Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Identification of molecular markers and signaling pathway in endometrial cancer in Hong Kong Chinese women by genome-wide gene expression profiling

Abstract

Endometrial cancer is the third most common gynecologic malignancy and the ninth most common malignancy for females overall in Hong Kong. Approximately 80% or more of these cancers are endometrioid endometrial adenocarcinomas. The aim of this study was to reveal genes contributing to the development of endometrioid endometrial cancer, which may impact diagnosis, prognosis and treatment of the disease. Whole-genome gene expression analysis was completed for a set of 55 microdissected sporadic endometrioid endometrial adenocarcinomas and 29 microdissected normal endometrium specimens using the Affymetrix Human U133 Plus 2.0 oligonucleotide microarray. Selected genes of interest were validated by quantitative real-time-polymerase chain reaction (qRT-PCR). Pathway analysis was performed to reveal gene interactions involved in endometrial tumorigenesis. Unsupervised hierarchical clustering displayed a distinct separation between the endometrioid adenocarcinomas and normal endometrium samples. Supervised analysis identified 117 highly differentially regulated genes (4.0-fold change), which distinguished the endometrial cancer specimens from normal endometrium. Twelve novel genes including DKK4, ZIC1, KIF1A, SAA2, LOC16378, ALPP2, CCL20, CXCL5, BST2, OLFM1, KLRC1 and MBC45780 were deregulated in the endometrial cancer, and further validated in an independent set of 56 cancer and 29 normal samples using qRT-PCR. In addition, 10 genes were differentially regulated in late-stage cancer, as compared to early-stage disease, and may be involved in tumor progression. Pathway analysis of the expression data from this tumor revealed an interconnected network consisting of 21 aberrantly regulated genes involved in angiogenesis, cell proliferation and chromosomal instability. The results of this study highlight the molecular features of endometrioid endometrial cancer and provide insight into the events underlying the development and progression of endometrioid endometrial cancer.

Introduction

Endometrial cancer is the third most common gynecologic malignancy and the ninth most common malignancy for females overall in Hong Kong. Its documented incidence in 2002 was 12.5 per 100 000 women (Hong Kong Cancer Stat., 2005). Although it has a comparatively low mortality rate compared with other gynecologic cancers, a proportion of cases display aggressive characteristics. Approximately 80% or more of endometrial cancers are endometrioid adenocarcinoma whereas only about 10% are serous adenocarcinomas and clear cell tumors (Ilvan et al., 2004; Amant et al., 2005). All three histotypes are associated with a distinct prevalence, etiology and prognosis. Serous endometrial carcinomas occur almost exclusively in postmenopausal women (>60 years), are unrelated to estrogen stimulation, and evolve from atrophic rather than hypertrophic epithelium. Endometrioid endometrial carcinomas are associated with hormonal imbalance, hyperlipidemia and obesity (Hendrickson et al., 1982). Precancerous hyperplasia lesions have been identified in both serous and endometrioid types of endometrial carcinoma, however, they are more frequently seen in association with malignant endometrioid histology. Clear-cell carcinoma of the endometrium is a relatively rare malignancy that is considered to be one of the most aggressive types of endometrial carcinoma (Abeler and Kjorstad, 1991).

Currently, a number of somatic gene mutations are being studied in order to provide insight into the pathogenesis of sporadic endometrial cancer. Although mutations in the mismatch repair genes are found in inherited endometrial cancers and confer a lifetime risk of developing the disease that approaches 60%, these are rarely seen in sporadic cancers. Mutations in the tumor suppressor gene TP53 have been reported in over 90% of serous papillary endometrial cancers (Tashiro et al., 1997). In contrast, none were documented in endometrioid tumors (Lax et al., 2000). These findings suggest that subgroups of endometrial cancers undergo unique genetic events during pathogenesis.

Gene expression profiling has proven invaluable in the identification of molecular pathways contributing to carinogenesis and metastasis. Comparing gene expression patterns in cells or tissues representing distinct phenotypes, for example, normal and malignant cells, can help link cellular phenotype with the transcriptome. Identifying genes that are differentially expressed in endometrial cancer is important for understanding the biology of endometrial tumorigenesis and progression. In this study, we utilized microdissected endometrioid tumor specimens and normal endometrium to identify differentially expressed genes apart from intervening stromal cells and lymphocytic infiltrate. Using the Affymetrix Human U133 Plus 2.0 oligonucleotide microarray, which contains over 55 000 probe sets representing more than 47 000 transcripts derived from approximately 39 500 human genes, we demonstrated a distinct separation between tumor and normal endometrial specimens. We identified genes that were highly deregulated in this common type of endometrial malignancy. Subsequent pathway analysis revealed signaling events, which might contribute to the development and progression of endometrioid endometrial cancer in Hong Kong Chinese women.

Results

Hierarchical clustering of endometrioid tumors and normal endometrium

Genome-wide expression profiles for 84 microdissected endometrioid endometrial tumors and normal isolates were generated using the Affymetrix U133 Plus 2.0 oligonucleotide microarray platform. Data filtering identified 2305 probe sets possessing informative signal intensity values in 20% of the arrays. To characterize the relationships between the tumor and normal endometrial samples, unsupervised hierarchical clustering was applied to all 2305 probe sets. A dendrogram possessing two distinct arms separating endometrioid tumors from normal endometrium was identified (Figure 1). Despite the existence of well-defined clusters between normal and tumor specimens, a single endometrioid tumor specimen was incorrectly classified. Furthermore, the cancers in the same stage were not clustered together as shown in the dendrogram of gene expression profiling.

Figure 1
figure1

Dendrogram is plotted from hierarchical clustering of 696 probe sets in 55 endometrial samples, and divides them into two major groups: cancer and normal control group. The sample is labeled by sample code. C, endometrioid endometrial adenocarcinoma; N, normal endometrium; I, stage I tumor; II, stage II tumor; and III, stage III tumor.

