Abstracts

Rationale: The Epilepsy Multiplatform Variant Prediction (EpiMVP) Center Without Walls is a highly-integrated, multi-center, collaborative project aimed at developing prediction tools for variants of uncertain significance (VUS) in epilepsy genes. We focus on the most common non-ion channel, non-receptor epilepsy-associated genes with substantial VUS impacting clinical practice. The goal is to establish precise algorithms to correctly classify VUS as pathogenic or benign using machine learning algorithms and in silico modeling, combined with biological data from in vitro and in vivo platforms. To this end, our initial studies involve the STXBP1 gene that is implicated in a range of severe childhood-onset epilepsies. Here we describe work using 2D and 3D human pluripotent stem cell (hPSC) models to establish STXBP1 loss-of-function (LoF) and variant phenotypes. These platforms will be used for subsequent assays of VUS rescue effects after forced expression on a STXBP1 LoF background.

Methods: We used CRIPSR/Cas9 genome editing to generate homozygous and heterozygous STXBP1 knockout (KO) H9 female human embryonic stem cells (hESCs) with isogenic controls. We also obtained iPSC lines that were reprogrammed from a STXBP1 patient with a pathogenic variant and from a heathy individual. We used two approaches to establish STXBP1 LoF and pathogenic variant phenotypes. One approach (2D model) is to differentiate STXBP1 hPSCs to cortical-like excitatory neurons by directly inducing expression of the transcription factor Neurogenin-2 (NGN2) in hPSCs, which results in homogenous and synchronized neuronal differentiation, termed iNeurons. The second approach (3D model) is to differentiate STXBP1 hPSCs to human brain organoids that partly recapitulate in vivo cortical development.

Results: To establish STXBP1 LoF phenotypes, we made multiple CRIPSR STXBP1 KO H9 female hESC lines including both homozygous and heterozygous lines, and are currently using these KO and isogenic control lines to generate iNeuron lines. Differentiating KO and control lines to brain organoids is in progress as well. To assess possible phenotypes caused by pathogenic STXBP1 variants, we generated iNeuron lines from a patient (p.Gly544Val fs*2) and a control iPSC line. We are currently differentiating them to iNeurons and will assess STXBP1 variant related phenotypes such as neurodegeneration, morphological abnormalities or impaired synaptic function. We also generated human cortical organoids from the STXBP1 patient and control iPSC lines. We found evidence of STXBP1 protein aggregation in patient organoids at day 58. We also observed that patient organoids were smaller than control organoids, and neurons in patient cortical organoids appeared less robust than those in control organoids.

Conclusions: We successfully generated STXBP1 LoF and patient 2D and 3D hPSC models, and found preliminary phenotypes caused by a pathogenic STXBP1 variant in patient-derived brain organoids. Future assays such as examining synapse formation and function, and multi-electrode array recordings of network activity in 2D and 3D cultures will be exploited to determine epilepsy-like phenotypes.

Funding: Please list any funding that was received in support of this abstract.: This work was supported by NIH (NINDS) U54NS117170.