JOURNAL OF PATHOLOGY, VOL.
166: 7-12 (1992)
THE EPIDERMAL GROWTH FACTOR RECEPTOR IN HUMAN PANCREATIC CANCER NICHOLAS R. LEMOINE*, CHRISTINE M. HUGHEST, CLAIRE M. BARTON*, RICHARD POULSOM;, ROSEMARY E. JEFFERY:, GUNTER KLOPPEL~, PETER A. HALL)/ AND WILLIAM J. GULLICK?
*Molecular Pathology Laboratory and tkfolecular Oncology Laboratory, ICRF Oncology Group, Hammersmith Hospital, Du Cane Road, London W12 OHS, U.K.; :ICRF Histopathology Unit, Lincoln's Inn Fields, London WCZA 3PX, U.K.; §Department of Pathology, Academic Hospital Jette, Free University of Brussels, Laarbeeklaan 101. B-1090 Brussels, Belgium; /I Department of Histopathology, Royal Postgraduate Medical School, Hammersmith Hospital, Du Cane Road, London WI2 OHS, U.K. Received I3 May 1991 Accepted2 July I991
SUMMARY The epidermal growth factor receptor (EGFR) and its ligands are thought to be important in the control of proliferation of many epithelial systems, including the exocrine pancreas. Abnormalities in expression of two of the known ligands ofthe EGFR, transforming growth factor alpha and epidermal growth factor, occur frequently in ductal adenocarcinoma of the human pancreas. We have examined an archival series of cases of pancreatic pathology for expression of the EGFR using the anti-EGFR antiserum 12E and found that there is almost ubiquitous overexpression of EGFR in pancreatic cancer and in chronic pancreatitis. Southern blot analysis showed no evidence of amplification or rearrangement of the EGFR gene. We conclude that an autocrine loop involving the EGFR system may be involved in the genesis of both neoplasia and reactive hyperplasia of pancreatic ductal epithelium. KEY WORDS-Epidermal
growth factor receptor, pancreatic cancer, autocrine loop, immunohistochemistry.
INTRODUCTION Pancreatic cancer is a major clinical problem and reports from several groups have recently shed light on the molecular events involved in its pathogenesis. Mutation of the Kirsten rus gene at codon 12 is present in about 75 per cent of advanced pancreatic cancers (reviewed by Shibata et ul.' and Lemoine and Hall') and represents an early event, occurring in a similar proportion of ductal carcinoma in s i m 3 The c-erbB-2 proto-oncogene is overexpressed in at least 20 per cent of pancreatic c a n c e r ~and , ~ this may be associated with amplification and/or rearrangement of this gene in some cases ( C . M. Barton. unpublished results). The gene encoding the epidermal growth factor receptor (EGFR) is closely related to the c-erbB-2 proto-oncogene and we have recently reported that at least two of the ligands of the EGFR, transforming growth factor-alpha Addressee for correspondence: Dr Nicholas Lemoine. Molecular Pathology Laboratory. ICRF Oncology Group. Hammersmith Hospital. Du Cane Road. London W12 OHS.
U.K.
0022-3417192iO 1000746 $05.00 0 1992 by John Wiley & Sons, Ltd.
(TGF-alpha) and EGF/urogastrone, are frequently overexpressed in pancreatic carcinoma.' Autocrine stimulation of tumour cells could occur if the receptor is expressed in the same cells, since expression of TGF-alpha or EGF/urogastrone is mitogenic only in cell lines possessing EGF receptors.6 A low level of EGFR expression has been reported in normal pancreatic cells' and simultaneous overexpression of EGF and EFGR, and of TGF-alpha and EGFR, has been detected in pancreatic cancer cell lines, supporting this hyp~thesis.'.~ In a small series of frozen samples, EGFR was found to be overexpressed in 30-50 per cent of cases of pancreatic cancer." One group has correlated EGFR overexpression with alterations in chromosome 7." A new antipeptide polyclonal antiserum (1 2E) that recognizes a carboxy-terminal domain of the EGF receptor" in routinely-fixed, paraffinembedded tissue has very recently become available, and this stimulated us to examine a retrospective series of 84 pancreatic cancers in which we correlate expression of E G F receptor with expression of its ligands TGF-alpha and EGFjurogastrone. In a
8
N. R. LEMOINE ET AL.
series of 23 snap-frozen specimens of pancreatic cancer, we could also use the well-characterized anti-EGFR antibody EFGR1,I3.l4allowing us to validate our results with the new 12E antibody.
