TDP-43 was recently identified as the major disease protein in neuronal inclusions in frontotemporal lobar degeneration with ubiquitin-positive inclusions (FTLD-U). TDP-43 is not only linked to disease mechanisms in FTLD-U, but it is also the most robust marker for the specific detection of neuronal inclusions in FTLD-U. In this study, we describe additional TDP-43 pathology in the white matter as a characteristic feature in a series of 38 FTLD-U cases including 3 cases with mutations in the progranulin gene. White matter pathology was most abundant in frontal and temporal lobes, but it was also detectable in brainstem and spinal cord. Based on morphology and double-labeling experiments, white matter cells with TDP-43-positive inclusions most likely represent oligodendrocytes. Biochemically, hyperphosphorylated and truncated TDP-43 was detectable in insoluble brain extracts from affected white matter regions in FTLD-U, similar to the biochemical signature observed in FTLD-U gray matter. Taken together, these results expand the spectrum of TDP-43 pathology in FTLD-U, suggesting that white matter pathology might contribute to the neurodegenerative process and clinical symptoms in FTLD-U.
Frontotemporal dementias (FTDs) are a clinically, genetically, and pathologically heterogeneous group of neurodegenerative disorders accounting for up to 20% of presenile dementia cases. The clinical presentation is characterized by progressive changes in social, behavioral, and/or language function (1-3). Additionally, patients may display movement abnormalities such as parkinsonism or motor neuron disease (4,5). Genetic, immunohistochemical, and biochemical data are incorporated into the current nosology of FTDs, which broadly divides cases into those with tau-positive inclusions (e.g. Pick disease, corticobasal degeneration [CBD], and progressive supranuclear palsy [PSP]) versus those with ubiquitin-positive and tau- and α-synuclein-negative inclusions, designated as frontotemporal lobar degeneration with ubiquitin-positive inclusions (FTLD-U) (2), the most common form of FTDs (4,6-8). More than 30% of FTDs are familial, and mutations in the microtubule associated protein tau (MAPT) gene are pathogenic for hereditary FTD characterized by tau pathology (9,10), whereas mutations in the progranulin gene (GRN) are pathogenic for familial FTLD-U (11,12).
Recently, TDP-43, a widely expressed nuclear protein with presumed functions in transcription regulation and exon skipping (13-15), was identified as the major disease protein ubiquitinated in the inclusions in FTLD-U including familial FTLD-U cases with GRN mutations (16). After initial identification, TDP-43 was rapidly confirmed as a disease protein in FTLD-U (17). Thus, TDP-43 is implicated in FTLD-U pathogenesis and is the most sensitive and specific marker for FTLD-U pathology. Furthermore, based on morphology and laminar distribution of ubiquitin and TDP-43-positive cytoplasmic, nuclear, and neuritic inclusions in affected cortical brain regions, at least three pathologic subtypes of FTLD-U can be distinguished (16,18,19).
In the present study, immunohistochemical analysis for TDP-43 in a series of 38 FTLD-U cases, including 3 cases with GRN mutations, revealed that in addition to neuronal pathology, FTLD-U is also characterized by TDP-43 pathology in the white matter of affected brain regions. Pathologic accumulation was predominantly found in oligodendrocytes. Biochemically, TDP-43 in affected white matter regions becomes hyperphosphorylated and truncated, similar to pathologic alterations of TDP-43 in affected gray matter regions, thereby suggesting that glial pathology might contribute to the neurodegenerative processes involved in FTLD-U.
Materials and Methods
Brain Tissue Collection and Neuropathologic Assessment
Frozen brain tissues and fixed, paraffin-embedded tissue blocks were obtained from the Center for Neurodegenerative Disease Research Brain Bank at the University of Pennsylvania and the Center for Neuropathology and Prion Research Brain Bank at the University of Munich, Germany. Consent for autopsy was obtained from legal representatives for all subjects in accordance with local institutional review board requirements.