Identification of differentially regulated genes distinguishing endometrioid specimens from normal endometrium

In order to identify differentially expressed genes, which might contribute to the development or progression of endometrioid endometrial cancer, a supervised analysis was completed. Using a two-sided Welch modified two-sample t-test, 696 of the 2305 probe sets satisfying the filtering criteria were differentially regulated in the endometrial tumors at a P-value <0.01 and possessed a fold change 2.0 (90% lower bound) with a mean difference 100. Permutation analysis returned a 0.0% median false discovery rate for the 696 probe sets under consideration.

To reduce the complexity of the data set, 133 probe sets possessing a ±4.0-fold change were evaluated further. These 133 probe sets comprised 117 unique genes and expressed sequence tags (ESTs), of which 25 were upregulated (Figure 2) and 92 downregulated (Figure 3) in the 55 endometrioid endometrial adenocarcinomas. In all, seven of the 117 differentially expressed genes have previously been reported as differentially expressed in endometrial cancer (Table 1). The remaining 110 genes and ESTs have not been reported to be associated with the endometrial malignancy (Table 2).

Figure 2
figure2

Genes upregulated in endometrial cancer (C) compared to normal endometrium (N).

Figure 3
figure3

Genes downregulated in endometrial cancer (C) compared to normal endometrium (N).

Table 1 Genes previously reported as differentially regulated in endometrial cancer that were identified in the current study
Table 2 Previously unreported genes differentially regulated ±4-fold in endometrioid endometrial cancer

In this set of endometrial cancers, there were 41 stage I, four stage II and 10 stage III cases. When early-stage (I–II) and late-stage (III) tumors were compared, 10 transcripts were found to be downregulated in late-stage cancers (Table 3).

Table 3 Differentially expressed genes identified in late-stage endometrioid endometrial cancer as compared to early-stage disease

Microarray validation

To validate the microarray results, nine upregulated genes, DKK4, ZIC1, KIF1A, SAA2, LOC16378, ALPP2, CCL20, CXCL5 and BST2, and three downregulated genes, OLFM, KLRC1 and MBC45780, that displayed significant differential expression between the tumor specimens and normal endometrium, but have not been reported previously in endometrial cancer, were selected for quantitative RT-PCR (qRT-PCR) analysis on an independent set of 56 cancer and 29 normal samples. In addition, GADD45G, which was downregulated in late- stage versus early-stage disease, was also evaluated. For all 13 genes, changes in expression observed by qRT-PCR agreed favorably the array data (Figure 4a and b). The first 12 genes measured by qRT-PCR were also shown to be significantly upregulated or downregulated in cancers when compared to normal counterparts (P<0.01). GADD45G was significantly underexpressed (−5.59 folds) in 10 late-stage cancers compared to 46 early-stage cancers in this independent set of 56 tumors (P<0.01).

Figure 4
figure4

qRT-PCR analysis of 13 genes in endometrial cancer. (a) Twelve genes, up/downregulated in endometrical cancer compared to normal endometrium; (b) Gene GADD45G is downregulated in late stage of endometrial cancer compared to early stage of cancer. Black bars represent the average fold change obtained by microarray analysis, whereas the gray bars denote fold change data obtained by qRT-PCR.

Pathway analysis

To identify signaling pathways that are associated with endometrial tumorigenesis, we analysed our microarray expression data using PathwayAssist version 3.0 software. Inclusion in the pathway required differential expression of ±2.0-fold relative to normal endometrium (P<0.01). Among the 696 probe sets analysed, genes represented by two or more probe sets were averaged to establish a composite fold change value during analysis. Figure 5 contains 21 differentially regulated genes encoding proteins linked to angiogenesis, cell-cycle progression and chromosomal instability in endometrioid endometrial cancer. In addition, a number of estrogen-regulated genes including CCNA1, CCNE1 and metalloproteinase (MMP)-9 were identified in the analysis.

Figure 5
figure5

Pathway analysis of differentially regulated genes identified in endometrioid tumors as compared to normal endometrium. Pathway diagrams were generated using PathwayAssist software overlayed with gene expression data. Genes included in the analysis were required to have a ±2.0-fold change relative to normal endometrium (P<0.01). Multiple probe sets were averaged for each gene. Differentially expressed genes identified for endometrioid endometrial tumors and their associated interactions are shown. Upregulated genes are denoted in green, whereas downregulated genes are indicated in red. Genes that did not display differential regulation are colored gray.

Discussion

A number of studies have utilized microarray-based platforms to identify genes that are aberrantly expressed in endometrial tumors as compared to normal endometrium (Mutter et al., 2001; Moreno-Bueno et al., 2003; Risinger et al., 2003; Cao et al., 2004; Saidi et al., 2004; Smid-Koopman et al., 2004; Ferguson et al., 2005; Maxwell et al., 2005). These analyses have provided useful data, but all have been limited by a number of factors including heterogeneous tumor populations, an insufficient number of tumor and/or control specimens and microarray formats with relatively few features. To determine a more global assessment of differential gene expression in endometrial tumors in Chinese Hong Kong women, we used the whole-genome Affymetrix U133 Plus 2.0 microarray, and focused on a large microdissected tumor set containing specimens from the most common histologic subtype of endometrial cancer, endometrioid endometrial adenocarcinoma. As normal endometrium is a dynamic organ under hormonal regulation, we evaluated gene expression in proliferative, secretory and postmenopausal specimens in almost equal numbers. This provided a composite view of gene expression in normal endometrium for comparison to the endometrial malignancies.

Unsupervised hierarchical clustering clearly demonstrated that endometrioid endometrial cancer is distinct from normal endometrium. However, the early and late-stage diseases were not hierarchically well clustered, respectively. Despite successful classification of a majority of the samples, a single early stage, well-differentiated tumor clustered with the normal specimens.

