Integrative analysis of PTEN‐related hub genes and validating drug targets for colorectal cancer

Colorectal cancer (CRC) is a heterogeneous disease and one of the most prevalent malignancies worldwide. Previous research has indicated that phosphatase and tensin homolog (PTEN)‐related genes found in CRC may serve as potential biomarkers for individualized treatment options. The present study aimed to examine the association between PTEN‐related genes and the prognosis of CRC patients by evaluating the significance of PTEN‐related hub genes and determining potential mechanisms and genes associated with them. Gene expression profiles and clinical information were obtained from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. At present, PTEN mutations have been identified in 7% of CRC patients, according to the most recent TCGA data. Differential expression analysis revealed 54 genes as differentially expressed genes (DEGs) between PTEN‐related genes and GEO databases (GSE39582 and GSE6263). Further gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted on PTEN‐related DEGs. The prognostic efficacy of the PTEN‐related DEG signature was assessed using Kaplan–Meier survival and receiver operating characteristic curve analyses. Bioinformatics methods were utilized to analyze the correlation between PTEN‐related DEGs and CRC prognosis, survival, and drug efficacy. Through these analyses, eight prognostic‐related PTEN‐related hub genes (PPARGC1A, NTRK2, ANK2, PLCB4, STC2, PLAU, CDKN1A, and HPGDS) were identified and a risk prognosis model was constructed. Notably, NTRK2 and HPGDS were found to affect drug treatment response in CRC. Targeting these prognostic‐related PTEN‐related hub genes can regulate cell death signaling, which may benefit the prognosis of CRC patients and improve drug sensitivity.

Initially discovered as a tumor suppressor gene frequently mutated on chromosome 10q23, PTEN mutations are estimated to occur in 50%-80% of several sporadic tumors, 6 including endometrial carcinoma, glioblastoma, and prostate cancer. 7Breast, colon, and lung tumors have mutation rates ranging from 30% to 50%. 8The hemizygous loss of PTEN leads to cancer progression in the prostate, colon, lymphatic system, mammary gland, and endometrium. 9,10udies conducted on mice have demonstrated that even a slight decrease in PTEN expression is sufficient to enhance cancer susceptibility. 11Therefore, different PTEN statuses may influence cancer progression, disease prognosis, treatment strategies, and patient responses to treatment measures.By exploring alterations in patients with PTEN mutations and evaluating their significance in disease progression, we can gain a better understanding of CRC pathogenesis and provide additional evidence for individualized treatment options.
In this study, we employed bioinformatics analysis approaches to analyze an RNA sequencing (RNA-Seq) data set of CRC and the GDSC database. 12We aimed to identify key pathways and genes associated with PTEN mutations and evaluate their relevance in drug selection.
We expect that these findings will uncover the potential role of PTEN mutations in CRC as predictive factors for prognosis and individualized treatment options.

| RNA-Seq data
Answer: The somatic mutation data and the clinicopathological information of patients with colorectal adenocarcinoma were obtained from the cBioPortal platform (http://www.cbioportal.org/datasets).The copy number alteration (CNA), mutation types, and alteration frequency of PTEN across colorectal adenocarcinoma can be observed under the "TCGA, Firehose Legacy" section.The data on CNA acquired from 220 colorectal adenocarcinoma patients (TCGA, Firehose legacy) were retrieved and downloaded from the cBioPortal (http://www.cbioportal.org).Three sets of samples were analyzed as follows: colon adenocarcinoma (n = 129, 58.6% percentage of samples), rectal adenocarcinoma (n = 68, 30.9% percentage of samples), mucinous adenocarcinoma of the colon and rectum (n = 23, 10.5% percentage of samples).In total, 7% PTEN mutation was observed in CRC patients.Two external independent data sets, namely, GSE6263 and GSE39582, were downloaded from the GEO database and quantitated by Affymetrix Human Genome U133 Plus 2.0 Array.Two external independent data sets, namely, GSE6263 and GSE39582, were downloaded from the GEO database and quantitated by Affymetrix Human Genome U133 Plus 2.0 Array.In the meantime, co-expression genes in PTEN mutation of 220 CRC patients were downloaded from the cBioPortal (https:// cbioportal.org/)database as the independent external validation cohorts.

