1986;46:5887-5892. Cancer Res   Alonzo H. Ross, David Pleasure, Kenneth Sonnenfeld, et al.   Neurofibroma Schwann CellsExpression of Melanoma-associated Antigens by Normal and

  Updated version

  http://cancerres.aacrjournals.org/content/46/11/5887

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

  .permissions@aacr.orgDepartment at

To request permission to re-use all or part of this article, contact the AACR Publications

on June 4, 2014. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from on June 4, 2014. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

(CANCER RESEARCH 46, 5887-5892, November 1986]

Expression of Melanoma-associated Antigens by Normal and NeurofibromaSchwann Cells1

Alonzo H. Ross,2 David Pleasure, Kenneth Sonnenfeld, Barbara Atkinson, Barbara Kreider,3 Donna M. Jackson,

Ingrid Taff, Elio Scarpini, Robert P. Lisak, and Hilary KoprowskiThe Wistar Institute of Anatomy and Biology [A. H. R., D. M. J., H. K.]; The Children's Hospital of Philadelphia, (D. P., B. K.J; Hospital of the University of

Pennsylvania [B. A., R. P. L.J Philadelphia, Pennsylvania 19104; Mt. Sinai School of Medicine, New York, New York 10029 ¡K.S., I. TJ; and University of MilanSchool of Medicine, Milan, Italy [E. S.J

ABSTRACT

The cell surface antigen distribution on traumatic neuroma Schwanncells and neurofibroma Schwann-like cells was characterized using monoclonal antibodies that define melanoma-associated antigens. Immunoflu-orescence staining of cultured cells, immunoprecipitation of radioiodi-nated antigens from cells placed in short-term cultures, and immunoper-oxidase staining of frozen tissue sections revealed most of the melanoma-associated antigens tested on traumatic neuroma and neurofibromaSchwann cells and on fetal and adult femoral nerve. The cross-reactivityof the antibodies with neural cells may reflect the common neural crestembryological origin of Schwann cells and ntelanocytes. Cell sorteranalysis of neurofibroma cells using a monoclonal antibody directedagainst the melanoma nerve growth factor receptor resulted in cellcultures highly enriched for Schwann-like cells which may bear thegenetic defect responsible for neurofibromatosis. The antigen detected bythis monoclonal antibody is the neurofibroma nerve growth factor receptor and the antibody was a potent inhibitor of nerve growth factor bindingto neurofibroma cells.

INTRODUCTION

Schwann cells are the major glial elements in the peripheralnervous system and are embryologically derived from the neuralcrest (1). In adults, these cells surround all peripheral nervefibers and, except in the case of injury or disease, have a verylow rate of division (2). Neurofibroma is a relatively commonautosomal dominant disorder most often involving benign dermal lesions containing Schwann-like cells, fibroblasts, mastcells, and neuronal processes (3). The primary proliferating cellis thought to be the Schwann-like cell (4).

Little is known of the molecular or cellular properties of theSchwann-like cells, as compared with those of melanoma cells(5). A large number of anti-melanoma MAbs4 are available thathave been used to characterize the distribution of melanoma-associated antigens among normal and transformed tissues (6,7). This information is important for applications of MAbs totherapy and diagnosis as well as to the determination of themelanoma-associated antigen function and role in the transformed phenotype. Many MAbs do not bind to normal tissuemelanocytes but do bind to nevi, the benign melanocytic lesions,and to neural-related tumors such as astrocytomas and neuro-blastomas. This binding pattern may be related to the neuralcrest embryological origin of melanoma (1). We previously

Received 4/29/86; revised 7/31/86; accepted 8/5/86.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.

1This investigation was supported by USPHS Grants CA-25874, NS-21716,CA-21124.CA-10815, RR-05540. NS-08075, NS-11037, and HD-08536, and bygenerous support from the National Neurofibromatosis Foundation, the MuscularDystrophy Association, and the Dysautonomia Foundation.

2To whom requests for reprints should be addressed.3 Research Postdoctoral Fellow of the National Neurofibromatosis Founda

tion.4 The abbreviations used are: MAb. monoclonal antibody; NGF, nerve growth

factor; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis;MEM, minimal essential medium: PBS, phosphate-buffered saline; gp260. gpl 20,a glycoprotein with a molecular weight of 260,000 or 120,000, respectively.

reported that the staining pattern of human peripheral nervesand neurofibromas with anti-NGF MAbs suggested that normalSchwann cells and Schwann-like cells in neurofibromas expressed the NGF receptor (8) and that Schwann-like cellsspecifically bind NGF (9).

