Tumor-suppressive microRNA-29a inhibits cancer cell migration and invasion via targeting HSP47 in cervical squamous cell carcinoma

Our recent studies of microRNA (miRNA) expression signatures indicated that microRNA-29a (miR-29a) was significantly downregulated in several types of human cancers, suggesting that miR-29a may be a putative tumor-suppressive miRNA in human cancers. The aim of this study was to investigate the functional significance of miR-29a in cervical squamous cell carcinoma (SCC) and to identify novel miR-29a-regulated cancer pathways and target genes involved in cervical SCC oncogenesis and metastasis. Restoration of miR-29a in cervical cancer cell lines (CaSKi, HeLa, ME180 and Yumoto) revealed that this miRNA significantly inhibited cancer cell migration and invasion. Gene expression data and in silico analysis demonstrated that heat-shock protein 47 (HSP47), a member of the serpin superfamily of serine proteinase inhibitors and a molecular chaperone involved in the maturation of collagen molecules, was a potential target of miR-29a regulation. Luciferase reporter assays showed that miR-29a directly regulated HSP47. Moreover, silencing of the HSP47 gene significantly inhibited cell migration and invasion in cancer cells and the expression of HSP47 was upregulated in cancer tissues and cervical intraepithelial neoplasia (CIN), as demonstrated by immunostaining. Downregulation of miR-29a was a frequent event in cervical SCC and miR-29a acted as a tumor suppressor by directly targeting HSP47. Recognition of tumor-suppressive miRNA-regulated molecular targets provides new insights into the potential mechanisms of cervical SCC oncogenesis and metastasis and suggests novel therapeutic strategies for treatment of this disease.


Introduction
Cervical cancer is the second most common cause of cancer death in women worldwide and ~500,000 new cases of cervical cancer are diagnosed each year, with 280,000 deaths (1). Cervical squamous cell carcinoma (SCC) is the most frequent type of cervical cancer and the most important risk factor for cervical-SCC is persistent human papilloma virus (HPV) infection (2)(3)(4). High-risk HPVs contain oncoproteins, i.e., E6 and E7, which contribute to the oncogenesis of cervical SCC by silencing tumor-suppressive p53 and Rb proteins and several cancer-related genes (5). Therefore, recent research on cervical SCC has focused on E6 and E7 oncoproteins. However, the molecular mechanisms of cervical SCC initiation, development, and metastasis have not yet been fully elucidated.
The discovery of non-coding RNAs in the human genome was an important conceptual breakthrough in the post-genome sequencing era (6). A growing body of evidence indicates that miRNAs are key regulators that contribute to the initiation and development of various types of cancer (7). In cancer pathways, normal regulatory mechanisms are disrupted by altered expression of tumor-suppressive or oncogenic miRNAs. Therefore, identification of differentially ��������� ��RNA� �� a� ������a�� ���� �� �������a����g human oncogenesis.

Materials and methods
Clinical specimens. A total of 18 primary cervical SCC specimens and 11 non-cancer specimens were collected from patients who had undergone surgical treatment at Chiba University Hospital. The samples were processed and stored in liquid nitrogen until RNA extraction. Patient information is summarized in Table I. Our study was approved by the Bioethics Committee of Chiba University; prior written informed consent and approval was given by each patient. HPV status was examined by L1 consensus primers and typespecific real-time PCR primers, as described previously (15).
RNA isolation. Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. RNA concentrations were determined spectrophotometrically. RNA quality was confirmed using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific, USA).
Cell proliferation, migration and invasion assays. Cell proliferation was determined using XTT assays (Roche Applied Science, Tokyo, Japan) according to the manufacturer's instructions. Cell migration assays were performed using modified Boyden Chambers (Transwells, Corning/Costar no. 3422, USA). Cells were transfected with 10 nM miRNA by reverse transfection and plated in 10-cm dishes at 8x10 5 cells/ dish. After 48 h, 1x10 5 cells were added to the upper chamber of each migration well and were allowed to migrate for 48 h. After gentle removal of the non-migratory cells from the filter surface of the upper chamber, the cells that migrated to the lower side were fixed and stained with Diff-Quick (Sysmex Corp., Japan). The number of cells migrating to the lower surface was determined microscopically by counting four areas of constant size per well. Cell invasion assays were carried out using modified Boyden chambers in 24-well tissue culture plates at 1x10 5 cells per well (BD Biosciences, USA). All experiments were performed in duplicate.
Statistical analysis. The relationships between two variables and numerical values were analyzed using the Mann-Whitney U test and the relationships between three variables and numerical values were analyzed using the Bonferroni-adjusted Mann-Whitney U test. Expert StatView analysis software (ver. 4; SAS Institute Inc., Cary, NC, USA) was used in both analyses. In the comparison of three variables, an unadjusted statistical level of significance of P<0.05 corresponded to the Bonferroni-adjusted level of P<0.0167.

