Forty CRC and 40 matched non-carcinoma tissue samples were collected from patients from the First Affiliated Hospital of Jinan University (Guangzhou, China), and subjected to histopathological and clinical examination. The study protocol gained approval from the Institutional Research Ethics Committee of Jinan University (JNUKY-2023-0062) and was carried out following the Declaration of Helsinki. Informed consent was obtained from patients.

HCT116 and HT29 cells were cultivated within DMEM (Gibco, #11965, Shanghai, China) that contained 10% fetal bovine serum (FBS, Biological Industries, #04-001-1A, Shanghai, China) as well as 1% penicillin-streptomycin (Sangon Biotech, #E607011, Shanghai, China) under 37°C and 5% CO₂ conditions. Human umbilical vein endothelial cells (HUVECs), provided by the Institute of Precision Oncology Medicine and Pathology, School of Medicine, Jinan University, were cultured in ECM medium (Solarbio, #YZ-1001, Beijing, China) that contained 10% FBS and 1% penicillin-streptomycin under 5% CO₂ and 37°C conditions. All cells used in this study were free of mycoplasma contamination.

IHC assay was conducted on 4-µm-thick, formalin-fixed, paraffin-embedded tissue sections, which underwent xylene deparaffinization, gradient ethanol rehydration, and membrane permeabilization through incubation with 0.1% Triton X-100 in PBS for 10 min at room temperature. Antigen retrieval was subsequently performed through heat-induced epitope retrieval in citrate buffer (pH 6.0) for 15 min. After cooling, sections were incubated using 3% hydrogen peroxide (H₂O₂) contained in methanol at room temperature for a 15 min duration to quench endogenous peroxidase activity. The 5% normal serum was added to block sections for reducing nonspecific binding, followed by overnight section incubation under 4°C using primary antibodies: anti-USP13 antibodies (Santa Cruz Biotechnology, Dallas, TX, USA; #SC144160; 1:200), anti-human cell adhesion molecule-1 (CD31) antibodies (Servicebio, Wuhan, China; #GB113151; 1:500), and anti-human VEGFA antibodies (Proteintech, Wuhan, China; #19003–1-AP; 1:1000). Subsequently, the sections were incubated with biotinylated secondary antibodies (1:500) (Abcam, Goat Anti-Rabbit and Anti-Mouse IgG H&L (Biotin, ab6720 and ab6788, Shanghai, China) under ambient temperature for a 1-h duration, followed by staining with an avidin-biotin complex. Visualization was performed using diaminobenzidine (Sigma, D12384, Shanghai, China), which yielded a brown precipitate. Hematoxylin counterstaining (Sigma, H9627, Shanghai, China) was completed to enhance cellular contrast. The light microscope (Mshot, MF52-N, Guangzhou, China) was adopted to examine these sections, and image analysis was conducted using ImageJ (version 1.46, National Institutes of Health, Bethesda, MD, USA) for quantifying positive staining intensity and positive cell percentage. Statistical analyses were implemented to evaluate whether the positively stained sections were of significance, which contributed to our understanding of the expression patterns of target antigens in cancerous tissues.

The IHC staining pattern was scored to assess protein expression in tissue samples, primarily for detecting tumor markers. This scoring system comprises two primary components: staining intensity and positive cell proportion. The staining intensity is rated by a scale from 0 (non-staining) to 3 (strong staining), whereas positive cell percentage is rated from 0 (0% positive) to 4 (76%–100% positive). The IHC score is derived by multiplying these two values, yielding a score typically ranging from 0 to 12. To categorize high and low expression of USP13 in this study, we established the following threshold for IHC scores: scores ≥ 6, high expression; scores < 6, low expression.

By specific protocols, cells were subjected to either VEGFA siRNA (Thermo Fisher Scientific, 284703, Shanghai, China), PTEN siRNA (Cell Signaling Technology, #6251, Shanghai, China) or control siRNA (5^′-UUCUCCGAACGUGUCACGUdTdT-3^′) transfection by using Lipofectamine 2000 (Thermo Fisher Scientific, 11668030, Shanghai, China). The shRNA lentiviral vector pLKO.1 targeting USP13 was constructed by utilizing the target sequence: 5^′-CGATTTAAATAGCGACGATTA-3^′. For the negative control, a vector containing scrambled shRNA was obtained from Addgene (USA). USP13 overexpression was achieved by transfecting cells with a plasmid encoding the full-length human USP13 cDNA (pCMV-USP13, Sino Biological, HG10001-UT). Briefly, cells (2 × 10⁵/well) were inoculated into 6-well plates and transfected with Lipofectamine 3000 (Thermo Fisher, L3000015) following specific instructions. At 48 h later, transfected cells were subjected to selection using 500 μg/mL hygromycin B (Sigma, H0654) for 7 days. Overexpression efficiency was confirmed via Western blot using an anti-USP13 antibody (Proteintech, 14617-1-AP). Control cells were transfected with an empty vector (pCMV-3 × Flag, Sigma, E7033) under identical conditions. Subsequently, real-time quantitative polymerase chain reaction (RT-qPCR) and Western blotting (WB) assay were conducted to validate transfection efficacy.

