Ferroptosis Related Genes And PET/CT-based Prognostic Models For DLBCL And Mechanism Of PRMT1 In Regulating Ferroptosis In Acute Myeloid Leukemia

Chapter 1 A ferroptosis-related gene signature predicting overall survival for DLBCLIntroductionDiffuse large B-cell lymphoma(DLBCL)is the most common B cell lymphoma which is highly heterogeneous.Multi-omics research is becoming increasingly popular in DLBCL.Molecular markers are potential targets for DLBCL therapy and tools for prognostic analysis.Ferroptosis,an iron-dependent form of programmed cell death,is caused by lipid peroxidation.Cell lines related researches found that DLBCL cell lines were sensitive to ferroptosis inducers.Further xenograft tumor studies showed that ferroptosis inducers had therapeutic effects in DLBCL,indicating that ferroptosis may be involved in the development and progression of DLBCL.A comprehensive understanding of the relation between DLBCL and ferroptosisrelated genes is necessary.Materials and methods1.A search about DLBCL patients’ samples was conducted in GEO Datasets.Datasets including DLBCL samples and normal control samples were chosen for differential expression analysis.Datasets including follow-up information were selected for the construction and validation of the predictive model.2.Two datasets(GSE56315 and GSE25938)underwent differential expression analysis.Differential expression genes(DEGs)were calculated using Limma with these cutoff values(P cutoff: 0.05,log2 FC cutoff: 2).The intersection of DEGs and ferroptosis genes was considered differential expression ferroptosis-related genes(DEFRGs).3.GO/KEGG enrichment analysis and PPI network were performed using the Metascape website and its PPI and MCODE plug-ins.4.Univariate Cox regression was conducted to analyze the association between DEFRGs and survival.5.LASSO-Cox regression was conducted to seek the independent factor and to build a prognostic model for OS.6.Calculating the risk score of each case,patients were divided into two risk groups(low-risk group and high-risk group).Survival curves and time-dependent ROC curves were plotted.7.GSE34171 was analyzed to validate our model.8.R and several R packages were used to plot all figures(including volcano plots,PCA plots,and forest plots).Results1.We found 59 datasets containing DLBCL patients’ samples.2.There were 3752 genes significantly up-regulated and 5041 genes down-regulated in GSE56315 and 2270 and 554 genes up-and down-regulated in GSE25638,respectively.3.The intersection of DEGs and ferroptosis genes consisted of 152 DEFRGs.4.GO enrichment analysis of 152 DEFRGs,the first three biological processes were cellular response to chemical stress,response to nutrient levels,and regulation of autophagy,respectively.KEGG pathway enrichment analysis of these DEFRGs,the first three pathways were ferroptosis,autophagy-animal,and lipid and atherosclerosis.Seven MCODEs were identified in the PPI network.5.After univariate Cox regression on GSE31312,41 DEFRGs were associated with OS.6.After LASSO-Cox regression,eight independent risk factors for OS were identified(SLC7A5,MT3,AURKA,AKR1C3,PEBP1,NGB,TUBE1,and GPX2).IDH1,ATG16L1,and CISD2 were independent protective factors.7.We built a ferroptosis-related gene predictive model using GSE31312.The 1-year AUC,3-year AUC,and 5-year AUC for OS were 0.778,0.787,and 0.809,respectively.Our model has a preferable predictive ability.8.We validated the predictive model using GSE34171.The 1-year AUC for OS was 0.870.ConclusionForty-one ferroptosis-related genes are related to overall survival for DLBCL,including 8 independent risk factors and four independent protective factors.Based on these genes,we built a predictive model for OS,which has a preferable performance.Our predictive model can be used in the risk stratification of DLBCL and the genes mentioned may be new targets for immunotherapy.Chapter 2 Prognostic value of baseline 18F-FDG PET/CT semi-quantitative parameters in patients with newly diagnosed diffuse large B cell lymphomaIntroductionDiffuse large B-cell lymphoma(DLBCL)is highly heterogeneous in its clinical features,treatment response,and prognosis.The International Prognostic Index(IPI)was widely used in clinical practice,but fails to meet the demand in rituximab era.New prognostic tools are required.PET/CT is a widely used tool for staging,treatment response evaluation,and prognosis prediction in lymphoma patients.The PET/CT images are usually analyzed with Deauville criteria.Many potential features remain undetected.The predictive