To identify genes contributing to the clustering result, a supervised analysis was performed. A total of 696 probe sets were identified as differentially regulated ±2.0-fold (P-value <0.01) in endometrioid tumors versus normal endometrium. Included in this data set were 117 genes possessing a±4.0-fold change in expression. Among them seven genes have been reported previously as differentially expressed in endometrial tumors. Lactotransferrin (LTF) was upregulated (Walmer et al., 1995), whereas six genes including SFRP4, DCN, IGFBP6, HGF, WT1 and WNT4 were downregulated in the endometrial tumors (Rutanen et al., 1994; Bui et al., 1997; Yoshida et al., 2002; Muller-Tidow et al., 2003; Smid-Koopman et al., 2004; Acs et al., 2004; Hrzenjak et al., 2004). Of these genes, only WT1 was been specifically associated with endometrioid cancer (Acs et al., 2004). Repeated identification of these genes as aberrantly regulated in endometrial tumors suggests that they may play a significant role in the development and progression of the disease.

The remaining 110 genes have not been described in endometrioid endometrial cancer. Out of the 110 genes, 36 are associated with non-endometrial tumors (data not shown) whereas the remaining 74 genes are novel cancer-related transcripts that involved in multiple biological processes including cell cycle, protein metabolism, proteolysis and peptidolysis etc. From the 110 genes, 12 including DKK4, ZIC1, KIF1A, SAA2, LOC16378, ALPP2, CCL20, CXCL5, BST2, OLFM1, KLRC1 and MBC45780 were validated in an independent set of tumor and normal endometrium specimens. qRT-PCR data was significant for all 12 genes with fold change values approximating the microarray data. Among these genes, CCL20 together with its unique receptor may contribute to aspects of tumor cell growth and invasion through autocrine and paracrine mechanisms (Schutyser et al., 2003). In addition, concentrations of CXCL5 in peritoneal fluid are markedly elevated in the endometriosis patients, as compared with the controls, especially in women with severe endometriosis, whereas KLRC2 is associated with an increased susceptibility to cancer (Miyashita et al., 2004; Suzumori et al., 2004). These novel, validated genes may provide important insights into the biology underlying endometrioid endometrial cancer.

To reduce the dimensionality of the data set and uncover signaling pathways contributing to tumorigenesis and progression, pathway analysis was completed for all 696 probe sets identified as differentially regulated at a fold change 2.0. Co-regulated signaling events contributing to angiogenesis, cellular proliferation and chromosomal instability were identified. The proangiogenic protein MMP9 featured prominently in the analysis. MMP9 plays a critical role in the development of tumor-associated vasculature and can enhance the invasive potential of tumor cells (Pepper, 2001; Hojilla et al., 2003). Upregulation of this gene may be linked to increased levels of HMGA1 and TFAP2A, which are both positive regulators of MMP9 (Liu et al., 1999; Sivak et al., 2004). In breast cancer patients, TFAP2A is highly associated with MMP9 expression (Pellikainen et al., 2004). NRC3C1 is a nuclear receptor that can inhibit expression of MMP9 (Cha et al., 1988). In conjunction with elevated levels of HMGA1 and TFAP2, reduced expression of this gene may lead to increased MMP9 activity.

Critical modulators of MMP9 were also differentially regulated including lipocalin 2 (LCN2), tissue inhibitors of metalloproteinases (TIMP)2, TIMP3, thrombospondin 2 (THBS2) and proenkephalin (PENK). In human breast carcinoma cells LCN2 can protect MMP9 from degradation preserving its enzymatic activity (Yan et al., 2001). MMP9/LCN2 complexes have also been identified in the urine of breast cancer patients suggesting that this marker may be useful for predicting disease status in endometrioid endometrial cancer patients (Fernandez et al., 2005). TIMP family members are endogenous inhibitors of MMP proteinase activity (Murphy and Knauper, 1997). Downregulation of these genes in tumor cells stimulates MMP9-mediated degradation of the extracellular matrix. TIMP2 has also been shown to negatively regulate angiogenesis through a mechanism that is independent of its anti-MMP activity (Stetler-Stevenson and Seo, 2005). Thrombospondin 2 is another potent inhibitor of angiogenesis (Streit et al., 1999). This effect may be due to its ability to inhibit activation of pro-MMP9 protein, as well as downregulate MMP9 expression through a receptor-mediated event (Bein and Simons, 2000; Kamochi et al., 2003). Decreased THBS2 levels may be the result of enhanced MYB expression, which can antagonize transcription of THBS2 (Bein et al., 1998). The opioid peptide PENK may also mitigate MMP9 expression (Takeba et al., 2001). A coordinated increase in MMP9 expression, along with decreased inhibitor levels, suggests that MMP9 may play a critical role in endometrioid endometrial cancer. Secretion of this protein into the extracellular matrix may help support an extracellular environment that is amenable to vascularization, as well as tumor invasion.

Cancer cells are typically associated with aberrant cell-cycle regulation. As expected, a number of previously described proteins involved in cell proliferation were upregulated in endometrioid endometrial cancer (Milde-Langosch et al., 2001; Muller-Tidow et al., 2003). CCNA1 is the cyclin required for progression from S phase into G2, whereas CCNE1 is necessary for transition through the G1/S and G2/M checkpoints (Schafer, 1998). Increased CCN1A expression may be stimulated by the transcription factors MYB and HMGA1 (Muller et al., 1999; Reeves et al., 2001). In addition, downregulation of WT-1 is associated with increased expression of cyclin E1 (Loeb et al., 2002). WT-1 is also able to negatively regulate expression of MYB, whereas MAF can reduce MYB activity through heterodimeriziation (McCann et al., 1995; Hedge et al., 1998). In vivo and in vitro assays assessing tumor growth and cellular proliferation have demonstrated that 3-hydroxy-3-methylglutaryl coenzyme A reductase can stimulate CCNE1/CDK2 activity (Tanaka et al., 1998; Duncan et al., 2004). In contrast, ZAK, which encodes a serine/threonine kinase, is associated with reduced cyclin E1 levels and cell-cycle arrest (Yang, 2002). SPARC, a secreted matricellular protein, can also impede cell-cycle progression by inhibiting cyclin E1 and cyclin A1 in tumor and vascular smooth muscle cells (Dhanesuan et al., 2002; Motamed et al., 2002). Together these signaling events contribute to a cellular context favoring tumor expansion.

