Effect of Fluvoxamine on Amyloid-β Peptide Generation and Memory

Abstract

Alzheimer’s disease is characterized by abnormal amyloid-β (Aβ) peptide accumulation beginning decades before symptom onset. An effective prophylactic treatment aimed at arresting the amyloidogenic pathway would therefore need to be initiated prior to the occurrence of Aβ pathology. The SIGMAR1 gene encodes a molecular chaperone that modulates processing of the amyloid-β protein precursor (AβPP). Fluvoxamine is a selective serotonin reuptake inhibitor and a potent SIGMAR1 agonist. We therefore hypothesized that fluvoxamine treatment would reduce Aβ production and improve cognition. We firstly investigated the impact of SIGMAR1 on AβPP processing, and found that overexpression and knockdown of SIGMAR1 significantly affected γ-secretase activity in SK-N-MC neuronal cells. We then tested the impact of fluvoxamine on Aβ production in an amyloidogenic cell model, and found that fluvoxamine significantly reduced Aβ production by inhibiting γ-secretase activity. Finally, we assessed the efficacy of long-term treatment (i.e., ∼8 months) of 10 mg/kg/day fluvoxamine in the J20 amyloidogenic mouse model; the treatment was initiated prior to the occurrence of predicted Aβ pathology. Physical examination of the animals revealed no overt pathology or change in weight. We conducted a series of behavioral tests to assess learning and memory, and found that the fluvoxamine treatment significantly improved memory function as measured by novel object recognition task. Two other tests revealed no significant change in memory function. In conclusion, fluvoxamine has a clear impact on γ-secretase activity and AβPP processing to generate Aβ, and may have a protective effect on cognition in the J20 mice.

Keywords

Alzheimer’s disease amyloid-β peptide cognition fluvoxamine J20 SIGMAR1

INTRODUCTION

Alzheimer’s disease (AD) is the most common progressive neurodegenerative disorder, affecting approximately 13% of the population at 80 years of age (reviewed in [1]). AD is diagnosed definitively by two key neuropathological features, senile plaques and neurofibrillary tangles [1]. Senile plaques in the extracellular spaces between neurons consist of central cores of amyloid-β (Aβ) peptides and fibrils, surrounded by degenerated neurites and glia. The Aβ peptide consists of 39 to 43 amino acids and is derived by sequential proteolytic cleavage, via a series of enzymes known as secretases, of the amyloid-β protein precursor (AβPP), encoded by the APP gene [1].

Current clinically approved drugs for AD include cholinesterase inhibitors and memantine, a glutamatergic antagonist. These drugs only alleviate symptoms associated with the disease. However, there are next-generation therapies based on modulating the amyloidogenic pathway that include γ-secretase inhibitors and monoclonal antibodies directed against Aβ peptide. Unfortunately, clinical trial data show that blocking Aβ after its accumulation into senile plaques is unlikely to reverse the clinical symptoms of disease (reviewed in [2]). While both vaccine and passive immunization trials have shown that they can reduce amyloid levels in vivo, there has been no significant effect observed on the patient’s dementia status [2]. Researchers hypothesized that the failure of these strategies is due to the fact that abnormalities in amyloid metabolism start 10 to 15 years prior to the onset of symptoms, and act as a trigger for downstream AD-related neuropathology and neuronal loss [3]. Therefore, drugs that inhibit the amyloidogenic pathway would be ineffective after diagnosis of dementia, but could be therapeutically beneficial if given prophylactically.

The Sigma-1 receptor (SIGMAR1) protein is primarily located at the endoplasmic reticulum in the central nervous system and peripheral organs and modulates a variety of cellular functions consistent with a role as a ligand-regulated molecular chaperone [4]. Thus, SIGMAR1 could play a role in altering the localization and biological activity of key neurodegenerative disease molecules, including those involved in the production of Aβ. The association of SIGMAR1 polymorphisms with AD has been equivocal. Studies have focused on the potentially functional polymorphism rs1800866, whose minor allele results in a Gln to Pro amino acid substitution at codon position 2. The Gln2Pro substitution is located within the N-terminus endoplasmic reticulum retention signal of SIGMAR1, which may affect the subcellular transport and localization of the molecule [4]. Fluvoxamine is a selective serotonin reuptake inhibitor (SSRI) and has been in clinical use since 1984 to treat a variety of psychiatric disorders [5]. Of interest, fluvoxamine is also a potent ligand for the SIGMAR1 protein with a Kd of 36 nM [5].

In this study, we examined the role of the SIGMAR1 polymorphism rs1800866 and fluvoxamine on γ-secretase activity and Aβ secretion in cellular models. We then examined the behavioral effects of long-term chronic treatment with fluvoxamine in the J20 amyloidogenic mouse model which overexpresses the human APP gene containing both the Swedish (K670N/M671L) and Indiana (V717F) mutations (APPSwInd) [6].

MATERIALS AND METHODS

Materials

Commercially available sigma-1 ligands fluvoxamine maleate (F2802) and NE-100 (4-methoxy-3-(2-phenylethoxy)-N,N-dipropyl-benzeneethanamine hydrochloride, SML0631) were obtained from Sigma-Aldrich (USA). NE-100 is a high affinity antagonist of the SIGMAR1 receptor, but with >2000 fold lower affinity for D1, D2, 5-HT1A, 5-HT2 and phencyclidine (PCP) receptors [7].

