Ras Mutations in Human Prolactinomas and Pituitary Carcinomas*

0021-972x/94/7801-M)89$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright 0 1994 by The Endocrine Society

Ras Mutations in Human Pituitary Carcinomas*

Vol. 78, No. 1 Printed in U.S.A.

Prolactinomas

WEN YI CAI, JOSEPH M. ALEXANDER, E. TESSA BERND W. SCHEITHAUER, J. LARRY JAMESON, NICHOLAS T. ZERVAS, AND ANNE KLIBANSKI

and

HEDLEY-WHYTE,

Neuroendocrine Unit (W.Y.C., J.M.A., A.K.), Department of Pathology (T.H. W.), Thyroid Unit (J.L.J.), Department of Neurosurgery (N. TX), Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02114; and the Department of Pathology, Mayo Clinic (B. W.S.), Rochester, Minnesota 55905 ABSTRACT

and

role in the pathogenesis of PRL-secreting pituitary tumors and/or may be a marker for tumor invasiveness and malignant transformation. Therefore, we investigated 78 pituitary tumors (59 prolactinomas, 13 invasive prolactinomas, and 6 pituitary carcinomas) for activating point mutation in the three ras genes using oligonucleotide-specific hybridization. In contrast to the relatively high frequency of ro.s mutations in many different tumor types, no ras mutations were identified in either prolactinomas or pituitary carcinomas. Our data indicate that ~0s mutations are rare in prolactinomas and pituitary carcinomas. (J Clin Endocrinol Metab 78: 89-93,1994)

Pituitary adenomas have been shown to be clonal in origin, indicating that one or more somatic mutations underlie tumor pathogenesis. Mutated oncogenic forms of ra.s protein have been identified in a number of human neoplasms, including thyroid adenomas and carcinomas. However, the potential role of activated ros in the development of specific human pituitary tumor phenotypes has not been determined. Although rus mutations were not found in glycoprotein hormonesecreting or somatotroph adenomas, we recently identified a mutation in the H-ras gene (Gly-Val) at codon 12 in a highly invasive prolactinoma. These data raise the possibility that ras mutations might play a

R

ECENT studiesusing molecular genetic techniques have demonstrated the clonal nature of human pituitary tumors (1). These data indicate that one or more somatic mutations are key events in the pathogenesisof such tumors. Well characterized mutations in the GTP-binding protein, GS~,have been shown to be requisite events in the development of approximately 40% of somatotroph adenomas(2, 3). However, G,a mutations in pituitary tumors appear to be restricted to somatotrophs, and little is known about the mechanismsof tumor formation in other pituitary cell types. It is likely that prevalent underlying somatic mutations may differ in specific pituitary tumor phenotypes. Members of the GTP-binding protein family have been shown to be potent stimulators of cellular proliferation. Previous work has shown that ras codons specifying residues 12 and 61 represent critical sites, which, when mutated, create oncogenie gene products in benign as well as malignant tumors (4, 5). Ras protooncogenes are important factors in cellular proliferation and differentiation, and have been shown to play a role in early stagesof tumorigenesisas well as tumor progression(4, 6). Therefore, data characterizing rus mutations in pituitary tumors could further identify cellular factors associatedwith tumor growth. The investigation of protooncogene abnormalities in en-

docrine neoplasia has only recently been initiated. Ras mutations have been identified in benign as well as malignant thyroid tumors, and their prevalence may reflect specific histological tumor findings (7). There are few data available regarding rus mutations in pituitary neoplasia. In our preliminary study of 19 pituitary tumors, we identified a mutation in codon 12 of the YRS gene in a highly invasive prolactinoma (8). No rus mutations were seenin somatotroph or glycoprotein hormone-secretingpituitary tumors. Therefore, the identified ras mutation may be unique to the pathogenesis of lactotroph adenomas. Activating ras mutations may also be associatedwith invasiveness or malignant transformation of pituitary neoplasms.For example, in colon tumors, there is a striking difference between the mutation prevalence of the K-ras gene depending on the degree of tumor invasiveness (9). To determine whether the presence of ras mutations underlie prolactinoma development and whether such mutations can be used as a marker for tumor invasiveness, we investigated 78 pituitary tumors (59 prolactinomas, 13 invasive prolactinomas, and 6 pituitary carcinomas)for activating point mutation in the three ras genesusing oligonucleotidespecific hybridization. Materials Isolation

