Acta Neuropathologica

, Volume 118, Issue 2, pp 303–311

TDP-43 pathology in familial British dementia

Authors

  • Claudia Schwab
    • Department of Psychiatry, Kinsmen Laboratory of Neurological ResearchUniversity of British Columbia
  • Tetsuaki Arai
    • Department of Psychogeriatrics, Tokyo Institute of PsychiatryTokyo Metropolitan Organization for Medical Research
  • Masato Hasegawa
    • Department of Molecular Neurobiology, Tokyo Institute of PsychiatryTokyo Metropolitan Organization for Medical Research
  • Haruhiko Akiyama
    • Department of Psychogeriatrics, Tokyo Institute of PsychiatryTokyo Metropolitan Organization for Medical Research
  • Sheng Yu
    • Department of Psychiatry, Kinsmen Laboratory of Neurological ResearchUniversity of British Columbia
    • Department of Psychiatry, Kinsmen Laboratory of Neurological ResearchUniversity of British Columbia
Case Report

DOI: 10.1007/s00401-009-0514-3

Cite this article as:
Schwab, C., Arai, T., Hasegawa, M. et al. Acta Neuropathol (2009) 118: 303. doi:10.1007/s00401-009-0514-3

Abstract

Trans-activation-responsive DNA-binding protein 43 (TDP-43) is a component of pathological inclusions in amyotrophic lateral sclerosis and several forms of sporadic and familial frontotemporal lobar degeneration. This has suggested defining a new class of diseases known as TDP-43 proteinopathies. However, it has been reported more recently that TDP-43 positive inclusions occur in other neurodegenerative disorders such as Alzheimer’s disease, Dementia with Lewy Bodies and Parkinsonism dementia complex of Guam. Here we report the occurrence of TDP-43 inclusions in one other neurodegenerative disorder: familial British dementia. Using a variety of antibodies against phosphorylated and non-phosphorylated TDP-43 epitopes, we found intense accumulation occurred in the form of dystrophic neurites, neuronal cytoplasmic inclusions and was also occasionally associated with neurofibrillary tangles. Double immunostaining revealed that TDP-43 and tau aggregates were rarely directly colocalized, but co-existed in the same neurons as separate inclusions. Double staining with ubiquitin showed a direct colocalization with TDP-43. The phosphorylation-dependent TDP-43 antibodies proved superior to phosphorylation-independent antibodies in revealing pathological inclusions since the former did not stain non-phosphorylated TDP-43 in normal nuclei. Our results support the concept that TDP-43 pathology is not narrowly restricted, but is involved in the etiology of many neurodegenerative disorders.

Keywords

TDP-43Familial British dementiaABriUbiquitinIntracellular inclusionsPhosphorylation-dependent TDP-43 antibodies

Introduction

Trans-activation-responsive (TAR) DNA-binding protein 43 (TDP-43) has been identified as a major component of ubiquitinated tau-negative inclusions in several forms of sporadic and familial frontotemporal lobar degeneration (FTLD) and amyotrophic lateral sclerosis (ALS) [2, 25]. For this common trait, these disorders were grouped together as a new entity of neurodegenerative diseases, TDP-43 proteinopathies [6, 20]. TDP-43 is ubiquitously expressed and involved in regulating transcription and alternative splicing [4, 5, 26]. TDP-43 has also been linked to cytoskeletal stability and axonal transport by a recent study, which showed that TDP-43 regulates hNFL RNA stability [32].

More recently an association of TDP-43 with other inclusion-forming proteins has been described. TDP-43 inclusions have been shown to co-exist with tau aggregates in a subset of Alzheimer’s disease (AD) cases, particularly those with hippocampal sclerosis [1, 3, 13, 15, 22, 33], Parkinsonism dementia complex of Guam (PDC-G) [9, 11, 21] as well as argyrophilic grain disease (AGD) [8] and coexistence with tau deposits also occurs in some cases of Pick’s disease and corticobasal degeneration [7, 13, 33]. Association with alpha-synuclein containing inclusions has been described in Lewy body dementia (LBD) and Parkinson’s disease (PD) [3, 13, 22]. Additionally we recently reported colocalization of TDP-43 with huntingtin inclusions in Huntington’s disease (HD) [30].

