Downregulation of microRNA-181d had suppressive effect on pancreatic cancer development through inverse regulation of KNAIN2

Abstract

We explored the expression and function of miR-181d (microRNA-181d) in human pancreatic cancer. Quantitative real-time polymerase chain reaction was used to probe miR-181d expression in both pancreatic cancer cell lines and human pancreatic carcinoma. Pancreatic cancer cell lines, PANC-1 and AsPC-1 cells, were engineered to stably downregulate endogenous miR-181d through lentiviral transduction. The mechanistic effects of miR-181d downregulation on pancreatic cancer development were tested by proliferation, migration, fluorouracil chemosensitivity assays in vitro, and explant assay in vivo. Possible miR-181d downstream gene, NKAIN2 (Na+/K+ transporting ATPase interacting 2), was tested by dual-luciferase activity assay and quantitative real-time polymerase chain reaction. Functional involvement of NKAIN2 in miR-181d-regulated pancreatic cancer development was tested by small interfering RNA–mediated NKAIN2 knockdown in miR-181d-downregulated PANC-1 and AsPC-1 cells. MiR-181d was upregulated in both pancreatic cancer cell lines and human pancreatic carcinoma. Lentivirus-induced miR-181d downregulation decreased pancreatic cancer proliferation, migration, and fluorouracil resistance in vitro and inhibited the growth of cancer explant in vivo. NKAIN2 was directly targeted by miR-181d in pancreatic cancer. Small interfering RNA–mediated NKAIN2 knockdown reversed the inhibition of miR-181d downregulation on pancreatic cancer development. MiR-181d is aberrantly overexpressed in pancreatic cancer. Inhibiting miR-181d may suppress pancreatic cancer development, possibly through the inverse regulation on NKAIN2.

Introduction

Pancreatic cancer is one of the most deadly malignancies in the world.^1,2 In developing country of China, not only had the estimated cases of pancreatic cancer surpassed the estimated numbers in the United States but also the mortality rates are very high, almost close to the diagnosis rates.^3,4 Even worse, pancreatic carcinomas are often highly metastatic, and the prognosis of patients with pancreatic carcinoma is extremely poor, with mean 5-year survival rate to be smaller than 10%.^1,3 Therefore, it is very important to decipher the molecular mechanisms of pancreatic cancer tumorigenesis, proliferation, and metastasis in order to develop efficient therapeutic reagents to target pancreatic carcinoma growth and improve patients’ overall survivals.

MicroRNAs (miRNAs) are groups of short-length (18–22 n.t. long) revolutionarily conserved noncoding RNAs that post-transcriptionally suppressed gene or protein production by attaching to the complimentary 3′–untranslated regions (3′-UTR) of downstream targeting genes.^5,6 In human cancers, including pancreatic cancer, miRNAs have been demonstrated to play critical roles in almost all aspects of cancer development and regulations.^7,8 Among them, family of human mature miR-181 (microRNA-181), including miR-181a, miR-181b, miR-181c, and miR-181d, has been shown to be aberrantly expressed in various types of human cancers and played important roles in regulating cancer development, such as metastasis, apoptosis, as well as predicting prognosis among cancer patients.^9–11 In human pancreatic cancer, microarray study demonstrated that members of miR-181 family were very likely to be upregulated in pancreatic tumorous tissues.^12,13 In addition, a recent study demonstrated that one member of miR-181 family, miR-181a, was a tumor oncogene in pancreatic cancer and actively regulated cancer proliferation and migration. ¹⁴ However, it is not clear, for other members of miR-181 family, what the expression patterns are and what the modulatory mechanisms do they have.

NKAIN2 (Na⁺/K⁺ transporting ATPase interacting 2) is a gene that encodes a trans-membrane protein interacting with β-subunit of a Na⁺/K⁺-ATPase and is implicated in complex ion transportation among neural populations. ¹⁵ In various human cancers, NKAIN2 gene is found to be recurrently truncated, possibly located in the 6q commonly deleted region. ¹⁵ It was suggested that NKAIN2 might act as a tumor suppressor in human cancer as it induces chromosome loss, mutation, and promoter methylation.^15,16 Interestingly, in another human cancer, NKAIN2 was found to be highly expressed in cancer cells of neuroblastoma and subsequently downregulated in retinoic acid differentiation, thus suggesting an oncogenic role in neuroblastoma. ¹⁷ Nevertheless, the role of NKAIN2 in human pancreatic cancer, whether to be tumor suppressor or oncogene, was never elucidated.

