Potential Roles of IGF2BP3 in Bladder Cancer

The IGF2BP3 can affect cancer cells through the regulation of m6A. In this study, we combined the RNA sequencing and the RNA methylation sequencing to investigate the role and the mechanism of the IGF2BP3 in bladder cancer.TCGA data showed that the intersection genes of RBP differentially expressed and RBP involved in m6A modication were IGF2BP1 and IGF2BP3, only IGF2BP3 was signicantly associated with short survival time (P < 0.01). Then immunohistochemistry and qRT-PCR detection showed that IGF2BP3 is highly expressed in bladder cancer (P < 0.05). The T24 knockout IGF2BP3 cell line was successfully constructed, with cell proliferation decreased and the apoptosis increased (P < 0.05). RNA sequencing results of clinical bladder tissue from 53 cases showed that the expression of IGF2BP3 samples were associated with the high levels of proliferation-related genes. The analysis of combined RNA and M6a level showed the differential genes related to cell growth and proliferation: SPHK1, POMT2, MRPS18a in T24 and T24-KO. Our study suggests that the IGF2BP3 is highly expressed in bladder cancer and related to tumor metastasis and invasion, the IGF2BP3 expression is related to cell proliferation genes, which affects the growth, proliferation, and apoptosis of T24 cells, and SpHK1, POMT2, and MRPS18a may be potential targets of IGF2BP3. the most of the specimens we muscle-invasive urothelial carcinoma; a few are nonmuscle invasive urothelial carcinoma; and some are adenocarcinoma, squamous cell and Immunohistochemical experiments are carried out on the collected clinical tissues to verify the expression level of IGF2BP3 in clinical samples in TCGA database. Results show the abnormal expression of the IGF2BP3 gene in and almost no expression in normal tissues (P < 0.01). At the same time, the expression level of the IGF2BP3 is detected in qRT-PCR results show that the expression level of IGF2BP3 in cancer tissues is higher than that in adjacent tissues (P < 0.05). Clinical and pathological data show that the positive rate of the IGF2BP3 is higher in lymph node metastasis and invasive specimens (P < 0.01). However, due to small sample size, the relationship among the IGF2BP3, lymph node metastasis, and bladder cancer invasion should be further proven. The survival and the prognosis of IGF2BP3 patients are dicult to analyze because of incomplete collection of clinical samples and pathological data. Also, the expression of the IGF2BP3 in nonmuscle invasive urothelial carcinoma is signicantly correlated with advanced tumor stage (P < 0.001), advanced tumor grade (P = 0.004), and tumor recurrence (P < 0.001) 37 . The IGF2BP3 expression in the nonmuscle invasive urothelial carcinoma is an independent unfavorable prognostic factor for disease-free survival. Given that most of the specimens we have collected are muscle-invasive bladder cancer, the expression of the IGF2BP3 in the nonmuscle invasive urothelial cancer and its impact on survival and prognosis should be increased in the future. After the expression of the IGF2BP3 in clinical tissues is veried, the biological function of the IGF2BP3 in bladder cancer cells should be further studied. Considering the increased expression of the IGF2BP3 in bladder cancer to facilitate the stability of subsequent cell experiments, the CRISPR/cas9 system is used to construct stable cell lines with the IGF2BP3 knockout. invasion tests show that the cell migration and invasion abilities do not change signicantly. This phenomenon may be related to the characteristics of cell lines. Various cell lines should be selected to analyze cell biological function changes comprehensively and verify the cell function. The IGF2BP3 gene plays a vital role in many tumors. In tumor development, the IGF2BP3 has shown the role of promoting cancer by inuencing the tumor cell proliferation, apoptosis, migration, invasion, and drug resistance. The abnormal expression of IGF2BP3 in bladder tumors is determined in the rst part of this study. The IGF2BP3 knockout T24 cell line is constructed, the verication of the cell biological function showed that the IGF2BP3 was associated with cell population dependence, proliferation, and apoptosis. In the second part of the study, the RNA sequencing is used to analyze the expression and the signicance of the IGF2BP3 in clinical tissue samples. Given the increase in the IGF2BP3 expression, the expression of proliferation-related genes, such as ANLN, ASF1B, CCNB1, and CCNA2, also increases. Only SNAI2, CCN4, and FAP are consistent with the expression of the IGF2BP3. This result is consistent with our cell experiment. Finally, the RNA sequencing and the m6A methylation sequencing of cell samples are performed. Some differences are observed between T24 and T24– KO in the transcriptome, m6A-modied gene sets, and GO and KEGG enrichment analysis pathways. At the transcriptional