Results
Expression of miR-29-family miRNAs in clinical cervical SCC specimens. The sequences and chromosomal locations of miR-29-family miRNAs (miR-29a/b/c� �� �h� h��a� g����� are shown in Fig. 1A. These miRNAs were clustered at two different human genomic loci, miR-29b-1 a�� miR-29a a� 7q32.3 and miR-29b-2 a�� miR-29c at 1q32.2. We evaluated the expression of miR-29-family miRNAs in 18 clinical specimens and 11 non-cancer tissues. The expression levels of miR-29a a�� miR-29c were significantly lower in tumor tissues than in non-cancer tissues. However, there was no significant difference in the expression of miR-29b (Fig. 1B). When we compared two miRNAs (miR-29a a�� miR-29c) after normalization to the expression of RNU48, miR-29a was more abundantly expressed in both normal and cancer tissues.
Identification of miR-29a-regulated putative target genes. T� identify genes regulated by miR-29a, w� ���� in silico a�� genome-wide gene expression analyses. The strategy for the selection of miR-29a-target genes is shown in Fig. 3. First, to gain further insight into which genes were affected by miR-29a, we performed genome-wide gene expression analysis using miR-29a-transfected CaSKi cells; 986 genes were identified as downregulated (log 2 ratio <-1.0) by miR-29a transfection. Next, we used the TargetScan database to examine whether these g���� ����a���� miR-29a binding sequences in their 3'UTRs. Finally, the gene set was analyzed with a publicly available gene expression data set in the GEO (accession no. GSE6791), and genes upregulated (log 2 ratio >1.5) in cervical SCC were chosen. A total of 29 genes were candidate miR-29a-regulated oncogenic genes in cervical SCC (Table II).
As a result of our selection strategy, we identified HSP47 as one of the most highly upregulated genes in cervical SCC   HSP47 was directly regulated by miR-29a. We performed qRT-PCR and western blotting in HeLa cells to investigate whether HSP47 expression was reduced by restoration of miR-29a. The mRNA and protein expression levels of HSP47 were significantly repressed in miR-29a ��a������a��� �� comparison with mock-or miR-control-transfected cells ( Fig. 4A and B). T� ��������� wh��h�� �h� 3'UTR �� HSP47 �RNA ha� an actual target site for miR-29a, we performed luciferase reporter assays using a vector encoding the 3'UTR of HSP47 mRNA. We found that the luminescence intensity was significantly reduced by transfection with miR-29a and the vector carrying the wild-type 3'UTR of HSP47, wh���a� ��a��������� with a deletion vector blocked the decrease in luminescence (Fig. 4C).

Immunohistochemistry of HSP47 in a tissue microarray.
We confirmed the expression levels of HSP47 in normal cervical tissues, CIN tissues, and cancer tissues by immunohistochemical staining. Very low expression of HSP47 was observed in normal tissues (Fig. 7A). In CIN, weak expression �� HSP47 was observed (Fig. 7B). In contrast, HSP47 was more strongly expressed in several tumor lesions compared to normal tissues and CIN tissues ( Fig. 7C and D). The expression score of HSP47 in cervical SCC was significantly higher than that in CIN (P= 0.0001) and normal tissues (P<0.0001; Fig. 7E).

