Opposite effects of two estrogen receptors on tau phosphorylation through disparate effects on the miR-218/PTPA pathway

The two estrogen receptors (ERs), ERα and ERβ, mediate the diverse biological functions of estradiol. Opposite effects of ERα and ERβ have been found in estrogen-induced cancer cell proliferation and differentiation as well as in memory-related tasks. However, whether these opposite effects are implicated in the pathogenesis of Alzheimer’s disease (AD) remains unclear. Here, we find that ERα and ERβ play contrasting roles in regulating tau phosphorylation, which is a pathological hallmark of AD. ERα increases the expression of miR-218 to suppress the protein levels of its specific target, protein tyrosine phosphatase α (PTPα). The downregulation of PTPα results in the abnormal tyrosine hyperphosphorylation of glycogen synthase kinase-3β (resulting in activation) and protein phosphatase 2A (resulting in inactivation), the major tau kinase and phosphatase. Suppressing the increased expression of miR-218 inhibits the ERα-induced tau hyperphosphorylation as well as the PTPα decline. In contrast, ERβ inhibits tau phosphorylation by limiting miR-218 levels and restoring the miR-218 levels antagonized the attenuation of tau phosphorylation by ERβ. These data reveal for the first time opposing roles for ERα and ERβ in AD pathogenesis and suggest potential therapeutic targets for AD.


Introduction
Alzheimer's disease (AD), the most common form of dementia, was first reported in 1906 by Alois Alzheimer. Extensive research has established the two most prominent pathological hallmarks in AD brains: senile plaques and neurofibrillary tangles (NFTs). The degree of cognitive impairment has been shown to significantly correlate with the presence of NFTs (Braskie et al., 2010).
Hyperphosphorylated tau, which forms paired helical filaments, is the major component of NFTs (Johnson & Jenkins, 1999). Although the precise role of tau phosphorylation in the toxicity remains unclear, the abnormalities caused by hyperphosphorylated tau have been well studied. For example, abnormal tau hyperphosphorylation converts normal tau from a microtubule assembly-promoting to a microtubuledisrupting protein (Alonso et al., 1994). In AD brains, tau is hyperphosphorylated about three times more than that in normal brains, and it promotes misfolding of normal tau and coaggregates with it into filaments (Alonso et al., 1996). The levels of tau phosphorylation are positively correlated with cognitive deficits in multiple animal models and patients with AD (Mitchell et al., 2002;Zhu et al., 2004;Stancu et al., 2014). Therefore, the development of tau-based therapeutic drugs for AD-related tauopathies will require the elucidation of the underlying mechanisms of how the abnormal phosphorylation is regulated.
Previous studies have shown that the incidence of AD in women is significantly higher than that in men , and this difference has been attributed to the loss of estrogen and a variety of related mechanisms at the molecular, cellular, and hormonal levels. Subsequent studies have elucidated the neuroprotective roles of estrogen against AD-related pathology and have proposed that the beneficial effects of estrogen on AD are directly linked to its ability to reduce amyloid-b peptide and tau aggregates (Vest & Pike, 2013). There are two known estrogen receptors (ERs), usually referred to as ERa and ERb and both widely distributed in the brain (Perez et al., 2003). In the brain of patients with AD, both ERa and ERb are abnormally regulated. For example, the mitochondrial ERb is reduced in the frontal cortex of female patients with AD (Long et al., 2012), and the alternative splicing of ERa mRNA is diminished in the AD brain especially in female cases (Ishunina & Swaab, 2012). Paradoxically, in the hippocampus of patients with AD, the ERa-expressing neurons are decreased (Hu et al., 2003), while the ERb immunoreactivity is increased (Savaskan et al., 2001). As previously reported, neuroprotection against b-amyloid toxicity by estrogen administration requires the expression of ERa or ERb, as well as activation of the mitogen-activated protein kinase pathway (Fitzpatrick et al., 2002).
