USP48 functions as a suppressor of colorectal cancer via SELENBP1 stabilization

Highlights

• •USP48 and SELENBP1 expression is positively correlated with clinical outcome of CRC.
• •Overexpression of USP48 suppresses the malignant behaviors of CRC cells.
• •USP48 functions as a suppressor of colorectal cancer by stabilizing SELENBP1.
• •USP48 could serve as a potential biomarker for the malignancy degree of CRC.

Abstract

Aim

Colorectal cancer (CRC) is a leading cause of cancer-related death. SELENBP1 is a tumor suppressor in CRC. However, the mechanism regulating the stability of the SELENBP1 protein remains elusive.

Methods

Here, we explored multiple ubiquitin-specific proteases (USPs) that may control SELENBP1 stability. In vitro, USP48 was overexpressed in CRC cell lines to observe its effect on malignant behaviors of cells. The interaction between USP48 and SELENBP1 was investigated using immunoprecipitation.

Results

Our data showed that USP48 exhibits a robust effect in stabilizing SELENBP1. Results of bioinformatics analysis showed that although USP48 expression was variable across datasets, its prognostic signal was consistent, and it was positively correlated with the prognosis of CRC patients. USP48 expression is also decreased in multiple CRC cells, and overexpression of USP48 impaired the malignant behaviors and stemness of HCT-116 and LOVO CRC cells. At the molecular level, USP48 interacts with SELENBP1 in CRC cells. USP48 overexpression substantially decreased the ubiquitination level of SELENBP1 and enhanced its stability. Depletion of SELENBP1 in USP48-overexpressing CRC cells improved their malignant behaviors and stemness, suggesting that SELENBP1 is a key factor in mediating the tumor suppressor function of USP48 in CRC.

Conclusion

Together, these observations deepen our understanding of the post-translational regulation of SELENBP1 and underscore the diagnostic and therapeutic potential of the USP48-SELENBP1 axis in CRC.

Graphical abstract

USP48 mediates deubiquitination of SELENBP1, blocking its ubiquitin-proteasome degradation. Stabilized SELENBP1 then inhibits PI3K/AKT signaling, suppressing CRC cell proliferation, migration, invasion, and stemness, while promoting apoptosis.

Introduction

Colorectal cancer (CRC) poses a significant health challenge worldwide, standing as one of the primary contributors to cancer-related illness and death [1]. The complexity of its molecular mechanisms contributes to its aggressive nature and resistance to conventional therapies [[2], [3], [4], [5], [6]]. Consequently, a better understanding of the etiology and pathology of CRC is warranted for novel diagnostic and treatment development to enhance the outcome predictions for patients with CRC.

Ubiquitination is a ubiquitous and essential post-translational modification, and its key role in tumorigenesis has been well documented [[7], [8], [9]]. Ubiquitination involves the addition of ubiquitin molecules to client proteins to label them for degradation or other biological events. It is a reversible process that is tightly controlled by the E1-E2-E3 cascades and deubiquitinases such as ubiquitin-specific proteases (USPs) [7,10].

The role of USPs in CRC has been intensively investigated. For instance, USP4 is overexpressed in CRC tissues and promotes CRC tumorigenesis by stabilizing PRL-3 [11]. USP47 facilitates CRC progression by regulating the deubiquitination and stabilization of YAP [12]. By contrast, USP44 curbs the growth and triggers the programmed cell death of by deubiquitinating the Wnt pathway repressor Axin1 [13]. However, given the large number of USPs in the family, the function of other USPs in CRC tumorigenesis remains incompletely understood.

SELENBP1, a selenium-binding protein implicated in cancer progression, was initially identified as a marker of mature adipocytes and promotes adipogenesis [14,15]. SELENBP1 is involved in cellular stress responses and has been linked to various malignancies. For example, the abnormal expression of SELENBP1 in head and neck cancer [16], oral squamous cell carcinoma [17], bladder cancer [18], and breast cancer [19] was associated with poor prognosis of patients. SELENBP1 has been reported to influence oxidative stress and act as a tumor suppressor in prostate cancer [20,21]. The role of SELENBP1 in CRC has been documented as well, the suppression of SELENBP1 expression is a common and late event in tumorigenesis and is closely associated with rapid tumor progression in CRC. Inducing SELENBP1 can significantly inhibit the growth and metastasis of CRC cells [[22], [23], [24]]. However, the mechanisms regulating the expression of SELENBP1, particularly at the post-translational level in CRC, remain largely unexplored.

In this study, we began with the screening of USPs that may affect the stability of SELENBP1. Our data identified USP48 as a potent stabilizer of SELENBP1. Subsequent bioinformatic analysis revealed that USP48 expression is downregulated in CRC samples and that low USP48 expression predicts poor survival of CRC patients. Next, we investigated the role of USP48 in two CRC cell lines, HCT-116 and LOVO. By overexpressing USP48 in these CRC cells, we examined the impact of USP48 overexpression on the viability, cell cycle, apoptosis, migration/invasion, and stemness of these CRC cells. Furthermore, we explored the mechanism by which USP48 stabilizes SELENBP1 in CRC cells, focusing on its effect on SELENBP1’s ubiquitination level. Finally, we depleted SELENBP1 in USP48-overexpressing CRC cells to determine whether USP48’s tumor suppressor function in CRC cells is mediated by SELENBP1. We also conducted single-cell sequencing analysis for the expression patterns of USP48 and SELENBP1 in different cell type tumors compared to normal cells. The findings of this study indicate that USP48 exerts a tumor-suppressive function in CRC by deubiquitinating and stabilizing SELENBP1, and USP48 expression is associated with consensus molecular subtypes (CMS) subtypes in tumor epithelial cells of CRC.

Materials and methods

Cell lines

HEK293T (#IM-H222), CCD-841CoN (#IM-H477), CaCo2 (#IM-H092), SW480 (#IM-H111), HCT-116 (#IM-H098) and LOVO (#IM-H105) cell lines were all obtained from Immocell Biotechnology (China), HEK293T and CCD-841CoN cell lines were cultured in DMEM with 10% fetal bovine serum (FBS, Gibco, USA) and 1% penicillin-streptomycin (Solarbio, China). CaCo2, SW480, HCT-116 and LOVO cell lines were cultured with the corresponding specific medium which obtained from Immocell Biotechnology, all of these cell lines are derived from colon (or colorectal) tissue. All of the cells were kept in a CO₂ incubator at a temperature of 37 °C.

