Antisense epidermal growth factor receptor RNA transfection in human malignant glioma cells leads to inhibition of proliferation and induction of differentiation

Neuropathology and Applied Neurobiology (1998), 24, 389–396

Antisense epidermal growth factor receptor RNAtransfection in human malignant glioma cells leadsto inhibition of proliferation and induction ofdifferentiationX. X. Tian*, P. Y. P. Lam*, J. Chen†, J. C. S. Pang*, S. S. T. To‡, E. Di-Tomaso* and H. K. Ng*

*Department of Anatomical & Cellular Pathology, Prince of Wales Hospital, The Chinese University of Hong Kong, Shatin, HongKong, China, †Department of Pathology, Peking Union Medical College Hospital, Beijing, China and ‡Department of Health Sciences,The Hong Kong Polytechnic University, Hong Kong, China

X. X. Tian, P. Y. P. Lam, J. Chen, J. C. S. Pang, S. S. T. To, E. Di-Tomaso and H. K. Ng (1998) Neuropathology andApplied Neurobiology 24, 389–396Antisense epidermal growth factor receptor RNA transfection in human malignant glioma cellsleads to inhibition of proliferation and induction of differentiation

The epidermal growth factor receptor (EGFR) is a proto- transfected cells. Flow cytometric analysis revealed thatthe proportion of cells in G

0/G

1phases of the cell cycleoncogene that is frequently observed with alterations in

late stage gliomas, suggesting an important role of this in the antisense clones increased by up to 31% comparedwith control cells, whereas the proportion of cells in Sgene in glial tumorigenesis and progression. In this study

we evaluated an antisense EGFR approach as an alterna- phase decreased by up to 58%. In addition, the antisenseEGFR-transfected cells showed higher expression of glialtive therapeutic modality for glioblastomas. We

transfected U-87MG cells with an antisense EGFR con- fibrillary acidic protein and a more diCerentiated form,with smaller cell bodies possessing fine tapering cellstruct and obtained several clones stably expressing

lower or undetectable levels of EGFR protein. These processes. These results suggest that EGFR plays a majorrole in modulating cell growth and diCerentiation inclones were found to have impaired proliferation as well

as a reduced transforming potential to grow in soft glioblastoma cells. Our experimental model of antisenseEGFR provides a basis for future development of anti-agarose. The number of cells positive for the cell cycle-

specific nuclear antigen Ki-67 was also significantly sense EGFR oligodeoxynucleotides in treatment ofglioblastomas.decreased (P<0.05) in antisense EGFR-transfected

clones compared with parental or empty vector-

Keywords: human malignant glioma, epidermal growth factor receptor, antisense, transfection

tyrosine kinase domain [25,30]. Binding of EGF orIntroduction

transforming growth factor to EGFR results in receptordimerization, autophosphorylation of the receptor itselfThe epidermal growth factor receptor (EGFR) is a trans-

membrane glycoprotein of 170 kDa and is composed of and phosphorylation of cellular substrates leading to celldivision and proliferation [30].an extracellular ligand-binding domain, a single hydro-

phobic membrane-spanning domain and a cytoplasmic In addition to serving a normal physiological function,EGFR has also been linked to transformation. There areseveral lines of evidence suggesting that aberrant regu-

Correspondence: Dr H. K. Ng, Department of Anatomical & Cellularlation of EGFR expression is tumorigenic. First, EGFR isPathology, Prince of Wales Hospital, The Chinese University of Hong

Kong, Shatin, Hong Kong, China. homologous to the transforming gene product of the

389© 1998 Blackwell Science Ltd

390 X.X. Tian et al.

avian erythroblastosis virus, v-erbB [4]. Second, over- Grand Island, NY, USA) supplemented with 10% fetalbovine serum, 100 mg/mL streptomycin and 100 U/mLexpression of normal human EGFR can transform cells

in culture [9]. Third, excessive activation of EGFR is penicillin, in a humidified atmosphere of 5% CO2

at37°C. Cells were transfected with empty vector, the sensefound in many human tumours including brain, breast

and lung [12,18,26]. or the antisense EGFR constructs by use of Transfectamreagent (Promega Corp., Madison, WI, USA) as describedIn glioblastomas, EGFR is found to be overexpressed

in up to 50% of all cases [12,14,23]. Genetic abnormali- by the manufacturer. Individual clones selected in1 mg/mL puromycin (Sigma Chemical Co., St Louis, MO,ties such as gene rearrangement and amplification are

