ANLN-induced EZH2 upregulation promotes pancreatic cancer progression by mediating miR-218-5p/LASP1 signaling axis

Five pancreatic cancer cell lines were acquired from the Cell Bank Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). The human pancreatic duct epithelial cell line (hTERT-HPNE) was obtained from the American Type Culture Collection (Manassas, VA, USA). The AsPC-1 and BxPC-3 cells were cultured in RPMI 1640 medium (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Gibco, USA). The PANC-1 and MIA PaCa-2 cells were grown in DMEM (Gibco, USA) containing 10% fetal bovine serum. The SW1990 cell line was maintained in L-15 medium (Gibco, USA) supplemented with 10% fetal bovine serum. The hTERT-HPNE cell line was cultured in ATCC-suggested complete growth medium (1 V Medium M3, 3 V glucose-free DMEM, 5% fetal bovine serum, 10 ng/ml EGF, 5.5 mM D-glucose and 750 ng/ml puromycin).

This study was approved by the Ethical Committee of Army Medical University, China. All patients provided written informed consent to participate in the study. The study methodologies conformed to the standards set by the Declaration of Helsinki. Eighty pancreatic cancer tissues and ten cancer-adjacent normal tissues were obtained from the Southwest Hospital, Army Medical University (Chongqing, China) between December 2009 and June 2011. All tumor tissues were diagnosed as pancreatic cancer independently by two pathologists. Patients who had received chemotherapy or radiotherapy were excluded. The overall survival (OS) was defined as the interval between the date of definite diagnosis and the date of death or the last follow up. Collection of follow-up data was ceased in July 2016.

Tissue sections were dewaxed in xylene for 20 min at 37 °C and then rehydrated with a series of graded alcohols. To eliminate endogenous peroxidase activity, the sections were incubated with endogenous peroxidase blocking buffer (Beyotime, Shanghai, China). The sections were subsequently treated with a citrate antigen retrieval solution (Beyotime, China) for 20 min at 98 °C. After washing, the sections were blocked with normal goat serum (Boster, Wuhan, China) for 20 min at room temperature. The sections were incubated with a mouse monoclonal antibody against human ANLN (1200; Abcam, Cambridge, United Kingdom) overnight at 4 °C. The sections were then washed and incubated with goat anti-mouse secondary antibodies (Boster, Wuhan, China) for 30 min at room temperature. Finally, the expression levels of ANLN were analyzed according to methods reported [38]. The degree of immunostaining of the sections was defined by the sum of a proportion score and an intensity score. The proportion score was defined as follows: 0, no positive immunoreactive cells; 1, ≤10% positive immunoreactive cells; 2, 10 to 50% positive immunoreactive cells; or 3, > 50% positive immunoreactive cells. The intensity score was defined as follows: 0, no immunoreactive staining; 1, weak immunoreactive staining; 2, intermediate immunoreactive staining; or 3, strong immunoreactive staining. The scores were independently evaluated by 2 pathologists.

SiRNAs against ANLN and EZH2 were designed and synthesized by Sesh-biotech (Shanghai, China) (Additional file 1: Table S1). The pCMV3-LASP1 CDS (NM_006148) and pCMV3-EZH2 CDS (NM_004456) expression plasmids were acquired from Sino Biological Inc. (Beijing, China). Mimic control (con), miR-218-5p mimics (miR-218-5p), inhibitor control (anti-con) and miR-218-5p inhibitors (anti-miR-218) were obtained from GenePharma (Shanghai, China). For transfection, BxPC-3 and SW1990 cells were cultured in 6-well plates. When the BxPC-3 and SW1990 cells were 80% confluent, they were transfected with the negative control siRNA (NC), ANLN siRNA (ANLN RNAi) or EZH2 siRNA (EZH2 RNAi) using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. For the rescue experiments, ANLN siRNA together with the pCMV3-LASP1 expression plasmid (LASP1) or pCMV3-EZH2 expression plasmid (EZH2), or miR-218-5p mimic together with the pCMV3-LASP1 expression plasmid (LASP1), or EZH2 siRNA together with the miR-218-5p inhibitor were transfected to BxPC-3 and SW1990 cells. To establish the stable ANLN-silencing BxPC-3 cell line, short hairpin RNA (shRNA) oligonucleotide sequences that targeted ANLN were cloned into the pLV-hU6-shRNA-CMV-puromycin lentiviral vector by Sesh-biotech (Shanghai, China). Lentiviral Packaging System was then used for lentivirus packaging. The shRNA sequences are listed in Additional file 1: Table S1. BxPC-3 cells were infected with lentivirus at an MOI (multiplicity of infection) = 15 and selected with 3 μg/ml puromycin for 15 days.

