Menin Promotes the Wnt Signaling Pathway in Pancreatic Endocrine Cells

Gao Chen, Jingbo A, Min Wang, Steven Farley, Lung-Yi Lee, Lung-Ching Lee, and Mark P. Sawicki

David Geffen School of Medicine at University of California at Los Angeles and the Greater Los Angeles VA Medical Center, Los Angeles, California

Abstract Introduction Menin is a tumor suppressor protein mutated in Multiple endocrine neoplasia type 1 (MEN1) is an patients with multiple endocrine neoplasia type 1. We autosomal dominant disease characterized by the development show that menin is essential for canonical Wnt/B- of parathyroid hyperplasia, pancreatic islet cell tumors, and catenin signaling in cultured rodent islet tumor cells. anterior pituitary endocrine tumors (1). Besides these classic In these cells, overexpression of menin significantly tumors, several other endocrine and nonendocrine tumors may enhances TCF gene assay reporter activity in response develop, including lipomas, dermatofibromas, collagenomas, to B-catenin activation. Contrastingly, inhibition of ependymomas, meningiomas, schwannomas, carcinoids, adre- menin expression with Men1 siRNA decreases TCF nal cortical tumors, and thyroid tumors. The MEN1 gene is reporter gene activity. Likewise, multiple endocrine highly conserved and encodes for a 67-kDa nuclear protein, neoplasia type 1 disease associated missense called menin, which has been implicated in DNA metabolism mutations of menin abrogate the ability to increase TCF and transcription regulation (2). MEN1 patients have trunca- reporter gene activity. We show that menin physically tion mutations and missense mutations that are predicted to interacts with proteins involved in the canonical Wnt inactivate menin function consistent with a tumor suppressor À/À signaling pathway, including B-catenin, TCF3 (TCFL1), protein (3). Homozygous knockout mice (Men1 )die and weakly with TCF4 (TCFL2). Menin overexpression in utero (E11.5-13.5) with multiple developmental defects, increases expression of the Wnt/B-catenin downstream suggesting that menin that has an essential role in early target gene Axin2, which is associated with increased development (4). Similar to MEN1 patients, heterozygous À/+ H3K4 trimethylation of the Axin2 gene promoter. knockout mice (Men1 ) develop normally, but the adult mice Moreover, inhibition of menin expression by siRNA eventually grow endocrine tumors (5). abrogates H3K4 trimethylation and Axin2 gene To determine how the loss of menin promotes endocrine expression. Based on these studies, we hypothesized tumor development, most investigations have focused on the that Wnt signaling could inhibit islet cell proliferation role of menin in transcription regulation. Menin interacts with because loss of menin function is thought to increase several transcription factors including members of the activator endocrine tumor cell proliferation. TGP61 rodent protein-1 (JunD), transforming growth factor-h/bone morpho- islet tumor cells treated with a glycogen synthase genetic protein (RUNX2, SMAD1, SMAD3, and SMAD5), and kinase 3B inhibitor that increases Wnt pathway nuclear factor-nB (p65, p50, and p52) signal transduction signaling had decreased cell proliferation compared pathways (6-11). More recently, menin was identified in MLL with vehicle-treated cells. Collectively, these and MLL2 histone methyltransferase complexes and found data suggest that menin has an essential role in essential for the histone methyltransferase activity (12-16). Wnt/B-catenin signaling through a mechanism that Pancreas development, in particular islets, is mediated by eventually affects histone trimethylation of the expression of a specific program of growth factors including the downstream target gene Axin2, and activation of Wnt proteins (17-24). In general, Wnt proteins regulate Wnt/B-catenin signaling inhibits islet tumor cell development through actions on cell proliferation, differentia- proliferation. (Mol Cancer Res 2008;6(12):1894–907) tion, cell fate decisions, apoptosis, axial polarity, axonal guidance, and adhesion (25). In the canonical Wnt pathway, the extracellular Wnt ligand (20 different genes in mammalian cells) binds the cell surface receptor frizzled (10 FZD genes in mammalian cells) and the coreceptor LRP5/6 (low density Received 12/14/07; revised 8/4/08; accepted 8/25/08. lipoprotein receptor–related proteins 5 and 6). This initiates a Grant support: NIH grant R01CA095148 and VA Merit grants (M.P. Sawicki). cascade of intracellular mediators that results in h-catenin 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 stabilization and subsequent h-catenin nuclear localization to accordance with 18 U.S.C. Section 1734 solely to indicate this fact. regulate target gene transcription (26). In the absence of Wnt Note: Supplementary data for this article are available at Molecular Cancer ligand, h-catenin is phosphorylated in a cytoplasmic complex Research Online (http://mcr.aacrjournals.org/). Requests for reprints: Mark P. Sawicki, David Geffen School of Medicine, CHS with casein kinase 1, glycogen synthase kinase 3h (GSK3h), 72-215, Los Angeles, CA 90095-6904. Phone: 310-268-3298; Fax: 310-268- axin 1, and the tumor suppressor adenomatous polyposis coli. 3026. E-mail: [email protected] h Copyright D 2008 American Association for Cancer Research. Phosphorylated -catenin is ubiquitylated and targeted for doi:10.1158/1541-7786.MCR-07-2206 degradation by the proteasome. Wnt binds the Fzd receptor and

stops phosphorylation of h-catenin. The hypophosphorylated (dephosphorylated) h-catenin is not altered by menin over- h-catenin translocates to the nucleus where it associates with expression. the T-cell–specific transcription factor/lymphoid enhancer binding factor 1 (TCF/LEF) family transcription factors and Menin Is Essential for b-Catenin Activation of a TCF regulates Wnt responsive genes expression. Responsive Reporter Components of the Wnt pathway regulate pancreas deve- To determine whether menin is involved with Wnt/h-catenin lopment, although the mechanism is not fully understood. signaling, we studied mouse TGP-61 islet tumor cells trans- Numerous members of the Wnt pathway are expressed during fected with a TCF responsive luciferase reporter vector the development of the mouse pancreas and in the adult organ (Super8XTOPflash, M50) containing eight TCF/LEF binding (27). In particular, activated (dephosphorylated) h-catenin and sites or a control vector (Super8XFOPflash, M51) that has the parahox protein Pdx-1, which signals pancreas lineage mutant TCF/LEF binding sites (32). As expected, over- development, are both expressed in the pancreas epithelium expression of either wild-type h-catenin or the constitutively during development (20). Pdx-1 promoter directing Wnt activated phosphorylation mutant h-catenin (S37A; ref. 33) misexpression in a transgenic mouse model results in pancreas increased TCF reporter gene assay activity (Fig. 2A). Menin agenesis (28). A h-catenin conditional knockout mouse, alone did not activate the reporter gene, but cotransfection of however, loses normal pancreatic acinar tissue development, menin and either h-catenin or activated h-catenin (S37A) but islet development is preserved (22). Whereas it is clear that strongly stimulated the reporter gene activity severalfold above Wnt signaling and h-catenin are important for pancreas that seen with either h-catenin or h-catenin (S37A) alone. One development, its role in pancreatic endocrine tumor develop- possible explanation for the increased reporter activity with ment is unknown. menin overexpression could be altered h-catenin expression, Based on the role of menin in multiple cell signaling but Western blot analysis shows similar expression of h-catenin pathways and the broad role of Wnt signaling in tumor when menin is co-overexpressed with h-catenin or h-catenin development, we hypothesized that menin may be important for (S37A; Fig. 2A, right). To determine whether menin was Wnt/h-catenin signaling in endocrine tumors. In this report, we affecting the reporter promoter through a mechanism not investigate the role of menin in Wnt/h-catenin signaling in a involving the TCF response elements, similar experiments were mouse islet cell line. These data suggest menin is important for done comparing the response of the control reporter gene, regulation of Wnt signaling in the endocrine pancreas through a Super8XFOPflash (plasmid M51 labeled ‘‘F’’ in Fig. 2B), mechanism that involves histone H3K4 trimethylation. which has mutant TCF binding sites, with the response from the nonmutated reporter Super8XTOPflash (plasmid M50 labeled Results ‘‘T’’ in Fig. 2B). Menin and h-catenin did not activate the Expression of Endogenous b-Catenin, TCF3, and TCF4 in negative control reporter gene Super8XFOPflash that has Endocrine and Nonendocrine Cell Lines and the Effect of mutant TCF binding sites (Fig. 2B). Expression of the Menin Overexpression overexpressed proteins is shown in Fig. 2B (right), indicating Canonical Wnt pathway protein expression in endocrine that the transfections of multiple plasmids did not significantly tumor cell lines has not been previously reported. Therefore, we interfere with each other. We then wondered whether Wnt/h- first determined the expression of activated (dephosphorylated) catenin signaling was dependent on menin function. We tested h-catenin in selected islet tumor endocrine (TGP-61 and this hypothesis by decreasing menin expression with transfec- InR1G9) and nonendocrine cell lines (HEK 293T). Endogenous tion of mouse specific menin-optimized siRNA (TGP61 is a h-catenin protein was expressed in all three cell lines (Fig. 1A). mouse derived cell line). Several different menin siRNAs were The amount of activated h-catenin protein expression, however, tested (see Materials and Methods) with the most effective was variable, with highest expression in the mouse islet tumor siRNA chosen for subsequent experiments. Decreased menin cell line TGP-61 (29), slightly less in human embryonic kidney expression significantly inhibited reporter gene activity in cells HEK 293T (30), and barely detectable in the hamster islet response to transfected h-catenin (label ‘‘M’’ in Fig. 2C) cell line InR1G9 (31). We then examined the expression of two compared with control siRNA (label ‘‘C’’ in Fig. 2C). Western of the TCF/LEF family members. TCF3 (also called TCF7L1) blot analysis confirmed that menin protein expression was and TCF4 (also called TCF7L2) were both highly expressed in reduced by the siRNA transfection (Fig. 2C, right). Similar TGP-61 cells, but TCF4 was predominantly expressed in both studies in human embryonic kidney (HEK 293T) cells showed HEK 293T and InR1G9 cells. Whereas the expression patterns that menin more modestly augmented reporter gene activity in of TCF4 and TCF3 differ in these cell lines, these proteins are the presence of h-catenin (data not shown). functionally similar (25). Overexpression of menin did not We then wondered whether disease-associated menin muta- affect the expression of TCF4, TCF3, or activated (dephos- tions would lose the ability to augment Wnt/h-catenin–mediated phorylated) h-catenin in any of the cell lines studied. In cells signaling. All of the disease-associated menin missense and expressing fluorescent epitope–tagged h-catenin, TCF3, and COOH-terminal deletion mutants studied (mutants illustrated in TCF4, menin overexpression did not affect the nuclear Fig. 3A) lost the ability to stimulate h-catenin–mediated reporter localization of these proteins (Fig. 1B). These data suggest gene activity (Fig. 3B and C) consistent with a role for menin in that proteins important for canonical Wnt signaling are canonical Wnt/h-catenin–mediated signaling. The menin trun- expressed in these cell lines, in particular the mouse islet cation mutants (meninDB-E) had lower expression compared tumor cell line TGP-61, and their expression is not altered by with wild-type menin (Fig. 3C, right). To adjust for this effect, exogenous menin. Specifically, the amount of activated we transfected more plasmid DNA and increased expression of