Antibodies used in this study included mouse monoclonal antibody (mAB) 1510 anti-ubiquitin (Chemicon, Temecula, CA), affinity-purified rabbit polyclonal anti-TDP-43 (ProteinTech Group, Chicago, IL), mouse mAB 2E2-D3 anti-TDP-43 (Abnova, Taipei, Taiwan), and mAB tau-2 anti-tau (Sigma-Aldrich, St. Louis, MO). For double-labeling experiments, mouse mAB anti-glial fibrillary acidic protein (GFAP) (Dako, Glostrup, Denmark) was used as a marker for astrocytes, mouse mAB anti-CD68 (Dako) as a marker for activated microglia, and mouse mAB anti-CD57 (Leu-7; Chemicon) and mouse mAB SMI91 anti-2‘,3’-cyclic nucleotide 3′-phosphodiesterase (CNPase; Sternberger Monoclonals, Lutherville, MD) as a marker for oligodendrocytes.
The harvesting, fixation, and further processing of the tissue specimens used in this study were conducted as described previously (18). Briefly, tissue blocks from representative brain regions were fixed with either 70% ethanol in 150 mmol/L NaCl or phosphate-buffered 3.65% formaldehyde and embedded in paraffin. Immunohistochemistry was carried out using the avidin-biotin complex detection system (Vector Laboratories, Burlingame, CA) and 3,3′-diaminobenzidine as chromogen. Antigen retrieval was done by boiling the sections in 10 mmol/L citrate buffer (pH 6.0) in a microwave oven (three times for 5 minutes). Double-labeling immunofluorescence was performed using Alexa Fluor 488 and 594 conjugated secondary antibodies (Molecular Probes, Eugene, OR).
Sequential Biochemical Fractionation and Immunoblot Analysis
Frozen tissues from frontal and temporal lobes as well as from cerebellum were used for the sequential extraction of TDP-43 with buffers of increasing stringency, as described (16). Gray and white matter were separated and processed individually. Briefly, gray and white matter was extracted at 5 mL/g (v/w) with low-salt buffer (10 mmol/L Tris, pH 7.5, 5 mmol/L EDTA, 1 mmol/L dithiothreitol, 10% sucrose, and a cocktail of protease inhibitors), high salt-Triton X buffer (low salt buffer + 1% Triton X-100 + 0.5 mol/L NaCl), myelin flotation buffer (Triton X buffer containing 30% sucrose), and Sarkosyl buffer (low salt buffer + 1% N-lauroyl-sarcosine + 0.5 mol/L NaCl). The detergent-insoluble materials were extracted in 0.25 mL/g of urea buffer (7 mol/L urea, 2 mol/L thiourea, 4% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, 30 mmol/L Tris, pH 8.5). For Western blot analysis, equal volumes of urea fractions (10 μL, equivalent to 40 mg of starting material) from different samples were resolved by Tris-glycine 5% to 20% gradient sodium dodecyl sulfate-polyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membranes (Millipore, Billerica, MA), and probed with antibodies to TDP-43. Primary antibodies were detected with alkaline phosphatase-conjugated anti-mouse or anti-rabbit IgG (Dako) and visualized by incubation with nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (Roche Molecular Biochemicals, Mannheim, Germany).
TDP-43-Positive White Matter Pathology in Frontotemporal Lobar Degeneration With Ubiquitin-Positive Inclusion Cases
Anti-TDP-43 immunohistochemistry was performed on frontal and temporal cortex as well as medulla or spinal cord, if available, from 38 FTLD-U cases, including 3 cases with GRN mutations (Table). In addition to the neuronal inclusions, TDP-43 staining allowed the detection of FTLD-U-specific TDP-43 accumulation in the cytoplasm of glial cells, predominantly located in the white matter (Fig. 1A, B) but also scattered in the cortex and motor nuclei in the medulla and anterior horn of the spinal cord (Fig. 1C, arrow). Glial inclusions were predominantly found in cells of oligodendroglial morphology with small, round nuclei. The shape of TDP-43-positive inclusions varied from round to triangular to comma-shaped, and they occasionally extended into processes (Fig. 1D-F). Moreover, few TDP-43-positive thread-like inclusions in cell processes were detectable (Fig. 1G), but no intranuclear inclusions were noted in glial cells. No pathologic TDP-43 staining was observed in the white matter of controls and subjects with other neurodegenerative disorders, such as Alzheimer disease (AD), PSP, CBD, and multiple system atrophy (MSA) (data not shown), arguing that the observed TDP-43-positive white matter pathology is FTLD-U specific. To further characterize the glial cells bearing TDP-43 inclusions, we performed double-labeling experiments. Anti-GFAP used as a marker for astrocytes and CD68 used as a marker for microglia showed no colocalization with TDP-43-positive glial cells (Fig. 2A-F), giving further support to the hypothesis that the predominantly affected cell type with TDP-43 inclusions in the white matter is oligodendrocytes. Markers used to label oligodendroglia, including CNPase and CD57 (Leu7), gave only weak immunoreactivity of some of the cells with inclusions due in part to the well-recognized lack of robust cellular markers for mature oligodendrocytes (20). Ubiquitin staining revealed numerous small, dot-like structures in the white matter in FTLD-U (Fig. 2H), control, and AD brains, a common finding described in normal aging (21). However, TDP-43-positive glial inclusions were not stained by ubiquitin (Fig. 2G-I), in contrast to TDP-43-positive neuronal inclusions (16), and there was no labeling of TDP-43-positive glial inclusions with tau-2 (data not shown).
