Long non-coding RNA KTN1-AS1 promotes progression in pancreatic cancer through regulating microRNA-23b-3p/high-mobility group box 2 axis

High KTN1-AS1 expression in PC

To identify the potential lncRNAs which were involved in the modulation of PC progression, we performed the RNA-seq analysis in clinical PC tissues and adjacent non-tumor tissues. Among the differently expressed lncRNAs, we identified that KTN1-AS1 presented a most significant elevation in the PC tissues (Figure 1A). We then selected KTN1-AS1 for the further analysis. Moreover, quantitative detection of KTN1-AS1 in PC tissues detected by qRT-PCR exhibited that KTN1-AS1 in PC samples was obviously elevated compared with para-tumor tissues, (Figure 1B), and likewise, KTN1-AS1 was observably enhanced in PC cells relative to normal epithelial pancreatic cells. (Figure 1C). The PANC-1 and BxPC-3 cell lines were selected in the subsequent experiments. We demonstrated that the AUC of KTN1-AS1 was 0.946 in the Receiver operating characteristic (ROC) curve (Figure 1D). According to the median expression of KTN1-AS1 (1.68), samples were divided into 48 cases with high KTN1-AS1 expression and 39 cases with low KTN1-AS1 expression. It turned out that KTN1-AS1 was correlated with tumor size, LNM, differentiation degree, and pathological stage. We identified that high KTN1-AS1 expression was correlated with the poor survival rate of the PC patients. (Table 1, Figure 1E).

Figure 1. KTN1-AS1 expression and significance in PC. (A) Heatmap of differentially expressed lncRNAs in the RNA-seq analysis of clinical PC tissue and paracancer tissues, fold change ≥ 2 and P value < 0.05. (B) KTN1-AS1 expression in PC tissue. (C) KTN1-AS1 expression in pancreatic cancer cells. (D) ROC curve of KTN1-AS1 in the diagnosis of PC. (E) Effects of different KTN1-AS1 expression levels on the survival rate of patients with PC. a indicates P < 0.05.

KTN1-AS1 induced oncogenetic function in PC cells

We transfected Si-KTN1-AS1, sh-KTN1-AS1 and control NC into PANC-1 and BxPC-3 to observe the effect of KTN1-AS1 on PC cells. It was noticed that the KTN1-AS1 expression in Si-KTN1-AS1 group cells was markedly down-regulated but was increased in Sh-KTN1-AS1 group cells (P < 0.05). Knockdown of KTN1-AS1 suppressed the proliferation and invasion of PC cells but accelerated their apoptosis, observably up-regulated pro-apoptotic Bax and Caspase-3 expression, and remarkably down-regulated the anti-apoptotic bcl-2 expression (P < 0.05). While on the contrary, KTN1-AS1 could further induce the proliferation/invasion and inhibit apoptosis of PC cells, and meanwhile, markedly declined Bax and Caspase-3 expression, and dramatically enhanced bcl-2 expression. (Figure 2 and Supplementary Figure 1) The expression of Bax, Bcl-2, and caspese-3 was verified by IHC in the clinical PC tissues and paracancer tissues. (Supplementary Figure 2).

Figure 2. Effects of KTN1-AS1 on PC cells. (A) Expression of KTN1-AS1 in PC cells after transfection. (B) Effects of KTN1-AS1 on the proliferation of PC cells. (C) Effects of KTN1-AS1 on the invasion of PC cells. (D) Effects of KTN1-AS1 on the apoptosis rate of PC cells. (E) Effects of KTN1-AS1 on the apoptosis-related proteins of PC cells. a indicates P < 0.05.

KTN1-AS1 inhibited miR-23b-3p expression by binding to it

To explore the potential downstream miRNAs of KTN1-AS1, we performed the RNA-seq analysis in PANC-1 cells transfected with sicontrol or siKTN1-AS1. Among the differently expressed miRNAs, we identified that miR-23b-3p presented a most significant enhancement by the depletion of KTN1-AS1 in PANC-1 cells (Supplementary Figure 3). We then selected miR-23b-3p for the further analysis. Moreover, we found the potential binding target of KTN1-AS1 and miR-23b-3p. To further verify it, we employed DLR and RIP experiments to confirm their correlation.

