Overexpression of Epidermal Growth Factor and Hepatocyte Growth Factor Receptors in a Proportion of Gastrinomas Correlates with Aggressive Growth and Lower Curability
Total Page:16
File Type:pdf, Size:1020Kb
Vol. 8, 2273–2285, July 2002 Clinical Cancer Research 2273 Overexpression of Epidermal Growth Factor and Hepatocyte Growth Factor Receptors in a Proportion of Gastrinomas Correlates with Aggressive Growth and Lower Curability Paolo L. Peghini, Michiko Iwamoto, Furthermore, each is overexpressed in a minority (15–20%) Mark Raffeld, Yuan-Jia Chen, of the gastrinomas, and the overexpression correlates with Stephan U. Goebel, Jose Serrano, and aggressive growth and lower curability. Robert T. Jensen1 INTRODUCTION Digestive Diseases Branch, National Institute of Diabetes and 2 Digestive and Kidney Diseases [P. L. P., M. I., S. U. G., J. S., R. T. J.] NETs (PETs and carcinoids) are generally slow-growing and Hematopathology Section, Laboratory of Pathology, National neoplasms; however, a subset, the proportion of which varies Cancer Institute [M. R.], NIH, Bethesda, Maryland 20892 with different NET tumors, demonstrates an aggressive growth pattern with a markedly decreased survival (1–4). The most common malignant, symptomatic PET (2, 5) is gastrinoma, and ABSTRACT in 25% of patients, the gastrinoma has an aggressive growth Purpose: Growth factor receptor expression and acti- pattern with the development of liver metastases and a 10-year vation, particularly for epidermal growth factor (EGF) and survival of 30% (6). In contrast, the 10-year survival rate is hepatocyte growth factor (HGF), in many endocrine and Ͼ95% in the 75% of patients in which the gastrinomas show nonendocrine tumors is important in determining tumor indolent growth (6). Furthermore, NETs in different patients recurrence, growth, and aggressiveness. Whether this is true vary considerably in their surgical curability (1, 2, 4). For of neuroendocrine tumors such as gastrinomas is unclear. example, only 30–40% of patients with gastrinomas after gas- Experimental Design: To address this question, we an- trinoma resection remain disease free for the long term (7). The alyzed the extent of EGFR and HGFR expression in gastri- factors determining this variable growth pattern, curability, and nomas from 38 patients with Zollinger-Ellison syndrome tumor aggressiveness with gastrinomas as well as other PETs and correlated it with clinical and tumor characteristics. are largely unknown (8). This has occurred because the molec- mRNA levels were ular pathogenesis of NETs is largely unknown and appears to (22 ؍ and HGFR (n (38 ؍ EGFR (n determined by competitive PCR, and immunohistochemis- differ significantly from other nonendocrine malignant tumors try was performed on a subset. in which the molecular pathogenesis has been well studied (i.e., Results: In each of the gastrinomas studied, detectable colon cancer, breast cancer, and pancreatic cancer; Refs. 8–11). levels of EGFR and HGFR mRNA were present. Low levels Alterations of common oncogenes (ras, myc, and others) or of EGFR protein expression were detected in 40% of gas- alterations in common tumor suppressor genes (p53 and retino- trinomas and HGFR protein expression in 90%. EGFR blastoma gene), which are frequent in many common nonendo- mRNA expression varied by 1050-fold and HGFR by 375- crine malignancies, are uncommon in typical PETs and carci- fold. Eighteen percent of gastrinomas overexpressed EGFR noid tumors (8–11). Alterations in the MEN1 tumor suppressor mRNA and 14% overexpressed HGFR mRNA, compared gene on chromosome 11q13 occur in 16–42% of sporadic with normal pancreas. Maximal EGFR and HGFR mRNA gastrointestinal NETs (8, 12, 13), and inactivation of the levels were 4- and 1.2-fold increased and correlated with the p16INKa/CDKN2A tumor suppressor gene on chromosome 9p21 -and decreased has been reported in 17–92% of various PETs (14–16). How (0.034 ؍ presence of liver metastases (P -but not tumor location, size, ever, the presence of p16INK4a/CDKN2A or MEN1 gene muta (0.027 ؍ long-term curability (P or tumor functional characteristics. tions in the subset of sporadic tumors has not been shown to Conclusions: These above results indicate that EGFR correlate with the aggressiveness of the NET (12, 15). and HGFR mRNA are universally expressed in gastrinomas. Many studies on different nonendocrine and endocrine tumors show that the expression of various growth factor recep- tors including those for vascular endothelial growth factor, platelet-derived growth factor, EGF, insulin-like growth factor, Received 12/12/01; revised 3/11/02; accepted 4/26/02. