22 January 2010. TDP-43, normally a mild-mannered nuclear protein, morphs into a cytoplasmic super villain in TDP-43 proteinopathies. Two recent papers add to the quickly mounting pile of data (see ARF related news story) on how this transition might occur. Alas, a clear picture is not yet in sight. Writing in the January 13 issue of the Journal of Neuroscience, researchers from the University of California in San Francisco report that cytoplasmic localization, but not aggregation, is linked with neurotoxicity. And in a paper posted online January 4 by Molecular and Cellular Proteomics, scientists from Emory University in Atlanta, Georgia, report on post-translational modifications of the protein: when overexpressed, it picks up both ubiquitin and small ubiquitin-like modifier (SUMO).
TDP-43 causes trouble in the neurons of people with amyotrophic lateral sclerosis (ALS) as well as a subset of frontotemporal lobar dementia cases that include neural ubiquitin-positive inclusions (FTLD-U) (Arai et al., 2006). Mutations in TDP-43, mostly in the glycine-rich carboxyl-terminal domain, have been linked to both diseases (see ARF related news story on Sreedharan et al., 2008 and Gitcho et al., 2008), but TDP-43 also aggregates in sporadic cases where the person has no known mutation. TDP-43 aggregates are seen in postmortem pathology of both sporadic and familial Alzheimer disease, other forms of dementia, even Down syndrome (Kadokura et al., 2009; Schwab et al., 2009; Lippa et al., 2009), though its importance in these other diseases is as yet unclear.
TDP-43 is primarily nuclear, though it does visit the cytoplasm when it shuttles mRNAs across the nuclear envelope. In people with ALS and FTLD-U, TDP-43 all but deserts the nucleus for the cytoplasm and forms insoluble aggregates throughout the cell.
The work in the Journal of Neuroscience paper was led by first author Sami Barmada of the University of California, San Francisco, working in the laboratory of Steven Finkbeiner at the Gladstone Institute of Neurological Disease in San Francisco, with collaboration from the laboratory of Jane Wu at the Northwestern University Feinberg School of Medicine in Chicago, Illinois. Finkbeiner’s group has developed tools to study the localization and aggregation of huntingtin (see ARF related news story on Arrasate et al., 2004). When news of TDP-43’s role in ALS and FTLD-U broke in 2006 (see ARF related news story on Neumann et al., 2006), Barmada, who had treated people with FTLD as a clinician, approached Finkbeiner about applying those same methods to TDP-43 proteinopathies.
When they started the project, all the researchers knew was that TDP-43 accumulated in dying neurons—but it could be a cause or a consequence of that neurodegeneration. They expressed GFP-tagged wild-type TDP-43 or TDP-43 with an ALS-linked A315T mutation (see ARF related news story on Kabashi et al., 2008) in primary rat cortical neurons. They also transfected the cells with mCherry, which only fluoresces in living cells, to track cell viability. The scientists used automated time-lapse microscopy to image multiple fields of interest, encompassing hundreds of neurons, over 10 days (Arrasate and Finkbeiner, 2005). Using this system, they watched to see what would happen first: TDP-43 redistribution to the cytoplasm and inclusion formation, or signs of cell death such as membrane blebbing and loss of mCherry signal. “You have the longitudinal power of following what happens to each cell, but then you also have the power of statistics,” Barmada said. “It gives you the capability to make more clearly defined cause-and-effect conclusions.”
Wild-type TDP-43 did not increase the neurons’ risk of cell death, compared to cells expressing only GFP, but the mutant TDP-43 increased cell death risk by a factor of 1.2. (In comparison, disease-associated huntingtin increases cell death risk by a factor of 1.5.) This small increased risk of cytotoxicity might be compounded by other factors, genetic or environmental, to make neurons in people carrying the mutation vulnerable to disease.
Cells expressing high levels of TDP-43, particularly the mutant form, were more likely to exhibit inclusion bodies. However, these aggregates were not associated with a higher risk of cell death compared to cells with diffuse TDP-43. Perhaps, the authors suggest, TDP-43 inclusion bodies are fairly inert repositories for a protein that is toxic when in solution.
Next, the researchers wondered if the nuclear or cytoplasmic location of TDP-43, regardless of inclusion body presence, could be involved in cytotoxicity. They selected cells for analysis that had only nuclear or only cytoplasmic TDP-43. Having TDP-43 in the cytoplasm increased risk of cell death by a factor of two- to threefold, regardless of whether the TDP-43 was wild-type or mutant, although cells expressing the mutant protein were more likely to exhibit cytoplasmic localization. The data suggested that it is the cytoplasmic location, and not the mutation itself, that is truly toxic. Nuclear levels of TDP-43 did not affect cell survival.
Barmada and colleagues then hypothesized that by controlling TDP-43 localization, they could change its toxicity. They changed the nuclear localization sequence of wild-type TDP-43, forcing it to remain cytoplasmic, and found that it became as deadly as the TDP-43 A315 mutant, with a risk of cell death 1.3 times that of wild-type TDP-43 with normal localization. Conversely, when the scientists disrupted the nuclear export sequence of A315-mutant TDP-43, it remained nuclear and the cells survived as well as cells expressing wild-type TDP-43.
