RuvBL1/2 reduce toxic dipeptide repeat protein burden in multiple models of C9orf72-ALS/FTD

Abstract

A G4C2 hexanucleotide repeat expansion in C9orf72 is the most common cause of amyotrophic lateral sclerosis and frontotemporal dementia (C9ALS/FTD). Bidirectional transcription and subsequent repeat-associated non-AUG (RAN) translation of sense and antisense transcripts leads to the formation of five dipeptide repeat (DPR) proteins. These DPRs are toxic in a wide range of cell and animal models. Therefore, decreasing RAN-DPRs may be of therapeutic benefit in the context of C9ALS/FTD. In this study, we found that C9ALS/FTD patients have reduced expression of the AAA+ family members RuvBL1 and RuvBL2, which have both been implicated in aggregate clearance. We report that overexpression of RuvBL1, but to a greater extent RuvBL2, reduced C9orf72-associated DPRs in a range of in vitro systems including cell lines, primary neurons from the C9-500 transgenic mouse model, and patient-derived iPSC motor neurons. In vivo, we further demonstrated that RuvBL2 overexpression and consequent DPR reduction in our Drosophila model was sufficient to rescue a number of DPR-related motor phenotypes. Thus, modulating RuvBL levels to reduce DPRs may be of therapeutic potential in C9ALS/FTD.

Figure S1.. Schematic illustration of codon-optimised AUG-driven V5-tagged synthetic dipeptide repeat constructs.

(A, B, C) The expression cassette from plasmids containing V5-tagged 100x repeating units (non-G4C2) of GA (A), GR (B), and PR (C) are depicted, with the CMV promoter, Kozak-ATG start site, and stop codons also shown.

Figure 1.. RuvBL overexpression reduces C9orf72-associated dipeptide repeat (DPR) proteins in vitro.

(A, B, C, D) HeLa cells transfected with empty vector control (Ctrl), FLAG-RuvBL1, or HA-RuvBL2 (A, B, C, D) were co-transfected with empty vector (A) or AUG-driven synthetic codon-optimised V5-tagged 100 repeat poly(GA) (B), poly(GR) (C), or poly(PR) (D) DPR expressing constructs. RuvBL overexpression was confirmed via immunoblot with GAPDH indicating equal loading of samples. Levels of V5-tagged DPRs were determined via dot-blot. DPR levels were normalised to GAPDH and plotted relative to empty vector transfected control (mean ± SEM; one-way ANOVA with Tukey post-test: *P ≤ 0.05, **P ≤ 0.005, ***P ≤ 0.001; N = 3 independent experiments). (E) HeLa cells transfected with empty vector control (ev), FLAG-RuvBL1, or HA-RuvBL2 were co-transfected with empty vector or with 45 uninterrupted sense GGGGCC repeats (45xG4C2) with V5-tags in all three reading frames. RuvBL overexpression was confirmed via immunoblot with GAPDH indicating equal loading of samples. Levels of repeat-associated non-AUG translated V5-DPRs were determined via dot-blot. DPR levels were normalised to GAPDH and plotted relative to empty vector transfected control (mean ± SEM; one-way ANOVA with Tukey post-test: *P ≤ 0.05, ***P ≤ 0.001, ****P ≤ 0.0001; N = 4 independent experiments).

Figure S2.. RuvBL overexpression does not affect PR100 dipeptide repeat or EGFP levels.

(A) HeLa cells transfected with 0.5 μg empty vector control (Ctrl), FLAG-RuvBL1, or HA-RuvBL2 were co-transfected with 0.5, 0.25, or 0.125 μg AUG-driven synthetic codon-optimised V5-tagged 100 repeat poly(PR). RuvBL overexpression was confirmed via immunoblot with GAPDH indicating equal loading of samples. Levels of V5-PR100 were determined via dot-blot using anti-V5 antibodies. V5-PR100 levels are plotted relative to the empty vector control-0.5 μg V5-PR100 sample (mean ± SEM; one-way ANOVA with Tukey post-test: ns, non-significant, N = 3 independent experiments). (B) HeLa cells transfected with empty vector control (Ctrl), FLAG-RuvBL1, or HA-RuvBL2 were co-transfected with pEGFP-C2 plasmid. RuvBL and EGFP overexpression was confirmed via immunoblot with GAPDH indicating equal loading of samples. EGFP levels were normalised to GAPDH and plotted relative to empty vector transfected control (mean ± SEM; one-way ANOVA with Tukey post-test: ns, non-significant, N = 3 independent experiments).

