Abstract
Background/Aim: Our recent miRNA analyses revealed that miR-30a-5p has tumor-suppressive activity in pancreatic ductal adenocarcinoma (PDAC). Herein, we sought to identify tumor-suppressive genes controlled by miR-30a-5p, emphasizing on genes that are closely involved in the molecular pathogenesis of PDAC. We uncovered several novel findings regarding the pathogenesis of this disease. Materials and Methods: In silico analyses were used to identify the putative target genes of miR-30a-5p and assess their expression levels. Direct regulation of RRM2 by miR-30a-5p and its oncogenic functions were evaluated in PDAC cell lines. Overexpression of RRM2 was demonstrated in clinical samples. Results: A total of 24 putative targets were identified by in silico database analysis. High expression of 4 genes (CBFB, RRM2, AHNAK, and DCBLD1) was significantly associated with shorter survival of patients with PDAC. Functional assays demonstrated that knockdown of RRM2 attenuated the malignant phenotype of PDAC cells. Conclusion: The miR-30a-5p/RRM2 axis facilitated the malignant transformation of PDAC cells.
- microRNA
- miR-30a-5p
- RRM2
- pancreatic ductal adenocarcinoma
- tumor-suppressor
Pancreatic ductal adenocarcinoma (PDAC) is an extremely aggressive cancer with the highest mortality rate among various human cancers (1, 2). The 5-year survival rate for patients with PDAC is around 8%, and approximately 459,000 people were diagnosed with PDAC and 432,000 died in 2018 (3). A notable feature of PDAC is that patients have few symptoms in the early stages of disease. By the time symptoms appear, many patients have reached an advanced stage in which there is invasion of surrounding tissues or distant metastasis (4, 5). The prognosis for advanced cases (surgically unresectable cases) is extremely poor (most patients die within a year) (5).
Cancer cell development and progression are linked to the accumulation of genetic mutations, epigenetic dysregulation and aberrant activation of oncogenic signaling (6, 7). Driver mutations are only a small fraction of the genetic alterations identified in patients. The remaining genomic changes likely contribute to oncogenic transformation (8). There are three types of RAS proteins, KRAS, NRAS and HRAS. Mutations in these genes are frequently detected in human cancers, and they are closely involved in human oncogenesis (9). KRAS mutations are detected in most patients with PDAC. Thus, KRAS mutations are driver mutations in this disease (10). Controlling the activation of KRAS-mediated oncogenic signaling is an important issue, but therapeutic drugs that control KRAS have not yet been developed (11, 12).
A vast number of studies indicate that non-coding RNAs (ncRNAs) are involved in a wide range of biological processes, e.g., cell proliferation, apoptosis, development, the epithelial-to-mesenchymal transition (EMT) and chromatin modelling (13-15). In human cancer, aberrant expression of ncRNAs can contribute to malignant transformation, metastasis and drug resistance (16). However, the biological roles and functions of non-coding RNAs are mostly uncharacterized.
MicroRNAs (miRNAs) are small ncRNAs that control the expression of RNA transcripts in both normal and disease cells in a sequence-dependent manner (17). A single miRNA can influence numerous transcripts of protein coding and non-coding genes, accounting for approximately 60% of cellular RNAs (18). During the development of cancer cells, the abnormal expression of microRNAs may disrupt RNA networks. There is currently a consensus that aberrantly expressed miRNAs contribute to cancer cell development, progression, EMT, metastasis, and drug resistance (14, 18).
Based on a PDAC miRNA signature obtained by RNA sequencing, we have identified tumor-suppressive miRNAs and their oncogenic targets that are closely associated with PDAC pathogenesis. There are multiple miRNAs with the same mature sequence, and they are grouped in “families”, e.g., the miR-29-family, miR-30-family, miR-200-family etc. Focusing on the miR-30-family, we have been searching for oncogenic genes that are controlled in pancreatic cancer. Our recent study showed that both strands of pre-miR-30a (miR-30a-5p: the guide strand and miR-30a-3p: the passenger strand) were down-regulated in PDAC tissues. Ectopic expression assays demonstrated that both strands of pre-miR-30a had tumor-suppressive roles, suggesting these miRNAs controlled oncogenic genes in PDAC cells (19).
