ML385

Pterostilbene alleviates Aβ1-42-induced cognitive dysfunction via inhibition of oxidative stress by activating Nrf2 signaling pathway

Jikai Xu 1, 2, Jingyu Liu 1, 2, Qing Li 1, Yan Mi 1, Di Zhou 3, Qingqi Meng 1, Gang Chen 3, Ning Li 3,*, Yue Hou 1, 2,*

Abstract

Scope: In the present study, we investigated effect of pterostilbene on Aβ1-42 induced cognitive impairment in mice and explored its possible mechanism of action.
Methods and results: The behavior results showed that pterostilbene alleviated Aβ1-42-induced cognitive dysfunction assessed using the Y-maze test, novel object recognition task, Morris water maze test, and passive avoidance test. Pterostilbene alleviated neuron loss and accumulation of reactive oxygen species (ROS) in Aβ1-42 treated mouse brain. Additionally, pterostilbene promoted Nrf2 nuclear translocation and enhanced the transcription and expression of antioxidant genes such as HO-1 and SOD both in vivo and in vitro. Nrf2 inhibitor ML385 reversed the antioxidant function of pterostilbene in SH-SY5Y cells. Nrf2 is the master regulator of oxidative homeostasis and can be activated by p62. Pterostilbene promoted the binding of Keap1 and p62 which enhanced activation of Nrf2.
Conclusion: The present study reported that pterostilbene alleviated Aβ1-42-induced cognitive dysfunction in mice. The mechanism of pterostilbene can be associated to the inhibition of oxidative stress through the Nrf2 signaling pathway. The present study reported that pterostilbene alleviated Aβ1-42-induced cognitive dysfunction in mice. Its mechanism of action may be associated to the inhibition of oxidative stress through the p62/Keap1/Nrf2 signaling pathway.

Keywords: Pterostilbene, Aβ1-42, Nrf2, p62, Cognition, Oxidative stress.

1. Introduction

Alzheimer’s disease (AD) is the major cause of dementia. Histopathological features of AD include senile plaques formed by β-amyloid (Aβ) aggregation and neurofibrillary tangles formed by hyperphosphorylation of tau [1, 2]. Oxidative stress is involved in the pathogenesis of AD [3, 4]. Abnormally elevated reactive oxygen species (ROS) levels and neuronal oxidative stress damage were found in the early pathological process in AD patients [5]. There are various factors that cause oxidative stress injury in AD. Among them, Aβ is a pro-oxidant that contributes to the production of excessive ROS [4]. Moreover, ROS increases Aβ production, thus forming a vicious circle [6]. An excessive ROS level causes a redox homeostasis imbalance and oxidative stress in the brain, which leads to neurodegeneration and loss of neurons [7, 8].
Nuclear factor-E2 p45-related factor 2 (Nrf2) is the master regulator of cells, that maintains oxidative homeostasis under oxidative stress [9]. By combining with antioxidant response element (ARE), Nrf2 initiates the transcription of a variety of cytoprotective and antioxidant enzymes/proteins such as heme oxygenase-1 (HO-1) and superoxide dismutase (SOD) [10, 11]. Physiologically, Kelch-like
ECH-associated protein 1 (Keap1) binds to Nrf2 and keeps Nrf2 in the cytoplasm [12]. As a repressor of Nrf2, Keap1 binds to Nrf2 and causes Nrf2 degradation in the cytoplasm. Substrate adaptor sequestosome-1 (also named p62), the target substrate for autophagy, can bind to Keap1 and positively modulate the activity of Nrf2 [13]. The interaction between p62 and Keap1 promotes degradation of Keap1, which leads to Nrf2 accumulation and nuclear translocation eventually [13].
Pterostilbene (3′,5′-dimethoxy-4-stilbenol) is a natural stilbenoid, which is mainly found in berries and grape. In blueberries, the content of pterostilbene can reach 520 ng/g dry sample [14]. The content of pterostilbene in grapes can reach 5 μg/g fresh weight [15]. Pterostilbene possesses multiple pharmacological activities, including anticancer activity and anti-liver fibrosis and regulates the gut microbiota [16, 17, 18]. Evidence suggests that pterostilbene may regulate Nrf2 signaling pathway and exert antioxidant effect [19].
It is reported that a Nrf2-dependent antioxidant mechanism may play a key role in the pathogenesis of AD [20]. We hypothesized that pterostilbene may exert an anti-AD effect by regulating Nrf2 signaling pathway. Thus, in this study, we investigated the effect of pterostilbene on Aβ1-42-induced AD animal model and explored the Nrf2-dependent antioxidant mechanism in the presence of pterostilbene on Aβ1-42-treated mice.

