Bioinformatic analysis

Microarray expression data (GSE107943) were retrieved from the Gene Expression Omnibus. Clinical information and bulk RNA-sequencing data for CCA were obtained from The Cancer Genome Atlas. Differential expression analysis was performed using the limma R package (Ritchie et al. 2015). Pearson correlation coefficients were calculated to assess the relationship between NSUN5 and GLS expression levels. A genetic association analysis of NSUN5 and GLS was conducted utilizing data from the FinnGen study (Kurki et al. 2023).

Human samples, cell culture and transfection

A total of four intrahepatic CCA tissues and four paracancerous tissues were collected from Shanghai Electric Power Hospital. Human samples were obtained with informed consent, and experiments involving human specimens were approved by our institution's research ethics committee.

Human intrahepatic CCA cells (HCCC-9810) were sourced from Procell Biotechnology Co., Ltd. (Wuhan, China) and cultured in DMEM supplemented with 10% fetal bovine serum (v/v), penicillin G (100 units/mL), and streptomycin (100 μg/mL). The cell line was identified by short tandem repeat analysis (Supplemental file 1). Lentiviral particles containing NSUN5 overexpression (OE), NSUN5 knockout (KO), GLS knockout (KO), and double mutants of NSUN5 were procured from Genechem (Shanghai, China). Transfections were carried out according to the manufacturer's protocol. In brief, cells were seeded at a density of 1 × 105 per well in a 24-well plate. Before transfection with lentivirus, the cell number was adjusted to approximately 2 × 105 per well. The frozen virus and its corresponding empty vector control were thawed on ice. The amount of virus was calculated based on the MOI value, and the virus solution along with Polybrene (5 µg/mL) was added to the culture medium. After mixing well, the culture system was incubated for 12 h. The cell status was observed, and if there was no significant change, the medium was replaced with fresh medium after 12 h. After 72 h of transfection, antibiotics were added to the culture medium for cell selection. Subsequently, the cells were observed regularly and the medium was changed until the cells with stable viral transfection were selected. For the transfection of NSUN5, puromycin (5 µg/mL) was used for selection. For the transfection of GLS, neomycin (500 µg/mL) was used for selection.

Polymerase reaction (PCR) and Western blot

Cells were lysed using TRIzol reagent, and PCR was conducted to assess mRNA expression levels as previously described (Li et al. 2024). GAPDH served as a control, and relative mRNA expression levels were quantified using the 2–ΔΔCt method. Primer sequences are provided in Supplemental Table 1.

RNA immunoprecipitation assays were performed following the manufacturer's instructions using the RNA Immunoprecipitation Kit (BersinBio, China). Briefly, 1 × 107 cells were collected and lysed in 0.9 mL polysome lysis buffer supplemented with protease inhibitor and RNase inhibitor. The lysates were treated with DNase to remove DNA contamination. For immunoprecipitation, the lysates were incubated with antibodies (1: 200) overnight at 4 °C using a vertical rotator mixer at 10 rpm to ensure thorough mixing. Protein A/G magnetic beads (20 µL) were then added to capture the antibody-RNA–protein complexes, and the mixture was incubated for an additional one hour at 4 °C with gentle rotation using the same vertical rotator mixer. The beads were washed sequentially with polysome washing buffer 1 and polysome washing buffer 2 to remove nonspecific binding. The RNA–protein complexes were eluted using polysome elution buffer. The immunoprecipitated RNA was purified by phenol–chloroform extraction and ethanol precipitation. The purified RNA was subsequently analyzed by PCR for downstream RNA detection.

For Western blot analysis, cells were lysed in a buffer containing proteinase and phosphatase inhibitors. Cell lysates were subjected to SDS-PAGE, transferred onto membranes, and incubated with specific antibodies for protein visualization. GAPDH served as a control, and the antibodies used are detailed in Supplemental Table 2.

RNA pull-down assays were conducted using the RNA pulldown Kit (BersinBio, China). Biotin-labeled mRNA or a random oligo probe was first denatured at 90 °C for 2 min and then cooled on ice for 2 min to form secondary structures. The probe was then incubated with 40 µL of streptavidin magnetic beads in a total volume of 300 µL containing 2 × TES buffer at room temperature for 30 min with gentle rotation. The cell lysates, prepared in RIP buffer supplemented with protease inhibitors and DNase, were incubated with the probe-beads complex at 4 °C overnight to allow RNA-associated proteins to bind to the RNA. After three washes with NT2 buffer at 4 °C, the RNA–protein complexes were eluted from the beads using protein elution buffer containing DTT at 37 °C for 2 h with intermittent mixing. The eluted complexes were analyzed by Western blot.

