Abstract
OBJECTIVE To study the difference in gene expression between human ovarian carcinoma and normal ovarian tissues, and screen the novel associated genes by cDNA microarrays.
METHODS Total RNA from 10 cases of ovarian cancer and from normal ovarian tissues were extracted by a single step method. The cDNA was retro-transcribed from an equal quantity of mRNA derived from the 10 cases of ovarian carcinoma and normal ovarian tissues, followed by labeling the cDNA strands with Cy5 and Cy3 fluorescence as probes. The mixed probes were hybridized with BiostarH 8464 dot human somatic cell genes. Fluorescence signals were assessed by a ScanArray 4000 laser scanner and the images analyzed by Gene Pix Pro 3.0 software with a digital computer.
RESULTS By applying the cDNA microarray we found a total of 185 genes for which expression levels differed more than 5 times comparing human ovarian carcinoma with normal ovarian epithelium. Among these genes 86 were up-regulated >5 times and 99 were down regulated <0.2.
CONCLUSION The cDNA microarray technique is effective in screening the differential gene expression between human ovarian cancers and normal ovarian epithelium. It is suggested that these genes identified are related to the genesis and development of ovarian carcinoma.
keywords
Differential gene expression in various tissues may be measured with parallel analysis by gene chips, a technique which has greatly improved traditional experiments in which only a single or several gene expressions can be examined. In recent years more and more cDNA microarray methods have been applied in the study of gene expression. In this paper, a gene chip technique was used to analyze the difference in gene expression patterns between human ovarian carcinomas and normal ovarian epithelium. Exploration of tumor-associated gene-clusters and their function in initiation and development of ovarian carcinoma will be helpful to understand the molecular mechanism of cell transformation, and provide molecular markers and target genes for clinical diagnosis, prevention, prognosis and treatment of ovarian carcinoma.
Materials and Methods
Materials
The tumor tissue specimens were from 10 ovarian carcinoma patients treated in the Gynecology Department of our hospital. Clinical pathologic data are shown in Table 1. Normal ovarian epithelium was taken from patients with other diseases. For each sample, a portion was frozen immediately in liquid nitrogen after surgical resection and the other part was used for histopathological examination. Control ovarian tissues were examined microscopically to verify that the ovarian epithelia were normal. The ovarian carcinoma types are classified in Table 1.
cDNA chip
Human gene expression chips (BiostarH-80s ) were obtained from the Shanghai Biostar Ltd.Co.
Extraction and purification of total RNA and isolation of mRNA
Total RNA of cancer cells and normal ovarian epithelium was extracted by a single step method.[1] The frozen ovarian carcinomas and normal tissues were completely ground in a ceramic mortar into a powder while adding liquid N2. The power was then transferred into a tube containg TRIzol agent and the power homogenized. After centrifugation, the supernatant was extracted twice with phenol: chloroform (1:1), then once with sodium acetate and acidic phenol: chloroform (5:1). The aqueous phase was precipitated by adding an equal volume of isopropanol. The precipitate was centrifuged and dissolved with Milli-Q water that was RNAse free. After further purification by a LiCl precipitating method, UV analysis and electrophoresis showed that the purification of the RNA was adequate. mRNA was isolated and purified using an oligotex mRNA Midi Kit (Quagen, Inc, USA).
Labeling the probes
Fluorescent-labeled cDNA probes were prepared by retro-transcription and purification using the method of Schena et al. [2] The probes from normal ovarian epithelium were labeled with Cy3-dUTP, and those from the cancer tissues with Cy5-dUTP. The probes were mixed (Cy3-dUTP control + Cy5-dUTP ovarian carcinoma), precipitated with ethanol and then dissolved in 20 μl of hybridization solution (5× SSC + 0.2%SDS).
Hybridization and washing
Pre-hybridization: The hybridizing probes and chips were denatured in a 95 °C bath for 2 min, then the chips were put into the pre-hybridizing solution and placed in a 95 °C water bath for 30 s. The chips were then placed into anhydrous alcohol for 30 s, dried at room temperature, and the pre-hybridizing solution added on the array area of the chip and the chip covered with glass. The chip was pre-hybridized in a sealed chamber at 42 °C for 5~6 h. Hybridization: After removing the cover glass, the slide was washed with double distilled water. Then the probe was put into the 95 °C bath for denaturing for 2 min after which it was taken out and put on ice. The chip was then put into the 95 °C bath for denaturing for 30 s, followed by putting the chip into anhydrous alcohol for 30 s. The probe was then placed on the chip and the chip covered with glass. The chip was hybridized in a sealed chamber covered with parafilm in a hybridizing box at 42 °C overnight for 16~18 h. Washing: After removing the coverglass, the slide was washed in a solution of 2 × SSC+ 0.2%SDS, 0.l%× SSC + 0.2%SDS,0.l% × SSC for 10 min, and dried at room temperature.
