Secreted Frizzled-Related Protein 4 Regulates Two Wnt7a Signaling Pathways and Inhibits Proliferation in Endometrial Cancer Cells

Kendra S. Carmon and David S. Loose

Department of Integrative Biology and Pharmacology, University of Texas Health Science Center Houston, Houston, Texas

Abstract Introduction In the endometrium, hormonal effects on epithelial The uterine endometrium is a complex tissue that responds cells are often elicited through stromal hormone to numerous hormones and undergoes dynamic reorganization receptors via unknown paracrine mechanisms. Several during the menstrual cycle in preparation for implantation (1). lines of evidence support the hypothesis that Wnts The development, growth, and functional differentiation of the participate in stromal-epithelial cell communication. endometrium are dependent on a dynamic collaboration Wnt7a is expressed in the luminal epithelium, whereas between its constituent stromal and epithelial cells (2). the extracellular modulator of Wnt signaling, secreted Hormonal effects on epithelial cells are often elicited through frizzled-related protein 4 (SFRP4), is localized to the stromal hormone receptors via paracrine mechanisms (3, 4). stroma. Studies have reported that SFRP4 expression is Recombination experiments using estrogen receptor a knock- significantly decreased in endometrial carcinoma and out and wild-type tissues showed that it is the stromal cell that both SFRP4 and Wnt7a genes are differentially estrogen receptors that are responsible for mediating epithelial regulated in response to estrogenic stimuli. Aberrant proliferation induced by estrogen (5). Progesterone acting on Wnt7a signaling irrevocably causes organ defects and the stromal compartment opposes estrogen-induced epithelial infertility and contributes to the onset of disease. proliferation and regulates paracrine mediators that influence However, specific frizzled receptors (Fzd) that bind epithelial gene expression. However, the complete nature of Wnt7a and the particular signal transduction pathway the paracrine signaling and the molecular basis of how stromal each Wnt7a-Fzd pair activates have not been identified. and epithelial cells communicate to one another remain largely Additionally, the function of SFRP4 in the endometrium unknown. Wnt7a is expressed in the endometrium and levels has not been addressed. We show here that Wnt7a have been shown to fluctuate in response to circulating steroid coimmunoprecipitates with Fzd5, Fzd10, and SFRP4 in hormones during the menstrual cycle and pregnancy (6). Ishikawa cells. Wnt7a binding to Fzd5 was shown to Wnt7a is required for proper differentiation and gland activate B-catenin/canonical Wnt signaling and increase formation during uterine development. Following postnatal cellular proliferation. Conversely, Wnt7a signaling growth, Wnt7a expression becomes restricted primarily to the mediated by Fzd10 induced a noncanonical c-Jun luminal epithelium and is responsible for maintaining expres- NH2-terminal kinase–responsive pathway. SFRP4 sion of other Wnts in the stroma (6-9). Given that Wnts suppresses activation of Wnt7a signaling in both an undergo hormone regulation and different Wnts are generally autocrine and paracrine manner. Stable overexpression localized to different cell types in the endometrium suggests of SFRP4 and treatment with recombinant SFRP4 protein that Wnt signaling may play a key role in the action of steroid inhibited endometrial cancer cell growth in vitro. These hormones and in mediating intercellular communication in the findings support a mechanism by which the nature of the endometrium. Wnt7a signal in the endometrium is dependent on the Members of the Wnt signaling family are highly conserved, Fzd repertoire of the cell and can be regulated by SFRP4. secreted glycoproteins that are vital to mammalian development (Mol Cancer Res 2008;6(6):1017–28) and many aspects of adult tissue homeostasis. Wnts regulate important cellular processes, such as proliferation, differentiation, and cell fate specification (10). The binding of a Wnt ligand to its corresponding seven-transmembrane frizzled Received 2/8/08; accepted 3/11/08. Grant support: Training fellowship from the Keck Center for Interdisciplinary receptor (Fzd) leads to the activation of at least one of the Bioscience Training of the Gulf Coast Consortia NIH grant #5 T90 DK070109-03 three possible signaling pathways: (a)theh-catenin/canonical and 5 R90 DK071505-03, the American Legion Auxiliary Fellowship, and pathway, (b) the planar cell polarity/c-Jun NH2-terminal kinase the NIH Uterine Cancer Specialized Programs of Research Excellence grant 2+ P50-CA098258. (JNK) pathway, and (c) the Ca -mediated pathway (10, 11). The costs of publication of this article were defrayed in part by the payment of The complexity of mammalian Wnt signaling is exemplified by page charges. This article must therefore be hereby marked advertisement in the fact that there are currently 19 known mammalian Wnts and accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Requests for reprints: David S. Loose, University of Texas Health Science 10 Fzds (10, 12). Despite the numerous studies that have been Center Houston, 6431 Fannin Street, MSB 5.132A, Houston, TX 77030. Phone: done to date, it has yet to be completely elucidated which 713-500-7440; Fax: 713-500-7455. E-mail: [email protected] Copyright D 2008 American Association for Cancer Research. particular Wnts bind to which Fzds and what signal doi:10.1158/1541-7786.MCR-08-0039 transduction pathway each Wnt-Fzd pair activates.

