Connectome Dysfunction Of Brain Network System In Major Depressive Disorders And Its Association With Gene Expression Profiles

Objective:Major depressive disorder（MDD）is one of the most common mental disorders in the world.It has high incidence rate,high recurrence rate and high disability rate.The lifetime prevalence rate is 12%.MDD is mainly manifested in emotional disorders,decreased interest,impaired cognitive function,and even suicidal thoughts or behaviors.It is a disease that seriously threatens human health.However,the pathophysiological mechanism of MDD remains unclear.Research shows that MDD is a mental disease with abnormal brain connection.Therefore,exploring the neuropathological mechanism of MDD from the perspective of brain connection has important scientific significance and clinical value.In recent years,the rapid development of resting state functional magnetic resonance（rs-f MRI）imaging technology provides an important means for noninvasive exploration of human brain function.Using the method of graph theory,the brain of MDD can be modeled as a brain network composed of nodes and edges.Brain network is served as the best explanatory model to clarify the biological mechanism of MDD.Resting f MRI technology combined with graph theory analysis provides a favorable means to characterize various attributes of human brain network.Based on this research framework,the healthy human brain functional network can be described as an optimized modular network architecture,which is composed of spatially separated functional modules.The connections within the modules are dense,but the connections between modules are sparse.This modular network organization plays a vital role in maintaining the balance between the separation and integration of brain functions and maintaining individual cognitive and behavioral abilities.More and more evidence shows that the modular structure of healthy human brain network is damaged in MDD patients.The specific reason is the connection barrier within and between the functional systems involving advanced network and primary network,which leads to the decline of related cognitive processing and emotion control ability.Network connector node plays an important role in modular structure.It is closely connected with different modules.It plays a vital role in cross module information transmission,maintaining functional integration between network modules and coordinating network integrity.It is found that the modular structure destruction of MDD preferentially targets the network connector nodes mainly involving intersecting communication and multiple cognitive components.Therefore,the research on the integration between MDD brain function network modules based on network nodes will be more helpful to explore the modular changes related to diseases,so as to find the abnormal connection mode of MDD patients’brain function network system,which will help us to understand the biological mechanism behind MDD.MDD is a moderate genetic disease with heritability of 31-42%.There is increasing evidence that genetic factors play an important role in brain connectome.Twin studies and genome-wide association studies（GWAS）also show that resting state functional connectivity（rs FC）in human brain is moderately or highly inherited.Genetic neuroimaging studies have shown that the risk genotypes of MDD affect brain structure and function through molecular and cellular mechanisms.These risk genes are related to different brain connection patterns and indirectly regulate phenotypes at the behavioral level.Therefore,the combination of MDD brain function network and genomics will provide new clues for exploring the pathophysiological mechanism of MDD and the gene regulation mechanism behind MDD.The whole brain gene expression map builds a bridge between brain network and transcriptome.The Allen Atlas of the human brain（AHBA）microarray dataset can be mapped to the same anatomical space as human imaging data,which makes it possible to explore how the molecular mechanism of genome-wide transcription is related to brain f MRI dynamics and connectivity.At present,AHBA microarray data set has been used to identify transcriptome related to human neuroimaging,indicating an association between gene expression and function related loops.Combined with neuroimaging and Allen Brain Atlas gene transcription data,studies have explored the relationship between brain structural and functional abnormalities related to a variety of mental diseases and gene expression.However,there is no study using transcriptome neuroimaging correlation analysis to explore the gene expression related to the modular structure of brain functional network in patients with MDD.This study will establish a fusion model of connectomics and transcriptomics.Firstly,the abnormal mode of brain function network system connection in MDD patients was explored by the method of system participation coefficient（PC）;Then,combined with the whole gene expression data of AHBA,we further explore the relationship between the abnormal connection mode of brain function network system and gene expression profile in patients with MDD,so as to provide a new perspective and new ideas for exploring the pathophysiological mechanism of MDD.Methods:A total of 559 subjects aged 18-45 years were collected in this study,including 262 MDD patients and HC297 matched by gender and age.Part I:The cortical functional network was constructed by resting-state functional MRI.Through the system-based PC method,the PC groups were compared to find the different brain regions of PC,and to clarify the changes of functional integration between brain functional network systems in patients with MDD.To further study the abnormal connection mode between modules of PC abnormal brain areas in patients with MDD,we took PC abnormal brain areas as seed points to analyze the differences between them and other modules and functional connectivity（FC）between the two groups,to clarify the abnormal connection mode of brain functional network system in patients with MDD.This study further analyzed the correlation between brain network indicators（PC,FC）and behavioral indicators（including disease severity,cognitive impairment and course of disease）in patients with MDD,so as to clarify the relationship between abnormal connection mode of MDD brain function network system and clinical symptoms.Part II:The correlation between brain network indexes（PC,FC）and the whole gene expression data set of AHBA was analyzed.The partial least squares regression models of case-control PC value,case-control FC value and gene data were established by partial least squares regression analysis to find the related genes of PC change mode and FC change mode.Further,go and KEGG enrichment analysis were carried out for the genes that contribute greatly to the changes of PC and FC,respectively,to clarify the molecular and biological functions of the related genes.Results:Part I:（1）There are PC elevation connector nodes in bilateral visual network of MDD patients,including right primary visual cortex（V1）,left second visual area（V2）,bilateral third visual area（V3）,left third visual auxiliary area（V3A）,bilateral ventromedial visual areas（VMV3,VMV2,VMV1）,right posterior striate cortex（Pro S）and right parahippocampal gyrus 1（PHA1）,P_FDR<0.05.