ESCC tissue specimen and high-throughput sequencing

ESCC tissues and paired adjacent non-tumor samples were preserved at -80 °C at Changhai Hospital, Navy Medical University, following approval by the Medical Ethics Committee(approval no.CHEC2023-018). The Declaration of Helsinki was followed for informed consent. Circular RNA expression profiles were analyzed by RNA sequencing in four pairs of ESCC and adjacent non-cancerous tissue samples. Total RNA was extracted using TRIzol reagent, and its integrity and purity were verified. An Illumina HiSeq 4000 platform was used to sequence the RNA-seq libraries at Shanghai OE Biotech Co. Ltd. With bioinformatics analyses, including circRNA prediction via CIRI2R software peforemd subsequently. Additionally, 32 pairs of ESCC tissue samples were collected to analyze clinical features and validate tissue expression levels. A detailed description of clinical characteristics presented Supplementary Table S1.

Cell culture

Cell lines obtained from esophageal cancer, including KYSE30, KYSE150, KYSE510, TE-1, and Eca109, as well as the HEK 293 cell line, were provided by Procell Life Science & Technology Co. Ltd(Wuhan,China). Meanwhile, normal esophageal epithelial cells (HET-1A) were sourced from Immocell Biotechnology Co.Ltd. KYSE30, KYSE150, KYSE510, TE-1, and Eca109 cultured in RPMI 1640 medium; DMEM for HEK293T; and a BEGM kit for HET-1A. Each medium type was supplemented with 10% fetal bovine serum to maintain cellular growth.

Real-time quantitative PCR

We isolated RNA samples from ESCC cell lines and tissues utilizing TRIzol reagent(Sigma Aldrich). The cDNA was synthesized from 1000 ng of RNA and analyzed via qRT-PCR with SYBR Green Master Mix. Using the QIAGEN miRNA First-Strand cDNA Kit and miRNA qPCR Kits, miRNA was analyzed through qRT-PCR analysis. GAPDH was used as the normalization value for circRNAs and mRNA expression and U6 for miRNA expression. qRT-PCR was conducted using the LightCycler® 480 Instrument following the manufacturer's guidelines. Gene expression levels were quantified using A two-fold Ct method, with sequences of primer detailed in Supplementary Table S2.

Lentiviral transduction and plasmid construction

The shRNA targeting hsa_circ_0001165 was suppressed via lentiviral transduction, with a non-targeting shRNA as the control, both synthesized by OBiO Technology (Shanghai, China). KYSE30 and KYSE150 cells at 50% confluence were infected at an MOI of 20 and then selected with 2 mg/mL of puromycin (Yeason, China).Genechem Co. Ltd (Shanghai, China) synthesized the hsa_circ_0001165 and the TNS3 overexpression plasmid, which were transfected into KYSE30 and KYSE150 using Lipofectamine 3000 (Invitrogen). In addition to mimics, inhibitors, and negative control sequences for MiR-381-3p, Lipofectamine RNAiMAX (Invitrogen) was used for transfection. qRT-PCR confirmed the success of the transfections.

Western blotting

For the analysis of proteins, cells were initially lysed using a pre-cooled RIPA buffer for 15 min. Following this, the samples were centrifuged for ten minutes at 12,000 rpm, and the protein concentration was measured using the BCA. Following denaturement for ten minutes, proteins were separated onto 10% SDS-PAGE gel. A polyvinylidene fluoride membrane (Millipore, USA) was used to isolate proteins after electrophoresis. The membranes were treated with non-fat milk for one hour to block them, followed by an overnight incubation at 4 °C with primary antibodies: anti-EIF4A3 (1:1200), anti-TNS3 (1:1000), and anti-GAPDH (1:50,000), all obtained from Proteintech, USA. After the incubation, the membrane visualization was carried out using an ECL reagent and detected with an Amersham Mager 600 amplification detector following incubation with a 1:5000 dilution of a secondary antibody (Proteintech, USA).

Nucleocytoplasmic separation experiment

Invitrogen's Ambion® PARISTM Kit was used to isolate RNA samples from nuclear and cytoplasmic fractions. qRT-PCR was performed to detect circular RNA in cDNA from nuclear and cytoplasmic samples with U6 and GAPDH serving as controls.

CCK-8 and clone formation assay

After seeding 3,000 cells in each 96-well plate, we introduced 10 µL of CCK-8 solution at 24, 48, 72, and 96 h following cell adherence. Following a two-hour incubation, the absorbance of the samples was evaluated at 450 nm using a SpectraMax i3 microplate reader. Each experimental condition was assessed in triplicate wells.

To conduct colony assays, introduce 1500 cells into each well of a 6-well plate containing pre-warmed complete medium. Maintain incubation at 37 °C in an atmosphere of 5% CO₂ for a duration of 10 to 14 days. Afterward, rinse the wells using PBS, fix the samples with 4% paraformaldehyde for 10 to 15 min, and subsequently stain with 0.5% crystal violet diluted in methanol for 15 min. Finally, utilize a microscope to count colonies consisting of more than 50 cells and carry out the proliferation rate with ImageJ software (version 1.8.0).

