Hexanucleotide expansions in the C9ORF72 gene are the most common genetic cause of amyotrophic lateral sclerosis and frontotemporal dementia, but how the dipeptide repeats (DPRs) translated from the expansions incite disease still puzzles researchers. A study published March 16 in the EMBO Journal proposes a mechanistic link between the dipeptides and cytoplasmic aggregates of TDP-43—the pathology thought to trigger neurodegeneration.

  • Poly-GA dipeptide repeats travel between neurons, at least in culture.
  • The peptides derail the work of the proteasome.
  • Proteasome deficits trigger TDP-43 mislocalization, aggregation.

Researchers led by Dieter Edbauer at the German Center for Neurodegenerative Diseases in Munich reported that poly-GA peptides travel between neurons in cell culture. In both donor and recipient neurons, the dipeptides disrupted proteasomal function. This, in turn, derailed the normal trafficking of TDP-43 into the nucleus, rendering this protein prone to aggregation in the cytoplasm. The findings connect DPRs and mislocalized TDP-43, two pathologies rarely spotted cohabitating in the same cell.

Using a noncanonical translation mechanism called repeat associated non-ATG (RAN) translation, DPRs are translated from C9ORF72 hexanucleotide transcripts in six reading frames, producing five unique DPRs. Poly-GA is the most abundant. Previously, Edbauer and colleagues reported that poly-GA dipeptides triggered mislocalization of TDP-43, preventing its entry into the nucleus (Khosravi et al., 2017). While this offered a potential causal link between DPRs and TDP-43, it did not explain the observation that in patient tissue, cytoplasmic aggregates of TDP-43 are usually spotted in different cells than are DPR inclusions. Might DPRs somehow bungle TDP-43’s nuclear transport from afar?

To address this question, first author Bahram Khosravi and colleagues set up a neuronal co-culture. They transduced rat primary hippocampal neurons with a gene encoding 175 poly-GA dipeptide repeats linked to green fluorescent protein. After growing the “donor” cells on coverslips for four days, they washed them and positioned them next to a coverslip of untreated, “recipient” neurons that did not themselves express the poly-GA dipeptide. Four days later, the researchers observed poly-GA aggregates in both the donor and recipient cells. What’s more, they found cytoplasmic TDP-43 in both donor and recipient cells, even in cells in which the fluorescent poly-GA aggregates were undetectable. Addition of a poly-GA antibody to the culture system blocked transmission of poly-GA aggregates as well as TDP-43 mislocalization, suggesting that the dipeptide repeats were directly responsible for TDP-43 pathology, even at exceedingly low levels.

Traveling Trouble. Compared with “donor” cells transduced with GFP alone (top left panel), donor cells transduced with Poly-GA-GFP (bottom left) had GA aggregates (green) and some TDP-43 excluded from the nucleus (red). The same effects occurred in non-transduced receiver cells (right panels). [Courtesy of Khosravi et al., EMBO Journal, 2020.]

Poly-GA pumped from neighboring cells also coaxed the formation of cytoplasmic aggregates of TDP-43 in cells expressing a version of the protein lacking a nuclear localization sequence (NLS). Again, poly-GA antibodies blocked this effect. Together, the findings suggested that poly-GA dipeptides instigate the mislocalization and aggregation of TDP-43 in the cytoplasm, in the cells expressing the dipeptides as well as in cells stationed nearby.

How could poly-GA be doing this? In a series of experiments, the researchers zeroed in on the proteasome. Previous structural studies had spotted poly-GA dipeptides gumming up proteasomes (Feb 2018 news). Using a reporter of proteasome function, the researchers confirmed that indeed, poly-GA dipeptides not only derailed proteasome function in the cells expressing the dipeptides, but also in receiver cells nearby in the co-culture. This proteasomal shutdown created a backlog of proteasome substrates—in particular, ubiquitinated TDP-43—in the cytoplasm. Interestingly, the researchers also found that poly-GA triggered a buildup of TDP-43 ubiquitinated on lysine-95—a residue smack dab in the middle of the nuclear localization sequence (NLS). Ubiquitination on this residue directly blocked TDP-43 from interacting with nuclear import machinery. Treating cells with the proteasome activator rolipram countered poly-GA’s promotion of TDP-43 mislocalization and aggregation.

