Introduction
Microtubule associated protein tau (tau) is a cytoskeletal stabilizing protein involved in microtubule maintenance, fast axonal transport, and other physiological functions in neurons. Tau protein deposition is a characteristic of many neurodegenerative diseases that are collectively referred to as tauopathies. It has been shown that pathological species of tau protein are abnormally phosphorylated at multiple residues [
1] and that this is linked to a decrease in tau’s ability to bind to and stabilize microtubules [
2,
3] with accompanying cytotoxicity [
4]. While the isoform composition of insoluble tau deposits and the structural formation of the protein aggregates differs amongst the tauopathies, there are several phosphorylation sites that are thought to be universally important in the induction of a tauopathy.
One potentially important phosphorylation site that has gone relatively unstudied is threonine
175 (Thr
175). First identified in Alzheimer’s disease as a unique phosphoepitope [
5], it was then determined that this site could be phosphorylated by multiple kinases linked to tau protein pathology, including GSK3β, JNK, ERK2, and p38 [
6]. pThr
175 tau was then identified in amyotrophic lateral sclerosis with cognitive impairment (ALSci) [
7] and characterized in further detail in the context of this disease [
8‐
12]. Importantly, pThr
175 tau has been shown to induce tau fibril formation and cell death in vitro [
13]. Unlike other widely accepted pathological phosphorylation sites on tau, such as pThr
231 and pSer
202, pThr
175 has not been observed in the fetal brain where tau protein is hyperphosphorylated [
14‐
16], suggesting that this site may be uniquely associated with pathological processes. pThr
175 tau has been shown to induce GSK3β activation in cell culture and may therefore act as a destabilizing event resulting in enhanced phosphorylation of tau at other residues, resulting in dissociation from microtubules and neuronal toxicity [
17]. In order to understand the extent to which this pathway of pThr
175 mediated tau aggregate formation underlies a broad range of tauopathies, we have used a panel of phospho-specific antibodies to characterize tau protein pathology with specific interest in the expression of pThr
175 tau across a broad range of tauopathies.
Materials and methods
Diseases studied included Alzheimer’s (AD; 3 cases), vascular dementia (VD; 2 cases), ALS (5 cases), ALSci (6 cases), dementia with Lewy Bodies (DLBD; 2 cases), DLBD with mixed pathology (mDLBD; 3 cases including 2 with DLBD/VD and 1 with DLBD/AD)), frontotemporal lobar dementia (FTLD-TDP; 3 cases including one with a pathological C9orf72 hexanucleotide expansion with Type B pathology; a single case with Type A pathology and a single case with Type B pathology, FTLD-Tau; 1 case with familial history and no known mutations) [
18], multiple system atrophy (MSA; 6 cases), Parkinson’s disease (PD; 5 cases), Pick’s disease (1 case), and corticobasal degeneration (CBD; 2 cases) (Table
1). The institutional research ethics board approved the protocol and consent was given for use of all tissue used in this study. All neuropathological diagnosis were performed by a neuropathologist (RH, LCA) and conformed to international neuropathological criteria [
19‐
21]. For all comparisons, we grouped the staining according to ALS (
n = 5), ALSci (
n = 6), or other tauopathy (
n = 22).
Table 1
Case demographics
AD | 72 ± 8 | 3 (2) |
VD | 78 ± 11 | 2 (1) |
ALS | 56 ± 16 | 5 (4) |
ALSci | 64 ± 11 | 6 (5) |
DLBD | 68 ± 1 | 2 (2) |
mDLDB | 83 ± 6 | 3 (3) |
FTLD | 64 ± 9 | 4 (1) |
MSA | 69 ± 12 | 6 (3) |
PD | 77 ± 2 | 5 (4) |
Pick’s | 70 ± 2 | 2 (2) |
CBD | 71 ± 1 | 2 (1) |
Control 1 | 55 ± 2 | 6 (4) |
Control 2 | 64 ± 2 | 6 (4) |
Control 3 | 75 ± 3 | 8 (4) |
To assess the extent of pThr
175 tau, pThr
231 tau and tau oligomer pathological inclusions as a function of ageing, three groups of controls were studied, encompassing the 6
th (
n = 6), 7
th (
n = 6), and 8
th (
n = 8) decades of life (Table
1). Hippocampal sections from each group were stained for pThr
175 tau, pThr
231 tau and oligomeric tau (T22). These cases have been previously characterized in a study examining age-dependant tau deposition in the frontal and entorhinal cortices and were shown to be free of neurodegenerative disease [
22].