SIGMAR1 expression constructs

A full-length wild type (wt) SIGMAR1 cDNA expression construct (pcDNA-SIGMAR1 wt) was made as described previously [8]. The wt sequence for rs1800866 corresponds to the major A allele. The minor C allele of rs1800866 was introduced into the SIGMAR1 cDNA by site-directed mutagenesis (QuikChange Site-Directed Mutagensis kit, Agilent Technologies, CA, USA) according to manufacturer’s instructions. A Stealth RNAi oligonucleotide (HSS145543, Invitrogen) was used to knockdown endogenous SIGMAR1 gene expression in a cell line. The RNAi has been previously validated to knockdown SIGMAR1 gene expression in SK-N-MC cells [8]. A high GC RNAi control (12935-400, Invitrogen, CA, USA) was used as negative control.

Gamma-secretase luciferase reporter assay

In addition to the SIGMAR1 expression constructs or fluvoxamine, the two reporter constructs comprising MH100 and C99-GVP plasmids [9] was used to assay for a specific cleavage activity within a fusion protein containing the AβPP derived γ-secretase site, and is ideal for the examination of the effects of exogenous genes or small molecules on this enzymatic activity. The reporter plasmids were co-transfected with either the SIGMAR1 cDNAs or RNAi into SK-N-MC (ATCC HTB 10) cells using Lipofectamine 2000 reagent (Invitrogen) according to manufacturer’s instructions. The cells were plated at a density of 3X 10⁴ cells per well in 24 well plate (Corning Incorporated, NY, USA), and left to recover for 24 h before transfection. When examining the effects of fluvoxamine (46 nM) or NE-100 (25 nM), the small molecules were added immediately after the transfection process. Cells within each well were lysed in 200μl of 1×Glo Lysis buffer (Promega) after 48 h and luciferase activities were assayed using the Readi-Glo reagent (Promega) according to manufacturer’s instructions.

Aβ secretion and quantification of APP gene expression

Chinese hamster ovary (CHO) cells that stably express the human AβPP695 (CHO-AβPP) [10] were a gift from Prof Andrew Hill (La Trobe University, VIC, Australia). To examine AβPP processing, CHO-AβPP cells were solubilized in lysis buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, and protease and phosphatase inhibitors) and equal amounts of protein were electrophoresed on SDS-PAGE gels and transferred onto 0.45μm nitrocellulose membranes. Western blotting of secreted Aβ was carried out as previously described [10]. Briefly, Aβ in the culture medium was separated on 12% SDS-PAGE gels and transferred onto 0.2μm nitrocellulose membranes. Membranes were probed with WO2, a mouse monoclonal antibody against the N-terminus of AβPP (a gift from Prof Colin Masters and Dr Qiao-Xin Li, University of Melbourne, VIC, Australia; diluted 1:1000) followed by rabbit anti-mouse horseradish peroxidase-conjugated secondary antibody. The western blotting of cellular AβPP (CTF-β) was carried out as follows. Protein concentrations were determined using Bradford colormetric protein assay (Bio-Rad, CA, USA) and 40μg of protein were separated on SDS-PAGE gels and transferred onto 0.45μm nitrocellulose membranes at 100 volts for 30 min. Membranes were blocked overnight at 4°C in PBS containing 5% non-fat dry milk and probed with WO2 (1:1000 dilution) at 4°C overnight. The membranes were washed three times in PBS containing 0.1% Tween-20 and then incubated with horseradish peroxidase-conjugated secondary antibody (Dako, Carpinteria, CA, USA, 1/2000 dilution) for 2 h. ECL signal were detected using enhanced chemiluminescence and X-ray films (GE Healthcare). The signal intensity was quantified using NIH ImageJ software.

J20 amyloidogenic mouse model

The generation of the J20 line [JAX Stock No. 006293: B6.Cg-Tg(PDGFB-APPSwInd)20Lms/2Mmjax] has been described previously [6] and was a gift from Prof. Lennart Mucke (Gladstone Institute of Neurological Disease and Department of Neurology, University of California, USA). Genotypes were determined after weaning by tail biopsy and polymerase chain reaction as described previously [6]. All J20 transgenic mice were heterozygous with respect to the transgene and backcrossed to C57BL/6J for >10 generations. C57BL/6J littermates served as non-transgenic controls. Mice were bred and housed in independently ventilated cages (Airlaw, Smithfield, Australia) at Australian BioResources (Moss Vale, Australia). Following transport to the holding facility of Neuroscience Research Australia at the age of ∼13 months, mice were pair-housed in Polysulfone cages (1144B: Tecniplast, Rydalmere, Australia) with minimal environmental enrichment [11]. Mice were kept under a 12:12 h light:dark schedule [light phase: white light (illumination 124 lx); dark phase:red light (illumination <2 lx)]. Food and water were available ad libitum. Behavioural phenotyping commenced at least two weeks after the arrival of the test animals. Research and animal care procedures were approved by the University of New South Wales Animal Care and Ethics Committee in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes.