To better understand the growth and immunological properties of normal and neurofibroma Schwannian cells and theircorrelation, if any, with tumor progression, we have analyzedthe antigenic profiles of these cells using anti-melanoma MAbs.We report here that traumatic neuroma Schwann cells andneurofibroma Schwann-like cells express many of the antigensdetected with anti-melanoma MAbs. These anti-melanomaMAbs also exhibited reactivity in femoral nerve frozen tissuesections. The cross-reacting antigens detected on the Schwanncells are similar to the corresponding melanoma-associatedantigens as judged by SDS-PAGE. These results define newcross-reactivities of the anti-melanoma MAbs with peripheralnerves which are relevant to future therapeutic and diagnosticapplications of the MAbs, identify tumor-associated antigenspresent on neurofibroma cells, provide a means by cell sortingof preparing pure cultures of neurofibroma Schwann-like cells,and demonstrate that two tumors of similar embryologicalorigin, neurofibroma and melanoma, express a similar antigenicprofile.

MATERIALS AND METHODS

MAbs. The production and characterization of the anti-melanomaMAbs have been described (6, 8, 10). MAbs P3X63Ag8, ME31.3,ME19-19, ME20.4, ME75-29, ME061, ME13-17, and ME491 areIgGl isotype and MAbs MENu4B and ME37-7 are IgG2a isotype.Table 1 summarizes the antigens detected and cross-reactivities of theMAbs. The proteoglycan antigen identified in this study with MAbME31.3 has also been extensively studied by other groups using MAbs9.2.27 (15), 155.8 (16), 225.28S (17), and B5 (18). The M, 97,000glycoprotein identified in this study with MAb ME061 has been extensively studied using MAbs 96.5 (19, 20) and L235 (18) and has beenshown to be structurally related to transfemn (21).

Tissue Samples. Neurofibromas removed for therapeutic purposesand femoral nerves removed postmortem from adults with no apparentneuropathies were quick frozen. Fragments of traumatic neuromas wereobtained at the time of delayed secondary nerve repair. Fetal femoralnerves were obtained from therapeutic abortions at 19 or 22 wk gestation. The fetal neural samples were rapidly frozen by immersion inisopentane cooled with liquid nitrogen.

Cell Cultures. The derivation of melanoma cell lines WM-9 andA875 has been described (22, 23). Single cell suspension were preparedfrom freshly excised neurofibromas and traumatic neuromas. Tissuesamples were minced in Hanks' balanced salt solution with 1 mM

EDTA and lacking magnesium and calcium, then transferred to RPMI1640 medium containing 15% fetal calf serum-25 mM 4-(2-hydroxy-ethyl)-l-piperazineethanesulfonic acid, dispase, 1.25 units/ml (crude;Boehringer)-0.05% (w/v) collagenase (Worthington)-0.1 % (w/v) hya-luronidase (Sigma), and incubated overnight at 37°Cin 5% CO2/95%

air. The next day the medium was removed and replaced with RPMI1640 without serum containing dispase, 1.25 units/ml and 0.05%

5887

on June 4, 2014. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SCHWANN CELL SURFACE MARKERS

Table 1 Properties of anti-melanoma MAbs and corresponding antigens

Cross-reactivitywithMAbME31.3andME19-19MENu4BME20.4ME75-29ME061AntigenM,

260,000 glycoprotein + M,500,000proteoglycanM,

105,000, 130,000glycoproteinNGF

receptor, M, 75,000 glycoproteinM,

120,000glycoproteinM,

97,000 glycoproteinNormal

adulttissuesNevus,

sweat gland, basal keratinocytes,connective tissue in colon, capillary en-

dotheliumNevusNevus,

adrenal medulla, peripheralnervesNevusNevusTumorsAstrocytomaAstrocytomaPheochromocytoma, neuroblastoma,melanomaAstrocytoma,

carcinoma,sarcomaAstrocytoma,

carcinoma, lymphoma,Ref.5-7,5-685-«5-711-12

ME491

ME37-7 andME13-17

M, 30,000 to 60,000 glycoprotein

HLA-DR

Nevus, some epithelial and hematopoieticcells

Nevus, B-lymphocytes, macrophages

Astrocytoma, some carcinomas 13

Lymphoma, some carcinomas 14

collagenase. After 30 min at 37°C,the suspension was pipetted vigor

ously, then spun down and the pellet washed three times with RPMI1640 medium containing 15% (v/v) fetal calf serum. The cells wereseeded on poly-L-lysine coated glass coverslips.