Familial British dementia (FBD) is another neurodegenerative disorder characterized by tau inclusions and amyloid deposits. While tau aggregates similarly to the neurofibrillary tangles (NFT) in AD, the amyloid plaques in FBD are created by ABri, and not by β-amyloid of AD, and the distribution of plaques is different to that of AD.

Familial British dementia is a rare familial disorder caused by a point mutation in the BRI gene [34]. This mutation changes a stop-codon and leads to an 11-amino acid extension of Bri-precursor protein. Its 22 C-terminal amino acids are cleaved off as ABri peptide and aggregate in the form of various types of amyloid plaques and angiopathy [14]. The deposition of the amyloidogenic ABri is accompanied by tau pathology, in the form of AD-type NFT, dystrophic neurites (DN) and neuropil threads [10, 28, 35]. ABri amyloid deposition in FBD is generally limited to the limbic system, the cerebellum and brain stem areas. Tau pathology is found in the same regions, except the cerebellum [14].

Since association of TDP-43 with tau aggregates has been reported in AD and PDC-G [1, 9, 11, 13, 22, 33], we examined if such an association also exists in FBD and whether the distribution of TDP-43 pathology correlates with that of ABri amyloid and tau aggregates.

Materials and methods

For this study the following cases were selected from our brain bank at the University of British Columbia: one case of familial British dementia (female, 64 years, postmortem delay 4 h) and two elderly control cases (female, 87 years, 4 h postmortem delay and male, 72 years, 4.5 h postmortem delay). The FBD case had been identified in a family described by Plant and colleagues [28] and appears as V44 in Fig. 1.
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-009-0514-3/MediaObjects/401_2009_514_Fig1_HTML.jpg
Fig. 1

Photomicrographs of immunohistochemistry in amygdala (ad), CA1 (eh), entorhinal cortex (il), midtemporal cortex (mp) and cerebellum (qt) of a case with familial British dementia. The first column (a, e, i, m, q) shows that ABri-labeled deposits appear in all regions except midtemporal cortex, where only amyloid angiopathy is found. Abnormal tau inclusions are most dense in amygdala and entorhinal cortex (b, j), a moderate number is found in CA1 and midtemporal cortex (f, n), and none in the cerebellum (r). Distribution if ubiquitin is similar (c, k, o, s), with the exception of CA1 where many extracellular NFTs are stained (g). Immunostaining of TDP-43 parallels that of tau (d, h, l, p, t). Calibration bar 100 μm, except in ad and il: 50 μm

Immunohistochemistry was carried out on brain tissue using a panel of antibodies against TDP-43, tau, ubiquitin, β-amyloid and ABri. Their source and dilutions used are shown in Table 1. The phosphorylated TDP-43 antibodies were raised against phosphorylated sequences of human TDP-43 [12, 17] and have the advantage of not recognizing normal TDP-43.
Table 1

Antibody source and dilutions

Antibody

Source

Catalog #

Antigen/specificity

Host, dilution (DAB/fluorescence)

AT8

Pierce, Rockford, IL

MN1020

PHF tau pS202

Mouse, 1:3,000/1:500

A0024

Dako, Carpinteria, CA

A0024

C-term. of recombinant human tau

Rabbit, 1:2,000/1:1000

ABri

Dr. H. Akiyamaa [31]

NA

ABri 22-34

Rabbit, 1:10,000/1:5,000

6F3D

DAKO, Mississauga, ON

M0872

Synthetic β-amyloid, 8-17

Mouse, 1:500/ND

Ubiquitin

DAKO, Mississauga, ON

Z0458

Ubiquitin from cow erythrocytes

Rabbit, 1:1,000/1:200

TDP-43

ProteinTech, Chicago, IL

10782-2-AP

Recombinant protein

Rabbit, 1:300/ND

TDP-43

Abnova, Taipei, Taiwan

H00023435-A01

Full-length recombinant protein with GST tag (1-261)

Mouse, b1:300/200

TDP-43 pS403/04

Dr. M. Hasegawaa [12]

NA

TDP-43 (398-408, pS403/404)

