Correlation of TGF-alpha and EGF-Receptor Expression with Proliferative Activity in Human Astrocytic Gliomas

PATHOLOGY

RESEARCH AND PRACTICE

© Gustav Fischer Verlag

Correlation of TGF-alpha and EGF-Receptor Expression with Proliferative Activity in Human Astrocytic Gliomas Peter von Bossanyi 1, Jorg Sallaba 2 , Knut Dietzmann 1, Michaela Warich-Kirches 1 and Elmar Kirches 1 'Department of Neuropathology, University of Magdeburg, Germany, 2Department of Internal Medicine, District Hospital, Gardelegen, Germany

Summary Fifty-nine paraffin-embedded astrocytic gliomas (four WHO grade 1,21 WHO grade 2,17 WHO grade 3 and 17 glioblastomas, WHO grade 4) were immunohistochemically investigated for expression of transforming growth factor-alpha (TGF-alpha), epidermal growth factor receptor (EGF-R) and oncoprotein c-erbB-2 by semiquantitative assessment. Proliferative activity was simultaneously analyzed by using the antibody Ki-67 (MIB-l). Immunostaining in neoplastic cells was quantified by image analysis. Concerning the antibodies used, the percentage of immunoreactive cells increased with histologic malignancy. There was no expression of EGF-R and c-erbB-2 in the majority of low-grade astrocytomas. However, small focal expressions of TGF-alpha and EGF-R were observed in several lowgrade astrocytomas (11125), suggesting an early stimulation of malignant transformation. With regard to percentage, a strong positive correlation between TGFalpha and EGF-R-stained cells was found, indicating an autocrine stimulation of the mitogenic pathway of the TGF-alphalEGF-R system. Likewise, indices of EGF-R and c-erbB-2 positive cells correlated significantly. Less significant correlations were also seen between EGF-R, c-erbB-2 frequencies and the Ki-67 labeling index. However, there was no correlation between TGF-alpha and Ki-67 indices. The results suggest that TGF-alpha expression is not directly related to the proliferative potential as judged by the Ki-67 labeling index. Furthermore, besides EGF-R and c-erbB-2, other growth factors and their receptors or mutant EGF-R might participate in the proliferative activity of gliomas. Pathol. Res. Pract. 194: 141-148 (1998)

Key words: Glioma - Transforming growth factor alpha - Epidermal growth factor receptor - c-erbB-2 oncoprotein - Ki-67

Introduction Polypeptide growth factors play an important role in the stimulation of cellular proliferation, and appear to be involved in pathologic processes including cancer. Besides the epidermal growth factor (EGF), the transforming growth factor-alpha (TGF-alpha), a 50 amino acid protein of 55 kD, is the best characterized. TGF-alpha has 42% amino acid homology with EGF [41J. TGFalpha is secreted not only in a variety of tumors [45J, but also in embryos, fetuses [20J and many normal adult cells, such as keratinocytes, gut and breast epithelial cells, neurons and glial cells [7, 33]. In human gliomas, a positive correlation was seen between tumor grade and TGF-alpha immunoreactivity [29,31]. Almost all experimental data indicate that TGF-alpha binds with high affinity to the epidermal growth factor receptor (EGF-R), a transmembrane 170 kD glycoprotein [12, 22]. The extracellular part of EGF-R contains the ligand-binding domain [18]. The binding of ligands induces dimerization of EGF-R with activation of the receptor's intrinsic tyrosine kinase. The resultant autophosphorylation of tyrosine residues mediates growth Address for correspondence: P. von Bossanyi, Department of Neuropathology, University of Magdeburg, Leipziger StraBe 44, D - 39120 Magdeburg, Germany. Fax: (0391) 671-

3300

0344-0338/98/0194-0141 $5.00/0

142 . P. von Bossanyi et al.

