Inhibiting α1-adrenergic Receptor Signaling Pathway Ameliorates Alzheimer’s Disease Type Pathologies and Behavioral De cits in an AD Mouse Model


 Background The role of α1 adrenergic receptors (α1-ARs) signaling pathway in the pathogenesis of Alzheimer’s disease (AD) has rarely been investigated. Clarifying pathophysiological functions of α1-ARs in the AD brain is helpful for better understanding the pathogenesis and screening novel therapeutic target of AD. Methods This study included 2 arms of in vivo investigations: 1) 6-month-old female APPswe/PS1 mice were intravenously treated with AAV-PHP.eB-shRNA (ARs)-GFP or AAV-PHP.eB-GFP for 3 months. 2) 3-month-old female APPswe/PS1 mice were daily treated with 0.5 mg/kg terazosin or equal saline for 6 months. SH-SY5Y cell lines bearing human Amyloid precurssor protein were treated with terazosin or saline for investigating possible mechanisms. Results α1-ARs knockdown mice exhibited improved behavioral performances than control mice. α1-ARs knockdown mice had significantly lower brain amyloid burden, as reflected by soluble Aβ species, compact and total plaques, than control mice. The α1-ARs inhibitor terazosin substantially reduced Aβ deposition, attenuated downstream pathologies including Tau hyperphosphorylation, glial activation, neuronal loss, synaptic dysfunction, and rescued behavioral deficits of APPswe/PS1 mice. In vitro investigation demonstrated that α1-ARs inhibition down-regulated BACE1 expression, and promoted ser9 phosphorylation of GSK-3β, thus reduced Aβ production. Conclusions This study indicates that inhibition of α1-ARs signaling pathway might represent a promising therapeutic strategy for AD.


Introduction
Alzheimer's disease (AD) is the most popular form of neurodegenerative diseases among the older adults (1). AD has two major pathological hallmarks, including the intracellular senile plaques consisted of Amyloid-beta (Aβ) plaques and the intracellular neuro lament tangles (NFT) formed by hyperphosphorylated Tau (2). It is suggested that the loss of noradrenergic neurons in the locus coeruleus (LC) represents an early pathological change in the AD brain. Noradrenergic degeneration in the AD brain simultaneously leads to compensatory changes, such as modi cations of neuronal anatomy, neurotransmitter, and noradrenergic receptors (3,4), and enhanced responses to norepinephrine (NE) that likely underlie aggressive behaviors, agitation, and sleep disturbance et al. in AD patients (5).
Previous studies have well documented the role of α2-and β-adrenergic receptors in the pathogenesis of AD (6, 7). However, the pathophysiological functions of α1 adrenergic receptors (α1-ARs) in the AD brain have been rarely investigated. A panel of studies have suggested that the expression of α1-ARs is increased in the AD brain (8-10), with the pathophysiological signi cance unknown. In this study, we investigated the effects of pharmacological and genetic inhibition of α1-ARs on behavioral and pathological phenotypes in APP/PS1 mice.

Animals and treatment
APPswe/PS1dE9 transgenic mice (AD mice) were obtained from Jackson Laboratory and C57BL/6 wildtype (Wt) mice were obtained from the animal center of Daping hospital, Third Military Medical University. 3-month-old AD mice were daily treated with 0.5 mg/kg terazosin or equal saline intraperitoneally from 3 months old to 9 months old (n = 8 per group). 6-month-old AD mice were treated with AAV-PHP.eB-shRNA (α1-ARs)-GFP for 3 months. Control AD mice were treated with a same amount of AAV-PHP.eB-GFP. At 9 months of age, mice were subjected to behavioral tests and were sacri ced for biochemical and histological experiments. All animal experimental protocols were approved by Third Military Medical University Animal Welfare Committee and performed according to the guidelines of laboratory animal care and use.

