Int. J. Cancer (Pred. Oncol.): 69, 165-169 (1996) 0 1996 Wiley-Liss, Inc. Publicationof the InternationalUnion Against Cancer Publicationde I'llnion lnternationaleContre le Cancer ASSESSMENT OF EGFR AND TGF-ALPHA EXPRESSION IN RELATIONSHIP TO HPV STATUS AND KI-67 DISTRIBUTION IN CERVICAL INTRAEPITHELIAL NEOPLASMS Athanassios DELLAS1s3,Elisabeth SCHULTHEISS~, Alfonso C. ALMENDRAL', Joachim TORHORST? and Fred G U D A T ~ Department of Obstetrics and Gynecology and 21nstituteof Pathology, University of Basel, Basel, Switzerland. ' Expression of epidermal-growth-factor receptor (EGFR), transforming growth factor alpha (TGF-a) and Ki-67 proliferation antigen in cervical intra-epithelial neoplasms were analyzed. To examine the interrelationshipof TGF-a, EGFR, Ki-67 and HPV status in dysplasia and carcinoma in situ, formalin-fixed tissue sections of 92 women were immunostainedwith monoclonal antibodies to EGFR, TGF-a and Ki-67. The presence of HPV was assessed by in situ DNA hybridization. The highest positive TGF-a expression was seen in the group of mild dysplasia. The difference was significant between the relatively high expression in mild dysplasia and the low occurrence in severe dysplasia and carcinoma in situ as well. The same relation could be found between TGF-a expression in papilloma-virus-negative dysplasia and those with the presence of HPV 16/ 18. In contrast to these findings, the Ki-67 proliferation marker was intensely detectable in severe dysplasia and carcinoma in situ. Ki-67stained neoplastic cell n d e i were found in a significantly higher percentage of HPV-positive than in HPV-negative lesions. TGF-a over-expression is obviously combined with low proliferating activity and vice versa. Irrespective of the grade of dysplasia or HPV status, EGFR was expressed abnormally as compared with normal squamous epithelium. Over-expressionof TGF-a in mild dysplasia could be associated with the autocrine pathway of cell-growth regulation. In the presence of HPV 16/ 18 the EGFR/TGF-a pathway for growth stimulation is probably not involved. o 1996 Wiley-Liss,Inc. The concept that simultaneous production of a growth factor and expression of its specific receptor by the same cell could result in self-stimulation is based on the landmark description by De Larco and Todaro (1978) of the autostimulation pathway of EGF-receptor activation in cultured tumor cells (Sporn and Todaro, 1985). Although TGF-a and EGF act primarily by stimulating cell proliferation, they also can have inhibitory effects. In high concentrations, they inhibit the growth of cultured tumor cell lines that express extraordinarily high levels of EGF receptors. The mechanisms that determine these 2 opposed pathways are not yet determined (Mendelsohn and Lippman, 1993). Many types of epithelial malignancies, such as cancers of the lung, breast and bladder, display increased EGF receptors on their cell-surface membranes (Harris et al., 1989; Neal et aL, 1985; Ozanne et al., 1986; Sainsbury et al., 1985; Veale et al., 1989). Increased receptor expression is often associated with increased production of TGF-a by the same tumor cells (Luetteke and Lee, 1990). Additional data are derived from using anti-EGF-receptor monoclonal antibodies (h4Abs) in nonmalignant and malignant cell lines. The data from these studies are best presented by considering 3 types of response to anti-EGF-receptor MAbs (Bates et al., 1990; Ennis et al., 1989). Cells that can proliferate without exogenous TGF-a and are strongly inhibited by MAbs presumably have an active and obligatory autocrine pathway. Second, other cell lines proliferate in response to exogenous TGF-a and are inhibited by MAbs. The third type of response represents cells that express EGF receptor, do not respond to TGF-a and ignore the presence of MAb. It is suggested that they do not require receptor activation for growth. Similar pathways could be demonstrated using