2. Methods

2.1 Animals

Male Swiss-Kunming mice (6-8 week) weighing 18-22 g were obtained from the Experimental Animal Center of Shenyang Pharmaceutical University. All animals were housed under a 12 h light/dark cycle and 22 ± 2°C room temperature. Animals had ad libitum access to food and water. All Animal use procedures followed the Regulations of Experimental Animal Administration issued by the State Committee of Science and Technology of the People’s Republic of China. The ethical permission number is SYPU-IACUC-C2017-11-10-102.

2.2 Surgical operation

Aβ1–42 peptide (Sigma Aldrich, St. Louis, Missouri, USA) was treated as previously described [21] to obtain Aβ oligomers. HFIP-treated Aβ1–42 was resuspended in dimethylsulfoxide (DMSO) and dissolved at a concentration of 1 μg/μL in PBS. Aβ1–42 was incubated at 4°C for 24 h. This stock-solutions were stored at -80°C. Before intracerebroventricular injection, the stock-solutions was diluted to 100 μM in ice-cold PBS and incubated at 4°C for 24 h. The preparation was centrifuged (14,000 × g, 4°C, 10 min) and the supernatants containing soluble Aβ1–42 oligomers were collected. For intracerebroventricular injection, mice were anesthetized with avertin (2.5%) and gently restrained on a stereotaxic frame. A small incision was performed on the scalp. Aβ1–42 oligomers were injected into the lateral ventricle (AP: -1.1 mm, ML: -0.5 mm, DV: -3.0 mm relative to the bregma). Three microliters of aggregated Aβ1–42 oligomers (10 pmol) were injected (1 μL/min), and the syringe was kept for 5 min to allow diffusion [22, 23].

2.3 Dosage Information/Dosage Regimen

Pterostilbene (purity ≥ 98.0%, Chengdu Pufeide Biotechnology Co., Ltd, Chengdu, China) was suspended in 1.0% (w/v) sodium carboxy methyl cellulose (CMC-Na). The dose of pterostilbene in our present study is 10, 20 and 40 mg/kg daily referring our previous study [16]. The human equivalent dose is 0.95, 1.90 and 3.81 mg/kg daily calculated based on body surface area. In vitro, the concentrations of pterostilbene (1, 3 and 10 μM) was used based on our previous study [20].

2.4 Grouping

Mice were randomly divided into five groups: control, intracerebroventricular injection (ICV) of Aβ1–42, ICV Aβ1–42 + 10 mg/kg pterostilbene, ICV Aβ1–42 + 20 mg/kg pterostilbene and ICV Aβ1–42 + 40 mg/kg pterostilbene. Ten mice were used for each group. One day after intracerebroventricular injection, mice in control and ICV Aβ1–42 group were orally administered with 0.5% CMC-Na and mice in ICV Aβ1–42 + pterostilbene group were orally administered with pterostilbene daily. Behavior tests were started on the 7th day of pterostilbene treatment and followed by the schedule presented in Fig. 1.

