4 February 2010. Rodents and fruit flies and worms, oh my! The new animal models for TDP-43 proteinopathies, a series of related neurodegenerative diseases, are coming out at a fast and furious rate. The latest to debut, published online this week by PNAS, are mouse models from researchers at the VIB-University of Antwerp in Belgium. The scientists report a set of mice that express wild-type, human TDP-43 in neurons. The work was led by first author Hans Wils and senior authors Christine Van Broeckhoven and Samir Kumar-Singh in Van Broeckhoven’s laboratory.
“This is the first mouse model that has TDP-43 inclusions,” Kumar-Singh said. “This is going to be a very powerful series of animal models that can be used for eventual drug targeting.” The mice are also the first wild-type TDP-43 overexpression strains to be published; last October Robert Baloh and colleagues at Washington University in St. Louis, Missouri, reported on a mouse carrying human TDP-43 with the disease-linked mutation A315T (see ARF related news story on Wegorzewska et al., 2009). Other animal models under development include rats, zebrafish, fruit flies, and nematodes (see ARF related news story). “The avalanche has begun,” quipped Brian Kraemer of the University of Washington in Seattle. Kraemer was not involved with the current study but has his own TDP-43 knockout mouse in the works. The fish, flies, and worms offer powerful genetic platforms to explore the normal functions of TDP-43; the rodents, it is hoped, will provide systems more akin to human biology.
In people, TDP-43 proteinopathies encompass a spectrum of diseases including some forms of amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD). Researchers have discovered several TDP-43 mutations in some people with ALS, but in most cases of TDP-43 proteinopathy, it is the wild-type protein that aggregates and is linked to neurodegeneration. The new mouse models may best mimic these sporadic cases, Kumar-Singh said. In that way it parallels a Drosophila model published recently, which also showed neurodegeneration from the wild-type protein (see ARF related news story). Ronald Klein and colleagues at the Louisiana State University Health Sciences Center in Shreveport also showed that the wild-type protein could be destructive in rodents. They reported last year on a rat model, produced by injecting a viral vector carrying the wild-type TDP-43 gene into the substantia nigra (see ARF related news story on Tatom et al., 2009).
Wils and colleagues created transgenic mice expressing human TDP-43 driven by the neuron-specific mouse Thy-1 promoter, which activates approximately one week after birth. They used multiple mouse lines to analyze the animals’ response to different “doses” of human TDP-43. Of the lines they initially made, the two highest expressers made 1.9- and 2.8-fold more TDP-43 than non-transgenic mice, according to immunohistochemistry analyses. The researchers then self-crossed each line to create homozygotes that expressed approximately double that again, 3.8- and 5.1-fold more TDP-43 than non-transgenics. (At the RNA level, quantitative PCR analysis showed that the animals had 0.6- to 1.2-fold of the human TDP-43 gene, compared to the endogenous mouse gene.)
Mice with the greatest TDP-43 load developed their first symptom, abnormal hindlimb reflexes, by just two weeks of age. Symptoms progressed to include stumbling, impaired performance on the rotarod test, and facial muscle spasms. These animals died within a month of birth. The mice expressing 3.8-fold normal TDP-43 evinced hindlimb symptoms at two months, and survived for nearly seven months. The 2.8-fold animals had normal reflexes until 14 months of age. Kumar-Singh and colleagues euthanized these animals before determining their survival rate, but in more recent studies, some of them have survived to 20 months or more, he told ARF in an e-mail. The researchers have not noticed any symptoms in the lowest, 1.9-fold expressers, which they have followed for 19 months. Degeneration of motor neurons, characteristic of ALS, and cortical neurons, as in FTLD, was similarly dose dependent.
TDP-43 appears to be a Goldilocks protein: too much or too little is lethal; a certain amount is just right. Other researchers attempting to make TDP-43 mice have found the project challenging due to the narrow window between getting no phenotype and getting a lethal one. “It tells me that TDP-43 is very well controlled,” Kumar-Singh said. Baloh noted that using the Thy-1 promoter as Wils and colleagues did is a promising approach because the transgene is not activated until after birth, so the excess TDP-43 does not interfere with early development.
How might these models relate to the human disease? There is no evidence that people with TDP-43 proteinopathy have abnormally high levels of the protein, but Kumar-Singh theorized that just a little extra TDP-43, over the course of a lifetime, could make neurons vulnerable. Similarly, he noted, increased levels of amyloid precursor protein—due to duplication or trisomy—can increase the protein’s dose and cause Alzheimer disease. Because levels of the protein are so tightly controlled, simple gene duplication is unlikely to explain TDP-43 proteinopathies. However, Kumar-Singh suggested that polymorphisms in the untranslated regions flanking the TDP-43 gene, or in a gene that influences its expression, might cause increased production. Researchers have already found a variant in the TDP-43 untranslated region in a person who had FTLD (Gitcho et al., 2009).
