The researchers took advantage of the particular characteristics of their model system, Schmid noted in an interview with Alzforum. The RNA interference screens Liachko performed are easy to do in nematodes; moreover, unlike some other animal TDP-43 models, the worm actually mimics the TDP-43 phosphorylation seen in people. Zebrafish allowed Schmid to cleanly knock out TDP-43 and analyze stages of development that are difficult to study in mammals. Mouse embryos die before birth without TDP-43, for example, but that only tells researchers that the gene must be crucial for development. Fish embryos lacking TDP-43 die eight days after fertilization; however, since fish eggs are fertilized outside the mother’s body and develop into translucent embryos, Schmid was able to investigate why the knockouts expire.

Something Fishy
Zebrafish carry two homologues of the human TDP-43 gene, TARDBP and a “TARDBP-like” gene called TARDBPl. The former protein contains the two RNA recognition motifs and a carboxy-terminal glycine-rich domain found in human TDP-43, whereas the latter protein lacks the glycine domain. Mutations that cause ALS tend to cluster in that region. When the researchers knocked out one or the other fish gene, “the result was really boring,” complained Haass in his talk. Nothing seemed to happen to the motor neurons or any other part of the fish embryo.

At first, the researchers, assuming that the glycine-rich domain was crucial to TDP-43 function, were surprised that fish lacking TARDBP were healthy. But looking more closely at the sequence of the TARDBP-like protein gene, they saw a potential answer. The glycine-rich code was present, but usually spliced out. A Western blot showed that the glycines were included in TARDBPl when TARDBP was not present. “This glycine-rich domain is so extremely important for the function that the fish makes a backup version,” Schmid concluded.

To obtain a full TDP-43 knockout phenotype, then, Schmid had to delete both TARDBP and TARDBPl genes. In these embryos, Schmid saw shortened motor neuron axons, as might be expected from deleting a gene linked to ALS. Furthermore, she saw that muscles degenerated and red blood cells pooled in the embryo’s yolk instead of coursing through the circulatory system. Blood vessels were wildly disorganized in the brain and body of the double mutant. “Probably TDP-43 is, at least developmentally, required to maintain muscles, the vasculature, and axonal outgrowth,” Haass concluded. Schmid suggested that researchers studying ALS and FTD in people might find it fruitful to look for muscle and blood vessel defects.

Aaron Gitler of Stanford University, who was not involved in the study, noted that this is the first vertebrate animal model with TDP-43 genetically removed from the genome—not eliminated by RNA interference, which tends to leave some of the protein around. Even if the defects in vasculature and musculature turn out to be specific to fish, studying the affected cells could still inform researchers about TDP-43 pathology that might occur in human neurons, Gitler added.

Worm Work
Liachko, who works in the laboratory of Brian Kraemer at the University of Washington in Seattle, studies worms that express mutant human TDP-43. Their motor neurons degenerate, and they are “severely uncoordinated,” Liachko said in her presentation. The worms tend to coil up and rarely stray from where they hatched. She previously reported that phosphorylation at serines 409 and 410 contributes to the toxicity of mutant TDP-43 in C. elegans (Liachko et al., 2010). This contrasts with some mouse TDP-43 models, which have not exhibited the same post-translational modification, although some newer models do, Liachko told Alzforum.

Where there is phosphorylation, there must be a kinase, and Liachko described her search for it. Using short interfering RNAs, she knocked down 186 different enzymes in worms with mutant TDP-43, looking for animals that moved normally. She found 12 that did, and checked whether the silenced genes directly affected TDP-43. In three hits, phosphorylation of the protein fell.

The candidates were the homologues of human CDC7, TTBK1, and TTBK2. Pairing the purified human kinases with TDP-43 in vitro, Liachko found that only CDC7 directly phosphorylated TDP-43. Being a cell cycle regulator, CDC7 is not an enzyme scientists would expect to find active in non-dividing neurons. Even so, Liachko showed that the kinase appears in the cell bodies and nuclei of neurons in the human brain. CDC7 clearly has functions beyond the cell cycle, Liachko said, although she does not know what they might be.

Could inhibiting CDC7 prevent TDP-43 toxicity? Using the small molecule inhibitor PHA767491, Liachko reduced not only TDP-43 phosphorylation, but also neurodegeneration in her C. elegans model. The drug also limited TDP-43 phosphorylation in mouse NSC-43 cells, which are a hybrid motor neuron-neuroblastoma line where TDP-43 phosphorylation does occur.

CDC7 inhibitors were tested in clinical trials to treat cancer. They interfere with the tumor cells’ cycling, causing apoptosis. Liachko noted that attacking a crucial cell cycle protein is one thing in acute chemotherapy with unpleasant side effects, but quite another in progressive neurodegenerative diseases such as ALS and FTD. “You cannot chronically downregulate CDC7 in a person and expect a good outcome,” she said. But based on her data, she believes, “the idea of preventing TDP-43 phosphorylation is promising.”

Schmid told Alzforum the study was elegant and worthy of follow-up. It demonstrates the value of screens in model systems, Gitler added. Both scientists suggested the next step should be to analyze TDP-43 phosphorylation in rodent models. “We will be able to do that soon,” Liachko said, since now there are mice that recapitulate the phosphorylation phenotype.—Amber Dance.