ARNTL2 regulated the oncogene c-myc and promoted the progression of esophageal cancer through activating ANXA2 transcription

Abstract

Objective

Aryl hydrocarbon receptor nuclear transporter-like 2 (ARNTL2) can bind to clock circadian regulator (CLOCK) to regulate gene expression and is abnormally expressed in various cancers. Nevertheless, its effects on esophageal cancer (ESCC) are unclear. This work can uncover the intriguing mechanism of ARNTL2 in ESCC.

Methods

Malignant phenotypes including cell proliferation, invasion, migration, and epithelial mesenchymal transition (EMT), were investigated. We established a BALB/c nude mouse (5−6 weeks) model with ESCC to verify the influence of ARNTL2/ANXA2/C-MYC axis. ESCC tissues (n = 100) and paired adjacent normal tissues (n = 100) from patients with ESCC were collected. The recruitment of ARNTL2 and CLOCK in ANXA2 promoter was studied by ChIP and dual-luciferase reporter assay. RIP and RNA pulldown were used to explore the relationship between ANXA2 and C-MYC mRNA.

Results

Compared to adjacent normal tissues, ESCC tissues developed the significant increase ARNTL2, ANXA2, and C-MYC. ARNTL2, which interacts with CLOCK, was recruited in ANXA2 promoter and elevated ANXA2. ARNTL2 silence reduced cell proliferation, migration and invasion and inhibited EMT, which was reversed by ANXA2 overexpression. ANXA2 can bind to the 3ʹUTR of C-MYC transcript; further assays confirmed that ANXA2 increased the protein abundance of C-MYC. ANXA2 knockdown resulted in a decrease in malignant phenotypes, whereas C-MYC overexpression reversed these changes. ARNTL2 silence inhibited the formation, growth and EMT of subcutaneous tumors and suppressed C-MYC; ANXA2 overexpression reversed these alterations.

Conclusion

ARNTL2 activated the transcription of ANXA2, which interacts with C-MYC transcript, promoting the development of malignant behaviors of ESCC cells.

Introduction

Esophageal cancer (ESCC) is one of the common cancers in the world and is the seventh leading cause of cancer death.^1,2 There are more than 51 million new cases of ESCC worldwide in 2022.¹ ESCC is highly insidious, and many patients are found in the advanced stage at the time of diagnosis.³ Current treatments, such as radical resection or targeted therapy cause the non-significant effect on increasing the 5-y survival of patients at advanced stages. Understanding the mechanism regulating the occurrence and development of ESCC is instrumental in promoting advances in novel treatments.⁴

Aryl hydrocarbon receptor nuclear transporter-like 2 (ARNTL2) is a member of PAS (PER, ARNT, and SIM) superfamily characterized by a helix-loop-helix structure.^5,6 ARNTL2 is differentially expressed in various cancers, such as non-small cell lung cancer, bladder cancer and triple-negative breast cancer.^6-8 As a transcription factor, ARNTL2 can form a heterodimer with clock circadian regulator (CLOCK); this complex activates transcription by binding to E-box sequences in the promoters of its target genes.⁹ However, the potential targets of ARNTL2 in the pathogenesis and progression of ESCC remain unclear.

Annexin 2 (ANXA2) is an important member of the calcium-dependent phospholipid-binding protein family.^10,11 As an RNA-binding protein, ANXA2 can interact with target RNA to affect cell growth and signal transduction in the cellular process.^12,13 ANXA2 is capable of affecting tumorigenesis and tumor metastasis, including triple-negative breast cancer, glioma and gastric cancer.^14-17 In ESCC, ANXA2 is capable of regulating Wnt/β-catenin pathway, phosphorylation of ERK and the expression of TTK protein kinase to elevate the hyperproliferation and malignant expansion of cancer cells.^18-20 ANXA2 is overexpressed in ESCC, and nevertheless, the cause of its aberrant expression is still not fully understood.

The oncogene C-MYC has been widely investigated in the pathogenesis of cancers, which displays the implication in various cellular processes, such as cell cycle, apoptosis, differentiation and proliferation.²¹ The dysregulation in C-MYC has been found in 70% of human cancers.²² Overexpression of C-MYC promotes the tumorigenesis and immune escape in ESCC and is positively correlated with poor overall survival and lymph node metastasis in patients with ESCC.^23,24 C-MYC expression may be implicated by ANXA2. Previous studies reported that ANXA2 could directly interact with the 3ʹUTR of C-MYC transcript, which resulted in increased protein abundance of C-MYC.^25,26 Unfortunately, the interaction of ANXA2 and C-MYC transcript in ESCC is unclear.

Here, we report that ANXA2 can be upregulated by the transcription factor ARNTL2 on the promoter of ANXA2, which promotes the interaction between ANXA2 and the C-MYC transcript, consequently contributing to the malignant progression of ESCC.

Methods

Cells

Human ESCC cells, including TE-1 (well-differentiated ESCC cells) and KYSE70 cells (poorly differentiated ESCC cells), and normal human esophageal epithelial cell line Het-1A were purchased from ATCC (USA), which were cultured with 10% fetal bovine serum (Gibco, USA)-containing RPMI-1640 medium (Gibco, USA) at 37 °C with 5% CO₂.

Clinical tissues from patients with ESCC

ESCC tissues (n = 100) and paired adjacent normal tissues (n = 100) from patients with ESCC were collected to detect the expression of ARNTL2, ANXA2, and C-MYC using qPCR. All patients were diagnosed with primary esophageal carcinoma, with no history of recurrence, no history of chemoradiotherapy, no history of immunotherapy, no systemic diseases and no esophageal-related diseases, such as reflux esophagitis.

Plasmid transfection

The siRNAs of ARNTL2 and ANXA2 and the negative control of siRNA (si-NC) were purchased from Santa Cruz Biotechnology (ARNTL2 siRNA: sc-141718; ANXA2 siRNA: sc-29683). The upregulation vector of ARNTL2, ANXA2, or C-MYC based on pcDNA 3.1 (oe-ARNTL2, oe-ANXA2 or oe-C-MYC) were generated by Sangon Biotech (Shanghai, China), and the empty plasmid of pcDNA 3.1 was considered as the negative control (oe-NC). Lipofectamine 3000 was used for transfection into the cells. After 48-h transfection, the transfection efficiency was determined by qPCR.

