Main

MicroRNAs are approximately 18–23 nucleotides long small non-coding RNA molecules, which post-transcriptionally regulate the expression of more than 30% of protein coding genes by translational repression.1 MicroRNAs regulate the expression of several putative target genes by binding to a complementary sequence predominantly in their 3′UTR. This post-transcriptional regulation within the context of tumor development has been reported for the regulation of genes which have an impact on cell differentiation, apoptosis and neoplastic transformation.2 This led to the categorization of cancer-specific miRNAs known as oncomiRs (ie microRNAs with oncogenic or tumor-suppressive activity).3 The central role of miRNAs in cell homeostasis and tissue-specific expression profiles envisage their utility as makers of therapeutic success during cancer treatment, in addition to their role as discriminators of cancer from non-cancerous tissue. For instance, previous studies have reported that the expression ratio of miR-196 and miR-217 can-not only distinguish normal from cancer but also chronic pancreatitis from each other. Furthermore, miR-196-expression also harbors a prognostic relevance in pancreatic carcinomas.4

Pancreatic cancer is the fourth leading cause of cancer-related deaths in the Western world owing to late diagnosis, fast tumor progression and low response rate to recent chemotherapeutic strategies. Hence, further investigation of pancreatic cancer biology is warranted to gain plausible insights into the patho-physiologic mechanisms underlying PDAC-development. An in-depth investigation of the functional role of deregulated microRNAs in PDAC is essential for a deeper understanding of this complex disease. Mir-148a has been earlier described to be down-regulated in several types of solid cancers including gastrointestinal,5 colon, breast, lung, head and neck cancer, as well as melanoma6 and hepatoblastoma.7 The generally observed reduction in miR-148a abundance in several cancers lead to the assumption that miR-148a confers a tumor suppressive role and is important for tumorigenesis. To date there are only limited reports of direct miR-148a targets namely: TGIF2,6 DNMT1,8 DNMT3b9 and MSK1.10

CDC25-proteins are highly conserved dual specificity phosphatases that are vital for appropriate cell cycle progression. In mammals, the CDC25-family consists of three different isoforms (CDC25A, CDC25B and CDC25C) that control the specific activation of distinct CDK/cyclin complexes at different time points during the cell cycle by removing inhibitory phosphate residues from target cyclin-dependent kinases (Cdk).11, 12, 13 Among these, the CDC25B isoform activates the CDK1/cyclinB complex at the G2/M checkpoint, thereby initiating mitosis.14, 15 The tight regulation of CDC25B is pivotal to temporal orchestration of cell cycle progression and hence its over-expression correlates not only with DNA-damage checkpoint response abrogation but also with increased genomic instability.12 Not surprisingly, the CDC25B-induced elevated genomic instability correlates with its oncogenic effects,11, 16, 17 although CDC25B over-expression alone does not lead to malignant tumor formation. CDC25B, in combination with other oncogenes (eg, HRAS mutation or the loss of RB1 in mouse fibroblasts) does serve as an important factor for transformation.18

We investigated if miR-148a directly targeted the 3′UTR of CDC25B and if overexpression of miR-148a impacted the expression of CDC25B on the protein level. We also focused on the impact of miR-148a re-expression in IMIM-PC2 cells and its potential role in proliferation and anchorage-independent growth.

MATERIALS AND METHODS

Cell Culture

The human embryonic kidney (HEK) 293T cell line, the human pancreatic cancer cell line IMIM-PC2 and the human osteosarcoma cell line U2-OS were cultured in DMEM-media supplemented with 10% fetal calf serum, 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen-Gibco, Karlsruhe, Germany). Cells were maintained in a humified incubator with 5% CO2 at 37 °C.

Microdissection and RNA Isolation

Fresh frozen and microdissected normal ductal cells (n=10), and pancreatic adenocarcinoma cells (n=9) were included in our study, respectively. The sample collection was performed according to a protocol approved by the ethics committee of the Ruhr-University Bochum (permission no. 3534–09 and 2392–04). All samples were reviewed by a pathologist (JBM).

