., and Reisfeld, R. A. Characterization of monoclonal antibody 155.8 and partial characterization of its proteoglycan antigenon human melanoma cells. J. Immunol., 132: 2096-2104. 1984.

6. Brown, J. P., Woodbury, R. G., Hart, C. E., Hellstrom, !.. and Hcllstrom.K. E. Quantitative analysis of melanoma-associated antigen p97 in normaland neoplastic tissues. Proc. Nati. Acad. Sci. USA, 78: 539-543, 1981.

7. Atkinson. B., Ernst, S. C, Christ, B. F., Herlyn, M., Blaszczyk, M., Ross,A. H., Herlyn, D., Steplewski, Z., and Koprowski. H. Identification ofmelanoma-associated antigens using fixed tissue screening of antibodies.Cancer Res., 44: 2577-2581, 1984.

8. Berd, D., Maguire, H. C., Jr., and Mastrangelo, M. J. Induction of cell-mediated immunity to autologous melanoma cells and regression of métastases after treatment with a melanoma cell vaccine preceded by cyclophos-phamide. Cancer Res., 46: 2572-2577, 1986.

9. Berd, D., and Mastrangelo, M. J. Active immunotherapy of human melanoma exploiting the immunopotentiating effects of cyclophosphamide. Cancer Invest., 6: 335-347, 1988.

10. Peters, L. C., Brandhorst, J. S., and Hanna. M. G.. Jr. Preparation ofimmunotherapeutic autologous tumor cell vaccines from solid tumors.Cancer Res., 39: 1353-1360, 1979.

11. Schroff, R. W., Woodhouse, C. S., Foon, K. A., Oldham, R. K., Farrell, M.M.. Klein, R. A., and Morgan, A. C., Jr. Intratumor localization of monoclonal antibody in patients with melanoma treated with antibody to a250.000-dalton melanoma-associated antigen. J. Nati. Cancer Inst., 74:299-306, 1985.

12. Burchie!, S. W., Martin, J. C., Imai, K., Ferronc, S., and Warner, N. L.Heterogeneity of HLA-A.B, la-like, and melanoma-associated antigenexpression by human melanoma cell lines analyzed with monoclonal antibodies and flow cytometry. Cancer Res.. 42: 4110-4115, 1982.

13. Ross. A. H.. Grob. P., Bothwell. M.. Elder, D. E.. Ernst, C. S., Maraño,N.,Christ. B. F., Slemp, C. C., Herlyn, M., and Atkinson. B. Characterizationof nerve growth factor receptor in neural crest tumors using monoclonalantibodies. Proc. Nati. Acad. Sci. USA, 81: 6681-6685, 1984.

14. Tsuchida, T., Ravindranath. M. H., Saxton, R. E.. and Irie. R. F. Ganglio-sides of human melanoma: altered expression in viro and in vitro. CancerRes., 47: 1278-1281, 1987.

15. Cheresh, D. A., Pierschbacher. M. D.. Her/ig, M. A., and Mujoo, K.Disialogangliosides GD2 and GDJ are involved in the attachment of humanmelanoma and neuroblastoma cells to extracellular matrix proteins. J. CellBiol.. 102: 688-696. 1986.

16. Lindmo. T.. Davies. C.. Rofstad. E. K.. Fodstad. ().. and Sundan. A. Antigenexpression in human melanoma cells in relation to growth conditions andcell-cycle distribution. Int. J. Cancer, 33: 167-171, 1984.

17. Leong, S. P. L.. Bolen. J. L., Chee, D. O., Smith. V. R.. Taylor. J. C..Benfield. J. R.. and Klevecz. R. R. Cell-cycle-dependent expression of humanmelanoma membrane antigen analyzed by flow cytometry. Cancer (Phila.).55: 1276-1283. 1985.

18. Lindmo. T.. Davies. C.. Fodstad. O., and Morgan, A. C. Stable quantitativedifferences of antigen expression in human melanoma cells isolated by flowcytometric cell sorting. Int. J. Cancer. 34: 507-512. 1984.

