Having a few excess polyglutamine codons—but not a lot—in the gene ataxin-2 predisposes a person to amyotrophic lateral sclerosis, and scientists are just starting to figure out how. In the July 4 Journal of Neuroscience, Michael Hart of the University of Pennsylvania in Philadelphia and Aaron Gitler, now at Stanford University in Palo Alto, California, report that those intermediate-length polyQ repeats ramp up caspase activity and make a toxic turncoat out of the ALS protein TDP-43, at least in cell culture. Accumulation of activated caspase-3 appears to be a hallmark of the disease when ataxin repeats contribute to the pathology, the scientists found in autopsy tissue samples.

Researchers in Gitler’s laboratory previously discovered that carrying between 27 and 33 glutamine codons in the ataxin-2 gene is a risk factor for ALS (see ARF related news story on Elden et al., 2010). Most people have 22 or 23 repeats. Curiously, having more than 34 repeats causes a higher likelihood not of ALS, but of another disease, spinocerebellar ataxia.

Ataxin-2 protein associates with TDP-43 in stress granules. These are transient protein accumulations that cells create in times of trouble. Stress granules have been repeatedly linked to ALS (see ARF Webinar; Liu-Yesucevitz et al., 2010). Pathogenic TDP-43 in cytoplasmic stress granules tends to be both fragmented and phosphorylated. Hart and Gitler had a hunch that the extra repeats might enhance the ataxin-2-TDP-43 interaction, effectively imprisoning TDP-43 in stress granules and promoting disease.

Hart tested the hypothesis in three cell types: the HEK293T human embryonic kidney line, lymphoblasts from people with ALS, and differentiated BE(2)-M17 human neuroblastoma cells. The lymphoblasts included samples from people with ALS who had either normal or intermediate-length expansions, as well as one person with spinocerebellar ataxia due to 40 repeats. In the case of the cell lines, Hart transfected them with ataxin-2 toting 22, 31, or 39 repeats.

Simply adding ataxin-2, of any repeat size, did not affect TDP-43. To stress the cells and to force TDP-43 out of its normal nuclear location into cytoplasmic stress granules, Hart heated the cultures to 42 degrees Celsius for one hour. Then he immunoblotted with an antibody specific for phosphorylated TDP-43. Heat shock drove up the amount of phosphorylated, insoluble protein, which was present as a fragment of approximately 30 kilodaltons. The intermediate-length repeats approximately doubled the concentration of this pathogenic TDP-43, compared to ataxin-2 with longer or shorter repeat sequences.

Since only fragmented TDP-43 was phosphorylated, Hart suspected that the cleavage would precede, and might even promote, the phosphate acquisition. Other researchers have shown that caspase-3 cleaves TDP-43 (see ARF related news story on Zhang et al., 2007), and indeed caspase activation is a normal function of short-repeat ataxin-2 (Wiedemeyer et al., 2003). This led Hart to examine his immunoblots with antibodies specific to activated caspases.

The cells with intermediate-length ataxin repeats turned on caspase-3; cells expressing longer or shorter versions did not. It appears that the medium-length repeats somehow lower the cell’s threshold for stress, making it prone to activate caspase and place TDP-43 in stress granules. Treating the cells with a caspase inhibitor diminished accumulation of phosphorylated TDP-43 fragments—suggesting that caspase activation occurs upstream of the TDP-43 aggregation.

Finally, Hart examined spinal cord sections from eight people who died of ALS. Specifically in the samples from four people with intermediate ataxin expansions, he observed cytoplasmic inclusions of activated caspase-3 in the motor neurons. “We think this is a new feature of ALS pathology that is specific to cases of ALS that harbor ataxin-2 expansions,” Gitler concluded.

The paper supports the current hypothesis that accumulated TDP-43 fragments cause disease, and offers a pathological pathway specific to people with the ataxin-2 risk factor, said Randal Tibbetts of the University of Wisconsin, Madison, who was not involved in the paper. In people without the expansions, other genetic or environmental factors, such as TDP-43 mutations, could lead to the same “common endpoint” of TDP-43 pathology, he suggested.

Subduing the caspase might treat ALS, Gitler suggested. Unfortunately, caspase inhibitors have fared poorly in ALS trials (Gordon et al., 2007). Perhaps the therapy would be effective for the subpopulation of people with ataxin expansions, Gitler speculated. Alternatively, Tibbetts suggested, scientists might try to alter the pathway between ataxin-2 repeats and caspase activation, but researchers still have to work out that chain of events.

