Up-Regulation of SH3TC2 Induced by YTHDF1 Predicts Poor Outcome and Facilitates Cell-Cycle Progress in Colorectal Cancer

N6-methyladenosine (m6A) modification plays a crucial role in determining the fate and function of RNA transcripts in tumor cells. Nevertheless, how m6A regulates the expression of key molecules and coordinates its involvement in the development of colorectal cancer (CRC) remains largely unclear. Here, we report that the m6A reading protein YTHDF1-mediated up-regulation of SH3TC2 promotes CRC growth both in vitro and in vivo. In a pan-cancer analysis across more than thirty types of cancer, we found that SH3TC2 was dysregulated in nine cancers, including BLCA, CHOL, COAD, LAML, PAAD, READ, SKCM, BRCA, and TGCT, and was closely associated with patient prognosis in four cancers, including COAD, MESO, PAAD, and READ. In particular, SH3TC2 was overexpressed in CRC as confirmed by six independent study cohorts. Clinically, high expression of SH3TC2 predicted worse disease-free survival (DFS) in CRC patients. SH3TC2 showed fascinating diagnostic value and was correlated with immunosuppression in CRC. Functionally, RNA-sequencing combined with experiments revealed that knockdown of SH3TC3 significantly inhibited cell-cycle progress of CRC, impairing cell growth. Mechanistically, YTHDF1 protein directly binds with SH3TC2 mRNA and promotes its elevation in an m6A-dependent manner. Thus, our findings provide a mechanism to target the YTHDF1/SH3TC2 axis for CRC therapy.


Introduction
Colorectal cancer (CRC) is one of the most prevalent malignant tumors of the digestive system worldwide, which poses a serious challenge to human health. Sung and colleagues published the Global Cancer Statistics 2020, which reported that CRC ranked among the top three out of 36 human cancers in terms of morbidity and mortality [1]. Over the past decade, the number of CRC-related deaths worldwide has increased from 600,000 in 2008 to 900,000 in 2020 [1,2]. Te survival of CRC patients has been prolonged as a result of improved early screening techniques and clinical treatment strategies [3,4]. However, tumor recurrence and progression mediated by unknown mechanisms remain major factors afecting clinical outcomes in CRC patients [5,6]. Terefore, it is urgent to explore the molecular mechanism driving the occurrence and development of CRC.
RNA m6A modifcation is a common form of epigenetic modifcation that occurs in eukaryotic cells. Under normal conditions, methyltransferase-mediated methylation and demethylase-mediated demethylation are in relative balance, and once the intracellular m6A modifcation state is out of balance, the biological function of the cell is also changed [7,8]. Recent studies have shown that tumor cells require m6A modifcation to maintain their malignant growth, progression, and therapeutic resistance [9,10]. As one of the m6A reading proteins, YTHDF1 determines the fate of m6A-labeled RNA transcripts, afecting their stability, localization, or translation efciency [11,12]. However, how YTHDF1 regulates the expression of a key molecule in CRC has not been well characterized.
SH3TC2, located on chromosome 5 with a 26,468 bp transcript, was initially identifed as a gene closely associated with neurodevelopmental abnormalities in humans [13,14]. But the clinical relevance and biological function of SH3TC2 in human cancer remain largely unclear. In this study, we for the frst time delineated the expression and prognostic landscapes of SH3TC2 in more than 30 cancers. SH3TC2 was shown to be abnormally expressed and signifcantly associated with diseasefree survival (DFS) in many cancers, especially CRC. We then focused on the biological role of SH3TC2 in CRC in this work later. Functionally, the deletion of SH3TC2 signifcantly attenuated the growth ability of CRC in vitro and in vivo. Mechanistically, the m6A recognition protein YTHDF1 binds to and induces up-regulation of SH3TC2 in CRC. Tus, our study advances the understanding of SH3TC2's role in pancancer and provides a reference for developing SH3TC2 as a clinical biomarker and/or target for CRC.

GEPIA Database.
GEPIA is a comprehensive tumor database containing gene expression and prognostic information for more than thirty common human cancers [15]. In this study, the database was used to analyze the expression of related genes in tumor and normal samples, the relationship with DFS of tumor patients, and the expression correlation between genes.

