Abstract
Ovarian endometrioid adenocarcinomas (OEAs) frequently exhibit constitutive activation of canonical WNT signaling, usually as a result of oncogenic mutations that stabilize and dysregulate the β-catenin protein. In prior work, we used microarray-based methods to compare gene expression in OEAs with and without dysregulated β-catenin as a strategy for identifying novel β-catenin/TCF target genes with important roles in ovarian cancer pathogenesis. Among the genes highlighted by the microarray studies was MSX2, which encodes a homeobox transcription factor. We found MSX2 expression was markedly increased in primary human and murine OEAs with dysregulated β-catenin compared to OEAs with intact β-catenin regulation. WNT pathway activation by WNT3a ligand or GSK3β inhibitor treatment potently induced MSX2, and ectopic expression of a dominant negative form of TCF4 inhibited MSX2 expression in ovarian cancer cells. Chromatin immunoprecipitation studies demonstrated that β-catenin/TCF directly regulates MSX2 expression via binding to TCF binding elements in multiple regions of the MSX2 gene. Notably, ectopic MSX2 expression was found to promote neoplastic transformation of the rodent RK3E model epithelial cell line and to enhance the invasiveness of immortalized human ovarian epithelial cells in vitro and ovarian carcinoma cells in vivo. Inhibition of endogenous MSX2 expression in ovarian endometrioid cancer cells carrying a β-catenin mutation using shRNA approaches inhibited neoplastic properties of the cells in vitro and in vivo. Expression of MSX2 in selected ovarian carcinoma cells induced changes suggestive of epithelial-mesenchymal transition (EMT), but based on analysis of ovarian cell lines and primary tumor tissues, effects of MSX2 on EMT appear to be complex and context-dependent. Our findings indicate MSX2 is a direct downstream transcriptional target of β-catenin/TCF and has a key contributing role in the cancer phenotype of OEAs carrying WNT/β-catenin pathway defects.
Keywords: MSX2, ovarian cancer, endometrioid, WNT signaling, β-catenin
INTRODUCTION
Ovarian endometrioid adenocarcinoma (OEA) is the second most common type of ovarian cancer. We and others have previously shown that 16-38% of ovarian endometrioid adenocarcinomas have constitutive activation of the canonical (β-catenin-dependent) WNT signaling pathway, usually as a result of oncogenic mutations in the β-catenin protein or, less commonly, due to inactivating mutations in key negative regulators of β-catenin such as the APC and AXIN tumor suppressor proteins (Catasus et al., 2004; Gamallo et al., 1999; Sagae et al., 1999; Wright et al., 1999; Wu et al., 2001). The mutations lead to stabilization of β-catenin and increases in the “free” pool of β-catenin in the cytoplasm and nucleus. In the nucleus, β-catenin binds to TCF/LEF transcription factors, leading to altered transcription of β-catenin/TCF-regulated target genes, a subset of which likely play critical roles in neoplastic transformation.
Many candidate β-catenin/TCF-regulated target genes have been proposed (summarized at http://www.stanford.edu/~rnusse/pathways/targets.html). However, the role of the vast majority of candidate β-catenin/TCF target genes in the pathogenesis of OEAs is poorly understood. As a strategy for identifying novel WNT pathway target genes that have a functional role in ovarian cancer development and progression, we have used high-density oligonucleotide microarrays to compare gene expression profiles of OEAs with and without dysregulated β-catenin (Schwartz et al., 2003). Using this approach, two members of the fibroblast growth factor family, FGF9 and FGF20, both of which showed highly up-regulated expression in the WNT pathway-defective tumors, were shown to be downstream targets of β-catenin/TCF with important roles in OEA pathogenesis (Chamorro et al., 2005; Hendrix et al., 2006).
Another gene highlighted by our array-based approaches was MSX2, a member of the homeobox protein family that is also known as HOX-8. MSX2 is a transcriptional regulator with an essential role in embryonic development (Satokata et al., 2000). Previous studies have suggested MSX2 can function as a downstream effector and upstream regulator of WNT and BMP signaling (Hussein et al., 2003; Ramos and Robert, 2005; Shao et al., 2005; Song et al., 2009; Willert et al., 2002), though in-depth insights into the means by which MSX2 could function in these varied roles were not offered. It has also been suggested that MSX2 may contribute to the malignant phenotype of breast and pancreatic carcinomas, perhaps through induction of an epithelial-mesenchymal transition (EMT) (di Bari et al., 2009; Satoh et al., 2008). The role of MSX2 in ovarian cancer pathogenesis is as yet unexplored. In the current study, we present data indicating that MSX2 is a direct β-catenin/TCF target gene with a key role in the molecular pathogenesis of a substantial fraction of OEAs.
RESULTS
MSX2 transcripts and protein are highly expressed in primary OEAs with WNT signaling pathway defects
As a means of identifying novel candidate β-catenin/TCF-regulated target genes with roles in OEA pathogenesis, we used high-density oligonucleotide microarrays to compare gene expression patterns in primary human OEAs with and without dysregulated β-catenin. Tumors were classified as having “dysregulated” β-catenin based on immunohistochemical stains showing nuclear β-catenin expression and/or identification of biallelic inactivating APC mutations (one case) or activating CTNNB1 mutations (all other cases). Tumors were classified as having “intact” β-catenin regulation if immunohistochemical stains showed membranous expression of β-catenin protein and DNA sequencing failed to identify CTNNB1 mutations (Schwartz et al., 2003; Wu et al., 2007). MSX2 transcripts were greater than 9-fold higher in the OEAs with dysregulated β-catenin relative to tumors with intact β-catenin regulation (Figure 1A). MSX1 transcripts were also approximately 3-fold higher, on average, in OEAs with dysregulated β-catenin, but several tumors with intact β-catenin regulation also showed high levels of MSX1 expression (Supplemental Figure S1A). To evaluate MSX2 protein expression in OEAs with and without β-catenin dysregulation, we performed immunohistochemical staining for MSX2 using formalin-fixed, paraffin-embedded tissue sections from 30 of the 38 OEA samples used for the earlier expression profiling studies. β-catenin immunostaining and mutational analyses for CTNNB1, PIK3CA, and APC have been reported previously for these tumors (Wu et al., 2007; Wu et al., 2001; Zhai et al., 2002). Cases with mutations predicted to dysregulate β-catenin showed nuclear accumulation of β-catenin protein, and of these, the vast majority showed moderate or strong MSX2 expression (Figure 1B and Table 1). Two OEAs with presumptive activation of canonical WNT signaling on the basis of nuclear accumulation of β-catenin protein, but lacking CTNNB1 exon 3 mutations, also showed strong (++) MSX2 expression. Only one of thirteen OEAs with intact β-catenin regulation and exhibiting membranous β-catenin expression showed similarly strong MSX2 expression. Hence, in primary OEAs, strong MSX2 expression is highly correlated with dysregulated β-catenin (p=.0024, two-tailed Fisher's exact test). We also evaluated MSX2 expression in several ovarian carcinoma-derived cell lines by Western blot (Figure 1C). Highest expression of MSX2 was identified in the OEA-derived cell line TOV112D, which has an activating mutation in CTNNB1 (Wu et al., 2001). MSX2 expression was also readily observed in the SKOV3 and TOV21G cell lines, and low but detectable levels of MSX2 were observed in MDAH2774, OVCA429, and OVCA433. MSX2 expression was not detectable in the five remaining lines (Figure 1C). In addition to strong correlation of dysregulated β-catenin with elevated MSX2 expression, we noted that primary OEAs (Table 1) and ovarian carcinoma-derived cell lines known to have PIK3CA mutations (SKOV3, TOV21G, and OVCA429) (Kuo et al., 2009; Whyte and Holbeck, 2006) also showed elevated expression of MSX2. However, several of these samples also harbored CTNNB1 mutations, and the collection of PIK3CA mutant tumors and cell lines was too limited to determine if PI3K activation was independently correlated with increased MSX2 expression.