Plasmid construction

The coding sequence of SELENBP1 was inserted into the pmirGLO-G4S-Nluc backbone, and the resulting plasmid, which encodes luciferase-conjugated SELENBP1 was named SELENBP1-Nanoluc. The His-tag-conjugated ubiquitin overexpression plasmid (His-Ub) was purchased from Miaoling Biotechnology (China). To construct the Flag-SELENBP1 plasmid, the coding sequence of SELENBP1 was cloned into the pcDNA3.3-Flag backbone. USP overexpression constructs, cloned into the pcDNA3.1-HA vector, were obtained from Xiamen Anti-hela Technology Co., Ltd (China). For the construction of the overexpression plasmid, the gene information was first retrieved from the NCBI database (https://www.ncbi.nlm.nih.gov/). Subsequently, primers were designed to amplify the coding sequence (CDS) of the target gene, with restriction enzyme sites incorporated at both ends. Following PCR amplification, the target band was purified by gel extraction. Both the vector plasmid and the PCR product were digested with the restriction enzymes and ligated using T4 DNA ligase. The ligation product was then transformed into chemically competent Escherichia coli cells. Positive clones were screened by PCR. After identification, positive clones were inoculated into fresh liquid medium and cultured overnight. Plasmid extraction was performed, followed by restriction enzyme digestion for verification. The digested products were analyzed by agarose gel electrophoresis to confirm the correct fragment sizes. Finally, the constructed plasmid was validated by sequencing.

For knockdown assays, short hairpin RNAs targeting SELENBP1 and USP48 (shSELENBP1s, shUSP48s) and a negative control shNC, were respectively inserted into the pLKO.1-TRC-PGK-puro plasmid. All plasmids were subjected to Sanger sequencing to confirm sequence accuracy before further experiments. The primers used for plasmid construction are listed in Table 1.

Table 1.

Primers for plasmid construction.

Plasmid name	Primer name	Sequence (5′-3′)
shSELENBP1–1	Forward	CCGGGGACGAGCTGCATCACTCAGGCTCGAGCCTGAGTGATGCAGCTCGTCCTTTTT
	Reverse	AATTAAAAAGGACGAGCTGCATCACTCAGGCTCGAGCCTGAGTGATGCAGCTCGTCC
shSELENBP1–2	Forward	CCGGGCTTCCTCTACTTCAGCAACTCTCGAGAGTTGCTGAAGTAGAGGAAGCTTTTT
	Reverse	AATTAAAAAGCTTCCTCTACTTCAGCAACTCTCGAGAGTTGCTGAAGTAGAGGAAGC
shSELENBP1–3	Forward	CCGGGGGAGACGTTCGAGGTGAAGGCTCGAGCCTTCACCTCGAACGTCTCCCTTTTT
	Reverse	AATTAAAAAGGGAGACGTTCGAGGTGAAGGCTCGAGCCTTCACCTCGAACGTCTCCC
shNC	Forward	CCGGTTCTCCGAACGTGTCACGTTTCTCGAGAAACGTGACACGTTCGGAGAATTTTT
	Reverse	AATTAAAAATTCTCCGAACGTGTCACGTTTCTCGAGAAACGTGACACGTTCGGAGAA
SELENBP1-Nanoluc	pmirGLO-SELENBP1-F	TGGAGCCACCGAATTCATGGCTACGAAATGTGGGAATTGTG
	SELENBP1–3*G4S-R	CCCGAGCCACCGCCACCAGAAATCCAGATGTCAGAGCTACAATCGC
Flag-SELENBP1	Forward	TACCGAGCTCGGATCCGCCACCATGGCTACGAAATGTGGGAATTGTG
	Reverse	GCCCTCTAGACTCGAGAATCCAGATGTCAGAGCTACAATCGC
HA-USP48 (USP48 OE)	Forward	CTAGAGAATTCGGATCCATGGCCCCGCGGCTGCAGC
	Reverse	AGCTTCCATGGCTCGAGTTAATGTCCAAGAAGACCAG
HA-USPL1	Forward	TAGAGAATTCGGATCCATGGCTACTACAAAAGATCCTGC
	Reverse	GCTTCCATGGCTCGAGTCAATAATTCTCAAACAGATATTC
HA-USP52	Forward	TAGAGAATTCGGATCCATGAACTTTGAGGGTCTGGACCCTGG
	Reverse	GCTTCCATGGCTCGAGTCAGAGCGCCAGCACTGAGG
shUSP48–1	Forward	CCGGGGTGAATGGTATAAGTTTAATCTCGAGATTAAACTTATACCATTCACCTTTTT
	Reverse	AATTAAAAAGGTGAATGGTATAAGTTTAATCTCGAGATTAAACTTATACCATTCACC
shUSP48–2	Forward	CCGGAGATGGTGATGCAGAACAAAGCTCGAGCTTTGTTCTGCATCACCATCTTTTTT
	Reverse	AATTAAAAAAGATGGTGATGCAGAACAAAGCTCGAGCTTTGTTCTGCATCACCATCT
shUSP48–3	Forward	CCGGCGAGAACTGTCAAAGCAAACACTCGAGTGTTTGCTTTGACAGTTCTCGTTTTT
	Reverse	AATTAAAAACGAGAACTGTCAAAGCAAACACTCGAGTGTTTGCTTTGACAGTTCTCG

Transcriptomic data collection and bioinformatic analysis

RNA expression and clinical survival data of patients with CRC were obtained from The Cancer Genome Atlas (TCGA-COAD, COAD stand for colon adenocarcinoma) database (https://portal.gdc.cancer.gov/). The expression data or clinical survival data of GSE44861, GSE103512, and GSE75500 [[25], [26], [27]] were collected from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). Differentially expression of USP48 and SELENBP1 was determined by the limma package (version 3.56.2) in R (version 4.3.0). The 'survival' package (version 3.5.0) in R was employed to perform Kaplan-Meier (KM) survival analysis.

Luciferase assay

HEK293T cells were seeded into 24-well plates at a density of 1.5 × 10⁴ cells per well. Upon reaching 70% confluence, the cells were co-transfected with 0.3 μg of plasmids encoding each USP, 0.3 μg of SELENBP1-Nanoluc plasmids, and 0.3 μg of pRL-TK vectors using Lipofectamine 2000 (Invitrogen, USA) according to the manufacturer's instructions. Following 48 hours after transfection, the cells underwent lysis, and the luciferase activity was evaluated using a dual-luciferase reporter assay kit (Promega, USA), the fluorescence values were measured using the GloMax® 20/20 Luminometer (Promega), according to the instructions provided by the manufacturer.