TISIDB Database.
TISIDB is an integrated database for customers to analyze the interactions between tumors and the immune system [16]. It provides a valuable resource for users to explore the immunotherapy of tumors. In this study, we used this database to investigate the relationship between SH3TC2 expression and tumor immune cell infltration in pan-cancer.

Gene Expression Omnibus (GEO)
Datasets. SH3TC2 expression data in four independent CRC cohorts, GSE9348 (12N vs. 32T), GSE32323 (17N vs. 17T), GSE21510 (25N vs. 123T), and GSE39582 (19N vs. 566T), were obtained from the GEO database. According to the expression level of SH3TC2 in tumor and normal samples, the receiver operating characteristic (ROC) curve was used to analyze its diagnostic value in each GEO dataset.

CRC Tissue Specimens and Immunohistochemistry (IHC).
Two independent cohorts were used for validating the expression of SH3TC2 in CRC. Cohort one included parafnembedded tissue samples from 12 CRC patients for IHC staining that were collected from Zhengzhou Central Hospital Afliated to Zhengzhou University between 2018 and 2020 (ZZCH cohort), and this study was approved by the Medical Ethics Committee. Cohort two was a tissue microarray (TMA) containing 30 pairs of CRC samples purchased from Alenabio Company (Xi'an, China) for IHC test (TMA cohort). IHC analysis of parafn-embedded tissues was supported by Servicebio Company (Wuhan, China). Te primary antibodies used in this study were anti-SH3TC2 (Abcam, USA).
2.6. RNA Sequencing. HCT116 cells infected with sh-NC lentivirus or sh-SH3TC2#1 lentivirus were used for transcriptome sequencing (3 replicates per sample). Te sequencing process was similarly described in the previous study [17]. Firstly, total RNAs were extracted by the Trizol method, RNA purity was detected by spectrophotometer, and RNA integrity was analyzed by agarose gel electrophoresis and the Agilent 2100 BioAnalyzer. Te library was constructed using Illumina's NEBNext ® UltraTM RNA Library Prep Kit. Initial quantitative analysis was performed using a Qubit2.0 fuorometer, and the library was diluted to 1.5 ng per μL. Agilent 2100 BioAnalyzer and qRT-PCR were used to detect library quality. Ten, the Illumina platform was used for library sequencing, and a 150 bp paired terminal reading was generated to obtain the sequence information of the fragment to be measured. After quality control and sequence alignment based on the reference genome, DESeqv2 software was used to analyze the diferentially expressed genes (DEGs) between the two groups. Finally, the DEGs were used for gene enrichment analysis based on gene ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG).

Cell Cycle Analysis.
Propidium Iodide (PI) staining kit (7Sea, China) and fow cytometer (Beckman, USA) were used to detect the changes of cell cycle in CRC cells with SH3TC2 deletion. In short, the adherent cells were digested with trypsin solution to collect cell deposits. Te cells were immobilized with precooled 70% ethanol at 4°C for 3 h. Finally, cells were incubated with PI staining solution at 37°C for half an hour and detected by fow cytometry.

Cell Growth Assays.
A colony formation experiment was used to evaluate the in vitro nonpopulation-dependent growth of CRC cells with SH3TC2 defciency. Te cells were seeded in 12-well plates with 800 cells in each well. About 9 to 12 days later, visible cellular colonies appeared in the plates. After fxation and crystal violet staining, the colonies were photographed and analyzed. For the CCK-8 growth assay, the transfected cells were inoculated into 96well plates, and 10 μL of CCK-8 solution was added into the wells and incubated at 37°C for 3 to 4 hours. Finally, the light absorption value of each well at 450 nm was measured by using a microplate reader.
2.9. Animal Experiment. Te in vivo growth ability of CRC cells with SH3TC2 ablation was investigated by a subcutaneous tumor-forming experiment in nude mice. HCT116 cells infected with sh-NC lentivirus or sh-SH3TC2 #1 lentivirus were used to inoculate nude mice with 5 mice in each group. Each BALB/c mouse was injected with 5 × 106 cells. When obvious tumors appeared subcutaneously in nude mice, the formula length × width × width × 3.14/6 was used to calculate the tumor volume. Finally, tumors from nude mice were isolated and used for further analysis. Te study was reviewed and approved by the Ethics Committee of Zhengzhou Central Hospital Afliated to Zhengzhou University.

