Rational design, cognition and neuropathology evaluation of QTC-4-MeOBnE in a streptozotocin-induced mouse model of sporadic Alzheimer’s disease

Alzheimer’s disease (AD) is a multifactorial pathology characterized by amyloid deposits, neurofibrillary formation, oxidative stress and cholinergic system dysfunction. In this sense, here we report the rational design of a multi-target directed ligand (MTDL) for AD based on virtual screening and bioinformatic analyses, exploring the molecular targets β-secretase (BACE-1), glycogen synthase kinase-3β (GSK-3β) and acetylcholinesterase (AChE). After this screening, the compound with higher molecular docking affinity was selected, the 1-(7-chloroquinolin-4-yl)-N-(4-methoxybenzyl)-5-methyl-1H-1,2,3-triazole-4 carboxamide(QTC-4-MeOBnE). To further our studies, the protective effect of QTC-4-MeOBnE (0.1 and 1 mg/kg for 20 days) on STZ-induced sporadic AD mice was determined. QTC-4-MeOBnE pretreatment attenuated cognitive and memory deficit induced by STZ in an object recognition test, Y-maze, social recognition test and step-down passive avoidance. The mechanisms underlying this action might be attributed to the reduction of lipid peroxidation and reactive species formation in the prefrontal cortex and hippocampus of mice submitted to STZ. In addition, QTC-4-MeOBnE pretreatment abolished the up-regulation of AChE activity and the overexpression of GSK 3β and genes involved in amyloid cascade such as BACE-1, protein precursor amyloid, у-secretase, induced by STZ. Moreover, toxicological parameters were not modified by QTC-4-MeOBnE chronic treatment. This evidence suggests that QTC-4-MeOBnE exerts its therapeutic effect through multiple pathways involved in AD.

Alzheimer Disease (AD) is the most common cause of dementia, with a prevalence of more than 50 million cases in 2015 1 . AD is a neurodegenerative disease, characterized by progressive memory loss and impaired cognitive ability. The predominant pathological hallmarks are the deposition of β-amyloid (Aβ) in senile plaques and neurofibrillary tangles (NFT) composed by hyperphosphorylated TAU 2 . These alterations occur together with extensive oxidative stress, chronic neuroinflammation, excitotoxicity, and mitochondrial dysfunction. Together, these events are considered the main effectors of synaptic dysfunction and the neurodegenerative progression 3,4 .
From this perspective, AD is the major source of unmet medical needs in neurology with high rate of failed clinical trials 5 . These failures can be attributed, in part, to the selectivity of the tested drugs against a single target, while AD is a complex and multifactorial disease 6 . At the same time, several biological targets for potential therapeutics have been identified, suggesting the research strategy of development, a multi-target approach. This

QTC-4-MeOBnE is the result of virtual screening.
The central pharmacological target of this study was BACE-1 inhibitors, which are considered a potential treatment for AD, due to their ability to reduce Aβ deposition. However, one of the limiting characteristics of these drugs is the reduced oral bioavailability and the low capacity to cross the BBB. Therefore, we performed a virtual screening based on drug-like properties (molecular weight <500 KDa, Log P, number of hydrogen donors (<5) and acceptors (<10)) to avoid poor pharmacokinetics. This chemical scaffold was utilized to create 79 molecules whose binding affinity to BACE-1 was individually evaluated by molecular docking.
Multiple clinical trials suggested that BACE-1 inhibitors attenuate Aβ accumulation but fail to improve dementia. Thus, AChE inhibitors, currently approved drugs to treat AD, have been shown to reduce cognitive manifestations, without delaying or halting the progression of the disease 16 .Thus, we evaluated the QTC-4-MeOBnE binding affinity with AChE on CAS and on PAS. CAS is composed of Trp84, Tyr130, Phe330 and Phe331. This motif is responsible for binding the substrate 17 . QTC-4-MeOBnE had a docking score of −11.1 kcal/ mol and in this position, interacted with the entire site: Trp84, Tyr130, Phe330, Phe331 and with an additional hydrogen bond with Tyr121 (Fig. 1c). PAS is located at the entry to the active site gorge and is responsible for extra activities, as induce Aβ oligomerization. This site consists of five amino acids: Tyr70, Asp72, Tyr121, Trp279 and Tyr334 17,18 . Our results indicate that QTC-4-MeOBnE is a dual binding site cholinesterase inhibitor, because in this position, the compound had a docking score of −11.1 kcal/mol. As can be seen in Fig. 1d, the compound interacts with Tyr 121, Tyr335, Tyr70 and continues to form a hydrogen bond with Tyr130.
