Esophageal squamous cancer cell line (KYSE-150) was purchased from Procell Company (Wuhan, China). It has been identified by short tandem repeats (STR) and proved to be free from mycoplasma pollution. Cells were cultured in RPMI 1640 growth medium supplemented with 10% fetal bovine serum (FBS, Cell-Box, Australia) in a humidified atmosphere of 5% CO₂ at 37 °C. Cells were transfected with pcSLenti-EF1-EGFP-P2A-Puro-CMV-RNF180-3xFLAG-WPRE after seeding on dishes by Lipofectamine 3,000 (Thermo Fisher Scientific, United States). In order to check drug sensitivity, either TAXOL (a mitotic inhibitor, MedChemExpress, Cat. No. HY-B0015) or 5-fluorouracil (5-FU, thymidylate synthetase inhibitor, MCE, Cat. No. HY-90006) were used to treat cells. To investigate the effects of lipogenesis on chemosensitivity, we transfected pcDNA3.1-ACC1-3×Flag or pcDNA3.1-ACLY-3×Flag (HonorGene, China) into RNF180 transfectants.

A Cell Counting Kit-8 (CCK-8) kit (CT01B, Cellcook Biotech, Guangzhou, China) was employed to assess cell viability/cytotoxicity. Briefly, 3,500 cells/well were seeded into the 96-well plates, and cultured in medium for 0–48 h, then 10 μL of CCK-8 test kit were added for following 90-min incubation, and the absorbance was tested at 450 nm wavelength by a Multiskan™ FC Microplate Photometer (Thermo Fisher Scientific).

We used 7-amino-actinomycin (7-AAD) combined with phycoerythrin (PE)-labeled Annexin V (559,763, BD Pharmingen, United States) to detect the externalization of phosphatidylserine by flow cytometry (Beckman Coulter, Brea, CA, United States). In brief, 1 × 10⁵ cells were collected, washed with PBS, and pelleted. PE-labelled Annexin V (5 μL, final concentration: 1 ug/mL) and 7-AAD (5 μL, final concentration: 50 μg/mL) were added to cell suspension, mixed, incubated for 15 min, and examined by flow cytometry. Finally, the results were analyzed by Flowjo V10 software.

8.5 × 10⁵ cells/well were seeded into a 6-well culture plate. When cells reached 80% confluence, surface was throughly scraped off in line with pipette tips. Subsequently, the cells were rinsed thrice with phosphate buffered saline (PBS) to ensure the complete removal of any broken cells. Finally, cells were cultured with FBS-free RPM1640. Cell migration photos were taken at the same location at 0 h and 24 h.

Transwell chambers (3,422, Costar, Corning, United States) pre-coated with Matrigel^® Basement membrane matrix (356,234, Corning Life Sciences, MA, United States) were used for the cell invasion assay. 4.6 × 10⁵ cells were seeded in the upper layer of the chamber, and the medium containing 14% FBS was added as an inducer in the lower layer. After 48 h of incubation at 37°C, the chamber was washed and fixed with 4% paraformaldehyde (Beyotime, China). After crystal violet staining (Beyotime, China), cells were photographed under the microscope. For the migration experiment, except for adding Matrigel (Corning, United States) to the upper layer and incubating for 48 h, the steps are the same as those of the invasion experiment. ImageJ was used to analyze the data.

4,500 cells were seeded on a glass slide, cultured for 12 h, washed with PBS three times, fixed with 4% paraformaldehyde, and dyed with Nile red dye (MCE, United States) for 15 min, and subsequently 4′, 6′-diamidino-2-phenylindole (DAPI, C1006, Beyotime, Shanghai, China) for 5 min, and sealed with anti-fade mounting medium (MCE, United States). The average fluorescence intensity of cells was analyzed by ImageJ.

Minute nuclear or cytosolic proteasome enrichment kit (Cat. No.PN-040, Invent Biotechnologies, Plymouth, MN, United States) was used to extract proteasomes. Briefly, we collected and resuspended 5 × 10⁶ cells in 450 µL buffer A. After that, we immediately transferred the cell suspension to the filter cartridge. And then, we inverted the sample for a few times and centrifuged at 16,000 g for 30 s. After the filter was removed, we vortexed briefly and centrifuged the collection tube at 16,000 g for 30 min. The supernatant (400 µL) was transferred into a fresh 1.5 mL microfungi tube with 400 µL buffer B added. The sample was vortexed for 120 s and followed by 10-min incubation on ice. After centrifugation at 10,000 g for 10 min, we completely removed supernatant, and added 50–150 µL of Minute™ Denaturing Protein Solubilization Reagent (WA-009) to resuspend the precipitate for the following Western blot.