Received July 20, 1993. Accepted September 7, 1993. Address all correspondence and requests for reprints to: Anne Klibanski, M.D., Neuroendocrine Unit, Jackson 1021, Massachusetts General Hospital, Boston, Massachusetts 02114. * This work was supported in part by NIH Grant DK-40947 and an American Cancer Society IRG Award.

of

and Methods

DNA from tumor tissue

Prolactinoma samples were obtained from both paraffin-embedded specimens and resected tumor tissue. Tumor diagnosis was verified in all tissue samples used for genomic DNA extraction and ras mutation analysis. Genomic DNA was then isolated and purified for DNA amplification by polymerase chain reaction, as previously described (10, 11).

89

TABLE

JCE& M

CA1 ET AL.

90 1. Synthetic oligonucleotide

probes used to screen for ra.s mutations

Oligonucleotide name(s) H-ra.s-12 and -13/normal H-ras-12/2-4

sequence GTGGGCGCCGGCGGTGTGGG NGC

H-ras-1215-7

GNC

H-rwGl/normal (antisense) H-ra.s-61/2-4 (antisense) H-ras-61/5-7

TACTCCTCCTGGCCGGCGGT CTN

(antisense)

CNG

H-ra.s-61/8,9 (antisense) K-ras-12 and -13/normal (antisense) K-ra.+12/2-4 (antisense) K-ras-12/5-7

NTG CCTACGCCACCAGCTCCAAC ACN

(antisense)

ANC

K-m-s-13/2-4 (antisense)

GCN

K-ra.s-13/5-7 (antisense)

GNC

K-ra.s-Gl/normal (antisense) K-ra.s-61/2-4 (antisense)

TACTCCTCTTGACCTGCTGT TTN

(antisense)

TNG

K-ras-61/8,9 (antisense)

NTG

N-ra.s-12 and -13/normal N-ras-1212-4

GGAGCAGGTGGTGTTGGGAA NGT

K-ras-61/5-7

Vol%.Nol

GNT

N-m-12/5-7 N-ro.s-13/2-4

NGT

N-ras-13/5-7

GNT

N-ra.s-Gl/normal (antisense) N-roe-61/2-4 (antisense)

TACTCTTCTTGTCCAGCTGT TTN

N-ras-61/5-7

(antisense)

TNG

K-roe-61/8,9

(antisense)

NTG

Amino acid GUY Arg

Ser CYS Ala Asp Val Gln Gln stop LYS Leu Arg Pro His His ‘JY CYS Ser Arg Val Asp Ala CYS Ser As Val Asp Ala Gln Gln stop LYS Ax Leu Pro His His Glr Ser Arg CYS Asp Ala Val Ser Arg CYS Asp Ala Val Gln Glu stop LYS Arg Leu Pro His His

-1994

RAS MUTATIONS TABLE

IN PROLACTINOMAS

2. Summary of clinical features of patients with prolactinomas Tumor type Prolactinoma

Sex

AgeW

(n = 59)

Invasive prolactinoma

(n = 13)

16-65 (median, 28) 18-84 (median, 34)

Sex

Age

1 2 3 4

M F F

69 60 68 65

5 6

M M

63 31

IlO.