level, the differentially expressed genes in the knockout and the control groups are related to cell growth, differentiation, development, cancer, inammation, and other related pathways. At the m6A modication level, the differentially expressed genes are related to cell growth, endoplasmic reticulum protein processing, splicing, cancer, and other related pathways. These results suggest that the IGF2BP3 may regulate the cell biological function in different ways at the transcription and the m6A levels. The essential genes involved in these two regulatory pathways are signicant. After crossing the differential genes of the two groups and the differential genes of m6A, there were 191 genes signicantly different online analysis tool GEPIA. Results show that SPHK1, MRPS18A, and POMT2 are positively correlated with IGF2BP3, and the expression levels of SPHK1, MRPS18A, and POMT2 in carcinoma are higher than those in normal tissues. The sphingosine-1-phosphate (S1P), a bioactive metabolite of sphingosine, is associated with tumor progression by enhancing the cell proliferation and motility. The SPHK can catalyze the formation of S1P, regulate sphingolipid rheostat, and promote tumor growth in various cancers. Related studies have shown that the increased expression of SPHK1 can promote tumor cell growth by increasing a variety of carcinogenic processes, including proliferation and self-renewal in vitro and tumor growth in vivo 42-44 . The POMT2 plays a crucial role in embryogenesis and development. Some studies have shown that the POMT2 can change the expression of E-cadherin, the essential protein involved in cell adhesion 45 . The increased expression of MRPS18A in human breast cancer cells can be explained by the increased energy metabolism of cancer cells, which indicates that MRPS18A may be involved in tumorigenesis 46 . In conclusion, after the IGF2BP3 knockout in T24 cells, the level and the m6A expression of SPHK1, MRPS18A, and POMT2 genes are downregulated and may be the target genes of IGF2BP3. In this study, we showed that the IGF2BP3 was highly expressed in bladder cancer tissue and cell lines. This nding was related to tumor metastasis and invasion and growth, proliferation, and apoptosis of T24 cells. We mainly revealed that SPHK1, POMT2, and MRPS18a might be potential targets of IGF2BP3 in T24 cells at RNA and m6A levels by sequencing, This result needs to be veried in the next project, but it provides valuable research ideas for further research. was used for differential expression analysis, and the absolute value of FoldChange (FC) ≥ 1.5 and adjusted P ≤ 0.05 were set as the cutoff criteria. The Gene Ontology (GO) functional annotation and the gene pathway (KEGG) analyses were used for further study. For the MeRIP-seq, after sending the cell samples to Guangzhou Epi-Biotech (Guangzhou, the total RNA extraction, RNA fragmentation, methylated RNA enrichment, methylated RNA recovery, sequencing library construction, sequencing, and data quality control were performed. The analysis process included peak calling statistics, peak annotation, motif analysis, and differential peak analysis. The correlation analysis was used to analyze the differential genes of m6A and differential expression genes and visualize GO and KEGG analyses.


Introduction
A total of 2.7 million new patients with bladder cancer are diagnosed or treated in the world every year. Thus, bladder cancer has become the 11th most common cancer worldwide. Bladder cancer has the highest recurrence rate among all malignancies 1,2 .
The RNA-binding protein (RBP) regulates the transcriptional and the post-transcriptional levels of most eukaryotic genes and endows each gene with a unique expression pro le under speci c conditions 3,4 . The insulin-like growth factor-2 (IGF2) messenger RNA (mRNA)-binding protein (IGF2BP) is a kind of mRNA-binding protein (IMP) that regulates the IGF2. The IGF2BP has a high expression in many malignant tumors, such as pancreatic, liver, lung, breast, and colorectal cancers 5-10 . Therefore, the IGF2BP can be used as a valuable tumor marker in human cancer. However, reports show considerable differences in the expression frequency among different types of cancer [11][12][13][14][15][16][17][18][19][20][21][22][23][24] , which may be due to the speci city of tissues, use of different antibodies, staining schemes, and scoring standards in experiments. Ramaswamy Suhasini suggest that the IGF2BP3 may promote the expression of the target IGF2 mRNA in an mTOR-controlled manner and promote the tumor's growth 25,26 . Rivera Vargas T et al. have shown that the IGF2BP3 can promote the tumor cells' proliferation through a synergistic effect with HNRNPM in the nucleus, resulting in increased expression of cyclin 9(Rivera Vargas et al., 2014). Furthermore, in triple-negative breast cancer research, the IGF2BP3 can affect the ability of tumor invasion and metastasis 27 .
