ALS-linked TDP-43 mutations produce aberrant RNA splicing and adult-onset motor neuron disease without aggregation or loss of nuclear TDP-43

Abstract

Transactivating response region DNA binding protein (TDP-43) is the major protein component of ubiquitinated inclusions found in amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD) with ubiquitinated inclusions. Two ALS-causing mutants (TDP-43(Q331K) and TDP-43(M337V)), but not wild-type human TDP-43, are shown here to provoke age-dependent, mutant-dependent, progressive motor axon degeneration and motor neuron death when expressed in mice at levels and in a cell type-selective pattern similar to endogenous TDP-43. Mutant TDP-43-dependent degeneration of lower motor neurons occurs without: (i) loss of TDP-43 from the corresponding nuclei, (ii) accumulation of TDP-43 aggregates, and (iii) accumulation of insoluble TDP-43. Computational analysis using splicing-sensitive microarrays demonstrates alterations of endogenous TDP-43-dependent alternative splicing events conferred by both human wild-type and mutant TDP-43(Q331K), but with high levels of mutant TDP-43 preferentially enhancing exon exclusion of some target pre-mRNAs affecting genes involved in neurological transmission and function. Comparison with splicing alterations following TDP-43 depletion demonstrates that TDP-43(Q331K) enhances normal TDP-43 splicing function for some RNA targets but loss-of-function for others. Thus, adult-onset motor neuron disease does not require aggregation or loss of nuclear TDP-43, with ALS-linked mutants producing loss and gain of splicing function of selected RNA targets at an early disease stage.

Fig. 1.

Generation of multiple lines of transgenic mice expressing wild-type, Q331K, or M337V-mutant human TDP-43. (A) Schematic of the PrP-TDP-43 transgene. (B) Immunoblotting using an antibody that recognizes both mouse and human TDP-43 with equal affinity in whole-organ lysates from transgenic animals. Br, brain; Cb, cerebellum; SC, spinal cord; Hrt, heart; Mu, skeletal muscle; Liv, liver; Kid, kidney. Equal protein amounts were loaded per lane. (C) Immunoblotting of whole spinal cord lysates from transgenic mice with the mouse/human TDP-43 antibody. (D) Accumulated transgene-encoded mRNA levels of in spinal cords of PrP-TDP-43 mice. (E) Representative single-plane confocal images of the ventral horn of lumbar-level spinal cord from 2-mo-old animals using a rabbit polyclonal antibody (Sigma C3956) to the myc tag on the transgene encoded protein and costained for the neuronal marker NeuN.

Fig. 2.

TDP-43^Q331K and TDP-43^M337V mice develop age-dependent, progressive motor deficits. (A and B) Motor deficits measured by rotarod. n ≥ 11 for each genotype and each time point. **P < 0.01 and ***P < 0.001 using one-way ANOVA at each time point with Bonferroni’s post hoc test. (C) The 10- to 12-mo-old TDP-43^Q331K high-expressing animals developed hindlimb weakness as measured by a hindlimb grip-strength assay (P = 0.0002 by Student’s t test). Data shown are the average ± SEM. n = 12 per genotype.

Fig. 3.

Electrophysiological measures identify lower motor neuron deficits in aged mutant-expressing TDP-43 mice. (A) Schematic of measurement of MEPs elicited by electrical stimulation of motor cortex and extrapyramidal system, with responses recorded from the exposed T12 spinal segments (spinal cord surface MEPs) or from the gastrocnemius muscle (MMEPs). (B) Resting EMG recording from the gastrocnemius muscle in isoflurane anesthetized animals in the absence of any stimulus. (C) MMEP recordings from the gastrocnemius muscle. **P < 0.01, ***P < 0.001 by Student's t test. (D) Spinal cord surface MEP recordings in nontransgenic, TDP-43^Wild-Type and TDP-43^Q331K mice.

Fig. 4.

TDP-43^Q331K transgenic mice develop age-dependent lower motor neuron degeneration. (A) Quantification of ChAT-positive α-motor neurons in lumbar spinal cords from TDP-43 transgenic mice (average ± SD, n = 3 per genotype per time point; *P = 0.04). (B) Quantification of motor axons shown in D (at 10–12 mo of age, n ≥ 3 per genotype per time point; ***P < 0.001). (C) Distributions of diameters of L5 motor axons determined from images in D. (D) Representative images of the L5 motor axon roots from TDP-43^Wild-Type or TDP-43^Q331K transgenic animals stained with Toluidine blue. (arrowheads) Degenerating axons in TDP-43^Q331K mice. (E) Quantification of average neuromuscular junctions per section in 10- to 12-mo-old TDP-43 transgenic animals (± SEM; *P = 0.02, n ≥ 3 per genotype). (F) Representative neuromuscular junctions from nontransgenic and TDP-43^Q331K–expressing animals. FM, fluoromyelin. (G) Hematoxylin and eosin staining of grastrocnemius muscles. (arrowheads) Atrophied fibers of varying sizes; (arrows) regenerating fibers with centralized nuclei. All statistics performed using one-way ANOVA with Bonferroni’s post hoc test.

Fig. 5.

