The Role of Melatonin on Behavioral Changes and Concomitant Oxidative Stress in icvAβ1-42 Rat Model with Pinealectomy

One of the pathological hallmarks of Alzheimer’s disease (AD) associated with its progression that contributes to β-amyloid (Aβ) generation is oxidative stress (OS). Clinical data suggest that melatonin is a potent antioxidant that might be effective in the adjunctive therapy of this neurodegenerative disease. The present study aimed to explore the role of melatonin on behavioral changes and markers of OS in three rat models, namely, pinealectomy (pin) model of melatonin deficit, intracerebroventricular (icv)Aβ1-42 model of AD, and combination of both pin and Aβ1-42 model (pin+icvAβ1-42). The chronic injection with vehicle/melatonin (50 mg/kg, i.p. for 40 days) started on the same day of sham/pin and icv vehicle/Aβ1-42 infusion procedures. Anxiety in the open field and the elevated plus-maze test and cognitive responses in the object recognition test were tested between the 30th–35th day after the surgical procedures. Markers of OS in the frontal cortex (FC) and hippocampus were detected by the ELISA method. Melatonin treatment corrected the exacerbated anxiety response only in the pin+icvAβ1-42 model while it alleviated the cognitive impairment in the three models. Pinealectomy disturbed the antioxidant system via enhanced SOD activity and decreased GSH levels both in the FC and hippocampus. The Aβ1-42 model decreased the SOD activity in the FC and elevated the MDA level in the two brain structures. The pin+icvAβ1-42 model impaired the antioxidant system and elevated lipid peroxidation. Melatonin supplementation restored only the elevated MDA level of icvAβ1-42 and pin+icvAβ1-42 model in the hippocampus. In conclusion, our study reveals that the pin+icvAβ1-42 rat model triggers more pronounced anxiety and alterations in markers of OS that may be associated with melatonin deficit concomitant to icvAβ1-42-induced AD pathology.


Introduction
The main pathological characteristics of Alzheimer's disease (AD) are the formations of extracellular Aβ plaques and intraneuronal deposits of neurofibrillary tangles (NFTs). Oxidative stress (OS) is an important pathogenic mechanism of AD which appears as a prominent and early feature of this neurodegenerative disease and leads to the formation of free radicals and subsequent damage of specific brain regions, including the frontal cortex (FC) and hippocampus [1][2][3][4][5]. Oxidative stress is an imbalance between the reactive oxygen species (ROS) and the antioxidant system that removes them. This state may be due to overproduction of ROS or a reduction in the antioxidant defense system, which has a high oxidative capacity while its ability to fight the OS is limited [6,7]. Brain tissue has multiple potential sources of ROS [8,9]. One of the most relevant hypotheses describing OS as the main factor in AD pathophysiology is because neuronal cells have a higher intake of oxygen, elevated lipid content, and a lower content of antioxidant enzymes such as Cu/Zn-superoxide dismutase (SOD), glutathione (GSH) and catalase compared to other types of cells, making them more vulnerable to changes in OS [10][11][12][13]. Together with the presence of Aβ and NFT, these factors lead to mitochondrial dysfunction and neuronal cell death [14]. Therefore, one of the promising therapies of AD is the reduction of OS through antioxidant treatment [15][16][17].
Melatonin (N-acetyl-5-methoxytryptamine) is a crucial endogenous indoleamine secreted by the pineal gland and other extrapineal tissues such as the gastrointestinal tract [18]. In addition to its critical role in regulating circadian rhythms, melatonin and its metabolites have well-known antioxidant properties in the central nervous system (CNS) [19,20]. Furthermore, along with the antioxidant activity of melatonin to CNS, there is evidence that suggests the efficacy of melatonin in preventing oxidative stress in cells as erythrocytes in a concentration-related manner [21]. Clinical evidence suggests that the diminished function of this hormone represents an earlier biomarker of AD [22][23][24][25]. Changes in melatonin (MT) receptors associated with this melatonin deficit have also been reported in patients with AD [26,27]. Melatonin can quickly cross the blood-brain barrier, making the exogenous melatonin administration a drug of choice against circadian desynchronization, emotional disturbance, and memory decline in patients with AD [28][29][30][31][32].
Pinealectomy was established as an experimentally induced melatonin-deficient animal model to evaluate the occurrence of scoliosis in rodents [33] and appears to be a good model to study melatonin deficiency and pathway dysfunction in the pathogenesis of neurodegenerative diseases like multiple sclerosis, AD, and Parkinson's disease [34]. Most AD models were performed using synthetic Aβ, Aβ 1-40 , or Aβ 1-42 analogous to peptides found in neuritic plaques in AD patients [35]. The central administration of Aβ induced learning, and memory deficits in rats [36], cholinergic dysfunction [37], neuronal apoptosis [38], OS [39], and neuroinflammation [40]. Specifically, microinjections of Aβ 1-42 impaired memory in various memory tasks, including Y-maze, radial arm-maze, Morris water-maze tests, and passive avoidance test [41][42][43].
In the present study, we aimed to introduce a new AD animal model concomitant with melatonin deficiency obtained by a combination of pinealectomy and icv infusion of Aβ 1-42 (pin+icvAβ , to check whether: (i) melatonin deficit can exacerbate the expected icvAβ 1-42 -related alterations in behavior and markers of OS and (ii) melatonin supplementation can correct them in a rat pin+icvAβ 1-42 model.