Fluorescence scanning and analysis
The chip was scanned with a ScanArray 4000 type Scanner. Cy3 and Cy5 overall intensities were normalized and corrected by a coefficient based on the ratios of 40 house-keeping genes. The acquired images were analyzed further by Gene Pix Pro 3.0 software with a digital computer to determine the intensities of fluorescent signals and the Cy5/Cy3 ratio. The differentially expressed genes were defined as follows: (1 )The signal values of Cy3 and Cy5 were all more than 200, or one value was more than 800. (2) The signal Cy5/Cy3 ratio was between 0.1 and 10.
Results
Total RNA analysis
The ratio of absorption of the 260 nm and 280 nm wavelength (A260/A280) was 1.71~1.81, and the electrorphoretic band of a hot-stability assay was conducted at 70 °C for 1 h. The results showed that bands at 28 s and 18 s were not degraded (Fig. 1).
An electrophoretogram of RNA from ovarian carcinoma cells and normal ovarian epithelium.
Scatter plot of hybridization signals on the gene chip
Fig.2 shows a scatter plot for the X and Y coordinates displaying the Cy3 and Cy5 fluorescent intensity values. Each data point represents one gene; Red spots represent a ratio of Y to X values between 0.5 and 2.0, indicating non-differential expression. Yellow spots distributed far away from a 45° diagonal line indicate the presence of a gene in which expression has changed in the ovarian carcinoma, i.e. their signal intensities were 2 times more than that of the normal control.
The scatter plot of the Cy3 and Cy5 fluorescent signal intensity values.
Expression patterns by scanning analysis
A double-color fluorescent comparison profile of the ovarian carcinoma and normal ovarian epithelium is shown in Fig.3. Analysis of the hybridization results demonstrated; (1)There were 1545 genes where there was an expression differences of more than 2 times. In the present study we have focused on 185 genes in which the expression differences were more than 5 times. Of these, 86 genes showed >5 fold up regulation (their fluorescence intensity was increased), and 99 genes showed<0.2 down regulation (their fluorescence intensity was decreased). These genes were divided into 16 groups (Table 2).
The double-color fluorescence merged image of ovarian
(2) Comparison of gene expression in ovarian carcinoma versus normal ovarian epithelium showed that there were 38 genes with an expression difference of more than 10 times. Among those, 15 genes were up-regulated (Table 3) and 23 genes were down-regulated (Table 4).
(3)In the present study, there were 12 genes that displayed a difference in expression which were similar to our previous research work on the highly metastatic human ovarian cancer line HO-8910PM and its parent cell line (HO-8910)[3]. In those studies the cells were compared to normal ovarian epithelium. Of these 12 genes, 10 genes were up-regulated or down-regulated in a similar manner. Only 1 gene (U72936) was up-regulated in human ovarian cancer, but down-regulated in both cell lines. The NM-016039 gene was down-regulated in the human ovarian carcinoma tissues, but it was up-regulated in both cell lines (Table 5).
DISCUSSION
Progression of early-stage ovarian cancer is asymptomatic, so most ovarian cancer patients are diagnosed at an advanced stage with a poor prognosis. Ovarian carcinoma has the highest mortality among gynecological carcinoma, resulting from a series of molecular changes, caused by abnormal expression of tumor-associated genes or the inactivation of some tumor suppressor genes. It is therefore necessary to study the changes in thousands of genes, which make up the normal genome, as opposed to studying just one or several genes. Investigating gene expression levels may help us to understand the interrelation between genotype and phenotype. Gene chip assays may detect gene expression differences in various specimens by parallel analysis on a large scale. The greatest advantage of this technique is that it improves traditional experiments in which only a single or several gene-expression differences can be observed in one procedure. Therefore, more and more cDNA microarrays are being used to study gene expression. Detection of the changes in ovarian cancer gene-expression patterns may reveal the cause of pathologic changes, but also may provide new targets for diagnosis and opportunities for prevention, and for further providing a new science base for genetic diagnosis and treatment.
