Ten PDAC and 13 IPMN tissue samples were obtained from patients who underwent surgical resection in Miyagi Cancer Center Hospital from 2010 to 2012. Histologic diagnosis of each sample was performed by two pathologists who were not informed about the present study. Histologically, cancer duct cells and non-tumor cells (mixed with acinar cells, inflammatory duct cells and stromal cells) from PDAC tissues (n=10), and non-invasive IPMN cells (adenoma and carcinoma in situ, n=13) were microdissected and subjected to RNA extraction. These paraffin-embedded tissues were cut into 10-μm sections and ~10 sequential regions from the same paraffin block were microdissected using a Leica CIR MIC system (Leica Microsystems, Wetzkar, Germany). Total RNA including miRNA was extracted using the Recover All Total Nucleic Acid Isolation kit (Ambion, Austin, TX, USA), according to the manufacturer’s instructions.

Between April 2011 and March 2013, the plasma samples of 32 PDAC patients, 30 healthy controls (HC) and 12 IPMN patients were collected at Miyagi Cancer Center. Patient characteristics with respect to age, sex, and stages of disease are described in Table II. The stage of PDAC was assessed according to the Union for International Cancer Control (UICC) Classification (43). All patients were pathologically diagnosed as having PDAC using surgical (n=8) or biopsy specimens [EUS-FNA (n=20), endoscopic transpapillary biopsy (n=3) and biopsy from duodenal invasion (n=1)]. As a control, plasma samples were collected from 30 HC. HC were 22 medical staff members and 8 hospitalized patients with benign disease such as asymptomatic cholecystolithiasis or choledocholithiasis. They underwent medical examinations and did not have any pancreatic disease. All IPMN patients were the branch duct type of IPMNs (BD-IPMN). In this study, the definition of BD- IPMN was based on imaging as a grapelike multilocular cystic lesion without mural nodules communicating with the MPD by computed tomography (CT), magnetic resonance cholangiopancreatography (MRCP), endoscopic ultrasonography (EUS), and/or endoscopic retrograde cholangiopancreatography (ERCP).

Blood samples were collected from patients and controls in sodium EDTA-Na tubes and were immediately centrifuged at 3500 rpm for 10 min at room temperature. Then the plasma supernatant was collected and stored at −80°C until further analysis. Total RNA containing small RNA was extracted from 300 μl plasma samples using a mirVana PARIS kit (Ambion), undiluted into 100 μl of preheated (95°C) elution solution according to the manufacturer’s instructions. To normalize sample-to-sample variation in the RNA isolation step, 2 μl of synthetic C. elegans miR-39, which lacks sequence homology to human miRNAs, was added to each denatured sample.

The amount of miRNAs was quantified in duplicate via quantitative RT-PCR using the human TaqMan MicroRNA Assay kit (Applied Biosystems, Foster City, CA, USA). The reverse transcription reaction was carried out with a TaqMan MicroRNA Reverse Transcription kit (Applied Biosystems) in 15 μl containing 5 μl of RNA extract, 0.15 μl of 100 mM dNTPs, 1.5 μl of 10× reverse transcriptase buffer, 1 μl of Multiscribe reverse transcriptase (50 U/μl), 0.19 μl of RNase inhibitor (20 U/μl), 3 μl of gene-specific primer, and 4.16 μl of Nuclease-free water. For the synthesis of cDNA, reaction mixtures were incubated at 16°C for 30 min, at 42°C for 30 min, and at 85°C for 5 min, and then held at 4°C. Next, 15 μl of cDNA solution was amplified with 25 μl of TaqMan 2× Universal PCR Master Mix with no AmpErase UNG (Applied Biosystems), 2.5 μl of gene-specific primers/ probe, and 7.5 μl of nuclease-free water in a final volume of 50 μl. Quantitative RT-PCR was run on a LightCycler^® 480 real-time PCR system (Roche), and reaction mixtures were incubated at 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec, and 60°C for 1 min. Cycle threshold (Ct) values were calculated with the LightCycler 480 for Software Version 1.5 (Roche). The relative expression of the mature miRNAs was calculated using the comparative CT (2^−2ΔΔCt) method with miR-16 and RNU6B for plasma and tissue samples, respectively, as the endogenous control to normalize the data (44,45).

Student’s t-test or Mann-Whitney U test was used to evaluate differences in the miRNA expression between PDAC or IPMN cases and controls. Receiver operating characteristic (ROC) curves were constructed and the area under the curve (AUC) was calculated to evaluate the sensitivity and specificity for predicting cases and controls based on the expression of each individual miRNA and their combinations. Fisher’s exact test or Pearson’s χ² test was used to determine if there was a significant association between the clinicopathological factors and the relative plasma expression levels of miRNAs. Survival rate was estimated by the Kaplan-Meier method. The difference in survival rates between the groups was tested for significance using the log-rank test. The overall survival was defined as the period from the date of diagnosis for PDAC to the date of death or last follow-up. All tests of statistical significance were two-sided. A P-value of <0.05 was considered statistically significant. All statistical analyses were done using Excel 2010 software.

The Institutional Review Board of the Miyagi Cancer Center (MCC) approved this study protocol, and written informed consent was obtained from each patient (permit MCC-24-38).

