Introduction

Due to a lack of early diagnostic strategies and its aggressive nature, pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal cancers known to medicine [1]. Treatment options for locally advanced or metastatic PDAC are limited, and the median life expectancy is 6ā€“11 months and 3ā€“6 months for patients presenting with locally advanced disease or metastatic disease, respectively [2, 3]. Searching for new therapeutic targets and developing useful prognostic molecular markers are important goals to improve treatment outcomes of PDAC.

MicroRNAs (miRNAs) are small noncoding RNAs 19ā€“24 nucleotides in length. They regulate gene expression by repressing translation or by cutting mRNAs in a sequence-dependent manner [4,5,6]. A single miRNA species is capable of modulating many protein-coding and noncoding RNA transcripts [7,8,9]. Thus, aberrantly expressed miRNAs can disrupt normal cell function, including supporting cancer pathogenesis [7,8,9].

Based on our original miRNA expression signatures by current genomic approaches, including that for PDAC, we have identified RNA networks that are controlled by antitumor miRNAs in several cancers [10,11,12,13,14,15]. In PDAC cells, our previous studies demonstrated that miR-375, miR-216b-3p, miR-217, miR-148a, and miR-124-3p were downregulated in PDAC tissues and these miRNAs had tumor suppressing functions, including controlling various oncogenes in PDAC cells [14, 16,17,18,19].

For example, expression of anillin (ANLN), actin-binding protein was directly controlled by miR-217 and aberrant expression of ANLN promoted to cancer cell migration and invasion capabilities of PDAC cell lines [17]. Ectopic expression of miR-124-3p attenuated cancer cell aggressiveness through targeting oncogenic signaling via FAK, AKT, and ERK in PDAC cells [19]. Integrin Ī±3 (ITGA3) and integrin Ī²1 (ITGB1) were direct targets of miR-124-3p regulation in PDAC cells [19]. These findings suggest that analyses of antitumor miRNAs that regulate RNA networks will enhance understanding of PDAC molecular pathogenesis.

In this study, we focused on the passenger and guide strands of the miR-130b duplex (miR-130b-5p, the passenger strand and miR-130b-3p, the guide strand) based on miRNA expression signature of PDAC by RNA sequencing. Involvement of passenger strands of miRNAs is a new concept of miRNA biogenesis and these miRNAs provide the opportunity to find new regulatory networks in cancer cells. Here, we investigated the antitumor roles of miR-130b-5p, and their regulated oncogenic genes in PDAC pathogenesis.

Materials and methods

Human PDAC clinical specimens and cell lines

The present study was approved by the Bioethics Committee of Kagoshima University (Kagoshima, Japan; approval no. 160038 28-65). Written prior informed consent and approval were obtained from all of the patients.

In this study, 31 PDAC clinical samples were collected from PDAC patients who underwent resection at Kagoshima University Hospital from 1997 to 2016. Fifteen normal pancreatic tissue specimens were collected from noncancerous regions. The clinical samples were staged according to the American Joint Committee on Cancer/Union Internationale Contre le Cancer (UICC) TNM classification. Clinical features in PDAC specimens are shown in Supplemental TableĀ 1.

We used two PDAC cell lines: SW1990, purchased from the American Type Culture Collection (Manassas, VA, USA), and PANC-1, purchased from RIKEN Cell Bank (Tsukuba, Ibaraki, Japan).

Quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR)

The procedure for qRT-PCR has been described previously [17,18,19,20,21]. TaqMan qRT-PCR probes were obtained from Thermo Fisher Scientific (Waltham, MA, USA)Ā as follows: miR-130b-5p (product ID: 002114), miR-130b-3p (product ID: 00456) and EPS8 (product ID: Hs00610286_mH). GUSB (product ID: Hs99999908_m1) and RNU48 (product ID: 001006) were used as internal controls.

Transfection of mimic and inhibitor miRNA, small interfering RNA (siRNA) into PDAC cells

The following mature miRNAs and siRNAs were transfected into PDAC cells (PANC-1 and SW1990): miR-130b-5p (product ID: PM12970, Applied Biosystems, Foster City, CA, USA), miR-130b-3p (product ID: PM10777, Applied Biosystems) and Stealth Select RNAi siRNA, EPS8 siRNAs (product IDs: HSS103325 and HSS103326, Invitrogen, Carlsbad, CA, USA). The transfection procedures were described in previous studies [20,21,22,23,24,25].