The majority of tau-based therapeutic strategies have focused on modulating tau hyperphosphorylation to reduce the formation of NFT. As GSK-3β is the main kinase responsible for the phosphorylation process of TAU, we performed a molecular docking on its ATP binding pocket, in the motif where the inhibitors described interact 19 . QTC-4-MeOBnE had a high docking score value of −9.2 kcal/mol, interacting with residues of ATP binding pocket: Asp200, Val70, Val110, Ala83, Leu 132 and Cys199 (Fig. 1e). Previous studies reported that interactions with Cys199 and Asp 200 are crucial for GSK3β inhibition 20 .
ADMET properties reveal that QTC-4-MeOBnE crosses the blood brain barrier (BBB) possessing a passive blood-brain partitioning (Log BB) of −0,215, and also had 75% of predictive activity in the central nervous system (CNS). Our results show that QTC-4-MeOBnE is probably almost completely absorbed in human intestine and is highly metabolized by CP450 3A4 enzyme. The compound has low oral toxicity, probably with a DL50 of 500-5000 mg/kg.

QTC-4-MeOBnE protects memory and cognitive impairment induced by ICV injection of STZ.
Due the promising characteristics, the effect of pretreatment with QTC-4-MeOBnE (0.1 and 1 mg/kg for 20 days) was evaluated on a STZ-induced mouse model of sporadic AD. First, mice were evaluated in Open Field Test (OFT) to avoid any false results induced by altered locomotor activity in the other behavioral tests. The two-way ANOVA revealed the absence of statistically significant interaction between STZ x QTC-4-MeOBnE for crossings [F (2,36 ) = 1.078, p = 0.35] (Fig. 2a) and rearings [F (2,36 ) = 0.1577, p = 0.85] (Fig. 2b) demonstrating that treatment with QTC-4-MeOBnE did not induce psychostimulant disturbances.
The short-term memory, in the Object Recognition Test (ORT), was defined as 90 min after the presentation of identical objects. The ORT two-way ANOVA indicated significant interaction of STZ x QTC-4-MeOBnE [F (2,36) = 12.75, p < 0.001]. STZ (3 mg/kg, 2 days) infusion-induced cognitive deficits (Fig. 2c) demonstrated by a significant reduction in the object recognition index, when compared to the sham group. QTC-4-MeOBnE protected against the memory decline in both doses tested (0.1 and 1 mg/kg) when compared to animals treated only with STZ and vehicle (canola oil).
The same reduction of the recognition index in the ORT of animals treated with STZ was observed in long-term memory, defined in the present study as 24 h after the habituation with identical objects. Two-way ANOVA analyses showed a significant interaction of STZ x QTC-4-MeOBnE [F (2,36) = 9.911, p < 0.001]. QTC-4-MeOBnE treatment, in both doses tested, was effective to protect against the impaired recognition index compared to mice treated with STZ and canola oil (Fig. 2d).
Two-way ANOVA of Y-maze revealed significant interaction between STZ x QTC-4-MeOBnE for spontaneous alternation [F (2,30) = 12.37, p < 0.001] and no interaction for the number of entries [F (2,30) The STZ administration induced a decline in spatial and working memory, otherwise, the treatment with QTC-4-MeOBnE at doses of 0.1 and 1 mg/kg prevented memory impairment, defined by an increasing percentage of spontaneous alternation behavior (Fig. 2e). This increase was independent of number of arm entries, excluding any false positive related to altered locomotor function in this test (Fig. 2f). The Step-Down Passive Avoidance (SDPA) test is extensively used to evaluate learning and aversive memory through the latency time in stepping down from the platform, after an electric shock 21 . In this study, the two-way ANOVA indicates a statistically significant interaction between STZ x QTC-4-MeOBnE on acquisition latency time [F (2, 36) = 5.958, p = 0.006] and no significance in transfer latency time [F (2, 36) = 1.422, p = 0.25]. STZ induced memory impairment when compared to the sham group and the pretreatment with QTC-4-MeOBnE 1 mg/kg was efficient efficient in preventing this increase (Fig. 2g). This result indicates that only the highest compound dosage can protect against learning deficit through aversive memory formation.