| Identification of differentially expressed genes
Effects of PTEN mutation on gene expressions in CRC cells were downloaded from Gene Expression Omnibus (GEO, Reference Series: GSE6263 and GSE39582), and differentially expressed genes (DEGs) were analyzed by GEOR2.DEGs were identified with the following criterion: jfold change (FC)j >1; both the p value and false discovery rate (FDR) <0.05.The DEGs were used for further bioinformatics analysis.

| Venn diagram
PTEN-related genes were acquired from the GeneCards database (https://www.genecards.org/), and a total of 6002 PTEN-related genes were identified.In the meantime, interactions of DEGs identified from GSE6263 and GSE39582, and PTEN-related genes were shown using the Venn diagram.

| Enriched pathway analysis
After taking the intersection by Venn diagram, 54 differentially expressed PTEN-related genes, and genes down-or upregulated in 54 PTEN-related DEGs were finally identified by heatmap and PCA, which were used for functional annotation of gene ontology (GO) and pathway enrichment analysis of Kyoto Encyclopedia of Genes and Genomes (KEGG).GO enrichment was carried out at three levels: cellular component (CC), biological process (BP), and molecular function (MF).The FDR q < 0.01 was considered statistically significant.Enrichment analysis was performed using the OECloud tools at https://cloud.oebiotech.com.

| Establishment and validation of PTEN-related hub gene by prognostic signature
To further evaluate the predictive ability of the PTEN-related hub genes, using the "ROC curves" package, we plotted receiver operating characteristic (ROC) curves of PTEN-related genes.We established a PTEN-related gene prognostic signature comprising 18 critical genes.
Next, top genes were used to construct correlation heatmaps.To determine the mechanism by which PTEN-related hub gene affected the prognosis of CRC patients, 18 differentially expressed PTENrelated genes were subjected to Cox regression analysis.We evaluated the prognostic performance of PTEN-related gene prognostic signature using Kaplan-Meier (KM) analysis and compared them in the high-and low-risk groups.Furthermore, the critical PTEN-related hub genes were identified for GO Classification analysis and pathways.

| Establishment and validation of hub genes for prognostic signature
We then analyzed the single-cell type of PTEN-related hub genes in the HPA database, which is divided into 10 types of cell clusters.In addition, using the "ROC plotter," we plotted time-dependent ROC curves to evaluate the predictive effect of the risk signature in predicting the survival of CRC patients.Subsequently, univariate and multivariate Cox regression analyses were used to detect whether the risk score could be utilized as an independent risk factor for survival prediction in CRC patients.

| Molecules interaction study
Molecular docking is used to predict the binding conformation of a ligand with the suitable target protein.In the present study, molecular docking was performed using the software of AutoDock Maestro 12.8.The structures of NTRK2 (PDB ID = 4asz) and HPGDS (PDB ID = 6w58) were obtained from the Protein Data Bank.First, we explored suitable binding sites on NTRK2 and HPGDS dimmers and performed the docking studies with the flexible ligand (larotrectinib or tranilast) in Drugbank.Three-dimensional docking analysis binding NTRK2 with larotrectinib.The two-dimensional docking analysis binding larotrectinib in the active site of NTRK2, and the formed hydrogen bonds with amino acid residues near the active site were used for evaluation of its biological functions.The interaction forces between individual molecules were analyzed using the computer software.that were formulated into liposomes (Oligofect AMINE, Life Technologies, Carlsbad, CA).Then, the final volume of culture medium was added to 2.0 mL per well.The cells were monitored by microscopy and harvested for biochemical analyses at 24, 48, and 72 h after transfection.

