TDP-43 Related Neuropathologies and Phosphorylation State: Associations with Age and Clinical Dementia in the Cambridge City over-75s Cohort

Abstract

Pathologies associated with the Tar-DNA binding protein 43 KDa (TDP-43) are associated with neurodegenerative diseases and aging. Phosphorylation of cellular proteins is a well-accepted mechanism of biological control and can be associated with disease pathways. Phosphorylation state associated with TDP-43 associated pathology has not been investigated with respect to dementia status in a population representative sample. TDP-43 immunohistochemistry directed toward phosphorylated (TDP-43P) and unphosphorylated (TDP-43U) was assessed in sections of hippocampus and temporal cortex from 222 brains donated to the population representative Cambridge City over-75s Cohort. Relationships between dementia status and age at death for TDP-43 immunoreactive pathologies by phosphorylation state were investigated. TDP-43 pathologies are common in the oldest old in the population and often do not conform to MacKenzie classification. Increasing age is associated with glial (TDP-43P) and neuronal inclusions (TDP-43P and TDP-43U), neurites, and granulovacuolar degeneration (GVD). Dementia status is associated with GVD and glial (TDP-43 P) and neural inclusions (TDP-43 P and U). Dementia severity was associated with glial (TDP-43P) and neuronal inclusions (TDP-43U and TDP-43P), GVD, and neurites. The associations between dementia severity and both glial cytoplasmic inclusions and GVD were independent from other pathologies and TDP-43 neuronal cytoplasmic inclusions. TDP-43 pathology contributes to dementia status and progression in a variety of ways in different phosphorylation states involving both neurons and glia, independently from age and from classic Alzheimer-related pathologies. TDP-43 pathologies as cytoplasmic inclusions in neurons or glia or as GVD contribute independently to dementia.

Keywords

Aging dementia hippocampus phosphorylation population study TAR-DNA binding protein of 43 kDa

INTRODUCTION

Within the older population, the dementia syndrome is a heterogeneous, clinically defined condition leading to progressive cognitive impairment and loss of function in everyday life. Dementia-related neuropathologies, including deposition of aggregated proteins such as the amyloid-β protein, tau, and the Tar DNA binding protein 43 (TDP-43), are also associated with aging including in those who do not have the syndrome of dementia during life [1 –4]. Additionally, mixed vascular and neurodegenerative pathologies are common in the older population both with and without the dementia syndrome [3, 4], suggesting that relationships between neuropathology, aging, and dementia before death are complex. Intracellular inclusions of TDP-43—a DNA and RNA binding protein with roles in the regulation of transcription, RNA splicing [5], and stress response [6]—are associated with various neurodegenerative diseases including frontotemporal dementia (FTLD) [7], amyotrophic lateral sclerosis [7], old age hippocampal sclerosis [8], Alzheimer’s disease (AD) [2 , 8–10], and corticobasal degeneration [10]. Various autopsy or clinic series [11], community-based studies [12, 13], and population representative cohorts [2, 14] find that TDP-43 pathologies are common in the oldest old and that TDP-43 pathology is associated with clinically defined dementia, often mimicking AD [11, 15]. The contributions of TDP-43 to dementia in the older population have recently been formally defined as Limbic-predominant Age-related TDP-43 Encephalopathy (LATE) [16].

Neuronal TDP-43 can be deposited as fibrous, irregular, or spherical cytoplasmic inclusions, lentiform intra-nuclear inclusions, and neurites [7]. Both cytoplasmic and nuclear deposits of TDP-43 are also seen in glia [7]. Inclusions can be detected using anti-TDP-43 antibodies raised against either unphosphorylated TDP-43 (TDP-43U) [7] or phosphorylated TDP-43 (TDP-43P) [17, 18]. Neurons with neuronal cytoplasmic inclusions marked by anti TDP-43U immunostaining are also associated with clearance of TDP-43U seen as loss of both normal nuclear and cytoplasmic staining [7, 19].

In addition to the well-recognized intracellular neuronal and glial inclusions and neurites, various diffuse and granular intracellular deposits in neurons and glia are reactive with anti TDP-43P antibodies. Abnormal diffuse granular deposition has been observed in neurons of frontal cortex in FTLD linked to chromosome 9 [19] and TDP-43P has been associated with the dense bodies of granulovacuolar degeneration (GVD) [20, 21] but not neurofibrillary tangles (NFT) [10]. Some granular deposits have been interpreted as “pre-inclusions” [22].

Phosphorylation and de-phosphorylation of proteins are well-accepted biochemical mechanisms regulating protein function and interactions. Changes in phosphorylation state of a protein can lead to changes in cellular processes. Differences seen in neuropathologies immunoreactive with antibodies raised against different phosphorylation states of TDP-43 in those with and without dementia will add to knowledge of TDP-43-related disease processes involving loss as well as gain of function [5, 23].

The Cambridge City over 75s Cohort (CC75C) [24] is a population representative clinico-pathological study of brain aging with a brain donation program. This study allows neuropathological lesions to be investigated with respect to clinically defined dementia and the resultant estimates of the relationships between pathologies and dementia status are directly translatable to the older population, among whom most dementia occurs.

The characterization of TDP-43 neuropathologies by phosphorylation state and the clinical significance of pathologies seen with anti-TDP-43P or anti-TDP-43U immunoreactivity in relation to aging and dementia in the population are currently unknown. This study aimed to address the lack of knowledge regarding the range of pathologies immunoreactive with TDP-43P and TDP-43U seen in the hippocampus from brains donated in CC75C [24] and investigate their associations with clinical dementia and age. It is critical to identify and quantify these relationships to understand the underlying biology of aging and dementia, which is increasingly being recognized as far more complex than the traditional markers of amyloid and tau. Further, this study adds to the body of knowledge relating to the role of TDP-43 and phosphorylation state which is currently not characterized in the older population [16].

MATERIALS AND METHODS

Cohort description

CC75C is a population-based longitudinal study of aging and dementia with a brain donation program (http://www.cc75c.group.cam.ac.uk/) [24]. Each participant included gave consent to participate in the original baseline study and later for brain donation. Consent for brain donation was also given by next of kin. Each phase of the CC75C study has been approved by the Cambridge Research Ethics Committee (05/Q0108/308). At the time of this study, there were a total of 237 brain donations, of which n = 222 were included in this study. Exclusions were due to tissue being unavailable.

A consensus diagnosis for dementia status at death consistent with Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) criteria [25] was agreed by two clinicians blinded to neuropathology report using postmortem review of all interviews including proxy informant data, death certificates, and retrospective informant data after death. In life, participant interviews included cognitive assessments using the Mini-Mental State Exam [26] and where possible the Cambridge Diagnostic Examination for the Elderly (CAMDEX) assessment [27]. Dementia status was defined as present or absent. Dementia severity was rated as none, mild cognitive impairment (MCI), minimal, mild, moderate, and severe. The brain donor sample includes the full spectrum of cognitive abilities and the distribution of MMSE scores in the baseline sample closely matches those of the brain donor sample [28].

