LncRNA HCG11/miR-579-3p/MDM2 axis modulates malignant biological properties in pancreatic carcinoma via Notch/Hes1 signaling pathway

Expression profiles of HCG11 in pancreatic carcinoma tissues and cells

To identify the expression level of HCG11 in pancreatic carcinoma, qPCR experiment was utilized to detect the expression of HCG11 in 20 pairs of pancreatic carcinoma tissues. Significantly, a marked augmentation of HCG11 expression was observed in pancreatic carcinoma tissues when compared with the corresponding non-cancerous tissues (Figure 1A, p<0.001). Next, we performed qPCR to detect HCG11 expression in various pancreatic carcinoma cancer cell lines. Compared with HPDE6-C7 cells, we observed that HCG11 expression was highly increased in pancreatic carcinoma cells including BxPC-3, Capan-2, SW1990, PANC-1 and AsPC-1 (Figure 1B, p<0.001). Due to the high expression rates of HCG11 in PANC-1 and AsPC-1 cells, these two cells were used for subsequent experiments.

Knockdown of HCG11 inhibited growth, movement and caused apoptosis in pancreatic carcinoma cells

To examine the effect of HCG11 in pancreatic carcinoma, si-HCG11-1 or si-HCG11-2 was applied to knockdown of HCG11 in pancreatic carcinoma cells. Both si-HCG11-1 and si-HCG11-2 significantly reduced the expression of HCG11, but si-HCG11-1 was more efficient, so it was selected for subsequent experiments (Figure 1C, 1D, p<0.01). After knockdown of HCG11, we observed that the OD values of pancreatic carcinoma cells were decreased significantly (Figure 2A, 2B, p<0.01). Moreover, the data from flow cytometry assay indicated that suppression of HCG11 in pancreatic carcinoma cells controlled the cell cycle by inducing G0/G1 arrest compared with the siRNA NC group (Figure 2C, 2D). Besides, we also discovered that depletion of HCG11 notably reduced the number of clones in pancreatic carcinoma cells (Figure 2E, p<0.05). The results insinuated that depletion of HCG11 suppressed the growth of pancreatic carcinoma cells. Thereafter, an apoptosis experiment was conducted to detect the effect of HCG11 on pancreatic carcinoma cells apoptosis. As expected, the apoptotic percentage of pancreatic carcinoma cells was significantly increased after knockdown of HCG11 (Figure 2F, p<0.0001). Besides, we also detected that depletion of HCG11 reduced the invaded number of pancreatic carcinoma cells (Figure 2G, p<0.05). All these observations illustrated that HCG11 played an active role in the progression of pancreatic carcinoma.

HCG11 acted as a ceRNA by direct sponging of miR-579-5p

A plenty of reports presented that lncRNAs can serve as ceRNAs, eliminating the endogenous inhibitory effect of these miRNAs on their targeted transcripts. Then, we used the bioinformatics software starBase v2.0 (http://starbase.sysu.edu.cn/) to explore the miRNAs that targeted with HCG11. The data showed an underlying combination of HCG11 and miR-579-3p, the predicted targeting sequences were presented in Figure 3D. To further identify the interaction, the HCG11 sequences including the wt or mut miR-579-3p targeting sites was inserted into the downstream of pGL-3 vector. Next, the function of miR-579-3p mimic on the luciferase activity was observed. In PANC-1 and AsPC-1 cells, the luciferase activity in wt-pGL3-HCG11 group was obviously reduced than the control group after miR-579-3p mimic treatment (Figure 3E, 3F, p<0.05). Nonetheless, no obvious change of luciferase activity has been observed in mut-pGL3-HCG11 group when treated by miR-579-3p mimic. Data from Figure 3A, 3B showed that miR-579-3p expression was obviously decreased in pancreatic carcinoma samples when compared with the adjacent normal tissues and HPDE6-C7 cells. Moreover, miR-579-3p expression was significantly negatively correlated with the expression of HCG11 (Figure 3C, p<0.01). Additionally, the qPCR results from PANC-1 and AsPC-1 cells showed that depletion of HCG11 increased miR-579-3p expression (Figure 3G, 3H, p<0.01). Inversely, overexpression of HCG11 obviously reduced miR-579-3p expression (Figure 3G, 3H, p<0.01), further validating the interaction between HCG11 and miR-579-3p.

