Upregulation of ATG7 attenuates motor neuron dysfunction associated with depletion of TARDBP/TDP-43

Abstract

A shared neuropathological hallmark in amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) is nuclear clearance and cytoplasmic aggregation of TARDBP/TDP-43 (TAR DNA binding protein). We previously showed that the ability of TARDBP to repress nonconserved cryptic exons was impaired in brains of patients with ALS and FTD, suggesting that its nuclear depletion contributes to neurodegeneration. However, the critical pathways impacted by the failure to repress cryptic exons that may contribute to neurodegeneration remain undefined. Here, we report that transcriptome analysis of TARDBP-deficient neurons revealed downregulation of ATG7, a critical gene required for macroautophagy/autophagy. Mouse and Drosophila models lacking TARDBP/TBPH in motor neurons exhibiting age-dependent neurodegeneration and motor deficits showed reduction of ATG7 and accumulation of SQSTM1/p62 inclusions. Importantly, genetic upregulation of the autophagy pathway improved motor function and survival in TBPH-deficient flies. Together with our observation that ATG7 is reduced in ALS-FTD brain tissues, these findings identify the autophagy pathway as one key effector of nuclear depletion of TARDBP that contributes to neurodegeneration. We thus suggest that the autophagy pathway is a therapeutic target for ALS-FTD and other disorders exhibiting TARDBP pathology.Abbreviations: ALS: amyotrophic lateral sclerosis; ANOVA: analysis of variance; ChAT: choline acetyltransferase; CTSD: cathepsin D; FTD: frontotemporal dementia; LAMP1: lysosomal associated membrane protein 1; NMJ: neuromuscular junction; RBFOX3/NeuN: RNA binding fox-1 homolog 3; SQSTM1: sequestosome 1; TARDBP/TDP-43: TAR DNA binding protein 43.

Keywords: ALS; ATG7; Drosophila; SQSTM1/p62; TARDBP/TDP-43; autophagy; frontotemporal dementia; mouse.

Figure 1.

Autophagy-related genes are affected by Tardbp deletion in both neurons and skeletal muscle. (a) Venn diagram illustrating the intersection between three gene subsets: all autophagy related genes (Gene Ontology ID GO:0006914) and genes with at least one transcript in which the absolute value fold change is > 1.5 for either brain or skeletal muscle knockout of Tardbp. Of the manually created 50 genes found to overlap between the brain and muscle subsets, 2 genes perform crucial functions in the autophagy process: Tecpr1 and Atg7. (b) Immunoblot analysis of hippocampal lysates of Camk2a-Cre;tardbp^F/F mice using antisera specific to TARDBP and ATG7. Levels of ATG7 were significantly reduced in both p25 and 3-months-old Camk2a-Cre;tardbp^F/F mice as compared to that control littermates. (c) Immunofluorescent staining of hippocampus CA3 region in 3 months old Camk2a-Cre;tardbp^F/F mice (SQSTM1: green, TARDBP: red, TO-PRO3: blue). Scale bar: 20 μm. Note the significantly increased accumulation of large SQSTM1 inclusions (arrows) in TARDBP-depleted neurons of Camk2a-Cre;tardbp^F/F mice but not in control littermates. (d) Quantifications of % neurons (RBFOX3/NeuN⁺) with increased SQSTM1 puncta in the hippocampus (see Figure S2(a)) at 3 months of age (n: wild-type = 3, tardbp knockout = 3, Mann-Whitney test, *p < 0.05) .

Figure 2.

Progressive motor dysfunction in ChAT-IRES-Cre;tardbp^F/F mice in lower motor neurons. (a), Breeding strategy to generate ChAT-IRES-Cre;tardbp^F/F mice (see Methods). (b), Immunofluorescent staining of ventral spinal cord of p21 ChAT-IRES-Cre;Tardbp^+/+ and ChAT-IRES-Cre;tardbp^F/F mice (ChAT: green, TARDBP: red, DAPI: blue). Scale bar: 50 μm. Note the depletion of nuclear TARDBP staining in neurons of motor neurons. Quantifications of % neurons (ChAT-positive) expressing detectable TARDBP showing TARDBP was lost in the majority of motor neurons (n = 6 mice per age group, Mann-Whitney test, **p < 0.01). (c), Reduction in body weight of ChAT-IRES-Cre;tardbp^F/F mice (two-way ANOVA with Bonferroni post-test, *p < 0.05, **p < 0.01). (d), Kaplan-Meier survival curves for ChAT-IRES-Cre;Tardbp^+/+ (n = 7) and ChAT-IRES-Cre;tardbp^F/F (n = 10) mice over 300 days. (e, f), Footprint analysis of 8–14 week-old ChAT-IRES-Cre;tardbp^F/F mice to assess gait (n: wild-type = 3; tardbp knockout = 4); note ChAT-IRES-Cre;tardbp^F/F mice displayed a significantly shorter stride length at 11 weeks of age in both forelimb and hindlimb, and this effect became more pronounced as the mice aged. (g), Strength of 6–21 week-old ChAT-IRES-Cre;tardbp^F/F mice was determined using the hanging wire test; note the hang time of these mice was significantly reduced as compared to that of control littermates and the time to fall became shorter as the ChAT-IRES-Cre;tardbp^F/F mice aged (n = 11 mice per age group). Two-way ANOVA with Bonferroni post-test, *p < 0.05, **p < 0.01.

Figure 3.