In the traditional canonical pathway, Wnts released from or milieu within the cell dictates which signal Wnt7a elicits within presented on the surface of signaling cells act on target cells the endometrium and that SFRP4 is a key regulator of Wnt7a by binding with high affinity to Fzd, which forms a complex action. with Lrp5/6 coreceptor (13). This interaction results in the cytoplasmic accumulation of stabilized h-catenin, which trans- Results locates to the nucleus where it can associate with members of Identification of Wnt Components in Ishikawa Cells the lymphoid enhancer factor/T-cell factor class of transcription Using microarray analysis, we assessed the relative factors and induce transcription of genes, including c-myc and expression levels for each Wnt, Fzd, and Wnt signaling cyclin D (14-17). In the absence of a Wnt signal, cytoplasmic antagonist from grouped sample data of the well-differentiated h -catenin levels are otherwise kept low through a proteasome- human endometrial adenocarcinoma Ishikawa cell line. Results mediated degradation (18). It has been established that the listed in Table 1 indicate gene symbol, detection signal, and b -catenin gene is mutated at a frequency rate of 30% in type I P value scores where significance level is P V 0.01 based on a b endometrial carcinoma (19, 20). Mutated -catenin can resist t test following background subtraction and rank invariant degradation and subsequently act as an oncogene affecting cell normalization. Our survey revealed that Wnt3, Wnt3a, and adhesion and constitutively activating Wnt target gene Wnt7a were expressed at the highest levels. It was anticipated expression. In the noncanonical planar cell polarity/JNK that Wnt7a would be expressed because it is of epithelial cell pathway, Wnt interaction with Fzd facilitates the recruitment of small GTPases RhoA/Rac/cdc42 and triggers the consequent phosphorylation and activation of c-Jun (12, 21). The Ca2+- TABLE 1. Microarray Analysis of Wnt Component Gene Expression mediated pathway involves heterotrimeric G proteins and 2+ subsequent activation of protein kinase C, Ca -calmodulin– Wnt Component Ishikawa Cells dependent kinase II, nuclear factor associated with T cells, and other downstream effectors (22-24). Gene Accession Average P Significance There are a host of diverse proteins that function as Symbol Signal extracellular modulators of Wnt signaling, including members Ligands WNT1 NM_005430 4.6 0.210 NS of the secreted frizzled-related protein (SFRP) family (25). WNT2 NM_003391 À1.0 0.457 NS SFRPs share sequence homology with the cysteine-rich domain WNT2B NM_024494 1.5 0.401 NS WNT3 NM_030753 54.0 0.001 S (CRD) of Fzd and are thought to antagonize Wnt action by WNT3A NM_033131 27.3 0.009 S competing with Fzd for the binding of Wnts and preventing WNT4 NM_030761 4.2 0.216 NS Wnt-Fzd interaction (25). The identification of specific Wnt- WNT5A NM_003392 2.6 0.364 NS WNT5B NM_030775 17.2 0.037 NS SFRP associations is another important, but ill defined, area of WNT6 NM_006522 3.9 0.226 NS Wnt research that requires a further in-depth exploration. WNT7A NM_004625 133.3 <0.001 S À SFRP4 is restricted to the endometrial stroma and its WNT7B NM_058238 1.5 0.616 NS WNT8A NM_031933 4.0 0.234 NS expression has been shown to peak during the proliferative WNT8B NM_003393 7.7 0.120 NS (or estrogen dominated) phase of the menstrual cycle. WNT9A NM_003395 8.3 0.125 NS WNT9B NM_003396 2.7 0.280 NS Quantitative reverse transcription-PCR (RT-PCR) studies done WNT10A NM_025216 4.5 0.214 NS in our laboratory found that in postmenopausal women treated WNT10B NM_003394 6.3 0.196 NS with estrogen for 3months SFRP4 expression in the WNT11 NM_004626 À6.5 0.828 NS 1 WNT16 NM_016087 2.9 0.297 NS endometrium increased 9-fold. SFRP4 expression levels are Receptors FZD1 NM_003505 101.5 <0.001 S significantly decreased in endometrial cancer (26-29). Down- FZD2 NM_001466 437.7 <0.001 S regulation of SFRP4 and associated promoter hypermethylation FZD3 NM_017412 17.0 0.030 NS FZD4 NM_01219347.1 0.003 S has been reported in several other cancers, including mesothe- FZD5 NM_003468 89.0 0.001 S lioma, chronic lymphocytic leukemia, and esophageal adeno- FZD6 NM_003506 374.1 <0.001 S FZD7 NM_003507 79.7 0.001 S carcinoma (30-32). Although it is recognized that SFRP4 plays FZD8 NM_031866 1.5 0.364 NS a role in adult uterine function and is a potential biomarker for FZD9 NM_003508 185.4 <0.001 S certain cancers, its Wnt binding partners and overall mechanism FZD10 NM_007197 9.5 0.089 NS Antagonists SFRP1 NM_003012 207.0 <0.001 S of action remain unclear. SFRP2 XM_050625 13.6 0.044 NS In the present study, we report that Wnt7a activates h- SFRP3 NA NA NA — À catenin/canonical signaling and functions to induce endometrial SFRP4 NM_003014 3.0 0.711 NS SFRP5 NM_003015 3.9 0.240 NS cell proliferation in the presence of Fzd5 but not Fzd10. Wnt7a DKK1 NM_012242 À5.6 0.812 NS signaling mediated by Fz10 was shown to stimulate the DKK2 NM_014421 À0.8 0.528 NS DKK3 NM_013253 109.3 <0.001 S noncanonical JNK-responsive pathway. Activation of both DKK4 NM_014420 1.2 0.372 NS Wnt7a-induced pathways was inhibited through the addition WIF1 NM_007191 3.7 0.248 NS of SFRP4. We also show that SFRP4 inhibition of endometrial CER1 NM_005454 0.5 0.430 NS cancer cell growth signifies its potential role as a tumor NOTE: Results of Illumina microarray analysis used to identify gene expression suppressor. These observations indicate that the Fzd receptor of Wnt components. Analysis is representative of four samples. Biological replicate data were grouped and normalized with rank invariant statistics using BeadStudio 3software. Abbreviations: S, significance of P V 0.01; NS, nonsignificance; NA, not 1Unpublished data. available.

origin in the mammalian endometrium. Among the Fzd Fzd10 successfully pulled down with Wnt7a (Fig. 2A). As a receptors, Fzd1, Fzd2, Fzd4, Fzd5, Fzd6, Fzd7, and Fzd9 negative control, lysate from cells overexpressing Fzd5 was were significantly detected. Expression of these Wnts and Fzds immunoprecipitated with anti-Wnt7a to verify that Fzd5 did not was validated using an independent set of primers by RT-PCR bind nonspecifically (Fig. 2A, middle). We also did a reverse (data not shown). Wnt signaling antagonists determined to be experiment to show that Wnt7a ligand coimmunoprecipitates appreciably expressed in Ishikawa cells included SFRP1 and with receptor by using a Fzd5 antibody (Fig. 2A, right). The DKK3, which were expressed to a level that reached signif- experiment was not repeated for Fzd10 due to the unavailability icance. The fact that SFRP4 was not expressed in Ishikawa of a quality Fzd10 antibody. These data corroborate that Wnt7a cells suggests that its expression is confined solely to the stroma interacts with both Fzd5 and Fzd10. of the endometrium. Wnt7a Activation of the b-Catenin/Canonical Pathway Localization of Wnt7a, Fzd5, Fzd10, and SFRP4 Occurs via Fzd5 Our previous quantitative RT-PCR screening studies have To determine if the Wnt7a interaction with Fzd5 and Fzd10 revealed mRNA expression of Fzd5 and Fzd10 in both human could induce h-catenin/canonical Wnt signaling, we used the 1 and rat endometrium. We investigated the cell type–specific luciferase reporter construct TopFlash. The TopFlash construct expression of Wnt7a, Fzd5, Fzd10, and SFRP4 in tissue contains T-cell factor–binding sites, which are directly sections taken from rat uteri. In situ hybridization using activated by the T-cell factor/h-catenin complex. Alternatively, antisense probes revealed specific distributions of the Wnt7a the FopFlash construct, which contains mutated T-cell factor– transcripts in the luminal epithelium and SFRP4 in the stroma binding sites, served as a negative control. Ishikawa cells were (Fig. 1). Fzd5 expression was apparent in both stromal and transiently transfected with TopFlash or FopFlash and Wnt7a, glandular epithelial cell types, with minimal expression in the Fzd receptor, or Wnt7a plus Fzd receptor. A Renilla-TK luminal epithelium. Fzd10 localization was restricted mainly to luciferase reporter was cotransfected for assay normalization. the stroma (Fig. 1) and was not detectable in the luminal Cotransfection of Wnt7a and Fzd5 was shown to significantly epithelium. Control sections treated with sense probes specific induce TopFlash activity 5.6-fold (P V 0.05; Fig. 2B). to each individual Wnt component showed no detectable Transfection of Fzd5 alone yielded a slight increase in TopFlash signals. These findings indicate a complex pattern of tissue but did not reach significance. Fzd10, Wnt7a, and Wnt7a- expression within the endometrium. Fzd10 failed to induce any reporter activity. Because Ishikawa cells contain endogenous Wnts, including Wnt7a, it is likely Interaction of Wnt7a with Fzd5 and Fzd10 that the binding of endogenous Wnts to Fzd5 accounts for the To address the question of whether Wnt7a can associate increase in TopFlash activity when Fzd5 is transfected alone. physically with the extracellular Wnt binding region of Fzd5 Altogether, it is apparent that Wnt7a activation of h-catenin/ and/or Fzd10, we did a series of coimmunoprecipitation canonical Wnt signaling is mediated through Fzd5 and not experiments. Cells were transfected with empty vector (EV) Fzd10 in Ishikawa cells. and Wnt7a or cotransfected with Wnt7a and Fzd5 or Wnt7a and As a means to test whether Wnt7a could activate h-catenin/ Fzd10 and then lysates were coimmunoprecipitated with an canonical signaling in a concentration-dependent manner, antibody directed against Wnt7a. Western blot analysis using TopFlash experiments were done using Ishikawa cells trans- antibody against the V5 tag confirmed that both Fzd5 and fected with Fzd5 and increasing amounts of the Wnt7a

