Differential Expression of ras Oncogene Products Among the Types of Human Melanomas and Melanocytic Nevi

Hidemi Yasuda, M.D., Hitoshi Kobayashi, M.D., Akira Ohkawara, M.D., and Noboru Kuzumaki, M.D. Department of Dermatology (HY, HK, AO), Laboratory of Molecular Genetics (NK), Cancer Institute, Hokkaido University School of Medicine, Sapporo, Japan.

Ras oncogene expression of malignant melanoma and me­lanocytic nevus have been immunohistochemically analyzed on formalin-fixed and paraffin-embedded tissues from 26 melanomas and 24 melanocytic nevi with a monoclonal anti­body that was generated against Harvey sarcoma virus-de­rived ras oncogene products (p21"u). We found distinct dif­ferences of p21 ras expressions by the type of melanoma. Nodular melanoma, epithelioid cell type melanoma, and deeply invading melanoma revealed higher reactivity with anti-p21 ras monoclonal antibody than the other types. The reactivity of melanomas appeared to correlate with the de­gree of malignancy of the melanoma. It was also demon-

A number of monoclonal antibodies (MoAb) against melanoma have been reported since the early report of Koprowski et al [1], and most of them have been raised against cell membranes or intracellular organelles of melanoma cells. In recent years, the role of oncogenes

in carcinogenesis has been extensively studied and MoAb against oncogene products have been produced to investigate the biologic function of oncogenes and their clinical significance [2-10]. Among various oncogenes, a member of the ras family has been intensively studied, because it was demonstrated that ras oncogenes have a strong transforming capacity [11]. Three human ras genes have been identified: Ha-(Harvey), Ki-(Kirsten) and N-ras. Ras gene products are known to comprise a family of proteins of molec­ular weight 21 kD designated as p21"", and to be located on the cytoplasmic side of the cell membrane [12]. Furth et al [2] used rat cells transformed by Harvey sarcoma virus as immunogens to gener­ate MoAb against p21''''. They found that human colonic adenocar-

Manuscript received June 16, 1988; accepted for publication December 14,1988.

This study was partially supported by Grant-in-Aid for Cancer Research from the Ministry of Education Science and Culture, and for Comprehen­sive 10-year Strategy for Cancer Control from the Ministry of Health and Welfare, Japan.

Reprint requests to: Dr. Hiderni Yasuda, Department of Dermatology, Hokkaido University School of Medicine, Kita-1S, Nishi-7, Kita-ku, Sap­poro 060, Japan.

Abbreviations: ABC: 3-amino-9-ethyl-carbazole ALM: acrallentiginous melanoma MoAb: monoclonal antibody NM: nodular melanoma p21'''; ras oncogene products of mol wt 21 kD PBS: phosphate-buffered saline rp-12: anti-p21'" monoclonal antibody SSM: superficial spreading melanoma

strated, however, that part of melanocytic nevi reacted with anti-p21 ras monoclonal antibody with a relatively strong in­tensity. Melanocytic nevi with junctional activity and nevus cells located in the epidermis in compound nevi did not show the positive reaction in contrast to dermally located nevus cells that had relatively strong reactivity. The different p21 ras

expression among the type of tumors may represent the state of tumor cells differentiation with greater expression with more immaturity in the melanocyte lineage. p21 ras expression does not appear to represent a marker of malignant transfor­mation.] Invest Dermato193:54-59, 1989

cinomas in general showed a similar staining intensity to that seen in normal mucosa. Adenomas, however, showed consistently high p21 ,u expression, suggesting that elevated p21 nu expression may be important in the development of adenomas, but that high levels need not to be sustained in the conversion to invasive carcinoma [3]. Schlom et al used a synthetic peptide reflecting positions 10-17 of the Ha-ras protein product from the T24 bladder carcinoma as im­munogens, to generate MoAbs against p21'u, one of which reacted highly with human colon carcinomas l4,5], mammary carcinomas [6], bladder carcinomas [7], prostate carcinomas [8], and stomach adenocarcinomas [9]. p21'u expression correlated with the depth of carcinoma within the bowel wall, suggesting that p21'u expression may be a relatively rae event in colon carcinogenesis [5].

