Background
Tauopathies consist of intracellular accumulation of the microtubule-associated protein tau in the somatodendritic compartment associated with hyperphosphorylation and aggregation of the protein. Tau dysfunction can lead to neurodegeneration, motor dysfunction, and behavioral deficits in animal models that express mutated forms of the protein [
1‐
4]. One of the most common tauopathies includes Alzheimer's disease (AD). Consequently, numerous studies targeting different components of the disease have been initiated to reduce tau pathology as well as amyloid-β. These include inhibition of kinases which phosphorylate tau, such as glycogen synthase kinase-3β (GSK3β) [
5‐
8], manipulating heat shock proteins [
9‐
11], immunotherapy which targets the tau peptide and subsequently reduces p-tau levels, [
12,
13], and modifying p-tau by manipulating the immune response [
14]. Some reports indicate that pathological tau induces inflammation [
15] and that inflammation modifies tau [
16,
17].
Inflammation arguably plays a significant role in the progression of AD pathology. The microglial activation state contributes to many of the ongoing debates, and it is believed that microglia can cause both beneficial and detrimental effects, depending on the microenvironment and cytokines involved. Recently, more attention has been paid to the functional status of microglia rather than generalized activation by generic markers. As more studies emerge identifying selective markers that represent different phenotypic activation states of microglia, an association of their role during disease pathology is being revealed. In the classical or M1 state, pro-inflammatory cytokines produce tissue damage and pathogen destruction, whereas the alternative activation state (M2) dampens this response and directs tissue repair and healing responses [
18]. Some reports suggest that in chronic neurodegenerative diseases like AD, a hybrid activation state exists including markers of both M1 and M2 phenotypes [
19]. Other reports argue that beginning stages of AD pathology in animal models of amyloid-β deposition resemble an M2 state that switches to a more classical response with age [
20]. It is possible the phenotypic state of microglia in response to amyloid-β deposition influences certain aspects of tau pathology as well. Thus, it is important to identify the components of inflammation that promote vs. reduce tau pathology in order to design better therapeutic strategies which target the immune response.
In previous studies [
21‐
25], intracranial LPS, which induces both M1 and M2 markers, activates microglia and reduces Aβ pathology in APP transgenic models of amyloid deposition. This requires microglial activation and can be suppressed by dexamethasone administered systemically. Importantly, there is no indication for systemic inflammation in AD patients [
26]. Herein, we similarly provoked central microglial activation by LPS to evaluate phospho-tau species and pathology in the rTg4510 mice. This model develops tangle pathology in the higher forebrain cortical layers and hippocampus coupled with cognitive deficits and neuronal loss [
1,
4,
27]. To our knowledge, this is the first report showing that activation of inflammation in the brain exacerbates tau phosphorylation.
Methods
Mouse breeding, tissue preparations, and animal treatments
The rTg4510 mice, lines carrying the parental tau mutations and the tetracycline-controlled transactivator protein (tTA) were used [
3]. For age-related microglia activation, brains were harvested from 1, 5, or 9 month old rTg4510 mice and their non-transgenic littermates. For lipopolysaccharide (LPS) studies, male and female mice were aged 4.5 months and a volume of 2 μl of LPS (5 μg/μl; Salmonella abortus equii, Sigma-Aldrich, St. Louis MO) in was unilaterally injected into the hippocampus and anterior cortex (frontal cortex area3) of rTg4510 and non-transgenic littermates. Stereotaxic coordinates from bregma were +1.7 mm anteroposterior, -2.2 mm lateral and -2.5 mm vertical for frontal cortex, and, -2.7 mm anteroposterior, -2.7 mm lateral and -3.0 mm vertical for hippocampus. The solution was dispensed at a constant rate of 0.5 μl/min. Seven days post injection; mice were weighed and overdosed with 100 mg/kg of pentobarbital. Mice were then perfused intracardially with 25 ml of 0.9% saline. The brain was removed, and immersion fixed in 4% paraformaldehyde in 100 mM PO
4 buffer (pH 7.4) for 24 hours. The tissue was cryoprotected in a series of 10%, 20% and 30% sucrose solutions. Horizontal sections were cut at 25 μm using a sliding microtome and stored at 4°C in Dulbecco's phosphate buffered saline containing 100 mM sodium azide for immunohistochemistry.
