C9orf72 poly(GR) aggregation induces TDP-43 proteinopathy

Abstract

TAR DNA-binding protein 43 (TDP-43) inclusions are a pathological hallmark of frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS), including cases caused by G₄C₂ repeat expansions in the C9orf72 gene (c9FTD/ALS). Providing mechanistic insight into the link between C9orf72 mutations and TDP-43 pathology, we demonstrated that a glycine-arginine repeat protein [poly(GR)] translated from expanded G₄C₂ repeats was sufficient to promote aggregation of endogenous TDP-43. In particular, toxic poly(GR) proteins mediated sequestration of full-length TDP-43 in an RNA-independent manner to induce cytoplasmic TDP-43 inclusion formation. Moreover, in GFP-(GR)₂₀₀ mice, poly(GR) caused the mislocalization of nucleocytoplasmic transport factors and nuclear pore complex proteins. These mislocalization events resulted in the aberrant accumulation of endogenous TDP-43 in the cytoplasm where it co-aggregated with poly(GR). Last, we demonstrated that treating G₄C₂ repeat-expressing mice with repeat-targeting antisense oligonucleotides lowered poly(GR) burden, which was accompanied by reduced TDP-43 pathology and neurodegeneration, including lowering of plasma neurofilament light (NFL) concentration. These results contribute to clarification of the mechanism by which poly(GR) drives TDP-43 proteinopathy, confirm that G₄C₂-targeted therapeutics reduce TDP-43 pathology in vivo, and demonstrate that alterations in plasma NFL provide insight into the therapeutic efficacy of disease-modifying treatments.

Fig. 1.. Poly(GR) directly accelerates and enhances TDP-43 aggregation.

(A) TDP-43–MBP (5 μM) was incubated with buffer, 2 μM poly(GR), or 2 μM poly(GA) in the presence or absence of TEV protease (1 μg/ml). Aggregation was assessed by turbidity measured at an absorbance of 395 nm. Values are normalized mean ±SEM (n = 6). (B) Quantification of the area under the curve in turbidity for each condition in the in vitro aggregation assay (n = 6). (C) TDP-43 aggregation was initiated as described above, with the addition of TEV protease (10 μg/ml) with or without co-incubation with poly(GR) or poly(GA) for 30 min. Aggregation was visualized by differential interference contrast microscopy. Scale bars, 10 μm. (D) Representative electron micrographs of TDP-43 with or without poly(GR) or poly(GA). Samples were processed for EM at the end point of the turbidity assay. Scale bars, 2 μm (top) and 0.4 μm (bottom). (E) EM quantification of TDP-43 aggregate area and intensity per aggregate area with or without poly(GR) or poly(GA) (n = 9). Data shown are the mean ± SEM, one-way ANOVA, Tukey’s post hoc analysis. In (B), **P = 0.0027 and ****P < 0.0001. In (E), * (left to right) P = 0.0300 and P = 0.0263, ** (left to right) P = 0.0027 and P = 0.0027, NS (not significant; left to right) P = 0.5707 and P = 0.6080. a.u., arbitrary units.

Fig. 2.. Poly(GR) mediates sequestration of cytosolic full-length TDP-43 into the inclusions.

(A) IEM using an anti-poly(GR) antibody labeled with gold particles (18 nm) in HEK293T cells expressing GFP(GR)₁₀₀. The selected region in the low-magnification image (left) is shown at high magnification (right). Scale bars, 1 μm (left) and 0.1 μm (right). (B) Triple-immunofluorescence staining for poly(GR), TDP-43, and TIA-1 in HEK293T cells expressing GFP-(GR)₁₀₀ and various Myc-tagged TDP-43 species. Scale bars, 5 μm. (C) Quantification of the number of cytoplasmic TDP-43 aggregates in HEK293T cells co-expressing various Myc-tagged TDP-43 constructs with either GFP or GFP-(GR)₁₀₀ (n = 3 independent experiments). (D) Proximity ligation assay (PLA) for GFP-(GR)₁₀₀ and Myc-tagged TDP-43 species in HEK293T cells, with the PLA signal being indicative of their interaction. Scale bars, 5 μm. (E) Quantification of the PLA signal in HEK293T cells expressing GFP-(GR)₁₀₀ and Myc-tagged TDP-43 species (n = 14 to 18 images). Data shown are the mean ± SEM. In (C), ****P < 0.0001 and NS P = 0.9959, two-way ANOVA, Tukey’s post hoc analysis. In (E), ****P < 0.0001, one-way ANOVA, Tukey’s post hoc analysis. WT, wild type.

Fig. 3.. Poly(GR) aggregation is sufficient to induce mislocalization and aggregation of endogenous TDP-43 in vivo.