MAbs against TGF-a in cultures with EGF/TGF-a-dependent malignant epithelial cells (Ciardiello et al., 1990). To explore the role of the TGF-a/EGF-receptor pathway in the growth of cervical cancer, Brown et al. (1994) analyzed 4 cervical cancer cell lines for EGFR content, TGF-a-gene expression and growth response to treatment with an anti-EGFR MAb that blocks TGF-a binding to EGFR. Their findings provide evidence for an autocrine-growth-stimulatory pathway involving TGF-a/ EGFR. In clinical material, the immuno-histochemical evidence of the EGFR/TGF-a pathway is particularly lacking. A MAb, Ki-67, has been utilized to demonstrate proliferating cells in the GI, S, G2 and M phases of the cell cycle (Gerdes et al., 1984). Cells in quiescent phase Go consistently lack this antigen. Ki-67 stained an antigen present in the nucleoli of proliferative interphase cells as well as the condensed chromatin in mitotic cells. This MAb has since been applied to study the growth fraction and cytokinetic activities in various cancers, e.g., lung cancer (Gatter et al., 1986), breast cancer (Gerdes et al., 1986) and cervical cancer (Brown et al., 1988). Its use has provided useful prognostic information. It was found that a high Ki-67 score correlates with histological tumor grade and with early recurrence in cancer. The present study describes results for clinical material. We attempted to identify EGF receptor and TGF-a expression in dysplasia and in carcinoma in situ of the uterine cervix. We also compared the presence of EGF receptor and TGF-a with histological grade and HPV status and assessed neoplastic-cell proliferation by analyzing the Ki-67 staining. SUBJECTS AND METHODS Samples of 20 specimens with normal cervical squamous epithelium and 92 specimens of dysplasia and carcinoma in situ (CIS) of the uterine cervix were obtained prior to surgical treatment or from hysterectomy specimens in the Department of Gynecology. The intraepithelial neoplasms were classified and graded according to the WHO criteria. Immunostaining of EGFR and TGF-a The tissue samples were fixed in 4% buffered formalin, embedded in paraffin and routinely stained to determine histology. Serial sections 4-pm thick were cut from each selected block, de-waxed and processed for enzyme digestion with 0.1% pronase (protease type XIV; Sigma, St. Louis, MO) for 5 min at 37°C in a water bath. To abolish endogenous peroxidase activity, sections were immersed in 0.3% H202 methanol. They were buffered in 0.3 M PBS at pH 7.3 and then incubated sequentially with normal horse serum (Vector, Burlingame, CA) for 30 min at room temperature, and afterwards with the primary MAb (mouse) at 4°C overnight. The primary mouse MAbs used were obtained from Triton Diagnostics (Alameda, CA; anti-EGFR, clone 31G7; diluted to 1:150) and from Oncogene Science [Uniondale, NY; anti3To whom correspondence and reprint requests should be sent, at Department of Obstetrics and Gynecology, University of Basel, Schanzenstrasse 46, CH-4031 Basel, Switzerland. Fax: (+41) 61-3259191. Received: October 20,1995 and in revised form January 24,1996. 166 DELLAS ETAL TGF-a (Ab-2), clone 213-4.4; diluted to 1:1000]. After overnight incubation, the slides were incubated with biotinylated anti-mouse IgG (1:200) as secondary antibody, for 30 min at room temperature. The streptavidin-biotin-peroxidase preformed complex (ABC Elite Standard, Vector) was subsequently applied for 30 rnin at room temperature, diluted at 150. In order to obtain a negative control, appropriate slides were incubated with PBS instead of primary mouse MAbs. The immunologic reactions for the EGFR and TGF-(Y.antibodies were developed with 3,3’-diamino-benzidine (DAB) tetrahydrochloride (Sigma). To obtain this solution, 3 tablets of DAB tetrahydrochloride (1 tablet = 10 mg) were dissolved in 60 ml 0.05 M Tris-buffered saline (pH 7.6). To DAB, 40 pI 30% hydrogen peroxide were added. The