The Potassium Channel KCa3.1 Represents a Valid Pharmacological Target for Astrogliosis-Induced Neuronal Impairment in a Mouse Model of Alzheimer’s Disease

Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by progressive decline of cognitive function. Astrogliosis plays a critical role in AD by instigating neuroinflammation, which leads ultimately to cognition decline. We previously showed that the intermediate-conductance Ca2+-activated potassium channel (KCa3.1) is involved in astrogliosis-induced by TGF-β in vitro. In the present study, we investigated the contribution of KCa3.1 channels to astrogliosis-mediated neuroinflammation, using TgAPP/PS1 mice as a model for AD. We found that KCa3.1 expression was increased in reactive astrocytes as well as in neurons in the brains of both TgAPP/PS1 mice and AD patients. Pharmacological blockade of KCa3.1 significantly reduced astrogliosis, microglial activation, neuronal loss, and memory deficits. KCa3.1 blockade inhibited astrocyte activation and reduced brain levels of IL-1β, TNF-α, iNOS, and COX-2. Furthermore, we used primary co-cultures of cortical neurons and astrocytes to demonstrate an important role for KCa3.1 in the process of astrogliosis-induced neuroinflammatory responses during amyloid-β (Aβ)-induced neuronal loss. KCa3.1 was found to be involved in the Aβ-induced activated biochemical profile of reactive astrocytes, which included activation of JNK MAPK and production of reactive oxygen species. Pharmacological blockade of KCa3.1 attenuated Aβ-induced reactive astrocytes and indirect, astrogliosis-mediated damage to neurons. Our data clearly indicate a role for astrogliosis in AD pathogenesis and suggest that KCa3.1 inhibition might represent a good therapeutic target for the treatment of AD. Highlights: (1) Blockade of KCa3.1 in APP/PS1 transgenic mice attenuated astrogliosis and neuron loss, and an attenuation of memory deficits. (2) Blockade of KCa3.1 attenuated Aβ-induced indirect, astrogliosis-mediated damage to neurons in vitro via activation of JNK and ROS.


INTRODUCTION
Alzheimer's disease (AD) is a neurodegenerative disorder characterized by progressive decline of cognitive function. The proposed mechanisms of cognitive impairment include synaptic dysfunction triggered by β-amyloid (Aβ), neuronal death, oxidative stress, tau pathology, and glutamate excitotoxicity. It is widely acknowledged that reactive gliosis plays a significant role in the progression of AD. Accumulation of reactive astrocytes and activation of microglia in affected brain regions are apparent in both AD patients and the majority of transgenic rodent models of AD. Garwood et al. (2011) reported that reactive astrogliosis is involved in regulating Aβ-induced neurotoxicity and tau phosphorylation. Reactive astrocytes release a variety of cytokines and pro-inflammatory mediators, which activate intracellular signaling pathways including extracellular signalingrelated kinase (ERK), c-Jun N-terminal kinase (JNK), protein kinase C and PI3 kinase (Matos et al., 2008).
It has been reported that brief (5-15 min) application of Aβ to mixed disassociated hippocampal cultures evoked Ca 2+ influx and up-regulated reactive oxygen species (ROS) in astrocytes, but not in neurons, whereas longer (24 h) exposure led to cell death, especially in neurons (Abramov et al., 2003). Although these data indicate that the primary target of Aβ might be astrocytes and that Aβ-induced oxidative stress and calcium overload in astrocytes may lead to neuronal death, the mechanism of these effects is not well understood.
In immune cells, airway smooth muscle cells and fibroblasts, the intermediate-conductance Ca 2+ -activated potassium channel KCa3.1 regulates membrane potential through K + efflux, thus facilitating Ca 2+ entry (Toyama et al., 2008;Di et al., 2010;Yu et al., 2013;Yi et al., 2016a). In the central nervous system (CNS), KCa3.1 plays a role in neuroprotection following traumatic brain injury, stroke and spinal cord injury (Bouhy et al., 2011). Up-regulation of KCa3.1 has been detected in reactive astrocytes in a mouse model of spinal cord injury. We have previously reported that blockade of KCa3.1, or gene deletion, attenuated TGF-β-induced astrogliosis by regulating intracellular Ca 2+ in primary astrocytes (Yu et al., 2014). Yi et al. (2016a) reported that KCa3.1 is involved in scratch-induced migration of reactive astrocytes mediated by the JNK/c-Jun pathway. More recently, we demonstrated that blockade of KCa3.1 in senescenceaccelerated mouse prone 8 (SAMP8) mice resulted in a decrease in astrogliosis and, moreover, an attenuation of memory deficits (Yi et al., 2016b). Evidence was found for astrogliosis in AD patients with mild cognitive impairment in a positron emission tomography study (Carter et al., 2012). Most recently, Schöll et al. (2015) reported that measure of astrogliosis in autosomal dominant AD could be observed decades before symptom onset using multi-tracer positron emission tomography, possibly coinciding with early fibrillar Aβ plaque deposition. Although reactive astrogliosis has long been recognized as a pathological feature of AD, the role of astrogliosis in the process of cognitive decline in AD is still poorly understood.
In the present study, we showed that blockade of KCa3.1 attenuated Aβ-induced changes in the biochemical profile of reactive astrocytes, including activation of JNK MAP kinase (MAPK) and production of ROS. An experimental KCa3.1 blocker, TRAM-34, attenuated Aβ-induced, indirect, astrogliosis-mediated damage to neurons in vitro, decreased astrogliosis and neuronal loss, and attenuated memory deficits in APP Swe /PS1 A246E transgenic (Tg APP/PS1 ) mice in vivo.