Because the abundance of glial inclusions varied from case to case, we asked whether the extent of TDP-43 white matter pathology correlated with the severity of TDP-43 gray matter pathology and whether there are differences among the three subtypes of FTLD-U described (16,18). Therefore, cases were separated into FTLD-U subtype 1 (cases 1-10), subtype 2 (cases 11-22), and subtype 3 (cases 23-35) according to Sampathu et al (18) on the basis of the morphology and laminar distribution of ubiquitin and TDP-43-positive neuronal inclusions. Briefly, subtype 1 was characterized by predominance of long neuritic profiles over cytoplasmic inclusions in superficial cortical layers, subtype 2 showed predominance of cytoplasmic inclusions in both superficial and deep cortical layers, and in subtype 3 ring-shaped cytoplasmic as well as short neuritic profiles were present predominantly in the superficial cortical layers. The amount of TDP-43 gray and white matter pathology in frontal and temporal cortex was assessed semiquantitatively (Table). Glial inclusions were most widespread and abundant in FTLD-U subtypes 2 and 3 and in familial FTLD-U with GRN mutations but were rare or absent in subtype 1. In subtype 2 and 3 cases, there was a slight correlation between severity of gray matter and white matter pathology. Glial inclusions in the medulla and spinal cord were only found in cases with motor neuron involvement as detected by TDP-43-positive inclusions in motor neurons of medulla and spinal cord (Fig. 1C).
Pathologic Alterations of TDP-43 in White Matter of Frontotemporal Lobar Degeneration With Ubiquitin-Positive Inclusion Cases
A disease-specific biochemical signature of pathologically altered TDP-43 was reported in detergent-insoluble, urea-soluble extracts of gray matter from sporadic and familial FTLD-U cases (16). To biochemically characterize TDP-43 in the white matter of FTLD-U, samples of frontal and temporal gray and white matter from FTLD-U cases were sequentially extracted with buffers of increasing strength and the proteins in these extracts were analyzed by TDP-43 immunoblot methods. Cerebellar gray and white matter from FTLD-U cases as well as frontal gray and white matter from AD brains were used as controls. Pathologic alterations of TDP-43 similar to those detected in frontal and temporal gray matter were observed in the urea extracts of affected frontal and temporal white matter from FTLD-U cases, namely additional bands of ~25 and 45 kDa as well as a high molecular weight protein smear (Fig. 3). Frontal gray and white matter from AD brains as well as cerebellar gray and white matter from FTLD-U brains only showed the full-length TDP-43, which migrated as an ~43-kDa band (Fig. 3), consistent with the lack of pathologically altered TDP-43 and inclusions formed by TDP-43.
The presence and amount of pathologic TDP-43 species obtained by immunoblot analyses varied among cases and regions. However, the intensity of the modified bands corresponded roughly with the density of TDP-43-positive inclusions demonstrated by immunohistochemistry. For example, case 15, with high density of inclusions in gray and white matter, showed strong pathologic TDP-43 bands by immunoblot in these regions, whereas in case 26 with only rare inclusions in the temporal white matter, no pathologic TDP-43 species were detectable by immunoblot in this region.
TDP-43 is the major disease protein in FTLD-U that forms the signature intraneuronal cytoplasmic and intranuclear inclusions predominantly in the frontal and temporal cortex (16, 17) and therefore represents the most robust marker for the specific neuropathologic diagnosis of FTLD-U.
