PTBP3 modulates P53 expression and promotes colorectal cancer cell proliferation by maintaining UBE4A mRNA stability

The RNA binding protein PTBP3 was recently reported to play a critical role in multiple cancers, and the molecular mechanisms involved RNA splicing, 3′ end processing and translation. However, the role of PTBP3 in colorectal cancer (CRC) remains poorly explored. Herein, PTBP3 was upregulated in CRC and associated with a poor prognosis. PTBP3 knockdown in colorectal cancer cell lines restricted CRC proliferative capacities in vitro and in vivo. Mechanistically, PTBP3 regulated the expression of the E3 ubiquitin ligase UBE4A by binding the 3′ UTR of its mRNA, preventing its degradation. UBE4A participated in P53 degradation, and PTBP3 knockdown in colorectal cancer cell lines showed increased P53 expression. UBE4A overexpression rescued PTBP3 knockdown-induced inhibition of CRC cell proliferation and P53 expression. Our results demonstrated that PTBP3 plays an essential role in CRC cell proliferation by stabilizing UBE4A to regulate P53 expression and may serve as a new prognostic biomarker and effective therapeutic target for CRC.


INTRODUCTION
Colorectal cancer (CRC) is the most common malignancy worldwide and the second leading cause of cancer-related death worldwide [1,2]. Although survival has improved because of advances in surgical techniques, malignant growth and metastasis remain major causes of cancer-related death [3]. Therefore, increasing our understanding of the mechanisms that drive CRC progression is urgent.
PTBP3 was first identified as an essential RNA-binding protein in 1999 [4] and is a member of the PTB family containing three paralogs -namely, PTBP1, PTBP2 and PTBP3 [5]. Previous studies have shown that PTBP3 is dysregulated and promotes the progression of numerous cancers, including breast cancer [6], hepatocellular cancer [7], gastric cancer [8] and CRC [9]. The functions of PTBP3 include RNA splicing, translational activation and mRNA stability [6][7][8][9][10]. However, the specific function of PTBP3 and its potential mechanism in CRC proliferation remain largely unknown.
E3 ubiquitin ligase ubiquitination factor E4A (UBE4A) is a U-boxcontaining ubiquitylation enzyme [11]. Similar to UBE4B, UBE4A belongs to the human homologs of the yeast UFD2 ubiquitination factor family, whose members share a conserved domain of approximately 70 amino acids named the U box [12][13][14]. Previous studies have implicated UBE4B as a regulator of P53 [15][16][17][18]; the U box catalytic domain is closely related to the RING-finger domain of MDM2 and is responsible for its E3 activity. These two enzymes are considered significant regulators of P53 through the ubiquitination process [17]. However, the function of UBE4A, as a U box catalytic domain-containing E3 ubiquitin ligase, remains largely unknown. Previous studies have reported that PCBP1, BMP7 and Viperin are UBE4A-ubiquitinated substrates [19][20][21]. In gastrointestinal disease, UBE4A dysregulation in enteroendocrine cells is associated with Crohn's disease [22], but its function and regulatory mechanism in CRC remain unclear.
In this study, we attempted to identify the function and regulatory mechanism of PTBP3 in CRC. PTBP3 was upregulated in CRC patients with a poor prognosis and led to CRC proliferation, suggesting that PTBP3 may be a crucial factor in CRC development. Mechanistically, PTBP3 regulated the expression of the E3 ubiquitin ligase UBE4A by binding to its 3′ untranslated region (UTR) to prevent its mRNA degradation. Additionally, UBE4A promoted CRC progression and participated in P53 degradation. Our results provide new evidence that PTBP3 exerts an oncogenic function and that the PTBP3/UBE4A/P53 axis might serve as a potential therapeutic target for CRC.

MATERIALS AND METHODS Tissue collection and ethics statement
Clinical material was obtained from patients treated at the Third Xiangya Hospital of Central South University (Changsha, China) with informed consent and approval of the Medical Ethics Central South University (No:2020-S095). Tissue specimens were snap frozen and stored in liquid nitrogen until further use.

Patients and tissue sampling
All the clinical material was obtained from patients who had undergone surgical resection for CRC at the Third XiangYa Hospital of Central South University (Changsha, China) after informed consent and approval from the Medical Ethics Central South University (No:2020-S095).