Figure 1. MSX2 transcripts and protein are highly expressed in human and murine OEAs with dysregulated β-catenin.
A) Relative expression of MSX2 in OEAs with dysregulated β-catenin (right) was compared with OEAs with intact β-catenin regulation (left), based on Affymetrix U133A microarray analysis (Probe Set1: 205555_s_at, Probe Set2: 210319_x_at); B) Representative immunohistochemical staining of MSX2 and β-catenin expression in OEAs with (OE-2T) or without (OE-20T) dysregulated β-catenin; C) Western blot analysis of MSX2 and β-actin protein expression in ovarian cancer-derived cell lines. D) Relative expression of Msx2 transcripts in murine APC−/PTEN− and KRASmut/PTEN− ovarian tumors and control ovaries based on Affymetrix Mouse430_2 microarray analysis.
Table 1.
Expression of β-catenin and MSX2 in ovarian endometrioid adenocarcinoma
Case | Tumor ID | β-catenin expression |
MSX2 expression |
Documented or presumptive Wnt pathway defect |
PIK3CA mutation |
||
---|---|---|---|---|---|---|---|
N | C | M | N | ||||
1 | OE-2T | +++ | ++ | β-cat Mut | ND | ||
2 | OE-13T | + | + | ++ | β-cat Mut | ND | |
3 | OE-17T | + | + | − | β-cat Mut | ND | |
4 | OE-18T | ++ | ++ | ++ | β-cat Mut | ND | |
5 | OE-19T | ++ | − | β-cat Mut | ND | ||
6 | OE-21T | + | + | ++ | β-cat Mut | Mut | |
7 | OE-36T | ++ | ++ | + | β-cat Mut | Mut | |
8 | OE-37T | + | + | ++ | β-cat Mut | ND | |
9 | OE-44T | + | + | +++ | + | β-cat Mut | ND |
10 | OE-47T | ++ | + | β-cat Mut | ND | ||
11 | OE-48T | + | + | + | ++ | β-cat Mut | ND |
12 | OE-53T | + | − | β-cat Mut | ND | ||
13 | OE-55T | + | ++ | β-cat Mut | Mut | ||
14 | OE-71T | ++ | ++ | β-cat Mut | Mut | ||
15 | OE-32T | ++ | + | ++ | APC Mut | ND | |
16 | OE-102T | + | ++ | β-cat WT, NP | ND | ||
17 | OE-103T | ++ | + | ++ | β-cat WT, NP | ND | |
| |||||||
18 | OE-11T | + | + | ND | ND | ||
19 | OE-12T | + | ± | ND | ND | ||
20 | OE-16T | + | ± | ND | ND | ||
21 | OE-20T | ++ | ± | ND | ND | ||
22 | OE-23T | ++ | ++ | + | ND | ND | |
23 | OE-24T | ++ | ± | ND | ND | ||
24 | OE-25T | + | − | ND | ND | ||
25 | OE-31T | + | + | ++ | ND | Mut | |
26 | OE-39T | ++ | +++ | − | ND | ND | |
27 | OE-64T | + | − | ND | ND | ||
28 | OE-69T | + | − | ND | ND | ||
29 | OE-76T | + | + | + | ND | ND | |
30 | OE-79T | + | + | + | ND | ND |
N: nuclear; C: cytoplasmic; M: membranous; ND: not detected; NP: APC and AXIN2 mutation analysis not performed
MSX2 expression is markedly up-regulated in APC−/PTEN− but not KRASmut/PTEN− murine ovarian endometrioid carcinomas
We have previously reported that conditional activation of the WNT/β-catenin and PI3K/Pten/Akt signaling pathways in the murine ovarian surface epithelium induces tumors (APC−/PTEN−) with similar biological behavior and morphology to human OEAs (Wu et al., 2007). Using a similar approach, Dinulescu et al. reported that conditional activation of a mutant Kras allele and biallelic inactivation of Pten promoted OEA (KRASmut/PTEN−) development in mice (Dinulescu et al., 2005). We used Affymetrix Mouse430_2 oligonucleotide microarrays to compare gene expression in APC−/PTEN− (n=7) and KRASmut/PTEN− (n=6) murine tumors, as well as in normal ovaries from each model (n=4 each). Data from the APC−/PTEN− tumors has been previously published (GSE5987) (Wu et al., 2007), while the full data set from the KRASmut/PTEN− tumors will be analyzed and published separately. We consistently found marked up-regulation of Msx2 transcripts in the APC−/PTEN− murine tumors compared to whole normal ovary samples or the KRASmut/PTEN− tumors (Figure 1D). Hence, data from these two murine models of OEA provide further support for the notion that Msx2 is a transcriptional target of β-catenin/TCF signaling and not simply up-regulated in ovarian cancer independent of the particular gene mutations which promote OEA development.
MSX2 is regulated by WNT/β-catenin signaling in ovarian epithelial cells
In an effort to establish that MSX2 expression is regulated by WNT-dependent signaling, we evaluated whether treatment of human ovarian epithelial cells with a prototypical WNT ligand that activates the canonical WNT signaling pathway could up-regulate MSX2 expression. IOSE80, an immortalized but non-tumorigenic ovarian epithelial cell line, was treated with 50 ng/ml of recombinant mouse WNT3a for 8 hr, and significantly increased MSX2 transcripts were observed after WNT3a treatment (Figure 2A). As expected, expression of several well-established β-catenin/TCF target genes (AXIN2, LGR5 and BMP4) also increased following WNT3a treatment of IOSE80 cells (Supplemental Figure S2A). GSK3β plays a central role in the destruction complex regulating the free pool of β-catenin in the absence of WNT ligand stimulation. Hence, we next assessed MSX2 expression in MDAH2774 cells following treatment with the GSK3β inhibitor SB216763 (Bommer et al., 2010; Coghlan et al., 2000). These cells were chosen because they express low, but detectable levels of MSX2 at baseline. A time-dependent increase of MSX2 transcripts was observed following SB216763 treatment (Figure 2B). To further investigate whether MSX2 is a downstream transcriptional target of β-catenin/TCF, we evaluated MSX2 expression in TOV112D cells stably expressing a dominant-negative version of TCF4 (dnTCF4), which is known to diminish the transcriptional activity of TCF-responsive promoters (Kolligs et al., 2002; Leung et al., 2002). TOV112D cells express mutant (activated) β-catenin protein and have high endogenous levels of MSX2 transcripts and protein. Consistent with the results expected for a β-catenin/TCF-regulated gene, MSX2 transcript and protein levels were reduced in TOV112D cells expressing dnTCF4 compared with empty vector-transduced control cells (Figure 2C). dnTCF4 had a similar effect on expression of known β-catenin/TCF-regulated genes AXIN2, LGR5, and BMP4 in TOV112D cells (Supplemental Figure S2B). Notably, MSX1 transcripts were not increased by stimulation of IOSE80 cells with Wnt3a, and expression of dnTCF in TOV112D cells did not inhibit MSX1 expression (Supplemental Figures S1B and S1C). Hence, unlike MSX2, MSX1 does not appear to be a direct downstream target of activated Wnt signaling in our system
Figure 2. MSX2 expression is induced by WNT signaling pathway activation and is a direct β-catenin/TCF target.