Western Blot.
A Western blot was used to detect protein levels of corresponding genes in CRC cells or nude mouse tumor tissues. In brief, total protein was extracted using a lysate mixture containing RIPA and protease inhibitors. After concentration determination, an equal amount of total proteins from each sample was used for polyacrylamide gel electrophoresis. After membrane transfer, blocking, antibody incubation, and coloration, protein bands on PVDF membranes were used for detection and analysis. Primary antibodies used in this study include anti-GAPDH (Bioworld, China), anti-SH3TC2 (Bioss, China), anti-CDK4 (Bioworld, China), anti-Cyclin D1 (Bioworld, China), and anti-YTHDF1 (Abcam, USA).

Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR).
Te mRNA expression of SH3TC2 in CRC cells following YTHDF1 silence was detected by qRT-PCR. Te Trizol method was utilized to extract total RNAs from CRC cells. After concentration and purity determination, total RNAs were used to synthesize cDNAs. Ten, the SYBR Green mix kit (DBI, Germany) and 7500 Fast system were utilized for qRT-PCR detection. Te housekeeping gene GAPDH was used as an internal reference, and gene expression wasanalyzed by the 2−ΔΔCt method. Primers are shown in Supplementary Table 1. 2.12. RNA m6A Prediction of SH3TC2. Te RMBase v2.0 online server (https://rna.sysu.edu.cn/rmbase/) based on epitranscriptome sequencing data was employed to predict the m6A modifcation on SH3TC2 mRNA. Te m6A reading protein that may bind with SH3TC2 was analyzed by the m6A2Target algorithm (http://m6a2target. canceromics.org/).

Methylated RNA Immunoprecipitation (MeRIP)
Analysis. Te m6A modifcations on SH3TC2 mRNA were detected by using the MeRIP kit (Millipore, USA). First, the Trizol method was used to extract total RNAs from CRC cells, and the RNA concentration was adjusted to 1 μg per μL for each sample. Te RNA was then segmented with a crushing bufer and incubated with magnetic beads coupled with IgG or m6A antibodies at 4°C for 3 h. Finally, PCR and agarose gel electrophoresis were used to detect m6Alabeled SH3TC2 transcripts. Te primer for MeRIP-PCR is displayed in Supplementary Table 1. 2.14. RNA-Protein Immunoprecipitation (RIP) Assay. Te direct interaction between YTHDF1 protein and SH3TC2 mRNA was detected by the RIP kit (Millipore, USA). In summary, CRC cells were cleaved using a mixture of bufer containing RNA inhibitors and protease inhibitors. Cell lysates and magnetic beads coupled with IgG or YTHDF1 antibodies were then incubated in immunoprecipitation bufer at 4°C for 2 to 4 h. Finally, the immunoprecipitated samples containing RNA were purifed and analyzed by PCR and agarose gel electrophoresis.

Statistical Analysis.
Te student's t-test was used to compare the data diferences between any two groups, and a p value less than 0.05 was considered statistically significant. Te Kaplan-Meier method was utilized for patient survival analysis. SPSS 19.0 and GraphPad 9.0 were used for statistical analysis and data graphing.

Te Expression Panorama of SH3TC2 in Pan-Cancer across 31 Human Cancers.
To explore the potential role of SH3TC2 in human cancer, we for the frst time analyzed the expression landscape of SH3TC2 in 31 cancer types by interrogating the comprehensive cancer database GEPIA. For this, we defned a fold change greater than 1.5 and p value less than 0.05 as a signifcant diference in SH3TC2 expression between tumor and normal samples in each cancer type. We found that SH3TC2 was signifcantly overexpressed in 7 cancers, including bladder urothelial cancer (BLCA), cholangio cancer (CHOL), colon adenocarcinoma (COAD), acute myeloid leukemia (LAML), pancreatic adenocarcinoma (PAAD), rectum adenocarcinoma (READ), and skin cutaneous melanoma (SKCM), while downregulated in 2 cancers, breast invasive carcinoma (BRCA), and testicular germ cell tumors (TGCT), compared with the corresponding normal tissue samples (Figures 1(a)-1(i)). However, there was no signifcant diference in SH3TC2 expression levels between normal and tumor samples of the other 22 cancers (Supplementary Figures 1A-1C). Tese results suggest that SH3TC2 may play a carcinogenic role in some specifc cancers, such as BLCA, CHOL, etc.

