Abstract
Epilepsy affects millions worldwide with many failing to respond to standard treatments. Drug-refractory epilepsy requiring surgical intervention often presents due to developmental aberrations, such as focal cortical dysplasia (FCD). In FCD type II, somatic mutations in the mTOR pathway allow excessive activation of the mTOR kinase, causing unregulated cell growth and differentiation. Ex vivo studies of resected brain tissues offer a direct window into epileptogenic brain and disease pathophysiology. Recent studies implicate abnormal cell fate specification in FCD, but it is unclear if mutant cells are directly epileptogenic and/or whether they impact neighboring cells. Using a custom 43-antibody mass cytometry panel, we broadly assessed cell lineage and functional state across multiple sequential resections from a female child with FCD harboring a mosaic gain-of-function AKT3 mutation. Our approach used unsupervised machine learning tools to define protein-level features of dysplasia-associated cells and revealed cell populations associated with clinical outcomes. While AKT3 mutational burden did not correlate with clinical outcomes, protein profiling identified a distinct population of CUX1-high cells correlating with seizure control. Furthermore, we identified a similar cell population across multiple additional cortical malformation cases. This work demonstrates the advantage of deep single-cell protein profiling of surgically resected epilepsy tissue and provides insight on disease mechanisms that can be overlooked in traditional genetic or transcriptional studies. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s40478-026-02273-3.