TDP-43 Is a Developmentally Regulated Protein Essential for Early Embryonic Development*

TDP-43 is a DNA/RNA-binding protein implicated in multiple steps of transcriptional and post-transcriptional regulation of gene expression. Alteration of this multifunctional protein is associated with a number of neurodegenerative diseases including amyotrophic lateral sclerosis and frontotemporal lobar degeneration with ubiquitin positive inclusions. Whereas a pathological link to neurodegenerative disorders has been established, the cellular and physiological functions of TDP-43 remain unknown. In this study, we show that TDP-43 is a nuclear protein with persistent high-level expression during embryonic development and with progressively decreased protein levels during postnatal development. In mice where the TDP-43 gene (Tardbp) was disrupted using a gene trap that carries a β-galactosidase marker gene, heterozygous (Tardbp+/−) mice are fertile and healthy, but intercrosses of Tardbp+/− mice yielded no viable homozygotic null (Tardbp−/−) mice. Indeed, Tardbp−/− embryos die between 3.5 and 8.5 days of development. Tardbp−/− blastocysts grown in cell culture display abnormal expansion of their inner cell mass. The pattern of β-galactosidase staining at E9.5 Tardbp+/− embryos is predominantly restricted to the neuroepithelium and remains prominent in neural progenitors at E10.5–12.5. TDP-43 is detected in spinal cord progenitors and in differentiated motor neurons as well as in the dorsal root ganglia at E12.5. β-Galactosidase staining of tissues from adult Tardbp+/− mice shows widespread expression of TDP-43, including prominent levels in various regions of the central nervous system afflicted in neurodegenerative disorders. These results indicate that TDP-43 is developmentally regulated and indispensible for early embryonic development.

TDP-43 is a DNA/RNA-binding protein implicated in multiple steps of transcriptional and post-transcriptional regulation of gene expression. Alteration of this multifunctional protein is associated with a number of neurodegenerative diseases including amyotrophic lateral sclerosis and frontotemporal lobar degeneration with ubiquitin positive inclusions. Whereas a pathological link to neurodegenerative disorders has been established, the cellular and physiological functions of TDP-43 remain unknown. In this study, we show that TDP-43 is a nuclear protein with persistent high-level expression during embryonic development and with progressively decreased protein levels during postnatal development. In mice where the TDP-43 gene (Tardbp) was disrupted using a gene trap that carries a ␤-galactosidase marker gene, heterozygous (Tardbp ؉/؊ ) mice are fertile and healthy, but intercrosses of Tardbp ؉/؊ mice yielded no viable homozygotic null (Tardbp ؊/؊ ) mice. Indeed, Tardbp ؊/؊ embryos die between 3.5 and 8.5 days of development. Tardbp ؊/؊ blastocysts grown in cell culture display abnormal expansion of their inner cell mass. The pattern of ␤-galactosidase staining at E9.5 Tardbp ؉/؊ embryos is predominantly restricted to the neuroepithelium and remains prominent in neural progenitors at E10.5-12.5. TDP-43 is detected in spinal cord progenitors and in differentiated motor neurons as well as in the dorsal root ganglia at E12.5. ␤-Galactosidase staining of tissues from adult Tardbp ؉/؊ mice shows widespread expression of TDP-43, including prominent levels in various regions of the central nervous system afflicted in neurodegenerative disorders. These results indicate that TDP-43 is developmentally regulated and indispensible for early embryonic development.
The gene coding TDP-43, 2 or TAR DNA-binding protein 43 (Tardbp), is highly conserved throughout evolution and is found in all higher eukaryotic species including distant species Drosophila melanogaster, Xenopus laevis, and Caenorhabditis elegans (1,2). In humans, Tardbp is located at the chromosomal locus 1p36. 22 and is comprised of six exons, five of which encode a ubiquitously expressed, predominantly nuclear, 43-kDa protein that contains two RNA recognition motifs and a glycine-rich C-terminal domain, characteristic of the heterogeneous nuclear ribonucleoprotein class of proteins (3). The RNA recognition motif domains of TDP-43 are highly homologous among species; however, the glycine-rich sequence varies significantly among all species, reflecting species-specific functions in the different organisms.
