Activity-Regulated Transcription: Bridging the Gap between Neural Activity and Behavior

Abstract

Gene transcription is the process by which the genetic codes of organisms are read and interpreted as a set of instructions for cells to divide, differentiate, migrate, and mature. As cells function in their respective niches, transcription further allows mature cells to interact dynamically with their external environment while reliably retaining fundamental information about past experiences. In this Review, we provide an overview of the field of activity-dependent transcription in the vertebrate brain and highlight contemporary work that ranges from studies of activity-dependent chromatin modifications to plasticity mechanisms underlying adaptive behaviors. We identify key gaps in knowledge and propose integrated approaches toward a deeper understanding of how activity-dependent transcription promotes the refinement and plasticity of neural circuits for cognitive function.

Figure 1.. Schematic of signaling mechanisms driving activity-dependent transcription of immediate early genes and late-response genes

Neurotransmitter signaling leads to the generation of action potentials in the neuron. Membrane depolarization induces the opening of L-type voltage-sensitive calcium channels (L-VSCCs). Stimulus-dependent calcium entry via L-VSCCs preferentially leads to the activation of the Ras-MAPK pathway, calcium/calmodulin-dependent protein kinases, and calcineurin-dependent signaling. Note that calcium influx can also occur via activation of NMDA receptors. Decades of work from numerous laboratories have elucidated the molecular mechanisms of these calcium-dependent signaling cascades, which have been simplified in this schematic. Cell type-specific differences in signaling mechanisms that are not illustrated here have also been described (e.g., see Cohen et al., 2016). These pathways lead to activation of preexisting transcription factors CREB, SRF/ELK, and MEF2, which regulate the expression of immediate early genes (IEGs) such as Fos. Many IEGs encode transcription factors that regulate a subsequent wave of late-response genes, which have now been shown to be cell type-specific.

Figure 2.. Cell-type specificity of the activity-dependent gene programs

(A) Single-cell RNA-sequencing technologies have enabled unprecedented characterization of the diversity and cell-type specificity of experience-dependent transcriptomes in the brain.(B) Divergent activity-dependent transcriptional responses in neuronal and non-neuronal cell types depicted by heat map of 611 stimulus-regulated genes (horizontal black lines) grouped into early-response and late-response genes by cell type. Exc, excitatory neurons; Int, interneurons; Olig, oligodendrocytes; endo/SM, endothelium, smooth muscle; Micro, microglia. This figure is adapted with permission from Hrvatin et al. (2018).

Figure 3.. Model of developmental specification of cell type-specific activity-dependent gene programs by AP-1 and cell type-specific pioneer factors

(Top) Current model of stimulus-dependent enhancer selection posits that during differentiation of a given cell type, AP-1 sites across the genome are occluded by nucleosomes and thus inaccessible. Adjacent to each AP-1 site it is often possible to identify additional sequence motifs that are more cell type-specific (CTSE, cell-type specific element) and predicted to bind cell type-specific so-called pioneer factors (CTSF). Once Fos/Jun complexes are expressed in response to extracellular stimuli, they most likely cooperate with CTSFs to recruit the chromatin remodeling BAF complex to cell type-specific enhancer elements. This leads to chromatin remodeling, enhancer selection, and thus activation of late-response gene transcription in a cell type-specific manner. (Bottom) Once these cell type-specific enhancers have been specified during development, even after Fos and Jun decay away, the enhancers are thought to remain primed, via specific post-translational histone modifications. These enhancers are then ready for activation in the mature brain the next time Fos/Jun complexes are induced in response to neuronal activity.

Figure 4.. Model of MeCP2-dependent regulation of gene repression

(A) Simplified schematic showing that in the absence of neuronal activity, MeCP2, a reader of mCA, acts as a repressor of gene transcription in part through the recruitment of repressive co-factors (e.g., NCoR complex). In response to neuronal activity, MeCP2 is rapidly phosphorylated at multiple sites. This modification of MeCP2 potentially leads to the regulated release of NCoR, thus relieving gene repression.(B) MeCP2 loss-of-function models suggest that MeCP2 represses highly methylated genes that contain a large number of mCA sites within these genes. DNA methyltransferases (DNMT) deposit methylation across the genome.

Figure 5.. Cell type-specific activity-dependent gene programs are tailored to specific function of cell within a neural circuit

Each cell type in the brain possesses a distinct set of activity-regulated genes (top panel) that allow each cell type to interact with and modify specific synaptic inputs within their resident neural circuit (middle panel, right). For example, IGF1 is secreted from VIP-expressing interneurons, and recruits inhibitory inputs onto VIP-expressing interneurons themselves, while BDNF is secreted from excitatory neurons and recruits inhibitory inputs onto the cell bodies of excitatory neurons. Activity-regulated genes have been shown to regulate cellular processes such as dendritic growth and restriction, and synapse formation, maturation, elimination, and strength (middle panel, left). For example, Arc and Homer1a have been shown to function as negative regulators of AMPA receptor expression at synapses (bottom panel).

Figure 6.. Mutations in specific components of the activity-dependent transcriptional pathway have been implicated in various neurodevelopmental and neurological disorders in humans

Simplified schematic depicting mutations in L-VSCCs have been implicated in Timothy syndrome, Rsk2 in Coffin-Lowry syndrome, and CREB-binding protein CBP in Rubenstein-Taybi syndrome. Mutations in multiple subunits of the BAF complex have been implicated in various neurological disorders including sporadic autism and intellectual disability (see Ronan et al., 2013). Mutations in MeCP2 lead to Rett syndrome. Bdnf is an example of a late-response gene whose defects in expression or function have been associated with impaired episodic memory and depression, among others.