Fluvoxamine treatment

Plaque pathology has been reported in 50% of the J20 mice by 5 months of age [6]. Therefore, long-term chronic treatment of J20 mice with fluvoxamine was initiated at 5 months of age for 8 months. Fluvoxamine was dissolved in water at a final concentration of 0.038 mg/mL and offered to the treated J20 mice ad libitum. Fresh fluvoxamine was given twice a week. Females mice were used to reflect the higher prevalence of AD in women [12]. Moreover, female transgenic mice of AD tend to accumulate amyloid at an earlier age than age-matched male mice [13]. Based on the average weight of female C57BL/6J mice (23 grams) and liquid consumption per day (6 mL), we estimated that the treatment group received a dose of 10 mg/kg/day of fluvoxamine. There was no significant difference in water consumption between the untreated and treated groups (data not shown).

Behavioral testing

Test animals were J20 transgenic females treated with fluvoxamine (n = 10) or untreated (n = 10) and non-transgenic controls untreated (n = 5). Animals were assessed in a battery of established cognitive tasks [14, 15] using an inter-test interval of at least six days (the least aversive/disruptive cognitive tasks were carried out first). All devices (and objects) were cleaned thoroughly with 70% ethanol in between trials and sessions. Each of the behavioral tests used is described below.

Novel object recognition task

The distinction between familiar and unfamiliar objects is an index of recognition memory, and its measurement is aided by the innate preference of rodents for novel over familiar objects [16]. The novel object recognition task (NORT) was conducted over 3 days. On day 1, mice were placed in an empty Perspex square arena and allowed to explore the arena freely for 10 min in both trials. On day 2, mice were placed in the empty arena for 10 min in trial 1. In the second trial, two identical objects were placed in the center of the apparatus and mice were allowed to explore freely for 10 min. On day 3, mice were exposed to two identical objects (which differed from those on day 2) for 10 min in trial 1 (sample trial), and then one familiar and one novel object for 5 min in trial 2 (test trial). An inter-test interval of 1 h was used to characterize the mice for short-term memory. The frequency and duration of nosing the objects were recorded offline using Any-Maze™ tracking software (Stoelting Co., Wood Dale, USA). The percentage of time spent nosing the novel object was calculated using [(novel object time / time for both objects)×100].

Y-maze

The version of Y-maze used for this study assesses short-term memory of the familiarity to a specific context [17]. The apparatus consisted of three arms placed at 120° with respect to each other, and equipped with different internal visual cues. Corn-cob bedding (Able Scientific, Canning Vale, Australia) covered the apparatus floor and was changed in between sessions. The Y-maze test consisted of two trials (training and test), with a 1 h inter-trial interval (ITI). The trial duration for training and test was 10 and 5 min, respectively [18]. During training, one arm was blocked off (novel arm); mice were placed facing the end of one of the other two accessible arms (start arm). In the test trial, all arms were accessible, and mice were placed facing the end of the start arm then allowed to explore the apparatus freely. Time, entries and distance travelled in arms was recorded using Any-Maze™. An arm entry was scored whenever an animal entered an arm with more than half of its body length. The percentage of novel arm time was calculated using [(novel arm time/total arm time)*100]. The corresponding calculations were performed for novel arm distance travelled and novel arm entries.

Cheeseboard

The cheeseboard paradigm was employed as a less stressful dry-land equivalent of the Morris Water Maze [19]. It measures spatial reference memory. Mice were trained to find a food reward over a number of days; spatial reference memory was indexed by a decreased latency (and distance) to find the reward over days. A camera was mounted above the cheeseboard to measure distance travelled, latency to find the reward as well as time spent in cheeseboard zones using Any-Maze™. Latency to find the target was measured using a stopwatch. During habituation (three days to the blank side of the cheeseboard) three 2 min trials were conducted each day with a 10 min inter test interval. Mice were food-restricted one day prior to habituation and kept at 85–90% of their pre-test body weight throughout testing by feeding mice for 1-2 h per day post training/testing. The cheeseboard test was carried out as previously described [20].

Biochemical analysis of brain tissue

Triton X-100 detergent-soluble protein fractions were isolated from frozen frontal cortical brain tissue of J20 mice. Briefly, 100 mg of tissue was homogenized in Tris-buffered saline (TBS: 50 mM Tris.HCl (pH 7.4), 150 mM NaCl, 1 mM PMSF, 1X complete protease inhibitor cocktail (Complete Mini, Roche Applied Science, Penzberg, Germany)) in microcentrifuge tubes on ice. The lysates were centrifuged at 16,000 g for 20 min at 4°C, and the supernatant was saved as the soluble protein fraction. The pellets were resuspended in TBS-T buffer (50 mM Tris.HCl (pH 7.4), 150 mM NaCl, 1 mM PMSF, 1X complete protease inhibitor cocktail and 1% Triton X-100). The lysates were centrifuged at 16,000 g for 20 min at 4°C, and the supernatants were saved as Triton X-100-soluble protein fractions. For western blot analysis of total AβPP levels, approximately 25μg of Triton X soluble protein were electrophoresis on a 10% SDS-PAGE gel and transferred to a nitrocellulose membrane (Trans-blot transfer medium, Bio-Rad) and probed with WO2 antibody (diluted 1:1000). Densities of chemiluminescence bands were quantified using the Bio-Rad Chemidoc system. Differences in protein levels between samples were normalized using β-actin levels (MAB1501, Millipore, diluted 1:5000). For ELISA of full length Aβ species, approximately 100μg of Triton X-100 soluble protein was used to assay for Aβ₄₀ and Aβ₄₂ using the Aβ₄₀ (KHB3481) and Aβ₄₂ (KHB3441) Human ELISA Kits (Life Technologies-Thermo Fisher Scientific Australia, VIC, Australia) according to manufacturer’s instructions.