These studies indicate the potential roles for ERa and ERb in the pathogenesis of AD. However, the specific effect of ERa or ERb in tauopathy is still elusive. Here, using specific human ERa and ERb plasmids and small interfering RNA transfection, we demonstrated that ERa positively, whereas ERb negatively, regulated the phosphorylation levels of tau protein. Interestingly, the distinct regulation of tau phosphorylation by ERa and ERb resulted from their opposite regulatory role on miR-218, resulting in differential changes in protein tyrosine phosphatase (PTP) a. The abnormal expression of PTPa resulted in the aberrant tyrosine phosphorylation and thus function of glycogen synthase kinase-3 (GSK-3) and protein phosphatase 2A (PP2A), causing a disruption in the phosphorylation balance of tau protein. Our data provide a novel mechanism for the epigenetic regulation of tau phosphorylation in AD, which may suggest new therapeutic targets.

ERa and ERb differentially regulated tau phosphorylation
To explore the potential roles of ERa and ERb in tau phosphorylation, we first examined the relevance of ERa and ERb in tau phosphorylation in the prefrontal cortex of 18-month-old Tg2576 mice (pathogenic stage), which is a widely used AD mouse model. By analyzing fluorescence intensity, we found that ERa was negatively correlated with Tau1, which is a nonphosphorylated tau located at Ser198/199/202 sites, while the intensity of ERb was negatively correlated with AT8, which is a phosphorylated tau at Ser202/Thr205 (Fig. 1A,B). In addition, ERa was positively correlated with AT8, and ERb was positively correlated with Tau1 ( Fig. 1C,D). These correlations are specific: First, neither ERa nor ERb is correlated with total tau (Tau5) in the pathogenic stage ( Fig. S1a, b). Second, the elevation of ERa and decreasing of ERb are only been detected in the pathogenic stage (Fig. S1c) but not in the nonpathogenic stage of Tg2576 mice (3 months old) (Fig. S1c). Third, the alterations of ERs are not age-dependent because the ERa is not altered and ERb displays a minor decrement in aged wild-type mice (Fig. S1c). Fourth, given that there were no significant changes in ERa in aged wild-type mice, we only examined the correlation of tau phosphorylation with ERb and did not observe similar positive correlation of ERb with Tau1 (Fig.  S1d). These data suggested that ERa and ERb might differentially regulate tau phosphorylation.
To validate the exact role of ERa and ERb in tau phosphorylation, we overexpressed human ERa and ERb in HEK293/tau cells to examine the phosphorylation levels of tau. The phosphorylation of tau at Thr205, Ser214, Thr231, Ser396, and Ser404 sites dramatically increased in response to ERa overexpression, while significantly decreased to ERb overexpression ( Fig. 1E-H). These data suggested that ERa promoted tau phosphorylation, while ERb delayed tau phosphorylation.
We then asked whether blocking ERa and ERb could reverse the abnormal tau phosphorylation induced by ERa and ERb overexpression. ICI 182,780 (ICI), which is a nonspecific ER antagonist, reduced ERacaused tau hyperphosphorylation and precluded ERb-dependent attenuation in tau phosphorylation (Fig. S2a-c). Meanwhile, ICI treatment alone did not alter the tau phosphorylation level (Fig. S2d,e), as well as the protein level of ERa and ERb (Zhao et al., 2007;Zou et al., 2009).
We then applied an effective small hairpin (sh) RNA plasmid that specifically targeted mouse ERa or ERb. We transfected the effective shRNA with its scrambled control into Neuro2A cells and examined tau phosphorylation by Western blot. shRNAs selectively decreased the protein levels of ERa or ERb but not altered another receptor level . Silencing ERa reduced the phosphorylation of tau at multiple sites, while silencing ERb facilitated the phosphorylation of tau at multiple sites ( Fig. S3a-d). Together, these findings demonstrated that ERa and ERb differentially regulated tau phosphorylation.
As abnormally phosphorylated tau usually aggregates to form paired helical filaments, which are the dominant component of neurofibrillary tangles in AD, we then examined the tau aggregation in the HEK293/tau cells upon ERa or ERb treatment. We found that ERa overexpression significantly increased the phosphorylation of tau in the insoluble fraction but not the soluble fraction, as well as the amount of aggregated tau in the insoluble fraction ( Fig. S4a-c). In cells overexpressed with ERb, we detected much weaker immunoreactivities on tau phosphorylation in the soluble fraction and no signals  in the insoluble fraction. Those data further confirmed the critical roles of ERs in the tauopathy in AD.