CCK-8 and flow cytometry

In CCK-8 assays, cells were seeded at a density of 1 × 102 cells per well in a 96-well plate. Subsequently, 10 µl of CCK-8 reagent was added at specific time points, and the optical density (OD) was measured at 450 nm after a one-hour incubation at 37 °C.

For apoptosis detection, cell suspensions were prepared in phosphate-buffered saline, and the Annexin V-FITC Apoptosis Detection Kit (Vazyme, China) along with flow cytometry were employed. Cells located in the right quadrants (Annexin V positive) were classified as apoptotic.

Cell migration and invasion assay

For the migration assay, 1 × 104 cells were seeded in the upper chamber and incubated with serum-free medium, while culture medium containing 10% FBS was added to the basolateral chamber. After 24 h, cells were fixed with 4% paraformaldehyde for 10 min, rinsed with PBS, and stained with 0.1% crystal violet for 10 min. Stained cells were photographed in three random visual fields and analyzed using ImageJ software.

For invasion assays, Matrigel from Corning (USA) was thawed at 4 °C overnight. A 100 µL Matrigel solution diluted in serum-free medium was applied to the upper chamber and allowed to solidify at 37 °C for 30 min. Subsequently, 1 × 104 cells were placed in the upper chamber for the invasion assay.

m5C quantification and Immunofluorescence

The global m5C level was assessed using the MethylFlash™ 5-mC RNA Methylation ELISA Easy Kit (Fluorometric) (EpiGentek, USA). In each well, 100 µl of binding solution and 200 ng of total RNA sample were added, followed by incubation at 37 °C for 90 min to allow RNA binding. Subsequently, 50 µl of m5C Detection Complex Solution, containing the m5C antibody, was added to each well after washing. After a 50-min incubation at room temperature, the diluted m5C antibody was removed. The wells were then incubated with Fluorescence Development Solution at room temperature for 2–4 min, away from direct light. The fluorescence signal was read by a microplate reader set at 530ex/590em nm within 2 to 10 min.

For immunofluorescence, cells were washed with phosphate-buffered saline and fixed with 4% paraformaldehyde for 10 min at room temperature. Subsequently, cells were permeabilized with 0.5% Triton X-100 for 20 min at room temperature, blocked, and incubated with the primary antibody overnight at 4 °C. Following this, cells underwent a washing step and were then incubated with a fluorochrome-conjugated secondary antibody for 1 h at room temperature, followed by DAPI staining. Image acquisition was performed using confocal microscopy.

RNA stability assays and molecular docking

For RNA stability assays, cells were plated in 6-well dishes and exposed to actinomycin D (5 μg/mL) (Sigma-Aldrich, USA) for 1, 3, and 6 h. In the negative control group, DMSO was added at 0 h. Subsequently, total RNA was extracted and analyzed via PCR.

The sequence of human NSUN5 was retrieved from the UniProt database and modeled through AlphaFold 3. The mRNA sequence of human GLS was obtained from the National Center for Biotechnology Information. The transcript analyzed in this study is NM_014905, with a length of 4840 bp. AlphaFold 3 was used to assess the interaction between NSUN5 and GLS mRNA (Abramson et al. 2024).

Xenograft tumor model in nude mice

Four-week-old male BALB/c nude mice were procured from Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). These animals were housed in a temperature-controlled environment maintained at 23–25 °C with a 12-h light–dark cycle, and they had ad libitum access to both water and food. The nude mice were subcutaneously inoculated with 1 × 105 stably transfected HCCC-9810 cells or control cells. Two weeks post-inoculation, all mice were euthanized, and the xenografts were excised and weighed. Tumor volume was determined using the following formula: Volume (mm3) = length × width2 × 0.5. The animal experimentation was conducted with the approval of the research ethics committee of our institution (Approval No. 2024–008).

Statistical analysis

All analyses were conducted using R software (version 4.4.0). Data are presented as mean values ± standard deviation from three or more independent experiments. Group differences were evaluated using either a two-tailed Student’s t-test or a Mann–Whitney U test. Statistical significance was defined as P < 0.05.