Pathological disruption of CELF2 shuttling causes neuronal hyperactivity, learning deficits, and seizures
Michelle Hua, et al.
Michelle Hua, et al.
View: Text | PDF
Research Article Clinical Research Development Genetics

Pathological disruption of CELF2 shuttling causes neuronal hyperactivity, learning deficits, and seizures

Abstract

De novo heterozygous variants in CUGBP Elav-like family member 2 (CELF2) have recently been associated with a rare neurodevelopmental disorder, yet the mechanisms linking specific variants to distinct clinical phenotypes remain poorly understood. Here, we reported a cohort of 18 individuals and provided evidence that variants causing CELF2 mislocalization, but not protein-null variants, were associated with seizures. Using proband-derived human cortical neurons and transgenic mouse models, we demonstrated that CELF2 underwent activity-dependent nucleocytoplasmic shuttling in excitatory neurons and that its cytoplasmic retention caused neuronal hyperactivity, elevated seizure susceptibility, and learning and memory deficits. We further found that cytoplasmic CELF2 regulated mRNAs critical for synaptic function and neuronal excitability and implicated in epileptic seizures and intellectual disability. Drug screening further identified AKT signaling as a key regulator of CELF2 nucleocytoplasmic shuttling and a candidate target for reversing neuronal hyperactivity. Together, our findings expand the clinical and genetic spectrum of CELF2-related neurodevelopmental disorders and establish a variant-specific mechanism that links CELF2 mislocalization to neuronal hyperactivity, seizures, and cognitive impairment.

Authors

Michelle Hua, Mohamad-Reza Aghanoori, Melissa J. MacPherson, Yi Ren, Shehani V. Siripala, Yifan Yang, Yvonne Yan Yan Or, Malea Nguyen, Robert Duba-Kiss, Daniel Feng, Laura Williams, Christopher J. Gafuik, GengYi Wang, Chloe Quelin, Boris Keren, Sarah Schuhmann, Georgia Vasileiou, Alexia Bourgois, Antonio Vitobello, Christophe Philippe, Zornitza Stark, Richard J. Leventer, George McGillivray, Frederic Tran Mau-Them, Marine Tessarech, Clément Prouteau, Phillis Lakeman, Mahdi M. Motazacker, Donald R. Latner, Raymond C. Caylor, Yvette van Ierland, Eloise Prijoles, Angie Lichty, Evangelos Theodorou, David A. Sweetser, Edward Steel, Jan Cobben, Majed J. Dasouki, Daniel G. Calame, Bertrand Isidor, Benjamin Cogné, Mitchell Kesler, Brooke Rackel, Isabel Clark, Deborah M. Kurrasch, G. Campbell Teskey, James Ellis, Guiqiong He, Scott D. Ryan, Douglas J. Mahoney, A. Micheil Innes, Jonathan R. Epp, Guang Yang

×

Figure 1

CELF2 missense variants causing CELF2 mislocalization are associated with epileptic seizures.

Options: View larger image (or click on image) Download as PowerPoint
CELF2 missense variants causing CELF2 mislocalization are associated wit...
(A) Schematic showing the positions of identified variants in the CELF2 protein, containing 3 RNA recognition motifs (RRM1–3). PTVs are shown in blue and missense variants in red. Colored dots indicate effect of missense variants on CELF2 subcellular localization; black circles denote seizure occurrence in corresponding individuals. (B) Heatmap of identified variants and associated clinical features, including ES (epileptic seizures), GDD (global developmental delay), and SD (speech delay). A feature’s presence is indicated by a black box; a gray box denotes unavailable information. (C and D) Bar graphs showing normalized mRNA levels of minigene reporters with the indicated variants, compared with WT, in HEK293 cells (C) or hiPSCs (D), treated with or without cycloheximide (CHX; 4 hours) as determined by qPCR. Data are presented as means ± SEM, normalized to WT. Each dot represents 1 experiment. One-sample 2-tailed t test. (E) Confocal images of HEK293 cells expressing WT EGFP-CELF2 (green) or the indicated variants. White box areas are shown at higher magnification below. Nuclei were counterstained with Hoechst 33258 (blue) and are outlined with dashed white lines. “N” denotes the nucleus. (F) Quantifications of cytoplasmic/nuclear ratio of CELF2 levels, from E. n = 5 (100 cells each). One-way ANOVA, Dunnett’s post hoc test, compared with WT. (G) Schematic of CELF2 protein variants tested in E, marked by vertical red bars, grouped by their subcellular localization pattern (nucleus “N” vs. cytoplasm “C”) and associated seizure status. Truncated protein lacking RRM3 is depicted without the domain. Scale bars, 5 μm.