20 January 2011. It is so easy to point the finger at conglomerates. In the brain, large inclusions of proteins—be they amyloid-β, α-synuclein, or TDP-43—just look like trouble. But lately, they are starting to look less culpable. In the case of the ubiquitinated aggregates of the RNA-regulating protein TDP-43—a pathologic hallmark of frontotemporal lobar dementia and amyotrophic lateral sclerosis—inclusions may be irrelevant to disease, conclude Virginia Lee and colleagues at the University of Pennsylvania, Philadelphia. In a paper published online January 4 by the Journal of Clinical Investigation, they report that in transgenic mice expressing the human protein, inclusions rarely formed, and did not correlate with neurodegeneration. The researchers did observe a correlation between neurodegeneration and the disappearance of mouse TDP-43 from the nucleus. They suggest that this vanishing act, and its effect on the many genes TDP-43 modulates, is the crux of the disease. The researchers also confirmed that among the many unknown TDP-43-regulated mRNAs is the sequence for TDP-43 itself.
TDP-43 inclusions in people are mainly cytoplasmic, leading scientists to suspect that the normally nuclear DNA- and RNA-binding protein leaves the nucleus during disease (see ARF related news story on Barmada et al., 2010). To study the effects of TDP-43 localization, Lee, first author Lionel Igaz, and colleagues used a construct lacking the protein’s nuclear localization sequence (ΔNLS), forcing it to remain in the cytoplasm. Another key tool in the study was a new antibody specific for mouse TDP-43, the first of its kind, which allowed the scientists to differentiate between the endogenous protein and the transgene-encoded human TDP-43.
Lee and colleagues have already studied the ΔNLS TDP-43 mutation in cell culture, where they found the construct accumulated in cytoplasmic aggregates (see ARF related news story on Winton et al., 2008). To ask the same question in mice, they used an inducible system to turn on human transgenes—either ΔNLS or wild-type TDP-43 (WT)—in forebrain neurons. At four weeks of age, the researchers activated the genes and then observed the mice for up to six months.
Following the activation of either transgene, the researchers noticed neuron loss in the cortex, hippocampus, and olfactory bulb. The forebrain expression led to neurodegeneration more reminiscent of FTLD than ALS, although motor neurons and the upper corticospinal tract were also affected. In animals with WT human TDP-43, in whom the protein was primarily nuclear, degeneration and the accompanying gliosis began after approximately one month, and progressed for at least six months. To quantify the neurodegeneration, the researchers counted neurons in the hippocampal dentate gyrus; 20 percent of neurons disappeared in a month. In contrast, in ΔNLS mice, half of the dentate gyrus neurons—in which TDP-43 was cytoplasmic—degenerated within the same timeframe.
The pathology in Lee’s mice parallels that in many other mouse models. Most have few ubiquitinated TDP-43 aggregates (see ARF related news story on Wegorzewska et al., 2009), although several mouse models evince mitochondrial aggregates devoid of TDP-43 (see ARF related news story on Shan et al., 2010; ARF related news story on Xu et al., 2010). In Lee’s ΔNLS animals, TDP-43 inclusions formed in less than 1 percent of cortical neurons, and in even fewer WT TDP-43 neurons. When the researchers identified neurons that were undergoing degeneration, they saw no correlation with the presence or absence of TDP-43 aggregates.
The results were surprising, said Robert Baloh of Washington University in St. Louis, Missouri. The ΔNLS construct “was the one, if people were going to bet, that would form inclusions,” said Baloh, who was not involved in the study. Lee suggested that a second hit, such as cellular stress, might cause aggregate formation. Indeed, researchers have found that TDP-43 can reside in stress granules (Liu-Yesucevitz et al., 2010).
Although aggregates appeared not to kill neurons, Igaz and colleagues did observe, using their mouse-specific antibody, that the disappearance of endogenous TDP-43 correlated with neurodegeneration. The authors suggest that this loss of nuclear, endogenous TDP-43 is what makes the mice sick. “There is definitely a lot of evidence for loss of function being a very bad thing,” said Brian Kraemer of the University of Washington in Seattle, who was not involved with the study. “I think the fact that you can have toxicity even without any nuclear-localized TDP-43 is an important finding.”
