Epilepsy-Linked RAPGEF5 Variant Shifts Protein Out of the Nucleus in Belgian Shepherds

A tiny genetic insertion that raises the risk of idiopathic epilepsy in Belgian shepherd dogs appears to do its damage by misplacing a protein inside the cell.

The variant in question is a 3-base-pair insertion in exon 1 of the RAPGEF5 gene. It was previously identified through a genome association study of well-phenotyped Belgian shepherds whose epilepsy status was classified at the International Veterinary Epilepsy Task Force’s Tier 1 confidence level. That study revealed a four-base risk haplotype on canine chromosome 14, located upstream of RAPGEF5, and dogs carrying the risk haplotype were fully concordant with the presence of the disruptive 3-bp insertion. At the population level, the variant was found in both related and unrelated dogs at a frequency of 19.6 percent, and the average age of idiopathic epilepsy onset in dogs carrying it was 42 months.

RAPGEF5 encodes a guanine nucleotide exchange factor that participates in the Wnt signaling pathway, an evolutionarily conserved network critical for normal embryonic development. Wnt family members help establish cell polarity, define mitotic capacity, and determine cell fate during embryogenesis, and the pathway also maintains tissues in adulthood. The most well-described branch, the canonical Wnt pathway mediated through β-catenin, regulates proliferation and neurodevelopment. Signaling through this branch requires β-catenin to translocate into the nucleus, where it interacts with transcription factors to drive Wnt-associated gene expression. Without nuclear translocation, β-catenin is degraded and Wnt signaling is impaired. RAPGEF5 has been shown to facilitate this translocation by maintaining RAP proteins in their GTP-bound, activated state, allowing those activated RAP proteins to physically interact with β-catenin and promote its movement into the nucleus.

Disruptions in Wnt/β-catenin signaling have been linked to a range of neurological problems, including altered left-right patterning in the embryo, neurodegeneration, and both acute and chronic epilepsies. Genetic studies of human epilepsy, including subjects diagnosed with idiopathic epilepsy, have uncovered associations with members of the Wnt pathway, and a potential role for RAPGEF5 in epilepsy has been implicated in humans and in mouse models. In one RAPGEF5 knockdown model, decreased nuclear and increased cytoplasmic β-catenin were observed alongside reduced Wnt gene expression. These converging lines of evidence suggested that the canine insertion variant might be more than a harmless marker linked to a nearby disease gene.

The insertion itself is small but structurally significant. It introduces one additional alanine amino acid into a stretch of ten consecutive alanines, adjacent to the conserved Dishevelled, Egl-10, and Pleckstrin domain, known as the DEP domain. Homorepeats like this are often engaged in protein-protein interactions and subcellular trafficking, and they can be prone to expansion through DNA strand slippage, with a propensity to aggregate and lead to disease. Expanded homorepeats have been associated with neurodegenerative disease, and comparable alanine expansions in other genes cause recognized disorders. In humans and mice, a polyalanine expansion in the Aristaless Related Homeobox gene causes intellectual disability, epilepsy, and brain malformations, and that gene positively regulates the Wnt/β-catenin pathway with a similar spatial expression profile to RAPGEF5. The DEP domain, meanwhile, directs RAPGEF5 to cellular membranes, permitting interactions with binding partners, so the insertion also has the potential to change the protein’s configuration in that region.

The dogs included six idiopathic epilepsy cases and seven healthy controls over the age of seven years, some related as parents and offspring, half-siblings, or cousins sharing grandparents. Whole blood samples were collected by the dogs’ own veterinarians, DNA was extracted and sequenced on an Illumina NovaSeq 6000 platform at an average of 12-fold coverage, and the consensus cDNA was aligned to the CanFam3.1 reference genome. The SNPEff annotation program characterizes the insertion as a disruptive inframe insertion with moderate effect, and in silico modeling with the program Phyre2 predicts it changes the protein’s secondary structures.