How can just a few excess amino acids make such a striking difference in ataxin’s activity? The simplest explanation is that they alter the protein’s conformation, suggested Leonard Petrucelli of the Mayo Clinic in Jacksonville, Florida, who was not involved in the study. The expansion might cause a form of toxic gain of function, allowing ataxin to interact differently with normal partners or even recruit new binding partners. For his part, Gitler suspects that the intermediate repeats enhance some normal function of ataxin, such as its caspase-activating ability, while the long repeats that cause spinocerebellar ataxia create a new, poisonous property.—Amber Dance


  1. I think the overall idea is quite interesting. In a nutshell, it is that TDP-43 normally shuttles between the nucleus and the cytoplasm. Under stress, it moves temporarily to the cytoplasm to form stress granules to help cells cope with the stress. Versions of ataxin-2 with polyQ expansions that are longer than normal but not long enough to cause spinocerebellar ataxia enable it to cooperate with TDP-43 and cause neurodegeneration. Both proteins are involved in stress granule formation, which would be the logical place for an interaction that might cause synergistic toxicity. But there also seems to be a role for activation of caspase-3, which can cleave TDP-43, and presumably promote cell death. The biology is interesting in part because only intermediate polyQ lengths seem to confer these properties—normal polyQ stretches or ones that cause SCA2 don't work.

    Whereas the amount of ALS explained by this biology is probably small, it does give us additional insights into how TDP-43 might work and lead to neurodegeneration. Since TDP-43 pathology is found in a number of neurodegenerative diseases, those insights have broader implications. We had shown a few years ago that ALS-causing mutations in TDP-43 increase its levels in the cytoplasm. The extent to which TDP-43 is in the cytoplasm, whether a disease-causing mutation is present or not, seems to predict whether and when neurons will die. The results of the Gitler paper are consistent with that and suggest a model in which ataxin-2 with intermediate polyQ lengths might lead to an interaction with TPD-43 in the cytoplasm that has the net effect of stabilizing it there. The work provides important insights into the mechanisms by which a genetic modifier might mediate its activities.

    It's too early to tell whether there is a therapeutic angle here. We need to understand the biology better, and the one thing we can say for sure is that the levels and localization of TDP-43 appear to be meticulously and intricately regulated by cells.

    View all comments by Steven Finkbeiner

Make a Comment

To make a comment you must login or register.


News Citations

  1. ALS—A Polyglutamine Disease? Mid-length Repeats Boost Risk
  2. Progranulin Controls Cutting of Inclusion Protein

Webinar Citations

  1. Stress Granules and Neurodegenerative Disease—What’s the Scoop?

Paper Citations

  1. . Ataxin-2 intermediate-length polyglutamine expansions are associated with increased risk for ALS. Nature. 2010 Aug 26;466(7310):1069-75. PubMed.
  2. . Tar DNA binding protein-43 (TDP-43) associates with stress granules: analysis of cultured cells and pathological brain tissue. PLoS One. 2010;5(10):e13250. PubMed.
  3. . Progranulin mediates caspase-dependent cleavage of TAR DNA binding protein-43. J Neurosci. 2007 Sep 26;27(39):10530-4. PubMed.
  4. . Ataxin-2 promotes apoptosis of human neuroblastoma cells. Oncogene. 2003 Jan 23;22(3):401-11. PubMed.
  5. . Efficacy of minocycline in patients with amyotrophic lateral sclerosis: a phase III randomised trial. Lancet Neurol. 2007 Dec;6(12):1045-53. PubMed.

Further Reading


  1. . Distinct TDP-43 pathology in ALS patients with ataxin 2 intermediate-length polyQ expansions. Acta Neuropathol. 2012 Aug;124(2):221-30. PubMed.
  2. . Ataxin-2 polyQ expansions in FTLD-ALS spectrum disorders in Flanders-Belgian cohorts. Neurobiol Aging. 2012 May;33(5):1004.e17-20. PubMed.
  3. . The modulation of Amyotrophic Lateral Sclerosis risk by ataxin-2 intermediate polyglutamine expansions is a specific effect. Neurobiol Dis. 2012 Jan;45(1):356-61. PubMed.
  4. . Evaluating the prevalence of polyglutamine repeat expansions in amyotrophic lateral sclerosis. Neurology. 2011 Jun 14;76(24):2062-5. PubMed.
  5. . Ataxin-2 repeat-length variation and neurodegeneration. Hum Mol Genet. 2011 Aug 15;20(16):3207-12. PubMed.
  6. . Neurodegeneration: An expansion in ALS genetics. Nature. 2010 Aug 26;466(7310):1052-3. PubMed.

Primary Papers

  1. . ALS-associated ataxin 2 polyQ expansions enhance stress-induced caspase 3 activation and increase TDP-43 pathological modifications. J Neurosci. 2012 Jul 4;32(27):9133-42. PubMed.