5 January 2007. Pathogenic proteins bearing expanded polyglutamine (polyQ) repeats are causative factors in nine distinct neurodegenerative diseases to date, the most common of which is Huntington’s. But just how they are toxic is poorly understood, and a burning question in the field remains whether the expansions interfere with a normal function of the protein, or confer a new, toxic function.
The dominant nature of the mutations has favored the gain-of-toxic-function scenario. But new results from Huda Zoghbi and colleagues at Baylor College of Medicine in Houston, Texas, show that one such protein, ataxin1, doesn’t have to hook up with new partners to damage neurons. Instead, they propose that polyQ-expanded ataxin1 acts through a normal physiological protein partner, the transcriptional repressor protein Capicua (CIC). Published in the December 29 issue of Cell, the work at once identifies a physiological role for ataxin1 in transcriptional repression, and indicates a mechanism for the toxicity of this particular polyQ-expanded protein. It also advances the field’s understanding of how similar genetic defects can cause wildly divergent phenotypes in different diseases.
PolyQ expansion of ataxin1 causes spinocerebellar ataxia1, in which Purkinje cells in the cerebellum degenerate. To better understand the normal physiological functions of ataxin1, Zoghbi’s lab collaborated with Harry Orr at the University of Minnesota in Minneapolis to isolate and analyze native soluble ataxin-containing complexes from the cerebellum of normal mice. Most of the ataxin1, first authors Yung Lam and Aaron Bowman found, was in large, stable complexes. The complexes also contained two isoforms of the human homolog of the Drosophila protein CIC, a protein that came up as an ataxin1 partner in a previous interactome study (see ARF related news story).
The investigators went on to show that ataxin1 and CIC directly associate in vitro and are coexpressed in brain; both are particularly high in the nuclei of Purkinje cells where ataxin mutations wreak their damage. At the protein level, normal ataxin1 appeared to stabilize CIC via a direct protein-protein interaction between the AHX domain of ataxin1 and a conserved N-terminal sequence in Capicua.
PolyQ expansion of ataxin1 did not affect its interaction with CIC in vitro, or in cerebellar extracts. But a key experiment linking CIC to toxicity demonstrated that a nontoxic mutant of poly Q-expanded ataxin1 was absent from the CIC-containing complexes. Previously, Zoghbi and colleagues had shown that an S227A mutant of ataxin1 with 82 repeats was less toxic because it was not phosphorylated and did not interact with 14-3-3 proteins (see ARF related news story). The absence of the S227A mutant protein from CIC-containing complexes suggests that the complex formation is necessary for toxicity.
To look at CIC function, the investigators turned to Drosophila genetic models. They found that CIC was required for 82Q-ataxin1 to be toxic in a Drosophila model of neurodegeneration. On the flip side, 82Q-ataxin1 expression interfered with the normal repression function of CIC during wing development, which in turn allowed aberrant overexpression of target genes. Similar effects appeared in mammalian cells, where overexpression of normal ataxin1 increased CIC repressor activity on a reporter construct. The 82Q-expanded protein also supported repression, but less so than did the wild-type protein. These results indicate that the polyQ expansion interferes with ataxin’s ability to regulate Capicua function in a normal way.
“We conclude that ataxin1 acts in a transcriptional repressor complex and that polyglutamine-expanded ataxin1 exerts neurotoxicity through its native complexes containing CIC rather than through aberrant interactions with novel proteins,” the authors write.
This insight could explain how different polyQ-expanded proteins cause distinct neurological diseases. As Nan Liu and Nancy Bonini of the University of Pennsylvania in Philadelphia write in an accompanying preview, “The implication is that each disease has individual molecular pathologies depending on each disease protein’s interaction network, sharing only the common initiating event of a polyQ expansion.”
If so, the results could steer researchers in new directions in the hunt for targets and treatments for individual polyQ expansion diseases. One example is the recent finding that heat shock protein 90 (HSP90) inhibitors, developed as anti-cancer agents, can prevent neurodegeneration in a mouse model of spinal and bulbar muscular atrophy (SBMA) caused by a polyQ-expanded androgen receptor (Waza et al., 2005; Waza et al., 2006). HSP90 complexes with the androgen receptor (AR), and depending on what other proteins are around as well, it can either stabilize the receptor or speed its degradation. The HSP90 inhibitor geldanamycin blocks the stabilizing effects while enhancing degradation of HSP90-bound proteins. The researchers, in Gen Sobue’s lab at Nagoya University in Japan, found that both normal AR and polyQ-expanded AR are subject to proteasome-mediated degradation after HSP90 inhibition. In fact, the polyQ AR appears more susceptible to the inhibitor than wild-type protein. Treating SBMA mice with a nontoxic analog of geldanamycin (17-AAG) partly blocked motor neuron degeneration and increased the mice’s mobility and survival, while reducing 97Q-AR protein levels. 17-AAG is currently in phase 1 and 2 clinical trials against a number of cancers, sponsored by the National Cancer Institute. Sobue has been proposing that the compound might be useful in any disease involving proteins that depend on HSP90 for stability. This category currently encompasses over 100 proteins, including many kinases that regulate neurotoxic pathways, such as tau phosphorylation.
The alteration of normal pathways might occur with other polyQ proteins, and also in diseases including Alzheimer disease and Parkinson disease, where overexpression of normal amyloid precursor or α-synuclein is sufficient to cause neurodegeneration, Zoghbi and coauthors speculate. That makes the study of the normal function of neurotoxic proteins all the more important for understanding disease.—Pat McCaffrey.
Lam YC, Bowman AB, Jafar-Nejad P, Lim J, Richman R, Fryer JD, Hyun ED, Duvick LA, Orr HT, Botas J, Zoghbi HY. ATAXIN-1 Interacts with the Repressor Capicua in Its Native Complex to Cause SCA1 Neuropathology. Cell. 2006 Dec 29;127(7):1335-47. Abstract
Liu N, Bonini NM. Hosting neurotoxicity in polyglutamine disease. Cell. 2006 Dec 29;127(7):1299-300. Abstract