The researchers then synthesized custom cDNA constructs of both the wildtype and the risk variant, each including the 5′ untranslated region, and cloned them into an expression vector carrying enhanced green fluorescent protein to create fusion proteins that could be tracked visually. These were transiently transfected into Madin-Darby Canine Kidney cells, a canine cell line chosen to ensure species-relevant expression since RAPGEF5 is naturally expressed in kidney cells, though not at high levels. Cells were plated at 50,000 cells per well to reach 70 to 80 percent confluency, transfected with 2.5 micrograms of purified plasmid using Lipofectamine 2000, and counterstained with Hoechst 33342 to mark the nuclei in blue. Positive and negative controls, consisting of cells treated with a standard GFP plasmid and untreated cells, were plated in triplicate across two plates in two separate experiments.

Seventeen to twenty-three hours after transfection, before cell division, live cells were imaged at 40x magnification using a confocal fluorescence microscope. Fifteen images per treatment well were captured, and the fluorescent localization of every cell in each well was quantified as nuclear, cytoplasmic, or both. The results were striking. Wildtype RAPGEF5 protein was found abundantly in the nucleus as well as in the cytoplasm, and in a large proportion of cells it appeared in both compartments simultaneously. The risk variant, by contrast, was detected in the cytoplasm 1.5 times more often than the wildtype protein, a difference that reached statistical significance at p less than 0.05. Combined nuclear and cytoplasmic localization was 1.7 times more frequent in cells carrying the wildtype protein than in those with the variant, though this difference did not reach significance at p equal to 0.06.

The authors interpret these findings through the lens of β-catenin biology. Because the variant did not abolish nuclear-associated RAPGEF5 entirely, some functionality would be maintained, but a protein with reduced nuclear presence could impair Wnt signaling by reducing β-catenin gene transcription. The effects could manifest during development, creating aberrations in neural network circuitry, and continue through adulthood through reduced neurogenesis and tissue maintenance. They also note that seizures have a bimodal effect, promoting both neuronal apoptosis and neurogenesis through the Wnt/β-catenin pathway, and that in humans seizure episodes potentiate additional seizures, so a defective RAPGEF5 protein could alter normal neuronal circuitry development and also impair seizure recovery, potentially resulting in greater seizure severity or frequency with increasing age.

The study has limitations the authors acknowledge directly. The GFP tag is a large protein added to the N-terminus of RAPGEF5, which could affect distribution in the transfection model, though the observed distribution of the labeled wildtype protein was consistent with published literature, suggesting the construct is acceptable and the results reflect the true altered distribution of the variant. The work was performed in cultured kidney cells rather than neurons, and the degree to which the variant reduces nuclear β-catenin activity in a given dog remains unclear. Further studies using genetically modified mice carrying the RAPGEF5 variant could yield additional insight into its role in epileptogenesis. Still, the finding that a single alanine insertion shifts the protein’s cellular address provides a functional explanation for why this locus, identified through genomic association, is a true risk locus for idiopathic epilepsy, and it further implicates Wnt/β-catenin signaling in both canine and human epilepsy. Because more than 70 percent of genes associated with seizures in a human infantile seizure expression study were in the Wnt/β-catenin pathway, the parallel between species offers potential avenues for interventions to mitigate idiopathic epilepsy in both, whether through selective breeding in dogs or, eventually, therapeutic approaches targeting β-catenin signaling.

Subject of Research: Cellular localization of a RAPGEF5 risk variant associated with idiopathic epilepsy in Belgian shepherd dogs

Article Title: Cellular localization of a variant RAPGEF5 protein associated with idiopathic epilepsy risk in the Belgian shepherd

Article References: Cayabyab, D. D., Belanger, J. M., Xu, C., Maga, E. A., & Oberbauer, A. M. (2024). Cellular localization of a variant RAPGEF5 protein associated with idiopathic epilepsy risk in the Belgian shepherd. Canine Medicine and Genetics, 11(1), Article 4. https://doi.org/10.1186/s40575-024-00138-3

Image Credits: AI Generated

DOI: 10.1186/s40575-024-00138-3

Keywords: idiopathic epilepsy, Belgian shepherd, RAPGEF5, Wnt signaling, beta-catenin, 3-bp insertion, cellular localization, canine genetics, MDCK cells, GFP fusion protein, DEP domain, alanine homorepeat