FIGURE 1. Expression and colocalization of menin, h-catenin, and TCF3/4 in endocrine and nonendocrine cell lines. A. Western blot shows protein expression of endogenous h-catenin, activated (dephospho)h-catenin, and TCF3/4 in various cell lines transfected with either empty vector or pCMV- SPORT-menin (designated À and +, respectively). Forty-eight hours after transient transfection with either control vector or a menin-expressing plasmid (pCMV-Sport-menin), plated cells (HEK 293T, TGP-61, and InR1G9) were harvested and the protein lysate was immunoblotted and probed with antibodies to h-catenin, activated h-catenin (recognize h-catenin dephosphorylated on Ser37 or Thr41), and TCF3/4. Anti-menin antibody was used to confirm overexpression of menin in the pCMV-Sport-menin – transfected cells, and anti – h-tubulin was used to show equal protein loading. For each protein detected such as menin, all images from different cell lines were from the same Western blot and exposure, but the images were separated for publication purposes. B. TGP-61 mouse pancreas islet tumor cells expressing fluorescent epitope – tagged expression constructs show colocalization of menin with h-catenin and TCF3/4. ECFP images are pseudocolored red to distinguish these images from Hoechst dye nuclear staining. ECFP-h-catenin is predominantly nuclear due to overwhelming the proteasome degradation pathway by high levels of ECFP-h-catenin protein expression, but some cytoplasmic protein is still visualized. All images were obtained with 63Â objective.

the truncated proteins, but this further decreased reporter activity overexpressed proteins may aggregate and produce artificial as if truncated menin was acting in a dominant negative fashion interactions detected by immunoprecipitation, endogenous (data not shown). Overexpression of the missense mutants was menin/h-catenin interactions were investigated (Fig. 4A). comparable to wild-type menin overexpression (Fig. 3C, right). HEK 293T cells were treated with vehicle or a GSK3h Like the menin deletion mutants, there was loss of TCF gene assay inhibitor IX (BIO) to activate the canonical Wnt signaling reporter activity augmentation with these menin missense mutants cascade, and endogenous menin/h-catenin coimmunoprecipita- compared with overexpression of wild-type menin (Fig. 3C, left). tion was done. Treatment with the GSK3h inhibitor increased expression of activated h-catenin (see Fig. 4A, bottom). The COOH Terminus of Menin Interacts with b-Catenin, Immunoprecipitation with anti–activated h-catenin that recog- TCF3, and Weakly with TCF4 nizes h-catenin dephosphorylated on Ser37 or Thr41 confirmed Because Wnt signaling is mediated by h-catenin binding that endogenous menin coimmunoprecipitated with activated members of the TCF/LEF transcription factor family at target h-catenin with GSK3h inhibitor treatment. This does not prove gene promoters (34), we wondered whether menin interacted that h-catenin dephosphorylation is necessary for the menin with either h-catenin, TCF3, or TCF4. Because exogenously interaction, but more likely that nuclear localization increases

FIGURE 2. Menin promotes TCF luciferase reporter gene activity by activation of the Wnt/h-catenin pathway in TGP-61 mouse islet cells. All experiments were done with triplicate (n = 3) transfections, with the results graphed as the mean F SD. In all of the experiments, the amount of plasmid DNA transfected was kept constant by adding empty vector DNA as needed for each sample. All samples were transfected with the control vector pRL-hTK (Renilla luciferase) to adjust the results for transfection efficiency. A. Overexpression of menin promotes TCF luciferase gene reporter activity when cotransfected with h-catenin. The TCF luciferase reporter vector M50 (Super8XTOPflash) was cotransfected with pCMV-SPORT-menin alone, pCMV-SPORT-h-catenin alone, pCMV-SPORT-h-catenin/pCMV-SPORT-menin, or activated h-catenin mutant (S37A)/pCMV-SPORT-menin. The control sample was transfected with the reporter vector M50 with added empty pCMV-SPORT plasmid DNA. TCF reporter gene assay activity was measured by the Dual Luciferase Assay (Promega) and the results of the firefly luciferase/Renilla luciferase (FL/RL) ratio are graphed for each sample set. Expression of transfected constructs for each experiment is shown on the right. Protein loading was confirmed by h-tubulin expression. B. The augmentation of TCF luciferase gene reporter activity by menin is abrogated by mutation of the TCF enhancer elements of the gene reporter. To prove that activation of the TCF luciferase gene reporter was specifically due to the TCF enhancer, TGP-61 cells were transfected with the TCF reporter vector Super8XTOPflash (M50; label ‘‘T’’ in B) or mutant reporter vector Super8XFOPflash (M51; label ‘‘F’’ in B) either alone or cotransfected with h-catenin – or h-catenin/menin – expressing plasmids as above. Expression of transfected constructs for each experiment is shown on the right. Protein loading was confirmed by h-tubulin expression. C. Menin expression knockdown abrogates the TCF luciferase gene reporter response to overexpressed activated h-catenin. To show that menin knockdown inhibited h-catenin activation of the TCF gene reporter (M50), TGP-61 cells were transfected with either siRNA (Dharmacon) specifically optimized for mouse menin ‘‘M’’ or control ‘‘C’’ siRNA 24 h prior the transfection with the reporter vector M50 and either empty vector or activated h-catenin mutant (S37A) – expressing plasmid. Control sample has reporter vector and siRNA. Expression of transfected constructs for each experiment is shown on the right. Protein loading was confirmed by h-tubulin expression.

the likelihood of interaction because endogenous and overex- Based on these data, menin could either directly complex with pressed menin is largely located within the nucleus. Similar h-catenin or indirectly interact as part of a large complex. We interaction between endogenous menin and h-catenin was shown tested this hypothesis by performing in vitro pull-down assays. in TGP61 cells by coimmunoprecipitation (data not shown). For these interaction experiments, an NH2-terminal deletion

