Mammary Gland Tissue Targeted Overexpression of Human Protease-Activated Receptor 1 Reveals a Novel Linkto B-Catenin Stabilization

Research Article

Yong-Jun Yin,1 Vered Katz,1 Zaidoun Salah,1 Myriam Maoz,1 Irit Cohen,1 Beatrice Uziely,1 Hagit Turm,1 Sorina Grisaru-Granovsky,1 Hiromu Suzuki,2 and Rachel Bar-Shavit1

1Department of Oncology, Hadassah-Hebrew University Hospital, Jerusalem, Israel and 2Department of Public Health, Sapporo Medical University S1, Chuo-ku, Sapporo, Japan

Abstract receptor PAR1 in hemostasis, thrombosis, and vascular biology, it Protease-activated receptor 1 (PAR1) is emerging with distinct emerges with distinct assignments in tumor biology and angio- assignments in tumor biology. We show that tissue targeted genesis (3–7). We have previously shown that PAR1 is involved both overexpression of hPar1 in mice mammary glands results in in malignant breast carcinoma tumor progression (3) and the precocious hyperplasia,characterized by a dense network of physiologic invasion of placenta trophoblasts into the uterine ductal side branching and accelerated proliferation. These decidua (4). A direct correlation exists between levels of hPar1 glands exhibit increased levels of wnt-4 and wnt-7b and expression and tumor advancement in both clinically obtained a striking B-catenin stabilization. Nuclear localization of biopsy specimens and differentially metastatic cell lines (3, 8). In- B-catenin is observed in hPar1 transgenic mouse tissue fact, hPar1 plays an active role in breast carcinoma invasion sections but not in the wild-type,age-matched counterparts. because antisense silencing of the gene abrogates efficiently metastatic breast carcinoma cells from invading Matrigel-coated PAR1 induces B-catenin nuclear localization also in established epithelial tumor cell lines of intact B-catenin system filters in vitro (3). (transformed on the background of mismatch repair system; In parallel, a cDNA expression library screen based on the loss of RKO cells). We propose hereby that PAR1-mediated B-catenin anchorage-dependent growth and focus-forming activity in NIH stabilization is taking place primarily via the increase of Wnt 3T3 cells has identified PAR1 as a novel oncogene (9). With these expression. Enforced expression of a specific Wnt antagonist observations, PAR1 joins other G protein–coupled receptors, family member,secreted frizzled receptor protein 5 (SFRP5 ), including mas and g2a, which behave as oncogenes (10, 11). The efficiently inhibited PAR1-induced B-catenin stabilization. oncogenic properties of PAR1 accompany a collection of data Likewise,application of either SFRP2 or SFRP5 on epithelial showing that hPar1 is overexpressed in a wide range of epithelial tumor cells completely abrogated PAR1-induced B-catenin tumors, pointing altogether to the central role of PAR1 in nuclear accumulation. This takes place most likely via carcinoma invasion. inhibition of Wnt signaling at the level of cell surface (forming The canonical wnt/Wingless signaling pathway directs cell fate in a neutralizing complex of ‘‘Receptors-SFRP-Wnt’’). Further- many cell types and plays a central role in development and in more,depletion of hPar1 by small interfering RNA (siRNA) tumor progression. Wnt proteins are soluble glycoproteins vectors markedly inhibited PAR1-induced Wnt-4. The striking initiating cell signaling through binding to receptor complexes stabilization of B-catenin,inhibited by SFRPs on one hand and composed of Frizzled proteins and LDL receptor–related protein Wnt-4 silencing by hPar1 siRNA on the other hand,points to a (LRP) 5/6 (12, 13). The core of the Wnt pathway is the stability of h h novel role of hPar1 in Wnt-mediated B-catenin stabilization. -catenin. Accumulation of -catenin in the cytoplasm leads This link between PAR1 and B-catenin may bear substantial ultimately to its transport to the nuclei where it forms functional implications both in developmental and tumor progression complexes with lymphoid enhancer factor (LEF)/T-cell factor processes. (Cancer Res 2006; 66(10): 5224-33) (TCF) transcription factors (14, 15). Because the LEF/TCF DNA- binding proteins are incapable of activating gene transcription h Introduction alone, -catenin acts as a bridging cofactor, enabling the performance of LEF/TCF (for review, see ref. 16). A growing list Protease-activated receptors (PAR) form a family of G protein– of genes has presently been identified as downstream targets of coupled receptors encoding their own ligands and uniquely h-catenin nuclear activity. Among these are c-Myc (17) and cyclin activated via proteolytic cleavage (1). Each of the four PAR family D1 (18–20). members is activated via proteolytic cleavage, exposing an internal To gain further insight into the causal relationship between ligand distinct for every PAR protein (2). PARs act as sensitive hPar1 expression, breast tumor formation, and mammary gland sensors to the constantly changing extracellular proteases regard- development, we have established a line of mice carrying MMTV- less of whether they are present in a soluble or microenvironment- long terminal repeat (LTR)-SV40-driven hPar1 designed to over- immobilized forms. In addition to the classic role of thrombin express in the mammary glands. Whereas mammary tissues can be used to study discrete developmental remodeling aspects of the breast, they also provide an opportunity to dissect the contribution Requests for reprints: Rachel Bar-Shavit, Department of Oncology, Hadassah- of individual genes in normal and malignant mammopoiesis. We Hebrew University Hospital, P.O. Box 12000, Jerusalem 91120, Israel. Phone: 972-2-677- examined the phenotype of hPar1 transgenic (tg) mice with respect 7563; Fax: 972-2-642-2794; E-mail: [email protected]. I2006 American Association for Cancer Research. to breast morphogenesis and evaluated levels of distinct Wnt gene doi:10.1158/0008-5472.CAN-05-4234 expression and h-catenin stabilization. We hereby propose a novel

