amplification and overexpression of the egf receptor gene in primary ...

14 downloads 0 Views 8MB Size Report
Roman numerals denote cDNA fragments used as hybridization probes (I, ..... U llrich , A., Coussens, L., Hayflick, J. S., D u ll, T. J., Gray, A., Tam, A. W., Lee, J.,.
J . Cell Sci. Suppl. 3, 161-172 (1985) Printed in Great Britain © The Company of Biologists Limited 1985

161

AM PLIFICATION AND OVEREXPRESSION OF T H E EGF R EC EPTO R GENE IN PRIMARY HUM AN GLIOBLASTOMAS T O W IA A. L IB E R M A N N 1, H A R R IS R . N U S B A U M 1, N I S S IM R A Z O N 2, R IC H A R D K R I S 1, I R I T L A X 1, H ER M O N A S O R E Q 3, N IG E L W H I T T L E 4, M IC H A E L D . W A T E R F IE L D 4, A X E L U L L R I C H S a n d JO S E P H S C H L E S S IN G E R 1’ * 1Department o f Chemical Immunology, The Weizmann Institute o f Science, Rehovot, Israel 2Department o f Neurosurgery, Ichilov Hospital, Tel Aviv M edical Center, Tel Aviv, Israel 3Department o f Neurobiology, The Weizmann Institute o f Science, Rehovot, Israel 4Protein Chemistry Laboratory, Imperial Cancer Research Fund, Lincoln’s Inn Fields, London WC2A3PX, U.K. 5Department o f Molecular Biology, Genentech, Inc., 460 Point San Bruno Boulevard, South San Francisco, California 94080, U.SA.

SUMMARY

The expression of epidermal growth factor (EGF) receptor in brain tumours of glial origin was studied at the protein, mRNA and genomic levels. Four out of 10 glioblastomas that overexpress EG F receptor also have gene amplification. The amplified genes appear to be rearranged, generating an aberrant mRNA in at least one of these tumours. Such receptor defects may be relevant to tumorigenesis of human glioblastomas.

INTRODUCTION

Epidermal growth factor (EG F) is one of a structurally diverse series of poly­ peptides that, through interaction with specific cell surface receptors, generate a pleiotropic response that by an ill-defined mechanism can induce a mitogenic response in target cells (Carpenter & Cohen, 1979; Schlessinger et al. 1983). Subversion of the EG F mitogenic signal through expression of a truncated receptor (Downward et al. 1984; Ullrich et al. 1984; Yamamoto et al. 1983) seems to be involved in transformation by the avian erythroblastosis virus (AEV) oncogene erbB, suggesting that similar EG F receptor defects may be found in human neoplasias. Overexpression of EG F receptors has been reported on the epidermoid carcinoma cell line A431 (Fabricant, DeLarco & Todaro, 1977) and in various primary brain tumours (Libermann et al. 1984). In A431 cells this is related to amplification of the receptor gene and is accompanied by gene re-arrangement, which causes production of a truncated extracellular EG F binding domain (Ullrich et al. 1984; Merlino et al. 1984; Weber, Gill & Spiess, 1984; Lin et al. 1984). * Author for correspondence.

162

T. A. Libermann and others

The target cell specificity for different growth factors is in part governed by the expression of high-affinity cell surface receptors. These receptors transduce the signal generated by ligand binding and, since the expression of altered receptors can in the case of the oncogene v-erbB of AEV contribute to the transformation process, the expression of receptors in neoplasias is under investigation. We have obtained cDNA clones that have provided the complete primary structure of the human E G F receptor (Ullrich et al. 1984). These probes show that the human A431 epidermoid carcinoma cell line has amplified normal and rearranged receptor genes (Ullrich et al. 1984). Amplification of other proto-oncogenes has been recently detected in several tumours and tumour-derived cell lines (Ullrich et al. 1984; Brodeur et al. 1984; Pelicci et al. 1984; Collins & Groudine, 1982). It was proposed that oncogene amplification might play a role in tumorigenesis by leading to excess production of the gene products. Hence, we have used the EG F receptor complementary DNA to examine various neoplasias for abnormal receptor expression at the protein, mRNA and genomic level.

M A T E R I A L S A ND M E T H O D S

Cloning of E G F receptor cDNA clone p8 mRNA from A431 cells was isolated with the guanidine thiocyanate/caesium chloride method (Maniatis, Fritsch & Sambroot, 1982). cDNA synthesis, cloning into pUC9 and colony screening were carried out according to published protocols (Helfman et al. 1983). Briefly, a 17-mer probe, involving a mixture of 256 oligonucleotides (3' A T A / G T T A / G G G X T G X T G X A T 5 ') based on the amino acid sequence of a tryptic peptide of A431 EG F receptor, was endlabelled with bacteriophage T 4 polynucleotide kinase (New England Biolabs) and [y-32P]ATP (Amersham, 3000 Cimmol-1). Colony screening hybridization was carried out in 6 x S S C (SSC is 0-15 M-NaCl, 0-15M-sodium citrate, pH7-0), 5XDenhardt’s solution (D S: 0-2% (w/v) Ficoll, 0-2% (w/v) polyvinylpyroliclone, 0-2% (w/v) bovine serum albumin), 1 % sodium dodecyl sulphate (SD S), 100 fig ml-1 salmon sperm DNA at room temperature for 36 h. Filters were washed at 45 °C with 3X SSC , 0-1 % D S. Nucleotide sequence analysis was carried out according to Maxam & Gilbert (1980). p8 is a 2 -5 x l0 3base-pair (bp) cDNA clone derived from the 2-8X103bp variant mRNA from A431 cells described by Ullrich et al. (1984). p8-4 is a Pst fragment (399 bp) derived from clone p8.