For Co-IP, cells were lysed in ice-cold RIPA reagent (Merck, United States) using a cell disruptor. Cell lysates (>1000 μg) were incubated with 500 μg the antibody against RNF180, ACC1 or ACLY, and then incubated with the Protein G agarose beads (37478S, Cell Signaling Technology, Shanghai, China). Beads were washed with RIPA reagent four times, collected, and mixed with the sample. After heating (95°C, 10 min), the protein samples were eluted and separated from the beads. The sample supernatant was used for following Western blot.

All the tissues and plasma were obtained from the First Affiliated Hospital of Jinzhou Medical University. All tissues were preserved in 10% neutral formalin, embedded in paraffin, sectioned at 4 μm, and stained by hematoxylin and eosin (HE). Under the microscope, the representative parts of solid tumors were selected. Using tissue microarray (TMA), 2 mm tissue cores were punched out from each donor block and transferred to recipient blocks with 48 tissue cores (Azumaya kin-1, Tokyo, Japan). Tissue and cDNA microarrays of esophageal cancer and normal mucosa was also purchased from Shanghai Outdo Biotech (Shanghai) and used for immunohistochemistry and RT-PCR respectively. Informed consent was obtained from the patients in writing. No treatment was taken in these EC patients before operation. The Ethical Committee of the First Affiliated Hospital of Jinzhou Medical University approved our protocol.

Cells and homogenized tissues were lysed in RIPA buffer and measured by Coomassie Brilliant Blue Protein Assay Kit (Biorad). After that, the protein was separated by electric field migration on SDS- polyacrylamide gel, and transferred to the Polyvinylidene fluoride (PVDF) membrane (Merck, United States). After being blocked by 4% milk for 1 h, it was incubated with the primary antibody (Supplementary Table S1) at 37°C for 2.5 h. The membrane was washed with Tris-Buffered Saline Tween-20 (TBST) for 13 min and incubated with the HRP (horseradish peroxidase)-labelled IgG antibody (Proteintech, United States) corresponding to the primary antibody. ECL detection reagent (Beyotime, China) was used to detect the membrane signal, and ImageJ was used to analyze the signal intensity. Here, GAPDH is glyceraldehyde 3-phosphate dehydrogenase in glycolysis and β-actin comprises along with microtubules, a major component of the cytoskeleton, both of which are regarded as internal control for cytosolic fraction. Lamin B is an important nuclear layer protein belonging to the nuclear fibronectin family, mainly involved in maintaining the structure and function of the nuclear membrane, which is considered as internal control for nuclear fraction. PSMC1 is a proteasome 26 S subunit, ATPase 1 and acts as an internal control for proteasomal fraction.

Four-μm-thick slides of TMA were heated at 65°C, deparaffinized with xylene, rehydrated with alcohol, and then retrieved in target retrieval solution (Beyotime, China). After that, slides were incubated in methanol containing 2.5% hydrogen peroxide to block endogenous peroxidase activity, and to stop nonspecific binding by TBST containing 4.5% bovine serum albumin for 18 min. TMA slides were incubated with anti-rabbit RNF180 antibody (Genetix, United States), washed with TBST, and incubated with HRP-labelled IgG antibody (DAKO, United States) for 1.5 h, and the binding sites were visualized with diaminobenzidine (DAB, yellow color). Sections were dehydrated by ethanol, cleared by xylene, and mounted by neutral balsam after counterstaining with Mayer’s hematoxylin.

We performed ELISA with RNF180 ELISA Kit (Jianglai, China) to determine the plasma RNF180 concentration. In short, 45 μL of standard and diluted serum samples were added to polystyrene microtiter plates containing anti-RNF180 antibody, and then 90 μL of detection antibody labeled with HRP was added to each well. After that, we discarded the liquid and wash the plate using washing buffer. Reaction substrate was dispensed into the plate and then stop buffer was added. The absorbance was detected at 450 nm wavelength.