Cancer Res 2006; 66: (10). May 15, 2006 5224 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2006 American Association for Cancer Research. hPar1 Induced b-Catenin Stabilization link between hPar1, h-catenin, and Wnt generation. This is based Materials and Methods on the combined analyses of mammary gland tissue samples and Generation of MMTV-hPar1-tg epithelial cancer cell lines. By the use of either secreted frizzled The coding sequence of the full-length human Par1 gene from pcDNA3- receptor proteins (SFRP), members of the Wnt antagonist family, or hPar1 was subcloned into MMTV-SV40-BSSK (kindly provided by Dr. R.G. hPar1 small interfering RNAs (siRNA), we provide evidence on the Pestell, Department of Developmental and Molecular Biology and Medicine, involvement of Wnts in PAR1-induced h-catenin stabilization. Albert Einstein College of Medicine, New York, NY). Briefly, pcDNA3-hPar1 Enforced expression of SFRP5, as also application of either SFRP2, was digested with HindIII and EcoRI to isolate the 1.4-kb full-length hPar1. SFRP5, or both, effectively abrogated PAR1-induced h-catenin In parallel, MMTV-SV40-BSSK was digested with HindIII and EcoRI, and stabilization. This mode of inhibition points to the presence of an full-length hPar1 was ligated into digested MMTV-SV40-BSSK after the autocrine Wnt signaling loop initiated by PAR1. Ectopic over- MMTV-LTR followed by the SV40 poly(A) site (see Fig. 1). The product was expression of members of SFRP antagonist sequesters the PAR1- named the MMTV-hPar1 construct. The purified MMTV-hPar1 plasmid was prepared for microinjection by digestion with SpeI. It was then injected into induced Wnts and binds Frizzled receptors through the homology the pronucleus of fertilized C57BL/6 mouse oocytes and transferred to a site of cysteine-rich domain. Thus, SFRPs antagonize Wnt signaling pseudopregnant CB6/F1 mice uterus (initial animal manipulations were on the level of cell surface. Our data are in line with the elegant carried out by the Transgenic Facility Unit of Hadassah-Hebrew University studies by Bafico et al. (21) on the Wnt autocrine loop in human Medical School, Jerusalem, Israel). Mice were maintained on a CB6/F1 tumorigenicity. In parallel, hPar1 siRNA constructs depleting hPar1 background. For timed pregnancies, male and female mice were mated levels efficiently inhibited Wnt-4 expression. Altogether, we conc- overnight and female mice were scored for vaginal plugs the next morning, lude that PAR1-induced h-catenin stabilization is mediated representing pregnancy day 1. primarily via the induced Wnt generation. The novel link between Southern blot Analysis of Genomic DNA PAR1, h-catenin stabilization, and Wnt may impinge significantly Genomic DNA was prepared using proteinase K (10 Ag) and cut both on developmental and tumor progression processes. (overnight at 37jC) with BamHI and EcoRI (see Fig. 1B), separated by

Figure 1. Generation of MMTV-hPar1-tg mice (hPar1-tg) targeted to the mammary glands. A, schematic representation of MMTV-SV40-BSSK-hPar1construct for mammary-specific expression (MMTV-LTR-hPar1). The long form of the MMTV-LTR is used to drive selective targeted expression and the SV40 splicing and polyadenylation fragment enhances export and translation. Full-length human Par1DNA was inserted between the sites of HindIII and EcoRI. B, genotyping of the founder mice was carried out by Southern blotting. The tail DNA was digested with BamHI and EcoRI, and a transgene fragment was detected via Southern blot. Estimated copies of hPar1 gene 300 pg (30 copies), 100 pg (10 copies), lines 1-4 (L1-L4), and wt. C, whole-mount hematoxylin staining of wt and hPar1-tg mammary glands at different developmental stages. The epithelial tissue derived from the hPar1-tg mammary glands displays increased lateral branching and pervasive intraductal hyperplasia in virgin (V-13w) and pregnant mammary glands (days 8 and 12 of pregnancy, respectively) compared with age-matched wt mice. D, histologic analyses of H&E staining showing the fine histology of the same stages depicted in the whole-mount staining of wt and hPar1-tg mammary glands. V, virgin; P, pregnancy. www.aacrjournals.org 5225 Cancer Res 2006; 66: (10). May 15, 2006

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2006 American Association for Cancer Research. Cancer Research electrophoresis on a 1% agarose-TAE gel, and transferred to Hybond-N+ were centrifuged at 15,000 Â g (15 minutes at 4jC) and the protein contents membranes (Amersham Pharmacia Biotech UK Limited, London, United of the supernatants were evaluated. Kingdom). The blots were then hybridized (overnight at 65jC) with [a-32P]dCTP probe (Rediprimer II, Amersham Biosciences UK) for human Imunoprecipitation Par1 (1.9 kb BamHI- and EcoRI-digested fragment of MMTV-hPar1). Blots Mammary gland lysates (600 Ag) of total protein were used for were washed and exposed to X-ray films. The approximate transgene copy immunoprecipitation using goat anti-mouse Wnt-4 antibody (10 Ag/mL; number was determined by adding 100 pg (10 copies) or 300 pg (30 copies) R&D Systems, Inc., Minneapolis, MN). After overnight incubation, protein of MMTV-hPar1 to genomic DNA from a wt mouse. G-sepharose beads (50 AL; Amersham Pharmacia Biotech, London, United Kingdom) were added to the suspension, which was subsequently rotated at Plasmid Purification 4jC for 1 hour. Elution of the reactive proteins was done by resuspending The plasmids pCMV-HA, pCMV-HA-SFRP5, and pcDNA3-Myc-wt-h- the beads in 2Â protein sample buffer followed by boiling for 5 minutes. catenin (22) were purified using the Qiagen plasmid purification kit (Qiagen The supernatant was then resolved on a 10% SDS-polyacrylamide gel and GmbH, Hilden, Germany): treated as indicated above for Western blotting.