slides were then incubated with this solution for 6 rnin at room temperature. For additional chromogen enhancement, osmium tetroxide (Simec, Birsfelden, Switzerland) was applied for 2 min at room temperature, diluted at 1 : l O in PBS. Sections were counterstained with hematoxylin (Shandon, Zeist, The Netherlands), dehydrated and mounted with Eukitt (0.Kindler, Freiburg, Germany). Ki-67 immunostairiing The tissue samples were de-waxed and rehydrated. They were placed for 60 min at 90°C in 10 mM citrate buffer (pH 6.0). This step was performed using a H2500 microwave processor (Energy Beam Sciences, Agawam, MA). Following the microwave procedure, all slides were treated with 0.3% H 2 0 2methanol and subsequently with normal horse serum for 30 min at room temperature. To assess the nuclear Ki-67 proliferation antigen, the mouse MAb MIB 1 (Dianova, Hamburg, Germany) was used, diluted to 15300. With the applied MIB 1, the slides remained overnight at 4°C in a humid chamber. Thereafter, the slides were incubated with the bridging anti-mouse antibody, the avidin-biotin-peroxidase complex, and then with 3-amino-9-ethylcarbazole (AEC) chromogen (BioGenex, San Ramon, CA) to obtain strong red Ki-67 staining. The sections were counterstained with hemalum. This procedure leads to blue-stained Ki-67-negative nuclei. We chose the AEC chromogen to obtain a better contrast between red- and blue-stained nuclei. Finally, all sections were mounted with Crystal/Mount (Biomeda, Foster City, CA). Assessment of stainingfeatures Immunohistochemical staining of TGF-a and EGFR were scored as follows: 0, negative; 1, sporadic reaction on single cells; 2, heterogenous pattern with a less than 50% of positive cells; 3, staining present in the majority of neoplastic cells; 4, homogeneous staining in all neoplastic cells. The score was converted in values expressed by percentage for better presentation purposes. For EGFR, only membrane but not cytoplasmatic staining was scored; for TGF-a, cytoplasmatic and perinuclear staining was scored. Over-expression of EGFR was defined as staining in excess of normal cervical epithelium, which shows staining only in basal cells and the lower third of squamous epithelium. Staining of TGF-a was considered as positive when immunoreactivity occured only in neoplastic cells. A quantitative method was used to count the proportion of Ki-67-positively stained nuclei. In cervical intraepithelial lesions, 500 neoplastic cells were counted at randomly chosen fields at ~ 4 0 0magnification. The growth fraction of a neoplasm can be expressed as the percentage of positively stained neoplastic nuclei in the total number of nuclei counted. Analysis of the HPVstatus For the detection and typing of human papillomavirus (HPV) DNA in the specimens we used an in situ DNAhybridization procedure (PathoGene DNA Probe Assay, Enzo, Farmingdale, NY) to identify HPV types 6/11, 16/18 or 31/33/51. The formalin-hed, paraffin-embedded tissue sections were de-paraffinized, rehydrated and treated with 0.0125% proteinase K (to make the specimen DNA accessible to the biotinylated DNA probes) and with 3% hydrogen peroxide in buffered NaCI/EDTA (to eliminate endogenous peroxidase activity). In small biopsy specimens, 0.0125% proteinase K was applied for 5 rnin on a 37°C heating block, whereas on relatively large tissue samples the incubation time was extended to 15 min. The sections were then dehydrated and dried at room temperature. In accordance with instructions for the assay, one drop (40 KI) of each used vial of HPV DNA probe reagent was applied to the slides. They were placed on a 95°C heating block for 3 min and then moved to the 37°C slide warmer for 15 min. They remained overnight in a humid chamber at 37°C to obtain a stronger signal. During the post-hybridization procedure, 0.5 ml buffered NaCl containing formamide were applied to each slide for 10 min. Before application of the detection