Neuropathological assessment

After death, brains were removed as soon as possible in the local mortuary. The brains were bisected in the sagittal plane. One cerebral hemisphere was dissected coronally into approximately one cm slices, macroscopically examined and snap frozen at –80°C. The other half of the brain was fixed in 10% neutral buffered formalin. For diagnostic purposes, tissue blocks for paraffin embedded histology were taken from posterior hippocampus (at the level of the lateral geniculate body), anterior hippocampus and transentorhinal cortex (at the level of the mammillary body), frontal, temporal, parietal, and occipital lobes, basal ganglia, thalamus, pons, medulla, cerebellum, and from two levels of the midbrain. Postmortem interval (PMI) was defined as the time between death and postmortem (mean PMI = 30 h 56 min, SD = 29 h 28 min, n = 239). Fixation time was defined as time in weeks between dates of postmortem and tissue processing to paraffin embedded tissue blocks (mean fixation time = 27.22 weeks, SD = 27.61 weeks, n = 230).

Sections cut from the paraffin-embedded brain tissue samples were assessed blind to clinical status using the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) protocol [29, 30] and Braak staging [31, 32]. Ten micrometer sections were stained with anti-tau monoclonal antibody (11/57 supplied by The Cambridge Brain Bank UK, (Cambridge Brain Bank Cat# 11/57, RRID:AB_2810220) to visualize NFT, neuritic plaques, and dystrophic neurites. Amyloid-β protein (Aβ) deposits were visualized with Congo red or anti Aβ antibody (Agilent Cat# M0872, RRID:AB_2056966), with formic acid rinse for antigen retrieval. Ratings for NFT, neuritic plaques, and amyloid deposits per section were graded as none, mild, moderate, or severe [29, 30]. Tau pathology was also assessed according to Braak staging [31, 32]. Slides stained with anti-ubiquitin antibody (Agilent Cat# Z0458, RRID:AB_2315524, first 174 cases) or anti-α-synuclein (Enzo Life Sciences Cat# BML-SA3400-0025, RRID:AB_2050693, remaining cases) and slides stained with hematoxylin and eosin were used to detect Lewy bodies which were graded as none, mild, moderate, or severe. All immuno-stained sections for diagnosis were counterstained with Ehrlich’s hematoxylin with diaminobenzidine as the chromagen. All slides were produced by the Cambridge Brain Bank UK and assessments were performed blind to clinical status by neuropathologists at Addenbrooke’s Hospital, Cambridge UK.

Nine micrometer thick sections from the hippocampal formation and temporal cortex were stained with anti-phosphorylated TDP-43P antibody (Cosmo Bio Co LTD. Cat# CAC-TIP-PTD-P02, pS409/410-2 RRID:AB_1961898) with antigen retrieval via microwave in a citrate buffer. Nine micrometer thick sections from the hippocampal formation were stained with anti-non-phosphorylated TDP-43U antibody (ProteinTech Group, R Cat# 18280-1-AP, RRID:AB_2240312) with antigen retrieval via formic acid rinse by the Cambridge Brain Bank UK. Slides were counterstained with Harris’ hematoxylin with diaminobenzidine as the chromagen. To avoid immunohistochemical artefacts due to prolonged fixation (>3.2 years), all slides immunostained for TDP-43 came from blocks taken during the diagnostic process. Cases were not included if these tissues were already used up.

Neuronal and glial cytoplasmic inclusions, neuronal and glial nuclear inclusions, GVD bodies, and neurites were assessed on TDP-43P and TDP-43U slides; absence of normal TDP43U staining from the cytoplasm and nucleus was assessed on TDP-43U slides. Pathologies were semi-quantitatively scored as none (no inclusions), minimal (one inclusion per region of interest), mild (more than one inclusion in up to half the fields of view per region of interest), moderate (a few inclusions in over half the fields of view per region of interest), or severe (a few inclusions in most fields of view per region of interest) at x200 magnification and an adapted TDP-43 staining protocol [7]. Care was taken to identify glial inclusions versus larger cross sectioned neurites or neuronal inclusions though without double staining for markers specific to glia this measure has a degree of uncertainty. Any staining that was ambiguous was not recorded. Cases were typed according to Mackenzie et al. [33] for all cases that presented with cortical TDP-43P neuronal cytoplasmic inclusions or neurite pathology in either the parahippocampal or temporal cortex. When cortical TDP-43P neuronal cytoplasmic inclusions or neurite pathology was present, but no subtype could be unambiguously assigned, subtype was recorded as “unclassifiable”.

Reliable assignment of TDP-43 stage according to the protocol defined by Josephs staging [34, 35] is not possible with our current resources. CC75C currently lacks TDP-43 immunohistochemistry for the amygdala for reliable assignment of none or stage 1, and further cortical and subcortical areas for the assignment of stages 3 and higher. Only the assignment of stage 2 is possible. Further work is required to reliably investigate Josephs TDP-43 staging in this cohort.

Separate scores for the hippocampal regions, dentate, Cornu Ammonis (CA)4, CA2/3, and CA1 were generated. Scoring was performed blind to clinical status by SHu. 10% (n = 23) were scored by SHo blind to previous readings. Inter-rater analysis using Gwet’s Agreement Coefficient 2 (AC2) (Agreestat 2011.3 program; Advanced Analytics, Gaithersburg, MD, USA) is shown for TDP-43P and TDP-43U in Supplementary Tables 1–4. The extent of agreement was assessed using the benchmark proposed by Landis and Koch [36]; a coefficient > 0.6 indicating substantial agreement and a value > 0.8 near-perfect agreement. Agreement was either substantial or near perfect for all variables except glial inclusions immunoreactive with anti-TDP-43U which were not included in analyses.

Statistical analysis

Cortical TDP-43 pathology was assessed for Mackenzie subtype. Analyses of the range of TDP-43 pathologies considered here and their associations with concomitant other pathologies, dementia, and age were confined to the hippocampus as this was the only area scored for all the pathologies considered. Since the distributions of TDP-43 scores did not support reliable analysis by specific hippocampal area, scores for each pathology were collapsed over the entire hippocampus by selecting the highest score from any hippocampal area. All analyses were performed with TDP-43 immunoreactivity dichotomized as present (mild, moderate, and severe) or absent (none and minimal). The relationships between age and TDP-43 immunoreactive pathologies (dependent variables) were tested using logistic regressions with age as the independent variable. The relationships between dementia severity and presence of TDP-43 immunoreactive pathologies were investigated by performing ordinal logistic regressions with dementia severity as the dependent variable and the presence of TDP-43 immunoreactive pathologies as independent variables (Model I), while controlling for age at death (Model II) or the presence of other pathologies (Model III). A dichotomous variable was constructed for “other pathologies” defined as Braak stage 4 or higher, or neuritic plaque score was moderate or severe, or amyloid deposition score was moderate or severe. The AD Neuropathic change score (none, low, intermediate, or high) [37] is not possible with this cohort as we currently lack Thal phase for amyloid deposition. Model IV was similar to Model III with addition of controlling for neuronal TDP-43P inclusions (the traditional TDP-43 pathology employed in the context of dementia). Lewy bodies (Supplementary Table 5) were not included in analyses due to their rarity. All tests were 2-tailed and alpha < 0.05 was considered statistically significant. Statistical analyses were performed using statistical package STATA, version 12.