detected in late stage gliomas suggesting an important USA) for 3 weeks were examined for the presence ofEGFR insert by PCR. The forward and reverse PCRrole played by EGFR in tumour proliferation and malig-

nant progression [12,17,28]. The oncogenic role of EGFR primers were 5∞-TGGATATTCTGAAAACCGTAA-3∞ (cor-responding to EGFR cDNA position 1361–1381), andin glioblastomas has led to extensive studies of its

relationship to glioma cell proliferation, diCerentiation, 5∞-AGTCCCTTATACACCGTGC-3∞ (position 2375–2357)[25], respectively. These primers are derived frompatient survival as well as its potential as a target of

alternative therapeutic modalities [7,16,21]. Several sequences located on two separate exons, and thus onlythe EGFR cDNA fragment from the construct but notstudies have exploited EGFR as a target for therapeutic

procedures using anti-EGFR antibodies [8], toxin- the endogenous gene is amplified by PCR. Positive cloneswere subcloned by serial dilution.conjugated ligand [2], and tyrosine kinase-specific inhibi-

tors [29]. However, these approaches have their limi- To distinguish between cells transfected with emptyvector and cells that developed spontaneous resistancetations in clinical applications [8,22]. The aim of this

study was to evaluate an alternative approach in tar- to puromycin, DNA dot blotting was performed on emptyvector-transfected cells using radioactively labelledgeting EGFR by antisense transfection. We transfected

U-87MG glioblastoma cells with an antisense EGFR empty vector as a probe. Only those clones producingpositive hybridization signals to the probe were selectedconstruct and assessed the proliferative, diCerentiative,

and transforming properties of such transfectants. for use as controls in subsequent experiments.

Materials and methods Western blot analysis

Aliquots of whole-cell lysate (50 mg) were resolvedConstruction of antisense-EGFR vector

on 4–15% gradient sodium dodecyl sulphate-polyacrylamide gels, transferred electrophoretically toA 1024 bp EGFR cDNA fragment (BamHI-BamHI)

derived from plasmid PE7 was inserted in reverse orien- nitrocellulose membranes, and incubated with rabbitanti-human EGFR polyclonal antibody (15200, Santatation at the BamHI site and under the control of the

Moloney murine leukaemia virus LTR promoter of the Cruz Biotechnology, Santa Cruz, CA, USA) or rabbitanti-human GFAP polyclonal antibody (15200, DAKO,5.05 kb pBabe-puro retroviral vector (kindly provided by

Dr N. R. Lemoine, Imperial Cancer Research Foundation, Carpinteria, CA, USA) prior to incubation with horserad-ish peroxidase-conjugated secondary antibody (Bio-Rad,UK) [13]. This cDNA corresponds to the last 256 amino

acid residues of the extracellular domain, the entire Hercules, CA, USA) at 151000 dilution. Antigen-antibody complexes were detected using an ECL chemi-transmembrane domain and the first 61 amino acid

residues of the cytoplasmic domain of EGFR [25]. luminescence reagent (Amersham, UK) according to themanufacturer’s procedure. Quantification of bands onAnother construct with the EGFR cDNA fragment cloned

in the forward orientation was used as the sense control. autoradiograms was carried out using laser scanningdensitometry.

Cell culture and transfection

ImmunohistochemistryThe human glioblastoma cell line U-87MG (AmericanType Culture Collection, Rockville, MD, USA) was grown Cells were fixed in cold methanol and incubated with

mouse anti-human Ki-67 monoclonal antibody (usedin Minimum Essential Medium alpha medium (Gibco,

© 1998 Blackwell Science Ltd, Neuropathology and Applied Neurobiology, 24, 389–396

EGFR RNA transfection in gliomas 391

neat; Zymed, South San Francisco, CA, USA), rabbit as a 10.5 kb transcript, were observed in parentalU-87MG cells and in cells transfected with empty vectoranti-human EGFR polyclonal antibody (15200) or rabbit

anti-human GFAP polyclonal antibody (15200). Bound (U-87MG/pBabe) or sense EGFR construct (U-87MG/S-pBabe). In contrast, this band was absent from fourantibody was visualized by the avidin biotin-peroxidase

complex method or peroxidase-antiperoxidase method. antisense clones (the most inhibited ones) and very weakfrom six other antisense clones (intermediately inhibitedFor Ki-67 staining, the area with the highest number of

labelled cells was identified. In this area, the percentage clones) (data not shown). Two of the most inhibited(denoted as AS-1 and AS-2) and one of the intermediatelyof Ki-67 positive nuclei (expressed as labelling index, LI)

was determined by counting about 1000 nuclei in inhibited clones (denoted as AS-3) were then subclonedby serial dilution, and used in subsequent experimentscontiguous microscopic fields at ×200 magnification

with the assistance of an eyepiece graticule. in this study. PCR analysis was repeated to confirm thepresence of the EGFR cDNA in these clones as shownin Figure 1.