Figure S3.. Schematic illustration of the repeat-associated non-AUG-dependent 45xG4C2 repeat construct with 3xV5-tags in all ORFs.

As described previously, concatemerized and annealed G4C2/C4G2x15 DNA oligonucleotides blunted with Mung bean nuclease were cloned into the Klenow-filled EcoRI site of pcDNA3.1 to build pcDNA3.1-45xG4C2. A NotI/XbaI cassette encoding 3xV5-tags and three stop codons in all frames was digested from a synthetic custom-synthesised plasmid (Thermo Fisher Scientific) and subcloned into the NotI/XbaI site downstream of 45xG4C2. Note the absence of canonical Kozak-ATG start codons from the transcription start site leading to repeat-associated non-AUG-dependent translation of C9orf72 sense repeat dipeptide repeat proteins. The frame, codon usage and the resulting poly dipeptide are highlighted in the green box. Sense dipeptide repeats are produced in the three reading frames leading to poly(GA), poly(GP), and poly(GR), each with their own V5-tag.

Figure S4.. Detection of RuvBL1 and RuvBL2 with endogenous antibodies.

(A, B, C, D, E) HeLa cells transfected with empty vector control (Ctrl), FLAG-RuvBL1, or HA-RuvBL2 (A, B, C, D, E) were co-transfected with empty vector (A), AUG-driven synthetic codon-optimised V5-tagged 100 repeat poly(GA) (B), poly(GR) (C), poly(PR) (D), or V5-tagged 45xG4C2 repeats (E). Samples correspond to samples in Fig 1, and were probed with anti-RuvBL1 antibodies (top panels) and anti-RuvBL2 antibodies (bottom panels) to demonstrate the level of overexpression compared with endogenous RuvBL1/2 levels. α-Tubulin was used to demonstrate equal loading.

Figure 2.. C9ALS/FTD patient cells have reduced levels of RuvBL proteins.

(A, B, C) RuvBL1 (left immunoblot) and RuvBL2 (right immunoblot) protein levels from 3 C9orf72-ALS/FTD patient iNPC lines and their age and sex-matched controls were determined by immunoblot. Levels of RuvBL1 (left graph) and RuvBL2 (right graph) were normalised to GAPDH and are shown relative to the age and sex-matched control (mean ± SEM; unpaired t test: *P ≤ 0.05, ***P ≤ 0.001, ****P ≤ 0.0001, ns, non-significant; N = 3 independent experiments). (D, E, F) Expression of RuvBL1 (upper graphs) and RuvBL2 (lower graphs) transcripts were quantified by RT-qPCR using 18S as a housekeeping gene (mean ± SEM, N = 3 independent experiments; unpaired t test: **P ≤ 0.005, ***P < 0.001, ****P < 0.0001).

Figure S5.. Expression of repeat-associated non-AUG dipeptide repeat producing V5-tagged 45xG4C2 repeats does not affect endogenous RuvBL levels.

(A, B) Endogenous RuvBL1 (A) and RuvBL2 (B) levels were determined in HeLa cells transfected with empty vector control (Ctrl) or with 45 uninterrupted sense GGGGCC repeats (V5-45xG4C2) via immunoblot. The presence of V5-tagged dipeptide repeats was confirmed via immunoblot using anti-V5 antibodies. GAPDH was used to indicate equal loading. RuvBL1 and RuvBL2 levels were normalised to GAPDH (mean ± SEM; unpaired t test: ns, non-significant, N = 3 independent experiments).