RESULTS

In total, 222/237 brains were included in the sample. Among them, 156 (70%) were women. Mean age of death was 91.1 (±4.7) years. Assignment of dementia severity at death was not possible in 11 (5%) participants, 61 (27%) did not have dementia, 17 (7%) had MCI, and 145 (65%) had dementia at death. Additional exclusions were due to tissue unavailability. Table 1 shows the distribution of hippocampal TDP-43 immunoreactive pathologies across the semi-quantitative TDP-43 severity categories. Glial, neuronal, and neuritic TDP-43 immunoreactive pathologies, GVD, and complete loss of TDP-43U staining are illustrated in Fig. 1.

Table 1

Distribution of number of cases with each TDP-43 immunoreactive pathology seen in the hippocampus

TDP -43 pathology	Pathological severity n (%)
	None	Mild	Moderate	Severe	Total
GVD	69 (32)	105 (49)	39 (18)	3 (1)	216
Neurites	164 (75)	19 (9)	7 (3)	30 (14)	220
GCI-P	197 (89)	20 (9)	5 (2)	0 (0)	222
NCI -P	160 (72)	34 (15)	13 (6)	15 (7)	222
NCI-U	174 (78)	28 (13)	12 (5)	8 (4)	222
TDP-43U loss	101 (47)	95 (44)	13 (6)	8 (4)	217

GVD, granulovacuolar degeneration; GCI, glial solid cytoplasmic inclusions; NCI-P, neuronal solid cytoplasmic inclusions immunoreactive with anti-TDP-43P; NCI-U, neuronal solid cytoplasmic inclusions immunoreactive with anti-TDP-43U. Numbers do not always add up to 222 as some cases do not have immunohistochemistry or specific areas were not present on the slide in all cases.

Fig.1

Hippocampal neuropathologies immunoreactive with antibodies raised against either phosphorylated or unphosphorylated TDP-43. A) Hippocampal dentate stained with anti- phosphorylated TDP-43 showing numerous neuronal cytoplasmic inclusions (>). B) Hippocampal dentate stained with anti- unphosphorylated TDP-43 showing numerous neuronal cytoplasmic inclusions with nuclear clearance (>). A and B are from the same case. C) Hippocampal dentate stained with anti- phosphorylated TDP-43 showing no neuronal cytoplasmic inclusions. D) Hippocampal dentate stained with anti- unphosphorylated TDP-43 showing nuclear clearance without neuronal cytoplasmic inclusions (>). C and D are from the same case. E) Hippocampal CA1 stained with anti- phosphorylated TDP-43 showing neurites (1), GVD (2), glial cytoplasmic inclusion (3) and neuronal cytoplasmic inclusions (4). F) Hippocampal CA1 stained with anti- unphosphorylated TDP-43 showing fibrous neuronal cytoplasmic inclusions with nuclear clearance and unmarked GVD bodies without nuclear clearance. G) Hippocampal CA2 stained with anti- phosphorylated TDP-43 showing GVD bodies ranging from well to faintly marked. H) Hippocampal CA2 stained with anti- unphosphorylated TDP-43 showing lack of marked GVD bodies and normal nuclear staining. I) Hippocampal CA2 neurons with complete loss of staining associated with unphosphorylated TDP-43. J, K) Possible GVD “ghosts” seen with anti- phosphorylated TDP-43 antibody. Scale bars in A-D, F-H, J, and K = 20μm. Scale bars in E and I = 50μm.

Comparison of neuronal cytoplasmic inclusions with anti-TDP-43P and anti-TDP-43U

Agreement between TDP-43P and TDP-43U regarding neuronal inclusions staining was assessed by calculating the Gwet’s AC2, which was 0.94 (95% CI 0.933; 0.956; p < 0.001), demonstrating near-perfect agreement. To verify if TDP-43P and TDP-43U are giving the same information regarding neuronal inclusions, we used the McNemar’s test, which showed that the proportion of brains stained only by TDP-43P was significantly higher than those stained only by TDP-43U (7% versus 1%, χ²(1) 10.9, p < 0.001).

Mackenzie subtype

90/222 (41%) CC75C cases presented with neuronal cytoplasmic inclusions or neurite TDP-43 pathology in at least one of the cortical areas investigated. Out of these, four cases had only parahippocampal but no temporal cortex sections available, 27/90 cases had neuronal cytoplasmic inclusions or neurite pathology in the parahippocampal cortex but not the temporal cortex, and 12/90 cases presented with neuronal cytoplasmic inclusions or neurite pathology only in the temporal cortex but not in the parahippocampal cortex. These 12 cases were also negative for hippocampal neuronal cytoplasmic inclusions and neurite pathology. Table 2 presents the frequency of neuronal cytoplasmic inclusions and neurites in the parahippocampal and temporal cortex.

Table 2

Frequency of TDP-43P immunoreactive neuronal solid cytoplasmic inclusions and neurites in the parahippocampal cortex and temporal cortex

TDP-43 pathology	Parahippocampal cortex (n = 222)	Temporal cortex (n = 215)
	n (%)	n (%)
NCI	70 (32)	48 (22)
Neurites	51 (23)	31 (14)

NCI, neuronal cytoplasmic inclusions.

36% (n = 32/90) of the cases with cortical neuronal cytoplasmic inclusions or neurite pathology were marked as Mackenzie type A, characterized by many short neurites and neuronal cytoplasmic inclusions, especially in cortical layer II. 7% (n = 6/90) were evaluated as type B, characterized by a moderate number of neuronal cytoplasmic inclusions and few short neurites in all cortical layers. However, most cases (58%, n = 52/90) were unclassifiable, mainly presenting with sparse TDP-43 pathology. Cases with glial cytoplasmic inclusions and/or GVD but no neurites or neuronal cytoplasmic inclusions are unclassifiable according to Mackenzie type [33], as the categories are based on neurite and neuronal inclusion pathology. No dystrophic long neurites (typical of FTLD-TDP subtype C), were seen in this aged population. Only 2/222 cases presented with lentiform “cat’s eye” neuronal nuclear inclusions (typical of FTLD-TDP subtype D) illustrated in Supplementary Figure 1.

Relationship between age and hippocampal TDP-43 immunoreactive pathologies

The relationships between age and presence of TDP-43 immunoreactive pathologies were tested using logistic regressions with TDP-43 immunoreactive pathologies as dependent variables and age at death as the independent variable (Table 3).

Table 3

Logistic regression analyses investigating the relationship between each hippocampal TDP-43 immunoreactive pathology and age at death

TDP-43 pathology	OR(95% CI)	p
GVD	1.13 (1.06; 1.21)	<0.001
Neurites	1.14 (1.06; 1.22)	<0.001
GCI-P	1.15 (1.05; 1.26)	0.002
NCI-P	1.14 (1.07; 1.22)	<0.001
NCI-U	1.14 (1.06; 1.22)	<0.001
TDP-43U loss	0.96 (0.91; 1.02)	0.170

Increasing age was significantly associated with the presence of glial cytoplasmic inclusions, neuronal cytoplasmic inclusions (TDP-43P and TDP-43U), neurites, and GVD; ORs varied between 1.13 and 1.15. Complete loss of TDP-43U staining representing clearance of unphosphorylated TDP-43 was not associated with age.

Relationships between dementia and TDP-43 immunoreactive pathologies

The distributions of TDP-43 immunoreactive pathologies scored as present or absent according to dementia severity is shown in Table 4.

