Screening of prognostic genes associated with ferroptosis in esophageal adenocarcinoma and the establishment of its prognostic model

Fa-Zhang Chen ,Jun-Song Wu ,Ye Li ,Xue-Lian Zhang ,Xiao-Lan Zhang ,Ru-Yi Yang

1Medical institute of Qinghai University,Qinghai University,Xining 810000,China.2Affiliated Hospital of Qinghai University,Xining 810000,China.

Abstract We analyzed three gene microarray datasets by GEO2R and obtained differential genes associated with ferroptosis in esophageal adenocarcinoma by obtaining the FerrDb database to obtain ferroptosis-related genes for the intersection.To further elaborate on the functions of differentially expressed genes (DGEs),this study performed gene ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analysis on DEGs.We used the Kaplan-Meier plotter database to verify the effect of DGEs genes on the overall survival of esophageal adenocarcinoma.We performed univariate/multifactorial COX regression analysis of DGEs genes associated with esophageal adenocarcinoma prognosis by R language to obtain ferroptosis-associated independent prognostic genes.To further understand the relationship between the upstream molecules of independent prognostic genes and ferroptosis,we obtained the upstream regulatory molecules miRNAs and LncRNAs of prognosis-related ferroptosis genes with the help of the miRWalk database,Oncomi database and StarBase database.we obtained a total of 75 DEGs.These DGEs were mainly enriched in the cellular response to lipids,and negative regulation of intracellular.These DGEs were mainly enriched in the negative regulation of intracellular signaling,positive regulation of cell death,cellular autophagy,HIF-1 signaling pathway,microRNAs in cancer,and ferroptosis.We performed prognostic analysis and univariate/multifactorial COX regression analysis on 75 ferroptosis-related genes and established four independent genes for esophageal adenocarcinoma,ATF3,ATM,ATG5,and HMGB1.The study also established hsa-miR-876-5p,hsa-miR-186-5p,hsa-miR-421,hsa-miR-505-3p,hsa-miR-503-5p,hsa-miR-299-3p and hsa-miR-191-5p,seven miRNAs with upstream regulation of LncRNAs.these miRNAs can be competitively bound by LncRNAs to prevent the inhibition of translation of target gene expression by miRNAs.In conclusion,our study identified four esophageal adenocarcinoma independent prognostic genes and their upstream regulatory molecules.These genes are involved in the ferroptosis regulation of cells and also play an important role in tumor therapy and drug resistance as one of the disease therapeutic targets.The study suggests that by targeting ATF3,ATM,ATG5,HMGB1 and their upstream regulatory molecules is a new direction for the treatment of esophageal adenocarcinoma.

Keywords:Esophageal adenocarcinoma,bioinformatics,ferroptosis,non-coding RNA

Background

Esophageal cancer is an aggressive gastrointestinal malignancy that occurs in the esophageal epithelial tissues.The global incidence of esophageal cancer is about 65%,mortality is about 38% and about 500,000 new cases per year.In 2018,incidence and mortality of esophageal cancer ranked 7th and 6th globally,respectively [1].The incidence of esophageal cancer in China accounts for about 46.6% of esophageal cancer patients worldwide,making it one of the highest incidence and mortality countries worldwide [2–3].According to the different pathological characteristics of esophageal cancer,we can divide it into esophageal phosphorous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC).It is currently believed that the occurrence of EAC is the result of long-term chronic damage to the esophageal mucosa,and is associated with obesity,smoking,reflux gastroesophageal lesions and Barrett's esophagus [4].For patients with early-stage esophageal cancer,endoscopic treatment is the current effective treatment option.Surgery and neoadjuvant chemoradiotherapy are suitable for patients with existing metastasis or no metastasis [5].Because of its diagnosis difficulties,high metastasis and recurrence rate,the 5-year overall survival rate of esophageal adenocarcinoma patients,undergoing surgery,is also only 16% to 34% [6–7].Therefore finding suitable and highly effective treatments to inhibit tumor cell proliferation and metastasis is the primary problem in esophageal cancer research.

