A central goal in neuroscience is determining how the functioning of individual neurons goes awry in neurological disorders such as autism and brain tumors. We have demonstrated that REST, a transcription factor first identified in our lab, is a master regulator of the neuronal phenotype. We showed recently that REST controls more than 8,000 genes critically important to neuronal function and often dysregulated in diseases of the nervous system. Unlike most transcriptional regulators, which can both repress and activate gene expression, REST seems only to be a repressor. Indeed, the decision to become a neuron rather than some other cell type is critically dependent on the repressor function of REST. For example, prior to formation of the nervous system, REST is expressed everywhere throughout the embryo. As the nervous system develops, REST is turned off selectively in neurons, releasing neuronal genes from its repression. We have exploited this feature to understand more about the process of differentiation and, more recently, about Rett syndrome. This neurological disorder is caused by the methyl CpG–binding protein 2 (MeCP2), which collaborates with REST to repress neuronal genes.
REST as a Model for Genomic Reprogramming
Stem cell research holds promise for curing diverse human diseases, particularly if embryonic stem cells can be "back generated" from a patient's own terminally differentiated skin or blood cells. One problem with this approach, however, is that the generation of totipotent embryonic stem cells, which can generate all tissue types, from unipotent fibroblasts or from lymphocytes, is still extremely inefficient. Scientists interested in this problem do not yet have a good enough understanding of the molecular mechanisms involved in the dedifferentiation process to improve the efficiency. We discovered that REST is expressed to unusually high levels in mouse embryonic stem cells. Furthermore, it represses neuronal genes by a completely different mechanism in stem cells and the terminally differentiated fibroblasts and lymphocytes used for reprogramming. Thus, study of REST regulation in these two cellular contexts is an excellent way to explore how a repressor complex can be rebuilt with fidelity during reprogramming.
Using a new methodology for biochemically purifying REST complexes, we have elucidated the complete complex of REST-interacting proteins in embryonic stem cells that establishes a "quiet" state for neuronal genes. These cells are poised for subsequent expression in neurons. We are now investigating the biochemical bases for a permanent "off" state for the same genes by performing a similar analysis in fibroblast cells. In this manner we will flush out the proteins that need to be removed or altered to reprogram a cell back to the embryonic state. The identification of the roadblocks that occur at the level of the genome will aid the efficient reprogramming of differentiated human cells back to the plastic stem cell state.
REST Regulates RNAs as well as Protein-Coding Genes
Small regulatory mRNA molecules, called microRNAs (miRNAs), are present across the phyla and bind to target mRNAs in the cell that produce proteins. As a result of this interaction, translation of the proteins specified by the mRNAs can be blocked, degraded, or amplified. This activity plays important roles in development, in the conversion of normal cells to metastatic cancer cells, and in the differentiation of stem cells, but how it relates to the activities of transcription factors, which also regulate gene expression, is not clear. We identified a family of miRNAs that are repressed by REST in fibroblasts. In neurons, where REST was absent, the miRNAs are expressed. In collaboration with Gregory Hannon (HHMI, Cold Spring Harbor Laboratory), we developed a novel strategy to identify the complete set of mRNA targets associated with the brain-specific miRNA 124a. Others had shown that miRNA 124a can degrade mRNAs, but the biological context for this activity remained unknown. We showed that REST prevents miRNA 124a expression in fibroblasts. Conversely, the absence of REST results in high-level expression of miRNA124a. Furthermore, miRNA124a both destroys and blocks nonneuronal mRNA functions that would be deleterious in the neurons. Thus, REST participates in the regulation of both the neuronal and nonneuronal phenotypes: in neurons, REST works through miRNA; in nonneuronal cells, it acts directly. This REST-miRNA circuit (see the figure) reinforces and amplifies the neuronal phenotype in a terminally differentiated neuron and perhaps helps block interconversion of neurons to other types of cells. (Grants from the National Institutes of Health provided support for aspects of this work.)
Is Rett Syndrome a Complex Disorder of Both Neurons and Supporting Cells?
Rett syndrome (RTT), a neurological disorder related to autism, affects approximately 1 in 10,000–15,000 girls. Afflicted individuals exhibit an outwardly normal developmental period of 12–18 months, after which time they begin to regress, losing verbal communication, purposeful hand motions, and normal walking gait. Their brains are smaller, they acquire severe respiratory problems, and they often suffer from seizures. In 1999, Huda Zoghbi (HHMI, Baylor College of Medicine) identified the culprit gene as MECP2. Mouse models for RTT have revealed that loss of MeCP2 from neurons contributes significantly to the disease in mice, and there are similarities in the phenotypes of afflicted mice and humans. Despite these findings, no substantive changes in gene expression, as a result of the absence of MeCP2, have been shown to account for RTT. We found recently that loss of MeCP2 in glia, a nonneuronal cell type, is a significant contributory factor to RTT. Our work is now focused on determining the mechanisms through which glia and MeCP2 function influences neuronal function, in particular the avenues of communication between glia and nerve. (Grants from the National Institutes of Health provided support for aspects of this work.)
