Intellectual disabilities (ID) are a heterogenous group of neurodevelopmental disorders affecting cognition, adaptive learning and memory. The ability to study commonalities and differences in genetically-linked ID within or across patient populations can advance knowledge on the molecular and cellular mechanisms contributing to clinical phenotypes and appropriate treatment options. Chapter 1 addresses types of disease models used to understand neurodevelopmental disorders and key considerations for developing models to benefit understanding on disease pathology and connect clinical observations. Chapter 2 focuses on developing a human neurodevelopmental model for the ID risk gene, PPP2R5D by reprogramming patient-derived fibroblasts into induced pluripotent stem cells (iPSCs) and subsequently differentiating into multiple neuronal cell states. The work represented in this dissertation will focus on developing a model for a heterozygous de novo missense mutation which changes glutamic acid (E) to lysine (K) at the 198 amino acid (E198K) in PPP2R5D. Apart from a higher number of E198K cortical neurons, differences in neuronal morphology or complexity between the E198K or isogenic neurons were not observed. Some PPP2R5D variants have also been linked to Early-Onset Parkinson’s Disease (EOPD) phenotypes which led to a midbrain neuronal differentiation to understand if dopamine-related dysregulation could be captured in a neurodevelopmental model. RNA sequencing revealed upregulation of genes associated with dopamine secretion in midbrain neurons which narrows the scope of what molecular or cellular pathways could be linked to PPP2R5D variants. In addition, this disease model expresses key markers of midbrain cell-fate and shows differences in neuronal complexity between the E198K and isogenic neurons, suggesting the disease model can be used as a tool to decipher PPP2R5D related molecular and morphological dysregulation during midbrain development. In Chapter 3, a novel CRISPR-based approach for editing single nucleotides at the transcript level is evaluated in E198K neural stem cells (NSCs). PspdCas13b-ADAR2DD RNA editing of the E198K variant is dependent on sgRNA length and A-C mismatch. An sgRNA of 190nt or 230nt in length is critical to forming the dsRNA substrate for ADAR2DD adenosine-to-inosine conversion at the PPP2R5D E198K site. In addition, conferring the target adenosine nucleotide within the E198K site with the “A-C” mismatch in the center of the sgRNA is optimal. A screen of available PspdCas13b-ADAR2DD constructs confirmed variable editing efficiency at the E198K site and a previously published site. The work described in this dissertation shows that RNA editing in a stem cell line is possible and at an upwards of 5.3% frequency based on a transient method, thus supporting future work to develop a more translational CRISPR RNA editing platform. In Chapter 4, the discussion is focused on considerations for improving the CRISPR RNA editors as well as the importance of expanding disease models to reflect multiple PPP2R5D variants in an effort to advance understanding on the molecular and cellular mechanisms and how to develop an appropriate candidate therapeutic for individuals with PPP2R5D mutations.

Huntington’s Disease (HD) is an autosomal dominant neurodegenerative disorder caused by a trinucleotide repeat in exon 1 of the Huntingtin (HTT) gene. This expansion leads to protein misfolding that causes widespread cellular dysfunction and ultimately leads to neuronal cell death. The advent of nuclease-deficient CRISPR-Cas9 (CRISPR-dCas9) gene regulation technologies allows targeting of the causative gene and subsequent downregulation via fused effector domains that induce heterochromatin at the epigenetic level through DNA methylation (DNMT3A/L) and H3K9me3 deposition (KRAB), blocking transcription. Therefore, we propose using dCas9 epigenetic editing to downregulate HTT as a therapeutic approach for HD. HTT has large haplotype blocks that allow for allele-specific targeting based upon the presence of heterozygous single nucleotide polymorphisms (SNPs). These SNPs are in genomic regions devoid of NGG PAM sites, a requirement for spdCas9 binding. To address this bottleneck, a screen was conducted of multiple dCas9 variants fused to KRAB and DNMT3A/L with increasingly expanded PAM targeting to initially assess the ability to downregulate total HTT. Surprisingly, only spdCas9 was able to significantly knockdown HTT, while expanded PAM site variants dxCas9 and dCas9-VQR were less efficient in reducing HTT expression. ChIP-qPCR of the HTT promoter showed a decrease in the binding efficiency of dCas9 variants, likely leading to the decreased efficiency of HTT downregulation. We further investigated DNA methylation changes through reduced representation bisulfite sequencing, showing high on-target increases in DNA methylation and few off-targets. In addition, we demonstrate mitotically stable HTT silencing of up to 6 weeks in vitro in a rapidly dividing cell line. We then assessed total HTT knockdown in HD patient-derived fibroblasts and neuronal stem cells. We identified significant downregulation of HTT in our treatment group compared to an unguided control. RNAseq was used to identify differential gene expression between treatment and controls, and gene ontology analysis was performed to identify the rescue of biological processes involved in HTT molecular pathogenesis. Off-targets were assessed through overlaying genome-wide changes in H3K9me3 enrichment, dCas9 binding, and differential gene expression. An additional bottleneck is the delivery of large dCas9 epigenome editors to the CNS. Current studies are addressing the ability to do large scale-ups of AAV for transgene delivery, showing we can produce clinical levels of AAV for decreased costs compared to current methodologies. This approach holds great promise for those suffering from HD.

Introduction

Areas covered

Expert opinion