Gene expression profiles and associated clinical information for patients with CRC in The Cancer Genome Atlas (TCGA) cohort were obtained from the UCSC Xena database (https://xena.ucsc.edu/, accessed in March 2024). The data, with 332 primary CRC cases, were screened and samples without follow-up information, with 0-day survival or representing duplicate sequencing from the same patient were removed, which resulted in a final cohort of 282 primary CRC cases for the training dataset. The gene expression dataset GSE40967 (575 CRC samples and 10 normal tissue samples) was downloaded from the Gene Expression Omnibus (GEO) repository (https://ncbi.nlm.nih.gov/geo/) (30); this dataset served as the external validation set. Samples lacking survival data in the GSE40967 dataset were excluded from further analysis. To minimise technical variability between datasets, batch effects were corrected using the R ‘limma’ package (version 3.50.3) (31). Specifically, the ‘removeBatchEffect’ function was applied after merging tumour samples from the validation cohort in the UCSC Xena data with GSE40967, which were generated using the same microarray platform [GPL570 (HG-U133_Plus_2); Human Genome U133 Plus 2.0 Array; Affymetrix; Thermo Fisher Scientific, Inc.]. Subsequently, mRNA expression data from both datasets were normalised to ensure consistency for downstream analysis. The demographics and clinicopathological characteristics of these patients are displayed in Table I.

Pan-cancer functional analyses of DCAF13 were performed using the Cancer Single-Cell State TCGA (CancerSEA; http://biocc.hrbmu.edu.cn/CancerSEA/) database (32). Expression levels of DCAF13 across different types of cancer were explored using the Tumor Immune Estimation Resource (TIMER; cistrome.shinyapps.io/timer/) (33), which was used to compare DCAF13 expression levels between cancer tissues and matched normal samples.

For prognostic analysis, patients in the data from TCGA (https://www.cancer.gov/ccg/research/genome-sequencing/tcga) were stratified into high- and low-expression groups based on the median expression levels of DCAF13, which served as the optimal cut-off value. The median DCAF13 expression was 11.98325, with patients exhibiting expression levels ≥11.98325 classified into the high-expression group, and those <11.98325 into the low-expression group. The association between DCAF13 expression and overall survival (OS) was evaluated using Kaplan-Meier (KM) survival analysis via the Kaplan-Meier Plotter online tool (http://gepia.cancer-pku.cn/), using the CRC mRNA dataset. Patients were stratified into high- and low-expression groups based on the median expression of DCAF13, and OS was analysed using the log-rank test. The findings were validated using the GSE40967 dataset by performing the same survival analysis on this external cohort, with patients stratified into high- and low-expression groups based on the median DCAF13 expression level within the GSE40967 cohort.

To assess immune cell infiltration and immune checkpoint activity, several computational tools and bioinformatics algorithms were used. The ESTIMATE algorithm (https://bioinformatics.mdanderson.org/estimate/) was used to evaluate the stromal (StromalScore, SS), immune (ImmuneScore, IS), and overall ESTIMATE scores for each tumour sample in the data from TCGA (34). These scores were derived from gene expression data using the ESTIMATE R package (version 1.0.13). The relationships between DCAF13 expression and the SS, IS, and ESTIMATE scores were further analysed to gain insights into its role in immune regulation within the TME.

For further assessment of immune infiltration, the CIBERSORT software (https://sangerbox.com; accessed in March 2024) was used to estimate the relative proportions of 22 immune cell types in each tumour sample. This was achieved through the analysis of a normalised gene expression matrix (35). Differences in immune infiltration between the high- and low-DCAF13 expression groups were compared using the Wilcoxon rank-sum test (two-sided).

To explore the role of DCAF13 in regulating cancer immunity, its correlation with various immune checkpoints, including B and T lymphocyte attenuator (BTLA), lymphocyte-activating 3 (LAG-3), signal regulatory protein α (SIRPα) and V-domain Ig suppressor of T-cell activation (VISTA) was visualised using Gene Expression Profiling Interactive Analysis (GEPIA; http://gepia.cancer-pku.cn/) (36). Additionally, the correlation between DCAF13 expression and 122 immune checkpoint-related genes (ICPs) was analysed using Pearson's correlation analysis (37).

