Abstract
OBJECTIVE To study the gene expression profiles of human endometrial cancers at various differentia0ted grade levels and to identify the genes related to differentiation of the endometrial cancers.
METHODS cDNA microarray technology was used to analyze the differentially-expressed genes among different differentiated grades of 32 cases of endometrial cancer. Hierarchical cluster analysis (HCA) for the gene expression profiles of the cases was employed.
RESULTS The tissue samples were grouped based on the various differentiated tumor grades with 33 differentiation-related genes identified out (P<0.001). Based on the results from the HCA, the conformity rate was 91% among the 33 differentially-expressed genes and the analysis of pathological classification.
CONCLUSION Genes related to the differentiation of endometrial cancer can be identified by using gene chips to analyze the expression profiles of endometrial cancers at various differentiated grades; HCA of the gene expression profiles can be helpful for distinguishing high-risk endometrial cancers before surgery.
keywords
INTRODUCTION
Endometrial carcinoma (EC) continues to be the most common malignancy of the female genital tract over the past years. Every year, EC developes in about 142,000 women worldwide, and about 42,000 women die from this cancer. Approximately 75% of the EC patients acquire endometrioid endometrial adenocarcinoma (EEA). Abnormal uterine bleeding is the most frequent symptom allowing EC to be detected and diagnosed in an early stage, with a 5-year survival rate of 83%. However, about 10% to 20% of the early-stage EEA patients experience a recurrence or metastasis[1]. At present, many of the prognostic indicators are not adequate to accurately predict clinical outcomes and many patients may have to receive radio or hormonal therapy, regardless of whether or not the additive treatment is needed. In our 17-year practice, we have observed that the EEA patients with same condition of the disease (the FIGO staging, myometrial invasion and histological grading, etc.) can have a very different response to the treatment and overall outcome.
Microarrays have allowed a unitary analysis of gene expression and have provided an opportunity to identify patterns that underlie biological differences between cancers[2,3]. They have been widely used to identify novel cancer-related genes, classify human cancer subtypes[4-6], and predict outcomes[7,8]. In our study, we searched for a progression gene expression profile of high-risk EEA, by using a high-density cDNA microarray. In order to identify the gene-expression patterns in relation to high-risk EEA, with a possible future relapse or tumor metastasis, we profiled a total of 32 EEA, and examined the differences in gene expression patterns between the EEC cases with different histological grades.
MATERIALS AND METHODS
Sources of the materials
The experimental group included 32 patients with endometrioid endometrial adenocarcinoma (EEA), with a mean age of 61 years, who were hospitalized during a period from May 2002 to May 2004. According to the standards of the International Federation of Gynecology and Obstetrics (FIGO) for surgical pathologic staging, 16 cases were identified as Stage I, 3 Stage II, 10 Stage III and 3 Stage IV; pathologic grades were as follows: 14 cases were confirmed as G1, 8 G2 and 10 G3. The control group was composed of mixed endometrial specimens obtained from 36 cases who underwent a complete hysterectomy or diagnostic curettage because of a hysteromyoma or abnormal uterine bleeding. All endometrial tissues from the 36 control cases were pathologically diagnosed as normal. The tissue specimens, obtained during the operation, were immediately put into liquid nitrogen until used. All specimens were analyzed using frozen sections and H&E staining. Moreover, the ratio of tumor cells in each specimen was larger than 80% (80.2%~100%) of the specimens in the experimental group.
Reagents and chips
The dNTP, reverse-transcriptase SuperScript II and human Cot-1 DNA (small-part repetitive sequences of human tissue; with a length of 100 bp) were all purchased from the Invitrogen Co., US. The Cy3 and Cy5-labbled dNTP were obtained from the Amersham Co. UK. A RNA extraction kit and PCR purification kit were from the QIAGEN Co., Germany and RNA enzyme inhibitors from the Takara Co., Japan. Oligo-dT was synthetized by the Shanghai Boya Co. The buffer solution and RNAase-free water were independently prepared by the Shanghai Biochip Co., and the cDNA gene chip used in the analysis contained a human 14K gene-expression profile of 13,824 genes developed by the Shanghai Biochip Co. Ltd.
Total RNA extraction, determination and control test
Total RNA from the EEA and normal endometrial tissues was extracted using the TRIzol method, and the absorbance at 260 nm and 280 nm used for RNA analysis, after which the quality of the RNA values was assayed by 1% agarose gel electrophoresis and by the chip laboratory. The QIAGEN RNA extraction kit was used for further RNA purification, and the selection of RNA of high quality was used for the labeling experiment.
