PKM2 inhibitor

PKM2‐regulated STAT3 promotes esophageal squamous cell carcinoma progression via TGF‐β1‐induced EMT

1 | INTRODUCTION

Esophageal cancer (EC) is the eighth most prevalent malignancy and sixth highest cause of cancer‐related genes that contribute to metastasis and clarification of the underlying molecular mechanisms are essential for the development of effective therapeutic strategies.
Pyruvate kinase isoenzyme type M2 (PKM2), the key rate‐limiting enzyme catalyzing the final step of glycolysis, plays a critical role in glucose metabolism of tumors.

Several recent reports suggest that PKM2 serves as a potential protein kinase that promotes cancer cell metas- tasis and proliferation.5-7 In addition, PKM2 has been shown to induce epithelial‐mesenchymal transition (EMT), the early event of metastasis.8,9 However, the specific roles and precise underlying mechanisms of PKM2 in transform- ing growth factor β1 (TGF‐β1)‐induced EMT and ESCC
progression are yet to be clarified.

PKM2 exerts tumor‐promoting effects through regula- tion of multiple signaling pathways. In an earlier investigation, STAT3/Snail1 signaling was suppressed and expression levels of EMT‐related molecules, such as E‐cadherin, N‐cadherin, and vimentin, were altered after
treatment of human hepatocellular carcinoma cells with PKM2 short hairpin RNA (shRNA).10 Gao et al11 suggested that PKM2 played a critical role in STAT3 phosphorylation at Tyr705, leading to cell proliferation. Another study by Yang et al12 showed that PKM2 promoted cell migration through increased STAT3 gene transcription and phosphorylates STAT3 at Ser727. In view of these previous findings, we hypothesized that PKM2 could contribute to TGF‐β1‐induced EMT and ESCC progression through effects on STAT3. During TGF‐β1‐induced EMT, the regulatory effects of PKM2 on STAT3 and exact phosphorylation state of STAT3 in ESCC are critical factors that need to be established. Data from this study showed associations of PKM2 overexpres- sion with early lymph node metastasis and poor prognosis. In addition, high PKM2 expression in tumor tissues frequently coincided with the high pSTAT3Tyr705 expression and low E‐cadherin expression.

Based on the collective findings, we propose that PKM2 acts as a tumor promoter in ESCC by inducing phosphorylation of STAT3 at Tyr705 leading to TGF‐β1‐induced EMT and enhanced proliferation, migration, and invasion.

2 | METHODS
2.1 | Cell culture

Human ESCC cell lines, KYSE150 and TE‐1, were obtained from the Chinese Academy of Sciences (Shanghai, China), EC9706 and Eca109 cells were from Wuhan University (Wuhan, China), and T4 was kindly provided by digestive physician Zhiqiang Zhang practicing in the Department of Gastroenterology, the First Affiliated Hospital of Xinjiang Medical University. All cells were maintained in Roswell Park Memorial Institute (RPMI) 1640 medium (Gibco, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS; Gibco) and cultured in a 5% CO2 humidified
incubator at 37°C. Cells were exposed to 10 ng/mL TGF‐ β1 (PeproTech, Rocky Hill, NJ) in serum‐free medium for 48 hours or left untreated.

2.2 | Clinical tissues

In total, 139 ESCC and paired normal adjacent tissues (NAT) were examined. One tissue microarray consisted of 85 paired cases of ESCC and matched NAT (catalog number: # HEso‐Squ180Sur‐01; Outdo Biotech, Shanghai, China), and 54 paired cases of ESCC and NAT, subjected to esophagectomy, obtained from the First Affiliated Hospital of Xinjiang Medical University. None of the patients had received preoperative anticancer therapy. All samples were histopathologically confirmed by two experienced patholo- gists blinded to this study. Staging and grading were determined according to the World Health Organization classification and grading system. Informed consent was obtained from each patient involved, and the study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Xinjiang Medical University.