Animal

Healthy BALB/c nude mice (5−6 weeks old; Shanghai Jihui Laboratory Animal Care Co., Ltd.) were subcutaneously injected with KYSE70 cells under anesthesia using sterile syringes. After injection, the mice were housed individually. Once subcutaneous tumors formed, the mice were randomly assigned to three experimental groups (n = 7 per group) based on a random number table:

1. sh-NC + oe-NC: injected with negative control lentiviral particles and an empty overexpression vector;

2. sh-ARNTL2 + oe-NC: injected with ARNTL2 shRNA lentiviral particles and an empty vector;

3. sh-ARNTL2 + oe-ANXA2: injected with ARNTL2 shRNA lentiviral particles and an ANXA2 overexpression vector.

When the tumor volume reached approximately 100 mm³, lentiviral particles (ARNTL2 shRNA or negative control; Santa Cruz Biotechnology) and plasmids (ANXA2 overexpression or empty vector; Sangon Biotech) were intratumorally injected. The tumor dimensions were measured every 7 d using Vernier calipers until day 28. On day 28, the mice were euthanized by cervical dislocation, and the tumors were harvested for paraffin sectioning (5 μm thickness).

The sample size was determined using variance analysis, with E-values between 10 and 20. All procedures complied with ARRIVE guidelines. No animals were excluded. The researchers performing the measurements and analysis were blinded to the group assignments, while the operator was aware of the assignments. The mice were maintained under controlled conditions: a 12 h/12 h light‒dark cycle, an ambient temperature of 21 ± 1 °C, and a relative humidity of 60% ± 1%, with free access to food and water.

CCK-8

The cells were seeded in 96-well plates at a density of 2,000 cells per well in 100  μL of complete culture medium and allowed to adhere overnight at 37 °C in a 5% CO₂ incubator. Then, 10 μL of CCK-8 reagent (E606335, Sangon Biotech, Shanghai, China) was added at T = 0, 24, 48, and 72 h, and the cells were subsequently incubated at 37 °C for 1 h. The optical density (OD) value was read by the microplate reader (Thermo Fisher, USA) at 450 nm.

Colony formation

Cells were seeded in 6-well plates at a low density of 1000 cells per well and cultured for 14 d, with the medium being changed every 3−4 d. After incubation, the medium was aspirated, and the cells were gently washed twice with PBS. Subsequently, the cells were fixed with 4% paraformaldehyde for 30 min at room temperature, followed by two washes with PBS. The fixed cells were then stained with 0.1% (w/v) crystal violet (A600331, Sangon Biotech, Shanghai, China) solution for 15 min at room temperature. After staining, the plates were thoroughly rinsed with tap water and air-dried. Colonies containing more than 50 cells were counted manually.

Transwell

Cell invasion was assessed using 24-well Transwell plates with 8-μm pores (3428, Corning, USA). The upper chamber membranes were precoated with 50 μL of Matrigel (354234, Corning, USA) diluted 1:8 in serum-free cold medium and then allowed to polymerize in a 37 °C incubator for 1 h. A cell suspension (200 μL containing 1 × 10⁵ cells) in serum-free medium was added to the upper chamber. The lower chamber was filled with 500 μL of medium containing 10% FBS as a chemoattractant. After 16 h of incubation at 37 °C, noninvading cells on the upper surface of the membrane were removed by scrubbing with a cotton swab. The cells that had invaded the lower surface were fixed with 4% paraformaldehyde (P0099, Beyotime, Shanghai, China) for 30 min, stained with 0.1% crystal violet for 5 min, washed gently with PBS, and then photographed and counted under an optical microscope (at least 5 random fields per chamber).

EdU assay

EdU Cell Proliferation Kit was used in this work (C0071S; Beyotime, Shanghai, China). The cells (2000 cells/100 μL) were transferred into 6-well plates overnight at 37 °C, followed by incubation with 10 μmol/L EdU solution at 37 °C for 2 h. Subsequently, the cells were fixed with 4% polyformaldehyde at room temperature for 15  min and washed with PBS containing 3% bovine serum albumin (ST023, Beyotime, Shanghai, China). The cells were next incubated with 0.3% Triton X-100 (T9284, Millipore Sigma, Germany) buffer at room temperature for 10 min. Then, the click reaction solution was added to incubate cells for 30 min, protected from light. Hoechst 33342 solution (C1029, Beyotime, Shanghai, China) was added to stain the cells for 10 min, protected from light. The images were captured using a fluorescence microscope (Olympus, Japan), and the percentage of EdU-positive cells (red fluorescence) relative to the total number of cells (blue fluorescence) was calculated from at least three independent fields.

Wound healing

Cells were seeded in 6-well plates at a high density of 1 × 10⁷ cells per well and cultured until they reached 100% confluence. A sterile 200 μL pipette tip was used to create a straight scratch across the cell monolayer. The wells were gently washed twice with PBS to remove detached cells. Fresh serum-free medium was then added to minimize cell proliferation. Images of the scratch were taken at 0 h (immediately after scratching) and 24 h using an inverted microscope. The migration distance was measured using ImageJ software (National Institutes of Health, USA). The migration rate was calculated as follows: Migration rate (%) = (W0 − W24)/W0 × 100%, where W0 and W24 represent the scratch width at 0 and 24 h, respectively.

Dual luciferase report gene assay

The mutant of ANXA2 promoter was generated using QuikChange® Lightning Site-Directed Mutagenesis kit (Agilent, Santa Clara, CA, USA). The mutant or the wild type of ANXA2 promoter was cloned into pGL4 vector containing reporter gene and then co-transfected with oe-ARNTL2 and oe-NC, respectively. Then, Dual-Luciferase® Reporter Assay System (E1910; Promega, USA) was used to detect the luciferase activity. The relative luciferase activity was calculated.

