Alzheimer’s disease pathological lesions activate the spleen tyrosine kinase

Tg PS1/APPsw, Tg APPsw, Tg Tau P301S and wild-type mice were generated and maintained in a C57BL/6 genetic background as previously described [28]. All mice were maintained under specific pathogen free conditions in ventilated racks in the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC) accredited vivarium of the Roskamp Institute. All experiments involving mice were reviewed and approved by the Institutional Animal Care and Use Committee of the Roskamp Institute before implementation and were conducted in compliance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals.

All mice were humanely euthanatized and their brains were collected and fixed in 4% paraformaldehyde (PFA) for 48 h. The method of euthanasia used follow the AVMA (American Veterinary Medical Association) guidelines for the euthanasia of animals. Briefly, mice were rendered unconscious through inhalation of 5% isoflurane in oxygen using a vaporizer and a gas chamber. While under anesthesia, after verifying the absence of reflexes, mice were euthanatized by exsanguination (blood was withdrawn from cardiac puncture).

Subsequently, the hemispheres were processed in a Sakura Tissue-Tek VIP (Leica Biosystems Inc., IL, USA) vacuum infiltration processor. Brains were then embedded in paraffin with the Sakura Tissue-Tek (Leica Biosystems Inc., IL, USA) and stored at 4 °C for 2 days for subsequent cutting with a Leica RM2235 microtome (Leica Biosystems Inc., IL, USA). All brains were cut at a thickness of 12 μm. Sagittal slices were mounted on glass slides and dried for 48 h at 37 °C for subsequent immunofluorescence staining and confocal imaging.

Paraffin sections were washed in two baths of histoclear (National Diagnostics, USA) and progressively rehydrated with ethanol gradients and phosphate buffered saline (PBS, Sigma Aldrich, MO, USA). Brain sections were subjected to antigen retrieval for 7 min in citric acid buffer (pH 6) at 100 °C. All sections were treated with 0.05% Sudan Black in 70% ethanol to quench autofluorescence. Sections were then blocked in PBS containing 10% donkey serum (Abcam, MA, USA) for 1 h. Sections were incubated in PBS containing 1% donkey serum and the respective panel of primary antibodies overnight at 4 °C. The following antibodies were used: CP13 (anti(α)-phospho-tau (pTau) S202, 1:200, Dr. Peter Davies’ Lab), MC1 (α-conformational tau, 1:200, Dr. Peter Davies’ Lab), TOC1 (1:200, Dr. Lester Binder’s Lab), PHF-1 (α-pTau S396/404, 1:200, Dr. Peter Davies’ Lab), 9G3 (α-pTau Y18, 1:200, MediMabs Inc., QC, Canada), DA9 (α-total-tau (tTau), 1:200, Dr. Peter Davies’ Lab), α-BACE1 (1:200 Cell Signaling, MA, USA), α-sAPPβ with Swedish mutation (1:100 Immuno-Biological Laboratories Co, Ltd., Japan), α-Iba1 (1:300, Abcam, MA, USA), α-GFAP (1:5000, Aves Labs, OR, USA), α-pSyk (Y525/526, 1:200, Cell Signaling, MA, USA). In addition to the α-pSyk (Y525/526, 1:200, Cell Signaling, MA, USA), we used the α-pSyk (Y525/526, 1:100, Abgent, CA, USA) and obtained similar results. After three washing steps in PBS for 5 min, sections were incubated in a solution containing PBS, 1% donkey serum and the respective panel of secondary antibodies for 1 h in the dark at room temperature in a humidified chamber. The following secondary antibodies were used: donkey α-rabbit, α-goat, α-mouse conjugated to Alexa 488, 568 and 647, respectively (1:500, Life technologies). After three washing steps in PBS for 5 min, sections were mounted in Fluoroshield with or without DAPI (Sigma Aldrich, MO, USA). All images were acquired using the confocal microscope LSM 800 (Carl Zeiss AG, Germany), the ZEN Blue 2.1 (Carl Zeiss AG, Germany) software and a 20× or 63× objective. The acquisition settings were kept the same for all genotypes within the same experiment.