There have been several reports about the role of oncogenes in the tumorigenesis of melanoma [13-16]. Because it has been pointed out that the activation of ras gene by the point mutation [15] or enhanced transcription [16] may play some role in carcinogenesis of melanoma, p21'u expression of melanoma was immunohistochem­ically examined using monoclonal anti-p21'u antibodies (rp-12). Preliminary studies using an immunofluorescense technique showed that some melanomas express p21'u on their cell mem­branes but not melanocytic nevi [17], therefore, we used formalin­fixed and paraffin-embedded melanomas and melanocytic nevi specimens for immunohistochemical analysis (avidin-biotin-perox­idase complex method) to evaluate the role of ras oncogenes in melanocytic tumors.

MATERIALS AND METHODS

Subjects Surgically excised tissues were obtained from 26 mela­noma patients*: 14 acrallentiginous melanoma (ALM), 7 nodular

• Surgically excised samples were kindly supplied by Dr. K. Inoue (De­partment of Pathology. Hokkaido University Hospital), Dr. T. Sugihara (Department of Plastic Surgery, Hokkaido University School of Medicine), and Dr. T. Aoyagi (Otaru, Hokkaido, Japan).

0022-202X/89/S03.S0 Copyright © 1989 by The Society for Investigative Dermatology, Inc.

S4

VOL. 93, NO. I JULY 1989

melanoma (NM), 3 superficial spreading melanoma (SSM), and 2 metastasis. And from 24 patients with melanocytic nevus·: 6 junc­tional, 7 compound, and 11 dermal. Malignant melanomas were analyzed according to Breslow's classification, TNM classification, and dominant cell types.

Anti-p21- MoAb Proteins as immunogen for hybridoma prepa­ration were produced in Escherichia coli bearing the plasmid pJLcII­ras1 containing most of Harvey sarcoma virus ras(Ha-ras) gene. More than 10% of the total proteins were constituted by cII-ras fusion protein. Four hybridoma cell lines (rp-12, rp-28, rp-35, rp-38) producing anti-p21 r41 MoAb were established (NS-l

MELANOMA AND ras ONCOGENE 55

BALBjc mouse myeloma cells as parent cells). In the present study, rp-12 was used for immunohistochemical assays. Rp-12 recognize epitopes shared with p21 encoded by activated Ha-ras and Ki-ras genes derived from mice and humans and it was confirmed to be more specific than rp-35 and rp-38 by the preliminary study [17].

ImmuDoperoxidase Assays Immunohistochemical assays were performed using the modified methods ofHsu et al [18] as described previously. In brief, 4-f.Lm sections of formalin-fixed and paraffin­embedded tissus were deparaffinized and immersed in 80% metha­nol containing 0.3% H 20 2 to eliminate endogeneous peroxidase activity. The sections were rinsed in phosphate-buffered saline

Figure 1. Immunoperoxidase staining offormalin-fixed and paraffin-embedded human tissues with MoAb rp-12 using the ABC immunoperoxidase method. Counterstained with hematoxylin. a: Nodular melanoma, epithelioid cell type (63-yr-old man). The brownish-red reflects the reaction of the AEC substrate and thus MoAb, rp-12 binding to cytoplasmic p21 nu in melanoma cells. b: Nodular melanoma, epithelioid cell type (48-yr-old woman). Melanoma cells are strongly positive for cytoplasmic reactivity. c: Superficial spreading melanoma, epithelioid cell types (65-yr-old man). Melanoma cells are strongly positive for cytoplasmic reactivity. d: Acral lentiginous melanoma, spindle cell type (77-yr-old woman). Melanoma cells are slightly positive. t : Melanocytic nevus, junctional type (24-yr-old man). Basal cells of epidermis are positive for cytoplasmic reactivity, but nevus cells are negative.J Melanocytic nevus, dermal type (34-yr-old woman). Nevus cells are positive for cytoplasmic reactivity.

56 YASUDA ET AL

(PBS) and incubated in normal horse serum to suppress nonspecific binding of proteins to some tissues, particularly collagen. Rp-12 diluted 1 : 320 in PBS was applied for 30 min at room temperature. For the negative control, PBS was applied instead of MoAb. After washing in PBS, the sections were incubated with biotinylated horse anti-mouse IgM (Vector Laboratories, Inc., Burlingame, CA) for 30 min, and then treated with avidin dehydrogenase and biotin­ylated horseradish peroxidase H complex for 30 min. After another rinse in PBS, the sections were treated with 3-amino-9-ethyl-carba­zole (AEe) and were counterstained with hematoxylin.