| Quantitative real-time polymerase chain reaction analysis
The mRNA expression levels of PTEN mutation-related genes NTRK2 were detected in HCT116 cells.Quantitative real-time polymerase chain reaction (qRT-PCR) was used to analyze the quantitative expression of key prognostic genes.Total RNA was extracted from HCT116 cells, and the Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, Inc., Waltham, MA, USA) was used to measure and evaluate the concentration and quality of samples.The master mix Kit (Takara, Shiga, Japan) was used to obtain the cDNA by reverse transcription.Step one plus real-time PCR system (BIO-RAD, CFX96TM Real-time System, C1000TM Thermal Cycle, USA) was used to amplify the mRNA.β-actin was used as the internal reference.The sequences of the primers (Sangon Biotech Shanghai Co., Ltd.) were as follows: NTRK2 Forward 5 0 -CTGGTCTTGGGCTTCTGGAG-3 0 and Reverse 5 0 -CGTGATGT TCTCCGG GTCAACG-3 0 ; PTEN Forward, 5 0 -AAGGATCCCCAGACATGACAGCCAT C ATC-3 0 and Reverse 5 0 -CACAACTCG AGTCAGACTTTTGTAATTTGTG TATGC-3 0 .Actin: Forward 5 0 -GATCTGGCA CCACACCTTCT-3 0 , Reverse 5 0 -GGGGTGTTGAAGGTCTC AAA-3 0 .The PCR reactions were performed in triplicate, and the relative mRNA expression levels were normalized to the expression of β-actin.The specific amplification of target genes was validated using a dissociation curve.

| Western blot analysis
The protein levels of NTRK2, HPGDS, and PTEN in HCT116 cells that treated with larotrectinib (NTRK2 inhibitor derived from Loxo Oncology, Inc.) or tranilast blocker (HPGDS inhibitor, Apexbio, USA, A1663) were detected by Western blot.Briefly, proteins in treated cells were extracted using the whole protein extraction reagent.SDS-PAGE gel was used to separate proteins, and then proteins were transferred to PVDF membrane.After blocking with 5% BSA or 5% skim milk for 2 h, the membrane was incubated with primary antibodies overnight at 4 C: anti-NTRK2 and GAPDH (Cell Signaling Technology, Danvers, MA, USA), anti-HPGDS (ab183511, Abcam, UK), and anti-PTEN (ab170941, Abcam, UK).After incubation, the membrane was washed three times (5 min per time) using TBST, and then the membrane was incubated with the secondary antibody (anti-NTRK2, anti-HPGDS, and anti-PTEN) at room temperature for 1.5 h (the concentration of secondary antibody was 1:1000).Then, the membrane was washed three times (5 min per time) using TBST again.The bands were tested by enhanced chemiluminescent reagents (Vazyme, Nanjing, China) and exposed to x-ray films (Bio-Rad Laboratories, USA).

| Statistical analysis
The Student's t test was used to compare the mRNA expression levels between different groups.The KM method with log-rank test was used to calculate the clinical outcome between different PTEN groups by GraphPad.FDR in edgeR and GSEA were adjusted for multiple testing with the Benjamini-Hochberg procedure to control FDR, respectively. 13p Value <0.05 was considered statistically significant for all experiments.In total, 7% PTEN mutation was observed in CRC patients (Figure 1A), and the mutation types mainly included truncating, deep deletion, and missense mutations spanning over the entire gene (Figure 1B).
We next investigated the influence of PTEN mutation on CRC progression and prognosis.We first determined the PTEN expression in CRC based on sample types from the UALCAN website.Results showed that PTEN was downregulated in CRC tumors compared to normal tissue (Figure 1C).Analysis of the relationship between the PTEN status and disease prognosis showed that patients with PTEN mutation had worse prognosis (Figure 1D), which indicated that the PTEN mutation may contribute to CRC progression.

| GO and KEGG analyses
Twenty-one upregulated PTEN-related DEGs were used for GO classification analysis, which suggested that biological adhesion, biological regulation, cellular process, and cellular component organization or biogenesis were significantly enriched (Figure 2D,E).Furthermore, in the KEGG pathway classification analysis, 21 upregulated PTENrelated DEGs were significantly enriched in the Immune system, Cancer: overview, Signal transduction, and Cell growth and death (Figure 2F).In the KEGG Enrichment analysis, p53 signaling pathway and TGF-beta signaling pathway were significantly enriched (Figure 2G).
In contrast, 15 downregulated PTEN-related DEGs were used for pathway enrichments, and biological adhesion and biological regulation were significantly enriched (Figure 3A,B).Furthermore, in the KEGG pathway classification analysis, DEGs were also significantly enriched in pathways such as Cancer: overview, Signal transduction, and Cell growth and death (Figure 3C).In the KEGG Enrichment analysis, DEGs were mainly enriched in Focal adhesion, AGE-RAGE signaling pathway, Proteoglycans in cancer, and Pathways in cancer (Figure 3D).

