An Antagonist of Dishevelled Protein-Protein Interaction Suppresses B-Catenin–Dependent Tumor Cell Growth Naoaki Fujii, 1,3 Liang You, 3 Zhidong Xu, 3 Kazutsugu Uematsu, 8 Jufang Shan, 5,7 Biao He, 3 Iwao Mikami, 3 Lillian R. Edmondson, 4 Geoffrey Neale, 6 Jie Zheng, 5,7 R. Kiplin Guy, 1,2 and David M. Jablons 3 Departments of 1 Pharmaceutical Chemistry and 2 Cellular and Molecular Pharmacology; 3 Thoracic Oncology Laboratory, Department of Surgery; and 4 Genome Analysis Core, Cancer Center, University of California San Francisco, San Francisco, California; 5 Department of Structural Biology and 6 Hartwell Center for Bioinformatics and Biotechnology, St. Jude Children’s Research Hospital; 7 Department of Molecular Sciences, University of Tennessee Health Science Center, Memphis, Tennessee; and 8 Division of Medical Oncology, Tokai University School of Medicine, Isehara, Japan Abstract Recent progress in the development of inhibitors of protein- protein interactions has opened the door for developing drugs that act by novel and selective mechanisms. Building on that work, we designed a small-molecule inhibitor of the Wnt signaling pathway, which is aberrantly activated across a wide range of human tumors. The compound, named FJ9, disrupts the interaction between the Frizzed-7 Wnt receptor and the PDZ domain of Dishevelled, down-regulating canonical Wnt signaling and suppressing tumor cell growth. The antitumori- genic effects of FJ9 were pronounced, including induction of apoptosis in human cancer cell lines and tumor growth inhibition in a mouse xenograft model. FJ9 is thus among the first non-peptide inhibitors to show therapeutic efficacy through disruption of PDZ protein-protein interactions. [Cancer Res 2007;67(2):573–9] Introduction Aberrant activation of the Wnt signaling pathway is implicated in the development of a broad spectrum of tumors (1–3). Up- regulation of Wnt-associated genes has been shown to play a role in the development of human cancers. For example, Wnt16 is overexpressed in the pre-B subtype of acute lymphoblastic leukemia, which is characterized by a t(1;19) chromosomal translocation that results in the E2A-PBX1 fusion protein. Wnt16 , a target gene of E2A-PBX1, plays key a role in leukemogenesis in such cases (4, 5). However, although many Wnt genes (Wnt1, Wnt3a , and Wnt16 ) are thought to have oncogenic potential, others (Wnt5a and Wnt7a ) have the characteristics of tumor suppressor genes (2). For example, most lung cancer cell lines and tissues show loss of Wnt7a, and the restoration of Wnt7a expression up- regulates E-cadherin and inhibits proliferation of non–small cell lung cancer cells in a peroxisome proliferator-activated receptor g–dependent manner (6–8). Loss of expression of h-catenin and E-cadherin is considered a marker of poor prognosis for lung cancer patients (9–12). Taken together, the evidence suggests that Wnt pathway activation may have opposing functions that depend on the specific ligand/receptor combination. Because of the multiple functions of the Wnt family, inhibition of all Wnt signaling may not be a perfect strategy for tumor therapy. One solution would be to target only Wnt signaling molecules that contribute significantly to tumorigenesis. A promising way to investigate the role of specific Wnt molecules in tumorigenesis is to observe whether inhibition of their signaling causes tumor promotion or suppression. For example, we have shown that silencing of the Wnt16 and Wnt1 genes suppressed the growth of pre-B leukemia cells and lung cancer cells, respectively (4, 13). Small-molecule drugs that inhibit Wnt signaling can be used to test our hypothesis that Wnt signaling promote the growth of lung cancers. Such drugs can be rationally designed by targeting the Frizzled (Frz) family of Wnt receptors, which relay Wnt signaling to the h-catenin/Tcf pathway (14). Several observations suggest that different Wnt molecules use a different subtypes of Frz receptor. For example, Wnt7a relays signal via Frz5 (15) and Frz9 (8). On the other hand, Wnt signaling via Frz7 is reported to have oncogenic potential (16). Frz7 ectodomain (an antagonist form) expression inhibits tumor growth in a colon cancer cell line (17), and Frz7 is overexpressed in tumor cell lines (18) and tumor tissues (19). Therefore, we hypothesized that selective targeting of Frz7 will suppress oncogenic Wnt signaling without interfering with tumor-suppressive Wnt7a signaling. Frz7 interacts directly with a PDZ protein interaction domain of the Dishevelled (DVL) family (20). DVL3 is overexpressed in a wide spectrum of cancer cells (21). Wnt signaling in the h-catenin pathways seems to be induced by DVL overexpression (22, 23). Therefore, disrupting the Frz7-DVL protein-protein interaction (24) represents a promising strategy for cancer therapy. Previously, we have shown that dominant-negative DVL lacking the PDZ domain decreased cytosolic h-catenin levels, diminished Tcf-mediated transcription, and suppressed tumorigenesis of mesothelioma cells in vitro and in vivo (22). This study suggests that the PDZ domain of DVL represents an attractive cancer therapeutic target. The PDZ domain is a common protein-protein interaction module that recognizes short peptide motifs on the COOH termini or internal sequences that structurally mimic a sharp h- turn structure of the cognate ligand (25). PDZ domains mediate crucial protein-protein interactions that enforce localization and organization of proteins in a variety of submembranous complexes associated with cell signal mediators, including ion channels, transmembrane receptors, and regulatory enzymes (26). The therapeutic usefulness of PDZ protein-protein interaction Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). N. Fujii and L. You contributed equally to this work. Current address for N. Fujii and R.K. Guy: Department of Chemical Biology and Therapeutics, St. Jude Children’s Research Hospital, Memphis, TN 38105. Requests for reprints: Naoaki Fujii, Department of Chemical Biology and Therapeutics, St. Jude Children’s Research Hospital, M/S1000, 332 North Lauderdale Street, Memphis, TN 38105. Phone: 901-495-5854; Fax: 901-495-5715; E-mail: Naoaki. Fujii@stjude.org. I2007 American Association for Cancer Research. doi:10.1158/0008-5472.CAN-06-2726 www.aacrjournals.org 573 Cancer Res 2007; 67: (2). January 15, 2007 Research Article Research. on May 23, 2020. © 2007 American Association for Cancer cancerres.aacrjournals.org Downloaded from Research. on May 23, 2020. © 2007 American Association for Cancer cancerres.aacrjournals.org Downloaded from Research. on May 23, 2020. © 2007 American Association for Cancer cancerres.aacrjournals.org Downloaded from