Statistical analyses

All cellular model experiments were performed with n = 5 biological replicates. Comparisons between control and treatment groups from cell culture and biochemical analyses were performed using Student’s t-tests. Data from the NORT and Y-maze tests for the three experimental groups were examined by Student’s t-tests to determine the effects of the APP mutation and drug treatment (J20 untreated versus J20 treated with fluvoxamine). Repeated measures ANOVA were used for within-subjects factor ‘time’ in the cheeseboard test (i.e., training and task acquisition). Data are presented as means±standard error of the mean (SEM). Effects were regarded as statistically significant if p < 0.05. Analysis was performed using SPSS v. 23 (IBM, Armonk, NY).

RESULTS

SIGMAR1 alters γ-secretase activity

We examined the effects of SIGMAR1 gene expression and knockdown expression in SK-N-MC neuronal cells, using a sensitive luciferase reporter assay [9], which allowed us to measure the specific cleavage of the γ-secretase site of the AβPP reporter construct (Fig. 1A). Transfection with the SIGMAR1 cDNA with the minor C allele of rs1800866 significantly increased γ-secretase cleavage by 1.2-fold (p = 0.008) relative to the A allele (Fig. 1A). Conversely, knockdown of endogenous SIGMAR1 expression by short interfering RNA (RNAi) led to a 4-fold (p = 5.62×10^-11) decrease in γ-secretase-mediated cleavage compared with a scrambled RNAi control (Fig. 1A).

Fig.1

Effect of SIGMAR1 expression and fluvoxamine on γ-secretase-mediated cleavage of amyloid-β protein precursor (AβPP) reporter construct. A) SK-N-MC neuronal cells were assessed for γ-secretase activity after transfection with either SIGMAR1 cDNA corresponding to either the A or C allele of rs1800866, or with RNAi designed to knockdown SIGMAR1 expression. B) The effect of SIGMAR1 ligand fluvoxamine (46 nM) or the specific Sigma-1 antagonist NE-100 (25 nM) was also examined in SK-N-MC neuronal cells. Luciferase activity values from each biological replicate were normalized to the control treatment for each experiment (open columns), so that the luciferase levels of the A allele SIGMAR1 construct, the control RNAi, or the cells that have been treated with vehicle only, were expressed as ‘1’. Mean +/– SEM is indicated, *p < 0.05, **p < 0.001.

Fluvoxamine decreases γ-secretase activity via SIGMAR1 receptor pathway

We examined whether the SIGMAR1 ligand fluvoxamine has a similar impact on γ-secretase activity in two cell-based assays. Firstly, exposure of cells to 46 nM fluvoxamine, a concentration similar to the published SIGMAR1 Kd value for the compound (36 nM) [5], significantly decreased the reporter luciferase-activity by 1.2-fold (p = 0.021) in SK-N-MC cells (Fig. 1B). Addition of 25 nM of the specific SIGMAR1 antagonist, NE-100 [7], abrogated the effect of fluvoxamine on γ-secretase activity in SK-N-MC cells (Fig. 1B).

Fluvoxamine decreases Aβ secretion in CHO-AβPP cells

We then proceeded to dissect the effects of fluvoxamine on Aβ secretion in CHO cells that stably express the human AβPP695 (CHO-AβPP). This cell line has been used extensively in AβPP processing studies [10]. Fluvoxamine reduced Aβ levels in a dose-dependent manner (Fig. 2A). The signal intensity of the monomeric Aβ detected by western blotting indicated that fluvoxamine at 10μM significantly inhibited Aβ production by 2-fold (p = 0.015) (Fig. 2A). To determine if fluvoxamine had any effect on cell protein levels that could affect Aβ levels, we treated CHO-AβPP cells with increasing concentration of fluvoxamine and assessed the total cellular protein. At 100μM of fluvoxamine, the total protein levels were significantly decreased, which may be due to fluvoxamine’s inhibitory effect on cell growth at this concentration. Conversely, at 10μM of fluvoxamine the total protein levels were not significantly altered (p = 0.439) (Fig. 2B).

Fig.2

Effects of fluvoxamine on AβPP processing and Aβ secretion in CHO-AβPP cells. A) Dose response of fluvoxamine. Western blot analysis (upper panel) and densitometric quantification (lower panel) of monomeric Aβ species in cells treated with increasing doses of the SIGMAR1 ligand. B) Corresponding total protein levels from cells treated with increasing doses of fluvoxamine (fluvox). C) Quantitative RT-PCR of endogenous APP gene transcript levels in cells treated with fluvoxamine (10μM). D) Western blot analysis (upper panel) and densitometric quantification (lower panel) of full-length AβPP and its cleavage product, CTF-β, in cells treated with fluvoxamine (10μM). β-actin was used as an internal loading control. Mean +/– SEM is indicated. *p < 0.05, **p < 0.001.