Table 4

Distribution of the presence of hippocampal TDP-43 immunoreactive pathologies, according to dementia severity

TDP -43 pathology	Dementia severity n (%)^†
	None (n = 59)	MCI (n = 16)	Minimal (n = 23)	Mild (n = 23)	Moderate (n = 34)	Severe (n = 50)
GVD	31 (53)	9 (56)	14 (61)	17 (74)	27 (79)	42 (82)
Neurites	12 (20)	2 (13)	4 (17)	6 (26)	10 (29)	20 (38)
GCI-P	2 (3)	2 (13)	0 (0)	3 (13)	2 (6)	14 (26)
NCI-P	9 (15)	3 (19)	5 (22)	7 (30)	13 (37)	21 (40)
NCI-U	6 (10)	2 (13)	2 (9)	4 (17)	12 (34)	19 (36)
TDP-43U loss	32 (54)	11 (69)	8 (35)	12 (52)	17 (50)	31 (61)

^†Percentage of presence of TDP-43 immunoreactive pathologies among dementia severity category. GVD, granulovacuolar degeneration; GCI, glial solid cytoplasmic inclusions; NCI-P, neuronal solid cytoplasmic inclusions immunoreactive with anti-TDP-43P; NCI-U, neuronal solid cytoplasmic inclusions immunoreactive with anti-TDP-43U.

We investigated the relationships between TDP-43 pathology and dementia severity (Table 5) and dementia presence (Table 6). The analyses using clinical dementia severity as an outcome highlight those pathologies that may relate to dementia progression and the analyses using clinical dementia status (present/absent) estimate which pathologies may relate to presence of any clinical dementia.

Table 5

Comparison of ordered logistic regression analyses investigating the relationship between hippocampal TDP-43 immunoreactive pathologies and dementia severity (none, MCI, minimal, mild, moderate, versus severe) with and without controlling for age and other pathologies

TDP-43 pathology	Model I¹		Model II²		Model III³		Model IV⁴		I×II		I×III		I×IV
	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p	χ²(1)	p	χ²(1)	p	χ²(1)	p
GVD	2.75 (1.61; 4.72)	<0.001	2.71 (1.55; 4.74)	<0.001	2.08 (1.18; 3.64)	0.011	1.98 (1.12; 3.49)	0.018	0.04	0.843	11.98	0.001	18.91	<0.001
Neurites	2.06 (1.17; 3.62)	0.012	1.93 (1.07; 3.47)	0.028	1.86 (1.05; 3.31)	0.035	1.02 (0.39; 2.63)	0.968	0.64	0.425	17.02	<0.001	19.47	<0.001
GCI-P	5.09 (2.13; 12.11)	<0.001	4.76 (1.95; 11.61)	0.001	4.63 (1.89; 11.35)	0.001	3.46 (1.29; 9.30)	0.014	0.43	0.512	17.71	<0.001	19.58	<0.001
NCI-P	2.51 (1.46; 4.33)	0.001	2.39 (1.35; 4.23)	0.003	2.19 (1.25; 3.82)	0.006	–	–	0.33	0.564	16.82	<0.001	–	–
NCI-U	3.37 (1.84; 6.18)	<0.001	3.22 (1.73; 6.02)	<0.001	2.95 (1.58; 5.48)	0.001	–	–	0.36	0.550	16.34	<0.001	–	–
TDP-43U loss	1.12 (0.69; 1.83)	0.639	1.18 (0.72; 1.93)	0.508	1.16 (0.71; 1.90)	0.553	1.03 (0.63; 1.70)	0.906	2.12	0.145	20.63	<0.001	28.22	<0.001

¹Independent variables in model I: TDP-43 immunoreactive pathologies. ²Independent variables in model II: TDP-43 immunoreactive pathologies, controlling for age. ³Independent variables in model III: TDP-43 immunoreactive pathologies, controlling for classic dementia related pathologies. ⁴Independent variables in model III: TDP-43 immunoreactive pathologies, controlling for classic dementia related pathologies and neuronal cytoplasmic inclusions. GVD, granulovacuolar degeneration; GCI-P, glial solid cytoplasmic inclusions immunoreactive with anti-TDP-43P; NCI-P, neuronal solid cytoplasmic inclusions immunoreactive with anti-TDP-43P; NCI-U, neuronal solid cytoplasmic inclusions immunoreactive with anti-TDP-43U.

Table 6

Comparison of ordered logistic regression analyses investigating the relationship between hippocampal TDP-43 immunoreactive pathologies and dementia presence (absent versus present) with and without controlling for age and other pathologies

TDP-43 pathology	Model I¹		Model II²		Model III³		Model IV⁴		I×II		I×III		I×IV
	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p	χ²(1)	p	χ²(1)	p	χ²(1)	p
GVD	2.60 (1.43; 4.71)	0.002	2.56 (1.39; 4.74)	0.003	1.94 (1.02; 3.67)	0.043	1.87 (0.97; 3.57)	0.060	0.03	0.861	7.60	0.006	13.97	0.001
Neurites	1.82 (0.92; 3.610	0.084	1.71 (0.85; 3.44)	0.135	1.74 (0.86; 3.51)	0.121	0.62 (0.19; 2.05)	0.435	0.64	0.425	12.32	<0.001	17.32	<0.001
GCI-P	3.09 (1.02; 9.36)	0.046	2.84 (0.93; 8.72)	0.068	2.74 (0.89; 8.46)	0.080	1.51 (0.42; 5.41)	0.532	0.88	0.349	12.78	<0.001	16.92	<0.001
NCI-P	2.85 (1.41; 5.77)	0.004	2.71 (1.32; 5.60)	0.007	2.59 (1.26; 5.32)	0.010	–	–	0.32	0.571	11.86	0.001	–	–
NCI-U	3.29 (1.45; 7.44)	0.004	3.11 (1.35; 7.16)	0.008	2.90 (1.26; 6.69)	0.012	–	–	0.42	0.517	11.61	0.001	–	–
TDP-43U loss	0.79 (0.45; 1.38)	0.406	0.81 (0.46; 1.43)	0.465	0.76 (0.42; 1.37)	0.360	0.67 (0.37; 1.22)	0.188	1.14	0.285	14.53	<0.001	22.50	<0.001

¹Independent variables in model I: TDP-43 immunoreactive pathologies. ²Independent variables in model II: TDP-43 immunoreactive pathologies, controlling for age. ³Independent variables in model III: TDP-43 immunoreactive pathologies, controlling for other pathologies. ⁴Independent variables in model III: TDP-43 immunoreactive pathologies, controlling for other pathologies and TDP-43P neuronal solid cytoplasmic inclusions. GVD, granulovacuolar degeneration; GCI-P, glial solid cytoplasmic inclusions immunoreactive with anti-TDP-43P; NCI-P, neuronal solid cytoplasmic inclusions immunoreactive with anti-TDP-43P; NCI-U, neuronal solid cytoplasmic inclusions immunoreactive with anti-TDP-43U.

The relationships between dementia severity and presence of TDP-43 immunoreactive pathologies were investigated using ordinal logistic regressions with dementia severity as dependent variable and presence of TDP-43 immunoreactive pathologies as independent variables (Model I), while controlling for age at death (Model II) and other pathologies (Model III). χ² was used to test how the associations between dementia severity and TDP-43 immunoreactive pathologies (Model I) were modified by age (Models I and II) or other dementia related pathologies including neuritic plaques, tangles, and amyloid deposition (Models I and III) and additionally TDP-43 neuronal cytoplasmic inclusions (Model IV).

