Clemastine Attenuates AD-like Pathology in an AD Model Mouse via Enhancing mTOR-Mediated Autophagy

Background: Alzheimer’s disease (AD) is a neurodegenerative disorder with limited available drugs for treatment. Enhancing autophagy attenuates AD pathology in various AD model mice. Thus, development of potential drugs enhancing autophagy may bring benecial effects in AD therapy. Methods: In the present study, we showed clemastine, a rst-generation histamine H1R antagonist and being originally marketed for the treatment of allergic rhinitis, ameliorates AD pathogenesis in APP/PS1 transgenic mice. Chronic treatment with clemastine orally reduced amyloid-β (Aβ) load, neuroinammation and cognitive decits of APP/PS1 transgenic mice as shown by immunohistochemistry and behavioral analysis. We further analyzed the mechanisms underlying the benecial effects of clemastine with using the combination of both in vivo and in vitro experiments. We observed that clemastine decreased Aβ generation via reducing the levels of BACE1, CTFs of APP. Clemastine enhanced autophagy concomitant with a suppression of mTOR signaling. Conclusion: Therefore, we propose that clemastine attenuates AD pathology via enhancing mTOR-mediated autophagy.


Introduction
Alzheimer's disease (AD) is a neurodegenerative disease characterized by de cits in learning, memory, and other cognitive functions. Its major pathological characteristics are the presence of amyloid-β (Aβ) plaques formed by accumulated Aβ and neuro brillary tangles (NFTs) consisted of hyperphosphorylated tau (Lu et al., 2018). Although the pathological mechanisms of AD remain unclear, the accumulation of Aβ, which induces neuroin ammation and impaired synaptic plasticity, are believed as one of key initiators in AD pathogenesis (Lu et al., 2018;Uddin et al., 2018). Reducing accumulation of Aβ or suppressing neuroin ammation is a therapeutic strategy of AD. In this regard, autophagy has drawn great attention due to the following facts: Impaired autophagy has been extensively observed in the brains of AD patients and model mice. Enhancing autophagy attenuates AD pathology including the accumulation of Aβ, impaired synaptic plasticity and neuroin ammation (Hong et  Clemastine (Tavegil™), a rst-generation histamine H1R antagonist, is originally marketed for the treatment of allergic rhinitis. Clemastine contains several hydrophobic functional groups facilitating the crossing of the blood-brain barrier (BBB). Clemastine is currently on phase III clinical trial for patients with relapsing forms of multiple sclerosis due to its recently discovered remyelinating properties (Green et al., 2017;Li et al., 2015). Clemastine displays a bene cial effect in restoring spatial working memory in cuprizone-induced demyelinated mice (Li et al., 2015). Clemastine confers neuroprotection and induces an anti-in ammatory phenotype in SOD1G93A mouse model of amyotrophic lateral sclerosis (ALS) (Apolloni et al., 2016). Clemastine also exhibits a therapeutic effect in depression model mice via HeLa cells, HEK293 cells stably expressing human swedish mutant APP (HEK-APP), ATG5 -/and ATG5 -/-MEFs cells were cultured in high glucose Dulbecco's Modi ed Eagle Medium supplemented with 10% fetal bovine serum (FBS) (Gibco, USA). Primary neurons were dissected from embryonic day (E) 17 APP/PS1 transgenic mice were cultured in neurobasal (Gibco) medium supplemented with 2% B27, 2 mM Glutamax-1 for 9-10 days. Lipofecamine 3000 (ThermoFisher, USA) was used for transfection according to the manufacturer's protocols.

