4270

Chin Med J 2012;125(23):4270-4276

Original article
Profiling of microRNA-mRNA reveals roles of microRNAs in cervical cancer
MA Ding, ZHANG You-yi, GUO Yan-li, LI Zi-jian and GENG Li Keywords: cervical cancer; microRNA; gene expression; microarray
Background Cervical cancer is one of the most common malignant tumors in women. This study was designed to explore the expression profiles of microRNAs (miRNAs) and mRNAs and the gene regulation network in cervical tumorigenesis and to find candidate molecular markers and key tumorigenic genes in cervical cancer. Methods miRNAs and mRNAs expression microarrays were used to detect the expression of miRNAs and mRNAs in normal and cancer cervical tissues. TargetScan 5.0 database (UK) was used to predict the target genes of the miRNAs, analyze their intersection with differentially expressed mRNAs and negatively correlate the intersection with miRNAs. Bioinformatic approaches were used to analyze functions and pathways of the target genes and establish miRNA-gene network. Results Twenty-nine miRNAs and 2036 mRNAs were differentially expressed in normal and cervical tumor tissues. Among them, 13 miRNAs and 754 mRNAs were up-regulated in cervical tumor tissues and 16 miRNAs and 1282 RNA were down-regulated. The 327 target genes negatively related to miRNAs in the intersection were involved in functions and signal pathways. Down-regulated miRNAs targeted genes and up-regulated miRNAs targeted genes were involved in 415 and 163 functions, respectively, and in 37 and 17 significant pathways, respectively (P <0.05, false discovery rate (FDR) <0.05). We constructed the miRNAs-gene network and found that hsa-miR-15a, hsa-miR-106b and hsa-miR-20b were key nodes in the network. Conclusions The differentially expressed miRNAs and mRNAs in cervical cancer and related miRNA-gene network have been identified. They play important roles in cervical tumorigenesis and are involved in many important biological functions and signal transduction pathways. These findings lay a foundation for research on the molecular mechanism of miRNAs in the pathogenesis of cervical cancer. Chin Med J 2012;125(23):4270-4276

ervical cancer is one of the most common tumors in women. Its incidence ranks first in female reproductive tract tumors and its death rate ranks second in various female malignant tumors. On a global scale, approximately 500 000 new cases of cervical cancer are reported annually and about 230 000 women die of cervical cancer annually.1

MicroRNAs (miRNAs) are single-strand, non-coding 22nt RNA widely presented in eukaryotes. Although miRNAs are most often found to completely or incompletely complement with the 3 untranslated region (UTR) of their target molecules to degrade mRNAs or repress mRNAs translation, they can also bind to regions within coding regions and 5UTRs of their target genes.2,3 More than one thousand miRNAs involved in many biological processes such as cell growth, differentiation, proliferation and apoptosis have been identified. Abnormal and absent miRNAs expression has been found in a variety of human tumors after they were initially found in B-cell chronic lymphocytic leukemia (CLL).4 It is believed that imbalance resulted from abnormal inter-regulation between miRNAs and genes could lead to tumor development. Microarray technology has advantages of rapidity, high-throughput, high specificity, and high sensitivity. Therefore, using miRNA microarray

and DNA microarray as research tools to construct miRNA/mRNA interaction network, we can improve our understanding of cervical cancer pathogenesis. Although miRNA microarray and mRNA microarray can detect and compare the expression profiles of miRNAs and genes, currently only few studies have jointly used miRNA and mRNA microarray methods, especially in the field of cervical cancer. Rao et al5 utilized miRNA microarray to detect miRNAs expression level in 13 cases of cervical tumors and their adjacent tissues and found a large
DOI: 10.3760/cma.j.issn.0366-6999.2012.23.020 Department of Gynecology and Obstetrics, Peking University Third Hospital, Beijing 100191, China (Ma D, Guo YL and Geng L) Institute of Vascular Medicine, Peking University Third Hospital and Key Laboratory of Cardiovascular Molecular Biology and Regulatory Peptide, Ministry of Health, Beijing 100191, China (Zhang YY and Li ZJ) Correspondence to: Dr. GENG Li, Department of Gynecology and Obstetrics, Peking University Third Hospital, Beijing 100191, China (Email: gengli57@163.com); LI Zi-jian, Institute of Vascular Medicine, Peking University Third Hospital and Key Laboratory of Cardiovascular Molecular Biology and Regulatory Peptide, Ministry of Health, Beijing 100191, China (Tel: 86-10-82265519. Email: lzjgy1995@ 163.com) This work was supported by grants from the National Natural Science Foundation of China (No. 11072006, No. 81030001 and No. 81070078).

