28 February 2013. Like a knot in your shoelace, entwined mRNAs can interfere with the nucleic acid’s normal actions. In amyotrophic lateral sclerosis, could it be that the cell overreacts to excess RNA snarls by responding with a toxic anti-virus response? Tassa Saldi of the University of Colorado in Boulder put forth this theory at the Keystone Symposium “New Frontiers in Neurodegenerative Disease Research,” held 4-7 February 2013 in Santa Fe, New Mexico. She studied the Caenorhabditis elegans orthologue of the ALS gene TDP-43, called TDP-1. Unlike TDP-43-negative mammals, C. elegans lacking TDP-1 survive, allowing her to examine the downstream effects on mRNAs.
“Their study was a great example of the importance of simple model organisms in learning about the normal function of disease proteins,” commented Aimee Kao of the University of California, San Francisco, in an e-mail to Alzforum. “It will hopefully stimulate some new thinking about the function of TDP-43 and how mutations cause disease.”
TDP-43 regulates expression and splicing of thousands of RNAs. To understand TDP-1’s role in the nematodes, Saldi, who works in the laboratory of Christopher Link, sequenced the transcriptome of the deletion strain. She discovered that 1,200 transcripts were over- or underexpressed compared to normal worms, and 350 genes were differentially spliced. No major gene categories emerged from the gene set, however, leaving few hints as to TDP-1’s primary effects.
Going through the list of genes one by one, Saldi did discover a common theme. It was gene overlap, a phenomenon where a given nucleotide sequence in one gene is also expressed as part of another gene (Sanna et al., 2008). Many genes differentially regulated in TDP-1’s absence overlapped with other genes, and the common sequences ran in opposite directions. This could occur if genes on opposite DNA strands share antiparallel coding sequences. It could also happen when one gene’s very long intron contains a second gene. This is not the first time long introns have been linked to TDP pathology; extended introns in the mouse genome are among the top targets of TDP-43 activity (see ARF related news story on Polymenidou et al., 2011, and Tollervey et al., 2011).
In general, about 8 percent of the worm genome overlaps, Saldi said. In the 1,550 genes that depend on TDP-1 for proper regulation, 35 to 45 percent were overlappers. The dataset also contained an unusually high number of introns of three kilobases or longer.
How does the TDP-1 knockout affect overlapping genes? If both sides of the DNA are transcribed at the same time, then their mRNAs are at risk of annealing to form a double-stranded structure. In fact, loss of TDP-1 resulted in noticeable dsRNA buildup in the worms, Saldi found. She stained the animals or individual tissues with an antibody to dsRNA and observed large nuclear inclusions. She could clear those aggregates by adding double-strand-specific RNase, but not RNase for single-stranded RNA, confirming their double-stranded structure.
Double-stranded RNAs, labeled with an antibody (red), accumulate in the nuclei of C. elegans lacking TDP-1. Image courtesy of Tassa Saldi, University of Colorado, Boulder
Saldi hypothesized that TDP-1 might work in RNA editing, which has evolved to disrupt these dsRNAs. Adenosine deaminases swap adenines for inosines, which pair awkwardly with the guanine on the opposite strand, forcing the dsRNA to unwind. This repair typically happens in untranslated regions or introns, and so does not interfere with the protein code, Saldi said.
When Saldi examined the editing system in her animals, she observed that the adenine-to-inosine transition still occurred in the transcriptome of TDP-1-negative worms. In fact, the deletion strain had more inosines than normal. Therefore, the editing process is working properly, Saldi said. She suspects that TDP-1 normally acts upstream, destabilizing dsRNAs so they unwind without editing. Worms lacking TDP-1, then, would require extra editing. Another possibility, she added, is that TDP-1 participates in dsRNA degradation.
Worms, of course, are not people, but Saldi found hints that TDP-43 in human cells performs a similar function. When she knocked down TDP-43 in HeLa cervical cancer cultures, dsRNA inclusions formed in the nucleus. Similarly, dsRNA inclusions formed in TDP-43-deficient M17 neurons, but this time the inclusions were cytoplasmic.
How might dsRNA aggregates contribute to neurodegeneration? Double-stranded RNA in the cytoplasm should raise a red flag, said Saldi, because it could be evidence of infection with a dsRNA virus. A natural response would be the interferon immune pathway that fights infection, but this can also be toxic. Could that be how TDP loss damages neurons? Supporting this theory, Saldi noted that many of TDP-43’s mRNA targets are involved in the interferon response (Polymenidou et al., 2011).—Amber Dance.