Mammary Gland Whole-Mount Staining Reverse Transcriptase-PCR Whole-mount staining of mammary glands was done as described (23). Entire right no. 4 mammary tissues were removed and total RNA was Briefly, the entire left no. 3 mammary gland was removed and flattened on a isolated using the TRI Reagent (Molecular Research Center, Inc., tissue capsule, fixed in Telly’s fixative, defatted in three changes of acetone, Cincinnati, OH) according to the instructions of the manufacturer. First- hydrated in 95% ethanol, and stained with hematoxylin (0.65 g FeCl3, 67.5 strand cDNA was synthesized from 1 Ag of total RNA (1 hour, 42jC) using mL H2O, and 8.7 mL 10% hematoxylin in 95% ethanol, pH 1.25). Glands were MMLV reverse transcriptase and oligo-dT (both from Promega, Heidel- rinsed in water, destained in acid ethanol, dehydrated in increasing ethanol berg, Germany). cDNA was subjected to PCR amplification with Taq concentrations, stored indefinitely in methyl salicylate, and photographed polymerase (Bioline, London, United Kingdom) using a specific set of using a Zeiss microscope. primers in a total volume of 20 AL. PCR reaction was done by denaturation at 94jC for 3 minutes and 24 to 30 cycles of amplification Histology and Immunohistochemistry Assay (94jC for 30 seconds, 55-65jC for 1 minute, and 72jC for 1 minute). The entire left no. 4 mammary gland was removed and fixed with 4% All the PCR reactions were followed by a 10-minute extension at 72jCat A formaldehyde. Sections of 5 m were prepared and stained with H&E. For the final step. Glyceraldehyde-phosphate dehydrogenase (GAPDH) and L19 immunoperoxidase staining, paraffin-embedded sections were gradually primers were used as housekeeping genes for of the human and mouse dehydrated. Antigen retrieval was achieved by microwaving the sections control. Primers used were L19: 5¶-CTGAAGGTGAAGGGGAATGTG-3¶ (5 minutes), then incubating overnight (4jC) with the following antibodies: (sense), 5¶-GGATAAAGTCTTGATGATCTC-3¶ (antisense; 24 cycles); human proliferating cell nuclear antigen (PCNA; Santa Cruz Biotechnology, Inc., Santa Par1:5¶-GCCAGAATCAAAAGCAACAA-3¶ (sense), 5¶-GAGATGAATGCAG- h Cruz, CA), -catenin (610153, BD Transduction Laboratories, Franklin Lakes, GAAGTTGTTT-3¶ (antisense; 30 cycles); mouse wnt-4:5¶-AGGAGTGCCAA- NJ), and anti c-Myc (9E10, Santa Cruz Biotechnology) and visualized using TACCAGTTCC-3¶ (sense), 5¶-TGTGAGAAGGCTACTCCATA-3¶ (antisense; Broad Spectrom (AEC) Kit (Zymed Laboratories, Inc., San Francisco, CA). 30 cycles); and mouse wnt-7b:5¶-CAAGGCTACTACAACCAGGC-3¶ (sense), ¶ ¶ In situ Hybridizations 5 -CACCTCCACCTGCACCGCTG-3 (antisense; 30 cycles). The PCR prod- RNA probes were transcribed and labeled by T7 RNA polymerase ucts were run electrophoretically on a 1% agarose-TAE gel and visualized (for antisense orientation) or T3 RNA polymerase (for sense orientation as by ethidium bromide under UV light. control) using DIG-UTPlabeling mix (Boehringer Mannheim, Mannheim, siRNA Constructs Germany) as previously described (3). We used U6 promoter-driven and lentivirus (pLentilox 3.7)-mediated Western blot Analysis delivery cassette of siRNA, specific for hPar1. For this, a sequence of 19 Entire right no. 4 mammary gland tissues were removed, homogenized in nucleotides of the hPar1 coding region was selected for stem-and-loop radioimmunoprecipitation assay buffer, incubated (20 minutes, 4jC) for oligonucleotide siRNA. The selected sequences were submitted to a adequate lysis, and centrifuged at 10,000 Â g (20 minutes, 4jC). Lysates BLAST search against the human genome to ensure that it was not (50 Ag) were resolved by 10% SDS-PAGE and transferred to Immobilon-P targeted. To construct the hairpin, stem-and-loop siRNA expression membrane (Millipore, Bedford, MA). Membranes were then blocked and cassette, appropriate DNA oligonucleotides were synthesized. The oligos probed with polyclonal anti–cyclin D1 (Santa Cruz Biotechnology), were composed of the following: 19 bases of hPar1 coding sequence, the monoclonal anti-h-catenin (Transduction Laboratories, Lexington, KY), loop sequence linker (9 bases), reverse complement of the 19 bases of or monoclonal anti-h-actin (Sigma-Aldrich Israel Ltd., Rehovot, Israel) hPar1 coding region, and a terminator sequence poly(T). The sticky end primary antibodies in 1% bovine serum albumin in TTBS. After extensive of the XhoI site was added to the antisense strand oligos. Both sense ¶ washes, blots were incubated with the appropriate secondary antibodies and antisense sequences were phosphorylated at the 5 ends. The sense conjugated to horseradish peroxidase (Pierce, Rockford, IL). Immunoreac- sequence oligos were annealed to their respective antisense oligos. tive bands were detected by the enhanced chemiluminescence reagent siRNA cassette sequences were then ligated into pLentilox 3.7 vector using Supersignal (Pierce). (Van Parij’s laboratory). We created four such siRNA cassettes from the hPar1 gene. PAR1 Activating Peptide Preparation of Lentivirus SFLLRNP. H-Ser-Phe-Leu-Leu-Arg-Asn-Pro-NH2. The lentivirus particles were generated by a three-plasmid expression Nuclear Fraction system in which 293T cells were cotransfected with the following three Cells were washed with PBS, scraped, and collected. The cell pellet was vectors: packaging (CMVD R8.91), envelope (CMV-VSV-G), and the transfer resuspended in buffer A [10 mmol/L HEPES (pH 7.9), 10 mmol/L KCl, 0.1 vector Plentilox 3.7. One day before transfection, 293T cells were plated mmol/L EDTA, 1 mmol/L DTT, and protease inhibitors 5 mg/mL aprotinin, to f60% confluency. On the next day, the medium was changed to fresh cocktail, and 10 mg/mL leupeptin]. The cells were incubated on ice medium and cells were transfected with the three plasmids using FuGENE (15 minutes); after which, a 10% solution of NP40 was added. The 6 transfection reagent. Medium was changed to fresh medium 24 hours homogenate was centrifuged at 10,000 Â g (30 seconds at 4jC). The pellet posttransfection. On days 2 and 3 posttransfection, medium was collected was rotated (15 minutes at 4jC) with buffer C [20 mmol/L HEPES (pH 7.9), to recover viral particles. The collected medium was centrifuged for 420 mmol/L KCl, 1 mmol/L EDTA, and 1 mmol/L DTT, with protease 1 hour at 40,000 rpm to concentrate the viral particles and concentrated inhibitors 5 mg/mL aprotinin, cocktail, and 10 mg/mL leupeptin]. The tubes to 100Â.

Cancer Res 2006; 66: (10). May 15, 2006 5226 www.aacrjournals.org

Results mammary glands from wild-type (wt) CB6/F1 mice were compared Mammary gland tissue targeted overexpression of hPar1-tg with age-matched transgenic littermates (Fig. 1C and D). The mice show precocious hyperplasia. To assess the effects of mammary glands of hPar1-tg mice showed increased branching elevated hPar1 expression in mammary epithelial cells in vivo, with moderate to marked lobuloalveolar development. During we generated transgenic mice which overexpress hPar1 targeted to normal mammary development, rudiments of the ducts formed at the mammary glands. An expression cassette construct was birth and grow slowly until onset of puberty, when terminal end engineered in which the full-length cDNA for hPar1 was placed buds form and ductal elongation and bifurcation begin. Normally, under transcriptional control of the MMTV-LTR, followed by an in virgin wt animals, the degree of ductal side branching and the SV40 polyadenylation (SV40 poly A) site (Fig. 1A). The construct number of terminal end buds increase with age (until f6 weeks of was injected into the pronucleus of fertilized CB57BL/6 mouse age, then they begin to decline); however, the overall ductal oocytes and transferred into pseudopregnant CB6/F1 foster complexity observed in hPar1-overexpressing glands was consis- mothers. Of the 60 resulting pups, 4 transgenic founders passed tently greater than that of their age-matched wt counterparts the transgene to progeny, generating four independent transgenic (Fig. 1C, V-13w; 3W-10W of virgin mice not shown). During early lines (MMTV-LTR-hPar1 L1-L4, Fig. 1B). We chose lines L2 and L4, pregnancy, precocious development of hPar1-tg glands was even which exhibited high hPar1 levels by Southern blot analysis, for more striking. Lobuloalveolar structures become evident in wt further studies. females during early pregnancy (Fig. 1C, wt P8d and P12d). This To determine the effects of hPar1 expression on mammary gland phenotype commenced earlier in the hPar1-overexpressing glands, morphogenesis, histologic preparations of mammary tissues from exhibiting a high level of ductal network complexity and increased various developmental stages were analyzed. Whole mounts of alveolar buds (see Fig. 1C, hPar1-tg; P8d and P12d).