Chinese Medical Journal 2012;125(23):4270-4276

4271

number of differentially expressed miRNAs in cervical tumor tissues. Hudelist et al6 utilized the microarray method to examine differentially expressed genes between cervical cancer cells and normal cervical epithelium and found abnormal expression of a large number of oncogenes, tumor suppressor genes, growth factors, chemokines, integrin and transforming growth factor, beta (TGF) family members. The formation of cervical cancer is related to oncogene activation, tumor suppressor gene inactivation, and alterations of many gene clusters related to cell cycle, apoptosis, adhesion, immunity, and signal transduction.4,6 Therefore, comprehensive and systematic researches on the alteration of miRNAs expression profiles and their target genes in cervical cancer will provide important clues for revealing the molecular mechanisms of cervical cancer occurrence, development, diagnosis and treatment. We studied clinical cervical specimens, which can reflect the features of miRNAs and mRNAs expression in cervical cancer more objectively and accurately than animal models and cell lines. We utilized both miRNA and mRNA microarray methods combined with bioinformatics to predict the relationship between miRNAs and their target genes, to analyze miRNAs regulation mechanisms, biological functions of target genes and signal transduction pathways, and to establish the miRNA-gene network in the pathogenesis of cervical cancer. METHODS Clinical specimens Eight patients who were admitted to the Department of Obstetrics and Gynecology, Peking University Third Hospital from January 2008 to April 2010 and diagnosed with IB-IIB cervical cancer based on the classification criteria of the International Federation of Gynecology and Obstetrics and eight patients who had benign gynecological diseases such as uterine fibroids were enrolled in the study. Patients with cervical cancer who accepted radical hysterectomy and pelvic lymphadenectomy were included in the experimental group; patients with benign gynecological diseases who accepted hysterectomy were included in the control group. All patients provided written informed consent before surgery. Cervical tumor tissues and normal cervical tissues were intra-operatively obtained based on the sampling requirement for microarray and immediately frozen in liquid nitrogen except for a part which was HE stained and examined by two gynecological pathologists to confirm the diagnosis. The preoperative cervical secretions were collected and used to detect human

papillomavirus (HPV) infection using hybrid capture generation II technology. All patients did not receive pre-operational radiation, chemotherapy and immunotherapy. Table 1 lists the clinical and pathological data of all patients. The study was approved by the Ethics Committee of Peking University (approval No. IRB00001052-06058). Total RNA extraction and miRNAs and genes expression microarray detection Total RNA extraction Total RNA was extracted from two sets of tissue samples using Trizol reagent and purified using phenol/chloroform extraction and precipitation/washing according to instructions provided by the manufacturer. The obtained RNA was dissolved in RNase-free water and stored at 80C. miRNA microarray A measured amount of (100 ng) RNA was labeled using the Agilents MicroRNA Complete Labeling and Hyb Kit (p/n 51900456, USA) and directly used in microarrays (Agilent Human miRNA oligonucleotide microarray data were from the Sanger miRBase V12.0 version). The microarrays were carefully washed and scanned using a microarray scanner (Agilent p/n G2565BA, USA). Data were obtained using Agilent G4450AA Feature Extraction software 10.0. Gene expression microarray The expression of the mRNAs was detected using Agilent human whole genome microarray method (Agilent G4112F Design ID 014850, 444 K format). RNA was fluorescently labeled using the Agilent low RNA input fluorescent linear amplification kit, hybrided following the instructions of the gene expression hybridization kit (Agilent p/n 51885242), and scanned using a microarray scanner (Agilent p/n G2565BA) and version A7.0 Agilent scan control software. Data were extracted and standardized using Agilent G4450AA feature extraction software 10.0. Bioinformatic analysis differentially expressed miRNAs and mRNAs and construct miRNA-gene network Screening of differentially expressed miRNAs and mRNAs Random variance model (RVM) t-test was used to filter the differentially expressed miRNAs and mRNAs between the cervical tumor and control groups since it can effectively increase the freedom degree of small sized samples. The false discovery rate (FDR) was calculated to correct the P-value. miRNAs and mRNAs with P <0.05 and FDR <0.05 were considered as differentially