2.5 Behavioral assessment

2.5.1 Y-maze test

Short-term memory was assayed by Y-maze as previously described [24]. The Y-maze was made up of 3 the same arms (40 × 5 × 12 cm, 120° apart). Each arm had a distinctive visual clue at the terminal. Mice were sited at the center of the experimental equipment and permitted to explore for 8 min. The arm entries were recorded only when mouse completely entered one arm. The maze was cleaned thoroughly with 10% ethanol after each mouse was removed from the equipment. Successful alternation was defined as consecutive entries into all three arms. Total arm entries were evaluated. Spontaneous alternation was calculated as successful alternations/ (total arm entries – 2) × 100%.

2.5.2 Novel object recognition task

Novel object recognition (NOR) task is commonly used to measure the nonspatial memory basis on the neural circuitry between hippocampus and medial temporal lobe [25]. The apparatus was a plastic open field arena (44 × 44 × 44 cm). The task was divided into three phases: habituation, acquisition and test. In the habituation phase, mice were familiarized with the field by exploring for 3 min twice daily. In the acquisition phase, each mouse could explore the two identical objects for 5 min. Total exploring time of both objects was recorded. Recognition index was calculated as: time spent exploring one of the identical objects / Total time spent exploring both objects × 100%. In the test phase, one hour after one mouse explored the identical objects, the same mouse performed the test trial for a duration of 5 min. A different object was substituted for one of the identical objects. The time spent in exploring both objects was recorded. Discrimination index (DI) was calculated as: DI = [(Tn – Ti) / (Tn + Ti)] × 100% in which Tn represents the time spent detecting the new object and Ti represents the time spent detecting one of the identical objects in acquisition phase.

2.5.3 Morris water maze test

Morris water maze test is a well-validated method to determine the spatial learning and memory ability of animals [26]. The Morris water maze is a white circle pool (100 cm diameter × 60 cm height) and a platform in a fixed quadrant of the pool below the water. The experimental procedure was split into two divisions: place navigation and spatial probe. At the place navigation stage, two-time trials at four-hour interval per day were conducted for 5 days. Briefly, mouse could swim and find the platform, and then successfully stay on the platform for 10 s. If a mouse fails to find the platform within 60 s, it will be helped to stand on the platform for 10 s. The time for each mouse to land on the platform was recorded as escape latency. At the spatial probe stage, the platform was taken away, and the animals were placed into the pool. The swimming speed, numbers of platform crossing, swimming time in target quadrant and swimming distance in target quadrant were recorded. Additionally, the motion tracks of the mice were registered by video-tracking system.

2.5.4 Passive avoidance test

Passive avoidance test is widely used to measure passive avoidance memory [27]. The test device consists of dark and bright compartments with a door in the middle. The experiment was conducted in two days. On day 1, as a training trial, mouse was placed into bright compartment and given 3 min for habituation, then driven into the dark compartment by opening the door. Once all limbs of the mouse entered the dark room, an electric shock was delivered, and the mouse ran back to the bright compartment immediately. A successful trial was defined as remaining in the bright compartment for 5 min. On day 2, the mouse was placed into the bright room. The times of dark compartment entry and escape latency were recorded.

2.6 Tissue preparation and immunofluorescence

Mice were anesthetized with avertin (2.5%) and perfused transcardially. The body was infused with 4% paraformaldehyde in PBS until the whole body became stiff. The brain tissue was collected and fixed with 4% paraformaldehyde at 4°C for 24 h and cut into 14-µm-thick sections with Cryostat Microtome. Unspecific antibody binding was blocked by 5% goat serum. Sections were stained by anti-NeuN (1:1000, Abcam, Cambridge, United Kingdom) or Nrf2 (1:500, Abcam) antibodies overnight at 4°C and incubated with secondary antibodies at 37°C for 1 h. Nuclei were stained with 4’, 6-Diamidin-2-phenylindol (DAPI) in Nrf2 immunofluorescence. NeuN positive cell numbers or Nrf2 fluorescence intensity were counted by ImageJ (version 1. 4. 3. 67).

2.7 Dihydroethidium staining

Frozen sections were washed 3 times with PBS and then incubated with dihydroethidium staining solution (DHE, Beyotime, Shanghai, China) in the dark for 30 min. After incubation, the sections were washed 3 times with PBS and the red fluorescence was observed under a fluorescence microscope.