value of baseline PET/CT(PET0)semi-quantitative parameters such as standard uptake value(SUV),total metabolic tumor volume(t MTV),and total lesion glycolysis(TLG)for the survival of DLBCL patients has been reported.New PET0 features,including tumor blood pool ratio(TBR)and tumor liver ratio(TLR),have shown prognostic value in Hodgkin lymphoma,but their prognostic values in DLBCL remain unclear.Therefore,this study aim to investigate the prognostic role of TLR and TBR in DLBCL and develop new prognostic model.Materials and methodsPathologically confirmed DLBCL cases between 2010 and 2018 were enrolled.1.Data of patients’ characteristics and clinical features were retrospectively collected,and descriptive statistical analysis were performed.2.PET/CT semi-quantitative parameters,including the maximum normalization of FDG uptake for patient body weight(SUVmaxbw),the maximum normalization of FDG uptake for patient lean body mass(SUVmaxlbm),the maximum n ormalization of FDG uptake for patient body surface area(SUVmaxbsa),TLG,t MTV,the maximum standard uptake value in the blood pool(SUVbloodpool),the maximum standard uptake value in the liver(SUVliver)were extracted using LIFEx software.Subsequently,TLR and TBR were calculated.3.PET/CT semi-quantitative parameters were transformed into categorical variables using optimal cutoff values generated by X-tile software.4.After long-term follow-up on patients,we analyzed the overall survival(OS)and progression-free survival(PFS)using Kaplan-Meier method.Univariate Cox regression was conducted on clinical and PET/CT-related parameters to assess their correlation with survival.5.LASSO-Cox regression was conducted to identify independent prognostic factors for DLBCL,and predictive models for OS and PFS were built based on them.6.Time-dependent ROC curves were plotted to assess those models’ predictive power.7.R software(Version 4.1.0)was used for survival curves,nomograms,forest plots of univariate and multivariate Cox regression.Results1.One hundred and two DLBCL cases were enrolled.The median follow-up time is 55 months.Except for SUVliver,all clinical and PET/CT-related features have no statistically significant differences between the response group and the no response group.2.The optimal cutoff values generated by X-tile for SUVbloodpool,SUVliver,SUVmaxbw,SUVmaxlbm,SUVmaxbsa,TLR,TBR,TLG,t MTV and lesion numbers are 0.90,1.20,30.85,19.26,5.42,16.91,20.96,7841,62,12 9.48 and 5 for PFS,respectively.3.Univariate Cox regression show that 10 factors were significantly associated with PFS and OS,including Ann Arbor stage,LDH,B2 M,lesion number,SUVbloodpool,TBR,TLG,t MTV,TLR,and SUVliver.4.LASSO-Cox regression results show that TBR(P = 0.001)and lesion number(P = 0.001)were independent risk factors for PFS.In addition,TBR(P < 0.001)and lesion number(P < 0.001)were independent risk factors for OS(P < 0.001).5.The predictive model of PFS consists of four parameters(TBR,lesion numbers,LDH,and TLG).While the OS model was composed of TBR,lesion number,LDH,and B2 M.For the PFS model,the 1-year AUC,3-year AUC,5-year AUC are 0.720,0.747,and 0.759,respectively.As for the OS model,the 1-year AUC,3-year AUC,5-year AUC are 0.716,0.782,and 0.787,respectively.Both models have preferable performance s.ConclusionOur study showed that TBR is an independent risk factor for PFS and OS in DLBCL patients.Predictive models combining clinical and PET/CT semi-quantitative parameters for PFS and OS have a preferable predictive ability.Chapter 3 Mechanism of PRMT1 in regulating ferroptosis in acute myeloid leukemiaIntroductionAcute myeloid leukemia(AML)is a type of hematological malignancy that originates from hematopoietic stem/progenitor cells.While traditional standard treatments can achieve a certain level of remission,the issues of relapse and refractory cases still persist.Current research aims to delve deeper into the mechanisms of myeloid leukemogenesis and develop new therapeutic targets and treatments.Recent studies have increasingly highlighted the significant role of epigenetic abnormalities in AML,presenting promising opportunities for targeting epigenetic factors in the broader application of AML treatments.Simultaneously,ferroptosis has emerged as a novel approach for combating cancer,including AML.Ferroptos is is a distinct form of regulated cell death characterized by its dependency on iron and the accumulation of lipid peroxidation.However,there is still limited understanding regarding the regulation of ferroptosis by epigenetic factors in AML and whether