Ó Springer-Verlag 2001

J Cancer Res Clin Oncol (2001) 127: 449±454

ORIGINAL PAPER

Matthias P. A. Ebert á Sara Hernberg á Guo Fei Armin Sokolowski á Hans U. Schulz Hans Lippert á Peter Malfertheiner

Induction and expression of cyclin D3 in human pancreatic cancer

Received: 24 February 2000 / Accepted: 29 December 2000

Abstract Purpose: Cyclins play a key role in the control and regulation of the cell cycle. The role of cyclins in the pathogenesis of pancreatic cancer is largely unknown. Methods: Using Northern blot analysis, polymerase chain reaction (PCR) and immunohistochemistry, we examined the expression of cyclins D1, D2, and D3 in human pancreatic cancer and studied the induction of these cyclins by growth factors in pancreatic cancer cell lines. Results: We now report that cyclin D1 and D3 mRNAs are expressed in human pancreatic cancer cell lines, and that the expression of cyclin D3 is enhanced in pancreatic cancer cells by amphiregulin, a member of the epidermal growth factor family. Cyclins D1 and D3 are also expressed in normal and malignant pancreatic tissues. However, while the normal pancreas and pancreatic cancers express cyclin D2 as determined by reverse-transcriptase PCR, we could not detect cyclin D2 mRNA by either Northern blot analysis or reverse transcriptase PCR in the two pancreatic cancer cell lines. Immunohistochemical analysis revealed the expression of cyclin D3 in pancreatic cancer cells. Conclusions: These ®ndings suggest that D-type cyclins are di€erentially expressed in pancreatic cancer and that the aberrant activation of the EGF receptor in human pancreatic cancer by amphiregulin may lead to the progression of

M. P. A. Ebert (&) á S. Hernberg á G. Fei á P. Malfertheiner Department of Gastroenterology, Hepatology and Infectious Diseases, Otto-von-Guericke University of Magdeburg, 39120 Magdeburg, Germany Tel.: +49-391-6713156; Fax: +49-391-67190054 e-mail: [email protected] A. Sokolowski Institute of Clinical Chemistry and Pathobiochemistry, Otto-von-Guericke University of Magdeburg, 39120 Magdeburg, Germany H. U. Schulz á H. Lippert Department of General Surgery, Otto-von-Guericke University of Magdeburg, 39120 Magdeburg, Germany

the cell cycle via induction of cyclin D3 expression, thus contributing to the growth advantage of these transformed cells. Key words Cell cycle á Cyclin á Growth factor á Cancer á Pancreas

Introduction D-type cyclins comprise a family of three members, cyclin D1 being identi®ed as the putative proto-oncogene PRAD1 which is located on chromosome 11q13. As the cell cycle is controlled primarily during the G1 phase, Dtype cyclins and their cyclin-dependent kinases are required for the regulation and progression of the cell cycle (Hunter and Pines 1994). Cyclin D1 associates primarily with cyclin-dependent kinases 4 and 6, and thus leads to the phosphorylation of the retinoblastoma protein (Hunter and Pines 1994; Grana and Reddy 1995). Experimental studies involving the selective overexpression of cyclin D1 in transgenic mice resulted in mammary hyperplasia and carcinomas in the lactating mouse (Mueller et al. 1997). Furthermore, cyclin D1 has recently been demonstrated to be overexpressed in a number of malignancies, including pancreatic carcinoma and esophageal cancer (Gansauge et al. 1997; Hinds et al. 1994). Moreover, ampli®cation of CCND1/ PRAD1 has been reported in leukemia and breast cancer, while the overexpression of cyclin D1 in human malignancies is associated with advanced disease and poor prognosis (Gansauge et al. 1997; Hinds et al. 1994). Based on independent analysis, several groups have reported that the expression of D-type cyclins is highly growth-factor-dependent, e.g., in human ®broblasts cyclin D1 mRNA levels are induced by platelet-derived growth factor (Surmacz et al. 1992). The second member of the family of D-type cyclins is cyclin D2 which is encoded by the CCND2 gene and which is located on chromosome 12p13 (Hunter and Pines 1994; Grana and Reddy 1995). In a small series of colorectal cancers,

