Activity-dependent gene regulation in astroglia-derived induced neurons

Abstract

The limited regenerative capacity of the adult mammalian brain following neuronal injury has encouraged researchers to explore neuronal replacement strategies to repair neural circuits and to recover compromised behavioural functions. One strategy uses retroviruses that target proliferating glia cells to induce ectopic overexpression of neuronal transcription factors (TFs), leading to glia-to-neuron reprogramming. Although these induced neurons (iNs) have been shown to acquire a neuronal morphology and neuron-specific functional features, it is not known to which extent their molecular phenotype resembles that of endogenous neurons (eNs) and whether they display homeostatic synaptic plasticity, a feature that would allow them to functionally integrate into a pre-existing neuronal network. These two aims were addressed in this work by overexpressing the TF Neurog2 in postnatal cortical glia and coculturing these with cortical eNs that have already established a network in vitro to allow integration of newly generated iNs. After a period of two weeks, network activity was either pharmacologically inhibited or left unperturbed and single-nucleus (sn) RNA-sequencing was performed. Interestingly, the subsequent molecular characterization of the iN population pointed to two distinct sources of diversity among the iN population: molecular subtype identities, comprising a broad range of inferred cortical subtypes and distinct developmental stages. To assess if iNs undergo HSP, I investigated the transcriptional dynamics in response to activity inhibition and how these compare to the ones elicited in eNs. While eNs displayed regulation of gene signatures indicative of HSP, iNs displayed downregulation of dendritic and postsynaptic genes that correlated with a reduced morphological complexity following activity inhibition. Interestingly, a set of synapserelated genes that was upregulated by eNs in response to activity inhibition was found to be highly expressed in iNs under control condition. Furthermore, pharmacological network activity disinhibition resulted in minimal c-Fos upregulation in iNs compared to eNs, pointing to limited integration into the network and suggesting that the high basal expression of synaptic genes in iNs may reflect ongoing competition for synaptic input. Immunocytochemical stainings showed that GABAergic, but not glutamatergic synapses decorate iNs. Taken together, these data suggest that iNs are not yet functionally integrated into the network, despite wide expression of synaptic machinery-related genes. Lack of synaptic input may be constraining iN maturation, while their high basal expression of synapse-related genes may indicate an increased competition for synaptic input. In sum, this work provides the first snRNA-seq study comparing the molecular phenotype and activity-dependent transcriptome of glia-derived iNs and cortical eNs, and suggests avenues for refining the iN differentiation process towards a functionally more mature and responsive phenotype.