ERa and ERb differentially regulated the tyrosine phosphorylation of GSK-3b and PP2A by PTPa
To further understand the mechanisms underlying ER-regulated tau phosphorylation, we screened the main kinases and phosphatases that are involved in tau phosphorylation. The phosphorylation of Ser9 in GSK-3b was significantly decreased, while the phosphorylation of Tyr216 in GSK-3 and Tyr307 in PP2A was dramatically increased with ERa overexpression. In the cells overexpressing ERb, an obvious increment in the phosphorylation of Ser9 in GSK-3b and a decrement in the phosphorylation of Tyr216 in GSK-3 and Tyr307 in PP2A were observed. In addition, no alterations were found in the total levels of GSK-3b, PP2A, protein kinase (PK) Aa, PKAb, cdk5, and p35/25 ( Fig. 2A-D). Treatment of the cells with ICI reversed the ERa-induced tyrosine hyperphosphorylation and restored the ERb-induced tyrosine hypophosphorylation . As predicted, ICI alone did not induce the alterations of GSK-3 ( Fig. S5f-g). The above results suggested abnormal and differential regulation in tyrosine phosphorylation in response to ERa and ERb overexpression.
As reportedly previously, tyrosine phosphorylation is mostly mediated by tyrosine kinases and phosphatases. Among those kinases and phosphatases, Src, fyn, PTPa, and PTP1B have been implicated in the pathogenesis of AD or tauopathy. We then examined the levels of those enzymes. In response to ERa overexpression, the levels of total Src, fyn, and PTP1B were not changed and the Tyr416 of Src was increased, but the Tyr527 of Src and the level of PTPa were decreased, suggesting the activation of Src and the inhibition of PTPa (Fig. 3A,B). We then applied PP2, a specific Src inhibitor, and a Src shRNA (Fig. 3C,D, si-Src) to test whether they reversed the ERa overexpression-induced tau and tyrosine hyperphosphorylation. Both PP2 and si-Src treatment indeed rescued the tyrosine hyperphosphorylation of GSK-3b and PP2A (Fig. 3A,B,E,F) and ERa-caused phosphorylation of tau ( Fig. 3G-J). These data suggested that ERa facilitated tau phosphorylation by activation of GSK-3b and inactivation of PP2A through Src activation via PTPa inhibition.
In response to ERb overexpression, the levels of total Src, fyn, and PTP1B and the phosphorylation of Src were not changed, but the level of PTPa was increased, suggesting the activation of PTPa (Fig. S5a,e). Using a PTPa-specific small interfering RNA (Fig. 4A,B, si-PTPa), knocking down of PTPa effectively restored the tyrosine phosphorylation of GSK-3b and PP2A (Fig. 4C,D) and inhibited the attenuation of tau phosphorylation at multiple sites by ERb ( Fig. 4E-F). These data suggested that ERb weakened tau phosphorylation through the inhibition of GSK-3b and the activation of PP2A by promoting PTPa expression.
To further verify that ERa and ERb differentially regulate PTPa pathway, we examined the tyrosine phosphorylation of GSK-3b, PP2A, and Src and the protein level of GSK-3b, PP2A, Src, fyn, and PTPa upon the ERa or ERb silencing in N2a cells. We found that silencing ERa caused a decrease in tyrosine phosphorylation of GSK-3b and PP2A, which was accompanied with an elevation in PTPa protein level (Fig. S6a,b), while silencing ERb induced the increment of tyrosine phosphorylation of GSK-3b and PP2A, along with the suppression of PTPa protein level (Fig. S6c,d). Furthermore, the PTPa protein level was decreased specifically in the pathogenic Tg2576 mice, but neither in nonpathogenic mice nor in aged normal mice (Fig. S6e-f), which was consistent with ERa overexpression results. In addition, silencing PTPa alone enhanced the tyrosine phosphorylation of GSK-3b and PP2Ac ( Fig. S6g-h). The above data suggest that ERa and ERb differentially regulate the expression of PTPa, which in turn results in the differential tyrosine phosphorylation of GSK-3b and PP2A.

