Main text

Abnormal accumulation of TDP-43 has been identified as a pathological signature of amyotrophic lateral sclerosis (ALS), an adult neurodegenerative disease characterized by a selective loss of motor neurons, and a part of frontotemporal dementia (FTD) [1]. Cytoplasmic accumulation of TDP-43 with a loss of TDP-43 from nuclei, known as a TDP-43 pathology, is observed in almost all forms of ALS, including sporadic and familial ALS. To date, more than 50 mutations in the TARDBP gene, encoding TDP-43, have been identified in inherited and sporadic ALS, implicating TDP-43 dysfunction as a central component for ALS pathogenesis [2]. TDP-43 is a ubiquitously expressed DNA/RNA binding nuclear protein and plays multifunctional roles in RNA metabolism, including pre-mRNA splicing, translational control, and mRNA stability [3]. Of note, TDP-43 is known to control its own mRNA stability through binding to the 3′ UTR, indicating that the level of TDP-43 protein is tightly regulated [3]. Indeed, overexpression of wild-type TDP-43 in mice induces neurodegeneration, whereas elimination of TDP-43 leads to embryonic lethality [4, 5]. However, it is still unclear whether dysfunction in TDP-43 leads to neurodegeneration through a gain or loss of TDP-43 function. To model TDP-43-mediated neurodegeneration in mice, several lines of transgenic mice have been developed and reproduced some features of neurodegeneration observed in human ALS/FTD. However, the overexpression approach has a limitation in differentiating the role between wild-type and mutant TDP-43 in motor neuron health and disease in mice [4, 5].

Based on these backgrounds, we set out to create a knock-in mouse model carrying an ALS patient-derived mutation in the murine Tardbp gene. Of more than 50 known mutations, we chose TDP-43M337V mutation for the following reasons: TDP-43M337V protein has a long half-life in cells, the ALS patients with this mutation show earlier disease onset [6, 7], and an amino acid sequence of 241–414 including a methionine residue at position 337 is highly conserved among mouse and human. We engineered mice with n.1009 A > G (M337V) mutation in the murine Tardbp gene by utilizing CRISPR/Cas9 genome-editing technology (Additional file 1). Both homozygous and heterozygous mice carrying the allele of TDP-43M337V developed normally as recently reported (Fig. 1a, Additional file 1: Figure S1, S2) [8].

Fig. 1
figure 1

Characterization of TDP-43M337V knock-in mice. a Schematic illustration of introducing TDP-43M337V mutation into an endogenous murine Tardbp exon 6 (left panel). The representative genotyping result is also shown (right panel). Nhe I restriction site is introduced in the mutant allele, resulting in no change of the amino acid at Nhe I site. b The expression level of Tardbp mRNA was not altered in the brains of 700-days-old TDP-43M337V/M337V (M337 V/M337 V) mice and wild-type (WT) littermates. c Alternation in splicing of genes regulated by TDP-43. The level of mRNA containing exons included by TDP-43 (Kcnip2 exon 2 and 3) was increased, while the levels of mRNA containing exons excluded by TDP-43 (Sort1 exon 17b and Sema3f exon 5) were reduced, suggesting a gain-of-function mechanism in TDP-43M337V/M337V mice. Relative expression levels of mRNA normalized to the WT control are plotted with SD (n = 3 each (b, c)) and were analyzed by unpaired t-tests. d Representative immunoblots of TDP-43 and β-actin in the brains and spinal cords of 700-days-old TDP-43M337V/M337V mice and WT littermates. Asterisk denotes a non-specific band. e and f Representative immunofluorescence images of the anterior horn in lumbar spinal cords of 700-days-old TDP-43M337V/M337V mice and WT littermates stained with anti-TDP-43 (3H8, green) and anti-ChAT (red) antibodies along with the merged images. TDP-43 was not mislocalized in motor neurons of TDP-43M337V/M337V mice (e). Low magnification images stained with anti-ChAT antibody (f). Scale bars: 20 μm (e), 100 μm (f). g Quantification of the numbers of ChAT-positive motor neurons per each anterior horn (AH) in the lumbar spinal cords of 700-days-old TDP-43M337V/M337V mice and WT littermates. For quantification, 20 AHs in three animals per each genotype were counted, and data are plotted as mean ± SD, and were analyzed by an unpaired t-test. h and i TDP-43M337V/M337V mice did not show any motor dysfunction phenotypes in the measurement of clasping score (h) and rotarod test (i). Data are plotted as mean ± SD, and were analyzed with two-way ANOVA. n = 15 for WT and 14 for M337 V/M337 V