Incorporation of miR-130b-5p into the RISC: assessment by Ago2 immunoprecipitation

Agonaute-2 (Ago2) is an essential components of the RNA-induced silencing complex (RISC) that binds to miRNAs. miRNAs were transfected into PANC-1 cells and were isolated using a microRNA Isolation Kit, Human Ago2 (Wako Pure Chemical Industries, Ltd., Osaka, Japan) as described previously [22,23,24,25]. The expression levels of Ago2-conjugated miRNAs were assessed by qRT-PCR assay.

Cell proliferation, migration, and invasion assays

Functional assays for determining cell proliferation, migration, and invasion were described previously [16,17,18,19].

Identification of putative oncogenic target genes regulated by miR-130b-5p in PDAC cells

To identify miR-130b-5p-controlled oncogenes, the following data sets were used: genome-wide gene expression analyses using PDAC cells transfected with miR-130b-5p predicted putative target genes that have miR-130b-5p binding sites in their 3ā€™ untranslated regions (TargetScan database ver. 7.1) and gene expression data of PDAC clinical specimens (Gene Expression Omnibus dataset: GEO accession number, GSE15471). Gene expression data (miR-130b-5p transfected PANC-1 cells) were deposited into the GEO database (accession number: GSE115801). An outline of the approach is shown in Supplemental Fig.Ā 1 and was described in previous studies [20,21,22,23,24,25].

Exploration of downstream targets regulated by si-EPS8 in PDAC cells

Genome-wide gene expression and database oriented in silico analyses were applied to identify EPS8-mediated downstream genes. Outlines of the strategies were described in our previous studies [17, 18, 20, 21]. Our target search strategy in this study is shown in Supplemental Fig.Ā 2. Gene expression data were deposited in GEO database (accession number: GSE118966).

PDAC clinical data analysis by TCGA database

TCGA database was used to investigate the clinical significance of PDAC miRNAs and the genes they regulated (https://tcga-data.nci.nih.gov/tcga/). Gene expression and clinical data were obtained from cBioPortal (http://www.cbioportal.org/) and OncoLnc (http://www.oncolnc.org) (data downloaded on April 28, 2018). Detailed information on the databases were described in the previous papers [26,27,28].

Western blot analysis and immunohistochemistry

The procedures for western blotting and immunohistochemistry were described in previous studies [17,18,19]. These assays used the following antibodies: anti-EPS8 (product ID: #43114, Cell Signaling Technology,Ā Danvers, MA, USA) and anti-GAPDH (product ID: SAF6698,Ā Wako).

Tissue sections were incubated overnight at 4ā€‰Ā°C with anti-EPS8 antibodies diluted 1:400 (HPA003897; Sigma- Aldrich, St. Louis, MO, USA).

Luciferase reporter assays

The following 2 sequences were cloned into the psiCHECk-2 vector (C8021; Promega Corporation, Madison, WI, USA): the wild-type sequence of the 3'-untranslated regions (UTRs) of EPS8, or the deletion-type, which lacked the miR-130b-5p target sites from EPS8 (position 713ā€“719). The procedures for transfection and dual luciferase reporter assays were provided in previous studies [20,21,22,23,24,25].

Statistical analysis

To assess the significance of differences between 2 groups, we used Mannā€“Whitney U-tests. Differences between multiple groups were assessed by one-way ANOVA and Tukey tests for post-hoc analysis. We evaluated the correlations between the expression levels of miR-130b-5p and EPS8 using Spearmanā€™s rank test. Tests utilized Expert StatView version 5.0 (SAS Institute, Inc., Cary, NC, USA) and JMPPro 14.0.0 (SAS Institute, Inc., Cary, NC, USA).

Results

Downregulation of miR-130b-5p and miR-130b-3p in PDAC clinical specimens and cell lines

We performed qRT-PCR to evaluate the expression levels of miR-130b-5p and miR-130b-3p in PDAC tissues (nā€‰=ā€‰31) as well as in normal pancreatic tissues (nā€‰=ā€‰15) and in 2 PDAC cell lines (PANC-1 and SW1990). Clinical features of the patients are summarized in Supplemental TableĀ 1.