Five to six micrometer paraffin-embedded sections from the superior frontal gyrus, anterior cingulate (ACC), hippocampus, entorhinal cortex, dentate gyrus, amygdala, basal ganglia and substantia nigra were used for all immunohistochemical analyses.
Cases were stained by haematoxylin and eosin (H&E) and Gallyas silver stain for routine histological analysis and overall pathology characterization. Immunohistochemistry was conducted using a series of antibodies (Table
2) previously characterized in ALSci [
12], consisting of PHF tau (AT8; Thermo Fischer IL, Canada), pThr
175 tau, pSer
208,210 tau, pThr
217 tau (antibodies generated and designed in house [
12], pThr
175 commercially available through 21
st Century, MA, USA).
Tau pThr175
| Rabbit, polyclonal | 1:1000 | 1 | pThr175
| 21st Century |
Tau pThr217
| Rabbit, polyclonal | 1:1000 | 1 | pThr217
| 21st Century |
Tau pSer208, 210
| Rabbit, polyclonal | 1:1000 | 1 | pSer208, 210
| 21st Century |
PHF (AT8) | Mouse, monoclonal | 2.5 ug/ml | No | pSer202
| Thermo Fischer |
Tau pThr231
| Rabbit, polyclonal | 1:1000 | 2 | pThr231
| Thermo Fischer |
T22 | Rabbit, polyclonal | 1:500 | 2 | Tau oligomer | EMD Millipore |
Alexa Fluor 488 | Goat anti-rabbit | 1:200 | 2 | Secondary | Life Technologies |
Antigen retrieval was conducted as necessary (Table
2). Endogenous peroxidase was quenched with 3% hydrogen peroxide (BDH Chemicals, VWR, On, Canada). Primary antibody incubation was performed at 4 °C overnight in blocking buffer (5% BSA, 0.3% Triton-X 100 in 1 X PBS). After washing, secondary antibody (1:200 biotinylated IgG) incubation was performed for 1 h at room temperature in blocking buffer. Antigen:antibody complex was visualized with either horseradish peroxidase or alkaline phosphatase according to the manufacturer’s instructions (Vectastain ABC kit, Vector Laboratories CA, USA), followed by substrate development with either DAB plus NiCl
2 or AP substrate kit III (Vector Laboratories). Counterstaining was performed using haematoxylin or nuclear fast red. The extent of pathology was described topographically and semi-quantitatively as previously reported [
12]. Representative images were captured with a 20x lens under light microscopy (Olympus BX45) and subsequently used for semi-quantitative analysis. The semi-quantitative scale was manually applied for each type of pathology by an evaluator blinded to the underlying diagnosis (WY) (neuronal, neuritic, or glial) separately as follows: ‘-’ = none; ‘±’ = less than 5 inclusions; ‘+’ = less than 10 inclusions; ‘++’ = more than 20 inclusions with scattered distribution; ‘+++’ = more than 20 inclusions but with locally dense distribution; ‘++++’ = more than 20 inclusions with a diffuse distribution. Additionally, the case positive ratio was defined for each antibody used and brain region investigated as the number of cases showing any pathology (± or more) compared to the total number of cases stained.
Oligomeric tau and pThr231 staining
Rabbit anti T22 (EMD Millipore CA, USA) and rabbit anti tau pThr
231 (Thermo Fischer) were used to probe tau inclusions for the presence of oligomeric tau (T22) and for phosphorylation at Thr
231. Tau oligomeric species are currently hypothesized to be more toxic to neurons than the fibrillar inclusions themselves [
23], and pThr
231 is thought to be a key site in the regulation of tau protein folding and ability to interact with microtubules [
3,
24]. Double labeling was performed on hippocampus from one case each from AD, ALSci, FTD, MSA, DLDB, and mDLDB. Tau protein was probed for pThr
175 using rabbit primary antibody (1:1000) overnight at 4 °C and Alexa Fluor goat anti-rabbit 488 nm secondary (1:200, Thermo Fischer) for 1 h at room temperature. Rabbit anti tau pThr
231 antibody was then labeled using a Zenon primary antibody labeling kit with Alexa Fluor 555 nm dye (Thermo Fisher) and probed for 1 h at room temperature. Slides were stored overnight at 4 °C and visualized within 24 h of labeling by confocal imaging on a Zeiss LSM 510 Meta NLO multiphoton confocal microscope.