phosphate-buffered saline (PBS) for 5 min and then rinsing in distilled water. The immunohistochemical technique involved the sequential application of the following reagents: the primary antibody (12E, 14E, or EGFRI) in PBS containing 0.5 per cent bovine serum albumin (BSA) at the concentration stated MATERIALS AND METHODS above for 1 h at room temperature; a secondary biotinylated anti-rabbit (for 12E and 14E) or antiTumour material analysed mouse (for EGFRI) antibody (Dako) diluted 1500 Formalin-fixed, paraffin-embedded blocks of or 1:300, respectively, in PBS containing 0.5 per cent pancreatic adenocarcinoma ( n = 84), chronic pan- BSA for 30 mini and ABComplex/HRP (Dako) for creatitis (n = 14), and normal adult pancreas ( n= 10) 30 min. Each incubation was followed by rinsing the were obtained from the archives of Hammersmith tissue sections in PBS three times (5 min each). The Hospital. This series was previously examined for sites of immunocomplex were identified by light microscopy following treatment with a chromogen, expression of TGF-alpha and EGF/urogastrone.’ We also examined a series of 23 pancreatic 3,3’-diaminobenzidine tetrahydrochloride (Northcancers which had been snap-frozen in liquid nitro- east Biochem). The specificity of the two antipeptide antibodies was tested by inhibition of binding after gen at the time of surgery and stored at -70°C. These were collected at the Hammersmith Hospital pre-absorption of antibody with the immunizing and the Middlesex Hospital, London; at the Dudley peptide, and retention of staining after preabsorpRoad Hospital, Birmingham; and at the Academic tion with other non-specific peptides. Overexpression of the EGF receptor in this study is Hospital Jette, Free University of Brussels. Sections stained with haematoxylin and eosin defined as staining of cell membranes, often were examined and cases of pancreatic cancer accompanied by cytoplasmic staining. classified according to standard criteria.” Hybridization analysis of EGFR Immunohis tochemical detection of expression of Southern blot analysis of the EGFR gene was EGFR performed on eight pancreatic carcinoma cell To detect EGF receptor immunoreactivity in lines (BxPc3, ASPCI, PSN1, MiaPaCa2, Pancl, frozen and paraffin sections, an affinity-purified CaPan1, and CaPanZ, obtained from ATCC) and preparation of the rabbit antibody 12E (raised six primary pancreatic tumour biopsies with against a synthetic peptide representing residues paired normal DNA from the same patients using 1059-1072 in the cytoplasmic domain of the E G F previously described protocols.” In situ hybridization analysis of EGFR mRNA receptorI6) was used at a concentration of 4 pg/ml, was performed and controlled as described preas well as an affinity-purified preparation of the rabbit antibody 14E (raised against a synthetic peptide viously,18 using an EGFR riboprobe from of residues 1164-1176 of the EGF receptor16) at EcoRI-linearized phEGFR.4a, which we prepared 4 pg/ml. The monoclonal antibody EGFR 1 (which by subcloning the 913 bp Sac?-Hind?IIfragment of ’ ~ pGEM-3 (Promega). was raised against A431 cells and recognizes the ~ 6 2 . 1into native folded external domain of the EGF receptor;I3 it has been extensively characterized for use in immuRESULTS nohistocherni~try’~) was used a concentration of 10 pg/ml to detect EGF receptor immunoreactivity Normal Pancreas and chronicpancreatitis in frozen sections only. Only the islets of Langerhans showed EGFR Paraffin sections were dewaxed by passage through xylene and rehydrated through alcohols expression in histologically normal pancreas, and to water. Endogenous peroxidase activity was this was very weak. However, in chronic pancreatitis quenched by incubation in 0.3 per cent hydrogen (even if quite mild) the epithelial cells of the duct peroxide for 30 min and then rinsing in tap system showed strong cytoplasmic EGFR immuwater. For frozen sections, endogenous peroxidase noreactivity (Fig. 1); this affected the cells of the activity was quenched by incubation in 0.1 per small ducts (including the centroacinar cells) most cent phenylhydrazine hydrochloride (Sigma) in markedly but also involved the largest ducts in the
9
EGF RECEPTOR IN PANCREATIC CANCER
head of the pancreas. in such pancreatitis, the EGFR immunoreactivity of islet cells was accentuated, but the acinar cells and interstitium did not express detectable EGFR. This distribution of EGFR immunoreactivity is concordant with that of the ligand TGF-alpha previously demonstrated in these tissues.
cancer. Almost all cases showed coordinate expression of both EGFR and its ligand TGFalpha, but there were a few exceptions. One case was negative for EGF receptor expression but showed immunoreactivity for TGF-alpha, albeit relatively weak. The only case of clear cell carcinoma examined showed membrane immunoreactivity for EGFR, but was negative for TGF-alpha. All other cases were concordant for expression of both EGF receptor and TGF-alpha. Table I-Immunoreactivity for EGFR, TGF-alpha, and EGF in pancreatic adenocarcinoma
Tumour type Ductal carcinoma Grade 1 Grade 2 Grade 3
n
Irnmunoreactivi ty (cases positive) E G F R TGF-alpha E G F
30
28
29
4
31
30
30
19
18
18
4 2
Other exocrine pancreatic carcinomas 1 1 Clear cell Large cell 2 2 Signet ring 1 1 Total Fig. 1-EGF receptor immunoreactivityin chronic pancreatitis. There is specific staining ofthe epithelialcells of the ducts, including the centroacinar cells. Acinar cells and interstitiurn do not express detectable EGF receptor
Exocrine pancreatic cancer Eighty of the 84 cases (95 per cent) of exocrine pancreatic adenocarcinoma showed immunoreactivity for the EGF receptor with both anti-EGFR antibodies 12E and 14E. The immunoreactivity was mainly cytoplasmic but with accentuation at the plasma membrane of tumour cells. Although all malignant cells within a positive tumour showed immunoreactivity, there was some degree of heterogeneity with some cells being stained very densely and others relatively modestly. There was no apparent relationship to tumour grade or histological type (see Table I which presents the results of the present study together with those of our previous study’ for comparison). Figures 2 and 3 show the typical features of EGF receptor expression in pancreatic
84
80
0 2 1
80
0 0 0 10
As previously reported, immunoreactivity for EGF/urogastrone was detectable in 10 of the 84 cases (12 per cent), and all of these cases were also positive for TGF-alpha immunoreactivity and EGFR immunoreactivity In the frozen tissue sections of pancreatic carcinoma, EGF receptor expression was recognized by both the EGFRl monoclonal antibody and the 12E polyclonal antibody with the same sensitivity. All except one case (22123 cases) showed cytoplasmic immunoreactivity with these antibodies, and in most cases membrane staining was also seen. Southern blot analysis of the eight pancreatic cell lines and the six pancreatic tumour biopsies showed no evidence of rearrangement or amplification of the EGFR gene (data not shown). Hybridization in siru of EGFR mRNA detected no specific signal in any of the cases of pancreatic cancer examined. although a strong specific signal was present in the control gastric cancer material in which overexpression of EGFR was known to be very high.” Control hybridization with a specific riboprobe”
10
N.R. LEMOINE ET AL.