Western blot analysis
The cultured cells were lysed with an SDS sample buffer (50 mm Tris-HCl (pH 6.8), 10% glycerol, 2% SDS, 0.1% bromophenol blue, and 6% 2-mercaptoethanol) for 10 min on ice. Lysates from the cortex and hippocampus were extracted in lysis buffer (50 mM Tris (pH 7.4), 150 mM NaCl, 1% TritonX-100, 1% sodium deoxycholate, 0.1% SDS) containing protease and phosphatase inhibitor cocktail (Roche, Switzerland). Lysates were subjected to SDS polyacrylamide gel electrophoresis, followed by transferring to a polyvinylidene uoride membrane (Millipore, Darmstadt, Germany). The membrane was blocked in 5% BSA dissolved in Tris buffered saline containing Tween 20 (TBST) (20 mM Tris-HCL pH7.6, 150 mM NaCl, and 0.1% Tween 20) for 1 h at room temperature (RT), and then was incubated with appropriate antibody in 5% BSA diluted in TBST for overnight at 4•C. The membranes were washed with TBST and incubated with HRP-conjugated secondary antibodies (Sigma-Aldrich, USA) for 1 h at RT. The immunoreacted proteins were visualized using ECL Western Blotting Detection Reagents (Pierce, USA). The intensities of the bands were analyzed by Image J.

ELISA analysis
Human Aβ42, Aβ40 and levels were assessed by using sandwich ELISA. Frozen brain tissue was dissolved in TBS, and the soluble supernatant fractions were collected after centrifuged. The insoluble materials were then dissolved with Guanidine Hydrochloride. The soluble and insoluble Aβ were quanti ed using human Aβ42 and Aβ40 ELISA kits (KHB3441 and KHB3481, respectively; Invitrogen) following the manufacturer's instructions. The absorbance was read at 450 nm using a 96-well plate reader. Aβ levels were calculated from a standard curve and normalized to the total protein levels, which were determined by the BCA protein assay kit (ThermoFisher). All values were normalized to wet brain weight.
Tissue samples were then post xed overnight in 4 % paraformaldehyde and then cryoprotected in 30 % sucrose at 4 °C. Brain were cut in 12 μm thickness with a frozen microtome. For immuno uorescence staining, sections were blocked in PBS containing 10% normal goat serum and 0.3 % Triton X-100 for 1 h at room temperature. Brain sections were incubated with primary antibody in blocking buffer overnight at 4℃. After rinsing in PBS, sections were incubated with appropriate secondary Alexa Fluor-conjugated antibodies for 2 h at RT. All sections were counterstained with 4, 6-diamidino-2-phenylindole (DAPI) (Vector Laboratories) after being rinsed with PBS. Images were captured under a Leica confocal microscope.
HeLa cells line were xed in 4% paraformaldehyde for 30 min on ice, followed by blocking in 10% FBS for 1 h at RT. The cells were then incubated with primary antibody in blocking buffer overnight at 4•C. After rinsing in PBS with 0.1% Triton X-100, cells were incubated with appropriate secondary Alexa Fluor conjugated antibodies for 1 h at RT. Images were captured under a Leica confocal microscope.

Image analysis
For quanti cation of GFP + puncta, amyloid plaques, the densities of astrocytes and microglia, the images were converted into 8-bit images, and binarized after subtracting the background noise using Image J software. The numbers of GFP + puncta per cell were counted; The density of Aβ plaques was expressed as the numbers of Aβ plaques per square micrometer; The size of Aβ plaques was quanti ed as the area that was occupied by Aβ plaques divided by total area of the cortex and hippocampus; That GFAP + or Iba1 + signals divided by total area of the cortex and hippocampus was quanti ed as the volume of microglia and astrocytes respectively.

Morris water maze
Mice underwent 4 trials per day. A different starting position was used on each trial. The duration of one trial was 90 seconds. Escape latencies (time spent swimming from start point to the target) and path length (the distance from start point to the platform) before reaching the platform were recorded for 5 consecutive days. The escape latencies in the following training day were analyzed (escape latency in the following day/escape latency in the rst day) and labeled as learning trend. For probe trials, the platform was removed after the last trial of the acquisition period. Mice were tested 24 hours later to assess memory consolidation. The time spent in the target quadrant within 60 seconds was recorded. The latency to the rst target site was measured, and the numbers of platform-site crossovers were recorded.
Novel object recognition Mice were exposed to 2 identical objects for 10 minutes placed in 2 opposite corners of the apparatus 8.5 cm from the sidewall. Ninety minutes after the training session, the animal explored the open eld for 10 minutes in the presence of 1 familiar and 1 novel objects. Location preference = time exploring one of the identical objects/time exploring the identical object pairs × 100%. Recognition index (RI) = time exploring novel object/ (time exploring novel object + time exploring familiar object) × 100%.