reagent, the slides were gently washed in buffered NaCI/EDTA. Specific hybridization between the HPV DNA probes and DNA in the specimen was determined by the detection of biotin. Detection of biotin was accomplished in 2 steps. First, a streptavidin-horseradishperoxidase complex was bound to the biotin of the hybridized HPV DNA probes. In the second step, the entire complex was visualized after conversion of the substrate and chromogen into a localized brick-red precipitate within the nuclei of epithelial cells. As chromogen we used AEC mixed with hydrogen peroxide in acetate buffer. The slides were incubated with chromogen for 10 min at room temperature. Following counterstaining with hemalum and mounting with Crystal/ Mount (Biomeda), the stained cells were observed by light microscopy. For a specimen to be identified as positive for HPV, at least one positive brick-red nucleus must be present. Statistical analysis Results are given as mean and standard deviation (SD). One-way ANOVA was used to compare the groups of Ki-67, EGFR and TGF-a scores in relationship to histological grade and HPV status. RESULTS The expression of EGFR, TGF-a and Ki-67 proliferation antigen in the different histological grades of cervical intraepithelial neoplasia is shown in Table I. The EGF receptor is over-expressed in dysplasia, irrespective of the grade, and in carcinoma in situ of the uterine cervix. In contrast to dysplasia, EGF-receptor expression is lost in the upper two thirds of normal epithelium. Intense cell-surface expression of the receptor was demonstrated by the anti-EGFR MAb used, which only rarely presented a diffuse cytoplasmic pattern. No significant differences in EGFR expression could be found between different grades of intraepithelial neoplasms. Receptor immunoreactivity was not related to HPV status (Table 11). The highest positive TGF-a expression was seen in the group of mild dysplasia. Compared with expression in the carcinoma in situ, the difference was significant (Fig. la). The higher the grade of intraepithelial neoplastic cell transformation, the lower was the TGF-a immunoreactivity ( p = 0.001). The expression pattern was staining in a few selected dysplastic cell groups, and the pattern of TGF-a immunoreactivity showed cytoplasmic fine granularity, in most cases combined with intense perinuclear staining. In normal epithelium, TGF-a was not detectable; it was found particularly in HPV-negative cervical intraepithelial neoplasms (Fig. l b ) . The higher the incidence of HPV in intraepithelial neoplasms, the lower was the TGF-a expression (Table 11). We found a correlation between TGF-a expression in HPV-negative intraepithelial 167 TGF-ALPHA, EGFR, HPV AND KI-67 IN CERVICAL CANCER neoplasms and pre-cancerous lesions infected with HPV type 16/18 (Fig. 3a). No correlation was found between EGFR and TGF-a positivity (results not shown). In contrast to the results of the TGF-a analysis, we found a high density of Ki-67-positively-stainedneoplastic nuclei in severe dysplasia and carcinoma in situ (Fig. Za). The area of nuclear staining varied from whole nuclei to the nucleoli and to small clusters along the nuclear membrane. Cytoplasmic staining by Ki-67 was not observed. In normal cervical epithelium, Ki-67 was present in the nuclei of the basal-cell layer, as the most active cell layer. A relatively low count of Ki-67-positive cells occured in mild and moderate dysplasia (Table I). HPV-positive cervical intraepithelial neoplasms were associated with a significantly higher percentage of Ki-67-stained nuclei, compared with HPV-negative lesions (Table 11, Fig. 2b). Of 92 probes with cervical intraepithelial neoplasms, 63 were HPV-16/18-positive, 12 were HPV-31/33/51-positive and 2 were HPV-6/1l-positive. The analysis of Ki-67 staining showed a clear difference between HPV-negative and HPV16/18 probes (Fig. 3b). TANLE I RtLATIOh I3LTWFF.N HISTOI.0GIC' CiKAUL: OF C ' t K V I C " . l\\TKAEPllHI:I.IAL StOPl..ASSIS, TGF-a, EGbK ,\NVKI-07 STAIKIYG ( V A I LJES I\' M I 3 S = 'ill ,\RE Histologic EX1'KI.SPED AS I'tIR('ENT,\GE) Number made Mild dysplasia Moderate dysplasia Severe dysplasia Carcinoma in situ 12 20 19 41 Mild dysplasia Moderate dysplasia Severe dysplasia Carcinoma in situ 12 20 19 41 Mild dysplasia Moderate dysplasia Severe dysplasia Carcinoma in situ 12 20 19 41 Mean 2 p value SD TGF-a 45.6 ? 