ERa and ERb differentially regulated PTPa by miR-218
We then examined how ERa and ERb induced the differential expression of PTPa. We first examined the mRNA levels of PTPa in HEK293/tau cells overexpressing ERa or ERb and did not find any differences among the three groups (Fig. 5A), indicating that a posttranscriptional modification might be involved in the regulation of PTPa protein levels by ERs. Because microRNAs (miRNAs) are the major regulators of posttranscriptional modification, we then performed a bioinformatics prediction with the online tool Targetscan. miR-218 had a highly conserved site that bound to the 3ʹ untranslated region (UTR) of PTPRA, the gene for PTPa (Fig. 5B). To determine whether this site was targeted by miR-218, we constructed a luciferase reporter with wild-type and mutant 3ʹUTR segments. The wild-type reporter showed apparent inhibition while the mutant one did not when coexpressed with miR-218 ( Fig. 5B,C). Further, overexpression of hsa-miR-218 mimics suppression in protein levels of PTPa (Fig. 5D), suggesting that miR-218 directly targeted PTPRA. Realtime polymerase chain reaction was then performed to determine the levels of hsa-miR-218 in ERa-and ERb-overexpressing cells. The levels of miR-218 were increased in ERa-overexpressing cells, while the levels were decreased in ERb-treated cells (Fig. 5E). The application of anti-miR-218 restored and miR-218 mimics suppressed the levels of PTPa resulting from ERa or ERb overexpression ( Fig. 5F-I), suggesting that ERa and ERb differentially regulated PTPa through miR-218.
Finally, we tested whether correcting the miR-218 disturbances affected the tau phosphorylation levels via PTPa signals. We found that administration of anti-miR-218 to the ERa-overexpressing cells attenuated tau hyperphosphorylation, while the administration of miR-218 mimicked the effect ( Fig. 5F-I). Meanwhile, the anti-miR-218 application reduced the tyrosine hyperphosphorylation caused by ERa overexpression, while the miR-218 mimics restored the tyrosine hypophosphorylation induced by ERb overexpression ( phosphorylation of GSK-3 and PP2Ac, while miR-218 inhibitor treated alone suppressed the tau phosphorylation and tyrosine phosphorylation of GSK-3 and PP2Ac ( Fig. S7a-f). These data further demonstrated that ERa and ERb differentially regulated tau phosphorylation through the miR-218/PTPa pathway.

Discussion
The accumulation of Ab plaques and tau NFT are the two major pathological hallmarks in AD, which is the dominant form of dementia in aged people. The occurrence of AD and the global changes in AD pathology are known to significantly correlate with the loss of estrogen in women after menopause (Barnes et al., 2005). In a study of more than 5000 brain samples, females had more affected brain regions with NFT formations than males (Corder et al., 2004). Many studies have indicated that neurons are more susceptible to age-related neurodegenerative processes with declining levels of estrogen in the brain, suggesting the potential protective roles of estrogen against AD. Estrogen exerts its neuroprotective effects through various ERs, which consist mainly of the two isoforms, ERa and ERb. Both of these are enriched in the neocortex and hippocampus, which are two brain areas that are highly involved in AD. Many studies have described the potential roles of ERa and ERb in AD pathogenesis. For example, in 2-month-old ERb-knockout mice, b-amyloid deposits and apolipoprotein E are widely distributed in the brain (Zhang et al., 2004), suggesting that ERb signaling disruption results in Ab deposition. In W4 cells, estrogen treatment reduces Ab-induced cell death through ERa-dependent pathways (Kim et al., 2001). These lines of data strongly suggest the critical roles of ERa and ERb in Ab generation. In addition, ER signaling disruption has been implicated in tauopathy. Clinical studies have suggested that there is a positive correlation between tau expression in breast cancer cells and ER expression and that this is influenced by ER signaling (Andre et al., 2007;Pentheroudakis et al., 2009). Administering ICI 182,780 does not change tau phosphorylation but reverses the tau hyperphosphorylation that is induced by okadaic acid, suggesting the involvement of ERs in tau phosphorylation (Zhang & Simpkins, 