ATP11A promotes EMT by regulating Numb PRR^L in pancreatic cancer cells

Confirmation statement

We confirm that all experiments were carried out in accordance with the relevant guidelines and regulations.

Bioinformatic analysis

The data of 178 pancreatic cancer patients were retrieved from the TCGA (The Cancer Genome Atlas) database, and the data of 171 normal pancreas were retrieved from the GTE_X (Genotype-Tissue Expression) database. A differential expression analysis was performed between the cancer tissues and the normal issues using Sangerbox tools, a free online platform for data analysis (http://www.sangerbox.com/tool). The relationship between ATP11A and Numb was determined based on the data of 178 pancreatic cancer patients obtained from the TCGA database using the online tool, ENCORI (Li et al., 2014).

Cell culture and tissue samples

The Ethics Committee of the First Hospital of China Medical University approved this study (Committee Approval Number: AF-SOP-07-1.1-01). Human pancreatic carcinoma cell lines, SW1990 and PANC-1, were purchased from the Chinese Academy of Sciences Cell Bank (Shanghai, China). All cell lines were maintained in RPMI-1640 medium with 10% FBS (Gibco Invitrogen, Carlsbad, CA, USA). Cells were grown at 37 °C in a humidified 5% CO₂ incubator.

In addition, pancreatic cancer samples were acquired from patients who had surgical resections. It is worth noting that the application for the exemption of informed consent was approved by the Ethics Committee of the First Hospital of China Medical University.

Cell transfection

For cell transfection, Lipofectamine™ 3000 reagent protocol was used. ATP11A plasmids (Genechem, Shanghai, China), NC plasmids (Genechem, Shanghai, China), ATP11A siRNA (GenePharma, Shanghai, China), Numb PRR^L siRNA (GenePharma, Shanghai, China), and negative control oligonucleotides (GenePharma, Shanghai, China) were transfected using the Lipofectamine 3000 (Invitrogen, Waltham, CA, USA). First, seed cells to be 70–90% confluent at transfection. Second, dilute the Lipofectamine 3000 reagent in Opti-MEM medium and prepare the master mix of DNA by diluting the DNA in Opti-MEM medium, then add P3000 reagent. Third, add diluted DNA to the diluted Lipofectamine 3000 Reagent. Then incubate and add DNA-lipid complex to cells. Incubate cells for 48 h and then perform the subsequent experiments. To transfect the cells with siRNA, follow the protocol as described for DNA but do not add P3000 reagent when diluting the siRNA. File S1 contains these sequences.

Western blot

Pancreatic cancer cells were seeded in 12-well plates (1 × 10⁵ cells/well) and harvested 48 h after transfection. Proteins were extracted from tissues or cells using RIPA Lysis buffer supplemented with phenylmethylsulfonyl fluoride, protease inhibitor, phosphatase inhibitor. Protein concentration was quantified using bicinchoninic acid (BCA) protein concentration determination kit (TaKaRa, Japan). Samples were then resolved on 10% SDS–polyacrylamide gels and transferred to nitrocellulose. PVDF membranes were blocked with 5% skimmed milk for 2 h, then incubated overnight with primary antibodies against: E-cadherin (Abcam, Cambridge, UK), N-cadherin (Proteintech, Rosemont, IL, USA), Fibronectin (Proteintech, Rosemont, IL, USA), Vimentin (Proteintech, Rosemont, IL, USA), Snail2 (Proteintech, Rosemont, IL, USA), ZEB1 (Proteintech, Rosemont, IL, USA), Numb (Abcam, Cambridge, UK), ATP11A (Abcam, Cambridge, UK), SRPK2 (Abcam, Cambridge, UK) and GAPDH (Proteintech, Rosemont, IL, USA). The second day, secondary antibodies incubations were performed using horseradish peroxidase(HRP)-conjugated secondary antibodies (Proteintech, Rosemont, IL, USA). Finally, an ECL-chemiluminescent kit (Thermol Biotech Inc, Waltham, MA, USA) was used to detect the protein bands.