Immunohistochemistry and silver stain
Immunohistochemistry was performed on free floating sections as described in detail previously [
28]. Sections were incubated with primary antibodies rat anti-mouse CD45 (1:3000) (Serotec, Raleigh, NC), rat anti-major histocompatibility complex -II (1:5000) (MHCII; BD Pharmigen), rabbit anti-mouse chitinase 3-like-3 (1:3000) (YM1; StemCell Technologies, Vancouver, Canada), chicken anti-arginase 1 (1:50,000)(generous gift from Dr. S.M. Morris)[
29], rabbit anti-human phospho-tau ser199/202 (1:65,000)(Anaspec, Fremont, CA), rabbit anti-human phospho-tau ser396 (1:3000) (Anaspec, Fremont, CA), or rabbit anti- human full length-tau (1:3000) (H-150, sc-5587, Santa Cruz Biotechnology, Santa Cruz, CA), AT8 (1:5000) (Thermo Scientific, Rockford, IL overnight at 4°C, then incubated in the appropriate biotinylated secondary antibody (VectorLabs, Burlingame, CA) for 2 h followed by a 1 hr incubation in ABC (Vector Labs, Birlingame, CA). Color development was performed using 0.05% 3, 3'-diaminobenzidine (DAB; Sigma, St. Louis, MO) enhanced with 0.5% nickelous ammonium sulfate (J. T. Baker Chemical Company, Phillipsburg, NJ). For silver staining, a series of sections stained using Gallyas silver stain method [
2]. Briefly, sections were fixed in 4% paraformaldehyde in 100 mM PO
4 buffer (pH 7.4) for 24 hours, horizontally sectioned at 25 μm thickness, and stored at 4°C in Dulbecco's phosphate buffered saline containing 100 mM sodium azide. Free floating sections were mounted on slides and processed together using Gallyas silver stain method with the omission of a counter stain for quantitative analysis. It should be noted for time courses and LPS studies that all tissue sections for each immunohistochemical stain and the Gallyas silver stain that was analyzed together were processed together at the same time under the same conditions.
Immunofluorescence
Immunohistochemistry was performed on free floating sections as previously describe above with slight modifications to primary antibody concentrations. Sections were incubated with primary antibodies rat anti-mouse CD45 (1:1000), rabbit anti-human phospho-tau ser199/202 (1:10,000), rabbit anti-mouse chitinase 3-like-3 (YM1) (1:1000), rabbit anti-human phospho-tau ser396 (1:1000), rabbit anti- human full length-tau (1:2000) (H-150), biotinylated AT8 (1:5000) (Thermo Scientific, Rockford, IL) overnight at 4°C, washed and incubated with the appropriate secondary Alexa Fluor antibodies for 2 h (Invitrogen); goat anti-rabbit Alexa 488, goat anti-rat Alexa 488, Streptavidin Alexa 594, donkey anti-chicken Alexa 488, goat anti-rabbit Alexa 594. Sections were mounted on slides with Vectashield, (Vector Labs, Burlingame, CA).
Image analysis quantification and statistics
Immunohistochemical staining was quantified with Image Pro Plus (Media Cybernetics, Silver Spring, MD) image software. Positively labeled microglia or tau positive neurons were segmented using RGB intensity. Each brain section was imaged at 100× magnification in the anterior cortex centered on the injection site (frontal cortex, area 3), the CA1 or CA3 region of the hippocampus, and entorhinal cortex (caudomedial). Data were obtained as a percent area of the image field that was positively stained by immunochemical or histochemical reaction product. Some sections were digitized on the Zeiss Mirax slide scanner. All values (8 sections) obtained from a single mouse were then averaged to represent a single value for each brain region. Statistical analysis was performed using 2-way ANOVA (Age and Treatment), followed by Fisher's LSD post hoc means comparison test with p values of <0.05 considered significant using Stat View software version 5.0 (SAS Institute Inc, Cary NC). Graphs were generated using GraphPad Prism 4.0 (La Jolla, CA).
Discussion
In this study, we show that the phosphorylated tau species previously characterized in the rTg4510 mice [
30] are associated with age-related microglial activation as measured by CD45 Further activation of microglia by LPS enhances tau phosphorylation. Prior work [
30], demonstrated that young mice possess the ability to clear soluble phospho-tau species, showing reductions in these markers between 1 and 3 months. However, by 5.5 months, insoluble tau aggregates appear in parallel with accumulation of a 64 kD soluble tau species. Thus, microglial activation begins at this age when soluble and insoluble tau species are present. When microglial activation is provoked by LPS challenge at this point, there are clear increases in the phosphorylation of tau. Previous studies showed that LPS-induced microglial activation in APP mice clears amyloid-β pathology in the CNS as early as 3 days following intracranial injection [
21‐
24]. Using this same paradigm, LPS-induced microglial activation in rTg4510 mice exacerbates pre-tangle pathology as visualized by phospho-tau staining. This highlights the need to include mouse models of tau pathology as well as models of amyloid pathology when assessing the impact of potential treatments for translation in to clinical trials in Alzheimer cases.
Another previous study using a 3xTg-AD model (harboring APP
K670N; M671L , PS1
M146V , and Tau
p301L mutations) showed no changes in APP processing after 6 weeks of peripheral administration of LPS. However, phosphorylation of tau at specific sites (AT-180, p231/235; AT8, p202/205; but not PHF-1, p396/404) was increased within the hippocampus, in a cyclin kinase 5-dependent mechanism [
16]. Another pro-inflammatory stimulus, interleukin-1β, also resulted in microglial activation and tau phosphorylation in cortical neurons [
17]. Herein, we show that acute activation of microglia by LPS increased phospho-tau staining within one week, not only in the hippocampus and anterior cortex, but also in other tau-laden areas that were not injected including entorhinal cortex. Although the level of microglial activation also increased in the entorhinal cortex to a lesser extent than that of the hippocampus and anterior cortex, the increased phospho-tau species observed distal to the injection site is conceivable from previous findings of systemic inflammation and CNS effects on phospho-tau [
16] and supports the potential role for diffusible ligands and cytokines and their impact on tau pathology. Although these data suggest that acute inflammatory conditions may accelerate the course of neurodegenerative tauopathies or AD, other models of low level chronic neuroinflammation should be explored in a similar context [
31].