2.3 | Immunohistochemistry analysis

Expression of PKM2 and E‐cadherin was detected via Immunohistochemistry (IHC). The tissue microarray was dewaxed and rehydrated, submerged into EDTA antigenic retrieval buffer, and treated with 3% hydrogen peroxide. Slides were incubated at 4°C overnight with primary polyclonal rabbit anti‐PKM2 (15822‐1‐AP; Proteintech Chicago, IL) and E‐cadherin (20874‐1‐AP, Proteintech) antibodies at 1:100 dilution, followed by treatment with biotinylated antirabbit secondary antibody (CST) for 60 minutes at 37°C. A semiquantitative scoring technique was used to determine the PKM2 and E‐cadherin expres- sions. The evaluation system included determination of the staining intensity and percentage of positive cells. Four grades each were assigned for staining intensity (0, none; 1, weak; 2, moderate; and 3, strong) and percentage of positive cells (0, <10%; 1, 10%‐25%; 2, 25%‐50%; and 3, >50%). ESCC patients were classified into two groups according to total
score (staining intensity plus positive cell score), specifically, “low expression” (total score, 0‐2) and “high expression” (total score, 3‐6) for analysis of prognosis between groups.

2.4 | Lentiviral transfection

Lentiviruses carrying green fluorescent protein (GFP) along with scrambled (Lv‐shRNA‐control and Lv‐ overexpress‐control), PKM2‐overexpressing construct (Lv‐overexpress‐PKM2), STAT3‐overexpressing construct (Lv‐overexpress‐STAT3), or PKM2 shRNA (Lv‐shRNA‐ PKM2‐1, Lv‐shRNA‐PKM2‐2, and Lv‐shRNA‐PKM2‐3) were purchased from Shanghai Gene Pharma Co Ltd (Shanghai, China). For infection, KYSE150 and Eca109 cells were incubated with lentiviruses (multiplicity of infection of 15) using polybrene (5 μg/mL) and enhanced infection solution according to the manufacturer’s proto- col. After 72 hours, all fluorescent cells were screened via flow cytometry and transfection efficiency evaluated via Western blots. The shRNA target sequences were as follows: shPKM2‐1 (AGGGAAAGAACATCAAGATTA), shPKM2‐2 (TGGATAACGCCTACATG GAA), shPKM2‐3 (AGCAAGAAGGGTGTGAACCTT), and shRNA control (TTCTCCGAACGT TCACGT).

2.5 | 3‐(4, 5‐Dimethyl ‐2‐thiazolyl)‐2, 5‐diphenyl‐2H‐tetrazolium bromide assay

ESCC cells were seeded in 96‐well plates at a density of 4× 103 cells/well. At the designated time points, cells were incubated with 100 μl sterile 3‐(4, 5‐dimethyl ‐2‐thiazolyl)‐2, 5‐diphenyl‐2H‐tetrazolium bromide (MTT; Sigma‐Aldrich, St. Louis, MO) for 4 hours at
37°C. The reaction was terminated by removal of culture medium followed by addition of 100 μl DMSO (Sigma‐ Aldrich) for 0.5 hours to dissolve the formazan product.Finally, absorbance values were measured at 490 nm.

2.6 | Wound healing assay

Cells subjected to various treatments were seeded into six‐well plates. After reaching 90% confluency, an artificial wound was scraped on cell monolayers using a sterile pipette tip. Cells were washed twice using phosphate‐buffered saline and fresh serum‐free medium added for another 48 hours. Cells that had migrated onto the scratches were captured and counted under an Olympus inverted fluorescence phase‐contrast micro- scope (Tokyo, Japan).

2.7 | Transwell assay

A Transwell system with 8 μm pores (Corning, New York) was precoated with 50 µL Matrigel (0.77 μg/μL; BD, Bedford, MA). Cells at a density of 1 × 105 per well were seeded in the upper chambers in 200 μlL medium without FBS. The lower chambers contained RPMI 1640 with 10% FBS. After 24 hours incubation, cells invading the lower surface of the filter membrane were fixed in methanol and stained with 0.1% crystal violet. Invading cells in five random fields were counted and images obtained under a light microscope.