FIGURE 3. Mutant menin loses its ability to promote Wnt/h-catenin signaling. All experiments were done with triplicate (n = 3) transfections, with the results graphed as the mean F SD. In all of the experiments, the amount of plasmid DNA transfected was kept constant by adding empty vector DNA as needed for each sample. All samples were transfected with the control vector pRL-hTK (Renilla luciferase) to adjust the results for transfection efficiency. A. Diagram illustrates human MEN1 gene structure, disease-associated missense mutants, COOH-terminal deletion mutants, and NH2-terminal deletion mutants used in this article. Exons are numbered 1 to 10. Exon 10 is shortened for illustration purposes. Disease-associated menin missense mutants are designated above the gene structure. The COOH-terminal menin deletion mutants are designated by the amino acid that is mutated to a stop codon within the menin protein. The NH2-terminal deletion menin mutants show the amino acid residues expressed by the expression vector. B. Menin COOH-terminal deletion mutants lose the ability to promote TCF luciferase gene assay reporter activity in TGP61 cells. TGP-61 cells were transfected with the reporter vector M50 and cotransfected with pSport-h-catenin with or without pSport-menin in the presence of different menin COOH-terminal deletion mutants (meninDB-E). TCF reporter gene assay activity was measured by the Dual Luciferase Assay (Promega) and the results of the firefly luciferase/Renilla luciferase ratio are graphed for each sample set. Western blot analysis (B, right) shows decreased expression of the transfected menin deletion mutants compared with overexpressed wild-type menin. Increasing the deletion mutant menin plasmid expression by increasing the amount of transfected plasmid (meninDB-E) further decreased reporter activity (data not shown). Protein loading was confirmed by h-tubulin expression. C. Menin missense point mutants lose the ability to promote TCF gene assay reporter activity in TGP61. TGP-61 cells were transfected with the reporter vector M50 and cotransfected with pSport-h-catenin with or without pSport-menin or missense menin mutants (P12L, L22R, H139Y, A160P, A176P, and L286P). Right, corresponding Western blot for the transfected proteins. Protein loading was confirmed by h-tubulin expression.

FIGURE 4. Menin interacts with h-catenin, TCF3, and TCF4. A. Endogenous menin interacts with h-catenin, shown by coimmunoprecipitation analysis. HEK 293T cells (control) or cells treated with GSK3h inhibitor IX (5 Amol/L), to activate Wnt signaling by inhibiting h-catenin phosphorylation and proteasome degradation, were harvested and lysed in radioimmunoprecipitation assay buffer. Bottom, Western blot for the input lysates used in the coimmunoprecipitation. The GSK3h inhibitor increased expression of activated (dephosphorylated) h-catenin. The immunoprecipitation and coimmunoprecipitation are shown on top. Anti – activated h-catenin antibody (recognize h-catenin dephosphorylated on Ser37 or Thr41) was used for immunoprecipitation (IP) and samples were West- ern blotted (IB) with either anti – activated h- catenin or anti-menin antibody (top). The immu- nopreciptiation of activated (dephosphorylated) h-catenin is clearly shown (top, lower Western blot). Coimmunoprecipitation for the endogenous menin and h-catenin proteins (top, upper Western blot) is maximal when h-catenin is activated with the GSK3h inhibitor. Control antibody for the immunoprecipitation was mouse IgG (Jackson ImmunoResearch Laboratories). B. The menin COOH terminus interacts with h-catenin, TCF3, and TCF4 in vitro as shown by GST pull-down assay. GST-tagged meninD5s (contains menin COOH terminus) was expressed in bacteria and immobilized on glutathione-coated agarose beads. In vitro transcribed and translated (TnT System from Promega) and biotin-labeled (Transcend from Promega) h-catenin, TCF3, and TCF4 were used for pull-down assay and analyzed by Western blotting. Input represents 10% of the TnT product to control for the amount of protein used to show interaction. Biotinylated TnT products were detected by the Transcend Non-Radioactive Translation Detection System from Promega (labeled TNT in figure). mutant, meninD5s (contains amino acids 477-610), was expressed occurring disease-associated truncation mutations, therefore, as a glutathione S-transferase (GST) fusion protein in bacteria, would be predicted to lose the ability to interact with h-catenin. purified by immobilized glutathione gel, and incubated with We then wondered whether menin interacted with other members in vitro transcribed-translated (TnT) h-catenin, TCF4, and TCF3 of the h-catenin transcription complex and showed that menin (Fig. 4B). The proteins bound to the NH2-terminal deletion mutant could bind TCF3 or weakly with TCF4 either directly or as part and analyzed by Western blot analysis revealed specific pull-down of a larger complex by coimmunoprecipitation in HEK 293T of h-catenin, TCF3, and TCF4 by GST-meninD5s. These proteins cells (Fig. 5C). were not seen with the GST control protein. Menin, therefore, directly interacts with h-catenin, TCF3, and TCF4. Menin Regulates Wnt Downstream Target Axin2 Gene To define the menin interaction region, menin and h-catenin Expression were coexpressed in HEK 293T cells and coimmunoprecipitated Because menin significantly increased TCF reporter gene (Fig. 5A). Full-length menin specifically coimmunoprecipitated activity in the presence of activated h-catenin, we wondered with h-catenin. To determine the region of the menin protein that whether endogenous gene expression was likewise affected binds to h-catenin, HEK 293T cells were cotransfected with in TGP-61 cells. To identify a target gene modulated by different menin deletion mutants (see diagram Fig. 3A). The Wnt/h-catenin signaling in TGP-61 cells, we studied the mRNA coimmunoprecipitation studied showed that h-catenin binds to expression of several known downstream target genes: c-MYC, the COOH-terminal region of menin (Fig. 5A). COOH-terminal NKX2-2, Taspase, PDX1, LEF, MET, CDX1, HNF3B, HDAC9, menin truncation mutants (meninDB-D), on the other hand, LMX1,andAXIN2 (35).1 TGP-61 cells transfected with largely lost the ability to interact with h-catenin (Fig. 5B). The longest of these COOH-terminal deletion mutants, meninDE, weakly interacted with h-catenin. The majority of naturally 1 See www.stanford.edu/~rnusse/wntwindow.html.

FIGURE 5. The COOH terminus of menin interacts with h-catenin, TCF3, and TCF4, shown by coimmunoprecipitation analysis. A. The menin COOH terminus interacts with h-catenin. Western blots (IB; immunoblot antibody) were done on cell lysates or immunoprecipitates (IP; immnuoprecipitate antibody). HEK 293T cells were cotransfected with plasmids expressing FLAG-h-catenin and either full-length His-menin or NH2-terminal deletion mutants meninD2-D6 and coimmunoprecipitated to show interaction between the menin COOH terminus and h-catenin. The lower two Western blots show expression of the transfected expression constructs. The upper two Western blots show the results from immunoprecipitation. Menin D6 has low levels of expression but immunoprecipitates well. B. Deletion of the menin COOH terminus abrogates its ability to interact with h-catenin. HEK 293T cells were cotransfected with plasmids expressing FLAG-h-catenin and HA-menin and the COOH-terminal deletion mutants (HA-meninDB-DE). Immunoprecipitation shows interaction between full-length menin and h-catenin. There is a weaker interaction between h-catenin and menin DE (which contains one nuclear localization signal). The other menin deletion mutants do not interact strongly. C. The menin COOH terminus interacts with TCF3/4. HEK 293T cells were cotransfected with plasmids expressing HA-menin and HA-meninD5 transfected alone or with FLAG-h-catenin, FLAG-h-catenin (S37A), FLAG-TCF3, and FLAG-TCF4, respectively, showing that the COOH terminus of menin could interact with all of them, albeit weakly with TCF4. Asterisks represent immunoglobin (IgG) proteins fromthe immunoprecipitation.