The DLR assay revealed that KTN1-AS1-WT fluorescence activity was markedly inhibited by miR-23b-3p-mimics. RIP experiments showed that the KTN1-AS1 and miR-23b-3p levels interacted with Ago2 were greatly higher than those of IgG. In addition, miR-23b-3p was underexpressed in PC tissues. Correlation analysis indicated a negative relationship of KTN1-AS1 and miR-23b-3p in PC tissues. We then transfected miR-23b-3p-mimics, sh-KTN1-AS1+miR-23b-3p-mimics, and miR-NC, or sh-KTN1-AS1+miR-23b-3p-mimics, and miR-NC to cells. It turned out that after transfecting miR-23b-3p-mimics, the proliferation/invasion/migration of cells was observably inhibited and apoptosis was increased, while the biological function of cells after co-transfection with Sh-KTN1-AS and miR-23b-3p-mimics shown an opposite result. KTN1-AS1 inhibited miR-23b-3p expression by binding to miR-23b-3p, thereby promoting tumor growth. (Figure 3 and Supplementary Figure 1).

Figure 3. KTN1-AS1 inhibited its expression by binding to miR-23b-3p. (A) DLR assay. (B) RIP experiment. (C) MiR-23b-3p expression in PC tissue. (D) Correlation analysis between KTN1-AS1 and miR-23b-3p. (E) Effects of miR-23b-3p on the proliferation of PC cells. (F) Effects of miR-23b-3p on the invasion of PC cells. (G) Effects of miR-23b-3p on the apoptosis rate of PC cells. (H) Effects of miR-23b-3p on apoptosis-related proteins in PC cells. a indicates P < 0.05.

KTN1-AS1 upregulated HMGB2 through competitive binding with miR-23b-3p

To explore the potential downstream genes of KTN1-AS1 in PC cells, we performed the RNA-seq analysis in PANC-1 cells transfected with sicontrol or siKTN1-AS1. Among the differently expressed genes, we identified that HMGB2 presented a most significant downregulation by the depletion of KTN1-AS1 in PANC-1 cells (Supplementary Figure 3). We then selected HMGB2 for the further analysis. Moreover, we found the targeted binding sites between miR-23b-3p and HMGB2 using Targetscan7.2 online software. The DLR assay exhibited that the luciferase activity of HMGB2 was reduced by miR-23b-3p-mimics. What’s more, HMGB2 was observed to be highly expressed in PC tissues. Correlation analysis revealed a positive relationship of HMGB2 and KTN1-AS1, and a negative relation between HMGB2 and miR-23b-3p. We then transfected miR-NC, miR-23b-3p-mimcis, sh-KTN1-AS1+miR-23b-3p-mimics into cells to detect the expression of HMGB2. It was found that HMGB2 protein and mRNAs expression in the cells were inhibited after upregulation of miR-23b-3p, while the effects were reversed after the co-transfection of sh-KTN1-AS1 and miR-23b-3p-mimics. (Figure 4) The expression levels of HMGB2were verified by IHC in the clinical PC tissues and paracancer tissues. (Supplementary Figure 2).

Figure 4. KTN1-AS1 competitively bound to miR-23b-3p to up-regulate the expression of HMGB2. (A) DLR assay. (B) Effects of KTN1-AS1 and miR-23b-3p on HMGB2 protein expression. (C) HMGB2 expression in PC tissue. (D) Correlation analysis between HMGB2 and KTN1-AS1. (E) Correlation analysis between HMGB2 and miR-23b-3p. a indicates P < 0.05.

Down-regulation of HMGB2 expression inhibited growth and metastasis of PC cells

For the purpose of further exploring the impact of HMGB2, we silenced HMGB2, and found that proliferation/invasion was suppressed and apoptosis was elevated by Si-HMGB2, and the Bcl-2, N-cadherin, and vimentin expression was declined, and the bax and Caspase-3 expression was boosted. (Figure 5 and Supplementary Figure 1).

Figure 5. Effects of HMGB2 on PC cells. (A) HMGB2 expression in PC cells after transfection. (B) Effects of HMGB2 on proliferation of PC cells. (C) Effects of HMGB2 on invasion of PC cells. (D) Effects of HMGB2 on apoptosis rate of PC cells. (E) Effects of HMGB2 on apoptosis-related proteins in PC cells. a indicates P < 0.05.