The N6 methyladenosine (m6A) is the most abundant internal modi cation of mRNA and lncRNA in most eukaryotes and is involved in translation, decay, cell cycle regulation, cell differentiation, and circadian rhythm maintenance 28,29 . The m6A can affect the tumor progression, regulate the RNA stability, and promote the oncogene expression in acute myeloid leukemia, glioblastoma, lung cancer, and liver cancer [30][31][32][33] .
In this study, we have screened the differentially expressed RBP IGF2BP3 in bladder cancer. The IGF2BP3 is abnormally expressed in clinical bladder cancer tissues. The expression of IGF2BP3 is higher than that in adjacent tissues, whereas no expression in normal bladder tissues is observed. Therefore, we have used biological function experiments and gene sequencing to analyze the role of IGF2BP3 in bladder cancer; explore the primary mechanism of IGF2BP3 in bladder cancer; and provide theoretical bases for early detection, diagnosis, treatment, and monitoring of bladder cancer. Results 1. RNA-binding proteins analysis reveals an association between IGF2BP3 and bladder cancer by using TCGA database TCGA data were analyzed, and 206 RBPs with different expressions were selected in accordance with the |FC|>2 and the FDR value < 0.05 (Table 1). A total of 29 RBPs were found to be involved in the modi cation of m6A. The RBP was differentially expressed, and the R package analyzed the RBP involved in the modi cation of m6A. The overlapping genes were IGF2BP1 and IGF2BP3 (Fig. 1A). The TCGA online data analysis tool GEPIA was used, and the quartile was utilized as the cutoff value to analyze the effect of IGF2BP1 and IGF2BP3 expression on the survival time of patients with bladder cancer. The survival analysis showed that the survival time of patients with high expression of IGF2BP3 was signi cant shorter than that of patients with low expression of IGF2BP3. By contrast, the expression of IGF2BP1 had no signi cant relationship with the survival time of patients (Fig. 1B).  The qRT-PCR was used to detect the relative expression of IGF2BP3 in 27 pairs of clinical bladder cancer and corresponding paracancerous tissues. Results showed that the expression of IGF2BP3 in cancer tissues (C.A. blue) was higher than that in paracancerous tissues (C.P. red), and the difference was statistically signi cant (P < 0.05, Fig. 2B).
3. Construction of the CRISPR/cas9 recombinant vector and T24 cells with knockout IGF2BP3 To further explore the effect of IGF2BP3 on bladder cancer cells, we constructed T24 cells with knockout IGF2BP3. The expression of the IGF2BP3 in SW780, J82, 5637, and T24 bladder cancer cell lines was detected using Western blot. The IGF2BP3 was highly expressed in 5637 and T24 cells (Fig. 3A). In view of the easy transfection of T24 cells, subsequent knockout experiments were carried out in T24 cells. After cutting the bsmbi site on the vector (Fig. 3B), the plasmid DNA was extracted and sequenced. Results showed that at the RNA level, the construction of the vector was successful (Fig. 3C). The lentiviral vector plasmid was injected into the cells, and DNA ampli cation was sent to the company for sequencing, which showed a base mutation in the sequence near the knockout site (Fig. 3D). Results showed that at the protein level, compared with control cells, ko#3 cells almost did not express the IGF2BP3, which proved that ko#3 cells successfully knocked out the IGF2BP3 gene (Fig. 3E).