450

ampli®cation of cyclin D2 has been reported (Leach et al. 1993), and ampli®cation of the cyclin D3 gene has been reported in a glioblastoma cell line (Hunter and Pines 1994; Wang et al. 1995; Kuchiki et al. 2000). Pancreatic carcinoma is a devastating disease with poor prognosis (Warshaw and Fernandez-Del Castillo 1992). Recently, a number of genetic and molecular alterations have been reported, such as K-ras gene mutation, deletion of DPC-4 and p53 gene mutation (Hahn et al. 1996; Berrozpe et al. 1994). In addition, human pancreatic carcinomas overexpress a number of growth factors, e.g., epidermal growth factor and amphiregulin of the EGF family and platelet-derived growth factor of the PDGF family (Ebert et al. 1994, 1995; Yokoyama et al. 1995). While the K-ras gene mutation is considered to be an early event in pancreatic carcinogenesis, its presence is not associated with poor prognosis (Berrozpe et al. 1994). In contrast, the overexpression of growth factors is often correlated with poor prognosis, pointing to a role of growth factor-dependent autocrine and paracrine loops in the pathogenesis of this disease (Ebert et al. 1994, 1995; Yokoyama et al. 1995). However, while amphiregulin and transforming growth factor-alpha, which are aberrantly expressed in human pancreatic ductal carcinomas, both stimulate the growth of pancreatic cancer cells in vitro (Ebert et al. 1994; Yokoyama et al. 1995), the overexpression of platelet-derived growth factors in pancreatic carcinoma is not associated with the stimulation of pancreatic cancer cell growth by PDGF in vitro (Ebert et al. 1995). In an e€ort to assess the expression of cyclins D1, D2, and D3 in pancreatic cancer and to further elucidate the pathways which are associated with amphiregulin-dependent growth stimulation of pancreatic cancer cells in vitro, we analyzed the expression and induction of D-type cyclins by growth factors in pancreatic cancer. We now report that D-type cyclins are di€erentially expressed in pancreatic carcinoma and that cyclin D3 mRNA levels are increased by amphiregulin in a human pancreatic cancer cell line.

Materials and methods Fetal bovine serum (FBS), trypsin EDTA solution, and penicillinstreptomycin solution were from Sigma Chemicals (Deisenhofen, Germany); Dulbecco's minimal essential medium (DMEM) and RPMI 1640 medium were from Gibco-BRL (Eggenstein, Germany); Hybond membranes were from Amersham (Braunschweig, Germany); pGEM3Zf vector was from Promega Biotech. (Mannheim, Germany); reverse transcriptase was from Boehringer Mannheim (Mannheim, Germany); Taq polymerase was from Gibco-BRL; [alpha-32P]dCTP (3000 Ci/mmol) was from Amersham; human recombinant amphiregulin and platelet-derived growth factor were from R&D Systems (Wiesbaden, Germany); oligonucleotides were purchased from MWG-Biotech. (Ebersberg, Germany). All other materials were obtained from Sigma Chemicals and were of molecular biology grade. Cell culture CAPAN-1, CAPAN-2, and PANC-1 human pancreatic cancer cells were obtained from the American Type Culture Collection

