Membres du jury :
Carole Kretz-Rémy, présidente
Sergio Gascón, rapporteur
Sophie Jarriault, rapporteure
Christophe Heinrich, directeur de thèse
Le vendredi 15 décembre à 09h00, en salle de conférence du SBRI
Our adult mammalian brain lacks intrinsic regenerative capacity to replace lost neurons following injury/disease and induce functional recovery. The main goal of regenerative medicine is to replace lost neurons in order to restore lost functions and correct neurological deficits. Lineage reprogramming of cell identity is an emerging concept for the remodeling and restoration of organs with limited regenerative capacity. In the context of neurological disease, this concept opens the possibility of regenerating neurons from other brain-resident cell types such as glial cells for the repair of diseased brain circuits. Over the past years, we and others have shown that various types of glial cells can be converted into induced neurons (iNs) in vitro and in vivo by forced expression of neurogenic transcription factors. While lineage reprogramming holds promise as a neuron-replacement strategy, our understanding of this process remains very limited. Little is known about how reprogramming transcription factors impose on differentiated glial cells new transcriptional networks driving a new molecular program reassigning a neuronal identity. Whether glia-to-neuron reprogramming relies on transition through intermediate states, including reversal to developmentally immature states, also remains to be determined.
In the present study, we show efficient conversion of postnatal cortical astrocytes into GABAergic iNs in vitro by retrovirus-mediated expression of Ascl1 and Dlx2, or Dlx2 alone. Using single cell RNA sequencing, we revealed the transcriptional alterations underlying reprogramming of astrocytes into GABAergic iNs. RNA velocity and pseudotime ordering allowed us to identify two alternative reprogramming trajectories that comprised several intermediate cell states, including one state exhibiting transient expression of neural stem cell-related genes preceding neuronal differentiation and specification into GABAergic iNs. Moreover, we observed extensive metabolic changes that were dynamically regulated during reprogramming, including the decrease of glycolytic gene expression. In addition, we found that the expression of genes linked with epigenetics also underwent important remodeling, with emerging expression of transcripts linked with epigenetic regulation of neurogenesis along the reprogramming trajectory. We also observed that combining Ascl1 and Dlx2 co-expression appeared to convert astrocytes into more mature GABAergic iNs showing distinct interneuron subtype specification compared to Dlx2 alone-induced neurons. Finally, we identified a key enzyme, which plays a crucial role during the early stages of astrocyte-to-iN reprogramming. Indeed, the number of iNs and their maturation were drastically reduced upon inhibition of this enzyme. In fact, cells undergoing reprogramming were not capable of negotiating their neuronal reprogramming upon inhibition of this enzyme and essentially succumbed to cell death.
Together, our findings shed new light on the molecular underpinnings of astrocyte reprogramming into GABAergic iNs, and contribute to answer fundamental questions that need to be addressed before potential translation of lineage reprogramming into a neuron-replacement therapy.