Mutant GGGGCC RNA prevents YY1 from binding to Fuzzy promoter which stimulates Wnt/β-catenin pathway in C9ALS/FTD

Abstract

The GGGGCC hexanucleotide repeat expansion mutation in the chromosome 9 open reading frame 72 (C9orf72) gene is a major genetic cause of amyotrophic lateral sclerosis and frontotemporal dementia (C9ALS/FTD). In this study, we demonstrate that the zinc finger (ZF) transcriptional regulator Yin Yang 1 (YY1) binds to the promoter region of the planar cell polarity gene Fuzzy to regulate its transcription. We show that YY1 interacts with GGGGCC repeat RNA via its ZF and that this interaction compromises the binding of YY1 to the Fuzzy^YY1 promoter sites, resulting in the downregulation of Fuzzy transcription. The decrease in Fuzzy protein expression in turn activates the canonical Wnt/β-catenin pathway and induces synaptic deficits in C9ALS/FTD neurons. Our findings demonstrate a C9orf72 GGGGCC RNA-initiated perturbation of YY1-Fuzzy transcriptional control that implicates aberrant Wnt/β-catenin signalling in C9ALS/FTD-associated neurodegeneration. This pathogenic cascade provides a potential new target for disease-modifying therapy.

Fig. 1. Expression of mutant GGGGCC RNA causes downregulation of Fuzzy promoter activity.

a, b The Fuzzy/Fuzzy transcript (a) and protein (b) levels were downregulated in C9ALS/FTD patient iPSC-derived spinal motor neurons when compared to the healthy and isogenic control neurons. c Illustration of the Fuzzy promoter luciferase reporter and different GGGGCC expression constructs used in this study. d Overexpression of pAG3-(GGGGCC)₆₆ construct suppressed Fuzzy promoter luciferase activities at Fuzzy^-2732/+574, Fuzzy^-2032/+574 and Fuzzy^-1332/+574 regions, whereas the activity of Fuzzy^-632/+574 luciferase construct was not altered. The (GGGGCC)₆₆-responsive region, Fuzzy^-1332/-633, is in yellow. e Fuzzy promoter activity was downregulated in (GGGGCC)₁₀₆-RO-expressing cells, while such downregulation was restored upon treatment of BIND peptide. f The (GGGGCC)₆₆- and (GGGGCC)₁₀₆-RO-responsive element was further narrowed down to Fuzzy^-1332/-1143 region, which is in red. g Schematic representation of the locations and sequences of Fuzzy^YY1-a (in red font) and Fuzzy^YY1-b (in blue font) sites in Fuzzy^-1332/+574 promoter region. “+1” represents the transcriptional initiation (arrow) site. The Fuzzy^YY1-a and Fuzzy^YY1-b sites were mutated upon introducing a single nucleotide substitution from “T” to “C” (in green font). h, i Mutation of Fuzzy^YY1-a (h), but not Fuzzy^YY1-b (i) abolished the (GGGGCC)₆₆-mediated downregulation of Fuzzy^-1332/+574 promoter activity. One-way ANOVA followed by post hoc Tukey’s test was used for the comparisons between disease and healthy control neurons in panels (a) and (b). One-way ANOVA followed by post hoc Dunnett’s test was used for the comparisons between disease and isogenic control neurons in panels (a) and (b). Two-tailed unpaired Student’s t-test was used in panels (d–f), (h) and (i). The exact P values are listed in Supplementary Table 7. For panels (a) and (b), n = 3 biologically independent experiments. For the rest of the panels, n = 5 biologically independent experiments. Data is presented as mean ± SEM. The illustrations in panels (c) and (g) were created using Adobe Illustrator and BioRender.com, respectively. Source data are provided as a Source Data file.

Fig. 2. GGGGCC RNA perturbs the binding of YY1 to Fuzzy promoter, leading to Fuzzy downregulation.

