Background
Recent genetic evidence has established the linkage between the neurological disorders amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) [
1‐
5]. The key pathological feature that is shared between ALS and FTD is the cytoplasmic aggregation and nuclear clearance of an RNA binding protein called transactive response DNA binding protein 43 kDa (TDP-43,
TARDBP) [
6]. Since the discovery of TDP-43, a number of other human diseases have also been characterized with TDP-43 pathology [
7‐
12]. Of particular interest, however, is the pathogenesis of inclusion body myositis (IBM), which is believed to be primarily myogenic rather than neurogenic [
13,
14]. To understand the mechanisms of disease pathogenesis that will inform appropriate therapeutic strategies, it will be critical to determine whether the pathways affected by TDP-43 proteinopathy differ between neurons and myocytes.
We have recently found that TDP-43 plays a major role in repressing nonconserved cryptic exons [
15]. These cryptic exons are regions of the genome that are normally skipped by the spliceosome due to the presence of adjacent UG microsatellite repeats, the consensus binding site of TDP-43. When TDP-43 function is lost, these cryptic exons become activated and often lead to nonsense-mediated decay (NMD) of the associated mRNA. In our previous report [
15], we utilized an
in vitro inducible stem cell model of TDP-43 deletion. However, we have yet to establish the cell type-specific cryptic exons that arise
in vivo. Here, we generated conditional Tdp-43 knockout mice to specifically delete Tdp-43 in excitatory neurons and skeletal myocytes. We found that Tdp-43 cryptic exons are highly variable between cell types and that many distinct pathways are altered—novel findings that have mechanistic and therapeutic implications for human diseases exhibiting TDP-43 proteinopathy.
Methods
Mouse breeding strategy
We crossbred our conditional Tardbp knockout mice (Tardbp
F/+) with CamKIIa-Cre transgenic mice to obtain a cohort of CamKIIa-Cre;Tardbp
F/+ mice which were subsequently crossbred to Tardbp
F/+ mice to generate the final cohort: CamKIIa-Cre;Tardbp
+/+, CamKIIa-Cre;Tardbp
F/+ and CamKIIa-Cre;Tardbp
F/F mice. A similar strategy was applied when crossbreeding the MLC-Cre driver line to Tardbp
F/+ mice. All mouse experiments were approved by the Johns Hopkins University Animal Care and Use Committee.
Histology and immunohistochemistry
For the CamKIIa-Cre line, wildtype and floxed mice were anaesthetized and perfused with 4% paraformaldehyde. Brains were embedded into paraffin, cut into 10 μm sections and stained according to standard protocols. For the MLC-Cre line, wildtype and floxed mice were anaesthetized and sacrificed by decapitation. Muscle tissue was then rapidly dissected and flash frozen in liquid nitrogen cooled isopentane. Frozen cryosections were cut at 10 μm thickness and stained according to standard protocols. Immunoreactivity was visualized using the Vectastain ABC Kit and diaminobenzidine peroxidase substrate (Vector Laboratories). Images were obtained using Olyumpus BX53 microscope.
Immunoblot analysis
For the CamKIIa-Cre line, wildtype and floxed mice were anaesthetized and sacrificed by decapitation. Brain tissue was then rapidly dissected and manually homogenized in RIPA buffer (Sigma) containing an EDTA-free protease inhibitor cocktail (Thermo Scientific). For the MLC-Cre line, wildtype and floxed mice were also anaesthetized and sacrificed by decapitation. Muscle tissue was snap frozen in isopentane cooled with liquid nitrogen, manually ground into a powder, and then homogenized in RIPA buffer with protease inhibitor cocktail. Protein concentration was determined using the BCA assay (Pierce). Proteins were resolved using the NuPAGE 4-12% Bis-Tris Gel (Novex) with NuPAGE MES SDS Running Buffer (Novex), and transferred to PVDF membrane (Millipore) with NuPAGE Transfer Buffer (Invitrogen).
The following antibodies were used for protein blots, immunofluorescence, and immunohistochemical analyses: rabbit anti-TDP-43 (Proteintech 10782-2-AP and 12892-1-AP), anti-NeuN monoclonal antibody (Chemicon), anti-GAPDH monoclonal antibody (Sigma), Alexa Fluor 488-conjugated Donkey anti-Guinea Pig IgG (H + L) antibody (Jackson ImmunoResearch), Alexa Fluor 594- and 647-conjugated Donkey anti-goat and anti-rabbit IgG (H + L) antibodies (Life Tech.).
Total RNA was extracted from hippocampi of 3 month old female
CamKIIa-Cre;
Tardbp
F/F (neuronal knockout) and littermate control mice (
CamKIIa-Cre;
Tardbp
+/+) using TRIzol (Life Tech.) and RNeasy Mini kits (Qiagen). Total RNA from 2 month old male
MLC-Cre;
Tardbp
F/F (skeletal muscle knockout) and littermate control mice (
MLC-Cre;
Tardbp
+/+) was also extracted in a similar manner. For the
CamKIIa-Cre line, 3 control brains and 3 knockout brains were analyzed and all mice were female. For the
MLC-Cre line, 2 control quadriceps and 2 knockout quadriceps were analyzed and all mice were male. 100-bp paired end RNA-seq libraries were generated using Illumina Tru-seq kits and then sequenced on an Illumina HiSeq 2000. For RT-PCR analysis, total RNA was isolated using RNeasy Mini Kit (Qiagen). cDNA was synthetized using RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific) with random primers. RNA-seq analysis was performed using HISAT [
16] and Cufflinks [
17] software suites and visualized on the UCSC Genome Browser [
18]. Cryptic exons were identified as previously described [
14]. To identify common pathways between species, gene ontology analysis was performed on cryptic exon targets using manual annotation of genes with known functions in combination with the bioinformatics resource DAVID v6.7 [
19].