The aim of the present study was to identify genes controlled by tumor-suppressive miR-30a-5p, genes that are tightly associated with PDAC molecular pathogenesis. Herein, we found 4 genes (CBFB, RRM2, AHNAK, and DCBLD1) that significantly predicted abbreviated survival of PDAC patients. We focused on RRM2, and investigated the functional significance of this gene in PDAC oncogenesis.
Materials and Methods
Clinical specimens and PDAC cell lines. All the patients in this study provided written prior informed consent and approval. The present study was approved by the Bioethics Committee of Kagoshima University (Kagoshima, Japan; approval no. 160038 28-65, date of approval: 4 September 2016). We utilized the PDAC clinical samples of two patients who underwent surgery at Kagoshima University Hospital from 2012 to 2014. The clinical samples were staged according to the American Joint Committee on Cancer/Union Internationale Contre le Cancer (UICC) TNM classification.
The two PDAC cell lines used in this study, SW1990 and PANC-1, were purchased from the American Type Culture Collection (Manassas, VA, USA) and RIKEN Cell Bank (Tsukuba, Ibaraki, Japan), respectively. Both PANC-1 and SW1990 were maintained in RPMI-1640 medium with 10% fetal bovine serum in a humidified atmosphere of 5% CO2 and 95% air at 37°C. Further details are available in our previous study (19).
RNA extraction and quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR). Total RNA was extracted from frozen cell lines. We previously described our methods for extracting RNA from the cell lines, and methods used for qRT-PCR (19-21). TaqMan probes and primers used in this study were obtained from Applied Biosystems (Waltham, MA, USA).
Transfection of PDAC Cells with Mature miRNAs, small-interfering RNAs (siRNAs), and plasmid vectors. We have described the methods for transfection of miRNAs, siRNAs and plasmid vectors into PDAC cells previously (19-21). The pre-miR miRNA precursors were obtained from Applied Biosystems (Waltham, MA, USA) and stealth RNAi siRNA of RRM2 was obtained from Invitrogen (Waltham, MA, USA).
Cell proliferation, migration, and invasion assays in PDAC cells. The methods for functional assays of cancer cells (proliferation, migration and invasion) were described in our previous studies (19-21). PDAC cells were transfected with 10 nM miRNAs or siRNAs. For invasion and migration assays, SW1990 cells at 1.2×105 and PANC-1 cells at 1.0×105 were transfected in 6-well plates. After 72 h, SW1990 and PANC-1 cells were adjusted to 2.5×105 and were added into each chamber. The cells on the lower surface were counted for analysis after 48 h. All experiments were performed in triplicate. XTT assays were used to assess cell proliferation (19-21).
Identification of the miR-30a-5p targets in PDAC. To identify the putative target genes of miR-30a-5p, we used the strategy shown in Figure 1. We used a combination of genome-wide gene expression and in silico analyses as described before (19-21). TargetScanHuman ver.7.2 (http://www.targetscan.org/vert_72/) was used to find the target genes regulated by miR-30a-5p. Our microarray data (from miR-30a-5p transfected cells) were deposited in the Gene Expression Omnibus (GEO) database (GSE155659). To determine the genes up-regulated in PDAC clinical specimens, we obtained expression data from the Gene Expression Omnibus (GEO) database (GSE15471). The expression data of 39 pairs of specimens are contained in this dataset.
Clinical database analysis of miR-30a-5p target genes. The expression level of each putative target gene was extracted and analyzed from the Cancer Genome Atlas (TCGA) – pancreas adenocarcinoma database (TCGA-PAAD) through the GEPIA2 platform [http://gepia2.cancer-pku.cn/#index (accessed on 10 February 2021)] to analyze gene expression levels in 179 tumor samples versus 171 normal PDAC tissue samples. The methods of clinical database analysis of miRNA target genes between normal and PDAC cancer tissues were described in the previous papers (19-21). Plasmid construction and dual-luciferase reporter assays. Plasmid vectors containing RRM2 with the wild-type sequences of the miR-30a-5p binding sites in the 3’-UTR and without those sequences were prepared. We have described the methods for transfection and dual-luciferase reporter assays in our previous studies (19-21). Dual-Luciferase® Reporter Assay System (Promega) was used following the manufacturer’s instructions.