detected the presence of p-actin mRNA in the pancreatic cancers, indicating that the failure to detect EGFR expression in these cases was not the result of nucleic acid degradation. The amount of EGFR mNRA present is therefore below the detection limit of this in siru hybridization system, although EGFR protein is easily detectable by immunohistochemistry. DISCUSSION
Fig. 2-EGF receptor immunoreactivity in a pancreatic exocrine adenocarcinoma of intermediate grade (grade 2). There is heterogeneous staining of tumour cells and accentuation at the Cell membranes
Fig. 5-EGF receptor immunoreactivity in a poorly differentiated (grade 3) pancreatic adenocarcinoma. Many ofthe tumour cells show very strong cytoplasmic immunoreactivity
We have previously demonstrated that there is frequent overexpression of ligands for the EGF receptor in pathological conditions of the pancreas, both in neoplasia and in reactive hyperpla~ia.~ This study shows that in pathological conditions (pancreatitis and neoplasia) the duct epithelial cells express high levels of the EGF receptor. This is further support for the autocrine stimulation hypothesis in pancreatic pathology, since conditional expression of E G F receptors by pancreatic ductal epithelial cells may allow them to respond to TGF-alpha and EGF/urogastrone to facilitate repair and repopulation in pancreatitis, and tribute to proliferation and population expansion in neoplasia. The mechanisms of control of EGFR gene expression are currently an area of intensive study. Both positively acting transcription factors (such as the ubiquitous factor Spl as well as factors TCF, ETFl , and ETF2”-”) and at least one inhibitory transcription factor (factor GCFZ6)have been shown to interact with the EGFR promoter. The increased expression of EGFR in the pancreatic tumours is presumably due to increased transcription, since our Southern blot analysis showed no evidence for gene amplification. In view of our finding of increased expression of EGFR in ductal epithelial cells involved in chronic pancreatitis, it will be interesting to see how the activity of these various transcription factors is influenced by lymphokines which might mediate the inflammatory process. Many studies on other tumour types have shown that overexpression of the EGF receptor is associated with poor clinical prognosis. These include human gliomas in which there is much higher EGFR expression in tumours of aggressive histological type.”-3’ In breast cancer, high expression of EGFR is associated with short relapse-free interval and s u r ~ i v a lincluding ~ ~ ~ ~ * those with node-negative disease.39In our analysis, there was no difference in frequency of EGFR overexpressionin pancreatic tumours of different histological grades. EGFR
EGF RECEPTOR I N PANCREATIC CANCER
expression is not a useful prognostic marker in pancreatic cancer because essentially all cases have a poor prognosis and also usually overexpress thisprotein. Nor is the detection of EGFR overexpression likely to be of any value as a diagnostic marker, because overexpression also occurs at high frequency in the duct system of pancreata affected by chronic pancreatitis. However, for our understanding of the molecular basis of this disease it is important to assemble a full picture of the oncogenes and growth factor systems that are activated in order that we might identify new targets for therapy. In this regard, growth factor receptors are currently the focus of much activity, and many potential mechanisms for inhibition are being e ~ p l o r e d . ~ ' ACKNOWLEDGEMENTS
We gratefully acknowledge the support of the Imperial Cancer Research Fund. and thank Professor RCN Williamson, Mr RCG Russell and Mr J Neoptolemos for providing clinical material and encouragement. REFERENCES 1 Shibata D. Capella G. Perucho M. Mutational activation of the C-Kras gene in human pancreatic cancer. Bailliere's Clin Gasrroenrerol 1990:4 151-169. 2. Lemoine NR. H a l l PA. Oncogenes and growth factors in pancreatic cancer. Boilliere > C/in Gaslrornrerol 1990: 4 8 15-832. 3. Lemoine NR. Jain S. Hughes CM. el a/. Ki-ras oncogene activation in
4.
5. 6. 7. 8.
9.
10.
I I.
I2
invasivc ductal adcnocarcinoma. but not in ductal papillary hyperplasiaor intraductal papillaryneoplasiaofthepancreas.Gastroenterology (in press). Hall PA. Hughes CM. Staddon SL. Richman PI. Gullick WJ. Lemoine NR. The c-erb5-2 proto-oncogene in human pancreatic cancer. J Pa~hol1990:161: 195-200. Barton CM. Hall PA. Hughes CM. Gullick WJ. Lemoine NR. Transforming growth factor alpha and epidermal growth factor in human pancreaticcancer. JPatho/1991: 163: 111-1 16. Rosenthal A. Lundquist PB. Bringman TS. Goeddel DV. Derynck R. Expression in rat fibroblasts of a human transforming growth faCtOr-alpha cDKA results in transformation. Cell 1986~46:301-309. Damjanov I. Mildner B. Knowlcs BB. Immunohistochemical localisation of the epidermal growth factor receptor in normal human tissue. Lab Invesr 1986: 5 5 588-592. Chen YF. Pan G-Z. Hou X, er 01. Epidermal growth factor and its receptor in human pancreatic carcinoma. Pancreas 1990: 5: 278-283. Smith JJ. Derynck R. Korc M. Production of transforming growth factor-alpha in human pancreatic cancer cells: evidence for a superagonist autocrine cycle. Proc Null Acad Sci USA 1987: 84: 7567-7570. Kloppel G. Maillet B. Schewc K. Kalthoff H. Schmeigel WH. Immunocytochemical detection of epidermal growth factor receptors (EGFR) and transferrin receptor (TR) on normal. inflamed and neoplastic tissue. Pancreas 1989: 4 649 (abstract). Korc M . Meltzrr P. Trent J. Enhanced expression of epidermal grou,th factor receptor correlates with alterations of chromosome 7 in human pancreatic cancer. Proc. Null A r a d Sci GSA 1986: 8 3 5 14 1-5 144 Gullick WJ. Hughes CM. Mellon K. Neal DE. Lemoine NR. Immunohistochemical detection of thc epidermal growth factor receptor in paraffin-embedded human tissues. J Parhol164: 281-289.