Co-IP

The cells lysates extracted by the protease inhibitor-containing lysis buffer were divided into input and IP samples. ARNLT2 and CLOCK in the input samples were detected by immunoblotting. Meanwhile, the IP samples were incubated with ARNTL2 (sc-365469, dilution: 1:50; Santa Cruz Biotechnology, Inc., USA) or IgG (ab205719, dilution: 1:50; Abcam, USA) antibodies overnight at 4 °C, followed by the incubation with protein A/G beads (HY-K0202, MedChemExpress, USA) overnight at 4 °C. Then, these beads were collected using the Magnetic Separation Rack (FMS012; Beyotime, Shanghai, China). After the protein was eluted, ARNTL2 and CLOCK in the IP samples were detected by Immunoblotting.

ChIP assay

The cells were incubated with 4% polyformaldehyde at room temperature for 10 min, followed by the incubation with 10× glycine for 5 min at room temperature. After centrifugation, the precipitates were performed by the fragmentation using ultrasonic crushing. The DNA fragments were divided into input and ChIP samples. The ChIP samples were incubated with ARNTL2 or IgG antibody overnight at 4 °C, followed by incubation with ChIP-grade protein G magnetic beads at 4 °C for 2 h. Then, the beads were collected by magnetic separation rack. After elution and decrosslinking, the products were subjected to re-ChIP. In re-ChIP, IgG or CLOCK antibody were used to incubate the products overnight at 4 °C, followed by the incubation with protein G magnetic beads at 4 °C for 2 h. The beads were subsequently collected by magnetic separation rack. After elution, decrosslinking, and purification, the DNA was subjected to qPCR to detect the ANXA2 promoter.

RIP

The cell lysates were divided into RIP and input samples. ANXA2 in input samples was detected by immunoblotting. Protein A/G beads were preincubated with 5% BSA blocking buffer at 4 °C for 30 min, followed by the incubation with IgG or ANXA2 antibody at 4 °C for 4 h. RIP samples were incubated with IgG or ANXA2 antibody at 4 °C for 4 h, followed by incubation with Protein A/G beads overnight at 4 °C. The protein-beads complexes were collected by the Magnetic Separation Rack (FMS012; Beyotime, Shanghai, China). After proteinase K digestion, RNA extracted from RIP samples was analyzed by qPCR to examine the enrichment of C-MYC mRNA.

RNA pulldown

The 3ʹUTR sequence of C-MYC mRNA obtained from UCSC database was amplificated by PCR based on DNA templates extracted from ESCC cells using the DNA extraction kit (DP304; Tiangen, Beijing, China). The PCR products were used for the in vitro transcription through T7 high-yield RNA synthesis kit (B639253-0050; Sangon Biotech, Shanghai, China) and then biotinylation using Pierce® RNA 3ʹ End biotinylation kit (20160; Thermo Scientific, USA). The biotin-labeled RNA was incubated with streptomycin affinity beads at 4 °C for 6 h. Next, the cell lysates were incubated with the RNA-beads complex overnight at 4 °C. After being washed, the beads in the mixed solution consisting of PBS buffer and 5 × SDS loading buffer (4:1, v/v) were heated at 100 °C for 10 min. After the electrophoretic separation, ANXA2 was detected by immunoblotting.

TUNEL

Apoptosis in subcutaneous tumor tissues was detected using the TUNEL staining kit (C1088, Beyotime, Shanghai, China). Briefly, 5-μm-thick tissue sections were dewaxed and rehydrated. After permeabilization with 20 μg/mL proteinase K (provided in the kit) in PBS for 30 min at 37 °C, the sections were washed three times with PBS. The sections were then incubated with 50 μL of TUNEL reaction mixture for 60 min at 37 °C in the dark. Following incubation, the sections were washed three times with PBS to terminate the reaction. Appropriate controls were included: a positive control treated with DNase I (1 μg/mL) for 10 min prior to labeling and a negative control incubated with label solution only (without terminal deoxynucleotidyl transferase enzyme). The sections were mounted with antifade mounting medium and visualized under a fluorescence microscope (Olympus, Japan). The green fluorescence of TUNEL-positive cells (indicating apoptosis) was captured at excitation/emission wavelength of 450/515 nm. For each section, TUNEL-positive cells were counted in five randomly selected fields (200× magnification), and the apoptosis rate was expressed as the percentage of TUNEL-positive cells relative to the total number of cells (determined by DAPI or Hoechst counterstaining) in the same field.

Immunohistochemistry (IHC)

After dewaxing and hydration, the tissue sections were incubated with citrate buffer at 95 °C for 8 min, followed by the incubation with 3% H₂O₂ at room temperature for 10 min. Then, the sections were incubated with 5% goat serum-containing blocking buffer at room temperature for 20 min. The sections were incubated with ki-67 (ab15580, diluted 1:50; Abcam, USA), vimentin (ab8978, diluted 1:100; Abcam, USA), N-cadherin (ab98952, diluted 1:200; Abcam, USA), and E-cadherin (ab231303, diluted 1:100; Abcam, USA) overnight at 4 °C, followed by the incubation with biotin-linked secondary antibodies for 30 min at 37 °C. HRP-conjugated streptavidin (D111054-0001, diluted 1:200; Sangon Biotech, Shanghai, China) was used to incubate the sections for 20 min at 37 °C, and then DAB color developer (C520017-0005; Sangon Biotech, Shanghai, China) was added for 5 min at room temperature. The sections were subsequently stained by hematoxylin dye for 3 min and washed with PBS buffer. After hydration, transparency, and film sealing, the sections were observed using the optical microscope (Olympus, Japan). There were 3 mice in each group used for IHC detection.

qPCR

Total RNA extracted using Trizol reagent (R0016; Beyotime, Shanghai, China) was reverse transcribed into cDNA using Reverse Transcription Kit (RT31-020, QIAGEN, Germany), and then quantification for cDNA were performed in QuantStudio 6 Flex real-time PCR system (Applied Biosystems, USA) using SYBR Green- or Dye-based qPCR kit (204343, QIAGEN, Germany). The primer sequences were listed in Table 1. There were 6 mice in each group used for qPCR analysis.