Proliferation assay and anchorage-independentThe eCect of antisense EGFR RNA on the expression

growth assayof EGFR protein was evaluated by western blot analysisusing a rabbit anti-human EGFR polyclonal antibody.Cells (1×104 /well) were seeded on 24-well plates, and

proliferation was followed by cell counting using a Parental cells as well as cells transfected with emptyvector or sense EGFR construct expressed the expectedhaemocytometer after 2, 4, 6 and 7 days of culture. All

measurements were done in triplicate. 170 kDa protein characteristic of human EGFR(Figure 2). The protein level was reduced about fivefoldFor the anchorage-independent growth assay, cells

(5×103/well) were seeded in soft agarose on 6-well in clone AS-3 and was undetectable in AS-1 and AS-2.A similar expression pattern was confirmed usingplates by use of a double layer method [10]. Colonies

were scored on the 14th day. immunocytochemical staining of the cultured cells. Thusa direct correlation was observed between the level ofEGFR transcript and the abundance of corresponding

Flow cytometryprotein expression in cells transfected with the antisenseEGFR construct.Monoparametric conventional cell-cycle analysis was

performed using a Coulter Epics XL flow cytometer Southern blot analysis revealed that AS-1 and AS-2have more copies of the integrated antisense EGFR(Coulter Corporation, FL, USA). Two days after seeding

(1×106), cells were harvested, fixed in ethanol and constructs than AS-3 (data not shown), which mayexplain the undetectable levels of the 10.5 kb EGFRstained with 50 mg/mL propidium iodide (Sigma) con-

taining 0.1% Triton X-100 and 1% RNase A. The transcripts and consequent protein expression in themost inhibited clones.percentages of cells in diCerent phases of the cell cycle

were analysed using the Multicycle DNA Cell Cycleprogram (Phoenix, San Diego, CA, USA).

Results

Human U-87MG glioma cells expressing antisense

EGFR RNA

A total of 15 clones transfected with the antisense EGFR

1353 bp

1078 bp

872 bp

603 bp

M 1 2 3 4 5 6

construct were isolated. PCR analysis demonstrated that Figure 1. Demonstration of the presence of EGFR cDNA fragmentin transfected cells by PCR. Lane 1, antisense EGFR/pBabe10 clones contained the EGFR cDNA fragment, whichconstruct; lane 2, parental U-87MG cells; lane 3, empty vector-was absent from parental cells and cells transfected withtransfected cells (U-87MG/pBabe); lanes 4–6, antisense EGFR-the empty vector. The eCect of antisense EGFR RNA ontransfected clones AS-1 to AS-3, respectively. Lane M, wx174

expression of EGFR transcripts was assessed by northern DNA/Hae III markers. The expected size of the PCR product was1015 bp.blot analysis. High levels of EGFR mRNA, predominantly



a

b

Figure 2. Western blot analyses of protein expression of a, EGFRand b, GFAP. Lane 1, parental U-87MG cells; lane 2, empty vector-transfected cells (U-87MG/pBabe); lanes 3–5, antisense EGFR-transfected clones AS-1 to AS-3, respectively. a, Human EGFRprotein (170 kDa) was highly expressed in parental and controlcells, undetectable in AS-1, AS-2 and reduced by about fivefold inAS-3 compared with control cells. b, The expression of GFAP(52 kDa) was three times higher in antisense EGFR-transfected cellsthan in control cells.