We also evaluated the latency time to step down from the platform in the training phase, defined as exploration time. There was no difference in the transfer latency time among groups, validating the SPDA test and suggesting that these results were not related to animal exploratory activity (Fig. 2h).
The social recognition memory was assessed utilizing the SRT, which simulates the ability of animals to discriminate between familiar and unfamiliar stimuli. SRT Two-way ANOVA demonstrated a significant STZ x QTC-4-MeOBnE interaction [F (2,30) = 9.422, p < 0.001]. The treatment with QTC-4-MeOBnE, in both tested doses, was efficient in preventing memory decline induced by ICV injection of STZ (Fig. 2i). This result, is represented by an increase of social contact with the juvenile intruder, demonstrating a social memory recognition.
Overall, the treatment with QTC-4-MeOBnE (0.1 and 1 mg/kg) in mice subjected to ICV injection of saline did not present any significant difference between groups in behavioral tests performed. These results suggest that www.nature.com/scientificreports www.nature.com/scientificreports/ QTC-4-MeOBnE did not present effects of improved cognition per se, it just prevented the cognitive impairment induced by STZ.
Treatment with QTC-4-MeOBnA prevents the oxidative damage induced by STZ. It has long been established that biomarkers of lipid peroxidation are elevated in the brain of patients with AD, even in the preclinical phase of the disease, when compared with age-matched controls 22 , suggesting an early and relevant involvement of AD and lipid peroxidation. Two-way ANOVA indicated a significant STZ x QTC-4-MeOBnE interaction for prefrontal cortex (PFC) [F (2,36) = 7.355, p = 0.002] and hippocampus (HC) [F (2,36) = 6.006, p = 0.0056]. The ICV injection of STZ increased TBARS levels in both structures, when compared to the sham group (Fig. 3a,b). The lipid peroxidation induced by STZ was significantly prevented by the administration of QTC-4-MeOBnE (0.1 and 1 mg/kg). In groups treated with QTC-4-MeOBnE (0.1 and 1 mg/kg) and injected with saline there was no significant change in lipid peroxidation compared with the sham group.
Lipid peroxidation is often caused by excessive RS production in AD, accompanied by events such as oxidatively modified proteins and nucleic acids 3 . The intracellular RS levels in this study were measured by DCF units of fluorescence. Two-way ANOVA showed a significant STZ x QTC-4-MeOBnE interaction for PFC [F (2,36)   www.nature.com/scientificreports www.nature.com/scientificreports/ did not reduce the AChE activity alone probably due to the time between compound administration and mice euthanasia.  www.nature.com/scientificreports www.nature.com/scientificreports/ GSK3-β is the main kinase responsible for TAU hyperphosphorylation, these data can suggest a reduction in the formation of neurofibrillary tangles, which is usually induced by STZ 24 .
Toxicological analyses. For toxicological determination, we performed a different experiment, in which mice were divided into three groups 1) canola oil, 2) QTC-4-MeOBnE 0.1 mg/kg and 3) 1 mg/kg treated for 20 days. No rate of morbidity and no toxicological manifestations in weight, skin, fur, eyes, autonomic and central nervous system (CNS) were observed.
As shown in Table 1, there were no changes in hepatotoxicity biomarkers AST and ALT in animals treated with QTC-4-MeOBnE compared to the control. In addition, the chronic compound treatment did not alter the renal biomarkers as creatinine and urea, suggesting the absence of toxicology in QTC-4-MeOBnE treatment for 20 days in these toxicological parameters.