Te Prognostic Landscape of SH3TC2 in Human Pan-Cancer.
Given that SH3TC2 is dysregulated in some cancers, we wondered whether SH3TC2 expression could

SH3TC2 Correlates with Tumor Immunity.
Emerging studies have shown that infltrating immune cells in the tumor microenvironment afect tumor progression and therapeutic efect [18,19], and lymphocyte infltration may be used as a prognostic marker of CRC [20]. We are interested to know whether SH3TC2 has some correlation with tumor immune status. Terefore, we used the TISIDB database to comprehensively observe the relationship between SH3TC2 expression and immune cell infltration in pan-cancer. Te results showed that SH3TC2 expression was associated with immune activation in ACC, LUAD, LIHC, etc. and immunosuppression in COAD, READ, etc. (Supplementary Figure 2A). Tumor cells can achieve immune escape by regulating the expression of immune checkpoint molecules (ICMs), and blocking ICMs can enhance the efect of immunotherapy and beneft cancer patients [21,22]. We then interrogated the GEPIA database to analyze the expression correlation between SH3TC2 and a series of wellknown ICMs in the CRC. We found that SH3TC2 was signifcantly correlated with VTCN1, but not with CTLA4, PDCD1, CD28, or CD70 (Supplementary Figures 2B-2F), implying that SH3TC2 may be involved in CRC immunity through VTCN1.

SH3TC2 Serves as a Potential Diagnostic Biomarker in CRC.
Considering that SH3TC2 showed an obviously high expression status in CRC (COAD and READ, Figures 1(c) and 1(f )), we further included four CRC cohorts (GSE9348, GSE32323, GSE21510, and GSE39582) from the GEO database to verify SH3TC2 expression. All these four independent datasets showed that SH3TC2 expression was upregulated in CRC tumor samples (Figures 3(a) and 3(d)).       Next, we determined the diagnostic value of SH3TC2 in these four CRC cohorts by using the receiver operating characteristic (ROC) curve and indicating the area under the curve (AUC). Consistently, the four GEO datasets showed that SH3TC2 had a high diagnostic value in CRC, with AUC values ranging from 0.79 to 0.96 (Figures 3(e)-3(h)), suggesting that SH3TC2 is a potential clinical diagnostic marker of CRC. Subsequently, we detected protein levels of SH3TC2 in two additional CRC cohorts using IHC, and SH3TC2 expression was again elevated in CRC tissues compared to normal tissues (Figures 3(i)-3(j)). Collectively, the above fndings hint that the up-regulated SH3TC2 may play a role in CRC biology.

RNA Sequencing and Flow Cytometry Reveal the Participation of SH3TC2 in CRC Cell-Cycle Regulation.
To clarify the biological role of SH3TC2 in CRC, RNA sequencing was performed to observe the molecular expression changes in HCT116 cells after SH3TC2 deletion. We found that knockdown of SH3TC2 caused 868 genes to be downregulated and 790 genes to be up-regulated (Figure 4(a)). We then performed pathway enrichment analysis based on GO and KEGG for these 868 down-regulated genes. GO results showed that SH3TC2-associated genes were mainly enriched in DNA replication, cell cycle checkpoints, positive regulation of cell cycle processes, mitotic cell cycle checkpoints, etc. (Figure 4(b)). KEGG analysis depicted that SH3TC2 may participate in the cell cycle, colorectal cancer, DNA replication, mismatch repair, etc. (Figure 4(c)). Both GO and KEGG suggest that SH3TC2 may be involved in the regulation of the cell cycle in CRC. Subsequently, fow cytometry confrmed that deletion of SH3TC2 could arrest the cycle of HCT116 and SW480 cells at the G0/G1 phase (Figure 4(d)), indicating that SH3TC2 is required for maintaining cellcycle progress in CRC.