Quantitative real-time PCR assays
Total RNA from cells and tissues was extracted using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA). cDNA was generated using the ReverTra Ace qPCR RT Master Mix (TOYOBO, Osaka, Japan). Quantitative real-time PCR (qRT-PCR) was performed using a LightCycler 480 Real Time PCR instrument (Roche, Basel, Switzerland). GAPDH was used to normalize the qRT-PCR data. All the primer pairs were purchased from Sangon Biotech (Shanghai, China), and sequences were available in Supplementary Table 1.

Western blot assays
Whole cells and tissues were collected and lysed using 1× RIPA buffer (KeyGEN BioTECH) containing 1% PMSF (KeyGEN BioTECH) to harvest proteins. The protein samples were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), transferred to polyvinylidene fluoride membranes (PVDF) (Millipore, CA, USA), blocked with 5% skim milk for 2 h, incubated with primary antibodies at 4°C overnight and secondary antibodies for 1 h, and visualized using an Odyssey CLx Infrared Imaging System (LI-COR Biosciences, NE, USA). The antibodies used for western blotting (WB) are provided in Supplementary  Table 2.

Lentiviral vector and transfection
Lentiviruses against PTBP3 and UBE4A and their corresponding negative controls were obtained from Shanghai GenePharma Co., Ltd. To generate stable lentivirus-transduced lines, cells were infected with virus and polybrene following the manufacturer's recommendations, and stable cell lines were selected using 4 µg/ml of puromycin after 72 h of transfection. The efficiency in different cells was determined by the GFP intensity, qRT-PCR and WB. The shRNA sequences are listed in Supplementary Table 1.

Plasmid transfection
UBE4A plasmids were obtained from Shanghai GenePharma Co., Ltd. After transfection with the UBE4A plasmid, the cells were seeded into six-well plates and cultured for 24 h. When the density reached 50-60%, Lipofectamine 3000 reagent (Invitrogen, USA) was used to transfect the UBE4A plasmids according to the instructions.

Cell proliferation and colony formation assays
The CCK-8 assay (Dojindo, Kumamoto, Japan) was used to measure cell proliferation in 96-well plates. Approximately 2000 HCT116 cells, 2000 SW480 cells and 1000 LOVO cells were seeded per well, with six replicates for each condition. CCK8 was added at 0, 24, 48, and 72 h, followed by incubation at 37°C for 2 h. The absorbance values (A450) were detected using an EnVision microplate reader (PerkinElmer). For colony formation assays, approximately 100 HCT116 cells, 1000 LOVO cells and 1000 SW480 cells were seeded in each well of a 6-well plate in triplicate for each condition and incubated for 7 to 12 days. The colonies were fixed with methanol, stained with crystal violet, and counted. The average colony counts were calculated, and paired t-test was used to test statistical significance. Each experiment was repeated three times.

Flow cytometric analysis
Cell cycle analysis was measured by flow cytometry. A total of 100000 cells labeled with propidium iodide (PI; Sigma-Aldrich, USA) were prepared from each group and analyzed using a FACSCalibur flow cytometer (BD Biosciences). Next, the proportions of G0/G1, S and G2/M cells were calculated and compared using ModFit LT 3.1 software. The results were analyzed using a FACSCalibur flow cytometer (BD Biosciences).

Tumor xenografts
The nude mouse xenograft tumor growth model was developed according to the guidelines for experimental animal management established by Kagawa University and guidelines for the welfare and use of animals in cancer research [23]. Female BALB/c nude mice (4-5 weeks, 18-20 g) were obtained from the Department of Laboratory Animals of Central South University and maintained under specific pathogen-free conditions. We first weighed and numbered the nude mice, which were numbered 1-12 respectively. Then we wrote the numbers 1-12 on the paper, mixed them and extracted them in four groups, with three numbers in each group. The first group was HCT116 sh-NC,the second group was HCT116 sh-PTBP3,the third group was LoVo sh-NC,the last group was LoVo sh-PTBP3.Next, 3 × 10 6 HCT116-shNC (small hairpin carrying negative control RNA) cells and HCT116-shPTBP3 (small hairpin carrying PTBP3-specific RNA) cells and 2 × 10 6 LoVo-shNC cells and LoVo-shPTBP3 cells were harvested and injected subcutaneously into the left or right flank of the nude mice (n = 3 per group). Tumors were recorded using calipers and an electronic scale to estimate the tumor volume and weight every four days. Thirty days after the injection, the mice were killed by an overdose of pentobarbital (250 mg/kg; intraperitoneal injection), and the final tumor volume and weight results were recorded. The tumor volumes were calculated based on the formula: volume (mm3) = length (mm) × width (mm) × width (mm)/ 2. Tumors were further embedded in paraffin for H&E and immunohistochemistry (IHC).