In this study, we first used much accurate, much quantitative method, quantitative real-time polymerase chain reaction (qRT-PCR), to examine the expression pattern of one member of miR-181 family, miR-181d, in both pancreatic cancer cell lines and biopsy tissues from pancreatic cancer patients. We then applied lentivirus to stably downregulate miR-181d in pancreatic cancer cell line PANC-1 and AsPC-1 cells. Subsequently, we used several biochemical assays to examine the functional effects of miR-181d downregulation on pancreatic cancer development both in vitro and in vivo. We also hypothesized that NKAIN2 is the downstream target gene of miR-181d in pancreatic cancer. This hypothesis was examined by dual-luciferase activity assay and siRNA-mediated knockdown of NKAIN2 in miR-181d-downregulated pancreatic cancer cell lines.

Materials and methods

Pancreatic cancer cell lines and human tissues

In this study, a normal human pancreatic duct epithelial-like cell line, hTERT-HPNE, and pancreatic cancer cell lines PANC-1, AsPC-1, AsPC-2, Capan-1, Capan-2, Mia PaCa-2, SW1990, BxPC-3, HPAC, and HPAF were all purchased from Cell Bank of Type Culture Collection of Chinese Academy of Sciences in Shanghai, China. All cells were maintained in six-well plate with culture medium of RPMI 1640 (Thermo Fisher Scientific, USA) supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scientific), 100 µg/mL streptomycin, and 100 units/mL penicillin (Thermo Fisher Scientific, USA) in a 37°C Thermo Scientific Forma™ 310 Direct-Heat Incubator (Thermo Fisher Scientific) with 5% circulating CO₂.

Between June 2012 and June 2016, pancreatic carcinoma tissues and their adjacent non-carcinoma pancreatic tissues were surgically obtained from 37 patients with pancreatic cancer from the Department of Biliary and Pancreatic Surgery and Department of Oncology at Tongji Hospital of Tongji Medical College. Immediately after surgery, all obtained human tissues were snap-frozen in liquid nitrogen and stored in a −80°C bio-freezer until RNA extraction.

RNA extraction and qRT-PCR

From pancreatic cancer cell lines and pancreatic human tissues, RNA was extracted with a TRIzol kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. RNA concentration was measured with a NanoDrop ND-2100 spectrophotometer (Thermo Fisher Scientific). Complimentary DNA (cDNA) was reversely transcribed using a PrimeScript RT Reagent Kit (TaKaRa, China) along with miRNA-specific stem-loop primers (Applied Biosystems, USA). To determine human miR-181d expression, qRT-PCR was carried out using a TaqMan MicroRNA Assays Kit (Applied Biosystems) according to the manufacturer’s protocol. To determine human NKAIN2 messenger RNA (mRNA) level, an SYBR qRT-PCR Assay (TaKaRa) was applied. The loading control was small nucleolar RNA (snoRNA) combination of RNU38B/RNU49A. Relative gene expressions were calculated as folder changes relative to loading control and then normalized to control samples using the 2^−ΔΔCt method.

Lentiviral transduction assay

To inhibit hsa-miR-181d in pancreatic cancer cell lines, a pLV-[hsa-miR-181d-5p] locker lentivirus (L-miR181d-I) was purchased from Biosettia (USA). A corresponding pLV-miR-locker control lentivirus, L-C, was also purchased from Biosettia. Pancreatic cancer cell lines, PANC-1 and AsPC-1, were transduced with L-C or L-miR181d-I in the presence of 8 µg/mL polybrene for 48 h at multiplicity of infection (MOI) of 20–30; 48 h after transduction, lentivirus/polybrene-containing medium was replaced with fresh medium plus blasticidin (50 µg/mL). Cells were maintained for another 7–10 days to stabilize lentiviral transduction and viral selection. At the end of the stabilization, cells were re-suspended and re-seeded in six-well pate for further experiments. Transduction efficiency was then measured by qRT-PCR.

Cancer proliferation assay

PANC-1 and AsPC-1 cells were suspended and re-seeded in 100 µL in 96-well plate (3000 cells/well). Caner proliferation was measured using a Vybrant MTT Cell Proliferation Assay Kit (Molecular Probes, USA) for five consecutive days. Each 24 h, culture medium was mixed with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) stock solution for 4 h at 37°C, followed by another 4 h treatment of HCl-SDS to dissolve formazan. Optical density was then measured at 570 nm using a PowerWave XS spectrophotometer reader (BioTek, USA).