Discussion
In the present study, we performed a rational design of an MTDL drug based on a virtual screening, exploring the biological targets as AChE, BACE-1 and GSK3-β by molecular docking. The chosen compound, QTC-4-MeOBnE also demonstrated a favorable predicted pharmacokinetics profile and when tested in vitro, inhibited the activity of AChE and protected against lipid peroxidation induced by sodium nitroprusside and Fe(II)/EDTA in mice cortex and HP (data not shown). This evidence encouraged us to test the in vivo activity of QTC-4-MeOBnE in a model of dementia induced by STZ. www.nature.com/scientificreports www.nature.com/scientificreports/ In this sense, we demonstrated the chronic administration of QTC-4-MeOBnE (0.1 and 1 mg/kg) protected against STZ induced cognitive deficit in mice, through different behavioral tests: ORT, Y-maze, SDPA and SRT. In addition, QTC-4-MeOBnA was also able to protect from STZ induced oxidative damage and up-regulation of AChE activity, amyloid cascade and GSK3-β expression. It is worth highlighting that no changes in renal and hepatotoxicity parameters were observed, defining in the first instance, the absence of hepatic and renal toxicity in QTC-4-MeOBnE administration for 20 days.
Recently, the discovery of drugs for AD has gradually tended towards the development of MTDL drugs. The discovery of molecules which can modulate multiple pathways of the disease should significantly advance therapeutic strategies. In summary, based on our prior virtual screening, we have attempted to design a low molecular weight compound which interacts with several promising AD targets. These findings provide support for the translational value of MTDL directly modulating a wide range of therapeutic targets as BACE, GSK-3β and AChE with a possibly interesting pharmacokinetics profile. This suggests that this action mechanism may possess a disease modifying potential for AD.
Consistent evidence has shown that in the majority of cases, clinical manifestations of AD start 10-15 years after the neuropathology began, which makes the prevention of AD characteristics, as induced by STZ, a significant finding. Other research studies have already demonstrated the efficacy of different MTDL's in vivo in AD but with different action mechanisms such as CHF5074 (anti-inflammatory and γ-secretase inhibitor) 25 , ARN14140 (NDMA and AChE) 26 ; MT-031 (MAO-A and AChE/BuChE inhibitor) 27 ; M30 (propargylamine and chelating) 28 .
Nevertheless, the combination of two or several structural features with specific single-target activity into one structure, face a major challenge related to the structure-activity relationships (SAR) which sometimes makes it difficult to link together distinct pharmacophore groups without losing their associated functionalities 29,30 . In this context, virtual screening and molecular docking might be useful to reshape drug design strategies, to counter (n = 5-6). # p < 0.05 and ## p < 0.01 when compared with the control group. *p < 0.05, **p < 0.01 when compared to STZ group (two-way ANOVA followed by Newman-Keuls test). www.nature.com/scientificreports www.nature.com/scientificreports/ determinant steps of the neurotoxic cascade. Thus, the present study shows for the first time, the biological activity of a rational designed moiety, exploring BACE-1, GSK-3β, AChE and oxidative stress therapeutic targets. Although preliminary, these results indicate an interesting new direction for the search for multi-target-directed ligands against AD, as well, the chemical scaffold obtained might be useful for the design of more powerful drugs to overcome the limitations of current single-target drugs in a multifaceted disease.
When STZ is administered by ICV, it decreases cerebral glucose uptake, inducing hypometabolism, accompanied by pathological alterations close to AD, being considered a non-transgenic model of this disease, simulating sporadic AD-like pathology. These alterations include neuroinflammation, metabolic deregulation, cholinergic deficits, accumulation of β-amyloid and tau proteins, and oxidative stress as well as memory and learning impairment 31 . However, besides the STZ insulin resistant state of the brain, a mechanistic explanation for the STZ mode of action and its relation to AD is currently lacking 32 .
In our study two injections of STZ (3 mg/kg) by ICV route unilaterally significantly produced memory impairment linked to AD susceptibility without significant effects on blood glucose levels. The memory impairment was assessed by different behavioral tests trying to mimic different types of memory which are formed throughout life.
Thus, the short-term and long-term memory were assessed by ORT. This test is primarily based on visual recognition to acquisition of memory, based upon the preference of interacting with novel targets among familiar ones 33 . In this study, Y-maze test was utilized to evaluate working memory, which also depends on an associative process. However, the memory evaluated in the Y-maze test also implicates the spatial navigation conditioned to environment 34 . On the other hand, learning mediated aversive memory was evaluated by SDPA, where the mouse subjected learns to associate a context (stepping down) with the occurrence of an aversive event (electric shock) 35 .