TDP-43 has been implicated in the regulation of gene transcription, pre-mRNA splicing, mRNA stability, and mRNA transport (4). It was first identified to bind the TAR DNA of the human immunodeficiency virus 1 long terminal repeat region. Both in vitro and in vivo experiments showed that TDP-43 represses human immunodeficiency virus 1 proviral gene expression (5). Later, it was shown to enhance exon skipping of the cystic fibrosis transmembrane conductance regulator exon 9 through binding to a (UG) m (U) n motif near the 3Ј splice site of the cystic fibrosis transmembrane conductance regulator intron 8 (6). TDP-43 was also shown to be involved in splicing of the apolipoprotein A-II (7) and survival of motor neuron (8) genes. In addition, TDP-43 has been implicated in regulation of mRNA biogenesis (9) and shown to be localized to sites of mRNA transcription and processing in neurons (10). As the glycine-rich domain of TDP-43 has been shown to mediate interactions with other heterogeneous nuclear ribonucleoprotein proteins, the low homology of this particular domain may afford a multitude of interactions that allows for diverse biological functions (11).
TDP-43 has been identified as the primary protein of neuronal and glial inclusions of sporadic and familial frontotemporal lobar degeneration with ubiquitin positive inclusions (FTLD-U), as well as in sporadic and the majority of familial amyotrophic lateral sclerosis (ALS) cases (12,13). TDP-43, normally observed in the nucleus, is found in pathological inclusions mostly in the cytoplasm and in some cases accumulates in dense deposits in the nucleus. The inclusions consist prominently of TDP-43 C-terminal fragments of ϳ20 -25 kDa. Both full-length and C-terminal fragments of TDP-43 undergo abnormal phosphorylation and ubiquitination in diseased states (13). More recently, TDP-43 inclusions are found in patients with Alzheimer and Parkinson diseases implying a common mechanism of TDP-43-related pathologies (14,15). The relevance of TDP-43 and its pathological function was supported by the discovery of autosomal dominant mutations in Tardbp in patients with familial and sporadic ALS (16,17). Most of these mutations are found in the C-terminal region of TDP-43, supporting an important function of this region and involvement of the C-terminal derivatives in disease onset and progression (18,19). Furthermore, mice overexpressing the familial A315T mutation of TDP-43 result in ubiquitin aggregates with degeneration of cortical projection neurons and spinal motor neurons similar to that observed in ALS and FTLD-U (20). Alternatively, a loss of TDP-43 function rather than the production of C-terminal fragments may lead to TDP-43-associated pathologies. It has been reported that TDP-43-null flies display defi-cient locomotor behavior and have reduced life span and defects at the neuromuscular junction (21). Knockdown of TDP-43 by small interfering RNA in cellular models reveals disruption of cell proliferation (22), inhibition of differentiation, and neuronal cell death (23).
Whereas a pathological link to neurodegenerative disorders has been established, the cellular and physiological functions of TDP-43 remain unknown. As a first step toward the elucidation of the fundamental role of TDP-43 in vivo, we disrupted the Tardbp gene in mice and found that TDP-43 is necessary for embryonic development: no Tardbp Ϫ/Ϫ embryos survived after 3.5 days post coitus (dpc), whereas Tardbp ϩ/Ϫ mice were healthy and indistinguishable from their control littermates. We also found that TDP-43 is present throughout embryonic development and is expressed predominantly in the neuroepithelium and neural progenitors of the developing embryo. We show nuclear expression of TDP-43 in embryonic stem (ES) cell outgrowths and in hippocampal neurons and glia. Moreover, we observed in postnatal brains that TDP-43 holoprotein is developmentally regulated. In adult mice, TDP-43 shows widespread expression, including prominent levels in various regions of the central nervous system afflicted in neurodegenerative disorders.