KTN1-AS1-mediated miR-23b-3p/HMGB2 promoted tumor formation in nude mice

For the sake of observing whether KTN1-AS1 affected solid tumor through miR-23b-3p/HMGB2 axis, we carried out tumorigenesis experiment in nude mice. By injecting miR-NC, miR-23b-3p-mimcis, sh-KTN1-AS1+miR-23b-3p-mimics subcutaneously into nude mice, we found that the tumor size of nude mice injected with miR-23b-3p-mimcis decreased observably, while those of nude mice injected with sh-KTN1-AS1+miR-23b-3p-mimics did not differ notably from those of miR-NC. Further measurement of the expression levels of HMGB2 protein and mRNAs in the tumor tissue of nude mice displayed that the expression levels of HMGB2 protein and mRNAs in the tumor of nude mice injected with miR-23b-3p-mimcis was significantly inhibited, while the expression levels of HMGB2 protein and mRNAs in the tumor of nude mice co-transfected with sh-KTN1-AS1+miR-23b-3p-mimics was reversed. (Figure 6) Meanwhile, we validated that KTN1-AS1 overexpression enhanced the tumor growth and the depletion of KTN1-AS1 or HMGB2 repressed the tumor growth in the nude mice. (Supplementary Figure 4).

Figure 6. KTN1-AS1-mediated miR-23b-3p/HMGB2 promoted tumor formation in nude mice. (A) Changes of subcutaneous tumor volume in nude mice within 28 days. (B) Tumor volume of nude mice was on the 28th day. (C) Relative expression of HMGB2 protein and mRNAs in nude mouse tumor. a indicates P < 0.05.

Clinical specimens

After patients’ consents, paired PC tissues and paracancerous tissues were taken from 100 PC patients diagnosed and treated in the First Affiliated Hospital of China Medical University, and the specimens were then frozen in liquid nitrogen tanks for follow-up experiments. Inclusion criteria: Patients pathologically diagnosed as PC. Exclusion criteria: Patients who had received any chemoradiotherapy prior to this study; Patients with other malignancies; Patients with severe hepatorenal dysfunction; Patients with severe infection or immune system disorders. Having been approved by the Medical Ethics Committee of the First Affiliated Hospital of China Medical University, this study was carried out after the consent of all enrolled patients or their families, with written informed consent obtained. The expression levels of HMGB2, Bax, Bcl-2, and caspese-3 were verified by immunohistochemistry (IHC) in the samples.

Cell culture, passage and transfection

PC cell lines (PANC-1, SW1990, BxPC-3, AsPC-1) and human normal pancreatic duct epithelial cells HPDE were all purchased from the ATCC. The PC cell line was cultured at 37°C with 5% CO₂ in a medium containing 10% PBS (HyClone, Logan, UT, USA) DMEM (Invitrogen, USA). When digestion finished, the cells were cultured and for passage before transfecting. Since RT-PCR results showed that KTN1-AS1 had the most obvious changes in PANC-1 and BxPC-3, these two were selected for the research objects. MiR-23b-3p-mimics, miR negative control (miR-NC), pcDBA3.1 vector containing targeted inhibition KTN1-AS1 plasmid (si-KTN1-AS1), pcDNA3.1 vector containing KTN1-AS1 plasmid (sh-KTN1-AS1), pcDNA3.1 vector containing targeted inhibition HMGB2 plasmid (si-PBX3), and control sequence (Control) were transfected into cells with Lipofectamine™ 2000 kit, respectively. Strictly abode by the instructions, the operation was carried out. The sequences used were as follows: miR-23b-3p mimics, 5′-ATCACATTGCCAGGGATTACC-3; si-KTN1-AS1, 5′-CTTTCATTTCTGCTATCCC-3′; si-PBX3, 5′-CCGUCAAUUUCGCGGAAUUTT-3′.

RNA-Seq analysis

The RNA-seq analysis was performed to explore the potential correlation of lncRNAs with pancreatic cancer. Total RNA was extracted from 3 paired pancreatic cancer tissues and paracancer tissues. Meanwhile, the RNA-seq analysis was performed to explore the potential downstream miRNAs and mRNAs of KTN1-AS1. Total RNA was extracted from PANC-1 cells transfected with sicontrol or siKTN1-AS1. The RNA sequencing assay was performed to at KangChen Bio-Tech (Shanghai, China) using Illumina HiSeq 4000 (Illumina, San Diego, CA, USA). The differential expression was analyzed based on P value and fold change (fold change ≥ 2 and P value < 0.05).

qRT-PCR detection

RNA was isolated from cells and tissues by using a TRIzol reagent (Invitrogen, USA) in accordance with the protocol, followed by reverse-transcription to cDNA by using PrimeScript Master Mix (Takara, Japan). The levels of RNAs were quantified by using a SYBR Premix Ex Taq Kit (Takara). GAPDH and U6 were used as internal control for normalization of gene expression using the 2^–ΔΔCt method. The primer sequences are detailed in Table 2.

