GGGGCC (G4C2) repeat expansion in the first intron of C9ORF72 is the most common genetic cause of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) [3, 9]. A key pathological hallmark of C9ORF72-related ALS/FTD is the accumulation of dipeptide repeat (DPR) proteins synthesized from both sense and antisense repeat RNAs in affected neurons [2, 8, 14]. The toxicity of poly(GA), poly(GR) and some other DPR proteins in cell and animal models has been well documented, including in mouse models. However, it is poly(GR) whose distribution in patient brains seems to best correlate with neurodegeneration [10, 11].

How DPR proteins are synthesized in human C9ORF72 neurons is unknown. One widely cited hypothesis is repeat-associated non-AUG (RAN) translation, which posits a novel mechanism by which ribosomes directly initiate translation on expanded repeats [13]. Over time, the term “RAN translation” has evolved to also refer to the phenomenon, that various expanded repeat sequences such as G4C2 repeats can be translated in different frames. However, whether this truly represents a novel translation mechanism has not yet been vigorously tested in human neurons.

To investigate how DPR proteins are synthesized in C9ORF72 human neurons, we used CRISPR-Cas9 technology to generate a homozygous deletion in the first intron of C9ORF72, 5′ to the G4C2 repeats (Figs. 1a, S1 and S2) to assess the effect of this deletion on DPR production. This deletion contains a CUG start codon that was shown to play a role in poly(GA) production in reporter constructs [4, 12]. The induced pluripotent stem cell (iPSC) line used in this experiment contains ~ 1000 copies of the G4C2 repeats [1]. We selected two iPSC lines containing a deletion in intron 1 and differentiated them and the parental iPSC line into ChAT-positive motor neurons (Figs. 1b and S3) as described [5]. This intronic deletion did not compromise the splicing of the intron containing the expanded repeats, as the levels of mature mRNAs for C9ORF72 variants 1–3 and antisense repeat RNA did not decrease (Fig. 1c–f).

Fig. 1
figure 1

Poly(GA) production in C9ORF72 iPSC-derived motor neurons. a Schematic of the location of the CRISPR-Cas9-mediated deletion in the first intron of C9ORF72. b Schematic of the motor neurons differentiation protocol. Images show iPSC colonies, neurospheres, and 10-day-old ChAT+ (red) and MAP2+ (green) neurons (DAPI in blue). c–g Neurons from parental and intron deletion lines were analyzed for expression of C9ORF72-V1, C9ORF72-V2, C9ORF72-V3, C9ORF72-antisense mRNAs, and C9ORF72 pre-mRNA, n = 3 independent differentiations. h Schematic of the C9ORF72 locus in four types of iPSC-derived neurons. i–n Poly(GA), poly(GR), and poly(GP) levels in 1.5-month-old (n = 3 independent differentiations) and 3-month-old (n = 3 or 6 independent differentiations) control, parental, and deletion motor neurons measured by MSD immunoassays. o Relative expression levels of TARDBP and HSPA1B in parental and intron deletion lines, n = 3 independent differentiations. In all panels, each data point represents one independent differentiation. Values are mean ± SEM. *p < 0.05, **p < 0.01 (c–g and o, one-way ANOVA; i–n, Welch’s t-test)

Full size image

To determine whether this deletion affects DPR production, in addition to the intron deletion and the parental iPSC lines described above, an iPSC line derived from a control subject [1] and an isogenic line in which expanded G4C2 repeats have been deleted by CRISPR-Cas9 [5] (Fig. 1h) were differentiated three times into motor neurons. After 1.5 months, motor neuron cultures were collected and analyzed for poly(GA) content, in a blinded manner, with a new Meso Scale Discovery (MSD) immunoassay developed at Biogen Inc (see Suppl. Information). We found that poly(GA) production was abolished in neurons containing the intronic deletion (Fig. 1i). The absence of poly(GA) production was not due to the absence of the mRNA translation template, as repeat-containing introns were still spliced (Fig. 1c, e) and the level of the pre-mRNA was not significantly decreased (Fig. 1g). More importantly, both poly(GR) and poly(GP) were still produced (Fig. 1j, k). To confirm these results, we differentiated again the no-repeats, the C9ORF72 parental and the two intronic deletion lines 6 or 3 more times into 3-month-old motor neuron cultures. Once more, blinded poly(GA) measurements confirmed the absence of poly(GA) in intronic deletion lines (Fig. 1l), while poly(GR) and poly(GP) production was not decreased (Fig. 1m, n).

Next, we used this model to investigate the contribution of poly(GA) to molecular phenotypes found in C9ORF72 iPSC-derived neurons or patient brain tissues. Upregulation of the expression of heat shock proteins and TAR DNA binding protein 43 (TDP-43) have been reported in ALS/FTD patient brain tissues [6, 7]. Here, we show that both HSPA1B and TARDBP mRNA are upregulated in 4-month-old motor neuron cultures (Fig. 1o). However, in the absence of poly(GA), these molecular phenotypes were not rescued, suggesting that, at least for these phenotypes in this experimental system, poly(GA) was not a key toxic DPR protein.

Our results show that in C9ORF72 patient neurons, ~ 1000 copies of G4C2 repeats alone are not sufficient to direct ribosome entry to initiate translation of poly(GA) frame. What remains to be determined is whether other cis elements within the first intron initiate the synthesis of poly(GR) and other DPR proteins or whether their production results from bona fide “RAN translation” in C9ORF72 patient neurons. These mechanistic investigations will help in the design of therapeutic approaches that aim to decrease DPR protein production.