Discussion
Aberrant expression of the miR-29 family miRNAs has been reported in several types of human cancers; however, the expression status varies according to the cancer type. Decreased expression of the miR-29 family has been observed in cholangiocarcinoma, nasopharyngeal cancer, non-small cell  lung cancer, hepatocellular carcinoma, malignant peripheral nerve sheath tumors and mantle cell lymphoma. In contrast, upregulation of the miR-29 family was reported in breast cancer, colon cancer and acute myeloid leukemia (16). In cervical cancer, it was reported that miR-29 �a�g��� �h� HPV-related gene (17). In this study, our data demonstrated �ha� miR-29a was significantly downregulated in cervical SCC clinical specimens and cell lines, regardless of the type of HPV infection. Furthermore, restoration of miR-29a �� cervical cancer cells inhibited cancer cell migration and inva-����, ��gg�����g �ha� miR-29a ��������� a� a ����� ���������� and may contribute to metastasis in cervical SCC.
MiRNAs are unique in their ability to regulate many protein coding genes. Bioinformatic predictions have indicated that miRNAs regulate >30% of protein coding genes (20). Aberrant expression of miRNAs causes destruction of tightly regulated miRNA/protein-coding RNA networks in human cancer cells. Therefore, identification of aberrantly expressed miRNA-mediated cancer pathways and target genes is the first step in elucidating the role of miRNAs in human cancers.
HSP47, a 47-kDa heat-shock protein, was first identified in fibroblasts (21) and is located within the human chromosome 11q13.5 region, which is frequently amplified in several types of human cancers, including cervical SCC (22). Moreover, HSP47 is localized in the endoplasmic reticulum, a cellular organelle involved in the intercellular processing and secreting of procollagens (23). Many studies have demonstrated that HSP47 is overexpressed in fibrotic diseases, including kidney fibrosis, pulmonary fibrosis, cardiac fibrosis, and liver cirrhosis (24). Fibrosis is a common disease of organ dysfunction and is closely associated with ECM proteins, such as collagens, actins and fibronectins (25). Interestingly, members of the miR-29 family have been shown to be involved in regulating ECM proteins and multiple studies have indicated that aberrant expression of miR-29 family members contributes substantially to the development of disease (26). Thus, downregulation of the miR-29 family and dysregulation of HSP47 and ECM components are key events contributing to the pathogenesis of diseases, suggesting that these molecules are potential therapeutic targets.
Overexpression of HSP47 has been reported in pancreatic cancer, gastric cancer, and head and neck squamous cell carcinoma (27)(28)(29). Our present data also support previous reports, suggesting that silencing of miR-29a caused overexpression �� HSP47 and was an important event in the pathogenesis of cervical SCC, contributing to cancer cell migration and invasion in particular. The epithelial-to-mesenchymal transition (EMT) is a fundamental biological process whereby epithelial cells lose their polarity and undergo a transition to a mesenchymal phenotype (30). Initiation of the EMT requires external signals, such as epidermal growth factor (EGF), fibroblast growth factor (FGF), hepatocyte growth factor (HGF), a�� ��a��������g g��w�h �a����-β �TGF-β) (31). The TGF-β pathway is a prominent inducer of the EMT and expression �� �h� miR-29 family has been shown to have an inverse relationship with the TGF-β pathway. Restoration of miR-29 family members directly suppresses TGF-β1 a�� TGF-β2 a�� disrupts the expression of ECM proteins (32). Furthermore, ECM molecules, including collagen type I, promote the EMT through integrin and discoidin domain receptor-1 signaling (33,34). Thus, the understanding of molecular pathways and targets regulated by the tumor-suppressive miR-29 family may provide new insights into the EMT process in cervical SCC and facilitate the development of more effective strategies for future therapeutic interventions for this disease.
In conclusion, downregulation of miR-29a is a frequent event in cervical SCC. Moreover, tumor-suppressive miR-29a directly regulates HSP47, a molecular chaperone involved in the maturation of collagen molecules. Restoration of miR-29a or silencing of HSP47 inhibited cancer cell migration and invasion, suggesting that the miR-29a-HSP47 pathway contributes to the metastasis of cervical SCC. Identification of molecular targets regulated by tumor-suppressive miRNAs will provide insights into the potential mechanisms of cervical SCC oncogenesis and metastasis, facilitating the development of novel therapeutic strategies for the treatment of this disease.