Construction of stable knockdown cell lines

24 h before transfection, HEK293T cells were digested with trypsin, and the cell density was adjusted to about 5 × 10⁶ cells/15 mL, the cells were seeded into a 10-cm dish and cultured for 24 h. When the cell density reached 70%−80%, it could be used for transfection. Two hours prior to transfection, the culture medium was replaced with serum-free medium, and the prepared DNA solution (vector plasmid 9 μg, psPAX2 plasmid 6 μg, pMD2.G plasmid 3 μg) was added to a sterilized centrifuge tube, then 500 μL Opti-MEM was added and mixed. Subsequently, 500 μL Opti-MEM and 60 μL ExFect transfection reagent were quickly mixed into a sterilized centrifuge tube and incubated at room temperature for 5 min. The two mixtures were mixed together and allowed to stand for 20 min. The mixture was slowly added to the culture medium of HEK293T cells, mixed, and cultured for 7 h. The medium was discarded, and 10 mL fresh medium containing 10% FBS was slowly added and cultured for 72 h, the supernatant was collected. The cells to be transduced were seeded into 6-well plates and allowed to grow to approximately 80% confluence prior to virus infection. The medium was removed and a 20 MOI mixture of virus and medium was added to each well, containing 10 μg/mL polybrene, then centrifuged at 800 × g for 2 h. After centrifugation, the cells were incubated overnight. After two days of infection, the medium was changed to contain 2 μg/mL puromycin for drug screening, and after 7 d, the antibiotics were withdrawn and the culture was continued. The cells were then used for subsequent experiments.

Cell counting kit-8 (CCK-8) assay

After transfection with the corresponding plasmid for 48 h, cells were seeded into 96-well plates at a density of 2000 cells per well, and then incubated for 48 h at 37 °C. Subsequently, 10 μL of CCK-8 solution (Beyotime, C0037, China) was added to each well, followed by further incubation for 1 h. Afterwards, the optical density (OD) at 450 nm was measured using a microplate reader.

Methyl thiazolyl tetrazolium (MTT) assay

After transfection with corresponding plasmid for 48 h, cells were plated in 96-well plates at a density of 6000 cells per well, then incubated for 24, 48 and 72 h. To achieve a final concentration of 0.5 mg/mL, MTT reagent (Yeasen, China) was introduced into each well, followed by incubation of the plates at 37 °C for a duration of 4 h. Following incubation, the culture medium was removed. To dissolve the formazan crystals, 100 μL of dimethyl sulfoxide (Sigma, USA) was added. The optical density (OD) at 570 nm was subsequently measured using a microplate reader.

Annexin V-FITC/PI staining assay

To detect the level of apoptosis, adherent CRC cells with different treatments were digested with EDTA-free trypsin, then collected by centrifugation at 300 × g for 5 min and washed twice with pre-chilled PBS. A total of 1–5 × 10⁵ cells were collected. The PBS was aspirated, and the cells were resuspended in 100 μL of 1 × binding buffer. Then, 5 μL of Annexin V-FITC and 10 μL of PI staining solution (Yeasen, #40302ES20) were added, and the mixture was gently mixed. The mixture was incubated for 10–15 min at room temperature in the dark. Subsequently, 400 μL of 1 × binding buffer was added, and the mixture was gently mixed and placed on ice. Apoptosis was then assessed using a BD FACSCalibur flow cytometer. Data analysis was performed using FlowJo software (BD Biosciences).

7-amino-actinomycin D (7-AAD) assay

To detect the cell cycle, CRC cells with different treatments were digested using trypsin and then fixed in 95% ethanol at −20 °C overnight. Following fixation, cells were washed with PBS. 20 mg/mL of 7-AAD solution (Beyotime Biotechnology) was added and incubated at 37 °C for 10 min, stained cells were subsequently analyzed by flow cytometry, and the data were processed using FlowJo software (version 10.8.1).

Colony formation

HCT-116 or LOVO cells, either with or without USP48 overexpression, were plated at a density of 1000 cells per 6 cm dish. The medium was changed every 72 h. After two weeks of incubation, the medium was removed, and the cells were washed twice with PBS. The cells were then fixed with 95% ethanol for 5 min and stained with a 0.1% crystal violet solution (Sigma) for 8 min. Colonies were photographed by an inverted microscope (Olympus) equipped with a digital camera and quantified using ImageJ software (NIH).

Spheroid formation assay

Cells were seeded at a density of 2000 per well in ultralow attachment plates (Corning, USA) and cultured for 10 d in the corresponding medium, each supplemented with 4 mg/mL insulin (MedChemExpress, USA), B27 (1: 50, Gibco), 20 ng/mL EGF (Sigma) and 20 ng/mL basic FGF (Sigma). Primary spheres were passaged by digesting with 0.25% trypsin and resuspending in culture media (with the same supplements). Finally, the spheres were imaged by an inverted microscope (Olympus) and measured.

Transwell assay

After transfection with USP48 OE or shSELENBP1 plasmid, cells were plated at a density of 4 × 10⁴ cells per well in the upper chamber of Transwell plates with an 8 μm pore size (Corning), using serum-free medium. For the invasion assay, 100 μL of a 1: 8 dilutions of Matrigel (BD Biosciences, USA) were applied to the upper chamber and incubated at 37 °C for 12 h. The lower chamber was filled with 600 μL of medium. The cells were incubated at 37 °C with 5% CO₂ for 24 h. Following incubation, the upper chamber was washed three times with PBS. Cells on the lower surface of the membrane were fixed with 4% paraformaldehyde for 30 min, stained with a 0.1% crystal violet solution for 10 min, and then washed three times with PBS. After air-drying, the stained cells were photographed by an inverted microscope and quantified using ImageJ software. For the migration assay, Matrigel coating was omitted, and all other steps were identical to those described for the invasion assay.

Quantitative PCR (qPCR)

Total RNA was first extracted from HCT-116 and LOVO cells. Specifically, 1 mL of TRIzol (Invitrogen, USA) was added to 10⁷ cells for 5 min of incubation. Following this, 0.2 mL of chloroform was introduced, and the tubes were sealed and vortexed for 15 s. The samples were then centrifuged at 12,000 × g for 15 min at 4 °C, and the RNA-containing aqueous phase was carefully transferred to a fresh RNase-free tube. An equal volume of isopropanol was added to the aqueous phase, mixed by inversion, and incubated for 10 min. The tubes were centrifuged again at 14,000 × g for 12 min at 4 °C. The RNA pellet was washed with 1 mL of 75% ethanol and centrifuged at 7500 × g for 5 min at 4 °C. After removing the supernatant, the pellet was allowed to air-dry for 10 min and then dissolved in 20 μL of RNase-free water. Its concentration was determined using a Nanodrop (Thermo Fisher Scientific, USA). qPCR was conducted using the ChamQ SYBR qPCR Master Mix (Vazyme, China) on a StepOnePlus Real-Time PCR system (Applied Biosystems, USA), following the manufacturer’s protocols. The relative expression levels of target genes were calculated using the 2^−ΔΔCt method. 18S rRNA was used as a reference gene. The qPCR primers are listed in Table 2.

