A Multi-Component Bioinformatics Study On The Construction Of A Prognostic Signature Of Genes Associated With Diverse Programmed Cell Death (PCD) In Acute Myeloid Leukemia (AML) And A Multi-Perspective Mechanism Exploration

Objective: A bioinformatics approach was used to assess the prognostic value of programmed cell death(PCD)-related genes in patients with acute myeloid leukemia(AML)and to explore the mechanisms of AML development and immunomodulatory therapy from multiple perspectives.Method: Our study downloaded clinical data and transcriptome sequencing data of AML patients from TCGA,GEO and GTEX data.Then,we downloaded key regulatory genes for 12 PCD patterns(apoptosis,necroptosis,ferroptosis,pyroptosis,netotic cell death,entotic cell death,lysosome-dependent cell death(LDCD),autophagy-dependent cell death,alkaliptosis,oxeiptosis,and parthanatos)and randomly divided AML samples with complete clinical data into training and validation sets.A machine learning algorithm was used to create a programmed cell death-related risk score(PCD-RS)with 6 gene signatures to quantify AML.Its predictive power was validated in several different databases.Next,molecular subtypes associated with AML were identified by unsupervised clustering analysis.By combining PCD-RS with clinical features,column line graphs were constructed.In addition,PCD-RS was analyzed for correlation analysis with immune checkpoint genes,tumor microenvironment components,and drug treatment sensitivity.Results: We successfully constructed a prognostic model of six PCD-associated genes by machine learning algorithms and validated its predictive accuracy in multiple datasets.The results of the study showed that PCD-RS has a good predictive effect for AML patients,with the area under the curve value greater than 0.70 and up to 0.86 in both the training and validation sets.Next,we identified two AML-associated molecular subtypes with different important biological processes by unsupervised cluster analysis.By combining PCD-RS with clinical features,a nomogram with high predictive power was constructed.In addition,PCD-RS was analyzed for correlation with drug treatment sensitivity analysis.The results showed that high risk scores yielded resistance to AML chemotherapeutic agents(5-fluorouracil,axitinib,cisplatin,doxorubicin,emputinib,epirubicin,mitoxantrone,olaparib,oxaliplatin,rapamycin,vincristine,and zoledronic acid).Therefore,targeting these genes may be a potential therapeutic target for AML chemoresistant patients.Finally,we correlated PCD-RS with immune gene expression and tumor microenvironment components by a comprehensive analysis of the overall and single-cell transcriptome.Conclusion: In this study,we analyzed the risk prediction of PCD-related prognostic features in AML patients and constructed a model by synthesizing the genes associated with the PCD pattern,which can suggest the prognosis and sensitivity of drug treatment in AML patients.