To Explore The Influence Of Iron Status On The Occurrence Of Breast Cancer And Ferroptosis On The Prognosis Of Breast Cancer Based On Biological Information Mining

Background and ObjectiveThe association between serum iron status and the risk of breast cancer is unclear till now.In this study,two-sample Mendelian Randomization(TSMR)was used firstly to explore the causal relationship between iron status markers(serum iron,transferrin,ferritin,and transferrin saturation)and the risk of breast cancer,based on genome wide association study(GWAS)database of iron status and breast cancer.Most of the free iron in the cells produces lipid peroxides when they participate in the Fenton reaction,causing ferroptosis.Ferroptosis could induce death or proliferation of cancer cells.The relationships of ferroptosis-related genes and breast cancer prognosis are unknown.Hence,the risk prognosis model of breast cancer was constructed by ferroptosis-related genes in this study,providing reference for the prevention and treatment of breast cancer.Methods1.Single nucleotide polymorphisms(SNPs)being used as instrumental variables(p<5×10-8)for iron status were selected via the Genetics of Iron Status consortium(GIS).SNPs being used as outcome variables corresponding to instrumental variables and existing no associations with breast cancer(P>0.05)were selected via OncoArray network database.The F statistic value for instrumental variables was calculated to ensure that there was no influence of weak instrumental variables.GWAS Catalog database was used to search whether the tool variables were related to other confounding factors.The conservative method(SNPs had a strong correlation with the four iron states)and the free method(SNPs had a strong correlation with at least one iron state)were used in the selection of instrumental variables.In the conservative method,an inverse variance weighted(IVW)approach was use to conduct the TSMR analysis,while IVW,MR-Egger regression,weighted median and simple mode approach were used in the free method.Finally,the sensitivity analysis was performed on the results of two sample MR analysis to test the robustness of causal inference.The heterogeneity test and multiplicity test were used in the sensitivity analysis,in which the inverse variance weighting method was used in the heterogeneity test,and the MR-Egger regression method was used in the multiplicity test.In addition,in order to exclude the possibility that a single SNP has excessive influence on causal estimation,leave-one-out analysis was conducted to test the tool variables separately.2.The mRNA expression profiles and the corresponding clinical data of breast cancer patients were downloaded from The Cancer Genome Atlas(TCGA)database.Ferroptosis-related genes were obtained from Ferroptosis Regulators and Markers and Ferroptosis-Disease Associations(FerrDb)database.The ferroptosis-related gene expression matrix was constructed combing the TCGA breast cancer mRNA expression profile.Differential expression analysis and single factor prognostic analysis of ferroptosis-related genes were performed to extract differential genes with prognostic value.Protein interaction network and gene co-expression analysis were carried out based on differential genes with prognostic value.Least Absolute Shrinkage and Selection Operator(Lasso)and COX regression were used to construct a polygenic prognostic model in the TCGA cohort.Breast cancer patients data obtained from the Molecular Taxonomy of Breast Cancer International Consortium(METABRIC)database were used to validate the prognostic model.According to the prognostic model,patients with breast cancer were divided into high and low risk groups,and risk differences between groups were analyzed.Finally,Gene Ontology(GO)analysis,(Kyoto Encyclopedia of Genes and Genomics(KEGG)analysis and single sample gene set enrichment analysis(ssGSEA)were performed according to the risk differentially expressed genes.Results1.The F statistics of all instrumental variables were greater than 10,indicating no weak instrumental variables existed.In the conservative method,the effects of the four iron status markers on breast cancer and its subtypes were not statistically significant,that is,there was no causal relationship(P>0.05).In the free method,the results of simple mode approach showed a positive correlation between transferrin and estrogen receptor(ER)negative breast cancer,which indicated positive causality(OR:1.225;95%CI:1.064,1.410;P=0.030).However,there was no causal relationship between other iron status markers and breast cancer and its subtypes(P>0.05).The sensitivity analysis indicated that there was no heterogeneity and multiplicity which interfered with the causal estimation of exposure and outcome by TSMR analysis(P>0.05).It was proved that the causal inference of TSMR analysis was robust.Leave-one-out analysis showed that all SNPs had no excessive effect on the causal estimation in this study(P>0.05).2.In the TCGA cohort,there were 204 ferroptosis-related genes in the expression profile,and most ferroptosis-related genes(81.4%)were differentially expressed between the tumor and adjacent normal tissues.In the univariate Cox regression analysis,thirty-one differentially expressed genes were associated with prognostic survival(P<0.05).Through protein interaction network and gene co-expression network analysis,HRAS,GPX4 and PRDX1 were at the core of the two networks.Additionally,a prognostic model was constructed for twelve ferroptosis-related genes(Risk score=0.263211139×expression level of G6PD+0.233409549 × expression level of CHAC1-0.557790972 × expression level of IFNG+0.642562765 × expression level of ANO6+0.282910123 × expression level of MTDH+0.37094507 ×expression level of CISD1-0.356743892 × expression level of TP630.378520589 × expression level of BRD4-0.223831169 × expression level of AIFM2+0.303694159 × expression level of PROM2-0.325522056× expression level of SLC1A4+0.426539599×expression level of NGB).Patients were divided into high-risk and low-risk groups by risk scores.Compared with the low-risk group,the overall survival rate of the high-risk group was significantly reduced(P<0.001).The risk score was determined as an independent prognostic factor through the univariate and multivariate Cox regression analysis(P<0.001).GO and KEGG analysis showed that the risk differentially expressed genes were enriched in immune-related functions.Finally,the immune scores of sixteen kinds of immune cells and thirteen kinds of immune-related functions between high-risk and low-risk groups were calculated by ssGSEA.It was found that the scores of most immune cells and immune-related functions in the low-risk group were significantly higher than those in the high-risk group.Conclusion1.Through the two sample MR analysis,elevated transferrin concentrations were found to increase the risk of ER-negative breast cancer,suggesting that transferrin,as a breast cancer exposure factor,is a factor in the development of estrogen receptor-negative breast cancer.It may provide scientific basis for breast cancer prevention.2.The prognosis model constructed by 12 ferroptosis-related genes and verified by the external database can be used to predict the prognosis of breast cancer,which has certain reference significance for the clinical treatment and prognosis of breast cancer.