Poly(ADP-ribose) promotes toxicity of C9ORF72 arginine-rich dipeptide repeat proteins

Abstract

Arginine-rich dipeptide repeat proteins (R-DPRs), abnormal translational products of a GGGGCC hexanucleotide repeat expansion in C9ORF72, play a critical role in C9ORF72-related amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD), the most common genetic form of the disorders (c9ALS/FTD). R-DPRs form liquid condensates in vitro, induce stress granule formation in cultured cells, aggregate, and sometimes coaggregate with TDP-43 in postmortem tissue from patients with c9ALS/FTD. However, how these processes are regulated is unclear. Here, we show that loss of poly(ADP-ribose) (PAR) suppresses neurodegeneration in c9ALS/FTD fly models and neurons differentiated from patient-derived induced pluripotent stem cells. Mechanistically, PAR induces R-DPR condensation and promotes R-DPR-induced stress granule formation and TDP-43 aggregation. Moreover, PAR associates with insoluble R-DPR and TDP-43 in postmortem tissue from patients. These findings identified PAR as a promoter of R-DPR toxicity and thus a potential target for treating c9ALS/FTD.

Fig. 1.. Loss of Parp/PARP1 function suppresses neurodegeneration in fly models of c9ALS/FTD.

(A) Fly eyes expressing (G₄C₂)₃₀, under the control of GMRGAL4, with or without parp RNAi (parp i) or PARG cDNA. Right: Fly eye defects were scored using a published method (79). Briefly, points were added if there was a complete loss of interommatidial bristles, necrotic patches, retinal collapse, loss of ommatidial structure, and/or depigmentation of the eye. (B) Fly eyes expressing (GR)₃₆ or (PR)₃₆ with or without parp RNAi or PARG cDNA. Quantification of the eye scores at the bottom. (C) Eyes of flies expressing (G₄C₂)₃₀, (GR)₃₆, or (PR)₃₆ and fed with DMSO or PJ34. Quantified at the bottom. (D) Flight assays (see Materials and Methods). Oneway analysis of variance (ANOVA) with Dunnett’s tests (A, B, and D) and Student’s t tests (C). **P < 0.01 and ****P < 0.0001. Means ± SEM.

Fig. 2.. Loss of Parp/PARP1 function suppresses neurodegeneration in c9ALS/FTD iPSNs.

(A) Control line #1 (Ctrl#1) or c9ALS/FTD line #1 (c9#1) iPSNs pretreated with DMSO or niraparib (nira), treated with glutamate, and stained with propidium iodide (PI; red) and NucBlue (Nuc; blue). (B) Quantification of glutamate toxicity assays on Ctrl#1 and c9#1 iPSNs pretreated with DMSO, nira, or veliparib (veli). For each condition, three visual frames were imaged, and total numbers of cells were counted. The experiments were biologically replicated twice. (C and D) Immunoblots of lysates from three age and sexmatched pairs (C) and an isogenic pair (D) of Ctrl and c9 iPSNs. In (C), three replicates of three iPSN pairs (n = 3, total m = 9 data points) were used for statistical analyses. (E) LDH assays on Ctrl#1 iPSNs pretreated with DMSO or 5 μM nira or veli and treated with 10 μM DPRs. χ² tests (B), Student’s t tests (C and D), and oneway ANOVA with Dunnett’s tests (E). ns, not significant. *P < 0.05, **P < 0.01, and ****P < 0.0001. Means ± SEM.

Fig. 3.. PAR induces R-DPR condensation in vitro.

(A) Co-IP of PAR and DPRs in HEK293T cells. (B) Dot blot binding assays of PAR and DPRs. (C) TAMRA–R-DPR (red) mixed with PAR (10 to 50 μM MAR equivalent) or MAR. (D and E) FRAP assays on PAR and TAMRA–R-DPR (red) condensates. Yellow circles indicate bleached and analyzed areas. Total R-DPRs (1 μM) were labeled. Buffer for (C to E): 61.5 mM K₂HPO₄ and 38.5 mM KH₂PO₄

Fig. 4.. Loss of PARP1 activity suppresses R-DPR–induced SG formation.