Fluvoxamine regulates AβPP processing and not production

To determine if the decrease in Aβ generation observed was due to changes in AβPP processing or simply a reduction in AβPP production, we treated CHO-AβPP cells with fluvoxamine at a concentration of 10μM and measured the expression of AβPP gene at both mRNA and protein levels. We found that neither mRNA nor protein levels of AβPP of CHO-AβPP cells were significantly altered by fluvoxamine (Fig. 2C, D), indicating that the decrease in Aβ generation was caused by changes in AβPP processing and not by a reduction in AβPP production. We then measured the level of cellular AβPP C-terminal fragment β (CTF-β) to determine whether the decrease in Aβ generation was due to inhibition of γ-secretase activity. The CTF-β level was significantly increased by 1.6-fold with fluvoxamine (p = 0.017) (Fig. 2D). Increased cellular CTF-β levels, in association with reduced Aβ secretion, indicated that γ-secretase activity was inhibited by fluvoxamine.

Effect of fluvoxamine treatment on learning and memory

We have demonstrated in vitro that fluvoxamine reduces Aβ generation by inhibiting γ-secretase activity. We now extend our study in vivo by investigating the effect of fluvoxamine treatment on learning and memory in the amyloidogenic J20 mice. J20 is a well-established mouse model of AD that expresses the human AβPP containing both the Swedish and Indiana mutations [6]. As a prophylactic therapy, we treated J20 mice with fluvoxamine at the age of 5 months, ∼3 months prior to the appearance of Aβ plaque pathology [6]. The mice were treated with fluvoxamine for 8 months. At the age of ∼13 months (i.e., at the completion of the treatment) the mice were weighed and carefully examined. There were no significant differences in weight nor physical appearance of the fluvoxamine treated J20 mice (26.3 g±1.2) compared to the untreated J20 mice (27.8 g±0.8) or age-matched untreated non-transgenic (NTG) controls (27.6 g±1.4).

We then tested the mice for learning and memory using NORT. NORT assesses the ability of animals to recognize a novel object in a familiar environment [16, 21]. It is a well-established learning-and-memory measurement that is commonly used to test the efficacy of a treatment on memory function. In this test, the mice were exposed (familiarized) to a number of objects. Then, one of the objects was replaced by a novel object, and the frequency and duration of exploration (i.e., nosing) of the novel object were measured. Increases in either the frequency or duration of exploration of the novel object would indicate improvements in memory function. As expected, the frequency of exploration was significantly lower in the untreated J20 mice compared to the untreated non-transgenic control mice (Fig. 3A). The treatment with fluvoxamine significantly reversed this deficit in the J20 mice. Furthermore, the treatment with fluvoxamine significantly improved the duration of exploration in the J20 mice (Fig. 3A). The mice were also tested for short-term memory and spatial reference memory using Y-maze and cheeseboard, respectively. However, in both Y-maze and cheeseboard tests there were no significant differences between NTG controls, the untreated J20 mice, nor the treated J20 mice (Fig. 3B, C). These data suggest that fluvoxamine treatment improves memory function in J20 mice in a task-dependent manner.

Fig.3

Novel object recognition task assessment of fluvoxamine treated J20 amyloidogenic mice. A) The frequency of nosing novel object of non-transgenic control (NTG) untreated and J20 mice untreated or treated with fluvoxamine. B) The duration of nosing novel object of NTG untreated and J20 mice untreated or treated with fluvoxamine. Mean +/– SEM is indicated. *p < 0.05, **p < 0.001.

Biochemical analyses of brain tissue

Western blot analysis of detergent soluble protein extracted from the frontal cortices of NTG and J20 mice confirmed the presence of the human APP (hAPP) transgene only in the transgenic mice (data not shown). There was no difference in total human AβPP protein levels between vehicle (4.8 densitometry units±0.29) and fluvoxamine treated J20 mice (5.9 densitometry units±0.56) (p = 0.114). There was also no difference in the amount of Aβ₄₀ (p = 0.888), Aβ₄₂ (p = 0.667) or Aβ₄₂/ Aβ₄₀ ratio (p = 0.701) between the two groups (Table 1).

Table 1

Aβ peptide levels from Triton-X soluble fraction of frontal lobe protein

Cohort	Aβ₄₀ (pg/mL)	Aβ₄₂ (pg/mL)	Aβ₄₂/Aβ₄₀
J20 + H₂O	1.21±0.04	0.09±0.01	0.08±0.01
J20 + fluvox	1.20±0.02	0.08±0.02	0.07±0.01
	p = 0.888	p = 0.667	p = 0.701