Immunohistochemistry
Tissue sections from the nude mice were embedded in paraffin and then were deparaffined and rehydrated. Next, endogenous peroxidase activity was blocked by incubating the tissue sections with 0.3% hydrogen peroxide for 20 min. After that, the tissue sections were blocked in 10% BSA for 10 min and incubated with anti-human Ki-67 (1:100) antibodies and anti-human PTBP3 antibody (1:100) at 4°C for 12 h. The tumor sections were then incubated in biotinylated secondary antibodies for 20 min at room temperature. After that, the tissue sections were reacted with streptavidin-peroxidase conjugate for 10 min. Next, 3,3'-diaminobenzidine was added as the chromogen substrate. Images were captured using an inverted microscope system (IX73; Olympus, Japan).

RIP assay
RNA-binding protein immunoprecipitation (RIP) was performed using the EZ-Magna RIP Kit (Merck, KGaA, Darmstadt, Germany; Catalog No.  according to the manufacturer's instructions. Approximately 2 × 10 7 HCT116 and LOVO cells were washed with ice-cold PBS and resuspended in RIP lysis buffer containing a protease inhibitor mixture and RNase inhibitor. Next, magnetic bead protein A/G was incubated with 5 µg of IgG (negative control) (Merck KGaA) or PTBP3 (Santa Cruz) antibody for 30 min at room temperature. Cell lysis buffer and immunoprecipitation buffer containing EDTA and RNase inhibitor were added to the complexes. Thereafter, the complexes were incubated with rotation overnight at 4°C. The next day, the complex was washed with washing buffer containing proteinase K and 10% SDS and then heated at 55°C for 30 min. Finally, RNA was extracted and purified for RT-qPCR analysis. RIP assays were performed in biological triplicates and were detected by RT-qPCR. The primers are described in Supplementary Table 1.

RNA antisense purification assay
The RNA antisense purification assay was performed using the RNA Antisense Purification (RAP) Kit (Bersin BioTM, Guangzhou, China; CataLog Bes5103-3) according to the manufacturer's instructions. Approximately C. Xie et al.
2 × 10 7 HCT116 cells were washed with ice-cold PBS and resuspended in methanol and 1.375 M Glycine to cross-link cells. Next, lysis buffer containing the protease inhibitor mixture and RNase inhibitor was added for homogenization. Furthermore, DNase salt stock, DNase, EDTA, EGTA, and DTT were added to remove DNA. Thereafter, the probes were added to the processed sample, followed by hybridization at 37°C for 30 min (the probes are described in Supplementary Table 1), incubate degeneration at 50°C for 50 min, and hybridization again at 37°C for 180 min. Next, streptavidin beads were added to the complex and incubated for 30 min. After washing with wash buffer, the RAP mix bound to the beads was eluted and then resuspended in 60 μLof 1× loading buffer and boiled for 5 min, followed by Western blot detection.

Luciferase reporter assays
HCT116sh-NC cells, HCT116sh-PTBP3 cells, Lovosh-nc cells and Lovosh-PTBP3 cells were seeded at a density of 1 × 10 5 cells per well in a 24-well plate. The cells were transfected with the pRL-TK (Promega) Renilla plasmid using Lipofectamine 2000 according to the manufacturer's instructions (Invitrogen). The Renilla luciferase sequence in the pRL-TK vector (Promega, WI, USA) was used as an internal control. Dual luciferase reporter assays were performed according to the protocol using the Dual-Luciferase Reporter Assay System (cat. E1910; Promega). The firefly luciferase activity was normalized to the Renilla luciferase activity. The data were expressed as the percent of luciferase activity in control cells (100%).

Immunoprecipitation assay
Immunoprecipitation was performed according to the manufacturer's instructions (Thermo, PierceTM Classic Magnetic, Rockford, USA; LOT: UC283101). HCT116 cell lysates were prepared in immunoprecipitation lysis buffer (20 mM Tris-Cl, pH 8.0, 10 mM NaCl, 1 mM EDTA, 0.5% NP-40) containing a protease inhibitor cocktail (Sigma). Cell extracts (2 mg) were precleared with 50 μl of protein A/G-agarose (Santa Cruz) at 4°C for 2 h, and the supernatant was incubated with the corresponding antibodies with gentle shaking at 4°C overnight, followed by the addition of 50 μl of protein A/G-agarose for another 2 h. The beads were washed and then resuspended in 60 μl of 1× loading buffer, boiled for 5 min, and subjected to Western blot detection.