Behaviour tests
After treatment, mice were assessed for behavioral analyses using the Morris water maze (MWM), Ymaze, elevated plus maze and open-eld tests. All behavioral performances were tracked and analyzed with a computer tracking system (ANY-maze, Stoelting, Wood Dale, IL, USA).
In brief, the MWM test consisted of three platform trials per day for 5 consecutive days, followed by a probe trial. In the platform trial, the distance of the path taken and the escape latency were measured to test spatial learning ability. In the probe trial, the time spent in each quadrant and the number of annulus crossings were recorded to assess memory consolidation.
Y maze test with spontaneous alternation performance was conducted according to a previous study (11). Each mouse was placed in the center of the symmetrical Y maze and was allowed to explore freely through the maze for 5 minutes. The sequence and total number of arms entered were recorded.
Percentage of alternation was determined as follows: number of trials containing entries into all three arms/maximum possible alternations × 100. The maximum possible alterations = total number of arms entered − 2.
In the novel arm exploration, one arm (de ned as the novel arm) was closed. After freely exploration for 5 min and 25 min interval, mice were resubjected to apparatus for the retrieval trial, during which mice were allowed to explore all arms freely for 5 min. The number of entries and time spent in the closed arm (novel arm) were analyzed.
The open eld test was conducted to meaure the spontaneous locomotor activity. The apparatus was made of four transparent Plexiglass walls (60cm high) and a gray 40cm x 40cm oor. Mice were subjected to apparatus for freely exploring for 5 min. The total distance travelled and average speed were analyzed.
The Elevated Plus Maze test is conducted to assess anxiety-related behavior. The EPM apparatus consists of a "+"-shaped maze elevated above the oor with two oppositely positioned closed arms, two oppositely positioned open arms, and a center area. In brief, mice were subjected into the center of appratus and allowed for freely exploring for 5min. The number of entries into and time spent in the open arm were analyzed.

Brain sampling
All experimental mice were sacri ced at 9-month of age. After anesthetized, mice were intracardial perfusion with 0.1% NaNO 2 in 0.9% saline, the right brain hemisphere was dissected and xed with 4% paraformaldehyde (PFA). Left brain hemispheres were snap-frozen by liquid nitrogen, ground into a powder, and stored at -80°C for biochemical analysis.

Immunohistochemistry and immuno uorescence
Immunohistochemistry and immuno uorescence were conducted to our previous methods (13). Serial coronal brain sections (30 µm in thickness) were cut using cryostat. Five brain sections spanning the entire brain were stained with a 6E10 antibody or Congo red (Sigma) to detect Aβ plaques in the brain (6E10 antibody for compact and diffuse Aβ; Congo red for compact Aβ). Cerebral amyloid angiopathy (CAA) was manually assessed in Congo red-stained hippocampal sections under a microscope (14). Brain sections were also incubated with anti-CD68 antibody (Abcam) and anti-glial brillary acidic protein (GFAP) antibody (Abcam) to assess the microgliosis and astrocytosis, respectively. Double immuno uorescence staining for NeuN and microtubule associated protein (MAP)-2 (Abcam) were performed to detect the neuronal loss and neurite degeneration. Apoptosis of neurons in the CA3 area of hippocampus was assessed by double-immuno uorescence staining for NeuN (Abcam) and caspase-3 (Abcam).

ELISAs
Proteins were extracted from brain tissues in tris buffer solution (TBS), 2% sodium dodecyl sulfonate (SDS) to measure soluble and insoluble Aβ levels in brain. The levels of human Aβ42 and Aβ40 in the TBS and SDS fractions were measured using ELISA kits (Invitrogen, USA). Adeno-associated virus (AAV) transduction for α1-ARs interference in vivo 100µL AAV-PHP.eB-hsyn-shRNA (α1-ARs)-GFP-pA or AAV-PHP.eB-hsyn-GFP-pA (1x10 13 v.g/ml) was injected into tail vein of 6-month-old AD mice. The target sequence of α1-ARs is provided in AAV 5′→3′ orientation: TGCCAAGAATAAGACTCACTT, respectively. The e ciency of interfering of AAV was detected by the expression of α1-ARs using western blot.
Cell culture and treatment SH-SY5Y-APP695 cells were maintained in DMEM (H) supplemented with 10% fetal bovine serum, G418 (100µg/ml), and 1% penicillin-streptomycin. SH-SY5Y-APP695 cells were seeded onto the 6-well plate or 24-well plate with a cover slide. After attachment, the cells were treated with 50µM terazosin or equal volume saline for 4 h. SH-SY5Y-APP695 cells treated with terazosin or saline were collected for immuno uorescence detection and western blot analysis.

Statistical Analysis
All data were expressed as the mean ± SEM and analyzed by SPSS 22.0 software (SPSS, Chicago, IL, USA). Appropriate two-tailed independent t-tests and MannWhitney U test were conducted for the comparsion between two groups. One-way analysis of variance (ANOVA) followed by Tukey post hoc analysis was used for multiple comparisons. A P value < 0.05 was considered as statistically signi cant.

