RPMI-1640 medium and fetal bovine serum (FBS) were obtained from Thermo Fisher Scientific (Waltham, MA, United States). Matrigel was purchased from BD Biosciences (San Jose, CA, United States). Antibodies against Bax, Bax-2, caspase-3, caspase-9, p21 MMP-2, MMP-9, PRMT5, LKB1, AMPK, phosphorylation (p)-AMPK, mTOR, p-mTOR, and GAPDH and secondary antibody were purchased from Abcam (Cambridge, MA, United States).

Esophageal squamous cell carcinoma tissue samples and paired adjacent normal tissue samples were collected from the same 40 patients who underwent surgery at the Heze Municipal Hospital between June 2017 and December 2018. The tissues were determined by immunohistochemical analysis at the Heze Municipal Hospital. All patients have given written informed consents, and the study protocol was approved by Heze Municipal Hospital. The clinic information of patients is shown in Table 1; and we tested the expression level of PRMT5 in 40 patient samples and selected the median as the classification standard, low expression for those lower than the median, and high expression for those higher than the median.

The ESCC cell lines ECA109 and KYSE150 and the human normal esophageal epithelial cell line Het-1A were purchased from the Shanghai Institute of Cell Biology (Chinese Academy of Sciences, Shanghai, China). All cells were cultured in RPMI-1640 medium supplemented with 10% FBS at 37°C. Cells were tested for mycoplasma (MycoAlert Mycoplasma Detection Kit, Lonza) and certified mycoplasma free.

TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.) was used to extract total RNA from tissue samples and cells. Equal amounts of RNA were transcribed into cDNA using the cDNA First Strand Synthesis kit (Beyotime Institute of Biotechnology, Haimen, China). Total cDNA was used as a starting material for real-time PCR on a StepOne real-time PCR system (Life Technologies Corp.) using a FastStart Universal SYBR Green Master (Roche Applied Science, Mannheim, Germany). The 2^{–Δ Δ CT} comparative method was used to normalize the expression levels of each target gene to corresponding GAPDH threshold cycle (CT) value. The primers are as follows: PRMT5: forward 5′-TGTAGGGAGAAGGACCGTGA-3′; PRMT5: reverse 5′-ATGGCTGAAGGTGAAACAGG-3′; LKB1: forward 5′-AGGGATGCTTGAGTACGAACC-3′; LKB1: reverse 5′-GTCCTCCAAGTACGGCACC-3′; GAPDH: forward 5′-ACGGATTTGGTCGTATTGGGCG-3′; GAPDH: reverse 5′-CTCCTGGAAGATGGTGATGG-3′.

Experimental method was as described above (Chung et al., 2019). Briefly, radioimmunoprecipitation assay (RIPA) lysis buffer was used to lyse the total protein from tissues or cells. Samples were fractionated on a 10% sodium dodecyl sulfate (SDS)–polyacrylamide gels, transferred to polyvinylidene difluoride membranes, and blocked in 5% skim milk in TBST. Subsequently, the membrane was incubated with 1:1,000 dilutions (v/v) primary antibody overnight at 4°C. After the membrane was washed with TBST, it was incubated with a second antibody. Detection was performed using a ChemiDoc Imaging System (Bio-Rad Laboratories, Inc., Hercules, CA, United States). The relative protein expression was analyzed using GAPDH as a control.

Experimental method was as described above (Chung et al., 2019). Briefly, lung samples were fixed and cut into 4-μm-thick sections, dried, deparaffinized, and dehydrated in a graded ethanol series and finally incubated with H₂O₂. Following blocking with bovine serum, sections were incubated overnight with monoclonal antibodies (1:1,000 dilutions) of PRMT5, LKB1, and Ki-67. The secondary antibody (1:500 dilutions) was used to incubate the sections, and then the sections were incubated with horseradish peroxidase-conjugated streptavidin. A 3,3′-diaminobenzidine (DAB) substrate kit was used, and slides were counterstained with hematoxylin. The laser scanning confocal microscope (Olympus Corporation, Tokyo, Japan) was used to visualize fluorescence.