The biological function detect of T24 and IGF2BP3 T24-KO cells
Next, we tested the biological function of knockout T24 cells. The clone numbers of T24 and IGF2BP3 T24-KO cells were 152 ± 17 and 108 ± 16, respectively. According to the SPSS, a signi cant difference in colony-forming ability was observed between T24 and T24-KO cell lines (P < 0.001, Fig.4A). After the cells were seeded into 96-well plates, the cell proliferation activity was measured at 12, 24, 36, 48, and 72 h. On the second day, the proliferation ability of T24-KO cells was signi cantly lower than that of T24 cells (Fig. 4B). The apoptosis rates of T24 and T24-KO cells were detected using ow cytometry. Compared with that of T24 cells, the apoptosis of T24-KO cells was signi cantly increased (P < 0.01, Fig. 4C). The migration distance between T24-KO and T24 cells was observed under the microscope for 0 and 24 h.
Results showed no signi cant change in the migration distance between T24-KO and T24 cells (Fig. 4D). After 24 h, the transwell chamber was obtained for staining, and the number of invasive cells was calculated by randomly observing ve elds under the microscope. No signi cant difference was observed in the number of invasive cells between T24-KO and T24 cells (Fig. 4E).

IGF2BP3 was highly expressed associated with cell proliferation markers in clinical bladder cancer tissues
To further explore the mechanism of IGF2BP3 in clinical bladder cancer, We further further performed RNA sequencing on 53 clinical bladder tissues. The samples were rst arranged according to the expression of IGF2BP3, and the samples were grouped according to the molecular subtypes, then, the characteristic gene expression was clustered using the R software ( Fig 5A-B). Results showed no signi cant correlation between the expression level of IGF2BP3 and the six subtypes of bladder cancer (Fig. 5A). Compared with the low expression of IGF2BP3, the expression of molecular markers related to cell proliferation was higher in tissues with high expression of IGF2BP3 (Fig. 5B). Also, the expression of some genes related to EMT increased (Fig. 5C).
6. Analysis of differences in RNA levels between T24 and IGF2BP3 T24-KO cells The downstream pathway of IGF2BP3 was further analyzed by sequencing data. According to the list of expressed genes, the gene expression value of 0 was eliminated. A total of 1149 genes were screened in accordance with the fold change | FC | ≥ 1.5 and adj. P < 0.05.
A total of 847 genes were upregulated, and 302 genes were downregulated. The GO analysis showed ve pathways with abundant differential genes in BP (i.e., extracellular matrix organization, positive regulation of cell adhesion, extracellular structure organization, positive regulation of vascular development, and regulation of trans-synthetic signaling), CC (i.e., collagen-containing extracellular matrix, cell-cell junction, membrane raft, membrane microdomain, and membrane region), and MF (i.e., receptor-ligand activity, receptor-ligand activity, glycosaminoglycan binding, heparin binding, and growth factor receptor binding; Fig. 6A). The most signi cantly enriched KEGG pathway genes were TNF signaling pathway, AGE-RAGE signaling pathway in diabetic complexes, rheumatoid arthritis, and MAPK signaling pathway ( Fig. 6B).
7. m6A and RNA sequencing were combined to analyze T24 and IGF2BP3 T24-KO cells We combined RNA and m6A data for further comparison. The quality control of m6A data was quali ed ( Table 4). The genome distribution of the m6A peak was found using the IGF2BP3 motif (GGAC). No signi cant difference was observed between T24 and T24-KO cells (Fig. 7A).
The related genes with m6A modi cation were counted and compared between T24 and T24-KO groups. GO and KEGG databases were used for the enrichment analysis. The top 10 pathways enriched in GO analysis were related to gene transcription regulation, chromosome modi cation, and RNA splicing (Fig. 7B); and the top ve pathways enriched in KEGG analysis were Hippo signaling pathway, protein processing in the endoplasmic reticulum, neurotrophin signaling pathway, renal cell carcinoma, and insulin signaling pathway (Fig. 7C). The intersection of genes with different m6A levels (| FC | > 2, P < 0.05) and different transcription expression levels (| FC | > 1, P < 0.05) was obtained (Fig. 7D), and the 4-quadrant diagram of genes with the intersection part ( Fig. 7E) was demonstrated on the basis of the transcriptome sequencing. Results showed that the genes WDR6, SPHK1, MRPS18A, POMT2, UBE2G1, SAV1, MAP3K5, AJUBA, and FAM213 changed in gene expression and m6A level. TCGA was used to analyze the correlation of WDR6, SPHK1, MRPS18A, POMT2, FAM213A, and IGF2BP3 expression levels ( Fig. 7F) with the expression differences in clinical samples (Fig. 7G). Results suggested that the expression levels of SPHK1, POMT2, and MRPS18A were positively correlated with the IGF2BP3. The expression level of cancer was higher than that of normal tissue and was consistent with our sequencing results. After the IGF2BP3 knockout, the expression levels of SPHK1, POMT2, MRPS18A, and m6A in T24 cells decreased.