The presence of glial (TDP-43P) and neuronal cytoplasmic inclusions (TDP-43U and TDP-43P), GVD, and neurites were significantly related to dementia severity with (Model II) and without (Model I) controlling for of age (Table 5). When analyses were controlled for other AD-related pathologies (Braak stage, neuritic plaques, and amyloid deposition; Model III), glial and neuronal cytoplasmic inclusions, GVD, and neurites remained significantly associated with dementia severity. When the analyses were additionally controlled for TDP-43 immunoreactive neuronal cytoplasmic inclusions (Model IV), TDP-43P staining of both glial cytoplasmic inclusions and GVD retained significant associations with dementia severity; GVD OR = 1.58 (95% CI 1.010–2.29, p = 0.14) and glial cytoplasmic TDP-43 inclusions OR = 3.15 (95% CI 1.13–8.75, p = 0.28). Given the statistical effects are independent, the biological processes underlying GVD (immunostained by TDP-43P) and glial cytoplasmic TDP-43P inclusions may be additional to the more commonly considered neuronal cytoplasmic inclusions in the context of dementia progression.

When assessing dementia status as the outcome, as opposed to dementia severity, the pattern of results was slightly different (Table 6). GVD, glial cytoplasmic inclusions (TDP-43 P), and neural cytoplasmic inclusions (TDP-43 P and U) all displayed significant associations with dementia presence. GVD and neural cytoplasmic inclusions (TDP-43 P and U) but not glial cytoplasmic inclusions remained significantly associated when accounting for age (Model II) and other dementia-related pathologies (Model III). No pathology remained significantly associated with dementia presence when adjusted for the presence of neuronal cytoplasmic inclusions (Model IV).

Across the ordinal (dementia severity) and logistic (dementia presence) models, there were no significant differences when comparing Model I (constrained) with Model II (controlling for age), suggesting that that while TDP-43 deposition is more common with increasing age, the effects of age do not add greatly to the associations between TDP-43 pathologies and dementia severity or presence. The significant differences between Model I (constrained) and Model III (controlling for other pathologies) suggest that while TDP-43 deposition is significantly associated with dementia, the presence of other pathologies including tau neuritic plaques and tangles, and amyloid deposits contribute to dementia in addition to any effects from TDP-43. The significant differences between Model I (constrained) and Model IV (controlling for other pathologies and neuronal cytoplasmic inclusions seen with TDP-43P), suggest that while the presence of additional pathologies including TDP-43P-associated neuronal cytoplasmic inclusions additionally contribute to dementia severity, GVD, and glial cytoplasmic inclusions remain independently associated with dementia severity but not status.

DISCUSSION

We provide detailed characterizations of the relationships between the range of TDP-43-related pathologies and dementia, which are required to fully understand how TDP-43 contributes to dementia in the population. We investigated the relationships between both severity and presence of clinical dementia, age, and a range of pathologies associated with TDP-43 immunoreactivity and phosphorylation state. These include the well-recognized neuronal and glial cytoplasmic inclusions and neurites (Fig. 1A, B, E, F) and additional pathologies not previously investigated in a population representative sample including GVD and cytoplasmic clearance of TDP-43U without inclusion formation (Fig. 1C- I).

Different immunoreactivities are seen with TDP-43P and TDP-43U that have different associations with clinical dementia, highlighting important mechanistic contributions of TDP-43 phosphorylation state to dementia. The associations between TDP-43P immunoreactive GVD bodies and dementia suggest a significant interaction between GVD, TDP-43P, and dementia. Although relatively rare, glial inclusions in the older population associate with dementia more strongly than other TDP-43 associated pathologies suggesting the contributions of TDP-43 in glia to dementia should be investigated independently from neuronal contributions. The different associations between TDP-43 pathologies and either dementia severity or dementia status, suggest that TDP-43 in neurons and glia may contribute mechanistically in different ways to dementia initiation and progression. Our study adds to the characterization of TDP-43 pathology in the older population and adds timely detail to the characterization of limbic predominant age related TDP-43 encephalopathy [16].

CC75C is a well-characterized, population representative sample with mutually blinded clinical and neuropathological assessments. The estimates of clinicopathological associations presented here translate directly to the population. However, there are limitations. Our semi-quantitative estimates of pathology are not stereological. Stereological selection of tissue sections is not possible in a study such as this, as each brain area would need to be randomly sampled many times leading to loss of the tissue resource for this cohort, bias in future studies due to missing data as tissue is used up, and a much-increased cost. Also, tissue has been taken for other studies previously which could undermine any stereological sampling strategy. Instead, sections were taken from paraffin embedded blocks removed during the diagnostic process. Our measures for clinical dementia severity and status are not specific for any dementia type diagnosis and represent all dementia types seen in this cohort. Not all neurodegenerative disorders leading to dementia involve TDP-43 mechanistically, therefore our estimates of the associations between TDP-43 pathologies and dementia type involving TDP-43 will be underestimates.

TDP-43P immunostaining detected neuronal cytoplasmic inclusions in more brain donations than TDP-43U. Although neuronal cytoplasmic inclusions detected by either antibody showed an association with dementia, those detected using TDP-43U were more strongly associated with dementia than those detected using TDP-43P, and these associations remained stronger when age and other pathologies were controlled for. This could be due to the normally diffuse cytoplasmic and nuclear staining seen with anti-TDP-43U immunohistochemistry [7, 19] masking smaller inclusions that maybe more clearly seen with anti TDP-43P. However, given the loss of TDP-43U staining normally associated with inclusion formation, this explanation may not be adequate and further investigation to clarify the different associations of neuronal inclusions by phosphorylation state is required as phosphorylation state may have mechanistic implications.

We used only one antibody raised against TDP-43 specifically phosphorylated at serines 409/410-2, therefore our results do not represent pathology associated with any phosphorylation state only that associated with TDP-43 phosphorylated at the specific serines 409/410-2. However, we do show that in principle, phosphorylation of TDP-43 can lead to different pathological associations to those seen with unphosphorylated TDP-43, confirming the important role of phosphorylation in relation to disease mechanisms and therapeutic targets.

The CC75C neuropathological protocol samples only one hemisphere for formalin fixation and paraffin embedding. Therefore, the contralateral hippocampus is not available for this study. By sampling only one hemisphere we may have missed a number of cases with unilateral pathology. Although this has been shown to be important in population studies specifically relating to hippocampal sclerosis, where up to 49% of hippocampal sclerosis cases may present with unilateral neuronal loss [14], we assume that the different types of TDP-43 associated pathology seen will be similar unilaterally or bilaterally.

The classification of FTLD-TDP subtypes by Mackenzie et al. [33] is widely accepted and used. However, it was designed specifically based on FTLD-TDP cases, and the limitations become evident when applied to other TDP-43 proteinopathies and aging: cases may not follow any of the categories, others may present with characteristics fitting several categories, and the criteria only account for neuronal and neurite pathology while disregarding the variety of TDP-43 pathologies. FTLD subtypes are associated with younger age at onset and are associated with specific gene variants including valosin-containing protein (VCP), TDP-43 (TARDBP), and the non-coding region of C9ORF72 [38]. We do not have genetic data for these genes; however, nearly all brain donors were aged over 80 where genetically defined forms of FTLD may be much less common than in younger cohorts. Results from this study demonstrate the poor applicability of the current Mackenzie classification system to the older population. The reliable assignment of Josephs staging [34, 35] or the simple staging described in Nelson et al. [16] is not possible with our current dataset and further work investigating TDP-43 immunoreactive pathology in additional brain areas is required.