Figure 2. Tissue localization and pattern of hPar1expression in hPar1-tg mammary glands. A, in situ hybridization analysis of hPar1expression in sections of mammary gland in the hPar1-tg and wt mice. hPar1-specific expression is seen in 13w virgin mice and in pregnant mice at 4, 8, and 12 days. No hPar1 was observed in either wt virgin or through the pregnancy period (data not shown) and 12-day pregnant wt (bottom). High magnification shows hPar1expression confined specifically to the luminal epithelial layer of the mammary gland (arrow in the square). B, RT-PCR analysis of hPar1expression in different development stages of the transgenic mice (5w, 8w 10w, 13w, P4d, and P8d) compared with age-matched wt mice. C, RT-PCR analysis of hPar1expression in different organs of MMTV-LTR-hPar1 transgenic mice. Organs from hPar1-tg mice were removed, RNA was isolated, and RT-PCR analysis was done as described in Materials and Methods. PCR for hPar1and L19(control gene) was done. D, i, mammary glands overexpressing hPar1display increased epithelial proliferation. Mammary glands from wt and hPar1-tg mice were examined for PCNA expression by immunohistochemistry. The staining was done at 5w, 10w, and 13w of the virgin stage, and 4 and 8 days of pregnancy, respectively. ii, histogram indicates the proliferation index, defined as the number of PCNA-positive nuclei per total number of nuclei. The values represent the average proliferation index for three different mice mammary glands.

www.aacrjournals.org 5227 Cancer Res 2006; 66: (10). May 15, 2006

Figure 3. Increasing mouse Wnt and h-catenin expression in hPar1-tg mammary glands. A, RT-PCR analysis of mouse wnt-4 and wnt-7b expression in mammary glands of wt and hPar1-tg mice. B, immunoprecipitation analysis of mouse wnt-4 expression in hPar1-tg and wt mice. The transgenic mice show higher expression levels in 4- and 8-day pregnant mice compared with the wt mice. The same amount of protein was loaded and blotted with h-actin as control. The immunoprecipitation data seem to be specific because normal immunoglobulin G (IgG) did not show any protein precipitation. C, increased cyclin D1induction in hPar1-tg mice mammary glands. Western blot analysis of cyclin D1expression in wt and hPar1-tg mammary glands. h-Actin was used as a control for protein loading. The average relative expression levels are shown by a histogram where wt mice were given an arbitrary value of 1.00.

H&E staining of tissue sections revealed accelerated lobuloal- alveolar epithelia resumes also in the wt mice; however, we observe veolar development of the hPar1-overexpressing mammary glands an accelerated proliferation rate in the hPar1-overexpressing mice. as compared with wt mice (Fig. 1D). In situ hybridization analysis Wnt-4 and wnt-7b are overexpressed in hPar1-tg mammary with hPar1 Dig-labeled riboprobes revealed abundant expression of glands. Because one of the mechanisms of mammary gland hPar1 in the transgenic animals at all ages examined. This hyperplasia suggested an essential role for Wnt-4 in ductal side expression is confined to the luminal epithelium and is absent branching (24), we thought to screen for levels of mouse wnts in from the myoepithelial layer (see arrow in Fig. 2A, P4d, high normal wt versus transgenic mice mammary glands. Wnt encodes a magnification). It is of interest that the luminal site at which hPar1 large family of secreted glycoproteins and binds to a complex of is expressed is also the site of wnt-4, which was elegantly shown to Frizzled family of seven-transmembrane receptors (15, 16, 18) and play a major role in mammary gland ductal side branching (24, 25). LRP5/6 coreceptors (12, 13, 21). Whereas wnt exerts its effects on Reverse transcriptase-PCR (RT-PCR) analysis of RNA extracted embryonic cells and in pathologic carcinoma progression, the from the mammary glands of individual mice showed a similar proximal steps of the signaling cascade beyond Frizzled are less pattern: high hPar1 expression levels in the hPar1-tg mammary well understood. glands (with a slightly elevated expression during pregnancy) as The expression of wnt in mouse mammary glands was initially compared with no expression in age-matched wt animals (Fig. 2B). analyzed by RT-PCR. We have examined an entire spectrum of RT-PCR analysis revealed that expression of hPar1 in both lines mouse wnts (i.e., wnt-7a, -5b, -6, and -10b, known to be present in (e.g., L2 and L4) was confined to the mammary glands where it was the epithelia, as well as wnt-2 and -5a, present in mouse mammary present at high levels (see Fig. 2C). stroma). Our data revealed specific enhancement of wnt-4 and Whether increased ductal network complexity in hPar1-over- wnt-7b in hPar1-tg mammary glands. No change was detected in expressing mammary glands is attributed solely to ductal side wnt-2, -5a, -6, as well as in wnt-7a, -5b, and -10b. Normally, low branching or due to on-going proliferation was next addressed. For levels of wnt-4 are observed in the virgin wt mice, which increased this, we immunostained sections of normal and hPar1-over- slightly during pregnancy. In comparison, hPar1-overexpressing expressing mammary glands with an antibody to PCNA, a marker mice exhibit greater levels of wnt-4 in the virgin mammary glands, for S-phase cells (Fig. 2D, i and ii). The proliferation index is most visibly at 10 weeks of age. An additional increase in wnt-4 defined as the number of PCNA-positive nuclei of alveolar levels is observed throughout pregnancy in the transgenic animals epithelial cells divided by the total number of nuclei. Whereas (Fig. 3A). To substantiate the effects of hPar1 overexpression on increased proliferation is seen in normal wt mammary already at wnt-4 levels, we did immunoprecipitation analysis on mammary 5 weeks of age, the proliferation index in the transgenic mice shows gland protein lysates (Fig. 3B). The results confirm a profound a 3-fold increase. The difference in proliferation index between wt increase in the levels of Wnt-4 protein when compared with age- and hPar1-overexpressing animals is particularly pronounced and gestation-matched controls, especially in pregnant hPar1- between postnatal weeks 10 and 13 (e.g., virgin 10W and 13W) overexpressing mammary glands at 4 and 8 days of pregnancy. We when proliferation of alveolar epithelia is normally arrested in wt cannot currently explain why a regulated increase in wnt-4 levels animals (Fig. 2D, i and ii). During pregnancy, proliferation of is observed in hPar1-tg, low in virgin mice, and significantly