Radioiodination, Immunoprecipitation, and Electrophoresis. Cellswere labeled by the lactoperoxidase procedure and extracted withsolubilizing buffer [0.5% Nonidet P-40:140 mM NaCl:10 mM NaF:10mM Tris:5 mM EDTA:aprotinin (100 kallikrein IU/ml):l mM phenyl-methylsulfonyl fluoride (pH 7.5)] (24). Immunoprecipitations werecarried out using 100 n\ of hybridoma culture supernatant per sample.Antigen:mouse immunoglobulin complexes were bound with 50 n\ goatanti-mouse immunoglobulin agarose (Sigma) and washed three timeswith solubilizing buffer. The samples were analyzed by SDS-PAGEusing a 10% gel (25). Autoradiography was carried out using a CronexLightning-Plus intensifying screen (DuPont).

Immunofluorescence and Immunoperoxidase Microscopy. Cells to beanalyzed by immunofluorescence microscopy were incubated withoutprior fixation with undiluted hybridoma culture supernatant, washedwith MEM, overlaid with rhodamine-conjugated goat anti-rabbit immunoglobulin (Cappel), washed with MEM, fixed with ice-cold 5%acetic acid (v/v) in ethanol, washed five times with MEM, and washedonce with water. Preparation of frozen sections and immunoperoxidasestaining have been described (14, 26).

Fluorescence-activated Cell Sorting. Cells to be analyzed were released from the substratum with dispase, 1.25 units/ml and 0.05% (w/v) collagenase. The cells were washed with Dulbecco's PBS (lackingCa2+ and Mg2*) and resuspended at 5 x IO6 cells/ml in hybridomaculture supernatant diluted 1:2 with PBS for I h at 4°C.The cells werepelleted, washed three times with PBS, resuspended at 5 x IO6 cells/ml in fluorescein-conjugated rabbit anti-mouse immunoglobulin (60fig/ml) and incubated for l h at 4°C.The cells were washed in the samemanner and resuspended in PBS at 1 x IO7cells/ml. An aliquot of the

cells was analyzed by flow cytometry (Ortho Cytofluorograf 50 HH).The cutoff for positive staining was determined using cells stained inparallel with P3X63Ag8, a nonspecific control MAb. The remainderof the cells were sorted on a Becton Dickinson FACS IV retaining the50% most fluorescent of the positive cells for cell culture.

RESULTS

Detection of Melanoma-associated Antigens on TraumaticNeuroma Schwann and Neurofibroma Cells. Cells from a traumatic neuroma and from a dermal neurofibroma were placedinto short-term culture for 1 day and then radioiodinated bythe lactoperoxidase method. For comparison, the melanomacell line WM-9 was also iodinated. The cells were solubilized,and specific antigens were immunoprecipitated with anti-melanoma MAbs and analyzed by SDS-PAGE. Fig. 1 shows the

Cells

MAb

Mr (x IO'3)

200 —

Neuroma Neurofibroma Melanoma

Oì Oí OÕ(M^j-ç^r^cN^cor^cN^tinö»-'r^Lnö^-"r^ir>öf^rgrocor^cNcncor^cNJu i i)i jji in in n i ni ni in ut

I,,' » — gp260

— gp120

— — NGF-R

43 —

25.7 —

HLA-DR

1 2 3 4 5 6 7 8 9 10 11 12Fig. l. Immunoprecipitation of radioiodinated antigens with anti-melanoma

MAbs. Short-term cultured neuroma (lanes I to 4) or neurofibroma (lanes 5 toiV)cells or melanoma line WM-9 cells (lanes 9 to 12) were surface labeled with'•"'Iby the lactoperoxidase method and extracted with solubilizing buffer. The

specific antigens were immunoprecipitated with MAb ME31.3 (anti-gp260, proteoglycan; lanes I, 5, and 9), ME37-7 (anti-HLA-DR; lanes 2, 6, and JO), ME75-29 (anti-gpl20; lanes 3, 7, and //), or ME20.4 (anti-NGF receptor (A); lanes 4,8, and 12) and analyzed by SDS-PAGE and autoradiography.

