Characterization of the MicroRNA Expression Profile of Cervical Squamous Cell Carcinoma Metastases

Cervical cancer is one of the most commonly diagnosed gynecological cancers. According to the IARC GLOBOCAN worldwide cancer incidence and mortality database (http://globocan.iarc.fr/fact sheets/ cancers/cervix.asp), 530 000 new cases were reported in 2008. Developing countries accounted for more than 85% of the global burden, with China contributing around 75 000 of those cases. Detection of cancer cells in the lymph node (LN) is widely-recognized as one of the most important indicators of cervical cancer prognosis and guides the design of appropriate treatment strategies. Yet, the underlying molecular mechanisms that mediate the metastatic process leading to positive LN status in cervical cancer patients have yet to be fully elucidated. The recently identified regulatory mechanism of microRNA (miRNA)-mediated post-transcriptional gene expression has emerged as an intriguing possible regulator of cervical cancer metastasis. MiRNAs are


Introduction
Cervical cancer is one of the most commonly diagnosed gynecological cancers. According to the IARC GLOBOCAN worldwide cancer incidence and mortality database (http://globocan.iarc.fr/fact sheets/ cancers/cervix.asp), 530 000 new cases were reported in 2008. Developing countries accounted for more than 85% of the global burden, with China contributing around 75 000 of those cases. Detection of cancer cells in the lymph node (LN) is widely-recognized as one of the most important indicators of cervical cancer prognosis and guides the design of appropriate treatment strategies. Yet, the underlying molecular mechanisms that mediate the metastatic process leading to positive LN status in cervical cancer patients have yet to be fully elucidated.
The recently identified regulatory mechanism of microRNA (miRNA)-mediated post-transcriptional gene expression has emerged as an intriguing possible regulator of cervical cancer metastasis. MiRNAs are

Characterization of the MicroRNA Expression Profile of Cervical Squamous Cell Carcinoma Metastases
Hui Ding 1 , Yi-Lin Wu 1 *, Ying-Xia Wang 2 , Fu-Fan Zhu 1 small noncoding RNAs of 18-25 nucleotides in length, which bind to the 3'-untranslated region (UTR) of target mRNAs. This sequence-specific binding leads to target mRNA degradation or repression of its translation, consequently decreasing encoded protein levels (Lee et al., 1993;Bagga et al., 2005). In fact, miRNAs have already been implicated as regulators of many physiological and pathological processes, including development, proliferation, differentiation, apoptosis, and tumor progression and metastasis (Bartel et al., 2004;He et al., 2004;Esquela-Kerscher et al., 2006;Negrini et al., 2008;Tavazoie et al., 2008;Bartels et al., 2009;Garzon et al., 2009). However, the role of miRNAs in cervical cancer metastasis and the precise mechanism of its action in these tumor types remain largely unexplored.
Therefore, this study was designed to identify the specific miRNAs that are associated with cervical squamous cancer metastasis to the lymph node. A panel of miRNAs discovered by initial microarray screening of tumors from LN-positive and LN-negative cervical cancer patients was confirmed by quantitative reverse transcription polymerase chain reaction (qRT-PCR). The predicted target genes of these significantly differentially expressed miRNAs were then investigated to determine the potential carcinogenic mechanisms.

Specimens
We collected tumor samples from 10 cervical squamous carcinoma patients who underwent hysterectomy with LN dissection at the Department of Gynecology, Second Xiangya Hospital of Central-South University between July and October of 2010. All patients provided written informed consent, and the study was approved by the Institutional Ethics Committee. None of the patients had received preoperative chemotherapy or radiotherapy. Four of the patients were confirmed as pelvic lymph node (PLN)-positive, indicating metastasis. Of the remaining 6 patients that were confirmed as being PLN-negative (without metastasis), and were selected for use in the comparative analysis.
The total 10 patients used in the study ranged in age from 29 to 65 years old (median: 41.5 years). The patients were staged according to the Federation of Gynecology and Obstetrics (FIGO) 2000 staging system, by which five were classified as Ib1, one as Ib2, and four as IIa. The patients' histological findings (hematoxylin and eosin (H&E) stain) indicated that nine were of middle-grade and one was of low-grade. The PLN-negative group and PLN-positive group had no significant difference in stage, degree of differentiation, age, tumor size, or HPV infected status.
One-half of each tissue sample collected was divided into two tubes, frozen in liquid nitrogen immediately after surgical removal, and stored at -80°C.

