Imaging Genetics Research Of Parkinson's Disease Based On Multimodal Magnetic Resonance Imaging

BackgroundParkinson's disease?PD?is one of the most common neurodegenerative diseases.In addition to motor symptoms,there are also non-motor symptoms such as cognitive decline.Levodopa is still the main treatment for PD,which may lead to motor complications after long-term treatment.Previous studies also have reported multiple single nucleotide polymorphisms?SNPs?,which might increase the risk of developing cognitive decline and motor complications?MC?.Recently,imaging genetics analysis has become an important tool to study the mechanism of genetic variations.In this study,neuroimaging features were extracted as intermediate phenotype,and its correlation analysis with genetic data was carried out.Single SNP analysis and polygenic risk score?PRS?were used to explore the influence of gene variation on brain function and structure.We aimed to reveal the brain phenotype-genotype-clinical phenotype associations,investigate the genetic and nueroimaging basis of PD,and provide view and methods to explore the mechanism of PD related cognitive function,motor complications and other clinical symptoms.MethodsWe recruited 45 PD patients and 39 healthy controls?HCs?.Clinical data was collected and cognitive ability was evaluated in all subjects.Evaluation of five cognitive domains were calculated based on these three cognitive scales.Severity of PD symptoms were evaluated in PD patients.SNPs were screened using mass array SNP genotyping technique.Ultra-high field magnetic resonance imaging?MRI?images were acquired using a 7T MRI research system with prototype sequences including the resting-state functional MRI?rsf MRI?,diffusion tensor imaging?DTI?and structural MRI?s MRI?.Image processing was performed by the DPABI,FSL,SPM/VBM8softwares to obtained imaging features of brain function,white matter fiber and gray matter,which were regared as intermediate phenotypes.Then,we performed the imaging genetics analysis.Experiment 1:Imaging genetics analysis between PD cognition-related SNPs and neuroimaging intermediate phenotypeWe selected SNPs related to PD cognitive function.Age,sex,and education were entered as covariates due to their potential influence.Firstly,the main effect of genotypes and its interactive effect with PD disease status were explored in single SNP analysis.Then,the PRS score was calculated,and the correlation between the PRS score and imaging intermediate phenotype was analyzed.We extracted the average value of imaging features in the brain regions which survived after multiple comparison correction,and its correlations with cognition evaluation scales were analyzed.Experiment 2:Imaging genetics analysis between PD motor complications-related SNPs and neuroimaging intermediate phenotypePD patients were further divided into two groups according to whether they had motor complications or not.Selected PD motor complications related SNPs were screened.We also performed single SNP analysis and PRS analysis.Age and sex were entered as covariates due to their potential influence.We extracted the imaging features of brain areas with statistical significance which survived multiple comparison correction,and performed correlation analysis to evaluate the correlations between imaging features and motor symptom and MC severity.ResultsExperiment 1:Results of imaging genetics analysis between PD cognition-related SNPs and neuroimaging intermediate phenotypeIn single SNP analysis,the A allele of COMT-rs4680 was related to the amplitude of low-frequency fluctuation?ALFF?values of frontal lobe,lingual gyrus,fusiform gyrus and occipital lobe.And its interactive effect with PD disease status was found in the gray matter volume?GMV?value of temporal lobe.PD patients with A allele showed significantly positive correlations between GMV value of inferior temporal gyrus and PANDA score,Visuospatial score.The A allele of SNCA-rs356181 was negatively related to ALFF values of putamen and temporal pole,positively related to the Re Ho values of olfactory cortex,putamen,caudate nucleus and frontal lobe,and also positively related to the GMV values of temporal lobe and fusiform gyrus.At the same time,the ALFF values of inferior temporal gyrus and fusiform gyrus were also affected by the SNCA-rs356181 geneotype×PD disease state interaction.The SNCA-rs894278 variant altered the ALFF values of frontal lobe,angular gyrus,lingual gyrus and temporal lobe.In addition,it interacted with PD disease status to affect the f ALFF value of lingual gyrus and calcarine gyrus,FA value of anterior corona radiate.PD patients without C allele showed positive correlation between f ALFF value of calcarine gyrus and Visuospatial score,and HC subjects without C allele also showed positive correlation between FA value of anterior corona radiate and cognitive sacle.Similar results were found in the analysis of SNCA-rs3910105.Both main effect of this genotype and its interaction with PD