2.8 | Western blots

ESCC cells were grouped and lysed with Radioimmuno- precipitation assay (RIPA) buffer containing protease and phosphatase inhibitors (Roche, Santa Clara, CA). Equivalent amounts of protein lysates were separated electro- phoretically using sodium dodecyl sulfate‐polyacrylamide gels and transferred onto polyvinylidene difluoride (PVDF) membranes. Next, membranes were incubated with 5% nonfat milk and the following primary anti- bodies overnight at 4°C: anti‐PKM2 (1:1000, 15822‐1‐AP; Proteintech), anti‐E‐cadherin (1:1000, 20874‐1‐AP; Proteintech), anti‐vimentin (1:1000, 10366‐1‐AP; Protein- tech), anti‐Snail (1:1000, 3879; Cell Signaling Technology), anti‐STAT3 (1:1000, 10253‐2‐AP; Protein- tech), anti‐pSTAT3 (Y705, 1:500, 9145 S; Cell Signaling Technology, Beverly, MA), and anti‐glyceraldehyde 3‐ phosphate dehydrogenase (GAPDH) (1:500, 20874‐1‐AP; Proteintech). After incubation with IRDye 800CD sec-
ondary antibody (1:15000; Odyssey, NE) at room tem- perature for 1 hour, immunoreactive bands were visualized using the CLx Infrared Imaging System (Odyssey), with GAPDH as the loading control.