Surgical pancreatic resections were immediately placed on ice, and subsequently snap-frozen and stored at −80 °C. For the identification of normal ductal cells, 5 μm frozen sections prepared from tissue blocks of pancreatic parenchyma, in particular from the resection margins, were subsequently placed in RNase free 95% ethanol (Merck, Darmstadt, Germany), stained with H&E and diagnosed by a pathologist (JBM). For normal ductal cell isolation, medium-sized interlobular ducts were selected by preference in order to avoid contamination by acinar tissue. Tissue blocks containing the cells of interest (ie normal ductal cells and pancreatic adenocarcinoma cells, respectively) were serially sectioned (10 μm sections). The slides were stained with methyl green and immediately stored at −20 °C. Areas with cells of interest were manually microdissected under a microscope (BH2, Olympus, Wetzlar, Germany) using a sterile injection needle (size 0.65 × 25 mm). Microdissected cells were placed in 50 μl extraction buffer from the RNAqueous-Micro kit (Ambion/Applied Biosystems, Darmstadt, Germany) and kept on ice. For each sample microdissected cells were collected from 10 to 20 serial sections (each 10 μm). RNA isolation from microdissected cells was performed using the RNAqueous-Micro kit according to the manufacturer's instructions. Total RNA concentration and purity of total RNA samples were measured with the NanoDrop 1000 spectrophotometer (NanoDrop Technologies/Thermo Scientific).

Quantitative RT-PCR

Mature miRNA expression of micro-dissected PDAC-samples, normal ducts and IMIM-PC2-cells were analyzed by TaqMan miRNA-assays (Applied Biosystems). For cDNA-synthesis 50 ng of total RNA were reverse transcribed using the specific looped primer. All analyzed samples were run in triplicates using a 7500 Real-Time PCR System (Applied Biosystems) and were normalized as indicated. Relative fold changes were calculated based on the 2ΔΔCt method. For absolute quantification CT values were converted into copy numbers using a standard curve. The standard curve was generated by spiking a serial dilution of synthetic miR-148a RNA oligonucleotides ranging from 104 up to 108 molecules per reaction into 50 ng E. coli tRNA (Sigma-Aldrich, Taufkirchen, Germany) and analyzed by TaqMan miRNA-assays (Applied Biosystems). The copy number per cell was calculated according to Cheng et al.19

5-aza-2′Deoxycytidine (5-aza-dC) Treatment

IMIM-PC2-cells were treated with the DNA-methylase-inhibitor 5-aza-dC for 5 days. 4 × 105 cells were seeded on 10 cm plates. The following day, cells were treated with 2.5 μM 5-aza-dC and the medium was renewed daily. After 5 days, cells were washed with PBS and subsequently lysed with 3 ml of lysis-buffer (4 M Guanidiniumthiocyanate, 25 mM Sodium Citrate, 0.5% N-Lauroyl-Sarcosin, 0.72% β-mercaptoethanol). Total RNA was then purified by acidic chlorophorm-phenol extraction.

Stable Over-Expression of shmiR-148a in IMIM-PC2 cells

MiR-148a over-expression was performed by cloning the hairpin sense: 5′-ccggacaaagttctgtagtgcactgactcgagtcagtgcactacagaactttgttttttg-3′ and antisense: 5′-aattcaaaaaacaaagttctgtagtgcactgactcgagtcagtgcactacagaactttgt-3′ into the AgeI/EcoRI site of the pLKO.1 puro-vector. Lentiviruses were produced by transfecting packaging cells (HEK293T) with a 3-plasmid system. DNA for transfections was prepared by mixing 12 μg pCMVΔRR8.2, 1 μg pHIT G and 12 μg pLKO.1 plasmid DNA with 62 μl of 2 M CaCl2 in a final volume of 500 μl. Subsequently 500 μl of 2 × HBS phosphate buffer was dropwise added to the mixture and incubated for 10 min at RT. The 1 ml transfection mixture was then added to 50% confluent HEK293T-cells (seeded the day before) into a 6 well plate. Cells were incubated for 16 h (37 °C and 5% CO2), and the medium was changed to remove remaining transfection reagent. Lentiviral-supernatants were collected 36 h post-transfection and for each infection 3 ml supernatant containing 4 μg/ml polybrene was immediately used to infect target cells seeded the day before in 6 well plates to reach 70% confluency on the day of infection. Cells were incubated for 24 h, and then the medium was changed to remove virus particles. To control infection rate, a parallel infection under identical conditions and targeting the same cell line was prepared using a lentiviral-GFP expression control vector (pRRLU6-CPPT-pSK-GFP). Six days after infection 2 μg/ml puromycin was added to the cell culture media.