Ferroptosis is an iron-dependent mode of cell death.Ferroptosis is different from apoptosis,necrosis and autophagy in terms of biochemical response,cell morphology and gene expression [8].Ferroptosis has no nucleus enrichment,chromatin breakage and activation of autophagic factors.It is mainly manifested by mitochondrial shrinkage,increased density of mitochondrial membrane,decrease or disappearance of the mitochondrial ridge and decreased cell mitosis [9].At present,many studies believe that lipid oxidation is the cause of cell death caused by ferroptosis.It is regulated by glutathione peroxidase-4 (GPX4) and glutathione (GSH),cystine-glutamate antitransporter (Xc),iron metabolism,and P53.GPX4 is the only protease in glutathione peroxidase that can protect the cell membrane from oxidative damage[10–11].GPX4 can remove peroxides on cellular phospholipid membranes,prevent the accumulation of reactive oxygen species (ROS),and inhibit the occurrence of ferroptosis [12].GSH can reduce excessive intracellular ROS and reactive oxygen species under the action of GPX,and prevent the occurrence of lipid peroxidation[9].Xc is an amino acid transport complex on the cell membrane,a channel for cystine and glutamate to and from the cell membrane [13].Xc can transport extracellular cysteine into the cell to participate in GSH synthesis [14].Iron is an important trace element in the human body,involved in many physiological reactions of the body.And the imbalance of iron metabolism is also one of the causes of ferroptosis.Intracellular excess of Fe2+catalyzes Fenton which produces large amounts of hydroxyl radicals that cause intracellular lipid oxidation to cell death [15].P53 is the tumor suppressor gene that regulates cell apoptosis,proliferation,and DNA repair in vivo[16].P53 can inhibit the uptake of cystine by Xc,to ccausethe accumulation of cellular ROS and induces the occurrence of ferroptosis [17].There has been evidence that ferroptosis inhibits the proliferation and metastasis of tumor cells by inducing programmed cell death,whereas ferroptosis can also be used as a drug therapeutic target in the treatment of drug-resistant tumors [18–19].Viswanathan et al found that tumor cells with mesenchyme and dedifferentiation were more likely to undergo ferroptosis [20].There are still many problems with ferroptosis as a new tumor treatment modality.At present,the tumor cell ferroptosis marker molecules and the prognostic molecules have not been found.At the same time,non-coding RNA,which regulats ferroptosis-related genes,in the occurrence of ferroptosis is temporarily unclear.We obtained and screened the ferroptosis-related genes of esophageal adenocarcinoma through GEO database and the FerrDb database.We obtained miRNA and LncRNA,which is the upstream regulator of ferroptosis genes,through miRWalk,Oncomi and StarBase databases,and explored the role of these non-coding RNA in the ferroptosis pathway.

Materials and Methods

Data source

The study material was obtained from the GEO database ({GSE1420,GSE26886 and GSE92396) (https://www.ncbi.nlm.nih.gov/) and the FerrDb database (http://www.zhounan.org/ferrdb/).The GSE1420 contains 8 normal esophageal tissue samples,8 Barrett's esophageal tissue samples and 8 esophageal adenocarcinoma samples;the GSE26886 contains 20 Barrett's esophageal tissue samples,21 esophageal adenocarcinoma samples,9 esophageal squamous cell carcinoma tissue samples and 19 normal esophageal tissue samples;the GSE92396 contains 9 esophageal adenocarcinoma tissue samples and 12 normal esophageal tissue samples.Ferroptosis-associated genes are included in the FerrDb database,including the marker genes of ferroptosis,ferroptosis-promoting and ferroptosis-inhibiting genes.In our study,we only screen the differential genes between normal tissues and esophageal adenocarcinoma tissues and differentially expressed ferroptosis genes.

Screening of ferroptosis genes differentially expressed in esophageal adenocarcinoma

The study analyzed gene expression differences between adenoma and normal tissues in 3 datasets by the GEO2R online analysis tool.P-value corrected (adjustP-value) <0.05 and | log FC | >1 were used as differential gene screening criteria.And the genes with log FC>1 were up-regulated genes,and log FC <–1 were downregulated genes.We downloaded the validated ferroptosis-related genes through the FerrDb database and used the TBtools tool to obtain the intersection genes of three GEO data sets with ferroptosis-related genes.