results for MAbs ME31.3, ME37-7, ME75-29, and ME20.4.ME31.3 is specific for the melanoma-associated chondroitinsulfate proteoglycan and its core protein gp260 (27). Bothspecies were detected in the traumatic neuroma, neurofibroma,and melanoma samples (lanes 1, 5, and 9). The gp260 component is marked on the gel, and the proteoglycan component isvisible as a broad band in the stacking gel (delineated by arrows).Additional bands in lanes I to 4 are due to higher backgroundimmunoprecipitation for the traumatic neuroma. The ME37-7antigen, HLA-DR, was detected only in the melanoma sample(lane 10). The ME75-29 antigen, gpl20, was detected in theneuroma, neurofibroma, and melanoma samples (lanes 3, 7,and 11). The ME20.4 antigen, the NGF receptor, was alsodetected in all three samples (lanes 4, 8, and 12). Table 2

5888

on June 4, 2014. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SCHWANN CELL SURFACE MARKERS

Table 2 Immunoprec¡pilotionsfrom Schwannian cells using anti-melanomaMAbs

No. cultures positive/no, culturestestedMAbME31.3

MENu4BME20.4

ME75-29ME061ME491ME37-7Traumatic

neuromaSchwanncells2/2

1/1I/I2/20/10/10/2Neurofibroma

cells2/2

1/12/22/21/1 (weak)0/10/2

Mr (x IO'3)

200

92.5

68

43

25.7

18

3 4Fig. 2. Detection of ME491 antigen by Western blotting. Membrane extracts

from either a neurofibroma (lanes I and }) or melanoma cell line A87S (lanes 2and 4) were applied to an SDS-polyacrylamide gel and then electroblotted ontonitrocellulose, and incubated with either control MAb P3X63Ag8 (lanes I and 2)or MAb ME491 (lanes 3 and 4). The antigens were then visualized with an '"I-

labeled second antibody and autoradiography.

summarizes the immunoprecipitation experiments with theseand other anti-melanoma MAbs.

The ME491 antigen present in cultured neurofibroma cellswas identified by Western blotting (Fig. 2, lane 3). The neuro-fibroma antigen which binds MAb ME491 has a molecularweight similar to that of the melanoma antigen but appears tobe less heterogeneous. Longer autoradiographic exposures ofthe same blot revealed that the difference was not due to thegreater intensity of the melanoma ME491 band (lane 4). Theblot shown is representative of three experiments made withthe detergent extract of a single neurofibroma culture. TheME491 antigen was not detected by immunoprecipitation (Table 2) probably because most of the antigen is intracellular and,hence, not accessible to cell surface lactoperoxidase iodination(13).

The protein bound by MAb ME20.4 was demonstrated to be

the neurofibroma NGF receptor in binding studies with 125I-labeled NGF (Fig. 3). ME20.4 ascites fluid at a 10"6 dilutionstrongly inhibited I25l-labeled NGF binding. Two other MAbs,ME 19-19 and ME 13-17, which bind to other proteins on theneurofibroma cell surface, did not inhibit ' 25I-labeled NGF

binding at these dilutions.The fraction of neurofibroma cells expressing melanoma-

associated antigens was determined by flow cytometry. In aninitial study, we analyzed 11 cultures of dermal and plexiformneurofibroma cultures for NGF receptor expression (MAbME20.4). Between 16 and 66% of the cells were positive forNGF receptor [39.1 ±16.4% (SD)]. We also analyzed fourdermal neurofibroma cultures with MAbs ME20.4, ME491,ME31.3, and ME13-17 (Table 3). ME20.4 and ME491 consistently stained a large percentage of neurofibroma cells,whereas ME31.3 stained a smaller fraction. ME 13-17 reactivitywas variable, with staining of some cultures but not others. Incontrols with WM-9 melanoma cells, the ME20.4, ME491,and ME13-17 antigens were uneffected by the protease treatment used to prepare the cells for flow cytometry, but theME31.3 antigen was greatly reduced.

The identity of the melanoma-associated antigen-expressingcells in the neurofibroma cultures was determined by inuminofluorescence microscopy. Fig. 4, A to C shows that the cellsstained with MAbs ME37-7 and ME20.4 are the long spindle-shaped cells with oval nuclei (Schwann-like cells). The fibro-blastic cells did not fluoresce detectably. The Schwann-like cellsbut not the fibroblasts were positive for MAbs ME31.3 (5 of 6cultures), ME37-7 (6 of 6 cultures), ME20.4 (5 of 6 cultures),ME75-29 (3 of 6 cultures), MENu4B (3 of 6 cultures), and

Fig. 3. Inhibition of MAbs of '"I-labeled NGF binding to dissociated neurofibroma cells. Dissociated neurofibroma cells were incubated for 90 min at 37'Cin binding buffer |RPMI 1640 containing 25 itiM 4-(2-hydroxyethyl)-l-pipera-zineethanesulfonic acid and BSA, 5 mg/mlj containing '"1-labeled NGF (5 ng/ml) and several dilutions of MAb-containing ascites [ME20.4 (•).ME19-19 (O).and ME37-7 (A)]. Triplicate samples were centrifuged for 30 s to separate boundfrom free '"I-labeled NGF, and bound '"I-labeled NGF was measured in a

gamma counter. Nonspecific binding (approximately 30% of total binding) wasdetermined in parallel samples to which unlabeled NGF (10 fig/ml) was addedand has already been subtracted from the data shown. Binding in the absence ofM Ab was I600cpm/1.4 x IO4cells.