SH3TC2 Promotes CRC Growth Both in Vitro and in Vivo.
Tumor cells maintain their malignant growth through cellcycle regulation. We next investigated the efects of SH3TC2 defciency on CRC cell growth. Results from colony formation experiment showed that knockdown of SH3TC2 attenuated the growth of HCT116 and SW480 cells in vitro ( Figure 5(a)). In addition, the results of a subcutaneous tumorigenesis experiment in nude mice demonstrated that SH3TC2 knockdown signifcantly inhibited the growth of HCT116 cells in vivo, presenting a smaller tumor volume ( Figure 5(b)). We then detected the protein levels of SH3TC2 and two cell-cycle regulators, CDK4 and Cyclin D1, in the tumor tissues from nude mice by Western blot and found that knockdown of SH3TC2 caused a decrease in CDK4 and Cyclin D1 ( Figure 5(c)). Tese data indicate that ablation of SH3TC2 impairs CRC growth both in vitro and in vivo.

Te m6A Reader YTHDF1 Mediates Upregulation of SH3TC2.
Considering that SH3TC2 is highly expressed in CRC and is necessary for sustaining CRC growth, we intended to investigate the molecular mechanism that mediates SH3TC2 up-regulation. A number of recent groundbreaking studies have confrmed that m6A epigenetic modifcation plays a critical role in regulating the stability of intracellular RNA and maintaining RNA pools in cancer cells, which are mainly dependent on the m6A reading protein [23,24]. In view of this, it is interesting to explore whether the high expression of SH3TC2 in CRC is mediated by m6A modifcation. First, we used RMBase v2.0 software to predict and obtain a potential m6A modifcation site at 1669 bp from the 5′-end of SH3TC2 mRNA (Figure 6(a)). We then confrmed the presence of the m6A modifcation on the SH3TC2 transcript in CRC cells by MeRIP-PCR experiment ( Figure 6(b)). Moreover, the m6A enrichment of SH3TC2 transcript in normal colonic epithelial cell NCM460 was lower than that in CRC cells (Figure 6(c)). Next, we used the m6A2Target database to analyze RIP-seqbased data and found that three m6A readers may bind with the SH3TC2 transcript, namely YTHDF1, YTHDF2, and YTHDF3. Subsequently, we used the GEPIA database to investigate the expression levels of these three molecules in CRC and their correlation with the expression of SH3TC2. It was found that both YTHDF1 and YTHDF2, but not YTHDF3, were upregulated in CRC (Figures 6(d) and 6(e), Supplementary Figures 3A-3D). YTHDF1, but not YTHDF2 or YTHDF3, had a signifcant positive correlation with the expression of SH3TC2 (Supplementary Figures 3E-3G). Tis suggests that YTHDF1 may play a key role in regulating SH3TC2 expression. Further RIP-PCR analysis confrmed the direct binding of YTHDF1 protein and SH3TC2 transcript in CRC cells (Figure 6(f )). After silencing YTHDF1, both mRNA and protein levels of SH3TC2 were decreased signifcantly in CRC cells (Figures 6(g) and 6(h)) without afecting the m6A enrichment of SH3TC2 transcript (Supplementary Figure 3H). To further investigate the role of the YTHDF1/SH3TC2 axis in CRC proliferation, we conducted functional rescue experiments. Results showed that overexpression of wild-type YTHDF1 rescued, at least in part, the CRC growth inhibition caused by SH3TC2 knockdown, while transfection of mutant YTHDF1 plasmids did not (Figures 6(i) and 6(j)). Together, these data demonstrate that YTHDF1 regulates the expression of SH3TC2 in CRC and that YTHDF1/SH3TC2 contribute to CRC progression.