Brain Autopsy Material
Human brain sections were collected from six AD cases and six control cases (Netherlands Institute for Neuroscience, Amsterdam, Netherlands). Written informed consent for a brain autopsy to be used for research purposes after death had been obtained by the Netherlands Brain Bank.

Animals
Nine-month male transgenic mice (the Jackson Lab, no. 003378, Tg APP/PS1 ) were purchased from the Jackson lab. Mice were divided into the following three groups: littermate wild type (WT) mice (n = 20), transgenic (Tg) mice with vehicle treatment (n = 20) or TRAM-34 (120 mg/kg, intraperitoneal, Apptec, Wuxi, China) treatment (Tg + TRAM-34, n = 20). After 4 weeks of drug treatment, the mice were submitted to behavioral testing, and then the mice were euthanized and brain tissues were collected for Western blotting or histology analyses. The protocol of animal experiments was approved by the Animal Experimentation Ethics Committee of Shanghai Jiao Tong University School of Medicine.

Morris Water Maze Test
A modified Morris water maze test was carried out as previously described (Morris, 1984). The test requires the animals to find a visible or hidden platform in a large pool of opaque water and involves 5 days of training sessions (1 day with the platform visible and 4 days with the platform hidden), followed by a spatial probe trial, with no platform present, on day 6. During the hidden platform training, the mice were able to swim freely for 60 s to find a platform 1 cm below the water surface. During the spatial probe trial, the directness of the route taken to the area where the platform was previously located, together with the percentage of total time spent in this quadrant of the pool, was recorded using a video tracking system (Jiliang Software Technology Co., Ltd., Shanghai, China).

Open Field Test
The open field test was carried out as described previously (Wu et al., 2016). Briefly, the mouse was gently placed in the center of an open field chamber (40 cm × 40 cm × 40 cm) and was allowed to move freely for 5 min. The movement parameters of the mouse were monitored and analyzed via a video camera connected to a tracking system (Jiliang Software Technology Co., Ltd., Shanghai, China). After each test, the floor of the open field was cleaned with solution of 70% ethanol to hide animal clues. The ratios of distance and time in the center were measured.

Immunostaining and Data Analysis
The mice were anesthetized with chloral hydrate and perfused with 4% paraformaldehyde. The brain tissues were collected and cryoprotected with 30% sucrose in 0.1 M phosphate buffered saline (PBS). Brain sections (12 µm thick) were blocked with 3% bovine serum albumin in 0.1 M PBS for 1 h at room temperature and were then incubated with the following primary antibodies: mouse anti-KCa3.1 (1:200, Santa Cruz), rabbit anti-GFAP (1:1000; DAKO, Glostrup, Denmark), rabbit anti-NeuN (1:100, Millipore), rabbit anti-Iba1 (1:500, Wako) at 4 • C overnight. Brain sections were labeled with either Alexa Fluor 488-or 568-conjugated anti-mouse or rabbit IgG (1:1000, Invitrogen). A TCS SP8 confocal laser scanning microscope (Leica, Germany), equipped with an argon-ion laser source, was used to capture images using excitation wavelengths of 405, 488, and 568 nm for DAPI, Alexa 488 and Alexa 568, respectively. Using the same reference position for each brain slice, between three and five random 0.01 mm 2 microscopic fields were selected for quantification. Quantification was carried out in six slices of each brain (120 µm intervals), using immunoreactivity for GFAP, Iba1, and NeuN expression. The numbers of GFAP + , Iba1 + , and NeuN + positive cells were counted in six slices per mouse in a blinded manner, using Leica LAS AF Lite software to measure the areas.