EVect of antisense EGFR RNA on cellular

proliferation

Three proliferation parameters, growth rate, Ki-67 stain-ing, and flow cytometry, were used to assess the prolifer-ative ability of U-87MG cells transfected with theantisense EGFR construct. There was no significant

Figure 3. Cellular proliferation. a, Growth curves of parentalU-87MG cells, empty vector-transfected cells (U-87MG/pBabe),sense (U-87MG/S-pBabe) and antisense EGFR-transfected cells,AS-1 to AS-3. Cells were counted at days 0, 2, 4, 6, and 7 ofculture. Cell count values are presented as the mean±sd (verticalbars) of three independent experiments; each experiment wasperformed in triplicate. b, Ki-67 labelling index (LI) of U-87MGcells, empty vector-and antisense EGFR-transfected cells. Cells werestained with mouse anti-human Ki-67 monoclonal antibody asdescribed in ‘Materials and methods’, and the percentage of positivenuclei was calculated. Experiments were repeated four times, dataare presented as means±sd (vertical bars). Significance **,P<0.01; *, P<0.05. The Ki-67 LI was highest in the parental andcontrol cells indicating that these cells were at a higherproliferation status than cells containing the antisense EGFRconstructs. c, Flow cytometric analysis of U-87MG, empty vector-and antisense EGFR-transfected cells. Experiments were repeatedthree times in duplicate. Data are presented as means of the threeexperiments. The results demonstrated a shift of the cell populationfrom the proliferative S phase to the G

0/G

1phases in antisense

EGFR-transfected cells as compared with control cells. S, Open bar.G

2, Stippled bar. G

0/G

1, Solid bar.

7

60

0Days in culture

Nu

mb

er o

f ce

lls (

×104 )

10

2

50

40

30

20

0

3 4 5 61

U-87MGU-87MG/pBabeU-87MG/S-pBabeAS-1AS-2AS-3

a

100

0

U-87M

G

Ki-

67 la

bel

ling

ind

ex (

%)

40

80

20

60

b

U-87M

G/pBab

eAS-1

AS-2AS-3

****

*

100

0

U-87M

G

Cel

l cyc

le d

istr

ibu

tio

n (

%)

40

80

20

60

c

U-87M

G/pBab

eAS-1

AS-2AS-3



diCerence in the growth rate between U-87MG parentalcells, empty vector-and sense EGFR-transfected cells asdetermined by cell counting. However, the proliferativeability of the three antisense-EGFR clones was impairedand all three antisense clones grew at a slower rate thanthe control cells (Figure 3). AS-3, which expressed mod-erate levels of EGFR protein, grew slightly faster thanAS-1 and AS-2, albeit slower than the control cells. Onday 7, growth of antisense transfectants was inhibitedby 80%, 77%, and 56% in AS-1, AS-2 and AS-3 clones,respectively, relative to the empty vector-transfectedcells.

The Ki-67 monoclonal antibody reacts with a nuclearantigen expressed by proliferating human cells duringthe late G

1, S, M, and G

2phases of the cell cycle. Cells

transfected with the empty vector had a similar Ki-67LI as the parental cells. However, a significantly lowerKi-67 LI was observed in antisense EGFR-transfected

1200

0

U-87M

G

Nu

mb

er o

f co

lon

ies

1000

800

600

400

200

U-87M

G/pBab

e

U-87M

G/S-p

Babe

AS-1AS-2

AS-3

clones AS-1, AS-2 and AS-3 (by 32% [P<0.01], 39%Figure 4. Anchorage-independent growth of parental U-87MG

[P<0.01], and 19% [P<0.05], respectively) than in cells, empty vector-(U-87MG/pBabe), sense (U-87MG/S-pBabe) andempty vector-transfected cells (Figure 3). antisense EGFR-transfectants, AS-1 to AS-3. Five thousand cells

were plated in soft agarose with medium supplemented with 10%Flow cytometric analysis showed that empty vector-fetal bovine serum and grown for 14 days. Colonies were thentransfected cells had a similar cell cycle distribution tocounted and divided into small (<200 mm) or large (>200 mm)

the parental cells (Figure 3). The percentage of cells in colonies. Results are the average of three independent experimentsclones AS-1, AS-2 and AS-3 that were in G

0/G

1phases performed in duplicate. Solid bars represent large colonies; open

bars represent small colonies. Antisense EGFR-transfected cellswas increased by 31% (P<0.01), 30% (P<0.01), andformed fewer and smaller colonies than the control cells.23% (P<0.01), respectively, with a corresponding

decrease of 53% (P<0.01), 58% (P<0.01) and 45%(P<0.01) in the percentage of cells in the S phase, when significantly reduced ability to grow in soft agarose;

there were fewer and smaller colonies than those of thecompared with empty vector-transfected control cells.No significant diCerence in cell cycle distribution could control cells (Figure 4). Clone AS-3, expressing inter-

mediate levels of EGFR, formed more and larger coloniesbe detected among the three antisense EGFR-transfected clones. than did AS-1 and AS-2, which expressed the lowest

levels of EGFR. These results suggest that the transformedIn summary, the three proliferation parameters,growth rate, Ki-67 staining, and flow cytometry, had phenotype is dependent on the levels of EGFR in