Genotyping-Genomic DNA from ear biopsies, yolk sacs, and embryos were lysed in Quick Lysis Buffer (50 mM NaCl, 10 mM Tris-HCl, pH 8.3, 0.2% Tween 20, 0.4 mg/ml of proteinase K) at 55°C for 1 h and then 95°C for 10 min. The PCR contained primer-a (5Ј-GCTGGGTCTGTGGTGCACGTC-3Ј) and prim- Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control. D, identification of a single gene trap insertion using Southern blot. Genomic DNA was digested with EcoRV or StuI and hybridized with a 32 P-labeled neomycin phosphotransferase II probe and the predicted ϳ6or ϳ8-kb band, respectively, was detected in Tardbp ϩ/Ϫ mice only. E, genotyping results of offspring from intercrossed Tardbp ϩ/Ϫ mice.
er-c (5Ј-GCTCATGCTCCTGTCTCCCTCCTTC-3Ј), which correspond to sequences upstream and downstream, respectively, of the insertion point and primer-b (5Ј-GTACCGCACTGCCG-GTTTCCTCCACC-3Ј), which is specific to pGT1LxF. Genomic DNA (100 ng) and primers-a, -b, and -c (300 nM each) were placed in standard Taq buffer supplemented with 1 M betaine, 3.3% dimethyl sulfoxide, 1.5 mM MgCl 2 , 0.1 mg/ml of bovine serum albumin, 0.2 mM deoxynucleoside triphosphates, and 1.25 units of Taq polymerase (New England Biolabs) for 10 min at 94°C. After enzymatic amplification for 35 cycles, the PCR products were resolved on 2% agarose gel in 1ϫ Tris acetate-EDTA buffer. Primers-a and -b amplify a 600-bp band from mutant allele; primers-a and -c amplify a 400-bp band from the wild-type (WT) allele. Southern Blot-Genomic DNA was purified from 6-week-old mice (F4 generation), digested overnight with EcoRV or StuI, and analyzed by Southern blot with a 32 P-labeled neomycin phosphotransferase II probe.
LacZ Staining of Tissues and Embryos-Mice were anesthetized with avertin and perfusion-fixed with 4% paraformaldehyde (PFA). Tissues were harvested and post-fixed in 4% PFA overnight at 4°C, washed with PBS, immersed in 30% sucrose for 24 h at 4°C, and frozen in optimal cutting temperature compound for sectioning. ␤-Galactosidase activity was assessed by incubating 30 -50-m thick sections with LacZ staining solution (1.0 mg/ml of X-gal (Invitrogen), 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 2 mM MgCl 2 ) for 24 h at 37°C and fixed in 4% PFA. After counterstaining with nuclear fast red, the sections were examined and photographed with a Zeiss SteREO Discovery version 12 microscope.
Whole embryos were harvested and post-fixed in LacZ fixing solution (1% PFA, 0.2% glutaraldehyde, 2 mM MgCl 2 , 5 mM EGTA, pH 8.0) for 15 min at room temperature, washed with PBS containing 2 mM MgCl 2 , and incubated in LacZ staining solution for 30 min. Embryos were then post-fixed with 4% PFA for 30 min, immersed in 30% sucrose for 24 h at 4°C, and frozen in optimal cutting temperature compound for sectioning.
For 125 I quantification of proteins, proteins were processed as described above, transferred to a polyvinylidene difluoride membrane, and blocked in 5% milk and 5% goat serum for 1 h at room temperature. Membranes were then incubated with primary antibodies TDP-43 (748C) (1:1000) and VCP (1:500) over-  night at 4°C, and then at room temperature for 2 h. Membranes were washed 3 times for 30 min with Tris-buffered saline-Tween 20 (TBS-T), followed by incubation with 125 I-labeled anti-rabbit IgG (1:1000) diluted in 5% milk/TBS-T with 0.05% NaN 3 and incubated overnight. The blots were washed 5 times for 5 min with TBS-T and exposed to a phosphorimager screen for 3 days before analysis.
Quantitative PCR Analysis-Primers were designed using Primer Express software (Applied Biosystems) based on GenBank TM sequence data. Primer sequences were BLASTed against the NCBI mouse genomic sequence data base to ensure unique specificity. Primer sequences for TDP-43 include 5Ј-CGTGTCTCAGTGTATGAGAGGAGTC-3 and 5Ј-CTG-CAGAGGAAGCATCTGTCTCATCC-3Ј. Reverse transcription quantitative-PCR (10 l) contained 20 ng of cDNA, 150 nM of each primer, and 5 l of SYBR FastGreen PCR Master Mix (Applied Biosystems). All reactions were performed in triplicate on an Applied Biosystems Prism 7500Fast sequence detection system, and relative mRNA levels were calculated by the comparative threshold cycle method using U36B primers (5Ј-TGGGCATCACCACGAAAAT-3Ј and 5Ј-TATCAGCTGCA-CATCACTCAGAATT-3Ј) as the internal controls.  . 30 and 31). B, transverse sections of whole embryo in series. There is strong staining in the neuroepithelium (ne) of the neural tube and to a lesser extend the first branchial arch (fb), second branchial membrane (sb), atrial heart chamber (ac), vitelline artery (va), bulbus cordis heart region (bc), left and right pericardio-peritoneal canal (ppc), hindgut (hg), and somite (s). Also annotated are the neural folds in caudal neuropore (cn), neural lumen (nl), and peritoneal cavity (pc). The developing heart is indicated by arrows. Scale bars represent 200 m.
