The importance of RNA processing in ALS has been well established by the discovery of pathogenic mutations in key RNA processing proteins, most notably TDP-43 which is abnormally aggregated in neurons of almost all patients. This session described work by a number of groups attempting to characterise the mechanisms associated with RNA processing in ALS pathogenesis.
First Christian Haass described recently published work from his group on repeat associated non-ATG translation (RANT) of the C9orf72 hexanucleotide repeat expansion. A number of other studies have recently shown that RNA foci are formed from both the sense and the antisense repeat sequence; Mori, Haass et al have demonstrated that RANT protein is generated from both the sense and antisense transcripts. Moreover, dipeptide repeat proteins translated from sense and antisense transcripts were shown to co-aggregate with each other (despite distinct biophysical properties) within p62-positive inclusions which characterise ALS and particularly C9ORF72-ALS. Finally, by raising antibodies against the predicted C-terminal region of the RANT-protein it was suggested that RANT occurs through the full length of the repeat expanded transcript, at least in a sense direction. The next step in this story is perhaps the hardest: it remains to be determined which (or all) of these RNA and protein structures is key to the pathogenesis of ALS. This will require the careful generation and comparison of models of each and every iteration.
In the next talk Diane Moujalled described her work on hnRNP K. hnRNP proteins are growing in prominence in study of RNA processing in ALS, but hnRNP K is a newcomer to the discussion. By co-immunoprecipitation Diane and colleagues demonstrated an interaction with another hnRNP protein: TDP-43. This group has recently published work describing modulation of the ALS associated accumulation of TDP-43 in stress granules by cyclin-dependent kinases and GSK3; in her talk Diane described how TDP-43 does not have a consensus binding site for these kinases and hypothesised that their action may be mediated via hnRNP K. Indeed she described how in SH-SY5Y cells siRNA knockdown of hnRNP K inhibits movement of TDP-43 into stress granules.
Daniela Zarnescu described work of her group using a drosophila model of neurodegeneration produced by overexpression of wild type and mutant TDP-43. They demonstrated that overexpression of fragile X mental retardation protein (FMRP) rescues this phenotype. FMRP has an established role in RNA transport and translation. Daniela and collegues have discovered an RNA dependent interaction between FMRP and TDP-43, and suggest that overexpression of FMRP can reduce aggregation of TDP-43. This is fascinating work; the exact mechanism and the significance in the human disease remain to be determined and I look forward to the next step of the story.
MND Association / MRC funded Lady Edith Wolfson fellow Pietro Fratta next described transcriptome profiling in two mouse models with ENU-induced mutations of TDP-43. Neither mutation is found in human cases, however both provide insight into the function and dysfuntion of this key protein. First, the F201I mutation is located in the RRM1 domain and is shown to virtually abolish RNA binding by the protein; as such Peitro hypothesises that this is loss-of-function model. In contrast, the M323K mutation is in the glycine rich region which also contains the majority of human pathogenic mutations. Using RNA-sequencing Peitro and collegues have shown that transcriptome changes in the two mice are almost always in opposite directions. Comparison with the work of Polymenidou, Cleveland et al which described transcriptome changes in the mouse brain resulting from depletion of TDP-43, revealed that the F201I changes were very similar to TDP-43 depletion except in the case of transcripts containing long introns which have been particularly implicated in the pathogenesis of ALS. Pietro therefore suggested that this important interaction with long intron-containing transcripts was independent of the RNA binding ability of TDP-43, perhaps explaining why depletion of FUS and TDP-43 produce divergent effects except in the case of these transcripts.
Jochen Weischaupt presented analysis of miRNA expression in sporadic and familial ALS patients. Interestingly the most prominent signature was identified when all familial patients were merged into a single group and compared to controls. A progression of expression changes was seen when comparing controls, asymptomatic carriers of particular genetic changes and symptomatic familial ALS patients.
Finally Udai Pandey presented recently published work describing another drosophila model of neurodegeneration resulting from overexpression of mutant FUS, another RNA processing protein mutated in certain ALS patients. By abolishing the RNA binding ability of mutant FUS by an additional mutation to the RNA binding site (drawing parallels with the work of Petro in TDP-43) Daigle, Pandey et al ameliorated the toxicity of the mutant FUS and prevented its mislocalisation into cytoplasmic stress granules. So preventing the incorporation of mtFUS into stress granules is therapeutic in this model; Udai also suggested induction of autophagy could ameliorate toxicity of mtFUS in the same model, via disassembly of stress granules.
In conclusion the talks in this session highlighted progress that is being made in the understanding of the role of RNA processing proteins in ALS. Common themes emerge around the mislocalisation of FUS and TDP-43 into cytoplasmic stress granules and perhaps this will become an important therapeutic target. Particularly with respect to RNA processing in C9ORF72-ALS progress has been break taking; we await the models which will answer the relative toxicity question and point to the most important mechanism if there is one.