A) IOSE80 (immortalized human ovarian surface epithelial) cells were treated with 50ng/ml of WNT3A or PBS for 8 h. MSX2 mRNA levels were determined by qRT-PCR and normalized to HPRT. Three independent experiments were performed (bars, mean ± SD); B) MDAH2774 cells were treated with SB216736 (GSK3β inhibitor) and harvested at the time points shown. MSX2 expression was measured by qRT-PCR and normalized to HPRT. Four independent experiments were performed (bars, mean ± SD, asterisks denote significance ranging from p =0.013 to 2.1E-05 based on Student's t test); C) MSX2 transcript and protein (inset) levels are reduced in TOV112D cells stably expressing dnTCF versus cells expressing empty vector; D) Chromatin immunoprecipitation (ChIP) assay of extracts from TOV112D cells immunoprecipitated by anti-TCF4 antibody. Inset shows schematic diagram of MSX2 locus with location of consensus TCF binding sites. Immunoprecipitated DNA was analyzed by quantitative PCR and normalized to the amount of input. Negative controls (two irrelevant [non TCF4] binding sites on the same chromosome as MSX2) and at the RPL30 locus (different chromosome) were used to demonstrate specificity.
To determine whether MSX2 is likely to be a direct, rather than indirect β-catenin/TCF target gene, we searched the MSX2 genomic locus and flanking DNA sequences for consensus TCF/LEF-binding elements (WWCAAWG, W=A/T). We found several putative binding sites at the MSX2 locus (Figure 2D). Chromatin immunoprecipitation (ChIP) assays were performed using TCF4 antibody to analyze chromatin isolated from TOV112D cells. A known TCF/LEF binding site in the promoter of AXIN2, a well-established direct β-catenin/TCF target gene, was used as a positive control for the anti-TCF4 ChIP assay, while irrelevant sites approximately 500 kb downstream of the MSX2 locus (Non-1 and Non-2) and at the RPL30 locus on a different chromosome than MSX2 served as negative controls (Figure 2D, inset). DNA isolated from the anti-TCF4 ChIP of TOV112D cells was 7- to 10-fold enriched for three of nine MSX2 sequences containing consensus TCF4 binding sites (Figure 2D). Studies by others have shown that clusters of TCF binding sites spanning several kilobases of DNA, rather than individual TCF sites, act as Wnt response elements (Parker et al., 2008). In keeping with this notion, Jho et al. have shown that several TCF/LEF consensus binding sites in the promoter and first introns of AXIN2 (a well established β-catenin/TCF target gene), are important for TCF-mediated activation of AXIN2 expression (Jho et al., 2002). Hence, it is not unexpected that TCF4 was found to bind more than one TCF consensus binding site at the MSX2 locus. Collectively, the gene expression and ChIP data suggest that MSX2 is indeed directly regulated by β-catenin/TCF in OEA cells.
MSX2 promotes neoplastic transformation of epithelial cells
Like other β-catenin/TCF target genes with well known or presumptive oncogenic activity in cancers with defective β-catenin/TCF signaling, such as c-MYC, cyclin D1, and FGFs 9 and 20, we wished to test whether MSX2 might have key roles in cancer-associated phenotypes. To assess the ability of MSX2 to promote neoplastic transformation, we performed focus formation assays using the immortalized rat epithelial cell line RK3E. Cancer-derived mutant forms of β-catenin, as well as several other known oncogenes (e.g., GLI1, mutant KRAS), readily generate foci of morphologically transformed cells when introduced into RK3E cells (Hendrix et al., 2006; Kolligs et al., 1999; Kolligs et al., 2002). Although transformation assays traditionally have been conducted in NIH3T3 fibroblasts, we consider RK3E cells more suitable for our experiments because they are epithelial rather than fibroblastic. Ectopic expression of MSX2 efficiently induced foci in RK3E cells within 2 weeks (Figure 3A). MSX2-induced focus formation was minimally inhibited in RK3E cells expressing the dnTCF4 mutant protein, whereas the ability of S33Y mutant β-catenin to induce foci in RK3E was potently inhibited by dnTCF4. Inhibitory effects of dnTCF4 on β-catenin/TCF-dependent transcription in RK3E cells was confirmed by showing the expected down-regulation of target genes AXIN2, LGR5, and CD44 (Supplemental Figure S2C). RK3E cells expressing β-catenin (S33Y) showed the expected upregulation of target genes AXIN2 and LGR5 (Supplemental Figure S2D). These findings are consistent with the data showing that MSX2 activation is downstream of β-catenin/TCF. As such, a downstream oncogenic target gene that is expressed under control of heterologous regulatory elements would not be expected to be inhibited by dnTCF4. To further investigate the role of MSX2 in neoplastic transformation, individual foci of MSX2-transformed RK3E cells were picked, clonally expanded, and plated in soft agar. A subset (20%) of clonal RK3E lines expressing MSX2 (n = 10) exhibited anchorage-independent growth in soft agar (Figure 3B).
Figure 3. MSX2 promotes neoplastic transformation of RK3E cells.
A) RK3E/neo (empty vector) cells (left panel top) or RK3E cells expressing dominant negative TCF (left panel bottom) were transduced with replication defective retroviruses containing vector alone (pPGS), MSX2, or β-cat/S33Y. Two to three weeks after infection with the indicated retroviruses, plates were fixed, stained, and photographed. Quantified data shown (right panel) are from three independent experiments (bars, mean ± SD); B) RK3E cells with vector alone (pPGS), two clones of RK3E cells expressing MSX2 (MSX2-C1 and MSX2-C2) derived from a focus formation assay, and RK3E cells stably over-expressing β-cat/S33Y were assessed for their ability to grow in soft agar. After 4 weeks, dishes were stained with methylene blue and colonies were photographed (upper panel) and counted (bottom left panel) using the Image J program (http://rsb.info.nib.gov/ij/). Data shown represent the mean number of colonies from two independent experiments, each in triplicate; bars, mean ± SD. Western blot analysis of RK3E cells after transduction with MSX2, activated β-catenin (S33Y) or empty vector (bottom right panel). β-actin expression was evaluated as a loading control.
shRNA-mediated knockdown of MSX2 inhibits proliferation of ovarian carcinoma cells
To further establish the function of MSX2 in ovarian cancer pathogenesis, we utilized shRNA approaches to inhibit MSX2 in two human ovarian cancer cell lines with elevated endogenous MSX2 expression. TOV112D and TOV21G cells were each transduced with three independent shRNA constructs, targeting different regions of the MSX2 transcript. Stable polyclonal cell populations expressing the shRNAs were obtained, and substantial knockdown of MSX2 was achieved in each of the two cell lines using at least two independent shRNAs (Figure 4A). MSX2 knockdown significantly inhibited the in vitro growth of TOV112D and TOV21G cells (Figure 4B). Colony formation in soft agar was also significantly reduced in TOV112D cells expressing MSX2 shRNAs relative to the shRNA control (Figure 4C). In order to test whether MSX2 knockdown affected growth of ovarian carcinoma cells in vivo, polyclonal TOV112D cell populations stably expressing the two independent shRNAs were also tested for tumorigenicity in nude mice. Subcutaneous tumors formed at 5 of 7 sites injected with control-transduced TOV112D cells (TOV112D/shc); 3 of 6 sites injected with cells expressing sh1 (TOV112D/sh1); and 4 of 7 sites injected with cells expressing sh2 (TOV112D/sh2). Average tumor volume at three weeks was significantly smaller in tumors arising from TOV112D cells expressing the MSX2 shRNAs (Figure 4D).