Senile Plaque Staining
For senile plaques staining, coronal sections (12 µm) were cut using a Leica cryostat (Leica CM1850). Primary 6E10 antibody (SIG-39300, Covance, Princeton, NJ, USA) was used on the sections overnight at 4 • C. Sections were incubated with biotinylated secondary antibody (AK-6602, Vector) for 45 min. Sections were developed using the ABC elite kit (AK-6600, Vector). Image-Pro plus software (Media cybernetics, USA) was used to measure and recorded as the average plaque areas per field. Six slices per mouse were used to count the plaque number in a blinded manner.

Enzyme-Linked Immunosorbent Assay
ELISA was performed using the kit for TNF-α and IL-1β (Rapidbio Labs, Langka Trade Co. Ltd., Shanghai, China). The procedures were conducted according to manufacturer's protocols.

Primary Cultures
Primary cortical astrocyte cultures derived from neonatal (P0-P2) C57BL/6J mice were prepared from mixed glial cultures (10-14 days in vitro) as described previously (Wang et al., 2008). Astrocyte-conditioned medium (ACM) for treating neurons was obtained using the following procedure. Firstly, astrocytes were cultured in Dulbecco's Modified Eagle's medium (DMEM) contained 10% fetal bovine serum. The culture medium (CM) was changed to neurobasal medium with B27 supplement (Invitrogen) after confluent astrocytes were serum-free for 24 h. The NB/B27-based astrocytes were then treated with Aβ 1−42 oligomer (5 µM) for different lengths of time. The CM from the NB/B27-based cells was collected and used immediately.
Neuronal cultures were established by adding cytosine arabinoside (2 µM) to the mixed glial cultures to inhibit proliferation of glial cells. The neurons were cultured for at least 14 days before treatment with astrocytes CM.
Primary mixed cortical cells were isolated from neonatal (P0-P2) C57BL/6J mice and cultured as described previously (Rodriguez-Gonzalez et al., 2016). Briefly, astrocytes were collected and seeded in multiwell plates in DMEM. Four days after plating, neurons were obtained from brain cortices of P0 C57BL/6J mice and seeded on top of the astrocyte layer.

Neurite Outgrowth Assay
Aβ 1−42 oligomers were added to the primary cultured neurons with or without TRAM-34 and then stained with primary antibody microtubule associated protein 2 (MAP2) plus Alexa Fluor 555-conjugated secondary antibody and DAPI. Cellomics KineticScan reader was used to scan the MAP2 positive cells. Extended Neurite Outgrowth software (Thermo Scientific, Philadelphia, PA, USA) was used to analyze the data.

Statistical Analysis
Statistical significance was analyzed using the Student's ttest, one-way ANOVA followed by Dunnett's post hoc tests or two way ANOVA followed by Bonferroni post hoc test, depending on the case, using Prism software (GraphPad Software, Inc., La Jolla, CA, USA). Results are presented as Mean ± SEM, with p < 0.05 considered to be statistically significant.