(Rockville, Md., USA). CAPAN-1 cells were grown in RPMI 1640 medium containing 10% FBS, 1% penicillin-streptomycin, and 1% glutamine. CAPAN-2 and PANC-1 cells were grown in DMEM containing 10% FBS, 1% penicillin-streptomycin, and 1% glutamine. Agonists were added after culturing the cells in serum-free media, containing 0.1% BSA, insulin, and transferrin (5 lg/ml), and antibiotics for 48 h (Ebert et al. 1994). Tissue samples Pancreatic cancer tissues (six female, four male) were obtained from patients undergoing pancreatic surgery. Normal pancreatic tissues were obtained from ®ve individuals (two female) through an organ donor program. The median age of the patients with pancreatic cancer was 61.5 years (range 48±73). The median age of the organ donors was 36 years (range 25±54). Immediately after surgical removal, tissue samples were frozen in liquid nitrogen. The tumor samples were classi®ed as pancreatic ductal adenocarcinomas according to the TNM classi®cation for pancreatic tumors (Warshaw and Fernandez-Del Castillo 1992). PCR and Northern blot analysis Oligonucleotide primers were purchased from MWG-Biotech. Primer sequences were adapted according to Toledo et al. (Toledo et al. 1995). Repeated RT-PCR analysis revealed a PCR product of the predicted size. Primer sequences were (Toledo et al. 1995): cyclin D1: 5¢-AGGAGAACAAACAG ATCA-3¢; 5¢-GTTGTTGAAGGACAGGAT-3¢; cyclin D2: 5¢-TCATGACTTCATTGAGCA-3¢; 5¢GTCGTCCTACTCCTTCAC-3¢; cyclin D3: 5¢-ACATGATTTCC TGGCCTT-3¢; 5¢-ACAGGCCCCTACTCGAGT-3¢. cDNAs were synthesized from total RNA (2 lg/sample) isolated from human pancreatic cancers (n ˆ 4), normal pancreatic tissues (n ˆ 5), and pancreatic cancer cell lines (n ˆ 2), using oligo(dT) and reverse transcriptase. Following inactivation, 2 ll of the reaction mixture were incubated in bu€er containing 1.23 mM concentrations each of dATP, dCTP, dGTP, and dTTP, 600 nM each of oligonucleotide primers, and 10% dimethyl-sulfoxide in bu€er consisting of 16.6 mM (NH4)2SO4, 67 mM Tris-HCl (pH 8.0), 6.7 mM MgCl2, 10 mM b-mercaptoethanol, and Taq polymerase (Ebert et al. 1994). Reaction cycles consisting of 1 min at 94 °C, 1 min at 57 °C, and 1 min at 72 °C were repeated 35 times. The PCR products were size fractionated on 1% agarose gels and visualized by ethidiume bromide (Ebert et al. 1994). A 1.4-kb EcoRI fragment of the human cyclin D1 cDNA, a 550-bp PstI fragment of the human cyclin D3 cDNA, which were kindly provided by Dr. Beach (Cold Spring Harbour Laboratory, New York), and a 190-bp BamHI fragment of the mouse 7S cDNA that cross hybridizes with human 7S RNA were random labeled with [alpha-32P]dCTP (Ebert et al. 1994). Total RNA was extracted from both cancer cell lines and human pancreatic tissues and blotted onto nylon membranes. The blots were prehybridized, hybridized, and washed under high stringency conditions for a cDNA probe, as previously described (Ebert et al. 1994). The blots were exposed at )80 °C to Kodak XAR-5 ®lm with Kodak intensifying screens, and the intensity of the radiographic bands was determined by laser densitometry (Ultrascan XL; Pharmacia LKB Biotechnology, Uppsala, Sweden). Immunohistochemistry The presence of human cyclin D3 was assessed using paran-embedded tissue sections obtained from ten patients with pancreatic cancer undergoing pancreatic surgery. The human tissues were ®xed in Bouin's solution and paran embedded. Two cyclin D3 antibodies were used: the anti-cyclin D3 antibody (H-292) is an anity-puri®ed rabbit polyclonal antibody raised against a recombinant protein corresponding to amino acids 1±292 representing full-length cyclin D3 of human origin and was used at a 1:100 dilution (Santa Cruz, Calif., USA). The speci®city of the antibody was con®rmed by Western blotting, immunoprecipitation, and

