Integrative genomic analysis of Neurogenin2 reprogramming of human iPSCs

Direct cell reprogramming is a rapidly growing field that has challenged traditional concepts of cellular identity. The expression of Neurogenin 2 (NGN2) results in rapid reprogramming of human pluripotent stem cells (PSCs) into functional excitatory neurons. My lab has previously demonstrated that gene targeting the components of a Tet-On system overexpressing NGN2 into two separate safe harbour sites overcome gene silencing and results in optimised transgene expression in hiPSCs (Opti-Ox), and consequently yields highly homogenous cultures of neurons within less than four days. The mechanisms that mediate this remarkable cellular metamorphosis however remain poorly understood. To explore this, I first sought to establish a protocol for long-term culture of electrophysiologically functional NGN2 induced neurons (iNs). This was achieved by co-culturing with primary rat-derived glial cells, enriched for astrocytes. iNs demonstrated functional activity around two weeks post-induction and by three weeks, formed networks of synchronous bursts that were mediated by glutamatergic AMPA-receptors. Based on these phenotypical hallmarks, I designed a time course-based genomic analysis that investigated the transcriptional and chromatin accessibility states of cells undergoing NGN2 reprogramming. Bulk RNA and ATAC-sequencing were performed at Day 0, 6h, 12h, Day 1, 36h, Day 2, Day 3, Day 4, Day 14 and Day 21 post-NGN2 induction. I also performed a scRNA-seq of the same time points, except for 6- and 36-hours post-induction, to investigate any heterogeneity in the time course and complement findings from the bulk RNA-seq. In order to differentiate direct from indirect NGN2 down-stream effectors, ChIP-seq of NGN2 binding acquired on day 1 after induction was subsequently overlaid with bulk-seq data. In addition, to study the genome-wide effects of astrocyte-enriched rat glia, I performed the same assays on neurons co-cultured with glia at Day 4, 14 and 21. Together, they revealed rapid transcriptional and accessibility changes induced by NGN2 within 6 hours of reprogramming. The subsequent events show a stark similarity to the familiar stages of neurogenesis found in development or conventional differentiation protocols - shutting down of non-neuronal networks, in this case pluripotency, establishment of neuronal commitment in an NSC-like stage by Day 1 post-induction, cell cycle exit by Day 3 or Day 4 and subsequent onset of neuronal differentiation, followed by neuronal maturation. Up until now, this entire process was believed to be a highly homogenous occurrence, but findings from the scRNAseq analysis showed that in addition to glutamatergic neurons, our NGN2 iNeurons are made up of two additional types of neurons: cholinergic neurons with a visceral motor phenotype and neurons with a hybrid profile of cholinergic and glutamatergic transcription. Comparison with neurons co-cultured with glia found an enrichment for synaptic genes and ontologies. Specifically, there was an enrichment for neuronal activity regulated genes in the post-synaptic compartment. Candidate transcription factors crucial to these processes and the states described above were identified from the transcriptional and epigenomic datasets. It is likely that some of these factors could potentially be crucial regulators of neuronal differentiation and function, not only in this protocol, but in development as well. Therefore, one major aim for the future would be to further explore the role these genes play in NGN2 reprogramming through gene knockdown or overexpression experiments along with investigation of their genomic binding sites. Ultimately, this could also lead to the discovery of new reprogramming strategies for producing neurons with enhanced maturity and functionality without the need of human or rodent glia. By uncovering broad, yet essential neuronal processes such as neuronal fate commitment and maturation, the dataset I have generated can be used as a useful resource for studying these processes and designing relevant disease models using this simple, yet robust and efficient protocol for generating functional excitatory human neurons.