KCa3.1 Is Up-Regulated in Reactive Astrocytes of Tg APP/PS1 Mice and AD Patients
We recently showed that KCa3.1 expression is increased in both reactive astrocytes and neurons in the brains of SAMP8 mice, a model that is generally used to investigate the mechanisms of agerelated memory deficits (Yi et al., 2016b). In the present study, expression of KCa3.1 was detected in the brains of Tg APP/PS1 mice at 9 months of age and in AD patients. Age-matched WT littermates and healthy humans were used as controls. In WT mice ( Figure 1A) and healthy humans (Figure 2A), there was little co-localization between KCa3.1 and GFAP + astrocytes. In 9 month old Tg APP/PS1 mice ( Figure 1A) and AD patients (Figure 2A), KCa3.1 was detected in GFAP + hypertrophic astrogliosis. Kaushal et al. (2007) reported that KCa3.1 is expressed in rat neurons in vitro. In the present study, expression of KCa3.1 was shown to be higher in NeuN + neurons of Tg APP/PS1 mice ( Figure 1B) and AD patients ( Figure 2B) than in WT mice and control humans, where expression levels are low. Although there is convincing evidence for the expression of KCa3.1 on microglia in vitro, data supporting expression of KCa3.1 on microglia in vivo are scarce (Kaushal et al., 2007). In this study, rare co-localization of KCa3.1 and Iba1 + microglia was detected in brain slices of both WT and Tg APP/PS1 mice ( Figure 1C) and control humans and AD patients ( Figure 2C).
KCa3.1 was expressed at low levels, mainly in neurons, in WT mice and control humans. In Tg APP/PS1 mice and AD patients, KCa3.1 expression was up-regulated in neurons and was also observed in astrocytes.
intraperitoneally) was used to test whether pharmacological blockade of KCa3.1 would attenuate loss of memory and spontaneous motor activity in Tg APP/PS1 mice, and the dosage of TRAM-34 was used as previously described (Bouhy et al., 2011;Chen et al., 2016;Yi et al., 2016b). Nine months old Tg APP/PS1 mice were treated once daily with TRAM-34 (120 mg/kg, intraperitoneally) or vehicle for 4 weeks and the Morris water maze test and open field test were then used to assess memory and motor activity, respectively. In the hidden platform arm of the Morris water maze   = 4), at the age of 9 months. Measurement of (E) TNF-α and (F) IL-1β released by ELISA in homogenated cortex of WT and Tg APP/PS1 mice, with or without treatment with TRAM-34 (n = 4) at the age of 9 months. Data represent mean ± SEM. ## p < 0.01 compared with WT mice. * p < 0.05, * * p < 0.01 compared with Tg APP/PS1 mice.

test, the Tg APP/PS1
+ vehicle group showed memory deficits compared with WT mice. The Tg APP/PS1 + TRAM-34 group showed significantly improved spatial learning and memory compared with the Tg APP/PS1 + vehicle group (p < 0.05, Figure 3A). In the spatial probe trial without an escape platform, the Tg APP/PS1 + TRAM-34 group spent more time and swam for greater distances in the target quadrant than the Tg APP/PS1 + vehicle group (p < 0.05, Figures 3B,C). In the open field test, the Tg APP/PS1 + vehicle group had a decreased ratio distance (p < 0.01, Figure 3D) and ratio time (p < 0.01, Figure 3E) in the center compared to WT mice. In contrast, the Tg APP/PS1 + TRAM-34 group had a significantly greater ratio distance (p < 0.05, Figure 3D) and more ratio time (p < 0.05, Figure 3E) in the center than the Tg APP/PS1 + vehicle group.

Blockade of KCa3.1 Attenuates Gliosis and Loss of Neurons in Tg APP/PS1 Mice
Reactive astrogliosis is a common pathological feature of chronic neurological diseases associated with aging, such as AD (Khakh and Sofroniew, 2015). We have previously shown that inhibition of KCa3.1 reduces the reactive astrogliosis response (Yu et al., 2014). In the present study, both GFAP + -reactive astrocytes and Iba1 + -activated microglia were significantly increased in the Tg APP/PS1 + vehicle group compared with WT mice. The Tg APP/PS1 + TRAM-34 group, however, showed a significantly suppressed astrogliosis response (p < 0.05, Figures 4A,B) and reduced numbers of activated microglia (p < 0.05, Figures 4C,D) compared with the Tg APP/PS1 + vehicle group. The Tg APP/PS1 + TRAM-34 group also showed reduced neuronal loss in the cortex (p < 0.05, Figures 4E,F) and CA1 hippocampus (p < 0.05, Figures 4G,H). Significantly more NeuN + neurons were present in the Tg APP/PS1 + TRAM-34 group than in the Tg APP/PS1 + vehicle group. Because dendritic and synaptic damage, as well as neuronal loss, likely play important roles in the process of AD, we questioned whether astrogliosis damages dendrites and synapses. Western blotting demonstrated that treatment with TRAM-34 (120 mg/kg) significantly reduced levels of astrogliosis and increased expression of dendritic marker MAP2 and excitatory synapses marker PSD95, but not the pre-synaptic marker synaptophysin, in Tg APP/PS1 mice (Figures 4I,J).