expression constructs for menin and h-catenin (S37A) showed protein expression was increased in response to increased increased Axin2 gene expression in response to h-catenin in both h-catenin and menin/h-catenin expression (Fig. 6B). Expression TGP-61 and HEK 293T cells (Fig. 6A). There was no effect on of other Wnt regulated genes tested was not affected by Wnt b-actin gene (loading control) expression. Similarly, Axin2 signaling in these particular cell lines. This is not surprising

because the Wnt signaling response is likely tissue specific molecular role for menin is unknown. Because menin is known and the particular target genes for endocrine cell types are to bind MLL histone methyltransferase complexes (13, 15), we unknown. Based on the reporter gene assay data, we hypothe- hypothesized that menin could be involved in histone H3K4 sized that decreased menin expression would block h-catenin– trimethylation at the Axin2 gene promoter in response to Wnt/ induced Axin2 transcription. Indeed, h-catenin–induced Axin2 h-catenin signaling. Using the chromatin immunoprecipitation gene transcription was repressed by siRNA knockdown of menin assay with an antibody to trimethylated histone 3 lysine 4 (anti- expression in TGP-61 cells (Fig. 6C and D). To further confirm H3K4) and DNA amplification by PCR for the proximal Axin2 the role of menin in Wnt signaling, we performed Western blot promoter (genomic regions designated T2 and T3; ref. 36; analysis of SW480 colon cancer cells with constitutively active Fig. 7A), we showed increased histone H3 K4 trimethylation in h-catenin (due to adenomatous polyposis coli mutation) after the h-catenin (S37A)–transfected TGP-61 islet tumor cells siRNA knockdown of menin expression. Compared with control compared with empty vector–transfected cells (Fig. 7B). transfected SW480 cells, menin siRNA–transfected SW480 Histone H3K4 trimethylation was further increased when the cells had decreased Axin2 gene expression similar to our findings active h-catenin (S37A) was cotransfected with the menin in TGP61 cells (Supplementary Fig. S1). These data suggest that expression vector (Fig. 7B). To determine if menin was essential at least one endogenous canonical Wnt/h-catenin signaling target for Axin2 gene promoter H3K4 trimethylation, Men1 siRNA gene (Axin2) is dependent on menin expression. knockdown was done. Menin protein expression was reduced and an associated decrease in Axin2 promoter histone H3K4 Menin Is Essential for Histone H3K4 Trimethylation in the trimethylation was also seen (Fig. 7C). Hence, menin is directly Mouse Axin2 Gene Promoter in Response to Wnt/b- or indirectly essential for Axin2 gene promoter histone H3K4 Catenin Signaling trimethylation in response to Wnt/h-catenin signaling. Axin2 gene expression is regulated by the Wnt/h-catenin Increased Wnt/h-catenin signaling decreases TGP61 islet signaling pathway in TGP-61 mouse islet tumor cells and tumor cell proliferation in vitro. Based on the above studies, we menin is essential for this signaling (see above), but the exact hypothesized that increased Wnt signaling could inhibit islet

FIGURE 6. Menin is important for Wnt/h-catenin endogenous Axin2 gene expression. A. Wnt/h-catenin signaling and menin overexpression increase endogenous Axin2 gene expression. Agarose gels show RT-PCR semiquantitative Axin2 gene expression results after activation of the Wnt/h-catenin signaling pathway. HEK-293T and TGP-61 cells were transfected with control expression vector, pSPORT-CMV-h-catenin expression vector, or pSPORT- CMV-h-catenin and pSPORT-CMV-menin expression vectors. Forty-eight hours later, total RNA was extracted and analyzed by RT-PCR using Axin2 primers. h-Actin primers were used to confirm equal loading of RNA into the RT-PCR reactions. B. Western blot analysis of Axin2 protein expression corresponding to the experiment done in A. Western blot with anti – h-tubulin antibody was used as a control for protein gel loading. C. Menin knockdown abrogates the ability of Wnt/h-catenin signaling to increase endogenous Axin2 gene expression. Agarose gels showing semiquantitative RT-PCR gene expression results after Men1 siRNA transfection and activation of the Wnt/h-catenin signaling pathway. TGP-61 cells were first transfected with either control or Men1 specific siRNA. Twenty-four hours later, the cells were transfected with control expression vector or pSPORT-CMV-h-catenin expression vector. RT-PCR was done 48 h later, using Axin2, Men1, b-catenin,orb-actin primers as indicated. D. Western blot analysis of Axin2, menin, and h-catenin protein expression corresponding to the experiment done in C. Western blot with anti – h-tubulin antibody was used as a control for protein gel loading.

FIGURE 7. Menin is important for histone H3 K4 trimethylation in the mouse Axin2 gene promoter in response to Wnt/h-catenin signaling. A. Diagram illustrating the genomic organization of the mouse Axin2 gene 5¶ upstream region. Locations of TCF/LEF consensus binding elements (T2-T8) are depicted. B. Overexpression of menin increases Wnt/h-catenin signaling – associated Axin2 gene promoter H3K4 trimethylation. Top, agarose gels with PCR products from the chromatin immunoprecipitation assay. TGP-61 cells were transfected with empty expression vector (UV), active h-catenin (S37A) expression vector, or h-catenin (S37A)/menin expression vectors as shown. Forty-eight hours later, the chromatin immunoprecipitation assay was done with a trimethyl specific anti-H3K4 antibody. The T2/T3 promoter region was PCR amplified as depicted in the gene diagram above. Bottom, input DNA loading for the PCR. C. Menin expression knockdown abrogates Axin2 gene promoter H3K4 trimethylation in response to Wnt/h-catenin signaling. Agarose gels with PCR products from the chromatin immunoprecipitation assay showing the effect of Men1 siRNA on H3K4 methylation of the Axin2 gene promoter region T2/T3 in response to Wnt/h-catenin signaling are shown. TGP61 cells were first transfected with either control or specific Men1 siRNA. Twenty-four hours later, cells were then transfected with control expression vector (UV)orah-catenin (S37A) expression vector. Chromatin immunoprecipitation assays were done 48 h later with either anti – trimethyl H3K4 antibody or anti-menin antibody. The fixed cells were also lysed and analyzed by Western blotting (IB) with anti-menin antibody and anti – h-tubulin antibody to show reduced expression of menin with Men1 siRNA.

cell proliferation because loss of menin function is thought to this is due to an increase in G2-M phase arrested cells increase endocrine tumor cell proliferation. Indeed, TGP61 islet (Supplementary Fig. S3). tumor cells transfected with menin siRNA show increased cell proliferation compared with control transfected cells (Supple- Discussion mentary Fig. S2). To determine whether Wnt signaling affects In this study, we show that menin is essential for Wnt/h- TGP16 rodent islet tumor cell proliferation, TGP61 cells were catenin signaling in a mouse islet tumor cell line. The molecular treated with a GSK3h inhibitor (BIO, Calbiochem) that mechanism involves a direct interaction between h-catenin and increases Wnt pathway signaling by decreasing h-catenin the COOH terminus of menin at the target gene promoter proteasome–mediated degradation (Fig. 8A). Western blot (Axin2 in this study) and resultant histone H3K4 trimethylation. analysis shows that GSK3h inhibitor treatment increased Menin COOH-terminal truncation mutants lose their ability to activated h-catenin expression compared with vehicle-treated interact with h-catenin and their ability to increase TCF gene cells. In addition, cell counting at 0, 24, 48, and 72 hours reporter activity, which is consistent with a requirement for showed decreased cell proliferation compared with vehicle- inactivation of Wnt signaling during islet tumor development. treated cells (Fig. 8B). Flow cytometry analysis suggests that Although these studies show that histone H3K4 trimethylation

is affected at the Axin2 promoter, we do not know whether this been proposed to have a direct role in bone development is a direct or indirect effect. (reviewed in ref. 43). Because Wnt/h-catenin has also been Because menin is inactivated in MEN1-associated endo- implicated in bone development, it is possible that menin is crine tumors, we hypothesize that canonical Wnt signaling critical for proper osteoblast differentiation, in part, through could be diminished in these tumors. This contradicts the Wnt signaling and h-catenin/menin transcriptional activation of typical role of Wnt activation in several malignancies such as RUNX2, the master regulator of bone formation (see review in colon cancer (25, 37, 38) and recent data suggesting that ref. 44). Wnt signaling is important for h-cell proliferation (39, 40). In the absence of Wnt signaling, TCF/LEF transcription Most tumors with Wnt signaling abnormalities reported factors form a complex with Groucho/Grg/transducin-like reveal activation of this pathway, but a few tumors such as enhancer-of-split proteins and thereby function as a transcrip- salivary gland tumors have inactivation of the Wnt pathway tion repressor (reviewed in ref. 26). With Wnt signaling, h- (41). Menin inactivation could contribute to tumorigenesis by catenin displaces Groucho from the TCF/LEF complex and disabling the canonical Wnt pathway. Indeed, activation of thereby activates transcription. The mechanism for transcription Wnt/h-catenin signaling in TGP61 cells resulted in decreased activation is hypothesized to involve interaction with coac- cell proliferation. This hypothesis is supported by mouse tivators such as the histone acetylase, cyclic AMP-responsive models showing that Wnt/h-catenin inactivation results in the element binding protein–binding protein, and brahma-related loss of pancreas acinar cell development, but promotes or at gene 1, which is a component of the SWI/SNF family of the least preserves pancreas islet development (22). None of ATP-dependent chromatin remodeling complexes. Histone these transgenic models of Wnt signaling/h-catenin signaling methylation is also thought to be involved (45). Because menin develop endocrine tumors, suggesting that Wnt signaling may interacts with both h-catenin and the histone methyltransferase not be critical for MEN1-associated tumor development. A complex, we hypothesize that menin is important for recruit- Men1 knockout mouse model, however, showed that h- ment of MLL histone methyltransferase to the h-catenin/TCF catenin is predominantly cytoplasmic, rather than nuclear, in activator complex at the gene promoter. the islet tumors (42). Interestingly, parafibromin, which is the tumor suppressor Men1 homozygous knockout mice, however, do have protein mutated in hyperparathyroidism-jaw tumor syndrome 2, craniofacial bone development abnormalities, and menin has also binds h-catenin and regulates Wnt target gene transcription

