Generating morphologically and functionally diverse neurons requires a precisely orchestrated expression of more genes in the developing nervous system than in any other organ. As all neurons arise from a common pool of progenitors, a cascade of cell state-specific gene repression and activation events must take place to restrict non-essential genes from being expressed and to promote genes mediating pro-neuronal specification. Epigenetic marks provide fundamental and heritable modifications to control the ‘DNA readout’, thus presenting a persistent form of cellular memory whereby terminally differentiated neurons remember their phenotypes (Sweatt, 2009). Whilst DNA methylation is the predominant epigenetic mark regulating neural stem cell proliferation and lineage commitment, neuronal differentiation is frequently associated with histone modifications. Histones, the highly basic proteins organizing DNA within the nucleus, can undergo post-translational modification by acetylation, methylation, ubiquitination or phosphorylation. Histone acetylation at lysine residues is reversible and is determined by the net balance of histone acetyltransferase (HAT) and histone deacetylase (HDAC) activities. Fundamental neurodevelopmental signaling cascades (CREB, NF-κB) enable HATs to regulate local chromatin conformation (Mannervik et al., 1999). As a result, transcription factors interacting with lineage-specific gene promoters can activate gene transcription programs that determine cell fate (Lee et al., 2009).
In contrast, HDACs operate as components of the transcriptional repressor complex to silence gene expression (Renthal & Nestler, 2009). All four HDAC subclasses are tightly regulated through post-translational modifications that control their activity, subcellular localization and stability. Class II HDACs, including HDAC9, respond to various environmental stimuli: when phosphorylated by CaMKII during neuronal activity, HDACs recruit 14-3-3 chaperones, which trigger their nuclear export thus reducing their activity in the nucleus (McKinsey et al., 2000). An exciting concept in contemporary neurobiology implicates epigenetic mechanisms, including HDACs, in switching off neuronal plasticity during the closure of the experience-dependent critical period of cortical development (Putignano et al., 2007). It is, therefore, particularly timely that Noriyuki Sugo and colleagues propose a novel link between HDAC9 function and neuronal morphogenesis (Sugo et al., 2010, this issue): HDAC9 mRNA expression in layers 2/3 and 5 of the cerebral cortex coincides with the experience-dependent critical period of early postnatal development. HDAC9 accumulates in the nucleus of immature neurons. In turn, HDAC9 undergoes neuronal activity-driven, 14-3-3-dependent translocation to the cytoplasm in mature, excitable neurons, particularly prospective layer 2/3 pyramidal cells, during formation of their elaborate dendritic tree. This observation links HDAC9 to cytoskeletal rearrangements in developing neurons, providing an important parallel between HDAC9 and HDAC6; the latter being known to deacetylate tubulin, a cytoskeletal protein, and to shuttle misfolded proteins to the proteasome (Kawaguchi et al., 2003). Identification of this activity-dependent HDAC9 nuclear export mechanism, and generation of a nucleo-cytoplasmic translocation-deficient HDAC9 mutant incapable of recruiting its 14-3-3 chaperone, allowed Sugo et al. (2010) to identify c-fos as a candidate gene repressed by HDAC9 and impacting on neuronal differentiation. Although this finding suggests parallels between HDAC9-mediated repression of gene transcription and activity-dependent dendritogenesis, genome-wide analysis of HDAC9 targets will be required to formulate a causal relationship between HDAC9 function(s) and the acquiring of phenotypic neuronal morphologies.
The observation that disrupting the temporal dynamics of HDAC9 nuclear export coincides with the pruning of dendritic trees in pyramidal cells raises further questions: could the lack of cytoplasmic HDAC9 function, rather than HDAC9-induced repression of gene transcription, preclude the development of new dendrites? HDACs are in fact ‘lysine deacetylases’ capable of deacetylating many proteins (transcription factors, cytoskeletal proteins, enzymes) besides histones. Therefore, a mind-provoking experiment would be to test whether HDAC9 has a cytoplasmic target, like tubulin for HDAC6, that controls dendrite growth. Some of the data presented by Sugo et al. (2010) support this possibility, as overexpression of nuclear export-competent HDAC9-EGFP facilitates early acquisition of pyramidal cell-like dendrite complexity. Thus, their work advances our current understanding of the epigenetic control of neuronal morphology and suggests significant interactions between HDACs and the neuronal cytoskeleton.