Statistical analysis
All statistical analysis was performed using SPSS 20.0. Data were presented as mean ± SEM. All dates were collected from at least three independent experiments or from three mice. Data between multiple groups were analyzed by one-way analysis of variance (ANOVA). Comparisons between two groups were made by independent samples t-test. P< 0.05 was considered for the signi cance level for all analyses. * : P < 0.05; * * : P < 0.01; * * * : P < 0.001.

Results
Chronic clemastine treatment rescues cognitive de cits of APP/PS1 Transgenic Mice We rst examined whether clemastine might attenuate the cognitive de cits in APP/PS1 transgenic mice. APP/PS1 transgenic mice exhibit appearance of Aβ plaques from 3-month old. They develop AD-like pathology such as extensive Aβ plaques, neuroin ammation and cognitive de cits at 7-month old (Zhang et al., 2014). Thus, we have taken a prophylactic strategy as an early invention in which four-month-old APP/PS1 transgenic mice were administered clemastine mixed in the food for 3 months. Clemastine exhibits bene cial effect in social isolation-caused depression model mice and in SOD1-G93A mice, a transgenic mouse model of amyotrophic lateral sclerosis (ALS) at 10 mg/kg/day (Apolloni et al., 2016). Thus, to prove the concept that whether clemastine might attenuate AD pathology, APP/PS1 mice were treated orally with clemastine at 10 mg/kg/day. APP/PS1 mice and age-matched WT mice fed with normal food were used as the control. In Morris water maze test, APP/PS1 mice showed impaired learning, as indicated by the increased escape latencies (Fig. 1A) and swimming distances (Fig. 1B, C) in the consecutive trials compared with WT mice. In contrast, clemastine-treated APP/PS1 mice showed shorter escape latencies (Fig. 1A), and decreased swimming distances (Fig. 1B, C) compared with control APP/PS1 mice, even to a level comparable to WT mice. In the probe trials, clemastine-treated APP/PS1 mice exhibited improved memory retention as indicated that they spent longer time in target quadrant ( Fig. 1D) and swam to cross over the target site more times than control APP/PS1 mice (Fig. 1E), which are comparable to WT mice. The differences among these groups of mice were not due to the distinct swimming capability, since the swimming speed of these groups of mice was similar (Fig. 1F). In novel object recognition tests, no signi cant difference was observed in location preference during the training phase, indicating that the location of the objects does not affect the exploratory behavior of mice (Fig.  1H). In the testing phase, as demonstrated previously, control APP/PS1 mice displayed a reduced recognition index (RI) than WT mice, which was rescued by treatment with clemastine (Fig. 1H). It is noteworthy that clemastine treatment showed no effects on WT mice in the above-mentioned tests ( Fig.  1A-H). These results indicate that the chronic treatment with clemastine rescues cognitive de cits in APP/PS1 mice.
Chronic clemastine treatment attenuates Aβ accumulation in APP/PS1 mice Accumulation of Aβ is the central initiator of AD pathogenesis (Musiek and Holtzman, 2015). We thus examined whether clemastine treatment would decrease Aβ accumulation. The coronal section of the hippocampus and cortex of clemastine-treated APP/PS1 transgenic mice and the control mice were stained with an antibody against Aβ (6E10) ( Fig. 2A, D). Results showed that the numbers (Fig. 2C, F) and size (Fig. 2B, E) of Aβ plaques were decreased in the hippocampus ( Fig. 2A-C) and cortex (Fig. 2D-F) of clemastine-treated APP/PS1 transgenic mice, compared to that in control APP/PS1 transgenic mice.
These results indicate that chronic treatment with clemastine decreases the densities of Aβ plaques. We further examined whether the reduced densities of Aβ plaques were due to decreased Aβ concentrations in the brain. ELISA analysis showed that the levels of both soluble and insoluble Aβ42 and Aβ40 in the cortex and hippocampus of clemastine-treated APP/PS1 transgenic mice decreased, compared with those in control transgenic mice. These results indicate that chronic treatment with clemastine decreases accumulation of Aβ.
Chronic clemastine treatment attenuates neuroin ammation in APP/PS1 mice Neuroin ammation is an essential contributor to the pathogenesis of AD. Microglia and astrocytes, the main types of cells in the in ammatory response in the central nervous system, are activated in the brains of AD patients and AD model mice (Calsolaro and Edison, 2016;Shadfar et al., 2015). Activated microglia and astrocytes accumulate around Aβ plaques and produce pro-in ammatory cytokines and chemokines, which cause synaptic dysfunction and neurodegeneration (Calsolaro and Edison, 2016). Antiin ammatory therapy has therefore been credited as a strategy for reducing the risk or slowing the progression of AD (Shadfar et al., 2015). Clemastine has been reported to decrease microgliosis and expression of microglia-related in ammatory genes in the model mice of ALS (Apolloni et al., 2016). We thus examined whether clemastine treatment might also decrease neuroin ammation in APP/PS1 transgenic mice. The densities of astrocytes and microglia as indicated by the volume of GFAP + (a marker for astrocytes, Fig. 3A, C) and Iba-1 + (a marker for microglia, Fig. 3B, D) cells decreased in both the hippocampus and the cortex of clemastine-treated APP/PS1 transgenic mice. These results indicate that chronic treatment with clemastine attenuates neuroin ammation in the brains of APP/PS1 transgenic mice.
