Mechanisms of Inactivation of the Receptor Tyrosine Kinase EPHB2 in Colorectal Tumors Hafid Alazzouzi, 1 Veronica Davalos, 1 Antti Kokko, 2 Enric Domingo, 1 Stefan M. Woerner, 3 Andrew J. Wilson, 4 Lars Konrad, 3 Pa ¨ivi Laiho, 2 Eloi Espı ´n, 1 Manel Armengol, 1 Kohzoh Imai, 5 Hiroyuki Yamamoto, 5 John M. Mariadason, 4 Johannes F. Gebert, 3 Lauri A. Aaltonen, 2 Simo Schwartz, Jr., 1 and Diego Arango 1 1 Molecular Oncology and Aging Group, Molecular Biology and Biochemistry Research Center (CIBBIM), Valle Hebron Hospital Research Institute, Barcelona, Spain; 2 Department of Medical Genetics, Biomedicum Helsinki, University of Helsinki, Helsinki, Finland; 3 Institute of Molecular Pathology/Applied Tumor Biology, University of Heidelberg, Heidelberg, Germany; 4 Montefiore Medical Center, Albert Einstein Cancer Center, Bronx, New York; and 5 First Department of Internal Medicine, Sapporo Medical University, Sapporo, Japan Abstract The receptor tyrosine kinase EPHB2 has recently been shown to be a direct transcriptional target of TCF/B-catenin. Pre- malignant lesions of the colon express high levels of EPHB2 but the expression of this kinase is reduced or lost in most colorectal carcinomas. In addition, inactivation of EPHB2 has been shown to accelerate tumorigenesis initiated by APC mutation in the colon and rectum. In this study, we investigated the molecular mechanisms responsible for the inactivation of EPHB2 in colorectal tumors. We show here the presence of mutations in repetitive sequences in exon 17 of EPHB2 in 6 of 29 adenomas with microsatellite instability (MSI), and 101 of 246 MSI carcinomas (21% and 41%, respectively). Moreover, we found EPHB2 promoter hypermethylation in 54 of the 101 colorectal tumors studied (53%). Importantly, EPHB2 expression was restored after treatment of EPHB2 -methylated colon cancer cells with the DNA methyltransferase inhibitor 5-aza-2V -deoxycytidine. In conclusion, in this study, we elucidate the molecular mecha- nisms of inactivation of EPHB2 and show for the first time the high incidence of frameshift mutations in MSI colorectal tumors and aberrant methylation of the regulatory sequences of this important tumor suppressor gene. (Cancer Res 2005; 65(22): 10170-3) Introduction The great majority of colorectal tumors display a constitutive up-regulation of TCF/h-catenin transcriptional activity, most commonly caused by mutations in the tumor suppressor gene APC (1). Increased activity of the TCF/h-catenin pathway is therefore a hallmark of colorectal cancer. This complex transcrip- tionally up-regulates key genes that are important in the devel- opment of these tumors. However, potentially oncogenic changes are often counterbalanced by additional effects of these alterations that abrogate a possible growth advantage. For example, deregu- lation and amplification of the transcription factor c-MYC is one of the most common events in colorectal tumors. Although c-MYC can promote cell growth by regulating the expression levels of multiple cell cycle regulators, it can also induce apoptosis through the modulation of proapoptotic genes such as BAX (2). The receptor tyrosine kinase EPHB2 has recently been shown to be a direct transcriptional target of TCF/h-catenin and premalignant lesions of the colon express high levels of expression of this kinase (3–5). However, EPHB2 expression is reduced in colonic carcino- mas and low levels are correlated with tumor progression (5). In addition, inactivation of EPHB2 has been shown to accelerate tumorigenesis initiated by APC mutations in the colon and rectum of APC Min/+ mice (5), demonstrating that EPHB2 is an important tumor suppressor in the large intestine. Therefore, despite being up-regulated by TCF/h-catenin signaling, inactivation of EPHB2 seems to be an important requirement in the progression of colorectal tumors. However, it is currently not known how EPHB2 activity is lost during tumor progression. Approximately 15% of the tumors of the colon and the rectum display a microsatellite unstable phenotype (MSI). This is observed as frequent insertions and deletions within short repetitive sequences known as microsatellites. Mutations within coding regions of the target genes result in frameshifts that can disrupt protein function. Mutations that confer a growth advantage to the cells are selected and can be found in a significant percentage of colorectal tumors with an MSI phenotype. Genes frequently targeted by these mutation include those involved in molecular mechanisms important in the development of colorectal tumors, including the transforming growth factor-h pathway, Wnt signal- ing, and DNA damage repair and apoptosis pathways (6). Because loss of EPHB2 activity is an important step in tumor progression, and because this receptor contains an A9 track in exon 17 that could be a target for mutation in MSI tumors, we screened this region for alterations in MSI tumor cell lines as well as in primary adenomas and carcinomas with MSI. Hypermethylation of cytosines located within CpG islands in the promoter of tumor suppressor genes is emerging as an important mechanism of gene silencing in both microsatellite stable (MSS) and unstable colorectal tumors, and has been reported to disrupt important pathways in colorectal tumori- genesis, including the TP53 pathway (p14ARF ), the WNT signal- ing pathway (APC , E-cadherin), DNA repair (MGMT, hMLH1, BRCA1 ), apoptosis (DAPK ), and metastasis (E-cadherin, TIMP3 ; ref. 7). We identified a CpG island spanning the proximal EPHB2 promoter and the first exon, and investigated the possible contribution of aberrant methylation of this region in the regulation of EPHB2 expression. Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). Requests for reprints: Diego Arango, Molecular Oncology and Aging Group, Molecular Biology and Biochemistry Research Center (CIBBIM), Valle Hebron Hospital Research Institute, Passeig Vall d’Hebron 119-129, Barcelona 08035, Spain. Phone: 34- 93-489-4058; Fax: 34-93-489-4040; E-mail: darango@vhebron.net. I2005 American Association for Cancer Research. doi:10.1158/0008-5472.CAN-05-2580 Cancer Res 2005; 65: (22). November 15, 2005 10170 www.aacrjournals.org Priority Report © 2005 American Association for Cancer Research (AACR). Reprinted with permission from Cancer Research. 2005; 65: 10170-10173.