Table 3 Flow cytometry analysis of neurofibroma cellsCells from each of four neurofibroma cultures were incubated with anti-

melanoma MAb followed by a fluorescent-stained antibody and analyzed by flowcytometry for the percentage of cells expressing the melanoma-associated antigens. The percentage of positive cells with control MAb P3X63Ag8 (2%) wassubtracted from all values.

MAb

% of cells positive formelanoma-associated antigen

Mean ±SD(range)ME31.3ME20.4

ME491ME13-172

±2 (0-5)50 ±22(19-72)56 ±9 (47-69)13 ±13(2-31)

5889

on June 4, 2014. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SCHWANN CELL SURFACE MARKERS

(.*._

Fig. 4. Phase contrast and indirect immunofluorescence microscopy. Schwann-like cells as visualized in cells cultured from a traumatic neuroma (A and /'). adermal neurofibroma (C and D), or Schwann-like cells enriched from a dermal neurofibroma by cell sorting using anti-NGF receptor MAb (E and F) were analyzedfor melanoma-associated antigen expression. Cells were stained with MAb ME37-7 (A to D) or MAb ME20.4 (E and F). Bars, 20 ion (A to D) or 70 urn (E and F).

5890

on June 4, 2014. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SCHWANN CELL SURFACE MARKERS

ME061 (1 of 6 cultures). The Schwann-like cells did not stainwith nonspecific control MAbs (not shown) and no stainingwas observed for MAb ME491 (0 of 6 cultures). The positiveresults with ME491 using flow cytometry (Table 3) may reflectgreater accessibility to the antigen following protease treatmentto release the cells from the substratum.

Cell Sorting of Neurofibroma Cell Cultures. To prepare purecultures of the NGF receptor-positive cells (ME20.4), a smallfraction of the neurofibroma culture was analyzed by flowcytometry. The rest of the cells were then sorted, collecting the50% most fluorescent of the positive cells. An essential prerequisite for this procedure was the preparation of single cellsuspensions from the neurofibroma cultures using the dispaseprocedure since the presence of aggregates resulted in pooryields and impure cultures. The viability of the sorted cells was99% as judged by trypan blue exclusion immediately aftersorting and about 50% of these cells, when placed into culture,adhered and grew, resulting in cultures consisting of about 98%Schwannian cells (Fig. 4, E and F).

Immunoperoxidase Staining of Normal Peripheral Nerve andNeurofibromas. The expression of the melanoma-associatedantigens in vivo was assayed by immunoperoxidase staining offrozen tissue sections (Table 4). MAbs MENu4B, ME75-29,and ME061 bound to the femoral nerve fiber but not to neu-rofibromas. MAbs ME31.3, ME20.4, ME491, and ME37-7bound to both nerve fibers and neurofibroma Schwann-likecells. ME31.3 and ME37-7 appeared to bind to a subset of theneurofibroma Schwann-like cells, in agreement with flow cytometry measurements (Table 3). All of the MAbs except forME31.3 and ME37-7 also bound to fetal femoral nerve.

DISCUSSION

Using a combination of methods, we have detected melanoma-associated antigens on Schwann cells from traumaticneuromas and neurofibroma Schwann-like cells. Because thesecultures contain a heterogeneous mixture of cells, we verifiedthat the antigens are indeed associated with the morphologicallyidentified Schwannian cells by immunofluorescence microscopyof the cultured cells or by immunoperoxidase staining of frozentissue sections. With some MAbs, the antigens were detectedby one method but not by another, perhaps reflecting differingsensitivities of the two methods, differing abilities of MAbs tofunction in different assays, or differing sensitivities of themelanoma-associated antigens to proteases used to release thecells prior to some assays. In some cases, these cultures weresorted with anti-NGF receptor MAb to obtain nearly purecultures of Schwann-like cells. The continued detection of theNGF receptor after sorting demonstrates that the Schwann-likecells actually express the NGF receptor and do not simplyabsorb receptor synthesized and released by some other celltype.