Table 1.

Primer sequence.

Primer	Sequence
ARNTL2-reverse	5ʹ-GGA GTT GTT TCC ACC AAG TCC TT-3ʹ
ARNTL2-forward	5ʹ-GGC ATC GAA ATC CAA CTG TGC AC-3ʹ
ANXA2-reverse	5ʹ-TCG GAC ACA TCT GGT GAC TTC C-3ʹ
ANXA2-forward	5ʹ-CCT CTT CAC TCC AGC GTC ATA G-3ʹ
C-MYC-reverse	5ʹ-CCT GGT GCT CCA TGA GGA GAC-3ʹ
C-MYC-forward	5ʹ-CAGA CTC TGA CCT TTT GCC AGG-3ʹ
GAPDH-reverse	5ʹ-ACC ACC CTG TTG CTG TAG CCA A-3ʹ
GAPDH-forward	5ʹ-GTC TCC TCT GAC TTC AAC AGC G-3ʹ

Immunoblotting

Cell lysates or tissue homogenates were centrifuged at 13,000 × g for 15 min at 4 °C, and next BCA protein concentration detection assay (Beyotime, Shanghai, China) was used to measure the protein concentration in the supernatant. After the electrophoretic separation of SDS–PAGE, the protein was transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, USA). Then, the membranes incubated with blocking buffer containing 5% skim milk for 1 h was incubated with diluted primary antibodies (listed in Table 2) overnight at 4 °C. The membranes were then incubated with goat anti-rabbit antibody (HRP-linked) for 1 h at 37 °C. Chemiluminescence substrates (PK10003, Proteintech, Wuhan, China) were added to the membranes for 1 min. Blots were measured by the gel imaging system (Bio-Rad, USA). β-actin was used as the internal protein. There were 6 mice in each group used for immunoblotting detection.

Table 2.

Antibody dilution.

Antibody	Dilution	Item	Manufacturer
Anti-ARNTL2	1:1000	ab221557	Abcam, USA
Anti-vimentin	1:1000	ab92547	Abcam, USA
N-cadherin	1:5000	ab76011	Abcam, USA
Anti-E-cadherin	1:25	ab227639	Abcam, USA
Anti-CLOCK	1:2000	ab3517	Abcam, USA
Anti-ANXA2	1:5000	ab189473	Abcam, USA
Anti-C-MYC	1:1000	ab32072	Abcam, USA
Anti-β-actin	1:1000	ab213262	Abcam, USA
Goat anti rabbit antibody	1:10000	ab6721	Abcam, USA

Screening and identification of the differentially-expressed genes in ESCC

Clinical data were obtained from TCGA, including adjacent cancer samples and 153 esophageal cancer samples. The transcriptome data were used for the differential expression analysis by DESeq2 R package. The gene with | FoldChange| > 2 and adj. p-value < 0.05 was considered as the differentially-expressed gene. p-value was adjusted using Benjamini–Hochberg method. To further explore its potential implications, we employed the deconvo_CIBERSORT method from the R package IOBR to analyze the correlations among ARNTL2, ANXA2, and C-MYC with tumor-infiltrating immune cells. We also assessed the relationship between ARNTL2/ANXA2 and drug resistance using the GSCA database.

Statistical analysis

Statistical analysis was performed using SPSS software (version 25.0). Data visualization was conducted with GraphPad Prism (version 9.0). All experiments were performed with three independent biological replicates, each consisting of three technical replicates. The data are presented as the mean ± standard deviation (SD). Statistical significance between groups was determined using an unpaired two-tailed Student's t-test (comparison between two groups) or one-way ANOVA (≥3 groups), followed by Tukey's post hoc test for multiple comparisons. A p-value of less than 0.05 was considered statistically significant.

Results

ARNTL2 was significantly increased in ESCC

We previously screened the differentially expressed genes (DEGs) based on TCGA database containing 11 paracancerous tissue samples and 153 ESCC samples. A total of 1474 DEGs were identified. Importantly, we found that the transcript factor ARNTL2 was significantly upregulated in ESCC (log₂ FC = 1.252, p < 0.001) (Figure 1A). To further demonstrate the change in ARNTL2, we quantified ARNTL2 expression in ESCC cells, including TE-1 (hypofractionated) and KYSE70 (hyperfractionated) cells. ESCC cells developed the increased ARNLT2 (p < 0.05) (Figure 1B,C); and particularly, hypofractionated ESCC cells showed the elevated ARNTL2 than that in the hyperfractionated cells (p < 0.05) (Figure 1B,C). Overall, ARNTL2 was significantly upregulated in ESCC. Given that ARNTL2 developed the increased expression in ESCC cells with hyperfractionation, we selected KYSE70 cells as the research object in the following experiments.

Figure 1.

ARNTL2 was significantly decreased in ESC. The differentially expressed genes (DEGs) based on TCGA database containing 11 paracancerous tissue samples and 153 ESCC samples. ESCC cell lines, including TE-1 (well-differentiated) and KYSE70 (poorly differentiated), were used for the quantification of DEGs, and Het-1A cells were used for the normal control. (A), Differentially expressed genes in ESC based on TCGA database (blue plot: downregulated gene; red plot: upregulated gene; grey plot: gene without differential expression). ARNTL2 was upregulated in ESCC. (B), ARNTL2 expression in Het-1A, TE-1, and KYSE70 cells quantified by qPCR. (C), ARNTL2 protein abundance in Het-1A, TE-1, and KYSE70 cells was detected by Immunoblotting, normalized to that of β-actin.

ARNTL2 knockdown inhibited the malignant behaviors of KYSE70 cells

To investigate the role of ARNTL2 in ESCC progression, ARNTL2-specific siRNA was transfected to downregulate ARNTL2 (Figure 2A). ARNTL2 knockdown significantly inhibited cell proliferation, as evidenced by reduced cell viability, fewer colonies, and a lower percentage of EdU-positive cells (p < 0.05) (Figure 2B–D). Additionally, ARNTL2 silence resulted in the significant decrease in migration and invasion phenotypes (p < 0.05) (Figure 2E,F). The biomarkers of EMT were altered because of ARNLT2 siRNA, characterized by the downregulation of vimentin and N-cadherin and the enhancement in E-cadherin in si-ARNTL2 group (p < 0.05) (Figure 2G).