3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide (MTT) Assay

IMIM-PC2-cells stably over-expressing shmiR-148a and mock-transduced cells were seeded at 5 × 103 cells per well in 12-well plates, respectively. Cell growth activity was monitored after 48, 96 and 120 h by adding MTT reagent (Sigma-Aldrich) (5 mg/ml in PBS) to a final concentration of 0.8 mg/ml for 4 h at 37 °C. After incubation cells were lysed by adding 500 μl Triplex solution (10% (w/v) SDS; 5% (v/v) isobutanol; 12 mM HCl) and incubated overnight at 37 °C. Absorbance was measured at 562 nm with background substraction at 630 nm. All measurements were done in triplicates and normalized to the corresponding t0 value.

Anchorage Independent Growth Assay

For anchorage independent growth 2 × 105 IMIM-PC2-cells were seeded per well on 0.7%. agarose (Invitrogen) in a 0.3% agarose-solution in a 6 well plate. Cells were fed every third day and cultivated over 4 weeks. Cells were stained with Iodonitrotetrazolium-chloride solution overnight. The stained cells were scanned on an Epson scanner.

Luciferase Reporter Assay

Luciferase constructs were made by ligating 200 bp 3′ of the 3′UTR of CDC25B as well as the mutated version after the luc orf in the pGL3. Cells were triple-transfected using Attractene (Qiagen, Hilden, Germany) with 41.5 ng of pGL3 containing either the wild type or mutant CDC25B sequence, in combination with 6.5 nM miR-148a mimics, 20 ng pRL-TK (Promega, Mannheim, Germany) served as transfection control. As a negative control 6.5 nM AllStars siRNA oligos (Qiagen) was used. Firefly and Renilla activity was measured 32 h post transfection. The luciferase-signal was measured for 10 s (Tecan M2000, Crailsheim, Germany). The Renilla-signal was used for normalization. Mean values and s.e.m. were calculated in triplicates.

Transient Transfection of miR-148a Mimics

For transient knock-down experiments 1 × 105 pancreatic carcinoma cells (AsPC1, Capan-1, MIA PaCa-2, Colo357 and PaTu 8092) were seeded in 6 well plates. Cells were transfected in solution with 10 nM miR-148a mimics (Qiagen) using Attractene (Qiagen) according to the manufacturer’s recommendations; as control served 10 nM AllStars (Qiagen). After 24 h the media was replaced. Cells were harvested in RIPA-buffer, 96 h post transfection and subjected to Western blot analysis using CDC25B antibody.

Western Blot Analysis

ShmiR-148a stably over-expressing and mock transducted IMIM-PC2-cells were grown to 80% subconfluency and lysed in RIPA buffer (50 mM Tris-HCl; pH 7.6, 150 mM NaCl, 1% NP40 (Sigma-Aldrich), 0.1% SDS, 0.5% Sodium deoxycholate, 1 mM EDTA, and 1 mM EGTA) with protease inhibitors (Thermo Scientific, Karlsruhe, Germany), and sonicated (Bioruptor, Liége, Belgium) for 15 min with 30 pulses in iced water to prepare whole cell lysates. 30 μg whole cell lysates were subjected to 10% SDS-PAGE and transferred onto PVDF membranes (Millipore, Schwalbach, Germany) with 2 mA/cm2 for 1 h. After protein transfer membranes were blocked in PBS-T, containing 5% (w/v) skimmed milk, for 1 h and incubated with anti-CDC25B antibody (Cell Signalling, Frankfurt am Main, Germany) overnight (1:1000 in PBS-T). As loading control anti-α-tubulin antibody (Sigma-Aldrich) was used at 1:2500 dilution in PBS-T for 1 h at room temperature. Membranes were incubated for detection with secondary antibodies raised against rabbit labeled with CyDye800 (Li-cor, Bad Homburg, Germany) and mouse labeled with CyDye700 (Li-cor) for 1 h at room temperature. Signals were detected by Odyssey Scanner (Li-cor).

Immunohistochemistry

Two μm thick formalin fixed paraffin embedded (FFPE) sections were deparaffinized and rehydrated. Sections were subsequently cooked in a microwave oven for 10 min in target retrieval solution; pH 6.0 (DAKO, Hamburg, Germany). Pretreated sections were then blocked with 2% BSA in TBS-T for 30 min at room temperature before overnight incubation at 4 °C with the CDC25B antibody (1:200; Cellt Signalling, Frankfurt am Main, Germany). After washing with TBS-T the secondary antibody was incubated for 20 min at room temperature followed by coupling the alkaline phosphatase by the DAKO Real™ Detection System, Alkaline Phosphatase/Red, rabbit/mouse (DAKO) according to the manufacturer's manual. CDC25B was detected by chromogen red staining for 4 min. Sections were finally counter stained with Hematoxylin (DAKO).

