C9orf72 repeat expansion alters post-transcriptional gene regulation in a cellular model of amyotrophic lateral sclerosis

A non-coding GGGGCC (G4C2) repeat expansion in the C9orf72 gene is the major genetic determinant of Amyotrophic Lateral Sclerosis (ALS), a late‐onset neurodegenerative disorder characterized by the selective loss of upper and lower motor neurons (Cozzolino et al., 2012). A common feature of non-coding repeat expansion disorders is the accumulation of RNA repeats as RNA foci in the nucleus and/or cytoplasm of affected cells. These RNA foci can be toxic by sequestering RNA-binding proteins, whose decreased availability may then affect various steps of post-transcriptional gene regulation, such as alternative mRNA splicing, translational regulation, mRNA transport, or mRNA decay (La Spada and Taylor, 2010). In samples from C9orf72 patients as well as patient-derived iPSC, RNA foci containing the sense as well the antisense RNA repeat sequence are likewise detected (Gendron et al., 2013; Lagier-Tourenne et al., 2013; Mizielinska et al., 2013), indicating that the sequestration of RNA-binding proteins and hence a dysregulation in one of the steps of RNA metabolism might well play a role in ALS. The precise step that is affected, however, remains ill defined, and thus our understanding of C9orf72 toxicity in the disease pathogenesis is still elusive. To get insights into these mechanisms, we built a cellular model based on the overexpression of normal and expanded G4C2 repeats, and we observed that the expression of (G4C2)31 pure repeats is sufficient to induce the formation of intra-nuclear RNA foci in mouse motor neuron-like NSC34 cells as well as in human HeLa cells. We then used an in vitro-transcribed biotinylated RNA containing (G4C2)31 repeats to identify C9orf72 RNA binding proteins. Through pull down assays and mass spectrometry analysis we were able to identify many different factors involved in posttranscriptional gene regulation, such as members of the hnRNP family, which regulate alternative splicing, and translational regulators, including initiation and elongation factors. Since a significant, although not complete, sequestration of some of these factors into RNA foci was observed, we analysed whether alternative splicing process and protein translation could be affected in cultured cells expressing the expanded repeats. Indeed, we observed that the expression of (G4C2)31 repeats alters the alternative splicing pattern of a splicing reporter (pSMN2), through the interaction with the splicing factor hnRNP H. Moreover, the expression of (G4C2)31 is able to activate a stress response that lead to a general reduction of translation. Indeed, (G4C2)31 repeat widely affects the overall distribution of Pura and its binding partner FMRP, two well-known regulators of translation, that accumulate into intra-cytosolic granules which are positive for stress granules markers. However, translational repression is not achieved through eIF2a phosphorylation-mediated ternary complex inhibition, while C9orf72 repeats strikingly induce an abnormal nuclear accumulation of polyadenylated mRNAs, and this is associated to the nuclear relocalization of the cytosolic form of poly(A) binding protein (PABPc). Thus, our observations suggest that nuclear accumulation of mRNAs, as a consequence of C9orf72 repeats ability to impair nuclear mRNA export, might contribute to ALS pathogenesis. Overall, our results indicate that G4C2 repeats affect post-transcriptional gene regulation, thus supporting the concept that a selective vulnerability of motor neurons to alterations in mRNA metabolism may be a common denominator in ALS, as well as in other motor neuron disease.