451 immunohistochemistry (Motokura et al. 1992). It also cross reacts with cyclin D1 and cyclin D2. In addition, the anti-cyclin D3 antibody (D-7) was also used. This antibody is a mouse monoclonal antibody raised against a protein corresponding to amino acids 1± 292 representing full-length cyclin D3 of human origin. This antibody reacts with and is speci®c for cyclin D3 of mouse, rat, and human origin by Western blotting, immunoprecipitation and immunohistochemistry. This antibody was also used at a concentration of 1:100. Paran sections (4-lm thick) were deparanized and rehydrated. For negative controls, the primary antibody was omitted (not shown). Endogenous peroxidase activity was inhibited by immersing the sections in 0.3% H2O2 for 30 min. The sections were incubated with the antiserum at 37 °C for 1 h and washed with PBS bu€er. The reaction was detected using the standard streptavidin-peroxidase technique (LSAB kit, DAKO; Hamburg, Germany). The analysis was performed according to the manufacturer's recommendations and all reactions were performed at 23 °C. Finally, the sections were counterstained with Mayer's hematoxylin (Ebert et al. 1994).

Results Northern blot analysis of total RNA extracted from ®ve normal pancreatic and four pancreatic carcinoma tissues failed to demonstrate transcripts corresponding to cyclins D1, D2, and D3 (not shown). Therefore, we performed reverse-transcriptase PCR analysis of these cases, using primers which had previously been demonstrated to be speci®c for these transcripts (Toledo et al. 1995). Furthermore, in order to ensure integrity of the cDNAs we performed an additional PCR analysis using

Fig. 1 RT-PCR analysis of human pancreatic tissues and pancreatic cancer cell lines. RNA (2 lg/sample), isolated from ®ve normal pancreatic, four pancreatic cancer tissue samples and two pancreatic cancer cell lines (CAPAN1, CAPAN2), was used to generate cDNAs by reverse transcriptase. The cDNAs were ampli®ed using primers speci®c for cyclin D1, D2, D3, and E. Primer sequences are outlined in ``Materials and methods''. Ampli®cation products of the same cDNAs with primers corresponding to b-actin were visualized by ethidium bromide staining, and served as an internal control. M DNA ladder, ()) negative control, i.e., PCR performed in the absence of cDNAs

primers speci®c for b-actin mRNA. All cases exhibited a band corresponding to b-actin, con®rming the integrity of our cDNA synthesis (Fig. 1). Furthermore, RT-PCR analysis of D-type cyclin expression revealed a band speci®c for cyclin D1, D2, and D3 in 4/4 cases and for cyclin E in 3/4 cases of pancreatic carcinoma (Fig. 1). In addition, three pancreatic cancer cell lines exhibited a band speci®c for cyclin D1, D3, and E in our RT-PCR analysis. However, repeated PCR analysis failed to detect a cDNA fragment corresponding to cyclin D2 in CAPAN and PANC-1 cells. Again, a 661-bp cDNA fragment corresponding to b-actin was readily evident in the pancreatic cancer cell lines (Fig. 1), con®rming the integrity of the cDNAs in the samples which exhibited no cyclin D2 expression in the RT-PCR analysis. Immunohistochemical analysis using an antibody directed against human cyclin D3 revealed the presence of cyclin D3 in the nuclei and perinuclear region of pancreatic cancer cells (Fig. 2). In order to study more closely the induction and expression of D-type cyclins in pancreatic cancer, we performed Northern blot analysis of cells which had been incubated with either amphiregulin or plateletderived growth factor. While the expression of the D-type cyclins in human pancreatic tissues was below the level of detection in Northern blot analysis, cyclin D1 (approximately 1.5 kb) and D3 (approximately 2.4 kb) mRNA transcripts were detectable in both cell lines using Northern blot analysis, thus con®rming the expres-

