个体基因组与先兆子痫的交互作用对孤独症谱系障碍症状严重程度的影响 Translated title: Impact of interaction between individual genomes and preeclampsia on the severity of autism spectrum disorder symptoms

Abstract

目的

孤独症谱系障碍(autism spectrum disorder，ASD)是一种复杂的神经发育障碍。既往研究表明，遗传易感性和环境暴露，如母亲妊娠期先兆子痫(preeclampsia，PE)，在ASD的病理机制中扮演重要角色。然而，遗传因素和环境因素的相互作用对于ASD表型严重程度的具体影响尚不明确。本研究通过构建基因-环境模型，旨在探讨个体基因组中的新发变异(de novo variants，DNVs)和常见变异与母亲妊娠期PE暴露的交互作用对ASD表型严重程度的影响。

方法

从西蒙斯孤独症收集项目(Simons Simplex Collection，SSC)数据库中提取4~18岁的单纯特发性ASD患者的表型数据。根据ASD患者的生物学母亲是否患有PE(即患者是否PE暴露)，将研究对象分为PE ⁺组和PE ^-组；将未携带DNVs的研究对象分为DNV ^-PE ⁺组和DNV ^-PE ^-组，将携带DNVs的研究对象分为DNV ⁺PE ⁺组和DNV ⁺PE ^-组；将多基因风险评分(polygenic risk score，PRS)低于其中位数的研究对象分为PRS ^lowPE ⁺组和PRS ^lowPE ^-组，将PRS高于或等于其中位数的研究对象分为PRS ^highPE ⁺组和PRS ^highPE ^-组。ASD的核心表型评估包括适应和认知能力、社交互惠能力、语言和交流能力、刻板行为。适应和认知能力评分包括文兰适应行为量表第2版(Vineland Adaptive Behavior Scales, Second Edition，VABS-II)中自适应行为的综合评分、SSC数据库中评定的言语智商(verbal intelligence quotient，VIQ)、SSC数据库中评定的非言语智商(nonverbal intelligence quotient，NVIQ)。社交互惠能力评分包括孤独症诊断访谈量表-修订版(Autism Diagnostic Interview-Revised，ADI-R)中社会维度(ADI-R social domain，ADI-R SD)评分、孤独症诊断观察表(Autism Diagnostic Observation Schedule，ADOS)中社会情感维度(ADOS social affective domain，ADOS SA)评分、社交反应量表(Social Responsiveness Scale，SRS)的正态性分布评分。语言和交流能力评分包括ADI-R中言语交流维度(ADI-R verbal communication domain，ADI-R VC)评分、ADI-R中非言语交流维度(ADI-R nonverbal communication domain，ADI-R NVC)评分、ADOS中人际交往维度(ADOS communication and social domain，ADOS CS)评分。刻板行为评分包括ADI-R中刻板行为和局限兴趣维度(ADI-R restricted and repetitive behaviors domain，ADI-R RRB)评分、ADOS中刻板行为维度(ADOS repetitive domain，ADOS REP)评分、重复刻板行为量表修订版(Repetitive Behavior Scale-Revised，RBS-R)评分。通过构建线性回归模型，来探讨ASD患者PE暴露及其与个人基因组(包括DNVs和常见变异)的交互作用对其ASD核心表型的影响。进一步使用Metascape网站对与DNVs相关的ASD候选基因进行基因本体论(gene ontology，GO)富集分析，并使用来自BrainSpan数据库中RNA测序(RNA-seq)数据探讨基因的时空表达模式。

结果

纳入2 439例有DNVs信息记录并已明确是否存在PE暴露的ASD患者，其中PE ⁺组146例和PE ^-组2 293例，2组之间SRS( β=2.01， P=0.08)和ADI-R NVC( β=-0.62， P=0.09)的差异有倾向性。在2 439例记录DNVs信息并已明确是否存在PE暴露的研究对象中，DNV ^-PE ^-组、DNV ^-PE ⁺组、DNV ⁺PE ^-组和DNV ⁺PE ⁺组分别有1 454、90、839和56例。单独分析PE暴露的主效应，发现PE暴露在SRS( β=3.71， P=0.01)和RBS-R( β=4.54， P=0.05)方面的影响有统计学意义；分析DNVs与PE的交互作用，发现交互作用在SRS( β=-4.17， P=0.06)方面的影响有倾向性。在2 236例获得PRS的研究对象中，PRS ^lowPE ^-组、PRS ^lowPE ⁺组、PRS ^highPE ^-组和PRS ^highPE ⁺组分别有1 033、72、1 069例和62例。单独分析PE的主效应，发现PE在SRS( β=4.32， P<0.001)和RBS-R( β=5.87， P=0.02)方面的影响有统计学意义；分析PRS与PE的交互作用，发现交互作用在SRS( β=-4.90， P=0.03)和ADI-R NVC( β=-1.43， P=0.04)方面的影响有统计学意义，在NVIQ( β=9.61， P=0.08)和RBS-R( β=-6.20， P=0.08)方面的影响有倾向性。此外，PE暴露的患者所携带的DNVs基因富集于上皮细胞向间充质细胞转变的调节和DNA结合转录因子活性的调节过程。时空表达模式分析表明，这些调节过程富集的基因在出生前的表达水平高于出生后。

结论

作为影响ASD发生的环境因素，PE暴露与ASD表型严重程度的增加有关。同时，PE暴露和遗传因素的交互作用对于ASD表型的变化至关重要。在PE暴露的背景下，具有高遗传风险常见变异的ASD患者在社交互惠和语言交流能力方面可能表现出改善。相较而言，DNVs与PE的交互作用虽然也可能有助于改善症状，但效果不如常见变异显著。这种差异性表明，在同样的PE暴露条件下，携带DNVs或具有高遗传风险常见变异的ASD患者可能会表现出不同程度的症状变化。在ASD病理学研究中应考虑遗传和环境因素的共同作用。

Translated abstract

Objective

Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder.