Expression data of EC patients were obtained from NCBI GEO database (GSE45670; platform: Affymetrix-GPL570) and analyzed in R studio 4.1.0 software. Expression, methylation, differential, related genes and prognostic significance of RNF180 were analyzed with the xiantao platform (https://www.xiantaozi.com/), STRING (https://www. string-db. org/cgi/input?sessionId = bgxRenupC483&input_page_show_search = on), and UALCAN database (https://ualcan.path.uab.edu/analysis.html). The prognostic significance of RNF180 was also explored by Kaplan-Meier plotter (https://kmplot.com/analysis/index. php?p = serviceandcancer = pancancer_rase). We constructed PPI network for differential genes and screened out the top ten hub genes. GO-KEGG and GSEA analysis were used to predict signal pathways.

We conducted three independent experiments in triplicate, and the data are expressed as mean ± standard deviation. The means were differentiated using Mann-Whitney U. Spearman correlation analysis was used for rank data. Survival curves generated by Kaplan-Meier plots was compared using log-rank statistics. The model of Cox’s proportional hazards was used to perform multivariate analysis. Statistics were considered significant at p < 0.05.

There was a lower RNF180 mRNA expression in EC than in normal tissues according to xiantao (Figure 1A, p < 0.05), UALCAN (Figure 1B, p < 0.05) and GEO database (Figure 1C, p < 0.05). However, RNF180 mRNA expression was higher in N₁, N₂ than in N₀ patients, stage 2 and 3 than in stage 1 patients (Figure 1D, p < 0.05), and G₃ than in G₂ and G₁ patients (Figure 1D, p < 0.05) (Figure 1D, p < 0.05) by UALCAN. Based on UALCAN database, RNF180 mRNA expression was positively associated with a low overall survival rate, stratified by race and histological grading (Figure 1E, p < 0.05). Based on Kaplan-Meier plotter (Figure 1F), RNF180 mRNA expression was positively linked to a low overall survival rate of the stage-2 SCC patients and AD patients with a mutation burden high (p < 0.05). It was the same for the relapse-free survival of all, male, Asian, stage 2, and grade 2 SCC patients and all AD patients (p < 0.05).

RNF180 mRNA expression was negatively related to its promoter methylation at cg09839635, cg08521800, cg07850154, cg14591786, cg23008153, cg16485558, cg06776999, cg10372047 and cg17370,163 sites (Figure 2A, p < 0.05). RNF180 methylation was higher in EC than in normal tissues (Figure 2B, p < 0.05), higher in Caucasians than Asians patients (Figure 2B, p < 0.05), higher in stage 1 than in stage 2 patients (Figure 2B, p < 0.05), and higher in AD than in SCC (Figure 2B, p < 0.05).

Xiantao platform helped us find out which genes are differentially expressed between low and high levels of RNF180 mRNA in EC and built a volcano map based on this information (Figure 3A). Based on GSEA analyses, top signal pathways included olfactory transduction, focal adhesion, vascular smooth muscle contraction, calcium signal pathway, cell adhesion molecules (CAMs), muscle contraction, extracellular matrix (ECM) receptor interaction and collagen degradation (Figure 3B, p < 0.05). PPI pairs were identified using STRING (Figure 4A), and the top 10 hub genes were identified using cytoscape (Figure 4B). As shown by xiantao, FN1, IL1B, and COL1A1 were more expressed in EC than in normal samples (Figure 4C, p < 0.05). On the other hand, ALB showed the opposite pattern (Figure 4C, p < 0.05). FN1, IL1B, and COL1A1 contribute to the aggressiveness of EC by promoting invasion and metastasis (Fang et al., 2019; Ma et al., 2023; Chen et al., 2024). FN1 enhances cell adhesion, migration, and extracellular matrix (ECM) remodeling, facilitating tumor cell invasion (Ma et al., 2023). IL1B drives chronic inflammation, activates pro-tumorigenic signaling pathways (e.g., NF-κB), and induces epithelial-mesenchymal transition (EMT) (Chen et al., 2024). COL1A1 stiffens the tumor stroma, creating a pro-invasive microenvironment and supporting cancer cell proliferation and invasion (Fang et al., 2019). These molecules collectively exacerbate ECM dysregulation, inflammation, and tumor-stromal interactions, accelerating EC progression.