Cancer migration assay

Transwell inserts (8-µm pore size; Corning, USA) were pre-coated with Matrigel for 24 h and placed in a 24-well Transwell assay. PANC-1 and AsPC-1 cells were suspended and re-seeded in the upper chamber of Transwell insert in RPMI 1640 medium with 2% FBS (50,000 cells/well). The lower chamber was filled with RPMI 1640 medium plus 10% FBS; 24 h after the onset of Transwell assay, the upper chamber was discarded and medium aspirated. Migrating PANC-1 and AsPC-1 cells in lower chamber were fixed by 4% paraformaldehyde (4% PFA; Sigma-Aldrich, USA) and immune-stained with 0.1% crystal violet. Images were then captured using an Olympus inverted microscope at 40× magnification (Olympus, Japan). Cell numbers were tallied from three random fields (200 µM × 200 µM) in each well. For each experimental condition, migration capability was characterized as the percentage of cell numbers relative to the cell numbers under control condition.

Cancer chemosensitivity assay

PANC-1 and AsPC-1 cells were suspended and re-seeded in 100 µL in 96-well plate (3000 cells/well). Various concentrations of 5-FU (0–500 µM) were added into culture for 72 h, followed by an MTT assay. Cancer cell viability was then characterized as the percentage of optical density for each experimental condition relative to the optical density under control condition.

In vivo pancreatic cancer explant assay

PANC-1 cells were subcutaneously inoculated into the abdominal flanks of 6-week-old BALB/c nu/nu mice (1 × 10⁶ cells/inoculation). The left flanks were inoculated with cells transduced with L-C, and the right flanks with cells transduced with L-miR181d-I. The growth of in vivo pancreatic cancer explant was monitored for 5 weeks. Every week, explant volumes (mm³) were calculated for both flanks using the formula, L × W²/2, where L was the length and W the width of explant. Mice were euthanized using CO₂ inhalation 5 weeks after explant assay, and PANC-1 explants were extracted.

Dual-luciferase activity assay

The 3′-(UTR) of wild-type human NKAIN2 gene, including the putative binding sequence complimentary to mature human miR-181d, was amplified from a human cDNA library and then cloned into psiCHECK-2 firefly luciferase plasmid (Promega, USA). The miR-181d binding sequence on NKAIN2 3′-UTR was point-mutated. And the mutant NKAIN2 3′-UTR was cloned into another psiCHECK-2 plasmid to produce NKAIN2-Mutant luciferase plasmid. In a luciferase co-transfection system, HEK-293T cells were co-transfected with NKAIN2 or NKAIN2-Mutant luciferase plasmids, along with a miR-181d-mimics plasmid (pLV-[hsa-mir-181d]; Biosettia) or a control miRNA plasmid (pLV-[mir-control]; Biosettia). Relative luciferase activities were measured using a Dual-Luciferase Reporter Assay System (Promega) 48 h after co-transfection.

Western blot assay

Western blot analysis on NKAIN2 protein level was performed according the method described previously. ¹⁷ Briefly, PANC-1 and AsPC-1 cells were collected in 2 mL Eppendorf tubes and treated with a lysis buffer (Sigma-Aldrich, China). Equal amount of proteins (30 µg) of each sample was resolved on 8% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and then electro-transferred onto polyvinylidene fluoride (PVDF) membranes (DuPont, USA). The membranes were probed with rabbit polyclonal anti-NKAIN2 antibody (HPA045860; Sigma-Aldrich, China) and mouse monoclonal anti-beta-actin (A5441; Sigma-Aldrich, China), followed by probe with peroxidase-coupled secondary antibodies.

SiRNA transfection assay

Human NKAIN2-specific ON-TARGETplus siRNA was purchased from Dharmacon (Si-NKAIN2; GE Healthcare Dharmacon, USA). A non-specific ON-TARGETplus control siRNA was also purchased from Dharmacon (si-C; GE Healthcare Dharmacon). PANC-1 and AsPC-1 cells were transfected with 50 nM Si-NKAIN2 or Si-C in the presence of DharmaFECT 1 Transfection Reagent (GE Healthcare Dharmacon). NKAIN2 knockdown in PANC-1 and AsPC-1 cells was checked by qRT-PCR 24 h after transfection.