GVD is a common pathology in AD, aging, and Down’s syndrome and has been observed in a variety of other neurodegenerative diseases including Pick’s disease, corticobasal degeneration, Parkinson’s disease with dementia, and progressive supranuclear palsy [21]. The staining of GVD bodies by TDP-43P was extremely variable, ranging from no staining to strong staining of all bodies within a single neuron, with some cases showing very little staining across many neurons with GVD and other cases with strong staining of bodies across all neurons with GVD across the hippocampal areas (Fig. 1E, G). TDP-43U did not stain any GVD bodies (Fig. 1F, H). Additionally, TDP-43U clearance was not associated with GVD and cells with GVD showed normal nuclear staining (Fig. 1F, H), suggesting that “pre-inclusions” of TDP-43 can be distinguished from GVD by the presence of normal nuclear staining. In agreement with Kohler [21] and in contrast with others [20], our results suggest that while there is a common interaction with GVD, TDP-43P is not a consistent marker for GVD bodies. In agreement with previous findings of a relationship between GVD and dementia [39 –42], GVD stained with anti-TDP-43P were associated with dementia presence and severity, as well as age. However in contrast to the relationships seen in this cohort between GVD and tau [40], the association between GVD stained with anti-TDP-43P and dementia severity (but not status) did not lose statistical significance when controlling for other pathologies nor when additionally controlling for neuronal cytoplasmic inclusions immunoreactive with TDP-43P. This suggests that the involvement of TDP-43P in GVD may represent an independent neuropathological interaction related to dementia progression that requires further investigation. While tau is associated with both mature NFT and GVD bodies, neither TDP-43P nor TDP-43U immunostaining was found in association with tangles (data not shown) suggesting that it is not tau per se that associates with TDP-43 but GVD specifically. Our results do not support the hypothesis that GVD represent “pre-tangles” [21]. We also found structures marked by TDP-43P immunoreactivity that could be interpreted as ghost GVD (Fig, 1J, K) as previously seen with tau [40], indicating that GVD may be a complex pathology involving multiple disease associated pathways and is distinct from NFT [40]. Our evidence suggests that interpretations of previously reported tau-related “pre-tangles” [21, 43] and TDP-43-related “pre-inclusions” [22] pathologies may be confounded by GVD in neuropathological studies where GVD is not specifically accounted for in experimental designs. Further study in longitudinal mechanistic settings is required to clarify the meaning of these associations.

Glial cytoplasmic inclusions of aggregated TDP-43 are associated with various neurodegenerative diseases including old age hippocampal sclerosis [44], amyotrophic lateral sclerosis [45, 46], and FTLD [22]. Glial cytoplasmic inclusions stained with TDP-43P were most strongly associated with dementia severity but not dementia presence and this was independent from other pathologies including TDP-43 neuronal cytoplasmic inclusions. This finding highlights the additional contributions to dementia of TDP-43 in glia in addition to the contributions of TDP-43 in neurons. While every care was taken to identify glial pathologies, including considerations of nucleus size and structure and background neuritic pathology, differentiating glial cytoplasmic inclusions with certainty requires additional double staining for glial markers. We did not include double staining for a glial marker and scored only those pathologies that could be confidently assigned so we may not have identified all glial pathology present. However, AC2 agreements between both raters for glial pathologies immunostaining for TDP-43P were all 0.8 or above, assessed using the benchmark proposed by Landis and Koch [36] as near-perfect agreement. In contrast to staining with anti-TDP-43P, staining intensity with anti-TDP-43U was variable which, together with the commonly strong glial immunolabelling, confounded interpretations of glial inclusions. Data for glial cytoplasmic inclusions seen with TDP-43U failed the inter-rater analysis and were dropped and only scores for glial cytoplasmic inclusions seen with TDP-43P were used in analysis. Therefore, we cannot characterize the effect of phosphorylation state on glial pathology. Given that we show different associations with dementia for different neuronal immunoreactivities seen with TDP-43P and/or TDP-43U, the possible contributions of TDP-43 phosphorylation state in glia to dementia requires further investigation.

Increasing age was significantly associated with the presence of GVD, neurites, glial cytoplasmic inclusions, neuronal cytoplasmic inclusions seen with TDP-43P, and neuronal cytoplasmic inclusions seen with TDP-43U, but not with TDP-43U clearance. How these associations relate to disease processes cannot be determined from our cross-sectional sample and longitudinal modelling would be required to investigate these associations further.

The mechanisms relating TDP-43 to disease may involve a complex mix of gain and loss of function [23]. Toxic gains of function arise due to changes associated with aggregation, oligomerization and potential prion-like behavior [47]. Loss of function mechanisms are also described, e.g., nuclear clearance of TDP-43 is associated with increase in DNA double strand breaks in amyotrophic lateral sclerosis patients [48]. Our results show that nuclear clearance without inclusion formation in neurons was not associated with dementia (severity or status) or age. TDP-43U clearance was occasionally seen in dentate cells without formation of cytoplasmic inclusions (Fig. 1C, D). In agreement with previous work showing roles for TDP-43 in both apoptosis and necrosis [49] TDP-43U clearance was also seen in neurons with contracted nuclei and collapsed cell bodies that could be interpreted as neurons in the process of dying (Fig. 1I). In terms of disease modeling, our results suggest that complete loss of TDP-43U staining shows that TDP-43U clearance, while it is associated with neuronal cytoplasmic inclusion formation, may also be independent from this process.

The range of pathologies associated with TDP-43P and TDP-43U immunoreactivity described above supports the hypothesis that TDP-43 plays multiple roles in cellular systems [5] and that phosphorylation state is an important contributor to processes, such as GVD, that are known to be associated with disease. Further, our results support the difference in presentation of TDP-43 in FTLD compared to the older population, suggesting that different disease associated processes may be involved in different individuals [16]. It is not possible to translate these pathologies to any specific cellular process with our cross-sectional data and future mechanistic investigations linking processes and pathologies in different cell types would be useful to identify therapeutic targets. Our results suggest that TDP-43 pathology contributes to dementia status and progression via a variety of mechanisms in a range of phosphorylation states involving both neurons and glia independently from age and from the classic AD related pathologies involving tau and amyloid-β protein.

Footnotes

ACKNOWLEDGMENTS

We thank all the participants, their families, and carers in the CC75C study. CC75C is a member study of the Cambridgeshire and Peterborough Collaboration for Leadership in Applied Health Research and Care (CLAHRC). This work was funded by the Addenbrooke’s Charitable Trust. The Cambridge Brain Bank Laboratory (which processed all CC75C cases) is supported by the National Institute for Health Research, Cambridge Biomedical Research Centre. CB and SHu were supported by NIHR Senior Investigators grant. Shu is additionally supported by a Network Support Grant awarded by the ARUK and the Paul G. Allen Foundation. SHo was supported by an ARUK funded PhD studentship. HK was supported by NHMRC grants GNT1042889 and GNT1135676. Thais Minett was funded by an Academic Clinical Fellowship from National Institute for Health Research.