Knockdown of α1-ARs improved the behavioral de cits in AD mice
The expression of α1-ARs in brain was sigi cantly reduced 3 months after AAV-α1-ARs treatment ( Supplementary Fig. 1). Mice treated with AAV-α1-ARs dispalyed better performance in both Y maze test and open-eld test compared with AAV-GFP treated mice. In the Y maze test, mice treated with AAV-α1-ARs had more time spent in the novel arm and more entries into the novel arm, as well as a higher spontaneous alternation than AAV-GFP treated mice (Supplementary Fig. 2A), indicating a better performance of working memory and learning. AAV-α1-ARs treatment sigi cantly increased the distance travelled and the travelling speed of AD mice in the open-eld test ( Supplementary Fig. 2B and D).
Compared with AAV-GFP group, AAV-α1-ARs treated mice seemed to display fewer anxiety-like behaviours, although no signi cance has been achieved ( Supplementary Fig. 2C). Taken together, these ndigns demonstrated that the down-regulation of brain α1-ARs signi cantly improved behavioral de cits in AD mice.

Knockdown of α1-ARs decreased brain Aβ burden in AD mice
We further investigated the effect of AAV-α1-ARs treatment on Aβ burden in the brain of AD mice. 6E10 and Congo red staining were performed to detect the total Aβ plaques and compact Aβ plaques, respectively. Compared with control mice, mice treated with AAV-α1-ARs displayed a lower area fraction of 6E10-positive and Congo red-positive staining in the neocortex but not in the hippocampus (Fig. 1A).
Consistently, levels both Aβ42 and Aβ40 in the TBS and SDS fractions of brain homogenates in the AAV-α1-ARs group were lower than those in the AAV-GFP group (Fig. 1B).
We next investigated the expressions of proteins involved in Aβ production, transportation and degradation. There were no differences in the expression of full-length APP and CTF-α between mice treated with AAV-α1-ARs and AAV-α1-ARs. No signi cance was also observed in the expression of ADAM10 and PS1 between the two groups. However, levels of BACE1 and CTF-β in AAV-α1-ARs group were signi cantly lower than that in AAV-GFP group ( Fig. 1C and D). There were no signi cant differences in levels of proteins involved in Aβ transportation and degradation, including IDE, LRP1 and RAGE, between AAV-α1-ARs group and AAV-GFP group (Fig. 1E). The above ndings indicated that the decreased Aβ burden after knockdown of α1-ARs might be attributed to decreased Aβ production but not increased Aβ clearance in the brain.
Knockdown of α1-ARs attentuated other AD-type pathologies in AD mice Mice treated with AAV-α1-ARs exhibited lower area fraction and cell density of activated astrocytes (labeled by GFAP) in both neocortex and hippocampus ( Fig. 2A). Besides, positive area fraction and cell density of activated microglia (labeled by CD68) in the neocortex but not hippocampus was signi cantly lower in AAV-α1-ARs group than AAV-GFP group ( Fig. 2A). Western blot showed that the expression of phosphorylated tau protein, including PS396 and PS199, but not PT231, in AAV-α1-ARs group was lower than that in AAV-GFP group (Fig. 2B). Compared to AAV-GFP treated mice, mice treated with AAV-α1-ARs displayed higher expression of synapse-related proteins, including SYN and VAMP1, but not SNAP25 and PSD95 (Fig. 2C). Taken together, these ndings suggested that down-regulation of α1-ARs attenuated neuroin ammation, tau hyperphosphorylation and neurodegeneration in the brain of AD mice.
We further investigated whether terazosin, a speci c α1-ARs antagonist, had similar therapeutic effects to genetic knockdown of these receptors. Saline treated AD mice displayed an impaired learning ability and spatial reference memory in comparison with Wt mice, while mice treated with terazosin presented improved spatial learning ability, as re ected by a signi cant reduction in the escape latency (Fig. 3A). There were no differences in the time spent in the target quadrant and annulus crossing between terazosin and saline treated AD mice (Fig. 3A). In open-eld tests, Wt mice had a longer travelling distance than AD mice treated with saline, but AD mice treated with saline and terazoisn had no signi cant difference ( Fig. 3B and C). In open eld test, the ratio of time in the central area to that in the peripheral area was signi cantly higher in the terazosin treated group and the Wt group than that in the saline-treated group (Fig. 3C), re ecting decreased anxiety-like behaviours after terazosin treatment.
Furthermore, mice treated with terazosin also exhibited higher spontaneous alternation in the Y-maze test than saline-treated mice, re ecting improved spatial recognition memory after terazosin treatment (Fig. 3D).