To predict the response to immunotherapy, Tumor Immune Dysfunction and Exclusion (TIDE; https://tide.dfci.harvard.edu/; accessed in March 2024) analysis was applied to the high- and low-DCAF13 expression groups (38). In the present study, the TIDE score, dysfunction score, exclusion score and MSI score between the high- and low-DCAF13 expression groups were calculated to evaluate their potential responsiveness to immunotherapy.

In the present study, a total of 234 paraffin-embedded CRC tissue samples and their matched adjacent non-tumorous tissues, collected ≥5 cm away from the tumour margin, were obtained from patients who underwent surgical resection at the Department of Pathology, Affiliated Hospital of Inner Mongolia University (Hohhot, China) between January 2013 and June 2015. All samples were histologically confirmed as CRC and were obtained prior to any neoadjuvant therapy. Among the 234 patients, 146 were male and 88 were female. The age distribution showed 83 patients were aged ≤60 years and 151 patients were aged >60 years. The cohort included tumours originating from both the colon and rectum, including the ascending (n=76), transverse (n=31), descending/sigmoid (n=68) colon and rectum (n=59). Tumour histology was classified according to WHO criteria, with the majority being tubular or papillary adenocarcinomas (n=208), and the rest categorised as other types (n=26) (39). Tumour differentiation included well/moderately differentiated (n=141) and poorly differentiated (n=93) tumours. Staging was based on the AJCC 8th edition TNM classification, comprising stage I (n=26), stage II (n=62) and stage III/IV (n=146) cases (40). Additional T and N staging details were also recorded. The present study was approved by the Ethical Review Committee of Inner Mongolia Medical University (approval no. 2022513411; Hohhot, China) and was performed in accordance with the Declaration of Helsinki. Informed consent was written from all patients prior to inclusion in the study. Detailed clinicopathological characteristics are summarised in Table II.

All tissue samples were fixed in 10% neutral-buffered formalin (no. G2161; Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 24 h, routinely dehydrated, embedded in paraffin and stored under standard conditions. No tissues were frozen after paraffin embedding. Prior to use, the paraffin sections of the 234 CRC samples and their adjacent tissues were prepared at a thickness of 4 µm, deparaffinised in xylene, rehydrated through a graded ethanol series and washed with PBS (0.01 M, pH 7.0) (41). Endogenous peroxidase activity was quenched by incubation with 3% hydrogen peroxide (cat. no. H792073; Shanghai Macklin Biochemical Co., Ltd.) at room temperature for 10 min. Antigen retrieval was performed by boiling the slides in citrate buffer (0.01 M, pH 6.0) at 95–100°C for 15 min, followed by natural cooling to room temperature. For blocking of non-specific binding, the sections were incubated with 5% goat serum (cat. no. SL039; Beijing Solarbio Science & Technology Co., Ltd.) at 37°C for 30 min. The sections were then incubated at 4°C overnight with a rabbit anti-DCAF13 antibody (1:500; cat. no. bs-62879R; BIOSS). After washing with PBS, the sections were incubated with a goat anti-rabbit HRP secondary antibody (1:2,000; cat. no. ZDR-5306; Beijing Zhongshan Jinqiao Biotechnology Co., Ltd.) at 37°C for 1 h. For visualisation, the sections were incubated with DAB chromogen (cat. no. DA1010, Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 5 min, followed by counterstaining with haematoxylin solution at room temperature for 10 sec. Immunostained sections were independently evaluated by two experienced pathologists under blinded conditions using a light microscope (Olympus BX53; Olympus Corporation). The staining intensity and percentage of DCAF13-positive cells were quantitatively analysed using ImageJ software (version 1.53t; National Institutes of Health).

FHC, Caco2, SW480, HCT116, Lovo and RKO cells were used in the present study and cultured in a humidified atmosphere at 37°C and 5% CO₂. FHC cells, a normal intestinal cell line, were purchased from Zhejiang Meisen Cell Technology Co., Ltd. (cat. no. CTCC-001-0208) and cultured in their respective commercial medium (cat. no. CTCC-002-047; Zhejiang Meisen Cell Technology Co., Ltd). Caco2 cells were cultured in MEM medium (cat. no. PM150410; Procell Life Science & Technology Co., Ltd.) supplemented with 20% foetal bovine serum (FBS; cat. no. C04001; Shanghai VivaCell Biosciences, Ltd.) and 100 IU/ml penicillin/streptomycin (P/S; cat. no. 15070063; Thermo Fisher Scientific, Inc.). SW480, Lovo and RKO cells were cultured in DMEM medium (cat. no. PM150210; Procell Life Science & Technology Co., Ltd.) with 10% FBS and 100 IU/ml P/S. HCT116 cells were maintained in IMEM medium (cat. no. C12440500BT; Gibco; Thermo Fisher Scientific, Inc.) supplemented with 10% FBS and 100 IU/ml P/S.