Statistical Analysis

Data are expressed as mean±s.d. unless otherwise noted. MicroRNA expression of patient samples was statistical analyzed by Mann-Whitney U-Test. For tissue culture based assays the differences between groups were statistically analyzed by a two tailed unpaired Student's t-test. The level of statistical significance was set at P<0.05.

RESULTS

Quantitative Analysis of miR148a Expression in Ductal Pancreatic Adenocarcinoma

Using miRNA-array analysis we identified miR-148a to be frequently down-regulated in PDAC.20 To confirm and validate the miRNA-array data on miR-148a, we first compared the transcription level of miR-148a in microdissected samples of 9 patients with PDAC to 10 samples of normal pancreatic epithelia. The mean transcription level of miR-148a was significantly lower in PDACs compared with non tumor samples by a factor of four (P=0.008) (Figure 1a). To further investigate the potential mechanism underlying the miR-148a down-regulation in PDAC,6, 21 we investigated its expression in the pancreatic cancer tissue cell line IMIM-PC2 post 5-aza-treatment a demethylating reagent. After five days of 5-aza-dC treatment, IMIM-PC2-cells demonstrated a significant 8-fold increase in miR-148a-expression (P<0.05) (Figure 1b). This confirms that miR-148a is silenced in IMIM-PC2 cells through promoter-methylation enabling the use of this cell line as a model system for miR148a overexpression experiments.

Figure 1
figure 1

The miR-148a-expression is significantly down regulated in PDAC. (a) MiR-148a relative expression levels were determined using TaqMan real time qRT-PCR in normal ductal pancreas epithelia (n=10) and PDAC (n=9). (**P< 0.01 by Mann-Whitney U-Test) (b) Induction of miR-148a expression in IMIM-PC2-cells post 5-aza-dC-treatment for 5 days. Relative expression levels were determined using RNU44a as endogenous internal control. The histogram represents mean of three replicates. (**P<0.01 by unpaired two-tailed Student's t-test).

Full size image

MiR-148a Re-expression Suppressed Cell Growth in IMIM-PC2 Cells

As miR-148a-expression was silenced by promoter methylation in IMIM-PC2-cells, we analyzed the functional impact of miR-148a in the pancreatic cancer cell line IMIM-PC2 as a relevant in vitro model for PDAC. We generated stably miR-148a over-expressing IMIM-PC2-cells by lentiviral infection. The validity of miR-148a-expression was confirmed by quantitative RT-PCR: we observed a 160-fold increased miR-148a-expression in comparison to the respective mock transduced cells (Figure 2a). This relative fold change correlates with a re-expression of 1.27 × 105 miR-148a molecules/cell in IMIM-PC2 cells compared to 2.83 × 105 miR-148a molecules/cell in normal pancreatic ductal epithelia. Having shown that our miR-148a level did not exceed the physiological range in normal cells, we analyzed the impact of miR-148a re-expression on IMIM-PC2 cells using MTT and colony-forming assay in this system (Figure 2b and c). After 96 h the MTT assay revealed a significant change in the cell growth (P=0.003). At the end of the assay we observed a reduction in cell growth of 22.83% in the miR-148a over-expressing cells compared to the mock transducted ones (P=0.047). This leads to the assumption that miR-148a overexpression significantly reduces the tumor cell viability (Figure 2b). The colony forming assay showed a reduction of foci for the miR-148a overexpressing IMIM-PC2 cells in comparison to the mock transducted cells after 4 weeks (Figure 2c upper panel). Microscopical examination of the cells revealed that in addition to the increased number of foci the mock transducted cells also form bigger colonies (Figure 2c lower panel). Together, these findings led us to propose that miR-148a might be an important negative regulator in PDAC in terms of growth and proliferation.

Figure 2
figure 2

MiR-148a reduced cell growth and colony formation. (a) Average miR-148a relative expression levels were determined using TaqMan real time qRT-PCR in IMIMPC2-cells stably over-expressing miR-148a. (b) Growth kinetics of mock-IMIM-PC2 and miR-148a overexpressing-IMIM-PC2 over 5 days using MTT assay (*P<0.05 by unpaired two-tailed Student's t-test). (c) Anchorage independent growth of mock-IMIM-PC2 and miR-148a overexpressing-IMIM-PC2 over 5 days using colony forming assay. All experiments were performed in triplicates and mean values are shown with s.d. of triplicates.

