Nuclear pore complex dysfunction drives TDP-43 pathology in ALS

Abstract

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease characterized by progressive motor neuron degeneration and pathological aggregation of TDP-43. While protein misfolding and impaired autophagy are established features, accumulating evidence highlights the nuclear pore complex (NPC)as a vulnerable, redox-sensitive hub in ALS pathogenesis. Here, we show that selective loss of NPC components, particularly the scaffold proteins NUP107 and NUP93, and FG-repeat-containing components-is a consistent finding across ALS postmortem spinal cord, SOD1^G93A and TDP-43 mutant mouse models, and human cell systems.CRISPR-mediated depletion of NUP107 in human cells triggers hallmark features of ALS pathology, including cytoplasmic TDP-43 mislocalization, increased phosphorylation, and autophagy dysfunction. Conversely, TDP-43 knockdown perturbs NPC composition, suggesting a reciprocal regulatory loop. Crucially, we demonstrate that oxidative stress exacerbated NPC subunit mislocalization and enhanced TDP-43 aggregation. Using oxime blotting and DNPH assays, we show that FG-repeat subunits of NPC were direct targets of redox-driven carbonylation, indicating that oxidative modifications compromise NPC integrity thuspotentially affecting nucleocytoplasmic transport. Our findings established NPC dysfunction as a redox-sensitive driver of TDP-43 pathology in ALS and highlight nucleocytoplasmic transport as a promising therapeutic axis. The susceptibility of long-lived NPC proteins to oxidative damage provides a mechanistic link between redox stress, proteostasis collapse, and neurodegeneration.

Fig. 1

ALS is associated to altered nucleoporin content in spinal cord and brain cortex. A) Left panel shows a representative western-blot image of FG-repeats containing NUPs (reactive to monoclonal antibody MAb414) of lumbar spinal cord homogenates from ALS patients and healthy, age and sex matched controls. Right panel shows the quantified immunoreactivity. B) The left panel shows representative confocal microscopy images from isolated nuclei from spinal cord samples, quantified in the violin plots from the right panel, demonstrating in situ decreased nucleoporin content.C) Quantitative analyses of NUP107, FG-repeats, and NUP93 in isolated nuclei by flow cytometry analyses, showing decreased content of NUP107 in spinal cord samples from ALS patients. D) RNA Seq data in heatmap showing that the mRNAs content of scaffold NUPs and NUP153 (an FG-repeat containing NUP) was decreased in spinal cord from ALS patients.Right panel indicates TPM (transcrits per kilobase million). E) Left panel shows confocal microscopy images with anti-tubulin β III immunoreactivity in nuclei isolated from brain cortex (area 8) from ALS patients and healthy individuals. Middle panel indicates tubulin β III content distribution, with the vertical line situating median values. Right panel illustrates differences in relationship between the content of FG-repeat NPC and tubulin β III in nuclei, with reference to ALS and tubulin β III content. F) Left panel shows representative confocal microscopy images from isolated nuclei from spinal cord samples, quantified plots from the right panel, demonstrating a decreased amount of nuclei from neuronal cells, based on NeuN content. Bars show mean values ± SEM or % in F. ∗,∗∗, and ∗∗∗∗ indicate p < 0.05, p < 0.01 and p < 0.0001 with reference to control values by Student's T test, Chi-Square (F) or Mann-Whitney U test. In E) For controls, linear relationship equation Y = −0.2721∗X + 0.01782 (p < 0.0004) and for ALS, the equation is Y = 0.3195∗X + 0.008008 (p < 0.0001). Scale white bar length in B, E and F are 50 μm long. Ns: non statistically significant differences.

Fig. 2

NUPs are present in motor neurons and glial cells in human lumbar spinal cord, ventral horn. Representative immunohistochemical images of cellular distribution of the scaffolding NUP TPR (A–D) and the FG-repeat NUP107 (E,F). Samples from grey matter are shown in A,B,E and F, while as white matter cells (i.e. glial cells) are shown in C and D. Samples from healthy donors (A,C and E) and ALS patients (B,D and F) are presented. Arrows in A and B indicate non-nuclear staining of TPR, which were more frequent in ALS samples.

Fig. 3

SOD and TDP-43-based ALS preclinical models show altered nucleoporin content in spinal cord. A) Representative western-blot images of total homogenates of lumbar spinal cords from female and male mice (90 days of age) from non-transgenic and G93A SOD mice, quantified in the right panel box-plots. B)Left panel, representative western-blot images of total homogenates of lumbar spinal cords from female and male mice of non-transgenic and TDP-43^Q331K transgenic mice, quantified in the right panel box-plot. C) Uper panel, representative western-blot images of total homogenates from lumbar spinal cord from male non-transgenic and ΔNLS-TDP-43 transgenic mice, quantified in the low panels. Box plots show mean values ± min to max values, with bars showing mean values ± SEM from n = 5–7 different mice, with ∗∗, and ∗∗∗∗ indicate p < 0.01 and p < 0.0001 significant differences between non-transgenic and transgenic mice or between male and female mice by Student's T test, Mann Whitney U test or post-hoc LSD after two-way ANOVA. ns: non statistically significant differences.

