Glutamatergic Dysregulation and Compensatory Over-Pruning in Autism Spectrum Disorder: A Polygenic Pathway Analysis

Background: Autism spectrum disorder (ASD) is highly heritable and polygenic, with convergent evidence implicating glutamatergic neurotransmission and synaptic pruning in its etiology. However, the relative contributions and potential interdependence of these pathways remain unclear.Methods: We re-analyzed the largest ASD genome-wide association study (18,381 cases, 27,969 controls) using MAGMA for competitive gene-set testing, stratified linkage disequilibrium score regression for partitioned heritability, and Summary-PrediXcan for tissue-specific transcriptome-wide association in six brain regions. Custom annotations targeted core (23 genes) and expanded (130 genes) glutamatergic sets, shortened (38 genes) and expanded (262 genes) pruning sets, a pruning-specific set (225 genes after overlap subtraction), and negative controls.Results: Glutamatergic annotations showed strong heritability enrichment (1.45-fold for core set), while pruning pathways contributed independently (1.08-fold expanded, persisting at 1.05-fold after glutamatergic subtraction). Transcriptomic imputation revealed modest pruning inflation, highlighted by predicted GABBR1 downregulation in reward circuitry alongside cytoskeletal upregulation, suggesting hyperexcitability-driven over-pruning.Conclusions: ASD common-variant risk primarily disrupts glutamatergic plasticity, triggering secondary excessive pruning and circuit hypo-connectivity. This cascade model refines excitatory–inhibitory imbalance hypotheses, nominates timed interventions, and provides a framework for stratifying polygenic liability in neurodevelopmental disorders.