stimulation or neoplastic progression via signalling pathways [13]. Cells of human glioblastoma cell lines as well as stromal cells of capillary hemangioblastomas showed immunoreactivity with antibodies to TGFalpha and EGF-R, suggesting that a so-called autocrine growth stimulation may maintain an autonomous neoplastic growth [11, 27]. Expression of TGF-alpha, EGF-R and Ki-67 increased in malignant gliomas, which, however, could also be considered as a paracrine process [21]. An increased expression of EGF-R has been found in malignant human gliomas (anaplastic and glioblastomas), suggesting a relation to unfavorable prognosis [1, 9, 26, 36, 37 J, and expression of EGF-R and Ki-67 index is associated with malignant progression [15]. EGF-R is a member of a receptor family that also includes the c-erbB-2 (neu) oncoprotein. Both proteins have a homology of more than 80% in the intrinsic tyrosine kinase component [2J, but the ligand-binding domains differ considerably in structure [30]. The ligand of c-erbB-2, however, has not yet been identified, since heregulin/neu differentiation factor (NDF) does not directly bind to c-erbB-2 [19, 25J. Several studies suggest that overexpression of c-erbB-2 protein is associated with poor prognosis in some tumors such as breast cancer. In human gliomas, the immunoreactivity increased with the progression and dedifferentiation of tumors [3, 9, 32]. Using immunohistochemistry in 59 surgically sampled astrocytic brain tumors, we examined the expression ofTGF-alpha, EGF-R and c-erbB-2 by semiquantitative assessment. Moreover, we analyzed possible correlations between TGF-alpha and EGF-R, EGF-R and c-erbB-2 immunoreactivity, as well as between the expression of these proteins and proliferative potential as judged by antibody Ki-67.

Material and Methods Tissue samples from 59 astrocytic tumors with different grading types were analyzed. This material was composed of 4 pilocytic astrocytomas (WHO grade 1), 21 fairly-differentiated astrocytomas (WHO grade 2), 17 anaplastic astrocytomas (WHO grade 3) and 17 glioblastomas (WHO grade 4). Histologic diagnoses were based on the WHO brain tumor classification [16]. All patients had undergone neurosurgery at the Department of Neurosurgery, Otto-von-Guericke-University, Magdeburg. Five Ilm thick sections from formaldehyde-fixed and paraffin-embedded samples were stained according to Feulgen's method for determination of nuclear DNA content by cytophotometry. Measurements were performed on Feulgenstained nuclei (final magnification 400x) by using an OLYMPUS BH2- microscope which transmitted image data under a 585 nm bandpass filter to an AT-compatible computer equipped with a digital board (Leutron Vision AFP/AT). Nuclei were interactively selected, and the parameter SEXT (the factor expressing DNA content) was automatically calculated

