Hydroxysafflor yellow A ameliorates autism-like phenotypes and attenuates ADAM10-associated neuroimmune alterations: an integrative multi-omics and in vivo study

Autism spectrum disorder (ASD) is a heterogeneous neurodevelopmental condition characterized by synaptic dysfunction and neuroimmune imbalance. Current pharmacological treatments for ASD remain largely focused on managing associated symptoms rather than modifying core disease mechanisms, whereas the molecular regulators linking neuroimmune dysregulation to neural circuit abnormalities remain incompletely characterized. Identifying such regulators and exploring whether they are associated with pharmacologically responsive alterations may provide new insights into ASD pathophysiology and potential intervention strategies. We integrated Mendelian randomization, transcriptomic analysis, machine learning, and single-cell RNA sequencing to identify and prioritize ASD-associated candidate genes. Pseudotime, cell–cell communication, and functional enrichment analyses were performed to characterize the regulatory context of the prioritized candidates. Molecular docking and molecular dynamics simulations were used to assess the structural plausibility of the interaction between hydroxysafflor yellow A(HSYA) and A disintegrin and metalloproteinase 10 (ADAM10), followed by in vivo evaluation of HSYA treatment in a valproic acid (VPA)-induced ASD mouse model through behavioral, histopathological, and molecular analyses. Fourteen candidate genes were identified through the integration of Mendelian randomization and transcriptomic analyses. Among these candidates, ADAM10 was further prioritized based on its diagnostic performance and feature contribution in machine-learning analyses. Single-cell analysis showed that microglia were among the cell populations with relatively high proportions of ADAMA10 -expressing cells and revealed dynamic ADAM10 expression across microglial states. Functional enrichment analyses implicated inflammatory signaling, protein processing, and cell adhesion pathways. Cell–cell communication analysis indicated increased microglia–neuron communication in the high- ADAM10 -expression group. Molecular simulations supported a potential interaction between HSYA and ADAM10. In the VPA-induced ASD model, HSYA improved ASD-like behaviors, reduced neuronal damage and glial activation, and reduced ADAM10 protein expression. By integrating genetic, transcriptomic, machine-learning, cell-type-resolved, structural simulation, and in vivo evidence, this study prioritized ADAM10 as an ASD-associated regulatory candidate and further linked ADAM10 expression to microglial states, neuroimmune regulation, and intercellular communication. HSYA ameliorated ASD-like phenotypes and was associated with reduced ADAM10 protein expression in vivo, supporting further investigation of ADAM10-related neuroimmune alterations as potential targets of pharmacological modulation in ASD. These findings extend current evidence implicating ADAM10-related alterations in ASD by providing complementary multi-level evidence. However, the specific contribution of ADAM10 to HSYA-associated effects requires further mechanistic clarification.