To identify the CRC cells with the highest expression levels of DCFA13, protein lysates from CRC cells were extracted using RIPA lysis buffer (cat. no. DE101; TransGen Biotech Co., Ltd.) and quantified using a bicinchoninic acid assay kit (cat. no. PC0020; Beijing Solarbio Science & Technology Co., Ltd.) (42,43). Proteins (30 µg per lane) were resolved by 10% SDS-PAGE (cat. no. P1200; Beijing Solarbio Science & Technology Co., Ltd.) and transferred to PVDF membranes (cat. no. PVH00010; Merck KGaA) using 1× Tris-glycine buffer (cat. no. G2145; Wuhan Servicebio Technology Co., Ltd.). After blocking with 10% non-fat milk at room temperature for 1 h, the membranes were incubated at 4°C overnight with a rabbit anti-DCAF13 antibody (1:2,000; cat. no. bs-62879R; BIOSS), a rabbit anti-γ-H2AX antibody (1:2,000; cat. no. ab81299; Abcam), a rabbit anti-ATM antibody (1:2,000; cat. no. bsm-52360R; BIOSS), a rabbit anti-E-cadherin antibody (1:2,000; cat. no. AF0131; Affinity Biosciences), a rabbit anti-N-cadherin antibody (1:2,000; cat. no. AF6710; Affinity Biosciences), a rabbit anti-Vimentin antibody (1:2,000; cat. no. 60330-1-Ig; Proteintech Group, Inc.), a rabbit anti-PCNA antibody (1:2,000; cat. no. 10205-2-AP; Proteintech Group, Inc.) or a mouse anti-actin antibody (1:2,000; cat. no. CW0096M; Jiangsu CoWin Biotech Co., Ltd.). After three washes with Tris-HCl with 0.1% Tween-20, the membranes were incubated with goat anti-rabbit IgG H&L HPR antibodies (1:2,000; cat. no. S0001; Affinity Biosciences) or HRP-conjugated Affinipure goat anti-mouse IgG (1:2,000; cat. no. SA00001-1; Proteintech Group, Inc.) at room temperature for 1 h. Immunoblots were visualised using enhanced chemiluminescence (cat. no. W1001; Promega Corporation) using a ChampChemi system (Shanghai Champ Biomedical Technology Co., Ltd.). The densitometry analysis of the protein bands was performed using ImageJ software (version 1.53t; National Institutes of Health).

Lentiviral vectors expressing either a short hairpin RNA (shRNA) targeting the DCAF13 coding region (5′-CGAATCTTTCCTGTAGACAAA-3′) or a non-targeting control shRNA sequence (5′-TTCTCCGAACGTGTCACGT-3′), which does not match any known human gene, were supplied by Shanghai GeneChem Co., Ltd. The lentiviral plasmid containing the shRNA was packaged using a second-generation system, with 293T cells (American Type Culture Collection) used as the interim cell line. For transfection, 10 µg lentiviral plasmid was mixed with 10 µg lentivirus packaging plasmid (pMD2.G) and 10 µg packaging plasmid (psPAX2) in a 1:1:1 ratio. The mixture was transfected into 293T cells using Lipofectamine 3000 (cat. no. L3000008; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Transfection was carried out at 37°C for 6 h, and the medium was replaced with fresh growth medium after 6 h. Lentiviral particles were collected 48 h post-transfection, and the multiplicity of infection was adjusted to 5 for infection of RKO cells with high baseline expression of DCAF13. The cells were transduced in the presence of polybrene (8 µg/ml; cat. no. H9268; Sigma-Aldrich; Merck KGaA) for 12 h. Following transduction, cells were selected with puromycin (1 µg/ml; cat. no. 314Y0415; Beijing Solarbio Science & Technology Co., Ltd.) for 14 days to establish stable cell lines. Stable DCAF13 knockdown (shDCAF13) and the non-targeting control (shNC) cell lines were validated using RT-qPCR and western blotting.

