Introduction
Alzheimer’s Disease (AD) is characterised by the presence of two types of protein aggregates, amyloid-β (Aβ)-containing plaques and neurofibrillary tangles (NFTs), the latter containing the microtubule-associated protein tau in a hyperphosphorylated form [
1]. Regarding the latter, their distribution and density follows a distinct pattern through anatomically connected brain regions. NFTs constitute the basis for the Braak staging in AD [
2] that is routinely used for post-mortem diagnosis [
3]. This staging has led to the suggestion of a stereotypical spread of tau pathology, proposed to be a result of cross-synaptic spreading of a corruptive pathological protein species from a diseased to a healthy neuron in a “prion-like” manner along neuronal connections, and has been termed the spreading hypothesis [
4,
5]. In Prion disease it has been demonstrated that a pathological protein seed is capable of inducing misfolding of its native form and propagation of pathology between neurons, with severe down-stream functional consequences [
6]. In recent years, by extrapolation, cell-to-cell protein spreading and seeding have been suggested to underlie multiple neurodegenerative diseases [
7,
8] including AD, with a focus on a histological assessment and so far not so much on ensuing clinical changes [
9].
To address tau spreading in animal models, transgenic mice have been used that express mutant forms of tau found in familial cases of frontotemporal dementia (FTD), using promoter elements that confer broad expression of the transgene [
10]. Two principal approaches have been pursued. Firstly, by using inducible models and brain area-specific promoters to restrict the expression of FTD mutant tau to brain regions such as the entorhinal cortex, an early site of tau pathology in human AD, pathologically folded tau was found in neighbouring regions, indicative of tau spreading [
11,
12]. Although complete restriction of the transgene has been refuted in a subsequent study [
13], it has been argued that the changes seen in these neighbouring regions cannot be traced back to the low degree of ‘transgene leakage’. In a second approach, injection of mouse brain lysates isolated from FTD tau mutant mice into pre-symptomatic recipient mice has been shown to cause an accelerated, mature tau pathology both at the injection site and also, in neighbouring regions [
14]. Moreover, following the injection of the corresponding brain extracts, the hallmark lesions of tauopathies such as AGD, PSP and CBD have been recapitulated [
15]. While there has been indication that endogenous tau is not required to initiate tau pathology, recent data has demonstrated that a tau knock-out background does not affect the ability of tau to propagate but can reduce the pathological alterations induced by the ‘seed’ [
16]. Furthermore, injection of synthetic tau fibrils alone are able to induce the propogation of tau pathology [
17,
18]. However, the use of mature fibrillar tau seeds in previous seeding studies supersedes the intial seeding event causal of the primary aggregation events in the human disease state. Therefore, whether the effects observed in these studies are reliant on the introduction of a preformed mature tau seeds remains unknown.
To address this question we used a novel approach by creating a localised insult through injection of a single low dose of the protein phosphatase inhibitor okadaic acid (OA) in a very small volume, unilaterally targeting the amygdala, a limbic area affected early on in AD [
19]. Although OA predominantly inhibits PP1 and PP2A, it has a much higher affinity for the latter [
20]. OA inhibits the ability of PP2A to dephosphorylate tau resulting in tau hyperphosphorylation, which is exacerbated by an increase in kinase activity [
21]. Increases of endogenous PP2A inhibitors have been reported in human cases of AD, where they co-localise with phospho-tau and early NFT markers [
22]. Therefore, we propose that targeting the inhibition of PP2A will provide a spreading stimulus which precedes mature filament formation. Previous animal studies have primarily utilised OA at high doses or by chronic delivery to the hippocampus or ventricular system to recapitulate multiple features of both biochemical and behavioural dysfunctions of AD [
23‐
25], whereas we used a single low dose of 10 ng. Secondly, through using a low 130 nl volume of 100 μM OA we aimed to avoid spill-over, to ensure specific OA exposure. Finally, by using wild-type mice we were able to investigate the potential for endogenous tau phosphorylation to initiate spreading in the absence of tau mutations or external preformed tau seeds, to address whether these effects can be observed in brain regions anatomically distinct to that of the primary insult.