Establishment of the protein interaction network map and the GO and KEGG enrichment analysis

We construct the protein-protein interaction (PPI) network of the ferroptosis differential genes screened in 1.2 and annotate the gene function for these genes through the GeneMANIA database(http://genemania.org/).We perform gene ontology (GO)and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of ferroptosis differential genes through the Metascape database (https://www.kegg.jp/).APvalue of less than 0.01 and a minimum enrichment of 3 was used as the screening condition.We analyze pathways with the most enriched genes through the KEGG database(https://www.kegg.jp/).

Screening of genes for differential ferroptosis associated by prognosis in oesophageal adenocarcinoma

The study analyses the effect of ferroptosis differential genes on overall survival of esophageal adenocarcinoma through the Kaplan-Meier plotter database (http://kmplot.com/analysis/).We downloaded survival information of esophageal adenocarcinoma patients through the TCGA database(https://www.cancer.gov/).And then,we use R to perform univariate COX regression analysis and multivariate COX regression analysis of prognostic-related ferroptosis differential genes to establish independent prognostic genes in esophageal adenocarcinoma.We find the correlation between ferroptosis-related prognostic genes and clinical features,and analyze the model’s predictive ability,by nomogram and time-dependent ROC curve.

Screening of prognosis-related ferroptosis’s upstream regulated,miRNA and LncRNA

miRNA,the regulatory molecule of mRNA expressioncan,can regulate the protein translation of mRNA.This study obtained the miRNA of ferroptosis genes with independent prognostic markers and miRNA associated with esophageal cancer prognosis through the miRWalk database (http://mirwalk.umm.uni-heidelberg.de/) and the Oncomi database(http://www.oncomir.org/oncomir/).The intersection of the above miRNA is the miRNA associated with esophageal adenocarcinoma prognosis.To elucidate the function of these miRNA,we performed a functional analysis of the screened miRNA by the Funrich database.LncRNA can compete with mRNA for the binding site of miRNA,to avoid mRNA binding to miRNA and regulate mRNA translation.This study predicts the LncRNA,which is regulated upstream of the miRNA,through the StarBase v2.0 database(https://starbase.sysu.edu.cn/starbase2/index.php) and uses the highest selection reliability (very high stringency >5) as the screening criterion.

Results

ferroptosis genes differentially expressed in oesophageal adenocarcinoma

We perform the differential expression analysis on the three GEO datasets by GEO2R and screen the differential genes withP-value <0.05 and | log FC | >1 (Figure 1a,b,c).The GSE1420 dataset has 3,107 differential genes,including 1720 upregulated genes and 1,386 downregulated genes;There are 1,586 differential genes in the GSE92396 dataset,including 811 upregulated genes and 774 downregulated genes;The GSE26886 dataset has 8093 differential genes,including 4268 upregulated genes and 3859 downregulated genes;We obtained the genes associated with ferroptosis through the FerrDb database.And take the intersection with the 3 datasets,we can obtain 75 ferroptosis genes of esophageal adenocarcinoma(Figure 2).

Protein interaction networks of iron-death differential genes

The study constructs a protein-protein interaction network and annotates the gene function of 75 esophageal adenocarcinoma genes by the GeneMANIA database (Figure 3).Nodes in the PPI represent different proteins,and pie charts with different colors in the nodes represent the functions that the genes have.The proteins associated with iron metabolism in the PPI were HMOX1,FTL,CP,TF,TFRC,FTH1,SLC4DA1,TFR2,and HAMP (Table 1).

Figure 1 ferroptosis differential genes and PPI networks.a.GSE1420;b.GSE2688;c.GSE92396.

Figure 2 ferroptosis differential genes

Figure 3 The PPI networks of ferroptosis differential genes

Table 1 Genes enriched in iron metabolism in oesophageal adenocarcinoma

GO term and KEGG pathway enrichment analysis of differential genes in ferroptosis

The study performs KEGG pathway and GO term enrichment analysis for 75 ferroptosis differential genes by Metascape database.The filter criteria were set as follows:P-value less than 0.01 and a minimum number of enrichment as 3.GO enrichment analysis showed that the 75 ferroptosis differential genes are mainly enriched in cell response to lipids,negative regulation,cell death,and monocarboxylate metabolism (Figure 4a,b,c,d);genes are located ithe n HFE-transferrin receptor complex,membrane raft,and cell apical part;The function of these genes are lyase activity,transcription factor binding,and oxidoreductase activity.KEGG pathway enrichment analysis showed that ferroptosis genes in esophageal adenocarcinoma were mainly enriched in arachidonic acid metabolism,fluid shear stress and atherosclerosis,cell autophagy,HIF-1 signaling pathway,MicroRNAs,and ferroptosis in cancer.We analyzed the enriched ferroptosis pathways by the KEGG database.The upregulated genes of esophageal adenocarcinoma are marked in red and the downregulated genes are marked in green (Figure 5).