All these events are strongly dependent on activity in the hippocampus and cortex (perirhinal, entorhinal, and inferior temporal) being influenced by lesions and damage in these cerebral areas 36 . In support of this, the memory impairment observed in this study may be attributed to neurodegeneration caused by the ICV of STZ. In line with this, QTC-4-MeOBnE (0.1 and 1 mg/kg) is successful in preventing cognitive decline, and can be at least, related to its neuroprotective effect, preventing the oxidative damage induced by STZ in mice cortex and hippocampus, inferred by RS formation and MDA levels.
In addition, the protective effect of QTC-4-MeOBnE was further confirmed through gene expression, in which we studied the genes: APP, BACE-1, γsecretase, GSK3-β and synaptophysin. STZ significantly induced an over expression of amyloid cascade in HC but failed to achieve this goal in PFC. This event could be explained by amyloid deposition occurring initially and more prominent in HC 37 . Mice treated with QTC-4-MeOBnE showed a reduction of APP, BACE-1, γsecretase expression in HC but in PFC the level of γsecretase was reduced without reaching a statistically significant lower expression. This tendency might be confirmed with a higher number of animals.
This capacity of QTC-4-MeOBnE to modulate the amyloid cascade might result in the decreased generation of Aβ. In line with this, we also observed a reduction in GSK3-β expression which can suggest a reduction in TAU hyperphosphorylation, but further studies are needed to confirm this hypothesis. The reduction of Aβ and hyperphosphorylated TAU might explain the effect of QTC-4-MeOBnE in down regulating the activity of AChE, increased by STZ. Some studies have already pointed out these events as crucial for up-regulation of AChE activity in AD 23,27 . Unfortunately, the QTC-4-MeOBnE treatment at this dosage and time, was not able to show an AChE inhibitory activity in animals treated with saline, as we expected. These data can be explained by the time between the euthanasia and compound administration or can suggest a higher dosage or administration time in future studies involving QTC-4-MeOBnE and AChE.
Neuronal loss and synaptic disruption have been associated with cognitive decline in AD and many diseases associated with dementia. However, altered levels of synaptophysin, a major and widespread presynaptic protein, have shown that they are controversial in AD 38,39 . Previous studies reported that synaptophysin levels were not correlated with disease duration 40 . Moreover, recently, Poired et al. 41 found that synaptophysin expression was minimally altered by the progression of dementia, and just in severely demented patients. Furthermore, in this study we did not find statistically significant differences in synaptophysin expression in the PFC and HC of mice submitted to infusion of STZ. These results are in agreement with other previous reports utilizing an STZ-induced mouse model of AD, as in Gratuze et al. 42 and Guo et al. 24 , in which levels of synaptophysin or PSD-95 remain unchanged after the ICV (3 mg/kg) of STZ. This evidence suggests that the STZ-induced memory impairment might be attributed to loss of functional synapses or changes in the organization and distribution of synaptic networks without causing substantial synapse volume loss 43,44 . On the other hand, Rai et al. 45 and Xu et al. 46 , demonstrated that an ICV of STZ caused a reduction of PSD-95 levels, but not of synaptophysin levels, suggesting that STZ causes majority post synaptic neurotoxicity.
Moreover, drug toxicity is the most challenging drug property that characterizes one of the most significant reasons for many drugs failing to be approved to reach the market 47 . It is widely accepted that adverse reactions and toxicity, generally, are directly proportional to drug dosage 48 . QTC-4-MeOBnE demonstrated therapeutic effects at low dosages such as 0.1 and 1 mg/kg daily. These low drug doses can contribute to the absence of alteration in toxicity parameters tested, such as renal and hepatotoxicity. Besides, we also predict QTC-4-MeOBnE pharmacokinetics, which might suggest a good profile of absorption, distribution, metabolization and excretion. However, further experimental studies are needed to determine its pharmacokinetics, safely, and the pharmacodynamics exerted, which are part of our future study prospects.
In conclusion, we identified QTC-4-MeOBnE as an MTDL compound exploring its interaction with BACE-1, GSK-3β and AChE by molecular docking. The preliminary biological evaluation demonstrated the beneficial effect of QTC-4-MeOBnE in preventing cognitive decline, amyloidosis, oxidative stress, up regulation of AChE and GSK3-β of STZ-induced AD mice, without showing hepatic and renal toxicity. Our results emphasize the emerging strategy of MTDLs and contribute to indicating that these compounds can possibly play an important role in multifactorial diseases, representing a potential approach in anti-AD drug discovery. Furthermore, (2019) 9:7276 | https://doi.org/10.1038/s41598-019-43532-9 www.nature.com/scientificreports www.nature.com/scientificreports/ additional studies about the in vitro compound effect and biological validation are required to understand the complete mechanism of this compound, and its physiological effects in different AD models.