U-87MG cells.consistently demonstrated a direct correlation betweenEGFR expression level and cellular proliferation. All threeantisense clones containing reduced or undetectable level

Morphological changes and GFAP expressionof EGFR showed impaired proliferative ability.

Under the phase-contrast microscope, the empty vector-and the sense EGFR-transfected cells had the same

Transformation assaymorphology as the parental cells. The two most inhibitedantisense clones, AS-1 and AS-2, had smaller cell bodiesWe analysed the transforming property of antisense

EGFR-transfected cells by their ability to form colonies and longer processes than control cells, and most of thecells became bipolar (Figure 5). The AS-3 cells hadin soft agarose. The number and size of the colonies

were comparable in parental, empty vector and sense smaller cell bodies without appreciable extension of cellprocesses. Western blot analysis showed that parentalEGFR-transfected cells. However, antisense EGFR-

transfected cells expressing lower levels of EGFR had and control cells had very low levels of GFAP expression.



Figure 5. Photomicrographs of U-87MG cells.a, Empty vector-transfected cells. b, AntisenseEGFR-transfected AS-1 cells, which had smallercell bodies and much longer processes thancontrol cells, and most of the cells became

a

b bipolar. ×40.

In contrast, the level of GFAP expression of the antisense EGFR cDNA construct and assessing the consequentphenotypes of such transfectants. Suppression ofEGFR-transfected clones was about three times higher

than in control cells (Figure 2). Moreover, by immunohi- EGFR expression in the antisense EGFR clones resultedin impaired cellular proliferation, induced diCerentia-stochemical staining of the antisense EGFR-transfected

cells, an increased intensity of GFAP staining was tion, and reduced the transforming potential of thesecells.observed in these cells, which is consistent with the

western blot results. In the antisense EGFR construct, we have engineereda strong viral promoter to drive transcription of theantisense EGFR RNA. The level of antisense EGFR RNA

Discussionwas also dependent on the number of antisense plasmidsintegrated into the genome. Although the mechanism ofIn this study, we evaluated an antisense EGFR approach

as an alternative therapeutic modality in treatment of action is not clear, the antisense approach used in thisstudy demonstrated high specificity toward inhibition ofglioblastomas by transfecting U-87MG with an antisense



endogenous EGFR expression. Moreover, there is a direct One major goal of this investigation was to assess theeCect of antisense EGFR RNA on tumorigenicity of thecorrelation between the level of residual EGFR transcripts

and the corresponding protein expression. antisense clones. The reduced ability of the antisenseclones to form colonies in soft agarose indicated thatThe eCects of down-regulation of the EGFR level were

demonstrated by a decrease in cell proliferation. A higher these cells had lower transforming potential than theparental cells. Such an observation was also reported inproportion of antisense EGFR-transfected cells was found

to reside in the G0

/G1

phase rather than in the proliferat- previous studies using a similar approach in targetingEGFR [1,15]. It is of interest to find out if these antisenseive S phase of the cell cycle, reflecting the slow growth

rate of these cells. Antiproliferative eCects by antisense EGFR-treated cells would grow and form tumours inathymic nude mice.EGFR have also been reported in a variety of human

tumours including rhabdomysarcoma cells [3], colon In conclusion, it appears that although many othergenes are involved in glial tumorigenesis, the EGFR genecarcinoma cells [1], pancreatic carcinoma cells [13], and

epidermoid carcinoma KB cells [15,27]. Our results plays a major role in regulating cell proliferation anddiCerentiation. Abrogating the expression of EGFR leadssupport the notion that EGFR plays a pivotal role in

proliferation of malignant gliomas. to significant inhibition of cell growth and transformingability, and also to induction of diCerentiation. TheseIn addition to a lower level of cell proliferation, the

antisense EGFR-treated cells also displayed smaller cell results suggest that antiproliferation and reversion ofmalignant phenotypes by an antisense EGFR approachbodies possessing fine tapering cell processes and more