a The binding of recombinant YY1 protein to Fuzzy^YY1-a DNA is stronger than to Fuzzy^YY1-b DNA. The R² of the YY1–Fuzzy^YY1-a and YY1–Fuzzy^YY1-b binding curves are 0.9915 and 0.9818, respectively. The Fuzzy promoter region that carries one copy of Fuzzy^YY1-a or Fuzzy^YY1-b site was synthesized and used in the EMSA. The sequences of the Cy5-labelled Fuzzy^YY1-a DNA and Fuzzy^YY1-b DNA (each of 20 base pairs in length) probes are listed in Supplementary Table 4. b The ZF3 in YY1 is required for interacting with Fuzzy^YY1-a, while ZF4 mediates the interaction between YY1 and Fuzzy^YY1-b. Both Fuzzy^YY1-a and Fuzzy^YY1-b DNAs were amplified from the endogenous Fuzzy gene promoter. c Mutation of Fuzzy^YY1-a led to downregulation of Fuzzy promoter activity, whereas mutation of Fuzzy^YY1-b caused the opposite upregulation effect. d Knockdown of YY1 led to the inhibition of Fuzzy promoter activity. e Addition of GGGGCC RNA perturbed the formation of protein–DNA complexes between recombinant YY1 protein and Fuzzy^YY1-a/Fuzzy^YY1-b DNA, leading to the increase in levels of free Fuzzy^YY1-a and Fuzzy^YY1-b DNAs. The YY1–Fuzzy^YY1-b binding is more susceptible to GGGGCC RNA interference. f The binding of endogenous YY1 protein to Fuzzy^YY1-a and Fuzzy^YY1-b sites was impaired upon overexpression of the pAG3-(GGGGCC)₆₆ construct. The YY1–Fuzzy^YY1-b interaction was more prone to be disturbed in the presence of mutant GGGGCC RNA. Both Fuzzy^YY1-a and Fuzzy^YY1-b DNAs were amplified from the endogenous Fuzzy promoter. g The increase of the transfection amount of pAG3-(GGGGCC)₆₆ plasmid gradually inhibited Fuzzy promoter activity. One-way ANOVA followed by post hoc Tukey’s test and one-way ANOVA followed by post hoc Dunnett’s test were used in panels (b) and (c), respectively. Two-tailed unpaired Student’s t-test was used in panels (d), (f) and (g). The exact P values are listed in Supplementary Table 7. For panels (c), (d) and (g), n = 5 biologically independent experiments. For the rest of the panels, n = 3 biologically independent experiments. Data is presented as mean ± SEM. Source data are provided as a Source Data file.

Fig. 3. The GGGGCC RNA binds to YY1 protein and recruits it to RNA foci.

a The binding between recombinant His-YY1^full-length protein and (GGGGCC)₈ RNA was detected by EMSA. The R² of the YY1–(GGGGCC)₈ RNA binding curve is 0.9914. The sequence of the Cy5-labelled (GGGGCC)₈ RNA (48 bases in length) probe is listed in Supplementary Table 4. b The fluorescence polarisation assay was performed to demonstrate the binding between recombinant His-YY1^full-length protein and (GGGGCC)₈ RNA. The R² of the YY1–(GGGGCC)₈ RNA binding curve is 0.9881. The sequence of the FAM-labelled (GGGGCC)₈ RNA (50 bases in length) probe is listed in Supplementary Table 4. c The endogenous hnRNP H protein (green) was found co-localised with GGGGCC RNA foci (red) in (GGGGCC)₆₆-expressing SK-N-MC cells. Scale bars: 5 μm. No GGGGCC RNA foci were detected in (GGGGCC)₂-expressing cells or cells that received RNase treatment. d, e The co-localisation between endogenous YY1 protein (green) and GGGGCC RNA foci (red) was detected in (GGGGCC)₆₆-transfected SK-N-MC cells by two anti-YY1 antibodies raised against different epitopes. Scale bars: 5 μm. No GGGGCC RNA foci were detected in (GGGGCC)₂-transfected cells or cells that received RNase treatment. f is the quantification of (c–e). The number of GGGGCC foci-positive cells counted over three independent experiments were: hnRNP H: 180; YY1 (Abcam): 299; YY1 (Proteintech): 182. g, h The formation of GGGGCC RNA foci was detected when pAG3-(GGGGCC)₆₆-MS2, but not pAG3-(GGGGCC)₂-MS2, was co-transfected with MS2CP-YFP in SK-N-MC cells. The recovery of YY1-mCherry (g) and hnRNP H-mCherry (h) fluorescent signals after photobleaching was much less effective in GGGGCC foci formation cells. The circled regions indicate the region of interest (ROI) that underwent photobleaching. The pre-bleach and post-bleach fluorescent signals in the ROI were recorded. n represents the total number of cells examined over three independent experiments. Two-tailed unpaired Mann–Whitney U test was used in panels (g) and (h). The exact P values are listed in Supplementary Table 7. n = 3 biologically independent experiments and data is presented as mean ± SEM. Source data are provided as a Source Data file.

