Epigenetically upregulated oncoprotein PLCE1 drives esophageal carcinoma angiogenesis and proliferation via activating the PI-PLCε-NF-κB signaling pathway and VEGF-C/ Bcl-2 expression

Sources of antibodies against the following proteins were as follows: PLCE1 (Santa-sc-28,404, 1:200 for WB), PLCE1 (Sigma-Aldrich - HPA015598, 1:50 for IHC and IF), p65 (Abcam-Ab32536, 1:3200 for IHC and 1:10,000 for WB), Phospho-p65 (Abcam-Ab86299, 1:5000 for WB), IκBa (Abcam-Ab32518, 1:200 for IHC and 1:5000 for WB), Phospho-IκBa(Ser32) (CellSignalingTechnology-2859, 1:1000 for WB), IKKα (Abcam-Ab32041, 1:100 for IHC and 1:5000 for WB), phospho-IKKα/β(Ser176/180) (CellSignalingTechnology-2697, 1:1000 for WB), PKCα(Santa-sc-208, 1:1000 for WB), Bax (Abcam-Ab32503, 1:1000 for WB), caspase 3 (Proteintech-19,677, 1:1000 for WB), caspase 7 (Bioss-BAO088–1, 1:1000 for WB), cleaved PARP (Abcam-Ab32561, 1:10,000 for WB), vimentin (Proteintech-60,330, 1:1000 for WB), E-cadherin (Santa-sc-71,009, 1:200 for WB), VEGF-C (Proteintech-22,601-1-AP, 1:600 for WB), Bcl-2 (Beyotime- AB112, 1:1000 for WB), CD34 (ZSGB- ZM0046, 1:200 for IHC), Ki-67 (ZSGB- ZA0502, 1:200 for IHC and IF), and β-actin (Solarbio- RG000120, 1:1000 for WB). U73122 (112648–68-7) was purchased from MedChem Express. Bay11–7082 was purchased from Merck.

A total of 368 ESCC and 215 non-cancerous tissue samples were collected from patients and confirmed by pathological diagnosis. Aside from diagnostic biopsies, chemotherapy, or radiation therapy, no patient underwent previous surgery at the First University Hospital, Shihezi University School of Medicine, the People’s Hospital of Xinjiang Uyghur Autonomous Region, and the Xinjiang Yili Prefecture Friendship Hospital between 1984 and 2014. Tumor-node-metastasis stages of pathological diagnoses for all of the cases were evaluated based on the Cancer Stage Manual (7th Edition; issued in 2009 by the American Joint Committee on Cancer/Union for International Cancer Control). For the 368 ESCC specimens, 215 adjacent non-malignant tissues were controls. All of the patients were enrolled with written informed consent, and the study was approved by the Institutional Ethical Review Board at the First University Hospital, the Shihezi University School of Medicine, and Shihezi University.

In this assay, 236 samples, namely, 132 ESCC and 104 normal tissue samples, were used for PLCE1 methylation detection. The age was 60.36 ± 8.27 (mean ± SD) years for the cancer samples and 60.94 ± 8.61 years for the normal sample (P = 0.60). A total of 85 (64.4%) males and 47 (35.6%) females were selected for the case group and 68 (65.4%) males and 36 (34.6%) females were selected for the control group (P = 0.87). DNA was isolated from tissues using DNA Extraction Kit (Qiagen Inc., 56404). NanoDrop spectrophotometer (NanoDrop Technologies Inc.) and gel electrophoresis were used to ensure DNA purity and quality. Purified genomic bisulfite-converted DNA samples were successfully tested by PCR with human PLCE1 primers 5′-aggaagagagGTTGGGTATATTGATGGGGTTTAAT-3′ (forward) and 5′-cagtaatac-gactcactatagggagaaggctACCCCTAAAAACCATCCTTTCTAAC-3′ (reverse). Bisulfite was used to treat genomic DNA by using the EZ DNA Methylation Kit™ in accordance with the manufacturer’s protocol (Zymo Research, D5008). NCBI and CpG island prediction (http://​www.​ebi.​ac.​uk/​Tools/​seqstats/​emboss_​cpgplot/) were used to identify the sequence of the CpG islands. The analyzed region and CpG sites of the PLCE1 promoter are shown in Fig. 1h. Primer sets for the methylation analysis of the PLCE1 promoter were designed using EpiDesigner (http://​epidesigner.​com; Additional file 1: Table S1). Mass spectra were collected by MassARRAY Compact MALDI-TOF MS, and the methylation proportions of individual units were generated by EpiTyper 1.0.5 (SEQUENOM). Methylation level was expressed as the percentage of methylated cytosines over the total number of methylated and unmethylated cytosines. The methylation level of each sample was evaluated according to the average methylation values of all CpGs uinits within PLCE1 pomoter. We used the average methylation values 9.6% as a cutoff value to divide all ESCC samples into hypermethylation group and hypomethylation group.

