Pathogenesis of Polyglutamine Disorders
The spinocerebellar ataxias (SCAs) are dominantly inherited neurodegenerative disorders that belong to a family of diseases caused by dynamic mutations. In the case of SCA1, the mutation is an expansion of a translated CAG repeat that tends to expand further upon transmission to each subsequent generation (hence its description as a "dynamic" mutation). Precisely how mutant Ataxin-1 exerts toxicity has been a question in my lab and in that of my collaborator, Harry Orr (University of Minnesota), ever since we codiscovered the genetic basis of SCA1 on the very same day in 1993. All that was clear initially was that the expanded protein gains some function that is toxic to neurons.
Because proteins with expanded glutamine tracts tend to aggregate in the nuclei of neurons in both patient tissues and tissues from animal models, many in the field initially believed the aggregates themselves were pathogenic. My lab was the first to propose that the protein accumulations might represent misfolded proteins that could not be cleared efficiently. We demonstrated that the cellular distribution of protein-folding machinery as well as proteolytic machinery is altered in SCA1 tissues and that chaperone overexpression reduces the aggregates of the SCA1 gene product, Ataxin-1 (other labs showed these findings to be relevant to other triplet repeat disorders). Our animal model studies demonstrated that it is the misfolded protein that induces pathology, not the aggregates per se. In fact, our work contributed to reconceiving neurodegenerative diseases as "proteinopathies," a class that now includes diseases such as Parkinson's and Alzheimer's disease.
In collaboration with the Orr lab, we learned that residue S776 in Ataxin-1 is necessary for mutant Ataxin-1 to exert toxic effects: Mice bearing expanded Ataxin-1 with an S776A mutation live a normal, disease-free life span. We also learned, in collaboration with Juan Botas (Baylor College of Medicine), that Drosophila and mice overexpressing even wild-type Ataxin-1 develop a mild form of SCA1 neurodegeneration. This finding led us to think more deeply about whether the polyglutamine protein gains an entirely new function that is toxic or whether its normal functions are somehow intensified by its altered conformation or its level of abundance.
To better understand the normal function of Ataxin-1, as well as discover whether the dozens of diseases that cause ataxia operate along the same pathways, we developed an ataxia interactome that helped establish that the gain of function is due to alterations in the strength of interactions between mutant Ataxin-1 an its native protein partners. For example, the glutamine expansion intensifies Ataxin-1 interactions with RBM17-containing complexes. We then found that Ataxin-1 forms stable protein complexes with Capicua, a transcriptional repressor, and that the functions of these Ataxin-1/Capicua complexes are altered by the glutamine expansion such that some genes are hyper-repressed (because of enhanced activity of the Ataxin-1/Capicua complex), whereas other genes are de-repressed (because of partial loss of the Ataxin-1/Capicua complex function). Thus, the polyglutamine expansion in Ataxin-1 enhances its functions and interactions in some contexts and decreases them in others. Studies from other labs have provided evidence that loss- and gain-of-function mechanisms also come together in Huntington's disease and spinobulbar muscular atrophy.
Importantly, the phenotypes of SCA1 mice improve upon either decreasing interactions that mediate the gain of function or decreasing Ataxin-1 levels. Forward genetic screens in human cells and Drosophila are allowing us to identify genes that, when inhibited, either decrease Ataxin-1 levels or decrease Ataxin-1 phosphorylation at serine 776, which should provide new therapeutic entry points for SCA1.
Atoh1/Math1 and Neurodevelopment
To understand the neuronal basis of balance and coordination, we sought and identified the mouse homolog of the Drosophila gene atonal, which controls the development and function of the fly's chordotonal organs (the sensory elements that provide proprioception and hearing sense). Our studies of mouse atonal homolog 1 (Atoh1 or Math1) have opened several fascinating and entirely unexpected lines of inquiry in the lab.
Mice lacking Atoh1 lack cerebellar granule neurons, pontine neurons, hair cells in the vestibular and auditory systems, the D1 interneurons of the spinocerebellar tracts, and Merkel cells. They also die within minutes after birth because they cannot initiate or maintain respiration. That a single gene controls the genesis and/or differentiation of multiple components of the conscious and unconscious proprioceptive pathway was surprising. Even more surprising was our discovery that Atoh1 is essential for secretory cells in the gut (Paneth, goblet, and enteroendocrine cells) and the neurons critical for interoception. Our work tracing the fate of Atoh1-dependent neurons has allowed us to redefine the rhombic lip and its derivatives and revealed Atoh1's role in the differentiation of retrotrapezoid neurons and their role in neonatal breathing and chemosensitivity (Figure 1). We identified Atoh1’s transcriptional targets and revealed its critical role in regulating proliferation and differentiation of granule neuron precursors and how this regulation might go awry in sonic hedgehog–induced medulloblastoma. Our studies of Atoh1’s role during development is shedding light on a variety of human ills, including deafness, vestibular disorders, cancer, and central respiratory disorders.
Rett Syndrome and MeCP2-related Disorders
Rett syndrome is a crippling, delayed-onset autism spectrum disorder. After discovering that Rett is caused by mutations in the X-linked gene encoding methyl-CpG-binding protein 2 (MECP2), we learned that MECP2 mutations actually cause a broad spectrum of phenotypes beyond typical Rett: Females can present with isolated intellectual disability, autism, or milder forms of Rett if they have favorable X-chromosome inactivation; males with inactivating MECP2 mutations suffer severe neonatal encephalopathy and death in infancy; males with milder hypomorphic mutations develop intellectual disability, motor dysfunction, and psychosis.
MeCP2 is expressed in mature neurons, and the number of MeCP2-positive cortical neurons increases postnatally as the brain matures. As with humans, mice with hypomorphic MeCP2 mutations or loss of MeCP2 appear healthy at first, then develop a progressive phenotype that affects social behavior, cognition, and motor control (and even forepaw stereotypies that parallel the hand-wringing that is such a curious feature of Rett syndrome). Interestingly, turning off MeCP2 in adult animals reproduces these phenotypes, and over the same time period as animals that lack the protein from birth. These findings indicate that MeCP2 performs a maintenance function in the brain, which explains the delayed onset. Interestingly, transgenic mice that overexpress MeCP2 in the correct spatiotemporal distribution also develop a progressive neurological disorder, after an initial period of health. The mouse phenotype led us back to the clinic to look for a human parallel, which we found in children with duplications or triplications spanning the MECP2 locus.
To determine the neuroanatomical origins of the Rett phenotypes, we are selectively deleting Mecp2 in different neuronal populations and studying the corresponding gene expression changes in mice. These studies led to the definition of further clinical phenotypes not previously appreciated in MeCP2 disorders and revealed that neurons require the right dose of MeCP2 to have the optimal level of neurotransmitters and neuropeptides.
Grants from the National Institutes of Health, the International Rett Syndrome Research Foundation, and the Rett Syndrome Research Trust provided partial support for these projects.
As of March 12, 2013