Terazosin treatment signi cantly ameliorated Aβ burden in AD mice
Mice treated with terazosin displayed a lower area fraction of 6E10-positive plaques and plaque density in both neocortex and hippocampus. Congo red staining also showed that the area fraction of compact plaques in neocortex and hippocampus in terazosin treated group were lower than that in saline treated group (Fig. 4A), and the compact plaque density in terazosin treated group was lower than that in saline treated group in neocortex, but not in hippocampus (Fig. 4A). Mice treated with terazosin exhibited lower load of CAA, as indicated by Congo red staining, than saline treated mice (Fig. 4B). Consistently, both Aβ40 an Aβ42 concentrations in TBS and SDS fractions of brain homogenates in the terazosin group were lower than that in the saline group (Fig. 4C).
The expression of full-length APP and its metabolites, including sAPPα and CTFα was not signi cantly different between terazosin and saline treated group, while the levels of Aβ dimers and CTF-β in the terazosin group were signi cantly lower than those in saline treated group (Fig. 4D). We also found that the expression of BACE1 in terazosin treated mice was lower than that in saline treated mice (Fig. 4E). However, there were no signi cant differences in other APP metabolic enzymes, Aβ transport-and degradation-related proteins, between terazosin and saline group (Fig. 4E).

Terazosin attentuated other AD-type pathologies in AD mice
Tau 231 immunohistochemical staining and western blot showed that the levels of phosphorylation of Tau at multiple epitopes, including PS396 and PS199, in the terazosin group were signi cantly lower than those in the saline group ( Fig. 5A and B), whereas the expression of total tau (Tau5) didn't differ between two groups (Fig. 5B). Levels of synapse-related proteins, including SYN and SNAP25, but not PSD95, were signi cantly higher in mice treated with terazosin in comparison with mice treated with saline (Fig. 5C).
Compared with mice treated with saline, terazosin treated mice displayed higher area fractions of viable neurons (NeuN) and dendrites (Map-2), but lower area fractions of apoptosis-related protein (caspase-3) in the hippocampus (Fig. 5D). Additionally, mice treated with terazosin exhibited lower area fractions of activated astrocytes in both neocortex and hippocampus than saline treated mice ( Fig. 5E and F). Terazosin treated group had signi cantly lower area fractions and cell density of activated microglia in both neocortex and hippocampus than saline treated group (Fig. 5E and F). Overall, the above ndings suggested that terazosin could ameliorate neuroin ammation, tau hyperphosphorylation and neurodegeneration in brains of AD mice.
Terazosin reduced Aβ production through inhibiting GSK3β/BACE1 signaling pathway To explore the mechanism underlying reduced Aβ production by inhibition of α1-ARs, SH-SY5Y-APP695 cells, which overexpress human APP, were treated with 50µM terazosin for 4 h. Western blot showed that the terazosin did not affect the expression of the full-length APP, but decreased the ratio of CTF-β to CTFα (Fig. 6A). We also measured Aβ40 and Aβ42 levels in cell lysates and cell medium, and found that terazosin treatment decreased the levels of both Aβ40 and Aβ42 in cell lysates and cell medium (Fig. 6A).
In addition, the expression of BACE1, but not LRP1 and PS1, was signi cantly decreased after terazosin treatment ( Fig. 6B and C), indicating that the reduction of Aβ in SH-SY5Y-APP695 cells after terazosin treatment might be resulted from decreased expression of BACE1. Furthermore, we also found that terazosin increased the exression of pGSK3β, which could inhibit the activity of GSK3β and result in decreased BACE1 expression (Fig. 6D).