Statistical analysis
Statistical computations were performed using GraphPad Prism 8.0. (San Diego, CA). The data were presented as means ± s.d. of three independent experiments except where otherwise indicated. To compare the differences between two groups, Student's t-test was performed. Two-way analysis of variance (ANOVA) was used for comparisons between the different groups. The relationship between gene expression and clinicopathological indicators was examined using chi-squared test, and p < 0.05 was considered statistically significant.

RESULTS
PTBP3 is upregulated in CRC, and high PTBP3 expression correlates with a poor prognosis To identify the role of PTBP3 in CRC development, we first assessed the expression of PTBP3 in CRC tissues. The gene expression data for 471 colorectal cancer samples were downloaded from The Cancer Genome Atlas (TCGA) database. PTBP3 expression was first analyzed in 471 CRC tissues and 41 normal colorectal tissues, and PTBP3 was significantly upregulated in CRC tissues (Fig. 1A). Next, we downloaded and assessed two CRC gene expression datasets, GSE21510 and GSE44076, from the Gene Expression Omnibus (GEO) database, and the same trend was observed (Fig. 1B). To further confirm this conclusion, we detected PTBP3 expression in 30 matched pairs of human CRC tissues and adjacent nontumor tissues by qRT-PCR and 8 matched pairs of tissues by western blotting (WB). PTBP3 was significantly upregulated in human CRC tissues (Fig. 1C, D). Next, we summarized the clinicopathological characteristics of the 30 patients in the first cohort, and patients with high PTBP3 expression showed larger tumor sizes than those with low PTBP3 expression (Table 1). Additionally, PTBP3 expression was detected in 5 human CRC cell lines (HT-29, HCT116, SW480, SW620 and LoVo) and normal human colonic epithelial FHC cells, and PTBP3 expression was significantly higher in cancer cell lines than in FHC cells (Fig. 1E, F). HCT116 and LoVo cells with relatively high PTBP3 expression were selected for subsequent functional assays. Additionally, we examined the correlation between the PTBP3 expression level and prognosis of CRC patients using the Gene Expression Profiling Interactive Analysis (GEPIA) database. Kaplan-Meier survival analysis showed that patients with high PTBP3 levels had shorter overall survival times than those with low PTBP3 levels (Fig. 1G). Collectively, these results showed that PTBP3 was overexpressed in CRC patients with a poor prognosis.