For qualitative analysis of the pSyk burden in Tg PS1/APPsw and Tg APPsw mice compared to age-matched WT littermates (n = 6 for each genotype, equal amount of male and female), 116 ± 13.5 (avg. ± SEM) weeks of age were stained and analyzed as described above (Fig. 1).

For qualitative analysis of the pSyk burden in Tg Tau P301S mice compared to WT littermates, hippocampi and cortices of 16 male and female mice ranging from 8 to 56 weeks of age were stained and analyzed as described above.

For the quantitative analysis of the pSyk burden (Fig. 3), 140 randomly-selected microscopic fields of four non-consecutive brain slices (containing the hippocampus) from six animals per genotype (equal number of male and female) were acquired. The area covered with the pSyk immunopositive staining was quantified with Fiji [34] in microscopic fields containing Aβ plaques as well as in microscopic fields not containing Aβ deposits. The PS1/APPsw, APPsw and WT mice of the younger cohort were on average 45 ± 0.3 (avg. ± SEM) weeks old. The average age of the mice of the older cohort was 116 ± 13.5 weeks (±SEM). The pSyk burden of the transgenic mice was normalized to the level of pSyk burden quantified in wild-type littermates of the respective age-group. As a negative control, primary antibodies were omitted to determine background and autofluorescence (not shown).

For the quantitative analysis of the colocalization of pSyk and different tau epitopes (Fig. 8) between 400 and 570 cortical fields (50,000 μm² per field) from four male Tg Tau P301S animals (average age 47 ± 3.1 (SEM) weeks) were analyzed for each tau epitope. To quantify the percentage of the immunopositive neurons a total of 2546 microscopic fields and 21,800 neurons were counted using the Zen Blue 2.1 software (Carl Zeiss AG, Germany).

The fluorescence intensities (Figs. 9, 10, 11, 12 and 13) of 30 to 40 neurons immunopositive for pSyk, pTau or both (colocalized) were determined for each tau epitope (total of 90 neurons per epitope) using Zen Blue 2.1 (Carl Zeiss AG, Germany). The male Tg Tau P301S mice (n = 4) used for quantification were on average 47 ± 3.1 weeks old (avg. ± SEM).

In addition, the different immunostainings mentioned above were performed on paraffin-embedded tissue sections (10 μm, dorsolateral frontal cortex) from a 67-year-old, male patient with AD (Braak VI) and a 102-year-old, male non-demented control that were provided by Dr. Ann McKee (Boston University, MA, USA). Institutional review board approval for brain donation was obtained through the Boston University Alzheimer’s Disease Center (BUADC, Boston, MA, USA).

SH-SY5Y cells were purchased from American Type Culture Collection (VA, USA). SH-SY5Y cells were grown in DMEM/F12 medium (Thermo Fisher Scientific, MA, USA) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, MA, USA), GlutaMAX and 1% penicillin/streptomycin/fungizone.

A human cDNA ORF Clone of the human SYK gene (NM_003177, transcript variant 1) was purchased from OriGene Technologies (MD, USA). The cDNA fragment encoding human SYK was amplified by PCR using PfuUltra II Fusion HS DNA polymerase (Agilent Genomics, CA, USA) and subcloned into the p3xFLAG-Myc-CMV™-26 Expression Vector (Sigma-Aldrich, MO, USA) to generate the pCMV-SYK-Flag plasmid. The entire reading frame of the plasmid was confirmed by DNA sequencing. SH-SY5Y cells were maintained in advanced DMEM/F-12 medium supplemented with 10% fetal bovine serum, 1% GlutaMAX, 1% penicillin/streptomycin (Thermo Fisher Scientific, MA, USA) and incubated in a humidified 5% CO₂ atmosphere at 37 °C. For stable transfection, SH-SY5Y cells were grown in 6-wells cell culture plates until reaching 70-80% confluence and transfected with 3 μg of empty pCMV vector (control cells) or pCMV-SYK-Flag plasmids per well using lipofectamine 2000 (Thermo Fisher Scientific, MA, USA). After 48 h, the medium surrounding transfected cells was replaced with fresh medium containing 0.2 mg/ml of G418 for selection. After 14 days of selection, G418 resistant cells were trypsinized and expanded. The expression efficiency of SYK was analyzed by Western blot using antibodies against SYK (4D10 Syk antibody, Santa Cruz Biotechnology, TX, USA) and the Flag tag (Sigma-Aldrich, MO, USA).