2.8 Malondialdehyde assay

Malondialdehyde (MDA) production was measured by MDA assay kit (Beyotime). The test was carried out in accordance with the kit instructions. Briefly, brain tissue or SH-SY5Y cells were added to saline to prepare a 10% homogenate. Reagents were added according to the instructions, and the absorbance was measured at 532 nm.

2.9 Quantitative real-time PCR

Trizol reagent was used to extract total RNA from brain hippocampus. To obtain cDNA, cDNA synthesis kit was performed to reverse transcribed mRNA to cDNA. qRT-PCR was conducted with synthetic primers to amplifying HO-1, SOD and GAPDH. The specific primers are shown in Table 1. For qRT-PCR, samples were incubated at 50°C for 2 min and 95°C for 10 min, then involved in the reactions contained 40 cycles of 95°C for 3 s and 60°C for 30 s. The result was normalized to GAPDH.

2.10 Western blot

After blocking with 5% milk, anti-KEAP1 antibody (1:1000, CST, Boston, Massachusetts, USA), HO-1 (1:1000, CST), p62 (1:1000, CST), Beclin1 (1:1000, CST), LC3 (1:1000, CST) and anti-β-actin antibody (1:10000, CST) were respectively acted on the membranes at 4°C overnight. Then secondary response was performed for 2 h. The expression of each protein was quantified using LI-COR Odyssey infrared imaging system and normalized to that of β-actin.

2.11 Superoxide Dismutase activity assay

Superoxide Dismutase (SOD) activity in tissue samples or SH-SY5Y cells was detected using Total Superoxide Dismutase Assay Kit with NBT (Beyotime). Tissue samples were added to PBS, ground into homogenate at 4°C and centrifuged. After centrifugation, the supernatant was collected. For SH-SY5Y cells, similar protocol was followed. The results are expressed as SOD enzyme activity unit (U) / total protein (mg) in the sample.

2.12 Cell culture

SH-SY5Y neuronal cells were cultured in DMEM with 10% FBS and 1% penicillin-streptomycin in the medium and incubated at 37°C with 5% CO2.

2.13 Cell viability

Cell viability was detected using 3- (4,5-dimethylthiazol-2-yl) -2,5-diphenyltetrazolium bromide (MTT, Beyotime). SH-SY5Y cells were treated with pterostilbene (1, 3 and 10 μM) and/or Aβ1–42 (3 μM) for 24 h, and MTT (5 mg/mL) was added. DMSO was added after the medium was removed. The mixture was shaken in dark for 10 min, and the absorbance was measured at 490 nm.

2.14 Reactive Oxygen Species Assay

ROS content was assessed with Reactive Oxygen Species Assay Kit. SH-SY5Y cells were treated with pterostilbene and/or Aβ1–42 for 24 h. Then, the medium was changed to a serum-free medium containing 10 μM DCFH-DA. Cells were incubated in a 37°C for 20 min and then washed three times with PBS. The intensity of fluorescence was detected at 488 nm excitation wavelength and 525 nm emission wavelength.

2.15 Co-immunoprecipitation

Control SH-SY5Y cells or cells treated with pterostilbene for 6 h were lysed with Cell lysis buffer for Western and IP (Beyotime). Part of the supernatant was taken as “Input”. The remaining supernatant was added with 50% beads (Protein A+G Agarose, Beyotime) to remove unspecific antibody binding. Equal amount of p62 antibody (1:1000, CST) was added to the supernatant of control or pterostilbene-treated cells. The supernatant and antibody were mixed with an end-over-end mixer at 4°C overnight. After centrifugation, the supernatant was removed and the beads were washed twice with PBS. The washed beads were added to 1 × SDS loading buffer and placed at 100°C for 5 min. The mixture was analyzed by the method used for Western blot.