Discussion
In undertaking these studies, we were specifically interested in determining whether the pathogenic phospho-tau species recognized by antibodies against pThr
175 tau and pThr
231 tau as well as oligomeric tau (T22) were expressed across a broad range of tauopathies. We were also interested in determining whether these pathological tau species were colocalized in ALS and ALSci. It is known that the phosphorylation of tau protein at Thr
231 is of both physiological and pathological significance in mediating the dissociation of tau from microtubules [
3,
24,
25]. Thr
231 is phosphorylated by activated GSK3β physiologically and in pathological states [
25‐
29]. We have previously shown that pThr
175 tau induces GSK3β phosphorylation and that this in turn leads to Thr
231 tau phosphorylation resulting in tau fibril formation and cell death in vitro [
17].
Although the number of cases studied here is limited, the intent was not to undertake a detailed topographic analysis of tau deposition across all tauopathies, but rather to determine whether the proposed pathway of pThr
175 tau mediated induction of pThr
231 tau formation with its attendant pathological tau fibril formation (as recognized by T22) was evident. It is noteworthy therefore that we observed that in each tauopathy studied, pThr
175 tau, pThr
231 tau and T22 immunoreactivity co-localized to the same inclusion-containing neuronal populations. In each case, neuronal pThr
175 tau colocalized with pThr
231 tau. This, paired with the prior identification of pThr
175 tau in AD brain tissue but not controls [
30] and the lack of identified pThr
175 in fetal tau [
15,
16] suggests that pThr
175 is a key point in pathological tau metabolism, as it is not a physiologically utilized site involved in the regulation of tau function during development or microtubule reorganization. This suggests that the downstream events triggered by pThr
175 tau, including toxic monomer formation, are common to each of these diseases.
To further assess the pathogenicity of pThr
175 and pThr
231, we investigated each epitope in the hippocampus of control cases across three decades of life where tau pathology has been shown to increase with age [
22]. We observed no pThr
175 tau pathology in the 6
th decade with minimal immunoreactive neuronal inclusions in the 7
th decade. pThr
175 tau immunoreactivity was most prominent in the 8
th decade. In each case in which we observed pThr
175 tau immunostaining, we also observed T22 immunoreactivity. Similarly, we never observed T22 immunoreactivity in the absence of either pThr
175 tau or pThr
231 tau immunoreactivity. In contrast, pThr
231 tau immunoreactivity was frequently observed in the absence of either pThr
175 tau or T22 staining in younger individuals and when present, was within healthy appearing neurons and axonal processes. pThr
175 and T22 did not show pathology in hippocampal regions spared from pThr
231 pathology, and T22 was only positive in cases showing prominent pThr
175 pathology.
Glial pathology was recognized to a greater degree by pThr217 tau and PHF tau than by the pThr175 tau antibody, suggesting that different pathological processes are at play in these cells. This is supported by the lack of identifiable glial pathology by pThr231 tau and T22. This paired with the low frequency of pThr175 tau glial pathology further strengthens the correlation between pThr175 and pThr231 in the induction of neuronal pathology and provides evidence that this pair of phosphorylation sites may be exerting specific neuronal toxicity in the disease process across multiple tauopathies.
Although limbic regions universally presented tau pathology, frontal and ACC tau pathology was present mainly in AD, VD, ALSci, FTLD, mDLBD and MSA. This paired with the deeper layer pathology in this region may indicate that tau pathology did not originate here but instead propagated from other regions. If tau originates in limbic structures, propagating along the Papez circuit, it is possible that it would arrive in ACC through thalamic projections to layer IV and V which could act as a hub for propagation to other brain regions such as frontal cortex through this well connected region. Regardless of the induction cause or place, tau protein toxicity is undeniable once initiated [
23,
31], and must be considered when attempting to understand the underlying biology of many neurodegenerative diseases. This hypothesis also implies that disease entities such as primary age-related tauopathies (PART) [
32] may be in fact not age-related, but neuronal stress related, as increasing age would indicate longer time periods for stresses on neurons to become pathological through stochastic processes [
33]. Therefore, tau protein deposition should not be considered a simple function of normal ageing, but ageing should be considered a risk factor for tauopathy among a plethora of neuronal stresses. Of note as well is the frontal involvement in ALSci, which can be concluded is not likely a result of PART, which spares the neocortex by definition [
22,
32]. We cannot conclude, however if the layer distribution of tau pathology resembles PART, as this was not described in the consensus report.
Acknowledgements
Thanks to Dr. David Ramsay and Dr. David Munoz for their role in the neuropathological diagnosis of several cases used in this study. Research supported by an operating grant from the Ontario Neurodegenerative Disease Research Initiative and the Windsor-Essex ALS Association.