Transfected cells (2 × 10⁵) were cultured for 24 h at 37°C on a Petri dish and fixed in 4% formaldehyde. After phosphate-buffered saline (PBS) washes, cells were penetrated and blocked with 5% bovine serum albumin (BSA) for 1.5 h. Then, the cells were incubated with the primary antibody (1:1,000) overnight at 4°C and then incubated with the secondary antibody (1:500) for 30 min. DAPI (1:1,000) was used to stain the nuclei. The laser scanning confocal microscope was used to visualize fluorescence.

PRMT5 siRNA duplexes were obtained from Dharmacon RNA Technologies (Chicago, IL, United States). PcDNA-LKB1 and pcDNA3.1 NC were purchased from GenePharma Company (Shanghai, China). ECA109 and KYSE150 were seeded in six-well plates at a density of 2 × 10⁵ cells and transfected with siRNAs or pcDNA-LKB1 for 24 h at 37°C using Lipofectamine 2000 (Life Technologies). Transfected cells were cultured for 48 h and used for further experiments. The sequence of siRNA PRMT5: 5′-ACAAAGUCCGGAAGUUGUGCC-3′; 3′-CACAACUUCCGGACUUUGUGU-5′.

For Cell Counting Kit-8 (CCK-8) assay, transfected cells (2 × 10⁵) were seeded in 96-well plates and cultured for 0, 24, 48, and 72 h at 37°C. After treatment with CCK-8 (KeyGEN, Jiangsu, China), optical density (OD) was measured at 450 nm at each point using a microplate reader (Thermo Fisher Scientific, Waltham, MA, United States).

5-Ethynyl-2′-deoxyuridine (EdU) incorporation proliferation assay was carried out to evaluate the proliferation of ECA109 and KYSE150 cells. EdU staining was conducted using Cell-Light EdU Imaging detecting kit (Ruibo Biotech, Guangzhou, China) according to the manufacturer’s protocol. The ratio of EdU-positive nuclei to total nuclei was calculated as the proliferation rate of cells in five random high-power fields per well, and the cells were visualized using a fluorescence microscope.

The effect of PRMT5 on the apoptosis of ECA109 and KYSE150 cells was evaluated by flow cytometry using the Annexin V Apoptosis Detection kit I (BD Biosciences, San Jose, CA, United States), and the effect of PRMT5 on cell cycle was detected with Cell Cycle Kit. After transfection for 24 h at 37°C, ECA-109 and KYSE150 cells (2 × 10⁵) were digested and washed twice in PBS. Then they were incubated with fluorescein isothiocyanate (FITC) annexin V and propidium iodide (PI) in the dark. The percentage of apoptotic cells and cell cycle was evaluated by FACS Calibur (BD Biosciences, United States).

Transfected ECA109 and KYSE150 cells (2 × 10⁵) were seeded in a six-well plate at 37°C until the fusion reached over 90% and a 200-μl pipette was used to create a linear wound. Then cells were cultured in serum-free Dulbecco’s modified Eagle’s medium (DMEM) (Gibco; Thermo Fisher Scientific, Inc.) at 37°C. After 24 and 48 h, an inverted microscope was used to observe cell migration and to count the number of cells that migrated. Cell migration was assessed by calculating scratch width using the following formula: the relative scratch width = (number of cells at T48 - number of cells at T0)/number of cells at T0 × 100%, where T0 was 0 h and T48 was 48 h.