Open Field Test
The total motor activity was not affected by the three factors, Pinealectomy, Aβ, and Drug (p > 0.05) ( Figure 1A). Changes in the vertical activity (rears) and time spent in the center is the behavioral marker of the altered anxiety response. Post hoc test demonstrated that both, removal of the pineal gland, and addition of Aβ, diminished exploratory behavior (rears) in vehicle-treated groups compared to the C-sham-veh group (p = 0.0273; p = 0.014). The combination of both models showed identical to their individual effects (p > 0.05 Aβ-pin-veh group compared to C-pin-veh group and Aβ-sham-veh group). Melatonin treatment restored decreased vertical activity associated with pinealectomy (p = 0.034 C-pin-mel group compared to C-pin-veh group), Aβ accumulation (p = 0.012 Aβ-sham-mel group compared to Aβ-sham-veh group) and pin+Aβ treatment (p = 0.045 Aβ-pin-mel group compared to Aβ-pin-veh group) ( Figure 1B). A significant Pinealectomy x Aβ x Drug interaction [F 1,63 = 4.718, p = 0.035] was detected for the time spent in the aversive zone (center). Post hoc test confirmed that melatonin provoked an anxiolytic response in the Aβ-group with pinealectomy (p = 0.009 compared to Aβ-pin-veh group) ( Figure 1C). C-pin-veh compared to C-sham-veh group; * p = 0.014 Aβ-sham-veh compared to the C-sham-veh group, * p = 0.034 C-pin-mel group compared to C-pin-veh group; * p = 0.012 Aβ-sham-mel group compared to Aβ-sham-veh group; * p = 0.045 Aβ-pin-mel group compared to Aβ-pin-veh group (B); ** p = 0.009 Aβ-pin-mel compared to Aβ-pin-veh group (C).

GSH Activity in the Hippocampus
Like in the FC, the main effect of Pinealectomy [F 1,63 = 213.284, p < 0.001] with Pinealectomy × Aβ interaction [F 2,63 = 16.852, p < 0.001] for the GSH levels was demonstrated after Aβ accumulation. The post-hoc test showed a statistically significant decrease of the GSH levels both in the C-pin and the Aβ rats with pinealectomy (p < 0.005, C-pin-veh group compared to C-sham-veh group; p = 0.0065, Aβ-pin-veh group compared to C-pin-veh group; p < 0.001, Aβ-pin-veh group compared to Aβ-C-veh group) ( Figure 5B). Similar to the FC, melatonin did not affect GSH levels in the hippocampus.

SOD Activity in the FC
Three-way ANOVA revealed a main Pinealectomy effect for the activity of SOD in the FC [F 1,63 = 21.560, p < 0.001]. The post-hoc test demonstrated that the removal of the pineal gland increased the SOD activity in both the control and Aβ group (p < 0.001, C-pin-veh group compared to C-sham-veh group; p = 0.0002, Aβ-pin-veh group compared to Aβsham-veh group) ( Figure 5A). In the opposite, a decreased SOD activity of rats icv infused with Aβ was detected in the FC (p < 0.0013, Aβ-sham-veh group compared to C-sham-veh group). As for the GSH level, melatonin did not affect SOD activity in the FC.