Figure 4 Enrichment analysis of the intersection genes. a.GO biological processes;b.GO molecular function;c.GO cell components;d.The KEGG signaling pathway

Figure 5 ferroptosis signal pathway diagram

Survival analysis of genes for differential ferroptosis in oesophageal adenocarcinoma

In this study,the survival analysis of 75 ferroptosis genes in oesophageal adenocarcinoma was performed by the Kaplan-Meier plotter database.We found that the expression of NFE2L2,DPP4,PRKAA2,SCD,ENPP2,AKR1C1,ATF3,TLR4,FH,PML,HMGB1,ATM,SLC38A1,FTH1,MTDH,LAMP2,ATG5,MAPK3,and BAP1 was associated with the survival prognosis in esophageal adenocarcinoma(Figure 6).To further identify ferroptosis genes with independent prognostic markers of esophageal adenocarcinoma,we analyze the univariate COX regression analysis and multivariate COX regression analysis of ferroptosis genes associated with the prognosis of esophageal adenocarcinoma(Figure 7a).We found that the expression of ATF3,ATG5,and HMGB1 correlate negatively with a prognosis of esophageal adenocarcinoma,and ATM expression correlates positively(Table 2).The area under the ROC curve was used to compare the prognostic ability of ATF3,ATM,ATG5,and HMGB1.Averaged over the 1,3 and 5-year follow-up,The area,under the curve,of ATF3,ATG5 and HMGB1 was greater than 0.5.The area of ATM5 is above 0.5 in a follow-up of 5 years.The time-dependent ROC curve showed that ATF3,ATG5,HMGB1 and ATM expression correlated with the survival of esophageal adenocarcinoma patients within 5 years(Figure 7b,c,d).We performed the correlation analysis of ferroptosis-related prognostic genes and clinical characteristics by R language.We found that 1,3 and 5-year survival rates were inversely correlated with age,tumor stage,tumor metastasis and ATF3,ATG5 and HMGB1 expression,and positively correlated with ATM expression (Figure 7e,f).

Targeted miRNA of ATF3,ATM,ATG5,and HMGB1 and its functional enrichment

We obtained 1,725 targeted miRNA of ATF3,ATM,ATG5,and HMGB1 through the miRWalk database and 75 miRNA,which are associated with the prognosis of esophageal cancer,through the Oncomi database.The intersection of the above two types of miRNA is the Targeted miRNA of ATF3,ATM,ATG5,and HMGB1 (Figure 8,figure 9a).We performed functional enrichment of the screened miRNA by the Funrich software.The enrichment results showed that 44 miRNA related to prognosis in esophageal adenocarcinoma mainly existed in the cytoplasm and nucleus.The functions of these miRNA mainly enriched in the regulation and transport of nuclebases,nucleosides,nucleosides,and nucleic acid metabolism.These miRNA are involved in mediating the glycoprotein pathway,activation of MEK and the transport of mature mRNA of intron-containing transcripts(Figure 10).

LncRNA prediction of the upstream molecule of miRNA with a prognostic correlation in oesophageal adenocarcinoma

We predicted upstream LncRNA of the prognostic related miRNA in esophageal adenocarcinoma by the StarBase database.In the StarBase database,we found that only hsa-miR-876-5p,hsa-miR-186-5p,hsa-miR-421,hsa-miR-505-3p,hsa-miR-503-5p,hsa-miR-299-3p and hsa-miR-191-5p have upstream LncRNA (Figure 9b).This LncRNA may regulate the expression and translation of ATF3,ATM,ATG5,and HMGB1 by binding to miRNA to prevent mRNA degradation(Table 3).