2.9 | Statistical analysis

Statistical analyses were performed using SPSS 17.0 software (SPSS, Chicago, IL). Student t test or one‐way analysis of variance was used to evaluate the differences among groups, and chi‐square and Fisher’s exact tests applied to analyze correlations between PKM2/E‐cadher-
in expression and clinicopathological characteristics. The Kaplan‐Meier method and logrank test were used to plot survival curves and estimate survival rates. The multivariate Cox regression analysis was used to assess predictors related to survival. A two‐tailed P < 0.05 was regarded as significant in all tests. 3 | RESULTS 3.1 | High PKM2 expression was significantly correlated with metastasis and poor clinical prognosis in patients with ESCC To investigate the clinical significance of PKM2 in ESCC, we examined its expression patterns using IHC in 139 paraffin‐embedded human ESCC and paired NAT\ samples. Compared with the adjacent nontumor esophagus tissues in which PKM2 was undetectable or expressed at low levels, PKM2 was highly expressed in ESCC tissues (P = 0.000, Figure 1A, Table 1). Next, we analyzed the associations between PKM2 and clinicopathological parameters of ESCC. As shown in Table 1, upregulation of PKM2 was significantly correlated with lymph node metastasis (P = 0.023) while no significant correlation was observed between the PKM2 expression and age, sex, or degree of cell differentiation. Kaplan‐Meier survival analyses revealed a significantly shorter median survival time for patients with high PKM2 expression relative to patients with low PKM2 expression (P = 0.0036,Figure 1D). In multivariate Cox regression analyses, overall survival time was significantly dependent on PKM2 expression, N classification, and T classification, which may serve as independent prognostic factors for patients with ESCC (Table 2). FIGURE 1 Upregulation of PKM2 in ESCC tissues is markedly associated with poor prognosis. (A1) Normal epithelium showing weak or negative expression of PKM2. (A2) Representative positive staining of pyruvate kinase isoenzyme type M2 (PKM2) in esophageal squamous cell carcinoma (ESCC). (B1) Normal epithelium showing positive expression of E‐cadherin. (B2) Representative weak or negative staining of E‐cadherin in ESCC. (C1) Normal epithelium showing weak or negative expression of pSTAT3Tyr705. (C2) Representative positive staining of pSTAT3Tyr705 in esophageal squamous cell carcinoma (ESCC). The scale bar represents 10 μm. Magnification fold is 200×. (A3, B3, C3) Quantitative analyses of immunohistochemical data on PKM2, pSTAT3Tyr705, and E‐cadherin expressions per high‐power field (HPF) in ESCC and NAT (**P < 0.01 vs. paired NAT). (D) Kaplan‐Meier overall survival curves for all 139 patients with ESCC stratified by high and low expression of PKM2 (NAT, normal adjacent tissues). 3.2 | PKM2 promoted ESCC cell proliferation To establish the specific role of PKM2 in ESCC cell proliferation, we constructed PKM2 knockdown cells using three PKM2‐specific and control shRNAs. Western blot experiments showed that shRNA‐PKM2‐1 and shRNA‐PKM2‐2 successfully achieved significant depletion of PKM2 in the KYSE150 cell line (Figure 2B). In a subsequent MTT assay, KYSE150 cells treated with PKM2 shRNA exhibited markedly reduced absorbance values, compared with normal growth and control shRNA treatment groups at 48, 72, and 96 hours (P < 0.05), as shown in Figure 2C. We, in addition, induced overexpression of PKM2 in the Eca109 cell line (Figure 2D). MTT data revealed a significant increase in Eca109 cell numbers under conditions of PKM2 over- expression at 48, 72, and 96 hours (P < 0.05; Figure 2E), supporting the capability of PKM2 to promote tumor cell proliferation in ESCC. 3.3 | PKM2 promoted migration and invasion of ESCC cells through EMT induction We further examined the effects of PKM2 on ESCC migration and invasion via wound healing and Transwell assays. As shown in Figure 3A‐D, knock-down of PKM2 markedly suppressed migration and invasion of KYSE150 cells, whereas its overexpression had the opposite effects (Figure 3E‐H), clearly indicat- ing a significant role of PKM2 in these processes in ESCC cells. We further focused on the mechanisms underlying PKM2 activity in ESCC by examining the levels of Snail, E‐cadherin, and vimentin, the markers of EMT that are critical in cell invasion and migration. Notably, the PKM2 expression was negatively