This article presents independent research funded by the National Institute for Health Research. The views expressed are those of the authors and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Savva

, Wharton

, Ince

, Forster

, Matthews

, Brayne

(2009) Age, neuropathology, and dementia. N Engl J Med 360, 2302–2309.

Keage

, Hunter

, Matthews

, Ince

, Hodges

, Hokkanen

, Highley

, Dening

, Brayne

(2014) TDP-43 pathology in the population: Prevalence and associations with dementia and age. J Alzheimers Dis 42, 641–650.

MRC-CFAS (2001) Pathological correlates of late-onset dementia in a multicentre, community-based population in England and Wales. Neuropathology Group of the Medical Research Council Cognitive Function and Ageing Study (MRC CFAS). Lancet 357, 169–175.

Brayne

, Richardson

, Matthews

, Fleming

, Hunter

, Xuereb

, Paykel

, Mukaetova-Ladinska

, Huppert

, O’Sullivan

, Dening

(2009) Neuropathological correlates of dementia in over-80-year-old brain donors from the population-based Cambridge city over-75s cohort (CC75C) study. J Alzheimers Dis 18, 645–658.

Lee

, Lee

, Trojanowski

(2012) Gains or losses: Molecular mechanisms of TDP43-mediated neurodegeneration. Nat Rev Neurosci 13, 38–50.

Dewey

, Cenik

, Sephton

, Johnson

, Herz

, Yu

(2012) TDP-43 aggregation in neurodegeneration: Are stress granules the key? Brain Res 1462, 16–25.

Neumann

, Sampathu

, Kwong

, Truax

, Micsenyi

, Chou

, Bruce

, Schuck

, Grossman

, Clark

, McCluskey

, Miller

, Masliah

, Mackenzie

, Feldman

, Feiden

, Kretzschmar

, Trojanowski

, Lee

(2006) Ubiquitinated TDP-43 in frontotemporal lobar degeneration and amyotrophic lateral sclerosis. Science 314, 130–133.

Amador-Ortiz

, Lin

, Ahmed

, Personett

, Davies

, Duara

, Graff-Radford

, Hutton

, Dickson

(2007) TDP-43 immunoreactivity in hippocampal sclerosis and Alzheimer’s disease. Ann Neurol 61, 435–445.

Davidson

, Raby

, Foulds

, Robinson

, Thompson

, Sikkink

, Yusuf

, Amin

, DuPlessis

, Troakes

, Al-Sarraj

, Sloan

, Esiri

, Prasher

, Allsop

, Neary

, Pickering-Brown

, Snowden

, Mann

(2011) TDP-43 pathological changes in early onset familial and sporadic Alzheimer’s disease, late onset Alzheimer’s disease and Down’s syndrome: Association with age, hippocampal sclerosis and clinical phenotype. Acta Neuropathol 122, 703–713.

10.

Uryu

, Nakashima-Yasuda

, Forman

, Kwong

, Clark

, Grossman

, Miller

, Kretzschmar

, Lee

, Trojanowski

, Neumann

(2008) Concomitant TAR-DNA-binding protein 43 pathology is present in Alzheimer disease and corticobasal degeneration but not in other tauopathies. J Neuropathol Exp Neurol 67, 555–564.

11.

Pao

, Dickson

, Crook

, Finch

, Rademakers

, Graff-Radford

(2011) Hippocampal sclerosis in the elderly: Genetic and pathologic findings, some mimicking Alzheimer disease clinically. Alzheimer Dis Assoc Disord 25, 364–368.

12.

Kovacs

, Milenkovic

, Wohrer

, Hoftberger

, Gelpi

, Haberler

, Honigschnabl

, Reiner-Concin

, Heinzl

, Jungwirth

, Krampla

, Fischer

, Budka

(2013) Non-Alzheimer neurodegenerative pathologies and their combinations are more frequent than commonly believed in the elderly brain: A community-based autosy series. Acta Neuropathol 126, 365–384 .

13.

Wilson

, Yu

, Trojanowski

, Chen

, Boyle

, Bennett

, Schneider

(2013) TDP-43 pathology, cognitive decline, and dementia in old age. JAMA Neurol 70, 1418–1424.

14.

Kero

, Raunio

, Polvikoski

, Tienari

, Paetau

, Myllykangas

(2018) Hippocampal sclerosis in the oldest old: A Finnish population-based study. J Alzheimers Dis 63, 263–272.

15.

Nelson

, Smith

, Abner

, Wilfred

, Wang

, Neltner

, Baker

, Fardo

, Kryscio

, Scheff

, Jicha

, Jellinger

, Van Eldik

, Schmitt

(2013) Hippocampal sclerosis of aging, a prevalent and high-morbidity brain disease. Acta Neuropathol 126, 161–177.

16.

Nelson

, Dickson

, Trojanowski

, Jack

, Boyle

, Arfanakis

, Rademakers

, Alafuzoff

, Attems

, Brayne

, Coyle-Gilchrist

ITS

, Chui

, Fardo

, Flanagan

, Halliday

, Hokkanen

SRK

, Hunter

, Jicha

, Katsumata

, Kawas

, Keene

, Kovacs

, Kukull

, Levey

, Makkinejad

, Montine

, Murayama

, Murray

, Nag

, Rissman

, Seeley

, Sperling

, White Iii

, Yu

, Schneider

(2019) Limbic-predominant age-related TDP-43 encephalopathy (LATE): Consensus working group report. Brain 142, 1503–1527.

17.

Arai

, Mackenzie

, Hasegawa

, Nonoka

, Niizato

, Tsuchiya

, Iritani

, Onaya

, Akiyama

(2009) Phosphorylated TDP-43 in Alzheimer’s disease and dementia with Lewy bodies. Acta Neuropathol 117, 125–136.

18.

Neumann

, Kwong

, Lee

, Kremmer

, Flatley

, Xu

, Forman

, Troost

, Kretzschmar

, Trojanowski

, Lee

(2009) Phosphorylation of S409/410 of TDP-43 is a consistent feature in all sporadic and familial forms of TDP-43 proteinopathies. Acta Neuropathol 117, 137–149.

19.

Cairns

, Neumann

, Bigio

, Holm

, Troost

, Hatanpaa

, Foong

, White

3rd , Schneider

, Kretzschmar

, Carter

, Taylor-Reinwald

, Paulsmeyer

, Strider

, Gitcho

, Goate

, Morris

, Mishra

, Kwong

, Stieber

, Xu

, Forman

, Trojanowski

, Lee

, Mackenzie

(2007) TDP-43 in familial and sporadic frontotemporal lobar degeneration with ubiquitin inclusions. Am J Pathol 171, 227–240.

20.

Kadokura

, Yamazaki

, Kakuda

, Makioka

, Lemere

, Fujita

, Takatama

, Okamoto

(2009) Phosphorylation-dependent TDP-43 antibody detects intraneuronal dot-like structures showing morphological characters of granulovacuolar degeneration. Neurosci Lett 463, 87–92.

21.

Kohler

(2016) Granulovacuolar degeneration: A neurodegenerative change that accompanies tau pathology. Acta Neuropathol 132, 339–359.