disease status influenced the brain function activity of frontal lobe.In addition,the interactive effect also existed in corpus callosum,its FA value was positively correlated with the cognitive scores.In regards of the DRD3-rs6280,subjects with an G allele had significant lower ALFF value of temporal lobe,higher regional homogeneity?Re Ho?values of frontal lobe and precentral gyrus,higher GMV values of frontal lobe,hippocampus,parahippocampal gyrus,temporal obe,caudate nucleus,putamen and amygdala.In addition,the Re Ho value of frontal lobe was also affected by the interactive effect between DRD3-rs6280 and PD status.We calculated both the PRS_unweightedand PRS_weighted,considering the weak effect of multiple SNPs.In HCs,we found that the PRS_unweightedwas related to the ALFF value of prefrontal cortex and anterior cingulate gyrus.The PRS_weightedwas negatively correlated with the ALFF value of precuneus.And the ALFF value of precuneus was positively correlated with Language score and Visuospatial score in PD patients.At the same time,the PRS_weightedwas positively related the FA value of the superior corona radiate in PD patients,while negatively correlated with the FA value of the corpus callosum and the posterior corona radiate in HCs.Experiment 2:Results of imaging genetics analysis between PD motor complications-related SNPs and neuroimaging intermediate phenotypeIn single SNP analysis,the A allele of COMT-rs4680 was found to be related to higher ALFF value of supplementary motor area?SMA?,lower ALFF values of cerebellum and fusiform gyrus.The interactive effect between this SNP and PD disease status was shown in the ALFF value of cerebellum.The ALFF value of cerebellum was positively correlated with MDS-UPDRS part II and part III in PD-A-carrier patients,which was absent PD-non-carrier.In regards of the DRD3-rs6280,subjects with an G allele had lower ALFF values of putamen and insula,higher GMV values of caudate nucleus,insula and cerebellum.The DRD3-rs6280 geneotype×PD disease state interactive effect mainly affected the Re Ho value of SMA and GMV value of insula.In addition,the GMV value of insula was positively correlated with Levodopa equivalent dose daily?LEDD?in PD-G-carrier patients.Furthermore,the C allele of EIF4EBP2-rs1043098 had higher f ALFF values of SMA,precentral gyrus,and lower Re Ho value of cerebellum.The PRS_unweightedand PRS_weightedof PD patients were significantly higher than HC patients.However,PD-MC patients only showed a trend of higher PRS scores than those of PD-non-MC patients.MDS-UPDRS part IV assessed the severity of motor complications.No significant correlations between this score and PRS scores were found in total group of PD patients,while there was a significant negative correlation in PD-MC patients.Similar results were obtained in the analysis of these two PRS models.PRS scores of both PD and HC subjects were negatively correlated with ALFF value of cerebellum.In addition,ALFF values of caudate nucleus and putamen were positively related to the PRS scores in HC subjects.ConclusionsIn this study,we analyzed the dosage effect of SNPs related to PD cognition,and results showed that these five SNPs infulenced the imaging features of frontal lobe,temporal lobe,hippocampus,which played important roles in cognitive function.These indicated that SNPs were involved in the regulation of cognitive function in PD through its dosage effect on brain function and structure.Then we introduced the PRS score into imaging genetics study of PD for the first time,which comprehensively considers the cumulative effect of multiple SNPs.Although there was no correlations between PRS score and clinical scores,this score still affected the brain phenotypes.However,the brain phenotype-genotype-clinical phenotype association patterns were different in PD and HC.This study also aimed to investigate the effect of MC-related genetic factors on brain phenotypes.Single SNPs analysis showed that these three SNPs participated in the development of PD motor symptoms and motor complications by regulating the activity of motor cortex and the possible compensatory effect of cerebellum.At the same time,we also put forward the MC-related PRS score for the first time in imaging genetics study of PD.We found that the PRS scores of PD group were significantly higher than HC group,and the PRS scores of PD-MC were negatively correlated with the severity of motor complications.The scores also correlated with the neural activity of basal ganglia and cerebellum.The chief originality of this study is that we combined the ultra-high field 7T multimodal MRI,clinical features and genetic data,to investigate the dosage of a single SNP and the cumulative effect of multiple SNPs and reveal the brain phenotype-genotype-clinical phenotype association pattern.This study could provide a new way to screen genetic and neuroimaging biomarkers for PD,as well as to provide ore information for individualized and precise medical diagnosis and treatment.