ESCC cell lines (EC9706 and Eca109) were purchased from the Institute of Biochemistry and Cell Biology of the Chinese Academy of Sciences (Shanghai, China). Cell lines were cultured in Dulbecco’s Modified Eagle’s Medium (Gibco, C11995500) supplemented with 10% fetal bovine serum (Tian Hang Biotechnologies, 0005), 100 U/ml penicillin, and 100 μg/ml streptomycin. HUVECs were isolated from the umbilical cord (from the First University Hospital, Shihezi University School of Medicine; obstetrics, healthy, maternity, and postpartum) and were cultured in Endothelial Cell Medium (Sciencell, 1001). All cell lines were cultured at 37 °C with 5% CO₂.

LV-R-PLCE1-RNAi (sh-PLCE1) and relative negative scramble control CON053 (sh-sc) lentiviral plasmids were purchased from GeneChem. In most experiments, cells were treated with sh-PLCE1 at MOI of 15. Cell transfections were performed with polybrene (GeneChem, REVG0001) and enhanced infection solution (ENi.S.; GeneChem, REVG0002), HiPerFect transfection reagent (Qiagen, 301,705), or Lipofectamine 2000 reagent (Invitrogen, 11,668–027) by following manufacturers’ instructions. U73122 was also used to inhibit PLCE1, and our results indicated that U73122 exerted time-dependent and dose-dependent inhibition effects on PLCE1 in human esophageal cancer cells. Finally, we used 10 μM for 48 h in follow-up experiments.

Paraffin-embedded materials were sampled from 368 formalin-fixed esophageal cancer tissues and 215 normal tissues. Samples with 0.6 mm-diameter tissue cores were obtained with a tissue arrayer (ALPHELYS). Tumor samples were fixed with 10% formalin in phosphate-buffered saline (PBS). Paraffin-embedded 4 μm sections were baked at 65 °C for 60 min before rehydration with graded alcohols. Each 4 μm tissue section was deparaffinized and rehydrated. Sections were autoclaved in ethylenediaminetetraacetic acid (EDTA) buffer (pH 9.0) at 130 °C for 10 min in a microwave oven, cooled to 30 °C for 40 min, and incubated with fresh 3% H₂O₂ in methanol for 10 min at room temperature. Tissue sections were incubated at 4 °C overnight with antibodies. Negative controls were prepared by replacing primary antibodies with PBS. Tissues were washed thrice in PBS for 5 min before incubation with respective secondary antibody for 30 min at 37 °C. Subsequently, 3,3-diaminobenzidine was used for visualization. Tissue sections were counterstained with hematoxylin. The location of CD34 staining was vascular endothelial cells. Thus, the IHC result of CD34 staining is represented by the MVD. The average of the number of positive vascular endothelial cells was calculated as described previously [25].

Cells on glass coverslips (BD) were fixed with 2% paraformaldehyde and permeabilized with 0.2% Triton X-100 in PBS. Samples were then blocked in 5% goat serum in the presence of 0.1% Triton X-100 and stained with the appropriate fluorescence-coupled primary and secondary antibodies.