by Image C software (IMTRONIC, Germany). At least 250 non-overlapping tumor cell nuclei as well as 15-20 nuclei of peripheral leucocytes, used as standard diploid cells, were measured on each slide. Using densitometric SEXT parameter of the tumor and standard cells (2c ploidy level), the percentage of proliferating cells (S + Gz/M phase) was calculated by means of ploidy level 2.5c-4c for each specimen. Immunohistochemistry was performed on parallel sections with the avidin biotin peroxidase kits for mouse or rabbit biotinylated immunoglobulins, respectively (ExtrAvidin kits, Sigma, USA). A monoclonal non-crossreactive marker for TGF-alpha (clone 213-4.4) and a rabbit polyclonal antibody against EGF receptor (amino acid residues 1005-1016, internal domain) were purchased from Oncogene Science Inc., USA. Rabbit antibody for c-erbB-2 oncoprotein was obtained from DAKO, Denmark. Prediluted antibody Ki-67 (clone MIB 1, Dianova, Germany) was used after microwave pretreatment. Primary antibodies were diluted in phosphate buffered saline (PBS) as follows: anti-TGF-alpha 1:30, antiEGF-R 1:20 and anti-c-erbB-2 1:50. In brief, paraffin sections (4Ilm) were dewaxed and treated with 0.3% hydrogen peroxide in methanol followed by incubation in 10% normal goat serum. The sections were then incubated with primary antibodies. After rinsing in PBS, sections were incubated with biotinylated anti-Ig and avidin peroxidase. Immunoperoxidase reaction product was visualized with diaminobenzidine/hydrogen peroxide. The sections were counterstained with hematoxylin, dehydrated and coverslipped with Canada balsam. As provided by the manufacturers, samples of normal skin and human breast carcinoma were used as positive staining controls for antibodies to TGF-alpha and EGF-R, and samples of breast carcinoma for antibody to c-erbB-2. Sections of normal liver and kidney were also stained with all antibodies used in order to further determine the extent of unspecific labeling or cross reactivity. In the kidney, TGF-alpha immunostaining appeared in distal convoluted tubules and collecting ducts, EGF-R in distal and c-erbB-2 in proximal convoluted tubules. Immunoreactive TGF-alpha was localized in the bile duct epithelium in normal liver. C-erbB-2 expression was found in some small polygonal liver cells. However, EGF-R protein was not detected in normal liver. Negative controls included isotype-matched mouse IgG2a (DAKO) at the same concentration as TGF-alpha in place of antibody to TGFalpha. An irrelevant antibody (rabbit polyclonal antibody to ACTH) (DAKO) was used as a control for polyclonal antibodies to EGF-R and c-erbB-2, at a concentration equivalent to that applied for the antibody with the highest protein concentration (c-erbB-2). Furthermore, the antibody to EGF-R was pre-incubated with ten fold excess (by weight) of its corresponding peptide overnight before incubation with tissues. Using interactive microscopic image analysis VIDS V (Metagraphics Software Corporation), quantification of immunoreactivity was determined as percentages of positive tumor cells. In each section in most evident immunostained areas, immunoreactive and negative cells were counted in ten screens (microscopic magnification 40x). Using SPSS software, the relation between different variables was estimated by applying the Kruskal-Wallis test in connection with the Nemenyi test and Mann-Whitney's test. Furthermore, the correlation was analyzed by calculating Pearson's correlation coefficient. p < 0.05 was considered statistically significant.

Correlation ofTGF-alpha and EGF-Receptor Expression . 143

Results Protein expression of TGF-alpha, EGF-R and cerbB-2 was observed imunohistochemically at different frequencies mainly in the cytoplasm of small anaplastic and gemistocytic tumor cells (Fig. 1). The staining of EGF-R antibody is mostly located to the cytoplasm of neoplastic cells, since the antibody used reacts in the internal carboxy terminal region. Even in low-grade astrocytomas, expression of EGF-R and its ligand TGFalpha was only focal in some cases, whereas expression of c-erbB-2 was less frequent. The higher the malignan-

cy, the more homogeneous the distribution pattern of immunoreactive cells in tumors. TGF-alpha, EGF-R and c-erbB-2 immunoreactions of reactive astrocytes and neurons could be observed in perifocal marginal areas. These cells were not included in semiquantitative evaluations. Endothelial cells showed positive staining for TGF-alpha in 56%, and for EGF-R in 37% of the tumors investigated. Altogether, semiquantitative evaluation yielded a high index variation of TGF-alpha, EGF-R and c-erbB2 immunoreactive tumor cells in the tumors investigated, with the percentages ranging between low and high values (Fig. 2). The percentage of positive cells increased with increasing anaplasia grade in all antibody reactions (Table 1). In a similar manner, the percentage of proliferating cells (S + G/M phases) was increased. There were significant differences between low-grade astrocytomas versus grade 3 astrocytomas and glioblastomas. A strong significant correlation (r = 0.87; p < 0.001) was evident with regard to the percentage between TGF-alpha and EGF-R immunoreactive cells (Fig. 3). There was also a close correlation between the indices of EGF-R and c-erbB-2 positive cells (r = 0.68; p < 0.001). There were less significant correlations be-

100

Percent + "li 0

80 60

+ o

c

o +

2

+m +

o

+

oil

3

o +

+

+ +

+

4

Tumor grade TGF-alpha

+

100

o

+

++

000 00

EGF-R

o

Percent

80

o

+ 0

0 0

00

++ +

2

3 Tumor grade o

Fig. 1. Many tumor cells of different size in a grade 3 astrocytoma were stained with antibodies to A TGF-alpha and B EGF-R. ABC immunoperoxidase technique with hematoxylin counterstain, x330.