(A) Control (Ctrl) or PARP1 KO (ΔPARP1) HEK293T cells transfected with R-DPR–GFP (green) were immunofluorescently stained for G3BP1 (magenta), Ataxin-2 (blue), and DAPI 24 hours after transfection. Arrowheads indicate SGs. Note that (GR)₅₀, but not (PR)₅₀, localizes to SGs. (B and C) Quantification of R-DPR–GFP–expressing HEK293T cells that exhibit SGs. (B) Control (Ctrl) or PARP1 KO (ΔPARP1) cells. (C) Ctrl cells treated with DMSO, nira, or veli. DMSO or the PARP1 inhibitors were added to the cells 2 hours before transfection and remained in the culture media until fixation or cell lysis. For each condition, 20 to 30 visual fields from three biological replicates were randomly selected and imaged. The total numbers of GFP-positive cells with or without SGs were separately counted. χ² tests. ***P < 0.001 and ****P < 0.0001. (D and E) Western blots on HEK293T cells overexpressing R-DPR–GFP for 24 hours. (D) Ctrl or ΔPARP1 cells. (E) Ctrl cells treated with DMSO, nira, or veli.

Fig. 5.. PAR promotes G3BP1 condensation and interaction with poly(GR).

(A) Phase diagram and (B) images of Alexa 488–labeled recombinant G3BP1 (G3BP1-A488; green) mixed with PAR (2.5 to 50 μM MAR equivalents) or 50 μM MAR. (C and D) FRAP assays on G3BP1-A488 (green) and PARCy5 (blue) condensates. Yellow circles indicate bleached and analyzed areas. (E) Co-IP of (GR)₅₀-GFP and endogenous G3BP1 in control (Ctrl) or PARP1 KO (ΔPARP1) HEK293T cells or in Ctrl HEK293T cells with or without the PARG inhibitor PDD00017273 (PARGin+ or PARGin−, respectively) in the lysis and IP buffer. Quantified in (F). Student’s t tests. **P < 0.01; ****P < 0.0001. Means ± SEM. (G) Phase diagram and (H and I) images of TAMRA-(GR)20 (magenta) mixed with G3BP1-A488 (green) and PAR (2.5 to 25 μM MAR equivalents) or 25 μM MAR. Total molecules (1 μM) were fluorescently labeled. Buffer for condensation assays: PBS.

Fig. 6.. PAR promotes poly(GR)-induced TDP-43 aggregation.

(A) Phase diagram and (B) images of TAMRA-(GR)₂₀ (magenta) mixed with His₆-SUMO–TDP-43–A488 (green) and PAR (2.5 to 25 μM MAR equivalents) or 25 μM MAR. Total proteins were fluorescently labeled. Buffer: PBS. (C) The TDP-43 precipitation assay. (D) Quantification of A488 and TAMRA intensities. (E) The turbidity assay. (F) Quantification of increases in 395-nm absorbance. (G and H) Immunoblot of RIPA and urea fractions of control (Ctrl) or PARP1 KO (ΔPARP1) HEK293T cell lysate (G) or Ctrl HEK293T cell lysate with or without the PARG inhibitor PDD00017273 (PARGin+ or PARGin−, respectively; H), treated with (GR)₂₀. Quantification reported below. One-way ANOVA with Tukey’s tests (D and F) and Student’s t tests (G and H). *P < 0.05, **P < 0.01, and ****P < 0.0001. Means ± SEM.

Fig. 7.. PAR and protein aggregation in vivo.

(A) Dot blots of RIPA and urea fractions of head extract from flies expressing R-DPRs without (Ctrl) or with coexpressing parp RNAi (parp i) #1. (B) Quantification of (A) and (D). (C) Fly eyes expressing 44 G₄C₂ repeats with a C-terminal GFP tag in the poly(GR) reading frame [(G₄C₂)₄₄- GR-GFP] with or without coexpressing parp RNAi. (D) Immunoblot of RIPA and urea fractions of head extract from flies expressing (G₄C₂)₄₄-GR-GFP with or without coexpressing parp i #1. Student’s t tests. *P < 0.05 and ****P < 0.0001. Means ± SEM.(E) c9ALS/FTD patient information. (F) Association of insoluble PAR with insoluble poly(GR) (left) or insoluble TDP-43 phosphorylated at S409/410 (pTDP-43) (right) in frontal cortex lysates from C9ORF72-repeat expansion carriers with frontotemporal lobar degeneration (FTLD) with or without motor neuron disease (MND). CI, confidence interval. Multiple linear regression.