Cells were harvested, and the total cellular proteins were extracted using RIPA buffer (Sigma-Aldrich, St. Louis, MO, USA) that contained a protease inhibitor cocktail (Roche, Basel, Switzerland). A BCA protein assay kit (Beyotime, China) was subsequently used to quantify the protein concentration. The proteins were then denatured in SDS sample buffers for 10 min at 100 °C, separated via SDS-PAGE and blotted onto PVDF membranes (Bio-Rad Laboratories, Hercules, CA, USA). The membranes were blocked in Tris-buffered saline that contained 5% nonfat powdered milk for 1 h. The membranes were subsequently probed with a mouse monoclonal antibody against human ANLN (1:500; Abcam, United Kingdom), mouse monoclonal antibody against human β-actin (1:3000; Proteintech, Wuhan, China), rabbit polyclonal antibody against human LASP1 (1:3000; Proteintech, China) or rabbit polyclonal antibody against human EZH2 (1:2000; Proteintech, China) at 4 °C overnight. After washing, the membranes were incubated with goat anti-mouse/rabbit secondary antibodies (1:3000; Santa Cruz Biotechnology, Dallas, TX, USA) at room temperature for 2 h. Finally, the ECL system (Thermo Scientific, Rockford, IL, USA) was used to visualize the protein band signals. The band density was analyzed by Quantity One v4.6.2 (Bio-Rad Laboratories, USA).

For the cell proliferation assay, transfected BxPC-3 and SW1990 cells were seeded onto 96-well plates at a final concentration of 2000 cells/well and further incubated for 0, 1, 2, 3, and 4 d at 37 °C. Following incubation, the cell viability was measured using a Cell Counting Kit-8 (CCK-8) kit (Beyotime, China) according to the manufacturer’s instructions. The CCK-8 assays were repeated 3 times.

After 48 h of transfection, BxPC-3 and SW1990 cells were collected, seeded onto 6-well plates at a final concentration of 500 cells/well and cultured for an additional 14 days. The cells were then stained with 0.05% crystal violet for 20 min. The number of colonies was evaluated via light microscopy. The colony formation assays were performed 3 times.

The cell migration assay was performed using transwell chambers (BD Biosciences, USA) with a pore size of 8 μm. The cell invasion assay was performed using Matrigel-coated transwell chambers (BD Biosciences, USA) with a pore size of 8 μm. At 48 h after transfection, BxPC-3 and SW1990 cells were seeded onto the upper chambers at a final concentration of 5 × 10⁴ cells/well and cultured in 100 μl of serum-free medium. The lower chambers contained 700 μl of medium with 10% FBS. After 48 h of incubation, the cells migrated or invaded through the filter into the lower side of the chamber were fixed and stained with crystal violet for 30 min. The number of cells was counted under a microscope. Each experiment was performed in triplicate.

This study was approved by the Ethical Committee of Army Medical University, China. The stable ANLN-silenced BxPC-3 cells (LV-ANLN shRNA, 1 × 10⁶ cells in 100 μl of sterilized PBS) and the stable scramble control BxPC-3 cells (LV-NC, 1 × 10⁶ cells in 100 μl of sterilized PBS) were injected into the right and left dorsal flanks of 4-week-old BALB/c male nude mice (Animal Center of the Chinese Academy of Science, Shanghai, China), respectively. Next, all mice were raised in a pathogen-free condition. The lengths and widths of the tumors were measured every week, and the tumor volume was calculated as follows: tumor volume = (length × width²)/2. At 5 weeks after injection, all mice were euthanized, and their tumors were dissected. The tumors were subjected to IHC staining to analyze the expression levels of ANLN, EZH2 and LASP1 with primary antibodies against ANLN (1:200; Abcam, Cambridge, United Kingdom), EZH2 (1:200; Abcam, Cambridge, United Kingdom) and LASP1 (1:200; Proteintech, China).