Chronic clemastine treatment decreases β-amyloidosis of APP processing in vivo Chronic treatment with clemastine decreases Aβ accumulation, neuroin ammation and cognitive de cits of APP/PS1 transgenic mice. Among these pathological processes, Aβ accumulation is the upstream cause (Musiek and Holtzman, 2015). We thus examined the mechanisms underlying that clemastine reduces Aβ accumulation. To further con rm that clemastine affects Aβ generation in neurons, primary cortical neurons derived from embryonic 17 days (E17) APP/PS1 transgenic mice were treated with clemastine. The result showed that Aβ40 levels in culture medium were reduced by treatment with clemastine (Fig. 4A), con rming that clemastine can reduce Aβ levels in neurons, which is independent on glial cells. Moreover, the levels of neprilysin and insulin-degrading enzyme (IDE), two enzymes account for Aβ degradation, remained unchanged in the hippocampus and cortex of APP/PS1 transgenic mice upon treatment with clemastine ( Fig. 4B-D), suggesting that clemastine may not being involved in Aβ clearance. Since the ratio of Aβ40/Aβ42, which can be altered by γ-secretase cleavage of APP (Borchelt et al., 1996), remained identical levels in between clemastine-treated and control APP/PS1 transgenic mice (Fig. 4K), indicating that clemastine treatment does not alter γ-secretase activity. We then examined the cleavage of APP by α-/β-secretase. APP is cleaved by α-or β-secretase at the extracellular domain, generating two fragments called α-or β-CTF, respectively (Musiek and Holtzman, 2015). The levels of both α-and β-CTF, but not full-length APP, were decreased in the hippocampus and cortex of clemastine-treated APP/PS1 transgenic mice, compared to that in control transgenic mice (Fig. 4E-H). β-CTF is produced from cleavage of APP by BACE1 (Yan, 2016;Yan et al., 2016). Consistently, treatment with clemastine decreased BACE1 levels in the hippocampus and cortex of APP/PS1 transgenic mice (Fig. 4I, 4J). These results indicate that chronic treatment with clemastine reduces Aβ generation through suppressing cleavage of APP, especially by BACE1.
Clemastine treatment induces ATG5-dependent autophagy Autophagy is a highly conserved catabolic process in which proteins and organelles are engulfed in double-membraned vacuoles called autophagosomes and then transported to lysosomes for degradation. Autophagy plays broad functions in neurodegenerative disease (Nixon, 2013). Both α-/β-CTFs and BACE1 are degraded through autophagy (Tian et al., 2013;Wu et al., 2015). Moreover, cotreatment with H1R antagonist astemizole and histamine induces autophagy (Jakhar et al., 2016), suggesting that histamine signaling is involved in autophagy induction. Thus, to investigate the underlying molecular mechanisms of therapeutic effects of clemastine in AD pathogenesis, we examined the effect of clemastine on autophagy. Microtubule-associated protein light chain 3 (LC3) was used as a marker of autophagy induction, because cytosolic LC3-I is processed to its lipidated LC3-II form upon autophagy induction. LC3-II then locates to newly forming autophagophores and subsequently be present in mature autophagosomes (Wu et al., 2015). Therefore, processing of LC3 and a punctuate LC3 pattern represent the formation of autophagosomes and autophagic responses. HeLa cells transfected with LC3-GFP were treated with either 30 μM clemastine or DMSO. HeLa cells exhibited much more LC3-GFP + puncta 12 h after clemastine treatment (Fig. 5A, 5B), indicating that clemastine induces formation of autophagosomes. In addition, treatment with clemastine increases LC3-II levels in a dose-dependent manner (Fig. 5C, 5D). The increasement of LC3-II by clemastine was further enhanced in presence of chloroquine, a lysosomal inhibitor which inhibits fusion of autophagosomes to lysosomes, indicating that the increasement of LC3-II by clemastine is due to an enhanced autophagic in ux. Consistently, levels of P62, a substrate of autophagy, were decreased by clemastine dose-dependently (Fig. 5C, 5E).
ATG5 is an initial factor in autophagy induction. We then examined whether clemastine induced autophagy in a way dependent on ATG5. ATG5 +/+ and ATG5 -/-MEF cells were treated with clemastine. The results showed that, like in HeLa cells, clemastine treatment increased LC3-II levels (Fig. 5H, 5J), whereas decreasing P62 levels (Fig. 5H, 5J) indicating an enhanced autophagy in ux. In contrast, clemastine failed to do so in ATG5 -/-MEF cells (Fig. 5H-5J), indicating that clemastine induces ATG5dependent autophagy. Consistent with the fact that autophagy is impaired in AD (Nixon, 2013), APP/PS1 transgenic exhibited increased levels of LC3-II. In contrast, chronic treatment with clemastine increased LC3-II (Fig. 5K, 5M) while decreasing P62 levels (Fig. 5K, 5L), as it did in cultured cells. Therefore, these results indicate that clemastine enhances autophagy. However, clemastine failed to increase the LC3-II levels in WT mice, indicating clemastine enhances autophagy in a context-dependent manner.
Clemastine enhances autophagy via the mTOR pathway both in vivo and vitro We further explored molecular mechanisms underlying that clemastine induces autophagy. Target of rapamycin (mTOR) signaling, when being suppressed, is one of central pathways in autophagy induction (Zhu et al., 2019). We thus examined rst whether clemastine could affect mTOR signaling. HeLa cells treated with distinct concentrations of clemastine revealed that clemastine decreased levels of phosphorylated mTOR (p-mTOR, Ser2448) (Fig. 6A, B) and phosphorylated P70S6K (p-P70S6K) (Fig. 6A,  6C) in a dose-dependent manner. In contrast, the levels of total mTOR (Fig. 6A, 6D) and P70S6K (Fig. 6A,  6E) remained unchanged upon treatment with clemastine. Similar results were observed in clemastinetreated APP/PS1 transgenic mice. The levels of p-mTOR and p-P70S6K, but not total mTOR and P70S6K, were decreased in the hippocampus and cortex of clemastine-treated APP/PS1 transgenic mice, compared to control transgenic mice (Fig. 6F-6H). However, clemastine did not alter mTOR signaling in WT mice, suggesting that clemastine suppresses mTOR signaling in an environment-dependent manner. Thus, these results indicate that clemastine suppresses mTOR signaling, a central pathway inducing autophagy.