Beyond enhancing the proliferative potential of tumor cells, altered expression of cell-cycle-related proteins can also contribute to chromosomal instability. BUB1B and CDC20 are two regulators of the anaphase-promoting complex (APC). This complex of ubiquitin ligase proteins is responsible for the degradation of critical cell-cycle proteins ensuring checkpoint transitions occur in a normal manner (Harper et al., 2002). BUB1B is required for metaphase arrest in response to a lack of bipolar attachment. Overexpression of this protein is associated with a decrease in chromosomal stability and poor outcome in breast cancer patients (Warren et al., 2002; Dai et al., 2005). CDC20 is necessary for the degradation of an S-phase cyclin, which can antagonize APC activity (Harper 25 et al., 2002). Increased expression of CDC20 is also linked to increased chromosomal instability in breast cancer (Yuan et al., 2006). Phosphorylation of BIRC5 by cyclin B1 is essential for cell survival during mitosis. This observation suggests that BIRC5 may be play an important role in ensuring proper cell division occurs at the G2/M checkpoint (O'Connor et al., 2000). Abnormal regulation of these genes may lead to the accumulation of deleterious chromosomal alterations promoting tumor development and progression.

Estrogen-induced signaling plays a key role in the development and progression of endometrioid endometrial tumors. The necessity of estrogen-mediated signaling was evidenced by the differential regulation of known estrogen response genes including CCNE1, CCNA1, MMP9 and LTF. CCNE1 and CCNA1 have both been linked to estrogen stimulation (Tong and Pollard, 1999). Evidence that HMGA1 can enhance binding of the estrogen receptor to estrogen response elements may help explain increased CCNA1 levels, as well a increased MMP9 expression in the endometrial tumors (Massaad-Massade et al., 2004). The association of MMP9 transcription with estrogen signaling implies that this hormone may play a key role in the development of tumor vasculature required for disease progression (Mizumoto et al., 2002). Estrogen has also been implicated in the regulation of LTF2 (Liu and Teng, 1991). Upregulation of this gene in endometrial carcinoma cell lines has been linked to tamoxifen resistance suggesting a connection between estrogen signaling events and drug resistance (Albright and Kaufman, 2001). Given the hormone-dependent nature of these tumors, targeting estrogen-regulated genes for therapeutic intervention may facilitate the design of novel agents for the treatment of this disease.

To identify genes associated with the progression of endometrioid endometrial cancer, we compared the gene expression profiles of 45 stage I and II, and 10 stage III endometrial tumors. Late-stage tumors displayed similar expression profiles when compared to early-stage tumors, however 10 genes were significantly downregulated in late-stage compared to early-stage endometrial cancer. Among the 10 genes, GADD45G is involved in growth arrest and is inactivated in multiple tumor types (Kalsheker et al., 2002; Ying et al., 2005). Decreased expression of these genes in late-stage endometrial cancers may facilitate the identification of signaling pathways contributing to tumor progression.

Microsatellite instability (MSI) is a molecular phenotype present in 25% of endometrial cancers. Recently, Risinger et al. (2005) examined the global gene expression profiles in 24 early-stage endometrioid endometrial cancers including nine cases with and 15 cases without the MSI phenotype. Unsupervised principal component analysis of the expression data from these cases indicated two distinct groupings of cancers based on MSI phenotype. In particular, they found evidence that two members of the secreted frizzled-related protein family (SFRP1 and SFRP4) were more frequently downregulated in MSI cancers as compared with microsatellite stable cancers whereas the WNT target fibroblast growth factor 18 was found to be upregulated in MSI cancers (Risinger et al., 2005). The determination of MSI status in the cancer subjects of our present study, and then analysis of MSI-specific gene expression profiles are undergoing. We would like to further confirm which unique molecular signature exists in the MSI cancers and if it is able to classify histologically similar endometrioid endometrial cancers into two distinct groupings with implications affecting therapy and prevention based on such molecular signature(s).

In conclusion, we have identified differentially regulated genes distinguishing microdissected endometrioid endometrial adenocarcinoma from normal endometrium. In addition to novel cancer-related transcripts, signaling pathways contributing to angiogenesis, cell-cycle progression and chromosomal instability were identified. A number of these pathways may also be modulated through estrogen-related signaling events. Although late- stage an early-stage tumors retained similar gene expression profiles, differentially expressed genes were identified that may contribute to progression of the disease. Together, these findings provide a clearer picture of the pathways contributing to endometrioid endometrial cancer in Hong Kong Chinese women. Furthermore, they serve a starting point for the identification of novel clinical markers and therapeutic targets.

Materials and methods

Tissue samples

Study subjects were recruited at the Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong. One hundred and eleven tumor tissue specimens obtained from primary endometrioid endometrial adenocarcinoma were immediately embedded in optimal cutting temperature compound, snap-frozen in liquid nitrogen and stored at −80°C. Fifty-eight normal endometrium tissue specimens were collected from women who underwent a hysterectomy or endometrial curettage for unrelated conditions. Frozen sections from each specimen were stained by hematoxylin and eosin to verify the quality of the tissue sample and the extent of inflammatory infiltrate. Histologic typing of each tumor was according to World Health Organization (WHO) criteria, whereas clinical staging followed International Federation of Gynecology and Obstetrics (FIGO) standards. The mean age of patients with the endometrial cancer was 55.5 (36–82) years, whereas for the normal controls it was 52.8 (38–85) years. There was no statistical difference in age between tumor and control groups. All specimens, and their corresponding clinical information, were obtained under protocols approved by the Clinical Research Ethics Committee of The Chinese University of Hong Kong.

Microdissection and RNA extraction

Microdissection was performed to obtain homogeneous populations of neoplastic and normal glandular cells. Approximately 20 000 cells/specimen were accrued using laser capture microdissection (MicroBeam MicroLaser Systems, PALM Microlaser Technologies GmbH, Bernried, Germany). It is estimated that the epithelial components in both cancer and control samples reached to 90% or more, and were comparable.