SOD Activity in the Hippocampus
A main Pinealectomy effect [F 1,63 = 82.450, p < 0.001] with Pinealectomy × Aβ [F 2,63 = 5.144, p = 0.01] interaction was detected in the hippocampus. Like in the FC, the post hoc test showed that pinealectomy elevated the SOD activity in the sham group and more profoundly in the icvAβ-pin group (p = 0.0002, C-pin-veh group compared to C-sham-veh group; p = 0.05, Aβ-pin-veh group compared to sham-pin-veh; p = 0.0002 Aβ-pin-veh group compared to Aβ-sham-veh group) ( Figure 5B). Melatonin treatment was ineffective against pin-and pin + Aβ alterations of SOD activity in the hippocampus (p > 0.05).

MDA Level in the Hippocampus
The analysis of variance showed a main effect of Aβ [F 1,63 = 4.415, p = 0.044] as well as Pinealectomy × Aβ × Drug interaction [F 2,63 = 6.090, p = 0.019] for the MDA levels in the hippocampus. Pinealectomy per se did not affect MDA levels (p > 0.05 C-pin-veh compared to the C-sham-veh group). The lipid peroxidation was significantly increased in the hippocampus after the icvAβ infusion in both the sham and pin group (p = 0.002 Aβsham-veh group compared to the C-sham-veh group; p = 0.048 Aβ-pin-veh group compared to C-pin-veh group) ( Figure 6B). Melatonin was able to correct the increased MDA levels both in the Aβ-sham group (p = 0.0039 Aβ-sham-mel group Aβ-sham-veh group) and the Aβ-pin group (p = 0.002, Aβ-pin-mel group compared to Aβ-pin-veh group).