(yr)

Tumor type Corticotroph carcinoma Lactotroph carcinoma Corticotroph carcinoma Silent corticotroph carcinema Corticotroph carcinoma Corticotroph carcinoma

size

Serum PRL

F

Macro

Micro

22

37

43

16

8

5

13

Samples analyzed for ~a.5 mutations

;; 2 2 1’ 1

0 One primary pituitary tumor and one metastasis. * Two primary pituitary tumors, one recurrent tumor, and one metastasis. ’ One primary tumor. Ras mutation

Tumor

M

TABLE 3. Summary of clinical features of patients with pituitary carcinomas Case

91

analysis

Oligonucleotide primers were used to specifically amplify codons 121 13 and 61 from three ras alleles (N-ras, K-ras, and H-ras). Synthetic oligonucleotides specific for each ras mutation were used as controls for hybridization analysis. A panel of 55 oligonucleotide probes was used to detect ras mutations in each of the tumor samples and is detailed in

k/L)

20-7,000 (median, 230) 1291-24,000 (median, 4,000)

Table 1. Amplified ras from genomic DNA was blotted onto GeneScreen Plus (New England Nuclear, Boston, MA) nitrocellulose membranes and hybridized with specific 32P-labeled oligonucleotide probes, as previously reported (12, 13).

Results Prolactinoma

clinical

data

Clinical data from 72 patients with prolactinomas and invasive prolactinomas are summarized in Table 2. Positive immunohistochemical staining for PRL was detected in all tumor specimens.Fifty-nine patients with PRL-secreting adenomaswere studied. Patients ranged in age from 16-65 yr (median, 28 yr). There were 37 women and 22 men in this group. All patients had hyperprolactinemia, with PRL levels ranging from 20-7000 rg/L (median, 230 pg/L). Sixteen tumors were microadenomas,and 43 tumors were macroadenomas. The second group of patients (n = 13) had invasive macroprolactinomas, as defined by evidence of gross invasion of dura and/or bone at the time of neurosurgical resection.

Wild Type To 1

Gin (TTG)

FIG. 1. Hybridization of polymerase chain reaction-amplified DNA from 2’7 pituitary tumors to oligonucleotide probes for codon 61 of the N-W gene. A, Hybridization to the normal probe for codon 61 of N-ras. (Gin). B, Hybridization to two mutant probes for codon 61 of N-m-s., N61-3 (termination codon) and N61-4 (Lys). The arrows indicate control DNA containing normal or mutant codons for N-61.

7Norma) Control

1

(Nc,l-3)

1

(N&I-4)

Mutanl cant rols

CA1 ET AL.

92

Patients ranged in age from 18-84 yr (median, 34 yr old). There were five women and eight men in this group. All patients had hyperprolactinemia, with PRL levels ranging from 1,291-24,000 pg/L (median, 4,000 pg/L). Pituitary

carcinoma

clinical data

The clinical data of the six patients with pituitary carcinomas are shown in Table 3. All patients had evidence of intracranial or extracranial metastases,confirmed by surgical excision (casesl-4) or autopsy (cases5 and 6). Five of the six carcinomas were of corticotroph origin, and one was of lactotroph origin. The clinical histories of four patients (no. l-4) have been previously reported (14-16). Twelve specimens, including both primary and recurrent tumors as well asmetastases,were examined. Determination