corre- lated with that of E‐cadherin in clinical tissues of ESCC (P = 0.032; Figure 1B, Table 3). Immunoblot experiments confirmed that the PKM2 knockdown triggered significant upregulation of E‐cadherin and concomi- tant downregulation of vimentin in KYSE150 cells (Figure 3I). Conversely, PKM2 overexpression was associated with increased protein levels of vimentin and Snail along with decreased E‐cadherin expression (Figure 3J). Based on the collective results, we suggest that PKM2 contributes to tumor cell migration and invasion through promoting EMT. 3.4 | Phosphorylation of STAT3 by PKM2 contributed to EMT and affected cell functions To further establish whether STAT3 was involved in PKM2‐mediated EMT, we analyzed the associations between PKM2 and pSTAT3Tyr705 expression in clinical tissues of ESCC. As shown in Figure 1C and Table 3, high PKM2 expression in tumor tissues frequently coincided with high pSTAT3Tyr705 expression. Furthermore, ESCC cells were initially transfected with either shRNA‐PKM2 or overexpress‐PKM2 con- struct. As shown in Figure 4A‐D, the pSTAT3Tyr705/ STAT3 ratio was reduced in PKM2‐depleted KYSE150 cells and increased in PKM2‐overexpressing Eca109 cells, whereas no significant changes in total STAT3 expression were observed. STAT3 was additionally overexpressed via transfection with Lv‐overexpress‐ STAT3 in PKM2‐depleted KYSE150 cells. Overexpres- sion of STAT3 suppressed the effects of PKM2 knockdown on pSTAT3Tyr705, Snail, vimentin, and E‐cadherin expression (Figure 4E). Moreover, altered proliferation, migration, and invasion rates caused by PKM2 knockdown were clearly prevented in STAT3‐ overexpressing cells (Figure 4F‐J), supporting our hypothesis that phosphorylation of STAT3 plays a critical role in PKM2‐induced EMT and associated cellular functions. FIGURE 2 PKM2 promotes ESCC cell proliferation. A, Western blot of PKM2 levels in five ESCC cell lines (Eca109, KYSE150, EC9706, T‐4, and TE‐1). B, Western blot showing remarkable silencing efficiency in KYSE150 cell lines infected with shPKM2‐1 and shPKM2‐2. C, MTT assay showing that downregulation of PKM2 significantly reduces the growth of KYSE150 cells (**P < 0.01 vs. normal growth and shRNA control groups). D, Western blot of Eca109 cells transfected with PKM2‐vector and PKM2. E, MTT data showing that overexpression of PKM2 significantly enhances growth rates of the indicated cells (**P < 0.01 vs normal growth and overexpression groups). Abbreviations are defined in the legend to Figure 1. ESCC, esophageal squamous cell carcinoma; MTT, 3‐(4, 5‐dimethyl ‐2‐thiazolyl)‐2,5‐diphenyl‐2H‐tetrazolium bromide; PKM2, pyruvate kinase isoenzyme type M2; shRNA, short hairpin RNA. 3.5 | PKM2 knockdown inhibits TGF‐β1‐induced EMT, migration and invasion via STAT3 phosphorylation in ESCC cells To ascertain whether PKM2 affected TGF‐β1‐induced EMT, migration and invasion events via a pathway involving STAT3 phosphorylation, expression of PKM2, pSTAT3Tyr705 and EMT‐related markers in KYSE150 cells was examined via Western blot. We observed significant suppression of E‐cadherin and increase in PKM2, pSTAT3Tyr705, Snail, and vimentin expression levels in the presence of TGF‐β1 (Figure 5A). Interestingly, the effects of TGF‐β1 on expression patterns of these proteins were abolished upon depletion of PKM2 (Figure 5B). Moreover, altered migration and invasion rates in the presence of TGF‐β1 were clearly prevented in PKM2 knockdown cells (Figure 5C‐F). Accordingly, we propose that PKM2 mediates TGF‐β1‐induced EMT, migration and invasion through phosphorylation of STAT3 at Tyr705 in ESCC cells. 4 | DISCUSSION In this study, we showed that PKM2 was remarkably upregulated in ESCC compared with paired NAT samples. The high level of PKM2 was positively correlated with lymph node metastasis and poor overall prognosis. Furthermore, PKM2 expression was negatively correlated with that of E‐cadherin protein in ESCC. PKM2 promoted proliferation, migration, and invasion and was involved in TGF‐β1‐induced EMT through phosphorylation of STAT3 in ESCC cells. While extensive research has shown that PKM2 promotes gene transcription and cell proliferation in different cancer types, including human ovarian cancer, prostate cancer, and colorectal cancer,5,13,14 limited studies are available regarding its role in ESCC. In a previous study by our group,15 five upregulated proteins were identified via MALDI‐TOF‐MS