Table 4 Detection of melanoma-associated antigens by immunoperoxidasestaining of neurofibroma frozen sections

MAbME31.3

MENu4BME20.4

ME75-29ME061ME49IME37-7Immunoperoxidase

stainingofAdult

femoralnerve±

±±±Fetal

femoral Neurofibromanerve Schwann-likecells+

(few cells)

+ (focal)" ++, strong brown color; -I-,brown color. ±,slight brown color; —,no staining.

The antigens detected by MAbs ME31.3, MENu4B, ME20.4,ME75-29, and ME061 by immunoprecipitation were appar

ently identical from melanoma and neurofibroma as judged bySDS-PAGE. The neurofibroma ME491 antigen detected by

Western blotting was less heterogeneous than the corresponding melanoma antigen. Since this antigen is heavily glycosylated(13, 28), the difference may be the result of more homogeneousglycosylation in the neurofibroma. The ME20.4 antigen isindeed the NGF receptor in neurofibroma cells since MAbME20.4 completely blocked '"I-labeled NGF binding (Fig. 3).

We have also detected many of the melanoma-associatedantigens on frozen sections of adult and fetal peripheral nerves.Although the exact assignment of the melanoma-associatedantigen-bearing cells awaits more detailed histológica! analysis,possibly including electron microscopy, it is clear that normalelements of the peripheral nervous system express many of themelanoma-associated antigens. Preliminary studies using anti-

NGF receptor MAb indicate that cells of the perineurium whichare probably of neural crest embryological origin, consistentlyshow strong staining. Schwann cells appeared to stain in somecases but not in others,5 perhaps reflecting heterogeneity of

Schwann cells in different nerves.The cross-reactivity of these MAbs with normal tissue does

not rule out their use for melanoma immunotherapy. The anti-carcinoma MAb CO 17-1A, for example, has been used inclinical trials for patients with colon and pancreatic carcinomasand has resulted in remissions for some patients (29). Radioim-aging studies have demonstrated that CO 17-1A MAb localizesat the tumor in patients even though some epithelial cells innormal colon express the antigen (30, 31). This paradoxicalresult is apparently due to greater accessibility of the MAb toantigen in tumor than in normal colon.

The melanoma-associated antigen for which we have the mostinformation is the NGF receptor which is not detected in frozensections of normal melanocytes but is easily detected in sectionsof nevi and melanomas (8). Human neonatal melanocytes inshort-term culture under conditions inducing cell division ex

press low levels of the receptor (5, 6). Both Schwann cells andnerve soma in neonatal monkey dorsal root ganglia also showstrong staining.6 In the present study, fetal and adult peripheral

nerves were strongly stained with anti-NGF receptor MAb,perhaps reflecting the related neural crest embryological originof melanocytes, Schwann cells, and sympathetic and sensoryneurons. Schwann cells and melanocytes may derive from acommon bipotent progenitor cell (32). It has been shown thatthe expression of other tissue-specific genes is the result of acomplex interaction between multiple fra/w-acting factors (33).We suggest that the expression of the NGF receptor requiresat least two factors, the first of which is normally expressed inneural crest-derived cells and is necessary but not sufficient forexpression of NGF receptor. The proposed second factor required for NGF receptor expression is induced in stimulatedmelanocytic cells capable of undergoing cell division in cultureor in nevic or melanoma lesions but not in normal tissuemelanocytes, which rarely divide. In sensory nerves and someSchwann cells, this second factor or a substitute factor isexpressed. This model suggests many experiments using therecently cloned NGF receptor gene (34) to identify factors

" M. A. Bothwell and E. Scarpini, unpublished observations.* N. Maraño,B. Dietzschold, J. J. Earley, G. Schatteman. S. Thompson, P.

Grob, A. H. Ross, M. Bothwell. B. F. Atkinson, and H. Koprowski, Purificationand amino terminal sequencing of human melanoma nerve growth factor receptor.J. Neurochem., in press.

5891

on June 4, 2014. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

SCHWANN CELL SURFACE MARKERS

responsible for the tissue-specific expression of the NGF receptor and its normal expression in melanomas.

ACKNOWLEDGMENTS

We are very grateful to Drs. Allan Rubinstein (Department of Neurology, Mt. Sinai Medical School) and A. Lee Osterman (Departmentof Orthopedic Surgery, University of Pennsylvania) for their help inobtaining the tissue samples for this study. Marina Aquino providedexcellent technical assistance. We thank Jeffrey Faust for running thecell sorter, Marina Hoffman for editorial assistance, and MeenhardHerlyn for helpful discussions and preparation of hybridoma culturesupernatants.