Blockade of KCa3.1 Decreases Amyloid Deposition and Inflammatory Cytokine Production in Tg APP/PS1 Mice
To investigate whether inhibition of KCa3.1 channels can attenuate amyloid accumulation, cerebral senile plaques in the different groups of mice were visualized using 6E10 staining. Senile plaque formation in both cortex and hippocampus was significantly increased in the Tg APP/PS1 + vehicle group compared with WT mice (p < 0.01, Figures 5A-D). After treatment with TRAM-34 (120 mg/kg, i.p.) for 4 weeks, Tg APP/PS1 mice showed decreased senile plaque formation. Quantification showed that senile plaque formation was decreased in the cortex (p < 0.01; Figure 5B) but not in the hippocampus (p > 0.05; Figure 5D) in the Tg APP/PS1 + TRAM-34 group compared with the Tg APP/PS1 + vehicle group. Given the contribution of gliosis to Aβ production, the decrease in senile plaque formation seen in the cortex of Tg APP/PS1 + TRAM-34 group is likely attributable to a reduction in numbers of reactive astrocytes and active microglia, brought about by inhibition of KCa3.1 channels.
Neuronal death in AD is largely caused by the release of pro-inflammatory mediators during gliosis (Scuderi et al., 2014). Blockade of KCa3.1 inhibited the up-regulation of iNOS (Figures 6A,B) and COX-2 (Figures 6C,D), compared with vehicle-treated Tg APP/PS1 mice. Levels of IL-1β and TNF-α in brain homogenates from the different groups were measured by ELISA experiments (Figures 6E,F). Levels of both IL-1β and TNF-α released were attenuated in the Tg APP/PS1 + TRAM-34 group compared with the Tg APP/PS1 + vehicle group. These data demonstrate that KCa3.1 is involved in activation of glial cells in a mouse model of AD.
The molecular probe CM-H 2 DCFDA, which detects intracellular ROS levels, was used to determine whether KCa3.1 contributed to ROS production via Aβ-induced astrogliosis. After treatment with Aβ oligomer (5 µM) for 30 min, a rapid transient increase in the DCF signal was detected ( Figure 8G). ROS production at 30 min was inhibited by TRAM-34 (p < 0.01, Figure 8H). These data demonstrate that KCa3.1 is involved in reactive astrocytosis and oxidative stress via the JNK signaling cascades.

Blockade of KCa3.1 Attenuated Indirect Aβ-Induced Neurotoxicity Mediated by Astrocytes
It has been reported that astrocytes accelerate Aβ-induced neurotoxicity (Garwood et al., 2011). To determine whether or not KCa3.1 is involved in astrocyte-accelerated Aβ-induced neurotoxicity, neurons, astrocytes and mixed cultures (neurons with astrocytes) were treated with 5 µM Aβ oligomers and then cell viability was measured in the different groups at 48 or 72 h (Figures 9A-E). Treatment with 1 µM TRAM-34 alone did not significantly change the viability of any of these cell cultures. Pre-treatment of mixed cultures, but not neuronal cultures, with TRAM-34 before treatment with Aβ increased cell viability at both 48 and 72 h (Figures 9A,C), suggesting that TRAM-34 might reduce Aβ-induced neurotoxicity by suppressing astrogliosis. A significant decrease in cell viability was observed only after 72 h in Aβ-treated neuronal cultures (Figures 9B,D). Aβ oligomers did not cause decreased viability from astrocytes at 72 h (Figure 9E), indicating that the phenotype switch of astrogliosis significantly accelerates Aβ-induced neurotoxicity.
We next asked whether blockade of KCa3.1 might attenuate neurotoxicity induced by inflammatory mediators during astrogliosis. Pre-treatment with TRAM-34 (1 µM) was found to rescue neuronal viability in cultures treated with astrocytes CM (Figure 10A). As a negative control, neuronal viability was not significantly affected by cell-free astrocyte CM that had been incubated for 48 h with Aβ (5 µM; Figure 10B). Data represent mean ± SEM of cell viability relative to control (n = 4 cultures). * p < 0.05, * * p < 0.01, * * * p < 0.001.
Because damage to dendrites and synapses occurs before neuronal loss during the development of AD (Szegedi et al., 2006), we asked whether Aβ-induced astrogliosis-mediated by KCa3.1 could accelerate damage to dendrites and synapses. More signs of dendritic and synaptic damage were observed in neurons treated with Aβ-treated astrocytes CM than in those treated with control CM, as demonstrated by immunofluorescent staining for the dendritic marker MAP2 (Figure 11A). Incubation with Aβ-CM decreased total neurite length ( Figure 11B) and number of branch points ( Figure 11C). Pre-incubation with TRAM-34 (1 and 10 µM) markedly reversed the effect of Aβ-CM by increasing total neurite length and number of branch points (Figures 11A-C). The concentrations of TRAM-34 (1 and 10 µM) were used as previously described (Lallet-Daher et al., 2009;Yu et al., 2014). Western blots showed that blockade of KCa3.1 attenuated Aβ-CM-induced decreases in NeuN, MAP2, and PSD95, but not decreases in the pre-synaptic marker, synaptophysin (Figures 11D,E). Taken together, these results show that KCa3.1 is involved in Aβ-mediated damage to dendrites and synapses by an indirect, astrogliosis-mediated mechanism.