Homogenized tissues or cells were lysed at 4 °C in radioimmunoprecipitation assay buffer (Solarbio, R0010) mixed with protease and phosphatase inhibitors. Protein lysates were resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE) and transferred to Immun-Blot PVDF membranes (Solarbio). Equal amounts of protein were boiled, resolved by SDS-PAGE, and transferred to polyvinylidene difluoride membranes. Membranes were incubated in blocking buffer (5% milk, 0.1% Tween20 in Tris-buffered saline) for 1 h and probed overnight with primary antibody at 4 °C. Blots were rinsed thrice (0.1% Tween20 in Tris-buffered saline, 5 min each), followed by incubation with peroxidase-conjugated secondary antibody (Solarbio) for 2 h at room temperature. Proteins were detected using Pierce ECL Plus Western blotting substrate kit (ThermoFisher Scientific).

Cell lysates were prepared by incubating cells in NETN buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.2% Nonidet P-40, 2 mM EDTA) in the presence of Protease Inhibitor Cocktails (Roche) for 20 min at 4 °C. This step was followed by centrifugation at 14,000×g for 15 min at 4 °C. For IP, B500 mg of protein was incubated with control or specific antibodies (1–2 mg) for 12 h at 4 °C with constant rotation. A total of 50 ml of 50% protein G magnetic beads (Invitrogen) were then added, and incubation was continued for an additional 2 h. Beads were then washed five times using lysis buffer. Between washes, beads were collected by magnetic stand (Invitrogen) at 4 °C. Precipitated proteins were eluted from beads by re-suspending beads in 2 SDS-PAGE loading buffer and boiling for 5 min. Boiled immune complexes were subjected to SDS-PAGE, followed by immunoblot with the appropriate antibodies.

For ChIP-qPCR assays, cells were treated for 10 min with 5 mM dimethyl 3,30-dithiobispropionimidate-HCl (Pierce) in PBS at room temperature, rinsed with 100 mM Tris–HCl, 150 mM NaCl (pH 8.0), and crosslinked with 1% formaldehyde in PBS at 37 °C for 10 min. Total cell lysates were sonicated to generate 200–500 bp DNA fragments. All resulting precipitated DNA samples were quantified by qPCR. Data are expressed as percentage of input DNA.

Cells (3 × 10⁴) were seeded in triplicates in 48-well plates and allowed to settle for 24 h. A total of 0.1 μg of pNF-κBluc plasmid; or control-luciferase plasmid plus 1 ng of pRL-TK Renilla plasmid (Promega) were transfected into ESCC cells using Lipofectamine 2000 reagent (Invitrogen). After 48 h of transfection, luciferase and Renilla activities were measured using the Dual Luciferase Reporter Assay Kit (Promega).

For MTT assay, cells were seeded into 96-well plates at 4000 cells per well. After transfection and cell culture for different time periods, 20 μl of MTT (5 mg/ml) (Solarbio, M8180) was added to each well, and cells were incubated for an additional 4 h in the dark. Finally, 125 μl of DMSO (Solarbio, D8370) was added to each well, and optical density was measured at a wavelength of 490 nm.

For colony formation assay, after transfection, 1000 viable cells were plated in six-well plates in triplicate and maintained in complete medium for 15 days. Foci were fixed with 4% polyoxymethylene (Solarbio, P8430) and stained with 0.1% crystal violet (Solarbio, G1061).

To analyze apoptosis rate, 5 × 10⁴ cells were cultured in 24-well plates. After transfection for 48 h, cells were stained with AV/propidium iodide (PI) staining kits (Lianke, AP101–100-kit) and analyzed with a flow cytometer.

HUVECs (2 × 10⁴) were plated in the upper chamber of BioCoatTM Invasion Chambers (BD, Bedford, MA) and incubated with conditional medium collected from ESCC cells infected with shR-PLCE1, U73122, and Bay11–7082 at 37 °C for 22 h, followed by removal of cells inside the upper chamber with cotton swabs. Migratory cells on the lower membrane surface were fixed in 1% paraformaldehyde, stained with hematoxylin, and counted (ten random × 200 fields per well). Cell counts are expressed as the mean number of cells per field of view. Three independent experiments were performed, and data are presented as mean ± standard deviation (SD).

JC-1 is a red fluorescent dye which forms aggregates in normal mitochondria. However, JC-1 monomers emit green fluorescence in cells with mitochondrial dysfunction. Cells were seeded in 24-well plates and stimulated as indicated. After treatment, medium was discarded and replaced with 500 μL fresh medium containing 5 μg/mL JC-1. After incubating for 1 h, cells were washed twice with PBS and photographed at 100× magnification under a fluorescent microscope.