PTBP3 is required for the proliferation of CRC cells in vitro and in vivo
To assess the functional significance of PTBP3 in CRC cells, we first silenced PTBP3 expression in HCT116 and LoVo cells with a lentiviral vector carrying PTBP3-specific small hairpin RNA (shRNA). Control cells were transfected with a lentiviral vector carrying negative control shRNA. The transfection effect was observed according to green fluorescence protein expression, and the PTBP3 silencing effect was confirmed by qRT-PCR and WB (Fig.  S1a-c). Next, we investigated the role of PTBP3 in HCT116 and LoVo cells using the Cell Counting Kit-8 (CCK-8) assay and colony formation assay. PTBP3 silencing reduced cell proliferation activity ( Fig. 2A, B). Next, we assessed the effect of PTBP3 on the cell cycle using flow cytometry. Silencing PTBP3 expression increased the proportion of cells arrested in the G0/1 phase and decreased the proportion of cells in S phase for both HCT116 and LoVo cells (Fig.  2C). These findings suggested that PTBP3 plays an essential role in CRC proliferation in vitro. To further evaluate the role of PTBP3 in vivo, we injected HCT116 and LoVo cells with stable PTBP3 knockdown and their corresponding NC cells into nude mice. All the mice developed tumors at the injection site (Fig. S2a), but the average size and weight of the tumors generated by PTBP3 knockdown cells were significantly smaller than those generated by NC cells (Fig. 2D-F). Immunohistochemistry (IHC) showed lower Ki67 expression in the PTBP3 knockdown tumor tissue group than in the NC group (Fig. S2b), suggesting that PTBP3 knockdown suppressed the proliferation of cancer cells. All tumor tissues were verified by hematoxylin and eosin (H&E) staining (Fig. S2b). Together, these results confirmed the oncogenic activity of PTBP3 in CRC in vivo, a finding consistent with that observed in vitro.
PTBP3 binds to the UBE4A 3′ UTR and may stabilize UBE4A mRNA To study the mechanism related to the oncogenic activity action of PTBP3, we performed gene set enrichment analysis (GSEA) using mRNA expression data from TCGA CRC datasets (Table S3). PTBP3 mRNA expression correlated positively with ubiquitinmediated proteolysis gene signatures, indicating that this pathway is the most closely related to PTBP3 activity (Fig. 3A). However, in previous studies, PTBP3 was characterized as an RNA-binding protein without a protein domain that participates in ubiquitylation directly [27]. Therefore, we speculated that PTBP3 might impact some critical ubiquitylation-related genes to influence the pathway. Consequently, we first analyzed PTBP3-correlated genes in TCGA CRC datasets (Table S4) and downloaded PTBP3 RNA immunoprecipitation sequencing (RIP-seq) data (Table S5) [28].
Next, a Venn diagram was drawn, and 136 closely correlated genes were obtained (Fig. 3B). The 136 genes were then subjected to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis (Fig. 3C), which identified four genes (UBE4A, ITCH, CUL4B, and HERC4) distributed in the ubiquitin-mediated proteolysis pathway (Fig. S3A). To investigate whether PTBP3 affected the expression of these four genes, we first performed qRT-PCR in HCT116 and LoVo PTBP3 knockdown cells. Only the mRNA levels of UBE4A were decreased (Fig. 3D). Subsequent WB experiments confirmed this correlation at the protein level (Fig. 3e). A previous study showed that RNA-binding proteins maintain mRNA stability by preventing RISC-mediated mRNA degradation [29,30]. Similarly, PTBP3 functions in mRNA decay and mRNA stability [6,28]. Therefore, we speculated that PTBP3 might associate with the UBE4A-3′ UTR to affect its mRNA stability and expression. To verify this hypothesis, we first assessed the RNA decay rate of UBE4A in PTBP3 knockdown CRC cells and corresponding control cells. UBE4A mRNA expression was initially decreased, and the UBE4A mRNA half-life was consistently markedly shortened following PTBP3 knockdown (Fig. 3F). Next, we performed RIP assays in HCT116 cells, and primers targeting the UBE4A-3′ UTR were enriched with the anti-PTBP3 antibody significantly more than that with IgG (Fig. 3G, H). Additionally, we developed a psiCheck-2 reporter plasmid containing the UBE4A-3′ UTR sequence cloned downstream of Renilla luciferase (Rluc) and firefly luciferase driven by the HSV-TK promoter. Rluc activity in cells with PTBP3 knockdown was significantly higher than that in control cells (Fig. 3I). These results suggested that PTBP3 binds to the UBE4A-3′ UTR directly. Next, we designed 4 probes targeting the UBE4A-3′ UTR and performed a biotin-labeled RNA Antisense Purification Fig. 1 PTBP3 is overexpressed in CRC and is associated with a poor prognosis. A PTBP3 mRNA level in the TCGA database (two-tailed Student's t-test, P < 0.0001). B PTBP3 mRNA level in the GEO database (two-tailed Student's t-test, P < 0.0001). C PTBP3 mRNA levels in CRC tissues compared with those in adjacent normal tissues (n = 30; measured by qRT-PCR; GAPDH was used as an internal control) (two-tailed Student's t-test, P < 0.01). D PTBP3 protein levels in CRC tissues compared with those in adjacent normal tissues (n = 8; measured by Western blotting). E PTBP3 mRNA levels in colorectal cancer cell lines (HT29, SW480, HCT116, SW620 and LoVo) compared with those in FHC cells, a normal colon cell line (measured by qRT-PCR; GAPDH was used as an internal control) (two-tailed Student's t-test, P < 0.05). F PTBP3 protein levels in colorectal cancer cell lines (HT29, SW480, HCT116, SW620 and LoVo) compared with those in FHC cells (measured by Western blotting). G Kaplan-Meier estimated overall survival in patients with high or low PTBP3 expression. Group cutoff-points: 25% (high) and 75% (low). The results are presented as means ± s.d. and are representative of at least three independent experiments.
Assay. (Table S1). PTBP3 was present in the products for probes 1 + 3 (Fig. 3j). The presence of AGO2 (a core component of the RISC complex) further proved our speculation (Fig. 3J). These results suggested that PTBP3 may prevent UBE4A degradation by RISC to stabilize its mRNA. To provide more direct evidence of this speculation, we performed immunoprecipitation experiments and revealed interactions between PTBP3 and AGO2 (Fig. 3K, I). Collectively, our results suggested that PTBP3 stabilizes UBE4A mRNA by binding to its 3′ UTR to prevent mRNA degradation mediated by the AGO2-containing RISC complex.

