TDP-43 transgenic mice develop spastic paralysis and neuronal inclusions characteristic of ALS and frontotemporal lobar degeneration

Abstract

Neuronal cytoplasmic and intranuclear aggregates of RNA-binding protein TDP-43 are a hallmark feature of neurodegenerative diseases such as amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD). ALS and FTLD show a considerable clinical and pathological overlap and occur as both familial and sporadic forms. Though missense mutations in TDP-43 cause rare forms of familial ALS, it is not yet known whether this is due to loss of TDP-43 function or gain of aberrant function. Moreover, the role of wild-type (WT) TDP-43, associated with the majority of familial and sporadic ALS/FTLD patients, is also currently unknown. Generating homozygous and hemizygous WT human TDP-43 transgenic mouse lines, we show here a dose-dependent degeneration of cortical and spinal motor neurons and development of spastic quadriplegia reminiscent of ALS. A dose-dependent degeneration of nonmotor cortical and subcortical neurons characteristic of FTLD was also observed. Neurons in the affected spinal cord and brain regions showed accumulation of TDP-43 nuclear and cytoplasmic aggregates that were both ubiquitinated and phosphorylated as observed in ALS/FTLD patients. Moreover, the characteristic approximately 25-kDa C-terminal fragments (CTFs) were also recovered from nuclear fractions and correlated with disease development and progression in WT TDP-43 mice. These findings suggest that approximately 25-kDa TDP-43 CTFs are noxious to neurons by a gain of aberrant nuclear function.

Fig. 1.

hTDP-43 overexpression in mice leads to a dose-dependent motor neuron disease phenotype. (A) Schematic drawing of the Thy-1 expression vector used to generate transgenic TDP43^WT lines. (B) mRNA expression levels in brain measured by Q-RT-PCR for different transgenic lines (**P < 0.01). (C) Abnormal limb reflex in TDP43^WT lines at different ages. (D) A log-log correlation of hTDP-43 dosage in TAR4/4, TAR6/6, and TAR4 mice with the age at which the abnormal limb reflex was first noticed. (E) Disturbed footprint pattern in TAR4/4 mice with markedly wide-based stance, small stride, and frequent off-line stumbling in transgenic mice (arrow). (F) Significant 2-fold decrease in the stride of both forelimb (red) and hindlimb (black) (*P < 0.05). Data are presented as mean ± SD (G) Approximately 2.5-fold reduced performance of TAR4/4 mice on accelerating rotarod compared with Ntg littermates (**P < 0.01). (H) End-stage paralysis in TAR4/4 mice (Movie S1). (I) Survival curve of TAR 4/4 mice showed an average survival of 24 days, and 6.7 months for TAR6/6 mice.

Fig. 2.

NCI and NII brain pathology in TDP43^WT mice. (A) Ubiquitin pathology in layer V of the motor and somatosensory cortex of TAR4/4 but not in Ntg mice. A2 and A3 are higher magnifications. (Scale bars: 20 μm.) (B and C) Colabeling of ubiquitin-positive aggregates with p409/410 and human-specific TDP-43 antibodies showing NCIs (B, arrowheads) and NIIs (C, arrows). (Scale bars: 5 μm.) (D) Increased GFAP staining in cortical layer V of motor cortex that is absent in Ntg mice. (Scale bars: 20 μm.) (E) Quantification of the neuronal loss showed ≈30% reduction in cortical layer V of TAR4/4 mice compared with Ntg littermates (**P < 0.01).

Fig. 3.

Spinal cord NCI and NII pathology in TDP43^WT mice. (A) Ubiquitin pathology present in spinal motor neurons of TAR4/4 but absent in Ntg mice. (Scale bars: 20 μm.) (B1–B4) Lumbar spinal cord showing NIIs immunostained with phosphospecific p409/410 TDP-43 antibody that are remarkably similar to the round NIIs observed in FTLD-TDP patients (B5). B1 shows frequently observed occurrence of more than one NII within a neuron (arrows) and of phospho-TDP-43-positive granulovacuoles in the cytoplasm (arrowhead). (Scale bars: 5 μm.) (C and D) Colabeling of p409/410 TDP-43 with human TDP-43 and ubiquitin antibodies. Arrow points to a NII, and asterisk in C denotes TDP-43 clearing in the NII-bearing neuron. Note the NCI in D (arrowhead) is negative for p409/410 TDP-43 reactivity. (Scale bars: 5 μm.) (E and F) Ultrastructural microscopy of lumbar spinal cord showing two non-membrane-bound, loosely packed NIIs (asterisks) in E and dystrophic neurites observed in the ventral exit zone in F. (Scale bars: 500 nm.) (G) GFAP immunohistochemistry showing increased astrogliosis that is absent in Ntg mice. (Scale bars: 10 μm.) (H) Quantification of the neurons showed ≈25% reduction in motor neurons in lumbosacral spinal cord of TAR4/4 mice compared with Ntg littermates (**P < 0.01).

Fig. 4.

TDP-43 processing in TDP43^WT mice. (A) Immunoblot of brain lysates of TAR4/4 mice with a non-species-specific TDP-43 antibody showing increased full-length (FL) TDP-43 in the cytoplasmic but not in the nuclear fraction compared with Ntg littermates (see short exposed blots and quantification in Fig. S6). Approximately 35-kDa CTFs are decreased with increasing disease severity in both nuclear and cytoplasmic fractions (Fig. S6), whereas the ≈25-kDa CTFs were only recovered from the nuclear fraction and correlated with disease progression. GAPDH was used as a marker for the cytoplasmic fraction and lamin A/C for the nuclear fraction. Asterisk at ≈34 kDa likely represents a known splice form of endogenous mouse TDP-43. (B) 35-kDa and 25-kDa bands stained with a C-terminal TDP-43 antibody showing the same trend of 35-kDa decrease and 25-kDa CTFs increase with increasing disease severity. The 25-kDa CTF band is quantified below (***P < 0.001; error bars, SD). (C) Sequential extraction detected with a non-species-specific TDP-43 antibody showed the 35-kDa CTFs were mainly recovered from the LS fraction (and a small proportion in the Tx fraction), whereas the 25-kDa CTFs were recovered from the Tx fraction. LS, low salt; Tx, Triton X-100; SARK, sarkosyl. (D) Brain lysate of an FTLD-TDP patient showing that 25-kDa fragments are predominantly recovered from the nuclear fraction and in the Tx fraction in sequential extractions. #, non-disease-specific band. See also full sequential blot in Fig. S7. (E) Transfection of H4 and SH-SY5Y cells with 25-kDa TDP-43 CTF (Δ1–219) with a C-terminal GFP tag (TDP-25-GFP) showing that 25-kDa CTFs (green) predominantly localized in the nucleus (α-tubulin, red). (Scale bars: 10 μm.)