FIGURE 8. Proliferation of TGP61 cells is reduced by GSK3h inhibitor treatment. TGP61 cells were plated in triplicate for each time point and treated with either vehicle or GSK3h inhibitor IX (BIO from Calbiochem) as described in Materials and Methods. Cell cultures were split for counting and Western blot analysis. A. Western blot was done to determine the endogenous expression of activated h-catenin, menin, and h-tubulin after treatment with GSK3h inhibitor (5 Amol/L BIO, Calbiochem). Activated (dephosphorylated) h-catenin was detected with an antibody that recognizes h-catenin dephosphorylated on Ser37 or Thr41 (Upstate). Western blot with anti – h-tubulin antibody was used as a control for protein gel loading. B. Graph of TGP61 cell proliferation (triplicate cultures) shows cell counting at 0, 24, 48, and 72 h after plating. E, cells treated with GSK3h inhibitor (5 Amol/L BIO); ., vehicle-treated cells. Points, mean cell count for triplicate cultures; bars, SD.

through a mechanism thought to involve the PAF1 complex Laboratories, Inc.), probed with a random primed radiolabeled (Paf1, Cdc73, Leo1, Ctr9, and Rtf1; ref. 46). Hence, MEN1 and probe (Boehringer Mannheim), and PCR amplified (PCR hyperparathyroidism-jaw tumor syndrome 2 may share a Advantage, Clontech Laboratories) using the 500-bp genomic similar signaling pathway mechanism involving histone fragment from the first exon of human menin. The open reading methylation and Wnt signaling. frame of this full-length E phage cDNA clone (internal lab We found that Axin2 expression is regulated in our mouse designated clone E$-menin-1.2.3) was high-fidelity PCR islet tumor cell culture model for studying menin function. This amplified (PCR Advantage) with primers 5¶-gaatt- is not surprising because Axin2 is considered a possible cATGGGGCTGAAGGCC (small case: EcoRI restriction site) universal Wnt/h-catenin target, perhaps as part of autoregulation and 5¶-TCAGAGGCCTTTGCG, and then TA cloned into of this pathway (25). It will be important to identify islet-specific vector pCR2.1 (Invitrogen). Fidelity of the PCR DNA targets, if they exist, and further to understand the role of Wnt amplification and cloning was confirmed by fully sequencing signaling in islet cell biology as well as tumor development. the entire pCR2.1-menin (internal lab clone E4.1) insert in both directions. The pcDNA3.1/HisC-menin construct was created by subcloning the MEN1 open reading frame fragment from Materials and Methods pCR2.1-menin into the EcoRI site of the pcDNA3.1/HisC Reagents and Cell Lines expression vector (Invitrogen). An expression construct for Hamster islet tumor cells InR1G9 (kind gift from Dr. Craig hemagglutinin epitope–tagged menin (pcDNA3.1-HA-menin) Smith) and human embryonic kidney cells (HEK 293T from was created by PCR DNA amplification of full-length American Type Culture Collection) were cultivated in DMEM MEN1 open reading frame (pCR2.1-menin) using primers (Mediatech, Inc.) supplemented with 10% fetal bovine serum 5¶-gaattcGCCATGTACCCATACGATGTTCCAGATTACGCT- (Omega, Inc.), 2 mmol/L L-glutamine, 100 units/mL penicillin, TACCCATACGATGTTCCAGATTACGCTGGGCT- and 100 Ag/mL streptomycin (Omega). Mouse islet tumor cells GAAGGCCGCC (2xHA epitope underlined; EcoRI site shown (TGP-61 from American Type Culture Collection) were in small case) and 5¶-CCggatccTTCAGAGGCCTTTGCG cultivated in 50% DMEM/50% F12K medium (Mediatech) (small case: BamHI site not used for this construct). The supplemented with 10% fetal bovine serum and L-glutamine/ 2xHA-menin PCR fragment was TA cloned back into pCR2.1 penicillin/streptomycin. The rat monoclonal anti-hemagglutinin (pCR2.1-HA-menin). The EcoRI-EcoRI fragment from antibody (12CA5) was purchased from Roche Applied Science. pCR2.1-HA-menin was subcloned into the EcoRI site of The monoclonal anti-FLAG (M2) antibody was purchased from pcDNA3.1 (Invitrogen). Sigma-Aldrich. The rabbit polyclonal anti-menin antibody Hemagglutinin epitope–tagged menin COOH-terminal de- (BL342) and the goat anti-GST antibody were purchased from letion mutants, pcDNA3.1-HA-meninDB (delete amino acids Bethyl Laboratories. The mouse monoclonal anti-His G 263-610), meninDC (delete amino acids 341-610), meninDD antibody was purchased from Invitrogen Corp. The antimouse (delete amino acids 460-610), and meninDE (delete amino acids Axin2 antibody (ab32197) was purchased from Abcam. The 527-610), were generated from pCR2.1-HA-menin (see above) TCF/LEF reporter plasmid Super8XTOPflash (M50) and its by site-directed mutagenesis (48) and subcloning into the control mutant plasmid Super8XFOPflash (M51) were kind gifts expression vector pcDNA3.1. from Dr. Randall T. Moon. The human MEN1 full-length wild- The menin NH2-terminal deletion mutants pcDNA3.1/HisC- type and missense mutation expression constructs in the vector meninD2 (delete amino acids 1-197), meninD3 (delete amino pCMV-SPORT were a kind gift from Dr. Sunita Agarwal. acids 1-276), meninD4 (delete amino acids 1-382), and meninD5 (delete amino acids 1-443) were generated from siRNAand Transfection pCR2.1-menin (see above) by subcloning respectively the siRNAs were purchased from Dharmacon RNA Technol- blunted HincII-BamHI fragment, blunted BstEII-BamHI frag- ogies. Mouse Men1 siRNA SMARTpools were optimized to ment, blunted KpnI-BamHI fragment, blunted BstXI-BamHI find the best suppressor of menin expression (optimum mouse fragment, and blunted Avr II-BamHI fragment into pcDNA3.1/ Men1 siRNA sequence: GCUAAGACCUACUACCAGGUU) HisC. The pcDNA3.1HisB-meninD6 (delete amino acids 1-526) determined by reverse transcription-PCR (RT-PCR) and mutant was created by PCR DNA amplification using the primer Western blot analysis. Non-Specific Control-V siRNA was 5¶-CgaattcCCGAGGCCCTGAAGG (small case: EcoRI site) used as a siRNA control. Cells were transfected using and vector BGH reverse primer 5¶-TAGAAGGCACAGTC- Oligofectamine (Invitrogen) transfection reagent. For experi- GAGG, and then the PCR fragment was cloned into the ments requiring inhibition of Wnt signaling, the cells were pcDNA3.1HisB vector (Invitrogen). h mock treated with vehicle or treated by GSK3 inhibitor IX The GST-menin NH2-terminal deletion mutant pGEX4T2- (Calbiochem) 5 Amol/L for 1 h before harvesting the cells. meninD5s (delete amino acids 1-476) was created by subclon- Reduction of mouse Men1 mRNA expression was confirmed ing the SmaI-NotI fragment of pcDNA3.1-menin into the by RT-PCR and mouse menin protein expression was bacteria expression vector pGEX4T2 (GE Healthcare Bio- confirmed by Western blot analysis. Sciences Corp.) SalI-blunted/NotI restriction sites. Two menin missense mutant expression constructs, pCMV- Construction of Expression Plasmids SPORT-Men-P12L and pCMV-SPORT-Men-L22R, were gen- Recombinant expression plasmids were constructed as erated from pCMV-SPORT-menin (kind gift from Dr. Sunita described (47). Briefly, a full-length MEN1 cDNA was isolated Agarwal) using the splice overlap extension method (48) with by screening a human PBL E phage cDNA library (Clontech the primers P12Lf (5¶-GACGCTGTTCCTGCTGCGCTC),