11
13. Waterfield MD. Mayes ELV. Stroobant P. et al. A monoclonal antibody to the human rpidermal growth factor reseptor. J. CeN5iotltem 198220: 149-161. 14. Gullick WJ. Marsden JJ. Whittle N, er a/. Expression of epidermal growth factor receptors on human cervical. ovarian and vulva1 carcinomas. Cancer Rrs 1986: 46: 285-292. 15. Kloppel G. Pancreatic non-endocrine tumours. In: Kloppel G. Heitz PU. eds. Pancreatic Pathology. Edinburgh: Churchill Livingstone. 1981:79-113. 16. Gullick WJ. Downward J, Waterfield MD. Antibodies to the autophosphorylation sitesofthe epidermal growth factor receptor proteintyrosine kinease as probes of structure and function. E M 5 0 J 1985:4 2869-2877. 17. LemoineNR. Wyllie FS. LillehaugJR.etol.Absenceofabnormalities of the c-erbB-l and c-erbB-2 proto-oncogenes in human thyroid neoplasia. Eur J Cancer 1990: 2 6 777-779. 18. Wright NA. Poulsom R. Stamp GWH. er al. Epidermal growth factor (EGFIL'RO) induces expression of regulatory peptides in damaged human gastrointestinal tissues. J Pathol1990: 1 6 2 279-284. 19. Ullrich A. Coussens L. Hayflick JS. et al. Human epidermal growth factor receptor cDNA sequence and aberrant expression of the amplified gene in A431 epidcrmoid carcinoma cells. lvature 1984; 309: 418325. 20. Lcmoine NR. Jain S . Silvcstrc F. er al. Amplification and overexpression of the EGF receptor and c-erbB-2 proto-oncogenes in human stomach cancer. 5 r JCancer64: 79-83. 21. Porte P. Gunning P. Blau R. Kedes L. Human actin genes are single copy for alpha-skeletal and alpha-cardiac actin but multicopy for beta- and gamma-cytoskclctal genes: 3-untranslated regions arc isotopespecific but arc conscrved in evolution. Mol CeN Bio/ 1983: 3: 1783-1791. 22. Johnson AC. Ishii S. Jinno Y. Pastan 1. Mcrlino GT. Epidermal growth factor receptor gene promoter: deletion analysis and identificarion of nuclcar protein binding sites. J 5101 Chem 1988: 263: 5693-5699 23. Johnson AC. Jinno Y. Merlino GT. Modulation of epidermal growth factor receptor proto-oncogene transcription by a promoter sitesensitive to S1 nuclease. MolCel/BioI 1988:8:41744184. 24. Kageyama R. Merlino GT. Pastan I. Epidermal growth factor (EGF) receptor gene transcription: requirement for SPI and an EGF receptor factor. JBioI Chem 1988: 2 6 3 6329-6336. 25. Kageyama R. Merlino GT. Pastan 1. iiuclear factor ETF spccifically stimulates transcription from promoters without a TATA box. J Bio/ Chem 1989: 264: 15508-15514. 26. Kageyama R. Pasran I. Molecular cloning and characterization of a human DNA binding factor that represses transcription. Cell 1989: 59815-8?5.
27. Merlino GT. Epidermal growth factor receptor regulation and function. Semin Cancer B i d 1990: I: 227-284. 28. Libermann TA. Razon X . Barial AD. el a/. Expression of epidermal growth factor receptors in human brain tumors. Cancer Res 1984: 44: 753-760. 29. Libermann TA. Nusbaum HR. Razon N. er a/. Amplification. enhanced expression and possible rearrangement of EGF receptor eenc in primary human brain turnours of glial origin. Nature 1985: 513: 144-147. 30. Wong AJ. Bigner SH. Bigner DD. et a/. Increased expression of thc epidermal growth factor rcceptor gene in malignant gliomas is invariably associated with gene amplification. Proc f l u / / Acad Sci U S A 1987: 84: 689945903. 31. Hawkins RA. Killen E. Whittle IR. Jack WJL. Chetty 6. Prescotl RJ. Epidermal growth factor reccptors in intracranial and breast tumours: their clinical significance. 5 r JCancer 1991: 6 3 553-560. 32. Ekstrand AJ. James CD. Cavence WK. Seliger B. Pcttersson RF. Collins VP. Genes for epidermal growth factor receptor. transforming growth factoralpha. and epidermalgrowth factorand theirexpression i n humangliomasin uiro. ConcerRes 1991:51:2164-2172. 33. Sainsbury JRC. Malcolm AJ. Appleton DR. Farndon JR. Harris AL. Presence of epidermal growth factor receptor as an indicator of poor prognosis in paticnts with breast cancer. J Clin Paihol 1985: 38: 1225-1 22s.
12
N. R. LEMOINE €7 AL.
34. Sdinsbury JRC. Farndon JR. Needham GK, Malcolm AJ. Harris AL. Epidermal-growth-factor receptor status as predictor of early recurrenceofanddeath from breastcancer. Lancet 1987;i: 1398-1402. 35. Harris AL. Nicholson S. Sainsbury JRC. Farndon J, Wright C. Epidermal growth factor receptors in breast cancer: association with early relapse and death, poor response to hormones and interactions with neu. JSteroidBiochem 1989: 34: 123-131. 36. Nicholson S. Sainsbury JRC. Needham GK, Chambers P, Farndon JR, Harris AL. Quantitative assays ofepidermal growth factor receptor in human breast cancer: cut-off points of clinical relevance. h i J Cancer 1988: 4 2 3 6 4 1 . 37. Rios MA. Macias A. Perez R. Lage A. Skoog L Receptors for epidermal growth factor and estrogen as predictors of relapse in patients with mammary carcinoma. Anticancer Res 1988: 8: 173-176.
38. Foekens JA, Portengen H. van Putten WLJ. et al. Prognostic value of receptors for insulin-like growth factor 1. somatostatin and epidermal growth factor in human breast cancer. Cancer Res 198:4 9 7002-7009. 39. Nicholson S. Richard J. Sainsbury C. et a / . Epidermal growth factor receptor (EGFr); results of a 6 year follow-up study in operable breast cancer with emphasis on the node negative subgroup. Br J Cancer 1991:63 146-150. 40. Gullick WJ. Inhibitors of growth factor receptors. In: Carney D. Sikora K. cds. Genes and Cancer. Chichestcr: John Wiley, 1990; 263-273.