However, the study does not preclude the possibility that soluble TDP-43 in the cytoplasm is also toxic. “I think you have two different processes going on,” said Ben Wolozin of Boston University in Massachusetts, who was not involved in the study. Disease could arise from a combination of TDP-43 loss in the nucleus and the toxic actions of the cytoplasmic protein.
Previous work has indicated that TDP-43 manages its own translation (see ARF related news story on Sephton et al., 2011), and the current study confirms it at the protein level. The more human TDP-43 cells had, the less mouse protein was present. “Making the mouse antibody really nailed the question down,” Baloh said. Scientists have already discovered the mechanism: TDP-43 the protein binds to the 3’ untranslated region of TDP-43 the mRNA, destabilizing the transcript (Ayala et al., 2010). Since the transgene lacked the 3’UTR, only the endogenous protein was affected.
“TDP-43 itself is finely regulated,” Lee noted—too much will kill a cell, and so will too little. It is going to be tricky to target TDP-43 because drugs simply aimed at it would risk altering its concentration and potentially be damaging. Instead, Lee said, scientists should look for downstream targets of TDP-43 that are involved with disease, and might be safer to modulate as a treatment.
The study authors started to identify TDP-43 targets via gene expression microarrays in cortical neurons from mice two weeks after transgene activation. In the ΔNLS animals, 4,700 genes were differentially expressed compared to non-transgenic mice; 197 of those were changed more than twofold. The hTDP-43 WT transgenic animals evinced gene expression patterns in between that of non-transgenics and ΔNLS mice. The researchers note that not all cells had even begun to express the TDP-43 WT transgene at this point.
“The challenge with microarrays is that we do not know if those are direct or indirect effects,” Baloh commented; it is possible that many gene changes were generic to sick neurons and not specifically controlled by TDP-43. Lee noted that by collecting samples at two weeks post-induction, the scientists focused on changes relevant to the advent of disease. “We identified the timeframe when the transgene is expressed, and endogenous TDP replaced but minimal secondary changes,” she said. Another approach scientists are using to identify TDP-43 targets is to pull down RNAs attached to the protein, thus limiting their results to RNAs that actually bind TDP-43 (see ARF related news story on Sephton et al., 2011).
In other recent TDP-43 news, researchers from the Italian Amyotrophic Lateral Sclerosis Consortium found that many people with ALS in Sardinia share a TDP-43 mutation. The study, led by Adriano Chiò of the University of Torino, Italy, appeared online January 10 in the Archives of Neurology. Simultaneously, the Archives also published a paper by Hussein Daoud, Guy Rouleau, and colleagues at the University of Montréal in Québec, Canada, who hunted for new ALS-related mutations among 29 candidate genes involved in motor neuron differentiation and development (Arlotta et al., 2005). These scientists found suspicious variants in new genes including heavy neurofilament (NEFH), the proteoglycan lumican (LUM), and the thyroid hormone-binding protein CRYM, but more work is needed to confirm the findings.—Amber Dance.
Igaz LM, Kwong LK, Lee EB, Chen-Plotkin A, Swanson E, Unger T, Malunda J, Xu Y, Winton MJ, Trojanowski JQ, Lee VM. Dysregulation of the ALS-associated gene TDP-43 leads to neuronal death and degeneration in mice. J Clin Invest. 2011 Jan 4. pii: 44867. doi: 10.1172/JCI44867. Abstract
Chiò A, Borghero G, Pugliatti M, Ticca A, Calvo A, Moglia C, Mutani R, Brunetti M, Ossola I, Marrosu MG, Murru MR, Floris G, Cannas A, Parish LD, Cossu P, Abramzon Y, Johnson JO, Nalls MA, Arepalli S, Chong S, Hernandez DG, Traynor BJ, Restagno G; and the Italian Amyotrophic Lateral Sclerosis Genetic (ITALSGEN) Consortium. Large proportion of amyotrophic lateral sclerosis cases in Sardinia due to a single founder mutation of the TARDBP gene. Arch Neurol. 2011 Jan 1. Abstract
Daoud H, Valdmanis PN, Gros-Louis F, Belzil V, Spiegelman D, Henrion E, Diallo O, Desjarlais A, Gauthier J, Camu W, Dion PA, Rouleau GA. Resequencing of 29 candidate genes in patients with familial and sporadic amyotrophic lateral sclerosis. Arch Neurol. 2010 Jan 10. Abstract