2010). In the current study, we first revealed that ERa and ERb exerted adverse effects on tau phosphorylation. The overexpression of ERa caused tau hyperphosphorylation and aggregation, while the overexpression of ERb induced tau hypophosphorylation. The opposite effects of ERs on tau phosphorylation were mainly caused by the differential regulation of miR-218/PTPa signaling, which in turn disturbed the balance of GSK-3b/ PP2A. The differential role of ERa and ERb in other biological processes has been well studied. For example, ERa activates while ERb suppresses the gene expression of cyclin D1 (Liu et al., 2002). ERa mediates the cancer-promoting effects of estrogens, and ERb inhibits the proliferation with ERa with or without 5 lM PP2 for 1 h, and the samples were collected for Western blot. The representative blots for pS9-GSK-3b, pY216-GSK-3b, GSK-3b, pY307-PP2Ac, PP2Ac, pY416-Src, pY527-Src, Src, Fyn, PTPa, and PTP1B are shown in (A) and the quantitative analysis is presented in (B). *P < 0.05, **P < 0.01, vs. vector transfection group. # P < 0.05, ## P < 0.01, vs. ERa overexpression group (N = 3). (C, D) HEK293/tau cells were transfected with si-Src oligonucleotide (si-Src) or its scrambled control (ssi-Src), and the samples were subjected to Src antibody for Western blot (C) and quantitative analysis (D) (N = 3). (E, F) HEK293/tau cells were transfected with ERa plus ssi-Src (ERa+Ssi-Src) or ERa plus si-Src (ERa+si-Src) or the vectors (vector). Cell lysates were used for Western blot to detect the level of pY216-GSK-3b, GSK-3b, pY307-PP2Ac, PP2Ac, and Src (E), and the quantitative analysis was performed (F). *P < 0.05, **P < 0.01, vs. vector transfection group. # P < 0.05, ## P < 0.01, vs. ERa overexpression group (N = 3). (G-J) HEK293/tau cells were transfected with ERa with or without 5 lM PP2 for 1 h (G) or with ERa plus ssi-Src (ERa) or ERa plus si-Src or the vectors (H), and the samples were subjected for the detection of tau phosphorylation. Quantitative analysis was performed in (I) and (J). *P < 0.05, **P < 0.01, vs. vector transfection group. # P < 0.05, ## P < 0.01, vs. ERa overexpression group (N = 3). of breast cancer cells by repressing c-myc and cyclin A gene transcription (Paruthiyil et al., 2004). In HC11 mammary epithelial cells, ERa drives proliferation in response to E2, while ERb is growth inhibitory (Helguero et al., 2005). In the nervous system, ERa could impair memories for socially acquired food preferences, while ERb could enhance the acquisition of the task (Clipperton et al., 2008). Our data suggested the different roles of ERa and ERb in AD. ERa may have deleterious effects, but ERb may have protective effects.
As one of the most important noncoding RNAs, the roles of miRNA in AD pathogenesis have been well studied. After the first miRNA array report in AD appeared in 2007 (Lukiw, 2007), multiple studies have identified a number of disrupted miRNAs in AD. It has been reported that the decrement in miR-29a/b-1 results in the increased expression of bsecretase 1, which is the most important b-secretase, and the overproduction of Ab (Hebert et al., 2008). The depletion of Dicer, which is a RNase that is required for miRNA maturation, induces the hyperphosphorylation of tau, indicating the critical roles of miRNAs in tauopathy (Hebert et al., 2010). In progressive supranuclear palsy, the loss of miR-132 is associated with tau exon10 inclusion, which further induces an imbalance of the 4R/3R-Tau ratio in neuronal cells (Smith et al., 2011). In addition, two members of the miR-16 family, miR-15a and miR-15b, are downregulated in AD brain and cerebrospinal fluid, respectively    (Cogswell et al., 2008). Both of these have been suggested to target the 3ʹUTR of extracellular signal-regulated kinase 1 (Hebert et al., 2012), which is an important kinase for tau phosphorylation in AD (Ferrer et al., 2001). These lines of data provide the preliminary links between tau phosphorylation and miRNA dysfunction, but direct experimental evidence is missing. Here, we found that miR-218 acted as the axis pT231 pS9-GSK-3β   4). (F, G) HEK293/tau cells were transfected with ERa plus hsa-miR-218 inhibitors scrambled control (ERa+scr) or ERa plus hsa-miR-218 inhibitors (ERa+I-miR-218) or the control (Con), and the samples were used