BxPC-3 cells were transfected with NC or ANLN RNAi and collected after 3 days, and three biological replicates were utilized. Total RNA was extracted using TRIzol reagent (Invitrogen, Grand Island, NY, USA). The RNA quality was determined by a spectrophotometer at 260 and 280 nm. The RNA integrity was evaluated by electrophoresis (1% formaldehyde denaturing gel). The RNA was subsequently synthesized into cDNA, and converted into cRNA. Labeled cRNA was hybridized to the Affymetrix Gene Chip Human Gene 1.0 ST Array (Affymetrix, Santa Clara, CA, USA). Expression Console and Transcriptome Analysis Console v3.0 (Affymetrix, USA) were used to analyze differentially expressed genes. Gene ontology (GO) annotation analysis was performed using DAVID Bioinformatics Resources 6.8 (https://​david.​ncifcrf.​gov/​).

Total RNA was isolated from cells using TRIzol reagent (Invitrogen, USA), according to manufacturer’s instructions. A spectrophotometer was used to determine the RNA quality and quantity. RNA was then reverse-transcribed into cDNA using M-MLV Reverse Transcriptase (TaKaRa, Dalian, China) and BulgeLoop™ specific RT-primers and random primers for mRNA. The gene expression levels were analyzed using a SYBR Premix Ex Taq kit (TaKaRa, China). The relative expression of gene was calculated using the 2^-ΔΔCt method. β-actin or U6 was used as an internal control. This experiment was performed with three biological replicates. All primer sequences are listed in Additional file 1: Table S1.

The partial wild-type sequence of the LASP1 3′-untranslated region (UTR) (2262 bp) containing the three putative miR-218-5p binding sites (Site1:686–692, Site2: 1587–1593 and Site3 2080–2087) or the sequences having mutations of the miR-218-5p putative binding sites in LASP1 3’UTR were cloned into the downstream of the luciferase gene in the psiCHECK-2 vector (Promega, Madison, WI, USA).

According to the GENT database, ANLN expression was significantly upregulated in 174 pancreatic cancer tissues compared with that in 62 normal tissues (P < 0.001, Fig. 1a) [39]. In addition, results from cBioportal for Cancer Genomic (www.​cbioportal.​org/​) showed that the mRNA upregulation of the ANLN gene accounted for most of the alterations (QCMG, Nature 2016; TCGA, Pancancer Atlas; TCGA, Provisional; UTSW, Nat Commun) (Fig. 1b) [40, 41]. Moreover, fewer ANLN shallow deletion and high-level ANLN gene amplification (amplification) and more ANLN diploid and low-level ANLN gene amplification (gain) were observed in two pancreatic cancer datasets (Pancreatic Adenocarcinoma-TCGA, PanCancer Atlas and Pancreatic Adenocarcinoma-TCGA, Provisional) (Fig. 1c). The ANLN mRNA expression in the pancreatic cancer samples with low-level ANLN gene amplification (gain) was markedly increased compared with that in the pancreatic cancer samples with ANLN diploids (Fig. 1c). In addition, Kaplan-Meier plots summarizing the result from the Human Protein Atlas (www.​proteinatlas.​org/​) showed that the upregulation of ANLN was significantly correlated with worse overall survival (P = 0.000002, Fig. 1d). Consistent with these publicly available data, our results also showed significantly upregulated ANLN expression in pancreatic cancer tissues compared with that in their matched adjacent normal pancreatic tissues (Fig. 1e). We subsequently analyzed the association between the ANLN expression and clinicopathological features in 80 pancreatic cancer samples. We found that ANLN upregulation was significantly correlated with the tumor size (P = 0.002), tumor differentiation (P = 0.027), TNM stage (P < 0.001), lymph node metastasis (P = 0.004) and distant metastasis (P = 0.005) (Table 1). In addition, univariate analysis indicated that the TNM stage, lymph node metastasis and upregulated ANLN expression were associated with overall survival (P = 0.002, P = 0.016 and P = 0.034, respectively) (Table 2). Multivariate analysis showed that the TNM stage, lymph node metastasis and upregulated ANLN expression were independent prognostic factors for overall survival (P = 0.001, P = 0.021 and P = 0.002, respectively) (Table 2). Kaplan-Meier analysis showed that upregulated ANLN expression was significantly associated with shorter survival for patients with pancreatic cancer (P = 0.001, Fig. 1e). Moreover, we analyzed the ANLN protein expression in one normal human pancreatic duct epithelial cell line (hTERT-HPNE) and five pancreatic cancer cell lines (AsPC-1, PANC-1, BxPC-3, MIA PaCa-2 and SW1990). Our results showed that ANLN protein expression was significantly increased in the five pancreatic cancer cell lines compared with that in the hTERT-HPNE cell line (Fig. 1f).