Missing Link? Poly-GA released by a neighboring cell gums up a neuron’s proteasome, and TDP-43 piles up in the cytoplasm. Ubiquitination on lysine-95 blocks TDP-43 from entering the nucleus. [Courtesy of Khosravi et al., EMBO Journal, 2020.]

In all, the findings suggested that poly-GA dipeptides can clog the proteasome in nearby neurons. This would lead to a buildup of ubiquitinated TDP-43 in the cytoplasm. TDP-43 is not degraded by the weakened proteasome, and is also is blocked from entering the nucleus because of ubiquitination within its NLS.

The researchers speculate that in the human brain, some neurons, such as those in the cerebellum, are more adept at the noncanonical translation required to generate poly-GA peptides, but less susceptible to their toxicity. Others, including motor neurons, are less efficient at noncanonical translation and more susceptible to proteasomal inhibition triggered by even small amounts of traveling poly-GA. This could explain the different neuronal populations affected by DPR inclusions and TDP-43 pathology in patients, the researchers proposed.

The researchers found that only poly-GA, not other dipeptide repeats translated from different reading frames of C9ORF72, inhibited the proteasome. However, they speculate that RNA foci formed by C9ORF72 transcripts, as well as other DPRs translated from them, likely exert additional toxic effects, such as inhibiting nucleocytoplasmic transport.

The findings cast proteasomal activators, as well as anti-poly-GA antibodies, as potential therapeutic candidates.—Jessica Shugart


  1. Much study on C9ORF72-ALS has centered on toxicity mediated by dipeptide repeat proteins translated from the GGGGCC-repeat expansion. Despite this work, a fundamental question has remained over the relative absence of dipeptide-repeat protein pathology in postmortem tissue and the lack of correlation between dipeptide-repeat protein pathology and the presence of TDP-43 pathology. Cytoplasmic TDP-43 proteinaceous inclusions remain the only molecular marker that predicts neuronal loss. Moreover, clinically and pathologically, neurotoxicity spreads between contiguous areas of the nervous system and this, too, is unexplained.

    This paper from the Edbauer lab offers a proposal—that poly-GA dipeptide-repeat proteins can induce TDP-43 pathology via inhibition of proteasome activity; and that poly-GA proteins can transmit disease between adjacent cells such that a single cell containing a poly-GA protein aggregate may induce toxicity in all neighboring cells. The authors suggest that the lack of correlation between observed postmortem poly-GA and TDP-43 pathology is a result of a third independent variable, that is, baseline proteasome activity. It follows that in areas of the brain such as the cerebellum there is high baseline proteasome activity, so despite plentiful poly-GA pathology the proteasome is not successfully inhibited and TDP-43 pathology is rare. Alternatively, in spinal motor neurons, where baseline proteasome activity is low, a relatively low abundance of poly-GA pathology is sufficient to induce TDP-43 pathology and therefore neurotoxicity. 

    The authors have provided detailed mechanistic insight into this process. This included identification of a key K95-TDP-43 residue that is ubiquitinated in the presence of poly-GA mediated proteasome inhibition, resulting in impaired NLS function and thus cytoplasmic mislocalization of TDP-43.

    This is a well-thought-out hypothesis. Proof will require an in vivo demonstration that inhibition of poly-GA transmission can slow the development of TDP-43 pathology and clinical progression. Arguably, a suitable animal model for such an experiment does not exist currently. 

    Certain other questions remain, as well: poly-GA is absent in the majority of ALS patients who do not have a C9ORF72-expansion, but the clinical progression of C9ORF72-ALS is not significantly different to non-C9ORF72-ALS. How does this mechanism fit into a non-C9ORF72-ALS model? Finally, poly-GA pathology has been demonstrated years in advance of disease onset and the development of TDP-43 pathology (Vatsavayai et al., 2016). What explains the delay in development of TDP-43 mislocalization which preceeds disease onset?