Discussion
The IGF2BP3 is initially identi ed as a highly overexpressed gene in pancreatic cancer. Subsequently, increasing evidence shows that the IGF2BP3 is abnormally expressed in various tumor tissues and affects patients' survival and prognosis in varying degrees 34 . The IGF2BP3 is not expressed in benign, low-grade, and high-grade noninvasive urothelial carcinomas 35 . The expression levels of IGF2BP3 are 11.76% and 55% in low-and high-grade invasive urothelial carcinomas, respectively. Compared with that in noninvasive specimens, the expression of the IGF2BP3 in invasive urothelial carcinoma is statistically signi cant (P < 0.000). This result suggests that the expression of the IGF2BP3 is related to the invasion of bladder cancer. Another study found that, the 5-year progression-free and disease-free survival rates for IGF2BP3 negative patients were 91% and 94%, while those of IGF2BP3 positive patients were 64% and 76%. 36 . In patients with super cial invasive urothelial carcinoma, IGF2BP3 positive patients are more likely to have metastasis than IGF2BP3 negative patients (P = 0.0017).
The bladder urothelial carcinoma is usually divided into nonmuscle-and muscle-invasive urothelial carcinomas. Given the hospital treatment PCR results show that the expression level of IGF2BP3 in cancer tissues is higher than that in adjacent tissues (P < 0.05). Clinical and pathological data show that the positive rate of the IGF2BP3 is higher in lymph node metastasis and invasive specimens (P < 0.01). However, due to small sample size, the relationship among the IGF2BP3, lymph node metastasis, and bladder cancer invasion should be further proven.
The survival and the prognosis of IGF2BP3 patients are di cult to analyze because of incomplete collection of clinical samples and pathological data. Also, the expression of the IGF2BP3 in nonmuscle invasive urothelial carcinoma is signi cantly correlated with advanced tumor stage (P < 0.001), advanced tumor grade (P = 0.004), and tumor recurrence (P < 0.001) 37 . The IGF2BP3 expression in the nonmuscle invasive urothelial carcinoma is an independent unfavorable prognostic factor for disease-free survival. Given that most of the specimens we have collected are muscle-invasive bladder cancer, the expression of the IGF2BP3 in the nonmuscle invasive urothelial cancer and its impact on survival and prognosis should be increased in the future. After the expression of the IGF2BP3 in clinical tissues is veri ed, the biological function of the IGF2BP3 in bladder cancer cells should be further studied. Considering the increased expression of the IGF2BP3 in bladder cancer to facilitate the stability of subsequent cell experiments, the CRISPR/cas9 system is used to construct stable cell lines with the IGF2BP3 knockout.
The CRISPR/cas9 technology is recognized as a potentially transformative application in transcriptional interference, epigenetic regulation, base editing, high-throughput genetic screening, and animal or cell models of diseases 38 . Compared with ZFN and talen, the CRISPR/cas9 technology is convenient, economical, and e cient. This study has used the CRISPR/cas9 system to knock out the IGF2BP3 gene in the human bladder cancer cell line T24. After the knockout, the DNA sequencing shows that the gene sequence has a base mismatch and that the target sequence has an overlapping peak. This nding may be due to the tetraploid T24 cells and different base repairs after the gene knockout. Finally, the Western blot con rms that the knockout cell line almost does not express the IGF2BP3 protein, indicating that the gene knockout is successful.