in ESCC tissues. Among these, PKM2 was of particular interest as a potential critical molecule affecting ESCC progression. It is widely accepted that EMT plays a significant role in cancer progression and metastasis.16 Recent studies suggest that PKM2 promotes EMT and migration of hepatocellular carcinoma cell upon EGFR stimulation (Wei et al, 2016)17 further demonstrated that PKM2 significantly promotes leptin‐induced EMT and motility in breast carcinoma. On the basis of previous findings by our group,4 PKM2 was upregulated in ESCC relative to NAT, and significantly correlated with poor overall prognosis, in the present study. Here, we focused on the specific effects of PKM2 on EMT and tumor progression and the associated mechan- isms of action in ESCC. By expanding the quantity of samples, further results were obtained showing that PKM2‐positive staining is positively correlated with lymph node metastasis. High expression of PKM2 was negatively correlated with E‐cadherin expression and was positively correlated with pSTAT3Tyr705 expression, to a significant extent, supporting the theory that PKM2 can potentially serve as a predictor of prognosis and is associated with EMT and tumor progression in ESCC. On the basis of clinical data, it is reasonable to deduce that PKM2 regulates EMT as well as proliferation and motility of ESCC cells. Previous experiments by our group showed that siRNA‐induced PKM2 knockdown leads to inhibition of proliferation and motility in Eca109 cells.4 In the present study, KYSE150 cells were transfected with optimized shRNA, which was con- structed with high efficiency, exerting sustainable effects and contributing to less off‐target effects.18 Furthermore, we used gain‐ and loss‐of‐function methods to over- express and eliminate PKM2 expression in the ESCC cell lines, Eca109 and KYSE150, respectively. Consistently, stable knockdown of PKM2 in KYSE150 cells inhibited, while the elevation of PKM2 levels in Eca109 cells promoted ESCC cell proliferation, migration, and inva- sion. PKM2 was proposed to promote the proliferation, migration, and invasion capacities of ESCC cells via induction of EMT, as overexpression of PKM2 signifi- cantly suppressed E‐cadherin expression and increased Snail and vimentin expression, whereas its silencing exerted the opposite effects. These results suggest that, in addition to its vital role in cell proliferation, PKM2 promotes EMT/cell motility, and consequently, cancer metastasis. FIGURE 3 PKM2 promotes the migration and invasion capacities of ESCC cells via induction of epithelial‐mesenchymal transition (EMT). A, Variations in migratory activities of KYSE150 cells analyzed using the wound healing assay after transfection with shRNA sequences against PKM2 for 48 hours (magnification fold is 100×). B, Percentage of migration derived from A (**P < 0.01). C, Variations in invasive activity of KYSE150 cells evaluated with the Transwell method after transfection with shRNA sequences against PKM2 for 24 hours (magnification fold is 200×). D, Quantitative assay of Transwell experiments in C (**P < 0.01). E, Variations in the migratory activity of Eca109 cells analyzed using the wound healing assay after overexpression of PKM2 (magnification fold is 100 × ). F, Percentage of migration activity derived from E (**P < 0.01). G, Variations in invasive activity of Eca109 cells evaluated with the Transwell method after overexpression of PKM2 (magnification fold is 200×). H, Quantitative assay of the Transwell experiment in G (**P < 0.01). I, Protein levels of Snail, E‐cadherin, and vimentin in KYSE150 cells based on successful knockdown of PKM2; J, Protein levels of Snail, E‐cadherin, and vimentin in Eca109 cells after overexpression of PKM2 using GAPDH as a loading control. GAPDH, glyceraldehyde 3‐phosphate dehydrogenase; other abbreviations are defined in the legend to Figure 1. ESCC, esophageal squamous cell carcinoma; PKM2, pyruvate kinase isoenzyme type M2. FIGURE 4 Phosphorylation of STAT3 by PKM2 contributes to epithelial‐mesenchymal transition (EMT) and affects cell functions. A, After successful knockdown of PKM2, protein expression of PKM2, STAT3, and pSTAT3Tyr705 in KYSE150 cells was detected via Western blot, with GAPDH as a loading control. B, Quantitative assay of the pSTAT3Tyr705/STAT3 ratio in A (P < 0.01). C, Western blot of PKM2, STAT3, and pSTAT3Tyr705 protein levels in Eca109 cells under conditions