To validate the results of the previous comprehensive analysis, we first examined the miR-483-3p and miR-21 expression levels in a new series of PDAC and IPMN tissues. The results for these miRNAs, normalized with RNU6B are shown in Fig. 1. In microdissected PDAC cells, the mean level of miR-483-3p expression was 19.36±5.23 [miR-483-3p/RNU6B, mean ± standard error (SE)], and this expression level was significantly higher than that of non-tumor cells (3.26±1.22, P=0.0015). The miR-483-3p expression in PDAC cells was also significantly higher than in IPMN lesions (4.07±0.75, P=0.0010). Although there was no statistically significant difference between IPMN cells and non-tumor cells, the miR-483-3p level tended to be slightly higher in the IPMN samples than in the normal tissue, suggesting that the increased expression of miR-483-3p occurs in the transition from benign to malignant. Similarly, the expression of miR-21 was significantly higher in PDAC [23150.61±4493.76 (miR-21/RNU6B, mean ± SE)] than IPMN (1413.05±319.18, P=0.0001) and non-tumor cells (580.45±109.06, P=0.0002). On the other hand, miR-21 expression was significantly higher in IPMN compared to non-tumor cells, suggesting that miR-21 contributes to an early step of pancreatic tumorigenesis.

Next, we hypothesized that higher miR-483-3p expression in primary carcinoma cells would influence the plasma levels of miR-483-3p in PDAC patients. To determine the quantities of plasma miRNAs, we used cell-miR-39 as internal control miRNAs, and miR-16 was used for the normalization. Plasma miR-483-3p expression was calculated from the comparative Ct method by quantitative RT-PCR. Using this assay, circulating miR-483-3p was detectable in all plasma samples from 32 PDAC patients, 30 HC and 12 IPMN patients (Fig. 2A). The mean level of plasma miR-483-3p expression in the PDAC patients [16.50±6.49 (miR-483-3p/miR-16, mean ± SE)] was significantly higher than in HC (1.37±0.11, P=0.0006) and IPMN patients (1.90±0.39, P=0.0389) (Fig. 2B). We also evaluated the expression level of plasma miR-21, which was previously shown to be overexpressed in the plasma and tissue of PDAC (26,36–43). The mean expression level of plasma miR-21 (Fig. 2C) was significantly higher in the PDAC group [765.25±64.6 (miR-21/miR-16 ± SE)] than HC (414.54±44.7, P=0.0001). The plasma expression level of miR-21 in IPMN patients was also higher than in HC (P=0.0394), and no differences were seen between PDAC and IPMN (P=0.2979). While no difference was observed in the plasma miR-21 expression level between PDAC and IPMN patients, miR-483-3p expression in the plasma of PDAC patients was significantly higher than in that of IPMN patients. This result indicates that plasma miR-483-3p measurement has the potential to distinguish PDAC not only from healthy individuals, but also from the patient of IPMN.

To apply the expression value of plasma miRNAs for the diagnosis of PDAC, we then analyzed the ROC curves. The ROC curves for differentiating between PDAC and HC, and/or IPMN based on the expression level of miRNAs (miR-483-3p or miR-21) are shown in Fig. 3. The AUC of miR-483-3p (0.754) was slightly lower than that of miR-21 (0.790) for differentiating PDAC from HC. On the other hand, for discriminating PDAC from IPMN patients, the AUC of miR-483-3p (0.703) was higher than that of miR-21 (0.603), In addition, the AUC value based on the expression level of plasma miR-483-3p (0.740) and miR-21 (0.736) to distinguish PDAC patients from non-cancer individuals (HC and IPMN) was similar. Consequently, we compared the ability of measuring the plasma miR-483-3p and miR-21 to predict PDAC with that of the conventional tumor markers CEA and CA19-9. The respective AUC values of plasma miR-483-3p and miR-21 were higher than that of CEA (0.719), but lower than that of CA19-9 (0.866). Interestingly, the AUC (0.839) from the ROC curves for the combination of miR-483-3p and miR-21 was similar to that of CA19-9 for discriminating patients with PDAC from non-cancer individuals.

Finally, we evaluated the correlation between the plasma level of miRNAs and clinicopathological factors in 32 PDAC patients (Table III). We judged the miR-483-3p level as ‘high expression’ when it was higher than a cut-off value of 5. There was no significant relationship between the plasma miR-483-3p level and clinical factors with overall survival. The plasma miR-21 expression level was associated with advanced stage (P=0.023), metastasis to lymph node (P=0.007) and liver (P<0.001) when the cut-off value was defined as 850, as shown in Table IV. Moreover, overall survival was significant shorter in the high miR-21 expression group than in the low expression group when we analyzed by the Kaplan-Meier method in 24 unresectable PDAC patients, (3.0 months vs 13.8 months, P<0.001) (Fig. 4).

Since the plasma miR-21 level was not increased in operable patients (Table IV), we evaluated alterations in the plasma miR-483-3p level alone after resection of the tumor tissues. Fig. 5 shows representative cases of the plasma expression level of miR-483-3p and tumor markers before and after surgery. Case 1 was a 70-year-old female and case 2 was a 62-year-old female. In both cases, pancreaticoduodenectomy was performed for the diagnosis of pancreatic head cancer. By histological examination, the diagnosis was pancreatic adenocarcinoma and the classification stage IIA (T3N0M0) using UICC classification. As expected, a marked reduction of the plasma miR-483-3p expression as well as the CA19-9 value was observed after surgery, indicating that the expression of miR-483-3p was strongly associated with the presence of PDAC.