2.16 Statistical analysis

Statistical analysis was carried out by using Statistical Package for Social Sciences (SPSS) 19.0 software. Two-way analysis of variance (two-way ANOVA) was used to analyze the escape latency in navigation test of Morris water maze test. Student t test and one-way analysis of variance (one-way ANOVA) was performed followed by Fisher Least Significant Different or Dunnett’s T3 post hoc test to analyze other data. p < 0.05 was defined as statistically significant. Error bars on bar graphs represent standard error of the mean (SEM). 3. Results 3.1 Pterostilbene attenuated working memory impairment induced by Aβ1–42 Y-maze test was used to explore the working memory ability of the mice. The locomotor activity of the mice was unaffected, as no significant differences [F(4, 45) = 2.167, p = 0.088] were found in the total number of arm entries among groups (Fig. 2a). One-way ANOVA analysis showed significant differences [F(4, 45) = 3.922, p = 0.008] in spontaneous alternation among groups (Fig. 2b). Compared to the control group, ICV of Aβ1–42 significantly diminished the spontaneous alternation (p = 0.001), while treatment of pterostilbene significantly improved the spontaneous alternation (10 mg/kg: p = 0.009, 20 mg/kg: p = 0.002, 40 mg/kg: p = 0.010). No significant differences were found in the spontaneous alternation between all three pterostilbene group (post hoc test, p > 0.05).

3.2 Pterostilbene ameliorated short-time visual recognition ability

Novel object recognition task was carried out to test the short-time visual recognition ability of the mice. In the acquisition phases, mice spent similar time to explore objects [F(4, 45) = 0.830, p = 0.513, Fig. 2c] and did not have preference for either of the identical objects [F(4, 45) = 1.409, p = 0.246, Fig. 2d]. In the test phase, mice spent similar total time to explore objects [F(4, 45) = 0.810, p = 0.526, Fig. 2e], while mice in different groups showed significant differences in discrimination index [F(4, 45) = 3.092, p = 0.025, Fig. 2f]. ICV of Aβ1–42 induced worse discrimination (p = 0.002) between novel object and constant object, while pterostilbene 40 mg/kg (p = 0.018) significantly increased the discrimination index compared to that of Aβ1–42 groups, but not 10 mg/kg (p = 0.124) and 20 mg/kg (p = 0.056). The results showed no significant differences in the discrimination index between all three pterostilbene group (post hoc test, p > 0.05).

3.3 Pterostilbene improved fear-related memory impairment

Passive avoidance test was performed to determine fear-related memory ability of the mice. One-way ANOVA analysis showed significant difference in times of dark compartment entry [F(4, 45) = 4.181, p = 0.006, Fig. 2g] and escape latency [F(4, 45) = 5.056, p = 0.002, Fig. 2h] among groups. Compared to the control group, ICV of Aβ1– 42 induced significant fear-related memory impairment, which was evidenced by declined escape latency (p = 0.039) and times of dark compartment entry (p = 0.031). Treatment with pterostilbene significantly increased escape latency (p = 0.009) and times of dark compartment entry (p = 0.031). Pterostilbene 40 mg/kg showed better improvements in escape latency (10 mg/kg: p = 0.010, 20 mg/kg: p = 0.003) and times of dark compartment entry (10 mg/kg: p = 0.002, 20 mg/kg: p = 0.001).