DISCUSSION
The data presented herein demonstrate that blockade of KCa3.1 attenuates neuropathology by regulating neuroinflammation in a mouse model of AD and, moreover, that prevention of astrogliosis might be a promising strategy for the treatment of AD. Using Tg APP/PS1 mice as a model of AD, we have shown that pharmacological blockade of KCa3.1 significantly reduced astrogliosis, neuronal loss, and memory deficits. KCa3.1 blockade FIGURE 11 | Blockade of KCa3.1 attenuated Aβ-induced, indirect, astrocyte-mediated damage to dendrites and synapses. Levels of dendritic and synaptic markers were compared between hippocampal neurons treated with solvent only (control treatment), 5 µM Aβ, Aβ-CM, or CM from astrocyte cultures in which Aβ-induced activation was inhibited for 24 h by TRAM-34 (TRAM-34 + Aβ-CM). A Cellomics KineticScan HCS Reader was used to image the primary cerebral neurons. (A) For better visualization of individual dendrites, dendrites of the sparsely plated neurons were immunostained with MAP2 (red), and nuclei were stained with DAPI (blue). Extended Neurite Outgrowth bioapplication software was used to analyze (B) neurite length and (C) branch point counts. Data represent mean ± SEM (n = 3). * * p < 0.01, * * * p < 0.001 compared with Aβ-CM. ## p < 0.01 compared with control. Aβ-CM treatment significantly reduced neurite length and branch point counts. This reduction was prevented by inhibition of astrocyte activation by TRAM-34 (1 and 10 µM; n = 3). (D) Western blot analysis with the following primary antibodies: neuronal marker NeuN, dendritic marker MAP2, post-synaptic proteins PSD95, and pre-synaptic protein synaptophysin. (E) Results are presented as mean ± SEM (n = 3). # p < 0.05 compared with con-CM. * p < 0.05 compared with Aβ-CM. CM, conditioned medium.
inhibited astrocyte activation and reduced levels of IL-1β, TNFα, iNOS, and COX-2 in the brain. In vitro experiments with primary murine neuronal, astrocytic and mixed cortical cultures exposed to Aβ confirmed that blockade of KCa3.1 improved neuronal survival by reducing astrocyte activation and ROS production.
Deposits of Aβ could induce oxidative stress, glial activation, and local cell loss. In the progression of AD, astrocytes undergo a switch to a reactive phenotype, characterized by profound functional and morphological alterations (O'Callaghan et al., 2014). Up-regulation of the astrogliosis marker GFAP is a wellcharacterized example of these phenotypical changes and, in our cellular model, TRAM-34 was found to negatively modulate expression of both GFAP and KCa3.1.
There is a large body of data that implicates MAPK signaling pathways, especially JNK pathways, in oxidative stress (Matos et al., 2008). In the present study, blockade of KCa3.1 attenuated phosphorylation of JNK, which is activated by Aβ stimulation. Astrogliosis has been shown to contribute to inflammatory processes in AD through the release of ROS, pro-inflammatory factors, and cytokines (Heneka et al., 2001;Verkhratsky et al., 2010). For example, IL-1β, TNF-α, ROS, and NO have been found to regulate pro-and anti-inflammatory genes, including NOS-2 and COX-2 (Tuppo and Arias, 2005;Wyss-Coray, 2006).
We found that blockade of KCa3.1 attenuated Aβ-induced indirect neurotoxicity and decreased Aβ-mediated damage to dendrites and synapses by an indirect, astrogliosis-mediated mechanism. In the AD mouse model, levels of pro-inflammatory factors and cytokines were significantly up-regulated and these increases were inhibited by TRAM-34.