HUVEC tube formation assay was performed. Briefly, 200 μl of pre-cooled Matrigel (Collaborative Biomedical Products) was transferred to each well of a 24-well plate and polymerized for 30 min at 37 °C. HUVECs (2 × 10⁴) in 200 μl of conditioned medium were added to each well and incubated at 37 °C and 5% CO₂ for 20 h. Capillary tube structure was photographed under a × 100 bright-field microscope and quantified by measuring the total length of completed tubes. Each condition was assessed in triplicate.

Animal studies were conducted according to guidelines approved by the University of Shihezi Institutional Animal Care and Use Committee. Athymic nude mice were maintained in specific pathogen-free conditions. Animals were randomly distributed into groups. To induce tumor formation, athymic nude mice were subcutaneously infected with 2 × 10⁶ Eca109 cells into the left axillary area. Eca109 cells were steadily infected with pSIH1-H1-copGFP/shR-PLCE1 or pSIH1-H1-copGFP lentivirus after selection of puromycin. After treatment via injection, tumor length and width measurements were obtained with calipers. Tumor volumes were calculated with the following formula: tumor volume (mm³) = (major axial diameter × minor axial diameter 2 × 0.5) thrice weekly. On day 35, tumors were detected by an IVIS imaging system (Caliper Life Sciences, USA), and animals were sacrificed and photographed. Tumors were excised, weighed, photographed, and snap-frozen in liquid nitrogen or formalin-fixed and paraffin-embedded. Paraffin embedded tumor tissues underwent routine histological processing with hematoxylin and eosin (H&E) stain. Cell proliferation in tumors was detected by staining histological sections with a monoclonal antibody against Ki-67 (Zhongshan Biotechnology, ZM-0166); apoptosis was assessed by TUNEL assay (Beyotime, C1086).

IP3 expression was analyzed in cell-free extracts using enzyme-linked immunosorbent assay kits (Elabscience). DAG expression was assessed using enzyme-linked immunosorbent assay kits (Jianglai Shanghai, China). We performed the assay according to the manufacturer’s instructions. Cells were collected using trypsin and resuspended with PBS after treatment with shR-PLCE1, BIM (40 μM) and TPA (150 ng/ml) for 72 h, 24 h, and 36 h, respectively. Then, the cells were subjected to three rounds of repeated freezing and thawing and further broken by ultrasonication.

After the same treatment with ELISA, cells were incubated with 5 μM calcium fluorescence probe Fluo-3 AM (Solarbio) as above using D-Hanks solution. After washing, the fluorescence intensity of calcium was visualized on the scanning confocal microscope under a 63 × oil immersion objective at 490 nm wave. Images were subjected to semi-quantitative analysis using Image J software.

The relationship between PLCE1 levels and clinical features of ESCC was further examined in a test cohort with 150 paraffin-embedded, archived ESCC tissues. The end date of follow-up was the date of final contact or the date of death through July 2017. The mean follow-up period is 34.2 months, ranging from 0.5 months to 132 months. Correlation analysis of the cohort indicated that the positive expression of PLCE1 in ESCC was significantly associated with more aggressive tumor phenotypes. As shown in Fig. 1a, PLCE1 expression increased with advanced clinical stage in ESCC, whereas PLCE1 was only found at low levels in normal esophageal tissues. Kaplan-Meier analysis established that in the cohort, 27 months was the median disease-specific survival time for patients with high expressing PLCE1 compared with the 34 months for patients with low expressing PLCE1 (P = 0.002, log-rank test; Fig. 1b). As shown in Fig. 1c, the expression level of PLCE1 in GSE9982 esophageal cancer cell lines suggests high expression of PLCE1 in cancer cell lines (P = 0.0009; Fig. 1c). Analysis of 53 pairs of ESCC tissue specimens in the GSE23400 database revealed a high expression of PLCE1 in patients with ESCC in the cohort (P = 0.004; Fig. 1d). Altogether, these investigations indicate that PLCE1 may be a predictive biomarker for disease outcome in ESCC. Moreover, the GEPIA database in Fig. 1e show that compared to normal, PLCE1 significantly increased expression levels in ESCA, LAML, LIHC patients. Data from the Oncomine database (https://​www.​oncomine.​com/) also confirmed that PLCE1 expression was increased in esophageal carcinoma, hepatocellular carcinoma, glioblastoma and renal cell carcinoma in the four published datasets, GSE92396, GSE98383, GSE90886, and GSE6357 (Fig. 1f). These findings revealed an oncogenic role of PLCE1 in different cancers including ESCC.