166: 7-12 (1992)
THE EPIDERMAL GROWTH FACTOR RECEPTOR IN HUMAN PANCREATIC CANCER NICHOLAS R. LEMOINE*, CHRISTINE M. HUGHEST, CLAIRE M. BARTON*, RICHARD POULSOM;, ROSEMARY E. JEFFERY:, GUNTER KLOPPEL~, PETER A. HALL)/ AND WILLIAM J. GULLICK?
*Molecular Pathology Laboratory and tkfolecular Oncology Laboratory, ICRF Oncology Group, Hammersmith Hospital, Du Cane Road, London W12 OHS, U.K.; :ICRF Histopathology Unit, Lincoln's Inn Fields, London WCZA 3PX, U.K.; §Department of Pathology, Academic Hospital Jette, Free University of Brussels, Laarbeeklaan 101. B-1090 Brussels, Belgium; /I Department of Histopathology, Royal Postgraduate Medical School, Hammersmith Hospital, Du Cane Road, London WI2 OHS, U.K. Received I3 May 1991 Accepted2 July I991
SUMMARY The epidermal growth factor receptor (EGFR) and its ligands are thought to be important in the control of proliferation of many epithelial systems, including the exocrine pancreas. Abnormalities in expression of two of the known ligands ofthe EGFR, transforming growth factor alpha and epidermal growth factor, occur frequently in ductal adenocarcinoma of the human pancreas. We have examined an archival series of cases of pancreatic pathology for expression of the EGFR using the anti-EGFR antiserum 12E and found that there is almost ubiquitous overexpression of EGFR in pancreatic cancer and in chronic pancreatitis. Southern blot analysis showed no evidence of amplification or rearrangement of the EGFR gene. We conclude that an autocrine loop involving the EGFR system may be involved in the genesis of both neoplasia and reactive hyperplasia of pancreatic ductal epithelium. KEY WORDS-Epidermal
growth factor receptor, pancreatic cancer, autocrine loop, immunohistochemistry.
INTRODUCTION Pancreatic cancer is a major clinical problem and reports from several groups have recently shed light on the molecular events involved in its pathogenesis. Mutation of the Kirsten rus gene at codon 12 is present in about 75 per cent of advanced pancreatic cancers (reviewed by Shibata et ul.' and Lemoine and Hall') and represents an early event, occurring in a similar proportion of ductal carcinoma in s i m 3 The c-erbB-2 proto-oncogene is overexpressed in at least 20 per cent of pancreatic c a n c e r ~and , ~ this may be associated with amplification and/or rearrangement of this gene in some cases ( C . M. Barton. unpublished results). The gene encoding the epidermal growth factor receptor (EGFR) is closely related to the c-erbB-2 proto-oncogene and we have recently reported that at least two of the ligands of the EGFR, transforming growth factor-alpha Addressee for correspondence: Dr Nicholas Lemoine. Molecular Pathology Laboratory. ICRF Oncology Group. Hammersmith Hospital. Du Cane Road. London W12 OHS.
U.K.
0022-3417192iO 1000746 $05.00 0 1992 by John Wiley & Sons, Ltd.
(TGF-alpha) and EGF/urogastrone, are frequently overexpressed in pancreatic carcinoma.' Autocrine stimulation of tumour cells could occur if the receptor is expressed in the same cells, since expression of TGF-alpha or EGF/urogastrone is mitogenic only in cell lines possessing EGF receptors.6 A low level of EGFR expression has been reported in normal pancreatic cells' and simultaneous overexpression of EGF and EFGR, and of TGF-alpha and EGFR, has been detected in pancreatic cancer cell lines, supporting this hyp~thesis.'.~ In a small series of frozen samples, EGFR was found to be overexpressed in 30-50 per cent of cases of pancreatic cancer." One group has correlated EGFR overexpression with alterations in chromosome 7." A new antipeptide polyclonal antiserum (1 2E) that recognizes a carboxy-terminal domain of the EGF receptor" in routinely-fixed, paraffinembedded tissue has very recently become available, and this stimulated us to examine a retrospective series of 84 pancreatic cancers in which we correlate expression of E G F receptor with expression of its ligands TGF-alpha and EGFjurogastrone. In a
8
N. R. LEMOINE ET AL.
series of 23 snap-frozen specimens of pancreatic cancer, we could also use the well-characterized anti-EGFR antibody EFGR1,I3.l4allowing us to validate our results with the new 12E antibody.