P12Lr (5¶-GAGCGCAGCAGGAACAGCGTC), L22Rf (5¶- cells 24 h before transfection. Cells were transfected with GGTGCGCCGGTTTGCTGCCGA), and L22Rr (5¶-GCAG- different plasmids using Effectene Transfection Reagent CAAACCGGCGCACCACG; underlined and boldfaced letter (Qiagen, Inc.). For HEK 293T cells, 2 to 4 Ag of plasmid and is the mutated nucleotide). 15 AL of Effectene/Ag plasmid were used for each transfection. The cDNA coding for wild-type h-catenin was obtained from For TGP-61 and InR1G9 cells, 4 to 6 Ag of plasmid and 25 AL IMAGE clone 6151332 (Invitrogen), and the constitutively of Effectene/Ag plasmid were used for each transfection. active missense mutant pCMV-SPORT-h-catenin (S37A) was a Cotransfection with empty vector DNA was done as needed kind gift from Dr. Eric Vilain. The open reading frames of wild- to keep the amount of transfected DNA consistent between type h-catenin and mutant S37A were subcloned into the samples in each experiment. mammalian expression vectors p3XF10 (Sigma-Aldrich) and pECFP-N1 (Clontech Laboratories) by PCR-TA (Invitrogen) Total Cell Protein Extraction h cloning using the following primers: for p3XF10- -catenin, Plated cells were washed twice in cold 1Â PBS (without ¶ HCATf 5 -ggtaccATGGCTACTCAAGCTGATTTG (small case: calcium) buffer and then lysed in radioimmunoprecipitation ¶ KpnI site) and HCATr 5 -tctagaTTACAGGTCAGTAT- assay buffer containing 50 mmol/L Tris-HCl (pH 7.4), h CAAACCA (small case: XbaI site); for pECFP- -catenin, 200 mmol/L NaCl, 1 mmol/L EDTA, 1% NP40, 0.5% sodium CATNOr 5¶-gatatcCCAGGTCAGTATCAAACCAGGC (small deoxycholate, 0.5% SDS, 2 mmol/L NaVO4, 2 mmol/L NaF, case: EcoRV site). The full-length cDNAs for TCF3 (also called and 1 mmol/L phenylmethylsulfonyl fluoride supplemented TCF7L1) and TCF4 (also called TCF7L2) were purchased from with Complete Protease Inhibitor Cocktail as recommended by Invitrogen (IMAGE clones 6141641 and 5533185, respectively) the manufacturer (Roche Applied Science). After a 30-min and PCR-TA cloned into plasmids p3XF10, pcDNA3.1/HisB, incubation on ice, the cell lysate was sonicated twice on ice pEYFP-C1 (Clontech Laboratories), and pECFP-N1 using the (8 s), centrifuged (13,000 rpm at 4jC) for 15 min, and the ¶ following PCR primers: TCF3f, 5 -gaattcCACCATGCCC- supernatant was stored at À80jC for Western blotting. CAGCTCG (small case: EcoRI site); TCF3r, 5¶-gatatcT- TAGTGGGCAGACTTGGTG (small case: EcoRV site); TCF3Nr, 5¶-gatatcCGTGGGCAGACTTGGTGACC (small Coimmunoprecipitation and Western Blot Analysis case: EcoRV site); TCF4f, 5¶-gaattcAAAAATGCCGCAGCT- After 48 h posttransfection, the plated cells were harvested GAAC (small EcoRI case site); and TCF4r, 5¶-gatatcTAG- for protein as above. The protein lysates were precleared with TAAGCTTCCATCTGAAGA (small case: EcoRV site). protein G-Sepharose 4B (Sigma-Aldrich) for 30 min and then incubated for 2 h with rat monoclonal anti-hemagglutinin or GST Pull-Down Assay mouse monoclonal anti-HisG. Untransfected cells were treated h Pull-down protein interaction studies were done according to similarly but immunoprecipitated with anti–activated -catenin the manufacturer’s recommendations (ProFound Pull-Down antibody (Upstate clone 8E7). Immune complexes were GST Protein:Protein Interaction Kit, Pierce Biotechnology). captured with protein G-Sepharose 4B beads for an additional Briefly, the competent bacteria (BL21 strain) were transformed hour, centrifuged, washed four times in radioimmunoprecipita- with either pGEX4T2 or pGEX4T2-meninD5s plasmids and tion assay buffer, and then solubilized in SDS sample buffer. grown under selective conditions to produce the ‘‘bait’’ protein. Proteins were analyzed on precast 4% to 12% gel (Invitrogen) One colony from each transformants was scraped and grown in and immunoblotted with mouse monoclonal anti-FLAG suspension at 28jC overnight. Bacteria were diluted into fresh antibody. The immune complexes were detected by horserad- media the next day, and the protein expression was isopropyl-h- ish-conjugated secondary antibody and developed by enhanced D-1-thiogalactopyranoside induced when the bacterial culture chemiluminescence (Amersham Pharmacia Biotech). Images were digitized with Quantity One Software (Bio-Rad) using the attained an A600 z0.6. Induced bacteria were isolated by centrifugation and lysed in the provided buffer. GST-tagged Bio-Rad Versadoc Imaging System Model 3000. The images proteins were immobilized on the glutathione-coated agarose were prepared for publication with Adobe Photoshop and beads according to the manufacturer’s protocol. Biotinylated Illustrator software. (Transcend Non-Radioactive Translation Detection System from Promega) full-length h-catenin, TCF3, and TCF4 proteins Luciferase Assays were made by in vitro transcription/translation reaction of All luciferase assay experiments were done in triplicate, with pSPORT-h-catenin (construct from Invitrogen), pcDNA3.1/ data reported as mean F SD. Twenty-four-well plates were HisB-TCF3, and pcDNA3.1/HisB-TCF4 following the manu- seeded with 3 Â 104 TGP-61 cells 24 h before transfection. facturer’s recommendations (Promega, TnT Quick Coupled Cells were transfected with different plasmids (including M50, Transcription/Translation Systems) and added as ‘‘prey’’ to the M51, and pRL-hTK) using Effectene Transfection Reagent. An complex of beads with immobilized GST epitope–tagged equal amount of transfected plasmid DNA was ensured between ‘‘bait.’’ The bead complexes were washed and the bound different samples in each experimental group by adding empty ‘‘prey’’ proteins were eluted for SDS-PAGE and further control vector plasmid DNA as needed. Twenty-four hours after analysis by Western blot. transfection, the medium was changed to fresh medium containing 5% charcoal stripped serum. Twenty-four hours Transient Transfection later, the cells were lysed in 1Â PLB, and luciferase assays Culture dishes (100 mm) were seeded with either 2 Â 106 were done using the Dual-Luciferase Reporter Assay System as HEK 293T cells, 1 Â 106 TGP-61 cells, or 2 Â 106 InR1G9 recommended by the manufacturer (Promega). SDS-PAGE