SH-SY5Y cells were cultured in 24-well-plates for 24 h and subsequently lysed with mammalian protein extraction reagent (MPER, Thermo Fisher Scientific, MA, USA) containing Halt protease & phosphatase single use inhibitor/EDTA (Thermo Fisher Scientific, MA, USA) and 1 mM PMSF. Proteins of cell lysates were separated by 10% tris-glycine-SDS-PAGE using 1 mm Criterion TGX gels (Bio-Rad Laboratories, CA, USA) and electro-transferred onto 0.2 μm PVDF membranes (Bio-Rad Laboratories, CA, USA). Membranes were blocked in TBS containing 5% non-fat dried milk for 1 h and were hybridized with the primary antibody (αSyk (4D10, 1:1000, Santa Cruz, TX, USA), αpTau S396/404 (PHF-1, 1:1000, Dr. Peter Davies’ Lab), αtTau (DA9, 1:1000, Dr. Peter Davies’ Lab), αpTau Y18 (9G3, 1:1000, MediMabs Inc., QC, Canada,) overnight at 4 °C. Subsequently, the membranes were incubated for 1 h in HRP-conjugated αmouse secondary antibody (1:1000, Cell Signaling, MA, USA). Western blots were visualized using chemiluminescence (Super Signal West Femto Maxium Sensitity Substrate, Thermo Fisher Scientific, MA, USA). Signals were quantified using ChemiDoc XRS (Bio-Rad Laboratories, CA, USA) and densitometric analyses were performed using Quantity One (Bio-Rad Laboratories, CA, USA) image analysis software.

To investigate whether pathological Syk activation occurs in the brain of AD mouse models, we analyzed the brains of 116-week-old wild-type, Tg APPsw and Tg PS1/APPsw mice using high-resolution confocal microscopy and immunofluorescence. All transgenic mice (Fig. 1b-e) exhibit an increased Iba-1 and GFAP reactivity compared to wild-type littermates (Fig. 1a). Moreover,wild-type some of the activated amoeboid microglia that are observed in transgenic mice are also strongly positive for pSyk (Fig. 1b-d). By contrast, we did not detect any pSyk immunoreactivity in astrocytes (Fig. 1). In addition, we observed that pSyk immunoreactivity is upregulated near Aβ plaques but neither colocalizes with microglia nor astrocytes suggesting that it could be of neuronal origin. (Fig. 1e). We further investigated the cellular origin of these pSyk accumulations by immunofluorescence staining and confocal microscopy (Fig. 2).

To further characterize the cellular origin of pSyk accumulations near Aβ plaques, we tested different markers of dystrophic neurites (BACE-1 and sAPPβ) [31] and found a strong colocalization between pSyk and sAPPβ (Fig. 2a) associated with Aβ deposits. The sAPPβ staining clearly reveals dystrophic swellings of neurites (Fig. 2a) which are a known hallmark of AD. Most of the dystrophic neurites are positive for pSyk (Fig. 2a). Additionally, we found a strong colocalization between sAPPβ and BACE-1 (Fig. 2b) which are often used as markers of dystrophic neurites. Both sAPPβ and BACE-1 exhibit circular accumulations near Aβ plaques (Fig. 2b), highly reminiscent of the pattern observed for activated Syk.

In conclusion, activated Syk is not only found in microglia but also in neurons near Aβ deposits, particularly in dystrophic neurites of Tg APPsw and Tg PS1/APPsw mice supporting a possible role of Syk activation in the formation of dystrophic neurites. Dystrophic neurites are characterized by an accumulation of BACE-1 and sAPPβ [31] and our previous work [28] has shown that Syk regulates BACE-1 expression and sAPPβ levels suggesting that Syk upregulation in dystrophic neurites could contribute to the accumulation of BACE-1 and sAPPβ.