Statistical analysis

In our study, all procedures were carried out in triplicates or more biological repeats. Averaged values were then shown as mean ± standard errors. Statistical analysis was carried out on a windows-based SPSS software (version 12.0; SPSS, USA). For comparison between two independent data sets, student’s t-test was used. For comparison between three or more independent data sets, one-way analysis of variance (ANOVA) and Tukey post hoc test were used. Significant differences were termed if p < 0.05.

Results

MiR-181d is upregulated in pancreatic cancer

Recent microarray studies suggested that miR-181 family, including miR-181a/b/c/d, was highly expressed in pancreatic cancer.^12,13 In this work, we applied much quantitative and much accurate measurement of qRT-PCR to examine the exact expression pattern of miR-181d in both pancreatic cancer cell lines and pancreatic carcinoma in human patients. It showed that in nine immortal pancreatic cancer cell lines, miR-181d was predominantly upregulated, as compared to a normal human pancreatic duct epithelial-like cell line, hTERT-HPNE (Figure 1(a), *p < 0.05). We also compared miR-181d expression between pancreatic carcinoma tissues and non-carcinoma pancreatic tissues in 37 patients. In consistent to the findings in cell lines, it showed that miR-181d was significantly upregulated in pancreatic carcinoma (Figure 1(b), *p < 0.05).

Figure 1.

Analysis of qRT-PCR on miR-181d in pancreatic cancer. (a) Quantitative RT-PCR (qRT-PCR) was conducted to compare mature hsa-miR-181d expression between a normal human pancreatic duct epithelial-like cell line hTERT-HPNE and immortal pancreatic cancer cell lines PANC-1, AsPC-1, AsPC-2, Capan-1, Capan-2, Mia PaCa-2, SW1990, BxPC-3, HPAC and HPAF (*p < 0.05). (b) QRT-PCR was also performed to compare hsa-miR-181d expression between non-carcinoma pancreatic tissues and pancreatic carcinoma extracted from 37 patients diagnosed with pancreatic cancer (*p < 0.05).

MiR-181d downregulation had tumor-suppressive effect on pancreatic cancer development in vitro

Using two pancreatic cell lines, PANC-1 and AsPC-1 cells, we created pancreatic cancer cells with stable miR-181d downregulation through lentiviral transduction. After transduction was stabilized, qRT-PCR showed that in PANC-1 and AsPC-1 cells transduced with miR-181d inhibitor lentivirus (L-miR181d-I), endogenous miR-181d expression was significantly suppressed than in pancreatic cells transduced with L-C (Figure 2(a), *p < 0.05).

Figure 2.

MiR-181d downregulation suppressed pancreatic cancer development in vitro. (a) PANC-1 and AsPC-1 cells were transduced with miR-181d locker lentivirus (L-miR181d-I) to suppress endogenous miR-181d expression. In parallel, control pancreatic cancer cells were transduced with a control miRNA lentivirus, L-C. QRT-PCR confirmed that miR-181d expressions were significantly downregulated by miR-181d-I transduction in PANC-1 and AsPC-1 cells (*p < 0.05). (b) After lentiviral transduction was stabilized, PANC-1 and AsPC-1 cells were examined by an MTT proliferation assay (**p < 0.05, one-way ANOVA). (c) After lentiviral transduction was stabilized, PANC-1 and AsPC-1 cells were examined by a Transwell migration assay. Upper chambers were removed 24 h after plating. PANC-1 and AsPC-1 cells migrating into lower chambers were stained with crystal violet. (d) Migrating capabilities were compared between PANC-1 and AsPC-1 cells transduced with L-C, and those cells transduced with L-miR181d-I (*p < 0.05). (e) After lentiviral transduction was stabilized, PANC-1 and AsPC-1 cells were examined by an in vitro chemosensitivity assay with 72 h treatment of 5-FU at various concentrations from 1 to 500 µM. Cancer viability was measured using an MTT assay and compared between PANC-1 and AsPC-1 cells transduced with L-C, and those cells transduced with L-miR181d-I (**p < 0.05, one-way ANOVA).