Immunohistochemistry (IHC)

Paraffin-embedded tissues were dewaxing with dimethylbenzene and treated with ethanol at various concentrations. Antigen retrieval was performed using the high pressure method. A total of 3% hydrogen peroxidase was used to quench the endogenous peroxidase activity. The tissues were subsequently incubated with 10% ordinary goat serum. Next, sections were incubated overnight with anti-ATP11A primary antibodies (Abcam, Cambridge, UK, 1:200). Then secondary antibodies were added to incubate the sections. Diaminobenzidine (DAB) was utilized to visualize the sections, followed by counterstainining with hematoxylin. Sections were then observed and positive cells counted under a microscope. For each section, five high-power fields (magnification × 400) were randomly chosen. Staining intensity was scored 0 (negative) to 3 (strong intensity staining). Based on the percentage of staining, the extent was scored as 0 (<5%), 1 (5–25%), 2 (26–50%), 3 (51–75%), or 4 (>75%). IHC was evaluated by three independent pathologists. These two scores were then multiplied to make a final IHC score ranging from 0 to 12. Positive expression was defined as a IHC score >8, due to the high expression of ATP11A in pancreatic cancer tissues. Setting the standard as eight allowed us to better distinguish the clinicopathological parameters of the high expression group from the low expression group.

Transwell invasion and migration assay

Transfected cells were collected 48 h later after transfection and then resuspended in the serum-free RPMI-1640 culture medium. A total of 3 × 10⁴ transfected cells were placed on the top compartment pretreated with Matrigel. The lower chamber was filled with medium containing 10% serum. Cold methyl alcohol was used to fix the cells on the lower membrane. Then cells were stained using 1% crystal violet stain solution (Solarbio, Beijing, China). Finally, invaded cells were counted in five arbitrarily chosen fields for each chamber. Fields were counted by different authors that were blind to conditions. The protocol for the cell migration assay was similar to the invasion assay protocol, with the exception of Matrigel coating.

RNA extraction and rt-PCR validation

RNA from tissue samples was extracted using TRIzol reagent (TAKARA, Japan). Complementary DNA (cDNA) was synthesized using the PrimeScript RT reagent Kit with gDNA Eraser (TAKARA, Japan). qPCR was performed with a TB Green Premix Ex Taq II (TAKARA, Japan) with a QuantStudio 3 Real-Time PCR Instrument (Applied Biosystems, Waltham, MA, USA). The sequences of PCR primers of Numb PRR^L are CTTCCAAGCTAATGGCACTG (Forward primer) and CTCTTAGACACCTCTTCTAACCA (Reverse primer). The PCR cycling conditions were as follows: 95 °C for 30 s, 45 cycles of 95 °C for 5 s, 70 °C for 45 s, dissociation at 95 °C for 15 s, 60 °C for 1 min and 95 °C for 1 s. The results were analyzed using the 2(−ΔΔC(T)) Method.

Immunoprecipitation

For immunoprecipitation, Santa Cruz Biotechnologys Immunoprecipitation protocol was used. Entire protein lysates were extracted using RIPA lysis buffer supplemented with 1 mM phenylmethylsulfonyl fluoride (PMSF), protease inhibitor, phosphatase inhibitor (Bimake, Houston, TX, USA). Lysates were precleared by adding a 1:2 protein A/G PLUS agarose bead slurry (Santa Cruz Biotechnology, Santa Cruz, CA, USA), incubating for 10 min with agitation, centrifuging, and retaining the resulting supernatant. The rabbit polyclonal antibody ATP11A (Abcam, Cambridge, UK) and mouse monoclonal antibody IgG (Santa Cruz Biotechnology, Santa Cruz, CA, USA) antibodies were preincubated with magnetic beads. Next, the antibody-beads complex was washed three times with lysis buffer and incubated with the soluble supernatants of the protein lysates overnight at 4 °C. After centrifuging the samples at 14,000 rpm, the pellet was washed with PBS, resuspended, and boiled for 5 min to remove immunocomplexes from the beads before centrifugation to pellet the beads. Subsequently, the material was eluted and performed a SDS-PAGE test and western blot analysis.