Analysis of Tardbp-deficient ES Cells-Female Tardbp ϩ/Ϫ mice were superovulated by injection with 7 units of pregnant mare serum gonadotropin, followed 48 h later by injection of 7 units of human chorinonic gonadotropin and then mated with stud male Tardbp ϩ/Ϫ mice. Fertilized eggs were collected 3.5 days post-impregnation and genotyped as described above or placed in 12-well cell culture dishes containing coverslips coated with 0.1% gelatin. Blastocysts were cultured for 7 days in ES cell medium (knock-out Dulbecco's modified Eagle's medium supplemented with 15% fetal bovine serum, penicillin (100 units/ml)/streptomycin, 2 mM L-glutamine, 1ϫ minimal essential medium non-essential amino acids, 100 mM ␤-mercaptoethanol).

RESULTS AND DISCUSSION
TDP-43 Protein Is Developmentally Regulated-Proteins relevant to central nervous system disease pathogenesis, for example, the Alzheimer disease-associated amyloid precursor protein (27), are often developmentally regulated or play important developmental roles. Taking cues from this observation, we examined subcellular localization and protein levels of the TDP-43 protein in developing embryos and postnatal brains of WT mice. Embryos harvested at the two-and six-cell stage (Fig.  1A, panel i) and at the blastocyst stage (E3.5, Fig. 1A, panel ii) from intercrossed WT mice were immunoreactive for the TDP-43 protein. At these stages of development TDP-43 was confined to the nucleus of cells. The level of TDP-43 holoprotein was maintained in the developing embryos (Fig. 1B), but in postnatal brains the expression gradually decreased after birth (Fig. 1, C and D). The progressive postnatal decrease of TDP-43 protein does not result from reduced mRNA transcripts. mRNA levels stay relatively constant in the late embryonic stages and several days after birth, then modestly increase in the adult brains (data not shown). Co-staining of hippocampal neurons and glia from mice with anti-TDP-43 and anti-p70 S6K , a duo-specific nuclear and processing body marker (28), show nuclear staining of TDP-43 (Fig. 1E).
TDP-43 Is Required for Early Embryogenesis-Having observed that TDP-43 is a developmentally regulated protein, we next determined the developmental and physiological function of TDP-43 in mice wherein Tardbp was disrupted using a gene trap (Fig. 2A). The gene trap vector, pGT1Lxf, creates an in-frame fusion between the 5Ј exons of the trapped gene and a ␤-geo (a fusion of ␤-galactosidase and neomycin phosphotransferase II) reporter gene. The insertional mutation in the RRB030 Tardbp ϩ/Ϫ mice was located within intron-2 of Tardbp and a genotyping approach was established (Fig. 2B). The insertional mutation, which is predicted to generate a fusion  FEBRUARY 26, 2010 • VOLUME 285 • NUMBER 9 transcript containing exons 1-2 of Tardbp and ␤-geo, was detected around ϳ175 kDa in the Tardbp ϩ/Ϫ targeted ES cell line (Fig. 2C). The fusion protein contains the first 79 translated amino acids of TDP-43, which is not predicted to produce a functional protein. Southern blot analysis shows a ϳ6 or ϳ8 kb band from genomic DNA digested with EcoRI or StuI, respectively, in heterozygous mice only, indicating no additional genes are trapped in these mice (Fig. 2D).