Figure 2.

ARNTL2 was capable of promoting the malignant behaviors of ESCC cells. The specific siRNA of ARNTL2 (si-ARNTL2) and its negative control (si-NC) were transfected into KYSE70 cells to investigate the role of ARNTL2 in ESCC progression. (A), ARNTL2 expression was quantified by qPCR after siRNA transfection in KYSE70 cells. (B–D), Cell proliferation was investigated by CCK-8 kit (B), Colony formation assay (C) and EdU kit (D). (E–F), Cell migration and invasion were evaluated by wound healing assay (E) and Transwell assay (F), respectively. (G), Protein abundance of ARNTL2 and EMT biomarkers, including vimentin, N-cadherin, and E-cadherin detected by Immunoblotting, normalized to β-actin.

ARNTL2 promoted the transcription of ANXA2

Intriguingly, ANXA2 might be the potential target of ARNTL2-CLOOK heterodimer according to CHEA database. ANXA2 was associated with ESCC progression, showing its significant increase in ESCC tissues as compared to paracancerous tissues (Figure 3A). ANXA2 was higher both in TE-1 and KYSE70 cells compared with that in Het-1A cells; its expression was correlated with poor differentiation in ESC cells (p < 0.05) (Figure 3B,C). Importantly, ARNTL2 silence, in ESCC cells, caused the downregulated ANXA2 (p < 0.05), indicating potential regulation between these two factors (Figure 3D,E). In ESCC cells, ARNTL2 can bind to CLOCK to form the heterodimer as shown in Figure 3F, suggesting that the ARNTL2/CLOCK heterodimer may regulate transcription. We observed that the ANXA2 promoter was highly enriched in ARNTL2, and ARNLT2 overexpression contributed to the increase in relative luciferase activity when there was the ANXA2 promoter loaded in the luciferase vector (Figure 3F,G). Meanwhile, we determined that the ANXA2 promoter was also highly enriched in CLOCK (Figure S1). To further determine the functional necessity of CLOCK in ANXA2 expression, we established a mutant of CLOCK (CLOCK-mut). We found that CLOCK-mut failed to bind to ARNTL2 (Figure S2A); subsequently, we determined overexpression of ARNTL2 could not enrich in ANXA2 promoter (Figure S2B). We then observed that ARNTL2 overexpression resulted in the upregulation of ANXA2; however, CLOCK silence reversed the abovementioned changes in ANXA2 (Figure S2C). Based on these results, ARNTL2 was capable of promoting the transcription of ANXA2 in ESCC.

Figure 3.

ARNTL2 promoted the transcription of ANXA2. Using database predictions, we confirmed the transcriptional regulation of ANXA2 by the ARNTL2/CLOCK heterodimer. This was achieved through the assessment of ANXA2 expression in both tissues and cells, ARNTL2 knockdown experiments, binding assays, and luciferase reporter assays. A, Differentially expressed genes in ESCC based on TCGA database. ANXA2 was increased in ESCC; B, ANXA2 expression in Het-1A, TE-1 and KYSE70 cells quantified by qPCR; C, ANXA2 protein abundance in Het-1A, TE-1, and KYSE70 cells detected by immunoblotting. D, ANXA2 expression after ARNTL2 silencing was quantified by qPCR; E, ANXA2 protein abundance after ARNTL2 silencing was detected by Immunoblotting. F, The interaction between ARNTL2 and CLOCK was investigated by Co-IP. G–H, The recruitment of ARNTL2 in ANXA2 promoter was evaluated by ChIP (G) and dual luciferase reporter gene assay (H).

ANXA2 mediated the pro-cancer role of ARNTL2 in KYSE70 cells

To investigate whether ANXA2 mediates the oncogenic effects of ARNTL2, ARNTL2 siRNA, and an ANXA2 overexpression vector were co-transfected into KYSE70 cells (p < 0.05) (Figure 4A,G). Then, we found that overexpression of ANXA2 largely rescued the impaired proliferation, migration, and invasion caused by ARNTL2 knockdown (Figure 4B–F). We subsequently examined the changes in EMT biomarkers during the cooperation of the increased ANXA2 and the decreased ARNTL2 and found that ANXA2 overexpression offset the ARNTL2 siRNA-induced changes in vimentin, N-cadherin, and E-cadherin (Figure 4G). These data demonstrate that ANXA2 is a key downstream mediator of ARNTL2's pro-oncogenic functions.

Figure 4.

ANXA2 mediated the procancer role of ARNTL2 in ESCC. We conducted rescue experiments in KYSE70 cells to explore how ARNTL2 enhances ANXA2 to promote malignancy in ESCC cells. (A), ARNTL2 and ANXA2 expression was quantified by qPCR. (B–D), Cell proliferation was investigated by CCK-8 kit (B), Colony formation assay (C) and EdU kit (D). (E–F), Cell invasion and migration were evaluated by Transwell assay (E) and wound healing assay (F), respectively. (G), Protein abundance of ANXA2, ARNTL2, and EMT biomarkers, including vimentin, N-cadherin, and E-cadherin detected by immunoblotting.