Detection of protein expression by western blot

The samples were homogenized by RIPA lysis buffer that contains proteinase inhibitors (Sigma) to extract total proteins. The proteins were quantified by BCA assay kit (Beyotime, China). An equal amount of proteins (30 μg) were divided on SDS-PAGE gel, shifted to NC membranes, blocked by 5% silk milk and hatched with specific primary antibodies against HMGB2 (1: 500), Bax (1: 500), Bcl-2 (1: 500), Caspese-3 (1: 500), and β-Actin (1: 1000) at 4°C overnight. Next day, the protein bands were incubated with specific HRP-conjugated anti-mouse and anti-rabbit secondary antibodies at room temperature for 45 minutes, visualized by using an ECL solution (Millipore, Germany) in a Gel Image system (Bio-Rad, Hercules, CA, USA). HMGB2, Bax, Bcl-2, Caspese-3 and β-Actin antibody were all acquired from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

Cell proliferation assessed by CCK-8

Cell proliferation was detected according to the instructions of CCK-8 kit (Promega, Madison, WI, USA). Forty-eight hours after transfection, cells were gathered and diluted into 3 × 10⁴cell/mL, and inoculated in 96-well plates, 10 μL CCK8 solution was added to each well and continued to culture in an incubator with 5% CO₂ at 37°C for 2 h. With the help of a microplate reader, the OD value at 450 nm was finally tested to detect cell proliferation. The experiment was duplicated three times.

Transwell detection

According to the kit instructions (Reanta Biotechnology Co., Ltd., Beijng, China), cells were collected 24 hours after transfection and inoculated on 24-well plates (3 × 10⁴ cells/well). For cell invasion, cells were suspended in FBS-free medium, and seeded in the upper chambers of Transwell plates (Corning, NY, USA), while the lower chambers were filled with complete medium. The upper chambers were collected after incubation for 48 hours, washed in PBS, fixed in methanol, and dyed with 1% crystal violet. Finally, the cell invasion was observed by a microscope.

Apoptosis test

On the basis of the instructions of Annexin V-FITC/PI apoptosis kit (Biolab Biotechnology Co., Ltd., Beijing, China), the transfected cells were digested with 0.25% trypsin before rinsing with PBS twice. Then, the cells (1 × 10⁶) were treated with binding buffer, and then AnnexinV-FITC and PI were added successively to incubate for 5 min away from light at room temperature. Finally, apoptosis was detected by FACSVerse flow cytometry (Becton Dickinson Company, United States), and the average value was obtained by repeating the experiment for 3 times.

Targeting correlation of miR-23b-3p between KTN1-AS1 and HMGB2 determined by DLR gene assay

The wild type (WT) and mutated (MUT) sequences of KTN1-AS1 and the 3′-UTR of HMGB2 were inserted into PmirGLO plasmids. Cells were transfected with the constructed vectors and miR-23b-3p mimics, along with the pRL-TK as internal control. The luciferase activity was determined after 48-hours incubation with a dual luciferase reporter assay system (Promega).

RNA binding protein immunoprecipitation (RIP experiment)

Magna RIP RNA-binding protein immunoprecipitation kit (Millipore) was used for RIP assay in accordance with manufacturer’s instruction. In short, cells lysates were obtained by using lysis buffer and sonication, then incubated with Ago2 antibody or IgG antibody-conjugated beads for 3 hours at room temperature. The precipitation was washed and eluted, the enriched KTN1-AS1 and miR-23b-3p were analyzed by qRT-PCR assay.

Tumorigenesis experiment in nude mice

Animal experiments were conducted under the approval of the Animal Research Ethics Committee of the First Affiliated Hospital of China Medical University. 4–6 weeks old male nude mice (Charles River, China) were selected, with 5 in each group. After transfection, 200 mL of PANC-1 cells (3 × 10⁶) were injected subcutaneously into the left side of the back of each nude mouse. Tumor size was measured regularly and calculated using formula 0.52 × tumor length × tumor short diameter². The animals were euthanized 30 d after injection to remove the tumor and measure its volume and weight.

Statistical analysis

Data were expressed as mean ± SD and analyzed by using SPSS 20.0 software. Comparison between two or multiple groups were determined by student’s t test or one-way ANOVA analysis followed by LSD-t, respectively. p < 0.05 was considered as statistically significant.