Phosphorylation of S409/410 of TDP-43 is a consistent feature in all sporadic and familial forms of TDP-43 proteinopathies

Abstract

Accumulation of hyperphosphorylated, ubiquitinated and N-terminally truncated TAR DNA-binding protein (TDP-43) is the pathological hallmark lesion in most familial and sporadic forms of FTLD-U and ALS, which can be subsumed as TDP-43 proteinopathies. In order to get more insight into the role of abnormal phosphorylation in the disease process, the identification of specific phosphorylation sites and the generation of phosphorylation-specific antibodies are mandatory. Here, we developed and characterized novel rat monoclonal antibodies (1D3 and 7A9) raised against phosphorylated S409/410 of TDP-43. These antibodies were used to study the presence of S409/410 phosphorylation by immunohistochemistry and biochemical analysis in a large series of 64 FTLD-U cases with or without motor neuron disease including familial cases with mutations in progranulin (n = 5), valosin-containing protein (n = 4) and linkage to chromosome 9p (n = 4), 18 ALS cases as well as other neurodegenerative diseases with concomitant TDP-43 pathology (n = 5). Our data demonstrate that phosphorylation of S409/410 of TDP-43 is a highly consistent feature in pathologic inclusions in the whole spectrum of sporadic and familial forms of TDP-43 proteinopathies. Physiological nuclear TDP-43 was not detectable with these mAbs by immunohistochemistry and by immunoblot analyses. While the accumulation of phosphorylated C-terminal fragments was a robust finding in the cortical brain regions of FTLD-U and ALS, usually being much more abundant than the phosphorylated full-length TDP-43 band, spinal cord samples revealed a predominance of full-length TDP-43 over C-terminal fragments. This argues for a distinct TDP-43 species composition in inclusions in cortical versus spinal cord cells. Overall, these mAbs are powerful tools for the highly specific detection of disease-associated abnormal TDP-43 species and will be extremely useful for the neuropathological routine diagnostics of TDP-43 proteinopathies and for the investigation of emerging cellular and animal models for TDP-43 proteinopathies.

Figure 1. Characterization of novel mAbs 1D3 and 7A9

a. ELISA performed with serial dilutions of OVA-coupled phospho- and non-phosphopeptides corresponding to amino acids 404-413 of human TDP-43 (range 2-250 ng/ml) demonstrating the high specificity of mAbs 1D3 and 7A9 for the detection of peptides phosphorylated at S409/410. b. Immunoblot analysis of urea fraction from FTLD-U brain separated by 10% SDS-PAGE with phosphorylation-independent polyclonal C-t TDP-43 antibody raised amino acids 394-414 of human TDP-43 and phosphorylation specific mAb 1D3. While C-t TDP-43 shows the physiological TDP-43 band ∼43 kD (arrow) in addition to the pathologic TDP-43 species detected as ∼25 kDa (*), 45 kDa (**) and high molecular smear (***), mAb 1D3 and 7A9 specifically label only the abnormal TDP-43 species. c. Immunoblot analysis of urea samples from FTLD-U dephosphorylated by alkaline phosphatase (AP) treatment show a complete lack of immunoreactivity confirming the specificity of mAb 1D3 against phosphorylated TDP-43 species. d. Immunohistochemistry of hippocampus from FTLD-U brain shows numerous cytoplasmic inclusions as well as nuclear staining of non-inclusion bearing cells with C-t TDP-43 antibody. In contrast, mAb 1D3 and 7A9 only reveal the pathologic inclusions but do not show any nuclear immunoreactivity. Scale bar corresponds to 10 μm.

Figure 2. TDP-43 is phosphorylated at S409/410 in the whole spectrum of sporadic and familial TDP-43 proteinopathies

Representative images from abnormal inclusions in the distinct subgroups of TDP-43 proteinopathies stained with 1D3 are shown. a. sporadic FTLD-U subtype 1: long neuritic profiles; b. sporadic FTLD-U subtype 2: cytoplasmic inclusions; c. sporadic FTLD-U subtype 3: cytoplasmic inclusions and small neuritic profiles; d. familial FTLD-U linked to chromosome 9: more diffuse cytoplasmic staining (“pre-inclusions”); e. familial FTLD-U with GRN mutation: numerous small neurites and cytoplasmic inclusions; f. familial FTLD-U with VCP mutation: small neurites and nuclear inclusions; g. sporadic ALS: skein-like inclusions in spinal motor neurons; h. AD with concomitant TDP-43: small neurites and cytoplasmic inclusions; i. CBD with concomitant TDP-43: abundant white matter pathology. Scale bar corresponds to 50 μm.

Figure 3. Double-label immunofluorescence

Sections from spinal cord (a-c, g-i) and motor cortex (d-f, j-l) from an ALS case were either double-labelled for mAb 1D3 (a, d) and phosphorylation-independent rTDP (b, e) or 1D3 (g, j) and ubiquitin (h, k). Merged images are shown in c, f, i and l. 1D3 and rTDP-43 show a complete overlap of staining of pathologic inclusions, such as skein-like inclusion in motor neuron (a-c), glial cytoplasmic inclusions (arrow in a-c) as well as more granular, diffuse cytoplasmic inclusions (“pre-inclusions”) in the motor cortex (d-f). Note the absence of nuclear staining with 1D3. In contrast with compact neuronal inclusions which are consistently double-labelled by 1D3 and anti-ubiquitin (g-i), 1D3-positive “pre-inclusions” are negative with anti-ubiquitin (j-k). Scale bar corresponds to 20 μm.

Figure 4. Biochemical analysis of S409/410 phosphorylated TDP-43

a. Sarcosyl-insoluble, urea-soluble protein fractions extracted from frontal cortex from distinct FTLD-U subgroups were separated by 15% SDS-PAGE and immunoblotted with mAb 1D3. The detection of C-terminal fragments (up to 4 distinct bands labelled as 1-4) between 20-25 kDa, phosphorylated full-length TDP-43 band (∼45 kD) and a high molecular smear was present in all tested sporadic and familial forms of FTLD-U. However, no clear correlation between distinct banding patterns of C-terminal fragments and FTLD-U subtypes was detectable. b. Urea-samples extracted from motor cortex (MCtx) and spinal cord (sc) samples of ALS cases and immunoblotted with 1D3. A similar biochemical profile as for FTLD-U brains was detectable in MCtx samples. However, in sc samples the predominant band detectable was phosphorylated full-length TDP-43. c. Urea-samples extracted from frontal cortex (fc) or hippocampus (hp) from CBD and AD cases with or without TDP-43 pathology. Note the absence of immunoreactivity with 1D3 in AD case without TDP-43 pathology.

Figure 5. Ratio of phosphorylated C-terminal fragments to full-length TDP-43

Immunoblots of FTLD-U and ALS cases probed with 1D3. Band intensities of full-length (FL) and C-terminal fragments (CTF) of TDP-43 were analyzed and the ratio between both values was calculated. CTF:FL ratios are depicted as a box and whiskers blot that gives the range of values, with the box being subdivided into the 25% and 75% quartiles by the median; circles represent outliers, filled rhombus represent the mean.