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 2 The abbreviations used are: NET, neuroendocrine tumor; PET, pan- 1 To whom requests for reprints should be addressed, at NIH/National creatic endocrine tumor; EGF, epidermal growth factor; EGFR, EGF Institute of Diabetes and Digestive and Kidney Diseases/Digestive Dis- receptor; HGF, hepatocyte growth factor; HGFR, HGF receptor; BAO, eases Branch, Building 10, Room 9C-103, 10 Center Drive, MSC 1804, basal acid output; MAO, maximal acid output; HPRT, human hypox- Bethesda, MD 20892-1804. Phone: (301) 496-4201; Fax: (301) 402- anthine phosphoribosyltransferase; MEN1, multiple endocrine neoplasia 0600. type 1; RT-PCR, reverse transcription-PCR. Downloaded from clincancerres.aacrjournals.org on September 26, 2021. © 2002 American Association for Cancer Research. 2274 EGFR and HGFR in Gastrinomas and HGF likely play an important role in the molecular patho- imaging studies provided the basis for assessment of tumor genesis of the tumor and the invasiveness and aggressiveness of growth, recurrence, or progression (7). Consistent absence of the tumors (17, 18). The EGF and HGF families of receptors and imaged lesions or lack of increase in size or number of lesions their ligands are likely particularly important in a number of over the follow-up period was defined as a tumor demonstrating neoplasms causing autocrine growth associated with increased no growth. An increase in size or number of lesions was defined invasiveness, and overexpression by the tumor is associated as evidence of tumor growth as described previously (26). The with decreased survival/poor prognosis (17–22). Little is known development of liver metastases identified by imaging studies about the expression of EGFR or HGFR in NETs and whether during follow-up served as the definition for the liver metastases the extent of expression correlates with aggressive behavior or group (26, 27). Liver metastases were confirmed by percutane- advanced disease. Amplification of the HER-2/neu proto-onco- ous liver biopsy. No patient received antitumor treatment gene gene, a member of the EGFR family of receptors, was (chemotherapy, ␣-IFN, or somatostatin analogues) before the reported in a small number of cases (23) of NETs (9); however, surgical resection or biopsy; however, after biopsy, 7 patients it was unclear whether this was associated with increased ex- received antitumor treatment for progressive metastatic disease pression or correlated with tumor aggressiveness. (3, chemotherapy; 5, ␣-IFN; 7, octreotide). The purpose of the present study, therefore, was to deter- Tumor cell lines were obtained from the American Type mine whether EGFR or HGFR overexpression occurred in a Tissue Collection (Manassas, VA). Two breast cancer cell lines well-studied group of patients with gastrinomas and to correlate which are reported to overexpress EGFR [MDA-MB-468 (28, the extent of expression with clinical and tumor characteristics, 29), BT-20 (28, 29)], the epidermoid cancer cell line A-431 including primary gastrinoma size, location, tumor extent, and (30), and the pancreatic cell line PANC1 (31) were used as curability. positive controls in the EGFR quantitative PCR assay. The pancreatic cancer cell lines, CAPAN-2 and BxPc3, which are MATERIALS AND METHODS reported to overexpress the HGFR (32), were used as positive controls for the HGFR quantitative PCR assay. All cancer cells Patients and Tumors. Thirty-eight patients who under- were grown in DMEM with 10% calf serum, except BT-20 went exploratory laparotomy for Zollinger-Ellison syndrome at breast cancer cells were grown in Eagle’s minimum essential the NIH between 1990 and 1999 were included in this study. medium, CAPAN-2 in McCoy’s 5A medium, and BxPc-3 in The study protocol was approved by the Clinical Research RPMI 1640, all with 10% FBS. Total RNA or mRNA from three Committee of the National Institute of Diabetes and Digestive and Kidney Diseases, and all patients gave informed consent. normal pancreases were obtained from commercial sources Preoperative measurement of BAO and MAO were performed [Clontech (Palo Alto, CA) and Invitrogen (Carlsbad, CA)]. as described previously (24). Preoperative serum gastrin levels Competitive PCR. Tumor samples were immediately were analyzed by RIA by Bioscience Laboratories (New York, snap frozen in liquid nitrogen either during surgery or after Ϫ NY) or Mayo Clinic Laboratories (Rochester, MN). The diag- harvesting and stored at 70°C. The histological diagnosis of nosis of Zollinger-Ellison syndrome was established as reported gastrinoma was made by using standard histological and immu- previously (24). The duration of disease was defined by the nohistological stains (chromogranin A and gastrin). Tumor clinical history from the time of diagnosis or from the time of mRNA was extracted from 8-m cryosections of the gastrino- disease onset (6). Detailed conventional imaging studies (com- mas after