Prior research suggests that genetic susceptibility and environmental exposures, such as maternal preeclampsia (PE) during pregnancy, play key roles in ASD pathogenesis. However, the specific effects of the interaction between genetic and environmental factors on ASD phenotype severity remain unclear. This study aims to investigate how interactions between de novo variants (DNVs) and common variants in individual genomes and PE exposure affect ASD symptom severity by constructing a gene-environment model.

Methods

Phenotypic data were obtained from the Simons Simplex Collection (SSC) database for idiopathic ASD patients aged 4-18. Subjects were divided based on maternal PE status: PE ⁺ (exposed) and PE ^- (unexposed) groups. Those without DNVs were divided into DNV ^-PE ⁺ and DNV ^-PE ^- groups, and those with DNVs into DNV ⁺PE ⁺ and DNV ⁺PE ^- groups. Based on polygenic risk scores (PRS), subjects below the median were classified into PRS ^lowPE ⁺ and PRS ^lowPE ^- groups, and those at or above the median into PRS ^highPE ⁺ and PRS ^highPE ^- groups. Core ASD phenotypic assessed included adaptive and cognitive abilities, social reciprocity, language and communication skills, and repetitive behaviors. Adaptive and cognitive abilities were scored using adaptive behavior composite scores from the Vineland Adaptive Behavior Scales, Second Edition (VABS-II), along with verbal intelligence quotient (VIQ) and nonverbal intelligence quotient (NVIQ) scores from the SSC database. Social reciprocity abilities were measured using the social domain scores from the Autism Diagnostic Interview-Revised (ADI-R SD), social affective domain scores from the Autism Diagnostic Observation Schedule (ADOS SA), and normalized scores from the Social Responsiveness Scale (SRS). Language and communication abilities were assessed through verbal communication domain (ADI-R VC), nonverbal communication domain (ADI-R NVC) scores from ADI-R, and the communication and social domain scores from ADOS (ADOS CS). Repetitive behaviors were measured using the restricted and repetitive behaviors domain scores from ADI-R (ADI-R RRB), the repetitive domain scores from ADOS (ADOS REP), and the overall scores from the Repetitive Behavior Scale-Revised (RBS-R). Linear regression models were constructed to explore the impact of PE exposure and its interaction with individual genomes (including DNVs and common variants) on core ASD phenotypes. Additionally, ASD candidate genes associated with DNVs underwent gene ontology (GO) enrichment analysis via Metascape, and temporal and spatial gene expression patterns were examined using RNA sequencing (RNA-seq) data from the BrainSpan database.

Results

A total of 2 439 ASD patients with recorded DNV information and confirmed PE exposure status were included, with 146 in the PE ⁺ group and 2 293 in the PE ^- group. There was a trend toward differences between these two groups in SRS ( β=2.01, P=0.08) and ADI-R NVC ( β=-0.62, P=0.09). Among the 2 439 participants, there were 1 454 in the DNV ^-PE ^- group, 90 in the DNV ^-PE ⁺ group, 839 in the DNV ⁺PE ^- group, and 56 in the DNV ⁺PE ⁺ group. Analysis of the main effect of PE exposure showed significant impacts on SRS ( β=3.71, P=0.01) and RBS-R ( β=4.54, P=0.05). Interaction analysis between DNVs and PE exposure revealed a trend toward significance in SRS ( β=-4.17, P=0.06). In the 2 236 participants with available PRS data, there were 1 033 in the PRS ^lowPE ^- group, 72 in the PRS ^lowPE ⁺ group, 1 069 in the PRS ^highPE ^- group, and 62 in the PRS ^highPE ⁺ group. Analysis of the main effect of PE exposure showed significant impacts on SRS ( β=4.32, P<0.001) and RBS-R ( β=5.87, P=0.02). The interaction between PRS and PE exposure showed significant effects on SRS ( β=-4.90, P=0.03) and ADI-R NVC ( β=-1.43, P=0.04), with trends in NVIQ ( β=9.61, P=0.08) and RBS-R ( β=-6.20, P=0.08). Additionally, DNV-enriched genes in PE-exposed patients were associated with regulatory of epithelial-to-mesenchymal transition and DNA-binding transcription factor activity. Temporal and spatial expression pattern analysis indicated that genes enriched in these regulatory processes showed higher expression levels prenatally compared to postnatally.

Conclusion

PE exposure, an environmental factor influencing ASD, is associated with increased ASD symptom severity. The interaction of PE exposure with genetic factors is crucial in modulating ASD phenotypes. Among PE-exposed individuals, ASD patients with high genetic risk for common variants may show improvements in social reciprocity and communication skills. In contrast, while DNVs may also aid in symptom improvement, their impact is less pronounced than that of common variants. These differences suggest that under similar PE exposure conditions, ASD patients with DNVs or high-risk common variants may exhibit varying degrees of symptom changes. ASD pathology research should consider the combined influence of genetic and environmental factors.