Table 3 Noncoding RNA of independent prognostic genes

Figure 8 Screening of non-coding RNA.miRWalk database and Oncomi database intersection miRNA

Figure 9 Screening and reticulgraphy of non-coding RNA. a.The miRNA-mRNA network map,with red circles representing miRNA and green circles representing mRNA;b.LncRNA-miRNA-mRNA network diagram,with red circles representing miRNA,green circles representing mRNA,and yellow triangle representing LncRNA.

Figure 10 GO and KEGG enrichment analysis of miRNA. a.GO biological processes;b.GO molecular function;c.GO cell components;d.KEGG signaling pathway

Table 2 Univariate/multivariate COX regression analysis of the prognostic genes for ferroptosis

Figure 6 Prognosis analysis of ferroptosis genes.a.NFE2L2;b.DPP4;c.PRKAA2;d.SCD;e.ENPP2;f.AKR1C1;g.ATF3;h.TLR4;i.FH;j.PML;k.HMGB1;l.ATM;m.SLC38A1;n.FTH1;o.MTDH;p.LAMP2;q.ATG5;r.MAPK3 s.BAP1

Figure 7 Independent prognostic gene and prognostic model for ferroptosis.a.Forest map for univariate analysis;b.ROC curve:AFT3;c.ROC curve:HMGB1;d.ROC curve:ATM;e.ROC curve:ATG5;f.Prognostic model nomogram;g.Prognostic model calibration

Discussion

The ferroptotic cell death pathway induces apoptosis,which maybe is a potential therapeutic target for tumor therapy and drug resistance in cancer.Hasssannia found that Withaferin can induce ferroptotic cell death,which can effectively inhibit the growth and recurrence of neuroblastoma[21].The knockout of GPX4 can lead to increased lipid ROS in renal cell carcinoma cells and induce ferroptotic cell death[22].Ma found that lysosomal damaging agents together with tyrosine kinase inhibitors can induce ferroptosis in breast cancer cells through the regulation of intracellular iron metabolism [23].Another study found that corosonic acid could induce lipid oxidation in renal cell carcinoma and thus develop tumor cell apoptosis [24].Studies have shown that artesunate causes the growth arrest and death of pancreatic cholangiocarcinoma cells by inducing the production of intracellular ROS [25].Recent studies have found that sulfasalazine can inhibit the XC to induce ferroptotic cell death [26].Dingguan found that sulfasalazine combined with traditional antitumor drugs could reduce the drug resistance of tumor [27].Guo demonstrated that the drug resistance of tumors can inhibit the original mechanism of action,but have no effect on the drug-induced ferroptosis pathway[28].The above studies demonstrate the therapeutic potential of ferroptosis in tumor treatment and the drug resistance of tumor.Targeted ferroptosis therapy is an important way for tumor treatment and improvement of drug resistance.

In this study,ferroptosis-related genes of esophageal adenocarcinoma were obtained by GEO database and FerrDb database,and prognostic and functional enrichment analysis of these genes was performed.We found that the proteins enriched for iron metabolism correlation in PPI were HMOX1,FTL,CP,TF,TFRC,FTH1,SLC4DA1,TFR2,and HAMP.We performed a pathway analysis of differential esophageal adenocarcinoma genes enriched in the ferroptosis pathway.The KEGG enrichment analysis showed that ferroptosis-related genes in esophageal adenocarcinoma were mainly enriched in arachidonic acid metabolism,fluid shear stress and atherosclerosis,cellular autophagy,HIF-1 signaling pathway,MicroRNAs,and ferroptosis.

To further understand the relationship between ferroptosis-related genes and tumor prognosis,we performed a prognostic analysis of 75 differentially expressed genes in esophageal adenocarcinoma and obtained 19 genes associated with the prognosis of esophageal adenocarcinoma.We performed the univariate and multivariate COX regression analysis on the 19 prognostic genes.We established four independent prognostic genes related to ferroptosis in oesophageal adenocarcinoma:ATF3,ATM,ATG5,and HMGB1.