Immunohistochemical Evaluation We used a modification of the method by Thor et al [5]. Each section was evaluated for the presence of AEC precipitate indicative of primary MoAb binding under a light microscope using the following scoring system: (-) negative; (+) clearly positive; (++) strongly positive. The approxi­mate percentage of cells expressing p21"" in melanoma tissues was scored according to the number of positive melanoma cells divided by the total number of melanoma cells. Results were statistically analyzed by the Chi-square test.

RESULTS

Immunoreactive p21"" was demonstrated as brownish-red granules within the cytoplasm of melanoma and nevus cells (Fig 1). These AEC precipitates were easily differentiated from melanin granules, and no AEC precipitates were observed in the control tissues. Be­sides melanoma and nevus cells, basal cells of the epidermis and eccrine glands revealed a positive reaction. The results of26 malig­nant melanomas and 24 melanocytic nevi were analyzed as follows.

Malignant Melanoma

Type oj Melanoma Among the types of malignant melanomas, dis­tinctive differences both in staining intensity and in the percentage of positive cells were demonstrated (Table I, Fig 2). In 14 cases of ALM, 1 was strongly positive,S were clearly positive, and 8 were negative. In 7 cases ofNM, 4 were strongly positive, 2 were clearly positive, and 1 was negative. This difference was statistically signifi­cant in which NM cells were more intensely stained than in ALM cells (X2 = 8.9, P < 0.02). In 3 cases of SSM and 2 cases of metas­tasis, all cases were clearly positive. The percentage of positive cells by rp-12 in NM was higher than in ALM or SSM.

Dominant Cell Type Melanoma cases were histopathologically clas­sified by their cytologic appearances into large epithelioid cell type, spindle cell type, and small cell type according to Lund and Kraus [19] (Table II, Fig 3). Usually, different cell types were admixed in the same section. A predominant cell type (more than 50% in the tissue) was used for the present analysis. It was also evident that reactivity of rp-12 was different among cell types. Of 16 cases of epithelioid cell type, 4 were strongly positive, 10 were clearly posi­tive, and 2 were negative. And in 8 cases of spindle cell type, 1 was strongly positive, 1 was clearly positive, and 6 were negative. It was statistically significant that epithelioid melanoma cells were more intensely stained than spindle cells (X2 = 9.6, P < 0.01). In 2 cases of small cell type, one was clearly positive and another was negative.

Table I. Immunoperoxidase Staining of Malignant Melanoma with MoAb rp-12 (Type of Melanoma)

Number of Cases ++ +

ALM 14 1 (7%) 5 (36%) 8 (57%) NM 7 4 (57%) 2 (29%) 1 (14%) SSM 3 o (0%) 3 (100%) o (0%) Metastasis 2 o (0%) 2 (100%) 0(0%) Total 26 5 (19%) 12 (46%) 9 (35%)

ALM = acral lentiginous melanoma; NM - nodular melanoma; SSM - sup<:rficial spreading melanoma.

Intensity score for irnmunoperoxidase staining: ++. strongly positive; +. clearly positive; - . negative .

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

(J) ..J ..J L&J U

L&J > E (J)

0 Q.

~ z L&J U a: L&J Q.

23% 57% 33%

%

100 - @

75 - @. @. @

I---l

50 - ... @ •

25 ,-

o

I---l

• f----f • ••

••• t.u.....u --'L

ALM NM SSM

• clea rl y positi ve @ st rongly positi ve

63%

•

~

•

Met a

Figure 2. Percentage of reactivity of MoAb rp-12 with formalin-fixed and paraffin-embedded malignant melanoma tissues using immunohistochemi­cal assay. Each symbol represents a different individual. The percentage of positive cells denotes the number of melanoma cells scoring p21"" positive divided by the total number of melanoma cells X 100 in each lesion (type of melanoma).

The percentage of positive cells in epithelioid cell type melanoma was also significantly higher than those in spindle cell type (X2 = 9.8, P = <0.05).

TNM Classification (stage) The reactivity of rp-12 was compared by the stage o(TNM classification. Most of the cases were stage I and stage II melanomas. We did not find any difference in staining intensity or the percentage of positive cells in each stage.

Breslow's Classification The group of more than 1.5-mm thickness melanomas was more strongly stained than the group of less than 1.5 mm thickness (X2 = 5.5, p < 0.1) (Table III, Fig 4). The per­centage of positive cells was also higher in the group more than t.5 mm in thickness.