Rabbit, 1:2,000/1:1,000

TDP-43 pS409/10

Dr. M. Hasegawaa [12]

NA

TDP-43 (405-414, pS409/410)

Rabbit, 1:2,000/ND

TDP-43 pS409/10 (m)

Dr. M. Hasegawaa [17]

NA

TDP-43 (405-414, pS409/410)

Mouse, b1:5,000/1:1,000

Anti-TDP-43N (3-12)

Dr. M. Hasegawaa

NA

TDP-43 (N-terminal)

Rabbit, b1:5,000/ND

Anti-TDP-43C (405-414)

Dr. M. Hasegawaa [12]

NA

TDP-43 (C-terminal)

Rabbit, 1:5,000/ND

aTokyo Institute of Psychiatry, Tokyo, Japan

bAntigen retrieval

ND Not determined, NA not applicable

Brain tissues had been fixed in 4% paraformaldehyde, and, after 3–4 days, transferred to a 15% buffered sucrose maintenance solution. For immunohistochemistry, 30 μm sections of selected brain areas were cut on a freezing microtome (American Optical Corporation, Buffalo, NY).

For light microscope immunostaining, the sections were treated for 30 min with 0.5% H2O2 solution in 0.01 M phosphate buffered saline, pH 7.4, containing 0.3% Triton X-100 (PBS-T), transferred into 5% skim milk in PBS-T for 30 min, and incubated for 72 h at 4°C or overnight at room temperature with one of the primary antibodies (Table 1). Sections were next treated with the appropriate biotinylated secondary antibodies (DAKO, Mississauga, ON, 1:2,000) for 2 h at room temperature, followed by incubation in avidin–biotinylated horseradish peroxidase complex (DAKO, Mississauga, ON, 1:10,000) for 1 h at room temperature. Peroxidase labeling was visualized by incubation in 0.01% 3,3-diaminobenzidine (DAB; Sigma, Oakville, ON) containing 1% nickel ammonium sulfate (Fisher Scientific, Ottawa, ON), 5 mM imidazole (BDH Laboratory Supplies, Poole, UK) and 0.001% H2O2 in 0.05 M Tris–HCl buffer, pH 7.6. When a dark purple/black color developed, sections were washed, mounted on glass slides, air-dried, and coverslipped with Entellan (EMD Biosciences, Gibbstown, NJ). Some sections were counterstained with Neutral Red (BDH Laboratory Supplies, Poole, UK) prior to coverslipping. Antigen retrieval as indicated in Table 1 was carried out by boiling the sections in PBS for 5 min prior to all other steps to improve the staining of inclusions.

For double immunofluorescence staining, sections were incubated in 5% skim milk in PBS-T for 30 min, and then incubated for 72 h at 4°C or overnight at room temperature with a combination of two primary antibodies. These combinations were: AT8/TDP-43(pS403/404), A0024/TDP-43(Abnova), TDP-43(mouse pS409/10)/ABri 1129 and TDP-43(mouse pS409/10)/Ubiquitin (for the source and concentration of antibodies used see Table 1). Sections were next incubated with a mixture of fluorophore-labeled secondary antibodies (Alexa Fluor 488 goat-anti-mouse and Alexa Fluor 546 goat-anti-rabbit, Invitrogen, Burlington, ON, Canada; 1:500) in the dark, counterstained with Hoechst 33258 (Invitrogen, Burlington, ON, Canada), and mounted on glass slides. To reduce lipofuscin autofluorescence, sections were next treated with a solution of 0.3% Sudan Black B (Gurr Ltd, London, UK) in 70% ethanol for 7 min and washed in PBS eight times [29]. Subsequently, sections were air-dried and coverslipped with Prolong Gold (Invitrogen, Burlington, ON). The staining pattern for each individual antibody following the double-fluorescence staining was identical to the pattern observed in single immunostaining experiments. Controls for immunostaining were performed by omitting the primary antibodies. No staining was observed in these controls.

Confocal images were captured with a spindisk confocal microscope (inverted Olympus widefield microscope with Carv Spindisk, Olympus, Center Valley, PA) at 60× objective magnification. Fluorescent images were colocalized with ImagePro software (Improvision Inc., Waltham, MA). To assist interpretation of the figures, TDP-43 antibodies were always assigned the color green and tau, ABri and ubiquitin antibodies were assigned the color red in the false color assignment with ImagePro.