Transcriptional activator 3 (ATF3) is a member of the ATF/CREB transcription factor family,involved in the regulation of various cellular stresses,such as DNA damage,oxidative stress,and cell damage[29].Knockdown of ATF3 inhibited the viability and invasion ability of colon cancer cells [30].Several studies have found that the expression of ATF3 is upregulated in erastin-induced ferroptosis [31].ATF3 may be a facilitator of the occurrence of cell ferroptosis.Liyuan Wang found that the expression of ATF3 inhibited cystine transport by XC and promoted the occurrence of cell ferroptosis [32].ATF3 can also promote the increase of intracellular hydrogen peroxide by upregulating the expression of NOX4 and SOD1 and inducing the occurrence of cell ferroptosis [33].Nuclear factor erythroid-derived 2(Nrf2) is a transcription factor that can regulate the intracellular reduction–oxidation response with iron metabolism to inhibit the occurrence of cell ferroptosis [34].ATF3 can form an ATF3–Nrf2 complex and block Nrf2 signaling [35].Dazhi Fu et al found another pathway in which ATF3-induced cell ferroptosis occurs throught experimentally validated.ATF3 blocked the signal of Nrf2/Keapl/xCT to induce the occurrence of ferroptosis of gastric cancer and reduce drug resistance in gastric cancer cells [36].We found that the expression of ATF3 is upregulated in oesophageal adenocarcinoma and is correlated with the prognosis of oesophageal adenocarcinoma.There are few studies on the treatment of ATF3 and oesophageal adenocarcinoma.Targeting ATF3 induces cell ferroptosis,which may be a new target for esophageal adenocarcinoma treatment and a new direction for the therapy of tumor drug-resistant.

Serine/threonine kinase (ATM) is a protein kinase for DNA damage repair and a key enzyme in Nrf2 phosphorylation,which also plays important regulatory roles in the cell cycle and oxidative stress pathways [37].Nrf2 is a defense factor in the intracellular oxidative stress response [38].Nrf2 can bind to the antioxidant response element (ARE),upregulate the expression of cellular defense enzymes and antioxidant proteins,and improve the tolerance of cells to oxidative stress response [39].When cells are under oxidative stress,Nrf2 is released from Keapl to binds to ARE.There is a formation of Nrf2-ARE complex,which induces the expression of detoxification enzyme genes with antioxidant properties [40].Ma et al found that silencing of ATM expression decreased the expression of Nrf2 and its downstream genes and caused decreased cell proliferation in cattle mammary glands associated with cell death [41].This suggests that ATM can regulate cell tolerance to oxidative stress response by regulating Nrf2 expression.Another study found that ATM could be activated by ROS,while ATM knockdown cells were more sensitive to oxidative stress inducers[42–43].Cosentino et al found that ATM can also increase the cellular antioxidant cofactors (NADPH) and nucleotides through the pentose phosphorylation pathway.It prevents the accumulation of cellular ROS and damage to cells [44].

The final outcome of ferroptosis is the cell death caused by intracellular lipid peroxidation.Interestingly,the present study has found that the downregulation of ATM expression inhibits the occurrence of cell ferroptosis and avoids the oxidative lysis of the cells[45].Knockdown or inhibiting the expression of ATM can increase the expression of ferritin,SLCTAL,and GPX4,reduce the expression level of intracellular unstable iron,and inhibit the occurrence of ferroptosis[45].

For the different biological properties of ATM in inhibiting oxidation and inducing ferroptosis,we believe that it may be different from the cellular components enriched by ATM.ATM is mainly enriched in the nucleus and involved in DNA damage repair processes.But its downstream signaling molecule-P53,can be expressed in the cytoplasm and involved in the regulation of ferroptosis[46-47].

The high-mobility family protein box-1 (HMGB1) is a transcription factor during chromatin remodeling and DNA repair,which encodes non-histone and DNA-binding proteins.This protein plays a role in multiple cellular processes,including inflammation,cell differentiation,and tumor cell migration [48].Qi Rong et al found that ferroptosis inducers promoted HMGB1 expression [49].This suggests that HMGB1 may be involved in the occurrence of cell ferroptosis.In diabetic nephropathy,HMGB1 can lead to the rise of intracellular ROS and promote the occurrence of ferroptosis by inhibiting Nrf2 expression.The present study has found that HMGB1 can also participate in the occurrence of ferroptosis in leukemia cells through RAS-JNK/p38 [50].Depletion of HMGB1 expression can result in the reduction of TfR1 expression and suppress the accumulation of intracellular ROS[50].Zhang Haiyan et al found that HMGB1 expression promoted the induction of doxorubicin on ferroptosis in rats [51].In acute liver failure,inhibition of HMGB1 expression reduces the level of cellular oxidative stress.Thereby,it inhibits the occurrence of cell ferroptosis [52].The expression of HMGB1 induces cell ferroptosis occurrence.Therefore,targeting HMGB1 treatment induces ferroptosis occurrence in tumor cells.And it may be one of the new targets for the treatment of esophageal adenocarcinoma.