The expression levels of miR-130b-5p and miR-130b-3p were significantly downregulated in cancer tissues (pā€‰=ā€‰0.0005 and pā€‰=ā€‰0.0009; Fig.Ā 1a). Spearmanā€™s rank test showed a positive correlation between the expression levels of miR-130b-5p and miR-130b-3p (ā€‰p<ā€‰0.0001, rā€‰=ā€‰0.875; Fig.Ā 1b).

Fig. 1
figure 1

The functional significance of miR-130b-5p and miR-130b-3p in PDAC cells. a Expression levels of miR-130b-5p and miR-130b-3p in PDAC clinical specimens and cell lines (PANC-1 and SW1990). RNU48 was used as an internal control. b Spearmanā€™s rank test demonstrated a positive correlation between the expression levels of miR-130b-5p and miR-130b-3p. cā€“e Effects of ectopic expression of miR-130b-5p and miR-130b-3p on PADC cells. c Cell proliferation was determined by XTT assays 72ā€‰h following transfection with miR-130b-5p or miR-130b-3p. d Results of cell migration assays. e Cell invasion activity was determined using Matrigel invasion assays. *, pā€‰<ā€‰0.05,Ā **,Ā pā€‰<ā€‰0.0001

Full size image

In 2 cancer cell lines, PANC-1 and SW1990, the expression levels of miR-130b-5p and miR-130b-3p were extremely low (Fig.Ā 1a).

Effects of ectopic expression of miR-130b-5p and miR-130b-3p on PDAC cells

To verify the antitumor roles of miR-130b-5p and miR-130b-3p, we conducted gain-of-function studies by miRNA transfection into PANC-1 and SW1990 cells.

In cell proliferation assays, the inhibition of cancer cell growth was only detected with miR-130b-5p transfection into PANC-1 cells (Fig.Ā 1c). Cell migration activities were reduced in the cells transfected with miR-130b-5p or miR-130b-3p (Fig.Ā 1d).

Matrigel invasion assays revealed that transfection with miR-130b-5p or miR-130b-3p significantly decreased cell invasive capacity (Fig.Ā 1e). However, no change was observed in miR-130b-3p transfection into PANC-1 cells (Fig.Ā 1e).

Incorporation of miR-130b-5p and miR-130b-3p into the RISC in PDAC cells

Ago2 is an essential component of the RISC. We hypothesized that the miR-130b-3p passenger strand in PDAC cells might be incorporated into the RISC where it could act as a tumor suppressor. To test that possibility, Ago2 was immunoprecipitated from PANC-1 cells that had been transfected with either miR-130b-5p or miR-130b-3p. Following isolation of Ago2-bound miRNAs, they were analyzed by qRT-PCR to determine whether miR-130b-5p or miR-130b-3p or both were associated. In transfectants, we observed higher levels of miR-130-5p expression than in mock transfectants or miR-controls or miR-130b-3p (pā€‰<ā€‰0.005) (Supplemental Fig.Ā 3).

Identification of putative target genes controlled by miR-130b-5p in PDAC cells

To predict putative target genes controlled by miR-130b-5p in PDACs, we combined data from the following: genome-wide gene expression data (miR-130b-5p transfected into PANC-1 cells; GEO accession number: GSE115801), gene expression data from PDAC clinical specimens (GSE15471) and TargetScan database. The selection strategy of miR-130b targets is shown in Supplemental Fig.Ā 1. A total of 103 genes were identified as putative miR-130b-5p controlled oncogenes in PDAC cells (TableĀ 1).

Table. 1 Identification of putative targets regulated by miR-130b-5p in PDAC cells
Full size table

To investigate the relationship between these target genes and the course of PDAC, we examined these genes with TCGA database. Among these targets, high expression of 9 genes (EPS8, ZWINT, SMC4, LDHA, GJB2, ZCCHC24, TOP2A, ANLN, and ADCY3) was associated with poor prognosis (5-year overall survival rates: pā€‰<ā€‰0.01) (Fig.Ā 2).