sample buffer was added to the remaining lysate, and Western TCGTAGG; 5hMET, 5¶-GATCAACTCATTAGCTGTGGC; blotting was done to ensure every transfected plasmid was 5mMET, 5¶-GCAGCAGCAAAGCCAAT; 3MET, 5¶-TGAA- expressed for each samples. AAGTCTGAGCATCTAGAGT; 5HDAC9, 5¶-GGAGCC- CATCTCACCTTTAGACC; 3HDAC9, 5¶-TGCCACTGCC- Chromatin Immunoprecipitation Assay CTTTCTCGTC; 5AXIN2, 5¶-GCCGATTGCTGAGAGGA- The chromatin immunoprecipitation assays were done ACTG; 3AXIN2, 5¶-AAAGTTTTGGTATCCTTCAGGTT- following the protocol described with some minor modifica- CAT; 5CMYC, 5¶-CAGGAACTATGACCTCGACTACGACT; tions (49). Briefly, after different transfections and treatments, 3CMYC, 5¶-TGTCTTGGCCAGCC; 5PDX1, 5¶-GGAGCAG- 37% formaldehyde (Fisher Scientific) was added directly to TACTACGCGGCCA; 3PDX1, 5¶-TGGCCTTTCCACG- culture medium in the plate to a final concentration of 1% to CGTGA; 5LMX1A, 5¶-CCAAGTCTGTCTGCGAGGGC; cross-link the proteins to the DNA. After incubation for 10 min 3LMX1A, 5¶-TGCAGGGCTTGGAGGATACTTC; 5HNF-3B, at 37jC, cells were washed twice using ice-cold PBS 5¶-GCCGGCCTGGGGATGAA; 3HNF-3B, 5¶-TGTTGGGG- containing protease inhibitor cocktail. Cells were scrapped CTCTGCTGGATG; 5CDX1, 5¶-CAGGGCCCAGCATGCG; from the plate and harvested into 1.5-mL Eppendorf tubes and and 3CDX1, 5¶-TCTTACCGCTGCCACCGC. then briefly centrifuged. Cell pellets were lysed in 800-AL SDS lysis buffer containing 1% SDS, 10 mmol/L EDTA, and 50 mmol/L Tris-HCl (pH 8.1) with freshly added protease Microscopy Analysis inhibitor cocktail. After a 10-min incubation on ice, the lysates Forty-eight hours after transfection with fluorescent protein were sonicated thrice for 10 s on ice and centrifuged. Then the expression constructs, TGP-61 cells grown on coverslips were supernatant (200 AL) was diluted with 1,800 AL of chromatin washed twice with cold PBS (Sigma) and fixed with 0.5% immunoprecipitation dilution buffer containing 0.01% SDS, paraformaldehyde for 30 min. Fixed cells were washed twice more 1.1% Triton X-100, 1.2 mmol/L EDTA, 16.7 mmol/L Tris-HCl with cold PBS and incubated with 1 ng/mL bisbenzimide H 33258 (pH 8.1), and 167 mmol/L NaCl, with freshly added protease (Sigma). Coverslips were mounted on slides using Vectashield inhibitor cocktail. An aliquot of diluted sample (200 AL) was (Vector Labs) and viewed under a microscope (Leica DMIL) with saved as input control DNA and, after adding 8 AL of 5 mol/L appropriate band-pass filters (Chroma) for ECFP, EYFP, and NaCl, reverse cross-linked at 65jC overnight. The remaining Hoechst stain using a 63Â objective. Images were captured with a samples were precleared with 50 AL of salmon sperm DNA/ Hamamatsu Orca II cooled charge-coupled device camera and data protein G-Sepharose-50% slurry for at least 30 min at 4jC with were analyzed with Metamorph (Molecular Devices) software to agitation. The supernatant fractions were collected and either create merged images. The final images were assembled in Adobe anti–trimethyl-H3K4 (Upstate USA, Inc.) or anti-menin Photoshop and Illustrator software to create the publication images. antibody was added for immunoprecipitation with rotation at 4jC overnight followed by adding 50 AL of salmon sperm DNA/protein G-Sepharose slurry for 1 h at 4jC with rotation to Cell Counting collect the immunocomplex. For a negative control, a no- All cell counting experiments were done in triplicate, with data Â 5 antibody immunoprecipitation was done by incubating the reported as mean F SD. TGP61 cells (2.4 10 ) were plated on supernatant fraction with 50 AL of salmon sperm DNA/protein six-well plates. The following day, the medium was changed to G-Sepharose slurry for 1 h at 4jC before washing the beads. fresh medium containing 5% charcoal-stripped serum. One set of The samples were washed and eluted, and the DNA was cells were trypsinized and counted on a hemacytometer, and the extracted with a commercial DNA extraction kit (Qiagen). The remaining cells were treated with vehicle (DMSO) or 5 Amol/L DNA was eluted in 30 AL Tris (pH 8.5) buffer, PCR amplified GSK3h inhibitor for 1, 2, and 3 d. Each day, the medium was with AccuPrime Taq polymerase (Invitrogen), and analyzed by changed with fresh medium containing 5% charcoal-stripped electrophoresis on agarose gel. The primers used to amplify the serum and treatment was continued with GSK3h inhibitor. mouse Axin2 promoter region, designated as T2/3, were mt23F, 5 ¶-GCGGCGGGATCACTGGCT, and mt23R, 5¶- Flow Cytometry TCCTCCGGGCGCTTCCAAC (36). Adherent cells were detached and collected with 0.25% trypsin (Sigma). Cells were pelleted by centrifugation and RT-PCR resuspended in propidium iodide staining buffer (sodium citrate HEK 293T cells and TGP-61 cells were plated in a six-well 250 mg, 0.75 mL Triton X-100, propidium iodide 25 mg, plate 24 h before transfection. Cells were transfected with RNase A 5 mg, and distilled H2O to final volume of 250 mL) different expression plasmids using Effectene Transfection and stored in the dark at 4jC until flow cytometry (<1 h). Cells Reagent. Two days later, the total RNA was extracted with were then passed through a FACScan cytometer (Becton- Trizol (Invitrogen). The first cDNA was synthesized using the Dickinson) controlled with CellQuest software and the data 1st cDNA Synthesis Kit (GE Healthcare Lifesciences), and were analyzed with the ModFit software. followed by RT-PCR with Taq DNA polymerase (Invitrogen). The primers used were 5LEF1, 5¶-GCCGAGATCAGTCATC- Statistical Analysis CCGA; 3LEF1, 5¶-CACCACGGGCACTTTATTTGAT; 5TAS- Microsoft Excel software was used for data management PASE, 5¶-AACGAGCTTGTCAGAAGGCAATT; 3TASPASE, (calculating mean and SD) and creating graphs. All luciferase 5¶-CCACCCTTTCTGCCAGCTCTA; 5NKX2-2, 5¶-CTGAC- assays were done with triplicate plasmid transfections, with data CAACACAAAGACGGGG; 3NKX2-2, 5¶-GCCGCTCCAGC- presented as mean F SD.