We then examined the effect of miR-181d downregulation on pancreatic cancer in vitro proliferation. We plated lentivirus-transduced PANC-1 and AsPC-1 cells in 96-well plates and conducted a 5-day MTT proliferation assay. The daily proliferation rates were measured as optical density at 570 nm and compared between L-C-transduced and L-miR-181d-In-transduced pancreatic cancer cells. The result of 5-day MTT assay showed that miR-181d downregulation significantly suppressed in vitro proliferation in both PANC-1 and AsPC-1 cells (Figure 2(b), **p < 0.05, one-way ANOVA).

A Transwell assay was also carried out to examine the effect of miR-181d downregulation on pancreatic cancer in vitro migration. Cells were plated in upper chamber after 24 h, and immunostaining of lower chambers showed that much less PANC-1 or AsPC-1 cells migrated into lower chamber while they were transduced with L-miR181d-I than they were transduced with L-C (Figure 2(c)). Measurement on relative migration demonstrated that migration capabilities in PANC-1 and AsPC-1 cells were significantly suppressed by 40%–50% (Figure 2(d), *p < 0.05).

In addition, a chemosensitivity assay was conducted to examine the effect of miR-181d downregulation on pancreatic cancer chemo-response to 5-FU. The result showed that at various concentrations of 5-FU, chemosensitivities of PANC-1 and AsPC-1 cells were significantly increased by miR-181d downregulation (Figure 2(e), **p < 0.05, one-way ANOVA).

Thus, the results of in vitro biochemical assays on lentiviral-transduced pancreatic cancer cells demonstrated that miR-181d downregulation had significant tumor-suppressive effect by inhibiting cancer cell proliferation, migration, and increasing 5-FU chemosensitivity.

MiR-181d downregulation had tumor-suppressive effect on pancreatic cancer explant in vivo

After lentiviral transduction, PANC-1 cells were subcutaneously inoculated into the abdominal flanks of 6-week-old BALB/c nu/nu mice. The volumes of in vivo pancreatic cancer explants were monitored for 5 weeks and compared between L-C-transduced and L-miR181d-I-transduced tumor explants. Similar to the in vitro tumor-suppressive effect, the in vivo pancreatic cancer explant essay also showed that miR-181d downregulation significantly suppressed the development of pancreatic cancer explant (Figure 3(a), **p < 0.05, one-way ANOVA). At the end of in vivo explant assay, mice were sacrificed. Subcutaneous explants were dissected out and compared. It clearly showed that L-miR-181d-I-transduced APNC-1 explants were much smaller than L-C-transduced explants (Figure 3(b)).

Figure 3.

MiR-181d downregulation suppressed pancreatic cancer explant in vivo. (a) After lentiviral transduction was stabilized, PANC-1 cells were subcutaneously injected into abdominal flanks of 6-week-old null mice. Tumor volumes were measured weekly for 5 weeks and compared between PANC-1 explants transduced with L-C and those transduced with L-miR181d-I (**p < 0.05, one-way ANOVA). (b) PANC-1 explants were dissected out from mouse abdominal flanks and imaged 5 weeks after transplantation.

NKAIN2 gene is regulated by and inversely correlated with miR-181d in pancreatic cancer

As we demonstrated that miR-181d downregulation had tumor-suppressive effect on pancreatic cancer development in vitro and in vivo, we wondered what the downstream targeting pathways would be for miR-181d-mediated regulation in pancreatic cancer. Through the investigation on several online miRNA targeting algorithms, we identified that human gene of Na⁺/K⁺ transporting ATPase interacting 2 (NKAIN2) is one of the candidate gene, as its 3′-UTR hosts a DNA sequence complimentary to miR-181d (Figure 4(a)).

Figure 4.

NKAIN2 is targeted by miR-181d in pancreatic cancer. (a) Complimentary binding was shown for miR-181d and 3′-UTR of human NKAIN2 gene. A mutant 3′-UTR of NKAIN2 was created to deactivate the binding of miR-181d. (b) HEK-293T cells were co-transfected with luciferase plasmids expressing either wild-type or mutant NKAIN2 3′-UTRs and miR-181d mimics or non-specific miRNA (Control). After 48 h, a dual-luciferase activity assay was conducted (*p < 0.05). (c) QRT-PCR was carried out to compare NKAIN2 mRNA levels between PANC-1 and AsPC-1 cells transduced with L-C and the cells transduced with L-miR181d-I (*p < 0.05). (d) Western blot was conducted to compare NKAIN2 protein levels between PANC-1 and AsPC-1 cells transduced with L-C and the cells transduced with L-miR181d-I. (e) QRT-PCR was conducted to compare NKAIN2 mRNA levels between normal pancreatic cell line hTERT-HPNE and immortal pancreatic cancer cell lines PANC-1, AsPC-1, AsPC-2, Capan-1, Capan-2, Mia PaCa-2, SW1990, BxPC-3, HPAC, and HPAF cells (*p < 0.05). (f) QRT-PCR was also performed to compare NKAIN2 mRNA levels between non-carcinoma pancreatic tissues and pancreatic carcinoma extracted from 37 patients with pancreatic cancer (*p < 0.05).