    . Timing and significance of pathological features in C9orf72 expansion-associated frontotemporal dementia. Brain. 2016 Dec;139(Pt 12):3202-3216. Epub 2016 Oct 22 PubMed.

  2. Accumulating evidence suggests that the transmission of glycine-alanine (GA) dipeptide repeats (DPRs) could be a relevant phenomenon in C9ORF72 ALS/FTD. This event has now been shown in four independent cell culture studies (Chang et al., 2016; Flores et al., 2016; Westergard et al., 2016; Zhou et al., 2017) and we also recently reported it in vivo using different neuronal populations of the fly brain (Morón-Oset et al., 2019). However, what exactly this transmission entails for the disease remains unclear.

    In this elegant study, the Edbauer lab provides compelling evidence that transmission of GA175-eGFP between primary rat neurons or Hela cells causes accumulation of cytoplasmatic TDP-43, which is a classical feature of ALS and FTD (Neumann et al., 2006). In agreement with a recent paper showing the close association between GA and the proteasome in cell culture (Guo et al., 2018), this novel study also shows that GA-mediated proteasome inhibition in both donor and recipient cells is key to cause cytoplasmatic retention of TDP-43, which can be rescued by activating the proteasome with the drug rolipram. These are very interesting findings and pose the question  whether this is a specific event for TDP-43 or whether GA would have a similar effect for other proteasome substrates. In addition, while it is still somewhat unclear whether GA prevents the degradation of TDP-43 during its physiological turnover or whether it somehow enhances its nuclear export, the retention of a few ubiquitinated TDP-43 molecules in the cytoplasm could suffice to create the first seeds that would then go on to cause the vast accumulation of TDP-43 aggregates in the cytoplasm and lack of nuclear TDP-43 typically observed in ALS and FTD patients at autopsy. Hence the great relevance of this study. It will be interesting to test whether this cascade of events can also occur under in vivo conditions, as well as whether it can lead to neuronal dysfunction and subsequent death.


    . The Glycine-Alanine Dipeptide Repeat from C9orf72 Hexanucleotide Expansions Forms Toxic Amyloids Possessing Cell-to-Cell Transmission Properties. J Biol Chem. 2016 Mar 4;291(10):4903-11. Epub 2016 Jan 14 PubMed.

    . Distinct C9orf72-Associated Dipeptide Repeat Structures Correlate with Neuronal Toxicity. PLoS One. 2016;11(10):e0165084. Epub 2016 Oct 24 PubMed.

    . In Situ Structure of Neuronal C9orf72 Poly-GA Aggregates Reveals Proteasome Recruitment. Cell. 2018 Feb 8;172(4):696-705.e12. Epub 2018 Feb 1 PubMed.

    . Glycine-alanine dipeptide repeats spread rapidly in a repeat length- and age-dependent manner in the fly brain. Acta Neuropathol Commun. 2019 Dec 16;7(1):209. PubMed.

    . Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Science. 2006 Oct 6;314(5796):130-3. PubMed.

    . Cell-to-Cell Transmission of Dipeptide Repeat Proteins Linked to C9orf72-ALS/FTD. Cell Rep. 2016 Oct 11;17(3):645-652. PubMed.

    . Antibodies inhibit transmission and aggregation of C9orf72 poly-GA dipeptide repeat proteins. EMBO Mol Med. 2017 May;9(5):687-702. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Are ALS Dipeptide Repeat Ribbons Entangling Proteasomes?

Paper Citations

  1. . Cytoplasmic poly-GA aggregates impair nuclear import of TDP-43 in C9orf72 ALS/FTLD. Hum Mol Genet. 2017 Feb 15;26(4):790-800. PubMed.

Further Reading

Primary Papers

  1. . Cell-to-cell transmission of C9orf72 poly-(Gly-Ala) triggers key features of ALS/FTD. EMBO J. 2020 Apr 15;39(8):e102811. Epub 2020 Mar 16 PubMed.