MiRNA microarray analysis
One of the two frozen tissue samples from each patient was submitted to KangChen-Biotech (Shanghai, China) for miRNA microarray analysis. At KangChen-Biotech, the total RNA (including miRNA) was isolated from individual samples by using the Trizol reagent (Invitrogen, Carlsbad, CA, USA), following the manufacturer's instructions. The purity and quantity of the isolated RNAs were assessed by 1% formaldehyde-agarose gel electrophoresis and spectrophotometry (Bio-Rad, Hercules, CA, USA). The RNAs from PLN-positive group and PLN-negative group have been respectively mixed and put into microarray analysis. After further processing, the isolated miRNAs were labeled with Hy3 by using the miRCURY Hy3/Hy5 Power labeling kit (Exiqon, Vedbaek, Denmark). The labeled miRNAs were then hybridized to individual miRCURYTM LNA microRNA Arrays (v 14.0; Exiqon), as described previously (Castoldi et al., 2006). Microarray images were acquired by an Axon Genepix 4000B microarray scanner (Axon Instruments, Union City, CA, USA), and processed and analyzed by the accompanying Axon Genepix Pro 6.0 software.
MiRNAs were considered as differentially expressed between cervical cancer tissues with lymph node metastasis and without lymph node metastasis if the difference reached at least 4-fold.

Bioinformatic prediction of miRNAs target genes
The differentially expressed miRNAs were investigated to predict their potential target genes by using the iterative algorithms of TargetScan (http://www.targetscan.org/), PicTar (http://pictar.mdc-berlin.de), and MiRanda (http:// mirdb.org/miRDB/). Target genes were selected based on prediction by at least two of the programs. The identified genes were further selected for those with functions previously associated to tumor processes.

QRT-PCR validation
Six of the miRNAs that had significantly different expression levels according to the microarray analysis and had tumor-associated target genes were selected for verification by qRT-PCR. The qRT-PCR assays were performed on a QPCR System (Applied Biosystems, Inc., Foster City, CA, USA) using RNA that had been Trizolextracted in our lab from the second frozen tissue sample. The isolated RNA was quantified by UV-spectroscopy and used as template to synthesize cDNA with the One Step PrimeScript miRNA cDNA Synthesis kit (TaKaRa, Dalian, China).
The qPCR assay was carried out with reagents from the SYBR Green Real-Time PCR kit (TaKaRa), and the reaction mix was composed of 10 μL SYBR Premix Ex Taq II, 0.8 μL PCR forward primer (Table 1), 0.8 μL Uni-miR qPCR primer , 2 μL cDNA, 0.4 μL ROX reference dye, and 6 μL deionized H2O. Reactions were incubated in a 48-well optical plate at 95˚C for 30 s, followed by 40 cycles of 95˚C for 5 s, and 60˚C for 34 s.
The threshold cycle (Ct) was determined by using the default threshold settings. All experiments were carried out in triplicate and repeated three times. U6 was used as an endogenous control, to which miRNA expression was normalized. Normalized data were analyzed using the 2 −ΔΔCT method (Livak et al., 2001). The results were then calculated as fold-change of each miRNA detected in the cervical cancer tissues with lymph node metastasis relative to that in the tissues without lymph node metastasis.

Statistical analysis
Statistical data analyses were performed with the rank sum test, using SPSS 17.0 statistical software (SPSS, Inc., Chicago, IL, USA), and p-values <0.05 were considered significant.

Significant differentially expressed miRNAs
Of the 1223 miRNA genes analyzed by miRNA microarray, 39 were identified as being differentially expressed by more than 4-fold. Of those, 22 were up-regulated and 17 were down-regulated in cervical carcinomas with LN metastasis (Table 2).

Predicted target genes of the differentially expressed miRNAs in metastatic cervical cancer
By using a combined approach of analyzing the miRNA sequence by bioinformatic neural nets and searching previously published studies regarding the particular genes' functions, we were able to predict tumorassociated gene targets of the significantly differentially expressed 39 miRNAs. Twenty-seven of those miRNAs (69%) were found to have tumor-associated putative target genes, including 17 of the up-regulated miRNAs and 10 of the down-regulated miRNAs (data not shown). Results from subsequent verification analysis, narrowed the list to six of those 27 miRNAs, which are summarized in Table  3.

Validation of microarray data by qRT-PCR analysis
The differentially expressed miRNAs detected on the miRNA microarray were found to have tumor-associated putative target genes, and six of these were verified by quantitative real-time-PCR (Table 4). We select these six miRNAs to testify whether they do express differently is that, of the research on target proteins of these six miRNAs, many have proved that these proteins impact on the tumor invasion and metastasis. MiR-126, miR-96, and miR-144, which were detected on microarray as being down-regulated in metastatic cervical cancer, also showed significantly lower expression than in non-metastatic cervical cancer by qRT-PCR. Likewise, the miR-657, miR-490-5p, and miR-323-3p, which were detected as up-regulated by the microarray experiments, were found by qRT-PCR to be significantly up-regulated.

