| Background:Depression and Alzheimer’s disease（AD）are common comorbidities among old adults,seriously threatening their life and survival.Genetic factors play a significant role in the occurrence and development of both diseases.Cognition and memory as endophenotypes of AD are often used in the genetic studies of AD due to the superiority of endophenotype in the genetic studies of complex mental diseases.Many studies have evaluated the heritability and genetic variants of depression,cognition,or memory,respectively,but there are few studies focusing on the genetic correlation of them,and the multivariate genome-wide association study（GWAS）of them has not been reported.Therefore,this study aimed to:（1）evaluate the genetic correlation of them by multivariable genetic model;and（2）identify the shared single nucleotide polymorphisms（SNPs）,genes,and pathways of them using multivariate GWAS.Methods:The participants were twins from the Qingdao Twins Registry,with informed consent.Depression,cognition,and memory were measured using the 30-item Geriatric Depression Scale（GDS-30）,the Montreal Cognitive Assessment（Mo CA）,and the digit span task of the Wechsler Adult Intelligence Scale,respectively.Blood samples were taken after the participants fasted overnight,and identification of zygosity was carried out by gender,blood type,and the techniques of microsatellite DNA gene scanning and typing.Mx was used to construct a multivariate Cholesky decomposition model to assess the shared genetic and environmental effects among the three phenotypes,adjusting for gender and age（cognition was also adjusted for education level when cognition scored）.Infinium Omni2.5Exome-8v1.2Bead Chip from Illumina was used for genotyping in dizygotic twins.On the basis of the linkage disequilibrium（LD）principle,IMPUTE2 software was utilized to impute un-typed SNPs with reference to the data of the third phase of the 1000 Genomes Project.Genome-wide efficient mixed model association（GEMMA）was used for SNP-based analysis.The regional association plots were drawn using Locus Zoom.Haplo Reg v4.1 and Regulome DB were used to annotate and score for predicting the possible regulatory SNPs when there were many non-coding SNPs with high LD in a region.Additionally,Haplo Reg v4.1 was also used for enhancer enrichment analysis.Versatile Gene-based Association Study-2（VEGAS2）and pathway scoring algorithm（PASCAL）software were applied to carry out gene-based analysis and evaluate pathway scores,respectively.Results:There were 382 twin pairs in the final full sample after the screen of inclusion and exclusion criteria.The medians（inter-quartile ranges）of age and depression,cognition,and memory scores for participants were 50（12）years and 7（7）,22（6）,and 12（3）points,respectively.The full and nested models of the multivariate Cholesky decomposition model were fitted.The best model was the AE model,and the genetic correlations（r G）（95%CI）of depression with cognition,depression with memory,and cognition with memory were-0.33（-0.48,-0.18）,-0.34（-0.49,-0.18）,and 0.65（0.54,0.75）,respectively.The multivariate GWAS was performed in 139 pairs of dizygotic twins,and there were1,338,905 and 7,399,084 SNPs included by strict quality control before and after imputing,respectively.The findings were as follows:（1）In the SNP-based study,the Q-Q plot suggested no evidence of population stratification.The Manhattan plot depicted that a total of 164 SNPs reached the level of suggestion significance(P<1×10-5),among them,rs3967317(P=1.21×10-8)on the CNTN4gene in chromosome 3 exceeded the genome-wide significance level(P<5×10-8).The following were rs9863698(P=7.80×10-8)and rs3967316(P=1.33×10-7)on the CNTN4 gene,rs9261381(P=5.68×10-7)on the TRIM31 gene,rs11577464(P=5.82×10-7)on the LINC02567gene,rs73198369(P=7.00×10-7)on the RNU6-1325P gene,etc.Many SNPs with high LD exceeded the level of suggestion significance in the 6p22.1region as shown in the regional association plot,and it predicted that rs2301751,rs76702759,rs9261377,rs9261379,and rs9261305 might influence regulation.Enhancer enrichment analysis found that depression-cognition-memory-related genetic variants were significantly enriched in six tissues and cells,including pancreatic islets,stomach mucosa,fetal intestine large,primary natural killer cells and T regulatory cells from peripheral blood,and liver（P<0.05）.After imputing un-typed SNPs referencing the data of the third phase of the 1000Genomes Project,the amounts of SNPs with suggestion significance grew,and a total of 457SNPs reached the level of suggestion significance,such as 6:24597173,rs12210323,rs12213116,rs61783213,rs9589468,rs67007022,etc.（2）In the gene-based study,no statistically significant gene was found(P<2.61×10-6),but 1,107 genes reached the suggestion significance level（P<0.05）,and 1,166 genes reached the level of suggestion significance after imputing.The results of most genes were consistent before and after imputing,reaching the suggestion significance level,such as TMEM257,TEX26,MEDAG,LSM10,ZMAT4,KIAA0232,etc.（3）In the pathway-based study,587 and 592 pathways were significantly associated with depression-cognition-memory before and after imputing（emp-P<0.05）,respectively.Most of the pathways could be repeated before and after imputing,reaching the significance threshold,such as METABOLISM OF AMINO ACIDS AND DERIVATIVES,SPHINGOLIPID DE NOVO BIOSYNTHESIS,SPHINGOLIPID METABOLISM,N＿GLYCAN ANTENNAE ELONGATION IN THE MEDIAL TRANS GOLGI,N＿GLYCAN ANTENNAE ELONGATION,IMMUNOREGULATORY INTERACTIONS BETWEEN A LYMPHOID AND A NON-LYMPHOID CELL,etc.Conclusions:（1）There was a moderate genetic correlation among depression,cognition,and memory in Qingdao twins,and（2）the three phenotypes shared multiple SNPs,genes,and pathways,which might have a shared genetic mechanism.The findings provide evidence for further evaluating the shared genetic basis of depression,cognition,and memory and clues for exploring the shared genetic pathogenesis of depression with AD. |
PDF Full Text Request