ANXA2 might bind to C-MYC transcript to promote the malignant phenotype in KYSE70 cells

In our study, the RIP assay revealed an increased enrichment of the C-MYC transcript on ANXA2 upon ANXA2 overexpression (p < 0.05) (Figure 5A). RNA pulldown analysis further confirmed that ANXA2 could directly interact with 3ʹUTR of C-MYC transcript (Figure 5B). We then demonstrated that ANXA2 overexpression led to the upregulation of C-MYC (p < 0.05) (Figure 5C). Thus, these results suggested that ANXA2 elevates C-MYC in ESCC via direct binding to the 3ʹUTR of C-MYC transcript. We subsequently determined that decreasing the amount of ANXA2 reduced the protein and mRNA abundance of C-MYC (p < 0.05) (Figure 5D,J). Intriguingly, C-MYC overexpression vector reversed the role of ANXA2 siRNA in C-MYC, and failed to affect ANXA2 expression in ESCC cells (p < 0.05) (Figure 5D,J), suggesting that C-MYC was the downstream of ANXA2 in ESCC. We determined that the decreased ANXA2 contributed to the dropped proliferation in ESCC cells, whereas the increased C-MYC reversed the role of ANXA2 siRNA in proliferation (Figure 5E–G). C-MYC upregulation increased the ANXA2 siRNA-decreased migrative and invasive cells (Figure 5I). Additionally, ANXA2 siRNA significantly inhibited EMT, characterized by the decrease in vimentin and N-cadherin and an increase in E-cadherin, which was reversed by C-MYC overexpression (p < 0.05) (Figure 5J). These findings suggest that ANXA2 promotes ESCC malignancy by binding to and stabilizing C-MYC mRNA.

Figure 5.

ANXA2 directly bound to the 3ʹUTR of c-myc mRNA to promote the proliferation, migration, invasion, and EMT in ESCC cells. To explore the regulatory link between ANXA2 and C-MYC in ESCC, we used RIP and RNA pulldown assays to assess their interaction. We performed ANXA2 overexpression and knockdown experiments, followed by C-MYC overexpression rescue studies. Functional assays, such as cell proliferation, migration, and invasion tests, were conducted, along with EMT marker analysis under different genetic conditions. (A), the interaction between ANXA2 and 3ʹUTR of c-myc mRNA determined by RIP. (B), the recruitment of ANXA2 in 3ʹUTR of c-myc mRNA detected by RNA pulldown. (C), the protein abundance of c-myc after ANXA2 overexpression was detected by immunoblotting. (D), ANXA2 and c-myc expression was quantified by qPCR; (E), Cell proliferation was investigated by CCK-8 kit (E), Colony formation assay (F) and EdU kit (G). (H), Cell migration was evaluated by wound healing assay. (I), Cell invasion was evaluated by Transwell assay. (J), Protein abundance of ANXA2, c-myc, and EMT biomarkers, including vimentin, N-cadherin, and E-cadherin detected by immunoblotting.

ANXA2 overexpression reverses the tumor-suppressive effects of ARNTL2 knockdown in TE-1 cells

To generalize our findings, we validated the ARNTL2/ANXA2/C-MYC axis in TE-1 cells. In these cells, the silencing of ARNTL2 resulted in the downregulation of both ANXA2 and C-MYC, whereas ANXA2 overexpression restored C-MYC levels (p < 0.05) (Figure 6A,B). Additionally, ARNTL2 knockdown significantly inhibited cell proliferation, an effect that was reversed upon ANXA2 overexpression (p < 0.05) (Figure 6C–E). The migratory and invasive capacities of TE-1 cells were also substantially diminished following ARNTL2 silencing, yet these effects were similarly mitigated by ANXA2 overexpression (p < 0.05) (Figure 6F,G). Figure 6H illustrates the alterations in epithelial‒mesenchymal transition (EMT) markers across the experimental groups: ARNTL2 knockdown led to an increase in E-cadherin and a decrease in N-cadherin and vimentin, changes that were again reversed by ANXA2 overexpression (p < 0.05). Based on the evidence summarized above, we propose that the ARNTL2/ANXA2/C-MYC axis plays a critical role in ESCC. To further explore its potential implications, we employed the deconvo_CIBERSORT method from the R package IOBR to analyze the correlations among ARNTL2, ANXA2, and C-MYC with tumor-infiltrating immune cells. We also assessed the relationship between ARNTL2/ANXA2 and drug resistance using the GSCA database. As shown in Figure S3A, in ESCC, C-MYC showed a significant negative correlation with memory B cells and Tregs but a positive correlation with M1 macrophages; ANXA2 was negatively correlated with naïve B cells and positively correlated with resting NK cells, M0 macrophages, and activated dendritic cells; ARNTL2 was significantly negatively correlated with memory B cells, plasma cells, CD8⁺ T cells, and Tregs, while also showing significant positive correlations with resting NK cells, M0 macrophages, and M1 macrophages. Additionally, Figure S3B demonstrates that both ARNTL2 and ANXA2 are significantly associated with resistance to multiple anticancer drugs.

Figure 6.

ANXA2 overexpression reversed the tumor-suppressive effects of ARNTL2 knockdown in TE-1 cells. To further demonstrate the impact of the ARNTL2/ANXA2/C-MYC axis on ESCC, we conducted experiments using TE-1 cells subjected to ARNTL2 knockdown followed by ANXA2 overexpression. *p < 0.05, **p < 0.01, ***p < 0.001. (A–B), ARNTL2, ANXA2, and C-MYC were detected by qPCR (A) and Western blotting (B), respectively. (C–E), Cell proliferation was investigated by CCK-8 kit (C), Colony formation assay (D) and EdU kit (E). (F), Cell invasion was evaluated by wound healing assay. (G), Cell migration was evaluated by Transwell assay. (H), Protein abundance of ANXA2, ARNTL2, and EMT biomarkers, including vimentin, N-cadherin, and E-cadherin, were detected by Immunoblotting, normalized to β-actin.