Discussion
In this study, we found that both genetical knockdown and pharmacological inhibition of α1-ARs alleviated behavioral de cits and pathological changes in an AD mouse model. Mechanistically, we suppose that inhibition of α1-ARs may exert neuroprotective effects against AD by inhibiting GSK3β/BACE1 signaling pathway, thus reducing Aβ production and subsequent neurodegeneration in the brain.
NE signaling plays a critical role in maintaining cognitive functions, but the noradrenergic system is suggested to be dysfunctional in the AD brain (3). There is profound neurodegeneration in the LC with AD progressing, but its pathological signi cance has not been thoroughly addressed to date. Enhancement in neuronal sprouting to different brain regions and NE production (4,9,15), and reduction in NE reuptake (16, 17) might compensate for LC neuronal loss to maintain NE levels in the AD brain. As a consequence of these compensatory changes, extracellular NE is suggested to be increased in the AD brain (18, 19).
Furthermore, the increase in extracellular NE has been shown to be associated with the severity of AD (20,21), with higher NE levels associating with more severe cognitive impairment. In addition, densities of α1adrenertic receptors are suggested to be increased in the AD brain (4,10). These ndings may imply that although NE is important for maintaining cognitive functions, over-compensation of NE during the process of AD could be detrimental. This notion is supported by studies that have shown cognitive impairment in response to stress, because under stress conditions NE release is elevated (22,23). Therefore, inhibition of the NE signaling pathway in the AD brain might be helpful for controlling the dysregulated NE system and aid in rescuing cognitive impairment. Previous studies have shown that signaling of α2 and β2-adrenergic receptors is detrimental, and inhibition of these adrenergic receptors improves behavioral de cits and AD-type pathologies in AD mouse models (6, 7). Although activation of the α1-ARs signaling pathway improves cognitive in non-AD animals (24-26), blockade of α1-ARs by prazosin also improves the cognitive de cits of an AD transgenic mouse model (27). In AD patients, activation of α1-ARs is suggested to contribute to the agitation behaviors (26), and prazosin can alleviate such symptoms (28). These results indicate that the activation of α1-ARs by NE may contribute to agitation and aggressive behaviors, which are commonly seen in AD patients. Mice overexpressing α1-ARs had enhanced expression of genes involved in apoptosis and neurodegeneration (29), further supporting the possible pathological effects of α1-ARs in the AD brain. In accordance with this hypothesis, we found in this study that genetical and pharmacological inhibition of α1-ARs alleviated behavioral cognitive de cits and attenuated AD-type pathologies, including Aβ deposition, Tau hyperphosphorylation, neuroin ammation and neuronal degeneration et al.
In this study, we found that terazosin inhibited GSK3β pathway in vitro, as re ected by increased pGSK3β expression after terazosin treatment. This nding is consistent with a previous study which found that doxazosin, another α1-ARs antagonist, also increased the levels of pGSK3β (30). It is well-recognized that activation of GSK-3β could increase the expression and activity of BACE1, thus promoting the amyloidogenic pathway (31,32). Indeed, we found in this study that inhibition of α1-ARs reduced BACE1 expression both in vivo and in vitro, thus reduced Aβ production. Therefore, inhibition of the α1-AR pathway might attenuate Aβ pathologies through promoting ser9 phosphorylation of GSK3β. Interestingly, it is suggested that Aβ could in return activate α1-ARs (33), therefore, α1-ARs activation and Aβ accumulation during the pathogenesis of AD may form a vicious circle, which could be disrupted by inhibition of α1-ARs. Furthermore, a study found that terazosin rescues pathologies of Parkinson's disease through enhancing glycolysis in the brain (34). Energy metabolism dysfunction and reduced ATP levels in the brain are common features of AD. Impaired glucose metabolism could signi cantly promote the expression of BACE1 (35). Therefore, we suppose that inhibition of α1-ARs signaling pathway may also reduce brain amyloid burden through improving energy metabolism.
In this study, both genetical and pharmacological inhibition of α1-ARs was found to reduce NFT formation and suppress Tau hyperphosphorylation at several sites. GSK3β activation also promotes Tau hyperphosphorylation, and ser9 phosphorylation of GSK3β inhibits this process (36), thus downregulation of Tau hyperphosphorylation after inhibiting α1-ARs might be directly associated with its effect on promoting ser9 phosphorylation of GSK3β. Besides, amelioration of Aβ pathology might also contribute to reduced NFT formation and Tau hyperphosphorylation (37). Furthermore, other pathologies subsequent to Aβ or Tau, including cerebral amyloid angiopathy (CAA), overactivation of glia cells, downregulation of synaptic proteins, dendritic damage and neuronal apoptosis et al. were also improved after inhibiting α1-ARs signaling pathways. We suppose that inhibition of α1-ARs may act on Aβ and suppress its toxicity of triggering subsequent pathological changes.