Mouse vs. Mouse
The TDP-43 field now has both wild-type TDP-43 and mutant mouse models, but “I would not say they are a matched pair,” Baloh said. They are closer than apples and oranges, more like “oranges and tangerines,” he said. Baloh and his colleagues used the mouse prion promoter PrP, which activates expression in glia as well as neurons. They saw TDP-43 levels approximately three times that of endogenous protein, although these data are not directly comparable to the Wils data because they are based on lysate analysis, not immunohistochemistry. But like Kumar-Singh’s, Baloh’s animals were ill. They suffered motor symptoms by three or four months and lived for an average of five months.
Although human TDP-43 is present throughout the brain in models from both groups, certain neuronal populations were particularly sensitive to the transgene. Layer V cortical neurons, including the primary motor cortex, as well as spinal motor neurons, were hard hit in both models. By confirming Baloh’s observations, the new publication shows that the effects on the cortex were not specific to one line. “That is really exciting,” Baloh said. “What is it about those neurons that are so vulnerable to TDP-43 toxicity?”
A key hallmark of TDP-43 proteinopathy is aggregates containing ubiquitin and phosphorylated TDP-43. The Baloh group observed aggregates, but the accumulations did not contain TDP-43. Van Broeckhoven and colleagues, in contrast, observed that nuclear aggregates, and some cytoplasmic aggregates, in their highest-expressing mice were positive for phosphorylated TDP-43 and ubiquitin. The difference may not be major, Baloh suggested. He noted that the Van Broeckhoven mice were much sicker than his at the time they were sacrificed for immunostaining. The two groups also used different antibodies, and Baloh thinks that with the right antibodies, he, too, might see the occasional TDP-43-positive aggregate in his mice.
In TDP-43 proteinopathy, the normally nuclear protein tends to abandon the nucleus for the cytoplasm. Baloh and colleagues noted occasional, but not widespread, nuclear clearing of TDP-43. The Kumar-Singh group noticed TDP-43 inclusions in both the nucleus and cytoplasm. Caspase-based cleavage of TDP-43, releasing a 25-kDa carboxyl-terminal fragment, has also been linked to the protein’s toxicity (see ARF related news story on Zhang et al., 2009). Both wild-type and mutant TDP-43 mice showed evidence of these damaging fragments as well. The Van Broeckhoven group compared nuclear and cytoplasmic fractions and discovered the fragments were primarily nuclear in their mice, an unexpected finding given the usual association between cytoplasmic localization and cytotoxicity.
“This is the best TDP-43 animal model to date, because it is a mouse and wild-type TDP-43,” Klein wrote of Kumar Singh’s work in an e-mail to ARF. This will certainly not be the last model, but with just Kumar-Singh’s and Baloh’s work to go on, the question remains: Did the A315T mutation in Baloh’s model cause disease, or are those mice sick simply because of overexpression of TDP-43 itself, as in Kumar-Singh’s strain? To solve that conundrum, Baloh said, will require a matched pair of mice with wild-type or mutant TDP-43, under the same promoter, at the same level of expression. Kumar-Singh reported that the group has an as-yet-unpublished mutant TDP-43 line that is similar to Baloh’s strain.
“There is never a single mouse that will do everything,” said John Trojanowski of the University of Pennsylvania in Philadelphia. “It is inevitably the case that you need more than one transgenic mouse to model a disease.” His laboratory is pursuing its own mouse model, with a different strategy. Their results are different from, but complementary to, the published mouse models, Trojanowski told ARF.
A single species cannot do everything, either, Kraemer noted. Mice will be useful for the final tests of promising therapeutics, but to find those potential treatments, researchers will need to perform high-throughput screens in vitro or with smaller, cheaper animals—such as worms, flies, and fish.
“The optimistic thing for patients and families is that the pace of research is going so much more rapidly now,” Trojanowski said. In comparison, he noted, tau was discovered to be a component of tangles in the mid-1980s, but mouse models were not available until the end of the 1990s. TDP-43 was linked to disease in 2006 (see ARF related news story on Neumann et al., 2006), and four years later scientists can pick and choose from a plethora of model systems.—Amber Dance.
Wils H, Kleinberger G, Janssens J, Pereson S, Joris G, Cuijt I, Smits V, Ceuterick-de Groote C, Van Broeckhoven C, Kumar-Singh S. TDP-43 transgenic mice develop spastic paralysis and neuronal inclusions characteristic of ALS and frototemporal lobar degeneration. Proc Natl Acad Sci U S A. 2010 Feb 23;107(8):3858-63. Abstract