of

ras mutations

Eighty-four tumor samplesfrom 78 patients were analyzed for YUS mutations by oligonucleotide-specific hybridization. Hybridization of a panel of 55 oligonucleotide probes to amplified tumor DNA was performed (Table 1). None of the probes detected ras mutations in any of the prolactinomas, invasive prolactinomas, or pituitary carcinomas studied. Wild-type oligonucleotide probes showed positive hybridization to all amplified tumor DNA. Figure 1 shows a representative specific hybridization to amplified N-ras alleles using both wild-type (TTG) and two mutant probes (TTGTTA and TTG-TTT) specific for the N-61 activating site. Discussion We have investigated whether constitutive ras activation occurs in human pituitary adenomas and carcinomas. In contrast to the relatively high frequency of ras mutations in many human neoplasms,no ras mutations were identified in either prolactinomas or pituitary carcinomas. Our previous study identified a mutation in codon 12 of the H-ras gene (Gly to Val) in a singleinvasive prolactinoma (8). Our current data, examining a large series of PRL-secreting cell adenomas, invasive prolactinomas, and pituitary carcinomas, indicate that ras mutations are rare and, consequently, do not play an important role in human pituitary tumorigenesis. The identification of mutant ras genes in benign tumors suggeststhat YUS somatic mutations may be involved in adenoma development by conferring a selective growth advantage after the initiation of tumorigenesis (5, 17). The presenceof ras mutations has been investigated in a number of endocrine tumors, and such mutations have been described in both benign and malignant thyroid neoplasms. Ras mutations have been reported in both microfollicular and macrofollicular adenomas as well as papillary carcinomas(13,18). In a study by Karga et al. (12), an H-ras mutation in codon 12 was identified in 1 of 15 papillary carcinomas, and an N-ras mutation in codon 61 was identified in 2 of 14 follicular carcinomas(12). Collectively, these studiesconfirm that 20-50% of thyroid tumors contain activating ras mutations. In contrast to thyroid neoplasms,ras mutations have

JCE & M.

1994

not been found in parathyroid adenomas (19). Therefore, activating ras mutations may play a role in a subset of endocrine neoplasms, but do not appear to be a general mechanismof pituitary tumorigenesis.In addition, such mutations may occur more commonly in malignant, rather than benign, endocrine tumors. Although one or more somatic mutations underlie pituitary tumor development, current data suggestthat a single common mutation does not contribute to the pathogenesisof all pituitary tumor phenotypes. The finding of G,a mutations in only a subset of somatotroph adenomas indicates that somatic mutations may differ according to pituitary tumor cell type. Little is known about the pathogenetic mechanisms leading to prolactinoma development. A well characterized oncogene, hst, a member of the fibroblast growth factor family, was derived from two of five PRL-secreting tumors (20). Genomic DNA from these tumors was shown to transform NIH 3T3 fibroblast cells in vitro and induce solid tumors in nude mice. Southern hybridization and sequenceanalysis revealed no homology to other known oncogenes.Amplification of the v-fos gene in a single prolactinoma has also been reported (21). However, genomic amplification of transforming oncogeneshas not been demonstrated to be a common pathogenetic mechanismin pituitary adenomasor other tumor types (22, 23). As shown in the present study, our previous finding of a ras mutation in a single highly invasive prolactinoma is not representative of a common mechanism in lactotroph tumor formation. Therefore, although a number of potential candidate oncogenes have been examined in prolactinomas, no mutation has been convincingly demonstrated in a large seriesof such tumors. Further studies are needed to identify the requisite genomic alterations that give rise to lactotroph neoplasia. References 1. Alexander JM, Biller BM, Bikkal H, Zervas NT, Arnold A, Klibanski A. 1990 Clinically nonfunctioning pituitary tumors are monoclonal in origin. J Clin Invest. 86:336-340. 2. Spada A, Arosio M, Bassetti M, Vallar L, Clementi E, Bazzoni N. 1991 Mutations in the alpha subunit of the stimulatorv regulatory protein of adenylyl cyclase (Gs) in human GH-secreting $tuitary adenomas. Biochemical, clinical, and moroholoeical asuects. Path01 I v I Res Pratt. 187:567-570. 3. Vallar L, Spada A, Giannattasio G. 1987 Altered Gs and adenylate cyclase activity in human GH-secreting pituitary adenomas. Nature. 330:566-568. 4. Barbacid M. 1987 Ras genes. Annu Rev Biochem. 56:779-827. 5. Bos JL, Fearon ER, Hamilton SR, et al. 1987 Prevalence of ras gene mutations in human colorectal cancers. Nature. 327:293-297. H, Steinberg JJ, and Pellicer A. 1988 H-ras 6. Leon J, Kamino activation in benign and self-regressing skin tumors (keratoacanthomas) in both humans and an animal model system. Mol Cell Biol. 8:786-93. 7. Fagin JA. 1992 Genetic basis of endocrine disease. III. Molecular defects in thyroid gland neoplasia. J Clin Endocrinol Metab. 75:1398-1400. 8. Karga HJ, Alexander JM, Hedley-Whyte ET, Klibanski A, Jameson JL. 1992 Ras mutations in human pituitary tumors. J Clin Endocrinol Metab. 74:914-919. 9. Forrester K, Almoguera C, Han K, Grizzle WE, Perucho M. 1987 Detection of high incidence of K-ras oncogenes during human colon tumorigenesis. Nature. 327:298-303. 10. Gross-Bellard M, Oudet P, Chambon P. 1973 Isolation of high-