Total RNA was extracted from microdissected target cells using a RNeasy Micro Kit (Qiagen, Valena, CA, USA) according to the manufacturer's protocol and stored at −80°C. Isolated RNA was analysed with an Agilent 2100 Bioanalyzer (Agilent Technologies Inc., Palo Alto, CA, USA) to verify sample integrity and quantify total RNA yield.

Total RNA amplification for Affymetrix GeneChip hybridization

Total RNA obtained from 55 endometrial cancer and 29 normal endometrium samples was amplified using the GeneChip Eukaryotic Small Sample Target Labeling Assay. As previously described, it utilizes two cycles of complementary DNA (cDNA) synthesis and in vitro transcription for the amplification of 50–100 ng of total RNA (Bonome et al., 2000). Labeled amplified RNA was then hybridized to the Affymetrix Human Genome U133 Plus 2.0 GeneChip (Affymetrix, Santa Clara, CA, USA), which contains over 55 000 probe sets representing more than 47 000 transcripts derived from well characterized genes. The arrays were scanned with a confocal laser GeneChip scanner. The resulting images were captured using GeneChip Operating Software (GCOS) (Affymetrix, Santa Clara, CA, USA).

Clustering and statistical analysis of expression data

CEL files generated in GCOS were analysed using DNA-Chip analyzer (dCHIP) (Version 1.3). (http://www.dCHIP.org) software platform (Li and Wong, 2003). The 84 CEL files generated by the Affymetrix Microarray Suite version 5.0 were converted into DCP files. Perfect match (PM) and mismatch (MM) probes were normalized before PM/MM model-based expression analysis. The generated gene list was filtered. Only those genes whose expression levels fulfilled the following criteria were included: 1<s.d./mean <10, gene called P (present) in 20% of samples. Global comparisons of data derived from cancer and normal endometrium samples were carried out as follows: Emean/Bmean>4 or Bmean/Emean>4; EmeanBmean>100 or BmeanEmean>100; and P for testing Emean=Bmean<0.01 (where E=experimental cancer samples and B=baseline normal endometrium samples). The t statistic was computed as (mean1−mean2)/[SE(mean1)2+SE(mean2)2] and its P computed based on the t distribution, and the degree of freedom was set according to the Welch-modified two-sample t-test. For the subsequent analysis, only probe sets differentially expressed >±2.0-fold in tumor versus normal endometrium were considered.

qRT-PCR validation

qRT-PCR was performed on total RNA obtained from an independent set of tumor (56 cases) and normal (29 cases) specimens for which microdissection was also completed. Each reaction was run in duplicate on an ABI Prism 7900 Sequence Analyzer (Applied Biosystems, Foster City, CA, USA). The comparative computed tomography method was used to calculate fold change values as specified by the manufacturer. In brief, 1 μg of total RNA from each sample was reverse-transcribed using SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA). An aliquot of cDNA from each reverse-transcribed sample (1 μl) was PCR amplified using gene-specific Assay-On-Demand TaqMan probes for DKK4, ZIC1, KIF1A, SAA2, LOC16378, ALPP2, CCL20, CXCL5, BST2, OLFM, KLRC1, MBC45780 or GADD45G and TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA, USA). Expression values for each gene were determined using a standard curve constructed from human universal RNA (Clontech, Palo Alto, CA, USA). The housekeeping gene glyceraldehyde phosphate dehydrogenase was selected for normalization. Student's t-test was used to assess the statistical difference of each gene expression level measured by qRT-PCR between the target and control groups.

Pathway analysis

To identify co-regulated pathways contributing to the distinct biology associated with endometrioid endometrial cancer, 696 probe sets identified as differentially regulated at least >±2.0-fold relative to normal endometrium (P<0.01) were analysed using PathwayAssist version 3.0 software (Iobion Informatics LLC, La Jolla, CA, USA). This software package contains over 500 000 documented protein interactions acquired from PubMed using the natural language processing algorithm MEDSCAN. This proprietary database was used to develop a biological association network and identify putative co-regulated signaling pathways.

Accession codes

Accessions

GenBank/EMBL/DDBJ

References

  1. Abeler VM, Kjorstad KE . (1991). Clear cell carcinoma of the endometrium: a histopathological and clinical study of 97 cases. Gynecol Oncol 40: 207–217.

    CAS  Article  Google Scholar 

  2. Acs G, Pasha T, Zhang PJ . (2004). WT1 is differentially expressed in serous, endometrioid, clear cell, and mucinous carcinomas of the peritoneum, fallopian tube, ovary, and endometrium. Int J Gynecol Pathol 23: 110–118.

    Article  Google Scholar 

  3. Albright CD, Kaufman DG . (2001). Lactoferrin: a tamoxifen-responsive protein in normal and malignant human endometrial cells in culture. Exp Mol Pathol 70: 71–76.

    CAS  Article  Google Scholar 

  4. Amant F, Moerman P, Neven P, Timmerman D, Van Limbergen E, Vergote I . (2005). Endometrial cancer. Lancet 366: 491–505.

    Article  Google Scholar 

  5. Bein K, Simons M . (2000). Thrombospondin type 1 repeats interact with matrix metalloproteinase 2. Regulation of metalloproteinase activity. J Biol Chem 275: 32167–32173.

    CAS  Article  Google Scholar 

  6. Bein K, Ware JA, Simons M . (1998). Myb-dependent regulation of thrombospondin 2 expression. Role of mRNA stability. J Biol Chem 273: 21423–21429.

    CAS  Article  Google Scholar 

  7. Bonome T, Lee JY, Park DC, Radonovich M, Pise-Masison C, Brady J et al. (2000). Expression profiling of serous low malignant potential, low-grade, and high-grade tumors of the ovary. Cancer Res 65: 10602–10612.

    Article  Google Scholar 

  8. Bui TD, Zhang L, Rees MC, Bicknell R, Harris AL . (1997). Expression and hormone regulation of Wnt2, 3, 4, 5a, 7a, 7b and 10b in normal human endometrium and endometrial carcinoma. Br J Cancer 75: 1131–1136.