Results

In the control cases, no immunostaining was found with the antibodies against ABri (ABri 1129) or with the phosphorylation-dependent antibodies against TDP-43 (TDP-43 pS403/04, TDP-43 pS409/10, mouse TDP-43 pS409/10). The phosphorylation-independent TDP-43 antibodies stained nuclei of neurons and glia cells. The antibodies against tau and ubiquitin revealed occasional DNs and NFTs in hippocampus, fusiform and temporal cortex.

The results of immunostaining with the ABri, tau, TDP-43 and ubiquitin antibodies in the FBD case are summarized in Table 2.
Table 2

Pathological pattern in FBD

Brain region

ABri

Tau

TDP-43

Ubi

Severity

Pattern

Severity

Pattern

Severity

Pattern

Severity

Pattern

Amygdala

++++

dPl, cPl, AA, IG

++++

iNFT, preNFT, DN, NPT, TANC

+++

NCI, iNFT, DN, TANC

++++

iNFT, eNFT, NCI, DN, TANC

CA1/subiculum

++++

dPl, cPl, AA, IG

+

DN, NPT, TANC

+

DN, TANC

++

iNFT, eNFT, DN

Dentate gyrus

++

dPl, cPl

+++

PreNFT, DN

+

NCI, DN

++

NCI, DN

Entorhinal cortex

+++

dPl, cPl, AA, IG

+++

iNFT, preNFT, DN, NPT, TANC

++

NCI, iNFT, DN, TANC

+++

iNFT, eNFT, NCI, DN, TANC

Midfrontal cortex

+

AA

-+

DN

 

±

DN

Temporal cortex

++

dPl, AA

+

iNFT, DN, NPT

 

+

iNFT, DN

Supramarginal cortex

+

dPl, AA

 

 

 

Visual cortex

+

AA

 

 

 

Thalamus

+

dPl, AA

+

iNFT

+

DN

+

iNFT, DN

Locus coeruleus

+++

dPl, AA

+

iNFT

 

+

iNFT, DN

Substantia nigra

 

 

 

±

DN, Marinesco B

Periaqueductal gray

+++

dPl, AA

+

iNFT

 

+

iNFT, DN

Pedunculi cerebri

++

dPl, AA

 

 

 

Cerebellum

++++

dPl, AA

 

 

 

Ubi Ubiquitin, dPl diffuse plaques, cPl compact plaques, AA amyloid angiopathy, IG intracellular glial aggregates, iNFT intracellular neurofibrillary tangles, preNFT pre-tangle neurons, DN dystrophic neurites, NPT neuropil threads, TANC tangle associated neuritic clusters, NCI neuronal cytoplasmic inclusions

The ABri antibody stained various types of amyloid deposits in the form of diffuse, compact and star-shaped plaques and vessel associated amyloid angiopathy (Fig. 1 first column, Fig. 3r). These deposits were most frequently found in the hippocampus, entorhinal cortex and cerebellum, although the cerebellum was devoid of compact and star-shaped plaques. The other areas examined contained amyloid angiopathy and only minor diffuse amyloid deposits. No staining was found with the 6F3D antibody against β-amyloid in the FBD case.

Pathological patterns revealed by the tau antibodies (AT8 and A0024) included numerous intracellular NFTs (iNFTs), pre-tangle neurons and a thick meshwork of DNs and neuropil threads (Fig. 1, column 2). These inclusions were most densely distributed throughout the amygdala and the entorhinal cortex. Many dentate granule neurons had tau-positive inclusions. Hippocampal CA1 and subiculum were almost devoid of surviving neurons but contained many extracellular NFTs (eNFTs). Therefore, very few iNFTs were present in these regions and stained by the tau antibodies. However, thick, elongated dystrophic neurites were frequently associated with the eNFT forming tangle associated neuritic clusters (TANCs). In brainstem regions, such as the locus coeruleus, periaqueductal gray and thalamus, tau pathology in the form of iNFTs was less frequent. The cerebellum was devoid of tau pathology.