Figure 5A shows the top positively-correlated genes of RNF180 in EC based on xiantao (p < 0.05). These genes were involved in gastric acid and insulin section, muscle and cardiomyopathy, glycoprotein binding, collagen and extracellular matrix (Figure 5B). Figure 5C shows the top negatively-correlated genes of RNF180 in EC (p < 0.05). These genes were involved in fat digestion and diabetes, PPAR signal pathway, and dipeptidase activity (Figure 5D). All the top 5 RNF180-related genes (TUBB4B, TMEM54, PHLDA2, NOP10, and ZNF593) were more expressed in EC than in normal tissues (Figure 5E, p < 0.05).

Using Western blot, we found that RNF180 expression was similar in normal tissues and EC (Figures 6A, P > 0.05). Plasma RNF180 level was lower in EC patients than in healthy volunteers after the adjustment of body surface area (BSA) (Figure 6B, p < 0.05). RNF180 protein was positively expressed in either nuclei or cytosol of esophageal squamous epithelial cells and cancer cells by IHC (Figure 6C). The positive rate of RNF180 expression was 64.0% (215/336) in esophageal normal tissues and 58.4% (208/356) in EC with no statistical significance (Table 1, p > 0.05). There was a positive correlation between RNF180 expression and TNM stage, N stage, and T stage of EC (Table 2, p < 0.05). On univariate analysis, sex, T stage, N stage, TNM stage, and RNF180 expression were positively correlated with unfavorable overall survival of EC patients (Figure 6D; Table 3, p < 0.05). In multivariate analysis, the expression of RNF180, TNM stage, and sex were found to be independent factors in EC patients (Table 3, p < 0.05).

After transfected with RNF180-expressing plasmid, KYSE-150 had the overexpression of RNF180 protein, evidenced by Western blot (Figure 7A). RNF180-overexpressing cells showed decreased proliferative capacity, compared to parental cells (Figure 7B, p < 0.05). Overexpression of RNF180 increased the chemosensitivity to 5-FU and TAXOL (Figure 7C). After ectopic RNF180 overexpression, high levels of apoptosis were observed in KYSE-150 cells (Figure 7D, p < 0.05). In wound healing (Figure 7E, p < 0.05) and transwell assays (Figure 7F, p < 0.05), RNF180 overexpression decreased the migration and invasion capacity of KYSE-150 cell. RNF180 expression caused less lipid droplets formation than parental cells by Nile red staining (Figure 7G, p < 0.05). In Figure 7H, we see that RNF180 overexpression decreased the expression of PI3K, Akt, Bcl-2, N-cadherin, Snail, Slug, MMP-2, MMP-9, ACC1, ACLY, ADRP, ACAT1, CIDEB, CIDEA, but increased the expression of Bax, Caspase-1, Gasdermin D, IL-18, IL-1β and E-cadherin.

In order to confirm the impact of ACC1 and ACLY on RNF180-induced drug resistance to chemotherapy and lipid droplets formation, we increased their expression levels in KYSE-150 cells, evidenced by Western blot (Figure 8A). According to CCK-8 (Figure 8B, p < 0.05) and Nile red staining (Figure 8C, p < 0.05), either ACC1 or ACLY increased the chemoresistance against 5-FU and TAXOL as well as lipid droplets formation in KYSE-150 cells.

After cycloheximide treatment, ACLY and ACC1 expression were low in RNF180 transfectants, compared to KYSE-150 cells (Figure 9A, p < 0.05). MG132 strengthened the expression of ACC1 and ACLY although their expressions were lower in KYSE-150 cells than in RNF180 transfectants (Figure 9B, p < 0.05). RNF180 transfectants had higher levels of ACC1 and ACLY protein in their nuclear proteasomes than KYSE-150 cells, which was strengthened by MG132 treatment (Figure 9C, p < 0.05). In cytosol proteasome, the expressions of ACC1 and ACLY were similar between RNF180 transfectants and their parental cells regardless of the treatment with MG132 (Figure 9C, p > 0.05). Ubiquitylated ACC1 and ACLY appeared more abundant in RNF180 transfectants than in KYSE-150 cells although MG132 reduced the levels in KYSE-150 cells and RNF180 transfectants according to Co-IP (Figure 9D, p < 0.05). ACC1 and ACLY bound more to RNF180 in RNF180 transfectants than KYSE-150 cells although MG132 reduced the levels in KYSE-150 cells and RNF180 transfectants according to Co-IP (Figure 9D, p < 0.05).