RT-PCR primers
Ap3b2-Forward | AGCCAGAATATGGCCACGAC | Neuron |
Ap3b2-Reverse | CACTATGATGGGCACACGGA | Neuron |
Camk1g-Forward | CTGGCCAAGATCACAGACTGG | Neuron |
Camk1g-Reverse | CTGTGTAGACACCACGCTCT | Neuron |
Sh3bgr-Forward | GGAGCAGAGGCTTGGATCAC | Muscle |
Sh3bgr-Reverse | AAAGCCCACCACTTCTTGCT | Muscle |
Tns1-Forward | CCTGGTCTATCAGCACTCCG | Muscle |
Tns1-Reverse | GGGCTCCCGATTTCGTTCAT | Muscle |
Discussion
We have found that Tdp-43’s nonconserved cryptic exons vary widely between cell types and affect many pathways that are critical for neuronal and muscle physiology. This suggests that in human disease, myogenic and neurogenic TDP-43 proteinopathies exhibit cell type-specific cryptic exons that could influence disease progression in unique ways. Although our RNA-seq data are based on a limited number of samples, future analysis to increase sample sizes would strengthen our findings. Identifying the cryptic exons that are specific to human neurons or myocytes will also help clarify the selective vulnerability associated with diseases such as IBM and ALS-FTD.
While it remains to be proven whether TDP-43 loss of function is a central driver of human disease, our data demonstrates that within neurons and myocytes, TDP-43 is the major splicing repressor for numerous nonconserved cryptic exons. In human disease, dysregulation of Tdp-43 function may impair other neuronal functions beyond mRNA splicing such as axonal trafficking, hyperexcitability, and liquid-liquid phase separation [
31‐
34]. Nevertheless, mouse models of Tdp-43 have demonstrated that constitutive deletion of
Tardbp results in embryonic lethality [
24,
25,
35,
36]. Conditional depletion of
Tardbp in adult mice also leads to metabolic deficits and premature death [
20] and significant neurodegeneration [
26,
37,
38]. Together, these studies demonstrate the importance of Tdp-43 for cell survival.
The current work clarifies the mechanisms of toxicity that underlie Tdp-43 loss of function in the context of cryptic exon repression [
15], a finding that has been replicated by other groups [
39‐
41]. Our results suggest that cryptic exons disrupt unique pathways depending on cellular context, although future studies are needed to understand the degree to which these splicing errors contribute to cell death. Furthermore, TDP-43 belongs to a family of proteins that repress cryptic exons, suggesting that these splicing factors perform a general function in the cell to maintain splicing fidelity [
42]. Thus, loss of TDP-43 splicing repression contributes to cell death and the pathways affected by cryptic exon incorporation are likely to be relevant for disease pathogenesis.
The question then becomes, how do we prevent incorporation of nonconserved cryptic exons? Therapeutic strategies that aim to directly interfere with cryptic exon splicing (e.g. anti-sense oligonucleotides) will be difficult to envision due to the sizeable number of nonconserved cryptic exons per cell. Furthermore, because nonconserved cryptic exons are different between mouse and human, testing splicing modulators for human cryptic exons in animal models is essentially impossible. However, the general splicing repression function of TDP-43 is conserved. Thus, it may be possible to use mouse models of TDP-43 deletion to specifically test therapeutic strategies that rescue TDP-43 mechanism of action rather than directly targeting individual cryptic exons. One strategy would employ gene therapy to introduce designer splicing factors—chimeric proteins that would couple the UG binding domain of TDP-43 with non-aggregating splicing repressor domains [
15]—into neurons or muscles. In principal, this approach would repress most of TDP-43’s nonconserved cryptic exons in a manner that would be species-independent.
If neuron loss or skeletal muscle degeneration can be attenuated, such a therapeutic strategy could be rapidly translated into the clinic. Moreover, the observation that cryptic exons can occasionally introduce inframe insertions into mRNA suggests that certain human TDP-43 cryptic exons could represent biomarkers for human disease. We envision the development of specific antibodies to detect neoantigens introduced by human inframe cryptic exons in CSF or blood from patients, serving as either diagnostic biomarkers or tools to monitor the efficacy of treatments in future clinical trials.
Conclusions
This study demonstrates that Tdp-43 represses a unique set of cryptic exons, depending on cellular context. Thus, the pathways impacted by Tdp-43 loss-of-function and cryptic exon incorporation are likely distinct for each cell type. These results have important implications for human disease, given that Tdp-43 proteinopathy can manifest in various tissues.
Acknowledgments
We thank V. Nehus for technical assistance. CamKIIaCre and MLC-Cre mice were kindly gifted, respectively, by P. Worley (Johns Hopkins University School of Medicine) and S. Burden (New York University School of Medicine).
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (
http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.