The ubiquitin antibody stained a dense network of DNs, neuronal cytoplasmic inclusions (NCIs) and NFTs (Fig. 1 column 3). Distribution and extend of staining was similar to that observed with the tau antibodies. But while the tau antibodies labeled a dense web of DNs and neuropil threads of various calibers, mostly thick elongated DNs were labeled by the ubiquitin antibody. Additionally to the above described pattern, Marinesco bodies in the SN were labeled only by the ubiquitin antibody.

The staining patterns with the three phosphorylation-dependent TDP-43 antibodies were very similar, labeling NCIs, DNs, and single fiber strands in iNFTs (Fig. 2). This TDP-43 pathology was predominantly found in the gray matter. Dystrophic neurites were mostly of a large caliber in various shapes and sizes, forming round and elongated hooks and loops. Spherical dystrophic neurites were associated with extracellular NFTs (TANCs). NCIs were of various sizes and shapes. Some neurons contained merely a few wisps of TDP-43 immunopositive material, others were filled with a dense cocoon-like net and some contained solid inclusions filling almost the entire perikaryon. The C-terminal and phosphorylation-dependent TDP-43 antibodies also labeled peculiar starfish-shaped inclusions (Fig. 2p, q). Intranuclear TDP-43 aggregates were not found. The distribution of TDP-43 inclusions paralleled that of pathological tau immunoreactivity. Intense pathology was present in the amygdala and the entorhinal cortex. Moderate pathology was present in the adjacent fusiform gyrus, the dentate gyrus, CA1 and subiculum and light pathology was present in brainstem and some isocortical regions. No TDP-43 pathology was found in the cerebellum (Fig. 1t). In the amygdala and CA1, the density of TDP-43 inclusions was similar to that of the tau inclusions, but it was lower in all other areas examined. While the staining density of the C-terminal phosphorylation-independent TDP-43 antibody was comparable to that of the phosphorylation-dependent TDP-43 antibodies, the antibody against the N-terminal of TDP-43 stained fewer inclusions (Table 3). In addition to the pathological inclusions, the phosphorylation-independent TDP-43 antibodies labeled normal appearing cell nuclei. In areas with a high load of pathology, this nuclear staining was greatly diminished, as has been described as ‘clearing’ of nuclei [25].
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-009-0514-3/MediaObjects/401_2009_514_Fig2_HTML.jpg
Fig. 2

Immunostainig with various TDP-43 antibodies in FBD at high magnification. Neuronal cytoplasmic inclusions (NCI) and dystrophic neurites (DN) (first columna, d, g, j), fibers in neurofibrillary tangles (NFT) (second columnb, e, h, k) and tangle associated neuritic clusters (TANC) (third columnc, f, i, l) are stained by all TDP-43 antibodies (ac: ProteinTech, df: Abnova, gi: pS409-410, jl: monoclonal pS409-10, m: anti-TDP-43N (3-12). Additionally skein-like (n, Abnova), small ring-shaped (o, Abnova) and starfish-shaped inclusions (p, q; pS403-403) were stained by several of the antibodies. Note that ProteinTech, Abnova and anti-TDP-43N (3-12) antibodies stain normal appearing nuclei in addition to abnormal inclusions and. Sections in df and oq were counterstained with Neutral Red. Calibration bar in l 30 μm

Table 3

Staining pattern with various TDP-43 antibodies

Antibody

Epitope

Nuclei

NCI

DN

NFT

TANC

SF

TDP-43 (Abnova)

Full-length recombinant protein

++

+++

++

+

++

0

TDP-43 (ProteinTech)

Recombinant protein

+++

+++

++

+++

++

0

TDP-43 pS403/04

pTDP-43 (398-408, pS403/404)

0

+++

+++

++

++

+

TDP-43 pS409/10

pTDP-43 (405-414, pS409/410)

0

+++

+++

+

++

+

TDP-43 pS409/10 (monoclonal)

pTDP-43 (405-414, pS409/410)

0

+++

++

++

++

+

Anti-TDP-43N (3-12)

TDP-43 (N-terminal)

++

++

+

+–

+

0

Anti-TDP-43C (405-414)

TDP-43 (C-terminal)