Cancer Res 2006; 66: (10). May 15, 2006 5228 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2006 American Association for Cancer Research. hPar1 Induced b-Catenin Stabilization increased during pregnancy because hPar1 levels are similarly as compared with none in age-matched wt animals (Fig. 4A). This expressed throughout these stages. It may be that hormones shows that hPar1 overexpression leads to stabilization of h-catenin present during pregnancy may be involved in the process. In that is ultimately imported to the nuclei, where it may further addition to the increase in wnt-4, we found enhancement in the regulate gene transcription. levels of wnt-7b but no effect on levels of other wnts normally To address the molecular interrelation between hPar1 and expressed in the mouse mammary glands (24). This suggests that h-catenin stabilization, we examined transformed epithelial cells in hPar1 specifically regulates the expression of at least the two parallel to the transgenic-mice tissues. It is important to note that mouse wnt genes either directly or indirectly. most breast cancer cell lines have impaired h-catenin signaling One target of transcriptional activation by the wnt pathway is machinery system as shown in a wide screening done by van de cyclin D1 (19, 26). Cyclin D1, the major G1 cyclin expressed in Wetering et al. (27). It has been reported that only 1 of the 15 breast mammary epithelial cells, is deficient in mice resistant to breast cancer cell lines analyzed showed TCF-mediated h-catenin activity cancers induced by the neu and ras oncogenes, implying its tight although most of them expressed h-catenin. We therefore chose to connection to the Neu-Ras pathway transformation (20). We found do analyses in a panel of cell lines, including colorectal cancer cell increased cyclin D1 expression in the hPar1-overexpressing lines, widely used for studies in wnt/h-catenin signaling machinery. mammary glands as compared with wt counterparts (Fig. 3C). These cell lines express high hPar1 levels. When HT-29 colon At 4 days of pregnancy, cyclin D1 levels are 4.5-fold higher in the cancer cells were processed to obtain discrete cell fractions of transgenic mice than in wt animals, and at 8 days of pregnancy, either membrane or cytoplasm after PAR1 activation, the following they remain elevated 2.5 times higher in the transgenic animals. pattern was seen. Whereas the majority of h-catenin was initially Accumulation of B-catenin in hPar1-overexpressing mice. localized at the membrane pool, a significant increase in h-catenin To explore the possible signaling pathway downstream of wnt,we cytoplasmic fraction was seen as soon as 1 hour after PAR1 acti- evaluated the expression levels of h-catenin in the mammary vation and remained elevated at 2 hours activation (Fig. 5A). A glands overexpressing hPar1 and age-matched controls. Western similar pattern of kinetics is obtained when HCT116 cells were blot analysis showed a striking accumulation of h-catenin in the analyzed (data not shown). When the nuclear fraction of HCT116 mammary gland tissues of both virgin and pregnant hPar1 trans- cells was isolated and analyzed for levels of h-catenin, an increase genes as compared with minimal levels in the mammary tissue of is seen as soon as 2 hours after PAR1 activation, a significant wt controls (Fig. 4B). Immunohistochemistry staining revealed elevation after 4 hours, and remaining enhanced up to 24 hours of abundant localization of h-catenin in the nuclei of alveolar examination (Fig. 5B). Similarly, when these cells were pretreated epithelial cells, taking place only in hPar1-overexpressing glands, with LiCl (a known glycogen synthase kinase 3h inhibitor) followed by SFLLRNP (a synthetic peptide of PAR1 internal ligand) for PAR1 activation, an induced enhancement of h-catenin stabilization was observed. The induction was seen as early as 30 minutes after activation (3-fold), remaining up to 5 hours (3.2-fold; Fig. 5C). Thus, we notice a further induction in h-catenin stabilization, following PAR1 activation, regardless of whether pretreated with LiCl or without. These colon cancer cell lines, however, harbor mutations in either antigen-presenting cells (APC; e.g., HT-29 cells, truncated APC) or h-catenin (e.g., HCT116, mutated amino acid 45), resulting in a constitutively activated h-catenin (28). We did, therefore, a control experiment using a transformed cancer cell line of intact h-catenin machinery system, the RKO cells. RKO is a colorectal cancer cell line transformed on the basis of gene instability (29). This cell line displays microsatellite instability for hypermethylation of the hMLH1 promoter but expresses wt APC, h-catenin, and p53 (28, 30). Thus, RKO is a colon cancer cell line on the background of a defective mismatch repair system. We analyzed the effect of PAR1 in RKO cells to eliminate the possibility that PAR1-induced h-catenin takes place on the background of a constitutively activated h-catenin/TCF-regulated transcription system. Whereas h-catenin was not detected initially in these cells, pretreatment with either LiCl or MG132 (an inhibitor of the proteasomal system) followed by PAR1 activation showed a further enhancement in nuclear h-catenin (Fig. 5Di). Semiquantitative RT-PCR analysis shows the levels of hPar1 in various types of epithelial tumor cell lines. The data show high hPar1 levels in HCT116 and HT-29, very low levels in RKO and SW480, and none in CaCO2 cells (Fig. 5Dii). Figure 4. hPar1-tg mammary glands show increased h-catenin expression. h A, immunohistochemical staining shows h-catenin localization in the mammary Essentially, induced -catenin levels are observed also in MDA231 glands of hPar1-tg and wt mice. At 10 days of pregnancy, nuclear h-catenin breast cancer cell lines. Activation of PAR1 either before or after LiCl staining is observed in hPar1-tg mice but not in wt mice (Â40). Bottom, higher pretreatment showed elevation in h-catenin levels (Fig. 5E). Taken magnification (Â100). B, Western blot analysis of h-catenin expression in hPar1-tg mammary glands at different developmental stages compared with together, we conclude that PAR1 activation leads to stabilization of age-matched wt mammary glands. h-catenin analyzed either in a breast cancer cell line or in colon www.aacrjournals.org 5229 Cancer Res 2006; 66: (10). May 15, 2006

Figure 5. Activation of PAR1induces h-catenin stabilization. A, cytoplasmic accumulation of h-catenin. Separate fractions of membrane and cytoplasm in HT-29 colorectal cells show marked accumulation of h-catenin in the cytoplasmic fraction after thrombin PAR1activation. Significant enhancement is observed 1hour after activation and remains elevated at 2 hours. B, nuclear accumulation of h-catenin. The nuclear fraction of HCT116 cells shows consistent accumulation of h-catenin after PAR1activation (as indicated by addition of SFLLRNP), initiated 2 hours after activation and remaining elevated up to 24 hours of examination. C, LiCl pretreatment (2 hours) of HT-29 cells followed by PAR1activation (for the indicated periods of time) results in enhanced levels of h-catenin in the cell nuclei. Histogram shows the estimated levels of h-catenin before and after PAR1activation, obtained either with or without LiCl pretreatment (for 2 hours). D, i, PAR1activation induces h-catenin levels in RKO cells. RKO cells pretreated with either LiCl or MG132 before (À) and after (+) PAR1activation. Whereas no h-catenin is observed in nontreated RKO cells, it is seen following both treatments. After PAR1activation, a further increase in h-catenin levels is noticed. ii, semiquantitative RT-PCR analysis of colon cancer epithelial cell lines. E, nuclear h-catenin levels in MDA 231breast cancer cells. PAR1activation induces h-catenin levels in breast cancer cells, similar to the effects seen in colorectal cell lines, before and after LiCl pretreatment. cancer cells transformed regardless of whether on the background extracellular domains of both receptors, the SFRPs are effective of intact h-catenin system or constitutively activated h-catenin. in antagonizing Wnt signaling machinery (21). Enforced overexpression of SFRP5 and SFRP2 abrogate B- In parallel, 293T cells were transfected with either SFRP5, SFRP2, catenin stabilization following PAR1 activation. To examine or both and the conditioned medium containing secreted SFRP5, whether PAR1 activation leads primarily to increased Wnt SFRP2, or both was collected, providing a source for SFRPs. Equal expression that ultimately induces h-catenin stabilization, we have transfection yields were observed following SFRP2 and SFRP5 analyzed members of the Wnt antagonist family, SFRPs, on PAR1- plasmid transfections (Fig. 6Bb). When these conditioned media induced h-catenin stabilization. For this purpose, we cotransfected were administered to HT-29 cells before and after PAR activation, the HCT116 cell line with Myc-wt-h-catenin and SFRP5 plasmids. the following outcome was observed: whereas enhanced stabiliza- As seen in Fig. 6A, PAR1 activation led to a marked increase in the tion of h-catenin was seen after PAR1 activation, it was efficiently ectopically transfected h-catenin detected by the Myc tagging of abolished in the presence of conditioned medium containing either nuclear h-catenin after PAR1 activation. This increase is abrogated SFRP5, SFRP2, or both (Fig. 6B). This indicates that SFRPs (at least in the presence of enforced overexpression of SFRP5.The SFRP2 and SFRP5) can efficiently inhibit PAR1-induced h-catenin neutralizing effect of SFRP indicates that PAR1 activation induced stabilization. h-catenin stabilization primarily via the induction of wnt. In the siRNA-mediated hPar1 silencing inhibits PAR1-induced presence of SFRP5, it sequesters Wnt, which is generated by the Wnt-4 generation. We next evaluated whether a direct link activation of PAR1, binds to the Frizzled cell-surface receptors between PAR1 and wnt-4 exists. For this purpose, we have examined through the sequence homology domain cysteine-rich domain, and levels of wnt-4 expression in both RKO and HT-29 cells. Enhanced forms a complex also with LRP5/6 coreceptor (21, 31–33). It is levels of wnt-4 are seen as soon as 4 hours after PAR1 activation, postulated that by forming a neutralizing complex with the which were significantly elevated by 12 hours (Fig. 6Ci). The same