5-Ethynyl-2’-deoxyuridine (EdU) Assay

The process involved incubating a total of 500 L of medium with 10 μmol EdU in a 24-well plate to evaluate the incorporation of EdU. Cells that had been transfected were added to the plates and incubated for two hours in a medium enriched with 10 μmol EdU, utilizing the Cell Proliferation Kit (Beyotime, China). In order to stabilize the cells, paraformaldehyde was utilized for a duration of 30 min. This was succeeded by permeabilization using 0.5% Triton-X-100 for 20 min, followed by a wash with 3% BSA. After performing staining with Alexa Fluor 555, DAPI was utilized for Hoechst staining. Fluorescence images were acquired with a fluorescence microscope (Leica, Germany) and subsequently analyzed with ImageJ software (version 1.8.0).

Assays for cell migration and invasion

For the assessment of invasion or migration, a 1:8 dilution of Matrigel was prepared in medium with serum-free and applied at a volume of 100 μL per well into a 24-well plate that housed inserts with pore sizes of 8 mm. A serum-free medium contained with 5 × 10^4 per well was in the upper chamber of each insert, while the lower compartment was supplemented with 20% fetal bovine serum, serving as a chemoattractant for the growth of ESCC cells. For the upper chamber, 48 h at 37 °C were incubated, followed by 20 min of fixative and staining with crystal violet. ImageJ software (v1.8.0) was used to analyze microscope images from Olympus (Japan) to determine cell migration and invasion rates.

RNA fluorescence in situ hybridization (FISH)

Employing the Fluorescent In Situ Hybridization Kit from RiboBio (Guang Zhou), probes labeled with Cy3 were synthesized for hsa_circ_0001165, while FAM-labeled probes were designed for miR-381-3p. Cells underwent fixation using paraformaldehyde for two hours at a temperature of 4 °C. For cell permeabilization, a 0.5% Triton X-100 solution was utilized for five minutes. After a prehybridization step at 37 °C for 30 min, 5 M probes were hybridized overnight, and imaging was carried out using a Leica confocal microscope.

RNA pulldown assays

Probes that were biotinylated and aimed at hsa_circ_0001165, along with corresponding controls, were generated by CloudSeq Biotech (Shanghai, China), with the specific probe sequences provided in Supplementary Table S3. For this assay, Thermo Fisher Scientific's RNA–Protein Pull-Down Kit was employed to attach biotinylated RNA to streptavidin-coated magnetic beads, and the whole-cell lysates were allowed to incubate overnight at 4 °C with the biotinylated RNA. After the incubation period, the RNA samples were extracted for the assessment of hsa_circ_0001165 and miR-381-3p levels.

RNA immunoprecipitation (RIP) experiment

In this experiment, KYSE30 and KYSE150 cells (2.0 × 10^7) coated with magnetic beads both anti-IgG (1000 μg/ml, 5 μg, Proteintech, USA) or anti-Ago2 (2000 μg/ml, 5 μg, Proteintech, USA) antibodies Millipore's MagnaRIP Kit (Millipore, 17–700 Sigma-Aldrich, Germany). For miR-381-3p and hsa_circ_0001165 enrichment analyses, RNA was extracted overnight and purified.

Double luciferase reporting experiment

The promoter vector for both wild-type and mutant variants of hsa_circ_0001165, along with the 3' UTR of TNS3 (Genecreat, Wuhan, China), features binding sites for miR-381-3p located adjacent to the luciferase coding region. The experiments were executed 24 h after transfection utilizing the Dual-Luciferase Reporter Assay Kit sourced from Vazyme Biotech Co. Ltd, with normalization of Renilla luciferase activity being carried out in GraphPad Prism version 9.0 to calculate luciferase activity.

Nude mouse xenograft tumor model

Throughout the experiments, the Animal Ethics Committee of Naval Medical University approved the use of BALB/c nude mice aged 4–6 weeks purchased from Gempharmatech Co. Ltd. (Suzhou, China) and housed in the animal experiment facility of Changhai Hospital. Experimental and control group using ten mice were randomly assigned to establish xenograft tumor models for ESCC. Each mouse received subcutaneous injections approximately 5 × 10^6 cells of KYSE30 with sh-circ_0001165 or sh-NC. Tumor volumes were measured every five days over a period of 30 days, length × (width/2)^2 was used to as the formula to calculate tumor volume. An assessment of tumor size and weight was made after euthanasia of the mice on day 30.

Public data acquisition and statistical analysis

The GSE53622 dataset contains gene expression profiles from (ESCC) samples(https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE53622). The dataset includes both tumor and adjacent normal tissue samples, providing paired data for analysis. Quality control measures were implemented to ensure reliable data, including normalization of the microarray data and verification of sample integrity. GSE53622 consists of 119 paired samples, offering a substantial dataset for identifying differentially expressed genes and developing predictive models for ESCC prognosis.

GraphPad Prism 9.0 software was used to analyze the data, with results from repeated experiments expressed as means ± standard deviation (SD). Student's t-tests were used to evaluate differences between groups, and one-way ANOVAs were used for multiple group comparisons. Fisher's exact test emphasized the clinical significance of molecular expression, and while Pearson correlation coefficients were utilized to analyzed the correlation among hsa_circ_0001165, miR-381-3p, and TNS3 expressions. Statistical tests significant when P-values below 0.05.