A Molecular Epidemiology Study Of Attention Deficit Hyperactivity Disorder Based On The Transcriptome-wide Association Strategy

Objective:Attention deficit hyperactivity disorder（ADHD）is one of the most common neurodevelopmental disorders during childhood and adolescence.As a multifactorial complex disease,its average heritability is as high as 74%.However,the genetic variants identified by genome-wide association studies（GWASs）so far can only explain about 22%of the heritability,and are mainly located in non-coding regions,of which the action mechanisms are difficult to be elucidated.Therefore,based on the transcriptome-wide association strategy,this study aimed to systematically identify regulatory genetic variants associated with the risk of ADHD,and to explore their biological mechanisms in ADHD,which could provide new theoretical bases for screening of high-risk individuals and individualized prevention and treatment of the disease.Methods:Firstly,based on 20 gene expression-genetic variant weight datasets derived from ADHD-relavant brain and non-brain tissues from databases such as the Genotype-Tissue Expression and Common Mind Consortium,and the summary statistics from the latest ADHD GWAS,a transcriptome-wide association study（TWAS）was performed to integrate genomic,transcriptomic,and disease phenotypic information to identify genes whose cis-genetically regulated expression components were associated with ADHD,which was validated and complemented by summary data-based Mendelian randomization（SMR）analysis;Then,annotated by multiple databases and annotation tools,functional genetic variants that regulate the expression of the identified susceptibility genes were selected.Next,a multicenter two-stage case-control study was designed to analyze the associations of the selected genetic variants with the risk of ADHD in Han Chinese children.Using the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders（DSM-IV）as the mian diagnostic tool,a total of 543 newly diagnosed ADHD children and560 controls were recruited in Wuhan,Hubei in the discovery stage;316 cases and 332controls were recruited in Changsha,Hunan in the validation stage.The Swanson,Nolan,and Pelham-IV（SNAP-IV）rating scale and the Integrated Visual and Auditory Continuous Performance test（IVA-CPT）were used to assess the clinical symptoms of ADHD.A self-designed questionnaire was applied to collect information on the subjects,their families and their parents.Peripheral blood of the subjects was collected and genotyped by Sequenom Mass ARRAY.The associations between genetic variants and ADHD susceptibility in different genetic models were analyzed by unconditional logistic regression,and analysis of variance was applied to compare the clinical characteristics among different genotypes of each variant.In addition,unconditional logistic regression model was used to analyze environmental factors associated with ADHD.Then,multi-factor gene-environment interaction was explored using classification decision tree and generalized multi-factor dimensionality reduction（GMDR）.Furthermore,unconditional logistic regression and additive interaction analysis excel sheet were used to analyze two-factor gene-gene and gene-environment multiplicative and additive interactions,respectively.Finally,for the identified risk variant rs3768046,dual-luciferase reporter gene assay was performed to explore its effect on the promoter activity of TIE1,and then electrophoretic mobility shift assay（EMSA）was conducted to analyze its influence on the binding affinity to nucleoprotein.For the other identified risk variant rs1054037,dual-luciferase reporter gene assay was conducted to analyze its effect on the binding of hsa-mi R-5591-3p to its target sequences at the 3′-untranslated region（3′UTR）of MANBA.Moreover,real-time fluorescence quantitative polymerase chain reaction（RT-q PCR）was applied to detect the RNA expression levels of TIE1 and MANBA in the peripheral blood samples from 20 ADHD children and 20 controls.Results:TWAS identified a total of 38 ADHD susceptibility genes,which were mainly distributed on chromosomes 1,4,and 11.SMR identified a total of 5 ADHD risk