    CAS  Article  Google Scholar 

  9. Cao QJ, Belbin T, Socci N, Balan R, Prystowsky MB, Childs G et al. (2004). Distinctive gene expression profiles by cDNA microarrays in endometrioid and serous carcinomas of the endometrium. Int J Gynecol Pathol 23: 321–329.

    Article  Google Scholar 

  10. Cha HJ, Park MT, Chung HY, Kim ND, Sato H, Seiki M et al. (1988). Ursolic acid-induced down-regulation of MMP-9 gene is mediated through the nuclear translocation of glucocorticoid receptor in HT1080 human fibrosarcoma cells. Oncogene 16: 771–778.

    Article  Google Scholar 

  11. Dai H, van't Veer L, Lamb J, He YD, Mao M, Fine BM et al. (2005). A cell proliferation signature is a marker of extremely poor outcome in a subpopulation of breast cancer patients. Cancer Res 65: 4059–4066.

    CAS  Article  Google Scholar 

  12. Dhanesuan N, Sharp JA, Blick T, Price JT, Thompson EW . (2002). Doxycycline-inducible expression of SPARC/Osteonectin/BM40 in MDA-MB-231 human breast cancer cells results in growth inhibition. Breast Cancer Res Treat 75: 73–85.

    CAS  Article  Google Scholar 

  13. Duncan RE, El-Sohemy A, Archer MC . (2004). Mevalonate promotes the growth of tumors derived from human cancer cells in vivo and stimulates proliferation in vitro with enhanced cyclin-dependent kinase-2 activity. J Biol Chem 279: 33079–33084.

    CAS  Article  Google Scholar 

  14. Ferguson SE, Olshen AB, Viale A, Barakat RR, Boyd J . (2005). Stratification of intermediate-risk endometrial cancer patients into groups at high risk or low risk for recurrence based on tumor gene expression profiles. Clin Cancer Res 11: 2252–2257.

    CAS  Article  Google Scholar 

  15. Fernandez CA, Yan L, Louis G, Yang J, Kutok JL, Moses MA . (2005). The matrix metalloproteinase-9/neutrophil gelatinase-associated lipocalin complex plays a role in breast tumor growth and is present in the urine of breast cancer patients. Clin Cancer Res 11: 5390–5395.

    CAS  Article  Google Scholar 

  16. Harper JW, Burton JL, Solomon MJ . (2002). The anaphase-promoting complex: it's not just for mitosis any more. Genes Dev 16: 2179–2206.

    CAS  Article  Google Scholar 

  17. Hedge SP, Kumar A, Kurschner C, Shapiro LH . (1998). c-Maf interacts with c-Myb to regulate transcription of an early myeloid gene during differentiation. Mol Cell Biol 18: 2729–2737.

    CAS  Article  Google Scholar 

  18. Hendrickson M, Ross J, Eifel P, Martinez A, Kempson R . (1982). Uterine papillary serous carcinoma: a highly malignant form of endometrial adenocarcinoma. Am J Surg Pathol 6: 93–108.

    CAS  Article  Google Scholar 

  19. Hojilla CV, Mohammed FF, Khokha R . (2003). Matrix metalloproteinases and their tissue inhibitors direct cell fate during cancer development. Br J Cancer 89: 1817–1821.

    CAS  Article  Google Scholar 

  20. Hong Kong Cancer Registry. (2005). Hong Kong Cancer Stat 2002. Hong Kong Authority, Hong Kong.

  21. Hrzenjak A, Tippl M, Kremser ML, Strohmeier B, Guelly C, Neumeister D et al. (2004). Inverse correlation of secreted frizzled-related protein 4 and beta-catenin expression in endometrial stromal sarcomas. J Pathol 204: 19–27.

    CAS  Article  Google Scholar 

  22. Ilvan S, Akyildiz EU, Calay Z, Celikoyar M, Sahinler I . (2004). Endometrial clear cell carcinoma metastatic to the paranasal sinuses: a case report and review of the literature. Gynecol Oncol 94: 232–234.

    Article  Google Scholar 

  23. Kalsheker N, Morley S, Morgan K . (2002). Gene regulation of the serine proteinase inhibitors alpha1-antitrypsin and alpha1-antichymotrypsin. Biochem Soc Trans 30: 93–98.

    CAS  Article  Google Scholar 

  24. Kamochi J, Tokunaga T, Tomii Y, Abe Y, Hatanaka H, Kijima H et al. (2003). Overexpression of the thrombospondin 2 (TSP2) gene modulated by the matrix metalloproteinase family expression and production in human colon carcinoma cell line. Oncol Rep 10: 881–884.

    CAS  PubMed  Google Scholar 

  25. Lax SF, Kendall B, Tashiro H, Slebos RJ, Hedrick L . (2000). The frequency of p53, K-ras mutations, and microsatellite instability differs in uterine endometrioid and serous carcinoma: evidence of distinct molecular genetic pathways. Cancer 88: 814–824.

    CAS  Article  Google Scholar 

  26. Li C, Wong W . (2003). DNA-Chip Analyzer (dChip): The Analysis of Gene Expression Data, Methods and Software. Springer-Verlag: New York.

    Book  Google Scholar 

  27. Liu WM, Guerra-Vladusic FK, Kurakata S, Lupu R, Kohwi-Shigematsu T . (1999). HMG-I(Y) recognizes base-unpairing regions of matrix attachment sequences and its increased expression is directly linked to metastatic breast cancer phenotype. Cancer Res 59: 5695–5703.

    CAS  PubMed  Google Scholar 

  28. Liu YH, Teng CT . (1991). Characterization of estrogen-responsive mouse lactoferrin promoter. J Biol Chem 266: 21880–21885.

    CAS  PubMed  Google Scholar 

  29. Loeb DM, Korz D, Katsnelson M, Burwell EA, Friedman AD, Sukumar S . (2002). Cyclin E is a target of WT1 transcriptional repression. J Biol Chem 277: 19627–19632.