FIGURE 1. Localization of Wnt components in the endometrium. In situ hybridization was carried out on tissue sections of the rat uterus using sense and antisense digoxigenin-labeled RNA probes. Wnt7a is localized to the luminal epithelium of the endometrium. Fzd5 expression is apparent in both the stroma and glandular epithelium, whereas Fzd10 and SFRP4 are confined solelyto the stroma. g, glands; le, luminal epithelium; s, stroma. Arrows designate a reas of expression.

FIGURE 2. Wnt7a coimmunoprecipitates with Fzd5 and Fzd10 and induces h-catenin/canonical Wnt signaling via Fzd5. A. Cell extracts from Ishikawa cells overexpressing Wnt7a plus or minus Fzd5 or Fzd10 were immunoprecipitated (IP) with antibodyto Wnt7a. Anti-Wnt7a immunoprecipitation of lysate from cells overexpressing Fzd5 served as a control for specificity. Cell extracts expressing Fzd5 or Fzd5 and Wnt7a were immunoprecipitated with anti-Fzd5. An EV transfection served as a control. Immunoprecipitates were immunoblotted (IB) with antibodyspecific to V5-tag to detect both Wnt7a and its associated Fzds. Fzd10 migrates at the anticipated molecular weight just above Fzd5. B. Cotransfection of Wnt7a and Fzd5 into Ishikawa cells promotes significant activation of the h-catenin TopFlash gene reporter. Wnt7a failed to activate TopFlash in the presence of Fzd10. Cells were transfected with expression vectors as indicated above along with TopFlash (or negative control construct FopFlash) and Renilla-TK for normalization. Luciferase activitywas measured 24 h after transfection. Results are representative of at least three separate experiments done in duplicate. Columns, mean; bars, SE. **, P V 0.01, compared with control EV cells based on ANOVA and Bonferroni post hoc for multiple comparisons. C. Western blot indicates expression of Fzd5 and increasing concentrations of Wnt7a protein in Ishikawa cells using anti-V5 and anti-h-actin for loading control. D. Wnt7a was shown to activate h-catenin in a dose- responsive manner. Cells were transfected with 1 Ag Fzd5 in the presence of increasing amounts of Wnt7a and TopFlash luciferase measurements were recorded 24 h after transfection. Points, mean; bars, SE. *, P V 0.05, Student’s t test; **, P V 0.01, Student’s t test, compared with cells transfected with Fzd5 alone.

expression construct. Western blot analysis verified that the Cotransfection of SFRP4 with Wnt7a and Fzd5 significantly level of Wnt7a protein expression increased with greater diminished TopFlash activity to near baseline levels (Fig. 3A). amounts of transfected Wnt7a construct, whereas Fzd5 receptor To confirm an interaction between Wnt7a and SFRP4, levels remained constant (Fig. 2C). The appearance of a slight coimmunoprecipitation experiments were done (Fig. 3B). Cells increase in Fzd5 expression with cotransfection of 2 and 4 Ag were transfected with Wnt7a or SFRP4 plus or minus Wnt7a. of Wnt7a can be attributed to overexposure. A longer exposure Because SFRPs are secreted proteins, transfected cells were time was necessary to visualize the smaller amounts of Wnt7a. treated with the hydrophilic cross-linking reagent 3,3¶-dithio- Luciferase measurements confirmed that 2 Ag Wnt7a cotrans- bis(sulfosuccinimidylpropionate) (DTSSP) to capture protein- fected with 1 Ag Fzd5 was sufficient to produce significant protein interactions occurring at the cell surface. Lysates were activation of h-catenin when compared with 1 Ag Fzd5 baseline immunoprecipitated with either anti-Wnt7a or anti-SFRP4 and (P V 0.05; Fig. 2D). Transfection of 1, 2, and 4 Ag increased Western blotted with anti-V5. Our results suggest that inhibition h-catenin signaling 1.7-, 2.8-, and 4.3-fold, respectively. This of Wnt7a-mediated h-catenin/canonical signaling in transfected illustrates that, in Ishikawa cells, h-catenin/canonical signaling cells occurs through a direct interaction of Wnt7a with SFRP4. can be induced by Wnt7a in a dose-responsive manner in the To assess the levels of accumulated h-catenin protein, presence of Fzd5. cytosolic extracts were prepared from Ishikawa cells transfected with EV, Wnt7a and Fzd5, or Wnt7a and Fzd5 plus SFRP4. SFRP4 Inhibition of b-Catenin/Canonical Signaling A cellular fractionation method was necessary to ensure that To detect potential inhibition of Wnt7a signaling, SFRP4 only cytosolic h-catenin protein stabilized due to Wnt signaling was transfected alone or cotransfected with Fzd5 and Wnt7a. would be detected and not the relatively stable pool of

h-catenin that is membrane associated with cadherins. A 9.8- coculture experiments were done. One cell population, fold increase in cytosolic h-catenin accumulation was observed referred to as effector cells, was either transfected with EV in the presence of Wnt7a and Fzd5 compared with control or manipulated to overexpress SFRP4. The target cell transfectants (Fig. 3C). A 2-fold reduction in the h-catenin population was transfected with EV or Fzd5 receptor plus levels was observed as a consequence of SFRP4. These data Wnt7a, Renilla-TK, and the TopFlash reporter gene. When indicate that the Wnt7a signal is initiated in the presence of these two cell populations were mixed, SFRP4 effector cells Fzd5 and can be at least partially ablated by SFRP4. significantly inhibited h-catenin/canonical signaling in both EV and Wnt7a/Fzd5 target cells by 2- and 2.8-fold, SFRP4 Suppresses b-Catenin/Canonical Signaling respectively (Fig. 3D). The coculture experiment was also through a Paracrine Mechanism done using a constitutively active form of h-catenin as a To determine whether SFRP4 could inhibit Wnt7a-induced negative control for SFRP4 inhibition and to show that SFRP4 canonical signaling in a paracrine fashion, a series of was not toxic to the target cell population. As shown in

