8 September 2010. In motor neuron disease, evidence is mounting that RNA processing lies at the heart of molecular pathogenesis. The August 24 Proceedings of the National Academy of Sciences carried a paper on the latest in a long string of mice overexpressing TAR DNA-binding protein-43 (TDP-43), which, when mutated, can cause amyotrophic lateral sclerosis (ALS) and frontotemporal lobar dementia. The new data suggest that TDP-43 is but one of at least three RNA-processing proteins implicated in motor neuron dysfunction. The researchers found TDP-43 in the same aggregates as fused in sarcoma (FUS), the RNA-processing protein of another gene linked to ALS. And excess TDP-43 also altered the localization of survival motor neuron (SMN)—the RNA-processing protein encoded by a gene that is involved in spinal muscular atrophy (SMA), another motor neuron disease.
In addition, the researchers, led by senior author Philip Wong of Johns Hopkins University School of Medicine in Baltimore, report that proper TDP-43 levels are necessary for the distribution not only of RNA-processing structures in the nucleus, but also of mitochondria in the cytoplasm.
Wong and colleagues work to understand the physiological function of TDP-43 in motor neurons by studying mice that lack or overexpress the gene. “We are not modeling any disease per se,” Wong said. “We believe that, in vivo, mature motor neurons will give us the most valuable information.” More recently, the same researchers published a TDP-43 conditional knockout mouse, which linked the protein to fat metabolism. Once TDP-43 was eliminated, the animals quickly lost fat and died (Chiang et al., 2010). In the current work, joint first authors Xiu Shan and Po-Min Chiang, both at Johns Hopkins, describe the counterpart to that knockout mouse: animals that express human, wild-type TDP-43 over and above the mouse version. The scientists placed the human transgene under control of the Thy1.2 promoter so it would activate in neurons following birth. Of the three transgenic lines they obtained, the researchers selected the lowest-expressing line for further study. Though this mouse exhibited somewhat stunted growth and tremors, it survived to adulthood, and the authors did not report any change in fat metabolism. However, they did find a redistribution of key motor neuron organelles.
“When we overexpressed TDP-43, we observed two dramatic consequences, one in the nuclear compartment and one in the cytoplasm,” Wong said. In the nucleus, the excess protein formed abnormal inclusions. FUS joined TDP-43 in these inclusions, providing in-vivo confirmation of an interaction already shown in vitro (see ARF related news story on Ling et al., 2010; Kim et al., 2010). “It suggests a unification of two distinct causes of ALS,” said Randal Tibbetts of the University of Wisconsin in Madison, who was not involved in the study.
The other consequence Wong’s group noticed was an abnormality in so-called Gemini of coiled bodies (GEMs), another nuclear structure. GEMs are sites of RNA processing that contain SMN, the protein associated with SMA. Motor neurons typically contain two GEMs in their nucleolus. But in the TDP-43-overexpressing mice, motor neurons hosted up to eight GEMs, Wong said, and they tended to cluster around the nucleolus instead of inside it. “There are probably gross alterations in RNA metabolism in these motor neurons,” Tibbetts said. In contrast, in the TDP-43 knockout mouse, motor neurons had no GEMs at all.
The researchers found a different story when they examined the cytoplasm. Human ALS cases at autopsy often evince TDP-43 inclusions in the cytoplasm, yet the researchers found little TDP-43 pathology outside of the nucleus in neurons of these new mice. They did see cytoplasmic inclusions of another sort, however, namely big aggregates of mitochondria. These eosinophilic, mitochondrial aggregates have shown up previously in other TDP-43 mice (see ARF related news story on Xu et al., 2010 and ARF related news story on Wegorzewska et al., 2009), and suggest to researchers that energy metabolism may be altered in motor neuron diseases.
To further examine this mitochondrial phenotype, Wong and colleagues crossed their TDP-43-overexpressing mice with a strain carrying a cyan fluorescent protein mitochondrial tag. In wild-type animals, they observed, mitochondria appear both in the cell body and in axonal processes. But in the double-transgenics, neuronal processes contained many fewer mitochondria than wild-type mouse neurons. A dearth of mitochondria at the tips of the motor neuron axons could conceivably weaken neuromuscular junctions. Indeed, the researchers found that the junctions were underdeveloped and the muscle fibers were smaller than normal. In fact, Wong said, the junction pathology looks a lot like junctions in SMA model mice (Kong et al., 2009). “The findings plausibly link these two motor neuron conditions,” Tibbetts said.
Then why are mitochondria missing from the ends of axons? To look for clues, the researchers analyzed mRNA from spinal cords of three TDP-43 and three control mice. They used differential cDNA sequencing to quantify mRNAs because it allows identification of alternatively spliced isoforms (see upcoming ARF Webinar). The researchers discovered 2,017 genes that were differentially expressed between the two mouse lines, and 313 with different splicing patterns in the TDP-43 mice. Notably, the TDP-43 mice carried fewer mRNAs for several proteins involved in forming the cytoskeleton and transporting cargo along it. “Maybe,” Wong speculated, “we are perturbing axonal transport.” Other researchers have proposed that that altered mitochondrial fission and fusion may underlie symptoms in TDP-43 mice (see ARF related news story on Xu et al., 2010). “That could make sense, too,” Wong said.
On the latter, Bob Baloh of Washington University in St. Louis, Missouri, who was not involved in the study, noted that mitochondrial problems have long been of interest in mice overexpressing human SOD1, another gene linked to ALS (e.g., see ARF related news story on Israelson et al., 2010). The presence of mitochondrial aggregates suggests one potential similarity between TDP-43 and SOD1 mice, Baloh said.
The new mouse also follows the trend in mouse models of TDP-43 proteinopathy. “They all look largely the same,” Baloh said, with similar disease symptoms and TDP-43 nuclear inclusions (see ARF related news story; ARF news story on Xu et al., 2010; ARF related news story on Wegorzewska et al., 2009; ARF related news story on Wils et al., 2010; Stallings et al., 2010). That’s a good thing, Baloh noted; when all models, each made in a slightly different way, are on the same track, it suggests that track is the right one to follow.—Amber Dance.
Shan X, Chiang PM, Price DL, Wong PC. Altered distributions of Gemini of coiled bodies and mitochondria in motor neurons of TDP-43 transgenic mice. Proc Natl Acad Sci U S A. 2010 Aug 24. Abstract