Materials and methods
Experimental design
The objective of this study was to investigate consequences of localised OA-induced PP2A inhibition on tau phosphorylation at both the injection site and across the brain. We injected 8 month-old female C57Bl/6 mice with a single low dose of 10 ng OA to the lateral amygdala. By targeting the lateral amygdala we are able to use areas previously reported as susceptible to tau seeding as secondary spreading targets, such as of the hippocampus. The co-ordinates were chosen in order to avoid passing through the ventricle or the hippocampal formation, and the lateral amygdala was chosen due to its confinement within the divergence of the external capsule, to aid in the restriction of OA diffusion. DMSO was used to resuspend and solubilise OA and was volume-matched as a control throughout. An OA-competitive ELISA was used to confirm injection and define diffusion from the injection site. At 24 h and 7 d post-injection, tau phosphorylation was quantified at the injection site and at anatomically distinct regions of interest throughout the brain using immunohistochemistry for the tau phospho-epitope AT180 (pTh231/pSer235) [
26]. Transition of tau into an insoluble state was quantified by Western blotting of lysates fractionated by solubility. Aggregation was assessed by reactivity with thioflavin-S that detects β-pleated sheet structures [
27].
Animals
Female wild-type C57BL/6 mice were used from 6 to 8 months of age. Female
MAPT KO mice [
28] were used as controls from 3 to 8 months of age. All experiments were carried out in compliance with ethical approval from the UQ animal welfare unit (QBI/412/14/NHMRC; QBI/027/12/NHMRC).
Stereotaxic surgery
Mice were anaesthetised with isoflurane vaporised in oxygen. The head was shaved and depilated before being placed securely in ear bars. Mice were injected unilaterally with 130 nl of 100 μM OA (Sigma) solubilised in DMSO or DMSO alone to the lateral amygdala (Bregma coordinates: anterior/posterior −1.94, medial/lateral −3.15, dorsal/ventral −4.5) at an infusion rate of 60 nl/min. The same volume of the dye Evans Blue was injected in the same manner to confirm the injection site. Following the injection, the needle was left in place for 5 min before being slowly withdrawn, and the scalp was securely sutured. Once reflexes returned, postoperative analgesia was administered subcutaneously (1 mg/kg torbugesic) and mice were returned to their home cages. Mice were left to recover for 30 min, 24 h or 7 d. For histology, mice were anaesthetised with a lethal dose of pentabarbitone before transcardial perfusion with 30 ml PBS followed by 30 ml 4 % paraformaldehyde. Brains were removed from the skull and post-fixed overnight at 4 °C. For protein extraction, mice were perfused with PBS alone before being snap-frozen in liquid nitrogen. For the OA ELISA, brains were snap-frozen without perfusion.
Histological tissue preparation and immunocytochemistry
Brains were dissected into forebrain, hindbrain and cerebellum before processing by paraffin embedding as described [
29]. Sections were analysed by immunohistochemistry using the phospho-specific antibody AT180 (Thermo Fisher) as described [
30].
Thioflavin-S staining
Slides were dewaxed and rehydrated, followed by a 1 min incubation in 70 % ethanol and a subsequent 1 min incubation in 80 % ethanol. Slides were incubated in filtered 1 % thioflavin-S (Sigma) in 80 % ethanol for 15 min at room temperature (RT), protected from light. Slides were washed in 80 % ethanol, 70 % ethanol and twice in milliQ water for 1 min each, before incubation in ice-cold high phosphate buffer (411 mM NaCl, 8.1 mM KCl, 30 mM NA2HPO4, 5.2 mM KH2PO4, pH7.2) for 15 min at RT, protected from light. Slides were washed twice for 1 min in milliQ water before mounting in 50 % glycerol sealed with nail varnish and stored at 4 °C in the dark.