Materials and Methods
In silico analysis. Virtual screening. The virtual screening was performed with BACE-1 as an initial target using the ZINC database (http://zinc.docking.org/). We established chemical parameters based on drug-like properties, such as molecular weight <500 KDa, Log P, number of hydrogen donors (<5) and acceptors (<10) to avoid low bioavailability, and poor pharmacokinetics 49,50 . The screening led to 100 drug-like molecules, which were downloaded and manually reviewed to identify a sub-set of candidates that share structural and chemical characteristics with those molecules studied by our research group. At the end of this analysis, 79 structures were created, with the possibility of synthesis using a simple and efficient methodology.
Molecular docking. The 79 selected molecules were drawn using ChemDraw and their geometry optimized with Avogadro 0.9.4 following the MMFF94 method 51 . The molecular docking simulation was used as a second screening for these molecules utilizing the Autodock Vina software 52 . We used different structures of BACE (PDB: 1SGZ and PDB: 2ZHU) from the Protein Data Bank (PDB) (http://www.pdb.org/). The CHIMERA 1.5.3 software was used to remove molecules, ions, water and to minimize the structure of proteins, using the Gasteiger charges with 500 steps of minimization 53 .
The ligand and enzyme structures were improved using Auto Dock Tools 1.5.4, in which all the rotatable bonds of ligands were freely rotated and the receptors were considered rigid 54 . For docking studies, we used the Auto Dock Vina (version 1.1.1), utilizing a grid box centered on the active site of the enzymes with high resolution, allowing the program to search for additional sites of probable interactions.
Based on promising results obtained, we explored molecular docking on two other important targets: GSK-3β (PDB: 1H8F) and AChE (PDB:1DX6). The evaluation of the interaction of QTC-4-MeOBnE with AChE was explored using a grid box on CAS and one on the PAS.

QTC-4-MeOBnE synthesis.
The synthesis of QTC-4-MeOBnE was performed according to our previous protocol using enamide-azide cycloaddition 55 . Thus, to a solution of 4-azido-7-chloroquinoline 1 (1.0 mmol) in DMSO (2.0 mL), was first added the N-(4-methoxybenzyl)-3-oxobutanamide 2 (1.0 mmol) and then the catalyst pyrrolidine (5 mol%). The reaction mixture was stirred at 70 °C for 24 hours in an open vial. After completion of the reaction, the crude product was purified by column chromatography on silica gel and the desired product 3 was obtained with an 82% yield (Fig. 6). The spectra and NMR can be visualized in a supplementary section.
The STZ was purchased from Sigma-Aldrich (St Louis, USA), dissolved in saline 0.9% and administered by ICV injection at a dose of 3 mg/kg, with a limit rate of 3 μl/site.
All other chemicals of analytical grade were obtained from standard commercial suppliers. ICV administration. The ICV was administered employing a "free-hand" method as described previously 56,57 using the bregma fissure as a reference point. Briefly, a microsyringe (25 μL, Hamilton) was inserted perpendicularly and following the coordinates: 0.8 mm posterior to the bregma, 1.0 mm lateral to the sagittal suture, and 3.0 mm beneath the surface of the brain. The needle was inserted unilaterally 1 mm into the midline point equidistant from each eye at an equal distance between the eyes and the ears and perpendicular to the plane of the skull. www.nature.com/scientificreports www.nature.com/scientificreports/ First, animals were immobilized under light isoflurane anesthesia and and a gauze soaked in 70% ethanol was used for asepsis of the injection site). The injection was given over 30 s, and the needle was kept in place for a further 30 s to avoid reflux of the injected solution. After ICV injection the peripheral glucose levels were measured using a glucose monitor (Accu-Chek ® Aviva) (data not shown). During the dissection of the animal brain, the success of the injection was examined macroscopically, discarding animals whose injection occurred in an inappropriate place or caused cerebral hemorrhage (<5%).

Experimental design.