Experimental method was as described above (Wang et al., 2020). (1) Migration assay: transfected ECA109 and KYSE150 cells (2 × 10⁵) were seeded to the top chamber at a density of 6 × 10³ cells, and medium supplemented with 20% FBS and added to the bottom chamber. Twenty-four hours after incubation at 37°C, cells were fixed with 4% formaldehyde and stained with 0.1% crystal violet solution (Servicebio, Wuhan, China). After PBS washing, a microscope was used to count the number of cells. (2) Invasion assays: filters were precoated with Matrigel (BD Biosciences, San Jose, CA, United States), and then transfected cells were plated into the upper chamber. Other operations are the same as above.

A total of 100 μl 2 × 10⁶ KYSE150 cells were dissolved in the Matrigel (BD Biosciences, San Jose, CA, United States) and stably expressed blank, control-siRNA, or PRMT5 siRNA. Then, the cells were injected into the axilla of the male Balb/c nude mice (total of 18; n = 6, 6 weeks old, 18–22 g), which were obtained from Beijing Vitalriver Experimental Animal Technology Co., Ltd. The animals were housed under room temperature (18–25°C) and a 12-h light/dark cycle with food and water ad libitum. Animal welfare and experimental procedures complied with national guide and was approved by the Heze Municipal Hospital. Animal health and behavior were monitored every day, and the weight of mice was measured every 4 days. When the mice show obvious decrease in activity, abnormal diet, emaciation, and continuous weight loss, they should be immediately anesthetized and sacrificed according to humane endpoints. The experimental mice were routinely monitored and sacrificed on day 32 by cervical dislocation after an intraperitoneal injection of sodium pentobarbital (100 mg/kg). The tumors were removed, and the length and width of the tumors were manually monitored using a Vernier caliper. For lung metastasis model, mice (total of 18; n = 6, 6 weeks old, 18–22 g) were injected 100 μl 2 × 10⁵ KYSE150 cells into their tail vein. After mice were sacrificed on day 32, the lung tissues were removed and weighed.

Tumor tissues were fixed in 4% paraformaldehyde, embedded in paraffin, and serially sectioned. This section was stained with hematoxylin/eosin (H&E).

With the use of a TUNEL assay kit (Roche, United States) to detect the apoptosis of tumor sections, briefly, sections were deparaffinized, then permeabilized with proteinase K, treated with 3% H₂O₂, and incubated with terminal deoxynucleotidyl transferase (TdT) enzyme. Then, sections were incubated with antidigoxigenin–peroxidase conjugate and using DAB to evaluate activity. The sections were examined under an optical microscope.

Statistical analysis was performed using GraphPad Prism 5 software. Data are shown as mean and SD. Chi-square test was performed for continuous or discrete data analysis, and the Student paired t-test was performed to evaluate significant differences between two independent groups of samples. One-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was applied to compare differences among multiple groups. P < 0.05 was considered statistically significant.

Initially, we found PRMT5 is an oncogenic lncRNA in ESCC based on The Cancer Genome Atlas (TCGA) databases. The result also revealed that PRMT5 was significantly upregulated in ESCC tissues compared with normal tissues by using the Gene Expression Profiling Interactive Analysis (GEPIA) database¹ (Figure 1A). Moreover, RT-PCR analysis showed the PRMT5 expression was remarkably stronger in ESCC tissues than the control tissues, as well as western blotting and immunohistochemistry (IHC) assays (Figures 1B–D). The correlations between the expressions of PRMT5 of the clinic pathological characteristics of 40 patients with ESCC were analyzed. We tested the expression level of PRMT5 in 40 patient samples and selected the median as the classification standard, low expression for those lower than the median, and high expression for those higher than the median. As shown in Table 1, there were significant associations between PRMT5 levels with tumor size, pathological staging, and metastasis. However, there was no association with age and gender. Meanwhile, compared with the HET-1A cell line, a human esophageal epithelial cell line, the PRMT5 expression, was higher in the ESCC cell lines (ECA109 and KYSE150) (Figures 1E,F). Western blotting showed the same result that PRMT5 was upregulated in ESCC tissues and cell lines (ECA109 and KYSE150). In addition, the result of IHC also revealed that the PRMT5 expression was remarkably stronger in ESCC tissues than the control tissues, while the result of immunofluorescence confirmed that PRMT5 was overexpressed in the ESCC cell lines (Figure 1G).