Importantly, CpGs residing within PLCE1 were strongly hypomethylated in tumor tissues compared to adjacent normal tissues (Fig. 1g, h). In addition to CpG_3 an d CpG_4, the mean methylation levels at CpG_2, CpG_5.6, CpG_7.8, and CpG_9.10 were significantly lower in patients with ESCC than those in the controls (Fig. 1i and Additional file 1: Table S2). Notably, a significant inverse correlation was observed for CpG_2 and CpG_5.6 methylation and PLCE1 protein expression (Additional file 1: Table S3). A negative relationship between global PLCE1 methylation and protein expression was also observed (Additional file 1: Table S4). In TCGA Illumina 450 k infinium methylation beadchip, we found that the methylation status of 6 CpG site (cg03088791; cg06539629; cg24553184; cg13506485; cg14741153; cg16986921) was all negatively correlated with PLCE1 expression (Fig. 1j). Furthermore, Kaplan-Meier analysis showed that esophageal carcinoma patients with CpG_5.6 hypomethylation have significantly shorter overall survival than those with relative hypermethylation (log rank P = 0.046) (Fig. 1k).

Angiogenesis is the basic and important stage in the process of ESCC tumorigenesis. Microvessel density (MVD) is the most recognized indicator to evaluate angiogenesis of solid tumors. CD34 is used to label MVD. As PLCE1 has been investigated to be an important predictive marker for ESCC, researchers have worked to clarify the relationship between tumor neovascularization and PLCE1 in ESCC specimens. The PLCE1 and Ki-67-positive expression and MVD, which was labeled by CD34, were significantly higher in the ESCC tissues than in normal esophageal tissues (Fig. 2a, b). Figure 2c, d show the results of the correlation analysis between PLCE1 with MVD and Ki-67. PLCE1 and MVD are positively correlated, as are PLCE1 and Ki-67. In the consecutive clinical specimens of ESCC, accompanying by the increased expression status of PLCE1, Ki-67, and CD34 staining, the intensity score of their staining was increased (Fig. 2e). Moreover, the MVD numbers and the score of Ki-67 expression were much higher in PLCE1(+) ESCC tissues than in PLCE1(−) ESCC tissues (Fig. 2f). Altogether, these results suggest that PLCE1 may paticipate in angiogenesis and promote proliferation in ESCC. Survival analysis showed that high MVD in ESCC patients meant significantly shorter survival compared to low MVD (Fig. 2g). Survival was reduced when Ki-67 was positive compared to those with a negative expression of Ki-67 (Fig. 2h). Thus, PLCE1 may contribute to an oncogenic role in angiogenesis and proliferation of ESCC.