Preparation and purification of the cDNA probe
Cy3 and Cy5 were used to label the experimental and control cDNA probes respectively while the RT was conducted. A 50 to 100 μg sample of purified total RNA from the experimental and control groups were separately taken for RT labeling of the cDNA, with a 1 to 2 μg of oligo-dT as the primer. A PCR purification kit was employed for further purification of the probes and a microplate scanning spectrophotometer was used to evaluate the content of the probes. Hybridization, washing, scanning, image analysis and primary treatment of the gene chips were all performed by the Shanghai Biochip Co. Ltd.
Analysis and processing of the data
Based on results of the pathological differentiation, the 32 EEA cases with the EEA were divided into three groups, i.e., the well, moderately and poorly-differentiated EEA and the gene expressions compared using analysis avariance (ANOVA). The differentially expressed genes in relation to the pathologic type were screened out, then based on the gene expression, a hierarchical cluster analysis (HCA) was conducted to assay the gene expression profile of the EEA, using Cluster 3.0 software. The results of the HCA were assessed using Java Treeview software[9,10].
Statistical analysis
ANOVA by SPSS11.0 software.
RESULTS
Analysis of the gene expression profile
Hybridization signals and scanning results from the gene chips
The intensity of fluorescent signals of the chips was high and uniform. From the 32 gene chips, the mean variation coefficient of 9 house-keeping genes was 15%, indicating a uniform chip experimental system. The detection rate of the 32 chips was 90%.
The differentially expressed genes from the EEA of various differentiated grades
After an ANOVA of the expression levels of each gene among the three groups, i.e., the well, moderately and poorly-differentiation groups, 33 differentially expressed genes in relation to the differentiations were screened out. There was a significant difference in comparing the expression among the three groups (P<0.001, Table 1).
The 33 EC differentially expressed genes from various differentiated grades.
Hierarchical cluster analysis (HCA)
HCA of the EEA was performed employing the 33 differentially expressed genes. Fig. 1 shows that the 32 EEA cases were divided into two categories, i.e., a) most of the EEA cases on the left side were poorly-differentiated EEA (PDEEA) (10 cases of PDEC were clustered in this category) and, b) well and moderately-differentiated EEA (WMDEEA) accounted for a majority on the right side of Fig. 1. Seven of the 8 who died were classified in the left-side category. The other 1 well-differentiated EEA (EC16) case was classified in the PDEEA group, but the patient died during follow-up. The coincidence between the results of HCA (high-risk and low-risk categories) and pathologic grading (WMDEEA and PDEEA categories) was 91% (29/32).
HCA dendrograms for the 32 EEA cases. The items on top of Fig.1 represent the tumor samples, and the items in the column on the right side of Fig. 1 indicate the differentially expressed genes among the three groups. After pathologic grading, the deceased patients were marked with a “d”; the red genes indicate an up-regulation of the EEA gene expression, the green show a down-regulation, and the genes in black are those showing no changes in the average expression level in the EEA compared to normal endometrium.
DISCUSSION
EC is one of the most frequent female genital tract tumors. Because the diagnosis of EC is usually conducted in an early stage, the prognoses for most patients has been favorable. However, recurrences still occur in about 20% of the patients. At present, the FIGO standard is still followed for surgical pathologic staging, and the staging system relies on the results of the postoperative histological examination. Moreover, with adoption of the staging for more than 10 years, there has been no obvious improvement in survivals of EC patients. The conditions that relate to the prognosis of patients with EC include the size, stage and tumor differentiation, patient age, and the standards of progesterone receptor and lymphatic metastasis. However, these clinicopathologic parameters may not precisely predict the prognosis for the EC patients[11].
Big flux and parallel studies on the change of gene expression profiles have been conducted using gene chip technology. Through investigations on the change of gene expression profiles in EC employing gene chips, such as the CD147 and TIMP-2, etc., Meng et al.[12] and Zhou et al.[13] found some differentially expressed genes in EC. However, because of limited samples, they were not able to compare the differences between the gene expression profiles of the EC with various differentiated grades. After initial analysis of the gene expression profiles in the EEA at various stages, we found some differentially expressed genes.
However, there was a poor coincidence (66%) between results of the cluster analysis and FIGO staging[14,15]. In our present study, the differences between the EEA gene expression profiles at various differentiated grades were determined, and the HCA of the gene expression profiles conducted using chips containing 13,824 genes.