Figure 4. shRNA-mediated knockdown of endogenous MSX2 inhibits cell proliferation and tumor growth.
A) Western blot analysis of lysates from TOV112D and TOV21G cells after stable transduction with virus expressing two different shRNAs targeting MSX2 and a non-silencing control shRNA (shc). MSX2 protein levels were quantified with the Image J program and normalized to β-actin. The ratios were normalized to control shRNA. B) Polyclonal TOV112D and TOV21G cells stably expressing either of two MSX2 shRNAs or control shRNA were plated at 1×104/cm2 in 6-well plates and grown in complete medium. Cell numbers were determined by counting with a hemocytometer on the indicated days. Three independent experiments were performed, each in triplicate (data shown represent mean ± SEM; C) TOV112D cells stably expressing either of two MSX2 shRNAs or control shRNA were grown in soft agar for 4 weeks. Number of colonies was counted by Image J after staining with methylene blue (left panel). Data shown (right panel) represent mean number of colonies from two independent experiments, each in triplicate; bars, mean ± SD. D) Xenograft tumor growth in nude mice after s.c injection of TOV112D cells expressing MSX2 shRNAs or control shRNA. Three or four mice per group were injected on both flanks. Data shown represent mean ± SEM.
MSX2 promotes invasion of ovarian epithelial cells
Prior work suggested that deregulation of MSX2 may play a significant role in promoting invasion in mouse mammary epithelial cell line-derived tumors (di Bari et al., 2009). We investigated whether MSX2 could promote invasion of ovarian epithelial cells using Matrigel-coated transwells. The invasive properties of polyclonal IOSE80 cell populations expressing MSX2 were compared to a control IOSE80 cell population transfected with vector alone. Ectopic MSX2 expression led to a significant (p=0.001) increase in the number of invasive cells (Figure 5A). To confirm effects of MSX2 on ovarian epithelial cell invasiveness, PEO4 cells (which lack endogenous expression of MSX2) expressing vector alone (neo) or exogenous MSX2 were examined for invasion into the chick chorioallantoic membrane (CAM). This in vivo assay closely mimics invasion of neoplastic epithelial cells through the basement membrane into underlying stroma. Ectopic expression of MSX2 in PEO4 cells enhanced tumor cell invasiveness (Figure 5B).
Figure 5. MSX2 stimulates invasion of IOSE 80 cells and is associated with EMT-like changes.
A) Representative photographs of control (neo) and MSX2 expressing IOSE80 cells that invaded through Matrigel in transwell invasion assays (upper panel). The number of cells penetrating the membrane was counted in five fields from two independent experiments each in triplicate (lower panel). Columns and error bars represent mean ± SD (p=.001 based on Mann-Whitney U-test). B) Representative photomicrographs of control (neo) and MSX2 expressing PEO4 ovarian carcinoma cells on the chick CAM are shown. CAM tissues were evaluated by light microscopy (H&E staining, left panels), or fluorescence microscopy (center and right panels: immunostaining for chick laminin, red; DAPI, blue; and GFP, green). Dashed lines in center panels indicate the CAM surface and dashed rectangles indicate portion shown at higher magnification in right panels. C) Phase-contrast photographs of cultured IOSE, PEO4, and OVCA433 cells showing fibroblast-like morphological change in cells expressing MSX2 compared to control (neo). D) Western blot analysis of N-cadherin, E-cadherin, Vimentin, MSX2 and β-actin (loading control) in PEO4, IOSE80 and OVCA433 cells expressing vector (neo) or MSX2.
We noted that ectopic expression of MSX2 in IOSE80 cells appeared to induce a fibroblast-like morphological change, with features akin to epithelial-mesenchymal transition. Similar, but less pronounced effects were noted in PEO4 and OVCA433 cells (Figure 5C). In IOSE80 cells, the morphological changes were accompanied by a modest increase in expression of vimentin, a marker of mesenchymal differentiation, compared to cells transduced with vector alone (Figure 5D). N-cadherin and E-cadherin levels following over-expression of MSX2 in IOSE-80 cells were unaffected. Ectopic expression of MSX2 in PEO4 cells resulted in increased N-cadherin and vimentin expression, and a slight reduction in E-cadherin expression (Figure 5D). Despite the data from IOSE80 and PEO4 cells suggesting that MSX2 might alter some markers of EMT, the findings could not be readily generalized to all cell lines studied. Overexpression of MSX2 in OVCA433, another ovarian carcinoma cell line with low levels of endogenous MSX2, led to reduced N-cadherin and vimentin expression and no apparent change in E-cadherin. Activation of Wnt signaling via WNT3a stimulation or expression of mutant β-catenin in IOSE80, PEO4, or RK3E cells (Supplemental Figure S3A) as well as knockdown of endogenous MSX2 in TOV112D cells (Supplemental Figure S3B) likewise resulted in variable effects on EMT markers. Importantly, levels of exogenously expressed MSX2 in IOSE80, PEO4, and OVCA433 were comparable to physiological levels of endogenous MSX2 in TOV112D (Supplemental Figure S3C).
In order to more thoroughly address the relationship of dysregulated β-catenin with expression of MSX2 and markers of EMT, we used immunohistochemical staining to evaluate expression of MSX2, E-cadherin, N-cadherin, Vimentin, Snail and Slug in primary OEAs with and without dysregulated β-catenin (n=6 per group). The additional data are summarized in Supplemental Table S2 and representative immunostaining is shown in Supplemental Figure S4. We found that five of six OEAs with dysregulated β-catenin and elevated MSX2 expression expressed vimentin compared to none of the OEAs with intact β-catenin regulation. Although the increased vimentin staining in the β-catenin mutant tumors is suggestive of EMT induced by MSX2, other EMT markers failed to show consistent patterns of differential expression between the two tumor groups. Moreover, tumors with wild type β-catenin but increased MSX2 expression (e.g., OE 79T) did not show increased vimentin expression. Hence, MSX2 expression is not consistently correlated with EMT markers in primary tumor specimens. Collectively, our data suggest that MSX2's role in promoting EMT appears to be complex and context-dependent.