Table 2.

Primers for qPCR.

Target gene	Primer- Forward (5′-3′)	Primer- Reverse (5′-3′)
18S rRNA	CGACGACCCATTCGAACGTCT	CTCTCCGGAATCGAACCCTGA
USP48	AAGCAGCAGTAAAGGGCGAT	ACCGTTCATCTTTCCGTTGC
SELENBP1	TCCCCAGTATTGCCAGGTCAT	CGACTTGGTGCTATCACCGAA

Western blotting

HCT-116 or LOVO cells were collected and lysed on ice with RIPA buffer containing protease inhibitors (Beyotime) for 15 min. The lysates were then sonicated for 5 min, at 30-second intervals. Following homogenization, the lysates were centrifuged to collect the supernatant. The protein concentration was determined using a BCA kit (Beyotime). Subsequently, the samples were denatured at 95 °C for 10 min, and 20 μg of protein from each sample was loaded onto 10% SDS-PAGE gels. Electrophoresis was conducted at 80 V for 1.5 h, followed by the transfer of proteins to PVDF membranes at 300 mA for 2.5 h. The membranes were blocked with 5% non-fat dry milk for 1 h, followed by overnight incubation at 4 °C with the primary antibody solution. Following this, the membranes were washed with TBST and then incubated with secondary antibody solution for 1 h at 25 °C. Finally, the membranes were washed with TBST, and the bands were visualized using an ECL reagent and captured with a Bio-Rad ChemiDoc imaging system. The band intensity was measured using ImageJ. Antibody details are provided in Table 3.

Table 3.

Antibodies for Western blotting.

Antibody	Provider	Catalog No.	Dilution
Beta-actin (Actin)	Proteintech	20,536-1-AP	1: 5000
OCT4		60,242-1-AP	1: 3000
KLF4		11,880-1-AP	1: 2000
c-Myc		10,828-1-AP	1: 3000
HA-tag		51,064-2-AP	1: 3000
Flag-tag		20,543-1-AP	1: 3000
SOX2		11,064-1-AP	1: 3000
PI3K		20,584-1-AP	1: 3000
AKT		10,176-2-AP	1: 3000
p-AKT		28,731-1-AP	1: 3000
His-tag		66,005-1-Ig	1: 3000
USP48		12,076-1-AP	1: 500
SELENBP1		27,449-1-AP	1: 500
HRP goat anti-mouse IgG (H+L)		SA00001–1	1: 10,000
HRP goat anti-rabbit IgG (H+L)		SA00001–2	1: 10,000
p-PI3K	Cell Signaling Technology	#4228	1: 3000

Co-immunoprecipitation (Co-IP)

To detect the interaction between USP48 and SELENBP1, after co-transfection of cells with Flag-SELENBP1 and HA-USP48 plasmids for 24 hours, the cells were harvested and lysed, followed by a 15-minute sonication. After centrifugation at 14,000 × g for 15 min at 4 °C, the supernatant was collected. The supernatant was then incubated with HA-tag (Beyotime, #P2121) or Flag-tag (Sigma, #M8823) antibody-conjugated beads at 4 °C for 4 h. Finally, the proteins were eluted from the beads and analyzed via Western blotting.

Immunoprecipitation (IP)

To detect the effect of USP48 on the ubiquitination level of SELENBP1, after co-transfection of cells with Flag-SELENBP1, HA-USP48 (or shUSP48) and His-Ub plasmids for 24 hours, cells were exposed to 10 μM MG132 for 6 hours, the cells were harvested and lysed, followed by a 15-minute sonication. After centrifugation at 12,000 × g for 20 min at 4 °C, the supernatant was collected. The supernatant was then incubated with Flag-tag antibody-conjugated beads at 4 °C for 4 h. Finally, the proteins were eluted from the beads and analyzed via Western blotting.

Single-cell sequencing analysis

The Seurat package was used to analyze the CRC single-cell RNA sequencing dataset GSE200997 [28] (including 16 cases of tumor tissue and 7 cases of adjacent normal tissue), which focused on the expression patterns of USP48 and SELENBP1 in various cell types, as well as their differences between tumor epithelial cells and normal epithelial cells. R software (v4.4.2) was used to read the downloaded gene expression matrix, and the Seurat package (v5.4.0) was used to construct single-cell objects using the CreateSeuratObject() function. Cells were filtered based on the criterion that the number of expressed genes was ≥ 200, and genes were filtered based on the criterion that they were expressed in ≥ 3 cells. Quality control metrics were visualized using the VlnPlot() and FeatureScatter() functions, and low-quality cells were removed using the subset() function with the following filtering criteria: 200 < nFeature_RNA < 3000, percent.mt < 10%, and nCount_RNA < 30,000. The Seurat package (v5.4.0) was used to normalize the single-cell data after quality control. Subsequently, the FindVariableFeatures() function was used to identify highly variable genes (HVGs), and the top 2000 HVGs were selected for downstream analysis. Principal component analysis (PCA) was performed on the HVGs using the RunPCA() function to reduce data dimensionality. Based on the ElbowPlot results, the top 20 principal components were selected for subsequent analysis. Cell types were annotated based on the expression characteristics of known classic marker genes. Epithelial cells were annotated based on KRT18 expression, T cells based on CD3D expression, B cells based on CD79A expression, myeloid cells based on LYZ expression, fibroblasts based on COL1A2 expression, and endothelial cells based on CLDN5 expression.

For the analysis of the expression trends of USP48 and SELENBP1, the VlnPlot() function was used to visualize the expression distribution of the target genes in tumor and normal groups. The stat_compare_means() function from the ggpubr package (v0.6.1) was used for statistical testing using the Wilcoxon rank-sum test to assess the significance of expression differences between the two groups. Subsequently, within each cell type, expression data for USP48 and SELENBP1 were extracted using the FetchData() function. After filtering for cells with expression values greater than 0, Pearson correlation analysis was performed using the cor.test() function to calculate the correlation coefficient and significance p-value.