34.7 23.5 2 35.6 4.7 k 13.9 8.2 k 18.6 EGFR 92.9 18.89 79.4 ? 20.68 90.8 i. 18.05 76.9 k 21.76 Ki-67 22.2 2 8.6 31.5 2 9.8 42.9 2 11.4 37.1 12.4 .001 * NS .0004 * TABLE I1 - RELATION BETWEEN HPV STATUS, TGF-a, EGFR AND Ki-67 STAINING (VALUES IN MEAN HPV status Number Negative Positive 15 77 Negative Positive 15 77 Negative Positive 15 77 2 SD EXPRESSED AS PERCENTAGE) Mean 2 SD p value TGF-a 37.8 2 37.83 11.15 2 23.38 EGFR 80.83 k 19.64 82.05 k 22.11 Ki-67 25.9 ? 8.7 37.5 12.7 .0032 NS .003 * DISCUSSION The growth factor TGF-a is present in cervical intraepithelial neoplasms and can be detected by immunohistochemistry. Although EGFR and TGF-a have been found both in normal and in neoplastic endometrium (Santini et al., 1994), and in different cervical-carcinoma cell lines (Brown et al., 1994). relatively little is known about the distribution and significance of both factors among normal squamous epithelium, precancerous cervical lesions and different histotypes of cervical carcinoma. In the present study, a significant correlation between grade of intraepithelial neoplasms, HPV status, Ki-67 staining and TGF-a expression, detectable particularly in HPV-negative mild dysplasia, was found. These relationships suggest TGF-a/EGFR-dependent neoplastic autocrine growth stimulation in some cervical pre-cancerous lesions. Any contribution of the TGF-a/EGFR pathway to autocrine growth is probably strictly confined to mildly dysplastic lesions only. Such growth is never substantial, since these early lesions have a relatively low proliferative capacity. In this study, TGF-a expression was associated with low Ki-67-proliferation-marker activity and with HPV-negative cervical lesions. Staining for E G F R and TGF-a was intense in the neoplastic cells, in contrast to weaker plasma-membrane or basal-cell distribution of E G F R and no staining for TGF-a in normal cervical tissues. In a TGF-a-positive tissue, presumably not all neoplastic cells are capable of producing TGF-a for autocrine growth. W e demonstrated this heterogeneity of neoplastic cells in TGF-a expression in cervical neoplastic epithelium. Lack of TGF-a staining in a few moderate dysplasia, in severe dysplasia and in carcinoma in situ suggests that the proliferative state of neoplastic cells is maintained through other mechanisms, possibly other autocrine pathways or oncogene amplifications. / 90 8 80 a 80 70 60 3 50 .E $ b 8 70 5 60 a 5 50 40 40 2 T 3 30 20 I 1 -20 -20 CIN I CIN II CIN I l l CIS ' H Pkpositive HPV-negative FIGURE1- ( a ) Distribution of TGF-a in relation to the histological grade of cervical intraepithelial neoplasms (CIN). CIN I, mild dysplasia; CIN 11, moderate dysplasia; CIN 111, severe dysplasia; CIS, carcinoma in situ. (b) Distribution of TGF-a in relation to HPV status. 168 DELLAS E T A L . a 60] b P P 50 - &- 4 0 - 8 .-C .L Y 8 30 - c C 8 20 - H - 3 100 ' 0 ' CIN II CIN I CIN Ill HP w o s i t i v e CIS H PV-negative FIGURE 2 - (a) Distribution of Ki-67 in different histological grades of cervical intraepithelial neoplasms (CIN). CIN I, mild dysplasia; CIN 11, moderate dysplasia; CIN 111, severe dysplasia; CIS, carcinoma in situ. (b) Relation between Ki-67 counts and HPV status ( p = 0.002). 