0

c-erbB-2

4

+

EGF-R

Fig. 2. Indices (%) of neoplastic cells with TGF-alpha and EGF-R (top) as well as EGF-R and c-erbB-2 (bottom) immunoreactivity in astrocytomas and glioblastomas.

144 . P. von Bossanyi et al.

Table 1. Indices (%) of neoplastic cells with TGF-alpha, EGF-R, c-erbB-2 and Ki-67 (MIB 1) expression, as well as percentage of the proliferating Cells (S + Gz/M phases) according to the histologic grade. Means and standard deviations Grade

TGF-alpha

EGF-R

c-erbB-2

Ki-67

S+Gz/M

Gl G2 G3 G4

1.5 ± 3.0 12.6 ± 13.0 42.4 ± 25.5 48.8 ±25.5

0 14.7 ± 17.8 47.9 ± 32.8 59.5 ± 30.7

0 2.5 ± 8.8 17.7 ± 18.9 29.6 ± 26.5

4.0 ± 3.8 5.5 ± 2.3 11.8 ± 6.3 12.1 ±4.9

8.5 ± 1.4 18.6 ± 11.2 28.0± 7.9 39.4 ± 5.2

100

EGF-R (%) 0

0

B

0

80

0

60

0

0 0

0

o

0

While EGF-R immunoreaction showed the highest frequencies of positive cells in all grades of anaplasia, the number of tumors with a positive reaction was lower than TGF-alpha positive gliomas, the percentages of reactive cells being slightly lower (Table 2). The minority of tumors expressed c-erbB-2, also showing the lowest index of immunoreactive cells. Like the percentages of immunoreactive cells, the number of positive tumors increased in all 3 immunoreactions with increasing grade of malignancy. In many tumors that were mainly of the low-grade type, TGF-alpha and EGF-R immunoreactions were not ascertained. Eight tumors showing no reaction with antibody TGF-alpha, were also EGF-R and c-erbB-2 negative. Of the 17 tumors without EGF-R immunoreaction, all were also c-erbB-2 negative and 16 exhibited no or only a weak reaction (up to 30% positive cells) with TGF-alpha. Thus the conclusion can be drawn that some gliomas exist that do not express all three proteins investigated, or, if they do, only with low index. The minority of these tumors, however, are grade 3 astrocytomas or glioblastomas. Using antibodies to EGF-R or c-erbB-2 we observed tumors without any immunoreaction; these tumors had a significantly lower index of Ki-67 positive cells, thus indicating a low proliferative potential (Table 3).

0

Jl

0

0 00

0

40

0

0

0 0

0 0 0

20

40

60

TGF-alpha (%)

80

100

Fig. 3. Scatterplot with regression line representing the significant correlation between percentages of TGF-alpha and EGF-R stained cells in gliomas (r = 0.87; P < 0.001).

tween the indices of Ki-67 and EGF-R (r = 0.31; P < 0.05) and between Ki-67 and c-erbB-2 (r = 0.34; P < 0.01) positive cells, but not between Ki-67 and TGFalpha. However, a significant correlation (r = 0.48; p < 0.001) existed between the indices of TGF-alphastained cells and proliferating cells (S + G/M phases) evaluated by cytophotometry.

Table 2. Tumor rate with expression ofTGF-alpha, EGF-R and c-erbB-2 according to the histologic grade Histologic grade

TGF-alpha

Negative (n = 8) Reactive (n=51) EGF-R

Negative (n = 17) Reactive (n = 42) c-erbB-2

Negative (n = 35) Reactive (n = 24)

Gl

G2

G3

G4

3 (75%) 1 (25%)

4 (19%) 17 (81%)

1 (6%) 16 (94%)

0 17 (100%)

4 (100%) 0

10 (48%) 11 (52%)

2 (12%) 15(88%)

1 (6%) 16 (94%)

4 (100%) 0

19 (90%) 2 (10%)

8 (47%) 9 (53%)

4 (24%) 13 (76%)