MDM2 was validated as a downstream target of miR-579-3p and was co-modulated by HCG11/miR-579-3p

To investigate the presumptive target genes of miR-579-3p, we used Targetscan to search for the candidate genes. Bioinformatics analysis revealed that miR-579-3p directly binds MDM2 (Figure 4E). To further affirm the association between miR-579-3p and MDM2, a wt or mut pGL3-MDM2 vector was generated, and wt or mut pGL3-MDM2 was co-transfected with miR-579-3p mimic or NC into PANC-1 and AsPC-1 cells. Compared with the corresponding control group, the luciferase activity was notably decreased in cells that co-transfected with wt pGL3-MDM2 luciferase vector and miR-579-3p mimic (Figure 4F, 4G, p<0.05). However, the inhibitory effect was eliminated when pGL3-MDM2 vector that included mut binding sequences were co-transfected with miR-579-3p mimic (Figure 4F, 4G, p<0.05). Next, the qPCR was conducted to analyze MDM2 expression in pancreatic carcinoma samples, indicating that a marked augmentation of MDM2 was presented in pancreatic carcinoma samples when in contrast with the adjacent samples and HPDE6-C7 cells (Figure 4A, 4B, p<0.01). MDM2 expression was actively related to the expression of HCG11 (Figure 4C, p<0.01). Inversely, MDM2 expression was negatively correlated with miR-579-3p (Figure 4D, p<0.05). Importantly, we also discovered that MDM2 expression both at mRNA and protein levels were obviously decreased in PANC-1 and AsPC-1 cells when miR-579-3p mimic treatment (Figure 4H, 4I), further confirming the interaction between miR-579-3p and MDM2. Also, we analyzed the influence of HCG11 on MDM2 expression. As shown in Figure 4J–4M, depletion of HCG11 reduced MDM2 expression at transcription and translation levels in pancreatic carcinoma cells. However, MDM2 expression were obviously increased in PANC-1 and AsPC-1 cells at both mRNA and protein levels when upregulation of HCG11. These results suggested that MDM2 was affirmed as a target of miR-579-3p and regulated by HCG11/miR-579-3p, insinuating that HCG11, miR-579-3p and MDM2 may establish a ceRNA network to regulate the progression of pancreatic carcinoma.

HCG11 promoted the biological behaviors of pancreatic carcinoma cells by targeting miR-579-3p/MDM2 axis and activating Notch/Hes1 pathway

To validate the network among HCG11, miR-579-3p and MDM2, we carried out a series of biological experiments to observe the effect of HCG11/miR-579-3p/MDM2 on AsPC-1 cells growth, apoptosis and mobility. Firstly, we observed that MDM2 expression was co-regulated by HCG11, miR-579-3p and MDM2 (Figure 5A, p<0.01). Then, data from CCK-8, flow cytometry, plate cloning formation and Transwell assays suggested that upregulation of miR-579-3p suppressed the pro-oncogenic effect of HCG11 in pancreatic carcinoma cells. However, the above results were inversed by MDM2 upregulation (Figure 5B–5F, p<0.05). These data revealed that the pro-oncogenic effect of HCG11 in pancreatic carcinoma was achieved by targeting miR-579-3p to upregulate MDM2. Thereafter, the protein expression levels of NICD and Hes1 were also examined. We discovered that upregulation of HCG11 promoted the expression of NICD and Hes1, which can be eliminated by miR-579-3p overexpression. However, the effect of miR-579-3p on NICD and Hes1 expression was rescued by MDM2 expression (Figure 5G, 5H, p<0.05). These data insinuated that HCG11/miR-579-3p/MDM2 functioned as a ceRNA network to promote the progression of pancreatic carcinoma cells by inactivating Notch/Hes1 pathway.

Knockdown of HCG11 inhibited the tumor growth by regulating a range of biological behaviors

To investigate the function of HCG11 depletion in the modulation of tumor growth, we detect the effect of HCG11 using AsPC-1 cells subcutaneous xenograft mouse model. After injection, the tumor volumes were tested every 7 days for 4 weeks. We observed that knockdown of HCG11 notably suppressed tumor growth (Figure 6A). The tumor weight of AsPC-1-HCG11-knockdown (KD) group was obviously lower than the control group (Figure 6B, p<0.001). Besides, we also observed a decrease of HCG11 expression in the serum of sh-HCG11 nude mice (Figure 6C, p<0.01). Moreover, the decreased HCG11 expression and the increased miR-579-3p expression were observed in HCG11-KD tumors (Figure 6D, 6E, p<0.001). Besides, the mRNA and protein levels of MDM2 reduced in HCG11-KD tumors (Figure 6F, 6G, p<0.05). Furthermore, the proliferation associated genes Ki67 and PCNA were obviously decreased in HCG11-KD tumors (Figure 6H, p<0.05). As well, the TUNEL-positive cells in HCG11-KD tumors showed significant changes, accompanied by obvious increase (Figure 6I, p<0.001). Additionally, we also discovered that in HCG11-KD tumors, E-cadherin protein level increased, while Vimentin level decreased (Figure 6J, p<0.05). All these results insinuated that depletion of HCG11 suppressed the tumor growth by targeting miR-579-3p/MDM2, thereby inhibiting tumor cells proliferation, promoting tumor cells apoptosis and blocking tumor cells movement.