Fig. 4. The YY1^ZF3 plays an important role in YY1–GGGGCC RNA binding.

a The binding affinity of YY1^ZF1+ZF2+ZF3–(GGGGCC)₈ RNA (K_d = 59.01 ± 1.23 nM) was comparable to that of YY1^full-length–(GGGGCC)₈ RNA (K_d = 46.76 ± 3.42 nM; Fig. 3a) and was stronger than that of YY1^ZF1+ZF2–(GGGGCC)₈ RNA (K_d = 129.43 ± 26.23 nM). The R² of the YY1^ZF1+ZF2+ZF3–(GGGGCC)₈ RNA and YY1^ZF1+ZF2–(GGGGCC)₈ RNA binding curves are 0.9871 and 0.9828, respectively. The sequence of the Cy5-labelled (GGGGCC)₈ RNA (48 bases in length) probe is listed in Supplementary Table 4. b The co-localisation between YY1 (green) and GGGGCC RNA foci (red) was diminished when the YY1^ZF3, but not YY1^ZF4, was deleted. Scale bars: 5 μm. c is the quantification of (b). The number of GGGGCC foci-positive and YY1-EGFP co-transfected cells counted over three independent experiments were: YY1^full-length-EGFP: 217; YY1^ZF1+ZF2+ZF3-EGFP: 187; YY1^ZF1+ZF2-EGFP: 106. One-way ANOVA followed by post hoc Tukey’s test was used in (c). The exact P values are listed in Supplementary Table 7. n = 3 biologically independent experiments and data is presented as mean ± SEM. Source data are provided as a Source Data file.

Fig. 5. The YY1 protein is recruited to GGGGCC RNA foci in C9ALS/FTD iPSC-derived spinal motor neurons.

a The endogenous YY1 protein (green) was recruited to the GGGGCC RNA foci (red) in C9ALS/FTD iPSC-derived spinal motor neurons. No such recruitment was detected in isogenic control spinal motor neurons or neurons that received RNase treatment. b is the quantification of Fig. 5a and Supplementary Fig. 8. The number of GGGGCC foci-positive neurons counted over three independent experiments were: C901-06: 127; C901-07: 103; C902-02: 122; C902-03: 152; C904-01: 176; C904-12: 143. Data are presented as mean ± SEM. No GGGGCC foci formation was detected in isogenic control iPSC-derived spinal motor neurons or neurons treated with RNase. Arrows indicate the co-localisation between the endogenous YY1 protein and GGGGCC RNA foci. The cell nuclei were stained with Hoechst 33342 (blue). Scale bars: 2 μm. Source data are provided as a Source Data file.

Fig. 6. YY1 ameliorates the cell death and synaptic defects in C9ALS/FTD models.

a The binding between YY1 and Fuzzy^YY1 was disrupted in C9ALS/FTD iPSC-derived spinal motor neurons, and such disruption was partially restored when YY1 was overexpressed. Overexpression of YY1 also caused the increase in YY1 and Fuzzy^YY1 binding in isogenic control iPSC-derived spinal motor neurons. Both Fuzzy^YY1-a and Fuzzy^YY1-b DNA fragments were amplified from the endogenous Fuzzy promoter. b The level of nascent Fuzzy transcript was reduced in C9ALS/FTD iPSC-derived spinal motor neurons relative to isogenic control neurons. Such reduction was mitigated upon overexpression of YY1. Overexpression of YY1 did not affect nascent Fuzzy transcript level in isogenic control neurons. YY1 overexpression caused the increase of nascent YY1 transcript in both isogenic control and disease neurons. c Overexpression of YY1 suppressed the (GGGGCC)₆₆-induced cell death. Meanwhile, it did not cause a dominant toxic effect in the untransfected or (GGGGCC)₂-expressing cells. d, e The climbing (d) and survival (e) deficits of (GGGGCC)₃₆ flies were alleviated when Pho was overexpressed. Overexpression of Pho did not affect the climbing ability and survival probability of (GGGGCC)₃ flies. n represents the total number of flies examined over three independent experiments. The genotypes of flies are listed in Supplementary Table 6. f–i Bassoon (f) and Homer1 (h) puncta numbers were reduced in C9ALS/FTD iPSC-derived spinal motor neurons. Such reduction was alleviated upon overexpression of YY1. The number of Bassoon (f) and Homer1 (h) puncta was not affected in isogenic control iPSC-derived spinal motor neurons when YY1 was overexpressed. Scale bars: 20 μm. g and i are the quantifications of f and h, respectively. n represents the total number of neurites examined over three independent experiments. One-way ANOVA followed by post hoc Tukey’s test was used in panels (a–d), (g) and (i). Log-rank (Mantel-Cox) test was used in (e). The exact P values are listed in Supplementary Table 7. n = 3 biologically independent experiments and data is presented as mean ± SEM. Source data are provided as a Source Data file.