Full size image

CDC25B is a Candidate Target of miR-148a

To elucidate the molecular mechanism underlying miR-148a mediated regulation of proliferation we used in silico analysis based on the computer-aided algorithms: PicTar,22 Targetscan23, 24, 25 and miRanda26 in conjunction with Gene Ontology-analysis for predicted target genes. To attribute a gene as a potential target of miR-148a, it had to be predicted independently by at least two of the three applied algorithms and simultaneously be an established positive regulator of cell proliferation. The most promising candidate fulfilling the above criteria was CDC25B, which also contained a putative miR-148a binding site at the end of its 3′UTR (Figure 3a). To prove the functionality of this miR-148a binding-site, we chose to perform a luciferase based reporter assay. Therefore we cloned the 200 bp 3′ of the CDC25B 3′UTR containing the predicted seed sequence downstream of the luciferase (luc) reporter-gene. 32 h post-transfection the luciferase-activity was monitored in the presence of miR-148a mimics or AllStar oligos as a negative control in the human osteosarcoma cell line U2-OS, respectively. The luciferase-reporter activity decreased approximately 38% in U2-OS cells containing the CDC25B wild-type 3′UTR fragment (Figure 3b). In contrast; we observed no significant change in the luciferase-activity in cells expressing the mutated 3′UTR (Figure 3b). These data confirm the direct interaction between miR-148a and the 3′UTR of CDC25B mRNA. Next, we determined the impact of miR-148a on CDC25B protein expression by Western blot analysis in miR-148a stable overexpressing IMIM-PC2 cells. Concordant with our earlier results, we observed that the over-expression of miR-148a suppressed the endogenous CDC25B protein level in IMIM-PC2 cells below the detection limit (Figure 4a). Furthermore, we opted to analyze the impact of miR-148a on the CDC25B expression in a panel of different pancreatic carcinoma cell lines. Therefore, we analyzed the CDC25B level in the following pancreatic carcinoma cell lines: AsPC1, Capan-1, MIA PaCa-2, Colo-357 and PaTu 8092 in the presence and absence of miR-148a mimics. After 96 h a distinct reduction in the CDC25B level was shown in three out of five analyzed cell lines (Figure 4a), whereas Capan-1 cells showed only a slight reduction. The cell line AsPC-1 showed no alteration in the CDC25B expression. In addition to the CDC25B repression in our tissue culture experiments we determined the CDC25B expression in formalin fixed paraffin embedded PDAC samples with different miR-148a expression levels. This analysis showed a predominant nuclear CDC25B staining in samples of PDAC with a reduced miR-148a level (Figure 4c), compared to normal ductal cells and to PDAC with high miR-148a expression (Figure 4b and d). Collectively, we were able to show that the microRNA miR-148a down-regulated the expression of CDC25B in different pancreatic cell lines (ie IMIM-PC2, Capan-1, MIA PaCa-2, Colo347 and PaTu 8092), as well as in patient samples of PDAC.

Figure 3
figure 3

Suppression of CDC25B by miR-148a. (a) Sequence complementarity between hsa-miR-148a and the 1135–1141 bp of the CDC25B 3′UTR. * indicates mutated base pairs. (b) Interaction between miR-148a and 3′UTR of CDC25B. U2-OS cells were transfected with wild-type (WT) and mutant (Mut) 3′UTR of CDC25B containing luciferase reporter vector together, miR-148a and control vector respectively as indicated. (**P<0.01 by unpaired two-tailed Student's t-test)

Full size image
Figure 4
figure 4

Alteration of the CDC25B expression by miR-148a. (a) Western blot analysis demonstrating reduction in CDC25B protein in IMIM-PC2 cells stably overexpressing shmiR-148a. For transient CDC25B knock-down experiments, the pancreatic carcinoma cell lines AsPC-1, Capan-1, MIA PaCa-2, Colo-357 and PaTu 8092 were transfected with miR-148a mimics or control (10 nM). After 96 h the cells were harvested and the CDC25B levels were determined by Western blot analysis. Anti-α-tubulin served as loading control. CDC25B immunohistochemical staining of normal ductual cells (b) compared to PDAC section with down-regulated miR148a (c) and PDAC with high miR148a transcription-level (d). The most intense CDC25B signal intensity was shown for section (c), further supporting the link between miR148a transcription and down-regulation of CDC25B expression.