We also quantified the pSyk burden observed in the cortex of Tg APPsw, Tg PS1/APPsw and wild-type (WT) littermates (Figs. 1 and 2). Two different age-groups were investigated: younger animals, 45 weeks of age and older animals 116 weeks of age in average. In 45-week-old Tg APPsw mice, we did not observe significant β-amyloidosis (only three Aβ plaques were found in the cohort of mice analyzed) (data not shown) showing that, at that age, the Aβ pathology is almost inexistent in these mice. We differentiated between microscopic fields containing Aβ deposits and microscopic fields not containing Aβ deposits for the quantification of the pSyk burden in Tg PS1/APPsw and older Tg APPsw mice. 45-week-old Tg APPsw and Tg PS1/APPsw mice do not show any significant difference in pSyk burden in fields without Aβ deposits compared to WT mice. The pSyk burden of 45-week-old Tg APPsw mice is identical to that of the WT mice (100 ± 6.76% compared to 80.85 ± 11.77%; Fig. 3a). The pSyk burden in fields not containing Aβ plaques in Tg PS1/APPsw mice is not statistically significantly elevated (153.48 ± 18.47%), compared to the WT littermates. As expected, 45-week-old Tg PS1/APPsw mice exhibited a significantly higher pSyk burden in fields containing Aβ plaques (410.19 ± 46.46%) compared to WT and Tg APPsw mice.

The analysis of the pSyk burden in the cortex of older animals (average age: 116 weeks) revealed large differences between genotypes. The pSyk burden of Tg APPsw (216.32 ± 45.23%) mice in microscopic fields without plaques is not significantly increased compared to WT mice (100 ± 7.78%) (Fig. 3b). In contrast, microscopic fields of older Tg APPsw mice containing Aβ deposits exhibit a strong increase in pSyk burden (799.95 ± 130.19%) compared to age-matched WT mice. Tg PS1/APPsw mice also exhibit a statistically significant increase in pSyk burden in microscopic fields that do not contain Aβ deposits (458.1 ± 109.68) compared to age-matched WT controls. In addition, a much greater pSyk burden is found in Tg PS1/APPsw in microscopic fields containing Aβ deposits. In these fields, the pSyk burden is increased by 1157.31 ± 129.68% compared to WT littermates (Fig. 3b).

In conclusion, our data show that the pSyk burden is highly associated with Aβ plaques and increases with age in Tg PS1/APPsw and Tg APPsw mice whereas no activation of Syk is observed in the brain of WT littermates. The upregulation of Syk activation observed in the brains of Tg APPsw and Tg PS1/APPsw is mainly attributable to pSyk accumulations in dystrophic neurites that are associated with Aβ plaques and increase with age and Aβ burden.

Having shown that Aβ-overexpressing mouse models of AD exhibit an increased Syk activation in microglia and dystrophic neurites, we investigated whether Syk activation also occurs in Tg Tau P301S mice (a pure model of tauopathy) using immunofluorescence and confocal microscopy. Hippocampal neurons of Tg Tau P301S mice exhibit a high level of tau hyperphosphorylation (Fig. 4b) as well as an accumulation of pathogenic tau conformers (MC1, not shown) compared to WT littermates (Fig. 4a). Most importantly, pathological tau species clearly colocalize with pSyk (Y525/526) in hippocampal neurons (Fig. 4b). The pSyk burden is particularly prominent in hippocampal neurons of Tg Tau P301S mice (Fig. 4b) whereas WT littermates do not exhibit any pSyk immunoreactivity in the hippocampus (Fig. 4a).

Cortical neurons of Tg Tau P301S mice also exhibit an increased level of tau hyperphosphorylation (Fig. 5b) compared to wild-type littermates (Fig. 5a). We observed a colocalization between pSyk and pTau (S202) immunoreactivities in cortical neurons. Interestingly, we also observed neurons that are singly immunopositive for tau or for pSyk. We addressed this observation by performing additional analyses (Figs. 8, 9, 10, 11, 12, 13, 14 and 15).