Table 1. Clinical and pathological characteristics of the patients

Pathological evaluation Squamous cervical cancer (n= 8) Normal (n=8) Age (years) 41.753.25 45.133.43 HPV positive 8 0 Lymph node metastasis 1 Differentiation (low, medium, high) FIGO stage (IB1, IB2, IIA, IIB) 3, 4, 1 2, 2, 3, 1

4272

Chin Med J 2012;125(23):4270-4276

expressed.7 Prediction and analysis of differentially expressed miRNAs target mRNAs Prediction of differentially expressed miRNAs target genes was performed using TargetScan5.0 database (UK). The intersection of the predicted target genes and the differentially expressed mRNAs was collected. Based on the negative regulation relationship between miRNAs and genes, a negatively related target genes intersection was obtained and used to analyze their intrinsic correlations. Gene ontology (GO) analysis The main function of the negatively correlated target intersection genes was analyzed using GO analysis, the key functional classification of NCBI. Generally, Fishers exact test and 2 test were used to classify the GO category, and FDR was calculated to correct the P-value. The smaller the FDR, the smaller the error in judging the P-value. FDR was defined as FDR=1Nk/T, where Nk refers to cases wherein the P-value of Fishers test is less than that of 2 test. Only GOs with P <0.05 and FDR<0.05 were selected.8 Pathway analysis Similarly, pathway analysis was used to determine significant pathways negatively correlated with the intersection of target genes according to the databases such as Kyoto encyclopedia of genes and genomes (KEGG, Japan), Biocarta (Germany) and Reatome (USA). Again, Fishers exact test and 2 test were used to select the significant pathways, and the threshold of significance was defined by P-value and FDR. Only pathways that had P <0.05 and FDR <0.05 were selected.9,10 miRNA function analysis The main function of the differentially expressed miRNAs was acquired according to the relationships between the significant GOs and the differential expression target genes and the relationships between miRNAs and genes.11 miRNA-gene network miRNA-gene network was constructed based on the targeted regulatory relationships between miRNAs and their target genes. The status of miRNAs and genes in the network was evaluated using graph theory method based on their degree in the network. The number of target genes regulated by an miRNA was defined as the degree of the miRNA. The greater the number, the greater the degree of the miRNA. Verification of the differentially expressed miRNAs using real-time RT-PCR Real-time reverse transcription-polymerase chain reaction (RT-PCR) was used to verify the differential expression of hsa-miR-15a, hsa-miR-106b and hsa-miR-20b which were key nodes of the miRNA-gene network. Total RNA

was extracted from cervical cancer samples (n=4) and control samples (n=4). cDNA was synthesized using Quant reverse transcriptase (TIANGEN, China), dNTP and miR-RT primer and 100 ng total RNA. PCR amplifications were performed in triplicate for each sample. Each real-time RT-PCR reaction (in 25 l) contained 2SYBR Green Realtime PCR Master Mix, 20 mol/L primers, and 1 l of template cDNA. The cycling conditions consisted of an initial 3 minutes at 95C, followed by 40 cycles of 15 seconds at 95C, 20 seconds at 60C and 30 seconds at 70C. The comparisons of two groups were performed with t-test (P <0.05). RESULTS Differentially expressed miRNAs and mRNAs in cervical cancer Screening of differentially expressed mRNAs helps us to identify the molecular markers and the key candidate genes in cervical cancer. miRNAs can inhibit mRNAs translation or directly degrade mRNAs, which, as one of the regulatory factors involved in gene expression, promotes the development of cervical cancer. This study utilized miRNA microarray to detect miRNAs expression in four cases of cervical carcinoma tissues and four cases of normal cervical tissues, and mRNA microarray to detect mRNA levels in eight cases of cervical cancer (including the above four cases) and eight cases of normal cervical tissue (including the above four cases). We found 29 differentially expressed miRNAs (13 up-regulated and 16 down-regulated) and 2036 differentially expressed mRNAs (754 up-regulated and 1282 down-regulated, P <0.05, FDR <0.05). Figure 1 shows the scatter plot of the results and Figure 2 shows the cluster diagram. Figure 1 indicates the expression specificity in cervical tissues and the reliability of microarray data. Figure 2 visually displays the miRNAs and mRNAs with similar expression patterns and contributes to the further study of miRNAs and mRNAs function. These results indicate that there are a large number of differentially expressed miRNAs and mRNAs in cervical cancer, which together promote the occurrence and development of cervical cancer.