The IGF2BP3 (IMP3) gene is an IGF2. Members of the IMP family play an essential role in the early development of the embryo, neuron, and The IGF2BP3 gene plays a vital role in many tumors. In tumor development, the IGF2BP3 has shown the role of promoting cancer by in uencing the tumor cell proliferation, apoptosis, migration, invasion, and drug resistance. The abnormal expression of IGF2BP3 in bladder tumors is determined in the rst part of this study. The IGF2BP3 knockout T24 cell line is constructed, the veri cation of the cell biological function showed that the IGF2BP3 was associated with cell population dependence, proliferation, and apoptosis. In the second part of the study, the RNA sequencing is used to analyze the expression and the signi cance of the IGF2BP3 in clinical tissue samples. Given the increase in the IGF2BP3 expression, the expression of proliferation-related genes, such as ANLN, ASF1B, CCNB1, and CCNA2, also increases.
Only SNAI2, CCN4, and FAP are consistent with the expression of the IGF2BP3. This result is consistent with our cell experiment. Finally, the RNA sequencing and the m6A methylation sequencing of cell samples are performed. Some differences are observed between T24 and T24-KO in the transcriptome, m6A-modi ed gene sets, and GO and KEGG enrichment analysis pathways. At the transcriptional level, the differentially expressed genes in the knockout and the control groups are related to cell growth, differentiation, development, cancer, in ammation, and other related pathways. At the m6A modi cation level, the differentially expressed genes are related to cell growth, endoplasmic reticulum protein processing, splicing, cancer, and other related pathways. These results suggest that the IGF2BP3 may regulate the cell biological function in different ways at the transcription and the m6A levels. The essential genes involved in these two regulatory pathways are signi cant. After crossing the differential genes of the two groups and the differential genes of m6A, there were 191 genes signi cantly different at RNA and M6a levels. The IGF2BP3 is considered a protein that recognizes and binds to the mRNA modi ed by m6A, which can stabilize the target mRNA and promote translation. WDR6, sphingosine kinase (SPHK) 1, MRPS18A, POMT2, and FAM213A, whose expression level is consistent with IGF2BP3, are selected, and their correlation with IGF2BP3 expression tissues is analyzed in TCGA database online analysis tool GEPIA. Results show that SPHK1, MRPS18A, and POMT2 are positively correlated with IGF2BP3, and the expression levels of SPHK1, MRPS18A, and POMT2 in carcinoma are higher than those in normal tissues. The sphingosine-1-phosphate (S1P), a bioactive metabolite of sphingosine, is associated with tumor progression by enhancing the cell proliferation and motility. The SPHK can catalyze the formation of S1P, regulate sphingolipid rheostat, and promote tumor growth in various cancers. Related studies have shown that the increased expression of SPHK1 can promote tumor cell growth by increasing a variety of carcinogenic processes, including proliferation and self-renewal in vitro and tumor growth in vivo [42][43][44] . The POMT2 plays a crucial role in embryogenesis and development. Some studies have shown that the POMT2 can change the expression of E-cadherin, the essential protein involved in cell adhesion 45 . The increased expression of MRPS18A in human breast cancer cells can be explained by the increased energy metabolism of cancer cells, which indicates that MRPS18A may be involved in tumorigenesis 46 . In conclusion, after the IGF2BP3 knockout in T24 cells, the level and the m6A expression of SPHK1, MRPS18A, and POMT2 genes are downregulated and may be the target genes of IGF2BP3.
In this study, we showed that the IGF2BP3 was highly expressed in bladder cancer tissue and cell lines. This nding was related to tumor metastasis and invasion and growth, proliferation, and apoptosis of T24 cells. We mainly revealed that SPHK1, POMT2, and MRPS18a might be potential targets of IGF2BP3 in T24 cells at RNA and m6A levels by sequencing, This result needs to be veri ed in the next project, but it provides valuable research ideas for further research.

RNA extraction and qRT-PCR
The total RNA was extracted in accordance with the instructions in the Magen's total RNA extraction kit. The mRNA expression level was detected, and mRNAs were reversely transcribed into cDNA by using the PrimeScript™ RT reagent kit (TaKaRa CACCGAGGCGCAGAGGCAAATCACA and Primer-R (5′ to 3′) AAACTGTGATTTGCCTCTGCGCCTC were subcloned into the lentiviral vector ( Fig.   2A). Cloning was performed in accordance with the standard protocols in HEK 293T cells. Lentiviral particles transduced immortalized human podocytes, and single-cell clones of CRISPR/Cas9 genome-edited podocytes were generated. The DNA sequencing and the Western blot analysis were used to to verify the establishment of t24-ko cells.