Fig. 2
figure 2

Clinical significance of the expression of 9 genes (EPS8, ZWINT, SMC4, LDHA, GJB2, ZCCHC24, TOP2A, ANLN, and ADCY3) based on The Cancer Genome Atlas (TCGA) database. Survival rate differences were analyzed by Kaplanā€“Meier survival curves and log-rank statistics. a Kaplanā€“Meier plots of overall survival and b disease-free survival with log-rank tests for genes with high and low expression from The Cancer Genome Atlas database

Full size image

Below, we focused on EPS8 (epidermal growth factor receptor kinase substrate 8) because its expression was the most significantly predicted poor prognosis of the PDAC patients (TableĀ 2, Fig.Ā 2).

Table. 2 Identification of EP8 mediated downstream genes in PDAC cells
Full size table

Expression of EPS8 in PDAC clinical specimens and its clinical significance

The levels of EPS8 mRNA were significantly upregulated in PDAC tissues (Fig.Ā 3a), with a negative correlation between the expression of EPS8 and miR-130b-5p (pā€‰=ā€‰0.0191, ā€‰rĀ =ā€‰āˆ’0.349; Spearmanā€™s rank tests, Fig.Ā 3b).

Fig. 3
figure 3

Aberrant expression of EPS8 in PDAC specimens and its clinical significance. a Expression levels of EPS8 in PDAC clinical specimens. GUSB was the internal control. b Spearmanā€™s rank test was used to evaluate the correlation between EPS8 expression and miR-130b-5p expression in PDAC clinical specimens. c Analysis of the expression levels of EPS8 in patients with PDAC using TCGA database. Forest plot of univariate Cox proportional hazards regression analysis of 5-year overall survival. Univariate and multivariate analyses for OS using TCGA database were carried out by Cox proportional hazards regression analyses. d Immunohistochemical analysis of PDAC clinical samples. EPS8 was strongly expressed in cancer lesions

Full size image

Cox hazard regression analyses assessed the clinical significance of EPS8 expression for OS in patients with PDAC. With multivariate analysis, we found that EPS8 expression was an independent predictive factor for OS (hazard ratioā€‰=ā€‰1.893, ā€‰p=ā€‰0.0053; Fig.Ā 3c).

PDAC clinical specimens were subjected to immunohistochemical analyses. The results indicated that EPS8 protein was strongly expressed in cancer lesions. In contrast, expression was infrequent and weak in normal pancreatic cells (Fig.Ā 3d).

Direct regulation of EPS8 by miR-130b-5p in PDAC cells

In cells transfected with miR-130b-5p, the levels of EPS8 mRNA and EPS8 protein were significantly lower than mock- or miR-control-transfected cells (Fig.Ā 4a, b). Binding sites for miR-130b-5p in the 3ā€™-UTR of EPS8 (positions 713ā€“719, Fig.Ā 4c, upper) were predicted by the TargetScan database. We used luciferase reporter assays with vectors carrying either the wild-type or deletion-type 3'-UTR of EPS8. We observed greatly reduced luminescence after transfection with miR-130b-5p and the vector carrying the wild-type 3ā€™-UTR of EPS8. Transfection with the deletion-type vector did not reduce luminescence intensities in PANC-1 and SW1990 cells. Thus, miR-130b-5p directly bound to EPS8 in the 3ā€™-UTR (Fig.Ā 4c).

Fig. 4
figure 4

Oncogenic function of EPS8 in PDAC cells. a, b miR-130b-5p directly regulated EPS8 in PDAC cells. Expression levels of EPS8 mRNA (a) or EPS8 protein (b) 72ā€‰h or 96 hĀ following transfection with 10ā€‰nM miR-130b-5p into cell lines. c miR-130b-5p binding site (positions 713ā€“719) in the 3'-UTR of EPS8 mRNA. Dual luciferase reporter assays using vectors encoding putative miR-130b-5p target sites in the EPS8 3'-UTRs for both wild-type and deleted regions. Renilla luciferase values were normalized to firefly luciferase values. *pā€‰<ā€‰0.005. dā€“f Effects of silencing EPS8 in PDAC cells. d Cell proliferation, e migration and f invasion assays. These assays showed that inhibition of migration and invasion were observed in si-EPS8-transfected cell lines (PANC-1 and SW1990). *pĀ <ā€‰0.005

Full size image

In addition, we investigated the effect of suppression of EPS8/EPS8 by miR-130a-5p (seed sequence is almost identical) in PDAC cells. Expression of EPS8/EPS8 was slightly suppressed by miR-130a-5p in PANC-1 and SW1990 (data not shown).