Developmental Role of TDP-43
Tardbp ϩ/Ϫ mice were indistinguishable from their control littermates; heterozygotes had normal body weight, growth rate, appearance, and fertility (data not shown), and there were no premature postnatal deaths in these mice (Fig. 2E). Furthermore, there were no gross tissue abnormalities in any tissue observed, including the brain, spinal cord, lung, ovaries, testis, kidney, heart, and liver up to 6 months of age (see Fig. 7). Quantitative assessment of the protein levels of TDP-43 in tissues from Tardbp ϩ/Ϫ mice using 125 I-labeled secondary antibody and densitometry showed that TDP-43 levels were unchanged between heterozygous and WT mice ( Fig. 3A and data not shown). Analysis of mRNA showed no significant decrease in TDP-43 mRNA in brain (Fig. 3B, n ϭ 3) and other tissues including testis, liver, heart, spinal cord, and lung of Tardbp ϩ/Ϫ mice (data not shown). The lack of changes in mRNA and protein levels suggests a compensatory mechanism, perhaps by increased TDP-43 mRNA stability or by an increased rate of transcription from the WT allele, or by post-translational regulation. The compensated levels of TDP-43 may explain why the Tardbp ϩ/Ϫ mice do not display any overt phenotype.
Mice heterozygous for the non-functional Tardbp allele were intercrossed, but no homozygous mice were identified of 216 pups born, suggesting embryonic lethality (Fig. 2E). Therefore, embryos were genotyped at different time points during development to determine the stage of lethality. At 8.5, 9.5, and 12.5 dpc no Tardbp Ϫ/Ϫ embryos survived (Fig. 4A). However, genotyping of blastocysts at 3.5 dpc revealed that Tardbp Ϫ/Ϫ embryos were present at the expected Mendelian frequency (Fig. 4A) and microscopy revealed normal morphology (data not shown). These results indicate Tardbp Ϫ/Ϫ embryos die between day 3.5 and 8.5 post-fertilization, implying no functional redundancy of TDP-43 with other proteins, including other heterogeneous nuclear ribonucleoproteins. Death of Tardbp Ϫ/Ϫ embryos suggests that TDP-43 may be essential for transcription or mRNA splicing or processing, all of which are critical for developmental stages of implantation, gastrulation, and early organogenesis.
To identify the effect of the Tardbp null mutation in ES cell development, blastocysts from heterozygous intercrossed mice were harvested at E3.5 and cultured in vitro for 7 days. Heterozygous and WT embryos grew at similar rates, attaching to the coverslip after 2 days, subsequently expanding their inner cell mass (ICM) and growing as a mound on top of the trophoblast giant cells (Fig. 4B). The expanded ICM of the heterozygous and WT embryos were all immunoreactive for TDP-43 FIGURE 7. TDP-43 expression in adult tissues. ␤-Galactosidase expression in adult Tardbp ϩ/Ϫ mice was observed in (i) whole brain with prominent staining in the neocortex (NC), (ii) cerebellar cortex staining in the granular cell layer (G) and Purkinje cells (P), and (iii) in the hippocampus (shown are granular dentate gyrus (DG) as well as CA1 and CA3 regions). There is also distinct staining in the (iv) olfactory bulb glomerulus (Gl). Staining is observed throughout the (v) spinal cord mostly in the dorsal horn (DH) and to a lesser extent the ventral horn (VH). Staining is in the bronchiole (B) region of the (vi) lung and throughout the (vii) ovary, particularly in the oocytes, as well as in the (viii) testis, at the periphery of the seminiferous tubules (ST) where Sertioli cells (SC) are found and with some interstitial staining of Leydig (L) cells. Staining in (ix) kidney is concentrated in the (C) cortical region (shown is the medulla (M)). Moderate levels of expression were in the (x) liver (shown is the central vein (CV)) and (xi) heart. Scale bars represent 100 m. (Fig. 4B). Homozygous embryos showed signs of aberrant development: these embryos attached to the coverslip, but their ICM did not expand leaving behind a monolayer of giant cells with large nuclei (Fig. 4B). These data suggest that Tardbp-null embryos experience developmental defects due to lack of ICM expansion prior to implantation. An independent study recently reported a similar defect in ICM expansion of TDP-43 knock-out embryos (29). Feiguin et al. (21) demonstrated that D. melanogaster lacking TDP-43 display deficits in locomotor behavior and have reduced lifespan. Although mouse and fly display high homology in the RNA recognition motif domains, the C-terminal region has only 33% identity, implicating this region in the less severe phenotype observed in Tardbp-null flies. More consistent with our data are reports that knockdown of TDP-43 in mammalian cell culture models can lead to loss of cell proliferation (22), inhibition of differentiation, and neuronal cell death (23).