We also analyzed the temporal changes of pSyk and tau levels in hippocampi and cortices of Tg Tau P301S mice between the age of 8 and 56 weeks (Figs. 6 and 7). WT mice do not exhibit any detectable tau phosphorylation (Fig. 6f) or tau oligomerization at any age (not shown). We then focused on the dentate gyrus of the hippocampus and found an age-dependent increase of tau phosphorylation (Fig. 6, S202, left panels) in Tg Tau P301S mice. Tau phosphorylation at S202 in the dentate gyrus was already detectable in 8-week-old Tg Tau P301S mice, however, pSyk immunoreactivity was not observed. Neurons immunopositive for pSyk (Y525/526) and pTau (S202) or tau conformers (MC1, not shown) were observed in the dentate gyrus of 42-week-old Tg Tau P301S mice (Fig. 6d). The neuronal pSyk burden also increases with age in Tg Tau P301S mice and is mainly restricted to the neuronal cell body (Fig. 6). Interestingly, microglia and neurites did not exhibit activated Syk in the hippocampus of Tau P301S mice (Fig. 6). Abnormal Syk activation seems to follow tau hyperphosphorylation (S202) in the hippocampus of Tg P301S mice (Fig. 6), as well as the formation of MC1-tau pathological conformers (data not shown here but MC1 and pSyk colocalization were quantified below).

Cortical neurons of Tg Tau P301S mice also show an increase in tau hyperphosphorylation and pSyk with age (Fig. 7). Interestingly, the onset of abnormal Syk activation occurs earlier (16 weeks of age) in the cortex than in the hippocampus (Fig. 7b compared to Fig. 6d). In conclusion, both pSyk and tau pathology in Tg Tau P301S mice increase with age but the progression is different in the hippocampus and the cortex. Many cortical neurons exhibit a colocalization of pSyk and pTau (S202) (Fig. 7b-c, e) but as mentioned earlier, there are also neurons that are singly immunopositive for pSyk or pTau.

We further quantified the number of neurons that are singly pSyk immunopositive, singly immunopositive for tau pathogenic species and neurons immunopositive for both pSyk and tau pathogenic species in the cortex of 47-week-old Tg Tau P301S mice (Fig. 8). We calculated the percentages of neurons singly immunopositive for either pSyk, pathogenic tau species or neurons immunopositive for both. The sum of all cortical neurons counted was considered 100% including neurons positive for pSyk and the respective tau epitope and neurons immunopositive for both. For all the tau epitopes tested, we found that only a small fraction of the neurons is singly immunopositive for pSyk (9.7 ± 4.4% (pTau, Y18); 2.5 ± 1.2% (pTau, S202); 9.2 ± 1.6% (MC1 pathogenic tau conformers); 9.6 ± 6.3% (pTau, S396/404); 4.8 ± 2.0% (TOC1 (tau oligomers)). Interestingly, a larger percentage of neurons is immunoreactive for both pSyk and tau pathogenic species (44.7 ± 8.6% (MC1); 39.7 ± 12.4% pTau Y18; 22.5 ± 18.6% (PHF-1, pTau S396/404); 12.4 ± 8.1% (TOC1, tau oligomers) but only 5.7 ± 2.2% for pTau (S202)). The neurons singly immunopositive for tau complement these observations with relative values of 46.1 ± 8.2% (MC1), 50.6 ± 16.3% (Y18), 67.9 ± 24.9% (S396/404), 82.8 ± 10.1% (TOC1), and 91.8 ± 3.2% (S202) (Fig. 8). The differences in relative colocalization between pSyk and specific tau pathologic species suggest that specific pathogenic forms of tau may have a different impact on Syk activation or either that Syk activation may contribute to the formation of specific tau pathogenic species (Fig. 8). We therefore subsequently measured the fluorescent intensities of pSyk and of the different tau epitopes to determine whether the level of Syk activation correlates with the amount of specific tau pathogenic species. In general, we found that neurons that exhibit a high level of pSyk immunoreactivity also demonstrate a higher level of tau pathogenic species whereas neurons that are weakly immunopositive for pSyk show less tau pathology (Figs. 9, 10, 11, 12 and 13). In addition, neurons that are singly immunopositive for tau pathogenic species (including hyperphosphorylated tau and misfolded tau) show also less intense tau pathologies, as measured by fluorescent intensities, than neurons that are displaying both pSyk and tau pathology, further supporting a role of Syk in the formation of tau pathogenic species. For instance, the level of pathogenic tau conformers (MC1) is significantly increased in neurons that are also strongly immunopositive for pSyk compared to neurons that are singly immunopositive for MC1 (Fig. 9d, p < 0.05). Interestingly, the level of pSyk is also significantly increased in neurons that display an accumulation of MC1 pathogenic conformers compared to neurons that are singly immunopositive for pSyk (Fig. 9e, p < 0.01). These data suggest that pathogenic tau conformers and Syk activation may promote each other. We found that tau phosphorylation at Y18 is significantly increased in neurons that are also immunopositive for pSyk (Fig. 10d, p < 0.05) which is consistent with previous data showing that in vitro Syk can phosphorylates tau at Y18. We have previously shown that Syk positively regulates GSK-3β activity in vitro. It is therefore consistent with our observation that the GSK-3β-dependent phospho-tau epitope (S396/404, PHF-1) is also increased in neurons that display Syk activation (Fig. 11d, p < 0.0001). The pSyk level, however, is not statistically significantly increased in neurons that are immunopositive for both PHF-1 and pSyk compared to neurons that are singly immunopositive for pSyk suggesting that PHF-1 phosphorylated tau species do not induce Syk activation (Fig. 11e). Similar observations were obtained for tau oligomers (Fig. 12, TOC1) and tau species phosphorylated at S202 (Fig. 13, CP13). Altogether, these data suggest that only certain pathogenic forms of tau (MC1, Y18) promote Syk activation, whereas Syk activation appears to directly induce tau phosphorylation at Y18 and to indirectly regulate tau phosphorylation at multiple epitopes (S396/404, S202) as well as tau misfolding (MC1, TOC1).