phosphate-buffered saline (PBS) for 5 min and then rinsing in distilled water. The immunohistochemical technique involved the sequential application of the following reagents: the primary antibody (12E, 14E, or EGFRI) in PBS containing 0.5 per cent bovine serum albumin (BSA) at the concentration stated MATERIALS AND METHODS above for 1 h at room temperature; a secondary biotinylated anti-rabbit (for 12E and 14E) or antiTumour material analysed mouse (for EGFRI) antibody (Dako) diluted 1500 Formalin-fixed, paraffin-embedded blocks of or 1:300, respectively, in PBS containing 0.5 per cent pancreatic adenocarcinoma ( n = 84), chronic pan- BSA for 30 mini and ABComplex/HRP (Dako) for creatitis (n = 14), and normal adult pancreas ( n= 10) 30 min. Each incubation was followed by rinsing the were obtained from the archives of Hammersmith tissue sections in PBS three times (5 min each). The Hospital. This series was previously examined for sites of immunocomplex were identified by light microscopy following treatment with a chromogen, expression of TGF-alpha and EGF/urogastrone.’ We also examined a series of 23 pancreatic 3,3’-diaminobenzidine tetrahydrochloride (Northcancers which had been snap-frozen in liquid nitro- east Biochem). The specificity of the two antipeptide antibodies was tested by inhibition of binding after gen at the time of surgery and stored at -70°C. These were collected at the Hammersmith Hospital pre-absorption of antibody with the immunizing and the Middlesex Hospital, London; at the Dudley peptide, and retention of staining after preabsorpRoad Hospital, Birmingham; and at the Academic tion with other non-specific peptides. Overexpression of the EGF receptor in this study is Hospital Jette, Free University of Brussels. Sections stained with haematoxylin and eosin defined as staining of cell membranes, often were examined and cases of pancreatic cancer accompanied by cytoplasmic staining. classified according to standard criteria.” Hybridization analysis of EGFR Immunohis tochemical detection of expression of Southern blot analysis of the EGFR gene was EGFR performed on eight pancreatic carcinoma cell To detect EGF receptor immunoreactivity in lines (BxPc3, ASPCI, PSN1, MiaPaCa2, Pancl, frozen and paraffin sections, an affinity-purified CaPan1, and CaPanZ, obtained from ATCC) and preparation of the rabbit antibody 12E (raised six primary pancreatic tumour biopsies with against a synthetic peptide representing residues paired normal DNA from the same patients using 1059-1072 in the cytoplasmic domain of the E G F previously described protocols.” In situ hybridization analysis of EGFR mRNA receptorI6) was used at a concentration of 4 pg/ml, was performed and controlled as described preas well as an affinity-purified preparation of the rabbit antibody 14E (raised against a synthetic peptide viously,18 using an EGFR riboprobe from of residues 1164-1176 of the EGF receptor16) at EcoRI-linearized phEGFR.4a, which we prepared 4 pg/ml. The monoclonal antibody EGFR 1 (which by subcloning the 913 bp Sac?-Hind?IIfragment of ’ ~ pGEM-3 (Promega). was raised against A431 cells and recognizes the ~ 6 2 . 1into native folded external domain of the EGF receptor;I3 it has been extensively characterized for use in immuRESULTS nohistocherni~try’~) was used a concentration of 10 pg/ml to detect EGF receptor immunoreactivity Normal Pancreas and chronicpancreatitis in frozen sections only. Only the islets of Langerhans showed EGFR Paraffin sections were dewaxed by passage through xylene and rehydrated through alcohols expression in histologically normal pancreas, and to water. Endogenous peroxidase activity was this was very weak. However, in chronic pancreatitis quenched by incubation in 0.3 per cent hydrogen (even if quite mild) the epithelial cells of the duct peroxide for 30 min and then rinsing in tap system showed strong cytoplasmic EGFR immuwater. For frozen sections, endogenous peroxidase noreactivity (Fig. 1); this affected the cells of the activity was quenched by incubation in 0.1 per small ducts (including the centroacinar cells) most cent phenylhydrazine hydrochloride (Sigma) in markedly but also involved the largest ducts in the
9
EGF RECEPTOR IN PANCREATIC CANCER
head of the pancreas. in such pancreatitis, the EGFR immunoreactivity of islet cells was accentuated, but the acinar cells and interstitium did not express detectable EGFR. This distribution of EGFR immunoreactivity is concordant with that of the ligand TGF-alpha previously demonstrated in these tissues.
cancer. Almost all cases showed coordinate expression of both EGFR and its ligand TGFalpha, but there were a few exceptions. One case was negative for EGF receptor expression but showed immunoreactivity for TGF-alpha, albeit relatively weak. The only case of clear cell carcinoma examined showed membrane immunoreactivity for EGFR, but was negative for TGF-alpha. All other cases were concordant for expression of both EGF receptor and TGF-alpha. Table I-Immunoreactivity for EGFR, TGF-alpha, and EGF in pancreatic adenocarcinoma
Tumour type Ductal carcinoma Grade 1 Grade 2 Grade 3
n
Irnmunoreactivi ty (cases positive) E G F R TGF-alpha E G F
30
28
29
4
31
30
30
19
18
18
4 2
Other exocrine pancreatic carcinomas 1 1 Clear cell Large cell 2 2 Signet ring 1 1 Total Fig. 1-EGF receptor immunoreactivityin chronic pancreatitis. There is specific staining ofthe epithelialcells of the ducts, including the centroacinar cells. Acinar cells and interstitiurn do not express detectable EGF receptor
Exocrine pancreatic cancer Eighty of the 84 cases (95 per cent) of exocrine pancreatic adenocarcinoma showed immunoreactivity for the EGF receptor with both anti-EGFR antibodies 12E and 14E. The immunoreactivity was mainly cytoplasmic but with accentuation at the plasma membrane of tumour cells. Although all malignant cells within a positive tumour showed immunoreactivity, there was some degree of heterogeneity with some cells being stained very densely and others relatively modestly. There was no apparent relationship to tumour grade or histological type (see Table I which presents the results of the present study together with those of our previous study’ for comparison). Figures 2 and 3 show the typical features of EGF receptor expression in pancreatic
84
80
0 2 1
80
0 0 0 10
As previously reported, immunoreactivity for EGF/urogastrone was detectable in 10 of the 84 cases (12 per cent), and all of these cases were also positive for TGF-alpha immunoreactivity and EGFR immunoreactivity In the frozen tissue sections of pancreatic carcinoma, EGF receptor expression was recognized by both the EGFRl monoclonal antibody and the 12E polyclonal antibody with the same sensitivity. All except one case (22123 cases) showed cytoplasmic immunoreactivity with these antibodies, and in most cases membrane staining was also seen. Southern blot analysis of the eight pancreatic cell lines and the six pancreatic tumour biopsies showed no evidence of rearrangement or amplification of the EGFR gene (data not shown). Hybridization in siru of EGFR mRNA detected no specific signal in any of the cases of pancreatic cancer examined. although a strong specific signal was present in the control gastric cancer material in which overexpression of EGFR was known to be very high.” Control hybridization with a specific riboprobe”
10
N.R. LEMOINE ET AL.