“It is always good to see more evidence that TDP-43 mutations are causing some sort of dominant neurotoxicity,” said Brian Kraemer of the University of Washington in Seattle, who was not involved with either study. However, he noted that TDP-43 was overexpressed in the cell culture model, and without knowing how much of the protein was present, it is hard to be confident that the localization reflects the protein’s normal activity. Barmada said that higher expression of TDP-43 was sufficient to re-route the protein to the cytoplasm; neurons with cytoplasmic localization generally had three- or fourfold more TDP-43 than those with nuclear localization. However, the researchers restricted their analysis to cells that had similar expression levels.
Dressed to Kill
In the second study, joint first authors Nicholas Seyfried and Yair Gozal spearheaded the research with principal investigators Junmin Peng and Allan Levey. The researchers, all part of the Center for Neurodegenerative Disease at Emory University, approached TDP-43 with an interest in proteomics and post-translational modifications.
TDP-43 comes in 10 different splice variants, and with the exception of the full-length form, all lack the glycine-rich carboxyl-terminal domain (Wang et al., 2004). In a recent paper (Chen et al., 2010), researchers found that this glycine-rich portion of TDP-43 seems to contribute to its aggregation.
The carboxyl terminus is cleaved by caspase-3 (see ARF related news story on Zhang et al., 2007), and the carboxyl terminal fragments form toxic cytoplasmic aggregates containing ubiquitin (see ARF related news story on Zhang et al., 2009).
The Emory researchers focused their study on the splice form TDP-S6. They overexpressed the protein in human embryonic kidney (HEK-293) cells and mouse primary hippocampal neurons, localizing it with immunofluorescence. Full-length wild-type TDP-43—either endogenous or overexpressed—remained soluble and nuclear, as expected. Despite lacking the aggregation-prone carboxyl end, TDP-S6 formed cytoplasmic and nuclear aggregates—rather surprising, given all the previous data pointing toward the carboxyl terminus as the aggregator.
To confirm their cell biology results, Seyfried and colleagues fractionated the transfected cells to separate soluble protein from insoluble aggregates, and used immunoblotting to confirm that TDP-S6 was present at higher levels than full-length TDP-43 in the insoluble fraction. Immunoblotting also revealed that cells transfected with either form of TDP-43 had increased levels of ubiquitin and SUMO in the insoluble fraction, compared to mock-transfected cells, suggesting these modifiers were a component of the aggregates. TDP-43 has already been identified as a SUMO partner in a previous study (Golebiowski et al., 2009). SUMO performs a similar function to ubiquitin in the nucleus, tagging misfolded proteins for destruction, so may interact with malformed TDP-43, the authors suggest.
Ubiquitin, once it attaches to a protein, then serves as a binding site for other ubiquitin molecules that attach to its lysine residues to make a poly-ubiquitin chain. When ubiquitins link up via lysine-48, this targets the protein for degradation by the proteasome. Lysine-63 linkages, in contrast, mediate inclusion formation, protein trafficking, and targeting to the lysosome for degradation. When the researchers used immunofluorescence to probe transfected cells with antibodies specific for ubiquitin linkages, they found that both full-length TDP-43 and TDP-S6 inclusions were decorated with both kinds of poly-ubiquitin chains. When the scientists quantified the results using mass spectrometry, they found that lysine-63 linkages in the insoluble fraction were increased in cells transfected with TDP-S6, suggesting this is the key modification. The ubiquitination of the lysine-63 site suggests that instead of targeting the protein to the proteasome, the ubiquitins may send it elsewhere, perhaps to the lysosome. Understanding TDP-43 turnover could lead to therapeutic targets, Seyfried suggested in an interview with ARF.
Seyfried did not dispute that the TDP-43 carboxyl terminus is important for its aggregation, but his study suggests the rest of the protein can cause trouble on its own. Most cases of ALS and FTD are sporadic, with no TDP-43 mutations, so this research suggests a possible mechanism by which the wild-type protein could lead to neurotoxicity. Perhaps, the authors write, dysregulation of TDP-43 splicing could cause disease. However, Seyfried pointed out that there is not yet evidence directly linking TDP-S6 with disease in people with ALS or FTD. “Whether this is biologically relevant needs further investigation,” he said.
So far, a growing body of data point to the protein’s localization as being important for toxicity, but exactly how it makes its move, what other proteins or modifications it might carry with it, and how it disrupts the cell’s health remain uncertain. The field is moving quickly, so stay tuned for more from this puzzling protein.—Amber Dance.
Barmada SJ, Skibinski G, Korb E, Rao EJ, Wu JY, Finkbeiner S. Cytoplasmic mislocalization of TDP-43 is toxic to neurons and enhanced by a mutation associated with familial amyotrophic lateral sclerosis. J Neurosci. 2010 Jan 13;30(2):639-649. Abstract
Seyfried NT, Gozal YM, Dammer EB, Xia Q, Duong DM, Cheng D, Lah JJ, Levey AI, Peng J. Multiplex SILAC analysis of a cellular TDP-43 proteinopathy model reveals protein inclusions associated with SUMOylation and diverse polyUb chains. Mol Cell Proteomics. 2010 Jan 4. Abstract