Assigning Targetable Molecular Pathways to Transdiagnostic Subgroups Across Autism and Related Neurodevelopmental Disorders

Abstract

Significant genetic, behavioural and neuroanatomic heterogeneity is common in autism spectrum- and related- neurodevelopmental disorders (NDDs). This heterogeneity constrains the development of effective therapies for diverse patients in precision medicine paradigms. This has led to the search for subgroups of individuals having common etiologic factors/biology (e.g., genetic pathways), thus creating potential uniformity in prognosis and/or treatment response. Despite NDDs having a strong genetic component, only ~15-20% of individuals will present with a specific rare genetic variant considered clinically pathogenic, and therefore, subtyping efforts tend to focus on using clinical, cognitive, and/or brain imaging phenotypes to group individuals. Here we delineated mechanisms via mouse to human translational neuroscience. Using MRI derived structural neuroanatomy and a spatial transcriptomic comparison, we linked subgroups of 135 NDD relevant mouse models (3,515 individual mice) separately to two human databases, with 1,234 and 1,015 human individuals with NDDs, composed of autism, attention-deficit/hyperactivity disorder (ADHD), obsessive compulsive disorder (OCD), other related NDDs, and typically developing controls. Subgroups were significantly linked by consistent neuroanatomy across all three datasets, mouse and human, indicating that direct cross-species subgrouping and translation is consistent and reproducible. Ultimately, four specific neuroanatomical clusters were found and linked to precise molecular mechanisms: two showing a chromatin/transcription motif, with one of those showing specific links to G-protein coupled receptors (GPCR) and Notch signalling, and another two being mainly synaptic in origin, with one off those showing specific connections to axon guidance and Wnt signaling. Assigning molecular pathways, and thus genetic information, from the mouse to individual participants provides an insight into undetected and/or related genetic variants that could be working in combination or interacting with an environmental influence. Moreover, the subgroups found are transdiagnostic, including participants with autism, ADHD, and OCD, which indicates that NDDs as a whole can be subdivided into consistent neuroanatomical clusters with cohesive underlying biological mechanisms. This work allows us to bridge the gap between preclinical models and human disorders, linking previously idiopathic human patients to pertinent genetics, molecular mechanisms, and pathways.