RAS MUTATIONS molecular-weight 36:32-38. 11.

DNA

from

mammalian

Goelz SE, Hamilton

SR, Vogelstein

human

15.

Suarez HG, du-Villard

of DNA Biophys

malignant

18.

thyroid

JA, Severino M, et al. 1990 Presence

in all three ras genes in human

thyroid

tumors.

of

Oncogene.

Pernicone PJ, Scheithauer

SW, Kovacs K, Sebo T. 1992 Pituitary

carcinoma: a clinicopathologic study of nine cases. Endocr

and proliferative Pathol. 3:536.

cell nuclear

antigen

Scheithauer BW, Randall RV, Laws Jr ER, Kovacs KT, Horvath E, Whitaker MD. 1985 Prolactin cell carcinoma of the pituitary. Clinicopathologic, immunohistochemical, of a case with cranial and extracranial 604.

16.

tissue.

Karga H, Lee JK, Vickery Jr AL, Thor A, Gaz RD, Jameson JL.

mutations 5:565-570. 14.

J Biochem.

B. 1985 Purification

embedded

1991 Ras oncogene mutations in benign and neoplasms. J Clin Endocrinol Metab. 73:832-836. 13.

Eur

17.

from formaldehyde and paraffin Res Commun. 130:118-126. 12.

cells.

IN PROLACTINOMAS

and ultrastructural metastases. Cancer.

malignant

chromophobe

tumor

20.

21.

22.

23. of the pituitary

and extraaanial

metastasis: report of a case. Proc Staff Mayo Clin. 37:31-42. Bos JL. 1989 Ras oncogenes in human cancer: a review. Cancer Res. 49:4682-4689. Namba H, Rubin SA, Fagin JA. 1990 Point mutations of rus oncogenes are an early event in thyroid tumorigenesis. Mol Endocrinol. 4:1474-1479. Friedman E, Bale AE, Marx SJ, et al. 1990 Genetic abnormalities in sporadic parathyroid adenomas. J Clin Endocrinol Metab. 71:293297. Gonsky R, Herman V, Melmed S, Fagin J. 1991 Transforming DNA sequences present in human prolactin-secreting pituitary tumors. Mol Endocrinol. 5:1687-1695. U HS, Kelley P, Lee WH. 1988 Abnormalities of the human growth hormone gene and protooncogenes in some human pituitary adenomas. Mol Endocrinol. 2:85-89.

Nagai MA, Habr-Gama Association oncogenes Dis Colon

study 55:598-

Scholz DA, Gastineau CF, Harrison Jr EG. 1962 Gushing’s syndrome with

19.

93

Sunderman mosomal induced

A, Oshima

of genetic alterations in colorectal tumors. Rectum. 35:444-451.

of c-WC, Frequency

CT, Brentani

MM.

1992

and c-Ha-ras urotoand clinical signifiiance.

c-fos,

Jr FW, Hopfer SM, Nichols WW, et al. 1990 Chro-

abnormalities and gene in rats by nickel subsulfide.

amplification in renal cancers Ann Clin Lab Sci. 20:60-72.