Inhibited endogenous PLCE1 expression significantly reduced cell proliferation compared with the controls and significantly increased the apoptosis-related moleculars expression (Fig. 3a-c). Colony formation assays revealed that the transfection of Eca109 and EC9706 cells with shR-PLCE1 or U73122 reduced growth compared with controls or blanks (Additional file 2: Figure S1a, b). Annexin V (AV)– FITC and TUNEL assays demonstrated that silencing PLCE1 promoted apoptosis in ESCC cells (Fig. 3d, Additional file 2: Figure S1c-e). In Eca109 and EC9706 cells, after treatment with shR-PLCE1 or U73122, red fluorescence emitted by JC-1 was reduced, but green fluorescence was elevated, indicating PLCE1 induced mitochondrial dysfunction in ESCC cells (Additional file 2: Figure S1f). Western blot confirmed that the expression of bax, caspase 3, caspase 7, and cleaved poly (ADP-ribose) polymerase (PARP), which positively regulate apoptosis, were upregulated in the PLCE1-silenced cells. Bcl-2 was downregulated and E-cadherin was upregulated in the PLCE1-silenced cells, but vimentin was downregulated (Fig. 3e). Thus, PLCE1 is key to tumorigenicity of ESCC cells in vitro. Real-time PCR Chip analysis demonstrated a positive relationship between PLCE1 expression and angiogenesis-related molecules (Fig. 3f). We also observed that silencing PLCE1 strongly inhibits Eca109 and EC9706 ESCC cell induction of human umbilical vein endothelial cells (HUVECs) proliferation (Additional file 3: Figure S2a-c). Inhibition of PLCE1 significantly reduced the ability of ESCC cells to induce tubule formation and migration by HUVECs in vitro (Fig. 3g, Additional file 3: Figure S2d). Multiple studies have reported that expression level of VEGF-C, a vital pro-angiogenic factor, correlates strongly with ESCC progression and acts as an independent prognostic factor for ESCC. Additional file 3: Figure S2e shows VEGF-C mRNA expression was downregulated in PLCE1-silenced cells, and Western blot data agreed with this (Fig. 3h). Thus, PLCE1 contributes to promoting angiogenesis in ESCC cells.

Moreover, the tumors formed by PLCE1-silenced cells were smaller compared with tumors formed by controls in Fig. 3i–k. IHC staining indicated that PLCE1-silenced tumors had low Ki-67, CD34, and Bcl-2 expression and decreased MVD (Fig. 3l, m). IF staining showed that PLCE1 knockdown increased TUNEL-positive apoptotic cells compared with the controls (Additional file 3: Figure S2f, g). Thus, PLCE1 contributes to progression of ESCC in vivo.

As growing body of evidence demonstrates that the NF-κB signaling pathway plays a central role in both angiogenesis and resistance to apoptosis in cancer, we investigated whether PLCE1 participates in regulation of the NF-κB signaling pathway in ESCC. As shown in Fig. 4a, Real-time PCR Chip analysis shows that inhibition of PLCE1 significantly reduced NF-κB downstream genes, which suggest that PLCE1 may contribute to activation of NF-κB signaling pathway-related molecules. We observed that the downregulation of PLCE1 decreased, subsequently decreasing phosphorylated-IKK-α/β and-IκBα protein. The level of Nuclear NF-κB p65 protein decreased in PLCE1-silenced cells (Fig. 4b), suggesting that PLCE1 regulates NF-κB transcriptional activity by promoting NF-κB “cytoplasmic-nuclear” translocation. To clarify the specific molecular mechanism on how PLCE1 affects NF-κB, we tested whether PI-PLCε signaling pathway, which hydrolyzes plasma membrane phosphatidylinositol 4,5-bisphosphate (PIP2) to produce inositol trisphosphate (IP3) and diacylglycerol (DAG), was involved. When the ESCC cells were treated with 12-O-tetradecanoylphorbol 13-acetate (TPA), a regular substitution of diacylglycerol (DAG), results showed it could rescue the side effect of PLCE1 knockdown in NF-κB-related proteins, phosphorylated-IKK-α/β, −p65, and-IκBα protein. When cells were treated with bisindolylmaleimide (BIM), a broad-specificity DAG-dependent activation of protein kinase Cα (PKCα) inhibitor, it showed a similar effect to PLCE1 knockdown (Fig. 4c). As the great cleaved products after PLCE1 activated, IP3 and DAG were measured by ELISA assays. The level of IP3 and DAG in ESCC cells showed a significant decrease after treated with shR-PLCE1 (Fig. 4d). IP3-dependent calcium release was also observed through laser confocal microscopy. PLCE1 knockdown also triggers the downregulation of Ca²⁺ content as well (Fig. 4e, Additional file 4: Figure S3a). Silencing of PLCE1 after bay-11-7082 treatment (individually or in combination) reduced NF-κB luciferase reporter gene activity (Fig. 4f). IκBα-S32 phosphorylation antibody showed increased phosphorylation after TNFα treatment. However, silencing PLCE1 expression significantly decreased expression of IκBα-S32 phosphorylation (Fig. 4g). Consistently, typical NF-κB signaling gene protein and the phosphorylation of p65 in ESCC cells were downregulated by U73122 (Fig. 4h). Endogenous p65 was localized to the nucleus of controls, whereas p65 nuclear localization significantly decreased in shR-PLCE1-infected Eca109 cells (Fig. 4i). Immunoprecipitation (IP) assay shows that anti-PLCE1 bound to a band in p65 and IκBα IPs, which migrated together with PLCE1 protein in cell lysates (Fig. 4j left). Conversely, p65 and IκBα antibodies bound to PLCE1 protein in anti-p65 and anti-IκBα antibody (Fig. 4j right). Thus, PLCE1 and p65 and IκBα proteins interact. Chromatin IP (ChIP) experiments shows that p65 binding to the VEGF-C promoter region (− 367/− 198) increased after PLCE1 exposure.These results suggest that PLCE1 induces the recruitment of p65 to the promoter region of endogenous VEGF-C gene, and leads to VEGF-C expression (Fig. 4k left). Similarly, we also observed that PLCE1 induces p65 to the promoter region (− 2445) of the endogenous bcl-2 gene and leads to bcl-2 expression (Fig. 4k right).