| The relationship between hub genes and clinical features in CRC
To further evaluate the predictive ability of the PTEN-related hub genes, the ROC curves of hub genes were performed.In the training cohort, the areas under the ROC curves (AUCs) were 0.67, 0.83, and 1.0, respectively, indicating that the PTEN-related 18 hub gene signature may serve as a prognostic indicator for CRC patients (Figure 4A-R).
In addition, according to the Pearson correlation test, there were prominently positive or negative correlations among PTENrelated DEGs (Figure 5A).Plots of dilution curves and Rank abundance charts in routine sample diversity analyses for the PTEN-related 18 hub gene (Figure 5B), which were constructed heatmaps or PCA diagrams were created to obtain the up-or downregulated DEGs in the presence or absence of PTEN in HCT116 cells (Figure 5C).

| GO and KEGG analyses
Furthermore, 18 critical PTEN-related hub genes were identified for GO classification analysis, and pathways such as biological adhesion and biological regulation were significantly enriched (Figure 5D,E).

| Correlation between PTEN-related key hub gene prognostic signature and clinical characteristics
To determine the mechanism by which PTEN-related hub gene affected the prognosis of CRC patients, 18 differentially expressed PTEN-related genes were subjected to Cox regression analysis.We evaluated the prognostic performance of PTEN-related gene prognostic signature using KM analysis and compared them in the high-and low-risk groups.PLCB4 and SCT2 expression was both significantly higher in CRC samples than in normal samples (Figure 6A), and the PPARGC1A, PLCB4, and SCT2 expression and gender were significantly correlated to overall survival (OS) (Figure 6B).Then, the heatmap and Pearson correlation diagram were created to show expression levels and correlations among hub genes (Figure 6C,D).On the basis of sample abundance data, it indicated that PPARGC1A, NTRK2, ANK2, PLCB4, and SCT2 were the abundant genes in CRC samples (Figure 6E).In the KEGG Enrichment top 20 analysis, PTEN-related key hub genes enriched in NTRK2 actives signals, TFAP2 family regulates transcription, activation of the PPARGC1A, FOXO-mediated transcription, and collagen formation (Figure 6F).

| Establishment and validation of hub genes for prognostic signature
Similarly, we performed analyses in the validation cohort and the training cohort.Furthermore, stratification analyses based on risk score showed that the OS of low-risk patients was better than that of high-risk patients (Figure 7A).According to the KM survival curves, high-risk of PLCB4, CDKN1A, PLAU, and HPGDS patients showed a lower OS than low-risk patients (Figure 7A), illustrating the distribution of risk scores and survival status of CRC patients.
Next, we analyzed the expression levels of PTEN-related hub genes in single-cell data in the HPA website, in which cells are divided into 10 different types in colon.As shown in Figure 7B, in T cell, the expressions of PLAU and HPGDS were higher than those of NTRK2, PLCB4, SCT2, and CDKN1A.These results suggested that PTEN-related hub genes may closely relate to immune cells in CRC.
In the training cohort, the AUC of NTRK2 for predicting the OS of capecitabine chemotherapy was 0.686, and fluoropyrimidines monotherapy OS of NTRK2 was 0.671, respectively, indicating that the NTRK2 may serve as a prognostic indicator for CRC patients' chemotherapy (Figure 7C).

| Molecular interaction analysis
First, we looked for suitable binding sites on NTRK2 and HPGDS dimmers and performed the docking studies with the flexible ligand (larotrectinib or tranilast) in Drugbank.Using AutoDock Maestro 12.8 software, we docked these core molecular targets (NTRK2 and HPGDS), and the predicted target of larotrectinib and NTRK2 is presented by three-dimensional structure from PDB (representative) (Figure 8A,B).Studies have shown that lower binding energy of receptors and ligands suggest more stable binding conformation and higher possibility of action.As shown in Figure 8C, larotrectinib may bind to NTRK2 and form hydrogen bonds with amino acid residues TRY-574, which is close to the active site.In addition, it can be observed that larotrectinib had the docking score (À4.509) with NTRK2, and larotrectinib exhibited the most stabilized interaction with NTRK2 (Figure 8C).The 2D and 3D interaction diagrams of tranilast within the active site of HPGDS demonstrate the formation of two hydrogen bonds between hydroxyl groups and TRP-104 and ALA-105, respectively (Figure 8D).Notably, tranilast exhibited a favorable docking score of À9.52 with HPGDS, indicating a highly stabilized interaction between the two entities (Figure 8E,F).