In the present study the gene chip technique was employed to analyze the differences of gene-expression patterns in human ovarian carcinoma compared with normal ovarian epithelium. The results showed that: There were a total of 185 genes in which expression levels differed more than 5 times in comparing human ovarian cancer with normal ovarian epithelium. Among these genes, 86 were up-regulated > 5 times (fluorescence intensity was increased) and 99 down-regulated < 0.2 (fluorescence intensity was decreased). These results suggest that these genes might be related to the occurrence and development of ovarian carcinoma.
Glutathione S-transferase (GSTs) is part of an isoenzyme family which possesses multi-functions in physiology, and takes part in a 2-phase detoxification reaction for a large number of a potential chemical, mutagens and carcinogens. Reduced glutathione combines with lipophilic compounds to decrease their toxicity, increase their water solubility and clearance, thus avoiding interaction with macromolecules. Huang et al.[4] reported that the GST gene was down-regulated in lung carcinoma. Results from Xu et al.[3] indicated that in HO-8910PM cells (0.3) down regulation was lower than in its parent cell line HO-8910 (0.45). They found the same results [5] by using flow cytometry. Expression in highly metastatic HO-8910PM (67.1%) was lower than its parent cell line HO-8910 (73.6%). Wang et al. [6] reported that the GST gene was down-regulated (0.17) in ovarian carcinoma. Our results also showed that GST was down-regulated (0.37), but on the other hand, Sakamoto et al. [7] indicated that GST was up-regulated in an ovarian carcinoma cell line which was resistant to chemotherapeutics.
Protein tyrosine phosphatase (PTPase) interacts with the action of numerous hormones, (including epidermal growth factor, insulin and insulin-like growth factor 1 and so on) which mediate signal transduction, energy metabolism, cell proliferation and promoting the expression of the MHC-1 antigen. So decreased activity of PTPase will decrease antigen expression on the cell surface, and may result in malignant cells escaping from immune surveillance. Wang et al. [6] reported that PTPase was down-regulated in ovarian carcinoma (0.15), and in the 2 ovarian cancer cell lines we have studied [3] it was also down-regulated to varying degrees (0.64 and 0.27). However, Sawiris et al.[8] showed that PTPase was up-regulated in ovarian carcinoma. In our research the PTPase receptor type K was up-regulated (2.386), the non-receptor type 6 was also up-regulated (2.441) and the non-receptor type 3 was up-regulated as well (6.970).
Vimentin is part of the interstitial space and in mesoderm is part of the cytoskeleton. It has a very important action for cell migration, signal transduction in cell division, energy transport, metabolic control and the formation of fibroid cells. Some experiments have indicated that the vimentin gene was inhibited during cellular differentiation similar to the oncogene c-myc. Moch et al. [9] reported that vimentin-gene expression was significantly different between a renal-cancer cell line and normal kidney tissues. Using immunohistochemistry, they reported that the vimentin level was related to patient prognosis and showed a negative correlation in 532 cases of kidney carcinoma. Xu et al.[3] indicated that vimentin was down-regulated both in the highly and lowly metastatic ovarian carcinoma cell lines (0.087,0.122,0.131,0.161 and 0.093,0.126,0.144,0.153). In a report by Wang et al. [6] the vimentin gene was down-regulated in ovarian carcinoma (0.13), similar to our result(0.13).
The zinc finger protein is part of a large super-family of proteins that specificly bind to DNA and RNA. Each member of this family has an important role in embryogenesis. Zinc finger proteins possess many functions, such as activating transcription, apoptosis control, DNA recognition, packaging RNA, protein packing and binding lipids etc. If the zinc finger protein is abnormal, there are more chances for neoplasia. Some authors view the C4HC3 zinc finger structure as a leukemia-associated protein. The zinc finger protein was reported to be down-regulated in ovarian carcinoma (0.25)[6] and in the paper by Xu et al.[3] they showed that the zinc finger proteins were all down-regulated in both highly and lowly metastatic ovarian carcinoma cell lines (0.16 and 0.24). In our study, there were 19 cases in which zinc finger proteins were abnormal. Among those 8 cases were up-regulated (2.03, 2.09, 2.122, 2.126, 2.174, 2.282, 2.303, 2.45), and 11 cases were down-regulated (0.039,0.051,0.275,0.32,0.331,0.4,0.408,0.415, 0.463,0.486,0.489).
