Abstracts

Rationale: Valproic acid (VPA) is commonly prescribed for epilepsy treatment. However, VPA increases the rate of neural tube defects (NTDs) such as spina bifida when administered to pregnant mothers. VPA is known to inhibit histone deacetylase (HDAC), folic acid metabolism, GSK3ß, and increase oxidative stress; however, the exact mechanism by which VPA leads to NTDs is undetermined. To model NTD teratogenicity, we utilized a novel brain organoid protocol we developed, the self-organizing, single-rosette cortical spheroid (SOSRS) method. Unlike previous techniques which pattern a 3D embryoid body, we pattern a 2D monolayer followed by induction of neurulation into 3D SOSRS. After establishing the utility of this system for modeling neuroteratogenic compounds, we investigated the cause of VPA-dependent dysmorphic changes in the neural tube-like single rosette SOSRS by utilizing inhibitors of four pathways known to be modulated by VPA.

Methods: We treated neurulating SOSRS with inhibitors of cytoskeletal reorganization, including the Rho-kinase inhibitor, Y-27632, and myosin inhibitor, blebbistatin for 48 hours. The human iPSC line used to generate SOSRS expressed a ZO-1/EGFP fusion protein that constitutively labelled the lumen. For each treatment, 20-24 SOSRS were whole-mount imaged for GFP and bis-benzamide, and area, circularity, and solidity were quantified for the lumen and whole SOSRS. We then applied this approach to VPA-exposed SOSRS and examined whether VPA effects are replicated by inhibitors of HDAC, folic acid, or GSK3ß signaling pathways, or attenuated by antioxidant or folate treatment. For all data, Krukal-Wallis one-way analysis was performed with Dunn’s multiple comparison post-test.

Results: Both Rho-kinase and myosin inhibitors increased lumen size (lumen area/total SOSRS area) and decreased circularity. Exposure of SOSRS to varying concentrations of VPA (25-400 µM) caused SOSRS lumens to become significantly enlarged in a dose-dependent manner, beginning at 200 µM (0.274 ± 0.57, n = 66 compared with 0.199 ± 0.045 for vehicle, n = 71, p < 0.0001) and increasing at 400 µM VPA (0.315 ± 0.088, n = 56, p < 0.001). HDAC (trichostatin-A) and folic acid pathway (aminopterin) inhibitors did not have a significant effect on lumen size (p = 0.8583 and p > 0.9999, respectively). Additionally, supplementing with folic acid or antioxidants (vitamin E and resveratrol) in combination with 200 or 400 µM VPA did not ameliorate the phenotype (p > 0.9999 each). Treatment with the GSK3ß inhibitor, CHIR99021 (10 µM) mimicked the enlarged lumen effects of VPA (p < 0.0001 compared to vehicle, p > 0.9999 compared to 400 µM VPA).

Conclusions: These data strongly suggest that inhibition of GSK3ß is the mechanism by which VPA leads to lumen structural dysfunction, indicating a likely pathway by which VPA causes NTDs. Future experiments will be needed to corroborate the role of GSK3ß in this process. This study also establishes the utility of our new model system to investigate compounds and pathways that lead to NTDs.

Funding: Please list any funding that was received in support of this abstract.: This work was funded by NIH/NINDS NS116250.