CRC cells (1.0 × 10⁶) were cultivated overnight into 6-well plates; subsequently, 2 mL of DMEM was added to every well. The CM was obtained 24 h later, followed by 15 min of centrifugation at 1500× g and storage under −80°C. This CM was utilized in enzyme-linked immunosorbent assay (ELISA), Transwell, and tube formation assays.

Transwell assays were conducted according to prior depiction [18]. Migration and invasion assays using HUVECs were carried out in a coated Transwell chamber (Corning, New York, NY, USA; #353097). HUVECs (1 × 10³) were seeded onto the top chamber (Corning; #354248, Shanghai, China) with or without the Matrigel matrix (Corning, #354230, Shanghai, China). The bottom chamber contained 200 µL basal ECM and 800 µL CM (with/without recombinant VEGFA [rVEGFA]). These chambers were subjected to 36 h of incubation under 37°C using 5% CO₂. Following fixation, the cells onto top membrane were eliminated using a cotton swab. Membrane-bound cells were then immobilized before 10 min of 0.1% crystal violet staining at room temperature. The light microscope (Mshot, MF52-N) was employed for image capturing, whereas ImageJ (version 1.46; National Institutes of Health) was adopted for quantifying invading cells in three random fields from each well.

HUVECs (cell density: 1.0 × 10⁴/well) were inoculated on the 96-well plate covered with 50 µL of Matrigel matrix and later cultivated with CM under 37°C and 5% CO₂ conditions for 4 h. A microscope (Mshot, MF52-N) was used to capture four random images, while ImageJ software (version 1.46) was applied to the analysis.

ELISA kits (NeoBioscience, Shenzhen, China; #EHC108 for VEGFA) were utilized to quantify protein levels of VEGFA in cell culture supernatants in accordance with the manufacturer’s instructions.

The RNA extraction kit (Solarbio; #R1200) was utilized to extract total RNA following specific protocols. To conduct single gene analysis, total RNA (1 μg) was used to prepare cDNA by PrimeScript RT Master Mix (Takara, Shiga, Japan; #RR036A) with the 20 μL reaction system for a 15-min duration under 37°C and for a 5-s duration under 85°C. PCR analysis was conducted using a PCR system (Bio-Rad Thermal Cycler (Model: T100), Hercules, CA, USA) with SYBR Green Master Mix (Vazyme, Nanjing, China; #Q121-02-AA) in CFX96 Touch™ real-time PCR system (Bio-Rad Thermal Cycler, CA, USA) following specific protocols. The cycling procedure was shown below: one cycle under 95°C for 5 min; 10 s of amplification under 95°C and 30 s under 60°C for 40 cycles. The target mRNAs included VEGFA, VEGFB, VEGFC, VEGFD, PDGFB, PDGFC, THBS1, WNT7B, PGF, bFGF, MMP2, and CXCL9; the ACTB mRNA was an endogenous reference. The primers utilized in RT-qPCR can be observed from Table A1.

Protein extracts were prepared from CRC cells transfected with USP13-specific shRNA or overexpression plasmids with RIPA lysis buffer (Beyotime, P0013B, Shanghai, China) that contained protease/phosphatase inhibitors (Roche, 04906837001, Shanghai, China). Protein content was analyzed via Bicinchoninic Acid (BCA, Thermo Fisher Scientific, A55864, Shanghai, China) assay. Protein aliquots (30 μg/lane) were subjected to 10% SDS-PAGE for separation before transfer onto Poly(vinylidene fluoride) (PVDF) membranes (Millipore, IPVH00010, Shanghai, China). 5% defatted milk contained in Tris-buffered saline Tween (TBST) was added to block membranes under ambient temperature for a 1-h duration, followed by overnight incubation using primary antibodies under 4°C: anti-USP13 (1:1000, Abcam, #ab109264), anti-PTEN (1:1000, CST, #9552), anti-AKT (1:2000, CST, #9272), anti-phospho-AKT (Ser473, 1:1000, CST, #4060), anti-VEGFA (1:500, Santa Cruz, sc-53463), and anti-β-actin (1:5000, Sigma, A5441). Membranes were washed before a further 1 h of incubation using HRP-conjugated secondary antibodies (1:5000, CST) under ambient temperature. Protein band visualization was completed with ECL substrate (Millipore, WBULS0100, Shanghai, China) and quantified by ImageJ software.