19. Hoover. H. C.. Jr.. Surdyke. M. G., Dangel. R. B.. Peters. L. C.. and Hanna.M. G., Jr. Prospectively randomized trial of adjuvant active-specific immunotherapy for human colorectal cancer. Cancer (Phila.). 55: 1236-1243,1985.

20. Cheung, N. K. V., Saarinen. U. M.. Neely. J.. Landmeier. B., Donovan. D..and Coccia. P. F. Monoclonal antibodies to a glycolipid antigen on humanneuroblastoma cells. Cancer Res.. 45: 2642-2649. 1985.

21. Thurin. J.. Thurin. M.. Kimoto. Y.. Herlyn. M.. Lübeck.M. D.. Elder. D.E., Smereczynska. M., Karlsson. D.-A.. Clark. W. M., Jr., Sleplewski. Z..and Koprowski. H. Monoclonal antibody-defined correlations in melanomabetween levels of GD2and GD, antigens and antibody-mediated cytotoxicity.Cancer Res.. 47: 1229-1233. 1987.

22. Mujoo, K.. Kipps, T. J., Yang, H. M., Cheresh, D. A., Wargalla, U., Sander,D. J.. and Reisfeld, R. A. Functional properties and effect on growthsuppression of human neuroblastoma tumors by isotype switch variants ofmonoclonal antigangliosidc GnJ antibody 14.18. Cancer Res.. 49: 2857-2861. 1989.

23. Cheresh, D. A., Reisfield. R. A., and Varki. A. P. O-Acetylation of disialo-ganglioside (.,,, by human melanoma cells creates a unique antigenic determinant. Science (Wash. DC). 225: 842-846. 1984.

24. Thurin. J., Herlyn, M.. Hindsgaul. O.. Strombcrg. N.. Karlsson, K-A.. Elder.D., Steplewski, Z., and Koprowski. H. Proton NMR and fast-atom bombardment mass spectrometry analysis of the melanoma-associated ganglioside 9-O-acetyl-GDj. J. Biol. Chem.. 260: 14556-14563, 1985.

25. Tsuchida, T., Saxton. R. E.. Morton. D. L.. and Irie, R. F. Gangliosides ofhuman melanoma. J. Nati. Cancer Inst., 78: 45-54, 1987.

26. Albino, A. P.. Lloyd, K. O.. Houghton, A. N.. Oettgen. H. F.. and Old. L. J.Heterogeneity in surface antigen and glycoprotein expression of cell linesderived from different melanoma métastasesof the same patient. J. Exp.Med., 154: 1764-1778, 1981.

27. Natali. P.. Bigotti. A.. Cavaliere. R.. Liao. S. K., Taniguchi. M.. and Ferrone.S. Heterogeneous expression of melanoma-associated antigens and HLAantigens by primary and multiple metastatic lesions removed from patientswith melanoma. Cancer Res.. 45: 2883-2889. 1985.

28. Morgan. A. C.. Jr.. Woodhouse. C.. Bartholemew. R.. and Schroff, R. Humanmelanoma-associated antigens: analysis of antigenic heterogeneity by molecular, serologie, and flow-cytometric approaches. Mol. Immunol.. 23: 193-200. 1986.

29. Herlyn. M.. Steplewski, Z., Herlyn, D., Clark, W. H., and Ross, A. H.Production and characterization of monoclonal antibodies against humanmalignant melanoma. Cancer Invest.. /: 215-224. 1983.

30. Thurin. J.. Thurin. M., Herlyn, M., Elder. D. E.. Steplewski, Z., Clark, W.H.. Jr., and Koprowski, H. GD: ganglioside biosynthesis is a distinct biochemical event in human melanoma tumor progression. FEBS Lett., 208:17-22. 1986.

6844

Research. on January 8, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

1989;49:6840-6844. Cancer Res   David Berd, Meenhard Herlyn, Hilary Koprowski, et al.   AntigensHeterogeneity of Expression of Human Melanoma-associated Flow Cytometric Determination of the Frequency and

  Updated version

  http://cancerres.aacrjournals.org/content/49/23/6840

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/49/23/6840To request permission to re-use all or part of this article, use this link

Research. on January 8, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from