FIGURE 3. SFRP4 binds to Wnt7a and promotes inhibition of h-catenin/canonical signaling in an autocrine and paracrine manner. A. Wnt7a canonical signaling via Fzd5 is inhibited bythe extracellular antagonist SFRP4. TopFlash (FopFlash) experiments were done as described previously.Columns , mean; bars, SE. *, P V 0.05, Student’s t test, compared with control cells; **, P V 0.01, Student’s t test, compared with Wnt7a-Fzd5 – transfected cells. B. SFRP4 directlybinds Wnt7a in endometrial cells. Ishikawa cells were transfected to overexpress EV, Wnt7a, SFRP4, or SFRP4 plus Wnt7a and treated with 2 mmol/L DTSSP hydrophilic cross-linker before lysis. Cell extracts were then immunoprecipitated with antibody to either Wnt7a or SFRP4. Anti-Wnt7a immunoprecipitation of lysate from cells overexpressing SFRP4 served as a control for specificity. Western blot of coimmunoprecipitates with total cell lysate and IgG control was probed with anti-V5. C. The Wnt7a-Fzd5 interaction increases cytosolic h-catenin levels, which can be alternativelyinhibited bySFRP4. Ishikawa cells were transfected to express the proteins as indicated and then hypotonically lysed for Western blot analysis of cytosolic fractions. Blots were probed with anti-V5 to confirm expression of transfected proteins, anti-h-catenin, and anti-h-actin for loading control. Quantification of h-catenin was normalized to levels of h-actin and designated as fold difference relative to control transfection. D. SFRP4 inhibits Wnt7a-Fzd5 – stimulated TopFlash activity in a paracrine assay. Target Ishikawa cells, cotransfected with EV or Fzd5 plus Wnt7a, TopFlash reporter, and Renilla-TK, were cocultured with EV or SFRP4 effector cells 24 h after transfection. Cells were assayed for luciferase activity 24 h after effector cell seeding. Inset, coculture experiment done using a constitutivelyactive h-catenin expression construct shows a negative control for SFRP4 inhibition of canonical signaling. Results are representative of at least three separate experiments (two for h-catenin control) done in duplicate. Columns, mean; bars, SE. **, P V 0.01, Student’s t test; #, P V 0.001, Student’s t test, compared with EV/EV coculture.

Fig. 3D (inset), SFRP4 coculture had no significant effect on increased the level of c-Jun phosphorylation, whereas h-catenin signaling. This is due to the fact that the constitutive coexpression of Wnt7a and Fzd10 with SFRP4 repressed h-catenin activity occurs downstream of SFRP4 inhibition. c-Jun phosphorylation (Fig. 4C). An equivalent amount of These data indicate that SFRP4 associates with Wnt7a in a total c-Jun was observed in each sample, indicating no change paracrine mechanism to suppress h-catenin/canonical signal- in c-Jun protein expression. These data indicate that Wnt7a is ing mediated by Fzd5 in endometrial cells. capable of activating planar cell polarity/JNK-dependent noncanonical signaling via Fzd10 and that the signal is Wnt7a Induction of JNK/Noncanonical Signaling via inhibited by SFRP4. Fzd10 and Repression by SFRP4 Wnt ligands can activate either canonical or noncanonical SFRP4 Protein Directly Binds Purified Wnt7a to Inhibit pathways, often in a cell- or tissue-dependent manner. Activation of b-Catenin/Canonical Signaling Because Wnt7a did not activate h-catenin/canonical signaling Previous reports have shown that SFRPs not only bind to via Fzd10 in Ishikawa cells, we wanted to test whether Wnt7a Wnts as a means to modulate Wnt signaling but can also bind could activate the noncanonical JNK pathway. To determine Fzd. Given this knowledge, we tested whether SFRP4 was whether Wnt7a activated the JNK pathway through Fzd10, directly binding to Wnt7a or Fzd5 to inhibit downstream we used a transcriptional reporter system to measure signaling. Purified Wnt7a and/or SFRP4 were incubated with activation of JNK. This assay measures the change in gene protein A/G agarose and a purified fusion protein consisting of transcription through phosphorylation of c-Jun when fused to the CRD of Fzd5 and the Fc portion of human IgG or Fc IgG as the GAL4 DNA-binding domain, resulting in luciferase a negative control. Western blot analysis confirmed that Wnt7a activity. We observed that Wnt7a significantly activated bound to the CRD of Fzd5 as expected, whereas SFRP4 did not JNK when cotransfected with Fzd10 as indicated by a 2.6- (Fig. 5A). Direct association between purified SFRP4 and fold induction of luciferase activity compared with control Wnt7a was shown by immunoprecipitation with anti-Wnt7a. cells (P V 0.05; Fig. 4A). Coexpression of SFRP4 diminished These findings verify that SFRP4 directly binds Wnt7a and not JNK activation mediated by Wnt7a-Fzd10. Cotransfection of the CRD of Fzd5 to inhibit canonical signaling. Wnt7a with Fzd5 and Fzd5 alone was unable to activate JNK We next tested whether our purified Wnt7a and SFRP4 signaling. Western blots of total cell lysates are shown in proteins were biologically active and capable of transducing Fig. 4B. Coexpression of Wnt7a and Fzd10 in Ishikawa cells and inhibiting the canonical signal, respectively. We found that,

FIGURE 4. Wnt7a activation of the noncanonical planar cell polarity/JNK pathwaymediated byFzd10 and inhibition bySFRP4. A. Ishikawa cells were transientlytransfected with Wnt7a, Fzd5, and Fzd10 as specified along with constructs from a Path- Detect c-Jun Trans-Reporting System and Renilla-TK for normalization. Cells were incubated for 24 h, and luciferase activities were then measured. Columns, relative light units (firefly/Renilla) of three independent experiments; bars, SE. Statistical significance is based on ANOVA and Bonferroni post hoc for multiple comparisons. *, P V 0.05, compared with the activityin control EV-transfected cells; **, P V 0.01, compared with Wnt7a-Fzd10 cells. B. Western blot indicates expression of transfected proteins using anti-V5 and anti-h-actin for loading control. It should be noted that Fzd10 tends to form hetero- dimers indicated bysmeared band. C. Western blots showing c-Jun phosphorylation with baseline response to EV expression, induction with Wnt7a plus Fzd10, and reduction with Wnt7a plus Fzd10 and SFRP4. The total amount of c-Jun protein was used to control for loading and anti-V5 was used to confirm expression of transfected proteins.