M elanocytic Nevi Nevus cells in 13 of24 cases (55%) of melano­cytic nevi revealed a positive reaction with rp-12, however, a dis­tinctive difference was pres~nt in the reactivity among the types of nevi (Table IV, Fig 5). In 6 cases of junctional type, all cases were negative with rp-12. In 7 cases of compound type, t was strongly

Table II. Immunoperoxidase Staining of Malignant Melanoma with MoAb rp-12 (Dominant Cell Type)

Dominant Number Cell Type of Cases ++ +

Epithelioid Cell Type 16 4 (25%) 10 (63%) 2 (12%) Spindle Cell Type 8 1 (13%) 1 (13%) 6 (74%) Small Cell Type 2 0(0%) 1 (50%) 1 (50%)

Intensity score for immunoperoxidase staining: ++. strongly positive; +. clearly positiye~ -, negative.

VOL. 93, NO. 1 JULY 1989

%

100 VI ...J ...J I.IJ U

I.IJ 75 >

i= VI 0 a... ....

50 z I.IJ U 0:: I.IJ a...

25

a

50% 16 % 13 %

- @

- e®-@e @

-I-..@l •

- ••• • I----i ~

•• I •• : •• • epi thelioid spindle small ceil type ceil type ceil type

• clearly positive

@ strongly positive

Figure 3. Percentage of reactivity of MoAb rp-12 with malignant mela­noma (dominant cell type) .

positive, 3 were clearly positive, and 3 were negative for cells lo­cated in the dermis. Epidermal nests in the compound nevi did not react with rp- 12 as well as those in junctional nevi. In 11 cases of dermal type, 3 were strongly positive, 6 were clearly positive, and 2 were negative. Therefore, nevus cells were significantly less stained injunctional than in the other two types (X2 = 9.5, P < 0.01). The percentage of positive cells injunctional nevi was also significantly less than in the other two types (X2 = 10.4, P < 0.02) .

DISCUSSION

Several MoAb directed against p21 "" have been generated and iden­tified with reactivity to many neoplastic tissues, nonneoplastic tis­sues, and normal tissues [2 - 10]. We used an immunohistochemical assay that employed a MoAb(rp-12) generated against p21"" to in­vestigate the relationship between p21'" expression and malignant transformation of melanocytes. We have previously shown that lung cancer tissues of all the different histologic types were strongly reactive with rp-35 and those of well or moderately differentiated adenocarcinoma had high reactivity with rp-28. These cancer tis­sues showed membrane-associated staining, but bronchial epithelial and glandular tissues and macrophages in pleural effusion showed

Table m. Immunoperoxidase Staining of Malignant Melanoma with MoAb rp-12 (Breslow's Classification)

Number of Cases ++ +

<0.75 mm 4 0(0%) 1 (25%) 3 (75%) 0.75 - 1.5 mm 3 0(0%) 1 (33%) 2 (67%) 1.5 - 3.0 mm 5 1 (20%) 3 (60%) 1 (20%) > 3.0 mm 12 4 (33%) 5 (42%) 3 (25%)

In~ensity score for immunoperoxidase staining: ++, strongly positive; +, clearly posItive; - . negative.

%

100 VI ...J ...J I.IJ U

I.IJ 75 > E VI 0 a... .... 50 z I.IJ U 0:: I.IJ

-

-

a... 25 ,-

o

MELANOMA AND ras ONCOGENE 57

6% 17% 50%

. @

• -• ~

~

••• •• • ::>0.15 0.15 1.5

mm -1.5 -3.0

• clearly positive

@ strongly posit ive

42%

@

@. @

@ •

~

•••

.. ~ 3.0-

Figure 4. Percentage of reactivity of MoAb rp-12 with malignant mela­noma (Breslow's classification) .

cytoplasmic staining [20]. And also, rp-28 highly reacted with gas­tric cancers. Cell membrane and cytoplasm of malig­nant cells showed positive immunoreactivity in gastric cancers, but p21 ,., was observed in cytoplasm of positive cells in benign tissues [21] .

In the present study, we could not identify any differences in the pattern of cellular localization of immunoreactive p21 ,., detected by AEC precipitates. Both melanoma cells and nevus cells had the positive reactivity in their cytoplasm. Although p21 ros is considered a membrane-associated protein, the immunohistochemical presen­tation of p21 ,., may depend on the fixation method, more easily demonstrated in formalin-fixed and paraffin-embedded tissues compared with freshly frozen sections.