Fig. 4

NUP107 silencing alters autophagy and induces TDP-43 pathology in human cells. A) Left panel indicates representative western-blot of different NUPs in CRISPR-mediated NUP-107 silencing of HEK293 cells. Right panel shows densitometry analyses of different experiments. B) Left panel shows representative western-blot of different autophagy and protein-turnover components in the same cells, with right panel indicating densitometric analyses. C) Confocal microscopy of TDP-43 immunoreactivity of HEK293 CRISPR-mediated NUP107 cells, showing altered TDP-43 immunoreactivity, quantified in the lower panel.D) Confocal microscopy of phospho-TDP-43 immunoreactivity of CRISPR-mediated NUP107-silenced HEK293 cells, showing altered TDP-43 immunoreactivity, quantified in lower panel microscopy. E) Western-blot analyses of TDP-43 and phospho-TDP-43 in CRISPR-mediated NUP107-silenced HEK293 cells, quantified in the right panel. F) Higher magnification of confocal microscopy images showing cytosolic aggregates of phospho-TDP-43 (yellow arrows) induced by CRISPR-mediated NUP107-silencing in HEK293 cells, quantified in the left panel (both size and number of phospho-TDP-43 aggregates). G)Left panel, representative western-blot analyses of NUPs and autophagy components in siRNA-mediated silencing of NUP107 in HeLa cells.Right panel shows the quantitative analyses. Bars show mean values ± SEM from N = 3–5 different experiments. ∗,∗∗, ∗∗∗ and ∗∗∗∗ indicate, respectively, p < 0.05, p < 0.01, p < 0.001 and p < 0.0001 significant differences between silenced or non silenced cells by Student's T test or Mann Whitney U test. In B and D, white scale bar lenght is 50 μm long and 30 μm in F. ns: non statistically significant differences.

Fig. 5

Loss of TDP-43 induces alteration in nucleoporin homeostasis. A) Representative immunofluorescence analyses showing NUP107 and MAb414 in TDP-43 silenced cells for a HeLa inducible silencing clone. B) Violin plot showing the quantitative analyses of nuclear (i.e. colocalizing with DAPI) content of NUP107 and MAb414 of cells in A. C) Left panel, western-blot analyses of TPR-content in HeLa cells with inducible TDP-43 silencing, with right panel showing quantitative analyses. D) Western-blot analyses of selected NUPs and LC3B in HeLA cells with doxycycline-inducible TDP-43 silencing under autophagy induction by 24 h of nutrient deprivation (Depriv). E) Quantitative analyses of western-blots shown in D. Bars show mean values ± SEM from N = 3–5 different experiments. ∗,∗∗, and ∗∗∗∗ indicate, respectively, p < 0.05, p < 0.01, and p < 0.0001 significant differences between silenced or non silenced cells by Student's T test, Mann Whitney U test or post-hoc LSD after two-way ANOVA. In A white scale bar lenght is 50 μm long. ns: non statistically significant differences.

Fig. 6

TDP-43 aggregating cell stressors impairs nucleoporin assembly with oxidative stress potentially modyfing FG-repeat containing NUPs in vitro. A) Left panel, representative confocal imaging of TDP-43, p-ERK1/2 and γ-H2AX foci in HEK293 cells after osmotic stress (Sorbitol), indicating extranuclear aggregates (yellow arrow). Right panel shows quantitative analyses of these traits. B) Left panel representative confocal imaging showing cytosolic location of FG-repeats nucleoporin (MAb414) and NUP93 in HEK293 cells after osmotic stress, with right panel showing quantitative analyses of these characteristics. C) Left panel, representative western-blot of MAb414 immunoprecipitated from spinal cord samples from ALS patients and healthy, age and sex-matched individuals, showing potentially oxidative modifications by oxime blot analyses. Right panel shows the quantitative analyses after densitometry. D) In vitro oxidative modification of FG-repeats containing NUPs as shown by anti-DNPH western-blot analyses. Arrows show the change of electrophoretic mobility, compatible with oxidative modification. E) Left panel shows confocal imaging of FG-repeat NUPs (MAb414) and TDP-43 distribution in SH-SY5Y cells after oxidative treatments, with right panel showing quantitative analyses. F) Left panel confocal imaging of TDP-43 aggregating cell stressors (ER stress, proteasome stress and oxidative stress) affecting FG-repeat subcellular distribution in Neuro2A cells, as quantified in right panel violin plots. Bars show mean values ± SEM from N = 4–6 different experiments (except for D, 3 experiments). ∗∗,∗∗∗, and ∗∗∗∗ denote, respectively, p < 0.01, p < 0.001, and p < 0.0001 significant differences between stressed or non stressed cells by Student's T test, Mann Whitney U test or post-hoc LSD after one-way ANOVA. In A and F, white scale bar lenght are 30 μm long, while as in B and E are 50 μm. ns: non statistically significant differences.