Correlation of TGF-alpha and EGF-Receptor Expression . 145 Table 3. Labeling index (%) of Ki-67 expression with regard to EGF-R, c-erbB-2 and TGF-alpha negative or positive gliomas n

Ki-67 index (%)

EGF-REGF-R+

42

17

6.62 ± 5.27 10.08 ± 5.43

*

c-erbB-2 c-erbB-2 +

35 24

7.75 ± 5.06 11.l6 ± 5.74

*

TGF-alphaTGF-alpha+

8 51

6.49 ±6.85 9.52 ± 5.35

NS

*= p < 0.05; NS = not significant; Mann-Whitney test

Discussion TGF-alpha polypeptide is a natural ligand of EGF-R and may induce tyrosine phosphorylation of EGF-R by binding to the receptor molecule [28]. Some studies have shown that the expression of TGF-alpha in gliomas increased with grade of malignancy [29, 31], which is in accordance with our observations made in the present study. An increase in the expression of EGFR protein was also seen in malignant gliomas [1, 8, 9, 15,26,36,37]. In a small number of samples, Maruno et al. [21] demonstrated that both TGF-alpha and EGF-R were expressed more frequently in malignant gliomas. The significant positive correlation shown in the present study indicates that in most gliomas overexpression of TGF-alpha ligand is accompanied by an increased expression of its EGF-R receptor. These results imply the possibility that the mitogenic pathway of the TGF-alphalEGF-R system in glioma cells is regulated by an autocrine growth stimulation [10, 35, 40]. Cells of human glioblastoma lines showed a cytoplasmatic staining for TGF-alpha and a membrane reaction for EGF-R [11]. As we failed to prove a concomitant presence of TGF-alpha and EGF-R on the same tumor cells, a paracrine or juxtacrine proliferative stimulation must also be considered. However, there were 9 TGF-alpha immunoreactive tumors without EGF-R positive staining, 3 of which had positivity for more than 20% of the cells. Seven of these tumors were low-grade astrocytomas, in which TGF-alpha expression could be involved in the shifting to malignancy, without detectable EGF receptors already being expressed. Furthermore, it is possible that there is an expression of mutant EGF-R particularly in both malignant gliomas, which, though binding to TGFalpha, cannot be detected with the antibody used [14]. In contrast to other authors [1, 8, 26, 29], we observed a relatively high number of low-grade astrocytomas

(11/25) with only focal expression of TGF-alpha and EGF-R in most cases. Jaros et al. [15], however, demonstrated EGF-R immunoreactions in 33% of the grade 2 astrocytomas, and Schlegel et al. [31] found cells with TGF-alpha staining in 5/13 of low-grade gliomas. This focal expression of EGF-R and its ligand TGF-alpha may indicate an early onset of autocrine stimulation of tumor growth by certain cell populations within the malignant transformation. Through signal transduction cascades, the phosphorylated tyrosine residues of EGF-R cause, among others, the activation of protein kinase C, p21 Ras and MAP-kinase, and the induction of c-fos and c-myc oncogenes [4, 13]. Phosphorylation ofMAP-kinases induces cell proliferation or differentiation [23]. EGF-R kinase inhibitors blocked autophosphorylation of the receptor and the EGF-dependent proliferation of A431 cells afterwards [44]. Growth of glioma cells in soft agar was correlated with higher levels of EGF-R expression, especially in response to EGF [39]. The increased expression of TGF-alpha and EGF-R in malignant gliomas is linked to high proliferative potential, assessed by the Ki-67 labeling index [21]. Astrocytomas that expressed EGF-R and glioblastomas with EGF-R gene amplification showed a significantly higher Ki-67 labeling index, and patients with these tumors had a reduced survival [15,38]. On the other hand, Reifenberger et al. [26] failed to prove a correlation between EGFR expression and proliferative activity (Ki-67 index) in gliomas. In the present study, the percentage of EGF-R positive cells increased with grade of malignancy as did the mean Ki-67 (MIB 1) labeling index, but the correlation between EGF-R and Ki-67 was less distinct, which indicates that the EGF-R phosphorylation does not stimulate only the proliferative activity. Or altered nonfunctional EGF-R protein was recognized by the antibody [26]. However, tumors without detectable EGF-R immunoreaction showed significantly lower Ki-67 indices. TGF-alpha did not correlate with Ki-67, and Ki-67 indices were not significantly different in TGFalpha negative gliomas. Thus, in our study, the conclusion can be drawn that TGF-alpha does not only influence immediately the proliferative activity via EGF-R in the tumors investigated. This suggestion is also supported by the fact that antibodies to TGF-alpha only slightly suppressed DNA synthesis in malignant glioma cell lines [17]. By contrast, Ki-67 antibody, suitable for paraffin sections, could have led to false staining results [42], since there was a significant correlation between the percentages of proliferating cells (S + G/M), determined cytophotometrically, and TGF-alpha positive cells. However, the method of image cytophotometry bears technical problems, since the evaluated ploidy level 2.5c-4c (S + G/M fraction) contains not only proliferating cells but also non-proliferating aneuploid and cut nuclei.