We then carried out a dual-luciferase activity assay to verify the targeting of miR-181d on NKAIN2. After 48 h, measurement on relative luciferase activities demonstrated that miR-181d indeed bound the 3′-UTR of wild-type NKAIN2 gene, instead of mutant NKAIN2 gene, which lacked the putative miR-181d binding sequence (Figure 4(b), *p < 0.05). In addition, we used qRT-PCR to probe the change of NKAIN2 gene in pancreatic cancer cells due to miR-181d downregulation. It showed that NKAIN2 mRNA levels were significantly upregulated in PANC-1 and AsPC-1 cells transduced with miR-181d-I than in cells transduced with L-C (Figure 4(c), *p < 0.05). Also, western blot analysis demonstrated that NKAIN2 protein levels were also upregulated in PANC-1 and AsPC-1 cells transduced with miR-181d-I than in cells transduced with L-C (Figure 4(d)).

Furthermore, we examined the NKAIN2 mRNA levels in pancreatic cancer without miR-181d manipulation. QRT-PCR probe demonstrated that NKAIN2 was significantly downregulated in both pancreatic cancer cell lines (Figure 4(e), *p < 0.05) and pancreatic carcinoma (Figure 4(f), *p < 0.05), confirming that NKAIN2 expression is inversely correlated with miR-181d expression in pancreatic cancer.

NKAIN2 is directly involved in miR-181d downregulation-mediated tumor suppression of pancreatic cancer

Finally, we selected PANC-1 and AsPC-1 cells with stable miR-181d downregulation (those transduced with L-miR181d-I) and transfected them with NKAIN2-specific siRNA (Si-NKAIN2) or a control siRNA (Si-C). After 24 h, analysis of qRT-PCR demonstrated that NKAIN2 was markedly downregulated in Si-NKAIN2-transfected pancreatic cells, rather than in Si-C-transfected pancreatic cells (Figure 5(a), *p < 0.05).

Figure 5.

NKAIN2 inversely regulated miR-181d-mediated regulation on pancreatic cancer in vitro redevelopment. (a) In miR-181d-downregulated PANC-1 and AsPC-1 cells, they were transfected with 50 nM control siRNA (Si-C) or NKAIN2 siRNA (Si-NKAIN2). qRT-PCR was carried out to examine NKAIN2 mRNA expression levels 24 h after transfection (*p < 0.05). (b) In siRNA-transfected, miR-181d-downregulated PANC-1 and AsPC-1 cells, a 5-day MTT assay was carried out to measure pancreatic cancer in vitro proliferation (**p < 0.05, one-way ANOVA). (c) A Transwell migration assay was also conducted. Cells migrating into lower chambers were stained with crystal violet and imaged. (d) Relative migrations were measured for siRNA-transfected, miR-181d-downregulated PANC-1 and AsPC-1 cells (*p < 0.05). (e) A chemosensitivity assay was also conducted for 72 h. Relative viabilities were then measured for siRNA-transfected, miR-181d-downregulated PANC-1 and AsPC-1 cells (**p < 0.05, one-way ANOVA).

After siRNA transfection, pancreatic cancer proliferations were examined by MTT assay. It showed that NKAIN2 downregulation reversed the inhibition of miR-181d downregulation on pancreatic cancer proliferation (Figure 5(b), p < 0.05, one-way ANOVA). In addition, a Transwell assay was performed on siRNA-transfected PANC-1 and AsPC-1 cells. It showed that NKAIN2 downregulation significantly promoted pancreatic cancer cells to migrate into lower chambers (Figure 5(c)). Quantification showed that migration capability of miR-181d-downregulated pancreatic cancer cells was increased by at least 300% (Figure 5(d), *p < 0.05). Furthermore, a chemosensitivity assay showed that NKAIN2 downregulation also significantly reduced 5-FU chemosensitivity in miR-181d-downregulated pancreatic cancer cells (Figure 5(e), p < 0.05, one-way ANOVA).