Discussion
In the present study, we demonstrate that chronic treatment with clemastine rescues cognitive de cits in APP/PS1 transgenic mice by reducing Aβ load and neuroin ammation. We further observe clemastine enhances autophagy concomitant with a suppression of mTOR signaling, which is one of the key negative regulators of autophagy. The present study describes a potential therapeutic role of clemastine in AD via enhancing mTOR-mediated autophagy. . We herein observe that clemastine enhances autophagy in an ATG5-dependent manner. Furthermore, clemastine suppresses mTOR signaling, which is a suppressor of autophagy, in both primary cultured neurons and the brains of APP/PS1 transgenic mice. These results suggest that clemastine enhances autophagy through suppressing mTOR signaling. Clemastine is a rst-generation histamine H1R antagonist. These results are also consistent with the facts that other histamine H1R antagonists induce autophagy (Hu et al., 2012;Steele and Gandy, 2013), suggesting that the antihistamine effect may contribute to the clemastine-enhanced autophagy. However, it is worth noting that in addition to H1R, clemastine acts to antagonize ve more other targets including the ve muscarinic acetylcholine receptor subtypes (Chrm1-Chrm5) (Kubo et al., 1987). Especially, clemastine enhances myelination via antagonizing Chrm1 in oligodendrocyte progenitor cells (OPCs) rather than via other four Chrm subtypes (Chrm2-5) (Mei et al., 2016). Thus, it remains further investigation in future the molecular mechanisms underlying that clemastine induces autophagy and attenuates AD pathology.
Clemastine exhibits bene cial effects in neurological disorders via acting on distinct neural cells. Clemastine suppresses microglia-mediated neuroin ammation, enhances myelination and protects neuronal survival. Through these distinct cellular mechanisms, clemastine exhibits bene cial effects on ALS, depression, anxiety and MS (Apolloni et al., 2016;Green et al., 2017;Li et al., 2015;Su et al., 2018). Consistent with these observations, we herein have observed that clemastine treatment attenuates accumulation of Aβ, neuroin ammation and synaptic loss. In this regard, since accumulation of Aβ is the key initiator of AD pathogenesis, the attenuated neuroin ammation and synaptic loss may be the secondary change caused by reduced Aβ. However, it is worth noting that clemastine enhances autophagy, which plays essential roles in neurodegenerative diseases via keeping cellular homeostasis in different types of cells. For example, microglial autophagy degrades Aβ and regulates the release of in ammatory cytokines (Cho et al., 2014). Defective astrocytic autophagy contributes to oxidative stress and neuroin ammation (Wang and Xu, 2020). Neuronal autophagy is involved in the production, clearance and releasing Aβ (Son et al., 2012). We have herein con rmed clemastine induces autophagy in neurons by suppressing mTOR signaling. This observation is consistent with the facts that clemastine decreases the levels of Aβ, α/β-CTFs and BACE1, all of which are substrates of autophagy (Son et al., 2012;Tian et al., 2013;Wu et al., 2015;Yang et al., 2017), and that the generation of Aβ by APP cleavage is occurred predominantly in neurons (Wu et al., 2015). However, we still cannot exclude the potential contribution of clemastine in glial cells to AD pathogenesis, considering the essential roles of glial autophagy in AD pathology (Cho et al., 2014;Wang and Xu, 2020). It requires further investigation that whether and how clemastine regulates microglia-and/or astrocytes-mediated neuroin ammation and their contribution to AD pathogenesis.