UBE4A promotes CRC cell proliferation in vitro
The above results indicated that PTBP3 regulates oncogenic activity in CRC and regulates UBE4A expression by binding to its 3ʹ UTR. However, the function of UBE4A in CRC remains unknown, and no previously published article is available concerning UBE4A in CRC. To assess the function of UBE4A in CRC cells, we silenced UBE4A expression in HCT116 and LoVo cells with a lentiviral vector carrying UBE4A-specific shRNA. Control cells were transfected with a lentiviral vector carrying negative control shRNA. The transfection effect was observed by detecting green fluorescence (Fig.  S4a), and the UBE4A silencing effect was confirmed by qRT-PCR and WB (Fig. S4b, c). Next, we used CCK-8 and colony formation assays to determine the role of UBE4A in mediating the malignant behavior of HCT116 and LoVo cells. The results showed that UBE4A silencing reduced cell proliferation activity (Fig. 4A, B), a finding that was consistent with the effect of PTBP3 in CRC. Similarly, we assessed the effect of UBE4A on the cell cycle using flow cytometry. Silencing UBE4A expression increased the proportion of cells arrested in the G0/1 phase and decreased the proportion of cells in S phase for both HCT116 and LoVo cells (Fig. 4C). These findings suggested that UBE4A promotes CRC cell proliferation in vitro, a finding that is consistent with the role of PTBP3 in CRC.
PTBP3 regulates P53 expression by facilitating ubiquitinmediated UBE4A degradation Previous studies have shown that the U-box catalytic domain of UBE4B is closely related to the RING-finger domain of MDM2 and is responsible for its E3 activity, and these two enzymes are considered significant regulators of P53 through the ubiquitination process [17,[31][32][33]. UBE4A, as a homolog of UBE4B [17], may also exert the same effect on P53. To verify this claim, we performed qRT-PCR and WB analysis in UBE4A knockdown CRC cell lines and their corresponding negative control cell lines. UBE4A knockdown had no effect on P53 expression at the mRNA level but significantly increased P53 expression at the protein level (Fig. 5A,  B), suggesting that UBE4A affects P53 protein stability. To verify this finding, we treated the HCT116 UBE4A knockdown cell line and corresponding negative control cell line with cycloheximide (CHX), a protein synthesis inhibitor. P53 expression was increased after UBE4A knockdown (Fig. 5C, left), and the rate of P53 degradation was slow at the indicated time after UBE4A knockdown (Fig. 5C, right). The stability of P53 was increased by UBE4A knockdown (Fig. 5C). Next, we treated the HCT116 UBE4A knockdown cell line and corresponding negative control cell line with MG132, a proteasome inhibitor. P53 expression was increased after MG132 treatment, suggesting that UBE4A affected P53 stability in the context of proteasome-dependent degradation (Fig. 5d). Immunoprecipitation experiments using the anti-UBE4A antibody, anti-P53 antibody and anti-MDM2 antibody revealed that the three proteins bound to each other (Fig. 5E). Collectively, these results proved our speculation that UBE4A may affect P53 stability via MDM2-P53 proteasome-dependent degradation to influence CRC proliferation. PTBP3 regulates P53 expression in hepatocellular carcinoma, but the specific mechanism remains unclear [7]. Here, we performed qRT-PCR and WB analysis of PTBP3 knockdown CRC cell lines, revealing that PTBP3 knockdown had no effect on the P53 mRNA levels but increased P53 protein expression (Fig. 5F, G). Therefore, we speculated that PTBP3 modulates P53 expression by mediating UBE4A expression. Next, we transfected UBE4A plasmids into PTBP3 knockdown HCT116 and LoVo cells (Fig. S4d, e). WB assays showed that UBE4A overexpression restored P53 expression in PTBP3 knockdown HCT116 and LoVo cells (Fig. 5H). Given that LoVo and HCT116 cells possess the wild-type (WT) P53 gene, we next used the SW480 cell line with mutant P53 to examine the effect of PTBP3 and UBE4A on CRC. We first silenced PTBP3 and UBE4A in SW480 cells separately. Transfection was detected using green fluorescence (Fig. S5a), and silencing was confirmed by qRT-PCR and WB (Figs. S5b, 5E). Next, we assessed cell viability and P53 expression in SW480 cells. Knockdown of either PTBP3 or UBE4A did not decrease SW480 cell viability or P53 expression (Fig. S5c-e). Collectively, these findings suggest that PTBP3 regulates WT P53 expression by facilitating the ubiquitin-mediated degradation of UBE4A in CRC.
UBE4A overexpression restores PTBP3 knockdown-mediated CRC cell proliferation To investigate the function of UBE4A in the PTBP3-mediated promotion of CRC proliferation, we transfected UBE4A plasmids into PTBP3 knockdown HCT116 and LoVo cells and examined the effect of UBE4A overexpression on proliferation activity and cell cycle distribution. UBE4A overexpression restored the proliferative activity and abnormal cell cycle of PTBP3 knockdown HCT116 and LoVo cells  ( Fig. 6A-C). These results further proved that PTBP3 promotes CRC cell proliferation by regulating UBE4A.