For the in vitro PTEN deubiquitination experiment, HCT116 cells underwent transfection using Flag-PTEN, HA-ubiquitin, MYC-USP13, or control shRNA. Following transfection, the cells were exposed to 6 h of 10 µM MG132 (Thermo Fisher Scientific, J63250MCR, Shanghai, China) treatment. Flag-PTEN was then immunoprecipitated using Flag beads (MCE, HY-K0207, Shanghai, China) and subsequently immunoblotted with anti-HA and anti-Flag antibodies.

A total of 16 Six-week-old female BALB/c mice were obtained in GemPharmatech Co., Ltd. (Nanjing, China) and housed in a specific pathogen-free (SPF) environment at the Experimental Animal Center of Jinan University. The mice were randomly assigned to distinct groups. HCT116 cells (1 × 10⁷) with either empty vector (EV) or USP13 overexpression were suspended in 200 μL PBS at a 1:1 ratio before subcutaneous injection in the right flank of mice. Measurements of tumor volume were conducted at 2-day intervals. For investigating whether inhibition of USP13 against angiogenesis is mediated by the PTEN pathway and VEGF, mice inoculated with EV-containing or USP13-overexpressing HCT116 cells were administered the PTEN-specific inhibitor SF1670 (MedChemExpress, #HY-15842, Shanghai, China) once every two days. Xenografts were collected after 3 weeks. The tumor tissue was paraffin-embedded and subjected to IHC assay. The animal experimental findings were acquired in a blinded manner. The animal experiments gained approval from the Institutional Animal Care and Use Committee of Jinan University (approval number: 20210720-010).

GraphPad Prism 5.0 (La Jolla, CA, USA) was employed for statistical analysis. The mean values of the two groups were compared by Student’s t-test. Each experiment was conducted thrice. The p-value of < 0.05 stood for statistical significance.

For investigating the relation of USP13 expression with MVD of CRC and adjacent tissues, serial sections of tumor and adjacent tissues from 40 clinical CRC patients were subjected to IHC assay for USP13 and CD31 staining. As shown in Fig. 1A,B, CD31, a transmembrane glycoprotein, is a common endothelial marker of MVD and is primarily expressed by endothelial cells and various hematopoietic cells [10]. The IHC assay revealed a distinct difference in CD31 expression between CRC tumors and normal tissues (Fig. 1A,C). A subset of 22 cases with low USP13 expression and 18 cases with high USP13 expression was selected for further analysis. Tumor tissues from patients with high USP13 expression exhibited significantly lower CD31 expression as compared to those with low USP13 expression (Fig. 1D). Furthermore, USP13 expression was inversely related to CD31 expression in tissues of CRC patients (Fig. 1E). Therefore, USP13 has an important effect on modulating angiogenesis in CRC.

For analyzing the role of USP13 in the angiogenesis of CRC tumors, we established stable knockdown cell lines, namely sh-USP13-HCT116 and sh-USP13-HT29, and stable overexpression cell lines, namely Flag-USP13-HCT116 and Flag-USP13-HT29, through lentiviral infection. The successful knockdown and overexpression of USP13 in CRC cell lines HCT116 and HT29 were confirmed by WB assay and qPCR analysis (Fig. 2A–D). Given that vascular endothelial cell migration and proliferation are essential processes of angiogenesis [19], we examined how USP13 affects HUVEC migration and proliferation. Notably, CM from sh-USP13 CRC cells markedly promoted HUVEC invasion and migration (Fig. 2E,F). We also assessed whether USP13 facilitates tube formation by HUVECs, which is crucial for all angiogenesis stages [20]. From Fig. 2G, CM in sh-USP13 CRC cells significantly increased the branch point number. Furthermore, stable USP13 overexpression in HCT116 and HT29 cells was utilized to complement loss-of-function studies. The HUVEC migration, invasion, and tube formation abilities were significantly decreased by CM from CRC cells overexpressing USP13 (Fig. 3A,B). Therefore, USP13 is important for HUVEC angiogenesis.