Figure 1. Volcano plot of the differentially expressed miRNAs and mRNAs in the cervical cancer group. A and B show significantly expressed miRNAs and mRNAs in cervical cancer tissues, respectively. The horizontal axis represents the fold change of miRNAs and mRNAs between cervical cancer tissues and normal tissues. The vertical axis represents the P-value of the t-test for the differences between samples.

Chinese Medical Journal 2012;125(23):4270-4276

4273

Figure 2. Cluster analysis of significantly expressed human miRNAs and mRNAs in cervical cancer tissues. 2A and 2B show 29 miRNAs and 2036 mRNAs significantly expressed in cervical cancer tissues, respectively. Red icons represent the expression of miRNAs and mRNAs increased in cervical cancer tissues and green icons represent decreased expression. Each row represents a single miRNA or mRNA and each column a sample. Figure 3. The number of target genes negatively related to miRNAs in the intersection. The left circle represents 2036 differentially expressed genes and the right represents 5117 predicted target genes of the 29 differentially expressed miRNAs in cervical cancer tissues. The intersection of the two circles represents 485 differential genes and 327 target genes negatively related to miRNAs in the intersection.

Prediction of differentially expressed miRNAs and their target genes Since there are many correlations between miRNAs and genes, the methods used to accurately identify their relationships are complex. Currently only a small number of target genes have been identified using bioinformatic methods. In this study, 5117 predicted target genes of the 29 differentially expressed miRNAs were identified using TargetScan 5.0 database. The intersection of the 5117 genes and the 2036 differentially expressed mRNAs had 485 genes after excluding those not regulated by differentially expressed miRNAs. Finally, application of the negative correlation principle between miRNAs and their target genes further obtained 327 target genes, as shown in Figure 3. These results suggest using the intersection of the predicted target genes and the detected differential mRNAs and the negative correlation analysis could reduce the false positive rate in target genes prediction and improve the accuracy of candidate genes in predicting cervical cancer. GO analysis functional analysis of the differential miRNAs regulated target genes To investigate the function of the target genes regulated by the differential miRNAs, we performed GO analysis of the 327 negatively related genes in the intersection and found that the up-regulated miRNAs were involved in 415 gene functions and down-regulated miRNAs were involved in 163 gene functions (P <0.05, FDR <0.05). Among the significantly up-regulated functions are cell proliferation, cell cycle, DNA dependent transcriptional regulation, mitosis, co-stimulation of T cells, white blood cells migration, apoptosis and inflammatory responses; cell proliferation and cell cycle changes were more prominent, indicating that abnormal cell cycle regulation was a main factor for cervical cancer. Taken together, these results indicate that differential miRNAs target genes were involved in a large number of biological functions, promoting cervical cancer development.