Because an activated Ha-ras gene was detected from cultured human melanoma cells by a DNA transfection method to NIH/3T3 cells [15), the role of ras genes may be important in melanoma oncogenesis. An interesting result in the present study was that both staining intensity and number of positive cells corre­lated with the tumor type, the cellular type, and the depth of inva­sion. The more malignant the melanomas were, the more inten­sively melanoma cells reacted with anti-p21'4SMoAb . Nodular

Table IV. Immunoperoxidase Staining of Melanocytic Nevus with MoAb rp-12

Number of Cases ++ +

Junctional 6 0(0%) 0(0%) 6 (100%) Compound 7 1 (14%) 3 (43%) 3 (43%) Dermal 11 3 (27%) 6 (55%) 2 (18%) Total 24 4 (17%) 9 (38%) 11 (45%)

Intensity score for immunoperoxid ... staining: ++, strongly positive; +. clearly positive: -. negative.

58 YASUDA ET AL

%

en 100 ...J -...J 1.&.1 (.)

1.&.1 > 75 ~

-en 0 a. ~ 50 z -1.&.1 (.) a:: 1.&.1 a. 25 -

a

0% 23%

@

• I

I • I

• • , ..... - •••

junctional compound

• clearly positive

@ strongly positive

39%

. @

• @.

• I

. @.

•• dermal

Figure S. Percentage of reactivity of MoAb rp-12 with melanocytic nevus tissues. The percentage of positive cells denotes the number of nevus cells scoring p21ras positive divided by the total number of nevus cells X 100 in each lesion.

melanoma is composed largely of an epithelioid type of cells, ALM tend$ to show a predominance of spindle-shaped cells [22]; NM has poorer prognosis than ALM and the prognosis correlates with tumor thickness [22,23]. Also it has been suggested that NM repre­sents the end stage of all types of melanoma instead of a distinctive entity [24). From these findings, it may be hypothesized that ras oncogene expression is involved in a relatively late event in mela­noma oncogenesis similar to colon carcinomas reported by Thor et al [5]. Because melanocytic nevi showed high reactivity with rp-12, however, other mechanisms must also be proposed for the expres­sion of p21'" in these melanocytic tumors.

The anatomic location of cells in melanocytic nevi does not ap­pear to influence their proliferative capacity but has a marked effect on their differentiation. Although cell kinetic studies of melanocy­tic nevus are limited, autoradiographs prepared from incubated nevus tissues showed that repricative activities were similar for all anatomic areas within melanocytic nevus [25] . The nevus cells dif­fered from one another by anatomic site irt cellular structure, mela­nin content, tyrosinase activity, and cholinesterase activity [26,27]' Nevus cells in the epidermis possess melanin granules and dendritic processes indistinguishable from dendrites of epidermal melano­cytes. Nevus cells in the papillary dermis commonly resemble epithelioid cells, aggregate in nests, demonstrate abundant tyrosin­ase activity and negative cholinesterase activity, and contain mela­nin. In contrast, nevus cells in the deep dermis may resemble fibro­blasts or Schwann cells, lie in single array, demonstrate minimal (but definite) tyrosinase activity, and usually do not contain mela­nin. Thus, nevus cells are homogeneous regarding proliferation but heterogeneous in their degree of differentiation comparable witb melanocyte differentiation from melanoblasts to mature melano­cytes. The different expressions of p21 , .. by the location of nevus cells may reflect this heterogeneity of nevus cells.

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

It may also be hypotbesized that the heterogeneous p21'" expres­sions in melanoma by various types represent the differentiation status of melanoma cells. Epithelioid cells usually present in NM are comparable to immature melanoblasts and possess abundant p21'''. Spindle or dendritic cells in ALM, on the other band, resemble mature melanocytes and possess less p21'''. Similar findings were demonstrated by a study using anti-melanoma MoAb in which the expressions of cell surface antigen correlated with the state of differ­entiation both in the evolution of melanocyte and in the type of melanoma [28].

