Epigenetics sits at the interface between the genome and the environment. The epigenome is composed of chromatin, its associated proteins and archetypal pattern of covalent modification of DNA by methylation. The epigenome is critical for setting up and maintaining gene expression programs. A hallmark of the mammalian genomes is the dynamic changes in DNA methylation during development, differentiation and in response to actively changing environment. It was originally believed that DNA methylation by virtue of possessing strong covalent bond is a very stable epigenetic modification. However, with accumulating evidence it is becoming increasingly clear that DNA methylation is a bidirectional process, and can be reversed by an active (i.e. enzymatic) demethylation process. DNA demethylation is a widespread phenomenon and occurs in plants and animals, during development and differentiation, in the adulthood, and during somatic cell reprogramming of pluripotency genes. These dynamic variations in DNA methylation are also considered to be integral and one of the driving factors behind the plasticity of nervous system.
This thesis studies the activity-dependent differential induction of both coding and non-coding transcripts and also the role for activation-induced cytidine deaminase (Aid) in mediating activity-dependent DNA demethylation in post-mitotic neurons both in culture and in vivo. In Chapter 2, I present data demonstrating dynamic and significant differential expression of several novel transcripts (both coding and non-coding) following neuronal activation. To achieve this we used in vitro primary cortical neuron culture and KCl-induced depolarization to mimic neuronal activity by KCl stimulation. The long non-coding RNA Gomafu identified through this screen was found to be involved in schizophrenia-associated alternative splicing. Thus, in chapter 2 we identified several novel transcripts associated with neuronal activation and validated their expression pattern by time-course analysis. These data also indicated critical time-points for peak gene expression and provided the foundation for downstream expression and epigenetic analyses.
Following this, I used the above-mentioned time course model system to explore the underlying the role of Aid in mediating active DNA demethylation. In chapter 3, I used Bdnf exon I and IV as candidate genes to investigate the relationship between Aid, active DNA demethylation and activity-dependent Bdnf mRNA expression. We found that lentiviral-mediated knockdown of Aid results in blocking active DNA demethylation of Bdnf P4 promoter and also leads to decrease in Bdnf exon IV mRNA transcripts in primary cortical neuron culture. To our knowledge, this is the first evidence indicating a role for Aid in mediating activity-dependent DNA demethylation of the Bdnf P4 promoter in post-mitotic neurons.
Further, based on the role of Aid in regulating active DNA demethylation at the Bdnf P4 promoter and importance of Bdnf exon IV mRNA in mPFC during learning we investigated whether selective lentiviral-mediated knockdown of Aid in mPFC leads to any behavioural outcomes. We found out that by knocking down Aid in mPFC after fear-conditioning training, mice exhibited stronger memory for fear-extinction. Therefore, we managed to demonstrate requirement for Aid, both in vitro and in vivo, in regulating gene expression and in brain function.
Together, the results indicate that neuronal activation engages both coding and non-coding transcripts, and that Aid is an important epigenetic modifier, which acts in an activity-dependent manner in post-mitotic neurons and is critical for active DNA demethylation at the Bdnf P4 promoter.