DISCUSSION
Homeobox (HOX) genes encode transcription factors which function as important regulators of embryonic morphogenesis and differentiation. Specific HOX genes also play roles in cancer pathogenesis, and their altered expression has been described in several different tumor types, including lung, prostate, breast, colorectal, colon and ovarian cancers (Shah and Sukumar, 2010). Some HOX genes regulate Müllerian duct differentiation and specify morphological identity within the female reproductive tract; these HOX genes may also play a key role in ovarian carcinogenesis as a result of their aberrant expression (Cheng et al., 2005). Vertebrate muscle segment homeobox (Msx) genes are one of the most highly conserved families of homeobox genes involved in developmental processes, and Msx genes may also be involved in tumorigenesis (Bendall and Abate-Shen, 2000; Satoh et al., 2004; Satokata et al., 2000). An earlier study showed MSX2 expression is elevated in a variety of carcinoma cell lines (Suzuki et al., 1993), and more recent studies have reported MSX2 might enhance pancreatic cancer cell aggressiveness through the induction of an epithelial-to-mesenchymal transition (EMT) (Satoh et al., 2008).
Interestingly, several studies have suggested that MSX2 can function as both a downstream effector and upstream regulator of WNT and BMP signaling (Hussein et al., 2003; Ramos and Robert, 2005; Shao et al., 2005; Song et al., 2009; Willert et al., 2002). In addition, MSX2 expression has been linked to activated MAPK signaling because antisense MSX2 cDNA was found to interfere with the transforming activities of KRAS (Takahashi et al., 1996). In the current study, we found MSX2 expression is significantly correlated with dysregulation of WNT/β-catenin/TCF signaling in OEAs. We also found that Msx2 mRNA levels are markedly up-regulated in APC−/PTEN− murine ovarian carcinomas, but not in KRASmut/PTEN− murine carcinomas or normal ovary samples. Hence, in contrast to the findings of Takahashi et al., our results indicate that MSX2 is a downstream target of activated β-catenin/TCF, but not KRAS/MAPK signaling, in ovarian epithelial cells. This argument is further supported by our in vitro studies showing that WNT signaling pathway activation, either via WNT3a ligand or GSK3β inhibitor treatment, enhances MSX2 transcript levels. Overexpression of a dominant-negative mutant of TCF4 inhibits MSX2 expression. Significantly, chromatin immunoprecipitation assays revealed that TCF4 binds several LEF1/TCF binding sites at the endogenous MSX2 locus. Collectively, our data provide strong support for the proposal that MSX2 is a direct β-catenin/TCF target gene and are consistent with previous studies suggesting that MSX2 is a direct downstream target of activated WNT/β-catenin signaling in human embryonic carcinoma cells and in mouse embryo development during lip formation or fusion (Song et al., 2009; Willert et al., 2002).
MSX2 has been reported to play a role in regulating the differentiation and/or proliferation of epithelial cells as well as osteogenic cells (Dodig et al., 1999; Jiang et al., 1999). In addition, MSX2 stimulates pancreatic cancer cell proliferation in vitro and promotes branching morphogenesis of mouse mammary ducts (Satoh et al., 2004; Satoh et al., 2008). We showed that MSX2 promotes neoplastic transformation of E1A-immortalized epithelial cells (RK3E). Diminished MSX2 expression by shRNA-mediated knockdown in ovarian carcinoma cell lines inhibited cell proliferation and anchorage independent growth in vitro, and tumor growth in vivo.
Previous studies have also addressed effects of MSX2 on cellular properties that are of likely significance in cancer progression, such as motility and invasiveness, and it has been suggested that MSX2 may have a role in induction of EMT (di Bari et al., 2009; Satoh et al., 2008). In our work, immortalized human ovarian surface epithelial cells ectopically expressing MSX2 indeed showed increased invasiveness and morphological changes reminiscent of EMT, accompanied by increased vimentin expression compared to cells transduced with vector alone. PEO4 cells transduced with MSX2 showed reduced expression of E-cadherin, a marker of epithelial differentiation, with concomitant increased expression of the mesenchymal markers N-cadherin and vimentin. Notably, similar findings were not observed in OVCA433 and we did not observe up-regulation of Cripto-1 or Twist in IOSE80 or PEO4 over-expressing MSX2 (data not shown), which has been reported in other systems (di Bari et al., 2009; Satoh et al., 2008). Clearly, much remains to be determined regarding the role of MSX2 in promoting EMT, as MSX2's effects appear to be context-dependent.
In conclusion, our findings provide evidence that MSX2 is a direct downstream target of WNT signaling in human OEAs and a mouse model of ovarian endometrioid carcinoma based on conditional WNT signaling activation. We have shown that MSX2 has robust oncogenic properties in epithelial cells, including the ability to promote neoplastic transformation, facilitate anchorage-independent cell growth, and enhance invasiveness. Taken together, the findings support a key role for MSX2 in the development and progression of OEAs arising in the context of deregulated WNT signaling.
MATERIALS AND METHODS
Cell lines and drug treatment
RK3E (E1A-immortalized rat kidney epithelial cells) and ovarian carcinoma cell lines SKOV3, ES2, TOV21G, MDAH2774, and TOV112D were obtained from the American Type Culture Collection (ATCC, Rockville, MD); OVCAR3, OVCAR5, and PEO4 were a gift from T. Hamilton (Fox Chase Cancer Center, Philadelphia, PA); OVCA429, OVCA432, and OVCA433 were a gift from D. Fishman (Northwestern University, Chicago, IL). The cells were grown in Dulbecco's modified Eagle's medium with 10% fetal bovine serum (FBS, Invitrogen, Carlsbad, CA) and penicillin/streptomycin. IOSE-80 cells (human ovarian surface epithelial cells immortalized with SV40 LgTAg), were a gift from N. Auersperg (University of British Columbia, Vancouver, Canada). Polyclonal TOV112D and RK3E cell lines expressing dominant-negative TCF-4 (dnTCF4) were obtained after retroviral transduction with supernatants from amphotrophic Phoenix cells transfected with pPGS-dnTCF4 as previously described (Kolligs et al., 2002; Leung et al., 2002; Schwartz et al., 2003). Stable polyclonal lines were maintained in G418 at a concentration of 0.4-0.5 mg/ml. To assess effects of activated WNT/β-catenin signaling on MSX2 gene expression, IOSE80 cells were treated with 50 ng/ml of mouse WNT3a (R&D Systems, Minneapolis, MN) or PBS for 8 h, then cells were harvested for RNA isolation. GSK3-β activity in the OEA-derived cell line MDAH2774 was inhibited by treating 50% confluent cells with a final concentration of 10 nM SB216763 (Sigma, St Louis, MO) while control cells were treated with DMSO. RNA was isolated at the indicated time points after initiation of treatment.