| In vitro validation of PTEN-related oncogenic function and expression in CRC cells
PTEN expression was knocked down using specific PTEN siRNA in HCT116 cells (Figure 9A).As shown in Figure 9B, we found that NTRK2 was decreased in HCT116 cells with PTEN knockdown.It suggested that PTEN silence caused a significant increase of NTRK2 in HCT116 cell lines.Moreover, NTRK2 siRNA selectively silences NTRK2 expression with upregulation of PTEN; however, pCMV3-NTRK2 plasmids selectively enhance NTRK2 expression in HCT116 cell lines with PTEN downregulation.
In order to further investigate the biological role of NTRK2 and HPGDS in CRC cells, our data showed that NTRK2 expression was decreased in HCT116 cells by treatment with larotrectinib and PTEN upregulation (Figure 9C).The results were also observed HPGDS downregulation in HCT116 cells that treated with tranilast (Figure 9D).Furthermore, the cell growth in HCT116 cells was distinctly inhibited by downregulating the expression of NTRK2 or HPGDS that treated by larotrectinib or tranilast, compared with the control cells (Figure 9E).The above data demonstrated that NTRK2 and HPGDS were involved in CRC progression by inhibiting cell growth.proliferation.Notably, PTEN protein loss is observed in approximately 30% of sporadic CRC cases.The impact of PTEN on CRC lies in its ability to modulate the expression of downstream genes that directly affect the progression of the disease. 14,15Deletions and/or mutations in the PTEN gene have been associated with various human cancers, including CRC.In CRC patients, those with high expression of the PTEN gene generally exhibit a favorable prognosis. 16Some researchers have suggested that the PTEN gene plays a crucial role in inhibiting tumor-regulated protein kinase B (PKB), which may explain the correlation between high PTEN expression in CRC patients and better prognoses. 17In this study, our objective was to evaluate the clinical significance of PTEN deletions and/or mutations in CRC progression, prognosis, and drug selection, in order to facilitate individualized treatment approaches (Figure 10).Among the 220 cases analyzed, we observed PTEN mutations in approximately 7% of patients, including amplifications, truncations, deep deletions, in-frame mutations, and missense mutations spanning the entire gene.The NTRK gene family, consisting of NTRK1, NTRK2, and NTRK3, encodes tropomyosin receptor kinases (TRK) that activate cell proliferation, differentiation, apoptosis, and neuronal survival pathways via PI3K, RAS/MAPK/ERK, and phospholipase C-gamma signaling transduction. 18,19Dysregulation of NTRKs gene function is widely recognized to play an oncogenic role in various cancers.Notably, NTRKs gene fusion is the most well-characterized aberration that stimulates tumorigenesis through the constitutive activation of downstream cell growth and proliferative pathways. 19In multiple cancers, NTRK2 has been shown to function as an oncogene, 20,21 and increased expression of NTRK2 has been linked with poor outcomes. 22,23ile NTRK gene fusions are highly prevalent in certain rare cancers such as secretory breast carcinoma (>90%), they are rare in more common cancers like lung, sarcoma, and CRC (<1%). 24Consequently, basket designs, which enroll patients across multiple tumor sites, were implemented to ensure an adequate patient population for clinical trial designs to accurately assess efficacy. 25In 2018, larotrectinib became the first FDA-approved drug targeting NTRK gene fusion-positive tumors.Pooled analysis of three phase I and II single-arm trials, involving a combined total of 55 pediatric and adult patients, demonstrated an overall response rate of 75% (95% confidence interval: 61%-85%, independent review) for larotrectinib. 26The efficacy of larotrectinib, a TRK-selective inhibitor, has been investigated in three clinical trials: a phase I trial in adults (NCT02122913), a phase I/II trial in pediatric patients (SCOUT, NCT02637687), and a phase II trial in adults and adolescents (NAVIGATE, NCT02576431). 27,28r findings indicate that NTRK2 serves as a tumor suppressor gene in CRC, which aligns with a recent study demonstrating the antitumor effects and cellular apoptosis induced by inhibiting TRKB, encoded by NTRK2. 29,30Furthermore, NTRK2 exhibited significant associations with OS and disease-free survival across multiple data sets, 22,23 suggesting its potential as a promising biomarker for CRC patients with PTEN mutations.Hematopoietic prostaglandin D synthase (HPGDS) is a σ class glutathione transferase 31,32 that was discovered 40 years ago.It plays a role in the arachidonic acid metabolic pathway and catalyzes the production of prostaglandin D2 (PGD2). 33HPGDS is involved in fatty acid metabolism. 34In lung cancer, HPGDS promotes tumor cell apoptosis and inhibits invasion. 35Recently, potent small-molecule H-PGDS inhibitors have been developed as potential therapies for PGD2-mediated diseases.These inhibitors incorporate an amide or imidazole "linker" between the pyrimidine or pyridine "core" ring and F I G U R E 1 0 The flowchart of study by bioinformatics analysis approaches to identify key genes associated with PTEN mutations and evaluate their relevance in drug selection.PTEN, phosphatase and tensin homolog.

