In Vivo Modeling of Pathogenic DDX3X Missense Variants During Corticogenesis Underscores Cellular and Molecular Heterogeneity of DDX3X Syndrome

Abstract

Heterozygous mutations in the X-linked RNA helicase DDX3X cause DDX3X syndrome. DDX3X syndrome is a rare neurodevelopmental disorder associated with intellectual disability and autism spectrum disorder, as well as cortical malformations and microcephaly. These clinical findings in the central nervous system implicate DDX3X in neurodevelopment. Indeed, loss-of-function (LoF) mouse models have revealed a key role for DDX3X in progenitors of the cerebral cortex: DDX3X controls these progenitors’ ability to generate excitatory neurons and regulates both cell cycle length and translation of a subset of mRNAs. However, amongst ~200 known pathogenic DDX3X variants, approximately half are predicted loss-of-function (LoF) variants, while the remaining half are missense variants. In addition, missense variants are associated with more severe clinical presentations of DDX3X syndrome, suggestive of additional pathogenic mechanisms not represented in LoF models. As such, how missense mutations in DDX3X impact corticogenesis in vivo is unknown. In this work, we identify and characterize mouse models for two representative pathogenic variants in DDX3X to provide the first understanding of DDX3X syndrome missense variants’ impacts on corticogenesis in vivo. First, we generate and validate a conditional mouse model of Ddx3xT532M, a clinically severe and recurrent DDX3X syndrome mutation found in affected females. Then, using Cre-mediated expression of Ddx3xT532M in cortical progenitors, we show that Ddx3xT532M alters corticogenesis. Ddx3xT532M conditional hemizygous (cHemi) males have severe microcephaly and apoptosis. In contrast, Ddx3xT532M conditional heterozygous (cHet) females exhibit mild reductions in cortical size and neurogenesis. Using polysome fractionation of Ddx3xT532M and Ddx3xLoF cHet female cortices, we discover that Ddx3xT532M affects translation in cHet females with qualitative differences from Ddx3xLoF cHet females. Collectively, these findings demonstrate that while the Ddx3xT532M and Ddx3xLoF cHet females have similar impacts on corticogenesis, they have distinct molecular targets. In addition, we validate and characterize a second publicly available conditional mouse model of the DDX3X syndrome pathogenic variant Ddx3xR488H, also observed exclusively in female patients and associated with a mild clinical presentation. Using this second model in combination with a cortex-specific Cre, we discover that heterozygous or hemizygous expression of this mild variant in cortical progenitors has no discernible impact on cortex size or excitatory neuron output, consistent with its mild clinical presentation in DDX3X syndrome. Finally, we leverage the conditional models established in Chapters 2 and 3 in combination with a ubiquitous Cre driver to generate full knock-in mice and, as previously done, characterize early postnatal brain size and excitatory neuron output of corticogenesis, as well as body size. Consistent with the clinical data, Ddx3xT532M/+ and Ddx3xR488H/+ present with microcephaly and low birth weight, respectively, compared to their littermate controls. Interestingly, these phenotypes are specific to the two missense models and are absent from Ddx3xlox/+ haploinsufficient mice, suggesting a mechanism of regulating overall embryonic development shared among the two missense models that differs from Ddx3xLoF. Together, this thesis characterizes corticogenesis in two mouse models of DDX3X syndrome pathogenic missense variants, the clinically severe variant Ddx3xT532M and Ddx3xR488H. In addition, this work leverages these conditional models in combination with a whole-embryo Cre driver to generate and provides initial insights into their corresponding DDX3X syndrome clinical models. These models have provided and will continue to provide key insights into the etiology underlying the clinical heterogeneity of DDX3X syndrome secondary to missense variants.