Immunohistochemical analysis of MSX2 and EMT marker expression in primary ovarian carcinomas
Formalin-fixed paraffin-embedded tissue sections from 30 primary human OEAs that were included in our prior gene expression profiling studies (Schwartz et al., 2002; Wu et al., 2007) were evaluated for MSX2 expression using immunohistochemistry. Tissues were obtained from the Cooperative Human Tissue Network/Gynecologic Oncology Group Tissue Bank and analyzed with approval from the Institutional Review Board of the University of Michigan Medical School (IRBMED 2002-0430). Five-micron tissue sections were mounted on Probe-On slides (Fisher Scientific, Itasca, IL), de-paraffinized in xylene, and then rehydrated in distilled H2O through graded alcohols. Antigen retrieval was performed by microwaving the slides in citrate buffer (pH 6.0, Biogenex, San Ramon, CA) for 10 min. Endogenous peroxidase activity was quenched by incubation with 3% hydrogen peroxide in methanol, then sections were blocked with 1.5% normal rabbit serum for 30 min. Sections were incubated with goat polyclonal anti-MSX2 (1:50 dilution, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) overnight at 4°C. Slides were washed in PBS, then incubated with a biotinylated rabbit anti-goat secondary antibody for 30 min at room temperature. Antigen-antibody complexes were detected with the avidin-biotin peroxidase method using Vector® DAB as a chromogenic substrate (Vectastain ABC kit, Vector Laboratories). Immunostained sections were lightly counter-stained with hematoxylin and then examined by light microscopy. Immunostaining was scored on a four tiered scale for intensity (−, absent; ±, weak; +, moderate; ++, strong) in the cell nucleus. Similar methods were used to evaluate E-cadherin, N-cadherin, Vimentin, Snail and Slug expression in representative primary OEAs with and without dysregulated β-catenin (n=6 per group). The following antibodies were employed for these studies: mouse monoclonal anti-Vimentin (1:1000, Dako, Carpinteria, CA), anti-N-cadherin (1:500, Upstate), anti-E-cadherin (1:200, Invitrogen), rabbit monoclonal anti-Snail or anti-Slug (1:50, Cell signaling). Immunostaining was scored on a four tiered scale for intensity (−, absent; ±, weak; +, moderate; ++, strong) at the cell membrane for E-cadherin and N-cadherin; in the cytoplasm for Vimentin, and in the nucleus for Snail and Slug.
MSX2 expression vector construction and transfection
Full length MSX2 cDNA (spanning nucleotides 89 to 892, GenBank accession #: NM-002449.4) was generated by RT-PCR using total RNA from TOV112D cells. PCR primers were designed to include a Myc-epitope tag at the gene product's C-terminus. PCR products were sub-cloned into the retroviral vector pPGS-CMV-CITE-neo and the cDNA sequence of individual clones verified by DNA sequencing. Selected cell lines were transduced with retroviral supernatant from amphotrophic Phoenix cells transfected with vector alone or vector with MSX2 cDNA. Stable polyclonal lines were generated by selection in G418 at a concentration of 0.8-1 mg/ml. After 1 week, the G418 concentration was reduced to 0.4-0.5 mg/ml, and expression of MSX2 protein was confirmed by Western blot analysis.
shRNA-mediated knockdown of MSX2 expression
To knock down MSX2 expression in the ovarian carcinoma cell lines TOV112D and TOV21G, three shRNAs targeting MSX2 were expressed using the lentiviral shRNAmir system (Human pGIPZ collection, Open Biosystems, Huntsville, AL) or pSIREN-RetroQ (Clontech, Mountain view CA). The following sequences were introduced into the pGIPZ vector: human MSX2/sh1 (sense 5′-CGGCGCTCATGTCCGACAAGAA-3′, anti-sense 5′-TTCTTGTCGGACATGAGCGCCT-3′); human MSX2/sh2 (sense 5′-AGCGCAAGTTCCGTCAGAAACA , antisense 5′-TGTTTCTGACGGAACTTGCGCC-3′); human MSX2/sh3 (sense 5′-CGCATCCTACCCGTTCCATAGA-3′, antisense 5′-TCTATGGAACGGGTAGGATGCT-3′). pGIPZ MSX2/sh2 and MSX2/sh3 sequences were also used for ligation into the pSIREN-RetroQ vector. Stable polyclonal TOV112D and TOV21G cell lines expressing individual MSX2-specific (TOV112D/sh1, TOV112D/sh2) (TOV21G/sh2, TOV21G/sh3) or control shRNA (non-silencing shRNAmir [Open Biosystems]) were generated by transduction with lentiviral or retroviral supernatant, and grown in the presence of puromycin at a concentration of 1-1.25μg/ml.
RNA isolation and quantitative RT-PCR
Total RNA from each cell line was extracted using standard procedures (Trizol reagent, Invitrogen). cDNA was synthesized from 4 μg total RNA using a high capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA ). qRT-PCR was performed with an ABI Prism 7700 Sequence Analyzer using SYBR green fluorescent protocol (Applied Biosystems). Each reaction was run in duplicate. qRT-PCR reactions for target and internal control genes were performed in separate tubes. The comparative threshold cycle (CT) method was used for the calculation of amplification-fold as specified by the manufacturer. Primer sequences for all target genes and HPRT1 (internal control) were shown in Supplemental Table S1. PCR was performed in 25μl reactions containing 1 × SYBR® Green PCR mix (Applied Biosystems), 200nM of each primer, and 0.5 μl of first strand cDNA. The PCR conditions were 50°C for 2 min, 95°C for 10 min, followed by 40 cycles of 95°C 15 seconds, 60°C 1 min. Each reaction was set up in duplicate and three separate measurements were carried out.
Western blot
Parental and stably transduced cells were lysed in cold RIPA lysis buffer containing proteinase inhibitors (complete proteinase inhibitors, Roche Applied Science, Indianapolis, IN). Whole-cell lysates were analyzed by Western blotting with mouse monoclonal anti-MSX2 antibody (Assay Designs) at 1:1000; anti-E-cadherin antibody (BD Biosciences, San Jose, CA) at 1:5000; anti-Vimentin antibody (Dako, Carpinteria, CA ) at 1:10000; rabbit monoclonal anti-N-cadherin antibody (Upstate, Lake Placid, NY) at 1:5000. Expression of β-actin was used as a loading control and was detected with anti-actin monoclonal antibody (Sigma).
Chromatin immunoprecipitation
Chromatin immunoprecipitation was performed using the Simple Chip™ Enzymatic Chromatin IP kit with magnetic beads according to the manufacturer's instructions (Cell Signaling, Beverly, MA). Chromatin was immunoprecipitated with rabbit monoclonal anti-TCF4 antibody (Cell Signaling) or IgG control antibody. Primers for the quantitative PCR are listed in Supplemental Table S1.
Growth kinetics, focus formation, soft agar and tumorigenicity assays
For cell proliferation assays, cells were seeded at 1×104/cm2 in 6-well plates in triplicate and were grown in complete medium for the duration of the assay. Medium was changed every 3 days. Cells were trypsinized and counted using a hemocytometer at the indicated time points. Three independent experiments were performed and error was calculated as the SE. Focus formation and soft agar assays were performed using standard methods as described previously with the only modification that cells were plated in 12-well plates for the soft agar assay (Hendrix et al., 2006) (Zhai et al., 2007). Mouse xenograft tumor growth was measured twice weekly after subcutaneous injection of 5 ×106 cells in both flanks of Nu/Nu mice (Charles River, Wilmington, MA). Tumor volume was estimated using the formula: (length × height × width)/2.
Matrigel transmembrane invasion assays
Transwell invasion chambers (BD Biosciences, Bedford, MA) with 8 μm pore inserts were used to perform in vitro invasion assays. Five × 104 cells were plated in the upper chamber of 24-well invasion chambers and allowed to invade across the Matrigel-coated membrane for 48 h. Complete 10% FBS medium was added to the bottom well as a chemoattractant. After non-invading cells were removed from the top of each membrane with wet cotton swabs, invading cells attached to the bottom of the membrane were fixed and stained with DAPI. The number of cells that penetrated the membrane was determined by counting the number of cells in five randomly selected 20X fields and calculating the mean. Experiments were performed in triplicate, and the assays were repeated at least once to verify results.