| DISCUSSION
the "tail" ring system.The amide linker may stabilize the synthase-GSH-inhibitor complex, as suggested by a lack of significant change in thermal shift observed with the imidazole analogs. 31CB4 encodes PLCβ4 protein, a phospholipase C (PLC) isoform.
PLCs catalyze PtdIns(4,5)P2 to generate diacylglycerol (DAG) and inositol 1,4,5 trisphosphate (InsP3), which are important second messengers in signal transduction. 36,37In CRC, downregulation of PLCB4 is associated with oxaliplatin (OXA) resistance.[40][41][42][43] STC2 is implicated in glucose homeostasis, phosphorus metabolism, and various malignancies.Studies have consistently reported higher STC2 levels in CRC tissues compared to normal tissues, with correlations observed with tumor size, histological grade, lymph node metastasis, lymphatic invasion, and tumor depth. 44Moreover, elevated serum STC2 levels have been correlated with poor patient survival and pathological grade, suggesting the potential of STC2 as a serum biomarker. 45Further investigations have shown that STC2 promotes epithelial-mesenchymal transition (EMT), migration, and chemoresistance in CRC, involving activation of the ERK/MEK and PI3K/ AKT signaling pathways. 45,46In addition, STC2 has been associated with anti-VEGF resistance and chemoresistance to oxaliplatin. 47,48ese findings highlight the clinical significance of STC2 in CRC and its involvement in EMT and chemoresistance through multiple potential pathways.Further studies are necessary to confirm STC2 as a therapeutic target. 49Notably, STC2 expression in high-grade serous cancer is significantly associated with tumor grade and histopathological type, indicating unfavorable outcomes. 50Overall, we report STC2 as an effective diagnostic biomarker for CRC, demonstrating comparable efficiency to hub genes.
In conclusion, our study highlights the potential roles of NTRK2, PLCB4, SCT2, CDKN1A, PLAU, and HPGDS genes in CRC cooccurrence, providing candidate targets and strategies for individualized

2. 8 |
Cell samples and quantitative real-time polymerase chain reaction 2.8.1 | Plasmid and PTEN RNA silencing For transfection-based co-immunoprecipitation assays, HCT116 cells were transfected with the indicated plasmids using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA), lysed in 0.5 mL of lysis buffer, and immunoprecipitated with Protein G Plus/Protein Agarose Suspension beads (Calbiochem-Novabiochem Corp, San Diego, CA, USA) for 3 h at 4 C. Independent siRNA sequences were used to silence PTEN expression.The controls included liposomes formulated in the absence of siRNA.The siRNA concentration was 0.42 μg per 1.0 Â 10 5 cells.The cells were subcultured and transfected using synthetic, high-performance liquid chromatography-purified PTEN siRNAs or pCMV3-NTRK2 plasmids (Invitrogen, Carlsbad, CA, USA)

3 | RESULTS 3 . 1 |
Data information and clinical impact of PTEN mutation in CRC Related information of 640 CRC patients and corresponding cancer tissue RNA-Seq data sets from the TCGA database were collected, and mutation and CNA data (220 samples/patients) with complete follow-up profiles of CRC samples were acquired from the cBioPortal.