The tripartite synapses in the CNS are formed by pre-and post-synaptic neuronal compartments and astroglial perisynaptic processes (Verkhratsky et al., 2010). A phenotype switch of astrocytes may lead to abnormal release of gliotransmitters, such as glutamate and GABA, which could lead to synaptic loss, excitotoxicity, and neurodegeneration in AD (Agulhon et al., 2012;Jo et al., 2014). Numerous clinical studies have confirmed a correlation between the degree of dementia and the extent of synaptic loss. In our studies, blockade of KCa3.1 rescued memory deficits, and synapse/neuron loss in Tg APP/PS1 mice. After treatment with TRAM-34, quantification showed that senile plaque formation was decreased in the cortex ( Figure 5B) but not in the hippocampus ( Figure 5D). We propose that the different efficacy with respect to different brain areas is that the hippocampus was the greatest amyloid burden region in mice model of AD (Reilly et al., 2003), and the processes of astrogliosis mostly observed surrounding amyloid plaques also were able to accumulate large amounts of senile plaque (Rodriguez et al., 2009). Reducing astrogliosis is a promising strategy for controlling harmful CNS inflammation in neurodegenerative disorders (Khakh and Sofroniew, 2015;Sofroniew, 2015).
In the present in vivo study, we found that KCa3.1 was up-regulated in reactive astrocytes of Tg APP/PS1 mice and AD patients compared with WT mice and control humans. It has also been reported that KCa3.1 is involved in several aspects of microglial activation in vitro (Kaushal et al., 2007;Maezawa et al., 2011;Ferreira and Schlichter, 2013). Maezawa et al. (2011) reported that KCa3.1 activity is required for Aβ-induced microglial neurotoxicity in vitro. In our study, KCa3.1 channels were rarely co-localized with microglia, although blockade of KCa3.1 reduced both astrogliosis and microglial activation in Tg APP/PS1 mice.
Based on these results, we propose that the expression of KCa3.1 is under the immunohistochemically detectable levels in microglia of both mice and human. Co-cultures of astrocytes and neurons were established to determine whether or not KCa3.1 is involved in astrocyte-accelerated Aβ-induced neurotoxicity. Our data suggest that TRAM-34 might reduce Aβ-induced neurotoxicity by suppressing astrogliosis.
The mechanism of the Aβ-induced phenotypic switch of astrocytes and its potential role in the progression of AD provide excellent opportunities for novel therapies such as TRAM-34, a small molecule blocker of KCa3.1. We have provided evidence that KCa3.1 regulates Aβ-induced astrogliosis and neurotoxicity. The anti-inflammatory and neuroprotective properties of KCa3.1 blockers have been demonstrated in many animal models, including traumatic brain injury, retinal ganglion cell degeneration, and multiple sclerosis (Mauler et al., 2004;Reich et al., 2005;Kaushal et al., 2007;Yi et al., 2016b). Our data show that KCa3.1 is also a promising target for reducing inflammatory damage in AD.

AUTHOR CONTRIBUTIONS
ZY supervised the entire project, designed research, and wrote the paper. HC conceived and designed the experiments, interpreted and analyzed data, supervised all the experimental procedure. TW and MY conceived and designed the experiments, performed research interpreted and analyzed data. QL performed research and analyzed data. WG and LH analyzed data and critically revised the manuscript.

FUNDING
This work was supported by Science and Technology Commission of Shanghai Municipality grant 16ZR1418700, National Natural Science Foundation of China grant 81503044.