To determine the function of ANLN in pancreatic cancer, we conducted siRNA-mediated gene silencing. As shown in Fig. 2a, ANLN expression was significantly reduced in the BxPC-3 and SW1990 cell lines after transfection with ANLN siRNA (ANLN RNAi) compared with that in the negative control group (NC) (P < 0.01). After confirming the suppressive effect of ANLN siRNA by Western blot, we performed CCK-8 assays and showed that ANLN downregulation significantly inhibited BxPC-3 and SW1990 cell proliferation when compared with that in the NC group (Fig. 2b). In the colony formation assay, the colony numbers were significantly decreased in cells transfected with ANLN siRNA compared with those transfected with the control (Fig. 2c). The migration and invasion assay indicated that knockdown of ANLN significantly suppressed cell migration and invasion (Fig. 2d and e). Moreover, we established the stable ANLN silencing BxPC-3 cell line using lentiviral vectors with puromycin. The knockdown effect of ANLN-silencing lentiviral vectors was confirmed by Western blot. The ANLN-silencing lentiviral vectors (LV-ANLN shRNA) clearly inhibited the ANLN protein expression in BxPC-3 cells compared with that resulting from the control lentiviral vectors (LV-NC) (Fig. 2f). Moreover, the xenograft tumor volumes of the LV-ANLN shRNA group were obviously reduced when compared with those of the LV-NC group (P < 0.01, Fig. 2g).

To explore the potential mechanisms of ANLN in regulating pancreatic cancer progression, gene microarray analysis was performed using RNA isolated from BxPC-3 cells transfected with NC or ANLN RNAi. A total of 1926 genes were downregulated with fold changes < − 2.5 (P < 0.05), while 631 genes were upregulated with fold changes > 2.5 (P < 0.05) (Fig. 3a and b). The downregulated and upregulated genes were subsequently submitted to gene ontology (GO) annotation analysis using the DAVID Bioinformatics Resources 6.8 (https://​david.​ncifcrf.​gov/​). The Go terms representing biological processes indicated that these differentially expressed genes were highly enriched in cell-cell adhesion and cell cycle-related biological processes (Additional file 2: Table S2, Fig. 3c). The GO terms representing cellular compartment and molecular functions were shown in Additional file 3: Table S3 and Additional file 4: Table S4. Interestingly, cell-cell adhesion-related genes are largely involved in regulating cell growth, proliferation and motility [42, 43]. Based on GO analysis, 54 cell-cell adhesion-related genes were shown in Fig. 3d. According to the clustering results shown in Fig. 3d, eleven candidate genes were selected. To validate the microarray results, qRT-PCR was used to determine the expression of the eleven genes. Similar to the microarray results, the expression of all eleven genes in BxPC-3 cells transfected with ANLN RNAi was significantly decreased compared with that in the NC group (Fig. 3e).