Statistical analysis

The charts and statistical analyses were performed using GraphPad Prism software (version 8.0). The corresponding statistical analysis methods are indicated in the legend of each graph. Each experiment was performed with three independent biological replicates. All data are presented as the mean ± standard deviation (SD), with a P-value <0.05 considered statistically significant.

Results

Screening for USPs regulating the stability of SELENBP1

To identify USPs that potentially regulate SELENBP1 stability, we overexpressed SELENBP1-Nanoluc along with individual USPs in HEK293T cells and then measured the luciferase activity. Our results showed that the overexpression of multiple USPs enhanced the luciferase activity. Among these, cells transfected with the USP48 overexpression plasmid (HA-USP48) exhibited the strongest luciferase activity (Fig. 1A). To further validate whether these USPs increase the level of SELENBP1, we overexpressed USP48, USPL1, and USP52 together with Flag-SELENBP1. Western blotting analysis revealed that USP48 and USP52, but not USPL1, enhanced the protein level of SELENBP1, whereas the effect of USP52 was relatively modest (Fig. 1B and C). Given the tumor suppressor function of SELENBP1, we proceeded to assess the effect of these USPs on the viability of LOVO cells. CCK-8 assay results revealed that USP19, USP35, USP48, USP52, and USPL1 all reduced viability, while USP3 had minimal influence on the viability of LOVO cells (Fig. 1D). It is suggested that USP48 and USP52 may inhibit CRC cell growth at least partially through SELENBP1. Given the strongest effect of USP48 on elevating SELENBP1 expression, USP48 was selected for subsequent experiments.

Fig. 1.

Screening of USPs that potentially regulate SELENBP1 stability. (A) A plot showing the normalized luciferase activity in HEK293T cells overexpressing SELENBP1-Nanoluc and individual USPs or empty vector (n = 3). (B) Western blotting data indicating the influence of HA-USP48, HA-USPL1, or HA-USP52 on SELENBP1 protein expression. (C) Quantification of the SELENBP1-Flag expression (n = 3). (D) CCK-8 assay outcomes illustrating the impact of USP3, USP19, USP35, USP48, USP52, or USPL1 overexpression on the viability of LOVO cells (n = 6). One-way ANOVA followed by Tukey's multiple comparisons test was used to calculate the significance of the data present in (C) and (D). The uncropped blot is presented in the Supplementary Material. Each experiment was performed with three independent biological replicates. n.s., not significant. RLU: relative light unit.

USP48 and SELENBP1 expression is positively correlated with clinical outcomes of CRC patients

To investigate the potential involvement of USP48, we evaluated the expression of USP48 and SELENBP1 mRNA levels in multiple transcriptomic analysis datasets. Our bioinformatic analysis revealed that SELENBP1 expression is downregulated in tumor in all three analyzed CRC cohorts, including TCGA-COAD, GSE44861, and GSE103512, while USP48 mRNA was downregulated in tumor in GSE44861 and GSE103512, but not in TCGA-COAD (Fig. 2A-C). Next, we explored the prognostic value of USP48 and SELENBP1 expression in CRC. Kaplan-Meier survival analysis was used to estimate overall survival (TCGA-COAD, GSE75500), and the results showed that CRC patients expressing higher levels of USP48 or SELENBP1 had a better prognosis compared to those displaying low levels of USP48 or SELENBP1 (P < 0.05) (Fig. 2D and E).

Fig. 2.

Explore the expression and prognostic value of USP48 and SELENBP1 in CRC patients. (A-C) The mRNA levels of USP48 and SELENBP1 in normal and CRC samples from the data sets of TCGA-COAD (A, Normal tissue: n = 41, Tumor tissue: n = 480), GSE44861 (B, NT: n = 55, CC: n = 56), and GSE103512 (C, NT: n = 12, CC: n = 57). Unpaired two-tailed Student’s t-test was employed to calculate P-values of the data in (A-C). (D, E) Kaplan-Meier curves showing the clinical outcomes of CRC patients in the cohort of TCGA-COAD (D) and GSE75500 (E) based on the levels of USP48 or SELENBP1, the optimal cutpoint was calculated using the R package survminer (v0.5.1), and the log-rank test P-value was determined. n.s., not significant. HR: hazard ratio. NT: normal tissue, CC: colon cancer.

Overexpression of USP48 suppresses the malignant behaviors and stemness of CRC cells

To directly examine the role of USP48 in CRC cells, we initially assessed its expression in various CRC cell lines. qPCR analysis revealed that, compared to the non-cancerous colon epithelial cell line CCD-841CoN, USP48 mRNA levels were downregulated in HCT-116, CaCo2 and LOVO cells (Fig. 3A). Consistently, Western blotting confirmed reduced USP48 protein levels in these cell lines (Fig. 3B). In contrast to HCT-116 and LOVO, SW480 cells exhibited an elevated USP48 mRNA level compared to CCD-841CoN cells (Fig. 3A). However, Western blotting data indicated that the USP48 protein level in SW480 cells was not significantly upregulated relative to CCD-841CoN cells (Fig. 3B), suggesting that USP48 expression in SW480 cells may be regulated at the post-transcriptional level.

Fig. 3.

USP48 controls the malignant behaviors and stemness of HCT-116 and LOVO cells. (A) qPCR data showing the mRNA expression of USP48 in the indicated cell lines. (B) Western blotting results indicating the protein level of USP48 in the indicated cell lines. One-way ANOVA followed by Tukey's multiple comparisons test was used to calculate the significance of the data in (A, B). (C–O) Malignant behaviors were compared between USP48-overexpressing and control vector-transfected HCT-116 and LOVO cells. The following assays were performed: MTT assay for cell viability (C), colony formation assay for cell proliferation ability (D, E), 7-AAD staining for cell cycle distribution (F, G), Annexin V-FITC/PI staining for apoptosis (H, I), Transwell assay for migration (J, K) and invasion (L, M), and spheroid formation assay for stem‑like properties (N, O). Unpaired two-tailed Student’s t-test was used for all statistical analysis of the data in (C—O). The uncropped blot is presented in the Supplementary Material. Each experiment was performed with three independent biological replicates. (n = 3). n.s., not significant. OD, optical density. 7-AAD, 7-Aminoactinomycin D. PI: propidium iodide. OE, overexpression. CCD-841CoN is the non-malignant control cell line in (A, B).