Signal transduction pathway analysis of differential miRNAs target genes Significant pathway analysis of the obtained 327 negatively related intersections of the target genes was performed using methods similar to the above-mentioned gene function analysis. We found that the up-regulated miRNAs target genes were involved in 37 signal pathways and the down-regulated miRNAs target genes were involved in 17 signal pathways (P <0.05, FDR <0.05). Calcium signaling pathways, cytokine-cytokine receptor interaction, neuroactive ligand-receptor interactions, tumor signaling pathways and MAPK pathways were significantly up-regulated. These results were similar to that of the GO analysis, indicating that the differential miRNAs target genes were involved in many cellular pathways, promoting cervical cancer occurrence and development. Regulatory networks between miRNAs and target genes miRNAs and target genes in cervical cancer formed a complex relation network. Figure 4 shows the miRNA-gene network constructed using negative regulation relations between miRNAs and the target genes in the intersection. hsa-miR-15a, hsa-miR-106b and hsa-miR-20b, which regulated 23, 21 and 21 target genes, respectively, were key nodes in the network. These results suggested an intricate regulatory relation network between miRNAs and their target genes. In addition, the miRNA-gene network system could exhibit the quantitive inhibition of miRNAs on mRNAs, lower false-positive error in sequence analysis and predict miRNAs and their target genes which play key roles in the network. Up-regulated hsa-miR-106b and hsa-miR-20b in cervical cancer Hsa-miR-15a, hsa-miR-106b and hsa-miR-20b were key nodes in the miRNA-gene network; so we confirmed the expression of the three miRNAs by real-time RT-PCR.

4274

Chin Med J 2012;125(23):4270-4276 Figure 4. MicroRNA-gene network. Blue box nodes represent up-regulated miRNAs in cervical cancer tissues, and red cycle nodes represent down-regulated mRNAs. Edges show the inhibitory effect of miRNAs on mRNAs. Three over-expressed miRNAs (hsa-miR-15a, hsa-miR-106b and hsa-miR-20b) show the most targeted mRNAs (degree 23, 21 and 21, respectively) in cervical cancer tissues. Figure 5. Comparisons of the expression of hsa-miR-15a, hsa-miR-20b and hsa-miR-106b in control and cervical cancer tissues by real-time RT-PCR. U6 is the reference gene. *P <0.05, compared with controls.

Compared with controls, the expression of hsa-miR-20b and hsa-miR-106b showed a statistical significant change (P <0.05). The expression of hsa-miR-15a was also higher than those of normal controls, but showed no statistical significance (P >0.05). Figure 5 shows the expression of the three miRNAs in normal tissues and in cervical cancer tissues. DISCUSSION Cervical cancer is one of the most common malignant tumors in the female reproductive tract. Screening of differential genes will help find molecular markers and key candidate genes in cervical cancer. As one of the regulatory factors involved in gene expression, miRNAs can inhibit mRNAs translation or directly degrade mRNAs, thus playing key roles in promoting the occurrence and development of cervical cancer. This study adopted high-throughput microarray technology and analyzed miRNAs and whole genome expression profiles in cervical carcinoma. We got 29 miRNAs and 2036 mRNAs differently expressed in cervical cancer. The abundance of hsa-miR-15a, hsa-miR-20b, hsa-miR-21 and hsa-miR-224 is obviously increased in cervical cancer tissues and hsa-let-7c, hsa-miR-143, hsa-miR-199a-5p, hsa-miR-203 and miR-145 is sharply reduced, in agreement with previous studies on cervical cancer.12,13 These differentially expressed miRNAs and mRNAs together promote the occurrence and development of cervical cancer. hsa-miR-19a and hsa-miR-141 are differentially expressed in normal and cervical tumor tissues. Hsa-miR-19a is over-expressed in this study and in many malignant human cancers, including B-cell lymphomas, lung cancers and esophageal squamous cell carcinoma