Fig. 7. Induction of Wnt/β-catenin target gene PITX2 contributes to cell death and synaptic defects in C9ALS/FTD models.

a When compared to the unexpanded (GGGGCC)₂-expressing cells, overexpression of pAG3-(GGGGCC)₆₆ in SK-N-MC cells stimulated the TOP flash, but not FOP flash luciferase activity. b, c The transcript levels of CCND1, FOSL1, and PITX2, but not other Wnt-responsive genes, were significantly upregulated in C9ALS/FTD patient iPSC-derived spinal motor neurons when compared to the healthy (b) and isogenic (c) control neurons. d The protein level of PITX2 was upregulated in C9ALS/FTD iPSC-derived spinal motor neurons compared to the healthy and isogenic control neurons. e The (GGGGCC)₆₆-induced cell death was suppressed upon knockdown of PITX2. No dominant cytotoxic effect was detected in the untransfected or (GGGGCC)₂-expressing cells when PITX2 was knocked down. f, g Knockdown of Ptx1 ameliorated the climbing (f) and survival (g) deficits of (GGGGCC)₃₆ flies, and it did not affect the climbing ability and survival probability of (GGGGCC)₃ flies. n represents the total number of flies examined over three independent experiments. The genotypes of flies are listed in Supplementary Table 6. h–k Knockdown of PITX2 rescued the reduction of Bassoon (h) and Homer1 (j) puncta numbers in C9ALS/FTD iPSC-derived spinal motor neurons. The number of Bassoon (h) and Homer1 (j) puncta was not affected in isogenic control iPSC-derived spinal motor neurons when PITX2 was knocked down. Scale bars: 20 μm. i and k are the quantifications of h and j, respectively. n represents the total number of neurites examined over three independent experiments. Two-tailed unpaired Student’s t-test was used in panels (a–c). One-way ANOVA followed by post hoc Tukey’s test was used in panels (d–f), (i) and (k). One-way ANOVA followed by post hoc Dunnett’s test was used for the comparison between disease and isogenic control neurons in (d). Log-rank (Mantel-Cox) test was used in panel (g). The exact P values are listed in Supplementary Table 7. n = 3 biologically independent experiments and data is presented as mean ± SEM. The illustration in panel 7a was created using Adobe Illustrator. Source data are provided as a Source Data file.

Fig. 8. Overexpression of BIND rescues dysregulation of YY1–Fuzzy–PITX2 signalling axis and suppresses synaptic defects in C9ALS/FTD spinal motor neurons.

a Overexpression of BIND rescued the defects in binding between YY1 and both Fuzzy^YY1-a and Fuzzy^YY1-b promoter sites. Overexpression of BIND-S failed to elicit a similar rescuing effect. Both Fuzzy^YY1-a and Fuzzy^YY1-b DNA fragments were amplified from the endogenous Fuzzy promoter. b, c Overexpression of BIND, but not BIND-S, restored the Fuzzy reduction (b) and PITX2 induction (c) in C9ALS/FTD spinal motor neurons. d–g Overexpression of BIND protein suppressed the reduction of Bassoon (e) and Homer1 (g) puncta numbers in C9ALS/FTD iPSC-derived spinal motor neurons, whereas BIND-S overexpression did not cause a similar suppression effect. Overexpression of neither BIND nor BIND-S affects the Bassoon (e) and Homer1 (g) puncta numbers in the isogenic control iPSC-derived spinal motor neurons. Scale bars: 20 μm. e and g are the quantifications of d and f, respectively. n represents the total number of neurites examined over three independent experiments. Two-tailed unpaired Student’s t-test was used in panel (a). One-way ANOVA followed by post hoc Tukey’s test was used in panels (b), (c), (e) and (g). The exact P values are listed in Supplementary Table 7. n = 3 biologically independent experiments and data is presented as mean ± SEM. Source data are provided as a Source Data file.

Fig. 9. Graphical summary of the current study.

In normal spinal motor neurons, YY1 binds to Fuzzy promoter at Fuzzy^YY1-a and Fuzzy^YY1-b sites to control the normal level of Fuzzy protein. The canonical Wnt/β-catenin pathway is not activated, and no neurodegeneration occurs. In C9ALS/FTD spinal motor neurons, YY1 binds to GGGGCC RNA. The binding of YY1 to the Fuzzy promoter is impaired, followed by the downregulation of the Fuzzy protein level. The Wnt/β-catenin pathway is activated. The induction of PITX2 proteins consequently contributes to the degeneration of disease neurons. The illustration was created using BioRender.com.