RNA extraction and qPCR were performed to assess the efficiency of DCAF13 knockdown. Total RNA was extracted from RKO cells in both experimental groups using TRIzol^® reagent (cat. no. 79306; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Reverse transcription of 1 µg of RNA was performed using a PrimeScript™ RT Reagent Kit (cat. no. RR047A; Takara Biotechnology Co., Ltd.), according to the manufacturer's protocol, to synthesise cDNA. For amplification, specific forward and reverse primers for DCAF13 were designed using PrimerBank (version 2.1; http://pga.mgh.harvard.edu/primerbank/) and synthesised by Sango Biotech Co., Ltd. qPCR was performed using the SYBR^® Premix Ex Taq™ Kit (cat. no. RR820A; Takara Biotechnology Co., Ltd.) in a 20-µl reaction volume, using a Roche LightCycler 480 II System (Roche Diagnostics). The following thermocycling conditions were used for qPCR: Initial denaturation at 94°C for 2 min; 40 cycles of denaturation at 95°C for 5 sec; annealing at 60°C for 15 sec; and extension at 72°C for 10 sec. Fluorescence signals were detected during the annealing-extension phase, and GAPDH was used as the internal reference for normalisation. The relative expression levels of DCAF13 in each group were calculated using the 2^−ΔΔCq method (44). The following primer pairs were used for qPCR: DCAF13 forward (F), 5′-ACTGCACAGCTAAAGAACCG-3′ and reverse (R), 5′-TCCCAGACTACTTCCAGTCAC-3′; and GAPDH F, 5′-TGAACGGGAAGCTCACTG-3′ and R, 5′-GCTTCACCACCTTCTTGATG-3′.

shNC and shDCAF13 RKO cells (2×10³ cells) were seeded into 6-well plates and cultured at 37°C in DMEM medium supplemented with 10% FBS and 100 IU/ml P/S overnight. A wound was created by scratching the monolayer with a 200 µl pipette tip, and the cells were cultured in DMEM medium without FBS or P/S for an additional 24 h. The degree of wound closure was observed and microscopically (Eclipse Ts2 inverted phase contrast microscope; Nikon Corporation) imaged with the cell migration rate calculated using the following formula: Migration rate (%)=[(initial wound width-wound width at 24 h)/initial wound width] × 100% (42,45).

shNC and shDCAF13 RKO cells (2×10³ cells) were seeded into the upper chamber of a Transwell migration assay plate (cat. no. CLS3422; Corning, Inc.) without Matrigel precoating. The medium in the upper chamber was MEM medium (cat. no. PM150410; Procell system) without serum. In the lower chamber, DMEM medium (cat. no. PM150210; Procell system) was supplemented with 20% FBS to attract migrating cells. After 24 h incubation at 37°C, non-migrating cells on the upper side of the membrane were removed with a cotton swab, and the migrating cells on the lower membrane were fixed in 4% methanol (cat. no. P1110; Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 15 min and stained with 0.1% crystal violet (cat. no. C0121; Beyotime Biotechnology) at room temperature for 15 min. Cells were examined using an inverted phase contrast microscope (Eclipse Ts2; Nikon Corporation) at 400× magnification. A total of four fields of view per chamber were randomly selected for quantification (42,45).

shNC and shDCAF13 RKO cells (1×10³ cells) were plated into 6-well plates and cultured for 7 days at 37°C in DMEM medium supplemented with 10% FBS and 100 IU/ml P/S. The media was changed every 2 days. After 7 days incubation, cells were fixed with 4% paraformaldehyde (cat. no. P1110; Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 15 min, colonies were stained with 0.1% crystal violet at room temperature for 15 min. Colonies, defined as clusters of >50 cells, were examined using an inverted phase contrast microscope (Eclipse Ts2; Nikon Corporation) and counted using the Image J software (version 1.53t; National Institutes of Health) (42,45).

shNC and shDCAF13 RKO cells (5×10³ cells) were plated in 10 µg/ml fibronectin-coated (precoated at 37°C for 1 h) 96-well plates. After 30 min plating at 37°C, the wells were washed with PBS to remove non-adherent cells. Adherent cells were fixed with 100% methanol at room temperature for 15 min and stained with 0.1% crystal violet at room temperature for 15 min. Adherent cells were examined using an inverted phase contrast microscope (Eclipse Ts2; Nikon Corporation) and counted using the Image J software (version 1.53t) (42,45).