Figure S6.. Knockdown of RuvBL does not affect dipeptide repeat protein levels in vitro.

HeLa cells treated with control siRNA (siCtrl), RuvBL-targeting siRNA (siRuvBL1), or RuvBL2-targeting siRNA (siRuvBL2) were transfected with empty vector control or 45 uninterrupted sense GGGGCC repeats (V5-45xG4C2). (A, B) RuvBL1 levels (A) and RuvBL2 levels (B) were determined via immuoblot with GAPDH used to indicate equal loading. RuvBL1 and RuvBL2 levels were normalised to GAPDH (mean ± SEM; one-way ANOVA with Tukey post-test: ***P ≤ 0.001, ****P ≤ 0.0001; N = 4 independent experiments). (C) Levels of repeat-associated non-AUG translated poly(GP) dipeptide repeats were determined via MSD-ELISA (one-way ANOVA with Tukey post-test: **P ≤ 0.01, ns, non-significant, N = 4 independent experiments).

Figure 3.. Lentiviral transduction with RuvBL2 reduces C9orf72-associated dipeptide repeats (DPRs) in C9-500 BAC primary cortical neurons.

Primary cortical neurons were extracted from E16.5 WT and C9-500 BAC transgenic (Tg) mouse embryos. At DIV4, Tg neurons were transduced with LV-GFP, LV-RuvBL1, or LV-RuvBL2 at an MOI of 10. At DIV11, proteins were extracted for immunoblot analysis and MSD-ELISA. (A, B, C) Transduction and overexpression of GFP (A), RuvBL1 (B), and RuvBL2 (C) was assessed by immunoblot and quantified relative to Tg non-transduced samples. (D) Levels of poly(GA) DPRs were assessed by MSD-ELISA (mean ± SEM, N = 3 independent embryos; one-way ANOVA with Tukey post-test: *P < 0.05, **P < 0.01, ****P < 0.0001. ns, non-significant). (E) Levels of poly(GP) DPRs were assessed by MSD-ELISA (mean ± SEM, N = 3 independent embryos; one-way ANOVA with Tukey post-test: **P < 0.01, ***P < 0.001, ****P < 0.0001, ns, non-significant).

Figure 4.. Lentiviral transduction with RuvBL2 reduces poly(GA) dipeptide repeats (DPRs) in C9orf72 patient iPSC-derived motor neurons.

(A) A timeline to illustrate the differentiation procedure of the iPSC motor neurons and the timepoint of transduction. iPSC-derived motor neurons from Control (Ctrl: CS14) and C9orf72 patient (ALS-52) were transduced at DIV28 with LV-GFP, LV-RuvBL1, or LV-RuvBL2 at an MOI of 10. 7 d post transduction proteins were extracted for analysis via immunoblot and MSD-ELISA. Transduction and overexpression of RuvBL1. (B, C) RA, Retanoic Acid; CHIR, CHIR99021; Pur, Purmophamine; MN, motor neuron (B) and RuvBL2 (C) were confirmed via immunoblot with GAPDH indicating equal loading. Levels of RuvBL1 or RuvBL2 were quantified relative to Tg GFP-transduced samples. (D) Levels of poly(GA) DPRs were assessed by MSD-ELISA (mean ± SEM, N = 3 independent experiments; one-way ANOVA with Tukey post-test: *P < 0.05, ns, non-significant). (E) Levels of poly(GP) DPRs were assessed by MSD-ELISA (mean ± SEM, N = 3 independent experiments; one-way ANOVA with Tukey post-test: ns, non-significant).

Figure S7.. C9-500 BAC primary cortical neurons have reduced RuvBL2 levels.

(A, B) Endogenous RuvBL1 (A) and RuvBL2 (B) levels were determined in non-transgenic control (NTg) and C9-500 BAC transgenic (Tg) primary cortical neurons extracted from E16.5 mouse embryos. GAPDH was used to indicate equal loading. RuvBL1 and RuvBL2 levels were normalised to GAPDH (mean ± SEM; unpaired t test: ns, non-significant, *P ≤ 0.05, N = 3 independent embryos).