a combination of ferroptosis inducers and epigenetic drugs holds therapeutic value for AML.This study aimed to explore the epigenetic regulation of ferroptosis sensitivity in AML cells and uncover the underlying molecular mechanisms involved.Materials and methodsIn this study,we employed a comprehensive approach integrating cell experiments,database analysis,animal experiments,and multi-omics analysis.To investigate the research objectives,we focused on AML cell lines(e.g.,NB4,HEL,MOLM-13)as the primary subjects for our experiments.Additionally,female nude mice were selected as the animal model for conducting xenograft tumor experiments.1.Cell viability was assessed using the Cell Counting Kit-8(CCK-8)method.2.Flow cytometry was utilized to measure the levels of lipid peroxidation and Fe2+ content in cells,employing BODIPY-C11(581/591)and Ferro Orange,respectively.3.The expression of PRMT1 in various tumors was analyzed by leveraging public databases such as TCGA,GTEx,and Blood Spot.Furthermore,gene correlation analysis between PRMT1 and ACSL1 was conducted using the GEPIA 2 database.4.CRISPR/Cas9 technology was employed to construct knockout cells for PRMT1,ACSL1,ACSL3,and ACSL5 genes,and the efficacy of the knockout was assessed via western blot(WB).5.The combination of GSK3368715 and RSL3 was evaluated for its tumor suppression effect using a xenograft tumor model.6.Nanopore sequencing technology was utilized to perform transcriptome analysis after treatment with GSK3368715.Differential gene expression was validated using quantitative real-time polymerase chain reaction(q RT-PCR)and WB.7.The distribution of the H4R3me2 a and H3R17me2 a markers across the entire genome and around the ACSL1 promoter region was investigated using the CUT&Tag technology.ResultsWe discovered that the type I PRMTs inhibitor,GSK3368715,could enhance the sensitivity of AML cells to ferroptosis inducers RSL3 and FIN56 through screening of epigenetic inhibitors.The synergistic effect was confirmed using the Bliss independent model.Flow cytometry staining with BODIPY-C11(581/591),and Ferro Orange validated that the combination of GSK3368715 and RSL3 effectively induced AML cell death and significantly increased the levels of key ferroptosis indicators.These effects were reversed by ferroptosis inhibitors Fer-1,Lip1,and DFO.The knockout of PRMT1 exhibited a similar effect to GSK3368715 in these cells,suggesting that GSK3368715 primarily promoted ferroptosis sensitivity in AML cells by targeting PRMT1.In the xenograft tumor model,the combination of GSK3368715 and RSL3 effectively inhibited subcutaneous tumor growth without causing significant weight loss.Mechanistically,ONT-seq analysis revealed that GSK3368715 altered the expression of several ferroptosis-related genes,including ACSL1,ACSL3,ACSL5,CYBB,POR,FTH1,PEX10,and PEX12.q RT-PCR and WB further confirmed the differential expression of ACSL1,ACSL3,and ACSL5 following GSK3368715 treatment.Knocking out ACSL1,ACSL3,and ACSL5 genes in AML cells demonstrated that only ACSL1 knockout could restore the pro-ferroptosis effect of GSK3368715.Moreover,PRMT1 knockout significantly increased the protein level of ACSL1 in AML cells.Additionally,analysis of the GEPIA 2 database revealed a negative correlation between PRMT1 and ACSL1 gene expression in AML and pan-cancer studies.Using CUT&Tag technology,we investigated the genome-wide changes in PRMT1-mediated H4R3me2 a and CARM1/PRMT4-mediated H3R17me2 a after GSK3368715 treatment.The CUT&Tag results indicated that H4R3me2 a and H3R17me2 a were primarily enriched near the transcription start site.GSK3368715 significantly reduced the overall abundance of H4R3me2 a distribution throughout the genome,while it had no significant effect o n H3R17me2 a.Furthermore,the abundance of H4R3me2 a on the ACSL1 promoter region significantly decreased following GSK3368715 treatment.ConclusionPRMT1 plays a crucial role in regulating the sensitivity of AML cells to ferroptosis,and inhibiting PRMT1 can enhance ferroptosis in AML cells by increasing the expression of ACSL1.Through our exploration of epigenetic modifications,we observed that GSK3368715 treatment leads to a reduction in the abundance of H4R3me2 a specifically in the ACSL1 promoter region.These findings provide valuable insights into the involvement of PRMT1 in myeloid leukemogenesis and the development of AML.Furthermore,our study suggests that the combination of targeted epigenetic modulation and ferroptosis inducers holds potential for application in AML treatment.