(ESCC). Liu et al14 found that over-expression of miR-19a down-regulated tumor necrosis factor- (TNF-) and promoted cellular growth in vivo and in vitro, while inhibition of miR-19a by antisense oligonucleotides (ONs) induced apoptosis. Banaudha et al15 reported a direct role of hsa-miRNA-141 in targeting the tumor suppressor gene deleted in liver cancer 1 (DLC-1, a Rho GTPase-activating protein), which is frequently deleted in hepatic cell cancer (HCC) and other solid human tumors. They found that efficient hepatitis C virus (HCV) replication required hsa-miR-141-mediated suppression of DLC-1. Down-regulation of hsa-let-7c in many human cancers indicates its role in tumorigenesis. Nadiminty et al16 demonstrated that hsa-let-7c suppressed androgen receptor expression and activity via regulating Myc expression in prostate cancer cells. Bioinformatic prediction is a rapid screening tool for investigating miRNAs target genes, but it presents a certain degree of false-positive apparently.17 This study utilized the intersection of the differentially expressed mRNAs obtained in microarray assay and the predicted target genes to gradually narrow the range of the candidate genes and then applied the negative correlation principle of miRNAs and their target genes expression at mRNAs level for extraction of the candidate target genes. By doing these we found a total of 327 target genes, making target gene analysis more credible and the prediction of key genes on cervical cancer more accurate. Large numbers of mRNAs involved in cervical carcinogenesis belong to different functions and pathways. Bioinformatics GO analysis of negatively related target genes intersection found that these genes regulate cell proliferation, cell differentiation, cell adhesion, development of nervous system and axon direction, gene transcription, etc. Among these, genes

Chinese Medical Journal 2012;125(23):4270-4276

4275

regulating cell proliferation seem to be the most significant functional ones. Song et al18 found that the occurrence of cervical cancer is the result of loss of control in proliferation, apoptosis and cell cycle and is related to abnormal changes of multi-genes, multi-factors and multi-stages, consistent with our findings. Excessive proliferation, abnormal differentiation, reduced adhesion ability and gene transcription changes are the characteristics of cervical cancer cells.19 The target gene function analysis indicates that miRNAs play important roles in cervical cancer by inhibiting mRNAs transcription and translation of target genes involved in cell development, proliferation, differentiation, apoptosis, metabolism and other biological processes. Similarly, anomalies and disorders of cell signal transduction pathways play a crucial role in the occurrence and development of cervical cancer. Signal transduction pathways related to cytokine-cytokine receptor interactions, neural active ligand-receptor interactions, oncology, proliferation including MAPK, Wnt and Jak-STAT, adhesion and gene transcription regulation are all differentially expressed. Some genes in the above mentioned pathways have been reportedly involved in the occurrence and development of cervical cancer.20-22 Enhanced expression of chemokines CXCL9, CXCL10 and CXCL11 and members of cytokine-cytokine receptor interaction signaling pathways have been reported in breast cancer. These chemotactic factors can bind to their receptor CXCR3 and promote directional tumor metastasis in breast cancer.23 Currently, although their functions in cervical cancer have not been extensively studied, it is speculated that they have a similar carcinogenic mechanism in cervical cancer. The use of high-throughput microarray and bioinformatic methods to comprehensively search for differentially expressed genes related to signaling pathways in cervical cancer will help clarify the mechanisms underlying cervical carcinogenesis. This study utilized the relationships of the combined results of two microarrays with the bioinformatics predicted miRNAs target genes to construct the miRNA-gene network and supplemented the upstream regulatory miRNAs in the traditional mRNAs-interaction single relationship network. The network not only quantifies the number of miRNAs regulated genes and positions the upstream regulatory miRNAs in the entire network core, but also systematically finds the most prominent gene expression regulatory pathways at both miRNAs and genes levels from a biological network perspective. In this network, miRNAs that regulate the largest number of target genes are hsa-miR-15a, hsa-miR-106b and hsa-miR-20b which are pivotal in the network and have the strongest regulatory effects on differential gene expression in cervical cancer. In addition, these miRNAs are the most highly expressed in our and other microarray studies in cervical cancer.24,25 Although the function of the three miRNAs in cervical