for the detection of tau phosphorylation and PTPa level (F, G). **P < 0.01, vs. control group. ## P < 0.01, vs. ERa plus hsa-miR-218 inhibitors scrambled control treated group (N = 3). (H, I) HEK293/tau cells were transfected with ERb plus hsa-miR-218 mimics control (ERb+scr) or ERb plus hsa-miR-218 mimics (ERb+M-miR-218) or the control (Con), and the samples were used for examination of tau phosphorylation and PTPa (H, I). **P < 0.01, vs. control group. ## P < 0.01, vs. ERb plus hsa-miR-218 mimics control treated group (N = 4). (J, K) HEK293/tau cells were transfected with ERa plus hsa-miR-218 inhibitors scrambled control (ERa+scr) or ERa plus hsa-miR-218 inhibitors (ERa+I-miR-218) or the control (Con), and the samples were used for the detection of pS9-GSK-3b, pY216-GSK-3b, GSK-3b, pY307-PP2Ac, PP2Ac, pY527-Src and Src (J, K). **P < 0.01, vs. control group. ## P < 0.01, vs. ERa plus hsa-miR-218 inhibitors scrambled control treated group. (L, M) HEK293/tau cells were transfected with ERb plus hsa-miR-218 mimics control (ERb+scr) or ERb plus hsa-miR-218 mimics (ERb+M-miR-218) or the control (Con), and the samples were used for the detection of pS9-GSK-3b, pY216-GSK-3b, GSK-3b, pY307-PP2Ac, PP2Ac, and Src (L, M). **P < 0.01, vs. control group. ## P < 0.01, vs. ERb plus hsa-miR-218 mimics control treated group.
in regulating tau phosphorylation upon ERa or ERb activation. Specifically, ERa overexpression increased miR-218 expression and tau phosphorylation, and suppression of the increased miR-218 levels rescued the tau hyperphosphorylation that was caused by ERa. ERb overexpression decreased miR-218 expression and tau phosphorylation, and supplementation of miR-218 mimics blocked the alleviation of the tau phosphorylation that was induced by ERb. As previously reported, miR-218 accumulates in the hippocampus (Bak et al., 2008) and is activated during neuronal differentiation (Sempere et al., 2004). A number of miR-218 targets have been identified to exert diverse functions in the brain. For example, miR-218 targets multiple components of receptor tyrosine kinase signaling pathways, and miR-218 repression increases the abundance and activity of multiple receptor tyrosine kinase effectors (Mathew et al., 2014). In our study, miR-218 specifically targeted the 3ʹUTR of PTPA, the gene for PTPa, and regulated the tyrosine phosphorylation of GSK-3b and PP2A. Our study thus extended the potential role of miR-218 in the brain. PTPa belongs to the protein tyrosine phosphatase family that regulates a variety of cellular processes, including cell growth, differentiation, and the mitotic cycle (Pallen, 2003). The expression of PTPa is accompanied by Src dephosphorylation and activation in the developmental stage of neurons (den Hertog et al., 1993), and the stable expression of PTPa will activate Src and mediate epidermal growth factor-induced neurite outgrowth (Yang et al., 2002). Moreover, PTPa combines with the neural cell adhesion molecule contactin to form a receptor complex that plays an important role in neuronal cell interactions (Berglund et al., 1999) and in hippocampal synaptic plasticity (Murai et al., 2002). Most importantly, PTPa has been implicated in tyrosine phosphorylation and in the regulation of the activity of its substrates. For example, the physical interactions of phosphoinositide 3-kinase and protein kinase Cd with PTPa play a role in the activation of mitogen-activated protein kinase (Stetak et al., 2001). Here, we first reported that the phosphorylation of tyrosine 216 in GSK-3b and tyrosine 307 in PP2A was regulated by PTPa and concordant with the alterations in GSK-3b and PP2A activity, which in turn resulted in abnormal tau phosphorylation levels. The activation of PTPa increased the dephosphorylation of tyrosine sites and induced the inhibition of GSK-3b and the activation of PP2A, which further suppressed the phosphorylation of tau protein. Although PTPa changes in AD have not been studied, the attenuation of tau phosphorylation by restoring PTPa levels is a potential therapeutic strategy.
Taken together, our study demonstrated the differential regulation of ERa and ERb on tau phosphorylation through miR-218/PTPa signals for the first time and provided data on the fundamental role of the miR-218/ PTPa pathway in tauopathy.