Western blotting and immunohistochemistry. Immunohistochemistry was done on formalin-fixed and paraffin-embedded clinical sections. The procedures for Western blotting and immunohistochemistry were described in our previous studies (19-21). Anti-RRM2 antibody was purchased from Cell Signaling Technology (Danvers, MA, USA) and GAPDH antibody was purchased from Wako (Osaka, Japan).
Statistical analysis. We used Mann–Whitney U tests for comparisons between two groups and used one-way analysis of variance and Dunnett’s test to compare multiple groups. JMP Pro 14 (SAS Institute Inc., Cary, NC, USA) was used to perform these analyses. All data are presented as the mean±standard error (SE). p-Values less than 0.05 were considered significant.
Results
To identify the putative oncogenic targets of miR-30a-5p in PDAC cells, we used the TargetScanHuman database (release 7.2) and two types of genome-wide gene expression analysis data, miR-30a-5p-transfected into PANC-1 cells, and up-regulated genes in PDAC clinical specimens. Our strategy for identifying miR-30a-5p target genes is shown in Figure 1. Our strategy successfully identified a total of 24 putative oncogenic targets subject to miR-30a-5p regulation in PDAC cells (Table I).
We validated the expression levels of those 24 target genes using TCGA-PAAD database through the GEPIA2 platform. Expression of 22 genes was up-regulated in PDAC tissues (n=179) compared to normal tissues (n=171) (Figure 2).
To determine the clinical relevance, clinicopathological analysis of these target genes was performed using TCGA-PAAD datasets. The expression levels of four out of 24 target genes (CBFB, AHNAK, RRM2 and DCBLD1) had significant impact on the prognosis of PDAC patients (Figure 3; p<0.05) where the higher levels of expression predicted poorer prognoses. Kaplan–Meier curves of the 5-year overall survival for each gene are presented (Figure 3).
Next, we performed multivariate analyses with four prognostic determinants: tumor stage, pathological stage, LN stage, and gene expression. Multivariate analysis revealed that the expression levels of three genes (CBFB, AHNAK and RRM2) were independent prognostic factors in PDAC (p<0.05; Figure 4).
In this study, we focused on RRM2 (ribonucleotide reductase regulatory subunit M2). To investigate the functional significance of RRM2 in PDAC cells, we asked whether the expression of RRM2 was regulated directly by miR-30a-5p in PDAC cells. The expression levels of RRM2 (both protein and mRNA expression) were significantly down-regulated following transfection with miR-30a-5p (Figure 5A and B, respectively).
We also performed dual-luciferase reporter assays using plasmid vectors carrying partial sequences of the RRM2 3’-UTR to confirm whether miR-30a-5p could directly bind to RRM2 in PDAC cells. The plasmid vectors included a “wild-type” RRM2 3’-UTR sequence containing the predicted miR-30a-5p target site, and a “deletion-type” sequence lacking the target site (Figure 5C). Transfection of miR-30a-5p and “wild-type” RRM2 vectors reduced the luciferase activity in the cells (Figure 5C). On the other hand, transfection with miR-30a-5p and the “deletion type” vector did not reduce luciferase activity (Figure 5C). Thus, we showed that miR-30a-5p was directly bound to the 3’-UTR region of RRM2 and suppressed RRM2 expression in PDAC cells.
Next, we assessed the oncogenic functions of RRM2 in PDAC cells by performing knockdown assays using siRNAs. The efficiency with which RRM2 expression was knocked down was evaluated using two types of siRNAs. It was confirmed that both siRNAs significantly reduced the expression levels of RRM2 (both protein and mRNA levels) (Figure 6A and B, respectively). Functional assays showed that siRNA-mediated knockdown of RRM2 expression attenuated aggressive features of PDAC cells (PANC-1 and SW1990 cell lines), e.g., cell proliferation, migration, and invasive abilities (Figure 6C, D and E, respectively).
We assessed the expression of the RRM2 in clinical specimens from patients with PDAC using immunohistochemistry. RRM2 was overexpressed in these cancer tissues as demonstrated by patchy nuclear signals compared to its noncancerous counterparts (Figure 7).
Finally, we investigated the genes differentially expressed by high and low RRM2 expression groups in a PDAC cohort from TCGA using gene set enrichment analysis (GSEA). A total of 7 gene sets were significantly enriched (FDR q-value <0.05) in the high RRM2 expression groups. The significantly enriched gene sets in the RRM2 expression groups included “E2F targets”, “G2M checkpoint”, “MYC targets V2”, “DNA repair”, oxidative phosphorylation”, “Interferon alpha response” and “MYC targets V1” (Figure 8). These results indicate that RRM2 plays diverse roles as an oncogene in PDAC via the cell cycle control pathways.