Neuropathological quantification
Stained sections (3 per condition) were imaged using the slide scanner (Zeiss) at 20 × magnification and cropped into individual sections for analysis using the ImageJ software. Designated regions of interest (ROIs) were drawn for each brain region and phospho-tau immunoreactivity was quantified as a percentage area. The threshold for positive labelling was determined using the normal distribution of immunoreactivity from the control group and set at two standard deviations from the mean.
Protein fractionation and western blotting
Samples were kept on dry ice and weighed individually. Proteins were extracted sequentially based on insolubility as previously described [
31,
32]. Breifly, brain tissue was homogenised in 6 × volume of RAB buffer (0.1 M MES pH7.2, 1 mM EGTA, 0.5 mM MgSO4, 0. 75 M NaCl, plus phosphatase/protease inhibitors) in the TissueLyserLT (Qiagen) for 6 min at maximum speed. The homogenate was spun at 21,000 g for 90 min at 4 °C and the supernatant collectedand stored at −80 °C as the RAB soluble fraction. The remaining pellet was homogenised in the same volume of RIPA buffer to extract remaining insoluble proteins (10 × from Cell Signalling, plus phosphatase/protease inhibitors) and spun at 21,000 g for 90 min at 4 °C. The supernatant was collected and stored at −80 °C as the RIPA soluble fraction. Protein concentration in samples was determined by a BCA assay (Pierce) and 30 μg of protein was run per sample on a 10 % gel. Proteins were transferred onto Immobilon-FL membrane (Millipore) using the Trans-blot Turbo (BioRad) and blocked in Odyssey blocking buffer for 1 h at RT before incubation overnight in primary antibody (total tau (DAKO) 1:1000, AT180 1:500, AT8 1:500, AT270 1:1000, nitrated tau 1:100 (Thermofisher), ser422 1:1000 (GeneTex), ser262 1:1000 (ProSci), s235 1:1000 (Novus biologicals)). The next day membranes were washed in TBS-Tween (0.05 %) and incubated in secondary fluorescent antibodies (Licor, 1:20,000) for 30 min. Membranes were further washed in TBS-Tween before being visualised using the Odyssey Fc (Licor). Protein levels calculated using image studio software (Licor) and were normalised to actin.
Competitive okadaic acid ELISA
The OA ELISA was performed by adapting a protocol for the detection of OA in mussels and salt water samples, using a competitive ELISA kit from Bioo Scientific. Briefly, frozen tissue was homogenised without buffer in the TissueLyserLT for 6 min at maximum speed. 50 % methanol was added at 1 ml/0.25 g of tissue. Samples were vortexed for 5 min, then centrifuged at 4000 rpm for 10 min. The supernatant was transferred to a new tube and heated at 75 °C for 5 min before a further 4000 rpm, 10 min centrifugation. The supernatant was collected and diluted 1:1 with 1 × extraction buffer/methanol (provided by the kit). 50 μl of OA standard or 50 μl of sample were added in duplicate to the provided secondary antibody-coated 96 well-plate. Subsequently, 50 μl of OA-HRP conjugate followed by 50 μl anti-OA antibody were added to each well, immediately followed by mixing by pipetting up and down once. The plate was manually rocked for one min and the sample incubated at RT in the dark for 30 min. The plate was washed 3 × with 200 μl of the provided wash solution. After the last wash the plate was inverted and gently tapped dry on paper towels. 100 μl TMB substrate was added, the plate manually rocked for one min and then incubated at RT in the dark for 15 min. 100 μl of stop buffer were added to stop the enzyme reaction and the plate was read on the POLARstar Optima (BMG Labtech) at 450 nm wavelength.