ARNTL2 modulate C-MYC expression and promoted ESCC progression via targeting ANXA2 in the nude mouse model of ESCC

Since ARNTL2-mediated transcription modulation affected ANXA2 expression that elevated C-MYC in ESCC cells, we further demonstrated the mechanism by which ARNTL2/ANXA2/C-MYC axis regulates tumor formation and growth in the nude mouse model of ESCC. We generated mice that were subcutaneously grafted with the ARNTL2-increased and ANXA2-decreased KYSE70 cells. ARNTL2 downregulation resulted in a decrease in ANXA2 and C-MYC, whereas ANXA2 overexpression could offset these changes (p < 0.05) (Figure 7A,B). In a xenograft mouse model, ARNTL2 knockdown markedly inhibited tumor growth, which was significantly reversed by concurrent ANXA2 overexpression (p < 0.05) (Figure 7C,D). As shown in Figure 7E, the proliferation biomarker ki-67 was reduced after ANRLT2 decrease in mice, and ANXA2 upregulation contributed to the increase in ki-67 and offset the effect of ARNTL2 on ki-67. Moreover, sh-ARNLT2 + oe-NC group developed the decrease in vimentin and N-cadherin and an increase in E-cadherin as compared to sh-NC + oe-NC group; subsequently, ANXA2 upregulation reversed these ARNTL2 shRNA-caused changes (Figure 7E). TUNEL staining demonstrated that ARNLT2 silence resulted in the increased apoptosis in subcutaneous tumor, which was significantly reversed by ANXA2 overexpression (Figure 7F). In mice, ARNTL2 downregulation resulted in a decrease in ANXA2 and C-MYC, whereas ANXA2 overexpression could offset these changes (p < 0.05) (Figure 7G). We further collected tumor tissues and their adjacent normal tissues from patients with ESCC and compared ARNTL2, ANXA2, and C-MYC in ESCC tissues (n = 100) and adjacent normal tissues (n = 100). Compared to adjacent normal tissues, ESCC tissues developed the significant increases in ARNTL2, ANXA2, and C-MYC (p < 0.05) (Figure 8A). ROC analysis showed that ARNTL2 (AUC = 0.973; 95% CI = 0.955–0.992), ANXA2 (AUC = 0.956; 95% CI = 0.928−0.981) and C-MYC (AUC = 0.926; 95% CI = 0.891−0.961) could be used for the diagnosis of ESCC (Figure 8B).

Figure 7.

ARNTL2 modulates c-myc expression and promotes ESCC progression via targeting ANXA2 in the nude mouse model of ESCC. (A), c-myc expression in ESCC cells quantified by qPCR; (B), Protein abundance of c-myc in ESCC cells detected by immunoblotting. (C), tumor growth in 4 weeks. (D), tumor weight at Day 28. (E), Ki-67, vimentin, N-cadherin, and E-cadherin in tumor tissues were detected by IHC. (F), Cell apoptosis in tumor tissues was detected by TUNEL staining. (G), ARNTL2, c-myc, and ANXA2 were measured by immunoblotting, normalized to β-actin.

Figure 8.

Expression of ARNTL2, ANXA2, and C-MYC in patients with ESCC. (A), ARNTL2, ANXA2, and C-MYC in ESCC tissues (n = 100) and paired adjacent normal tissues (n = 100) detected by qPCR. (B), ROC analysis.

Discussion

The diagnosis of ESCC is low, and patients are at an advanced stage at the time of diagnosis. Patients with advanced ESCC is difficult to benefit from current treatments. Understanding the regulatory mechanism in ESCC would be instrument in revealing the potential cancer-related targets. This work provides a novel mechanism driving ESCC progression, in which ARNTL2 binds to the promoter region of ANXA2 to trigger the latter transcription via interaction with CLOCK, subsequently promoting the interaction between ANXA2 and 3ʹ UTR of C-MYC transcript, which cause the increased protein abundance of C-MYC, which drives the malignant behaviors of ESCC cells.

ARNTL2 is dysregulated in various cancers and has the potential of evaluating the malignant tumor progression, such as triple-negative breast cancer and hepatocellular carcinoma.^6,27 In this work, we demonstrated that ARNTL2 was overexpressed in ESCC and ARNTL2 could be a biomarker for screening patients with ESCC at the early stage. ARNTL2, as a significant transcription factor, plays a core role in regulating cell processes.²⁸ In this study, ARNTL2 siRNA repressed the malignant phenotypes of ESCC cells, including the elevation in proliferation, migration, invasion, and EMT. Our data suggest that ARNTL2 may function as an oncogene in ESCC. Interestingly, our bioinformatic analysis also indicated a potential association between high ARNTL2 expression and tumor drug resistance, which is consistent with findings in other cancer types.²⁹ While these observations are correlative, they warrant further investigation into whether targeting ARNTL2 with siRNA or other modalities could represent a viable strategy for modulating tumor resistance in the future.

ARNTL2 can interact with CLOCK to inhibit or activate the transcription of targets.^9,30 We confirmed that ANXA2 was the target of the ARNTL2-CLOCK dimer in ESCC. First, we demonstrated that there was the interaction between ARNTL2 and CLOK in ESCC cells based on Co-IP assay; further studies confirmed that the ANXA2 promoter was highly enriched in both ARNTL2 and CLOCK, suggesting that the ARNTL2-CLOCK complex can bind to ANXA2 promoter in ESCC cells. Dual-luciferase reporter gene assay and RNAi-mediated experiment demonstrated that ARNTL2 positively regulated ANXA2 expression in ESCC. CLOCK was necessary for the ARNTL2-induced upregulation of ANXA2 since ARNTL2, with a CLOCK mutation, failed to increase ANXA2 in ESCC cells. The abovementioned results indicated that ARNTL2 was capable of activating the transcription of ANXA2 via its interaction with CLOCK.

ANXA2 is a well-known peripheral membrane-binding protein that develops the increased amplification in response to pathological conditions, such as pathogen and the tumor microenvironment.³¹ ANXA2 has been found to be upregulated in ESCC, which indicates a lower overall survival of patients with ESC.³² In this work, we determined that ANXA2 was significantly increased in ESCC and associated with the poor differentiation degree of cancer cells. The mechanism by which ARNTL2 activated ANXA2 transcription sheds some light on the reason for the upregulation of ANXA2 in ESCC. Further assays confirmed that ANXA2 might mediate the oncogenic role of ARNTL2 in ESCC progression. Our study revealed a notable association between ARNTL2 and ANXA2 and the tumor microenvironment. Through bioinformatic analysis, we identified correlations between the expression levels of ARNTL2 and ANXA2 and the infiltration of immune cells, including CD8⁺ T cells, regulatory T cells (Tregs), macrophages, and natural killer (NK) cells. These findings suggest that ARNTL2 and ANXA2 may play a role in modulating the tumor immune microenvironment via these immune cell populations, potentially contributing to mechanisms such as tumor immune escape. Furthermore, our observations indicated a significant association between ARNTL2 and ANXA2 and tumor drug resistance in various cancer types. These findings suggest that the upregulated expression of these genes may contribute to therapeutic resistance, potentially resulting in unfavorable patient prognoses. Overall, ARNTL2 can elevate ANXA2 to promote the occurrence and development of ESCC.