60 -.. P 4) E. 20-1 ......................... -20 b ' 0 ' HPV-negative HPV 16/18 HPV-negative HPV 16/18 F~GURE 3 - (a) TGF-a expression in HPV-negative and HPV-16/ 18-positive intraepithelial neoplasms. The difference between the 2 groups is significant ( p = 0.006). (b) Difference in Ki-67 counts of HPV-negative and HPV-lh/l8-positive intraepithelial neoplasms ( p = 0.0017). The infection with HPV, particularly type 16/18, possibly inhibits the process of TGF-a/EGFR-dependent autocrinegrowth stimulation. It seems that neoplastic cells in HPVinfected squarnous epithelium do not need this pathway. Our results show that Ki-67 scores have a significant correlation with histological grade of intraepithelial neoplasms and with HPV status, with highest scores present in the poorly differentiated pre-cancerous lesions and in HPV-positive, particularly HPV-16/18-infected, pre-cancerous lesions. A-Saleh et al. (1995) have described significantly higher densities of Ki-67antigen-positive cells in HPV-16/18-positive cervical biopsies than in low-risk HPV types. These results suggest that HPV, particularly HPV 16/ 18, induces intense proliferation activity in intra-epithelial neoplasms. Furthermore, HPV may induce growth stimulation by a viral-defective version of the erbBproto-oncogene, which encodes the EGF receptor (Mendelsohn and Lippman, 1993). Two modifications characterized in its structure explain its malignant potential. The oncogene encodes a receptor lacking the extracellular domain that is the binding site for TGF-a or EGF; it also lacks the carboxylterminal residues, which are presumed to have a regulatory function. It is hypothesized that the viral-encoded EGF receptor permanently mediates uncontrolled signal transduction to the cell nucleus (Watson et al., 1993). On the other hand, the HPV-16 E5 gene product can interact with the EGF receptor. This gene can stimulate the transforming activity of the receptor in the presence of exogenous ligand (Leechanachai et al., 1992). Pim et al. (1992) showed that the HPV-16 E5 protein renders cells more sensitive to lower concentrations of growth factors. The earlier observation that the HPV-16 E6 and E7 genes are preferentially retained in cervical carcinomas implies that any role E5 may have in the development of these tumors is in the early stages or in the pre-cancerous lesions. Possibly E5 sensitizes the cells to lower levels of growth factors early in the infection. Later, as the lesion progresses and E5 is often deleted, this may be overcome by the observed overexpression of EGFR. The activity of E5 may therefore be one of the first steps on the way to virus-mediated cell transformation, with the more apparent effects of E6 and E7 being later functions in maintenance of the transformed phenotype (Pim et al., 1992). We therefore suggest that there is no evidence of a TGF-a/ EGFR-dependent autocrine-growth-stimulation pathway in HPV-positive cervical carcinomas. Further investigations on invasive cervical carcinoma may provide additional understanding of growth-stimulation mechanisms. ACKNOWLEDGEMENTS This work was supported by a research grant from the Department of Obstetrics and Gynecology, University of Basel. TGF-ALPHA, EGFR, HPV AND KI-67 IN CERVICAL CANCER 169 REFERENCES AL-SALEH, W., DELVENNE, P., GREIMERS, R., FRIDMAN, V., DOYEN, J. and BONIVER,J., Assessment of Ki-67-antigen immunostaining in squamous intraepithelial lesions of the uterine cervix. Correlation with the histologic grade and human papillomavirus type. Amer. J. clin. Pathol., 104,154-160 (1995). BATES,S.E., VALVERIUS, E.M., ENNIS,B.W., BRONZERT, D.A., SHERIDAN,J.P.. STAMPFER, M.R., MENDELSOHN, J., LIPPMAN,M.E. and DICKSON, R.B., Expression of the transforming-growth-factor-alpha/ epidermal-growth-factor-receptor pathway in normal human breast epithelial cells. Endocrinology, 126,596-607 (1990). BROWN,C.L., RUBIN,M., MASUI,H. and MENDELSOHN, J., Growth inhibition by anti-EGF-receptor monoclonal antibody in squamous cervical carcinoma cells expressing TGF-a. Proc. Amer. Ass. Cancer Res., 35,381, A2270 (1994). BROWN, D.C., COLE,D.J., GATTER,K.C. and MASON,D.Y., Carcinoma of the cervix: an assessment of tumour proliferation using the monoclonal antibody Ki-67. Brit. J. Cancer, 57,178-181 (1988). F., MCGEADY,M.L., KIM, N., BASOLO,F., HYNES,N., CIARDIELLO, LANGTON, B.C., YOKOZAKI, H., SAEKI,T., ELLIOTT, J.W. and MASUI, H., Transforming-growth-factor-alpha expression is enhanced in human mammary e ithelial cells transformed by an activated c-Ha-ras proto-oncogene gut not by the c-neu proto-oncogene, and overexpression of the transforming-growth-factor-alpha complementary DNA leads to transformation. Cell Growth Difer. 