Next, we overexpressed USP48 in HCT-116 and LOVO cells and assessed their malignant behaviors. MTT assays revealed that USP48 overexpression significantly reduced the viability of HCT-116 and LOVO cells (Fig. 3C). Colony formation assay data showed that USP48 overexpression impaired the colony formation ability of HCT-116 and LOVO cells (Fig. 3D and E). 7-AAD staining assays uncovered that overexpression of USP48 prevented HCT-116 and LOVO cells in the G0/G1 phase from entering the S phase (Fig. 3F and G). Moreover, Annexin V-FITC/PI staining assay results illustrated that USP48 overexpression triggered enhanced apoptosis of CRC cells (Fig. 3H and I). Additionally, Transwell assay data indicated that the migration and invasion of HCT-116 and LOVO cells were disrupted upon USP48 overexpression (Fig. 3J-M). Spheroid formation assay revealed that USP48-overexpressing HCT-116 and LOVO cells formed smaller spheres compared to their control counterparts (Fig. 3N and O).

At the molecular level, overexpression of USP48 did not exhibit a significant impact on the mRNA level of SELENBP1 in HCT-116 and LOVO cells but upregulated the protein expression level of SELENBP1, as indicated by qPCR and Western blotting data (Fig. 4A-C). Furthermore, Western blotting results showed that USP48 overexpression caused decreased expression of OCT4, KLF4, c-Myc, and SOX2, which are key transcriptional factors maintaining stemness [29] (Fig. 4D and E). In addition, the phosphorylated-PI3K (p-PI3K) and p-AKT levels, which promote CRC progression [30], were downregulated in these CRC cells upon USP48 overexpression (Fig. 4D and E). Collectively, these findings support the notion that USP48 may act as a CRC suppressor.

Fig. 4.

USP48 overexpression represses the levels of stemness-related transcription factors and inhibits the PI3K/AKT pathway in HCT-116 and LOVO cells. (A) qPCR data showing the mRNA level of USP48 and SELENBP1 in control and USP48-overexpressing cells. (B, C) Western blotting results showing the protein level of USP48 and SELENBP1 in control and USP48-overexpressing cells. (D, E) Western blotting results indicating the levels of OCT4, KLF4, c-Myc, SOX2, PI3K, p-PI3K, AKT, and p-AKT in control and USP48-overexpressing cells. Multiple t tests were used for all statistical analyses. The uncropped blot is presented in the Supplementary Material. Each experiment was performed with three independent biological replicates. (n = 3). n.s., not significant. OE, overexpression.

USP48 interacts with and stabilizes SELENBP1 in HCT-116 and LOVO cells

To delve into the mechanism by which USP48 suppresses CRC, we focused on its impact on SELENBP1. Our Co-IP results showed that USP48 interacts with SELENBP1 in HCT-116 and LOVO cells (Fig. 5A). Given the deubiquitinase nature of USP48, we then asked whether it regulates SELENBP1’s ubiquitination level in CRC cells. By performing IP assays, we observed that USP48 overexpression markedly decreased the ubiquitination level of SELENBP1 in HCT-116 and LOVO cells (Fig. 5B), while USP48 knockdown increased the ubiquitination level of SELENBP1 (Fig. 5C). To further verify the effect of USP48 on the stability of SELENBP1, we overexpressed USP48 or knocked it down in these CRC cells and then treated them with the protein synthesis inhibitor cycloheximide (CHX). Western blotting results indicated that USP48 overexpression delayed the degradation of SELENBP1, while its knockdown accelerated the degradation of this protein (Fig. 5D-G). These observations confirm that USP48 stabilizes SELENBP1 by removing its ubiquitination in CRC cells.

Fig. 5.

USP48 deubiquitinates and stabilizes SELENBP1 in HCT-116 and LOVO cells. (A) Co-IP data showing the interaction between HA-USP48 and Flag-SELENBP1 in cells. (B, C) IP results depicting the impact of HA-USP48 (B) or shUSP48 (C) on the ubiquitination of SELENBP1 in cells. Following transfection with the indicated plasmids (HA-USP48 or shUSP48, His-Ub, Flag-SELENBP1), cells were exposed to 10 μM MG132 for 6 hours, and subsequently harvested for the measurement of ubiquitination levels. (D-G) Western blotting results showing the effect of USP48 overexpression (D, E) or knockdown (F, G) on the stability of the SELENBP1 protein in cells under the treatment with 100 μg/mL CHX for 0, 6, 9, and 18 hours. Two-way ANOVA was used for statistical analyses. The uncropped blot is presented in the Supplementary Material. Each experiment was performed with three independent biological replicates. (n = 3). OE, overexpression. sh, short hairpin. TCL, total cell lysate.

Knockdown of SELENBP1 strengthens the malignant behaviors and stemness of USP48-overexpressing CRC cells

To determine whether USP48 fulfills its anti-CRC function by regulating SELENBP1, we depleted SELENBP1 in USP48-overexpressing HCT-116 and LOVO cells. The efficiency of shSELENBP1s was verified by qPCR, shSELENBP1–3 was selected for subsequent analysis (Fig. 6A, Fig. S1). By performing MTT assays, we discovered that SELENBP1 deficiency significantly improved the viability of USP48-overexpressing HCT-116 and LOVO cells (Fig. 6B). Similarly, the colony formation ability of USP48-overexpressing HCT-116 and LOVO cells was bolstered by SELENBP1 knockdown (Fig. 6C and D). The cell cycle of USP48-overexpressing HCT-116 and LOVO cells was also facilitated by SELENBP1 depletion (Fig. 6E and F). Furthermore, the apoptosis of HCT-116 and LOVO cells with USP48 overexpression is attenuated by SELENBP1 knockdown (Fig. 6G and H). In addition, the migration and invasion of USP48-overexpressing HCT-116 and LOVO cells were enhanced by SELENBP1 depletion (Fig. 7A-D), and the reduced stemness capacity of these CRC cells was improved by deficient SELENBP1 (Fig. 7E and F).

Fig. 6.

SELENBP1 knockdown improves the impaired malignant behaviors of USP48-overexpressing HCT-116 and LOVO cells. (A) qPCR data indicating the mRNA level of SELENBP1 in HEK293T cells transfected with the indicated shSELENBP1s. One-way ANOVA followed by Tukey's multiple comparisons test was used to calculate the significance of the data. (B) MTT assay data showing the viability of cells with the indicated treatments. (C, D) Colony formation assay results depicting the colony formation ability of cells with the indicated treatments. (E, F) 7-AAD staining assay results indicating the cell cycle of cells with the indicated treatments. (G, H) Annexin V-FITC/PI staining outcomes illustrating the apoptosis of cells with the indicated treatments. One-way ANOVA followed by Tukey's multiple comparisons test was used for all statistical analysis of the data in (B-H). Each experiment was performed with three independent biological replicates. (n = 3). n.s., not significant. OD, optical density. 7-AAD, 7-Aminoactinomycin D. PI: propidium iodide. OE, overexpression. sh, short hairpin.