Overall, examinations on the functions of NKAIN2 downregulation in miR-181d-downregulated pancreatic cancer cells demonstrated that NKAIN2 reversed the tumor suppression of miR-181d downregulation on pancreatic cancer development in vitro.

Discussion

In previous studies, it was suggested that family of miR-181, including miR-181a, miR-181b, miR-181c, and miR-181d, was aberrantly overexpressed in pancreatic cancer cells.^12,13 In this study, we used qRT-PCR to confirm that miR-181d was indeed upregulated in both pancreatic cancer cell lines and pancreatic carcinoma dissected from human patients. Interestingly, in some other cancers, such as human glioma, miR-181d was downregulated in carcinoma tissues.^18,19 Therefore, it seems like the aberrant expression pattern of miR-181d may vary between different cancer types.

As we demonstrated, miR-181d was upregulated in pancreatic cancer, and we hypothesized that miR-181d may act as an oncogenic factor and downregulating miR-181d may yield suppressive effect on pancreatic cancer development. We then conducted several functional assays and confirmed this hypothesis. Using lentiviral transduction, we successfully created two pancreatic cancer cell lines, PANC-1 and AsPC-1, with stable miR-181d downregulation. Then, through in vitro probes of MTT assay, Transwell assay, and 5-FU chemosensitivity assay, we showed that mir-181d downregulation suppressed cancer proliferation and migration and increased 5-FU chemosensitivity. Furthermore, through in vivo probe of PANC-1 explant, we demonstrated that miR-181d also suppressed in vivo growth of pancreatic cancer. Thus, not only did we show definitive evidence of aberrant upregulation of miR-181d but also we first demonstrated functional role of miR-181d, as an oncogenic factor, in pancreatic cancer. This finding is in accordance with other study, showing miR-181d is an oncogene in human hepatic cancer stem cells by promoting cancer stem cell growth. ²⁰ However, as we mentioned earlier, miR-181d was found to be downregulated in glioma, and miR-181d upregulation inhibited cancer proliferation and induced cancer apoptosis among glioma cell lines. ¹⁸ Therefore, not only the expression pattern but also the functional role of miR-181d may vary among different cancer types.

The reason of miR-181d acting as either oncogene or tumor suppressor may be strongly associated with its downstream signaling pathways in various cancers. In our study, we found that human gene of NKAIN2 was the direct target of miR-181d in pancreatic cancer. Not only dual-luciferase activity assay confirmed the binding of miR-181d on NKAIN2 3′-UTR but also NKAIN2 was specifically upregulated in miR-181d-downregulated pancreatic cancer cells. Furthermore, through siRNA transfection, we showed that NKAIN2 downregulation reversed the suppressive effects of miR-181d inhibition on pancreatic cancer proliferation, migration, and 5-FU chemosensitivity. It is worth noting that in human hepatic cancer stem cells, miR-181d exerted oncogenic effect by directly targeting caudal-type homeobox transcription factor 2 and GATA binding protein 6 to facilitate pancreatic cancer proliferation. ²⁰ However, while miR-181d was shown to be downregulated and acting as tumor suppressor in glioma, its downstream target genes were oncogenes, including methyl-guanine-methyl-transferase, B-cell lymphoma-2, and K-Ras.^18,19 Therefore, it is very likely that the role of miR-181d, whether as oncogene or tumor suppressor, may be well determined by inverse regulation on its downstream target genes in different cancers. As for gene of NKAIN2, though it was implicated in previous studies that it might act as tumor suppressor,^15,16 or oncogene in human cancer, ¹⁷ our study is the first one to reveal its predominant tumor-suppressive role, as well as its close correlation with miR-181d regulation in pancreatic cancer.

Conclusion

Overall, in this study, we revealed a novel epigenetic signaling pathway in pancreatic cancer, showing that miR-181d was aberrantly upregulated and inhibition of miR-181d had tumor-suppressive mechanisms on cancer development both in vitro and in vivo. In addition, we revealed that the regulatory effect of miR-181d inhibition is through the inverse regulation on its downstream tumor suppressor NAKIN2 gene.