Screening And Preliminary Validation Of Biomarkers Related To Ferroptosis In Osteoporosis Based On Bioinformatics Methods

Research background and purpose:Osteoporosis（OP）is one of the most common bone diseases in the world,characte rized by low bone mineral density and susceptibility to pathological fractures,especially in postmenopausal women and elderly men.According to the epidemiological statistics of osteoporosis in China in 2018,the prevalence of OP in people over 50 years old is m ore than 19.2%,especially in women.The prevalence is much higher than the average p revalence,and the number of low bone mass in China is not in the minority^[1].OP is a c hronic disease.Once discovered,osteoporosis is often more serious,and brittle fractures are prone to occur.It has a high disability rate and mortality rate,which seriously affect s the quality of life of the elderly and brings serious burden to families and society.At pr esent,the treatment of osteoporosis is mainly based on oral drugs,including bisphospho nates,estrogen-related therapy,parathyroid hormone analogues,RANKL inhibitors,sal mon calcitonin and other drugs^[2].However,the treatment cycle is often longer,the effec t is slower,and there are certain toxic and side effects（such as liver and kidney function damage,jaw necrosis,etc.）.Therefore,the treatment of osteoporosis still needs to furth er explore new therapeutic drugs.Ferroptosis is one of the newly discovered forms of cell death regulated by genes in recent years.A large number of studies have shown that ferroptosis is closely related to many diseases,including renal cancer,lung cancer and breast cancer.In addition,there are also related studies in other diseases,including Parkinson’s disease,Alzheimer’s disease,myocardial injury,kidney injury and other diseases,and some research results have been obtained^[3].Moreover,with the in-depth study of the molecular mechanism of ferroptosis,people have gradually realized that ferroptosis may also play an important role in many diseases other than tumors.In addition,the treatment of ferroptosis has been gradually proposed.A large number of studies have shown that the use of ferroptosis agonists or inhibitors in tumor diseases can promote or inhibit the development of diseases^[4].Therefore,targeted ferroptosis has gradually become a research hotspot in the treatment of many diseases.However,there are few studies on ferroptosis and osteoporosis,and the mechanism of ferroptosis in osteoporosis is still unclear.Therefore,this study aims to mine biomarkers related to ferroptosis in osteoporosis from GEO database through bioinformatics technology,and conduct a series of analysis of biomarkers,trying to provide a new direction for the diagnosis and treatment of osteoporosis in the future.Materials and Methods:Firstly,gene expression information and clinical information were downloaded from the public database（Gene Expression Omnibus,GEO）,and the ferroptosis gene set was downloaded from the Ferr Db database and published articles.Differentially expressed genes（DEGs）were screened using R software with adjusted p-value<0.05as the screening criterion.The DEGs and ferroptosis genes were intersected and the Venn plot was drawn to obtain the ferroptosis-related DEGs.GO function and KEGG pathway enrichment analysis of ferroptosis-related DEGs were performed by R software.A random forest model was constructed using ferroptosis-related DEGs to further screen key ferroptosis genes,and R software was used to calculate the co-expression correlation between key ferroptosis genes,and Wilcox rank sum test was used for significance analysis.When P<0.05 was considered to be significant correlation.The lnc RNA-miRNA-m RNA-TF regulatory network was constructed in Cytoscape software using key ferroptosis genes.The disease diagnosis model was constructed by LASSO regression,the ROC curve was drawn by R software to evaluate the diagnostic efficiency,and the reliability of the diagnosis model was verified by external data set GSE7429.Small molecule compounds with good docking scores with all key ferroptosis genes were screened from the Drugbank database by molecular docking technology.The screening conditions were Affinity≤-7 kcal/mol.Finally,the m RNA expression levels of key ferroptosis genes in osteoporosis and normal models were verified by animal experiments.SD female rats were divided into Sham group（sham operation group）and OVX group（ovariectomy group）.SD female rats in OVX group underwent bilateral ovariectomy to construct osteoporosis model.Micro-CT was used to verify the osteoporosis of the model.Monocytes in peripheral blood of model rats were extracted for real-time fluorescence quantitative PCR（q RT-PCR）detection to verify the m RNA expression level of key ferroptosis genes in osteoporosis and normal models.Result:A total of 400 DEGs were screened from the GSE56815 dataset,including 282down-regulated genes and 118 up-regulated genes,and 6 ferroptosis-related DEGs were preliminarily screened.GO function and KEGG pathway enrichment analysis showed that the six ferroptosis-related DEGs were mainly enriched in maintaining iron ion homeostasis and copper ion binding in GO function analysis,and KEGG pathway analysis was mainly enriched in ferroptosis,porphyrin metabolism and other related pathways.A random forest model was constructed to further screen and identify five key ferroptosis genes,including CP,FLT3,HAMP,HMOX1 and SLC2A3.Gene correlation analysis showed that CP（cor=-0.41）and FLT3（cor=-0.27）were significantly negatively correlated with SLC2A3,and HMOX1 was significantly positively correlated with HAMP（cor=0.26）,but the correlation between the five key ferroptosis genes was low.The lnc RNA-miRNA-m RNA-TF regulatory network showed that BAZ1B could interact with four key ferroptosis genes（including CP,SLC2A3,HAMP and HMOX1）,while STAT3 could interact with three key ferroptosis genes（including FLT3,HAMP and HMOX1）.The area under the ROC curve（AUC）of the disease diagnosis model constructed by LASSO regression was 0.8956,and the area under the ROC curve（AUC）of the disease diagnosis model of the external validation set GSE7429 was 0.6900.The results showed that the diagnostic performance of the model was good.Three small molecule compounds were screened from the Drug Bank database by molecular docking technology,which had good docking scores with five key ferroptosis genes,including NADH,Midostaurin and Nintedanib compounds.Micro-CT results showed that compared with the Sham group,the Tb.N,BV/TV and BMD of the OVX group were significantly decreased,while the Tb.Sp and Tb.Pf were significantly increased,indicating that the rat osteoporosis model was successfully constructed.The q RT-PCR experiment showed that the m RNA expression levels of CP and FLT3 were significantly up-regulated in the OVX group compared with the Sham group,while the m RNA expression levels of HAMP,HMOX1 and SLC2A3 were significantly down-regulated.Therefore,q RT-PCR results were consistent with the results of bioinformatics data analysis.Conclusion:This study identified five key ferroptosis genes that can be used as biomarkers of osteoporosis,including CP,FLT3,HAMP,HMOX1 and SLC2A3,which can provide new clues for the early diagnosis and treatment of osteoporosis in the future.