146 . P. von Bossanyi et al.

Tyrosine phosphorylation of EGF-R-related protein cerbB-2 is induced by phosphorylation of erbB-3 or erbB4 receptors after heregulin binding [5,25,34]. Like other authors [3, 32J, we observed a remarkable increase in the percentage of c-erbB-2 immunoreactive cells and in the number of positive tumors in malignant gliomas. There was a strong correlation between the expression of cerbB-2 and EGF-R, indicating a coexpression of these two proteins, which was also proven by double immunostaining in the same tumor cells [9]. Similar to EGF-R, the percentage of c-erbB-2 positive cells correlated less distinctly with Ki-67 index. Ki-67 index of c-erbB-2 negative tumors was significantly lower. The results presented support the notion that the increased expression of TGF-alpha/EGF-R and c-erbB-2 is correlated with proliferative activity, but that these proteins might not be involved in the regulation of glioma cell growth at all. Other growth factors and their receptors such as platelet-derived growth factor (PDGF), not investigated here, may play a role [43]. The multiplicity of the growth factor-mediated pathways is probably important in the growth promotion of gliomas [17J. Furthermore, mutant EGF-R, constitutively activated even without ligand binding, may be involved in proliferative stimulation [6, 24].

References 1. Agosti RM, Leuthold M, Gullick WI, Yasargil MG, Wiestler OD (1992) Expression of the epidermal growth factor receptor in astrocytic tumours is specifically associated with glioblastoma multiforme. Virchows Arch A Pathol Anat 420: 321-325 2. Bargmann CI, Hung MC, Weinberg RA (1986) The neu oncogene encodes an epidermal growth factor receptor related protein. Nature 319: 226-230 3. Bernstein 11, Anagnostopoulos AV, Hattwick EA, Laws ER (1993) Human-specific c-neu proto-oncogene protein overexpression in human malignant astrocytomas before and after xenografting. 1 Neurosurg 78: 240-251 4. Briscoe 1, Guschin D, Muller M (1994) lust another signalling pathway. Current Bioi 4: 1033-1035 5. Carraway KL 3rd, Sliwkowski MX, Akita R, Platko IV, Guy PM, Nuijens A, Diamonti AI, Vandler RL, Cantley LC, Cerione RA (1994) The erbB3 gene product is a receptor for heregulin. 1 BioI Chern 269: 14303-14306 6. Collins VP (1995) Gene amplification in human gliomas. Glia 15: 289-296 7. Derynck R (1990) Transforming growth factor-alpha. Mol Reprod Dev 27: 3-9 8. Diedrich U, Lucius 1, Baron E, Behnke 1, Pabst B, Zoll B (1995) Distribution of epidermal growth factor receptor gene amplification in brain tumours and correlation to prognosis. 1 Neuro1242: 683-688 9. Dietzmann K, von Bossanyi P (1994) Coexpression of epidermal growth factor receptor protein and c-erbB-2 oncoprotein in human astrocytic tumors. Zentralbl Pathol