22.

Geser

, Martinez-Lage

, Robinson

, Uryu

, Neumann

, Brandmeir

, Xie

, Kwong

, Elman

, McCluskey

, Clark

, Malunda

, Miller

, Zimmerman

, Qian

, Van Deerlin

, Grossman

, Lee

, Trojanowski

(2009) Clinical and pathological continuum of multisystem TDP-43 proteinopathies. Arch Neurol 66, 180–189.

23.

Halliday

, Bigio

, Cairns

, Neumann

, Mackenzie

, Mann

(2012) Mechanisms of disease in frontotemporal lobar degeneration: Gain of function versus loss of function effects. Acta Neuropathol 124, 373–382.

24.

Fleming

, Zhao

, O’Connor

, Pollitt

, Brayne

(2007) Cohort profile: The Cambridge City over-75s Cohort (CC75C). Int J Epidemiol 36, 40–46.

25.

American Psychiatric Association (1987) Diagnostic and statistical manual of mental disorders. DSM-III-R, American Psychiatric Association, Washington, DC.

26.

Folstein

, Folstein

, McHugh

(1975) “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

27.

Roth

, Tym

, Mountjoy

, Huppert

, Hendrie

, Verma

, Goddard

(1986) CAMDEX. A standardised instrument for the diagnosis of mental disorder in the elderly with special reference to the early detection of dementia. Br J Psychiatry 149, 698–709.

28.

Brayne

, Richardson

, Matthews

, Fleming

, Hunter

, Xuereb

, Huppert

, O’Sullivan

, Paykel

, Dening

, Cambridge City Over-75s Cohort Cc75c Study Neuropathology Collaboration (2009) Neuropathological correlates of dementia in over-80-year-old brain donors from the population-based Cambridge City over-75s Cohort (CC75C) study. J Alzheimers Dis 18, 1875–8908.

29.

Mirra

(1997) Neuropathological assessment of Alzheimer’s disease: The experience of the Consortium to Establish a Registry for Alzheimer’s Disease. Int Psychogeriatr 9(Suppl 1), 263–268; discussion 269-272.

30.

Mirra

, Heyman

, McKeel

, Sumi

, Crain

, Brownlee

, Vogel

, Hughes

, van Belle

, Berg

(1991) The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD). Part II. Standardization of the neuropathologic assessment of Alzheimer’s disease. Neurology 41, 479–486.

31.

Braak

, Braak

(1995) Staging of Alzheimer’s disease-related neurofibrillary changes. Neurobiol Aging 16, 271–278; discussion 278-284.

32.

Braak

, Braak

(1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol 82, 239–259.

33.

Mackenzie

, Neumann

, Baborie

, Sampathu

, Du Plessis

, Jaros

, Perry

, Trojanowski

, Mann

, Lee

(2011) A harmonized classification system for FTLD-TDP pathology. Acta Neuropathol 122, 111–113.

34.

Josephs

, Murray

, Whitwell

, Parisi

, Petrucelli

, Jack

, Petersen

, Dickson

(2014) Staging TDP-43 pathology in Alzheimer’s disease. Acta Neuropathol 127, 441–450.

35.

Josephs

, Murray

, Whitwell

, Tosakulwong

, Weigand

, Petrucelli

, Liesinger

, Petersen

, Parisi

, Dickson

(2016) Updated TDP-43 in Alzheimer’s disease staging scheme. Acta Neuropathol 131, 571–585.

36.

Landis

, Koch

(1977) The measurement of observer agreement for categorical data. Biometrics 33, 159–174.

37.

Montine

, Phelps

, Beach

, Bigio

, Cairns

, Dickson

, Duyckaerts

, Frosch

, Masliah

, Mirra

, Nelson

, Schneider

, Thal

, Trojanowski

, Vinters

, Hyman

, National Institute on Aging, Alzheimer’s Association (2012) National Institute on Aging-Alzheimer’s Association guidelines for the neuropathologic assessment of Alzheimer’s disease: A practical approach. Acta Neuropathol 123, 1–11.

38.

Bigio

(2011) TDP-43 variants of frontotemporal lobar degeneration. J Mol Neurosci 45, 390–401.

39.

Keage

, Ince

, Matthews

, Wharton

, McKeith

, Brayne

(2012) Impact of less common and “disregarded” neurodegenerative pathologies on dementia burden in a population-based cohort. J Alzheimers Dis 28, 485–493.

40.

Hunter

, Minett

, Polvikoski

, Mukaetova-Ladinska

, Brayne

(2015) Re-examining tau-immunoreactive pathology in the population: Granulovacuolar degeneration and neurofibrillary tangles. Alzheimers Res Ther 7, 57.

41.

Ball

, Lo

(1977) Granulovacuolar degeneration in the ageing brain and in dementia. J Neuropathol Exp Neurol 36, 474–487.

42.

Thal

, Del Tredici

, Ludolph

, Hoozemans

, Rozemuller

, Braak

, Knippschild

(2011) Stages of granulovacuolar degeneration: Their relation to Alzheimer’s disease and chronic stress response. Acta Neuropathol 122, 577–589.

43.

Hoozemans

, van Haastert

, Nijholt

, Rozemuller

, Eikelenboom

, Scheper

(2009) The unfolded protein response is activated in pretangle neurons in Alzheimer’s disease hippocampus. Am J Pathol 174, 1241–1251.

44.

Hokkanen

SRK

, Hunter

, Polvikoski

, Keage

HAD

, Minett

, Matthews

, Brayne

(2018) Hippocampal sclerosis, hippocampal neuron loss patterns and TDP-43 in the aged population. Brain Pathol 28, 548–559.

45.

Mackenzie

, Bigio

, Ince

, Geser

, Neumann

, Cairns

, Kwong

, Forman

, Ravits

, Stewart

, Eisen

, McClusky

, Kretzschmar

, Monoranu

, Highley

, Kirby

, Siddique

, Shaw

, Lee

, Trojanowski

(2007) Pathological TDP-43 distinguishes sporadic amyotrophic lateral sclerosis from amyotrophic lateral sclerosis with SOD1 mutations. Ann Neurol 61, 427–434.

46.

Ince

, Highley

, Kirby

, Wharton

, Takahashi

, Strong

, Shaw

(2011) Molecular pathology and genetic advances in amyotrophic lateral sclerosis: An emerging molecular pathway and the significance of glial pathology. Acta Neuropathol 122, 657–671.

47.

Prasad

, Bharathi

, Sivalingam

, Girdhar

, Patel

(2019) Molecular mechanisms of TDP-43 misfolding and pathology in amyotrophic lateral sclerosis. Front Mol Neurosci 12, 25.

48.

Mitra

, Guerrero

, Hegde

, Liachko

, Wang

, Vasquez

, Gao

, Pandey

, Taylor

, Kraemer

, Wu

, Boldogh

, Garruto

, Mitra

, Rao

, Hegde

(2019) Motor neuron disease-associated loss of nuclear TDP-43 is linked to DNA double-strand break repair defects. Proc Natl Acad Sci U S A 116, 4696–4705.

49.

Chi

, Chang

, Sang

(2018) Neuronal cell death mechanisms in major neurodegenerative diseases. Int J Mol Sci 19, E3082.