EMT construction

TGFβ1 (10 ng/mL) (Peprotech, RockyHill, NJ, USA) was applied to PANC-1 and SW1990 cells for 48 h. Both cell lines received a control treatment of 1% bovine serum albumin (Sigma-Aldrich, St. Louis, MO, USA). Cells were cultured in growth media containing 1% fetal calf serum, as recommended. EMT was discovered by observing EMT-like cell morphology and invasion, as well as changes in E-cad, Vimentin, and N-cad protein levels.

Statistical analysis

All data are presented as the means ± SEM of at least three independent experiments. Differences between two groups were assessed using Student’s t-test, while differences between multiple groups were determined by one-way ANOVA. The correlation of Numb and ATP11A was analyzed using the Pearson correlation coefficient. The clinicopathologic significance of ATP11A expression was analyzed using the Chi-squared test. The overall survival was analyzed using Kaplan–Meier curves, while the log-rank test was used for comparison. The differences of WB analysis, cell migration and invasion assays were compared using Student’s t-test. SPSS software (SPSS 21.0; SPSS Inc., Chicago, IL, USA) was used for statistical testing, and a P-value of less than 0.05 was considered to be significant.

A bioinformatics analysis found that ATP11A was highly expressed in pancreatic cancer and had a strong correlation with Numb

Among 27 types of malignant tumors, ATP11A was highly expressed in 16 types of tumors, had a low expression in six types of tumors, and had no statistically significant difference in five types of tumors, indicating that ATP11A is generally overexpressed in tumors. In pancreatic cancer tissues, ATP11A expression was remarkably higher than in normal tissues (P < 0.001; Fig.1A). In addition, Numb was strongly correlated with the expression of ATP11A (R = 0.676; Fig.1B).

A western blot analysis verified the high expression of ATP11A in pancreatic cancer

The expression of ATP11A protein in 31 pancreatic cancer tissues and corresponding paracancerous tissues was measured using western blot. (Fig.1C). The results showed that the expression of ATP11A was significantly higher in pancreatic cancer tissues than in paracancerous tissues (P = 0.0009; Fig.1D).

Immunohistochemistry confirmed that ATP11A expression was enhanced in pancreatic cancer tissues and correlated with clinicopathological parameters

The immunohistochemical results of 81 pancreatic cancer cases showed that ATP11A was mainly expressed in the ductlike structure of pancreatic cancer, and was expressed in the cytomembrane and cytoplasm of pancreatic cancer cells. The expression of ATP11A was higher in cancer tissues (Figs. 1E, 1F ) than in paracancerous tissues (Figs. 1G, 1H), and was associated with the AJCC stage of pancreatic cancer (P = 0.045; Table 1). Moreover, the low ATP11A expression group had a better prognosis than the high ATP11A expression group (P = 0.0025) (Fig. 1I).

ATP11A could induce EMT and regulate the expression of Numb PRR^L and EMT-related proteins

After transfection of PANC-1 and SW1990 cells with ATP11A overexpression plasmid, significant changes were found in cell morphology (Figs. 2A, 2B ), which indicated that the overexpression of ATP11A promoted the occurrence of EMT in pancreatic cancer cells. WB experiments were conducted to verify our conjecture. According to the western blot analysis results, the overexpression of ATP11A led to a concomitant increase in the expression of Numb PRR^L, ZEB1, Fibronectin, N-cadherin, Vimentin, Snail2 and SRPK2, and decreased the expression of E-cadherin. On the other hand, ATP11A knockdown decreased the expression of Numb PRR^L, ZEB1, Fibronectin, N-cadherin, Vimentin, Snail2 and SRPK2 and increased the expression of E-cadherin (Figs. 2C, 2D).