DISCUSSION
PTBP3 is a protein-coding gene that plays an essential role in humans [34]. Recent studies have shown that PTPB3 is dysregulated in multiple tumors and plays vital roles in carcinogenesis [6][7][8][9][10], but few studies have elucidated the specific mechanisms, particularly in CRC. In the present study, we found that PTBP3 is upregulated in CRC patients with a poor prognosis and promotes CRC cell proliferation by meditating UBE4A mRNA stability to regulate P53 expression.
In the present study, we first analyzed the TCGA and GEO online databases. PTBP3 mRNA was upregulated in CRC. Next, we performed an assay using clinical tissues to further confirm this conclusion at both the mRNA and protein levels. Next, we analyzed the correlations between the PTBP3 expression level and clinical features and found that a high PTBP3 expression level was correlated with clinical features, particularly the tumor size. Ping et al.'s study also showed that PTBP3 is upregulated and associated with a poorer clinical prognosis [9], supporting our findings. Additionally, PTBP3 overexpression was found in hepatocellular cancer [7], gastric cancer [8], and breast cancer [6] and was correlated with a poor clinical prognosis, suggesting that PTBP3 may serve as an oncogene in multiple cancers. However, these studies, including ours, lacked a sufficient amount of data to draw reliable conclusions. Thus, further multicenter clinical trials are required. To further investigate the oncogenic role of PTBP3 in CRC, we silenced PTBP3 in HCT116 and LoVo cells. After PTBP3 knockdown, CRC cell proliferation was significantly inhibited in HCT116 and LoVo cell lines in vivo and in vitro. This finding was partly consistent with that reported by Ping et al [9], who used HCT116 and SW480 cells for research but only conducted the CRC cell proliferation assay in HCT116 cells and not in SW480 cells. Considering that PTBP3 regulated P53 expression in our study, we also silenced PTBP3 in the SW480 cell line, a TP53-mutant gene cell line. However, silencing PTBP3 in SW480 cells had no effect on CRC cell proliferation. Similarly, PTBP3 has no effect on gastric cancer proliferation but promotes breast cancer and hepatocellular cancer growth [6,7]. We speculated that PTBP3-mediated regulation of P53 expression might be an explanation. Mechanistically, we found that PTBP3 was positively correlated with the ubiquitin-mediated proteolysis signaling pathway according to GSEA. However, PTBP3 does not have a protein domain that participates in ubiquitination [27]. Therefore, we investigated four candidate genes related to the ubiquitin-mediated proteolysis signaling pathway by PTBP3 RIPseq and identified PTBP3-correlated genes using TCGA database analysis. PTBP3 knockdown decreased UBE4A expression at both the mRNA and protein levels. We hypothesized that PTBP3 may influence UBE4A mRNA stability. We examined the decay rate of UBE4A mRNA, the results of which proved our hypothesis. Next, we performed RIP and luciferase reporter assays to confirm PTBP3 binding to the UBE4A 3′ UTR, and the RAP assay further proved this finding. The presence of AGO2, a core component of RISC, in the RAP assay suggested that PTBP3 may stabilize UBE4A mRNA by preventing RISC-mediated degradation of UBE4A mRNA. Further immunoprecipitation assays showed that PTBP3 and AGO2 could bind to each other. Collectively, our results first proved that PTBP3 has a marked impact on the ubiquitin-mediated proteolysis signaling pathway by stabilizing UBE4A mRNA in CRC.
The ubiquitin-mediated proteolysis signaling pathway plays a role in every cellular function, including essential processes for carcinogenesis, such as proliferation, apoptosis, and angiogenesis, which commonly occur in CRC [35][36][37]. UBE4A, an E3 ubiquitin ligase, has seldom been reported in the context of tumorigenesis. UBE4A represses ILEI protein expression to inhibit prostate cancer progression [38]. However, in thyroid carcinoma, UBE4A was reported as a ubiquitin ligase that is inversely correlated with PCBP1 protein expression and promotes cancer progression [19]. Therefore, the specific function and mechanism of CRC must be addressed. In the present study, we first demonstrated that UBE4A knockdown promoted CRC proliferation and increased P53 protein expression in CRC cell lines expressing the WT P53 gene but not in CRC cell lines expressing the mutant P53 gene. Further experiments indicated that UBE4A might influence P53 stability via the MDM2-P53 pathway. However, we did not elucidate the specific mechanism concerning the relationship between UBE4A and mutant P53 expression. Additionally, PTBP3 knockdown increased WT P53 expression, and a similar finding was reported in hepatocellular carcinoma [7]. However, PTBP3 knockdown decreased UBE4A expression in CRC cell lines with both WT and mutant P53, suggesting that PTBP3-regulated UBE4A stability is a common phenomenon in CRC. Finally, we overexpressed UBE4A in PTBP3 knockdown CRC cells, revealing that UBE4A restored P53 expression and inhibited CRC cell proliferation. Collectively, these results suggested that PTBP3 mediates UBE4A mRNA stability to regulate WT P53 expression and promote CRC proliferation.