Autophagy protein 5 (ATG5) is a key regulator of autophagy.The protein,encoded by ATG5,is involved in various cellular processes,such as autophagic vesicle formation,mitochondrial quality control after oxidative damage,adipocyte differentiation,and apoptosis.Current research suggests that the activation of autophagy can degrade intracellular ferritin,increase iron levels,and induce the occurrence of cell ferroptosis[53].Qi Rong et al found the occurrence of ferroptosis by knockdown of ATG5 [49].Knockdown of ATG5 reduced the intracellular level of ferrous ion lipid oxidation and inhibited intracellular ferroptosis [54].Another study found that ATG5 induced intracellular autophagosome formation and increase lipid ROS and depleted the depletion of GPX4,to induce ferroptosis[55].ATG5 can regulate the occurrence of cellular autophagy associated with cell ferroptosis.At the same time,the cellular autophagy,induced by ATG5,inhibits the formation and development of liver cancer,lung cancer and melanoma [56].Currently,ATG5 is rarely studied in the treatment of esophageal adenocarcinoma.We found that ATG5 is a prognostic gene related to ferroptosis in esophageal adenocarcinoma,and the induction of tumor cell apoptosis and ferroptosis by targeting the expression of ATG5 is a new therapeutics target of esophageal adenocarcinoma.

Non-coding RNA is a long-chain RNA that does not encode proteins and is involved in the regulation of many cells biological processes.Many non-coding RNA have been shown to play an important role in the development of tumors and treatment [57].miRNA can target the 3 ′ UTR of the target mRNA to regulate the transcription and translation of the mRNA [58].We through the miRWalk database and the Oncomi database acquired 44 miRNA regulating ATF3,ATM,ATG5,and HMGB1.These miRNAs are correlated with the prognosis in oesophageal adenocarcinoma and can simultaneously regulate the expression of independent prognostic genes in oesophageal adenocarcinoma.These miRNAs exist in the cytoplasm and nucleus,and are involved in the regulation and transportation of nucleic acid metabolism,signal transduction,and cell communication.To search for therapeutic targets of miRNA regulatory target genes,we screened the upstream regulatory molecules of 44 miRNA.We found that 18 LncRNA can bind to hsa-miR-876-5p,hsa-miR-186-5p,hsa-miR-421,hsa-miR-505-3p,hsa-miR-503-5p,hsa-miR-299-3p and hsa-miR-191-5p to antagonize inhibition of target gene expression by miRNA.This indicates that the expression of ATF3,ATM,ATG5,and HMGB1 can be regulated by targeting LncRNA.

Conclusion

In this study,75 ferroptosis-related genes for esophageal adenocarcinoma were screened by bioinformatics analysis.We used the Kaplan-Meier method and univariate/multivariate COX regression analysis to establish 4 independent prognostic genes:ATF3,ATM,ATG5,and HMGB1.We found that ATF3,ATM,ATG5,and HMGB1 were involved in ferroptosis regulation in cells by a literature search.Meanwhile,ATF3,ATM,ATG5,and HMGB1 are one of the targets for disease treatment,and play an important role in tumor treatment and drug resistance.At present,ATF3,ATM,ATG5,and HMGB1 are rarely studied in the treatment and drug resistance of esophageal adenocarcinoma,and the treatment methods targeting ATF3,ATM,ATG5,and HMGB1 are new directions for the treatment of esophageal adenocarcinoma.Furthermore,we screened the miRNA,regulating four independent prognostic genes for esophageal adenocarcinoma,and its upstream LncRNA.We established seven miRNA with an upstream regulatory LncRNA.These miRNA can be competitively bound by LncRNA to prevent inhibition of translation of target gene expression.