22. Murtaugh LC, Law AC, Dor Y, Melton DA. B-catenin is essential for Disclosure of Potential Conflicts of Interest pancreatic acinar but not islet development. Development 2005;132:4663 – 74. No potential conflicts of interest were disclosed. Epub 2005 Sep 4628. 23. Dessimoz J, Bonnard C, Huelsken J, Grapin-Botton A. Pancreas-specific Acknowledgments deletion of B-catenin reveals Wnt-dependent and Wnt-independent functions We thank Drs. Gregory Brent, Dean Yamaguchi, Steven Pandol, Ilya Gukovsky, during development. Curr Biol 2005;15:1677 – 83. and Matthias Stelzner at the West Los Angeles VA and David Geffen School of 24. Ireland H, Kemp R, Houghton C, et al. Inducible Cre-mediated control of Medicine at University of California at Los Angeles for valuable scientific gene expression in the murine gastrointestinal tract: effect of loss of B-catenin. discussions. We are grateful for critical reagents provided by Drs. Sunita Agarwal Gastroenterology 2004;126:1236 – 46. (National Institutes of Diabetes, Digestive and Kidney Diseases, Bethesda, MD), B Randall T. Moon (University of Washington, Seattle, WA), Eric Vilain (University 25. Clevers H. Wnt/ -catenin signaling in development and disease. Cell 2006; of California at Los Angeles, Los Angeles, CA), and Craig V. Smith (University 127:469 – 80. of California at Los Angeles, Los Angeles, CA). 26. Stadeli R, Hoffmans R, Basler K. Transcription under the control of nuclear Arm/B-catenin. Curr Biol 2006;16:R378 – 85. 27. Heller RS, Klein T, Ling Z, et al. Expression of Wnt, Frizzled, sFRP, DKK References genes in adult human pancreas. Gene Expr 2003;11:141 – 7. 1. Schussheim DH, Skarulis MC, Agarwal SK, et al. Multiple endocrine neoplasia type 1: new clinical and basic findings. Trends Endocrinol Metab 2001; 28. Heller RS, Dichmann DS, Jensen J, et al. Expression patterns of Wnts, 12:173 – 8. Frizzleds, sFRPs, and misexpression in transgenic mice suggesting a role for Wnts in pancreas and foregut pattern formation. Dev Dyn 2002;225:260 – 70. 2. Yang Y, Hua X. In search of tumor suppressing functions of menin. Mol Cell Endocrinol 2007;265 – 6:34 – 41. 29. Pettengill OS, Memoli VA, Brinck-Johnsen T, Longnecker DS. Cell lines derived from pancreatic tumors of Tg(Ela-1-SV40E)Bri18 transgenic mice 3. Guo SS, Sawicki MP. Molecular and genetic mechanisms of tumorigenesis in express somatostatin and T antigen. Carcinogenesis 1994;15:61 – 5. multiple endocrine neoplasia type-1. Mol Endocrinol 2001;15:1653 – 64. 4. Bertolino P, Radovanovic I, Casse H, Aguzzi A, Wang ZQ, Zhang CX. 30. Pear WS, Nolan GP, Scott ML, Baltimore D. Production of high-titer helper-free retroviruses by transient transfection. Proc Natl Acad Sci U S A 1993;90:8392–6. Genetic ablation of the tumor suppressor menin causes lethality at mid-gestation with defects in multiple organs. Mech Dev 2003;120:549 – 60. 31. Takaki R, Ono J, Nakamura M, et al. Isolation of glucagon-secreting cell 5. Crabtree JS, Scacheri PC, Ward JM, et al. A mouse model of multiple lines by cloning insulinoma cells. In Vitro Cell Dev Biol 1986;22:120 – 6. endocrine neoplasia, type 1, develops multiple endocrine tumors. Proc Natl Acad 32. Veeman MT, Slusarski DC, Kaykas A, Louie SH, Moon RT. Zebrafish Sci U S A 2001;98:1118 – 23. prickle, a modulator of noncanonical Wnt/Fz signaling, regulates gastrulation 6. Agarwal SK, Guru SC, Heppner C, et al. Menin interacts with the AP1 movements. Curr Biol 2003;13:680 – 5. transcription factor JunD and represses JunD-activated transcription. Cell 1999; 33. Sadot E, Conacci-Sorrell M, Zhurinsky J, et al. Regulation of S33/S37 96:143 – 52. phosphorylated B-catenin in normal and transformed cells. J Cell Sci 2002;115: 7. Naito J, Kaji H, Sowa H, Hendy GN, Sugimoto T, Chihara K. Menin 2771 – 80. suppresses osteoblast differentiation by antagonizing the AP-1 factor, JunD. 34. Molenaar M, van de Wetering M, Oosterwegel M, et al. XTcf-3 transcription J Biol Chem 2005;280:4785 – 91. Epub 2004 Nov 4723. factor mediates B-catenin-induced axis formation in Xenopus embryos. Cell 1996; 8. Kaji H, Canaff L, Lebrun JJ, Goltzman D, Hendy GN. Inactivation of menin, a 86:391 – 9. Smad3-interacting protein, blocks transforming growth factor type B signaling. 35. Hallikas O, Palin K, Sinjushina N, et al. Genome-wide prediction of Proc Natl Acad Sci U S A 2001;98:3837 – 42. Epub 2001 Mar 3813. mammalian enhancers based on analysis of transcription-factor binding affinity. 9. Sowa H, Kaji H, Canaff L, et al. Inactivation of menin, the product of the Cell 2006;124:47 – 59. multiple endocrine neoplasia type 1 gene, inhibits the commitment of 36. Jho EH, Zhang T, Domon C, Joo CK, Freund JN, Costantini F. Wnt/h- multipotential mesenchymal stem cells into the osteoblast lineage. J Biol Chem catenin/Tcf signaling induces the transcription of Axin2, a negative regulator of 2003;278:21058 – 69. Epub 22003 Mar 21020. the signaling pathway. Mol Cell Biol 2002;22:1172 – 83. 10. Sowa H, Kaji H, Hendy GN, et al. Menin is required for bone morphogenetic 37. Moon RT, Kohn AD, De Ferrari GV, Kaykas A. WNT and h-catenin protein 2-and transforming growth factor B-regulated osteoblastic differentiation signalling: diseases and therapies. Nat Rev Genet 2004;5:691 – 701. through interaction with Smads and Runx2. J Biol Chem 2004;279:40267 – 75. 38. Nusse R. Wnt signaling in disease and in development. Cell Res 2005;15:28– 32. Epub 42004 May 40218. 39. D’Amour KA, Bang AG, Eliazer S, et al. Production of pancreatic hormone- 11. Heppner C, Bilimoria KY, Agarwal SK, et al. The tumor suppressor protein expressing endocrine cells from human embryonic stem cells. Nat Biotechnol K K menin interacts with NF- B proteins and inhibits NF- B-mediated trans- 2006;24:1392 – 401. activation. Oncogene 2001;20:4917 – 25. 40. Rulifson IC, Karnik SK, Heiser PW, et al. Wnt signaling regulates pancreatic 12. Yokoyama A, Wang Z, Wysocka J, et al. Leukemia proto-oncoprotein MLL h cell proliferation. Proc Natl Acad Sci U S A 2007;104:6247 – 52. forms a SET1-like histone methyltransferase complex with menin to regulate Hox gene expression. Mol Cell Biol 2004;24:5639 – 49. 41. Takeda H, Lyle S, Lazar AJ, Zouboulis CC, Smyth I, Watt FM. Human sebaceous tumors harbor inactivating mutations in LEF1. Nat Med 2006;12:395– 7. 13. Hughes CM, Rozenblatt-Rosen O, Milne TA, et al. Menin associates with a trithorax family histone methyltransferase complex and with the hoxc8 locus. Mol 42. Bertolino P, Tong WM, Herrera PL, Casse H, Zhang CX, Wang ZQ. h Cell 2004;13:587 – 97. Pancreatic -cell-specific ablation of the multiple endocrine neoplasia type 1 (MEN1) gene causes full penetrance of insulinoma development in mice. Cancer 14. Yokoyama A, Somervaille TC, Smith KS, Rozenblatt-Rosen O, Meyerson M, Res 2003;63:4836 – 41. Cleary ML. The menin tumor suppressor protein is an essential oncogenic h cofactor for MLL-associated leukemogenesis. Cell 2005;123:207 – 18. 43. Hendy GN, Kaji H, Sowa H, Lebrun JJ, Canaff L. Menin and TGF- superfamily member signaling via the Smad pathway in pituitary, parathyroid and 15. Milne TA, Hughes CM, Lloyd R, et al. Menin and MLL cooperatively osteoblast. Horm Metab Res 2005;37:375 – 9. regulate expression of cyclin-dependent kinase inhibitors. Proc Natl Acad Sci 44. Li YL, Xiao ZS. Advances in Runx2 regulation and its isoforms. Med U S A 2005;102:749 – 54. Epub 2005 Jan 2007. Hypotheses 2007;68:169 – 75. 16. Scacheri PC, Davis S, Odom DT, et al. Genome-wide analysis of menin binding provides insights into MEN1 tumorigenesis. PLoS Genet 2006;2:e51. 45. Sierra J, Yoshida T, Joazeiro CA, Jones KA. The APC tumor suppressor counteracts h-catenin activation and H3K4 methylation at Wnt target genes. 17. Habener JF, Kemp DM, Thomas MK. Minireview: transcriptional regulation Genes Dev 2006;20:586 – 600. in pancreatic development. Endocrinology 2005;146:1025 – 34. Epub 2004 Dec 46. Mosimann C, Hausmann G, Basler K. Parafibromin/Hyrax activates Wnt/Wg 1016. target gene transcription by direct association with h-catenin/Armadillo. Cell 18. Kim SK, MacDonald RJ. Signaling and transcriptional control of pancreatic 2006;125:327 – 41. organogenesis. Curr Opin Genet Dev 2002;12:540 – 7. 47. Chen G, Ray R, Dubik D, et al. The E1B 19K/Bcl-2-binding protein Nip3 is 19. Kim HJ, Schleiffarth JR, Jessurun J, et al. Wnt5 signaling in vertebrate a dimeric mitochondrial protein that activates apoptosis. J Exp Med 1997;186: pancreas development. BMC Biol 2005;3:23. 1975 – 83. 20. Papadopoulou S, Edlund H. Attenuated Wnt signaling perturbs pancreatic 48. Dieffenbach CW, Dveksler GS. PCR primer: a laboratory manual. Plainview growth but not pancreatic function. Diabetes 2005;54:2844 – 51. (NY): Cold Spring Harbor Laboratory Press; 1995. p. xii, 714. 21. Dessimoz J, Grapin-Botton A. Pancreas development and cancer: Wnt/B- 49. Becker PB. Chromatin protocols. Totowa (NJ): Humana Press, 1999. catenin at issue. Cell Cycle 2006;5:7 – 10. Epub 2006 Jan 2004. p. xv, 528.

Mol Cancer Res 2008;6(12). December 2008 Downloaded from mcr.aacrjournals.org on September 23, 2021. © 2008 American Association for Cancer Research. Menin Promotes the Wnt Signaling Pathway in Pancreatic Endocrine Cells

Gao Chen, Jingbo A, Min Wang, et al.

Mol Cancer Res 2008;6:1894-1907.

Updated version Access the most recent version of this article at: http://mcr.aacrjournals.org/content/6/12/1894

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2009/01/07/6.12.1894.DC1

Cited articles This article cites 47 articles, 16 of which you can access for free at: http://mcr.aacrjournals.org/content/6/12/1894.full#ref-list-1

Citing articles This article has been cited by 7 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/6/12/1894.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/6/12/1894. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.