140:335-341

10. Di Marco E, Pierce IH, Fleming TP, Kraus MH, Molloy Cl, Aaronson SA, Di Fiore PP (1989) Autocrine interaction between TGF alpha and the EGF-receptor: quantitative requirements for induction of the malignant phenotype. Oncogene 4: 831-838 11. Dipasquale B, Colombatti M, Tridente G (1990) Morphological heterogeneity and phenotype modifications during long term in vitro cultures of six new human glioblastoma cell lines. Tumori 76: 172-180 12. Groenen LC, Nice EC, Burgess AW (1994) Structurefunction relationships for the EGFffGF-alpha family of mitogens. Growth Factors 11: 235-257 13. Heldin CH (1995) Dimerization of cell surface receptors in signal transduction. Cell 80: 213-223 14. Humphrey PA, Gangarosa LM, Wong AI, Archer GE, Lund-lohansen M, Bjerkvig R, Laerum OD, Friedman HS, Bigner DD (1991) Deletion-mutant epidermal growth factor receptor in human gliomas: effects of type II mutation on receptor function. Biochem Biophys Res Commun 178: 1413-1420 15. laros E, Perry RH, Adam L, Kelly PI, Crawford PI, Kalbag RM, Mendelow AD, Sengupta RP, Pearson ADI (1992) Prognostic implications of p53 protein, epidermal growth factor receptor, and Ki-67 labelling in brain tumours. Br 1 Cancer 66: 373-385 16. Kleihues P, Burger SH, Scheithauer BW (1993) World Health Organisation-international histological classification of tumors. Histological typing of tumors of the central nervous system. 2nd edn. Springer, Berlin Heidelberg New York, pp 11-16 17. Kurimoto M, Nogami K, Ogiichi T, Nishijima M, Takaku A (1995) Antiproliferative effect of multiple autocrine loop blockade in human malignant glioma cell lines. Neurol Med Chir Tokyo 35: 723-727 18. Lax I, Bellot F, Howk R, Ullrich A, Givol D, Schlessinger 1 (1989) Functional analysis of the ligand binding site of EGF receptor utilizing chimeric chicken/human receptor molecules. EMBO 1 8: 421-427 19. Lee DC, Fenton SE, Berkowitz EA, Hissong MA (1995) Transforming growth factor alpha: expression, regulation, and biological activities. Pharmacol Rev 47: 51-85 20. Lee DC, Rochford RM, Todaro GS, Villareal LP (1985) Developmental expression of rat transforming growth factor-alpha mRNA. Mol Cell Bioi 5: 3644-3646 21. Maruno M, Kovach IS, Kelly PI, Yanagihara T (1991) Transforming growth factor-alpha, epidermal growth factor receptor, and proliferating potential in benign and malignant gliomas. 1 Neurosurg 75: 98-102 22. Massague 1 (1983) Epidermal growth factor-like transforming growth factor. II. Interaction with epidermal growth factor receptors in human placenta membranes andA431 cells. 1 Bioi Chern 258: 13614-13620 23. McCormick F (1994) Raf: the holy grail of Ras biology? Trends Cell Bioi 4: 347-350 24. Nishikawa R, Ii XD, Harmon RC, Lazar CS, Gill GN, Cavenee WK, Huang HI (1994) A mutant epidermal growth factor receptor common in human glioma confers enhanced tumorigenicity. Proc Natl Acad Sci USA 91: 7727-7731 25. Plowman GD, Green 1M, Culouscou 1M, Carlton GW, Rothwell VM, Buckley S (1993) Heregulin induces tyro-

Correlation ofTGF-alpha and EGF-Receptor Expression· 147

26.

27.

28.

29.

30.

31. 32. 33.

34.