DISCUSSION

This is the first study to demonstrate that the SSRI antidepressant fluvoxamine plays a role in AβPP processing via modulation of γ-secretase activity in vitro. We observed support for a protective effect of fluvoxamine on two human AβPP-associated phenotypes ascertained by the novel object recognition task, but not the Y-Maze or cheeseboard tests in a mouse model of AD. Genetic evidence for the association between SIGMAR1 and AD has been equivocal. Previous studies have focused on the candidate functional polymorphism rs1800866. In a Japanese cohort of 239 AD patients, the minor C allele was associated with reduced disease risk [22]. In contrast, a Hungarian cohort of 322 AD patients observed the opposite effect with the C allele significantly associated with increased disease risk [23]. Huang et al. demonstrated that there was a significant interaction between the C allele and the apolipoprotein E alleles to increase severity of plaque pathology in a series of 330 neuropathologically confirmed Chinese AD cases [24]. Finally, a Polish study comprising 219 AD cases did not detect a significant effect for the polymorphism on disease risk [25]. Our γ-secretase assay demonstrated a significant effect of the C allele in increasing cleavage of the reporter substrate. Moreover, knockdown expression of the SIGMAR1 gene resulted in a decrease in cleavage of the substrate. These studies provide evidence that rs1800866 is a functional polymorphism that impacts AβPP processing, a finding that can aid interpretation of future association studies. It should be noted that other functional polymorphisms in linkage disequilibrium with rs1800866, such as the SIGMAR1 promoter polymorphism (rs1799729) may also play a role in disease risk [23].

SSRIs have been shown to have neuroprotective properties, such as increasing the proliferation of neural progenitors cells in the hippocampus, and therefore may improve memory processes and cognition [26]. A retrospective study of over 1.4 million people in Denmark suggested that depressed individuals with chronic use of antidepressants, including SSRIs, had a lower risk of dementia [27]. A second study of amyloid plaque imaging in 186 cognitively normal participants noted that a history of antidepressant use (the majority taking SSRIs), was associated with significantly reduced plaque burden [28]. To date, three other SSRIs antidepressants—fluoxetine [29], paroxetine [30], and citalopram [28, 31]—have been found to be efficacious in ameliorating neuropathological and cognitive deficits in transgenic mouse models of AD. Modes of action attributed to these compounds include a direct effect on the serotonergic and ERK signaling pathways [28]. In our cell culture experiments, we demonstrated that the inhibitory effect of fluvoxamine on γ-secretase activity is due to its effect on the SIGMAR1 protein, as the specific antagonist NE-100 [7] was able to reverse the effects of fluvoxamine. However, this does not precludes the possibility that fluvoxamine may also have pleiotropic effects such as elevated neuritogenesis or impact on the serotogenic pathways [26].

Given that AD pathology develops over many years, we postulate that a long-term chronic administration of SSRI may be most efficacious as a prophylactic agent. While SSRIs such as citalopram and paroxetine were efficacious in rescuing different aspects of the pathogenic effects of each animal model, each study differ markedly in their treatment period (1 to 4 months) or mode of drug delivery (injection or dissolved in water) [28 –30]. Aβ plaques appear typically at the age of 5 months in the J20 mice [6]. As our aim was to determine whether long-term fluvoxamine was feasible as a prophylaxis, we treated the mice for a period of ∼8 months in order to have an end-point with significant pathology and cognitive deficits in the untreated mice. The treatment duration is greater than other comparable studies [28 –30], but importantly, was well tolerated by the mice.

In our behavioral phenotyping of J20 mice, we observed significant differences in the novel object recognition task between non-transgenic control and amyloidogenic J20 mice in terms of a lower preference for the novel object in the amyloidogenic mice, indicating a negative phenotype associated with the AD mutation. The AD associated phenotype was significantly reversed by fluvoxamine treatment with a higher preference for the novel object in the treated compared with untreated amyloidogenic mice. However, we did not observe significant group differences for the Y-maze or cheeseboard tests. A significant cognitive deficit was previously observed in aged (>15 months) male J20 mice for the cheeseboard test [32]. This discrepancy between this study and our study using female mice is not that surprising, as significant effects of gender on the cognitive performance of mice have been demonstrated in multiple studies [33 –35]. Inconsistent effects on different behavioral phenotypes has also been observed after long term treatment with paroxetine [36]. Another important observation is that for the three behavioral tests that we had examined in the J20 mice at 13 months, there was a significant difference between the non-transgenic mice and the untreated J20 mice only for the first test (NORT) (Fig. 3). Therefore, it is not unexpected that fluvoxamine treatment would only be informative for the NORT. Finally, we did not observe a significant effect of the fluvoxamine on Aβ levels in our overall biochemical analyses. With the length of the chronic SSRI treatment, the lack of significant differences in amyloid levels at the experimental end point could be due to pharmacological tolerance or tachyphylaxis, in which receptor-mediated effects decrease after constant and prolonged exposure to the ligand [37]. Another possibility is that the drug may have amyloid-independent effects, consistent with the pleiotropic effects of SSRIs on neuroprotection [28].

Our cell culture studies suggest that fluvoxamine exerts some protective effects on the development of AD via inhibition of γ-secretase activity. Our data from our animal model also suggest a protective effect of fluvoxamine on behavior in terms of novel object recognition task. These findings support further studies in other AD mouse models and in epidemiological samples to confirm the protective effect of fluvoxamine.

