Establishment And Validation Of A Risk Model For Biochemical Recurrence Of Prostate Cancer Based On Glutamine-Related Genes

Objective: This study aimed to explore the predictive ability of a prostate cancer biochemical recurrence(BCR)research model based on glutamine-related genes in a prostate cancer cohort,and to reveal the differences in mutation background,immune infiltration background,and drug sensitivity between the high-risk and low-risk groups.Furthermore,this study aimed to gain a deeper understanding of the potential mechanism of glutamine in prostate cancer,and provide a new perspective for the treatment of prostate cancer.Methods: 1.Acquisition of differential genes in prostate cancer cohort: The m RNA expression data of prostate cancer and normal prostate tissues were downloaded from The Cancer Genome Atlas(TCGA)database.Differential genes between prostate cancer and normal prostate tissues were generated using the "limma" package in R software(screening criteria: |log2 Fold Change(FC)|>1 and adjusted(adj.)P<0.05).2.Acquisition of differential genes related to glutamine in prostate cancer: Genes related to glutamine were downloaded from the Gene Set Enrichment Analysis(GSEA)database.The differential genes were obtained by intersecting with genes related to glutamine,resulting in differential genes of glutamine in prostate cancer and normal prostate tissues.3.Screening of glutamine differential genes related to prostate cancer prognosis: Differential genes significantly associated with Progression-Free Interval(PFI)were screened by single-variable Cox analysis and Lasso regression analysis(using "glmnet" package and "survival" package in R,respectively)(P<0.05).4.Construction of prediction model: The expression values of the genes obtained by Lasso regression analysis were multiplied by corresponding coefficients,resulting in risk scores for each patient.Patients were then divided into high-and low-risk groups based on the median risk score,thereby establishing the prediction model.5.Evaluation of model performance using ROC curve: The "time ROC" package in R was used to evaluate the predictive ability of the model.The "ggplot2" package was used to visualize the data.6.Independent prognostic analysis:Single-factor Cox analysis and multiple-factor analysis were performed using the "survival" package to determine whether the risk score could be an independent prognostic factor for BCRFS in prostate cancer patients.7.External validation of the model: An external dataset(GSE70769)from the Gene Expression Omnibus(GEO)database was used to validate the predictive ability of the model.8.Exploration of mutations in five key genes and their impact on prostate cancer prognosis: A search was conducted in the c Bio Portal database(www.cbioportal.org/)using "prostate cancer" as the keyword.The "TCGA,pancancer atlas;493 patients/samples" dataset was selected for analysis.Subsequently,the genetic changes and mutation locations of these five genes were explored,and the survival prognosis between mutation and non-mutation groups was analyzed.9.Immunological evaluation: The CIBERSORTx,ESTIMATE,and ss GSEA algorithms were used to evaluate differences in immune cell infiltration between high-and low-risk groups.10.Analysis of drug sensitivity: Differences in therapeutic drugs between high-and low-risk groups were explored using the Cancer Drug Sensitivity Genomics(GDSC)and CELLMiners databases.11.q RT-PCR clinical sample verification: The expression trend of differential glutamine genes in clinical samples was detected using real-time quantitative polymerase chain reaction(q RT-PCR)to determine whether it was consistent with the expression trend in the TCGA database.12.Knockdown of GLUL and ASNS in prostate cancer cell lines(DU145 and 22Rv1)using si RNA to observe cell proliferation in the presence or absence of glutamine.Results: 1.Screening and modeling of glutamine-related genes: In the TCGA prostate cancer cohort,a total of 5926 genes were obtained from the "limma" package that differed between prostate cancer tissue and normal prostate tissue.A total of 91 glutamine-related genes were obtained from GSEA database.Forty glutamine-related differential genes were obtained by taking the intersection set.KEGG and GO analysis suggested that these 40 genes were enriched in the process of amino acid,glutamine synthesis and metabolism.The risk coefficients of five glutamine-related genes(ATCAY,GLUL,ASNS,CAD,FPGS)were obtained by lasso regression analysis,and the expression and associated risk coefficients of these five genes were used to construct a risk score prediction model.For prostate cancer patients from TCGA and GEO(GSE70769),we classified all patients into high and low risk groups based on the median value of the sub risk score.In the TCGA training set cohort,the BCRFS was significantly shorter in the high-risk group than in the low-risk group,and similar results were obtained in the validation cohort.At the same time,it was found that the mutation of these five key genes would also lead to shorter overall survival,progression-free survival,and disease-specific survival of the mutated group(P<0.05).And through single-factor and multi-factor independent prognostic analysis,the risk score model can independently predict BCRFS of prostate cancer patients,P<0.05.2.Differential characteristics between high and low risk groups: immune assessment analysis revealed that the high-risk group had a higher proportion of infiltrating Tumor-associated macrophages(TAM),Regulatory(Tregs)cells(p<0.05);and the immune score,stromal score,and estimate score were all lower in the high-risk group than in the low-risk group(p<0.05).The tumor purity was higher in the high-risk group than in the low-risk group(p<0.05).Single-sample GSEA analysis found that in the high-risk group,many pathways were related to DNA damage repair,such as "base excision repair(BER)","nucleotide excision repair(NER)","homologous recombination(HR)",and "DNA mismatch repair(MMR)".And the high-risk group had higher scores of tumor mutational burden(TMB)and Microsatellite instability(MSI),which implied that patients in the high-risk group had higher sensitivity to immunotherapy.Drug sensitivity analysis revealed that patients in the high-risk group were able to achieve better survival benefit with chemotherapeutic agents such as docetaxel,doxorubicin,etoposide and mitomycin C.Thus,immunotherapy plus chemotherapy may be a new option for patients with high glutamine levels.3.Clinical specimen validation and cellular assays: in clinical samples,q PCR results detected that the expression of ATCAY and GLUL(p<0.05)was significantly reduced in prostate cancer patient tissues.Meanwhile,CAD and FPGS(p<0.05)were phenotypically overexpressed in prostate cancer patient tissues,which is consistent with our bioinformatics results.However,ASNS gene expression was not statistically different in the PCa and normal groups,a result that may be related to insufficient sample size.Additionally,knocking down the GLUL and ASNS genes and depriving the external glutamine in the 22Rv1 and DU145 cell lines led to a significant decrease in proliferation in both cell lines(p<0.05).Conclusion: In this study,we successfully established and validated a risk score model based on five glutamine-related genes,which can serve as an accurate prognostic biomarker for biochemical recurrence(BCR)in prostate cancer patients.We also revealed the potential mechanism of promoting prostate cancer progression through the reconstruction of the tumor microenvironment(TME)related to glutamine.And explored the important role of glutamine in prostate cancer.