Cancer Res 2006; 66: (10). May 15, 2006 5230 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2006 American Association for Cancer Research. hPar1 Induced b-Catenin Stabilization pattern is seen in HT-29 cells, but with a different kinetics. Increased generated mice overexpressing hPar1 in their mammary glands. levels of wnt-4 take place earlier, showing increased levels as soon as These glands exhibit precocious hyperplasia and recapitulate the 1 hour after PAR1 activation with a maximal induction observed by hallmarks of Wnt signaling pathway as shown by h-catenin 4 hours (Fig. 6Di). stabilization and nuclear localization. Altogether, the properties When constructs of anti-hPar1 RNA interfering sequences portrayed by these glands point to a novel molecular trail elicited (siRNA) ligated into pLentilox 3.7 viral vector (Van Parij’s by hPar1. The fact that SFRPs, members of Wnt antagonist family, laboratory) were infected into either RKO or HT-29 cells (Fig. 6Cii effectively abrogate PAR1-induced h-catenin stabilization shows and Dii, respectively), effective depletion of hPar1 is noted in two that hPar1 expression and activation lead primarily to Wnt gene- of the four constructs examined (Fig. 6Dii). Under these conditions ration, which further induces h-catenin stabilization via an of hPar1-depleted levels, a significant abrogation of PAR1-induced autocrine loop (21). Once generated, it is postulated that secreted wnt-4 enhancement is observed. Wnts reactivate their cognate cell-surface receptors affecting downstream cell signaling. In addition, we show that viral constructs of hPar1 siRNA, depleting hPar1 levels, inhibit Discussion effectively PAR1-induced Wnt-4 generation. This is the first To evaluate the causal relationship between hPar1 gene demonstration of a direct link between hPar1 and Wnt-4. expression, breast tissue morphogenesis, and hyperplasia, we Neutralization of hPar1 and, hence, the abrogation of Wnt-4 take

Figure 6. SFRPs suppress PAR1-induced h-catenin stabilization. A, HCT116 cells were cotransfected with Myc-wt-h-catenin and SFRP5. Nuclear fractions were isolated and levels of h-catenin were analyzed by Western blot analysis before and after PAR1activation. Detection of the ectopically inserted wt- h-catenin was followed by its Myc tagging (i.e., anti-Myc antibodies; 9E10, Santa Cruz Biotechnology). Induced nuclear accumulation in Myc-wt-h-catenin is seen after PAR1 activation (b) as compared with (a). Whereas no effect on h-catenin levels is seen following mock plasmid cotransfection (data not shown), the cotransfection of Myc-wt-h-catenin with SFRP5 to the cells following PAR1activation resulted in distinct inhibition of h-catenin levels. B, i, application of conditioned medium (CM) containing SFRP5 or SFRP2 or both efficiently abrogates PAR1-induced h-catenin stabilization. Nuclear fractions of HT-29 cells were isolated either before (a and d) or after (b, c, e, f, g, and h) PAR1activation, following incubation of some of the cells with either SFRP2-CM ( e, g, and h) or SFRP5-CM (d, e and g, h). Detection of endogenous h-catenin levels is evaluated by the use of anti-h-catenin antibodies. The amount of nuclear proteins applied is monitored by a nuclear housekeeping control gene, lamin. ii, RT-PCR analysis showing levels of SFRP2 and SFRP5 following transfection (+) and levels before transfection (À). iii, histogram analyses of SDS-PAGE band intensity (nearly 2-fold induction is observed by PAR1activation, which is reduced to a background level in the presence of SFRPs). C and D, hPar1siRNA inhibits PAR1-induced Wnt-4. RT-PCR analysis of Wnt-4 in either RKO ( C) or HT-29 cells (D) following PAR1activation. Levels of wnt-4 were compared with a housekeeping control gene, GAPDH. Infection of siRNA constructs in RKO (Cii) or HT-29 cells (Dii) potently inhibited PAR1-induced wnt-4 levels. Inhibition of PAR1-induced wnt-4 is obtained under conditions where hPar1 is depleted [using the siRNA constructs i and iv as shown for HT-29 cells (Dii); similarly seen in RKO-depleted hPar1cells (data not shown)]. www.aacrjournals.org 5231 Cancer Res 2006; 66: (10). May 15, 2006