    CAS  Article  Google Scholar 

  30. Massaad-Massade L, Tacine R, Dulauroy S, Reeves R, Barouki R . (2004). The functional interaction between HMGA1 and the estrogen receptor requires either the N- or the C-terminal domain of the receptor. FEBS Lett 559: 89–95.

    CAS  Article  Google Scholar 

  31. Maxwell GL, Chandramouli GV, Dainty L, Litzi TJ, Berchuck A, Barrett JC et al. (2005). Microarray analysis of endometrial carcinomas and mixed mullerian tumors reveals distinct gene expression profiles associated with different histologic types of uterine cancer. Clin Cancer Res 11: 4056–4066.

    CAS  Article  Google Scholar 

  32. McCann S, Sullivan J, Guerra J, Arcinas M, Boxer LM . (1995). Repression of the c-myb gene by WT1 protein in T and B cell lines. J Biol Chem 270: 23785–23789.

    CAS  Article  Google Scholar 

  33. Milde-Langosch K, Bamberger AM, Goemann C, Rossing E, Rieck G, Kelp B et al. (2001). Expression of cell-cycle regulatory proteins in endometrial carcinomas: correlations with hormone receptor status and clinicopathologic parameters. J Cancer Res Clin Oncol 127: 537–544.

    CAS  Article  Google Scholar 

  34. Miyashita R, Tsuchiya N, Hikami K, Kuroki K, Fukazawa T, Bijl M et al. (2004). Molecular genetic analyses of human NKG2C (KLRC2) gene deletion. Int Immunol 16: 163–168.

    CAS  Article  Google Scholar 

  35. Mizumoto H, Saito T, Ashihara K, Nishimura M, Tanaka R, Kudo R . (2002). Acceleration of invasive activity via matrix metalloproteinases by transfection of the estrogen receptor-alpha gene in endometrial carcinoma cells. Int J Cancer 100: 401–406.

    CAS  Article  Google Scholar 

  36. Moreno-Bueno G, Sanchez-Estevez C, Cassia R, Rodriguez-Perales S, Diaz-Uriarte R, Dominguez O et al. (2003). Differential gene expression profile in endometrioid and nonendometrioid endometrial carcinoma: STK15 is frequently overexpressed and amplified in nonendometrioid carcinomas. Cancer Res 63: 5697–5702.

    CAS  Google Scholar 

  37. Motamed K, Funk SE, Koyama H, Ross, Raines EW, Sage EH . (2002). Inhibition of PDGF-stimulated and matrix-mediated proliferation of human vascular smooth muscle cells by SPARC is independent of changes in cell shape or cyclin-dependent kinase inhibitors. J Cell Biochem 84: 759–771.

    Article  Google Scholar 

  38. Muller C, Yang R, Idos G, Tidow N, Diederichs S, Koch OM et al. (1999). c-myb transactivates the human cyclin A1 promoter and induces cyclin A1 gene expression. Blood 94: 4255–4262.

    CAS  PubMed  Google Scholar 

  39. Muller-Tidow C, Diederichs S, Schrader MG, Vogt U, Miller K, Berdel WE et al. (2003). Cyclin A1 is highly expressed in aggressive testicular germ cell tumors. Cancer Lett 190: 89–95.

    CAS  Article  Google Scholar 

  40. Murphy G, Knauper V . (1997). Relating matrix metalloproteinase structure to function: why the “hemopexin” domain? Matrix Biol 15: 511–518.

    CAS  Article  Google Scholar 

  41. Mutter GL, Baak JP, Fitzgerald JT, Gray R, Neuberg D, Kust GA et al. (2001). Global expression changes of constitutive and hormonally regulated genes during endometrial neoplastic transformation. Gynecol Oncol 83: 177–185.

    CAS  Article  Google Scholar 

  42. O'Connor DS, Grossman D, Plescia J, Li F, Zhang H, Villa A et al. (2000). Regulation of apoptosis at cell division by p34cdc2 phosphorylation of survivin. Proc Natl Acad Sci USA 97: 13103–13107.

    CAS  Article  Google Scholar 

  43. Pellikainen JM, Ropponen KM, Kataja VV, Kellokoski JK, Eskelinen MJ, Kosma VM . (2004). Expression of matrix metalloproteinase (MMP)-2 and MMP-9 in breast cancer with a special reference to activator protein-2, HER2, and prognosis. Clin Cancer Res 10: 7621–7628.

    CAS  Article  Google Scholar 

  44. Pepper MS . (2001). Role of the matrix metalloproteinase and plasminogen activator-plasmin systems in angiogenesis. Arterioscler Thromb Vasc Biol 21: 1104–1117.

    CAS  Article  Google Scholar 

  45. Reeves R, Edberg DD, Li Y . (2001). Architectural transcription factor HMGI(Y) promotes tumor progression and mesenchymal transition of human epithelial cells. Mol Cell Biol 21: 575–594.

    CAS  Article  Google Scholar 

  46. Risinger JI, Maxwell GL, Chandramouli GVR, Aprelikova O, Litzi T, Asad Umar A et al. (2005). Gene expression profiling of microsatellite unstable and microsatellite stable endometrial cancers indicates distinct pathways of aberrant signaling. Cancer Res 65: 5031–5037.

    CAS  Article  Google Scholar 

  47. Risinger JI, Maxwell GL, Chandramouli GV, Jazaeri A, Aprelikova O, Patterson T et al. (2003). Microarray analysis reveals distinct gene expression profiles among different histologic types of endometrial cancer. Cancer Res 63: 6–11.

    CAS  Google Scholar 

  48. Rutanen EM, Nyman T, Lehtovirta P, Ammala M, Pekonen F . (1994). Suppressed expression of insulin-like growth factor binding protein-1 mRNA in the endometrium: a molecular mechanism associating endometrial cancer with its risk factors. Int J Cancer 59: 307–312.

    CAS  Article  Google Scholar 

  49. Saidi SA, Holland CM, Kreil DP, MacKay DJ, Charnock-Jones DS, Print CG et al. (2004). Independent component analysis of microarray data in the study of endometrial cancer. Oncogene 23: 6677–6683.