Among the eleven candidate genes selected according to the clustering results, LASP1, RAB11B, RUVBL1 and MYO1B are upregulated in human malignant tumors and contribute to cancer progression [20, 44‐46]. Based on the GENT database, we showed that both LASP1 and RUVBL1 gene expression were significantly upregulated in pancreatic cancer tissues compared with that in normal pancreatic tissues, while RAB11B was not changed obviously, and MYO1B was significantly downregulated (Additional file 5: Figure S1A). In addition, both LASP1 and RUVBL1 expression were positively correlated with ANLN expression (Additional file 5: Figure S1B). Further investigation showed that LASP1 restoration partially reversed the effects of ANLN knockdown on pancreatic cancer cell proliferation (Additional file 5: Figure S1C). However, RUVBL1 restoration did not reverse the effect of ANLN downregulation on pancreatic cancer cell proliferation (Additional file 5: Figure S1C). Thus, ANLN may promote pancreatic cancer cell progression by regulating LASP1. To confirm this hypothesis, we rescued LASP1 expression using LASP1 overexpression plasmid vectors in the cells transfected with ANLN RNAi. As shown in Fig. 4a, the LASP1 protein expression in ANLN RNAi-transfected cells was restored by the LASP1 plasmid. In addition, the CCK-8 and colony formation assays indicated that the suppressive effects of ANLN knockdown on pancreatic cancer cell growth were restored by LASP1 re-expression (Fig. 4b and c). Moreover, LASP1 re-expression partially reversed the effect of ANLN knockdown on pancreatic cancer cell migration and invasion (Fig. 4d and e).

Previous reports have shown that microRNAs (miRNAs) play an important role in cancer initiation and tumor progression by directly regulating target genes expression [47]. Here, we found that a total of 49 miRNA precursors showed significant changes in gene expression profiles; 46 miRNA precursors were upregulated, and 3 miRNA precursors were downregulated (Additional file 6: Figure S2A). By combining the gene expression profiles and Targetscan (http://​www.​targetscan.​org/​), three miRNAs (miR-145-5p, miR-218-5p and miR-9-5p) were found both upregulated in ANLN downregulated BxPC-3 cells and contained binding sites of the 3’UTR of LASP1 (Additional file 6: Figure S2B). To determine the effect of miR-145-5p, miR-218-5p and miR-9-5p on LASP1 expression, miR-145-5p, miR-218-5p and miR-9-5p mimics were transfected into BxPC-3 cells, respectively. We found that the expression of miR-145-5p, miR-218-5p and miR-9-5p in transfected cells was significantly upregulated (Additional file 6: Figure S2C). Additional qRT-PCR showed that only miR-218-5p upregulation significantly repressed LASP1 mRNA expression in BxPC-3 cells (Additional file 6: Figure S2D). To further confirm the effect of miR-218-5p on LASP1 expression, the mimic control (con), miR-218-5p mimics (miR-218-5p) were transfected into BxPC-3 and SW1990 cells, respectively. As shown in Fig. 5a, the expression of miR-218-5p in transfected cells was significantly upregulated. Moreover, miR-218-5p upregulation significantly suppressed LASP1 protein expression in BxPC-3 and SW1990 cells (Fig. 5b). To determine whether miR-218-5p directly targets the 3’UTR of LASP1, the luciferase reporter vectors containing the 3’UTR of LASP1 (wt-LASP1) or its mutant version were constructed (Fig. 5c). We found that the luciferase activity of miR-218-5p-transfected cells was significantly inhibited compared with that of the con-transfected cells. In addition, site-directed mutagenesis of the miRNA binding sequences in the 3’UTR of LASP1 showed a slightly different trend. The conserved miRNA binding site (Site3) appeared to be slightly more effective in miR-218b-5p-induced luciferase activity repression because mutation of this site significantly reversed a decrease in the luciferase signal (Fig. 5d). However, the miR-218-5p-mediated repression of luciferase activity was abolished by mutating the three binding sequences (triple mutation) of the LASP1 3’UTR (Fig. 5d). To further determine the role of LASP1 in the suppressive effects of miR-218-5p on pancreatic cancer progression, we rescued the expression of LASP1 by LASP1 overexpression plasmid vectors in the cells transfected with miR-218-5p. As shown in Fig. 5e, LASP1 protein expression in miR-218-5p-transfected cells was restored by LASP1 plasmid. In addition, CCK-8 analysis and colony formation assay revealed that the suppressive effects of miR-218-5p upregulation in pancreatic cancer cell growth were partially restored by LASP1 re-expression (Fig. 5f and g). Moreover, LASP1 re-expression partially reversed the inhibition of cell migration and invasion caused by miR-218-5p upregulation (Fig. 5h and i).