（2）In MDD patients,the FC of the connector node with increased visual network PC decreased in vis,the FC with default network（DMN）,frontal parietal network（FPN）,ventral attention network（VAN）and LIM increased outside VIS,and the FC with sensorimotor network（SMN）decreased.PFDR<0.05.In the visual network of MDD patients,the connector nodes with increased PC are mainly FC decreases between V1,V2,V3 and VMV and between them and the sixth visual area（V6）,Pro S,etc;Outside the VIS network,it is mainly associated with the increase of FC in bilateral S32,bilateral orbitofrontal complex（OFC）,lateral area of bilateral orbitofrontal complex（13I）,right10d,left insular inferior frontal gyrus 4（FOP4）of VAN,left 9-46 and left 46 of FPN,bilateral H and other brain areas of LIM,and P_FDR<0.05.（3）There was a significant negative correlation between the total score or factor scores of HAMD-17 and FC among nodes of within visual network.The most significant combination of correlation:FC between right VMV2 and right Pro S was significantly negatively correlated with core depression factors,with correlation coefficient R=-0.205and P_FDR=0.0002;FC between right Pro S and right V3 was significantly negatively correlated with HAMD scores,with correlation coefficient R=-0.201 and P_FDR=0.0003.FC values between left Pros、left V2 and left VMV1 within the visual network were significantly positively correlated with M8（speech fluency score）in MCCB(the correlation coefficients were R=0.276,P_FDR=0.0425;R=0.273,P_FDR=0.0425).Part II:（1）PLS1（the first component of PLS regression）associated with PC change explained 20.1%of the variation of PC change mode of MDD（Permutation Test,P<0.05）,and there was no other PLS component.Association between the transcriptional pattern of PLS1 with high expression characteristics of VIS region and the PC abnormal pattern of MDD.This component has a significant spatial positive correlation with the PC abnormal pattern of MDD in spatial distribution（R=0.46,P<0.0001）.The PLS1 gene list is sorted according to the weight from the best positive one to the best negative one,and the genes with the absolute value of gene weight greater than 5 are selected as the most significant gene list for enrichment analysis.（2）PLS1（the first component of PLS regression）associated with the abnormal pattern of VIS connection explained 23.6%of the variation of the change pattern of VIS connection of MDD（Permutation Test,P<0.05）,and there was no other PLS component.PLS1 shows the association between the transcriptional pattern characterized by low expression in the visual network region and the abnormal pattern of VIS connection in MDD.The spatial distribution of this component was significantly positively correlated with the abnormal mode of VIS connection of MDD（Correlation coefficient with each connection R=0.32～0.66,P<0.0001）.The PLS1 gene list is ranked according to the weight from the best positive one to the best negative one,and the genes with the absolute value of gene weight greater than 5 are selected as the most significant gene list for enrichment analysis.（3）Enrichment of genes related to PC abnormal pattern:GO terms show that cell components are mainly related to axons,protein binding in molecular function,and various biological processes such as"trans-synaptic signaling","regulation of ion transport"and"metal ion transport"in biological processes.KEGG pathways mainly located in"cancer pathway","glutamatergic synapse"and"Ras signaling pathway".（4）Enrichment of genes related to VIS abnormal pattern:GO terms show that cell components are mainly related to axons,and in biological processes,they are mainly related to various biological processes such as"metal off transport","ion transport regulation"and"ion homeostasis",no molecular functional terms.It is mainly involved in"cancer pathway"and"calcium signaling pathway","c AMP signaling pathway"and other KEGG pathways.（5）Comparison of genes related to abnormal PC mode and abnormal VIS connection mode:because the PC change mode of MDD is highly consistent with the genes contributing significantly in the list of genes related to abnormal vis connection mode of MDD(847 genes contribute significantly to both,and the same genes account for 95.3%（847/889）of the genes contributing significantly to the PC change mode,It accounted for 41.4%（847/2046）of the genes significantly contributing to the abnormal connection pattern of vis.For the change mode of brain functional system connection in patients with MDD,the direction of gene expression weight is highly consistent.8 of the20 pathways annotated by the two GO are the same,of which 7 are in the first 10pathways,mainly including"axon","cross synaptic signal",etc;The KEGG pathway of the two are the same in four ways:"cancer pathway"and"glutamine synapse","calcium signaling pathway"and"axon guidance".Conclusions:Using large sample data and high-resolution connection analysis framework,this study found the abnormal connection mode of brain function network system in MDD patients through PC method,which is characterized by the high integration of visual network in MDD patients,mainly the wide integration of VIS（primary network）and advanced network（DMN,FPN and van）,and the high integration of vis and Lim,This means that there is a phenomenon of dedifferentiation in the organizational structure of brain functional network of MDD;The decreased flexibility of brain function network information processing and affective regulation disorder in patients with MDD may be related to the extensive cognitive changes and affective disorders in MDD.The integration of vis and primary network（SMN）decreased,which means that the primary perception of information interaction obstacles.The decrease of internal integration of visual network indicates that the degree of modular separation of visual network is weakened,and it is found that this change is related to the severity of depression in patients with MDD and the decrease of speech fluency in patients with MMD.In addition,this study found that the abnormal connection mode of brain functional network system in patients with MDD was significantly correlated with the gene expression mode of biological processes such as synapse,cross synaptic transmission and ion transport,revealing the molecular biological basis of abnormal connection of functional network system in patients with MDD at rest.It provides new imaging transcriptomic evidence for the research viewpoint that synapse related pathways and ion transport related pathways are related to the pathophysiology of depression.