Chick embryo CAM invasion assays
Fertilized eggs were purchased from local poultry farms (Michigan State University, Lansing, MI or Bilbie Aviaries, Ann Arbor, MI). Embryos were incubated for 11 days at 37-38°C with 60% humidity. Preparation of CAMs has been previously described (Zhai et al., 2007) Briefly, to grow tumor masses, 2.5 × 106 cells in 30μl of Hank's Balanced Salt Solution were applied on the CAM. Cells were first infected with lentivirus expressing green fluorescent protein (GFP). After 3 days, the tumor masses and underlying CAM were excised and fixed with 4% paraformaldehyde in PBS overnight at 4°C. Frozen sections were cut from CAM tissues after immersion in 30% sucrose. The frozen sections were stained with mouse anti-laminin antibody (3H11, diluted 1:10, Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA) for 2 hours, the incubated with Alexa Fluor 594 conjugated goat anti-mouse IgG secondary antibody (1:500; Invitrogen, Carlsbad, CA) for 60 minutes at room temperature. Sections were mounted with Vectashield mounting media with 4′-6′-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA), monitored under a fluorescent microscrope and photographed with a high-resolution digital camera. Serial sections were also prepared and stained with H&E for light microscopic examination.
Statistical analysis
Statistical analyses were performed using SPSS 10.0 (SPSS Inc. Chicago, IL). For transwell invasion assays, statistical differences in cell invasion between groups were determined by the Mann-Whitney U test. For the qRT-PCR studies, statistical differences of expression between groups were calculated with the Student's t-test. Values were expressed as mean ± SD (standard deviation) or mean ± SEM. In all cases, values of P < .05 were considered significant.
Supplementary Material
ACKNOWLEDGEMENTS
The authors wish to thank Rork Kuick for assistance analyzing the microarray data from human and mouse tissue samples. This work was supported by grants from the National Cancer Institute (RO1 CA94172 and RO1 CA85463).
Footnotes
CONFLICT OF INTEREST
None of the authors have any competing financial interests in relation to the work described in this manuscript.
SUPPLEMENTARY INFORMATION
Supplementary information will be made available at Oncogene's website.
References
- Bendall AJ, Abate-Shen C. Roles for Msx and Dlx homeoproteins in vertebrate development. Gene. 2000;247:17–31. doi: 10.1016/s0378-1119(00)00081-0. [DOI] [PubMed] [Google Scholar]
- Bommer GT, Feng Y, Iura A, Giordano TJ, Kuick R, Kadikoy H, et al. IRS1 regulation by Wnt/beta-catenin signaling and varied contribution of IRS1 to the neoplastic phenotype. J Biol Chem. 2010;285:1928–38. doi: 10.1074/jbc.M109.060319. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Catasus L, Bussaglia E, Rodrguez I, Gallardo A, Pons C, Irving JA, et al. Molecular genetic alterations in endometrioid carcinomas of the ovary: similar frequency of beta-catenin abnormalities but lower rate of microsatellite instability and PTEN alterations than in uterine endometrioid carcinomas. Hum Pathol. 2004;35:1360–8. doi: 10.1016/j.humpath.2004.07.019. [DOI] [PubMed] [Google Scholar]
- Chamorro MN, Schwartz DR, Vonica A, Brivanlou AH, Cho KR, Varmus HE. FGF-20 and DKK1 are transcriptional targets of beta-catenin and FGF-20 is implicated in cancer and development. EMBO J. 2005;24:73–84. doi: 10.1038/sj.emboj.7600460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cheng W, Liu J, Yoshida H, Rosen D, Naora H. Lineage infidelity of epithelial ovarian cancers is controlled by HOX genes that specify regional identity in the reproductive tract. Nat Med. 2005;11:531–7. doi: 10.1038/nm1230. [DOI] [PubMed] [Google Scholar]
- Coghlan MP, Culbert AA, Cross DA, Corcoran SL, Yates JW, Pearce NJ, et al. Selective small molecule inhibitors of glycogen synthase kinase-3 modulate glycogen metabolism and gene transcription. Chem Biol. 2000;7:793–803. doi: 10.1016/s1074-5521(00)00025-9. [DOI] [PubMed] [Google Scholar]
- di Bari MG, Ginsburg E, Plant J, Strizzi L, Salomon DS, Vonderhaar BK. Msx2 induces epithelial-mesenchymal transition in mouse mammary epithelial cells through upregulation of Cripto-1. J Cell Physiol. 2009;219:659–66. doi: 10.1002/jcp.21712. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dinulescu DM, Ince TA, Quade BJ, Shafer SA, Crowley D, Jacks T. Role of K-ras and Pten in the development of mouse models of endometriosis and endometrioid ovarian cancer. Nat Med. 2005;11:63–70. doi: 10.1038/nm1173. [DOI] [PubMed] [Google Scholar]
- Dodig M, Tadic T, Kronenberg MS, Dacic S, Liu YH, Maxson R, et al. Ectopic Msx2 overexpression inhibits and Msx2 antisense stimulates calvarial osteoblast differentiation. Dev Biol. 1999;209:298–307. doi: 10.1006/dbio.1999.9258. [DOI] [PubMed] [Google Scholar]
- Gamallo C, Palacios J, Moreno G, Calvo de Mora J, Suarez A, Armas A. beta-catenin expression pattern in stage I and II ovarian carcinomas: relationship with beta-catenin gene Mutations, Clinicopathological features, and clinical outcome. Amer J Path. 1999;155:527–36. doi: 10.1016/s0002-9440(10)65148-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hendrix ND, Wu R, Kuick R, Schwartz DR, Fearon ER, Cho KR. Fibroblast growth factor 9 has oncogenic activity and is a downstream target of Wnt signaling in ovarian endometrioid adenocarcinomas. Cancer Res. 2006;66:1354–62. doi: 10.1158/0008-5472.CAN-05-3694. [DOI] [PubMed] [Google Scholar]
- Hussein SM, Duff EK, Sirard C. Smad4 and beta-catenin co-activators functionally interact with lymphoid-enhancing factor to regulate graded expression of Msx2. J Biol Chem. 2003;278:48805–14. doi: 10.1074/jbc.M305472200. [DOI] [PubMed] [Google Scholar]
- Jho EH, Zhang T, Domon C, Joo CK, Freund JN, Costantini F. Wnt/beta-catenin/Tcf signaling induces the transcription of Axin2, a negative regulator of the signaling pathway. Mol Cell Biol. 2002;22:1172–83. doi: 10.1128/MCB.22.4.1172-1183.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jiang TX, Liu YH, Widelitz RB, Kundu RK, Maxson RE, Chuong CM. Epidermal dysplasia and abnormal hair follicles in transgenic mice overexpressing homeobox gene MSX-2. J Invest Dermatol. 1999;113:230–7. doi: 10.1046/j.1523-1747.1999.00680.x. [DOI] [PubMed] [Google Scholar]