It is well known that oncogenes present in normal cells (protoon­cogene) play an essential role in cellular functions, especially regu­lating cell proliferation or differentiation, or both, in addition to their role in carcinogenesis. p21'" has been shown to have partial bomology to GTP-binding protein [29] and have GTPase activity [30], suggesting the function of p21'" in the modulation of signals from membrane receptors into biologic cell responses. It has been demonstrated that the expressions of ras oncogenes are enhanced in normal tissues, such as regenerating liver [31] and fetal and neonatal development [32,33]. Some investigators have suggested that ras oncogenes are involved in cellular [34,35] and organ [36] differen­tiation as well. Although p21ras expression in normal melanocyte lineage has not been studied yet, it may be suggested that p21'" is expressed in nevus cells and melanoma cells that have similar char­acteristics, i.e. , less melanin production,less tyrosinase activity, and less dendritic processes. It is more likely the expression of p21'" is a marker of the undifferentiated stage of melanocyte lineage (evolu­tion) than it is a marker of malignant transformation. The present immunohistochemical study could not identify the role of ras onco­genes in the tumorigenesis of melanoma. It may be hypothesized, however, that malignant transformation is evoked in high fre­quency through ras gene activation because immature melanocytes have active expressions of ras genes.

REFERENCES

1. Koprowski H, Steplewski Z, Herlyn D, Herlyn M: Study of antibodies against human m"elanoma produced by somatic cell hybrids. Proc Nat! Acad Sci USA 75:3405 - 3409,1978

2. Furth ME, Davis L], Fleurdelys B, Scolhick EM: Monoclonal antibod­ies to the p21 products of the transforming gene of Harvey Murine Sarcoma Virus and of the cellular ras gene family.] Virol 43:294 -304, 1982

3. Will iams ARW, Piris], Spandidos DA, Wyllie AH: Immunohisto­chemical detection of the ras oncogene p21 product in an experi­mental tumour and in human colorectal neoplasms. Br ] Cancer 52:687 - 693, 1985

4. Hand PH, Thor A, Wunderlich D, Muraro R, Caruso A, Schlom J: Monoclonal antibodies of predefined specificity detect activated ras gene expression in human mammary and colon carcinomas. Proc Nat! Acad Sci USA 81:5227 - 5231,1984

5. Thor A, Hand PH, Wunderlich D, Caruso A, Muraro R, Schlom]: Monoclonal antibodies define differential ras gene expression in malignant and benign colonic diseases. Nature 311:562- 565, 1984

6. Ohuchi N, Thor A, Page DL, Hand PH, Halter SA, Schlom]: Expres­sion of the 21,000 molecular weight ras protein in a spectrum of benign and malignant human mammary tissues. Cancer Res 46:2511 - 2519,1986

7. Viola MV, Fromowitz F, Oravez S, Deb S, Schlom J: ras oncogene p21 expression is increased in premalignant lesions and high grade blad­der carcinoma.] Exp Med 161:1213 - 1218, 1985

8. Viola MV, Fromowitz F, Oravez S, Deb S, Finkel G, Lundy J, Hand P, Thor A, Sch lom J: Expression of ras oncogene p21 in prostate cancer. N Engl] Med 314:133 - 137,1986

9. Ohuchi N, Hand PH, Merlo G, Fujita], Mariani-Costantini R, Thor A, Nose M, Callahan R, Schlom]: Enhanced expression of c-Ha-ras p21 in human stomach adenocarcinomas defined by immunoassays using monoclonal antibodies and in situ hybridization. Cancer Res 47:1413 - 1420,1987

10. N gan BY, Chen-Levy Z, Weiss LM, Warnke RA, C leary M L: Expres­sion in non-Hodgkin's lymphoma of the bcl-2 protein associated

VOL. 93, NO. t JULY t 989

with the t(14;18) chromosomal translocation. N Engl ] Med 318:1638 - 1644,1988

11. Shih TV, Weeks MO: Oncogenes and cancer: the p21ras genes. Cancer Invest 2:109 - 123, 1984

12. Shih TV, Weeks MO, Gruss P, Dhar R, Oroszlan S, Scolnick EM: Identification of a precursor in the biosynthesis of p21 transforming protein of Harvey murine sarcoma virus.] Virol 42:253 - 261 , 1982

13. Greene MH, Goldin LR, Clark WH, Lovrien E, Kraemer KH, Tucker MA, Elder DE, Fraser MC, Rowe S: Familial cutaneous malignant melanoma: autosomal dominant trait possibly linked to the Rh locus. Proc Natl Acad Sci USA 80:6071 - 6075,1983

14. Linnenbach A), Huebner K, Reddy EP, Herlyn M, Parmiter AH, Nowell PC, Koprowski H: Structural alteration in the MYB pro­too nco gene and deletion within the gene encoding a-type protein kinase C in human melanoma cell lines. Proc Natl Acad Sci USA 85:74 - 78, 1988