genes on chromosomes 1,4,and 11.Combining the results of multiple bioinformatics analyses,including TWAS,SMR and functional annotation,32 regulatory genetic variants of 8protein-coding genes located on chromosomes 1,4 and 11 were preferentially selected.The results of the two-stage case-control study showed that TIE1 rs3768046 and MANBA rs1054037 were significantly associated with the risk of ADHD in the discovery,validation and combined stages.In the combined stage,TIE1 rs3768046（G>A）was significantly associated with a reduced risk of ADHD in the co-dominant,dominant and additive models(co-dominant model,GA vs GG:OR=0.64,95%CI:0.52-0.80,P_FDR<0.001;dominant model,GA+AA vs GG:OR=0.63,95%CI:0.51-0.78,P_FDR<0.001;additive model,AA vs GA vs GG:OR=0.68,95%CI:0.56-0.81,P_FDR<0.001).Meanwhile,there were significant differences in the attention deficit and hyperactivity/impulsivity scores based on SNAP-IV,visual,auditory and full-scale attention quotients acquired from IVA-CPT among different genotypes of rs3768046(P_FDR<0.05).MANBA rs1054037（C>T）was observed to be significantly correlated with an increased risk of ADHD in co-dominant,dominant and additive models(co-dominant model,TC vs CC:OR=1.51,95%CI:1.23-1.86,P_FDR<0.001,TT vs CC:OR=1.61,95%CI:1.2-2.15,P_FDR=0.002;dominant model,TT+TC vs CC:OR=1.53,95%CI:1.26-1.86,P_FDR<0.001;additive model,TT vs TC vs CC:OR=1.32,95%CI:1.15-1.52,P_FDR<0.001).In addition,the hyperactivity/impulsivity score obtained from SNAP-IV,visual,auditory and full-scale response control quotients based on IVA-CPT among different genotypes of rs1054037 were siginificantly different(P_FDR<0.05).No significant multiplicative or additive interactions were found between rs3768046 and rs1054037（P>0.05）.Logistic regression analysis showed that low paternal educational level（OR=1.29,95%CI:1.05-1.59,P=0.017）,maternal pre-pregnant overweight or obesity（OR=1.55,95%CI:1.21-2.00,P=0.001）,preterm birth（OR=1.98,95%CI:1.32-2.98,P=0.001）and childhood atopic disease（OR=1.35,95%CI:1.05-1.73,P=0.019）were risk factors for ADHD.Multi-factor interaction analysis found that the classification decision tree model including rs3768046,maternal pre-pregnant overweight or obesity,and childhood atopic disease had the best prediction performance.Meanwhile,GMDR analysis showed that the combination of attributes including rs3768046,pre-pregnant overweight or obesity,childhood atopic disease and preterm birth was the best interaction model.Two-factor interaction analysis showed that there was a multiplicative interaction between maternal pre-pregnant overweight or obesity and rs3768046（P=0.048）.Compared with children carried with rs3768046 AA/GA genotypes and whose mother was not overweight before pregnancy,children with rs3768046 GG genotype and whose mother was overweight or obese before pregnancy had a 1.7-fold increased risk of developing ADHD（OR=2.70,95%CI:1.89-3.84,P<0.001）.Furthermore,the dual-luciferase reporter gene assay showed that rs3768046 G allele enhanced the promoter activity of TIE1,and the result of EMSA suggested that rs3768046might bind some transcription factors in an allele-specific manner to alter the expression of TIE1.Moreover,the dual-luciferase reporter gene assay showed that rs1054037（C>T）disrupted the binding of hsa-mi R-5591-3p to MANBA 3′UTR,which upregulated the gene expression.In addition,the result of RT-q PCR showed that the expression levels of TIE1（P=0.001）and MANBA（P=0.019）in the peripheral blood of ADHD children were significantly higher than those of controls.Conclusions:A transcriptome-wide association strategy was adopted to identify ADHD susceptibility genes.Among them,TIE1 rs3768046 and MANBA rs1054037 were significantly associated with the risk of ADHD in Han Chinese children.Besides,there was an interaction between rs3768046 and maternal pre-pregnant overweight or obesity.Furthermore,rs3768046 may alter the transcriptional activity of TIE1 by affecting the binding of transcription factors to the promoter region,and rs1054037 may influence the binding of micro RNAs to the target sequence at MANBA to alter the gene expression,which might thereby increase the risk of ADHD.