Discussion
Due to the poor initial symptoms and the high malignancy of the cancer cells, patients with PDAC are in advanced stage at the time of diagnosis (1-5). Therefore, the prognosis of the patients who are not indicated for curative surgical treatment is extremely poor (1-5). The search for early diagnostic markers for PDAC is an important issue and is underway on a global scale using currently developed genomic approaches (22-26).
The current RNA sequencing technology is suitable for the construction of miRNA expression signatures. RNA sequence-based miRNA signatures provide information on dysregulated miRNAs in cancer tissues. Analysis of miRNA signatures has revealed that members of the miR-30 family are frequently downregulated in several types of cancers, including PDAC (21, 27, 28). In the human genome, the miR-30 family is composed of 5 species: (miR-30a, miR-30b, miR-30c, miR-30d and miR-30e). In addition, miR-30c is further subdivided into miR-30c-1 and miR-30c-2. The miR-30 family is encoded by six genes located on human chromosomes 1, 6 and 8 (29). The guide strands of the miR-30 family share the same seed sequence (GUAAACA). Several studies have shown that miR-30a-5p has a tumor suppressive function in PDAC cells (19, 30). For example, transfection of miR-30a-5p suppressed PDAC cell proliferation, cell-cycle progression and enhanced cancer cell apoptosis via targeting FOXD1. Moreover, low expression of miR-30a-5p is associated with poor prognosis of PDAC patients (31). Other studies showed that miR-30a-5p was associated with the gemcitabine response, and its expression directly targeted SNAI1 (31, 32). Overexpression of Yin-Yang 1 (YY1), a zinc-finger transcription factor, has been reported in a wide range of cancers, including PDAC (33, 34). It was found that YY1 regulated autophagy in PDAC cells, and overexpression of miR-30a-5p attenuated the pro-autophagic effects through direct control of YY1 expression (35). Moreover, our recent study confirmed that both strands of miRNAs derived from pre-miR-30a (miR-30a-5p and miR-30a-3p) acted as tumor-suppressors in PDAC (19). Notably, several genes controlled by miR-30a-3p significantly predicted a poorer prognosis of patients with PDAC (19). These findings indicate that the identification of genes controlled by miR-30a-5p improves our understanding of the molecular pathogenesis of PDAC.
In this study, we further illuminated the oncogenic properties of miR-30a-5p by identifying the genes that it controls in PDAC cells. We successfully identified four genes (CBFB, AHNAK, RRM2 and DCBLD1) and their significant impact on PDAC patient prognosis. All four genes were overexpressed in PDAC tissues compared to normal tissues. Importantly, upon multivariate analysis, the expression levels of CBFB, AHNAK and RRM2 were independent prognostic factors in PDAC.
The four putative target genes of miR-30a-5p regulation (CBFB, AHNAK, RRM2 and DCBLD1) are reported to have significant roles in cancer. Three of those genes are discussed below. CBFB encodes the beta-subunit of Runt domain transcription factor, and forms heterodimers with RUNX proteins (RUNX1, RUNX2 and RUNX3), thereby regulating transcriptional activity (36). The inter-play between CBFB and RUNX proteins has also been described in gastric, hepatocellular and breast carcinomas (37-39). Overexpression of RUNX1 and RUNX2 proteins and their roles in tumor progression has been described in PDAC (36, 40). A previous study showed that silencing of RUNX2 enhanced gemcitabine sensitivity in PDAC cells through stimulation of TP53, Tap63 and Tap73-mediated pathways (36). Another study demonstrated that knockdown of RUNX1 reduced the invasive ability of PDAC cells (40). Notably, RUNX1 bound promoter region of miR-93 and negatively regulated expression of miR-93 (40). Overexpression of miR-93 blocked migration and invasiveness abilities of PDAC cells (40).