To further investigate the impact of Syk on tau, we generated human neuronal-like (SH-SY5Y) cells overexpressing Syk (Syk-OX). Syk-OX SH-SY5Y cells show an approximate 17-fold increase in Syk expression compared to control SH-SY5Y cells transfected with the empty vector (Fig. 14b, p < 0.0001). Interestingly, Syk upregulation in SH-SY5Y cells leads to a significant increase (1.7-fold) in phosphorylated tau at Y18 (Fig. 14c, p < 0.01) and at S396/404 (Fig. 14d, 3-fold, p < 0.0001) compared to control cells. Total tau levels are also significantly increased following Syk overexpression (Fig. 14e, 4.2-fold, p < 0.0001). We analyzed the possible impact of Syk overexpression on Tau mRNA levels by quantitative RT-PCR and found that Syk overexpression does not affect Tau transcription (data not shown) suggesting that Syk may regulate tau degradation or tau protein translation. In summary, these results show that the accumulation of tau pathogenic species can trigger Syk activation, as shown in Tg Tau P301S mice (Figs. 8, 9, 10, 11, 12 and 13), whereas Syk itself appears to regulate total tau levels and tau phosphorylation at multiple epitopes (Fig. 14) therefore influencing the development of the tau pathology.

We also performed different immunostainings against Aβ, pSyk, GFAP, Iba-1, tau pathogenic conformers (MC1) and phosphorylated tau at Y18 using brain sections from human AD and non-demented controls. We found an increase in Syk activation in DNs surrounding Aβ deposits as well as in neurons displaying an accumulation of phosphorylated Tau at Y18 and elevated levels of MC1 pathogenic tau conformers in AD brain sections whereas only weak immunoreactivity for pSyk was observed in brain sections from a non-demented control (Figs. 15, 16 and 17). As observed in the AD mouse models, astrocytes did not exhibit Syk activation in neither the AD brain section nor the control. Only a subset of microglial cells exhibited a weak pSyk signal. Most of the detected pSyk signal was of neuronal origin and either localized in somata or DNs. These data complement our observation in AD mouse models and reveal an association between Syk activation and typical AD pathological lesions in the human brain. Further studies will be required using a larger sample of AD pathological specimen to further clarify the role of Syk activation in AD brains.