Human samples

Twenty pairs of surgical samples including pancreatic carcinoma tissues and normal pancreatic tissues were originally taken from patients in our hospital. Surgical samples were gathered for the calculation of HCG11, miR-579-3p and MDM2 expression. The protocol of this study was performed based on the Declaration of Helsinki and has been approved by the Ethics Committee of our hospital. Written informed consent was achieved from each participant.

Cell lines and culture

Pancreatic carcinoma cell lines including BxPC-3, SW1990, PANC-1, AsPC-1, Capan-2 were obtained from the American Type Culture Collection (ATCC; Manassas, USA). Human immortalized pancreatic duct epithelial cells (HPDE6-C7) were afforded by Shanghai Cell Bank of the Chinese Academy of Sciences (Shanghai, China). All cells were grown in DMEM including 10% fetal bovine serum (FBS) in an incubator (37° C, 5% CO₂₎.

Cell transfection and treatment

Small fragments including si-HCG11-1, si-HCG11-2, si-RNA negative control (NC), pcDNA3.1-vector, pcDNA3.1-HCG11, pcDNA3.1-MDM2, miR-579-3p mimic and NC were designed and synthesized by Sangon (Shanghai, China). Then, the fragments were transfected into cells by using Lipofectamine 2000 to regulate the expression of HCG11, miR-579-3p and MDM2, respectively.

RNA isolation and qRT-PCR

TRIzol solution (TaKaRa, Japan) was performed to isolate RNA from pancreatic carcinoma samples. Then, cDNA was formed with a reverse transcription kit following the supplier’s specification. Then, the ABI Prism 7700 sequence detection system was applied to perform the qRT-PCR process with the subsequent steps: 95° C for 5 min, 40 cycles of 95° C for 30 s, 60° C for 45 s, and 72° C for 90 s and 72° C for 10 min. The relative expression of HCG11, miR-579-3p and MDM2 were analyzed by 2^-ΔΔCT method. GAPDH was deemed as the standard control for detection of HCG11 and MDM2. U6 was deemed as the standard control for determination of miR-579-3p. The primers were listed as below:

HCG11, F: 5’-AATGGTGGTAGGAGGGAGGA-3’,

R: 5’-CACACAGGGGAATGAAGAGG-3’;

MiR-579-3p, F: 5’-GCACGGAACTTCCCTTGACGTC-3’,

R: 5’-GCTCTAGGGATCGTCGCCGAA-3’;

MDM2, F: 5’-ATGTGCAATACCAACATCTCTGTGTC-3’,

R: 5’-GCTGACTTACAGCCACTAAATTTC-3’;

GAPDH, F: 5’-CGGAGTCAACGGATTTGGTCGTAT-3’,

R: 5’-AGCCTTCTCCATGGTGGTGAAGAC-3’;

U6, F: 5’-CGCTTCGGCAGCACATATACTAA-3’,

R: 5’-TATGGAACGCTTCACGAATTTGC-3’.

Cell viability and colony formation assays

The viability of PANC-1 or AsPC-1 cells was estimated with the support of cell counting kit-8 (CCK-8). Briefly, the cells were implanted in a 96-well plate with a concentration of 5×10³ cells/well and treated with si-HCG11 or siRNA NC. Cell viability was analyzed with the assistance of CCK-8 kit (Beyotime, China) based on the supplier’s introduction. Each experiment was independently repeated three times.

For colony formation assay, 500 PANC-1 or AsPC-1 cells were implanted into a 60 mm dish, transfected with si-HCG11 or siRNA NC, and grew in medium, which replaced every three days. After 14 days, visible colonies that had been fixed were counted.