Fig. 7.

SELENBP1 knockdown promotes the migration, invasion, and stemness of USP48-overexpressing HCT-116 and LOVO cells. (A-D) Transwell assay data showing the migration (A-B) and invasion (C-D) of cells with the indicated treatments. (E-F) Spheroid formation assay outcomes indicating the stem‑like properties of cells with the indicated treatments. One-way ANOVA followed by Tukey's multiple comparisons test was used to calculate the significance of the data. Each experiment was performed with three independent biological replicates. (n = 3). n.s., not significant. OE, overexpression. sh, short hairpin.

By conducting Western blotting, we found that knockdown of SELENBP1 upregulated the levels of OCT4, KLF4, c-Myc, SOX2, p-PI3K, and p-AKT in USP48-overexpressing HCT-116 and LOVO cells (Fig. 8A and B). Collectively, these findings provide evidence supporting the role of SELENBP1 as a crucial mediator of the tumor suppressor function of USP48 in CRC cells.

Fig. 8.

SELENBP1 deficiency stimulates the expression of stemness-related transcription factors and increases the activity of PI3K/AKT pathway in USP48-overexpressing HCT-116 and LOVO cells. (A) Western blotting results indicating the levels of USP48, SELENBP1, OCT4, KLF4, c-Myc, SOX2, PI3K, p-PI3K, AKT, and p-AKT in cells with the indicated treatments. (B) Quantification data of the data shown in (A). One-way ANOVA followed by Tukey's multiple comparisons test was used to calculate the significance of the data. The uncropped blot is presented in the Supplementary Material. Each experiment was performed with three independent biological replicates. (n = 3). n.s., not significant. OE, overexpression. sh, short hairpin.

USP48 expression is associated with CMS subtypes in tumor epithelial cells of CRC

Then, we further explored the expression patterns of USP48 and SELENBP1 in different cell types through single-cell RNA sequencing. Cell clustering and annotation were performed using the CRC single-cell dataset GSE200997 (Fig. S2).

Afterwards, the expression patterns of USP48 and SELENBP1 were analyzed. The results showed that there was no significant difference in the expression of USP48 between tumor epithelial cells and normal epithelial cells, and SELENBP1 expression was significantly higher in normal epithelial cells than in tumor epithelial cells, and its expression pattern was consistent with bulk RNA-seq results (Fig. S3A). Correlation analysis across cell types revealed a significant positive correlation between USP48 and SELENBP1 expression specifically in epithelial cells, supporting epithelial cells as the primary context for studying this co-expression relationship in CRC, which is the same context used for functional testing (Fig. S3B and C).

Further analysis of the correlation between the expression level of USP48 in tumor epithelial cells and the CMS classification revealed that USP48 is highly expressed in low-malignancy subtypes such as CMS2 and CMS3, while it is lowly expressed in highly malignant subtypes such as CMS4. A chi-square test indicated a significant association between USP48 expression groups and CMS subtype distribution (Fig. 9).

Fig. 9.

Further analysis of USP48 and SELENBP1 in epithelial cells. (A) Perform a second classification of the epithelial cells. (B, C) The co-expression situation of USP48 and SELENBP1 in tumor epithelial cells and normal epithelial cells. (D) The chi-square test showed that the expression groups of USP48 were significantly correlated with the CMS classification.

Nonetheless, these current observations are correlative and do not constitute functional evidence supporting a causal role for USP48 in CMS subtype-specific biology. Further studies are still needed to validate these associations and to explore the potential biological significance of USP48-SELENBP1 co-expression in epithelial cells and its relationship to CMS subtypes.

Discussion

Dysregulation of tumor suppressor genes is frequently observed in various types of cancers [31]. While a significant portion of such dysregulation is caused by genetic mutations, abnormal post-transcriptional regulations, such as ubiquitination, of tumor suppressors also contribute to tumorigenesis [32]. Our screening data showed that multiple USPs may potentially affect the stability of SELENBP1 in CRC. Notably, a significant portion of USPs exhibit a negative effect on SELENBP1’s stability. Considering deubiquitination usually prevents the degradation of the targeted protein, these USPs are unlikely to directly influence the ubiquitination of SELENBP1. Instead, they may stabilize ubiquitination promoters, such as E3 ubiquitin ligase that ubiquitinate SELENBP1 to facilitate the degradation of SELENBP1. Out of these various USPs, our focus was specifically on USP48 due to its notable ability to stabilize SELENBP1. In the context of CRC, while decreased mRNA levels of SELENBP1 may be responsible for its decreased expression, the reduced expression of its protein stabilizer like USP48 could exacerbate its loss of function in CRC tissues, thereby facilitating CRC development and progression.

The role of USP48 in cancer biology has been implicated in various contexts. For instance, USP48 deubiquitinates and stabilizes oncogene HMGA2 in CRC cells and the ablation of USP48 mitigates the metastatic potential of CRC cells [33]. HEG1 promotes gastric cancer progression by stabilizing AKT1 and is functionally regulated by the deubiquitinase USP48 [34]. In contrast to this, our data show that USP48 likely functions as a CRC suppressor as its expression in CRC is downregulated and is positively associated with the prognosis of CRC patients. Notably, although USP48 expression was variable across datasets, its prognostic signal was consistent. Consistent with our findings, a recent study by Zhang et al. demonstrated that USP48 interacts with and stabilizes TAK1, leading to the inhibition of TAK1-triggered NF-κB activation and effectively suppresses CRC tumorigenesis [35]. USP48 can stabilize Gasdermin E, a key mediator of pyroptosis, to promote cancer cell death [36]. Overexpression of USP48 has been observed to suppress the proliferation of acute promyelocytic leukemia cells [37]. This functional discrepancy of USP48 can be explained from the following perspectives. First, the function of USP48 may be highly dependent on the specific CRC cell line and its genetic background. The CRC cell lines used in this study harbor different genetic mutations, which may directly affect the downstream signaling networks that determine the functional output of USP48. Second, as a deubiquitinating enzyme, USP48 can simultaneously stabilize multiple target proteins. The function of USP48 in a given cell is likely the additive effect of its actions on all its substrates. Third, CRC is a highly heterogeneous disease encompassing multiple molecular subtypes. Our analysis of public datasets suggests that USP48 dysregulation may occur only in specific CRC subtypes. USP48 may function as an oncogene in certain subtypes while acting as a tumor suppressor in others. Future studies stratifying patients by molecular subtypes will be necessary to fully understand the context-dependent role of USP48 in CRC.