452 Fig. 2 Immunohistochemical analysis of cyclin D3 expression in the malignant pancreas. Immunohistochemistry revealed the nuclear and perinuclear expression of cyclin D3 in ductallike cancer cells. ´400

sion of these cyclins as determined by RT-PCR analysis (Figs. 1, 3). Platelet-derived growth factor had no detectable e€ect on the expression of either cyclin D1 (Fig. 3A) or cyclin D3 (Fig. 3B) in either cell line. However, as determined by laser densitometry and standardization against the respective 7S signal, amphiregulin (10 ng/ml) caused a marked increase in cyclin D3 mRNA levels in CAPAN-2 cells (Fig. 3C), which was maximal at 48 h following agonist addition. In contrast, amphiregulin did not alter the expression of cyclin D1 in either CAPAN-1 or CAPAN-2 cells (not shown). Fig. 3 A E€ect of PDGF on cyclin D1 mRNA levels. Cells were incubated in the absence (0) or presence (1 ˆ 4 h, 2 ˆ 8 h, 3 ˆ 24 h) of 10 ng/ml PDGF. Northern blots (20 lg RNA/lane) were hybridized with either the cyclin D1 (1 ´ 106 cpm/ml; exposure time, 6 days) or 7S (5 ´ 104 cpm/ml; exposure time, 24 h) cDNA probes, as described in ``Materials and methods''. B E€ect of PDGF on cyclin D3 mRNA levels. CAPAN1 cells were incubated in the absence (0) or presence (1 ˆ 4 h, 2 ˆ 8 h, 3 ˆ 24 h, 4 ˆ 48 h) of 10 ng/ml PDGF. Northern blots (20 lg RNA/lane) were hybridized with either the cyclin D3 (1 ´ 106 cpm/ml; exposure time, 5 days) or 7S (5 ´ 104 cpm/ml; exposure time, 24 h) cDNA probes, as described in ``Materials and methods''. C E€ect of amphiregulin on cyclin D3 mRNA levels. CAPAN 2 cells were incubated in the absence (0) or presence (1 ˆ 4 h, 2 ˆ 8 h, 3 ˆ 24 h, 4 ˆ 48 h) of 10 ng/ml amphiregulin. Northern blots (20 lg RNA/lane) were hybridized with either the cyclin D3 (1 ´ 106 cpm/ml; exposure time, 4 days) or 7S (5 ´ 104 cpm/ml; exposure time, 24 h) cDNA probes, as described in ``Materials and methods''

Discussion Pancreatic cancers exhibit multiple genetic and molecular alterations, the majority of which have been identi®ed in the last few years (Hunter and Pines 1994; Hahn et al. 1996; Berrozpe et al. 1994; Ebert et al. 1994). It is now generally well accepted that K-ras gene mutations are early and frequent events in pancreatic carcinogenesis (Berrozpe et al. 1994). Furthermore, recently, deletions of the DPC4-gene have been described in pancreatic cancer. While the role of DPC4 remains to be elucidated, it is well established that ras functions as a signal transducer and its activation is closely linked to the activation of tyrosine kinase receptors (Schlessinger and Ullrich 1992). In the past, we have studied the expression of growth factor receptors and have demonstrated that the majority of these receptors are overexpressed in pancreatic cancer (Ebert et al. 1994, 1995; Yokoyama et al. 1995). Furthermore, the combined expression of growth factors and their receptors is associated with a poor prognosis (Ebert et al. 1994). In vitro analysis has con®rmed that these growth factors, such as transforming growth factor-alpha and amphiregulin of the EGF-family, are potent stimulators of pancreatic cancer cell growth and proliferation (Ebert et al. 1994, 1995; Yokoyama et al. 1995). In contrast,