    CAS  Article  Google Scholar 

  50. Schafer KA . (1998). The cell cycle: a review. Vet Pathol 35: 461–478.

    CAS  Article  Google Scholar 

  51. Schutyser E, Struyf S, Van Damme J . (2003). The CC chemokine CCL20 and its receptor CCR6. Cytokine Growth Factor Rev 14: 409–426.

    CAS  Article  Google Scholar 

  52. Sivak JM, West-Mays JA, Yee A, Williams T, Fini ME . (2004). Transcription Factors Pax6 and AP-2alpha interact to coordinate corneal epithelial repair by controlling expression of matrix metalloproteinase gelatinase B. Mol Cell Biol 24: 245–257.

    CAS  Article  Google Scholar 

  53. Smid-Koopman E, Blok LJ, Helmerhorst TJ, Chadha-Ajwani S, Burger CW, Brinkmann AO et al. (2004). Gene expression profiling in human endometrial cancer tissue samples: utility and diagnostic value. Gynecol Oncol 93: 292–300.

    CAS  Article  Google Scholar 

  54. Stetler-Stevenson WG, Seo DW . (2005). TIMP-2: an endogenous inhibitor of angiogenesis. Trends Mol Med 11: 97–103.

    CAS  Article  Google Scholar 

  55. Streit M, Riccardi L, Velasco P, Brown LF, Hawighorst T, Bornstein P et al. (1999). Thrombospondin-2: a potent endogenous inhibitor of tumor growth and angiogenesis. Proc Natl Acad Sci USA 96: 14888–14893.

    CAS  Article  Google Scholar 

  56. Suzumori N, Katano K, Suzumori K . (2004). Peritoneal fluid concentrations of epithelial neutrophil-activating peptide-78 correlate with the severity of endometriosis. Fertil Steril 81: 305–308.

    CAS  Article  Google Scholar 

  57. Takeba Y, Suzuki N, Kaneko A, Asai T, Sakane T . (2001). Endorphin and enkephalin ameliorate excessive synovial cell functions in patients with rheumatoid arthritis. J Rheumatol 28: 2176–2183.

    CAS  PubMed  Google Scholar 

  58. Tanaka T, Tatsuno I, Noguchi Y, Uchida D, Oeda T, Narumiya S et al. (1998). Activation of cyclin-dependent kinase 2 (Cdk2) in growth-stimulated rat astrocytes. Geranylgeranylated Rho small GTPase(s) are essential for the induction of cyclin E gene expression. J Biol Chem 273: 26772–26778.

    CAS  Article  Google Scholar 

  59. Tashiro H, Isacson C, Levine R, Kurman RJ, Cho KR, Hedrick L . (1997). p53 gene mutations are common in uterine serous carcinoma and occur early in their pathogenesis. Am J Pathol 150: 177–185.

    CAS  PubMed  PubMed Central  Google Scholar 

  60. Tong W, Pollard JW . (1999). Progesterone inhibits estrogen-induced cyclin D1 and cdk4 nuclear translocation, cyclin E- and cyclin A-cdk2 kinase activation, and cell proliferation in uterine epithelial cells in mice. Mol Cell Biol 19: 2251–2264.

    CAS  Article  Google Scholar 

  61. Walmer DK, Padin CJ, Wrona MA, Healy BE, Bentley RC, Tsao MS et al. (1995). Malignant transformation of the human endometrium is associated with overexpression of lactoferrin messenger RNA and protein. Cancer Res 55: 1168–1175.

    CAS  PubMed  Google Scholar 

  62. Warren CD, Brady DM, Johnston RC, Hanna JS, Hardwick KG, Spencer FA . (2002). Distinct chromosome segregation roles for spindle checkpoint proteins. Mol Biol Cell 13: 3029–3041.

    CAS  Article  Google Scholar 

  63. Yan L, Borregaard N, Kjeldsen, Moses MA . (2001). The high molecular weight urinary matrix metalloproteinase (MMP) activity is a complex of gelatinase B/MMP-9 and neutrophil gelatinase-associated lipocalin (NGAL). Modulation of MMP-9 activity by NGAL. J Biol Chem 276: 37258–37265.

    CAS  Article  Google Scholar 

  64. Yang JJ . (2002). Mixed lineage kinase ZAK utilizing MKK7 and not MKK4 to activate the c-Jun N-terminal kinase and playing a role in the cell arrest. Biochem Biophys Res Commun 297: 105–110.

    CAS  Article  Google Scholar 

  65. Ying J, Srivastava G, Hsieh WS, Gao Z, Murray P, Liao SK et al. (2005). The stress-responsive gene GADD45G is a functional tumor suppressor, with its response to environmental stresses frequently disrupted epigenetically in multiple tumors. Clin Cancer Res 11: 6442–6449.

    CAS  Article  Google Scholar 

  66. Yoshida S, Harada T, Iwabe T, Taniguchi F, Fujii A, Sakamoto Y et al. (2002). Induction of hepatocyte growth factor in stromal cells by tumor-derived basic fibroblast growth factor enhances growth and invasion of endometrial cancer. J Clin Endocrinol Metab 87: 2376–2383.

    CAS  Article  Google Scholar 

  67. Yuan B, Xu Y, Woo JH, Wang Y, Bae YK, Yoon DS et al. (2006). Increased expression of mitotic checkpoint genes in breast cancer cells with chromosomal instability. Clin Cancer Res 12: 405–410.

    CAS  Article  Google Scholar 

Download references

Acknowledgements

This work was supported by a Hong Kong Research Grant Council Earmarked Grant (RGC) (4427/03M).

Author information

Affiliations

Authors

Corresponding author

Correspondence to Y F Wong.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Wong, Y., Cheung, T., Lo, K. et al. Identification of molecular markers and signaling pathway in endometrial cancer in Hong Kong Chinese women by genome-wide gene expression profiling. Oncogene 26, 1971–1982 (2007). https://doi.org/10.1038/sj.onc.1209986

Download citation

Keywords

  • endometrioid endometrial cancer
  • signaling pathway
  • gene expression profiling

Search

Quick links