Limitation
Our present study has the following limitation: Although we have described that clemastine, as a novel autophagy inducer, decreases the levels of CTFs and BACE1, thus suppressing Aβ generation in neurons. We did not examine whether and how clemastine regulates glial cells in AD mouse brains. Clemastine is clinical drug used for the treatment of allergic rhinitis, which shows good bioavailability and pharmacokinetics (Schran et al., 1996). However, it still requires further investigation on the biodistribution and pharmacokinetic pro le of clemastine in AD transgnice mice, especially the e cacy of clemastine across BBB. The present study just proves the concept that clemastine exhibits a potentiality in AD therapy. However, to con rm whether clemastine is a potential drug for AD therapy, we should perform more extensive analysis in future such as different does, the period of treatment, even with different AD model mice.

Conclusion
Chronic treatment with clemastine attenuates AD pathology including accumulation of Aβ plaques, neuroin ammation and cognitive de cits. Clemastine reduces Aβ generation by decreasing the levels of CTFs and BACE1, all of which are autophagic substrates. Clemastine enhances autophagy via suppression of mTOR signaling, which is one of the key negative regulators of autophagy. Therefore, the present study describes a potential therapeutic role of clemastine in AD via enhancing mTOR-mediated autophagy. Considering that clemastine is already wildly used in the treatment of allergic rhinitis, the present study highlights that clemastine may be a drug candidate in AD therapy.