+

+++

++

++

++

+

Number of stained objects: 0, none; +, rare; ++, intermediate; +++, numerous

NCI Neuronal cytoplasmic inclusions, DN dystrophic neurites, NFT neurofibrillary tangles, TANC tangle associated neuritic clusters, SF star fish-shaped inclusions

TDP-43 and tau. Confocal microscopy of double-immunostained sections with TDP-43 and tau antibodies, revealed that TDP-43 and tau inclusions often coexisted within the same neurons (Fig. 3). However, the staining was only rarely directly colocalized. Instead, TDP-43 and tau-labeled separate parts of the same inclusions or more often completely separate inclusions in the same neuron. NFT fibers were labeled by the tau antibodies while TDP-43 labeled round inclusions nestled between the tau-positive strands. Exceptions to this separate staining pattern were occasional fibers in tau-immunoreactive NFTs which appeared to be coated with TDP-43 (Fig. 3e–h). In addition to neurons with both TDP-43 and tau-labeled inclusions, many neurons had only one type of inclusion, either positive for TDP-43 or tau. In the amygdala, half of the neurons with any type of inclusion contained both TDP-43 and tau aggregates, while the other half had either TDP-43 or tau inclusions (approximately 25% for each type). Both tau antibodies stained a dense network of dystrophic dendrites and neuropil threads, often obscuring neuronal inclusions. Dystrophic neurites labeled by TDP-43 were more sparse, thicker and contorted. In TANCs, tau and TDP-43 labeled different sets of DNs. Tau stained DNs were elongated while TDP-43 labeled round DNs. In addition to the pathological aggregates, the monoclonal TDP-43 (Abnova) antibody stained normal appearing nuclei in neurons that carried tau-NFT.
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-009-0514-3/MediaObjects/401_2009_514_Fig3_HTML.jpg
Fig. 3

Confocal images of double staining with TDP-43, tau, ABri and ubiquitin. al pS 403/404 (green: a, e, i) and AT8 (red: b, f, j). Sections were counterstained with Hoechst 33258. Tau and TDP-43 inclusions co-exist but are not colocalized (ad). A tau-labeled tangle contains a few TDP-43-labeled fibers (eh). Occasionally tau-NFTs contain small ring-shaped TDP-43 inclusions (il, arrow). mt TDP-43 Abnova (green: m) and A0024 (red: n). Although TDP-43 and tau inclusions co-exist in the same neuron, the staining is not directly co-localized (p). Double-fluorescence staining of TDP-43 (green: monoclonal pS409/10, q, u) with ABri 1129 (red: r) or ubiquitin (red: v). TDP-43 inclusions are not associated with ABri plaques (qt). Double staining with ubiquitin (ux) reveals that most TDP-43 labeled inclusions are stained by ubiquitin. Several inclusions (possibly NFT) are positive for ubiquitin alone (v), and some small DNs are stained by TDP-43 only (u, x, arrows). Sections were counterstained with Hoechst 33258. Calibration bar 20 μm in ap and 30 μm in qx

TDP-43 and ubiquitin. Almost all TDP-43 immunostained inclusions were also labeled by the ubiquitin antibody (Fig. 3u–x). In contrast to tau, ubiquitin and TDP-43 labeling were directly colocalized. Additionally, the ubiquitin antibody stained many inclusions that were unstained by TDP-43; the majority of these were NFTs and DNs. The number of ubiquitin-labeled DNs was approximately one order of magnitude higher than that of DNs stained by TDP-43. However, a small number of small round DNs was negative for ubiquitin staining and labeled with TDP-43 only. While in the amygdala and the entorhinal cortex the number of neurons containing TDP-43 inclusions was only slightly lower than the number of neurons with ubiquitin inclusions, in other areas like the parahippocampal gyrus and adjacent temporal cortex a much higher proportion was stained with ubiquitin alone.

TDP-43 and ABri. Although immunostaining with both TDP-43 and ABri antibodies was intense, there was no direct association of TDP-43-labeled DNs or NCIs with ABri plaques (Fig. 3q–t).