Recent evidence concluded that ANXA2 can function as an RNA-binding protein to affect the expression of nuclear targets at the posttranscriptional level.^33,34 After being activated, ANXA2 can bind to specific cis‐acting sequences of mRNA to modulate the intracellular localization and translation of RNA.³⁵ ANXA2 has been found to interact with 3ʹUTR of C-MYC transcript as a transaction factor, thereby resulting in the increased protein abundance of C-MYC.^36,37 We determined that ANXA2 can elevate C-MYC expression in ESCC via binding to 3ʹ UTR of C-MYC transcript. The well-known cancer-related gene C-MYC is deregulated in malignant tumor, which enhances genomic instability, signaling transduction dysregulation and the tumor microenvironment.^38,39 C-MYC overexpression is involved in the tumorigenesis, tumor growth and metastasis of ESCC.⁴⁰ ANXA2 promoted the malignant behaviors of ESCC cells by increasing C-MYC, which was determined by our study. Recent studies determined multiple miRNAs can bind to the 3ʹ-untranslated region (3ʹ-UTR) of C-MYC mRNA, thereby inducing its degradation.^41-43 However, ANXA2 may protect against miRNA-mediated degradation by interacting with the 3ʹ-UTR of C-MYC mRNA. Then, we found ARNTL2 and ARNTL2 could modulate C-MYC expression and promote ESCC progression via targeting ANXA2. Direct targeting of c-MYC has proven challenging.⁴⁴ Our study suggests that upstream targeting of ARNTL2 or ANXA2 might offer an indirect strategy to modulate c-MYC activity.

Our study has several limitations that should be considered. First, a key limitation of this study is the use of only two ESCC cell lines. Given the well-documented molecular heterogeneity of ESCC, our findings, while robust in these models, may not fully represent the diversity of the disease. The generalizability of the ARNTL2/ANXA2/C-MYC axis across different ESCC subtypes remains to be established. Second, the clinical correlations require validation in a large, prospective patient cohort to firmly establish ARNTL2 and ANXA2 as prognostic biomarkers. Most importantly, our work leaves key translational questions unanswered. We provide only associative data linking the axis to drug resistance and immune cell infiltration in the TME. Direct experimental evidence demonstrating a causal role in these processes, which are critical for patient outcomes, is lacking. Finally, while we show that ANXA2 binds to c-MYC mRNA, the exact mechanism by which it stabilizes the transcript warrants further elucidation. Future research will address the following four aspects: 1) construction of nanoparticles capable of loading ARNTL2-specific siRNA using nanotechnology, to assess the effects of this specific nanoparticle on ESCC tumor formation, growth, and metastasis, aiming to provide new therapeutic options for ESCC; 2) elucidation of the impact of the ARNTL2/ANXA2/C-MYC axis on the development of ESCC drug resistance, using multiple ESCC cell lines; and 3) exploration of the correlation between the ARNTL2/ANXA2/C-MYC axis and biological factors such as the tumor microenvironment.

In summary, this study identifies the ARNTL2/ANXA2/C-MYC axis as a novel pathway that contributes to the malignant behavior of ESCC cells. While our findings are primarily derived from preclinical models, these findings nominate that ARNTL2, ANXA2, and C-MYC as promising candidate biomarkers and therapeutic targets. Future clinical studies are essential to translate these mechanistic insights into therapeutic applications.

Ethics approval

This study was approved by the Ethics Committee of Bengbu Medical University (No. 2023-244). All procedures adhered to the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants.

All animal experiments received approval from the Ethics Committee of Bengbu Medical University (No. 2023-603).

Acknowledgements

Not applicable.

Author contributions

Yanzi Qin: Conceptualization, Investigation, Project administration, Supervision, Writing – Original Draft. Hongfei Ci: Data Curation, Formal analysis. Zhaoyi Wang: Investigation, Methodology. Yandie Zhang: Resources, Software. Xifeng Xu: Validation, Visualization. Qiang Wu: Writing - Review & Editing. All authors have read and approved the final version of this manuscript to be published.

Supplementary Material

Supplementary material

Figure S1. ChIP assay determined CLOCK was recruited in ANXA2 promoter.

Supplementary material

Figure S2. The functional necessity of CLOCK when ARNTL2 regulates the transcriptional activation of ANXA2. A, CLOCK-mut failed to interact with ARNTL2. B, CLOCK-mut decreased the enrichment of ARNTL2 in the promoter of ANXA2. C, CLOCK silence reversed the ARNTL2 overexpression-induced change in ANXA2

Supplementary material

Figure S3 Correlations of ARNTL2, ANXA2, and C-MYC with immune cell infiltration and drug resistance in ESCC. A, Immune infiltration analysis based on CIBERSORT. B, Correlation between drug sensitivity and the expression of ARNTL2 and ANXA2 (a positive correlation indicates that increased gene expression may confer drug resistance, while a negative correlation suggests that increased expression may be associated with increased drug sensitivity).

Funding Statement

This study was supported by the Key Laboratory of Basic Cancer Research and Clinical Laboratory Diagnosis of Bengbu Medical University, Natural Science Key Project of Anhui Provincial Department of Education (No. 2022AH051521), Postgraduate Scientific Research Innovation Natural Science Research Project of Bengbu Medical University (No. Byycx24019) and Natural Science Key Project of Bengbu Medical University (Nos. 2024byzd124, 2024byzd482).

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/15384047.2025.2574544.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

All data generated or analyzed during this study are included in this article. The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.