The mRNA and protein expression levels of USP48 vary across different CRC cell lines. In SW480 cells, USP48 mRNA levels are higher than those in control cells, whereas the protein levels do not show a corresponding increase, suggesting the existence of regulatory mechanisms at the translational or post-translational level. First, certain miRNAs that are highly expressed in SW480 cells may bind to the 3′ untranslated region of USP48 mRNA, inhibiting its translation without affecting mRNA stability. Studies have shown that in acute myeloid leukemia, elevated expression of miR-26b-5p negatively regulates USP48 expression [38]. Second, m⁶A modification may also be involved in this regulation. It has been reported that METTL14-induced m⁶A modification stabilizes USP48 mRNA, thereby suppressing glycolysis in hepatocellular carcinoma cells [39]. Of note, m⁶A modification may also influence protein translation efficiency, and its modification patterns differ across different cell lines. In addition, SW480 cells harbor mutations in genes such as KRAS or TP53 [40]. These genetic differences may affect the molecular networks that regulate USP48 expression at the post-transcriptional level. Future investigations using a broader panel of CRC cell lines are warranted to comprehensively elucidate the regulatory mechanisms of USP48 expression in CRC.

Furthermore, our data reveal that USP48 mRNA level is reduced in tumor tissues of two CRC cohorts but is not changed in TCGA-COAD, possible reasons for this discrepancy include: TCGA-COAD is a comprehensive cohort with samples covering a wide range of clinicopathological stages and molecular subtypes, reflecting an overall average trend. The expression of USP48 may be downregulated in certain CRC subtypes but unchanged in others, resulting in no significant difference when averaged across the entire cohort. In contrast, GSE44861 and GSE103512 may predominantly include patients with specific stages, grades, or molecular subtypes in which the transcriptional downregulation of USP48 is more pronounced. Additionally, TCGA employs RNA-seq technology, whereas most GEO datasets are based on microarrays, these two platforms differ in their detection sensitivity for gene expression and in their normalization methods. Consequently, whether USP48 expression is altered in CRC tissues requires to be verified in more CRC patients.

Although this study focuses on the role of USP48 in regulating SELENBP1, it is interesting to note that USPL1 also increased SELENBP1-Nanoluc activity, whereas USPL1 had no effect on SELENBP1 protein expression. This discrepancy may be attributed to the fact that luciferase assays typically only detect regulation at the transcriptional level and do not involve regulation of protein stability. USPL1 may enhance SELENBP1-Nanoluc activity by affecting transcription factors, promoter activity, or SELENBP1 mRNA stability, but these effects may not be reflected in endogenous SELENBP1 protein levels. Another possible explanation is that USPL1 may not function as a direct deubiquitinase for SELENBP1.

Our single-cell analysis revealed a cell-type-specific positive correlation between USP48 and SELENBP1 expression, which was restricted to epithelial cells in both normal and CRC contexts. Given that SELENBP1 exerts tumor-suppressive functions by regulating oxidative stress [41], and USP48 maintains the tumor-suppressive function of SELENBP1 by stabilizing its protein, thus their co-expression in epithelial cells may represent a functional module involved in the regulation of oxidative stress responses. The alteration of this co-expression relationship in tumor epithelial cells may reflect dysregulation of stress responses during malignant transformation. Currently, further studies are still needed to validate this association and to explore the potential biological significance of USP48-SELENBP1 co-expression in epithelial cells.

Several limitations should be acknowledged. Although our study proved the USP48 interacts with and deubiquitinates SELENBP1, the interacting domain of these proteins remains undetermined. Future studies should map the key interacting regions by constructing a series of truncated mutants of USP48 and SELENBP1 for Co-IP assays. Additionally, our conclusion is based on in vitro experiments and lacks support from in vivo functional assays. Further validation of our conclusions would require verifying the results in vivo models (such as animal or organoid), investigating the mechanisms that regulate USP48 expression in CRC (including epigenetic or post-transcriptional regulation), and exploring whether USP48 interacts with other substrates relevant to CRC progression. Future ubiquitinomics combined with mass spectrometry could systematically identify the ubiquitination substrates of USP48 in CRC cells. Furthermore, loss-of-function experiments, particularly in CRC cells that show a high expression of USP48, could lend further support to get a more comprehensive overview of USP48’s function in CRC.

In conclusion, our study uncovers USP48′s crucial role in CRC by stabilizing SELENBP1. The intricate interplay between USP48 and SELENBP1 offers novel insights into the molecular mechanisms governing CRC tumorigenesis and highlights the USP48-SELENBP1 axis as a promising target for prognosis and therapeutic intervention of CRC.

Data of availability

The authors confirm that the data supporting the findings of this study are available within the article and/or its supplementary materials, and from the corresponding authors upon reasonable request.

Ethics approval

Not applicable.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

CRediT authorship contribution statement

Hongyue Lin: Writing – original draft, Visualization, Validation, Supervision, Software, Methodology, Formal analysis, Data curation. Ying Huang: Writing – review & editing, Validation, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix. Supplementary materials

Supplementary Fig. Legends

Fig. S1. Validation of SELENBP1 knockdown efficiency. (A) The knockdown efficiency of SELENBP1 in HCT-116 cells was assessed by qPCR. (B) The knockdown efficiency of SELENBP1 in LOVO cells was assessed by qPCR. One-way ANOVA followed by Tukey's multiple comparisons test was used to calculate the significance. Each experiment was performed with three independent biological replicates. (n = 3). sh, short hairpin.

Fig. S2. The single-cell dataset GSE200997 of CRC was used for cell clustering and annotation. (A) Annotate the cell populations based on the marker gene. (B-C) The expression status of marker genes within each cell population. (KRT18: epithelial cells, CD3D: T cells, CD79A: B cells, LYZ: myeloid cells, COL1A2: fibroblasts, CLDN5: endothelial cells).

Fig. S3. Analysis of the expression patterns of USP48 and SELENBP1 in different cell type tumors compared to normal cells. (A) The expression patterns of USP48 and SELENBP1 in different cell types. (B) USP48 is expressed in both immune cells and epithelial cells, while SELENBP1 is only expressed in epithelial cells. (C) Correlation analysis of the expression of USP48 and SELENBP1 in various cell types.