carcinogenesis is unclear, studies have found that hsa-miR-106b is up-regulated in human laryngeal cancer cells and other tumors. Inhibition of hsa-miR-106b expression could inhibit RB gene induced G0/G1 arrest, and over-expression of hsa-miR-106b could promote proliferation of laryngeal carcinoma cells.26 In addition, hsa-miR-106b expression is reduced in prostate cancer after radiotherapy; transfection of hsa-miR-106b precursor can reverse radiation-induced P21 activationinduced G2/M arrest, and immediately reduce radiotherapy-induced cell growth inhibition.27 Studies on multiple mouse tumor cell lines found that miR-20b could precisely sense the change of oxygen concentration, promote adaption of tumor cells to hypoxic microenvironment and survive and enhance malignant phenotype of tumor cells by regulating the transcription factor hypoxia-inducible factor 1 (HIF-1) and vascular endothelial growth factor (VEGF).28 hsa-miR-106b can promote tumor cell proliferation by regulating cell cycle related genes. Interaction of miR-20b and HIF-1 can enhance the tolerance of tumor cells to the hypoxic environment. These miRNAs may have a similar cancer-promoting mechanism in cervical cancer. The network predicts that hsa-miR-106b and hsa-miR-20b can promote cell proliferation by suppressing disabled homolog 2 (DAB2) and reduce adhesion capacity of tumor cells by regulating adhesion genes semaphorin 5A (SEMA5A) and matrix metalloproteinase 2 (MMP2). These miRNAs regulate their target genes not only involving in cell cycle and functional classification, but also cell proliferation and adhesion signal transduction pathways. miRNAs and their target genes can form a complex network relationship in cervical cancer. This study reveals the core miRNAs affecting the biological traits and their target genes and helps to find the potential regulatory mechanisms of miRNAs and mRNAs network underlying the development of cervical cancer. Cervical cancer is a multi-factor, multi-step process, where gene expression abnormalities and cellular signal transduction pathway disorder play crucial roles. Combined application of miRNA and mRNA microarray has provided a basis for further investigating new molecular markers and key genes in cervical cancer development and opened a new means to study miRNAs function. miRNAs play important roles in many important biological functions such as cell proliferation and in signal transduction pathways such as calcium signaling pathways. hsa-miR-15a, hsa-miR-106b and hsa-miR-20b might be key molecules in the pathogenesis of cervical cancer. The differentially expressed miRNA-gene network helps us to understand the related molecular mechanisms of cervical cancer.
REFERENCES 1. Arbyn M, Castellsague X, de Sanjose S, Bruni L, Saraiya M, Bray F, et al. Worldwide burden of cervical cancer in 2008. Ann Oncol 2011; 22: 2675-2686.

4276 2. 3. 4. Kim VN, Han J, Siomi MC. Biogenesis of small RNAs in animals. Nat Rev Mol Cell Biol 2009; 10: 126-139. Bartel DP. MicroRNAs: target recognition and regulatory functions. Cell 2009; 136: 215-233. Calin GA, Dumitru CD, Shimizu M, Bichi R, Zupo S, Noch E, et al. Frequent deletions and down-regulation of microRNA genes miR15 and miR16 at 13q14 in chronic lymphocytic leukemia. Proc Natl Acad Sci USA 2002; 99: 15524-15529. Rao Q, Zhou H, Peng Y, Li J, Lin Z. Aberrant microRNA expression in human cervical carcinomas. Med Oncol 2012; 29: 1242-1248. Hudelist G, Czerwenka K, Singer C, Pischinger K, Kubista E, Manavi M. cDNA array analysis of cytobrush-collected normal and malignant cervical epithelial cells: a feasibility study. Cancer Genet Cytogenet 2005; 158: 35-42. Wright GW, Simon RM. A random variance model for detection of differential gene expression in small microarray experiments. Bioinformatics 2003; 19: 2448-2455. Dupuy D, Bertin N, Hidalgo CA, Venkatesan K, Tu D, Lee D, et al. Genome-scale analysis of in vivo spatiotemporal promoter activity in Caenorhabditis elegans. Nat Biotechnol 2007; 25: 663-668. Yi M, Horton JD, Cohen JC, Hobbs HH, Stephens RM. WholePathwayScope: a comprehensive pathway-based analysis tool for high-throughput data. BMC Bioinformatics 2006; 7: 30. Kanehisa M, Goto S, Kawashima S, Okuno Y, Hattori M. The KEGG resource for deciphering the genome. Nucleic Acids Res 2004; 32: D277-D280. Joung JG, Hwang KB, Nam JW, Kim SJ, Zhang BT. Discovery of microRNA-mRNA modules via population-based probabilistic learning. Bioinformatics 2007; 23: 1141-1147. Pereira PM, Marques JP, Soares AR, Carreto L, Santos MA. MicroRNA expression variability in human cervical tissues. PLoS One 2010; 5: e11780. Wang X, Tang S, Le SY, Lu R, Rader JS, Meyers C, et al. Aberrant expression of oncogenic and tumor-suppressive microRNAs in cervical cancer is required for cancer cell growth. PLoS One 2008; 3: e2557. Liu M, Wang Z, Yang S, Zhang W, He S, Hu C, et al. TNF-alpha is a novel target of miR-19a. Int J Oncol 2011; 38: 1013-1022. Banaudha K, Kaliszewski M, Korolnek T, Florea L, Yeung ML, Jeang KT, et al. MicroRNA silencing of tumor suppressor DLC-1 promotes efficient hepatitis C virus replication in primary human hepatocytes. Hepatology 2011; 53: 53-61. Nadiminty N, Tummala R, Lou W, Zhu Y, Zhang J, Chen X, et al. MicroRNA let-7c suppresses androgen receptor