Footnotes

ACKNOWLEDGMENTS

This study was funded by the Australian National Health and Medical Research Council (NHMRC Project Grant 1005769 to CDS and JBJK). TK is supported by a Career Development Fellowship (Level 2) from the NHMRC (1045643), NHMRC Project Grant (1102012), the NHMRC Dementia Research Team Initiative (1095215) and the Rebecca L. Cooper Medical Research Foundation Ltd. We thank Professor Andrew Hill (La Trobe University, VIC, Australia) for CHO-AβPP cells, and Prof Colin Masters and Dr Qiao-Xin Li (University of Melbourne, VIC, Australia) for the WO2 antibody.

Authors’ disclosures available online ().

References

Zheng

, Koo

(2011) Biology and pathophysiology of the amyloid precursor protein. Mol Neurodegener 6, 27.

Karran

, Mercken

, De Strooper

(2011) The amyloid cascade hypothesis for Alzheimer’s disease: An appraisal for the development of therapeutics. Nat Rev Drug Discov 10, 698–712.

Bateman

, Xiong

, Benzinger

, Fagan

, Goate

, Fox

, Marcus

, Cairns

, Xie

, Blazey

, Holtzman

, Santacruz

, Buckles

, Oliver

, Moulder

, Aisen

, Ghetti

, Klunk

, McDade

, Martins

, Masters

, Mayeux

, Ringman

, Rossor

, Schofield

, Sperling

, Salloway

, Morris

(2012) Clinical and biomarker changes in dominantly inherited Alzheimer’s disease. N Engl J Med 367, 795–804.

Ruscher

, Wieloch

(2015) The involvement of the sigma-1 receptor in neurodegeneration and neurorestoration. J Pharmacol Sci 127, 30–35.

Hindmarch

, Hashimoto

(2010) Cognition and depression: The effects of fluvoxamine, a sigma-1 receptor agonist, reconsidered. Hum Psychopharmacol 25, 193–200.

Mucke

, Masliah

, Yu

, Mallory

, Rockenstein

, Tatsuno

, Hu

, Kholodenko

, Johnson-Wood

, McConlogue

(2000) High-level neuronal expression of abeta 1-42 in wild-type human amyloid protein precursor transgenic mice: Synaptotoxicity without plaque formation. J Neurosci 20, 4050–4058.

Okuyama

, Imagawa

, Ogawa

, Araki

, Ajima

, Tanaka

, Muramatsu

, Nakazato

, Yamaguchi

, Yoshida M et

al.

(1993) NE-100, a novel sigma receptor ligand: In vivo tests. Life Sci 53, Pl285–Pl290.

Luty

, Kwok

, Dobson-Stone

, Loy

, Coupland

, Karlstrom

, Sobow

, Tchorzewska

, Maruszak

, Barcikowska

, Panegyres

, Zekanowski

, Brooks

, Williams

, Blair

, Mather

, Sachdev

, Halliday

, Schofield

(2010) Sigma nonopioid intracellular receptor 1 mutations cause frontotemporal lobar degeneration-motor neuron disease. Ann Neurol 68, 639–649.

Karlstrom

, Bergman

, Lendahl

, Naslund

, Lundkvist

(2002) A sensitive and quantitative assay for measuring cleavage of presenilin substrates. J Biol Chem 277, 6763–6766.

10.

White

, Du

, Laughton

, Volitakis

, Sharples

, Xilinas

, Hoke

, Holsinger

, Evin

, Cherny

, Hill

, Barnham

, Li

, Bush

, Masters

(2006) Degradation of the Alzheimer disease amyloid beta-peptide by metal-dependent up-regulation of metalloprotease activity. J Biol Chem 281, 17670–17680.

11.

Karl

, Duffy

, Scimone

, Harvey

, Schofield

(2007) Altered motor activity, exploration and anxiety in heterozygous neuregulin 1 mutant mice: Implications for understanding schizophrenia. Genes Brain Behav 6, 677–687.

12.

Vina

, Lloret

(2010) Why women have more Alzheimer’s disease than men: Gender and mitochondrial toxicity of amyloid-beta peptide. J Alzheimers Dis 20(Suppl 2), S527–S533.

13.

Wang

, Tanila

, Puolivali

, Kadish

, van Groen

(2003) Gender differences in the amount and deposition of amyloidbeta in APPswe and PS1 double transgenic mice. Neurobiol Dis 14, 318–327.

14.

Chesworth

, Downey

, Logge

, Killcross

, Karl

(2011) Cognition in female transmembrane domain neuregulin 1 mutant mice. Behav Brain Res 226, 218–223.

15.

Duffy

, Cappas

, Lai

, Boucher

, Karl

(2010) Cognition in transmembrane domain neuregulin 1 mutant mice. Neuroscience 170, 800–807.

16.

Dere

, Huston

, De Souza Silva

(2007) The pharmacology, neuroanatomy and neurogenetics of one-trial object recognition in rodents. Neurosci Biobehav Rev 31, 673–704.

17.

Hughes

(2004) The value of spontaneous alternation behavior (SAB) as a test of retention in pharmacological investigations of memory. Neurosci Biobehav Rev 28, 497–505.

18.

Pang

, Stam

, Nithianantharajah

, Howard

, Hannan

(2006) Differential effects of voluntary physical exercise on behavioral and brain-derived neurotrophic factor expression deficits in Huntington’s disease transgenic mice. Neuroscience 141, 569–584.

19.