Serine/threonine kinase 15 (STK15 ) is located on the No. 20 chromosome ql3.2 area. It has a regulatory action on the nucleus and other organelles in the cell. Schraml et al. [10] found that the STK15 DNA copy number was frequently amplified in the chromosome 20q area in about 55% of human ovarian carcinomas. Miao et al.[11] reported that STK15 was over expressed in the progressive stage of esophageal carcinoma. In the present study we found the STK15 was up-regulated (4.82 )in human ovarian carcinoma.
Matrix metalloproteinase (MMPs) is a member of a group of important (about 25 known members) cell matrix degradation enzymes. They act to degrade various fractions of the intracellular matrix, and are key enzymes for tumor metastasis and invasion. Sood et al.[12] found that the matrix metalloproteinase was important for forming the blood vascular-like network in cell cultures of ovarian cancer and Stadlmann et al. [13] reported that MMP-8 expression levels correlated with their tumor grade (P<0.01), tumor stage (P <0.01), and a poor prognosis (P <0.05). Li et al.[14] reported that MMP1, MMP2, MMP12 were all highly expressed in the tissues of gastric carcinoma. In addition, when there were lymph-node metastases, MMP was highly expressed. Wang et al.[6] reported that MMP7 was up-regulated (8.58) in ovarian carcinoma. In our research we found that MMP2 (2.155) and MMP16 (4.894) were up-regulated, but MMP1 (0.496) was down-regulated in human ovarian carcinoma.
The cyclin family includes 8 main members, namely: Cyclin A through H. Cyclins play an important role in the G2/M transition in cell division and are a associated with apoptosis, the vascular growth factor and its receptor, telomerase etc. Sawiris et al.[8] found that cyclin-dependent kinase 6 was up-regulated and cyclin-dependent kinase 7 and cyclin H were down-regulated in human ovarian carcinomas. Sakamoto et al.[7] reported that the cyclin-dependent kinase10 was down-regulated in a chemotherapy-resistant ovarian carcinoma cell line and Xu et al.[3] showed that the CCNB1 gene was significantly up-regulated in both highly and lowly metastatic ovarian carcinoma cell lines (6.566, 3.508 and 6.955, 4.203), whereas the CCN1 gene was down-regulated (0.148 and 0.273). A high expression of cyclin D1 has been shown to be associated with invasion, biologic behavior and poor prognosis in ovarian carcinoma.[15] Lambros et al.[16] reported using array-comparative genomic hybridization, that cyclin DI was amplified in the chromosome 1 lql3 region. Our research showed that 11 cyclins were abnormally expressed, 5 cyclins were up-regulated: P15RS (2.059), DMTF (2.086), CCNF (2.020), HMOX2 (2.292), CCNB1 (2.879), and 6 cyclins were down-regulated:CCND2 (0.316), CCNH (0.352), P57, kip2 (0.355), P27, kipl (0.410), CCN1 (0.433), P21, cipl (0.460).
The CD9 antigen is a 24~27kD glycoprotein, encoded by a gene mapping to the 12 pl3.3 area that is part of the cell surface protein of the TM4 super-family. Wang et al. [6] reported that the CD9 antigen was up-regulated (4.34)in human ovarian carcinoma and Sakamoto et al.[7] also found the CD9 antigen was up-regulated in a chemotherapy-resistant ovarian carcinoma cell line. Chang et al.[17] reported that the CD9 antigen was up-regulated in ovarian carcinoma. In our research of human ovarian carcinoma the CD9 antigen was also up-regulated (3.319). However, Furuya et al.[18] reported that the CD9 was down-regulated which would indicate that this ovarian carcinoma will easily disseminate and would result in a poor prognosis.
The application of the gene chip technique is a revolutionary research method that will have a great impact on life-science research. Our studies illustrate that analysis of gene expression patterns by the gene chip may provide a new direction for genetic diagnosis, therapy and prevention of human ovarian carcinoma.
Footnotes
This work was supported by grants from the National Natural Science Foundation (No.30471819).
- Received January 6, 2005.
- Accepted April 13, 2005.
- Copyright © 2005 by Tianjin Medical University Cancer Institute & Hospital and Springer
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.