shNC and shDCAF13 RKO cells were fixed overnight in 70% ethanol at 20°C. Following the manufacturer's protocol, cells were then stained with the cell cycle analysis working solution (cat. no. CA1510; Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 1 h. The staining solution contained propidium iodide (PI; cat. no. C0121; Beyotime Biotechnology) for DNA content analysis, and RNase A (cat. no. C1608; Beyotime Biotechnology) was used to degrade RNA prior to PI staining. Intracellular permeabilization was not required in this assay, as the fixation process with ethanol preserved the cell membranes. Cell cycle analysis was performed using flow cytometry with a FACScan flow cytometer (BD Biosciences) equipped with a 488 nm laser for excitation. PI fluorescence was detected at 620 nm (FL3 channel). Data were analysed using FlowJo software (version 10.8.1; BD Biosciences) (42,45).

For transcriptomic analysis, total RNA was extracted from RKO cells using TRIzol^® reagent (cat. no. 79306; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions (46). RNA quality and integrity were assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc.), and only samples with an RNA integrity number (RIN) ≥7.0 were used for subsequent library construction. mRNA was enriched using Oligo(dT) beads, fragmented into short fragments, and reverse-transcribed into cDNA using the NEBNext^® Ultra™ RNA Library Prep Kit for Illumina^® (cat. no. 7530; New England BioLabs, Inc.). The resulting double-stranded cDNA fragments were end-repaired and ligated to Illumina sequencing adapters, followed by purification using AMPure XP Beads (1.0×; Beckman Coulter, Inc.) and PCR amplification. The final libraries were quantified using a Qubit™ 4 Fluorometer (Thermo Fisher Scientific, Inc.), and molar concentrations were calculated. Libraries were diluted to a final loading concentration of 8 pM prior to sequencing. Sequencing was performed on an Illumina NovaSeq 6000 platform (Illumina, Inc.) using paired-end sequencing with a read length of 150 bp (PE150). Raw sequencing data were pre-processed using fastp (version 0.18.0; http://github.com/OpenGene/fastp) to remove adapter sequences and low-quality reads. Clean reads were aligned to the human reference genome (GRCh38) using HISAT2 (version 2.4; http://daehwankimlab.github.io/hisat2/). Transcript abundance was quantified using RSEM (version 1.3.3; http://deweylab.github.io/RSEM/). Differentially expressed genes (DEGs) were identified using DESeq2 (version 1.40.2; http://bioconductor.org/packages/DESeq2/), with a false discovery rate (FDR)-adjusted P value <0.05 and an absolute log₂ fold change ≥1 considered statistically significant. Functional enrichment analyses were performed using DAVID Bioinformatics Resources (version 6.8; http://david.ncifcrf.gov/). In addition, gene set enrichment analysis (GSEA) was conducted using GSEA software (version 4.3.2; Broad Institute; http://www.gsea-msigdb.org/gsea/) to identify significantly enriched biological processes and signalling pathways associated with DCAF13 regulation.

Statistical analyses were performed using SPSS (version 19.0; IBM Corp.) and R (version 4.3.1; Posit Software, PBC). Pearson's χ² test was used to examine the associations between categorical variables, including DCAF13 expression status (high vs. low) and clinicopathological parameters. Pearson's correlation coefficient (r) was used to assess correlations between continuous variables, including DCAF13 expression levels and immune-related gene expression or immune infiltration scores. Differences between two independent groups were analysed using the Wilcoxon rank-sum test (Mann-Whitney U test). Kaplan-Meier survival curves were generated and compared using the log-rank test. Variables with prognostic significance in univariate analysis were further analysed using multivariate Cox proportional hazards regression. A two-sided P<0.05 was considered to indicate a statistically significant difference.

DCAF13 expression was negatively correlated with angiogenesis (r=−0.29; P<0.05) and differentiation (r=−0.30; P<0.05) signatures in CRC samples (Fig. 1A). TIMER analysis indicated that DCAF13 expression levels were significantly elevated in CRC tissues compared with that of normal colorectal samples. Additionally, compared with adjacent normal tissues, DCAF13 mRNA expression levels were increased in 18 other types of cancer (Fig. 1B).