GFAP expression than the parental cells. GFAP is a glial- can be a viable mechanism for the treatment of glioblas-tomas. Our model of antisense EGFR treatment of gliobla-specific intermediate filament protein and has been impli-

cated in maintenance of normal morphological shape of stoma cells also provides a basis for designing futureshorter antisense oligodeoxynucleotides for experimentalastrocytes [6]. It has been suggested that the intensity

of GFAP staining of glioblastomas is inversely related to gene therapy of gliomas.their degree of anaplasia [5,20]. Thus, strong GFAPexpression and morphological changes such as extension

Acknowledgementsof glial processes are indicative of cell diCerentiation in

The technical assistance from Enders Ng, K. W. Lo andastrocytic cells. Such observations are also supported byHardy Ko was much appreciated. We also thank Dr N. R.other studies using diCerentiation-inducing agents suchLemoine for providing the pBabe-puro retroviral vector.as retinoic acid [19], sodium butyrate [7], and dimethyl-

formamide [11]. The mechanism of induced diCeren-tiation by all these agents including antisense EGFR ReferencesRNA is unclear. However, U et al. [24] has demonstrated

1 Chakrabarty S, Rajagopal S, Huang S. Expression of anti-that the responsiveness of glioblastoma cells to diCeren-sense epidermal growth factor receptor RNA downregulatestiation-inducing agents correlated inversely with thethe malignant behavior of human colon cancer cells. Clin

endogenous EGFR level. Cells with a high level of EGFR Exp Matastasis 1995; 13: 191–5expression responded to fewer diCerentiation-inducing 2 Chaudhary VK, FitzGerald DJ, Adhya S, Pastan I. Activity

of a recombinant fusion protein between transformingagents than those with less EGFR protein. Consistentgrowth factor type alpha and Pseudomonas toxin. Proc.with their results, our data demonstrated that lowNatl Acad Sci USA 1987; 84: 4538–42expression of EGFR triggers cell diCerentiation in glioblas-

3 De Giovanni C, Landuzzi L, Frabetti F et al. Antisensetoma cells, and provided direct evidence that EGFR plays

epidermal growth factor receptor transfection impairs thea role in glial diCerentiation. Since proliferation and proliferative ability of human rhabdomyosarcoma cells.diCerentiation are two opposing events in tumour devel- Cancer Res 1996; 56: 3898–901

4 Downward J, Yarden Y, Mayes E et al. Close similarity ofopment, overexpression of EGFR may be a mechanismepidermal growth factor receptor and v-erb-B oncogeneby which glioblastoma cells suppress diCerentiation whileprotein sequences. Nature (England) 1984; 307: 521–7obtaining a growth advantage. Thus the function of

5 DuCy PE, Huang YY, Rapport MM. Glial fibrillary acidicEGFR in glioblastomas is multifaceted; in addition to protein in giant cell tumors of brain and other gliomas: amediating cellular proliferation through EGF, EGFR is possible relationship to malignancy, diCerentiation, and

pleomorphism of glia. Acta Neuropathol 1980; 52: 51–7also involved in diCerentiation.



6 DuCy PE, Huang YY, Rapport MM. The relationship of morphology and expression of GFAP of an anaplastic astro-cytoma cell line. Int J Cancer 1988; 42: 419–27GFAP to the shape motility, and diCerentiation of human

astrocytoma cells. Exp Cell Res 1982; 139: 145–57 20 SchiCer D, Giordana MT, Germano I, Mauro A. Anaplasiaand heterogeneity of GFAP expression in gliomas. Tumori7 Engelhard HH, Duncan HA, Dal Canto M. Molecular charac-

terization of glioblastoma cell diCerentiation. Neurosurgery 1986; 72: 163–7021 Schober R, Bilzer T, Waha A et al. The epidermal growth1997; 41: 886–97

8 Faillot T, Magdelenat H, Mady E et al. A phase I study of factor receptor in glioblastoma: genomic amplification, pro-tein expression, and patient survival data in a therapeutican anti-epidermal growth factor receptor monoclonal anti-

body for the treatment of malignant gliomas. Neurosurgery trial. Clin Neuropathol 1995; 14: 169–7422 Stragliotto G, Vega F, Stasieck P, Gropp P, Poisson M,1996; 39: 478–83