detected the presence of p-actin mRNA in the pancreatic cancers, indicating that the failure to detect EGFR expression in these cases was not the result of nucleic acid degradation. The amount of EGFR mNRA present is therefore below the detection limit of this in siru hybridization system, although EGFR protein is easily detectable by immunohistochemistry. DISCUSSION
Fig. 2-EGF receptor immunoreactivity in a pancreatic exocrine adenocarcinoma of intermediate grade (grade 2). There is heterogeneous staining of tumour cells and accentuation at the Cell membranes
Fig. 5-EGF receptor immunoreactivity in a poorly differentiated (grade 3) pancreatic adenocarcinoma. Many ofthe tumour cells show very strong cytoplasmic immunoreactivity
We have previously demonstrated that there is frequent overexpression of ligands for the EGF receptor in pathological conditions of the pancreas, both in neoplasia and in reactive hyperpla~ia.~ This study shows that in pathological conditions (pancreatitis and neoplasia) the duct epithelial cells express high levels of the EGF receptor. This is further support for the autocrine stimulation hypothesis in pancreatic pathology, since conditional expression of E G F receptors by pancreatic ductal epithelial cells may allow them to respond to TGF-alpha and EGF/urogastrone to facilitate repair and repopulation in pancreatitis, and tribute to proliferation and population expansion in neoplasia. The mechanisms of control of EGFR gene expression are currently an area of intensive study. Both positively acting transcription factors (such as the ubiquitous factor Spl as well as factors TCF, ETFl , and ETF2”-”) and at least one inhibitory transcription factor (factor GCFZ6)have been shown to interact with the EGFR promoter. The increased expression of EGFR in the pancreatic tumours is presumably due to increased transcription, since our Southern blot analysis showed no evidence for gene amplification. In view of our finding of increased expression of EGFR in ductal epithelial cells involved in chronic pancreatitis, it will be interesting to see how the activity of these various transcription factors is influenced by lymphokines which might mediate the inflammatory process. Many studies on other tumour types have shown that overexpression of the EGF receptor is associated with poor clinical prognosis. These include human gliomas in which there is much higher EGFR expression in tumours of aggressive histological type.”-3’ In breast cancer, high expression of EGFR is associated with short relapse-free interval and s u r ~ i v a lincluding ~ ~ ~ ~ * those with node-negative disease.39In our analysis, there was no difference in frequency of EGFR overexpressionin pancreatic tumours of different histological grades. EGFR
EGF RECEPTOR I N PANCREATIC CANCER
expression is not a useful prognostic marker in pancreatic cancer because essentially all cases have a poor prognosis and also usually overexpress thisprotein. Nor is the detection of EGFR overexpression likely to be of any value as a diagnostic marker, because overexpression also occurs at high frequency in the duct system of pancreata affected by chronic pancreatitis. However, for our understanding of the molecular basis of this disease it is important to assemble a full picture of the oncogenes and growth factor systems that are activated in order that we might identify new targets for therapy. In this regard, growth factor receptors are currently the focus of much activity, and many potential mechanisms for inhibition are being e ~ p l o r e d . ~ ' ACKNOWLEDGEMENTS
We gratefully acknowledge the support of the Imperial Cancer Research Fund. and thank Professor RCN Williamson, Mr RCG Russell and Mr J Neoptolemos for providing clinical material and encouragement. REFERENCES 1 Shibata D. Capella G. Perucho M. Mutational activation of the C-Kras gene in human pancreatic cancer. Bailliere's Clin Gasrroenrerol 1990:4 151-169. 2. Lemoine NR. H a l l PA. Oncogenes and growth factors in pancreatic cancer. Boilliere > C/in Gaslrornrerol 1990: 4 8 15-832. 3. Lemoine NR. Jain S. Hughes CM. el a/. Ki-ras oncogene activation in
4.
5. 6. 7. 8.
9.
10.
I I.
I2
invasivc ductal adcnocarcinoma. but not in ductal papillary hyperplasiaor intraductal papillaryneoplasiaofthepancreas.Gastroenterology (in press). Hall PA. Hughes CM. Staddon SL. Richman PI. Gullick WJ. Lemoine NR. The c-erb5-2 proto-oncogene in human pancreatic cancer. J Pa~hol1990:161: 195-200. Barton CM. Hall PA. Hughes CM. Gullick WJ. Lemoine NR. Transforming growth factor alpha and epidermal growth factor in human pancreaticcancer. JPatho/1991: 163: 111-1 16. Rosenthal A. Lundquist PB. Bringman TS. Goeddel DV. Derynck R. Expression in rat fibroblasts of a human transforming growth faCtOr-alpha cDKA results in transformation. Cell 1986~46:301-309. Damjanov I. Mildner B. Knowlcs BB. Immunohistochemical localisation of the epidermal growth factor receptor in normal human tissue. Lab Invesr 1986: 5 5 588-592. Chen YF. Pan G-Z. Hou X, er 01. Epidermal growth factor and its receptor in human pancreatic carcinoma. Pancreas 1990: 5: 278-283. Smith JJ. Derynck R. Korc M. Production of transforming growth factor-alpha in human pancreatic cancer cells: evidence for a superagonist autocrine cycle. Proc Null Acad Sci USA 1987: 84: 7567-7570. Kloppel G. Maillet B. Schewc K. Kalthoff H. Schmeigel WH. Immunocytochemical detection of epidermal growth factor receptors (EGFR) and transferrin receptor (TR) on normal. inflamed and neoplastic tissue. Pancreas 1989: 4 649 (abstract). Korc M . Meltzrr P. Trent J. Enhanced expression of epidermal grou,th factor receptor correlates with alterations of chromosome 7 in human pancreatic cancer. Proc. Null A r a d Sci GSA 1986: 8 3 5 14 1-5 144 Gullick WJ. Hughes CM. Mellon K. Neal DE. Lemoine NR. Immunohistochemical detection of thc epidermal growth factor receptor in paraffin-embedded human tissues. J Parhol164: 281-289.