Figure 5.. RuvBL co-expression decreases dipeptide repeat levels in a Drosophila model of C9ALS/FTD.

(A, B, C, D) Drosophila pan-neuronally (nSyb-Gal4) co-expressing UAS-mKate2.CAAX (mKate), UAS-Pontin (Pontin), or UAS-Reptin (Reptin) with either UAS-GA(1020)eGFP (GA1000) (A), UAS-GR(1136)eGFP (GR1000) (B), UAS-PR(1100)eGFP (PR1000) (C), or UAS-PA(1024)eGFP (P1000) (D) were aged to 7 DPE before heads were taken for protein extraction. Dipeptide repeats were analysed by MSD-ELISA and are presented relative to the mKate control (mean ± SEM, N = 3 [PA1000], 4 [GA1000 and PR1000], or 5 [GR1000] independent experiments; one-way ANOVA with Tukey post-test: *P < 0.05, **P < 0.01, ***P < 0.001, ns, non-significant).

Figure S8.. A schematic timeline illustrating the different procedures for Drosophila experiments.

Proteins were extracted at 7 d post eclosion from Drosophila heads for dipeptide repeat analysis via MSD-ELISA. Negative geotaxis (climbing) assays were performed at 7 and 14 DPE on remaining flies. Individual flies were placed into a TriKinetic Drosophila Activity Monitor (DAM5M) system after the day 14 climbing analysis and left to acclimatise for 12 h. Activity was then recorded over a 24-h period. IR, infrared.

Figure 6.. RuvBL2 co-expression rescues age related motor impairments in Drosophila pan-neuronally expressing dipeptide repeats.

(A, B, C, D) The vertical distance climbed 5 s after startle-induced negative geotaxis was recorded in Drosophila pan-neuronally (nSyb-Gal4) co-expressing mKate or Reptin with either mCD8-GFP control (A), PA1000 (B), GR1000 (C), or PR1000 (D), at 7 and 14 DPE. The minimum number of flies in any one group was eight (mean ± SEM, flies were from at least three independent crosses per genotype; one-way ANOVA with Tukey post-test: **P < 0.01, ns, non-significant). (E) The activity of Drosophila pan-neuronally (nSyb-Gal4) co-expressing mKate or Reptin with either mCD8-GFP control, PA1000, GR1000, or PR1000, was assessed at 14 DPE over a 24-h period. The total number moves per hour are presented, with the 12-h dark cycle indicated in grey (mean ± SEM, a minimum of at least six flies were used per group). (F) The total number of moves per day per, of each individual animal for each genotype are presented (mean ± SEM, flies were from at least three independent crosses per genotype; one-way ANOVA with Tukey post-test: *P < 0.05, **P < 0.01, ****P < 0.0001). Flies were from at least three independent crosses per genotype. (G) The total time (in hours) each individual animal was classified as sleeping during daylight hours are presented for each genotype (mean ± SEM, flies were from at least three independent crosses per genotype; one-way ANOVA with Šídák’s multiple comparisons test: *P < 0.05, ****P < 0.0001).

Figure 7.. RuvBL overexpression slows the rate of dipeptide repeat (DPR) production and interacts with the translational machinery.