Chin Med J 2012;125(23):4270-4276 expression and activity via regulation of Myc expression in prostate cancer cells. J Biol Chem 2012; 287: 1527-1537. Lindow M. Prediction of targets for microRNAs. Methods Mol Biol 2011; 703: 311-317. Song JY, Lee JK, Lee NW, Jung HH, Kim SH, Lee KW. Microarray analysis of normal cervix, carcinoma in situ, and invasive cervical cancer: identification of candidate genes in pathogenesis of invasion in cervical cancer. Int J Gynecol Cancer 2008; 18: 1051-1059. Wang J, Cai J, Li Z, Hu S, Yu L, Xiao L, et al. Expression and biological function of N-myc down-regulated gene 1 in human cervical cancer. J Huazhong Univ Sci Technolog Med Sci (Chin) 2010; 30: 771-776. Woodworth CD, Michael E, Marker D, Allen S, Smith L, Nees M. Inhibition of the epidermal growth factor receptor increases expression of genes that stimulate inflammation, apoptosis, and cell attachment. Mol Cancer Ther 2005; 4: 650-658. Kulkarni S, Rader JS, Zhang F, Liapis H, Koki AT, Masferrer JL, et al. Cyclooxygenase-2 is overexpressed in human cervical cancer. Clin Cancer Res 2001; 7: 429-434. Uren A, Fallen S, Yuan H, Usubutun A, Kucukali T, Schlegel R, et al. Activation of the canonical Wnt pathway during genital keratinocyte transformation: a model for cervical cancer progression. Cancer Res 2005; 65: 6199-6206. Ma X, Norsworthy K, Kundu N, Rodgers WH, Gimotty PA, Goloubeva O, et al. CXCR3 expression is associated with poor survival in breast cancer and promotes metastasis in a murine model. Mol Cancer Ther 2009; 8: 490-498. Wang X, Tang S, Le SY, Lu R, Rader JS, Meyers C, et al. Aberrant expression of oncogenic and tumor-suppressive microRNAs in cervical cancer is required for cancer cell growth. PLoS One 2008; 3: e2557. Li Y, Wang F, Xu J, Ye F, Shen Y, Zhou J, et al. Progressive miRNA expression profiles in cervical carcinogenesis and identification of HPV-related target genes for miR-29. J Pathol 2011; 224: 484-495. Cai K, Wang Y, Bao X. MiR-106b promotes cell proliferation via targeting RB in laryngeal carcinoma. J Exp Clin Cancer Res 2011; 30: 73. Li B, Shi XB, Nori D, Chao CK, Chen AM, Valicenti R, et al. Down-regulation of microRNA 106b is involved in p21-mediated cell cycle arrest in response to radiation in prostate cancer cells. Prostate 2011; 71: 567-574. Lei Z, Li B, Yang Z, Fang H, Zhang GM, Feng ZH, et al. Regulation of HIF-1alpha and VEGF by miR-20b tunes tumor cells to adapt to the alteration of oxygen concentration. PLoS One 2009; 4: e7629.

17. 18.

5.

19.

6.

20.

7.

21.

8.

22.

9.

23.

10.

24.

11.

25.

12.

13.

26.

27.

14.

15.

28.

16.

(Received June 10, 2012) Edited by CHEN Li-min