Immunohistochemistry
Tissues were collected within 30 min after tumor resection and xed in formalin for 48 h. The sections were incubated with primary antibodies at 4 °C overnight (Abcam, USA). The corresponding secondary antibodies conjugated to horseradish peroxidase were incubated with the sections for 1 h at room temperature. After washing with PBS, the sections were reacted with diaminobenzidine prior to counterstaining with hematoxylin. Images were acquired using the Leica Microsystems (Leica, Germany).

Western blot analysis
Tissues and cells were lysed in RIPA buffer (Boster, China). Lysates were denatured at 95 °C for 5 min and cooled using ice. Lysates were loaded onto the sodium dodecyl sulfate-polyacrylamide gel (10% SDS-PAGE) and electrotransferred onto the polyvinylidene di uoride antibody (Invitrogen, USA) at room temperature for 2 h. The primary antibodies were as follows: rabbit monoclonal antibody to IGF2BP3 (Abcam, USA), mouse monoclonal antibody to GAPDH (Abcam, USA), and rabbit monoclonal antibody to β-actin (Abcam, USA).

Colony formation, MTT, ow cytometry, scratch, and transwell cell invasion analyses
Cells were seeded into 6-well plates and grown for 10 days. Visible colonies (≥ 50 cells) were counted after methanol xation and Giemsa staining. After transfection, cells were cultured into 96-well plates with 1000 cells/well. Each well was added with MTT solution and incubated at 37 °C for another 4 h. The medium was collected, and 100μl DMSO was added into each well. The microplate reader was used to detect the absorbance of the plate (490 nm). Cells were stained with annexin V-FITC and PI (BestBio, China) and analyzed using ow cytometry (BD. Biosciences). Apoptotic cells were detected and analyzed using the CellQuest software (BD. Biosciences). Only the early apoptosis rates of cells were determined to analyze the results. Approximately 3×106 cells were evenly spread, and the cell layer was carefully scratched with a sterile tip. Cells were incubated for 0 and 24 h, and a 24-well plate transwell chamber system (Corning, USA) was used to acquire images. Membrane cells were xed in 4% PFA for 15 min and stained with 0.1% Gimsa for 20 min. The stained chambers were left to dry and photographed. All experiments were repeated thrice.

RNA-seq and MeRIP-seq
For the RNA-seq, the total RNA was rst extracted from T24 with the stable IGF2BP3 knockout and sent to Mingma Technology Co., Ltd.
(Shanghai, China). The Fastqc was used for the quality control of sequencing data, and the Hisat2 software was used for sequence alignment. The Samtool and the Stringtie were used to convert the data format and assemble the transcript. The DESeq2 was used for differential expression analysis, and the absolute value of FoldChange (FC) ≥ 1.5 and adjusted P ≤ 0.05 were set as the cutoff criteria. The Gene Ontology (GO) functional annotation and the gene pathway (KEGG) analyses were used for further study. For the MeRIP-seq, after sending the cell samples to Guangzhou Epi-Biotech Co., Ltd. (Guangzhou, China), the total RNA extraction, RNA fragmentation, methylated RNA enrichment, methylated RNA recovery, sequencing library construction, sequencing, and data quality control were performed. The analysis process included peak calling statistics, peak annotation, motif analysis, and differential peak analysis. The correlation analysis was used to analyze the differential genes of m6A and differential expression genes and visualize GO and KEGG analyses.

Statistical analysis
Statistical analyses were performed using the IBM SPSS v17.0 (IBM Co., Armonk, NY, USA). The relationship between IGF2BP3 expression in normal tissues and pathological features of bladder cancer was analyzed by chi square test and Fisher's exact test. The RNA expression level of IGF2BP3 in corresponding paracancerous tissues of bladder cancer was compared by chi square test and one-way ANOVA. All data were expressed as mean ± standard deviation (n ≥ 3). P < 0.05 was considered statistically signi cant.. Figure 1 . Figure 2 . Figure 5 Page 13/13 . Figure 6 .