35.

sine phosphorylation of HER4/pI80erbB4. Nature 366: 473-475 Reifenberger G, Prior R, Deckert M, Wechsler W (1989) Epidermal growth factor receptor expression and growth fraction in human tumours of the nervous system. Virchows Arch APathol Anat 414: 147-155 Reifenberger G, Reifenberger J, Bilzer T, Wechsler W, Collins P (1995) Coexpression of transforming growth factor-alpha and epidermal growth factor receptor in capillary hemangioblastomas of the central nervous system. Am J Pathol147: 245-250 Reynolds FH, Todaro GJ, Fryling C, Stephenson JR (1983) Human transforming growth factors induce tyrosine phosphorylation of EGF receptors. Nature 292: 259-262 Samuels V, Barrett JM, Bockman S, Pantazis CG, Allen MB (1989) Immunocytochemical study of transforming growth factor expression in benign and malignant gliomas. Am J Pathol134: 895-902 Schechter AL, Stem DF, Vaidyanathan L, Decker SJ, Drebin JA, Greene MI, Weinberg RA (1984) The neu oncogene: an erb-B-related gene encoding a 185.000 Mr tumor antigen. Nature 312: 513-516 Schlegel U, Moots PL, Rosenblum MK, Thaler HT, Furneaux HM (1990) Expression of transforming growth factor alpha in human gliomas. Oncogene 5: 1839-1842 Schwechheimer K, Uiufle RM, Schmahl W, Knodlseder M, Fischer H, Hofler H (1994) Expression of neu/c-erbB2 in human brain tumors. Hum Pathol25: 772-780 Seroogy KB, Lundgren KH, Lee DC, Guthrie KM, Gall CM (1993) Cellular localization of transforming growth factor-alpha mRNA in rat forebrain. J Neurochem 60: 1777-1782 Sliwkowski MX, Schaefer G, Akita RW, Lofgren JA, Fitzpatrick VD, Nuijens A, Fendly BM, Cerione RA, Vandlen RL, Carraway KL 3rd (1994) Coexpression of erbB2 and erbB3 proteins reconstitutes a high affinity receptor for heregulin. J BioI Chern 269: 14661-14665 Sporn MB, Todaro GJ (1980) Autocrine secretion and malignant transformation of cells. N Engl J Med 303: 878-880

36. Strommer K, Hamou MF, Diggelmann H, De Tribolet N (1990) Cellular and tumoural heterogeneity of EGFR gene amplification in human malignant gliomas. Acta Neurochir 107: 82-87 37. Torp SH, Helseth E, Dalen A, Unsgaard G (1991) Epidermal growth factor receptor expression in human gliomas. Cancer Immunol Immunother 33: 61-64 38. Torp SH, Helseth E, Dalen A, Unsgaard G (1992) Relationships between Ki-67 labelling index, amplification of the epidermal growth factor receptor gene, and prognosis in human glioblastomas. Acta Neurochir 117: 182-186 39. U HS, Espiritu aD, Kelley PY, Klauber MR, Hatton JD (1995) The role of the epidermal growth factor receptor in human gliomas: I. The control of cell growth. J Neurosurg 82: 841-846 40. Untawale S, Zorbas MA, Hodgson C, Coffey RJ. Gallick GE, North SM, Wildrick DM, Olive M, Blick M, Yeoman LC, Boman BM (1993) Transforming growth factoralpha production and autoinduction in a colorectal carcinoma cell line (DiFi) with an amplified epidermal growth factor receptor gene. Cancer Res 53: 1630-1636 41. Van Zoelen EJJ, LenferinkAEG, Kramer RH, Van de Poll MLM (1996) Rational design for the development of epidermal growth factor receptor antagonists. Path Res Pract 192:761-767 42. Weis S, Protapapa D, Hischa A, Winkler PA, Reulen HJ, Mehraein P (1995) Significance of proliferation markers in recurrent meningiomas. Clin Neurosci 48/1. Suppl: 56 43. Westermark B, Heldin CH, Nister M (1995) Platelet-derived growth factor in human glioma. Glia 15: 257-263 44. Yaish P, Gazit A, Gilon C, Levitzki A (1988) Blocking of EGF-dependent cell proliferation by EGF receptor kinase inhibitors. Science 242: 933-935 45. Yeh J, Yeh YC (1989) Transforming growth factor-alpha and human cancer. Biomed Pharmacother 43: 651-659

Received: May 6, 1997 Accepted: January 8, 1998