Comparison between the gene expression profiles of the EEA with various tumor differentiated grades
A total of 32 cases of EEA were divided into 3 groups based on differentiated grades and 33 differentially expressed genes in relation to tumor differentiations were screened out. Among the genes studied, attention to two gene groups are need, i.e., a) the genes with an up-regulated expression in PDEEA and those with a down-regulation in WMDEEA, such as the p53 –induced ENC1, methylenetetrahydrofolate dehydrogenase (MTHFD2), septin 3(SEPT3), paraneoplastic antigen MA3 (PNMA3) and microchromosome maintenance deficient 2 protein (MCM2); b) the genes with a down-regulated expression in PDEC and with an up-regulation in WMDEC, such as the laminin-γ2 (LAMC2), ciliary axis dynein (DNAI1), urinary cortexin gene (UCN), transcription factor FOXJ1, sperm-related antigen 8 (SPAG8), non-transitional cell expressed protein (NME5), testicular tektin 2 (TEKT2), homeobox gene-1 (MSX1) 3′phosphoadenosine-5′-phosphosulfuric acid synthetase (PAPSS2), secretin family member SCGB2A1, foot cell specific protein (PODXL) and regulating factor X-related protein (RFXAP). Reports of these genes were also seen in related studies of EC in China and overseas[16-19]. The functions of these genes are mainly involved in DNA replication, transcription regulation, enzymatic reactions, and the roles in cell differentiation, movement, adhesion and anti-apoptosis, etc[20-23]. Further studies on the functions of these genes in development and invasion by the EC are expected.
HCA significance for the gene expression profiles for various EEA grades
Based on the HCA of the 33 differentially expressed genes in the 32 EEA cases, the coincidence between the HCA results and pathological stages was 91%. Based on the HCA, the EEA were divided into two categories, i.e., the one category on the left of Fig.1 belonged to the high-risk cases (most were PDEEA) and another on the right were the low-risk (WMDEEA). This risk evaluation, following the HCA based on gene expression profiles, was in agreement with the prognostic indicators for the known cell differentiation of the tumor, but with a higher accuracy. It can be seen in Fig. 2 that several WMDEEA cases, such as the EC16 (IIb, G1), EC23 (IVb, G2) and EC15 (IIIc, G2), were classified into the PDEEA category by HCA, suggesting that although these tumors showed a better cell differentiation, their biological quality might be close to the tumors with a poor differentiation and high malignancy, in which metastases, relapse and unfavorable prognosis often occur. The postoperative survival time of the EC16 and EC23 cases was 18 and 11 months respectively, and both died of tumor metastasis. An enlarged excision of the total uterus plus two adnexa, partial greater omentum and scavenge of pelvic lymph nodes was performed on the EC16 case in October, 2003, with a discharge diagnosis of Stage-IIb and grade I endometrial adenocarcinoma. The patient died of tumor metastasis in April, 2005. Until July 2005, the EC15 patient had survived for 40 months, without a sign of recurrence and metastasis. Follow-ups for long term survival are to be expected. Although the EC26 case of the category on right of the dendrogram was diagnosed as a Stage-IIa and grade-I case, the patient had a concurrent ovarian cancer. Moreover, a radical excision of a rectal cancer was conducted one year before surgical treatment of her EC. So, the patient belonged to the high-risk group and died 16 months after operation.
It has been shown in our initial research on the gene expression profiles in EEA of various grades, that an integrated study on the gene expression profiles of EEA cases cam be conducted employing the gene chip technology. Screening of the genes in relation to oncogenesis and tumor progression may be helpful to further understand roles of the differentially expressed genes. In addition, the HCA of gene expression profiles will allow for identification of high-risk EEA with a metastatic tendency or a poor prognosis. Also, preoperative acquirement of tissue samples for EC diagnosis followed by an analysis of the gene expression profiles can be of help to develop an individual regimen for the EC patients before operation.
Footnotes
This work was supported by a grant from the National Natural Science Foundation of China (No.30371481), the Natural Science Foundation of Shanghai (No. 06ZR14053) and the Key Project of the Shanghai Health Bureau (No.2005ZD002).
- Received January 3, 2007.
- Accepted January 16, 2007.
- Copyright © 2007 by Tianjin Medical University Cancer Institute & Hospital and Springer
REFERENCES
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.