By immunohistochemical and biochemical approaches, we observed additional TDP-43 pathology in the white matter, composed of glial cytoplasmic inclusions and rare thread-like inclusions, as a characteristic feature in FTLD-U brains. TDP-43-positive glial inclusions could be found throughout the frontal and temporal lobes in most FTLD-U cases studied (n = 38), with the highest density in FTLD-U subtypes 2 and 3, providing further evidence for pathologic heterogeneity in FTLD-U (18,19). Moreover, scattered glial inclusions could be found in brainstem and spinal cord regions in FTLD-U cases with motor neuron involvement. These consistent findings of TDP-43-positive glial inclusions in areas with neurodegeneration suggests that glial abnormalities might be involved in the pathologic processes of FTLD-U.
On the basis of morphology and double-labeling experiments demonstrating a lack of colocalization of inclusion-bearing cells with markers for astrocytes and microglia, affected glial cells most likely represent oligodendrocytes. The morphology of TDP-43-positive glial inclusions varied from round or cap-shaped, resembling the α-synuclein-positive oligodendroglial cytoplasmic inclusions in MSA (22,23), to comma-like shapes resembling the tau-positive coiled bodies described in several neurodegenerative diseases, such as PSP, CBD, and familial forms of FTD with MAPT gene mutations (24, 25). A role of glial pathology in FTLD-U has been discussed previously by description of tau-2- or p62-positive glial inclusions (26,27). However, because cross-reactivity of tau-2 with proteins expressed in microglia has been described, and p62 also labels pathologic inclusions in other neurodegenerative disorders, the interpretation of these data with regard to etiology and pathogenesis of FTLD-U is unclear. Notably, the TDP-43-positive glial inclusions described in this study were not labeled by tau-2. Furthermore, α-synuclein- and tau-positive glial inclusions in MSA, PSP, and CBD were TDP-43-negative, and no TDP-43-positive inclusions were detectable in PSP, CBD, and MSA, thereby demonstrating specificity of the TDP-43 glial inclusions described in this study for FTLD-U. The fact that the TDP-43-positive glial inclusions are not immunoreactive for ubiquitin, in contrast to the neuronal TDP-43-positive inclusions in FTLD-U (16), might suggest that ubiquitination of TDP-43 is a later phenomenon in the disease course and is not necessary for TDP-43 accumulation.
By biochemical analyses of affected cortical brain regions in FTLD-U, we showed that hyperphosphorylated, ubiquitinated, and N-terminally truncated TDP-43 species are detected in the detergent-insoluble, urea-soluble protein extracts of these brain samples (16). In the current study, we demonstrated that an identical disease-specific biochemical signature of TDP-43, consisting of hyperphosphorylated, ubiquitinated, and truncated TDP-43, was also detectable in urea extracts from FTLD-U frontal and temporal white matter, arguing that similar pathomechanisms might be involved in neuronal and glial TDP-43 inclusion body formation.
TDP-43 is a highly conserved, ubiquitously expressed nuclear protein (28) that may act as a transcription repressor and activator of exon skipping (13-15) as well as a scaffold for nuclear bodies through interactions with survival motor neuron protein (29). Whereas physiologic TDP-43 is localized primarily to the nucleus of neuronal and glial cells, our data showed that under pathologic conditions immunodetection of nuclear TDP-43 was reduced in inclusion-bearing neuronal and glial cells and that TDP-43 is redistributed to accumulate in neuronal and glial cell bodies and processes, a consequence of which may be loss of TDP-43 nuclear function. Future studies will be needed to address these and other mechanistic aspects of the aggregation of pathologic TDP-43 in neuronal and glial cells and their role in the pathogenesis of neurodegeneration in FTDL-U.
In summary, this study expands the spectrum of TDP-43 pathology in FTLD-U by demonstrating widespread and abundant involvement of glial cells, especially oligodendroglia, in the white matter of FTLD-U brains. Thus, glial TDP-43 white matter pathology is a characteristic feature of FTDL-U in addition to neuronal TDP-43 gray matter pathology, thereby implying that glial TDP-43 pathology also contributes to the neurodegenerative process and the cognitive and motor impairments seen in patients affected by FTLD-U.
The authors thank the families of patients who made this research possible.