Experimental procedures Antibodies and reagents
All the primary antibodies used in this study are list in Table 1. ICI 182,780 was purchased from Tocris Bioscience (Bristol, UK) and dissolved in DMSO to 100 lM as stocking solution. Specific inhibitor PP2 was purchased from Merck KGaA (Darmstadt, Germany) and dissolved in DMSO to 5 mM for stock. Lipofectamine 2000 was purchased from Invitrogen (San Diego, CA, USA). Cell culture media were from Gibco (San Diego, CA, USA). Plasmids containing the human ERa and ERb cDNAs were constructed according to the following sequences: NM_001122741 and NM_001437 to pEGFP-N1. shRNA-ERb plasmid toward mouse ERb (NM_207707.1) was constructed to vector GV102 by Neuron Biotech Inc. (Shanghai, China). si-ERa oligonucleotide toward mouse was synthesized using the sequence published before (Carbonaro et al., 2009). Si-Src and si-PTPa oligonucleotides toward human were synthesized according to previous publications (Zheng et al., 2008). The primers used for PTPa detection are as follows: forward, 5 0 -AGTG GTCTGATATGTGTCAGTGC-3 0 ; reverse, 5 0 -GGTTCTGCCGTTGATGAGT TA-3 0 . The primers for hsa-miR-218 detection, the hsa-miR-218 mimics (Cat. No. miR10000275-1-2) and inhibitors (Cat. No. miR20000275-1-2), as well as their scrambled controls, were purchased from Ribobio Co., Ltd (Guangzhou, China).

Animals and treatment
Eighteen-month-old male Tg2576 mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA) and housed in a room on a 12:12 hr light-dark cycle and 22 AE 2°C with water and food ad libitum for at least 2 weeks before the day of experimentation. All animal experiments were performed according to the 'Policies on the Use of  (FBS, vol/vol). Both cells were cultured in a humidified atmosphere of 5% CO 2 at 37°C. The cells were cultured for at least 24 h after plating, and when grown to 80-90% confluence, the culture medium was replaced with serum-and antibiotic-free DMEM prior to treatment. Plasmids used for transfection were amplified and purified by Qiagen kit (Qiagen, Hilden, Germany) according to the manufacturer's instruction. Briefly, HEK293/tau or N2a cells were seeded in six-well plates, grown to 60-70% confluence, and then cultured in serum-and antibiotic-free OPTI-MEM for 4 h. Plasmids were transfected with Lipofectamine 2000 (Invitrogen) according to the manufacturer's instruction. Cells transfected with GFP constructs were visualized at 48 h after transfection by an Olympus IX70 microscope with a 209LCPlanF1 lens (Olympus Corporation, Matsue, Shimane Japan). For double transfection, the plasmid and oligonucleotides were added to OPTI-MEM, respectively, at the last step of transfection. 48-60 h after transfection, cells were treated with 100 nM ICI 182,780 for 1 h or 5 lM PP2 for 1 h. Then, the media were removed and the cells were harvested and stored at À20°C for further experiments.

Immunofluorescence and confocal microscopy
A total of 5 mice for each group were sacrificed by overdose chloral hydrate (1 g kg À1 ) and perfused through aorta with 100 mL 0.9% NaCl followed by 400 mL phosphate buffer containing 4% paraformaldehyde. About 2 h later, brains were removed and postfixed in perfusate overnight and then cut into sections (15-20 lm) with vibratome (Leica, Nussloch, Germany; S100, TPI). The sections of mice brain were collected consecutively in PBS for immunofluorescence staining. Free-floating sections were incubated with bovine serum albumin (BSA) to block nonspecific sites for 30 min at 25°C. Sections were then incubated overnight at 4°C with primary antibodies rabbit polyclonal ERa or ERb antibody for 48 h, and after washing with PBS, sections were subsequently incubated with mouse monoclonal Tau1, Tau5, or AT8 for 48 h. After washed with PBS for 30 min, sections were subsequently incubated secondary antibodies Alexa Fluor 488 (donkey anti-mouse) or Alexa Fluor 546 (goat anti-rabbit) for 1 h at 37°C. The prefrontal cortex region was chosen for imaging using a laser confocal microscope (LSM710 Carl Zeiss, M€ unchen, Germany) (Chen et al., 2012).