of PKM2 overexpression, using GAPDH as a loading control. D, Quantitative assay of the pSTAT3Tyr705/STAT3 ratio in C (P < 0.01). E, Western blot of pSTAT3Tyr705, Snail, E‐cadherin, and vimentin protein levels in KYSE150‐shRNA‐PKM2 cells with or without overexpression of STAT3. F, MTT assay showing variations in the proliferation of KYSE150‐shRNA‐PKM2 cells overexpressing STAT3 (**P < 0.01 vs control group; ##P < 0.01 vs shRNA‐PKM2). G, Wound healing assay showing the migration dynamics of KYSE150‐shRNA‐PKM2 cells with or without STAT3 overexpression (magnification: 100×). H, Percentage of migrated cells derived from G (P < 0.01). I, Transwell method showing variations in invasive activity of KYSE150‐shRNA‐PKM2 cells overexpressing STAT3 (magnification: 200 × ). J, Quantitative assay of Transwell experiments in I (P < 0.01). GAPDH, glyceraldehyde 3‐phosphate dehydrogenase; other abbreviations are defined in the legend to Figure 1. ESCC, esophageal squamous cell carcinoma; MTT, 3‐(4, 5‐dimethyl ‐2‐thiazolyl)‐2, 5‐diphenyl‐2H‐ tetrazolium bromide; PKM2, pyruvate kinase isoenzyme type M2; shRNA, short hairpin RNA. FIGURE 5 PKM2 regulates TGF‐β1‐induced epithelial‐mesenchymal transition (EMT) via STAT3 phosphorylation in ESCC cells. A, Western blot of PKM2, STAT3, pSTAT3Tyr705, Snail, E‐cadherin, and vimentin expression in KYSE150 cells treated with TGF‐β1, using GAPDH as a control; B, Western blot of PKM2, pSTAT3Tyr705, Snail, E‐cadherin, and vimentin expression in KYSE150‐shRNA‐PKM2 cells treated with TGF‐β1, using GAPDH as a control. C, Wound healing assay showing migration dynamics in KYSE150‐shRNA‐PKM2 cells treated with or without TGF‐β1 (magnification: 100×). D, Percentage of migrated cells derived from C (P < 0.01). E, Transwell method showing variations in invasive activity between KYSE150‐shRNA‐PKM2 cells treated with or without TGF‐β1 (magnification: 200×). F, Quantitative assay of Transwell experiments in E (P < 0.01). Cells were cultured with 10 ng/mL TGF‐β1 for 48 hours or left untreated (control). GAPDH, glyceraldehyde 3‐phosphate dehydrogenase; other abbreviations are defined in the legend to Figure 1. ESCC, esophageal squamous cell carcinoma; TGF‐β1, transforming growth factor β1; PKM2, pyruvate kinase isoenzyme type M2; shRNA, short hairpin RNA. The precise mechanisms underlying PKM2 activity in EMT and tumor progression in ESCC are under investigation. STAT3 has been identified as one of the downstream targets of PKM2 in different cancer types. PTBP1 is reported to promote anaplastic large cell lymphoma oncogenesis via the PKM2/STAT3 pathway.19 PKM2 regulates transcription of MEK5 via phosphorylat- ing STAT3 at Tyr705, whereas in colon cancer, STAT3 is activated by PKM2 through promoting transcription and phosphorylation at Ser727.12 However, in ESCC, it is unclear whether PKM2 affects STAT3 gene transcription or phosphorylation, and the potential phosphorylation site of STAT3 regulated by PKM2 has not been established to date. In the present study, altered expression of PKM2 affected the pSTAT3Tyr705/STAT3 ratio but induced no significant changes in total STAT3 expression. Moreover, overexpression of STAT3 reversed the effects of shRNA‐PKM2 on proliferation, migration, and invasion as well as the expression of EMT‐associated marker proteins and pSTAT3Tyr705 levels in ESCC cells. Our findings indicate that PKM2 regulates proliferation, migration, invasion, and EMT via phosphorylation of STAT3 at Tyr705 in ESCC. Previous experiments have shown that EMT is triggered by TGF‐β1.20 Hamabe et al21 demonstrated that PKM2 is bound to TGIF2 during TGF‐β1‐mediated induction of EMT. Cheng & Hao22 further reported the elevated expression of PKM2/STAT3/Snail in TGF‐β1‐ induced EMT. In the present study, under conditions of TGF‐β1 stimulation, PKM2 promoted phosphorylation of STAT3 at Tyr705 to increase transcription of Snail that acted as a repressor of E‐cadherin, leading to EMT and metastasis of ESCC. Furthermore, this was abolished upon PKM2 knockdown. In conclusion, PKM2 may effectively serve as an independent prognostic factor for ESCC. The PKM2/ STAT3 pathway appears to contribute to the proliferation and motility of ESCC via TGF‐β1‐induced EMT. Our collective data provide insights into the molecular mechanisms underlying EMT regulation and further support PKM2 inhibitor the targeted inhibition of PKM2 as a future therapeutic strategy for ESCC.