Likewise, AHNAK2 is highly expressed in PDAC clinical tissues, and its expression significantly predicted a lower overall survival rate of patients (41). Meta-analysis of PDAC transcriptome data showed that a five-gene classifier (TMPRSS4, AHNAK2, POSTN, ECT2, and SERPINB5) discriminated PDAC and early precursor lesions from non-malignant tissue (42). The function of AHNAK2 in PDAC is not well understood. Knockdown assay of AHNAK2 in thyroid carcinoma cells showed that expression of AHNAK2 contributed to cancer cell metastasis and EMT-process through the Wnt/β-catenin pathway (43). In lung adenocarcinoma cells, overexpression of AHNAK2 enhanced malignant transformation (e.g., migration, invasion, and EMT) via the TGF-β/Smad3 pathway (44).
The DCBLD receptor family consists of two paralogous, DCBLD1 and DCBLD2, amino acid sequences are highly conserved across vertebrates (45). To date, the function of the DCBLD2 and its relationship with diseases are becoming clear. However, the function of the DCBLD1 remains uncharacterized (45). A previous study showed that single nucleotide polymorphisms of promoter region of DCBLD1 was associated with head and neck cancer and lung cancer in never-smoking women (46-49). Analysis of TCGA analysis demonstrated that expression of DCBLD1 was associated with worse prognosis in the patients with non-small cell lung cancer and invasive breast cancer (50). Notably, high expression of DCBLD1 was associated with the integrin signaling pathway (50). Expression of these genes is promising as a prognostic marker for patients with PDAC. Furthermore, detailed functional analysis of these genes will lead to the elucidation of the molecular mechanism of malignant transformation of PDAC cells.
In this study, we analyzed the oncogenic roles of RRM2 in PDAC cells. The RRM2 protein is one of the two subunits of the ribonucleotide reductase complex. This reductase catalyzes the formation of deoxyribonucleotides from ribonucleotides and is thus a key enzyme in DNA synthesis (51). The overexpression and oncogenic roles of RRM2 have been reported in a wide range of cancers, and high expression of RRM2 independently predicted the prognosis of several cancers, e.g., lung adenocarcinoma, oral cancer, breast cancer and prostate cancer (51, 52). In pancreatic adenocarcinoma, as in other solid tumors, RRM2 affects tumor growth and invasiveness. siRNA mediated knockdown of RRM2 effectively suppresses pancreatic tumor growth in vivo and in vitro (53, 54). Notably, aberrant expression of RRM2 contributes to chemotherapy-resistance in several types of cancers, including treatment with hydroxyurea, docetaxel, tamoxifen and gemcitabine (55). Previous studies showed that overexpression of RRM2 induced gemcitabine-resistance in PDAC cells (56). Knockdown of RRM2 and gemcitabine treatment synergistically inhibited PDAC cell growth and metastasis (57). Furthermore, the sensitivity to gemcitabine treatment can be predicted by measuring RRM2 expression in patients with PDAC and non-small cell lung cancer (58).
In conclusion, we successfully identified a total of 24 oncogenic targets regulated by tumor-suppressive miR-30a-5p in PDAC cells. Among these targets, the expression levels of 4 genes (CBFB, RRM2, AHNAK, and DCBLD1) significantly predicted shorter survival of PDAC patients. Furthermore, the oncogenic function of RRM2 in PDAC cells was confirmed, indicating that the miR-30a-5p/RRM2 axis plays a pivotal role in PDAC oncogenesis. Identification of novel tumor-suppressive miRNAs and their regulated oncogenic targets should improve our knowledge of the molecular pathogenesis of PDAC.
Acknowledgements
This study was supported by JSPS KAKENHI grant numbers, 19K09077, 20K21633, 20H03753, 21K15597, 21K16426, and 21K09577.
Footnotes
Authors’ Contributions
Conceptualization, N.S.; methodology, N.S.; validation, T.I. and T.A.; formal analysis, Y.H., T.T., N.P., and M.W.; investigation, N.P., Y.H., T.T., M.W., S.A. and C.M.; data curation, T.I., H.K., and T.O.; writing—original draft preparation, N.S. and N.P.; writing—review and editing, H.K., and T.O.; supervision, N.S and T.O.; project administration, N.S.; funding acquisition, M.W., T.I., H.K., N.S., and T.O. All Authors have read and agreed to the published version of the manuscript.
Conflicts of Interest
The Authors declare no conflicts of interest.
- Received August 3, 2021.
- Revision received August 30, 2021.
- Accepted August 31, 2021.
- Copyright © 2021 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.