With normal aging up to 9 months, CD45 positive microglia increased in parallel with tau pathology, yet the alternative activation marker YM1 was not detected at the protein level by immunohistochemistry. However, upon LPS challenge, YM1 was elevated to a significantly greater level in rTg4510 mice with pre-existing tau than in nontransgenic mice. Thus, not only does microglia activation appear to influence tau pathology but tau pathology appears to impact the phenotypic responses of microglia as well.
It has yet to be determined whether this exaggerated YM1 activation occurred in response to insoluble vs. soluble tau species or even other specific tau forms such as truncated tau. A recent report showed that human misfolded, truncated tau protein promoted the up-regulation of immune molecules in microglia/macrophages and caused the influx of antigen-presenting cells from blood into the CNS of transgenic rats [
15]. It is also unknown whether YM1 influences the pathological state of tau, however recent reports show YM1 protein expression localizes around the amyloid plaques of APP/PS1 transgenic mice, and AD brains contain increased YM1 mRNA compared to normal, aged -matched control brains [
19,
20]. This implies that amyloid pathology can increase YM1 expression. YM1, whose function is poorly understood, exists as a secretory protein transiently expressed in microglia/macrophages during hematopoiesis [
32], during parasitic infection or after interleukin-4/interleukin-13 cytokine stimulation [
33]. YM1 shows specific binding affinity to glucosamine, galactosamine, and heparin sulfate, which has been hypothesized as a mechanism for shielding or maintaining macrophage integrity during parasitic infection [
32,
34]. Interestingly, heparin sulfate and sulfated glycosaminoglycans prevent tau from binding to microtubules, promote microtubule disassembly, and stimulate tau phosphorylation by several kinases [
35]. Although, LPS induced inflammation and tau pathology seems to influence the expression of YM1, it is unclear if up-regulation of YM1 by cytokines or amyloid-β deposition directly impacts the tau pathology.
Gene delivery of the pro-inflammatory cytokine tumor necrosis factor alpha (TNF-α) into the CNS of 3xTg-AD mice resulted in accumulation of both Aβ42 and phospho-tau species [
36]. Conversely, chronic ibuprofen treatment in 3xTg-AD mice reduced oligomeric amyloid-β, hyperphosphorylated tau, and improved memory deficits [
37]. However, overexpression of the pro-inflammatory and M1-stimulating cytokine, interferon gamma, using AAV resulted in differential effects on amyloid-β and phospho-tau[
14]. The authors observed increased levels of amyloid-β but reduced levels of phospho-tau, contradicting results with INF-γ in 3xTg-AD mice [
14]. Given the evidence that the amyloid deposition drives tau pathology in this 3xTg-AD model [
12], it is unclear whether direct effects on tau or indirect effects on amyloid are responsible for changes in tau pathology.
Transgene regulated over-expression of the M1-stimulating cytokine, interleukin-1 in APP mice caused reductions in amyloid pathology [
38]. Similarly, AAV-mediated over-expression of interleukin-6 in TgCRND8 and Tg2576 mice elicited massive gliosis and reductions in amyloid-β pathology [
39]. The microglial phenotype in these mice included increased YM1, but not other M2 activation markers such as arginase 1. In our study, we observed a tau transgene facilitation of YM1 induction following LPS in both hemispheres, but this effect was not evident for arginase 1activation, another putative M2 activation marker. This raises questions regarding how particular M2 activation state markers are regulated by proinflammatory stimuli and why RNA and protein levels of YM1 are increased in AD patients and animal models of amyloid deposition, which are typically considered to be associated with a proinflammatory (M1) cytokine environment [
19,
20]. These observations are important in considering the ultimate goals for therapeutic tuning of the microglial phenotype in order to reduce amyloid and/or tau pathology. Some dichotomous effects in the different transgenic models and therapeutic treatments make interpretations and potential translation to AD challenging. These results suggest a more complex set of microglia phenotypes than the dipolar M1/M2 characterization, and suggest that the clearance of amyloid pathology and tau pathology may be mediated by distinct activation subtypes.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
DCL led the study and involved in all aspects from surgeries to analysis of data, and manuscript preparation. JR participated in surgeries, immunohistochemical staining, and data acquisition. MS and PR participated in tissue preparation, immunohistochemical/fluorescent staining and data acquisition, generation of figures. CK, AJ, and LB performed Gallyas Silver staining and maintained animal husbandry, breeding, and genotyping of mice and data collection. CD, MG, DM were involved in overall conception, secured funding for the project, experimental design, manuscript preparation and analysis. All authors have read and approved the final manuscript.