Neither wild-type nor mutant TDP-43 show aberrant cytosolic localization in the brain and spinal cords of TDP-43 transgenic mice. (A) Experimental scheme for the nuclear and cytosolic fractionation of spinal cords (B, i and ii) or brains (B, iii and iv) of nontransgenic and TDP-43 transgenic mice. (B) Immunoblotting of enriched cytosolic or nuclear extracts from the brain or spinal cord of 10- to 12-mo-old nontransgenic, wild-type, and mutant TDP-43 transgenic animals using an antibody recognizing both mouse and human TDP-43 with equal affinity. Note enrichment of Hsp90 in the cytosolic extract and FUS in the nuclear extract, indicating the efficiency of separation. (C) Immunofluorescent localization of endogenous ChAT and transgene-encoded TDP-43 in 10-mo-old TDP-43 transgenic animals.

Fig. 6.

Both enhanced activity and loss-of-function for specific RNA splicing targets directly bound by TDP-43^Q331K. (A) Experimental strategy to identify differentially regulated splicing events using Affymetrix splicing-sensitive microarrays to analyze RNAs extracted from cortices of 2-mo-old nontransgenic, TDP-43^Wild-Type, TDP-43^Q331K, or TDP-43^Q331K-Low transgenic mice. (B) TDP-43–regulated cassette exons identified from TDP-43 depletion experiments in mouse (11). Cassette exons are divided into direct targets, bound by TDP-43 within 2 kb, and indirect targets, not bound by TDP-43. (C) Bar plot displaying the fraction of overlap of TDP-43–regulated exons, direct (dark green) and indirect (light green), with exons that changed upon expression of TDP-43^Wild-Type, TDP-43^Q331K, and TDP-43^Q331K-low. The TDP-43^Q331K and TDP-43^Q331K-low overlapping exons show a significant enrichment (by χ² analysis: P < 0.0005) in the fraction that are direct TDP-43–regulated exons compared with the fraction that are indirect TDP-43–regulated exons. No enrichment was found in TDP-43^Wild-Type. (D) Overlap of significantly changed cassette exons observed in the cortices from TDP-43 transgenic mice with TDP-43–regulated cassette exons. Direct and indirect exons that are also regulated in TDP-43 transgenic mice are represented as exons that are changed in the opposite (colored gray) or same (included or excluded in both, colored white) direction of TDP-43 depletion. (E and F) RT-PCR validation for a subset of alternate cassette exons identified in cortex by splicing-sensitive microarrays and that are (E) mutant-dependent or (F) transgene dose-dependent and that change in the same or opposite direction following knockdown (KD) in TDP-43 following antisense oligonucleotide infusion within striatum. Bar plots show the ratio of inclusion to exclusion calculated from the mean intensities (n ≥ 3 biological replicates, ± SD). Representative gel analyses are images shown with duplicate biological replicates. *P < 0.05, **P < 0.01, and ***P < 0.001 with Student’s t test; in/ex, ratio of inclusion to exclusion. (G) Bar graph depicting the percentages of differentially included or excluded exons upon human TDP-43 expression (green dashed line). The fraction of unchanged exons in any cortex experiment (TDP-43^Wild-Type, TDP-43^Q331K, TDP-43^Q331K-low) that have TDP-43 binding. (i) “All Changing Exons” includes all exons that were either included or excluded (25% of these were direct TDP-43 targets). (ii) “Overlapping Exons with TDP-43 KD” includes only exons that changed following TDP-43 depletion (i.e., exons regulated by endogenous TDP-43). A significant increase was found in the percentage of direct targets (∼43% for TDP-43^Q331K and TDP-43^Wild-Type, and 53% for TDP-43^Q331K-low) for cassette exons that are excluded upon human TDP-43 expression, with only a modest ∼20% of included exons that are direct targets. (iii) “Non-overlapping exons with TDP-43 KD” includes only exons not previously shown to be regulated by endogenous TDP-43 depletion, in which 39% of excluded exons in the TDP-43^Q331K animals were direct targets.

Fig. 7.

Unique splicing alterations in the spinal cord of TDP-43^Q331K mice include changes in genes involved in neurological function and transmission. (A) Experimental strategy to identify differentially regulated splicing events in spinal cords of 2-mo-old nontransgenic, TDP-43^Wild-Type, or TDP-43^Q331K transgenic mice. Pie charts display total alternative splicing events significantly altered in the spinal cords of transgenic mice, relative to nontransgenic animals (colors representing the types of events on the array, defined at the right). (B) Diagram showing overlap between events changing in cortex and spinal cords of TDP-43^Wild-Type and TDP-43^Q331K animals. (C) Overlap of spinal cord TDP-43^Wild-Type and TDP-43^Q331K alternative cassette exons reveals a set of 1,060 common splicing events, 1,059 of which change in the same direction. (D) RT-PCR validation for a subset of spinal cord alternate cassette exons identified in A. Bar plots show mean intensities of n ≥ 3 biological replicates (± SD). Representative gel images shown, with duplicate biological replicates. (E) Modes of TDP-43 action in exon splicing. *P < 0.05, **P < 0.01, and ***P < 0.001 with Student's t test.