Full size image

DISCUSSION

Mir-148a has been reported as a microRNA with tumor-suppressor activity in several types of cancers in which it is down-regulated through epigenetic silencing by promoter hypermethylation.6, 21, 27 In colon, lung, breast, and head and neck carcinoma and melanoma, miR-148a-promoter hypermethylation has been correlated with a higher metastatic potential.6 Hanoun et al have reported that hypermethylation of the promoter region encoding miR-148a is responsible for its repression not only in PDAC samples but also in preneoplastic-pancreatic-intraepithelial-neoplasia (PanIN) and can serve as an ancillary marker for the differential diagnosis of PDAC and chronic pancreatitis.21 Consistent with this observation, we report that miR-148a is significantly down-regulated in micro-dissected PDAC samples. Therefore the expression of miR-148a in TGFβ non-responsive IMIM-PC2 pancreatic cancer cells is probably regulated by DNA-methylation. In the present study, we first established an in vitro model based on lentiviral-mediated stable miR-148a overexpression in the pancreatic carcinoma cell line IMIM-PC2 (in which cellular miR-148a is silenced by promoter-methylation). We subsequently exploited this in vitro model to validate the tumor-suppressor role of miR-148a in PDAC by demonstrating that miR-148a over-expression had an inhibitory influence on the growth potential of pancreatic cancer cells.

Recently Fujita et al have demonstrated that miR-148a, which is part of a reported DNA methylation-signature for human cancer metastasis, attenuates paclitaxel-resistance of hormone-refractory, drug-resistant PC3PR cells in part by regulating MSK1 expression.6, 10 However, as miRNAs are expected to have multiple targets, we sought to further investigate and expand the putative targets of miR-148a in PDAC. In silico analysis revealed five putative targets, which have positive effects on cell cycle progression and contain a seed sequence for miR-148a binding in their 3′UTR. Of these candidates, CDC25B emerged as the most promising novel candidate.

CDC25B is frequently dysregulated in a broad variety of cancers,12, 28 and its overexpression has been reported in PDAC.29, 30 Further CDC25B has been reported to have prognostic relevance in breast,31 prostate,32 colorectal33, 34 and gastric cancer,35 in which the overexpression of CDC25B seemed to correlate with a poor prognosis. However, the mechanism which leads to the dysregulation of the CDC25B within the specific context of tumor proliferation and progression has largely remained elusive. The poor correlation between CDC25B mRNA and its corresponding protein abundance has often led to the speculation that the deregulation of CDC25B may occur at any stage between its transcription and translation.12 Our results for the first time indicate a down-regulation of CDC25B by miR-148a at the post-transcriptional level by translational repression. This proposed interaction is evidenced from our reporter assay based studies where miR-148a has been directly shown to influence the expression of the luc reporter by binding to the predicted target sequence in the CDC25B-3′UTR. Interestingly, increased CDC25B level can lead to a premature exit of the G2/M checkpoint,16 resulting in the loss of DNA damage repair mechanisms or the initiation of PCD and hence miR-148a-CDC25B-association is consistent with a model of pancreatic cancer progression in which miR-148a is lost in PanIN pre-neoplastic lesions. Accordingly, in PDAC it has been shown that there was no correlation between the percentage of CDC25B positive cells and tumor stage/grade implicating dysregulation of CDC25B as an early event in the current multi-step PDAC progression model.30 Based on these collective observations, we propose a model in which the control of the CDC25B expression is epigenetically pre-regulated in PanINs via miR-148a promoter-methylation and continues to confer proliferative advantage in PDAC via CDC25B. Further this early repression of miR-148a and the resultant increase in CDC25B expression may additionally confer an increased propensity for genomic instability at an early stage in the PDAC progression model, thus facilitating a higher risk to accumulate genetic defects and subsequent enhanced tumor progression.

In summary, using an in vitro model based on lentiviral-mediated stable miR-148a over-expression in the pancreatic carcinoma cell line IMIM-PC2, we validate the tumor-suppressor role of miR-148a in PDAC. Further, our results for the first time demonstrate that CDC25B is a direct target of miR-148a. Our data link the postulated promoter-methylation of miR-148a at early stages of pancreatic-tumorigenesis (PanIN)21 to an increased CDC25B expression in PanINs30 which leads to a PDAC model in which the transcriptional loss of miR-148a and the resultant increase in CDC25B expression facilitates increased genomic instability at an early stage of tumor development.