Discussion
The crucial role of OS in the pathogenesis of AD has been largely investigated, and numerous studies have reported that plasma levels of oxidative products are increased in patients with AD and can lead to neurodegeneration and impaired cognitive responses [44,45]. Melatonin has been reported to have multiple roles in the CNS, including chronotropic and antioxidant activity, improving neurogenesis and synaptic plasticity, neuroprotective, suppressing neuroinflammation, and enhancing memory function [46]. The underlying mechanism of the antioxidant activity of melatonin can be either as a direct free radical scavenger or by regulating the antioxidant or prooxidant enzymes via gene expression stimulation or increased activity of enzymes [47]. Although most of the published experiments point to melatonin as a free radical scavenger and antioxidant, there is some evidence to highlight the role of melatonin as a pro-oxidant mainly in cancer cells. [48]. The idea is that melatonin can increase oxidative stress and thus cause tumor cell death [48].
In the present study, we elaborated three rat models as follows: melatonin deficiency rat model induced by pinealectomy, icvAβ 1-42 model of AD, and Aβ 1-42 model of melatonin deficiency concomitant to AD pathology (pin+icvAβ  to check whether the lack of melatonin in a condition of Aβ pathology could exacerbate model-related behavioral deficit and OS in the FC and hippocampus. In addition to verifying these rat models, we explored the potency of chronic melatonin treatment to correct the model-associated pathology. The present study's findings suggest that while melatonin deficit exacerbates icvAβ-induced anxiety response and produces memory impairment in the pin+icvAβ 1-42 model, chronic melatonin supplementation restores these behavioral deficits.
The enhanced anxiety resulting from pineal gland removal detected a month after the surgery procedure, was confirmed in two different tests-OF and EPM. However, our previous results and literature data demonstrated a divergence with a lack of effect on anxiety, anxiolysis, or anxiogenic response resulting from pinealectomy [49][50][51][52][53]. This suggests that the difference in the strain and the testing time after pinealectomy might be an important issue determining this behavioral response.
The detected icvAβ 1-42 -induced emotional disturbance associated with increased anxiety in rats was in line with our very recent study [46] as well as other reports [54].
The rats with melatonin deficiency concomitant to AD pathology also demonstrated impaired anxiety response which was more pronounced in the EPM than the icvAβ 1-42induced pathology. Moreover, melatonin supplementation successfully restored this behavioral deficit, preferably in the pin+icvAβ 1-42 model (time in the aversive zone in the OF and EPM). Thus, we can speculate that model-related up-regulation of MT receptors in limbic structures including the amygdala and hippocampus determines the efficacy of melatonin treatment as was reported in patients with AD [26,27]. However, this presumption needs future investigation to confirm the role of MT receptors in the effect of melatonin supplementation on anxiety.
In support of the hypothesis that OS is elevated in the AD model [55][56][57][58], the present study confirmed previous reports which announced that the icvAβ 1-42 infusion is associated with an elevated MDA level both in the FC and hippocampus. In turn, melatonin supplementation alleviated lipid oxidation.
However, the detected lipid peroxidation was not accompanied by significant adaptive changes of antioxidant markers GSH and SOD, respectively, as was reported by other authors in icvAβ 1-42 mouse models [57]. Surprisingly, most of the previous reports considering the role of OS in Aβ pathology were conducted in mice while the few rat Aβ 1-42 models did not explore changes of the antioxidant markers GSH and SOD.
On the other hand, although the model of melatonin deficiency in the present study did not affect the MDA level in the FC and hippocampus, it induced significant changes in the antioxidant system (SOD activity and GSH levels), suggesting that OS represents an essential part of brain pathology associated with pinealectomy. Several other studies have demonstrated that melatonin deficiency provokes OS in different rat tissues [59,60]. We confirmed a report by Tasdemir et al. [59] who show a decrease of GSH in the rat brain with pinealectomy, including in the FC and hippocampus. However, in contrast to Tasdemir et al. [59], who demonstrated decreased antioxidant markers in brain tissue, we found enhanced SOD activity in rats with pinealectomy. The discrepancy between our findings and those of Tasdemir et al. [59] is probably due to the difference in experimental procedures, brain structures examined, and measured time points of antioxidant markers after the surgery procedure. While the rats were decapitated a month after pinealectomy in our study, Tasdemir et al. [59] measured antioxidant markers six months later that might exacerbate the pathology associated with melatonin deficiency, including OS. In support of the presumption that the time window of pinealectomy might affect the expected pathology differently, our previous report shows that time of testing after pinealectomy is an important factor in determining responses [49]. The overproduction of SOD as a byproduct of oxygen metabolism plays a crucial role in the pathology of many diseases [61,62]. The idea that disturbed balance in SOD activity is toxic is associated with its main product H 2 O 2 that might also be elevated due to SOD over-reactivity [61].
The original finding of the present study is that concomitant to the melatonin deficit model icvAβ 1-42 infusion (pin+icvAβ 1-42 ) represents a more relevant model of AD to study its pathology associated with OS. We report that the combined pin+icvAβ 1-42 model is characterized by brain region-specific overproduction of SOD activity (more profound than in pinealectomy per se), as well as decreased GSH and lipid peroxidation. Furthermore, the supplementation with melatonin turned down the MDA levels in the pin+icvAβ 1-42 model most probably by a direct mechanism of neutralizing the free radicals [60] and not through activation of the antioxidant system. The last assumption needs further clarification in future studies.

Animals
Ten weeks old adult male Sprague-Dawley rats (Charles River Lab., Wilmington, MA, USA), obtained from the vivarium of the Institute of Neurobiology, BAS, were kept in standard cages in groups of n = 3-4 under an artificial 12-h light/dark cycle, the temperature at 22-23 • C, and relative humidity 45%. Except during behavioral tests, the standard laboratory chow and water were at disposal ad libitum.

