Transcriptome-wide Association Study Of Rheumatoid Arthritis Based On Genetic Variation To Predict Gene Expression

Background and ObjectiveRheumatoid arthritis(RA)is a chronic autoimmune disease characterized primarily by inflammation of the synovial lining of the joints,which can lead to joint deformity and loss of function.The prevalence of RA is about 0.5%-1.0% worldwide.RA is a complex disease caused by a combination of factors,and genetic factors play an important role in the development of RA.Evidence from twin studies shows that the heritability of RA is about 50-65%.The search for susceptibility loci and causative genes in RA and a better understanding of the underlying pathogenesis of RA will help develop effective therapeutic targets.In recent years,genome-wide association study(GWAS),a molecular epidemiological tool for studying genetic susceptibility to complex diseases,has been used to screen for single nucleotide polymorphisms(SNPs)significantly associated with diseases through a case-control study design.Currently,GWAS has identified more than100 genetic susceptibility loci for RA.However,these genetic susceptibility loci explain only about 15% of the heritability.In addition,most of the RA-related genetic variants identified by GWAS are located in non-protein-coding regions,and the specific biological mechanisms remain unclear.Based on the fact that gene expression level is an intermediate variable linking genotype and phenotype associations,researchers have proposed the transcriptome-wide association study(TWAS)approach.This method first trains a predictive model of the relationship between genotype and gene expression in a small sample population with genotyping and gene expression data to obtain an estimate of the effect of genotype on gene expression.The model was then applied to genotyping data or GWAS summary data from a large sample of the population to predict the gene expression levels of the corresponding individuals.Finally,the association between gene expression levels and disease phenotypes is evaluated to discover disease susceptibility genes.In this study,we aimed to explore susceptibility genes significantly associated with the risk of developing RA by the TWAS approach.MethodsThis study consisted of three main parts.In the first part,a multi-stage TWAS analysis was performed to identify RA susceptibility genes based on the largest sample size of RA GWAS data(14,361 RA cases and 43,923 healthy controls),with gene expression prediction models from the Genotype-Tissue Expression(GTEx)project and the Netherlands twin register(NTR)project.In the second part,firstly,for the internally validated RA susceptibility genes,the FOCUS method was used to estimate the posterior inclusion probability of genes and further identify genes with possible causal effects.Subsequently,we conducted MAGMA and SMR methods to validate the ability of the TWAS approach to identify RA susceptibility genes.Then,various bioinformatics analyses,including differential expression analysis,colocalization analysis,proteinprotein interaction analysis and pathway enrichment analysis,were performed to gain insight into the genetic structure of RA.In the third part,the expression levels of some of the newly identified genetic susceptibility genes for RA were measured by quantitative real-time polymerase chain reaction(q RT-PCR)in 48 pairs of RA cases and healthy controls.ResultsThe UTMOST method incorporated 17,290 genes with significant heritability and identified 262 candidate RA susceptibility genes at the P < 0.05 criterion after correction by the False discovery rate(FDR).The FUSION method modeled the prediction of 8125 genes in GTEx v8 whole blood data.47 genes met with FDR corrected P < 0.05 in both methods,13 of which were located in genetic susceptibility loci not previously reported in RA GWAS(PSD4,FBXO40,DAP,PGBD1,TRIM10,TRIM27,ZNF322,HIST1H2 BG,YAF2,RNF40,SPNS1,TSSK6,OPRL1).The FUSION method validated 11 of the 47 genes using the prediction model of the NTR data,and there was a high correlation between the results of both datasets(R = 0.60,P < 0.001).FOCUS identified 7 of the 47genes(MMEL1,INPP5 B,AFF3,PDHB,IRF5,GSDMB,TYK2)as possible causal genes for the risk of RA development.MAGMA and SMR-based methods also validated 40 of the 47 genes as RA susceptibility genes.The GSE93272 dataset was analyzed for differential expression of 47 genes,and 15 of them were found to be differentially expressed.8 genes(PTPN22,PUS10,PPIP5K2,ZNF322,IRF5,PSMD5,YAF2,TMEM50B)were upregulated in the RA case,and 7 genes(PGBD1,PHF19,PGAP3,GSDMB,ORMDL3,TYK2,UBASH3A)were decreased in the RA case.Colocalization analysis revealed rs3761959 on the FCRL3,rs1131017 on the RPS26,rs1893592 on the UBASH3 A,and rs2241976 on the PSD4 as shared signals driving gene expression in RA GWAS and whole blood.Protein interactions of 47 genes identified the PTPN22 gene showed the strongest interaction with 10 genes.The results of the pathway enrichment analysis suggested that the RA risk genes were significantly enriched in autoimmune disease-related pathways and inflammatory immune response-related pathways.q RTPCR experiments verified that the expression levels of FBXO40(Z =-3.566,P < 0.001),RNF40(Z =-7.562,P < 0.001),SPNS1(Z =-7.925,P < 0.001)differed between the RA case and the healthy control.ConclusionsIn this study,cross-tissue and single-tissue TWAS analyses were performed on RA GWAS data combined with gene expression weighting data.We analyzed the genetic susceptibility to RA at the transcriptome-wide level and identified 47 genetic susceptibility genes,13 of which were located in genetic loci of RA not previously reported in GWAS.The results of this study provide important evidence to support the uncovering of the genetic structure of RA and guide the development of functional studies.