1,407420 (1990). DE LARCO,J. and TODARO,G., Growth factors from murine sarcoma virus-transformed cells. Proc. nat. Acad. Sci. (Wash.), 75, 40014005 (1978). ENNIS,B., VALVERIUS, E., LIPPMAN,M., BELLOT,F., KRIS, R., SCHLESSINGER, J., MAW, H., GOLDENBERG, A,, MENDELSOHN, J. and DICKSON, R., Monoclonal anti-EGF-receptor antibodies inhibit the growth of malignant and nonmalignant human mammary epithelial cells. Mol. Endocrinol., 3,1830-1838 (1989). GATTER,K.C., DUNNILL,M.S., GERDES,J., STEIN,H. and MASON, D.Y., New approach to assessing lung tumours in man. J. clin. Pathol., 39,590-593 (1986). H., Growth fractions in breast GERDES,J., LELLE,R.J. and PICKARTZ, cancer determined in situ with the monoclonal antibody Ki-67. J. clin. Pathol., 39,977-980 (1986). GERDES, J., LEMKE,H., BAISCH,H., WACKER,H.-H., SCHWAB, U. and STEIN,H., Cell-cycle analysis of a cell-proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki-67. J. Immunol., 133,1710-1715 (1984). HARRIS,A.L., NICHOLSON, S., SAINSBURY, J.R.C., FARNDON, J. and WRIGHT,C., Epidermal-growth-factor receptor: a marker of early relapse in breast cancer and tumor-stage progression in bladdercancer interactions with neu. In: M. Furth and M. Greaves (eds.), The molecular diagnostics of human cancer, Vol. 7, pp. 353-357, Cold Spring Harbor Laboratory, NY (1989). LEECHANACHAI, P., BANKS,L., MOREAU,F. and MATLASHEWSKI, G., The E5 gene from human papillomavirus type 16 is an oncogene which enhances growth-factor-mediated signal transduction to the nucleus. Oncogene, 7,19-25 (1992). LUETTEKE,N.C. and LEE, D.C., Transforming growth factor alpha: expression, regulation and biological action of its integral membrane precursor. Semin. Cancer. Biol., 1,265-275 (1990). MENDELSOHN, J. and LIPPMAN, M., Principles of molecular cell biology of cancer: growth factors. In: V.J. DeVita, S. Hellman and S. Rosenberg (eds.), Cancer:principles andpractice of oncology. Vol. 1, pp. 114-133, J.B. Lippincott, Philadelphia (1993). NEAL, D.E., BENNETT,M.K. and HALL, R.R., Epidermal-growthfactor receptors in human bladder cancer: comparison of invasive and superficial tumors. Lancet, I, 366-368 (1985). OZANNE, B., RICHARDS, C.S., HENDLER,F., BURNS,D. and GUSTERSON, B., Over-expression of the EGF receptor is a hallmark of squamous-cell carcinomas. J. Pathol., 149,9-14 (1986). PIM,D., COLLINS, M. and BANKS,L., Human-papillomavirus-type-16 E5 gene stimulates the transforming activity of the epidermal-growthfactor receptor. Oncogene, 7,27-32 (1992). SAINSBURY. J.. MALCOLM. A.. APPLETON. D.. FARNDON. J. and HARRIS. A., Presence of epidermal-growth-factor receptor as ’an indicator of poor prognosis in patients with breast cancer. J. clin. Pathol., 38, 1225-1228 (1985). SANTINI, D., CECCARELLI, C., MARTINELLI, G.N., LEONE,O., MARABINI, A., ORLANDI, C. and MANCINI, A., lmmunocytochemical study of epidermal-growth-factor receptor, transforming growth factor alpha and “squamous differentiation” in human endometrial carcinoma. Hum. Pathol., 25,1319-1323 (1994). SPORN,M. and TODARO,G., Autocrine growth factors and cancer. Nature (Lond.), 313,747-751 (1985). VEALE,D., KERR,N., GIBSON, G.H. and HARRIS,A.L.. Characterization of epidermal-growth-factor receptor in primary human non-smallcell lung cancer. CancerRes., 49,1313-1317 (1989). WATSON,J., GILMAN,M., WITKOWSKI, J. and ZOLLER,M., Recombinant DNA, 2nd ed., pp. 315-346, W.H. Freeman, New York and Oxford (1993).