FIGURE 5. Purified SFRP4 directlybinds Wnt7a protein and not the Fzd5 CRD to inhibit activation of h-catenin/canonical signaling. A. Wnt7a and SFRP4 proteins were incubated with protein A/G agarose and a purified fusion protein consisting of the CRD of Fzd5 and the Fc portion of human IgG or Fc alone. Direct SFRP4/Wnt7a binding was shown byimmunoprecipitation with anti-Wnt7a. Western blot analysiswas done with anti-V5 and rabbit anti-human IgG . B. TopFlash dose-response experiment indicates activation of canonical signaling bypurified Wnt7a protein in Ishikawa cells ( gray columns) and to a greater extent in stable Fzd5-overexpressing Ishikawa cells (black columns) C. Concurrent addition of SFRP4 protein inhibits purified Wnt7a activation of the TopFlash reporter in Fzd5-overexpressing cells. when transiently transfected into Ishikawa cells, the TopFlash carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium salt reporter activated in response to treatment with purified Wnt7a (MTS) assay to assess cell proliferation. At 48 h, a statistically and exhibited dose dependency. As shown in Fig. 5B, Wnt7a significant increase in proliferation was observed when Fzd5 protein induces a 2- and 3.5-fold increase in reporter activation was cotransfected with Wnt7a (P V 0.001). Additionally, in control and cells stably expressing Fzd5, respectively. We we found that SFRP4 significantly inhibited cell growth also calculated the EC50 value to be between 60 and 65 ng/mL induced by Wnt7a/Fzd5 (P V 0.001; Fig. 6A). No relevant for both cell types. We then tested whether the addition of effects on Ishikawa cell proliferation were observed in the SFRP4 protein could inhibit Wnt7a-induced reporter activity. presence of Fzd10. This result indicates that proliferation Using a concentration of Wnt7a protein just above the EC50 induced by Wnt7a is thereby an effect of canonical signaling (100 ng/mL), SFRP4 protein was concurrently added to Fzd5 mediated by Fzd5 and not the Fzd10-mediated JNK pathway. cells with Wnt7a at increasing doses. As shown in Fig. 5C, We then aimed to parallel this finding using purified Wnt7a SFRP4 elicits a dose-responsive reduction in Wnt7a-mediated and SFRP4 proteins in Ishikawa cells stably overexpressing reporter activation. No effect was observed when SFRP4 was Fzd5. A statistically significant increase in proliferation was added alone, indicating that SFRP4 is suppressing h-catenin/ observed when Fzd5-overexpressing cells were treated with canonical signaling induced by the purified Wnt7a protein. 100 ng/mL Wnt7a (P V 0.05). More intriguingly, we found that SFRP4 significantly inhibited cell growth induced by Wnt7a in SFRP4 Functionally Inhibits Wnt7a-Induced Proliferation a dose-dependent manner and also inhibited proliferation in the in Endometrial Cancer Cells absence of exogenous Wnt7a (Fig. 6B). To determine the functional effect of Wnt7a and SFRP4 in We next assayed the ability of SFRP4 to suppress Ishikawa cells, we used a 3-(4,5-dimethylthiazol-2-yl)-5-(3- endometrial cancer cell growth. Ishikawa cells were selected

FIGURE 6. SFRP4 inhibits endometrial cancer cell proliferation. MTS assayof Ishikawa cells ( A) transfected with Wnt component constructs and (B) Fzd5-overexpressing Ishikawa cells treated with purified Wnt7a and/or SFRP4 proteins. Proliferation was measured 48 h after seeding for transfected cells and 3 d after treatment with purified proteins. Absorbance was measured at 490 nm 2 h after addition of MTS reagent. Columns, mean of three separate experiments in triplicate; bars, SE. Statistical significance is based on ANOVA and Bonferroni post hoc for multiple comparisons. *, P V 0.05, compared with EV or untreated controls; ***, P V 0.001, compared with EV or untreated controls; #, P V 0.01, compared with Wnt7a-treated cells; ##, P V 0.001, compared with Wnt7a-treated cells. C. Colonyformation assayof SFRP4-Null and SFRP4 Ishikawa stable cell lines after selection with G418 for 2 wk. Colonies were quantified bycrystalviolet extraction. Columns, mean of three individual experiments; bars, SD. * P V 0.001, for SFRP4 stable clones compared with each SFRP4-Null clone based on Student’s t test analysis.

as a model because they do not natively express SFRP4. Cells signal via Fzd5/Lrp6 in neuronal PC12 cells as measured by were transfected with a SFRP4 construct and individual increased h-catenin stability and activation of a T-cell factor SFRP4-expressing and SFRP4-Null (cells transfected with reporter construct (37). In situ hybridization of mouse embryos SFRP4 construct, but not SFRP4 expressing based on Western showed that Fzd10 is expressed in the same regions as Wnt7a in blot) stable clones were selected for colony growth assays. the neural tube, limb buds, and Mullerian duct (38). Coinjection SFRP4 expression markedly decreased colony formation by an of Wnt7a and Fzd10 synergistically induces Wnt target genes overall average of f65% (Fig. 6C). in Xenopus embryos (39). We conducted a microarray analysis to compile a panel of all Wnt components in the Ishikawa cell line and found significant expression of seven different Fzds, Discussion including Fzd5. Quantitative RT-PCR studies revealed positive Previous studies have established that Wnt7a is important in expression of both Fzd5 and Fzd10 in human endometrial the development and differentiation of the female reproductive tissue.1 Interestingly, little has been published on the localiza- tract. Although Wnt7a has been recognized as a major ligand in tion of Fzds in the endometrium. Our in situ hybridization the endometrium, no information is available about its cognate experiments confirmed that Fzd5 expression was restricted to Fzd partners or underlying signaling mechanism in the adult. both stroma and epithelium, whereas Fzd10 was localized Wnt7a has been shown to activate both canonical and solely to the stroma. Based on this information, Fzd5 and Fzd10 noncanonical pathways in other cellular environments were predicted to be likely candidates for binding Wnt7a. (33-36). Therefore, it could be expected that Wnt7a has Our work confirms that Fzd5 and Fzd10 function as Wnt7a multiple functions in the endometrium. Wnt7a was shown to receptors in the endometrium and shows that Wnt7a can signal

