Abstract

Amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD) are two fatal neurodegenerative disorders that share overlapping clinical and pathological features (Kato, Hayashi et al. 1993, Lomen-Hoerth, Anderson et al. 2002). Although both diseases occur mostly sporadically, several disease-associated mutations have been identified in >25 different genes, many of which are encoding RNA binding proteins (RBP), components of the ubiquitin proteasome system (UPS), and the autophagy pathway (Kapeli, Martinez et al. 2017, Bartoletti, Bosco et al. 2019). Cytoplasmic inclusions of the multifunctional RNA-binding protein TDP-43 (TAR DNA-binding protein) that normally resides predominantly in the nucleus are found in ~90% of ALS and ~45% of FTD cases. TDP-43 dyshomeostasis including aberrations in its nucleocytoplasmic shuttling and aggregation has been shown to induce toxicity, explaining a crucial role in the process of neurodegeneration (Araki, Minegishi et al. 2014, Leibiger, Deisel et al. 2018, Prasad, Bharathi et al. 2019). The most common pathogenic mutation found in ~10% of ALS/FTLD patients is the massive expansion of a hexanucleotide repeat (G4C2)n in the first intron of C9orf72 gene (DeJesus-Hernandez, Mackenzie et al. 2011, Renton, Majounie et al. 2011). C9orf72 patients show, in addition to typical TDP-43 pathology, nuclear RNA foci containing the repeat RNA from both sense and antisense transcripts (DeJesus-Hernandez, Mackenzie et al. 2011) and unique inclusions of five different species of dipeptide repeat (DPR) proteins (poly-GA/-GP/-GR/-PA and -PR). DPR inclusions are derived from an unconventional non-AUG translation of the intronic repeat RNA in all reading frames (Ash, Bieniek et al. 2013, Gendron, Bieniek et al. 2013, Mori, Arzberger et al. 2013, Mori, Weng et al. 2013, Zu, Liu et al. 2013). Nearly all inclusions are poly-GA positive and often contain also poly-GP/-GR and far less frequently poly-PA/-PR. In vitro, poly-GA sequesters large amounts of stalled proteasomes suggesting a deleterious effect on cellular proteostasis (Guo, Lehmer et al. 2018). DPR pathology has been shown to precede TDP-43 pathology in patients, thereby considered as a major driver of neurodegeneration associated with C9ORF72 expansion (Baborie et al. 2015; Mann 2015; Mori, Arzberger, et al. 2013; Mori, Weng, et al. 2013; Proudfoot et al. 2014). How C9orf72‐specific pathology triggers TDP‐43 pathology, despite not being spatially correlated in human studies (Schludi et al. 2017), was the primary objective of my PhD studies. When I started my PhD project, several groups reported impaired global nucleocytoplasmic transport possibly through a direct effect on the nuclear pore complex (NPC) in different C9orf72 models and connected it to poly-GR/PR (Jovicic, Mertens et al. 2015), repeat RNA (Zhang, Donnelly et al. 2015), or both (Freibaum, Lu et al. 2015), however in these studies trafficking of TDP-43 itself was not analyzed. Furthermore, artificial aggregating β-sheet proteins were shown to inhibit nucleocytoplasmic transport related to sequestration of the THOC complex and RNA binding proteins (Woerner, Frottin et al. 2016). Since GA15 peptides, but not 15-mers of the other DPR species, form amyloid-like fibrils (Chang, Jeng et al. 2016), I asked whether poly-GA may also impair nucleocytoplasmic transport. I therefore compared the impact of individual expression of poly-GA, poly-GR and poly-PR on the nuclear import, specifically of TDP-43 to understand the link between the C9orf72 mutation and TDP-43 pathology. I quantitatively analyzed cytoplasmic mislocalization of endogenous TDP-43 and of a reporter containing the established bipartite classical nuclear localization signal (NLS) of TDP-43. Interestingly, poly-GA blocked nuclear import of TDP-43 more robustly than poly-GR/PR, while none of the DPR proteins affected the localization of a reporter containing a transportin-dependent PY-NLS in our in vitro models arguing DPR proteins mainly impair nuclear transport through the classical importin α/β pathway mediating TDP-43 import. In addition, I found that overexpression of two NPC components (NUP54 and NUP62) can fully rescue nuclear localization of the TDP-43 reporter. NUP54 and NUP62 were interestingly shown to be essential for nuclear import of TDP-43 (Nishimura, Zupunski et al. 2010) and NUP62 knockdown enhances (PR)25 toxicity in flies (Boeynaems, Bogaert et al. 2016). Thus, inhibition of nuclear import of TDP-43 by poly-GA may link the C9orf72 mutation to TDP-43 pathology in C9orf72 ALS/FTD cases. Since poly-GA precedes symptom onset by many years (Vatsavayai, Yoon et al. 2016) and rarely co-localize with TDP-43 inclusions, chronic toxicity and possibly non‐cell‐autonomous effects were proposed (Edbauer and Haass 2016). For example, cell‐to‐cell transmission of cytoplasmic Tau and α‐synuclein aggregates results in stereotypic spreading during the progression of Alzheimer's and Parkinson's disease, respectively (Jucker and Walker 2018). Thus, I analyzed non‐cell‐autonomous effects of DPRs as a potential trigger of TDP‐43 pathology using co‐culture assays and antibody treatment experiments to inhibit poly-GA cell‐ to‐cell transmission. Importantly, I discovered that poly‐GA transmission inhibits the proteasome even in neighboring cells. Activating the proteasome pharmacologically (with the PDE4 inhibitor rolipram) or genetically (using PSMD11 overexpression) rescues nuclear import of the TDP‐43 NLS reporter. Therefore, I hypothesized that poly‐GA inclusions may block the import of TDP‐43 via ubiquitination directly within its NLS, which accumulate through impaired proteostasis. I could show that mutagenizing lysine 95 in the NLS largely prevents ubiquitination of TDP-43 and did not impair basal nuclear import, but completely prevented poly-GA mediated inhibition of TDP-43 import. In contrast, mutation of lysine 84 severely blocks interactions with nuclear import receptors such as importin‐α5/KPNA1, independent of poly-GA expression, which is in consistence with previous reports (Hans, Eckert et al. 2018). Importantly, poly-GA antibodies reduce poly-GA transmission and cytoplasmic TDP-43 mislocalization in HeLa cells and primary neurons, suggesting poly-GA antibodies could be potentially used to treat C9orf72 ALS/FTD. Taken together, my work shows that poly-GA promotes cytoplasmic mislocalization of TDP-43 cell- and non-cell-autonomously due to impaired proteasomal clearance of TDP-43 ubiquitinated within its NLS at lysine 95. My work also indicates that pharmacologically boosting proteasome activity (e.g. by rolipram) or inhibiting poly-GA transmission using antibodies are promising therapeutic approaches for C9orf72 ALS/FTD. Indeed, our group (Zhou, Mareljic et al. 2020) and others (Nguyen, Montrasio et al. 2020) have reported promising results using anti-GA antibodies or vaccination in mouse models confirming my data in vivo.