Flow cytometry

After treatment, the cell density was adjusted to 1×10⁶ cells/mL. Next, 1 mL of the cell suspension were centrifuged and the sediment was placed in the pre-cooling 70% ethanol for fixing at 4° C overnight. Subsequently, 100 μL of cell suspension was put into 50 μg of propidine iodide (PI) including RNAase and left to react for 30 min away from the light. Finally, the percentage of the cells in G0-G1, S and G2-M phases were counted and compared at 488 nm by a flow cytometry.

The apoptotic rate of cells was tested by Annexin V-FITC/PI double staining. Briefly, the cells were collected by trypsinization and rinsed with pre-cooling phosphate buffer saline (PBS). Next, the cells were cultured with Annexin V-FITC and PI for 15 min in the dark. In the end, the apoptosis rate was estimated at 488 nm by a flow cytometry.

Western blotting

Protein was extracted from pancreatic carcinoma samples and quantified by bicinchoninic acid (BCA) method. Next, total protein was isolated by SDS-PAGE and blotted onto PVDF membranes. Then, the membranes and contents were probed with the primary antibodies against MDM2 (#86934, 1:2000, CST), Ki67 (#9449, 1:1000, CST), NICD (ab128076, 1:1000, Abcam), Hes1 (ab71559, 1:1000, Abcam), PCNA (#2586, 1:2000, CST), E-cadherin (#14472, 1:2000, CST), Vimentin (#5741, 1:2000, CST) and GAPDH (#5174, 1:2000, CST) at 4° C overnight. After blotted with the secondary antibody, the proteins were developed by enhanced chemiluminescence and visualized by Image J software.

Dual luciferase reporter gene assay

Bioinformatics software predicted that the targeted association between HCG11 and miR-579-5p, as well as miR-579-5p and MDM2 were existed. The luciferase assay was conducted in PANC-1 and AsPC-1 cells to further confirm the association among HCG11, miR-579-5p and MDM2. Firstly, the corresponding plasmids (including wild type (wt)-pGL3-HCG11, mutant (mut)-pGL3-HCG11, wt-pGL3-MDM2 and mut-pGL3-MDM2) and miR-579-3p mimic/NC were co-transfected into cells with the assistance of Lipofectamine 2000 reagent. After incubation for 48 h, the dual-luciferase reporter assay system was carried out to analyze the luciferase activity.

Cell invasion assay

Cell invasion assay was implemented utilizing Transwell chambers (Corning, USA). After the matrigel (BD Biosciences, USA) was pre-coated in the upper chamber, PANC-1 or AsPC-1 cells (200 μL, 3×10⁵ cells/well) were implanted into the above chamber with medium excluding serum. Then, DMEM medium including 10% FBS was filled into the under chamber, which was placed in 37° C incubator with 5% CO₂ for 24 h. Finally, the stained cells were counted under a microscope to estimate the invasion capacity.

TdT-mediated dUTP-FITC nick end-labeling (TUNEL) assay

The apoptosis of pancreas was measured through TUNEL assay utilizing a commercially available In Situ Cell Death Detection Kit (Roche Diagnostics, Germany). First, paraffin-embedded sections were dewaxed, rehydrated, and then incubated in 3% H₂O₂ for 10 min to block endogenous peroxidase activity. Then, a mixture of TdT and dUTP was loaded to the slices and cultured at 37° C for 120 min. Added diaminobenzidine (DAB) chromogenic solution to the tissue sections and reverse stained with hematoxylin for 3 min. In the end, the slices were dehydrated by gradient ethanol (70%, 80%, 95%, and 100%), then dehydrated by gradient xylene and sealed with resin fixator. All sections were captured under a microscopy and processed by the Image-Pro Plus 6.0 software.

Animal model

A total of 10 male nude mice were used for tumor formation experiment. As-PC cells were collected and re-suspended in medium (excluding serum) at a density of 5×10⁷ cells/mL. AsPC-1 cells stably transduced with sh-HCG11 or shRNA NC were inoculated subcutaneously at the right side of per mouse. The tumor size was measured and recorded every 7 days. After 28 days of incubation, the mice were euthanized and tumors were collected to measure their size and weight. Tumor volume was calculated by (length × width²). Afterward, the tissues were used to do the corresponding tests. All animal studies were approved by the Animal Ethics Committee of our hospital.

Statistical analysis

SPSS statistics 20.0 and Graphpad Prism 6.0 software were applied to process the experimental data, which presented as mean ± standard deviation. The statistical difference of two or multiple group was processed by Student's t-test or one way analysis of variance accompanied by Bonferroni's post hoc test. P value less than 0.05 was deemed as statistically significant.