F
I G U R E 1 Data information and clinical impact of PTEN Mutation in CRC.(A, B) Mutation frequency (A) and types (B) of PTEN in CRC reproduced from The Cancer Genome Atlas database.(C) PTEN downregulated in primary CRC patients' tumor tissue.(D) Analysis of the relationship between the PTEN status and disease prognosis on survival.(E) The volcano plot, comparison of the expression of the DEGs in HCT116 human colon cancer cells according to GEO, Reference Series: GSE6263.(F) The volcano plot, comparison of the expression of the DEGs in the CRC samples (GSE39582).(G) The heatmap was constructed for the upregulated DEGs and downregulated DEGs in GSE6263.(H, I) The intersection by Venn diagram, 54 differentially expressed PTEN mutation-related genes were finally identified and were constructed heatmap to obtain the intersection of 54 DEGs.CRC, colorectal cancer; DEG, differentially expressed gene; GEO, Gene Expression Omnibus; PTEN, phosphatase and tensin homolog.

F
I G U R E 2 Legend on next page.WANG ET AL.

F I G U R E 2
The up-or downregulated DEGs among the CRC samples and functional enrichment analysis of upregulated PTEN-related DEGs.(A) The 54 differentially expressed PTEN mutation-related genes were constructed heatmaps to obtain all the up-or downregulated DEGs among the GEO data sets (GSE6263) and CRC patients (GSE39582), respectively.(B, C) PCA diagrams were created to obtain all the up-or downregulated DEGs among the two GEO data sets.(D) Analysis of the 21 upregulated PTEN-related DEGs were identified for GO Classification analysis.(E) GO enrichment terms of upregulated PTEN-related DEGs.(F) The KEGG pathway classification analysis of upregulated PTEN-related DEGs.(G) Analysis of the KEGG Enrichment top 20, upregulated PTEN-related DEGs ranked by degree were identified in the KEGG pathway.CRC, colorectal cancer; DEG, differentially expressed gene; GEO, Gene Expression Omnibus; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; PTEN, phosphatase and tensin homolog.

F
I G U R E 3 Functional enrichment analysis of downregulated PTEN-related DEGs for GO and KEGG pathway classification analysis.(A) Analysis of the 15 downregulated PTEN-related DEGs were identified for GO classification analysis.(B) GO enrichment terms of downregulated PTEN-related DEGs.(C) The KEGG pathway classification analysis of downregulated PTEN-related DEGs.(D) Analysis of the KEGG Enrichment top 20, downregulated PTEN-related DEGs ranked by degree were identified in the KEGG pathway.DEG, differentially expressed gene; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; PTEN, phosphatase and tensin homolog.F I G U R E 4 Construction of the PTEN-related hub gene prognostic signature gene prognostic signature.(A-R) Predictive ability of the prognostic signature in the training cohort and validation cohort.PTEN, phosphatase and tensin homolog.

F I G U R E 5
Pearson correlation test, heatmaps for PTEN-related up-or downregulated DEGs and functional enrichment analysis of PTENrelated 18 hub genes.(A) Pearson correlation test between the PTEN-related DEGs at the mRNA levels.(B) Plots of dilution curves and Rank_abundacne charts in routine sample diversity analyses for the PTEN-related 18 hub gene.(C) Heatmaps were created to obtain the up-or downregulated DEGs in the presence or absence of PTEN in HCT116 human colon cancer cells.(D) Analysis of the PTEN-related 18 hub genes were identified for GO classification analysis.(E) GO enrichment terms of PTEN-related 18 hub genes.(F) The KEGG pathway classification analysis of PTEN-related 18 hub genes.(G) Analysis of the KEGG Enrichment top 20, PTEN-related 18 hub genes ranked by degree were identified in the KEGG pathway.DEG, differentially expressed gene; PTEN, phosphatase and tensin homolog.