TDP-43 Gene Expression Is Prominent in Neuroepithelium and in Central Nervous System Regions Affected in ALS and
FTLD-U-The ␤-galactosidase reporter harbored in heterozygotic embryos and mice provides an ideal tool to examine spatial and temporal expression patterns of Tardbp, which should provide insight into how TDP-43 may contribute to development. At 9.5 dpc, X-gal staining was strongly detected throughout the neuroepithelium (Fig. 5, A and B), which contains neural progenitors cells that eventually form the central nervous system. Staining was also present to a lesser extent in the developing heart, first branchial arch, second branchial membrane, and somite (Fig. 5B). Using TDP-43 and ␤-galactosidase antibodies we show that the expression pattern of endogenous TDP-43 corresponds with that of the TDP-43-␤-galactosidase fusion protein in the developing spinal cord of E10.5 embryos (Fig. 6A). X-gal staining of whole E10.5 embryos reveals prominent staining in the mid and hindbrain and in the developing spinal cord as well as the developing limbs (Fig. 6B). Furthermore, X-gal staining has the same pattern of expression in the developing spinal cord as seen using TDP-43 and ␤-galactosidase antibodies (Fig. 6B). Because the staining X-gal is dependent on an enzymatic reaction, the intensity of staining is greater where there is more enzyme present, unlike antibody staining. Differentiation of the spinal cord occurs at E12.5 (30), X-gal staining is detected in spinal cord progenitors and differentiated motor neurons and in the dorsal root ganglia (Fig. 6C, panel i). X-gal staining is also prominent in progenitors of the developing brain (Fig. 6C, panel ii).
In contrast to the restricted expression pattern in the developing embryos (E9.5-10.5), X-gal staining of tissues from adult Tardbp ϩ/Ϫ mice revealed more widespread expression of TDP-43 (Fig. 7). In the brain, ␤-galactosidase expression was largely present in the hippocampus, cortex, and Purkinje cells of the cerebellum and the glomerular layer of the olfactory bulb (Fig. 7, i-iv). ␤-Galactosidase expression was in the gray matter of the spinal cord, with particularly intense staining in the dorsal horn (Fig. 7v). Staining patterns in the brain and spinal cord are consistent with in situ results from the Allen Brain Atlas. LacZ staining was also observed in the bronchiole of the lung, oocytes of the ovary, as well as in areas of the testes involved in spermatogenesis, cortex of the kidney, hepatocytes of the liver, and throughout the heart (Fig. 7, vi-xi).
Patients with ALS experience loss of motor function and have TDP-43 inclusions in the gray matter of the spinal cord, whereas patients with FTLD-U have TDP-43 inclusions in the cortex and the hippocampal regions of the brain (13). The prominent presence in adult mice of TDP-43 in these affected areas are consistent with their susceptibility to TDP-43 pathologies in humans. Considering embryonic TDP-43 expression is localized to the neuroepithelium, which contains neural progenitors that give rise to interneuron subtypes and motor neurons (31), it is conceivable that mutations or alterations in TDP-43 regulation may alter normal central nervous system development, making individuals more sensitive to TDP-43related pathologies later in life.
Concluding Remarks-In this study we produced and characterized a null mouse model of TDP-43, and analyzed the spatial and temporal distribution patterns of TDP-43. We observed that TDP-43 is required for early embryogenesis and that TDP-43 is prominent in the neuroepithelium in developing embryos and various adult central nervous system regions afflicted in neurodegenerative disorders. Moreover, we observed that TDP-43 protein levels are developmentally regulated in postnatal brain development. These observations raise the possibility that alterations of the regulation and normal function of TDP-43 during animal development may result in the pathological degeneration that occurs in ALS and FTLD-U. This work thus provides basic information for and novel insight into understanding the developmental, physiological, and biochemical roles of TDP-43, setting up a foundation for testing if and how alteration of TDP-43 leads to neurodegenerative diseases.