15. Sekiya T, Hirohashi S, Nishimura S, Sugimura T : Transforming activ­ity of human melanoma DNA. Tpn] Cancer Res 74:794 - 797,1983

16. Ogiso Y, Oikawa T , Kondo N. Kuzumaki N, Sugihara T, Ohura T : Expression of proto-oncogenes in normal and tumor tissues of human skin. ] Invest Dermatol 90:841- 844, 1988

17. Kuzumaki N, Oda A, Yamagiwa S, Taniguchi N, Kobayashi H, Oi­kawa T: Establishment of four mouse hybridoma cell lines produc­ing monoclonal antibodies reactive with ras oncogene product p21. ] N C I 77:1273 - 1279,1986

18. Hsu S, Raine L, Fanger H : Use of Avidin-Biotin-Peroxidase Complex (ABC) in immunoperoxidase techniques: a comparison berween ABC and unlabeled antibody (PAP) procedures. ) Histochem Cyto­chem 29:577 - 580, 1981

19. Lund HZ, Kraus )M: Melanotic tumors of the skin. Atlas of tumor pathology. Section I-Facicle 3, A F I P Washington, 1962

20. Dosaka H, Harada M, Kuzumaki N, Kobayashi H, Isobe H, Miyamoto H, Kawakami Y: Immunohistochemical analysis of human lung cancers with anti-ras p21 monoclonal antibodies. Int) Bioi Markers 2:75 - 82, 1987

21. Nakajima K: Immunohistochemical study of ras p21 expression in human gastric cancers and benign lesions. Hokkaido Igaku Zasshi 62:573 - 580,1987

22. Clark WH, From L, Bernardino EA, Mihm MC: The histogenesis and biologic behavior of primary human malignant melanomas of the skin. Cancer Res 29:705 - 726,1969

MELANOMA AND ras ONCOGENE S9

23. Breslow A: Thickness, cross-sectional areas and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 172:902- 908, 1970

24. Heenan P], Holman CD]: Nodular malignant melanoma: a distinct entity or a common end stage? Am) Dermatopathol 4:477 - 478, 1982

25. Bentley-Phillips C, Marks R: Cell division and metabolic activity of naevus cells - the relationships berween anatomy and behaviour in moles. Br] Dermatol 94:557 - 563, 1976

26. Winkelmann RK, Minn R: Cholinesterase Nevus: cholinesterases in pigmented tumors of the skin. Arch Dermatol 82:71 - 77, 1960

27. Mishima Y: Macromolecular changes in pigmentary disorders. Arch DermatoI91 :519-557,1965

28. Houghton AN, Eisinger M, Albino AP, Cairncross JG, Old LJ: Surface antigens of melanocytes and melanomas. ] Exp Med 156:1755-1766,1982

29. Hurley]B, Simon MI, Teplow DB, Robishaw JD, Gilman AG: Ho­mologies berween signal transducing G proteins and ras gene prod­ucts. Science 226:860- 862, 1984

30. Stern we is PC, Robishaw]D: Isolation of rwo proteins with high affin­ity for guanine nucleotides from membranes of bovine brain. ] Bioi Chem 259:13806 - 13813,1984

31. Goyette M, Petropoulos C], Shank PR, Faust N : Expression of a cellu­lar oncogene during liver regeneration. Science 219:51 0- 512, 1983

32. Muller R, Siamon D], Tremblay JM, Cline M], Verma 1M: Differen­tial expression of cellular oncogenes during pre- and postnatal devel­opment of the mouse. Nature 299:640 - 644, 1982

33. Siamon D], Cline M]: Expression of cellular oncogenes during embry­onic and fetal development of the mouse. Proc Natl Acad Sci USA 81 :7141 - 7145,1984

34. Hagag N, Halegoua S. Viola M: Inhibition of growth factor-induced differentiation of PC12 cells by microinjection of antibody to ras p 21. Nature 319:680 - 682, 1986

35. Studzinski GP, Brelvi ZS: Increased expression of oncogene c-Ha-ras during granulocytic differentiation of HL60 cells. Lab Invest 56:499 - 504,1987

36. Chesa PG, Retting WJ , Melamed MR, Old L], Niman HL: Expression of p21 ras in normal and malignant human tissues: lack of association with proliferation and malignancy. Proc Nat! Acad Sci USA 84:3234 - 3238, 1987