Therefore, we hypothesized that PLCE1 regulates apoptosis and angiogenesis via the NF-κB signaling pathway in ESCC cells. To test this hypothesis, we first used NF-κB inhibitor Bay11–7082 to inhibit the NF-κB signaling pathway. Data show that bay11–7082 exerted time- and dose-dependent inhibition of NF-κB signaling pathway in human ESCC (Fig. 5a). PLCE1 knockdown in infected Eca109 and EC9706 ESCC cells increased the effects of NF-κB inhibitor, reducing proliferation and increased apoptosis (P < 0.05, Fig. 5b-f, Additional file 5: Figure S4a, b). PLCE1 knockdown in infected Eca109 and EC9706 ESCC cells increased the effects of NF-κB inhibitor, reducing HUVEC proliferation and ESCC cells’ ability to induce tubule formation and migration by HUVECs in vitro (P < 0.05, Fig. 5g-k, Additional file 5: Figure S4c).

We next sought to determine whether NF-κB inhibition affected tumor growth in PLCE1-silenced Eca109 cells subcutaneous tumor model. Figure 6a shows bioluminescent images of xenografts at different time points and treatment. On day 25, we observed metastatic signals in controls and shR-PLCE1-Bay11–7082-treated groups and tumor number and size decreased significantly (Fig. 6a). Tumor-bearing mice treated with Bay11–7082 had reduced tumor volume and weight compared with controls (P < 0.05, Fig. 6b-e). During the experiment, few nude mice appearance cachexia status, and before the end of the experiment, treatment and control groups emerge different number of deaths and the shR-PLCE1-Bay11–7082-treated group has no die (Fig. 6f). We confirmed the decreased expression of VEGF-C and Bcl-2 by Western blot (Fig. 6g) and immunohistochemistry for PLCE1, p65, Ki-67, Bcl-2, and CD34 showed reduced expression of all three proteins in Bay11–7082-treated xenografts (Fig. 6h, i). These results confirm our in vitro observations.

Finally, we examined whether PLCE1/NF-κB axis was also clinically relevant. ESCC tissue specimens showed that PLCE1 expression was positively correlated with p65, IKK, and IκBα expression (P < 0.001, P < 0.001, and P < 0.001, respectively; Fig. 7a, b). Consistently, patients with higher IκBα expression had shorter survival (P = 0.115 × 10^− 3; Fig. 7c). Figure 7d shows PLCE1 expression in The Cancer Genome Atlas (TCGA) database of 198 collected clinical esophageal carcinoma samples correlated positively with IKKα expression (r = 0.523, P = 2.8E-15) and IKKβ expression (r = 0.229, P = 0.001) and negatively with IκBα expression (r = − 0.177, P = 0.013). These data suggest that epigenetically upregulated oncoprotein PLCE1 can activate the NF-κB signaling pathway and may promote angiogenesis and cause poor clinical outcomes in ESCC (Fig. 7e).