DCAF13 was found to be significantly upregulated in CRC tumour tissue compared with that in the paired adjacent normal tissue, in both the TCGA (P<0.005; Fig. 1C) and GSE40967 validation datasets (P<0.005; Fig. 1D). KM analysis of TCGA data showed that patients with CRC with high expression levels of DCAF13 had a shorter OS [hazard ratio (HR)=1.63; P<0.05; Fig. 1E] compared with that of the low expression group. Similarly, KM analysis of the GSE40967 validation dataset confirmed that patients with CRC with high expression levels of DCAF13 had a shorter OS compared with that of the low expression group (HR=1.91; P<0.005; Fig. 1F).

DCAF13 expression exhibited a significant negative association with both SS (P<0.05; Fig. 2A) and IS (P<0.005; Fig. 2B) in TCGA-COAD cohort after excluding samples with missing transcriptomic or ESTIMATE score data. Immune infiltration analysis using the CIBERSORT algorithm demonstrated the proportions of the 22 immune cell types in the TME of CRC in TCGA cohort (Fig. 2C). Comparison of immune cell infiltration between the high- and low-DCAF13 expression groups showed that macrophages (M0 and M1), activated mast cells, neutrophils and activated CD4⁺ T cells were significantly more prevalent in the high-DCAF13 expression group. Conversely, the infiltration levels of naive B cells, resting dendritic cells, monocytes, plasma cells, resting CD4⁺ T cells and CD8⁺ T cells were significantly lower in the high-DCAF13 expression group (Fig. 2D).

Using the GEPIA database, the correlation between DCAF13 and various immune checkpoints was evaluated. The results indicated a potential association between DCAF13 and immune checkpoints, including BTLA (P<0.05; Fig. 3A), LAG-3 (P<0.05; Fig. 3B), SIRPα (P<0.05; Fig. 3C) and VISTA (P<0.05; Fig. 3D). This association with immune checkpoints led to further investigation into the relationship between DCAF13 and immunomodulators. ICP analysis demonstrated that DCAF13 expression was significantly correlated with a large proportion of immunomodulators in CRC (Fig. 3E).

To predict the potential response of patients with CRC to immunotherapy, the TIDE algorithm was applied to compare predicted immunotherapy responses between patients stratified into high- and low-DCAF13 expression groups. The analysis demonstrated significantly higher TIDE scores in the high-risk group, suggesting that patients with high DCAF13 expression may have a worse response to immunotherapy compared with those in the low-risk group (P<0.05; Fig. 4A). Notably, there was no significant difference in T-cell dysfunction scores between the high- and low-risk groups; however, T-cell exclusion scores were significantly increased in the high-risk group, which indicated a greater potential for immune evasion in high-risk patients (Fig. 4B and C). Additionally, the high-risk group exhibited significantly increased MSI scores, which suggested that the high-risk CRC group may have a higher prognostic potential and poor immunotherapy response (Fig. 4D).

DCAF13 expression was predominantly localised to the cytoplasm in CRC tissues, whereas no notably high positive staining was observed in normal colon tissue (Fig. 5A). Quantitative analysis using the χ² test demonstrated that high DCAF13 expression was detected in 58.55% (137/234) of CRC tissues, significantly higher compared with the 41.45% (97/234) of non-cancerous tissues (P<0.05; Table II). Further investigation into the association between DCAF13 expression and the clinicopathological features of patients with CRC is presented in Table II. Elevated DCAF13 expression was significantly associated with poorer differentiation and advanced TNM stage in patients with CRC. Specifically, patients with well or moderately differentiated tumours exhibited a higher proportion of high DCAF13 expression (63.8%) compared with those with poorly differentiated tumours (50.5%; P<0.05; Table II), indicating a negative association between DCAF13 expression and tumour differentiation. In terms of TNM stage, high DCAF13 expression was more frequent in advanced stages (65.8% in stage III + IV vs. 46.2% in stage I; P<0.05; Table II), suggesting a positive association between DCAF13 expression and disease progression. However, no significant associations were observed between DCAF13 expression and clinicopathological factors such as age, sex, tumour location or tumour size.