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2006 American Association for Cancer Research. Cancer Research place regardless of whether on the background of intact h-catenin model. Our preliminary data show that orthotopic injection of system but defective in mismatch repair system or in cells with MCF-7-hPar1 stable clones to the mammary fat pads induces mutated APC on the background of constitutively activated h- tumors [data generated in our laboratory as also by Boire et al. catenin. Whether Wnt-4 promotes a canonical or a noncanonical (40)]. The orthotopic injections of MCF-7 clones overexpressing signaling machinery is controversial. Recently, Wnt-4 was shown various hPar1 forms (e.g., full-length, Y397Z superactive or to promote, at least in a specific tissue context of Madin-Darby truncated form devoid the entire cytoplasmic tail) show advanced canine kidney epithelial cells, the canonical h-catenin pathway tumor formation in the hPar1 full-length, compared with very (34). The association between hPar1 and wnt-7b remains to be little or none in truncated hPar1 or mock-transfected cells.4 It is fully elucidated. The molecular pathway by which PAR1 possible that cross-mating of the hPar1-tg mice with known expression and activation induce Wnt3 remains as well to be models of mammary gland tumor progression (e.g., transgenes shown. overexpressing the erBb 1-4 gene family) may result in accelerated It is intriguing to note that a recent study (35) also showed a rates of tumor formation (currently being explored in our mechanism of h-catenin stabilization for another class of G laboratory). protein–coupled receptors belonging to the lysophosphatidic acid The extracellular matrix barrier surrounding the myoepithelial receptors. Lysophosphatidic acid receptors are shown to play a mammary gland ducts plays an active role as a depot site for central part in several types of cancer (36, 37). Thus, PAR1 G bioactive molecules, among which are proangiogenic and growth protein–coupled receptor joins at least lysophosphatidic acid 2 and factors as well as adhesion molecules (41). Indeed, loss of the 3 receptors, acting via h-catenin stabilization and shown to induce myoepithelium and extracellular matrix is associated with invasive increased expression of c-Myc and cyclin D1, two of the characteristics in breast cancer (42, 43). hPar1 overexpression in downstream target genes of h-catenin transcriptional activity. A the mammary glands elicits increased lateral budding and common theme is thus emerging, pointing to the significance of precocious mammary hyperplasia, suggesting the induction of G-proteins as the proximal missing piece in the Frizzled seven- stochastic events by hPar1, a gene that plays a central part in the transmembrane receptor/s signaling puzzle (under our current invasive phenotype. investigation). We hypothesize that several PAR1 ligand activators are present The Wnt family of secreted glycoproteins act through two cell- either in the stroma microenvironment or in neighbor host cells surface receptors, the Frizzled and the single-transmembrane surrounding the invading cancer cells. For example, circulating protein LRP5/6 (or Arrow). These components are essential for prothrombin is abundantly present in wt mouse mammary glands efficiently transducing signals from Wnt, an extracellular ligand, (data not shown). It is possible that other candidates belonging to to the intracellular pathway that stabilizes h-catenin. The current the coagulation cascade (44–46) may functionally activate PAR1. model suggests that Wnt protein binds to the extracellular domains Recently, an intriguing observation suggests that matrix metal- of both LRPand Frizzled receptors, forming membrane-associated loprotease 1 (MMP-1), derived from the stromal microenviron- hetero-oligomers that interact with both Dishevelled (via the ment, activates PAR1 at the same thrombin cleavage/activation intracellular portion of Frizzled) and Axin (via the intracellular site (40). Immobilization of proteases to the stroma microenvi- domain of LRP; refs. 12, 13, 38). The molecular mechanism, ronment allows the sustained and localized functions of proteo- however, by which these two types of membrane receptor synergies lytic enzymes such as cathepsin G, elastase, and plasmin, as well as transmit Wnt signals is yet unknown. MMP-1, all of which may eventually have an effect on hPar1 The fact that exogenous administration of specific Wnt signaling (47–49). antagonists (i.e., SFRPs) inhibits PAR1-induced h-catenin stabiliza- We propose that hPar1 expression and activation elicit primarily tion, further afflicting downstream components, implies a Wnt wnt expression and the striking stabilization of h-catenin. Our autocrine loop. We show now that overexpression of SFRP5 confers studies highlight hPar1 gene as a potent target for cancer therapy inhibition of PAR1-induced h-catenin stabilization by either because neutralization of the gene may effectively inhibit initiation enforced SFRP5 expression or the ectopic application of SFRP5 of oncogenicity via wnt generation, h-catenin stabilization, and and/or SFRP2 or both. This result is in line with the above studies LEF/TCF transcriptional response (i.e., capable of inducing an (21, 22) showing an autocrine loop of Wnt signaling. It has been array of responsive genes-downstream). Elucidation of this shown, for example, that colorectal cancer develops on the fundamental core molecular pathway may lead nonetheless to a background of SFRP gene neutralization (following the hyper- better understanding of cancer progression. methylated SFRP promoter silencing; refs. 22, 39). Whereas in the presence of functionally active SFRP, Wnt signaling is attenuated, the normal equilibrium is disrupted when SFRPs are hyper- Acknowledgments methylated and silenced. Furthermore, functionally active SFRPs Received 11/29/2005; revised 2/9/2006; accepted 3/6/2006. suppress Wnt signaling also in cells possessing constitutively Grant support: U.S. Army grant DAMD17-00-1-0277, Israel Science Foundation, activated mutations of h-catenin (22). The discrete pattern of founded by the Israel Academy of Sciences and Humanities, and the Israel Cancer individual SFRP gene expression in hPar1-tg mice is under current Association and Israel Cancer Research Fund (R. Bar-Shavit). The costs of publication of this article were defrayed in part by the payment of page investigation. charges. This article must therefore be hereby marked advertisement in accordance The hPar1-overexpressing mammary glands show a phenotype with 18 U.S.C. Section 1734 solely to indicate this fact. We thank Dr. Uri Gat for the reagents and for stimulating discussions and Dr. Cliff of precocious hyperplasia. We cannot currently explain the lack Tabin (Department of Genetics, Harvard Medical School, Boston, MA) for the generous of tumor formation in the hPar1-tg mammary gland mice gift of the mouse wnt-4 plasmid.