Statistical analysis
Results are displayed as mean with error bars representing ± the standard error of the mean. Statistical analyses were conducted using the student’s unpaired t-tests or one way ANOVA test with appropriate post-hoc analyses for multiple comparisons. GraphPad Prism was used to perform statistical tests with statistical significance set to a P < 0.05.
Discussion
The spreading hypothesis of tau pathology in AD in its current form originated from observations regarding the chronological order of phospho-tau immunoreactivity using phosphorylation-dependent antibodies and NFT formation in
post mortem tissue, known as Braak staging [
2]. This hypothesis dictates that template-misfolding of hyperphosphorylated tau is able to seed further protein aggregation, which can be transmitted along anatomically connected neurons, leading to a systematic and chronological progression of disease pathology. Comprehensive in vivo studies have shown that injection both of whole brain lysates and isolated recombinant tau aggregates can accelerate tau pathology at both the injection site and in neighbouring, anatomically connected brain regions when done in pre-symptomatic tau transgenic mouse models [
14,
15,
17,
18]. Although this data convincingly supports the concept that pathological forms of tau can seed further aggregation of mature structural confirmation, it does not address how the initiation of primary aggregation events occur and how this triggers further dissemination throughout the neural system. Furthermore, the capacity for tau to seed and propagate aggregation in in vivo systems with endogenous tau levels remains unclear. Therefore, we used a unilateral injection of a very low volume and dose of the PP2A inhibitor OA to elicit a localised upregulation in tau phosphorylation, avoiding the introduction of preformed, mature tau seeds. In addition, wild-type mice were used to circumvent confounding interactions from over-expression of transgenic tau variants. Our results indicate that OA induces significant changes in tau phosphorylation, solubility and protein aggregation in non-exposed brain regions. Therefore, our approach could be a useful tool for further investigations into mechanisms underlying disease progression in AD.
The previous use of OA in mice involved the intracerebral injection of a tenfold higher dose of OA than used by us in a larger, 10 μl volume, in order to induce a behavioural deficit for therapeutic testing, with a focus on the oxidative stress profile rather than categorising tau phosphorylation [
25]. Previous studies in rats have shown that OA induces the cell death of hippocampal CA1 neurons and the activation of heat-shock proteins in both injected and also non-injected hippocampi [
23,
34]. By ELISA, we were able to detect OA in brain tissue (the first report of such a measurement in brain), which allowed us to determine the consequences of OA exposure, using DMSO injections as control. Our data provides evidence that in OA-injected mice, tau phosphorylation is increased not only in areas that were directly exposed to OA, but also in regions anterior to the injection site and in the contralateral non-injected hemisphere, where no significant OA levels were detected by ELISA. By targeting the lateral amygdala we were able to study effects on phosphorylation in brain areas significantly affected in AD as secondary sites for spreading, including the hippocampus, commonly the target in seeding experiments. Furthermore, confining OA exposure to one hemisphere allowed analysis of cross-hemispheric signalling. Interestingly, at 24 h post-injection, tau phosphorylation was significantly increased across both hemispheres, although the largest effect was observed in the injected hemisphere. Importantly, the ELISA confirmed no significant diffusion of the inhibitor between hemispheres over time, suggesting that the observed effects were due to indirect effects of OA on downstream signalling pathways to transmit tau phosphorylation. Furthermore, neighbouring brain regions displayed different responses in tau phosphorylation. For example, the perirhinal cortex had significantly increased phospho-tau immunoreactivity, whereas neighbouring striatal tissue within the same tissue showed no significant difference. Previously, the striatum has shown resistance to tau seeding when targeted directly with fibrillized recombinant tau in contrast to successful seeding in both the cortex and hippocampus [
17]. Therefore, our data supports the notion that certain brain areas are more vulnerable to developing tau pathology than others, which may depend on intrinsic cell properties irrespective of connectivity, highlighting cell-autonomous mechanisms. This aspect can be further investigated by injecting OA into different brain areas to address regional differences in the ability to translate a localised PP2A inhibition into widespread alterations to tau phosphorylation.