To evaluate the clinical relevance of DCAF13 expression, univariable analyses were performed to assess its relationship with various clinicopathological factors (Table III). High DCAF13 expression (HR=2.230; P<0.001) and advanced TNM stage (HR=1.464; P<0.05) were significantly associated with a worse OS in patients with CRC. Multivariable Cox proportional hazards regression analysis confirmed that high DCAF13 expression (HR=2.148; P<0.001) and TNM stage (HR=1.398; P<0.05) were independent prognostic factors for poor 5-year OS (Table III). KM survival analysis demonstrated that patients with CRC with high DCAF13 expression had a significantly shorter OS compared with those with low or no DCAF13 expression (Fig. 5B). Furthermore, survival analysis according to TNM stage revealed a clear prognostic stratification, with stage I patients showing the most favourable survival, stage II patients exhibiting intermediate outcomes and stage III/IV patients having the worst prognosis, and the differences among TNM stages were statistically significant (Fig. 5B). Additionally, western blotting analysis demonstrated that, compared with normal colon cells (FHC cells), DCAF13 expression was significantly elevated in CRC cell lines, particularly in RKO cells (Fig. 5C and D).

To confirm the pro-tumorigenic role of DCAF13 on CRC progression, lentivirus-mediated shRNA was used to knock down DCAF13 expression in RKO cells. RT-qPCR results confirmed that DCAF13 expression levels in the shDCAF13 were significantly reduced compared with control cells (Fig. 6A). As confirmed by western blotting analysis, DCAF13 expression levels also showed a similar reduction, validating the knockdown efficacy (Fig. 6B and C). Cell proliferation assays revealed that DCAF13 knockdown significantly impaired the proliferative potential of RKO cells (Fig. 6D). In parallel, the expression of the cell proliferation marker PCNA was significantly decreased in shDCAF13 cells, further confirming the reduction in proliferation (Fig. 6E and F). Migration assays, performed using scratch and Transwell assays, demonstrated that DCAF13 knockdown significantly abrogated the ability of RKO cells to migrate compared with that of the shNC group (Fig. 6G-J). Furthermore, the colony formation and adhesion assays showed significant reductions in the number of colonies and adherent cells following DCAF13 knockdown compared with that of the control group (Fig. 6K-N). Cell cycle distribution analysis revealed that DCAF13 knockdown induced a G₁ phase arrest, indicating a disruption in cell cycle progression (Fig. 6O and P). Corresponding western blotting analysis showed that, compared with shNC cells, knockdown of DCAF13 resulted in increased expression of the epithelial marker E-cadherin and decreased expression of the mesenchymal markers N-cadherin and vimentin in CRC cells (Fig. 6Q and R), further supporting the involvement of DCAF13 in regulating EMT progression in CRC cells.

To further uncover the pro-tumourigenic effect of DCAF13 on CRC progression, the transcriptome characteristics of DCAF13 knockdown RKO cells were analysed. The results confirmed that a total of 12,652 transcripts were mined in the RKO cells between the shNC and shDCAF13 groups. The volcano plot identified 423 DEGs between the shNC and shDCAF13 groups; this comprised 219 upregulated DEGs and 204 downregulated DEGs in the shDCAF13 group when compared with the shNC group (Fig. 7A). A distinct hierarchical clustering of genes in the shDCAF13 RKO cells compared with the shNC group was demonstrated in the heat map (Fig. 7B). Conventionally, the functional GO enrichment analyses were performed to determine the overall functional implications of these identified DEGs. In terms of biological processes (Fig. 7C), the identified DEGs were predominantly involved in the subcategories of ‘locomotion’, ‘growth’ and ‘biological adhesion’. KEGG pathway enrichment analysis indicated significant involvement of multiple oncogenic and immune-related pathways, such as ‘TNF signalling pathway’, ‘MAPK signalling pathway’, ‘cytokine-cytokine receptor interaction’ and ‘NOD-like receptor signalling pathway’, which highlighted the potential role of DCAF13 in modulating tumour biology (Fig. 7D). GSEA results further demonstrated the TNF signalling pathway and homologous recombination as central regulatory pathways that potentially mediate the pro-tumourigenic effects of DCAF13 in CRC progression (Fig. 7E and F), which was further validated by the increased expression levels of γ-H2AX and decreased ATM expression levels in the shDCAF13 group when compared with that of the shNC group (Fig. 8). These results suggest an association between DCAF13 depletion and elevated DNA damage signalling.