Southern blot analysis of genomic DNA from tumours and cell lines High molecular weight chromosomal DNA was isolated as described (Maniatis et al. 1982). DNA (15 g) was digested to completion with excess of either ZscoRI or Hin&lW (New England Biolabs), fractionated by electrophoresis through a 0-7% agarose gel and transferred to nitro­ cellulose paper. EGF-receptor cDNA inserts were radiolabelled with [a,-32P]dATP and [o'-32?] dCTP (Amersham) by the procedure of Taylor, Illmensee & Summers (1976). Hybridization with 107ctsmin-1 of 3 P-labelled probe was performed in 6XSSC , 5xDenhardt’s solution, 10% dextran-sulphate, 50 mM-sodium phosphate (pH 6-5) and 100 fig ml-1 salmon sperm DNA at 65 °C for 16 h. Filters were washed in 0-2XSSC, 0-1 % SD S at 65°C and autoradiographed for 1 day at —70°C using intensifier screens. Sizes were calculated using phage DNA cleaved with restriction endonuclease H in d lll as standards.

Southern blot analysis (Fig. 2) High molecular weight DNA (15fig) of A431 cells (A431), human placenta (HPL), glio­ blastomas (GM1, GM2) and meningioma (MEN1) was digested w ith£coRI (a) or/itndlll (b). In

EG F receptor expression in glioblastomas

163

Table 1. Properties of human brain tumours Tissue designation Controls HPL A431 Brain tumours GM1 GM2 (6) GM3 GM4 GMS GM6 GM7 (20) GM8 GM9 GM10 CGM1 (10) CGM2 ACII1 (2) ACII2 ODG1 ODG2 ODG3 MEN1 MEN2 MEN3 PNE

Source of DNA Human placenta Epidermoid vulval carcinoma

No. of E G F receptor gene copies

EG F receptor expression

1 15-25

High High

Glioblastoma multiforme 20-30 Glioblastoma multiforme 10-20*/35-45f 1 Glioblastoma multiforme 1 Glioblastoma multiforme 1 Glioblastoma multiforme 50-60 Glioblastoma multiforme 1 Glioblastoma multiforme 1 Glioblastoma multiforme 6 -9 f Glioblastoma multiforme \ Adjacent normal brain tissue 1 1 Glioblastoma multiforme 1 Cerebellar glioblastoma Cerebellar glioblastoma 1 Astrocytoma grade II 1 Astrocytoma grade II 1 1 Oligodendroglioma 1 Oligodendroglioma 1 Oligodendroglioma Meningioma 1 Meningioma 1 Meningioma 1 1 Primitive neurocotodermal tumour

High High ND Low ND High Low None ND ND ND Low ND Medium ND ND Low Medium Medium ND Medium ND

Tumour tissue was frozen in liquid nitrogen immediately after removal to avoid degenerative changes. As only small amounts could be obtained, the kinase assay provides the only practical assay for E G F receptor levels in these samples at present. The radioactive content of the 32Plabelled E G F receptor was determined and expressed semi-quantitatively as follows: low, < 2000c.p.m .; medium, 2000-20000c.p.m .; high, > 2 0 000c.p.m. For further details see Libermann et al. (1984). ND, not determined. Numbers in parentheses refer to tumours analysed previously. * 5 ' probe (II); f 3 ' probe (V) (see Fig. 2).

order to exclude technical artifacts due to incomplete digestions, the digestions were repeated several times with large excess of restriction enzymes, producing the same results. The DNAs were electrophoresed and blotted. The blots were hybridized to the nick-translated E G F receptorspecific cDNA inserts isolated from the cDNA clones depicted in Fig. 1a and Fig. 2 under highstringency conditions. The same blots were reutilized for the different probes without detectable loss of signal. For denaturation of the cDNA-genomic DNA hybrids for reutilization, filters were soaked with slight agitation in 0-5M-NaOH, 1-5 M-NaCl for lOmin at room temperature, rinsed with water, neutralized twice for lOmin each in 0-5M-Tris •HC1 (pH 7-0), 1-5 M-NaCl and washed in 3 X SSC. Filters were kept wet in Saran wrap, preincubated in hybridization buffer and reused as described above. Broken lines in Fig. 2 indicate the regions of the schematically depicted cDNA giving rise to the hybridization patterns shown. Arrowheads indicate the location of DNA fragments in the amplified EG F receptor gene of the glioblastoma tumours, which are undetectable in other DNAs. Roman numerals denote cDNA fragments used as hybridization probes (I, p64-4; II, p8-4; II I, p64-l; IV, p64-3; V, p62-3).

