Divergent phenotypes in mutant TDP-43 transgenic mice highlight potential confounds in TDP-43 transgenic modeling

Abstract

The majority of cases of frontotemporal lobar degeneration and amyotrophic lateral sclerosis are pathologically defined by the cleavage, cytoplasmic redistribution and aggregation of TAR DNA binding protein of 43 kDa (TDP-43). To examine the contribution of these potentially toxic mechanisms in vivo, we generated transgenic mice expressing human TDP-43 containing the familial amyotrophic lateral sclerosis-linked M337V mutation and identified two lines that developed neurological phenotypes of differing severity and progression. The first developed a rapid cortical neurodegenerative phenotype in the early postnatal period, characterized by fragmentation of TDP-43 and loss of endogenous murine Tdp-43, but entirely lacking aggregates of ubiquitin or TDP-43. A second, low expressing line was aged to 25 months without a severe neurodegenerative phenotype, despite a 30% loss of mouse Tdp-43 and accumulation of lower molecular weight TDP-43 species. Furthermore, TDP-43 fragments generated during neurodegeneration were not C-terminal, but rather were derived from a central portion of human TDP-43. Thus we find that aggregation is not required for cell loss, loss of murine Tdp-43 is not necessarily sufficient in order to develop a severe neurodegenerative phenotype and lower molecular weight TDP-43 positive species in mouse models should not be inherently assumed to be representative of human disease. Our findings are significant for the interpretation of other transgenic studies of TDP-43 proteinopathy.

Figure 1. Expression of human TDP-43 in iTDP-43^14A and iTDP-43^8A mice in the postnatal period.

Immunohistochemical detection of hTDP-43 expression in cortex (CTX), hippocampus (HIP) and striatum (STR) in iTDP-43^14A (A) and iTDP-43^8A (B). Western analysis of organs demonstrated specificity of hTDP-43 expression to the brain in both iTDP-43^14A (C) and iTDP-43^8A (D) (SC = spinal cord, He = heart, Lu = lung, Li = liver, Ki = kidney, St = stomach, SM = skeletal muscle, Sp = spleen, Br = brain). (E) Brain weight measurement of non-transgenic (NT) and iTDP-43^14A mice at postnatal stages until 2 months of age (P60) (*p<0.05, **p<0.01, *** p<0.001, unpaired two tailed T-test). (F) Expression of hTDP-43 at indicated postnatal time points for iTDP-43^14A. (G) Expression of hTDP-43 at indicated postnatal time points for iTDP-43^14A (14) compared to iTDP-43^8A (8).

Figure 2. Early degenerative phenotype in iTDP-43^14A mice at P5 in the absence of FTLD-like TDP-43 aggregation.

(A) Monoclonal antibody to human TDP-43 showed expression at P5 remained restricted to previously characterized regions of hippocampus, cortex and striatum. (B) Western blotting of brain lysate of P5 non-transgenic (NT) and iTDP-43^14A demonstrated increased levels of activated caspase 3 in iTDP-43^14A mice. (C) Abundant caspase 3 immunoreactivity in the cortex of iTDP-43^14A mice that was virtually absent in NT mice, suggestive of elevated cell death in iTDP43^14A compared to NT mice. iTDP-43^14A mice were also characterized by increased ubiquitin staining in the upper layers of the cortex compared to NT mice, which upon higher magnification appeared to be completely diffuse and cytoplasmic. (D) Immunohistochemistry for hTDP-43 and p403/404 and immunofluorescence using antibodies to total TDP-43 and p409/410 TDP-43. Significant amounts of cytoplasmic hTDP-43 were observed in iTDP-43 mice (arrowheads). Note that this cytoplasmic staining was also observed in NT mice (arrowheads) with antibodies to total TDP-43 (tTDP-43 Ab1) and TDP-43 phosphorylated at 409/410 (p409/410). Scale bars in D = 50 µm.

Figure 3. Biochemistry of iTDP-43^14A brain lysates at P5.

(A) Western blotting using two antibodies to total TDP-43 (tTDP-43 Ab1 and tTDP-43 Ab2) demonstrated increased levels of low molecular weight species at 35 kDa (arrow) and 25 kDa (arrowhead) in iTDP-43^14A mice relative to NT mice. These species were not observed using antibodies to the C-terminus (405–414) or N-terminus (3–12) of TDP-43. (B) Western blot analysis of high salt (HS), myelin floatation buffer (MFB), sarkosyl (SARK) and urea fractions using antibody to human TDP-43. Note that human TDP-35 (arrow) is present in the urea fraction but is absent from MFB and SARK fractions, N = non-transgenic, T = iTDP-43^14A. (C) Antibody to murine Tdp-43 demonstrated reduction of mTdp-43 in brain compared to NT mice. (D) Quantification of blot in (C), **p<0.01, unpaired two tailed t-test.

Figure 4. Loss of murine Tdp-43 and increased TDP-35 in 25 month (25 M) iTDP-43^8A.

(A) Analysis of transgene expression in 10 M and 25 M cohorts using hTDP-43 and tTDP-43 Ab2 antibodies, β-actin for loading. (B) Densitometric quantitation of TDP-43, TDP-35 and TDP-25 from blot in (A), represented as relative levels of TDP species following sample normalization to β-actin loading controls. (C) mTdp-43 expression in 10 M iTDP-43^8A, β-actin loading control. (D) mTdp-43 expression in 25 M iTDP-43^8A, β-actin loading control. (E) Densitometric quantitation of reduction in mTdp-43 protein levels from blots in (C) and (D), represented as relative levels of mTdp-43 following sample normalization to β-actin loading controls. (F) Quantitative PCR analysis of Tardbp transcript in 2 month old transgenic (T) iTDP-43^17D mice hippocampus (HIP) using Tardbp primers 1, 2 and 3 detected significant reduction in murine Tdp-43 mRNA levels compared to nontransgenic (N) mice and acted as a positive control for our methodology (N = 3 per genotype). No reduction in Tardbp mRNA was detected in the cortex (COR) and hippocampus of 2 month (2 M) iTDP-43^8A mice relative to nontransgenic mice using murine-specific Tardbp primer pair 1 (N = 3 per genotype). No downregulation of Tardbp was observed in the cortex of 25 M iTDP-43^8A (N = 3) relative to control (C) mice (N = 4), using Tardbp primers 1, 2 or 3. Values shown as Normalized Relative Quantity, see Materials and Methods for further details. Control at 25 M consisted of 2 non-transgenic and 2 tTA mice. In graphs (B),(E) and (F): *p<0.05, **p<0.01, unpaired t-test, error bars are SEM.

Figure 5. Neuropathology of 25 month old iTDP-43^8A mice.

Immunohistochemical detection of ubiquitin revealed rare cells bearing increased ubiquitin staining in the cortex of iTDP-43^8A mice (arrows, A) that was absent in NT animals (B, scale bar = 200 µm). Staining was detected in both nucleus and cytoplasm of affected cells (inset in A, scale bar = 10 µm). (C) In iTDP-43^8A animals hTDP-43 was predominantly nuclear, some cells displaying cytoplasmic localization without aggregation. Cytoplasmic localization was observed in NT and iTDP-43^8A mice using antibodies to total TDP-43 (tTDP-43 Ab1) and phosphorylated forms of TDP-43 (p403/404, p409/410).