453

platelet-derived growth factors do not stimulate the growth of pancreatic cancer cells in vitro; instead they act as mitogenic and chemoattractants towards ®broblasts, pointing to a role of this family of growth factors towards the stimulation of the formation of the extracellular matrix and, thus, desmoplasia in human malignant epithelial tumors (Ebert et al. 1995). In conclusion, while overexpression of growth factors is a common feature of pancreatic cancers, the in vitro analysis of their proliferative and stimulatory e€ect has revealed that only a subset of growth factors exerts a direct stimulatory e€ect on pancreatic cancer cells in vitro. In an e€ort to further study and analyze the pathways associated with growth-factor-dependent growth stimulation of pancreatic cancer cells, we studied the expression of D-type cyclins in pancreatic cancer. While D-type cyclins regulate the progression of the cell cycle during the G1 phase by association with cyclin-dependent kinases, recently several groups have reported that the expression of these cyclins is highly growth-factordependent (Surmacz et al. 1992). Thus, in human cultured ®broblasts the expression of cyclin D1 was induced by platelet-derived growth factor (Surmacz et al. 1992), and Yan et al. reported the induction of cyclin D1 expression by transforming growth factor-alpha in esophageal squamous epithelial cells (Yan et al. 1997). While we could not demonstrate an aberrant expression of D-type cyclins in human pancreatic cancers, we turned to growth induction studies using three wellcharacterized human pancreatic cancer cell lines. Interestingly, we did not observe an induction of cyclin D1 using either amphiregulin or platelet-derived growth factor. However, the expression of cyclin D3 was greatly enhanced after incubating a pancreatic cancer cell line with amphiregulin. To our knowledge, this is the ®rst report demonstrating a growth-factor-dependent induction of cyclin D3 expression in transformed epithelial cells. In contrast, PDGF did not alter the expression of cyclin D3 in these cells. Inasmuch as Yan et al. have demonstrated that the induction of cyclin D1 by transforming growth factor-alpha is mediated in part by the induction of the early growth response protein Egr-1 and its binding to the cyclin D1 promotor, the molecular steps involved in the induction of cyclin D3 expression remain to be elucidated (Yan et al. 1997). Our RT-PCR analysis revealed the expression of Dtype cyclins in all cases of pancreatic cancer and the expression of cyclin E in 3/4 cases of pancreatic cancer. In contrast, cultured pancreatic cancer cells did not express cyclin D2. Together with the analysis of D-type cyclin expression in the normal pancreas, our ®ndings point to the di€erential expression of D-type cyclins in human pancreatic cancer. The lack of cyclin E expression in one case of pancreatic cancer and in one case of normal pancreas may result from the limited sensitivity of our RT-PCR analysis. Thus, our ®ndings may re¯ect a low level of cyclin E expression in these two cases which was not detectable using our RT-PCR analysis. From our limited number of pancreatic cancer samples

that were used for our analysis of cyclin expression, a correlation of cyclin E expression with histological or clinical parameters, however, cannot be drawn. In conclusion, our analysis demonstrates that the expression of D-type cyclins seems to be low in pancreatic cancer cells; thus, using Northern blot analysis no apparent expression was detectable. Furthermore, cyclin D2 was not expressed in pancreatic cancer cells in vitro pointing to the di€erential expression of D-type cyclins in human pancreatic cancer. In addition, cyclin D3 expression was greatly enhanced by amphiregulin, raising the hypothesis that cyclin D3 may lead to the progression of the cell cycle via aberrant activation of the EGF receptor by amphiregulin. Acknowledgements This study was supported by the Rudolf-Bartling-Stiftung (Hannover, Germany) and by the Land SachsenAnhalt.