It showed that the PRMT5 expression in clinical ESCC tissues and ESCC cell lines was abnormal, and we next explored the effects of PRMT5 in ESCC cells by transfecting siRNA-PRMT5 in ECA109 and KYSE150 cells. Transfection with siRNA-PRMT5 markedly downregulated PRMT5 mRNA and protein expressions, compared with the control siRNA (Figures 2A–C).

The CCK-8 assay and EdU assay were used on ECA109 and KYSE150 cells in order to evaluate the impact of PRMT5 on ESCC cell proliferation. CCK-8 assay showed that knocking down PRMT5 obviously inhibited the cell proliferation by 80.9% and 85.2% at 72-h point in ECA109 and KYSE150 cells, compared with the control siRNA (Figure 2D). In addition, EdU assay was also revealed that the percentage of positive proliferative cells in siRNA-PRMT5 group was significantly weakened, which indicated that low expression of PRMT5 suppresses cell proliferation (Figure 2E).

To further explore the role of PRMT5 in ESCC cell apoptosis, flow cytometry analysis was applied. The cells in the upper-right (UR, Q2) and lower-right (LR, Q3) quadrants of the fluorescence-activated cell sorting (FACS) histogram represent apoptotic cells after culture for 48 h. As shown in Figure 3A, after being transfected with siRNA-PRMT5, the apoptosis rates of ECA109 and KYSE150 cells were 20% and 21%, respectively, which were significantly higher than those of the controls, indicating that down-expression of PRMT5 could induce cell apoptosis. In addition, compared with the controls, downregulating PRMT5 dramatically upregulated the expression levels of Bax, caspase-3, and caspase-9, while Bcl-2 level was downregulated in ECA109 and KYSE150 cells (Figure 3B). Taken together, PRMT5 could inhibit the cell apoptosis of ESCC.

Furthermore, flow cytometry was also used to test the activity of PRMT5 on cell cycle in ESCC cells. After being transfected with siRNA-PRMT5, the rates of G1 in ECA109 and KYSE150 cells were 70% and 72%, respectively, which were significantly higher than the control siRNA, indicating that knocking down PRMT5 could induce cell arrest in G1 phase in both ECA109 and KYSE150 cells (Figure 3C). Besides, knocking down PRMT5 could downregulate the level of p21, compared with the blank and control-siRNA group (Figure 3D).

The wound-healing and transwell assays were used to investigate the effect of PRMT5 in ESCC cell migration. It demonstrated that downregulation of PRMT5 inhibited the migration of ECA109 and KYSE150 cells (Figures 4A,B). Furthermore, invasion assay was performed to measure the effect of PRMT5 in ESCC cell invasion. As shown in Figure 4C, the number of invasion cells was reduced after transfection with siRNA-PRMT5 in both ECA109 and KYSE150 cells. To further determine the effect of PRMT5 in ESCC cell invasion, western blotting was adapted to determine the MMP-2 and MMP-9 levels. With downregulation of PRMT5, the levels of MMP-2 and MMP-9 were dramatically downregulated in ECA109 and KYSE150 cells, compared with the blank and control-siRNA group (Figure 4D). Overall, these findings indicated that PRMT5 promoted the cell migration and invasion in ESCC.

To further investigate the mechanism of PRMT5 in ESCC, the expressions of LKB1, AMPK, and mTOR were examined. It showed that the knocking down PRMT5 significantly promoted the expression of LKB1 and the p-AMPK expressions in both ECA109 and KYSE150 cells, whereas the p-mTOR expressions were weakened after transfection with PRMT5 siRNA, compared with the controls (Figure 5A). Furthermore, the result of immunofluorescence indicated that the expression of LKB1 could be promoted by downregulating PRMT5 (Figure 5B). It suggested that PRMT5 could inhibit LKB1/AMPK/mTOR signaling pathway.