REFERENCES

1. Le I)(MUMM.N. The Neural Crest. Cambridge, England: Cambridge University Press, 1982.

2. Brown, M. J., and Asbury, A. k. Schwann cells proliferation in the postnatalmouse: timing and topology. Exp. Neurol., 74: 170-186, 1981.

3. Riccardi, V. M. Von Recklinghausen neurofibromatosis. N. Engl. J. Med.305: 1617-1627, 1981.

4. Sobue, G., Sonnenfeld, K., Rubenstein, A. E., and Pleasure, D. Tissue culturestudies of neurofibromatosis: effects of axolemmal fragments and cyclic 3',5'-AMP. Analysis on proliferation of Schwann-like and fibroblast-like NF cells.Ann. Neurol., IS: 68-73, 1985.

5. Herlyn, M., Thurin, J., Balaban, G., Bennicelli, J. L., Herlyn, D., Elder, D.E., Bondi, E., Guerry, D., Nowell, P., Clark, W. H., and Koprowski, H.Characteristics of cultured human melanocytes isolated from different stagesof tumor progression. Cancer Res., 45: 5670-5676, 1985.

6. Herlyn, M., Steplewski, Z., Herlyn, D., Clark, W. H., Ross, A. H., Blaszczyk,M., Pak, K. Y., and Koprowski, H. Production and characterization ofmonoclonal antibodies against human malignant melanoma. Cancer Invest.,/.-215-224, 1983.

7. Real, F. X., Houghton, A. N., Albino, A. P., Cordon-Cordo, C, Melamed,M. R.. Oettgen. H. F., and Old, L. J. Surface antigens of melanomas andmelanocytes defined by mouse monoclonal antibodies: specificity analysisand comparison of antigen expression in cultured cells and tissues. CancerRes., 45:4401-4411, 1985.

8. Ross, A. H., Grob, P., Bothwell, M., Elder, D. E., Ernst, C. S., Maraño,N.,Christ, B. F. D., Slemp, C. C., Herlyn, M., Atkinson, B. F., and Koprowski,H. Characterization of nerve growth factor receptor in neural crest tumorsusing monoclonal antibodies. Proc. Nati. Acad. Sci. USA, 81: 6681-6685,1984.

9. Sonnenfeld, K. H., Bernd, P., Sobue, G., Lebwohl, M., and Rubenstein, A.E. Nerve growth factor receptors on dissociated neurofibroma Schwann-likecells. Cancer Res., 46: 1446-1452, 1986.

10. Atkinson, B. F., Ernst, C. S.. Christ, B. F. D., 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.

11. Hellstrom. L, Garrigues, J.. Cabasco, L., Mosely, G. H., Brown, J. P., andHellstrom, K. E. Studies of a high molecular weight human melanoma-associated antigen. J. Imnmiml., 130: 1467-1472, 1983.

12. Ross, A. H., Herlyn, M., Ernst, C. S., Guerry, D., Bennicelli, J., Christ, B.F. D., Atkinson, B. F., and Koprowski, H. Immunoassay for melanoma-associated proteoglycan in the sera of patients using monoclonal and poly-clonal antibodies. Cancer Res., 44:4642-4647, 1984.

13. Atkinson, B. F., Ernst, C. S., Christ, B. F. D., Ross, A. H., Clark, W. H.,Herlyn, M., Herlyn, D., Maul, G., Steplewski, Z., and Koprowski, H.Monoclonal antibody to a highly glycosylated protein reacts in fixed tissuewith melanoma and other tumors. Hybridoma, 4: 243-255, 1985.

14. Thompson, J. J., Herlyn, M.. Elder, D. E., Clark, W. H., Steplewski, Z., and

Koprowski, H. Expression of DR antigen in freshly frozen human tumors.Hybridoma, /: 161-168, 1982.

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

16. 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. luminimi., 132: 2096-2104, 1984.

17. Wilson, B. S., Imai, K., Natali, P. G., and Ferrane, S. Distribution andmolecular characterization of a cell-surface and a cytoplasmic antigen detectable in human melanoma cells with monoclonal antibodies. Int. J. Cancer,28: 293-300, 1981.

18. Houghton, A. N., Eisinger, M., Albino, A. P., Cairncross, J. G., and Old, L.J. Surface antigens of melanocytes and melanomas. Markers of melanocytedifferentiation and melanoma subsets. J. Exp. Med., 156: 1755-1766, 1982.