F
I G U R E 6 PTEN-related key hub gene prognostic signature and clinical characteristics.(A) The prognostic performance of PTEN-related gene prognostic signature that compared them in the high-and low-risk groups.(B) KM survival curves of OS for CRC patients between the training cohort and validation cohort.(C, D) The heatmaps or Pearson correlation diagram were created to obtain the PTEN-related key hub genes.(E) The sample abundance data for sample 3 indicated that PPARGC1A, NTRK2, ANK2, PLCB4, and SCT2 were the abundant genes.(F) The KEGG Enrichment top 20 analysis of PTEN-related key hub genes.CRC, colorectal cancer; DFS, disease-free survival; KM, Kaplan-Meier; OS, overall survival; PTEN, phosphatase and tensin homolog.F I G U R E 7 Legend on next page.
PTEN is a crucial regulator gene involved in the dephosphorylation process.It exerts an influence on the expression of numerous genes, including P53 and PDK1, among others.In the context of CRC, PTEN plays a pivotal role in the development of cancer cell F I G U R E 7 Establishment and validation of hub genes for prognostic signature.(A) The KM survival curves of hub gene patients illustrate the distribution of risk scores and survival status of CRC patients.(B) The single cell type of PTEN-related hub gene in the HPA database, which showed the PTEN-related six hub gene expressions and number of cell types in colon, with the distribution and number of various tumor microenvironment-related cells presented.(C) In the training cohort, the AUCs for the capecitabine chemotherapy OS of NTRK2 and Fluoropyrimidines monotherapy OS of NTRK2.AUC, area under the curve; CRC, colorectal cancer; KM, Kaplan-Meier; OS, overall survival; PTEN, phosphatase and tensin homolog.

F I G U R E 8
The molecular interaction analysis.(A) The chemical structure of larotrectinib.(B) Three-dimensional docking analysis binding NTRK2 with larotrectinib.Larotrectinib had the docking score (À4.509) with NTRK2, and larotrectinib exhibited the most stabilized interaction with NTRK2.(C) The two-dimensional docking analysis binding larotrectinib in the active site of NTRK2.Larotrectinib may bind to NTRK2 and form hydrogen bonds with amino acid residues TRY-574 and near the active site to exert its biological functions.(D) The two-dimensional docking analysis binding tranilast in the active site of HPGDS.Tranilast in the active site of HPGDS revealed the formation of two hydrogen bonds between hydroxyl groups with TRP-104 and ALA-105.(E) The chemical structure of tranilast.(F) Three-dimensional docking analysis binding HPGDS with tranilast.Tranilast had the docking score (À9.52) with HPGDS, and tranilast exhibited the most stabilized interaction with HPGDS.HPGDS, hematopoietic prostaglandin D synthase.

F I G U R E 9
The mRNA and protein expression levels of PTEN, NTRK2, and HPGDS in CRC cell lines.(A) Transfection efficiency of PTEN siRNAs in CRC cells.The mRNA and expression levels of PTEN and NTRK2 in CRC cell lines were tested by real-time PCR.(B) Transfection efficiency of NTRK2 plasmids and NTRK2 siRNAs in CRC cells.The mRNA and expression levels of PTEN and NTRK2 in CRC cell lines were tested by real-time PCR.*p < .05,**p < .01,versus control.The above data represent the means ± SD from triplicate measurements.(C, D) Western blotting analysis NTRK2, PTEN, and HPGDS expression in the HCT116 cells.(E) The biological role of larotrectinib or tranilast on CRC cell proliferation.CRC, colorectal cancer; HPGDS, hematopoietic prostaglandin D synthase; PCR, polymerase chain reaction; PTEN, phosphatase and tensin homolog.
treatment.These hub genes indicate that tumors with PTEN mutations tend to exhibit increased cell growth, cellular migration, and immune system activity.Our study was limited to investigating the influence of PTEN mutations on disease progression, prognosis, and drug selection in CRC.Further research in clinical and molecular biology experiments is needed to understand the mechanism and validation of PTEN mutations in CRC.Additionally, our study identifies key pathways and novel genes associated with PTEN mutations in CRC, offering insights that may improve CRC risk prediction and guide the development of therapeutic strategies for these specific subtypes of CRC.AUTHOR CONTRIBUTIONS Gang Wang designed experiments; Gang Wang carried out experiments; Zhu Zhi-Min prepared Figures 1-7.Xiao-Na Xu analyzed experimental results.Jun-Jie Wang analyzed sequencing data and developed analysis tools.Gang Wang wrote the manuscript; Feng Shi and Xing-Li Fu revised the manuscript.