Surgery and icv Injection of Aβ 1-42
The rats under deep anesthesia induced with ketamine (80 mg/kg, i.p.) and xylazine (20 mg/kg, s.c.) were inserted on a stereotaxic apparatus (Narishige Sci. Inst. Labs, Tokyo, Japan). First, the pineal gland was rapidly removed through thin forceps according to the procedure described previously [49,63] and according to the procedure described by Hoffmann and Reiter [64]. We verified pinealectomy as a relevant model of melatonin deficit with a lack of melatonin in plasma during the dark phase [63]. After removal of the pineal gland, two cannulas were implanted bilaterally according to the atlas of Paxinos and Watson [65] at the following coordinates: (AP = −0.8, L = ± 1.5, H = 3.8). Amyloid β 1-42 (100 µg; FOT Ltd., Sofia, Bulgaria) was dissolved in 100 µL PBS (vehicle solution) and incubated at room temperature for a week before using it to produce neurotoxic fibrils [44]. The Aβ 1-42 intracerebroventricular (icv) infusion was performed with a 5-µL Hamilton microsyringe at a rate of 1 µL/min for 5 min. The microsyringe was left in place for an additional 2 min before removal. Matched procedures were carried out on the sham-operated group where PBS was infused. Several days after surgery the rats were injected with lactated Ringer's and an antibiotic (gentamicin, s.c.).

Behavioral Tests
The behavioral tests were carried out between 10:00 a.m. and 12:00 p.m. in a separate soundproof room, under artificially diffused light, where the rats were accommodated for 30 min before the test. Also, the rats were repeatedly handled for a week before behavioral procedures to attenuate the effect of stress. For the open field (OF) a video tracking system (SMART PanLab software, Harvard Apparatus, USA) was used to analyze the collected data automatically. For the object recognition test (ORT) a video-recording system connected with DVR was used and the files were examined offline by two experienced observers.
Open field test was carried out as described previously [45,62]. In brief, each rat was placed in the central zone of the apparatus (100 × 100 cm × 60 cm) for a 5-min period. The following parameters were analyzed: the total distance traveled (cm), the vertical activity (number of rears), and the time spent in the center (cm). The apparatus was cleaned with 70% alcohol solution after each test.
Elevated plus maze test was performed as depicted in our previous reports [45,62]. The apparatus was cleaned with 70% alcohol solution after each test.
Object recognition test was executed in an open plastic box (50 cm × 50 cm × 50 cm), with grey walls. The test consisted of three sessions. The flow, walls, and objects were cleaned after each trial with 70% alcohol solution. In the first session, each rat was placed in the box for 10 min without objects (accommodation). Twenty-four hours later, the second session was performed, and the rat was exposed to identical objects made of glass (A1 and A2) for 3 min. The objects were positioned about 10 cm away from the opposite corners. One hour later, one of the objects was replaced with a new one made from plastic (the novel object B), and test the third session was carried out again for 3 min. The number (count) and time (sec) of exploration defined as the nose approach to the object in the third session were recorded. The discrimination index DI (count and time) of exploration for each object was calculated as follows: B/B + A.

Detection of Markers Related to Oxidative Stress in the Homogenates from the Frontal Cortex and Hippocampus
After behavioral tests, all rats were decapitated with a guillotine after mild anesthesia with CO 2 and the whole brains were removed. FC and hippocampi were carefully excised. Each of the frontal cortex and hippocampal samples were weighted and homogenized (1:10) with SONOPULS ultrasonic homogenizer (BANDELIN electronic GmbH&Co. KG, Berlin) in ice-cold 0.9% saline (6 mice/group). The homogenates were centrifuged at 3000× g for 10 min at 4 • C to obtain the supernatants. The supernatants were diluted with the appropriate buffer and were used for assays of GSH and MDA levels and SODspecific activities.

Statistical Analysis
Three-way ANOVA analyzed the behavioral and biochemical data were by with factors: Pinealectomy (sham and pin), Aβ (control, Aβ), and Drug (vehicle, melatonin) followed by post hoc Bonferroni test in case of detected significant difference (SigmaStat 11.0). When data were not homogenously distributed, nonparametric tests were applied (Kruskal-Wallis on ranks followed by the Mann-Whitney U test). The significant level was set at p ≤ 0.05.

Conclusions
In the present study, we introduced a new pin+icvAβ 1-42 rat model of AD with concomitant melatonin deficiency and compared its behavioral and OS pathology with the pin model and icvAβ 1-42 model, respectively. Our results suggest that the pin+icvAβ 1-42 rat model induces more pronounced anxiety and OS. The latter is better expressed in the hippocampus. Melatonin treatment minimizes these neurodegenerative features, and OS elevation most probably by a direct free radical scavenger mechanism.