3 Manuscript in preparation. 4 I. Cohen et al., manuscript in preparation.

Cancer Res 2006; 66: (10). May 15, 2006 5232 www.aacrjournals.org

References MYC as a target of the APC pathway. Science 1998;281: catenin activity in kidney epithelial cells. Exp Cell Res 1509–12. 2004;298:369–87. 1. Vu TK, Hung DT, Wheaton VI, Coughlin SR. Molecular 18. Nusse R. WNT targets. Repression and activation. 35. Yang M, Zhong WW, Srivastava N, et al. G protein- cloning of a functional thrombin receptor reveals a Trends Genet 1999;15:1–3. coupled lysophosphatidic acid receptors stimulate novel proteolytic mechanism of receptor activation. Cell 19. Shtutman M, Zhurinsky J, Simcha I, et al. The cyclin proliferation of colon cancer cells through the h-catenin 1991;64:1057–68. D1 gene is a target of the h-catenin/LEF-1 pathway. pathway. Proc Natl Acad Sci U S A 2005;102:6027–32. 2. Coughlin SR. Thrombin signalling and protease- Proc Natl Acad Sci U S A 1999;96:5522–7. 36. Kamrava M, Simpkins F, Alejandro E, et al. Lyso- activated receptors. Nature 2000;407:258–64. 20. Tetsu O, McCormick F. h-Catenin regulates expres- phosphatidic acid and endothelin-induced proliferation 3. Even-Ram S, Uziely B, Cohen P, et al. Thrombin sion of cyclin D1 in colon carcinoma cells. Nature 1999; of ovarian cancer cell lines is mitigated by neutraliza- receptor overexpression in malignant and physiological 398:422–6. tion of granulin-epithelin precursor (GEP), a prosurvival invasion processes. Nat Med 1998;4:909–14. 21. Bafico A, Liu G, Goldin L, et al. An autocrine factor for ovarian cancer. Oncogene 2005;24:7084–93. 4. Even-Ram SC, Grisaru-Granovsky S, Pruss D, et al. The mechanism for constitutive Wnt pathway activation in 37. Mills GB, Moolenaar WH. The emerging role of pattern of expression of protease-activated receptors human cancer cells. Cancer Cell 2004;6:497–506. lysophosphatidic acid in cancer. Nat Rev Cancer 2003;3: (PARs) during early trophoblast development. J Pathol 22. Suzuki H, Watkins DN, Jair KW, et al. Epigenetic 582–91. 2003;200:47–52. inactivation of SFRPgenes allows constitutive WNT 38. Mao B, Wu W, Li Y, et al. LDL-receptor-related 5. Henrikson KP, Salazar SL, Fenton JW, Pentecost BT. signaling in colorectal cancer. Nat Genet 2004;36:417–22. protein 6 is a receptor for Dickkopf proteins. Nature Role of thrombin receptor in breast cancer invasiveness. 23. Williams JM, Daniel CW. Mammary ductal elonga- 2001;411:321–5. Br J Cancer 1999;79:401–6. tion: differentiation of myoepithelium and basal lamina 39. Suzuki H, Gabrielson E, Chen W, et al. A genomic 6. Nierodzik ML, Chen K, Takeshita K, et al. Protease- during branching morphogenesis. Dev Biol 1983;97: screen for genes up-regulated by demethylation and activated receptor 1 (PAR-1) is required and rate-limiting 274–90. histone deacetylase inhibition in human colorectal for thrombin-enhanced experimental pulmonary metas- 24. Brisken C, Heineman A, Chavarria T, et al. Essential cancer. Nat Genet 2002;31:141–9. tasis. Blood 1998;92:3694–700. function of Wnt-4 in mammary gland development 40. Boire A, Covic L, Agarwal A, et al. PAR1 is a matrix 7. Yin YJ, Salah Z, Grisaru-Granovsky S, et al. Human downstream of progesterone signaling. Genes Dev 2000; metalloprotease-1 receptor that promotes invasion and protease-activated receptor 1 expression in malignant 14:650–4. tumorigenesis of breast cancer cells. Cell 2005;120: epithelia: a role in invasiveness. Arterioscler Thromb 25. Brisken C, Park S, Vass T, et al. A paracrine role for 303–13. Vasc Biol 2003;23:940–4. the epithelial progesterone receptor in mammary gland 41. Batsakis JG, el-Naggar AK. Myoepithelium in salivary 8. Grisaru-Granovsky S, Salah Z, Maoz M, et al. development. Proc Natl Acad Sci U S A 1998;95:5076–81. and mammary neoplasms is host-friendly. Adv Anat Differential expression of protease activated receptor 1 26. Yu Q, Geng Y, Sicinski P. Specific protection against Pathol 1999;6:218–26. (Par1) and pY397FAK in benign and malignant breast cancers by cyclin D1 ablation. Nature 2001;411: 42. Silberstein GB. Postnatal mammary gland morpho- human ovarian tissue samples. Int J Cancer 2005;113: 1017–21. genesis. Microsc Res Tech 2001;52:155–62. 372–8. 27. van de Wetering M, Barker N, Harkes IC, et al. 43. Xiao G, Liu YE, Gentz R, et al. Suppression of breast 9. Martin CB, Mahon GM, Klinger MB, et al. The Mutant E-cadherin breast cancer cells do not display cancer growth and metastasis by a serpin myoepithe- thrombin receptor, PAR-1, causes transformation by constitutive Wnt signaling. Cancer Res 2001;61:278–84. lium-derived serine proteinase inhibitor expressed in activation of Rho-mediated signaling pathways. Onco- 28. da Costa LT, He TC, Yu J, et al. CDX2 is mutated in a the mammary myoepithelial cells. Proc Natl Acad Sci gene 2001;20:1953–63. colorectal cancer with normal APC/h-catenin signaling. U S A 1999;96:3700–5. 10. Young D, O’Neill K, Jessell T, Wigler M. Character- Oncogene 1999;18:5010–4. 44. Riewald M, Kravchenko VV, Petrovan RJ, et al. Gene ization of the rat mas oncogene and its high-level 29. Brattain MG, Levine AE, Chakrabarty S, et al. induction by coagulation factor Xa is mediated by expression in the hippocampus and cerebral cortex of Heterogeneity of human colon carcinoma. Cancer activation of protease-activated receptor 1. Blood 2001; rat brain. Proc Natl Acad Sci U S A 1988;85:5339–42. Metastasis Rev 1984;3:177–91. 97:3109–16. 11. Zohn IE, Klinger M, Karp X, et al. G2A is an 30. Eshleman JR, Markowitz SD, Donover PS, et al. 45. Riewald M, Ruf W. Mechanistic coupling of oncogenic G protein-coupled receptor. Oncogene 2000; Diverse hypermutability of multiple expressed sequence protease signaling and initiation of coagulation by 19:3866–77. motifs present in a cancer with microsatellite instability. tissue factor. Proc Natl Acad Sci U S A 2001;98: 12. Cong F, Schweizer L, Varmus H. Wnt signals across Oncogene 1996;12:1425–32. 7742–7. the plasma membrane to activate the h-catenin 31. Leyns L, Bouwmeester T, Kim SH, et al. Frzb-1 is a 46. Ruf W. PAR1 signaling: more good than harm? Nat pathway by forming oligomers containing its receptors, secreted antagonist of Wnt signaling expressed in the Med 2003;9:258–60. Frizzled and LRP. Development 2004;131:5103–15. Spemann organizer. Cell 1997;88:747–56. 47. Kuliopulos A, Covic L, Seeley SK, et al. Plasmin 13. Liu G, Bafico A, Aaronson SA. The mechanism of 32. Lin K, Wang S, Julius MA, et al. The cysteine-rich desensitization of the PAR1 thrombin receptor: kinetics, endogenous receptor activation functionally distin- frizzled domain of Frzb-1 is required and sufficient for sites of truncation, and implications for thrombolytic guishes prototype canonical and noncanonical Wnts. modulation of Wnt signaling. Proc Natl Acad Sci U S A therapy. Biochemistry 1999;38:4572–85. Mol Cell Biol 2005;25:3475–82. 1997;94:11196–200. 48. Loew D, Perrault C, Morales M, et al. Proteolysis of 14. Bienz M, Clevers H. Linking colorectal cancer to Wnt 33. Rattner A, Hsieh JC, Smallwood PM, et al. A family of the exodomain of recombinant protease-activated signaling. Cell 2000;103:311–20. secreted proteins contains homology to the cysteine- receptors: prediction of receptor activation or inactiva- 15. Polakis P. Wnt signaling and cancer. Genes Dev 2000; rich ligand-binding domain of frizzled receptors. Proc tion by MALDI mass spectrometry. Biochemistry 2000; 14:1837–51. Natl Acad Sci U S A 1997;94:2859–63. 39:10812–22. 16. Willert K, Nusse R. h-Catenin: a key mediator of Wnt 34. Lyons JP, Mueller UW, Ji H, et al. Wnt-4 activates the 49. Sambrano GR, Huang W, Faruqi T, et al. Cathepsin G signaling. Curr Opin Genet Dev 1998;8:95–102. canonical h-catenin-mediated Wnt pathway and binds activates protease-activated receptor-4 in human plate- 17. He TC, Sparks AB, Rago C, et al. Identification of c- Frizzled-6 CRD: functional implications of Wnt/h- lets. J Biol Chem 2000;275:6819–23.

www.aacrjournals.org 5233 Cancer Res 2006; 66: (10). May 15, 2006

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2006 American Association for Cancer Research. Mammary Gland Tissue Targeted Overexpression of Human Protease-Activated Receptor 1 Reveals a Novel Link to β -Catenin Stabilization

Yong-Jun Yin, Vered Katz, Zaidoun Salah, et al.

Cancer Res 2006;66:5224-5233.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/66/10/5224

Cited articles This article cites 49 articles, 18 of which you can access for free at: http://cancerres.aacrjournals.org/content/66/10/5224.full#ref-list-1

Citing articles This article has been cited by 4 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/66/10/5224.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://cancerres.aacrjournals.org/content/66/10/5224. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.