Abstracts

This abstract has been invited to present during the Basic Science Poster Highlights poster session

Rationale: Approximately one-third of focal epilepsy patients show no structural abnormality on MRI (non-lesional focal epilepsy, NLFE) even though histopathology may reveal subtle cortical malformations. Many NLFE patients are resistant to anti-seizure medications (ASMs) and are also poor candidates for epilepsy surgery. Surprisingly, somatic loss-of-function (LOF) variants in SLC35A2 are a common finding in resected NLFE tissue as well as focal cortical dysplasia type I. SLC35A2 encodes UGT-1, a UDP-galactose transporter that transports UDP-galactose from the cytosol to the lumen of the Golgi apparatus. UGT-1 is a known modulator of protein glycosylation and SLC35A2 variants are associated with aberrant protein N-glycosylation in brain. However, the effects of SLC35A2 variants on cortical development are not well understood. We hypothesize that Slc35a2 knockout (KO) in fetal mouse brain results in altered cortical laminar structure in vivo.

Methods: We designed a CRISPR/CAS9 plasmid construct (targeting Slc35a2 exons 2/3) for Slc35a2 KO. Mouse Neuro2a cells (N2aC) were transfected using Lipofectamine in serum free media with either the Slc35a2 KO plasmid or a Scramble (Scr) plasmid to create KO and Scr cell lines. Slc35a2 KO was validated using Western blot (Anti-UGT-1; Sigma Aldrich) and qPCR analysis. In utero electroporation (IUE) at E14.5 was used to model the effects of somatic mutations in the fetal brain. Once mice pups were born, brains were extracted, formalin-fixed, paraffin embedded, and sectioned at 8 microns. Immunohistochemistry (IHC) was used to visualize Slc35a KO neurons expressing a reporter GFP (Thermofisher, 1:200) and cortical laminar structure using Ctip2 as a marker for layer V neurons (Abcam, 1:100) and SATB2 (abcam, 1:100) as a marker for layer II-IV neurons. All images were captured using a Keyence microscope. Number of sections with GFP+ neurons were used to calculate region of interest for the electroporation range of affected neurons. This range was used to represent variant allele frequency (VAF) and model range of brain mosaicism in human patients

Results: Slc35a2 KO showed no mRNA expression by qPCR and no UGT-1 protein expression (WB) in vitro confirming Slc35a2 KO. IUE KO pups displayed cortical neuronal malpositioning when compared to controls. Slc35a2 KO neurons were located in various layers throughout the cortical electroporated region as well as within the white matter, whereas the control eGFP and Scr neurons were appropriately located within layers II/III. Electroporated mice displayed a range of affected cortical space displaying GFP positivity, modeling the range of VAF seen in affected human patients.

Conclusions: Slc35a2 KO in vivo resulted in altered cortical laminar structure, displaying malpositioned neurons outside their proper cortical layer as well as within the white matter. These results provide new insights into how loss of UGT-1 may lead to establishment of an epileptic network, as well as a potential model for brain somatic mosaic mutations.

Funding: 1R01NS115017-01A1