In this study, we showed that miR-218-5p upregulation inhibited pancreatic cancer cell growth, migration and invasion by directly regulating LASP1 expression. Moreover, ANLN knockdown significantly induced the expression of miR-218-5p. Thus, ANLN may regulate LASP1 expression and pancreatic cancer progression by miR-218-5p. To determine whether miR-218-5p was involved in ANLN-induced LASP1 expression and pancreatic cancer cell growth, migration and invasion, miR-218-5p inhibitor (anti-miR-218) was used to reverse the expression of miR-218-5p upregulation caused by ANLN knockdown. As shown in Fig. 6a, anti-miR-218 obviously reversed the ANLN knockdown-induced miR-218-5p expression. In addition, the LASP1 protein levels were restored in the cells cotransfected with ANLN RNAi and anti-miR-218 compared with the protein levels in the cells cotransfected with ANLN RNAi and inhibitor control (anti-con) (Fig. 6b). In functional assays, miR-218-5p knockdown in pancreatic cancer cells transfected with ANLN RNAi rescued the inhibition of cell proliferation, colony formation, cell migration and cell invasion caused by ANLN knockdown (Fig. 6c-f). Collectively, these results demonstrated that ANLN promotes pancreatic cancer cell growth, migration and invasion by regulating miR-218-5p/LASP1 signaling axis.

In pancreatic cancer, EZH2 is upregulated and induces pancreatic cancer cell proliferation, cell migration and cell invasion in vitro and tumor growth and metastasis in vivo by inhibiting the expression of miR-218-5p [37]. In this study, the result of gene expression profiles showed that ANLN knockdown significantly inhibited the expression of EZH2 (Additional file 7: Figure S3A). To further confirm the effect of ANLN knockdown on EZH2, qRT-PCR and Western blot analysis was performed. Our results showed that ANLN knockdown significantly repressed the expression of EZH2 mRNA and protein (Additional file 7: Figure S3B and C). Therefore, ANLN may promote pancreatic cancer cell progression and miR-218-5p/LASP1 signaling axis by mediating EZH2. To determine whether EZH2 regulates LASP1, we transfected the BxPC-3 and SW1990 cells with the negative control siRNA (NC) or EZH2 siRNA (EZH2 RNAi). As shown in Fig. 7a, EZH2 downregulation significantly inhibited the expression of LASP1 protein. In addition, to further confirm whether EZH2 downregulation could inhibit LASP1 expression by inducing miR-218-5p expression, miR-218-5p inhibitor (anti-miR-218) was used to reverse the EZH2 knockdown-induced miR-218-5p expression. As shown in Fig. 7b, anti-miR-218 obviously reversed the EZH2 knockdown-induced miR-218-5p expression. Moreover, the LASP1 protein levels were restored in the cells cotransfected with EZH2 RNAi and anti-miR-218 compared with the protein levels in the cells cotransfected with EZH2 RNAi and inhibitor control (anti-con) (Fig. 7c). To further determine whether EZH2 was involved in ANLN-induced LASP1 expression and the inhibition of miR-218-5p, EZH2 expression plasmid vectors were used to rescue the inhibition of EZH2 caused by ANLN knockdown. As shown in Fig. 7d, the ectopic expression of EZH2 reversed the upregulation of miR-218-5p expression caused by ANLN knockdown. In addition, EZH2 re-expression reversed the inhibition of LASP1 caused by ANLN knockdown (Fig. 7e). In functional assays, EZH2 restoration in pancreatic cancer cells transfected with ANLN RNAi rescued the inhibition of cell proliferation, colony formation, cell migration and cell invasion caused by ANLN knockdown (Fig. 7f-h).

To confirm the effect of ANLN knockdown on the expression of EZH2 and LASP1, we detected the levels of ANLN, EZH2 and LASP1 expression in tissues from in vivo xenograft tumor models established with the stable ANLN-silenced BxPC-3 cells or the stable scramble control BxPC-3 cells by IHC and Western blot. IHC staining showed that the levels of EZH2 and LASP1 expression were repressed after ANLN knockdown (Fig. 8a). Consistent with these findings, Western blot results also showed that the levels of EZH2 and LASP1 expression were significantly inhibited in the LV-ANLN-shRNA group when compared to the LV-NC group (Fig. 8b). Taken together, ANLN may promote pancreatic cancer cell growth, migration and invasion by regulating EZH2/miR-218-5p/LASP1 signaling axis (Fig. 9).