Discussion
Although several miRNAs have been found to function as either tumor metastasis promoters or suppressors (Asangani et al., 2008;Korpal et al., 2008;Li et al., 2009;Tsai et al., 2009), the exact role that miRNAs play in cervical cancer metastasis is only beginning to be uncovered. Here, we describe our microarraybased investigation into the metastatic-related miRNA expression profile of cervical squamous cancer. Among these differently expressed 39 miRNAs, 27 had predicted target genes whose functions are known to be closely associated with tumor pathogenesis.
Quantitative RT-PCR experiments verified the differential expression microarray findings for the downregulated miR-126, miR-96, and, miR-144, and the upregulated miR-657, miR-490-5p, and miR-323-3p. Thus, our findings indicate that miRNA might play an important role in the metastasis of cervical squamous cancers.
In our study, the verified pattern of miR-126 downregulation in metastatic cervical cancer specimens suggested a possible tumor metastatic suppressor function for this particular miR-126 in cervical cancer. Previous research of colon cancer (Li et al., 2013), cervical cancer (Yu et al., 2013) and gastric cancers (Feng et al., 2010) also observed down-regulated expression of miR-126. Moreover, it has been shown that retrovirus-mediated restoration of miR-126 expression in murine primary breast cancer cells caused the suppression of overall tumor growth and metastasis to the bone and lung (Sohail et al., 2008). Our data from cervical cancer specimens are consistent with these previous studies, and support the theory that miRNAs, particularly miR-126, contribute to the underlying molecular mechanisms of tumor metastasis.
Since we were interested in gaining a more detailed understanding of the underlying molecular mechanisms of metastasis, we investigated the predicted target genes of the significantly deregulated miRNAs: miR-126, miR-96, miR-144, miR-657, miR-490-5p, and miR-323-3p. Not surprisingly, our initial approach using bioinformatic neural nets to predict target genes based on corresponding sequence pairing yielded hundreds of genes for each miRNA (Li et al., 2007). It is likely that miRNAs may specifically target a variety of genes according to the precise physiological context or tissue-specific intracellular milieu. In fact, it has been reported that almost one-third of the protein-coding genes in humans are susceptible to miRNA regulation (Lewis et al., 2005) and, as a consequence, identification of one or several important targets may provide useful insights into the context-and tissue-specific functions of certain miRNAs. We, therefore, focused our research efforts on the miRNAs with target genes that had been previously related to tumor pathogenesis in the peer-reviewed literature. Thus, six MiRNAs with predicted tumor-associated target genes were identified. The target genes of these miRNAs encoded proteins that are known to be involved in cell adhesion, cytoskeletal remodeling, cell proliferation, cell migration, and apoptosis. These findings suggest that a panel of miRNAs may regulate multiple and various steps of the metastasis cascade by targeting metastasisassociated genes. In support of this theory, a previous study by Valastyan et al. had shown that alteration of miR-31 modulated multiple target genes and affected multiple phenotypes, such as motility, invasion, and resistance to anoikis (Valastyan et al., 2009). Pacurari M et al. reported the miR-200 family and these potential targets were functionally involved in canonical pathways of immune response, molecular mechanisms of cancer, metastasis signaling, cell-cell communication, proliferation and DNA repair in Ingenuity pathway analysis (Pacurari et al., 2013). Multiple-to-multiple relationships between miRNAs and targets have been reported by transfection experiments involving human gastric cancer cells (Hashimoto et al., 2013). Therefore, we speculate that miRNAs regulation of multiple and various target genes contributes to cervical cancer metastasis.
In conclusion, we used miRNA microarray technology to analyze the different miRNA expression profiles between cervical cancer specimens with lymph node metastasis and without lymph node metastasis, and 39 significantly disregulated miRNAs were identified. The expression levels of miR-126, miR-96, miR-144, miR-657, miR-490-5p, and miR-323-3p were validated by qRT-PCR, and the results agreed with the patterns detected by microarray analysis. Target gene predictions indicated that differentially expressed miRNAs may play an important role in the metastasis of cervical cancer, and that identified specific tumor-associated genes may be involved in the pathogenic process. However, this set of experiments remain some insufficiencies, such as tumor specimens from patients were fewer; 39 miRNAs were identified by miRNA microarray as being differentially expressed by more than 4-fold, we only select these six miRNAs to testify. So we need to expand the sample to estimate the value of miRNAs for predicting lymph node metastasis of cervical squamous cell carcinoma. Further investigative focus on these miRNAs and their functions in the pathogenesis of cervical cancer will be necessary.