Discussion
Advances in high-throughput sequencing technology in recent decades have greatly improved our understanding of the molecular mechanisms of disease [25,26]. In the feld of molecular oncology, changes in molecular expression between tumor and adjacent normal tissues have been detected by microarray or sequencing technology, and hundreds of DEGs have been identifed, accelerating the development of therapeutic targets and diagnostic markers for cancer [27,28]. Nevertheless, most of the key molecules that drive tumor development have not been fully identifed. Here, we report that SH3TC2 is a potential tumor driver and clinical biomarker. Dysregulation of SH3TC2 expression is considered to be associated with clinical spinal malformation and nervous system structural integrity [13,14]. Based on in silico analysis, Yu et al. found that increased expression of SH3TC2-DT/SH3TC2 gene pairs was associated with shorter survival in LAML patients [29]. But the specifc function ofSH3TC2 in human cancer has not been reported. In order to fully investigate the expression status and clinical prognostic value of SH3TC2 in human cancers, we employed the integrated database GEPIA to analyze the expression profle of SH3TC2 in more than 30 types of cancer and its relationship with patient survival. SH3TC2 was found to be dysregulated in 9 cancers, including BLCA, CHOL, COAD, LAML, PAAD, READ, SKCM, BRCA, and TGCT, and correlated with patient DFS in 4 cancers, including COAD, MESO, PAAD, and READ. Tese fndings suggest that there is signifcant heterogeneity in the expression of SH3TC2 in diferent tumors and that SH3TC2 may serve as a therapeutic target or prognostic marker for some specifc cancers. In particular, we noted that SH3TC2 expression was signifcantly reduced in TGCT tumor samples compared to normal samples, and when the threshold was set to 2, SH3TC2 expression was able to distinguish well between 165 normal samples and 137 tumor samples (Figure 1(i)). In future studies, it is possible to develop SH3TC2 as an efective biomarker for early diagnosis of tumors, including but not limited to TGCT, by incorporating large-scale, multicenter clinical blood samples to detect SH3TC2 levels.
Te emergence of immunotherapy is regarded as a new hope in the fght against tumors, including CRC. However, inherent or acquired resistance prevents most CRC patients from benefting from immunotherapy, which is mainly due to the elusive immune escape/suppression mechanisms of    Journal of Oncology tumor [30][31][32]. By interrogating the TISIDB database, we found that SH3TC2 expression was associated with low levels of immune cell infltration in CRC, suggesting that SH3TC2 may be involved in the immunosuppression of CRC. VTCN1, also known as B7-H4, a recently discovered ICM, has been found to be associated with immune cell infltration and poor prognosis for CRC [33,34]. Our results showed that SH3TC2 was signifcantly positively correlated with VTCN1 expression in CRC. Although we have not yet elucidated the exact mechanisms by which SH3TC2 participates in the immunosuppression of CRC and the regulation of VTCN1 by SH3TC2, our bioinformatics results provide the frst clue to further explore the role of SH3TC2 in CRC immunotherapy.
Corresponding to DNA methylation and histone acetylation, m6A modifcation provides a new dimension for understanding posttranscriptional regulation of RNA during tumorigenesis [35]. In this study, we demonstrated that the m6A reader protein YTHDF1 directly binds to the SH3TC2 transcript and up-regulates its expression. Tis explains, at least in part, the molecular mechanism by which SH3TC2 expression is overexpressed in CRC. Recent studies have shown that YTHDF1 plays an important role in chemotherapy resistance and the selfrenewal of CRC [36,37], but the relevant mechanism remains unclear. Our results showed that SH3TC2 promoted the cell-cycle progression of CRC as well as tumor growth and that SH3TC2 was a new downstream target of the YTHDF1 molecule in CRC. In addition, overexpression of YTHDF1 rescued, at least in part, the inhibitory efects of SH3TC2 knockdown on CRC growth and cell-cycle progress. Tese fndings suggest that the YTHDF1/SH3TC2 axis may play a role in the development of CRC, and the regulation network mediated by YTHDF1 needs to be further studied.
In summary, our study reveals for the frst time the expression status and prognostic value of SH3TC2 in pancancer and demonstrates the critical role of SH3TC2 in maintaining the cell-cycle progress and growth of CRC. Moreover, RNA m6A modifcation and YTHDF1 binding may be responsible for the up-regulation of SH3TC2 expression in CRC. Our work provides a scientifc reference for targeting YTHDF1/SH3TC2 against CRC clinically.

Data Availability
Te datasets supporting our results are available in the public databases GEO and TCGA as well as data sources in the method. Te data of our in-house cohort are provided in supplementary tables.