Altered astrocyte morphology in pathogenic HEPACAM variants: a multipoint analysis and new machine learning framework for 3D Sholl analysis

Astrocytes are intricate glial cells that play crucial roles in brain development and function. Changes in astrocyte shape are linked to altered function and are seen in numerous neurological disorders. HepaCAM, an astrocyte-specific cell adhesion molecule, influences astrocyte branching, tiling, and coupling. Pathogenic HEPACAM variants, which disrupt hepaCAM's self-interaction, lead to megalencephalic leukoencephalopathy with subcortical cysts (MLC), a rare early-onset disorder causing brain swelling, seizures, and cognitive and motor impairment.

The effects of these variants on astrocyte shape have not been thoroughly examined. To address this, we introduced three dominant pathogenic HEPACAM variants into astrocytes from developing mouse cortex and conducted an extensive analysis of astrocyte shape. Employing established methods alongside a novel machine learning algorithm for efficient 3D Sholl analysis, we discovered minor yet significant increases in shape complexity for the G89S pathogenic variant, which hampers hepaCAM's self-association, and the Q56P variant, which affects hepaCAM's cross-interaction.

This outcome contrasts with the shape alterations seen in astrocytes lacking HEPACAM altogether, indicating that dominant variants might influence astrocyte shape through a gain-of-function mechanism rather than a loss-of-function one. Furthermore, our research introduces a fresh machine learning pipeline for streamlined 3D analysis of astrocyte branching complexity, along with a framework for conducting multivariate analysis of astrocyte shape metrics across various conditions.