- Kolligs FT, Hu G, Dang CV, Fearon ER. Neoplastic transformation of RK3E by mutant beta-catenin requires deregulation of Tcf/Lef transcription but not activation of c-myc expression. Mol Cell Biol. 1999;19:5696–706. doi: 10.1128/mcb.19.8.5696. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kolligs FT, Nieman MT, Winer I, Hu G, Van Mater D, Feng Y, et al. ITF-2, a downstream target of the Wnt/TCF pathway, is activated in human cancers with beta-catenin defects and promotes neoplastic transformation. Cancer Cell. 2002;1:145–55. doi: 10.1016/s1535-6108(02)00035-1. [DOI] [PubMed] [Google Scholar]
- Kuo KT, Mao TL, Jones S, Veras E, Ayhan A, Wang TL, et al. Frequent Activating Mutations of PIK3CA in Ovarian Clear Cell Carcinoma. Am J Pathol. 2009;174:1597–601. doi: 10.2353/ajpath.2009.081000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Leung JY, Kolligs FT, Wu R, Zhai Y, Kuick R, Hanash S, et al. Activation of AXIN2 expression by beta-catenin-T cell factor. A feedback repressor pathway regulating Wnt signaling. J Biol Chem. 2002;277:21657–65. doi: 10.1074/jbc.M200139200. [DOI] [PubMed] [Google Scholar]
- Parker DS, Ni YY, Chang JL, Li J, Cadigan KM. Wingless signaling induces widespread chromatin remodeling of target loci. Mol Cell Biol. 2008;28:1815–28. doi: 10.1128/MCB.01230-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ramos C, Robert B. msh/Msx gene family in neural development. Trends Genet. 2005;21:624–32. doi: 10.1016/j.tig.2005.09.001. [DOI] [PubMed] [Google Scholar]
- Sagae S, Kobayashi K, Nishioka Y, Sugimura M, Ishioka S, Nagata M, et al. Mutational analysis of beta-catenin gene in Japanese ovarian carcinomas: frequent mutations in endometrioid carcinomas. Jpn J Cancer Res. 1999;90:510–5. doi: 10.1111/j.1349-7006.1999.tb00777.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Satoh K, Ginsburg E, Vonderhaar BK. Msx-1 and Msx-2 in mammary gland development. J Mammary Gland Biol Neoplasia. 2004;9:195–205. doi: 10.1023/B:JOMG.0000037162.84758.b5. [DOI] [PubMed] [Google Scholar]
- Satoh K, Hamada S, Kimura K, Kanno A, Hirota M, Umino J, et al. Up-regulation of MSX2 enhances the malignant phenotype and is associated with twist 1 expression in human pancreatic cancer cells. Am J Pathol. 2008;172:926–39. doi: 10.2353/ajpath.2008.070346. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Satokata I, Ma L, Ohshima H, Bei M, Woo I, Nishizawa K, et al. Msx2 deficiency in mice causes pleiotropic defects in bone growth and ectodermal organ formation. Nat Genet. 2000;24:391–5. doi: 10.1038/74231. [DOI] [PubMed] [Google Scholar]
- Schwartz DR, Kardia SL, Shedden KA, Kuick R, Michailidis G, Taylor JM, et al. Gene expression in ovarian cancer reflects both morphology and biological behavior, distinguishing clear cell from other poor-prognosis ovarian carcinomas. Cancer Res. 2002;62:4722–9. [PubMed] [Google Scholar]
- Schwartz DR, Wu R, Kardia SL, Levin AM, Huang CC, Shedden KA, et al. Novel candidate targets of beta-catenin/T-cell factor signaling identified by gene expression profiling of ovarian endometrioid adenocarcinomas. Cancer Res. 2003;63:2913–22. [PubMed] [Google Scholar]
- Shah N, Sukumar S. The Hox genes and their roles in oncogenesis. Nat Rev Cancer. 2010;10:361–71. doi: 10.1038/nrc2826. [DOI] [PubMed] [Google Scholar]
- Shao JS, Cheng SL, Pingsterhaus JM, Charlton-Kachigian N, Loewy AP, Towler DA. Msx2 promotes cardiovascular calcification by activating paracrine Wnt signals. J Clin Invest. 2005;115:1210–20. doi: 10.1172/JCI24140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Song L, Li Y, Wang K, Wang YZ, Molotkov A, Gao L, et al. Lrp6-mediated canonical Wnt signaling is required for lip formation and fusion. Development. 2009;136:3161–71. doi: 10.1242/dev.037440. [DOI] [PubMed] [Google Scholar]
- Suzuki M, Tanaka M, Iwase T, Naito Y, Sugimura H, Kino I. Over-expression of HOX-8, the human homologue of the mouse Hox-8 homeobox gene, in human tumors. Biochem Biophys Res Commun. 1993;194:187–93. doi: 10.1006/bbrc.1993.1802. [DOI] [PubMed] [Google Scholar]
- Takahashi C, Akiyama N, Matsuzaki T, Takai S, Kitayama H, Noda M. Characterization of a human MSX-2 cDNA and its fragment isolated as a transformation suppressor gene against v-Ki-ras oncogene. Oncogene. 1996;12:2137–46. [PubMed] [Google Scholar]
- Whyte DB, Holbeck SL. Correlation of PIK3Ca mutations with gene expression and drug sensitivity in NCI-60 cell lines. Biochem Biophys Res Commun. 2006;340:469–75. doi: 10.1016/j.bbrc.2005.12.025. [DOI] [PubMed] [Google Scholar]
- Willert J, Epping M, Pollack JR, Brown PO, Nusse R. A transcriptional response to Wnt protein in human embryonic carcinoma cells. BMC Dev Biol. 2002;2:8. doi: 10.1186/1471-213x-2-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wright K, Wilson P, Morland S, Campbell I, Walsh M, Hurst T, et al. beta-catenin mutation and expression analysis in ovarian cancer: exon 3 mutations and nuclear translocation in 16% of endometrioid tumours. Int J Cancer. 1999;82:625–9. doi: 10.1002/(sici)1097-0215(19990827)82:5<625::aid-ijc1>3.0.co;2-2. [DOI] [PubMed] [Google Scholar]
- Wu R, Hendrix-Lucas N, Kuick R, Zhai Y, Schwartz DR, Akyol A, et al. Mouse model of human ovarian endometrioid adenocarcinoma based on somatic defects in the Wnt/B-catanin and PI3K/Pten signaling pathways. Cancer Cell. 2007;11:321–33. doi: 10.1016/j.ccr.2007.02.016. [DOI] [PubMed] [Google Scholar]
- Wu R, Zhai Y, Fearon ER, Cho KR. Diverse Mechanisms of beta-Catenin Deregulation in Ovarian Endometrioid Adenocarcinomas. Cancer Res. 2001;61:8247–55. [PubMed] [Google Scholar]
- Zhai Y, Kuick R, Nan B, Ota I, Weiss SJ, Trimble CL, et al. Gene expression analysis of preinvasive and invasive cervical squamous cell carcinomas identifies HOXC10 as a key mediator of invasion. Cancer Res. 2007;67:10163–72. doi: 10.1158/0008-5472.CAN-07-2056. [DOI] [PubMed] [Google Scholar]
- Zhai Y, Wu R, Schwartz DR, Darrah D, Reed H, Kolligs FT, et al. Role of beta-catenin/T-cell factor-regulated genes in ovarian endometrioid adenocarcinomas. Am J Pathol. 2002;160:1229–38. doi: 10.1016/s0002-9440(10)62550-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.