19. Brown, J. P., Wright, P. W., Hart, C. E., Woodbury, R. G., Hellstrom, K.E., and Hellstrom, I. Protein antigens of normal and malignant human cellsidentified by immunoprecipitation with monoclonal antibodies. J. Biol.Chem., 255:4980-4982, 1980.

20. Brown, J. P., Nishiyama, K., Hellstrom, I., and Hellstrom, K. E. Structuralcharacterization of human melanoma-associated antigen p97 with monoclonal antibodies. J. Immunol., 127: 539-546, 1981.

21. Brown, J. P., Hewick, R. M., Hellstrom, L, Hellstrom, K. E., Iinoliale. R.F., and Dreyer, W. J. Human melanoma-associated antigen p97 is structurally and functionally related to transferrin. Nature (Lond.), 296: 171-173,1982.

22. Herlyn, M., Clark, W. H., Mastrangelo, M., Guerry, D., Elder, D. E.,Larossa, D., Hamilton, R., Bondi, E., rutilili. R., Steplewski, Z., and Koprowski, H. Specific immunoreactivity of hybridoma-secreted monoclonalantimelanoma antibodies to cultured cells and freshly derived human cells.Cancer Res., 40: 3602-3609, 1980.

23. Giard, D. J., Aaronson, S. A., Todaro, G. J., Arnstein, P., Kersey, J. H.,Dosik, H., and Parks, W. P. In vitro cultivation of human tumors: establishment of cell lines derived from a series of solid tumors. J. Nati. Cancer Insi.,51:1417-1423, 1973.

24. Ross, A. H., Mitchell, K. F., Steplewski, Z., and Koprowski, H. Identificationand isolation of melanoma-associated antigens with monoclonal antibodies.Hybridoma, 1: 413-421, 1982.

25. Laemmli, U. K. Cleavage of structural proteins during the assembly of thehead of bacteriophage T4. Nature (Lond.), 227: 680-685, 1970.

26. Atkinson, B. F., Ernst, C. S., Herlyn, M., Steplewski, Z., Sears, H. F., andKoprowski, H. Gastrointestinal cancer-associated antigen in immunoperox-idase assay. Cancer Res., 42: 4820-4823, 1982.

27. Ross, A. H., Cossu, G., Herlyn, M., Bell, J. R., Steplewski, Z., and Koprowski, H. Isolation and chemical characterization of a melanoma-associated proteoglycan antigen. Arch. Biochem. Biophys., 225: 370-383, 1983.

28. Ross, A. H., Dietzschold, B., Jackson, D. M., Earley, J. J., Christ, B. F. D.,Atkinson, B. F., and Koprowski, H. Isolation and amino terminal sequencingof a novel melanoma-associated antigen. Arch. Biochem. Biophys., 242:540-548, 1985.

29. Koprowski, H., Herlyn, D., Lübeck,M., DeFreitas, E., and Sears, H. F.Human anti-idiotype antibodies in cancer patients: is the modulation of theimmune response beneficial for the patient? Proc. Nati. Acad. Sci. USA, 81:216-219, 1984.

30. Shen, S., Atkinson, B. F., Koprowski, H., and Sears, H. F. Binding of murineimmunoglobulin to human tissues after immunotherapy with anticolorectalcarcinoma monoclonal antibody. Int. J. Cancer, 33:465-468, 1984.

31. Mach, J-P., Chatal, J-F., Lumbroso, J-D., Buchegger, F., Forni, M., Rit-schard, J., Berche, C., Douillard, J-Y., Carrel, S., Herlyn, M., Steplewski,Z., and Koprowski, H. Tumor localization in patients by radiolabeled monoclonal antibodies against colon carcinoma. Cancer Res., 43: 5593-5600,1983.

32. Nichols, D. H., and Weston, J. A. Melanogenesis in cultures of peripheralnervous tissue. Dev. Biol., 60: 217-225, 1977.

33. Blau, H. M., Pavlath, G. K., Hardeman, E. C., Chiù,C.-P., Silberstein, L.,Webster, S. G., Miller, S. C., and Webster, C. Plasticity of the differentiatedstate. Science (Wash. DC), 230: 758-766, 1985.

34. Chao, M. V., Bothwell, M. A., Ross, A. H., Koprowski, H., Lanahan, A. A.,Buck, C. R., and Sehgal, A. Gene transfer and molecular cloning of thehuman NGF receptor. Science (Wash. DC), 232: 518-521, 1986.

5892

on June 4, 2014. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from