After three days of adaptation to colony room conditions, mice were randomly divided into six groups: Twenty days after the second ICV injection, behavioral tests were carried out to evaluate the learning and memory capacity of each group, between days 43 and 45. On day 46, mice were sacrificed, and the brain dissected for further analysis (Fig. 7).
The SDPA was evaluated in a different group of animals, due to the possible alteration caused by the electric shock. In order to reduce the number of animals, the gene expression was evaluated utilizing the highest dose of QTC-4-MeOBnE with its respective vehicles.

Behavioral tests.
Open-field test (OFT). The animals were subjected to OFT for 5 min. The test was conducted in a wooden box (30 × 30 × 15 cm) divided into nine equal squares. The locomotor (scored by the number of segments crossed with the four paws) and exploratory activities (expressed by the number of times the mice stood on their rear limbs) were counted 58 .
Object recognition test (ORT). The object recognition test was performed in an open-field apparatus and conducted according to Rosa et al. 59 . Plastic objects with different shapes, composed only by primary colors were placed at the center of the apparatus. One day before the object presentation, each animal was submitted to a habituation session in the absence of objects for 10 min. Firstly, animals were placed in the center of the apparatus containing two identical objects for 5 min as a training phase. After 90 min one object was replaced by another of different form and color, and the exploration of both objects was recorded for 5 min, configuring short-term memory. A second trial was conducted 24 h after the identical object presentation, with another object (different from the first and second in shape and color) comprising the long-term memory.
The results were expressed as exploratory preference according to the following formula: [time spent in the novel object/(time spent in the familiar object + time spent in the novel object)].
Y-maze test. The working memory was evaluated according to Sarter et al. 60 . Y-maze is a three-arm horizontal maze (40 cm long and 3 cm wide with walls 12 cm high). Mice were initially placed in one arm (A), and the arm entry sequence (e.g. ABCACB, where letters indicate arm codes) and the number of arm entries was recorded for each mouse over a 6 min period.
Step-down passive avoidance (SDPA). The non-spatial memory was investigated through the step-down passive inhibitory avoidance task in accordance with Sakagushi et al. 61 with some slight modifications. The apparatus www.nature.com/scientificreports www.nature.com/scientificreports/ consists of a single box, in which the floor was made of a metal grid connected to a shock scrambler with a safe platform where each animal was placed before starting the experiment. The experiment consists in a training phase and 24 h later, a test session was carried out. The animal was set on the safe platform and when it stepped down with its four paws on the grid floor an electric shock (0.5 mA) was delivered during 2 s. The acquisition memory was defined as 1 min on the safe platform. The test session was performed 24 h later following the same procedure but no shock was given after stepping down.
Social recognition test (SRT). The SRT defined by Kogan et al. 62 was conducted with some modifications. In this test, mice were placed into individual cages immediately prior to the experimental sessions to habituate to the new environment for 5 min. The cages used were identical to those in which the animals were normally housed. After the habituation, a male juvenile mouse (2 weeks old) was placed into the cage with the tested mice for an initial interaction trial, recorded for 5 min. After 90 min, the same procedure was carried out. The social recognition was determined as the time in which the adult spontaneously investigated the younger animal and determined by the index: C 2 /C 1 .

Biochemical analyses. Tissue preparation.
On the 46th day, mice were euthanized, and the brain was immediately removed and dissected into prefrontal cortex (PFC) and hippocampus (HC). The PFC and HC were separated into two hemispheres in order to submit each sample to all analyses. The right hemispheres were homogenized in 50 mM Tris-HCl, pH 7.4 (1:4, w/v) to determine reactive species (RS) formation and thiobarbituric acid reactive species (TBARS) levels.
In another group of animals, the right hemispheres were immersed in TRIzol, maintained at −80 °C and submitted to the quantitative real-time polymerase chain reaction (qRT-PCR). Both hemispheres were homogenized in 25 mM sucrose buffer (1:10, w/v) and designated to AChE activity.
The homogenates were centrifuged at 2400 g for 10 min to yield a pellet that was discarded and a low-speed supernatant (S 1 ) was used for further analysis. The protein concentration was measured by the method of Bradford et al. (1976), using bovine serum albumin (1 mg/ml) as the standard.
