ASH1L is Necessary for Normal Development of Upper Layer Cortical Neurons

Abstract

Over 100 genes have been implicated in the genetic etiology of autism spectrum disorder (ASD). Many high confidence ASD genes are involved in chromatin remodeling, histone modifications and DNA methylation. Among these is ASH1L (Absent, Small, Homeotic-Like 1), which catalyzes histone H3K36 methylation and has a role in activating transcription by counteracting Polycomb repression. Over 130 autistic individuals have heterozygous loss-of-function ASH1L variants, and population studies confirm it is a high-risk autism gene. Several studies report autism-like behaviors in Ash1l deficient mice and characterized aspects of the underlying neuropathophysiology. However, the underlying genetic aberrations caused by Ash1l loss-of-function are poorly understood in the context of cortical development. We used mice with a cre-inducible deletion of Ash1l exon 4, which results in a frame shift and premature stop codon (p.V1693Afs*2). Firstly, we assessed the impact of global Ash1l loss-of-function on survival and craniofacial skeletal development. The proportion of homozygous Ash1l germline knockout embryos was normal at e18.5, however no live Ash1l null pups were present at birth (e18.5: n = 77, P = 0.90; p0: Ash1l+/+ n = 41, P = 0.00095). Ash1l-/- had shortened nasal bones at e18.5 (n = 31, P = 0.017) compared to heterozygous and wildtype littermates. To assess Ash1l loss-of-function specifically in the cortex, we used a tamoxifen-inducible cre strain to knockout Ash1l early in cortical development (Emx1cre-ERT2; e10.5). We performed immunohistochemistry with antibodies for SATB2, an upper layer marker, and BCL11B, a layer V marker. We found Ash1lcKO had a greater number of SATB2 neurons and they were distributed through the cortical plate (n = 6/genotype, P = 0.0001), whereas BCL11B neurons were similar between genotypes. We performed birthdating of cortical neurons by injected pregnant females with IdU at e14.5 and EdU at e15.5; no differences in neuronal birthdating were identified (n = 4-6/genotype, P = 0.40 for e14.5 and P = 0.057 for e15.5). We utilized single cell RNA sequencing to compare cortical cell populations and identify differentially expressed genes between Ash1lcKO and Ash1lCtrl e18.5 embryos. This revealed numerous differences in gene expression that were sufficient to cluster control and mutant upper layer neurons separately. We found that Ash1lcKO upper layer neurons had altered pseudotime compared to controls, suggesting underlying differences in cell differentiation trajectory. Together, this analysis reveals that ASH1L serves important roles during cortical development, specifically guiding cell fate via gene regulation. Understanding these mechanisms will be fundamental in mitigating Ash1l-associated ASD outcomes and aid in treatment development.