Real-time PCR
The total RNA from the cells was extracted by TRIzol reagent (Invitrogen), and 1 lg RNA was reversely transcripted. qRT-PCR was performed on ABI StepOne Plus using SYBR Green â Premix Ex Taq (Takara, Tokyo, Japan). MicroRNA was extracted using miRNA isolation kit (Tiangen, Beijing, China). Reactions were prepared in a total volume of 10 lL containing 0.5 lL cDNA (100 ng lL À1 ), 1 lL of each 2 lM primer (300 mM each), 5 lL of SYBR Green, and 2.5 lL RNase/DNase-free sterile water. Blank controls were run in triplicate for each master mix.
The cycle conditions were set as follows: initial template denaturation at 95°C for 1 min, followed by 40 cycles of denaturation at 95°C for 5 s, and combined primer annealing/at 60°C for 30 s, and elongation at 72°C for 30 s. This cycle was followed by a melting curve analysis, ranging from 60 to 95°C, with temperature increasing by steps of 0.5°C every 10 s.

Western blotting
Cells were rinsed twice in phosphate-buffered saline at pH 7.5 and lysed with buffer containing 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 0.02% NaN3, 100 lg mL À1 PMSF, and 10 lg mL À1 each of the protease inhibitors (leupeptin, aprotinin, and pepstatin A) followed by boiling for 5-6 min, and then sonicated for 5 s on ice. The cell lysates were then centrifuged at 12 000 g for 5 min at 4°C; aliquots of supernatants were added to one-third volume of 49 sample buffer, 10% beta-mercaptoethanol (ME), and 0.05% bromophenol blue and then stored at À20°C or used immediately. Protein concentration was quantitated using the BCA Protein Assay Reagent Kit (Pierce, Rockford, IL, USA) (Jiang et al., 2011;Liu et al., 2015).
Equal amounts of protein were separated by SDS-polyacrylamide gel electrophoresis (10% gel) and transferred to nitrocellulose membrane. The membranes were blocked with 5% nonfat milk dissolved in PBS (50 mM Tris-HCl, pH 7.6, 150 mM NaCl) for 30 min-1 h and probed with primary antibodies overnight at 4°C. Then, the blots were incubated with goat anti-mouse or anti-rabbit conjugated to IRDye 800 (Rockland Immunochemicals) (1:15000) for 1 h at 25°C. The protein bands were visualized and quantified by the Odyssey infrared imaging system (LI-COR, Lincoln, Nebraska, USA). The levels of the phosphorylated protein tau, PP2Ac, GSK-3b, and Src were normalized against the total protein tau, PP2Ac, GSK-3b, Src, and PTPa. The amount of protein was expressed as relative level of the sum optical density against controls.
For tau aggregation analysis (Ishihara et al., 1999;Li et al., 2014), the cells were homogenized in cold RAB Hi-Salt buffer (0.1 M MES pH 7.0, 1 mM EGTA, 0.5 mM MgSO4, 0.75 M NaCl, 0.1 M EDTA) containing protease inhibitors (100 lg mL À1 PMSF) and centrifuged at 50 000 g for 40 min in 4°C, and the supernatants were saved as the RAB-soluble fraction. The RAB-insoluble pellets were sonicated in sample buffer containing 0.2 g mL À1 sucrose, 18.5 mM Tris (pH 6.8), 2 mM EDTA, 80 mM DTT, and 2% SDS and centrifuged at 50 000 g for 20 min in 4°C. The supernatant was discarded, and the pellet was homogenized in cold RIPA buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM EDTA, 50 mM natrium fluoride) with protease inhibitor and centrifuged at 50 000 g for 20 min in 4°C. The supernatant was saved as RIPA-soluble fraction, and the pellet was extracted in 70% formic acid as FA fraction. Fractions were analyzed by SDS-PAGE.

Statistical analysis
Data were expressed as mean AE SD and analyzed using SPSS 10.0 statistical software (SPSS Inc., Chicago, IL, USA). The one-way ANOVA procedure followed by LSD's post hoc tests was used to determine the different means among groups.

Funding
This work was supported in part by the National Natural Science Foundation of China (81361120245, 31201011, 81261120570,

Supporting Information
Additional Supporting Information may be found in the online version of this article at the publisher's web-site.