Lipid peroxidation. The lipid peroxidation was evaluated by TBARS formation, as previously described by Ohkawa et al. 63 . An aliquot of S 1 was incubated with 8.1% SDS, 0.8% TBA (pH 3.2) and acetic acid buffer (pH 3.4) at 95 °C for 2 h. Malondialdehyde (MDA), one of the main products of lipid peroxidation, was utilized as biomarker. Sample absorbance was measured at 532 nm, and the results were expressed as nmol MDA/g tissue.
Reactive species quantification. The quantification of RS was performed according to Loetchutinat et al. 64 . In this assay, an aliquot of S 1 was incubated with 1 mM dichloro-dihydro-fluorescein diacetate (DCHF-DA) and 10 mM Tris-HCl pH 7.4. The oxidation of DCFH-DA into fluorescent dichlorofluorescein (DCF) is evaluated for intracellular RS detection. The fluorescence intensity was recorded in a spectrofluorophotometer, configured with 520 nm emission and 480 nm excitation and the results were expressed as arbitrary units of fluorescence.
AChE activity. AChE activity was carried out according to the method of Ellman et al. 65 , using acetylthiocholine as substrate. In this assay the thiocholine (8 mM, acetyl choline analog) complexed to dithiobis-2-nitrobenzoate (DTNB, 10 mM), reduces DTNB to 5-thio-2-nitro-benzoic acid (TNB). TNB has a yellowish color proportional to AChE activity and can be read in the UV light spectrophotometer at 412 nm, at 30 s intervals. The AChE activity was expressed as nmol/mg protein/min. Gene expression. Total mRNA was extracted in HC and PFC right hemispheres using TRIzol (Invitrogen ™ , Carlsbad, USA) followed by DNase treatment with DNA-free ® kit (Ambion ™ , USA) and mRNA quantification. The cDNA synthesis was performed using a High Capacity cDNA Reverse Transcription kit (Applied Biosystems ™ , UK) according to the manufacturer's protocol.
UltraSYBR Mix (COWIN Bioscience Co., Pequim, China) with the Stratagene Mx3005P was used for amplification. The primers were synthetized by Exxtend (Brasil) and their sequence is described in Table 2. Gene expressions were normalized utilizing βactin for GSK-3β, APP and γ-Secretase. For BACE-1 and synaptophysin β-microglobulin was utilized as a reference gene. The conditions for the reaction included the cycles: 95 °C for 15 s, 60 °C for 60 s and 72 °C for 30 s. The 2ΔΔCT (Delta-Delta Comparative Threshold) method was used to normalize the fold change in gene expressions.  www.nature.com/scientificreports www.nature.com/scientificreports/ Sub-chronic toxicity determination. The toxicity of sub-chronic QTC-4-MeOBnE gavage administration was evaluated in a different experiment. Mice were divided into 3 treatment groups (canola oil, QTC-4-MeOBnE 0.1 mg/kg and 1 mg/kg) which received the corresponding treatment for 20 days. During the treatment, animals were weighed twice a week and observations related to toxicology parameters were done daily. These toxicological parameters include changes in skin, fur, eyes, autonomic (salivation, lacrimation, perspiration, piloerection, urinary incontinence, and defecation), and central nervous system (drowsiness, gait, tremors, and convulsions) changes.
24 h after the last QTC-4-MeOBnE administration, mice were anaesthetized (by inhalation of isoflurane) for blood collection by cardiac puncture. Blood was collected directly in a heparinized tube and centrifuged at 2500 × g for 10 min to obtain the plasma. In view of evaluated biochemical markers of hepatic damage, aspartate (AST) and alanine aminotransferase (ALT) activity were performed, which were expressed as U/l. The renal function was analyzed by determining plasma urea and creatinine levels, which were expressed as mg/dl.
All these parameters were determined by enzymatic colorimetric methods using commercial kits (Labtest Diagnostica, MG, Brazil) according to the manufacturer's instructions.
Statistical analyses. All measurements were performed by an independent investigator blinded to the experimental conditions. Data were presented as mean ± standard error of means (S.E.M). Differences among experimental groups were determined by one-way ANOVA for toxicity analyses or two-way ANOVA for the rest of the tests. The post hoc test utilized was Newman-Keuls multiple comparisons tests. All results were analyzed using GraphPad Prism software (version 5.0, Prism software for PC, GraphPad) and were considered statistically significant when P < 0.05.