Given the involvement of several cytokines and growth factors related to tumor angiogenesis [21], this study investigated if USP13 affects the production of essential cytokines that promote angiogenesis. To test the above hypothesis, several angiogenesis-related cytokines were analyzed for their mRNA levels in CRC cells with sh-USP13 or Flag-USP13 expression (Fig. 4A,B) and the mRNA levels of VEGFA, VEGFB, VEGFC, VEGFD, PDGFB, PDGFC, THBS1, WNT7B, PGF, bFGF, MMP2, and CXCL9 were quantified. The mRNA expression levels of VEGFA, VEGFB, VEGFD, PDGFB, PDGFC, THBS1, PGF, and CXCL9 were significantly upregulated in CRC cells with USP13 knockdown, while WNT7B and MMP2 mRNA expression levels were significantly downregulated (Fig. 4A). Overexpression of USP13 downregulated VEGFA, VEGFC, bFGF, and MMP2 of CRC cells (Fig. 4B). The trend in the change of VEGFA expression was consistent with that of USP13 in CRC cells. We subsequently analyzed VEGFA protein expression in CRC cells expressing sh-USP13 or Flag-USP13. The results of ELISA showed that USP13 overexpression remarkably decreased the VEGFA amount of CRC cells, while sh-USP13 markedly increased the VEGFA protein level (Fig. 4C,D). Additionally, the WB assay indicated a negative correlation between USP13 and VEGFA expression levels in CRC cells (Fig. 4E,F). These findings demonstrate the negative regulation of VEGFA expression in CRC cells by USP13.

Because we observed that USP13 negatively regulates VEGFA expression in CRC and USP13 inversely correlates with MVD, we investigated whether the regulation of HUVEC angiogenesis by USP13 is dependent on VEGFA expression. We found that rVEGFA enhanced HUVEC migration, invasion, and tube formation in comparison with CM from EV-supplemented CRC cells Moreover, USP13 overexpression reduced HUVEC migration, invasion, and tube formation and antagonized the rVEGFA-mediated effects (Figs. 5A,B and 6A). The above findings demonstrate that CM supplemented with rVEGFA effectively counteracts the inhibitory effect of CM from Flag-USP13 CRC cells against HUVEC angiogenesis.

Furthermore, VEGFA-specific siRNA (VEGFA siRNA) significantly reduced VEGFA protein expression of CRC cells (Fig. 6B). HUVEC migration, invasion, and tube formation were markedly enhanced by CM from sh-USP13 CRC cells; however, these properties of HUVECs were inhibited when using CM from sh-USP13 CRC cells after VEGFA siRNA treatment (Fig. 6C,D). These results provide strong evidence that VEGFA is essential for the USP13-mediated angiogenesis-inhibiting effect on HUVECs.

USP13 is a deubiquitinating enzyme of PTEN, and its absence in breast cancer cells may downregulate PTEN expression, thereby enhancing AKT phosphorylation [16]. Moreover, AKT signaling pathways, including NF-κB, play a role in regulating VEGFA expression [22,23]. Therefore, it was hypothesized that USP13 regulates VEGFA in CRC through the PTEN-AKT pathway. Subsequently, our experiments revealed that USP13 knockdown reduced PTEN protein expression and increased p-AKT levels in HCT116 and HT29 cells (Fig. 7A). Conversely, USP13 overexpression increased PTEN protein expression while decreasing p-AKT levels as compared to those in the EV group (Fig. 7B). The MYC-USP13 group also significantly reduced PTEN ubiquitination in comparison with control group in HCT116 cells (Fig. 7C). In contrast, USP13 knockdown significantly increased the ubiquitination of PTEN (Fig. 7D). Furthermore, PTEN overexpression rescued USP13 knockdown-mediated upregulation and phosphorylation of VEGFA (Fig. 7E). Conversely, PTEN knockdown antagonized VEGFA decrease and p-AKT upregulation induced by USP13 overexpression (Fig. 7F).

To further elucidate the effect of the USP13-mediated PTEN-AKT signaling pathway on tumor angiogenesis, HUVEC migration, invasion, and tube formation in CM were assessed following the transfection of sh-USP13 CRC cells with Flag-PTEN. As expected, the CM from the sh-USP13 group treated with Flag-PTEN suppressed the migration (Fig. 8A), invasion (Fig. 8B), and tube formation (Fig. 8C) of HUVECs. These findings collectively indicate that USP13 inhibits tumor angiogenesis through the PTEN-AKT signaling pathway.

USP13 was analyzed for its effect on tumor angiogenesis of CRC in vivo. USP13 overexpression remarkably decreased tumor volume and weight as compared with those in the EV group (Fig. 9A–C). Moreover, PTEN inhibitors mitigated USP13 overexpression-mediated decrease in tumor volume and weight (Fig. 9A–C). Additionally, IHC assay revealed that USP13 overexpression reduced CD31 and VEGFA expression and antagonized PTEN inhibitor-mediated upregulation of CD31 and VEGFA expression in the transplanted CRC tissue (Fig. 9D). These results suggest that USP13 can inhibit CRC angiogenesis by regulating PTEN to suppress VEGFA expression in vivo.

Lastly, we analyzed clinical samples to investigate the relationship between USP13 and VEGFA expression. These results suggested that USP13 was negatively correlated with VEGFA in CRC (p < 0.05) (Fig. 9E). Therefore, USP13 serves as the critical regulatory factor for angiogenesis inhibition of CRC.