Effects of silencing EPS8 on PDAC cells

Next, we transfected siRNAs into PANC-1 and SW1990 cells to examine the function of EPS8 in PDAC cells. The mRNA and protein expression levels of EPS8/EPS8 were decreased by si-EPS8 (Supplemental Figs.Ā 4A and 4B).

We examined the effects of knockdown of EPS8 in PDAC cells, and found that cell proliferation was not affected (Fig.Ā 4d). Cancer cell migration and invasive activities were significantly inhibited by si-EPS8 transfection into PDAC cells, PANC-1 and SW1990. However, silencing of EPS8 did not affect cell proliferation (Fig.Ā 4dā€“f).

Downstream genes affected by silencing of EPS8 in PDAC cells

To identity downstream genes controlled by EPS8, we used two sets of genome-wide gene expression data (si-EPS8 transfected cells: GSE118966 and PDAC expression data: GSE15471). Our selection strategy is shown in Supplemental Fig.Ā 2.

In total, 48 genes were identified as putative downstream genes controlled by EPS8 in PDAC cells (TableĀ 2). Among 8 genes, high expression affected overall survival rates (pā€‰<ā€‰0.05). Specifically, MET, HMGA2, CORO1C, FERMT1, RARRES3, PTK2, MAD2L1, and FLI1 were significantly associated with poor prognosis in patients with PDAC by TCGA analysis (TableĀ 2 and Supplemental FigureĀ 5).

Discussion

miRNAs have unique characteristics. For example, a single miRNA species can regulate vast numbers of RNA transcripts in normal and pathological cells. Expression of RNAs controlled by miRNA varies depending on the cell [7,8,9]. Therefore, identification of aberrantly expressed miRNAs and their targets is the first step in elucidating molecular pathogenesis in PDAC cells. Using our original miRNA signature of PDAC by RNA sequencing, the molecular network controlled by antitumor miRNAs in PCAD cells is being clarified [14, 16,17,18,19]. In this study, we focused on both strands of pre-miR-130b (miR-130b-5p and miR-130b-3p) because these miRNAs were significantly downregulated in our and other PDAC signatures [14, 29, 30].

Several miRNAs form families based on their seed sequences. The miR-130 family consists of 4 miRNAs: miR-130a (chromosome 11q12.1), miR-130b (22q11.21), miR-301a (17q22), and miR-301b (22q11.21) [31,32,33]. The seed sequences of passenger strands miR-130a-5p and miR-130b-5p are almost identical. The seed sequences of guide strands of all member of the miR-130 family (miR-130a-3p, miR-130b-3p, miR-301a-3p and miR-301b-3p) are identical (seed sequences are summarized in Supplemental FigureĀ 6). Previous studies showed that a number of miR-130 family (guide strands) were overexpressed in cancer tissues and their functions were involved in oncogenesis, e.g., bladder cancer, esophageal squamous cell carcinoma and lung cancer [32, 34, 35].

Contrary to these reports, miR-130a and miR-130b were downregulated in cancer tissues and they acted as antitumor miRNAs in ovarian cancer, prostate cancer, endometrial cancer, and papillary thyroid carcinoma [31, 33, 36, 37]. Our present data show that both strands (miR-130b-5p and miR-130b-3p) were significantly reduced in PDAC clinical specimens and cell lines. Furthermore, ectopic expression of these miRNAs inhibited malignant phenotypes in PDAC cells, suggesting that both miR-130b-5p and miR-130b-3p play antitumor roles in PDAC cells. In a previous study of PDAC cells, expression of miR-130b-3p was suppressed in cancer tissues and its expression was an independent prognostic predictor of the patientsā€™ disease course [38]. Overexpression of miR-130b-3p induced apoptotic cells through targeting of STAT3 in PDAC cells [38]. These findings showed that miR-130b has multiple functions, oncogenic or antitumor roles depending on the specific cancer cell. It is indispensable to elucidate the molecular mechanism controlling miR-130b expression in several types of cancer cells.