Age-dependent neurodegeneration in ChAT-IRES-Cre;tardbp^F/F mice. (a), Spinal motor neuron loss in aged ChAT-IRES-Cre;tardbp^F/F mice. Lumbar sections were stained with cresyl violet (upper panel) and immunostained using antisera against ChAT (lower panel). Note few ChAT-positive neurons remain in ventral horns of 7 months old ChAT-IRES-Cre;tardbp^F/F mice. Scale bar: 200 µm left two panels, 100 μm right two panels. (b), Significant neuron loss were detected in ventral horns of ChAT-IRES-Cre;tardbp^F/F mice starting at 3 months old, n = 4 mice per age group. Mann-Whitney test, *p < 0.05. (c), Transverse sections of L4 dorsal and ventral roots. There is massive loss of large motor axons in ventral root of 7 months old ChAT-IRES-Cre;tardbp^F/F mice. Note also the presence of myelin ovoids and Wallerian degeneration in motor axons. Scale bar: 50 μm top two panels; 20 μm third panel. (d), Axial sections of gastrocnemius muscle from 3 months of old ChAT-IRES-Cre;Tardbp^+/+ and ChAT-IRES-Cre;tardbp^F/F mice stained with H&E (top panels) and gomori trichrome (bottom panels). Note the grouped atrophy (arrows) and significantly reduced cross section area (CSA) occurring in ChAT-IRES-Cre;tardbp^F/F mice. Scale bar: 50 μm. (e), Quantification of the muscle fiber cross-sectional area (CSA) in 3 months old ChAT-IRES-Cre;Tardbp^+/+ and ChAT-IRES-Cre;tardbp^F/F mice, n = 3 mice per age group. Mann-Whitney test, **p < 0.01.

Figure 4.

Compromised autophagy in ChAT-IRES-Cre;tardbp^F/F mice. (a), Immunoblot analysis of ventral hemisections of lumbar spinal cord from ChAT-IRES-Cre;tardbp^F/F mice and littermate controls shows a reduction in ATG7 protein. GAPDH was used as a loading control. (b), Immunofluorescent staining of ventral spinal cord of 1 and 3 months old ChAT-IRES-Cre;tardbp^F/F mice (SQSTM1: green, TARDBP: red, TO-PRO3: blue). Scale bar: 20 μm. Note the accumulation of large SQSTM1 inclusions (arrows) in TARDBP depleted neurons of ChAT-IRES-Cre;tardbp^F/F mice but not in control littermates. (c), Co-localization of Ubiquitin in the SQSTM1-positive inclusions (SQSTM1: green, Ubi: red, TO-PRO3: blue). (d), Immunofluorescent staining of ventral lumbar spinal motor neurons of 1 month ChAT-IRES-Cre;tardbp^F/F mice (left) compared with littermate controls (green: RBFOX3/NeuN; red: CTSD). (e), Quantifications of % neurons (RBFOX3/NeuN⁺) expressing increased SQSTM1 puncta in the spinal cord at 1 and 3 months of age (Figure S2(b)). Likewise, the increase in CTSD fluorescence intensity normalized to background and puncta area in tardbp knockout mice were quantified (n = 120 lumbar motor neurons per animal, 6 mice per age group, Mann-Whitney test, *p < 0.05) .

Figure 5.

Decreased Atg7 protein expression in fruit fly lacking TBPH and human ALS-FTD brains. (a), Diagram of RT-PCR detection strategy to amplify across the common exon splice junctions of Atg7 isoforms. Both Atg7 mRNA and protein were drastically decreased in homozygous TBPH-deficient mutants by RT–PCR (b) and immunoblot analysis (c). (d, e), Decreased ATG7 protein in motor cortex of ALS-FTD cases (n = 11) that all displayed TARDBP inclusions compared to age matched control brains (n = 6). GAPDH was used to normalize the measurement. Mann-Whitney test, **p < 0.01. (f), Immunofluorescent staining of motor cortex from a ALS-FTD brain showing accumulation of large SQSTM1 inclusions (arrows) in nuclear TARDBP-depleted neurons of the ALS-FTD brain, but not in the age matched control brain (SQSTM1: green, TARDBP: red, DAPI: blue). Scale bar: 20 μm. Quantification showing the significantly increased number of SQSTM1-positive neurons in ALS-FTD brains (n = 4) compared with that in the control brains (n = 3). Mann-Whitney test, *p < 0.05.

Figure 6.

Genetic rescue of TBPH-null phenotypes in Drosophila. Ubiquitously expressing Drosophila Atg7 suppressed TBPH loss-of-function phenotypes, as shown in (a), elevated developmental viability (n: TBPH^Q367/Q367 = 347, TBPH^Q367/Q367;UAS-Atg7/Act5C-Gal4 = 201, TBPHΔ23/Δ23 = 1195, TBPHΔ23/Δ23;UAS-Atg7/Act5C-Gal4 = 351); (b), enhanced locomotive capability (n = 30 per genotype) and (c), extended life span (n: wildtype = 86, TBPHΔ23/Δ23 = 34, TBPHΔ23/Δ23;UAS-Atg7/Act5C-Gal4 = 43. Mann-Whitney test, *p < 0.05; log-rank test, ****p < 0.0001). (d), Immunofluorescent staining of third instar larval NMJ (SQSTM1: red, synaptic marker Fas2 [Fasciclin 2]: green). Scale bar: 20 μm. Note the accumulation of large SQSTM1 inclusions in TBPH depleted neurons of TBPHΔ23/Δ23 fly but not in control TBPHΔ23/+ line. (e), Atg7 overexpression significantly decreased the SQSTM1 accumulation (n = 6 per genotype, Mann-Whitney test, ****p < 0.0001) .