T. A. Libermann and others

164

Northern blot analysis of mRNAfrom tumours and cell lines A431 cells were grown in Dulbecco’s modified Eagle’s medium containing 10 % foetal calf serum in an atmosphere of 5 % CO2/95 % air at 37°C. RNA was isolated from frozen tissue of human placenta, and glioblastoma GM1 after pulverization in liquid N2and from fresh A431 cells with the guanidine thiocyanate/caesium chloride method as described (Maniatis et al. 1982). Samples (10/ig) of poly(A)-selected mRNA were heated at 60°C for 10 min in a solution containing 50% form am ide, 6 % formaldehyde and running buffer (20 mM-MOPS, pH 7-0, 5 mM-sodium acetate, 1 mM-EDTA). The samples were electrophoresed at 100 V for 4h in 1 % agarose gels containing 6 % formaldehyde at 1X running buffer. The RNA was transferred with 10X SSC to nitrocellulose filters, fixed by heating at 80°C for 2 h and hybridized at 42°C for 2 days with 2 x 106 cts min-1 m P 1 of nick-translated p64-3 probe in a solution containing 50% formamide, 5X SSC , 10% dextran sulphate, lxDenhardt’s solution, lOmM-sodium phosphate (pH6-8) and lOOjUgml-1 salmon sperm DNA. After washing at 65 °C with 0-1X SSC, 041 % SD S, the filters were autoradiographed for 2 days at —70°C using intensifier screens.

Immunoprécipitation offunctional E G F receptor from tumours and cell lines Equal amounts (25 mg) of tissue from glioblastomas GM1, GM7 and GM8, meningioma MEN1, oligodendrogliomas ODG2 and ODG3 and human placenta HPL were solubilized as described by Libermann et al. (1984). Equal amounts of protein were used for immuno­ précipitations with either polyclonal rabbit antibodies TLM , generated against membranes enriched with EG F receptor from the A431 cell line (Libermann et al.-1984) or polyclonal rabbit antibodies RK2, generated against a synthetic peptide derived from the C-terminal domain of the EG F receptor (Kris et al. 1985). Functional EG F receptor kinase was immunoprecipitated and detected by phosphorylation of the immunoprecipitate using [y-32P] ATP according to Libermann et al. (1984). The immunoprecipitates were dissolved in electrophoresis sample buffer and electrophoresed in a 5 % to 15 % SDS/polyacrylamide gel. The gel was dried and auto­ radiographed for 12 h at room temperature. A431 cells were used as a standard source of the E G F receptor.

R E S U L T S AND D I S C U S S I O N

Brain tumour tissue was obtained from the primary tumours of 21 patients at surgery. Genomic DNA was prepared from these tumours and from human placenta Fig. 1. Southern blot analysis of DNA from primary human brain tumours and human placenta. A. Schematic representation of human EG F receptor cDNA probes from A431 epidermoid carcinoma cells used in this study and a diagram describing the complete cDNA structure. Untranslated sequences are represented by a line, coding sequences are boxed. The open portion represents sequences that encode the signal peptide; hatched regions code for the mature receptor. The filled region demarcates the coding region for the putative transmembrane domain. The probes used for this study are drawn above the diagram. Their construction has been described (Ullrich et al. 1984), except for that of probe p8-4, which is derived from the recombinant plasmid p8 (Libermann et al. unpublished results). p64-4 and the other probes were previously designated pHER-A64-4, respectively (Ullrich et al. 1984). B. Amplification of the E G F receptor gene in glioblastoma and A431 DNA detected by Southern blot analysis of DNA from primary human brain tumours, A431 epidermoid carcinoma cells and human placenta. The DNA from these tumours was digested with £eoRI and probed with 32P-labelled p64-l DNA. kb = 103bp. GM1, GM6, GM7 and GM8 are different primary human glioblastomas, CGM2 is a cerebellar glioblastoma, ACII2 is an astrocytoma grade II, MEN1 is a meningioma, A431 is a human epidermoid carcinoma cell line. HPL is human placenta, which served as a control tissue. PNE is a primitive neuroectodermal tumour; ODG1, ODG2 and ODG3 are oligodendrogliomas.

165

E G F receptor expression in glioblastomas p 6 4 .4

p 6 4 .3

p64.

p 6 2 .3

r i------- 1

8 .4 -

E c o R IP s flP s ü

EcoRI

EcoRI

EcoRI EcoRL

_J__ L_ AAA 0

3-0

20

10

Fig. UJ cl

00

rO CD

cvl v ­ CD CD

Q

Q

e? o

Q

o o

£

^ CD

40

5-0 kb

1a



^ CD

oo H

O