ATP11A could significantly promote the invasion and migration ability of pancreatic adenocarcinoma cells

Cell function experiments were conducted to further explore the effects of ATP11A on the biological behavior of pancreatic cancer cells. The invasion and migration experiments showed that increased ATP11A expression significantly promoted the invasion and migration ability of pancreatic adenocarcinoma cells, while decreased ATP11A expression significantly decreased the invasion and migration ability of pancreatic adenocarcinoma cells (Figs. 3A, 3B).

Numb PRR^L is highly expressed in pancreatic cancer tissues, and is important for maintaining the aggressive pancreatic cancer cell phenotype and EMT

After confirming the association between ATP11A and Numb PRR^L and the promoting effect of ATP11A on EMT, we decided to explore the expression of Numb PRR^L in pancreatic cancer tissues and whether Numb PRR^L affects EMT. Rt-PCR results showed that Numb PRR^L expression in pancreatic cancer tissues was higher than that in paracancer tissues (Fig. 4A). In order to further investigate the effects of Numb PRR^L on the invasion and migration ability of pancreatic cancer cells, transwell assays were conducted and results showed that the invasion and migration ability of PANC-1 cells and SW1990 cells was significantly reduced after specific knockdown of Numb PRR^L (Fig. 4B). These results indicate that Numb PRR^L plays a major role in maintaining the malignant biological behavior of pancreatic cancer. Furthermore, Numb PRR^L can affect EMT in pancreatic cancer cells. After Numb PRR^L specific knockdown, the expression of E-cadherin was up-regulated, while the expression of N-cadherin, Vimentin, Fibronectin, Snail2 and ZEB1 were down-regulated (Fig. 4C), suggesting that Numb PRR^L can promote EMT in pancreatic cancer cells.

ATP11A regulated the expression of ZEB1 and Snail2 by regulating Numb PRR^L, thus affecting EMT

Rescue experiments were performed in order to explore the association between ATP11A, Numb PRR^L, and transcription factors ZEB1/Snail2. Numb PRR^L was specifically knocked down after the overexpression of ATP11A, then observing whether the effect of ATP11A overexpression on the biological behavior of pancreatic cancer cells was recovered. Results indicated that the increased invasion and migration ability induced by ATP11A was blocked after specific inhibition of Numb PRR^L (Figs. 5A, 5B ). In addition, western blot assay results showed that specific knockdown of Numb PRR^L did not affect the expression of ATP11A, but the up-regulation of ZEB1 and Snail2 induced by ATP11A was blocked (Figs. 5C, 5D). These results suggest that ATP11A up-regulates the expression of ZEB1 and Snail2 by regulating Numb PRR^L.

ATP11A coimmunoprecipitates with Numb PRR^L directly in PC cells

Immunoprecipitation assays can detect direct binding of two proteins in cells. Our immunoprecipitation results (Fig. 6) showed that ATP11A could specifically bind to Numb PRR^L, but could hardly bind to Numb PRR^S in PANC-1 and SW1990 cells. These results indicate that ATP11A and the Numb PRR^L protein have a direct binding effect in pancreatic cancer cells.

The effect of ATP11A on EMT is TGFB dependent

After induction with TGFB1, PANC-1 and SW1990, cells showed obvious morphological changes (Figs. 7A, 7B). To explore whether TGFB can promote the expression of ATP11A, cell proteins were extracted and WB experiments were conducted. The results showed that after TGFB induction, the expression of ATP11A significantly increased, accompanied by changes of EMT-related proteins: the expression of E-Cadherin was down-regulated, and the expression of N-Cadherin, Vimentin, Fibronectin, Snail2 and ZEB1 were up-regulated (Figs. 7C, 7D). This indicates that the promoting effect of ATP11A on EMT is TGFB dependent.