via the h-catenin/canonical or noncanonical JNK-mediated found that purified SFRP4 inhibits Wnt7a-mediated prolifera- pathway depending on which Fzd receptor it binds to on the tion and stable overexpression of SFRP4 functions to suppress cell. Wnt7a interaction with Fzd5 resulted in stabilization of endometrial cancer cell growth in vitro. cytoplasmic h-catenin and subsequent activation of h-catenin/ The intimate collaboration between endometrial stromal canonical signaling in a concentration-dependent manner. The and epithelial cells is essential for normal adult function as binding of Wnt7a to Fzd10 induced the noncanonical JNK- well as for mediating the hormone-driven remodeling that mediated pathway and increased phosphorylation of c-Jun. accompanies the menstrual cycle (3, 4). Given that Wnts are Because Fzd10 is localized to the endometrial stroma and our generally localized to different cell types and regulated by studies were done in epithelial cells, it may be of interest to both estrogen and progesterone, it is presumed that they are further test whether the Wnt7a-Fzd10 interaction induces JNK key participants in this stromal-epithelial molecular dialogue. signaling in endometrial stromal cells. The potential for one Wnt7a signaling mediated through its cognate Fzd5 and Fzd10 Wnt ligand to activate two separate signaling pathways in a cell receptors may provide a molecular dialogue between epithelial through different receptors has also been recently shown by and stromal cells, which can be modulated by SFRP4. We Mikels and Nusse (40). However, this has never been reported believe that disruption of this intercellular communication, for two individual Fzd receptors. The purification of large through loss of SFRP4 expression, leads to excessive Wnt quantities of soluble active Wnt protein is extremely difficult signaling, overactive h-catenin, and increased proliferation and the biochemical properties of all 19 individual Wnts are not that may ultimately contribute to the onset of endometrial completely understood. We provide evidence that treatment of cancer. Based on our current findings, we believe that under cells with purified recombinant Wnt7a protein is able to directly normal circumstances SFRP4 can function as a suppressor of stimulate canonical Wnt signaling in a dose-dependent manner cell growth and may prove to hold therapeutic potential for and induce proliferation. endometrial carcinoma. There are several secreted proteins that can also interact with Wnt, including members of the SFRP family. SFRPs compete with Fzds to bind Wnt ligands and, in turn, modulate Wnt Materials and Methods action. However, we should also note that previous publications Cloning of Rat Wnt7a, Fzd5, Fzd10, and SFRP4 have provided evidence for the formation of nonfunctional Rat uterine total RNA was isolated with Tri-Reagent complexes between SFRP and Fzds (41). The specific SFRPs (Molecular Research Center) and further purified using phenol that inhibit Wnt7a-Fzd interactions to provide feedback chloroform extraction and isopropanol precipitation. RNA was regulation in the endometrium have not been identified. In washed with ethanol and resuspended in RNase-free diethyl the present study, we have found that SFRP4 is localized to the pyrocarbonate water. Oligonucleotide primers were designed to endometrial stroma and its expression is significantly down- flank the entire coding regions of rat Wnt7a, Fzd5, Fzd10, and regulated in certain types of endometrial carcinoma (26, 28, SFRP4 for RT-PCR and subsequent ligation of recombinant 29).1 SFRP4 expression peaks with high levels of estrogen cDNA into a pcDNA3.1 directional V5/6xHis TOPO vector during the proliferative phase of the menstrual cycle and (Invitrogen). Primer sequences for Wnt7a, Fzd5, and SFRP4 significantly declines during the progesterone-dominated se- (Table 2) were designed against rat sequences in Genbank and cretory phase (26, 42). In contrast, epithelial expression of modified to incorporate a COOH-terminal V5/6xHis-tag. Fzd10 Wnt7a is much lower throughout the estrogen-driven prolifer- was cloned by alignment of human Fzd10 Genbank sequence ative phase (6). Thus, increased expression of SFRP4 in with the homologous region of the rat genome. Fzd10 cloning response to estrogenic stimuli may assist in the estrogen- primers were designed to include the stop codon (Table 2). Site- dependent down-regulation of Wnt7a through the binding of directed mutagenesis was carried out using the cloned Fzd10 excess Wnt7a. We have discovered that SFRP4 can directly pcDNA3.1 construct and primers shown in Table 2 to convert bind Wnt7a, and not the CRD of Fzd5, to prevent Wnt7a the stop codon to a serine and reincorporate the COOH-terminal signaling in both an autocrine and paracrine fashion. We also V5/6xHis-tag of pcDNA3.1.

TABLE 2. List of Primers

Forward (5¶-3¶) Reverse (5¶-3¶)

Cloning Wnt7a [XM_342723] CACCGGACTATGACCCGGAA CTTGCACGTATACATCTCTGTGCGCT Fzd5 [NM_173838] CACCGGCGATGGCTAGAC TACGTGCGACAGGGACACTTG Fzd10 [NM_007197] CACCGGCGCCACCATGCAACA TCACACGCAGGCTGAAGGCT SFRP4 [NM_053544] CACCACAGTGCGATGCTCCT GCTTGTACTTTTCTTTGGGTTTGA Site-directed mutagenesis Fzd10 pcDNA3.1 TCAGCCTGCGTGTCAAAGGGTCAAGAC GTCTTGACCCTTTGACACGCAGGCTGA In situ hybridization Wnt7a [XM_342723] CTGGGCCACCTCTTTCTCA AGGCCTGGGATCTTGTTACAG Fzd5 [NM_173838] CAAAACCTGAACTCGCTACGAG GACACGAAGCCTGCCAGA Fzd10 [NM_007197] CCCCGATTATGGAGCAGTT TCCATGCACAGGTAGTTGG SFRP4 [NM_053544] GCTTCAGGAACAGCAGAGAA AGGTTTGGAAGCAGGTGAC

NOTE: Wnt components listed with corresponding Genbank accession numbers.

Microarray Analysis Using BeadChip Arrays Western blot analysis and Coomassie staining, respectively. Microarray experiments were done on four Ishikawa Quantification was done using a 6xHis-tag–based ELISA biological replicates. RNA was isolated as described above protocol. and microarray analysis was done using HumanRef-8 BeadChip arrays (24,000 gene sequences/array) from Illumina. Briefly, TopFlash and Phospho-c-Jun Luciferase Assays RNA was amplified and cDNA was synthesized via reverse The TopFlash and FopFlash firefly luciferase reporter transcription. The cDNA was converted to cRNA containing constructs were transfected at 0.5 Ag/plate along with 0.1 Ag/ biotinylated UTPs and incubated with avidin-labeled Cy3dye. plate of the Renilla (pRL-TK) reporter to permit normalization Each sample was then hybridized to a separate array on the of transfection efficiencies. Cells were harvested 24 h follow- BeadChip and incubated at 58jC overnight. Next day, bead ing transfection in 1Â passive lysis buffer (Promega Corp.). chips were washed and scanned. BeadStudio 3software For coculture experiments, target cells were transfected in 60- (Illumina) was used for microarray data analysis. mm plates and then reseeded into 12-well plates 24 h after transfection. Cells were allowed to recover for 6 h and then In situ Hybridization of Endometrial Tissue EV pcDNA3.0 or SFRP4-transfected effector cells from near- Twenty-one-day-old female Sprague-Dawley rats were confluent 60-mm plates were seeded onto the target cells to a purchased from Harlan and housed in accredited facilities final ratio of 2:1. Cells were harvested 24 h after seeding of at the University of Texas Houston Medical School. All effector cells. Wnt7a protein dose-response experiments were federal guidelines and institutional policies about their done in a similar manner. In place of effector cell coculture, treatment were followed. Animals were ovariectomized via cells were treated with purified recombinant Wnt7a and a dorsal approach and uterine tissue was removed 5 d later. SFRP4 proteins as described in the text. For analysis of c-Jun A Tissue was formalin fixed and then cut into 10-Am paraffin- activity, cells were transfected in six-well plates with 1 g/ A embedded sections by the University of Texas Houston well of pFR-Luc reporter construct and 0.05 g/well pFA2-c- Histology Department. Digoxigenin-labeled mRNA sense and Jun from the PathDetect c-Jun Trans-Reporting System A antisense probes targeting the 5¶ end of rat Wnt7a, Fzd5, (Stratagene), and 0.05 g/well of the Renilla (pRL-TK) Fzd10, and SFRP4 transcripts were generated using primers reporter to permit normalization. All other expression A listed in Table 2 and in vitro transcription using the constructs were transfected at 0.5 g/well. Luciferase activity digoxigenin RNA labeling kit (Roche). In situ hybridization was measured by a Monolight 2010 luminometer (Analytical analysis was done using the methodology previously Luminescence Laboratory) using the Dual-Luciferase Reporter described (43). Images were gathered by means of bright- Assay (Promega) as indicated by the manufacturer. Each field microscopy at Â100 magnification. experiment was repeated at least thrice in duplicate to ensure reproducibility.