CONCLUSIONS
In summary, our results revealed that PTBP3 is overexpressed and correlated with a poor prognosis and plays an oncogenic role, contributing to CRC proliferation. Additionally, PTBP3 is closely related to the ubiquitin-mediated proteolysis signaling pathwayrelated gene UBE4A and may mediate its mRNA stability to regulate its expression. However, we demonstrate for the first time that Fig. 4 UBE4A promotes CRC cell proliferation in vitro. A Reduction in the proliferation ability of shUBE4A HCT116 and LoVo cells compared with that of control (NC) cells using the CCK8 assay (two-way ANOVA, P < 0.01). B Reduction in the colony formation ability of shUBE4A HCT116 and LoVo cells compared with that of control (NC) cells using the colony formation assay. The bar graph indicates the number of colonies (two-tailed Student's t-test, P < 0.05). C Cell cycle distribution of shUBE4A HCT116 and LoVo cells compared with that of control (NC) cells was analyzed by flow cytometry. The results are presented as means ± s.d. and are representative of at least three independent experiments (twotailed Student's t-test, P < 0.05). and LoVo cells compared with that of control (NC) cells (measured by qRT-PCR; GAPDH was used as an internal control) (two-tailed Student's t-test, P > 0.05). B Relative protein level of P53 in UBE4A-silenced HCT116 and LoVo cells compared with that of control (NC) cells (measured by Western blotting). C shUBE4A and shNC HCT116 cells were treated with 100 μg/mL of cycloheximide (CHX) and harvested at the indicated time points. P53 protein was detected by western blotting (left), quantified by densitometry, and plotted against time to determine P53 stability (right) (two-tailed Student's t-test, P < 0.05). D shUBE4A and shNC HCT116 cells were treated with 5 μmol/L of MG132 for 12 h, and P53 protein was detected by western blotting. E Immunoprecipitation assays showing that UBE4A、MDM2 and P53 three proteins bound to each other. F Relative mRNA expression level of P53 in PTBP3-silenced HCT116 and LoVo cells compared with that of control (NC) cells (measured by qRT-PCR; GAPDH was used as an internal control) (two-tailed Student's t-test, P > 0.05). G Relative protein level of P53 in PTBP3-silenced HCT116 and LoVo cells (measured by Western blotting). H PTBP3 knockdown HCT116 and LoVo cells were transfected with the UBE4A plasmid, and P53 expression was detected. (measured by Western blotting). The results are presented as means ± s.d. and are representative of at least three independent experiments.
UBE4A knockdown inhibits CRC proliferation, which may be associated with P53 stability. In conclusion, our results indicate that the PTBP3/UBE4A/P53 axis may be a prognostic marker and therapeutic target in CRC, potentially providing new insight into CRC progression and treatment.

DATA AVAILABILITY
All the data generated or analyzed during this study are included in this published article [and its supplementary information files]. Fig. 6 Overexpression of UBE4A restores cell proliferation in PTBP3 knockdown CRC cells. A PTBP3 knockdown HCT116 and LoVo cells were transfected with the UBE4A plasmid, and the proliferation ability was quantified using the CCK-8 assay (two-way ANOVA, P < 0.001). B PTBP3 knockdown HCT116 and LoVo cells were transfected with the UBE4A plasmid, and the colony formation ability was quantified using the colony formation assay (two-tailed Student's t-test, P < 0.01). C PTBP3 knockdown HCT116 and LoVo cells transfected with the UBE4A plasmid were analyzed by flow cytometry. The results are presented as means ± s.d. and are representative of at least three independent experiments (twotailed Student's t-test, P < 0.05).