9 Haley JD, Hsuan JJ, Waterfield MD. Analysis of mammalian Delattre JY. Multiple infusions of anti-epidermal growthfactor receptor (EGFR) monoclonal antibody (EMD 55,900)fibroblast transformation by normal and mutated human

EGF receptors. Oncogene 1989; 4: 273–83 in patients with recurrent malignant gliomas. Euro J Cancer1996;32A: 636–4010 Hamburger AW, Salmon SE, Kim MB et al. Direct cloning

of human ovarian carcinoma cells in agar. Cancer Res 23 Strommer K, Hamou MF, Diggelmann H, de Tribolet N.Cellular and tumoural heterogeneity of EGFR gene amplifi-1978; 38: 3438–44

11 Li XN, Du ZW, Huang Q, Wu JQ. Growth-inhibitory and cation in human malignant gliomas. Acta Neurochir 1990;107: 82–7diCerentiation-inducing activity of dimethylformamide in

cultured human malignant glioma cells. Neurosurgery 24 U HS Espiritu OD, Kelley PY, Klauber MR, Hatton JD. Therole of the epidermal growth factor receptor in human1997; 40: 1250–9

12 Libermann TA, Nusbaum HR, Razon N et al. Amplification, gliomas: II. The control of glial process extension and theexpression of glial fibrillary acidic protein. J Neurosurg 1995;enhanced expression and possible rearrangement of EGF

receptor gene in primary human brain tumors of glial 82: 847–5625 Ullrich A, Coussens L, Hayflick JS et al. Human epidermalorigin. Nature 1985; 313: 144–7

13 Liu T, Chen J, Zeng C. ECects of antisense epidermal growth growth factor receptor cDNA sequence and aberrantexpression of the amplified gene in A431 epidermoid carci-factor and its receptor retroviral expression vectors on cell

growth of human pancreatic carcinoma cell line. Chin Med noma cells. Nature 1984; 309: 418–2526 Veale D, Ashcroft T, Marsh C, Gibson GJ, Harries AL.J (Engl) 1995; 108: 653–9

14 Louis DN. A molecular genetic model of astrocytoma. Brain Epidermal growth factor receptors in non-small cell lungcancer. Br J Cancer 1987; 55: 513–6Pathol 1997; 7: 755–64

15 Moroni MC, Willingham MC, Beguinot L. EGF-R antisense 27 Wang S, Lee RJ, Cauchon G, Gorenstein DG, Low PS.Delivery of antisense oligodeoxyribonucleotides against theRNA blocks expression of the epidermal growth factor

receptor and suppresses the transforming phenotype of a human epidermal growth factor receptor into cultured KBcells with liposomes conjugated to folate via polyethylenehuman carcinoma cell line. J Biol Chem 1992; 267:

2714–22 glycol. Proc Natl Acad Sci USA 1995; 92: 3318–2228 Wong AJ, Bigner SH, Kinzler KW, Hamilton SR, Vogelstein16 Nagane M, Coufal F, Lin H, Bogler O, Cavenee WK, Huang

HJS. A common mutant epidermal growth factor receptor B. Increased expression of the EGF receptor gene in malig-nant gliomas is invariably associated with gene amplifi-confers enhanced tumorigenecity on human glioblastoma

cells by increasing proliferation and reducing apoptosis. cation. Proc Natl Acad Sci USA 1987; 84: 6899–90329 Yaish P, Gazit A, Gilon C, Levitzki A. Blocking of EGF-Cancer Res 1996; 56: 5079–86

17 Nishikawa R, Ji XD, Harmon RC et al. A mutant epidermal dependent cell proliferation by EGF receptor kinase inhibi-tors. Science 1988; 242: 933–5growth factor receptor common in human glioma confers

enhanced tumorigenecity. Proc Natl Acad Sci USA 1994; 30 Yarden Y, Ullrich A. Growth factor receptor tyrosinekinases. Annu Rev Biochem 1988; 57: 443–7891: 7727–37

18 Rajkumar T, Gullick WJ. The type I growth factor receptorsin human breast cancer. Breast Cancer Res Treat 1994; Received 28 November 199729: 3–9

Accepted after revision 7 April 199819 Rutka JT, DeArmond SJ, Giblin JR, McCulloch JR, Wilson

CB, Rosenblum ML. ECect of retinoids on the proliferation,


Documents

Antisense epidermal growth factor receptor RNA transfection in human malignant glioma cells leads to inhibition of proliferation and induction of differentiation