Cell Culture and Transfection Coimmunoprecipitation Ishikawa cells were grown in RPMI 1640 with glutamine, For the Wnt7a-Fzd coimmunoprecipitation experiment, without phenol red (Invitrogen-Life Technologies), and supple- Ishikawa cells were harvested with radioimmunoprecipitation mented with 10% fetal bovine serum, 0.01 mol/L HEPES, and assay buffer [150 mmol/L NaCl, 1% Triton X-100, 1.0% penicillin/streptomycin. Cells were split 2 d before transient deoxycholate, 0.1% SDS, 50 mmol/L Tris (pH 7.5), 5 mmol/L transfection and grown to 60% to 70% confluency. Transient EDTA, 0.5 mmol/L phenylmethylsulfonyl fluoride] 24 h fol- transfections of Ishikawa cells (unless otherwise indicated) lowing transfection. Cell extracts were immunoprecipitated were conducted in 60-mm culture plates using FuGene 6 with goat anti-Wnt7a or goat anti-Fzd5 (Santa Cruz (Roche) according to the manufacturer’s guidelines. All Biotechnology) using Protein A/G Plus-Agarose (Santa Cruz A constructs were transfected at 1 g/dish unless otherwise Biotechnology). specified. To detect interactions between Wnt7a and SFRP4, To generate cell lines stably expressing Wnt7a-V5/His, Ishikawa cells were treated with 2 mmol/L DTSSP cross- SFRP4-V5/His, or the Fzd5-V5/His receptor, cells were plated linker (Pierce) 24 h after transfection for 30 min to capture out in 1:40 dilution 24 h after transfection and cultured under protein-protein interactions at the extracellular membrane. A neomycin selection (600 g/mL G418) for 10 to 14 d. At least The cross-linking reaction was terminated by addition of 12 clones expanded from single cells were screened for 10 mmol/L Tris (pH 7.5) and cell lysis was done in Triton expression via RT-PCR and Western blot analysis. X-100 buffer (20 mmol/L HEPES, 1.0% Triton X-100, 250 mmol/L NaCl, 10% glycerol, 2 mmol/L EDTA, 1 mmol/L Purification of Recombinant Wnt7a and SFRP4 6xHis- phenylmethylsulfonyl fluoride, 2.5 Ag/mL aprotinin, 2.5 Ag/mL Tagged Proteins leupeptin, 100 Amol/L sodium orthovanadate, 50 mmol/L 6xHis-tagged proteins were purified from stable cells sodium fluoride). Coimmunoprecipitation was conducted expressing Wnt7a-V5/His or SFRP4-V5/His by loading using either goat anti-Wnt7a or goat anti-SFRP4 (Santa Cruz cleared lysate into Ni-NTA spin columns (Qiagen) and Biotechnology). For experiments using purified proteins, following protocol for mammalian cells under native con- 500 ng/mL Wnt7a and/or SFRP4 were immunoprecipitated ditions. Eluted samples were run through a second column in Triton X-100 buffer with either 500 ng/mL Fzd5CRD-Fc and then dialyzed against PBS (pH 7.4) with addition of 1.0% IgG (R&D Systems), Fc IgG (Bethyl Laboratories), or 1 Ag CHAPS. Protein expression and purity were verified via goat anti-Wnt7a.

Western Blot Analysis 5. Cooke PS, Buchanan DL, Young P, et al. Stromal estrogen receptors mediate mitogenic effects of estradiol on uterine epithelium. Proc Natl Acad Sci U S A Proteins isolated from coimmunoprecipitation experiments 1997;94:6535 – 40. were electrophoresed on SDS-polyacrylamide gels and Western 6. Miller C, Pavlova A, Sassoon DA. Differential expression patterns of Wnt blot analysis was done using mouse anti-V5 (Invitrogen) or genes in the murine female reproductive tract during development and the estrous rabbit anti-human IgG (Santa Cruz Biotechnology). Immuno- cycle. Mech Dev 1998;76:91 – 9. reactive proteins were visualized using secondary antibodies 7. Mericskay M, Kitajewski J, Sassoon D. Wnt5a is required for proper epithelial-mesenchymal interactions in the uterus. Development 2004;131: and the enhanced chemiluminescence Western blotting system 2061 – 72. (Amersham Biosciences). 8. Miller C, Sassoon DA. Wnt-7a maintains appropriate uterine patterning during To assess cytoplasmic h-catenin accumulation, transfected the development of the mouse female reproductive tract. Development 1998;125: Ishikawa cells were lysed by hypotonic lysis as previously 3201 – 11. h 9. Parr BA, McMahon AP. Sexually dimorphic development of the mammalian described (40). Western blot was done with mouse anti- - reproductive tract requires Wnt-7a. Nature 1998;395:707 – 10. h catenin (Santa Cruz Biotechnology) and mouse anti- -actin 10. Logan CY, Nusse R. The Wnt signaling pathway in development and disease. (Chemicon International) for loading control normalization. Annu Rev Cell Dev Biol 2004;20:781 – 810. Quantitation was done using GeneTools software (Syngene). 11. Mlodzik M. Planar cell polarization: do the same mechanisms regulate To determine c-Jun phosphorylation, Western blot was done on Drosophila tissue polarity and vertebrate gastrulation? Trends Genet 2002;18: 564 – 71. total cell lysate using mouse anti-phospho-c-Jun and rabbit anti- 12. Miller JR. The Wnts. Genome Biol 2002;3:REVIEWS3001. c-Jun antibody (Santa Cruz Biotechnology). Protein expression 13. Tamai K, Zeng X, Liu CM, et al. A mechanism for Wnt coreceptor of all transfected constructs was confirmed by Western blot activation. Mol Cell 2004;13:149 – 56. analysis using mouse anti-V5. 14. He TC, Sparks AB, Rago C, et al. Identification of c-MYC as a target of the APC pathway. Science 1998;281:1509 – 12. Cell Proliferation Assay 15. Itoh K, Krupnik VE, Sokol SY. Axis determination in Xenopus involves biochemical interactions of axin, glycogen synthase kinase 3and h-catenin. Curr Ishikawa cells were transfected with the Wnt component Biol 1998;8:591 – 4. constructs as indicated in 60-mm plates. Transfectants were 16. Kishida S, Yamamoto H, Hino S, Ikeda S, Kishida M, Kikuchi A. DIX reseeded at 5 Â 103 cells per well in a 96-well culture plate domains of Dvl and axin are necessary for protein interactions and their ability to h 24 h after transfection. Cells stably overexpressing Fzd5 were regulate -catenin stability. Mol Cell Biol 1999;19:4414 – 22. seeded at 2.5 Â 103 per well in a 96-well culture plate. Cells 17. Shtutman M, Zhurinsky J, Simcha I, et al. The cyclin D1 gene is a target of the h-catenin/LEF-1 pathway. Proc Natl Acad Sci U S A 1999;96:5522 – 7. were allowed to recover for 6 to 8 h and treated with 100 ng/mL 18. Ikeda S, Kishida S, Yamamoto H, Murai H, Koyama S, Kikuchi A. Axin, a Wnt7a and purified SFRP4 protein as indicated. The number negative regulator of the Wnt signaling pathway, forms a complex with GSK-3h of viable cells was evaluated 48 h after reseeding of transfected and h-catenin and promotes GSK-3h-dependent phosphorylation of h-catenin. cells and 3d after treatment with purified proteins by EMBO J 1998;17:1371 – 84. 19. Machin P, Catasus L, Pons C, Munoz J, Matias-Guiu X, Prat J. CTNNB1 incubating for 2 h with the CellTiter 96 Aqueous Non- mutations and h-catenin expression in endometrial carcinomas. Hum Pathol 2002; Radioactive Cell Proliferation MTS Assay Reagent (Promega). 33:206 – 12. Plates were read with a plate reader at a wavelength of 490 nm. 20. Prat J, Gallardo A, Cuatrecasas M, Catasus L. Endometrial carcinoma: pathology and genetics. Pathology 2007;39:72 – 87. 21. Li L, Yuan HD, Xie W, et al. Dishevelled proteins lead to two signaling Colony Formation Assay pathways. Regulation of LEF-1 and c-Jun N-terminal kinase in mammalian cells. Ishikawa stable SFRP4 and SFRP4-Null cell lines were J Biol Chem 1999;274:129 – 34. plated at 1,000 per well in six-well plate and grown with 22. Ahumada A, Slusarski DC, Liu XX, Moon RT, Malbon CC, Wang HY. selection in G418 (600 Ag/mL) for 2 wk. Cells were stained Signaling of rat Frizzled-2 through phosphodiesterase and cyclic GMP. Science 2002;298:2006 – 10. with crystal violet, which was extracted with 1% SDS and 23. Kuhl M. The WNT/calcium pathway: biochemical mediators, tools and quantificated at an absorbance of 590 nm. future requirements. Front Biosci 2004;9:967 – 74. 24. Sheldahl LC, Park M, Malbon CC, Moon RT. Protein kinase C is differentially stimulated by Wnt and Frizzled homologs in a G-protein-dependent Disclosure of Potential Conflicts of Interest manner. Curr Biol 1999;9:695 – 8. No potential conflicts of interest were disclosed. 25. Kawano Y, Kypta R. Secreted antagonists of the Wnt signalling pathway. J Cell Sci 2003;116:2627 – 34. Acknowledgments 26. Abu-Jawdeh G, Comella N, Tomita Y, et al. Differential expression of frpHE: We thank the following individuals for providing reagents: O. Destree and a novel human stromal protein of the secreted frizzled gene family, during the H. Clevers (University of Utrecht) for TopFlash and FopFlash and P.D. McCrea endometrial cycle and malignancy. Lab Invest 1999;79:439 – 47. (University of Texas M. D. Anderson Cancer Center) for constitutively active 27. Fujita M, Ogawa S, Fukuoka H, et al. Differential expression of secreted B-catenin pcDNA3.1. frizzled-related protein 4 in decidual cells during pregnancy. J Mol Endocrinol 2002;28:213– 23. References 28. Hrzenjak A, Tippl M, Kremser ML, et al. Inverse correlation of secreted 1. Cunha GR. Epithelial-stromal interactions in development of urogenital tract. frizzled-related protein 4 and h-catenin expression in endometrial stromal Int Rev Cytol 1976;47:137 – 94. sarcomas. J Pathol 2004;204:19 – 27. 2. Arnold JT, Kaufman DG, Seppala M, Lessey BA. Endometrial stromal cells 29. Risinger JI, Maxwell GL, Chandramouli GVR, et al. Gene expression regulate epithelial cell growth in vitro: a new co-culture model. Hum Reprod profiling of microsatellite unstable and microsatellite stable endometrial 2001;16:836 – 45. cancers indicates distinct pathways of aberrant signaling. Cancer Res 2005;65: 5031 – 7. 3. Cooke PS, Uchima FDA, Fujii DK, Bern HA, Cunha GR. Restoration of normal morphology and estrogen responsiveness in cultured vaginal and uterine 30. Lee AY, He B, You L, et al. Expression of the secreted frizzled-related epithelia transplanted with stroma. Proc Natl Acad Sci U S A 1986;83:2109 – 13. protein gene family is downregulated in human mesothelioma. Oncogene 2004; 4. Cunha GR, Cooke PS, Kurita T. Role of stromal-epithelial interactions in 23:6672 – 6. hormonal responses. Arch Histol Cytol 2004;67:417 – 34. 31. Liu TH, Raval A, Chen SS, Matkovic JJ, Byrd JC, Plass C. CpG island