Full size image

TDP-43 plays a pivotal role in regulating alternative splicing as well as controlling the level of TDP-43 mRNA itself by a negative feedback mechanism. Therefore, we first examined whether ALS-linked TDP-43M337V mutation affects the expression level of its own mRNA in mice. Analysis of TDP-43 mRNA levels in the brains of 700-days-old homozygous TDP-43M337V mice (TDP-43M337V/M337V) revealed that there was no difference in expression level between wild-type and TDP-43M337V/M337V mice (Fig. 1b). In addition, the mRNA levels of Notch1 and Nek1, known as TDP-43 target genes, were unaffected by homozygous M337 V mutation (Additional file 1: Figure S3). We next examined whether TDP-43M337V deregulates alternative splicing of mRNAs that are known as splicing targets of TDP-43. Among the several splicing targets examined, we found a 1.49-fold increase in inclusion of Kcinp2 exon 2/3, a 0.85-fold decrease in exclusion of Sort1 exon 17b, and a 0.63-fold decrease in exclusion of Sema3f exon 5 in the brain of TDP-43M337V/M337V mice (Fig. 1c). Although there were no significant changes in other splicing targets, Poldip3 and Eif4h (Additional file 1: Figure S4), changes in splicing of Kcinp2, Sort1, and Sema3f in TDP-43M337V/M337V mice are consistent with a gain of TDP-43 function [9, 10].

Since the mislocalization of TDP-43 protein in cytoplasm is a pathological signature of ALS, we examined subcellular localization of TDP-43M337V mutant protein in the affected tissue in TDP-43M337V/M337V mice. Both mutant and wild-type TDP-43 proteins expressed at the similar level, and were predominantly localized in nucleus of brain and spinal cords of 700-days-old TDP-43M337V/M337V and wild-type mice (Fig. 1d, e), suggesting that disease-causing missense mutation in TDP-43 alone did not alter the protein level itself and was insufficient to induce protein mislocalization in mice. Moreover, carboxyl-terminal (C-terminal) fragments of TDP-43, characteristic of TDP-43 pathology, were not detected in the brains and spinal cords of TDP-43M337V/M337V mice (Fig. 1d), and there was no detectable loss of motor neurons or reactive gliosis in TDP-43M337V/M337V mice (Fig. 1e-g, Additional file 1: Figure S5). Nuclear Gems, where SMN complex resides to control splicing, are known to be regulated by TDP-43 and FUS [11,12,13]. In ventral horn neurons of TDP-43M337V/M337V mice, the number of nuclear Gems was not altered (Additional file 1: Figure S6). We further examined whether TDP-43M337V/M337V mice show motor dysfunction with aging. Measurement of rotarod and clasping scores as well as body weights revealed no difference in those scores between TDP-43M337V/M337V and wild-type mice until 18 months old (Fig. 1h, i, Additional file 1: Figure S2).

The present study demonstrates that homozygous TDP-43M337V mice generated by CRISPR/Cas9 show splicing deregulation of some TDP-43 target mRNAs without apparent deterioration in motor function and pathology until 20 months old. Recently, homozygous TDP-43Q331K knock-in mice showed a reduced number of parvalbumin-positive interneurons and cognitive dysfunction with phenotypic heterogeneity [9]. Homozygous TDP-43G298S or TDP-43M337V knock-in mice showed very mild denervation of hindlimbs at 2.5 years of age [8]. Besides, heterozygous TDP-43M323K mice, generated by N-ethyl-N-nitrosourea (ENU) random mutagenesis, showed modest neurodegenerative phenotype [10]. These mutant mice uniformly show very mild phenotypes, likely because the 2-years-life span of rodents may be insufficient to induce neurodegeneration derived from splicing deregulation caused by mutant TDP-43.

Our TDP-43M337V/M337V mice showed splicing deregulation of TDP-43 target mRNAs, Kcinp2, Sort1, and Sema3f, suggesting that M337 V mutation causes a gain of function in TDP-43. Gain of TDP-43 function is also suggested in TDP-43Q331K and TDP-43M323K mice [9, 10]. All three missense mutations discussed here are located in the low complexity region at the C-terminal of TDP-43, suggesting that ALS-causing TDP-43 mutations in the C-terminal region may cause gain of TDP-43 function, at least, at an initial disease stage. This point makes a good contrast with the role of N-terminal TDP-43 fragment in dominant-negative function in TDP-43 [14]. In our study, the mRNA and protein levels of TDP-43 were unchanged in TDP-43M337V mice, while they were moderately upregulated in TDP-43Q331K [9]. This difference may explain the more modest phenotype of our TDP-43M337V mice.

All knock-in mice carrying ALS-linked missense mutations in TDP-43 do not show robust TDP-43 pathology even in homozygous mutant mice. Perhaps, additional conformational change of TDP-43 protein may be needed to develop TDP-43 pathology. Finally, our results from TDP-43M337V knock-in mice further strengthen the findings that mutations at the C-terminal region of TDP-43 likely cause a gain of TDP-43 splicing function at an initial stage of the disease, which may be followed by the loss of TDP-43 function due to a loss of TDP-43 proteins from nuclei.