Llano Lopez

, Hauser

, Feldon

, Gargiulo

, Yee

(2010) Evaluating spatial memory function in mice: A within-subjects comparison between the water maze test and its adaptation to dry land. Behav Brain Res 209, 85–92.

20.

Logge

, Cheng

, Chesworth

, Bhatia

, Garner

, Kim

, Karl

(2012) Role of Abca7 in mouse behaviours relevant to neurodegenerative diseases. PLoS One 7, e45959.

21.

Ennaceur

, Delacour

(1988) A new one-trial test for neurobiological studies of memory in rats. 1: Behavioral data. Behav Brain Res 31, 47–59.

22.

Uchida

, Ujike

, Tanaka

, Sakai

, Yamamoto

, Fujisawa

, Kanzaki

, Kuroda

(2005) A variant of the sigma receptor type-1 gene is a protective factor for Alzheimer disease. Am J Geriatr Psychiatry 13, 1062–1066.

23.

Feher

, Juhasz

, Laszlo

, Kalman

, Pakaski

, Kalman J

, Janka

(2012) Association between a variant of the sigma-1 receptor gene and Alzheimer’s disease. Neurosci Lett 517, 136–139.

24.

Huang

, Zheng

, Halliday

, Dobson-Stone

, Wang

, Tang

, Cao

, Deng

, Wang

, Zhang

, Wang

, Hallupp

, Kwok

, Chen

(2011) Genetic polymorphisms in sigma-1 receptor and apolipoprotein E interact to influence the severity of Alzheimer’s disease. Curr Alzheimer Res 8, 765–770.

25.

Maruszak

, Safranow

, Gacia

, Gabryelewicz

, Slowik

, Styczynska

, Peplonska

, Golan

, Zekanowski

, Barcikowska

(2007) Sigma receptor type 1 gene variation in a group of Polish patients with Alzheimer’s disease and mild cognitive impairment. Dement Geriatr Cogn Disord 23, 432–438.

26.

Takebayashi

, Hayashi

, Su

(2004) A perspective on the new mechanism of antidepressants: Neuritogenesis through sigma-1 receptors. Pharmacopsychiatry 37(Suppl 3), S208–S213.

27.

Kessing

, Sondergard

, Forman

, Andersen

(2009) Antidepressants and dementia. J Affect Disord 117, 24–29.

28.

Cirrito

, Disabato

, Restivo

, Verges

, Goebel

, Sathyan

, Hayreh

, D’Angelo

, Benzinger

, Yoon

, Kim

, Morris

, Mintun

, Sheline

(2011) Serotonin signaling is associated with lower amyloid-beta levels and plaques in transgenic mice and humans. Proc Natl Acad Sci U S A 108, 14968–14973.

29.

Wang

, Zhang

, Xu

, Zhu

, Wang

, He

, Zhang

, Guo

, Kong

, Huang

, Li

(2014) Fluoxetine improves behavioral performance by suppressing the production of soluble beta-amyloid in APP/PS1 mice. Curr Alzheimer Res 11, 672–680.

30.

Nelson

, Guo

, Halagappa

, Pearson

, Gray

, Matsuoka

, Brown

, Martin

, Iyun

, Maudsley

, Clark

, Mattson

(2007) Prophylactic treatment with paroxetine ameliorates behavioral deficits and retards the development of amyloid and tau pathologies in 3xTgAD mice. Exp Neurol 205, 166–176.

31.

Sheline

, West

, Yarasheski

, Swarm

, Jasielec

, Fisher

, Ficker

, Yan

, Xiong

, Frederiksen

, Grzelak

, Chott

, Bateman

, Morris

, Mintun

, Lee

, Cirrito

(2014) An antidepressant decreases CSF Abeta production in healthy individuals and in transgenic AD mice. Sci Transl Med 6, 236re234.

32.

Karl

, Bhatia

, Cheng

, Kim

, Garner

(2012) Cognitive phenotyping of amyloid precursor protein transgenic J20 mice. Behav Brain Res 228, 392–397.

33.

King

, Arendash

, Crawford

, Sterk

, Menendez

, Mullan

(1999) Progressive and gender-dependent cognitive impairment in the APP(SW) transgenic mouse model for Alzheimer’s disease. Behav Brain Res 103, 145–162.

34.

Oliveira

, Grigoletto

, Funck

, Ribeiro

, Royes

, Fighera

, Furian

, Oliveira

(2015) Evaluation of potential gender-related differences in behavioral and cognitive alterations following pilocarpine-induced status epilepticus in C57BL/6 mice. Physiol Behav 143, 142–150.

35.

, Qi

, Qiao

, Wang

, Zhou

(2013) Sex differences in ICR mice in the Morris water maze task. Physiol Res 62, 107–117.

36.

Olesen

, Bouzinova

, Severino

, Sivasaravanaparan

, Hasselstrom

, Finsen

, Wiborg

(2016) Behavioural phenotyping of APPswe/PS1deltaE9 mice: Age-rrelated changes and effect of long-term paroxetine treatment. PLoS One 11, e0165144.

37.

Coplan

, Gopinath

, Abdallah

, Berry

(2014) A neurobiological hypothesis of treatment-resistant depression - mechanisms for selective serotonin reuptake inhibitor non-efficacy. Front Behav Neurosci 8, 189.