methylation and expression of the secreted frizzled-related protein gene family in receptor complex and modulation by Dickkopf proteins. J Biol Chem 2003;278: chronic lymphocytic leukemia. Cancer Res 2006;66:653– 8. 37024 – 31. 32. Zou HZ, Molina JR, Harrington JJ, et al. Aberrant methylation of secreted 38. Nunnally AP, Parr BA. Analysis of Fz10 expression in mouse embryos. Dev frizzled-related protein genes in esophageal adenocarcinoma and Barrett’s Genes Evol 2004;214:144 – 8. esophagus. Int J Cancer 2005;116:584 – 91. 39. Kawakami Y, Wada N, Nishimatsu S, Nohno T. Involvement of Frizzled-10 33. Hirabayashi Y, Itoh Y, Tabata H, et al. The Wnt/h-catenin pathway directs in Wnt-7a signaling during chick limb development. Dev Growth Differ 2000;42: neuronal differentiation of cortical neural precursor cells. Development 2004;131: 561 – 9. 2791 – 801. 40. Mikels AJ, Nusse R. Purified Wnt5a protein activates or inhibits h- 34. Lucas FR, Salinas PC. WNT-7a induces axonal remodeling and increases catenin-TCF signaling depending on receptor context. PLoS Biol 2006;4: synapsin I levels in cerebellar neurons. Dev Biol 1997;192:31 – 44. 570 – 82. 35. Lyu J, Joo CK. Wnt-7a up-regulates matrix metalloproteinase-12 expression 41. Bafico A, Gazit A, Pramila T, Finch PW, Yaniv A, Aaronson SA. Interaction and promotes cell proliferation in corneal epithelial cells during wound healing. of frizzled related protein (FRP) with Wnt ligands and the frizzled receptor J Biol Chem 2005;280:21653– 60. suggests alternative mechanisms for FRP inhibition of Wnt signaling. J Biol 36. Winn RA, Marek L, Han SY, et al. Restoration of Wnt-7a expression reverses Chem 1999;274:16180 – 7. non-small cell lung cancer cellular transformation through frizzled-9-mediated 42. Ace CI, Okulicz WC. Microarray profiling of progesterone-regulated growth inhibition and promotion of cell differentiation. J Biol Chem 2005;280: endometrial genes during the rhesus monkey secretory phase. Reprod Biol 19625 – 34. Endocrinol 2004;2:54. 37. Caricasole A, Ferraro T, Iacovelli L, et al. Functional characterization of 43. Komminoth P. Digoxigenin as an alternative probe labeling for in situ WNT7A signaling in PC12 cells: interaction with a FZD5 center dot LRP6 hybridization. Diagn Mol Pathol 1992;1:142 – 50.

Mol Cancer Res 2008;6(6). June 2008 Downloaded from mcr.aacrjournals.org on September 30, 2021. © 2008 American Association for Cancer Research. Secreted Frizzled-Related Protein 4 Regulates Two Wnt7a Signaling Pathways and Inhibits Proliferation in Endometrial Cancer Cells

Kendra S. Carmon and David S. Loose

Mol Cancer Res 2008;6:1017-1028.

Updated version Access the most recent version of this article at: http://mcr.aacrjournals.org/content/6/6/1017

Cited articles This article cites 43 articles, 19 of which you can access for free at: http://mcr.aacrjournals.org/content/6/6/1017.full#ref-list-1

Citing articles This article has been cited by 8 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/6/6/1017.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/6/6/1017. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.