Conclusion
Our ndings prompt a possibility that over-compensation of the α1-ARs system in the AD brain might be detrimental, thus suitable down-regulating the expression and pharmacological inhibition of α1-ARs may be bene cial for AD. Furthermore, terazosin is a clinically used drug which has validated tolerability; thus, it may hold promise for the treatment of AD.

Abbreviations
Alzheimer's disease (AD); Amyloid-beta (Aβ); intracellular neuro lament tangles (NFT); locus coeruleus (LC); norepinephrine (NE); α1 adrenergic receptors (α1-ARs); neprilysin (NEP); receptor for advanced glycation end products (RAGE); lipoprotein receptor-related protein 1 (LRP1); glycogen synthase kinase 3 beta (GSK3β); synaptosomal-associated protein 25 (SNAP-25); vesicle-associated membrane protein 1 (VAMP1); postsynaptic density protein 95 (PSD95); Tris-buffered saline containing 0.5% Tween 20 (TBST); glial brillary acidic protein (GFAP); microtubule associated protein (MAP)-2; sodium dodecyl sulfonate (SDS); Adeno-associated virus (AAV); Declarations Figures Figure 1 Knockdown of α1-ARs decreased brain Aβ burden in AD mice. A. Immunostaining and quanti cation of Aβ plaques stained with 6E10 and Congo red in the neocortices and hippocampi of AD mice treated with AAV-GFP or AAV-α1-ARs. B. Comparison of Aβ40 and Aβ42 levels in TBS and SDS fractions of brain homogenates between the AAV-GFP and AAV-α1-ARs group. C-E. Representative western blots and quantitative analysis of the levels of APP-metabolizing and Aβ-degrading enzymes as well as Aβ-transporting receptors brain homogenates. n=7-8 per group; * indicates P<0.05 and ** indicates P<0.01. Scale bar in A is 500 µm. Data are represented as mean ± SEMs. Knockdown of α1-ARs attentuated the AD-related pathology in AD mice. A. Immunostaining of astrocytes and microglia with GFAP and CD68 in the neocortices and hippocampus of AD mice treated with AAV-GFP or AAV-α1-ARs. B. Representative western blot images and quanti cation of phosphorylated Tau protein and total Tau protein levels in brain homogenates. C. Representative images and quanti cation of synaptic proteins (PSD95, Synapsin-1, SNAP25 and VAMP1). n=7-9 per group; * indicates P < 0.05; ** indicates P < 0.01. The scale bars in A are 500 µm. Data are represented as mean ± SEMs.  Terazosin treatment attentuated brain Aβ burden in AD mice. A. Immunostaining and quanti cation of Aβ plaques stained with 6E10 and Congo red in the neocortices and hippocampus of AD mice treated with saline or terazosin. B. Representative images of immunostaining of 1A4 (green) and 6E10 (red) as well as the number of cerebral amyloid angiopathy (CAA) number visualized by Congo red staining. C.
Comparison of Aβ40 and Aβ42 levels in the TBS and SDS fractions of brain homogenates between the saline and terazosin group. D and E. Representative western blots and quantitative analysis of the levels of APP-metabolizing proteins and Aβ-degrading enzymes as well as Aβ-transporting receptors brain homogenates. n=7-8 per group; * indicates P<0.05 and ** indicates P<0.01. Scale bar in A is 500 µm and in B is 25µm. Data are represented as mean ± SEMs.

Figure 5
Terazosin treatment attentuated AD-related pathogenesis. A. Representative immunostaining images of NeuN&Tau231, NeuN&MAP2 and NeuN&Caspase3. B. Representative western blots and quantitative analysis of phosphorylated Tau protein (PS199 and PS396) and total Tau protein levels in brain homogenates. C. Representative images and quanti cation of synaptic proteins (PSD95, Synapsin-1, SNAP25 and VAMP1). D. Quanti cation of neurons (NeuN) and dendrites (MAP2) in the CA1 region of the hippocampus and neural apoptosis (Caspase3) in the CA3 region of the hippocampus. E. Representative immunostaining images of CD68&NeuN and GFAP&NeuN. F. Immunostaining of astrocytes and microglia with GFAP and CD68 in the neocortices and hippocampus of mice treated with saline or terazosin. n=6-9 per group; * indicates P<0.05; ** indicates P<0.01. The scale bars in A and E are 50 µm. Data are represented as mean ± SEMs.