The role of UPF3A and UPF3B in early development and neural differentiation

Abstract

The nonsense mediated mRNA decay (NMD) pathway degrades transcripts with premature termination codons (PTCs) to prevent the production of C-terminally truncated proteins that might have dominant negative properties. Loss of function mutations and copy number variation in genes required for NMD have also been implicated in neurodevelopmental disorders (NDDs) such as autism spectrum disorders (ASD), intellectual disability (ID), childhood onset schizophrenia and attention deficit hyperactivity disorder (ADHD). One such gene, UPF3B, located on the X-chromosome, is the only NMD factor that has a gene paralog UPF3A, located on chromosome 13. Both are implicated in NDDs. Loss of function mutations in UPF3B lead to UPF3A protein stabilisation, a phenomena proposed to compensate for loss of UPF3B in order to maintain residual NMD activity and during critical periods of development. The impact of loss of UPF3B function in neurodevelopment has been investigated in UPF3B patient lymphoblastoid cell lines and mice and leads to a deregulation of mRNA important for brain function and development. The role of UPF3A has not yet been elucidated. UPF3A was initially shown to act as a weak NMD activator until recently when it was shown to primarily act to inhibit NMD of a majority, but not all, NMD-targeted mRNAs. In mouse loss of UPF3A has been shown to be embryonic lethal. Thus far, the role of NMD and UPF3A and UPF3B in particular, have yet to be determined in human cells of the developing brain. To elucidate the roles of UPF3A and UPF3B in human cells in a neurodevelopmental model, UPF3A and UPF3B knockout (KO) human Embryonic Stem Cells (hESCs) were generated using the CRISPR/Cas9 genome editing technology. Whole genome sequencing (WGS) of the gene-edited clones revealed that none of the small nucleotide variants (SNVs) detected in each gene-edited clone overlapped with the predicted off-target sites, however, notable genetic variation in each clone was attributed to extended cell culture as part of the CRISPR/Cas9 editing process, presence of mosaicism in the parental cell line and normal passaging. Larger deletions and duplications (structural variants (SVs)) were detected in each gene-edited clones that overlapped with the predicted off-target sites. The hESCs were subsequently differentiated into neural stem cells (NSCs) and transcriptome wide mRNA sequencing analysis was performed both on the UPF3A and UPF3B KO hESCs and NSCs. Loss of UPF3A in hESCs did not have any significant effect on the transcriptome of hESCs but lead to deregulation of 3.03% of transcripts in NSCs, the majority of which (2.96%) were downregulated. Genes deregulated in UPF3A KO NSCs were involved in intracellular signalling pathways regulated by the TGF-β superfamily pathway such as extracellular matrix remodelling, blood vessel development and morphogenesis and regulation of cell adhesion. Mutations in genes that encode components of the extracellular matrix and the vascular system lead to embryonic lethality in mice. Loss of UPF3B had an impact at both the hESC and NSC stage. Loss of UPF3B in hESCs lead to a deregulation of 0.62% (of which 0.21% were upregulated and 0.41% were downregulated) while loss in NSCs resulted in a 0.29% deregulation (0.07% upregulated and 0.22% downregulated). Loss of UPF3B in hESCs lead to a deregulation in genes that are important in cell-cell adhesion, calcium ion binding, synapse assembly and regulation of signalling receptor activity and post synaptic membrane potential. In NSCs, loss of UPF3B lead to deregulation of genes important for neurodevelopment such as of ARX, an ID gene and neural function such as ROBO2 which is important for axonal guidance. Our results suggest that UPF3A could act as an NMD inhibitor and is important in regulating the TGF-β superfamily pathway, while UPF3B is important for normal neurodevelopment and function. The deregulated genes in UPF3A and UPF3B KO NSCs only had two genes that overlapped suggesting that these paralogs have separate roles in NMD. In conclusion the CRISPR/Cas9 genome editing tool and hESCs were efficient and faithful tools to model NDDs caused by mutations in UPF3A and UPF3B. They provided resources to investigate the loss of UPF3A and UPF3B in a suitable cell type that would otherwise be impossible to acquire or only able to acquire post mortem.