(A) Representative images of immunohistochemical analysis of poly(GR) in the cortex of 2-week-old GFP mice or GFP-(GR)200 mice (diffuse labeling noted by black arrows, aggregates indicated by black arrowheads). Scale bars, 20 μm. (B) Quantification of the number of poly(GR)-positive cells with either diffuse or aggregated poly(GR) in 2-week-old (n = 8) or 3-month-old (n = 7) GFP-(GR)₂₀₀ mice. (C) Representative images of immunohistochemical analysis of ataxin-2 (inclusions indicated by black arrowheads) in the cortex of 2-week-old GFP or GFP-(GR)₂₀₀ mice. (D) Representative images of immunohistochemical analysis of TDP-43 in the cortex of 2-week-old GFP-(GR)₂₀₀ mice. Arrowheads in GFP-(GR)₂₀₀ mice indicate cells with TDP-43–positive inclusions and either normal (left panel) or reduced (right panel) nuclear TDP-43. Scale bars, 20 μm. (E) Double-immunofluorescence staining for poly(GR) and ataxin-2 in the cortex of 2-week-old GFP-(GR)₂₀₀ mice (n = 6). Scale bars, 5 μm. (F) Triple-immunofluorescence staining for eIF3η, TDP-43, and poly(GR) in the cortex of 2-week-old GFP-(GR)₂₀₀ mice (n = 6). Scale bars, 5 μm. (G) Double-immunofluorescence staining for poly(GR) and pTDP-43 in the cortex of 2-week-old GFP-(GR)₂₀₀ mice (n = 6). Scale bars, 5 μm. (H) Quantification of the percentage of cells that are either nontransduced (NT) cells or transduced cells with either diffuse or aggregated poly(GR) that contain TDP-43 inclusions and exhibit depleted nuclear TDP-43 in 2-week-old GFP-(GR)₂₀₀ mice (n = 6). (I) Triple-immunofluorescence staining for poly(GR), eIF3η, and TDP-43 in the hippocampus of patients with c9FTD/ALS [case #1 (top), case #2 (bottom); see table S3 for patient information]. (J) Triple-immunofluorescence staining for poly(GA), eIF3η, and TDP-43 in the cortex of 3-month-old GFP-(GA)₅₀ mice (n = 3). Scale bars, 5 μm. Data shown are the mean ± SEM. In (B), ***P = 0.0001 and ****P < 0.0001, two-tailed unpaired t test. In (H), **** P < 0.0001, one-way ANOVA, Tukey’s post hoc analysis.

Fig. 4.. Poly(GR) aggregates sequester nucleocytoplasmic transport factors and NPC proteins in vivo.

(A) Double-immunofluorescence staining for poly(GR) and importins in nontransduced (NT), diffuse, or aggregated poly(GR) cells, including importin α5 (top three panels) and karyopherin α2 (KNPA2; bottom three panels) in the cortex of 2-week-old GFP-(GR)₂₀₀ mice (n = 6). Scale bars, 5 μm. (B) Quantification of the percentage of NT cells or of transduced cells with either diffuse or aggregated poly(GR) that contain abnormal KPNA2 in 2-week-old GFP-(GR)₂₀₀ mice (n = 6). (C) Double-immunofluorescence staining for poly(GR) and several NPC proteins, including NPC (top three panels) and NUP98 (bottom three panels) in the cortex of 2-week-old GFP-(GR)₂₀₀ mice (n = 6). Scale bars, 5 μm. (D) Quantification of the percentage of NT cells or of transduced cells with either diffuse or aggregated poly(GR) that contain abnormal NUP98 in 2-week-old GFP-(GR)₂₀₀ mice (n = 6). (E) Triple-immunofluorescence staining for NPC, poly(GR), and TDP-43 in the cortex of GFP-(GR)₂₀₀ mice (n = 6). Scale bars, 5 μm. (F) Quantification of the percentage of NT cells or of transduced cells with either diffuse or aggregated poly(GR) that contain abnormal NPC and TDP-43 in 2-week-old GFP-(GR)₂₀₀ mice (n = 6). (G) Triple-immunofluorescence staining for NPC, poly(GR), and eIF3η in the cortex of 2-week-old GFP-(GR)₂₀₀ mice (n = 6). Scale bars, 5 μm. (H) Quantification of the percentage of NT cells or of transduced cells with either diffuse or aggregated poly(GR) that contain abnormal NPC and eIF3η in 2-week-old GFP-(GR)₂₀₀ mice (n = 6). Data shown are the mean ± SEM; ****P < 0.0001, one-way ANOVA, Tukey’s post hoc analysis.

Fig. 5.. A G₄C₂-targeting therapeutic reduces poly(GR) and alleviates TDP-43 aggregation and neurodegeneration in vivo.

(A) Representative images of RNA FISH for the detection of sense RNA foci in (G₄C₂)₁₄₉ mice treated with PBS or c9ASO. Scale bars, 10 μm. (B) Immunohistochemical labeling of poly(GA), poly(GP), and poly(GR) in PBS- or c9ASO-treated (G₄C₂)₁₄₉ mice. Scale bars, 20 μm. (C and D) Representative images of immunohistochemical analysis of ataxin-2 (C) and pTDP-43 (D) in PBS- or c9ASO-treated (G₄C₂)₁₄₉ mice, with inclusions indicated by black arrowheads. Scale bars, 20 μm. (E) Quantitative analysis of the number of ataxin-2 or pTDP-43 inclusions per square millimeter of cortex in (G₄C₂)₁₄₉ mice treated with PBS (n = 18) or c9ASO (n = 12). (F) Quantitative analysis of the number of NeuN-positive neurons per square millimeter of cortex in (G₄C₂)₁₄₉ mice treated with PBS (149R-PBS, n = 18) or c9ASO (149R-c9ASO, n = 12) relative to (G₄C₂)₂ control mice treated with PBS (2R-PBS, n = 15). (G) Plasma neurofilament light (NFL) concentrations in mice measured using the Simoa HD-1 Analyzer (2R-PBS, n = 17; 149R-PBS, n = 18; 149R-c9ASO, n = 12). (H) The number of NeuN-positive neurons per square millimeter of cortex negatively correlated with plasma NFL concentrations in (G₄C₂)₁₄₉ mice (PBS- and c9ASO-treated mice indicated by black and red squares, respectively). (I) The number of pTDP-43 inclusions per square millimeter of cortex positively correlated with plasma NFL concentrations in (G₄C₂)₁₄₉ mice. Data represent the mean ± SEM. In (E), *P=0.0407 and ****P < 0.0001, unpaired two-tailed t test. In (F), * (left to right) P=0.0293 and P=0.0426, one-way ANOVA, Tukey’s post hoc analysis. In (G), *P=0.0166 and ***P=0.0009, one-way ANOVA, Tukey’s post hoc analysis.