(A) HeLa cells transfected with empty vector control (Ctrl), FLAG-RuvBL1, or HA-RuvBL2 were co-transfected with empty vector or with 45 uninterrupted sense GGGGCC repeats (45xG4C2) before treating with cycloheximide (CHX) for the indicated time to block further protein translation. RuvBL overexpression was confirmed via immunoblot with GAPDH indicating equal loading of samples, and cyclin D demonstrating efficacy of the CHX treatment. (B) Levels of repeat-associated non-AUG translated poly(GP) DPRs were determined via MSD-ELISA allowing for the monitoring of protein turnover. (C) HeLa cells transfected with empty vector control (ev), FLAG-RuvBL1, or HA-RuvBL2 were co-transfected with empty vector or with V5-tagged 45 uninterrupted sense GGGGCC repeats (V5-45xG4C2). Proteins were harvested at the indicated times post transfection to follow rate of production. (D) Levels of repeat-associated non-AUG translated poly(GP) DPRs were determined via MSD-ELISA and are presented relative to the empty control transfected 0 h sample. (E) The level of poly(GP) DPRs at 8 h post transfection with empty vector control, FLAG-RuvBL1, or HA-RuvBL2. Poly(GP) DPRs are presented relative to the empty vector control (mean ± SEM, N = 3 independent experiments; one-way ANOVA with Tukey post-test: *P < 0.05, **P < 0.01, ***P < 0.001). Lysates from HeLa cells transfected with empty vector control or FLAG-tagged RPL10A were subjected to immunoprecipitation with anti-FLAG antibodies. (F, G) Immune pellets were probed for RuvBL1 (F) and RuvBL2 (G) on immunoblot.

Figure S9.. The effect of RuvBL overexpression on transcription is not CMV promoter dependent.

(A) Schematic diagram indicating the location of qPCR primers (pink arrows) used to determine V5-45xG4C2 transcript levels. (B) HeLa cells transfected with empty vector control (Ctrl), FLAG-RuvBL1, or HA-RuvBL2 were co-transfected with EGFPC2 plasmid and EGFP transcripts quantified by RT-qPCR using GAPDH as housekeeping gene. EGFP transcript levels are shown relative to the EGFP+Ctrl samples (mean ± SEM, N = 3 independent experiments; one-way ANOVA with Tukey post-test: ns, non-significant, ****P < 0.0001). (C) HeLa cells transfected with empty vector control (Ctrl), FLAG-RuvBL1, or HA-RuvBL2 were co-transfected with 45 uninterrupted sense GGGGCC repeats (V5-45xG4C2). V5-45xG4C2 transcript levels were quantified by RT-qPCR using GAPDH as a housekeeping gene (mean ± SEM, N = 3 independent experiments; one-way ANOVA with Tukey post-test: *P < 0.05, ****P < 0.0001). (B, C, D) GAPDH transcript levels in samples from (B, C) were quantified by RT-qPCR using 18S as a housekeeping gene (mean ± SEM, N = 3 independent experiments; one-way ANOVA with Tukey post-test: all ns, non-significant).

Figure 8.. RuvBL overexpression reduces transcription of C9orf72 sense DNA.

(A) HeLa cells transfected with empty vector control (ev), FLAG-RuvBL1, or HA-RuvBL2 were co-transfected with empty vector or with 45 uninterrupted sense GGGGCC repeats (45xG4C2). RuvBL overexpression was confirmed via immunoblot with GAPDH indicating equal loading of samples. (B, C, D) The levels of transcription of the sense repeats (B), GAPDH (C), and endogenous C9orf72 (D) was quantified by RT-qPCR using 18S as a housekeeping gene (mean ± SEM, N = 4 independent experiments; one-way ANOVA with Dunnett’s post-test: ns, non-significant, ****P < 0.0001).

Figure S10.. Knockdown of RuvBL does not affect 45xG4C2 transcript levels.

HeLa cells treated with control siRNA (siCtrl), RuvBL1-targeting siRNA (siRuvBL1), or RuvBL2-targeting siRNA (siRuvBL2) were transfected with empty vector control or 45 uninterrupted sense GGGGCC repeats (V5-45xG4C2). (A, B) RuvBL1 transcript levels (A) and RuvBL2 transcript levels (B) were determined via RT-qPCR using 18S as a housekeeping gene. Levels of RuvBL1 and RuvBL2 transcripts are shown relative to the siCtrl sample transfected with empty vector control plasmid. (C) V5-45xG4C2 transcript levels were quantified by RT-qPCR using 18S as a housekeeping gene (mean ± SEM, N = 3 independent experiments; one-way ANOVA with Tukey post-test: ns, non-significant, ****P < 0.0001).