Drastic changes in tau phosphorylation were attenuated by 7 days post injection, consistent with a decrease in detectable OA in the injected quadrant, with tau phosphorylation predominantly persisting in the injected hemisphere at much lower levels than observed at 24 h. The histology correlated well with western blotting data demonstrating remaining tau was predominantly detectable in the insoluble fraction at a higher molecular weight. This shift in solubility and size is a correlative with the progression of tau into a pathological aggregate-prone state, which typically takes months to occur in transgenic models [
22] and as observed by histology, was detectable in both the injected and non-injected hemispheres. To address whether this change in solubility was reflected by protein aggregation we used the β-pleated sheet-sensitive dye thioflavin. This allowed us to detect thioflavin-positive structures at the posterior level of the injection in both the injected and non-injected hemisphere in both the cortex and amygdala at both 24 h and 7 days post-injection. Interestingly, thioflavin-reactivity was not observed in the hippocampus despite significantly elevated phospho-tau levels at 24 h, indicating that increased phospho-tau levels are not necessarily predictive of future aggregation (data not shown). In fact, transgenic human tau models present with significant levels of tau phosphorylation without developing end-stage thioflavin or Gallyas-positive lesions [
31,
32], unless the mice become very old [
35]. Qualitatively, these thioflavin-positive aggregates appear to increase in numbers over time, suggesting that while phosphorylation is dynamic and transient, once proteins adopt an aggregated state these forms tend to persist. Importantly, the presence of tau is crucial to the formation of thioflavin-positivity as evidenced by the lack of staining in OA-injected tau knock-out controls. Although the mechanism by which tauopathy crosses hemispheres in this model requires further investigation, our data presents evidence that effects of tau spreading can be observed over short-time frames, without the need for human transgene expression or injection of preformed seeds.
A limitation of our study is that the tau we observe is unlabelled, thus we are unable to track specific tau species as they propagate from neuron to neuron. This could be addressed in subsequent studies by using gene editing methods to generate a mouse strain that expresses tagged forms of endogenous tau with a restricted expression pattern, however, this would again introduce incomplete restriction of transgene expression as a possible confound. Currently, it is not known whether the phospho-tau species observed in our study are a result of a direct spreading of tau originating in the lateral amygdala, or a result of the communication of a stress signal translated into phosphorylation. A plausible explanation given the short time frame in which tau phosphorylation disperses throughout the brain is that the accumulation of pathological tau provides a metabolic insult within a neuron which causes downstream dysfunction through trans-synaptic signalling. However, one can assume that it is likely that both direct tau spreading and metabolic injury act concurrently in disease progression, both in our experimental paradigm and in a human disease setting. Approaches that utilise OA offer the potential to further explore this hypothesis.
Conclusions
In summary, the data presented here suggests that localised OA injections may be a viable avenue for better understanding the initiation and cell vulnerability underlying the spreading of tau. Our findings strongly suggest that tau phosphorylation can propagate rapidly across hemispheres following a confined unilateral insult, resulting in significant changes in tau solubility and tau-dependent protein aggregation. Moreover, these responses can be observed over short time frames in wild-type mice that only express endogenous tau. Future studies to better understand the underlying mechanisms are warranted, while application of OA to different brains regions may assist in clarifying cell-autonomous mechanisms necessary for the initiation of tau spreading.
Acknowledgements
This study was supported by the Estate of Dr Clem Jones AO, the State Government of Queensland, and by grants from the Australian Research Council (DP13300101932) and the National Health and Medical Research Council of Australia (APP1037746, APP1003150) to JG. We thank Tishila Palliyaguru and Linda Cumner for expert technical assistance, and Trish Hitchcock and her team for animal care.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
JG and SB designed the experiments, SB performed the experiments, and JG and SB analyzed the data and wrote the paper. Both authors read and approved the final manuscript.