11
13. Waterfield MD. Mayes ELV. Stroobant P. et al. A monoclonal antibody to the human rpidermal growth factor reseptor. J. CeN5iotltem 198220: 149-161. 14. Gullick WJ. Marsden JJ. Whittle N, er a/. Expression of epidermal growth factor receptors on human cervical. ovarian and vulva1 carcinomas. Cancer Rrs 1986: 46: 285-292. 15. Kloppel G. Pancreatic non-endocrine tumours. In: Kloppel G. Heitz PU. eds. Pancreatic Pathology. Edinburgh: Churchill Livingstone. 1981:79-113. 16. Gullick WJ. Downward J, Waterfield MD. Antibodies to the autophosphorylation sitesofthe epidermal growth factor receptor proteintyrosine kinease as probes of structure and function. E M 5 0 J 1985:4 2869-2877. 17. LemoineNR. Wyllie FS. LillehaugJR.etol.Absenceofabnormalities of the c-erbB-l and c-erbB-2 proto-oncogenes in human thyroid neoplasia. Eur J Cancer 1990: 2 6 777-779. 18. Wright NA. Poulsom R. Stamp GWH. er al. Epidermal growth factor (EGFIL'RO) induces expression of regulatory peptides in damaged human gastrointestinal tissues. J Pathol1990: 1 6 2 279-284. 19. Ullrich A. Coussens L. Hayflick JS. et al. Human epidermal growth factor receptor cDNA sequence and aberrant expression of the amplified gene in A431 epidcrmoid carcinoma cells. lvature 1984; 309: 418325. 20. Lcmoine NR. Jain S . Silvcstrc F. er al. Amplification and overexpression of the EGF receptor and c-erbB-2 proto-oncogenes in human stomach cancer. 5 r JCancer64: 79-83. 21. Porte P. Gunning P. Blau R. Kedes L. Human actin genes are single copy for alpha-skeletal and alpha-cardiac actin but multicopy for beta- and gamma-cytoskclctal genes: 3-untranslated regions arc isotopespecific but arc conscrved in evolution. Mol CeN Bio/ 1983: 3: 1783-1791. 22. Johnson AC. Ishii S. Jinno Y. Pastan 1. Mcrlino GT. Epidermal growth factor receptor gene promoter: deletion analysis and identificarion of nuclcar protein binding sites. J 5101 Chem 1988: 263: 5693-5699 23. Johnson AC. Jinno Y. Merlino GT. Modulation of epidermal growth factor receptor proto-oncogene transcription by a promoter sitesensitive to S1 nuclease. MolCel/BioI 1988:8:41744184. 24. Kageyama R. Merlino GT. Pastan I. Epidermal growth factor (EGF) receptor gene transcription: requirement for SPI and an EGF receptor factor. JBioI Chem 1988: 2 6 3 6329-6336. 25. Kageyama R. Merlino GT. Pastan 1. iiuclear factor ETF spccifically stimulates transcription from promoters without a TATA box. J Bio/ Chem 1989: 264: 15508-15514. 26. Kageyama R. Pasran I. Molecular cloning and characterization of a human DNA binding factor that represses transcription. Cell 1989: 59815-8?5.
27. Merlino GT. Epidermal growth factor receptor regulation and function. Semin Cancer B i d 1990: I: 227-284. 28. Libermann TA. Razon X . Barial AD. el a/. Expression of epidermal growth factor receptors in human brain tumors. Cancer Res 1984: 44: 753-760. 29. Libermann TA. Nusbaum HR. Razon N. er a/. Amplification. enhanced expression and possible rearrangement of EGF receptor eenc in primary human brain turnours of glial origin. Nature 1985: 513: 144-147. 30. Wong AJ. Bigner SH. Bigner DD. et a/. Increased expression of thc epidermal growth factor rcceptor gene in malignant gliomas is invariably associated with gene amplification. Proc f l u / / Acad Sci U S A 1987: 84: 689945903. 31. Hawkins RA. Killen E. Whittle IR. Jack WJL. Chetty 6. Prescotl RJ. Epidermal growth factor reccptors in intracranial and breast tumours: their clinical significance. 5 r JCancer 1991: 6 3 553-560. 32. Ekstrand AJ. James CD. Cavence WK. Seliger B. Pcttersson RF. Collins VP. Genes for epidermal growth factor receptor. transforming growth factoralpha. and epidermalgrowth factorand theirexpression i n humangliomasin uiro. ConcerRes 1991:51:2164-2172. 33. Sainsbury JRC. Malcolm AJ. Appleton DR. Farndon JR. Harris AL. Presence of epidermal growth factor receptor as an indicator of poor prognosis in paticnts with breast cancer. J Clin Paihol 1985: 38: 1225-1 22s.
12
N. R. LEMOINE €7 AL.
34. Sdinsbury JRC. Farndon JR. Needham GK, Malcolm AJ. Harris AL. Epidermal-growth-factor receptor status as predictor of early recurrenceofanddeath from breastcancer. Lancet 1987;i: 1398-1402. 35. Harris AL. Nicholson S. Sainsbury JRC. Farndon J, Wright C. Epidermal growth factor receptors in breast cancer: association with early relapse and death, poor response to hormones and interactions with neu. JSteroidBiochem 1989: 34: 123-131. 36. Nicholson S. Sainsbury JRC. Needham GK, Chambers P, Farndon JR, Harris AL. Quantitative assays ofepidermal growth factor receptor in human breast cancer: cut-off points of clinical relevance. h i J Cancer 1988: 4 2 3 6 4 1 . 37. Rios MA. Macias A. Perez R. Lage A. Skoog L Receptors for epidermal growth factor and estrogen as predictors of relapse in patients with mammary carcinoma. Anticancer Res 1988: 8: 173-176.
38. Foekens JA, Portengen H. van Putten WLJ. et al. Prognostic value of receptors for insulin-like growth factor 1. somatostatin and epidermal growth factor in human breast cancer. Cancer Res 198:4 9 7002-7009. 39. Nicholson S. Richard J. Sainsbury C. et a / . Epidermal growth factor receptor (EGFr); results of a 6 year follow-up study in operable breast cancer with emphasis on the node negative subgroup. Br J Cancer 1991:63 146-150. 40. Gullick WJ. Inhibitors of growth factor receptors. In: Carney D. Sikora K. cds. Genes and Cancer. Chichestcr: John Wiley, 1990; 263-273.