References Berrozpe G, Schae€er J, Peinado MA, Real FX, Perucho M (1994) Comparative analysis of mutations in the p53 and K-ras genes in pancreatic cancer. Int J Cancer 58: 185±191 Ebert M, Yokoyama M, Kobrin MS, Friess H, Lopez ME, BuÈchler MW, Korc M (1994) Induction and expression of amphiregulin in human pancreatic cancer. Cancer Res 54: 3959±3962 Ebert M, Yokoyama M, Kobrin MS, Friess H, BuÈchler MW, Korc M (1995) Induction of platelet-derived growth factors and overexpression of their receptors in pancreatic cancers. Int J Cancer 62: 529±535 Gansauge S, Gansauge F, Ramadani M, Stobbe H, Rau B, Harada N, Beger HG (1997) Overexpression of cyclin D1 in human pancreatic carcinoma is associated with poor prognosis. Cancer Res 57: 1634±1637 Grana X, Reddy E (1995) Cell cycle control in mammalian cells: role of cyclins, cyclin-dependent kinases (CDKs), growth suppressor genes and cyclin-dependent kinase inhibitors (CKIs). Oncogene 11: 211±219 Hahn SA, Schutte M, Hoque AT, Moskaluk CA, da Costa LT, Rozenblum E, Weinstein CL, Fischer A, Yeo CJ, Hruban RH, Kern SE (1996) DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1. Science 271: 350±353 Hinds PW, Dowdy SF, Eaton EN, Arnold A, Weinberg RA (1994) Function of a human cyclin gene as an oncogene. Proc Natl Acad Sci USA 91: 709±713 Hunter T, Pines J (1994) Cyclins and cancer: cyclin D and CDK inhibitors come of age. Cell 79: 573±582 Korc M, Chandrasekar B, Yamanaka Y, Friess H, BuÈchler M, Beger HG (1992) Overexpression of the epidermal growth factor receptor in human pancreatic cancer is associated with concomitant increase in the levels of epidermal growth factor and transforming growth factor-alpha. J Clin Invest 90: 1352± 1360 Kuchiki H, Saino M, Nobukuni T, Yasuda J, Maruyama T, Kayama T, Murakami Y, Sekiya T (2000) Detection of ampli®cation of a chromosomal fragment at 6p21 including the cyclin D3 gene in a glioblastoma cell line by arbitrarily primed polymerase chain reaction. Int J Cancer 85: 113±116 Leach FS, Elledge SJ, Sherr CJ, Willson JK, Markowitz S, Kinzler KW, Vogelstein B (1993) Ampli®cation of cyclin genes in colorectal carcinomas. Cancer Res 53: 1986±1989 Motokura T, Keyomarsi K, Kronenberg HM, Arnold A (1992) Cloning and characterization of human cyclin D3, a cDNA closely related in sequence to the PRAD1/cyclin D1 protooncogene. J Biol Chem 267: 20412±20415 Mueller A, Odze R, Jenkins TD, Shahsesfasei A, Nakagawa H, Inomoto T, Rustgi AK (1997) A transgenic mouse model with

454 cyclin D1 overexpression results in cell cycle, epidermal growth factor receptor, and p53 abnormalities. Cancer Res 57: 5542± 5549 Schlessinger J, Ullrich A (1992) Growth-factor signalling by receptor tyrosine kinases. Neuron 9: 383±391 Surmacz E, Reiss K, Sell C, Baserga R (1992) Cyclin D1 messenger RNA is inducible by PDGF in cultured ®broblasts. Cancer Res 52: 4522±4525 Toledo SM, Azzam EI, Gasmann MK, Mitchel REJ (1995) Use of semiquantitative reverse transcription-polymerase chain reaction to study gene expression in normal human skin ®broblasts following low dose-rate irradiation. Int J Radiat Biol 67: 135± 143

Wang Z, Zhang Y, Kamen D, Lees E, Ravid K (1995) Cyclin D3 is essential for megakaryocytopoesis. Blood 86: 3783±3788 Warshaw AL, Fernandez-Del Castillo C (1992) Pancreatic carcinoma. N Engl J Med 326: 455±465 Yan YX, Nakagawa H, Lee MH, Rustgi AK (1997) Transforming growth factor-alpha enhances cyclin D1 transcription through the binding of early growth response protein to a cis-regulatory element in the cyclin D1 promotor. J Biol Chem 272: 33181± 33190 Yokoyama M, Ebert M, Funatomi H, Friess H, BuÈchler MW, Johnson GR, Korc M (1995) Amphiregulin is a potent mitogen in human pancreatic cancer cells: correlation with patient survival. Int J Oncol 6: 625±631