Discussion

In the present study, we report severe TDP-43 pathology in a case of FBD. Most of the pathological inclusions were found in the amygdala, hippocampus, entorhinal cortex and parahippocampal gyrus, paralleling the distribution of tau inclusions and ABri amyloid in this disorder. The TDP-43 antibodies labeled DNs, NCIs and fibers in NFTs. Confocal analysis revealed that the TDP-43 immunoreactivity was almost always directly colocalized with ubiquitin immunostaining. The exception was rare DNs which were stained by TDP-43 alone. Double staining of TDP-43 with tau revealed that these were rarely directly colocalized although many neurons contained both types of inclusions. No immunostaining for β-amyloid was detected, ruling out confounding AD-type pathology.

This report is based on the single case available to us. However, the occurrence of TDP-43 in FBD parallels other reports on conditions where there is co-deposition of TDP-43 with tau. They include AD, PDC-G, AGD and Pick’s disease [1, 79, 11, 13, 21, 22, 33]. As in these other ‘tauopathies’, TDP-43 in FBD is rarely directly colocalized with tau, but is frequently found as separate inclusions in the same neurons [1, 9, 11, 13]. The rate of TDP-43/tau double-labeling in FBD was higher than reported in AD by Hasegawa and colleagues [11]. Approximately half of neurons with TDP-43 had also tau inclusions in AD [11], while in the amygdala of FBD, approximately two-thirds of TDP-43 inclusions were accompanied by tau inclusions. This could be explained by differences in the pathoetiology or variations in severity of pathology in the individual cases. Although the selection of brain regions available for this FBD case was limited, most severe pathology was found in the amygdala, indicating that the anatomical pattern and progression of TDP-43 pathology may be similar to the pattern observed in AD and DLB [3, 15].

The types of TDP-43 inclusions found in FBD were NCIs, DNs and occasional fibers in NFTs; similar to the inclusions described in AD and PDC-G [1, 9, 11, 13]. Glial TDP-43 inclusions have been described in PDC-G [9], but were not found in FBD. This parallels reports of prominent glia pathology in PDC-G [21, 27] but not in FBD. Lack of glia pathology in FBD may be the reason for predominance of TDP-43 inclusions in gray matter compared to white matter. In PDC-G the severity of TDP-43 pathology has been described as equal in white and gray matter or even more severe in white matter [9, 11].

In contrast to other TDP-43 proteinopathies (ALS and FTLD), in this case of FBD, no TDP-43 containing intranuclear inclusions (NII) were found. Even in most types of FTLD, with the exception of cases with VCP mutations, NII are “never very numerous” [19, 24]. Presence of NII was reported only in a fraction of cases in AD and PDC-G [11, 13] and not found in LBD [22]. Lack of intranuclear TDP-43 inclusions could indicate differences in the mechanism of the disorders.

In FBD, antibodies which recognize TDP-43 phosphorylated at S403/404 and S409/10 exclusively labeled pathological inclusions. This is in agreement with previous reports that TDP-43 is phosphorylated in abnormal aggregates [2, 11, 12, 17, 23, 25].

We found that the antibodies directed against TDP-43 C-terminal epitopes (including phosphorylation-dependent epitopes) label a higher number of pathological inclusions compared to the antibody against the N-terminus. Although variations in antibody affinity could be the cause, enrichment of C-terminal fragments in inclusions has been previously reported [16, 18, 23, 25].

It is interesting to note that the distribution of TDP-43 inclusions parallels that of tau, and not that of ABri amyloid. Despite having a heavy amyloid load, the cerebellum is free of TDP-43 (and tau) inclusions. This could point to differences in the propensity of the particular neuron types to accumulate and develop inclusions, be it tau, TDP-43 or alpha-synuclein.

Co-deposition of TDP-43 with tau, alpha-synuclein or huntingtin does not necessarily imply a direct interaction between these molecules. Instead it may suggest that common pathological mechanisms act which lead to the aggregation of various abnormal proteins in susceptible neurons.

Acknowledgments

We greatly appreciate the expert support with confocal imaging of Dr. Henry G. S. Martin (Vancouver, Canada). This research was supported by the Pacific Alzheimer Research Foundation.

Conflict of interest statement

The authors declare that they have no conflict of interest.

Copyright information

© Springer-Verlag 2009