To further explore the mechanism of LKB1 on ESCC, we performed an experiment by overexpression of LKB1 in ECA109 and KYSE150 cells. The transfection efficiency of LKB1 overexpression vector was identified by RT-PCR, and results have shown that the PRMT5 expressions were dramatically increased (Figures 6A,B). CCK-8 assay demonstrated that overexpression of LKB1 could suppress cell proliferation (Figure 6C). In addition, the result of flow cytometry analysis showed that overexpression of LKB1 could enhance the cell apoptosis rates in both ECA109 and KYSE150 cells, compared with the controls (Figure 6D). Furthermore, compared with the control, overexpression of LKB1 markedly upregulated the expression levels of Bax, caspase-3, and caspase-9, while Bcl-2 expression level was downregulated in ECA109 and KYSE150 cells (Figure 6E). Taken together, overexpression of LKB1 could promote the ESCC cell apoptosis.

Furthermore, the effect of LKB1 in ESCC cell migration and invasion was experimented on. And ECA109 and KYSE150 cells transfected with LKB1 overexpression vector could decrease the number of migrated cells and the invading cells (Figures 7A,B). Besides, with overexpression of LKB1, the MMP-2 and MMP-9 expressions were obviously downregulated in ECA109 and KYSE150 cells, compared with the blank and empty vector groups (Figure 7C). Last, we examined the expressions of key proteins from LKB1-related signaling. As a result, overexpression of LKB1 significantly upregulated the p-AMPK expression while downregulating the p-mTOR expressions in both ECA109 and KYSE150 cells, compared with the blank and empty vector groups (Figure 7D). Overall, it revealed that LKB1 could inhibit ESCC cell proliferation, migration, and invasion and induce cell apoptosis.

To further test and verify the effect of PRMT5 on the ESCC in vivo, KYSE150 cells transfected with blank, control siRNA, and PRMT5 siRNA were injected into the axilla of the male Balb/c nude mice. In the PRMT5 siRNA group, compared with the controls, tumor volume and tumor weight were significantly increased but have no significant difference in the weight of mice (Figures 8A–C). Furthermore, with the analysis of histological and IHC in tumor sections, the levels of TUNEL and Ki-67 were lower in the PRMT5 siRNA group than the control siRNA, which indicated that knocking down PRMT5 could inhibit the tumor growth and induce the tumor cell apoptosis in vivo (Figure 8D). Besides, it also showed that the expression of LKB1 was significantly upregulated along with the lower PRMT5 levels in the PRMT5 siRNA group than the blank and control siRNA groups (Figure 8D). Additionally, western blotting showed that knocking down PRMT5 could dramatically upregulate the expression of LKB1 and the p-AMPK in tumor sections, while downregulating the expression of p-mTOR (Figure 8E). These results suggested PRMT5 induced the tumor growth of KYSE150 cells in vivo via LKB1/AMPK/mTOR signaling pathway.

The lung metastasis model was performed to detect the effect of PRMT5 on ESCC metastasis in vivo. The lung metastatic foci and lung/total weight dramatically reduced in the PRMT5 siRNA group (Figures 9A–C). Furthermore, the expressions of PRMT5 and LKB1 in lung tissue dramatically downregulated in the PRMT5 siRNA group (Figure 9D). It showed that the levels of LKB1 and phosphorylation expressions of AMPK were obviously upregulated, while phosphorylation expressions of mTOR were obviously downregulated in the PRMT5 siRNA group (Figure 9E). Besides, knocking down PRMT5 obviously downregulated the MMP-2 and MMP-9 expression levels (Figure 9F). Taken together, these results provide direct evidence supporting our hypothesis that PRMT5 promotes lung metastasis in mice via the LKB1/AMPK/mTOR signaling pathway.