Previous study showed that hyper-methylation of the promoter region of miR-130b was observed in ovarian cancer clinical tissues and cell lines and methylation cased to downregulation of miR-130b expression [31]. In prostate cancer, downregulation of miR-130b/miR-301b cluster was detected in clinical specimens and cell lines [33]. Methylation levels of their promoter region were significantly higher in prostate cancer tissues compared to normal tissues [33]. Expression levels of miR-130b and miR-301b were upregulated by treatment of demethylation drugs [33]. These findings showed that downregulation of miR-130b was mediated by aberrant methylation on its promoter region. For miR-130 family, comprehensive analysis of the molecular mechanism of suppressing their expression in PDAC cells is indispensable.

This is the first report that miR-130b-5p (the passenger strand) acted as an antitumor miRNA in PDAC. Therefore, we focused on miR-130b-5p to investigate its control of oncogenes involved in PDAC pathogenesis. In this study, a total of 103 putative oncogenes were identified that were regulated by miR-130b-5p in PDAC cells. Among these targets, overexpression of 9 genes (EPS8, ZWINT, SMC4, LDHA, GJB2, ZCCHC24, TOP2A, ANLN, and ADCY3) were closely associated with poor prognosis of the patients with PDAC. Interestingly, aberrant expression of ANLN (actin-binding protein anillin) was detected in PDAC clinical specimens [17]. Knockdown assays of ANLN expression markedly inhibited cancer cell migration and invasive capabilities of PDAC cell lines [17]. In addition, ANLN was directly controlled by antitumor miR-217 in PDAC cells [17].

Furthermore, we investigated the functional significance of EPS8 (epidermal growth factor receptor pathway substrate 8) in PDAC cells. EPS8 binds several adaptor proteins and acts as a substrate for receptor and non-receptor tyrosine kinases, e.g., EGFR, FGFR, VEGFR, and Src [39]. Other studies showed that EPS8 has the actin-binding ability and it acts by capping barbed ends and promoting bundling [40]. Furthermore, EPS8 forms a trimer (EPS8, Abi-1 and SOS-1) and this complex acts as a guanine nucleotide exchange factor (GEFs) in Rac signaling and contributes to Rac-based actin polymerizing processes [41]. Aberrantly expressed EPS8 was reported in colon cancer, breast cancer, hematologic malignancies, and cervical cancer, and its overexpression was closely involved in the malignant phenotype [42,43,44,45]. Our present data revealed that EPS8 regulated cancer cell migration and invasion and its expression is promising as a diagnostic marker for PDAC. Aberrant expression of EPS8 might be a promising therapeutic target for PDAC.

Finally, to investigate the EPS8-mediated oncogenic genes and pathways in PDAC cells, we applied genome-wide gene expression analyses using knockdown of EPS8 in cells. A total of 48 genes were identified as putative EPS8-mediated targets in PDAC cells. Surprisingly, aberrant expression of 7 genes (MET, HMGA2, FERMT1, RARRES3, PTK2, MAD2L1, and FLI1, pā€‰<ā€‰0.05) was closely associated with poor prognosis of patients with PADC. In this study, it was revealed that many of the genes controlled by antitumor miR-130b-5p and ESP8-mediated downstream genes were closely involved in the molecular pathogenesis of PDAC. Elucidation of novel RNA networks controlled by antitumor miRNAs will accelerate comprehensive understanding of molecular pathogenesis of PDAC.

In conclusion, our results showed that expression of both strands of the pre-miR-130b duplex were significantly downregulated in PDAC clinical specimens and thus the miR-130b-duplex could act as an antitumor miRNA in such cells. A total of 9 genes (EPS8, ZWINT, SMC4, LDHA, GJB2, ZCCHC24, TOP2A, ANLN, and ADCY3) were closely associated with PDAC pathogenesis. Among these targets, the aberrant expression of EPS8 enhanced cancer aggressiveness, suggesting that EPS8 could be a promising therapeutic target for PDAC. Our approach, discovery of antitumor miRNAs and their target RNAs, will contribute to exploring the causes of this malignant disease.