3.4 Pterostilbene declined spatial learning and memory impairment

The Morris water maze test was used to evaluate the spatial learning and memory ability of the mice. Two-way ANOVA analysis showed that time and treatment groups were two independent factors [Fgroup × day (16, 225) = 0.940, P = 0.524] during the training period in the Morris water maze test. On the fifth day in place navigation phase, mice in the Aβ1–42 group swam for a longer time [F(4, 45) = 4.217, p = 0.006, post hoc test, p = 0.001] to find the safe platform than that of the control group (Fig. 3a), while pterostilbene 40 mg/kg significantly reduced the escape latency (p = 0.003) in Aβ1–42-stimulated mice. No significant differences were found in the escape latency between all three pterostilbene group (post hoc test, p > 0.05).
In the spatial probe phase, all mice swam at similar speed [F(4, 45) = 1.126, p = 0.356, Fig. 3b], indicating there were no significant difference in swimming ability in mice. One-way ANOVA analysis showed significant difference in the number of platform crossings [F(4, 45) = 3.688, p = 0.011, Fig. 3c], swimming time in target quadrant [F(4, 45) = 3.489, p = 0.015, Fig. 3d], and swimming distance in target quadrant [F(4, 45) = 3.127, p = 0.024, Fig. 3e] among groups. ICV of Aβ1–42 reduced the number of platform crossing (p = 0.014), swimming time in target quadrant (p = 0.001), and swimming distance in target quadrant (p = 0.008) compared to those of the control group. On the contrary, treatment of pterostilbene significantly increased the number of platform crossing (20 mg/kg, p = 0.004, 40 mg/kg p = 0.010), swimming time in target quadrant (20 mg/kg, p = 0.013, 40 mg/kg p = 0.007), and swimming distance in target quadrant (20 mg/kg, p = 0.007, 40 mg/kg p = 0.006) in mice. No significant differences were found in the number of platform crossing, swimming time in target quadrant and swimming distance in target quadrant between all three pterostilbene group (post hoc test, p > 0.05). Pterostilbene improved the spatial learning and memory ability of AD-like mice, and the swimming trace was shown in Fig. 3f.

3.5 Pterostilbene prevented the neuronal damage in Aβ1-42-treated mice

To detect neuronal damage, mature neurons were labeled with NeuN. The results of immunofluorescence experiments showed that the number of mature neurons was significantly lower in the hippocampal CA1 [F(4, 70) = 37.438, p < 0.001, post hoc test, p < 0.001, Fig. 4a, 4b] and cortical regions [F(4, 70) = 121.781, p < 0.001, post hoc test, p < 0.001, Fig. 4c, 4d] of the mouse model compared to those of the control group, while pterostilbene significantly inhibited neuronal death in in the hippocampal CA1 (20 mg/kg: p = 0.001, 40 mg/kg: p < 0.001) and cortical regions (10 mg/kg: p = 0.007, 20 mg/kg: p < 0.001, 40 mg/kg: p < 0.001). Pterostilbene 40 mg/kg showed better protective effect than 20 mg/kg (p = 0.011) and 10 mg/kg (p < 0.001) in CA1 and better protective effect than 10 mg/kg (p < 0.001) in cortical regions. 3.6 Pterostilbene decreased ROS content in Aβ1-42-treated mouse brain Aβ1-42 could cause oxidative stress and neuronal damage by increasing ROS content. The ROS content was higher in hippocampal CA1 [F(4, 70) = 15.000, p < 0.001, post hoc test, p < 0.001, Fig. 5a, 5b] and cortical regions [F(4, 70) = 21.971, p < 0.001, post hoc test, p < 0.001, Fig. 5c, 5d] of mice injected with Aβ1-42 compared to that of the control group. Pterostilbene significantly reduced ROS content in Aβ1-42-treated mouse brain (CA1: 20 mg/kg, p < 0.001, 40 mg/kg p < 0.001, cortex: 20 mg/kg, p = 0.031, 40 mg/kg p < 0.001). Between the three different dosages, pterostilbene 10 mg/kg showed higher ROS content than 20 mg/kg (p < 0.001) and 40 mg/kg (p < 0.001). Excessive ROS oxidizes lipids and produces MDA. Aβ1-42 promoted the generation of MDA in mouse brain [F(4, 70) = 16.682, p < 0.001, post hoc test, p < 0.001, Fig. 5e], and pterostilbene dose-dependently inhibited the MDA production (20 mg/kg: p = 0.033, 40 mg/kg: p = 0.002). 3.7 Pterostilbene activated the Nrf2 signaling pathway in Aβ1-42-treated mouse brain Nrf2 pathway is a key regulator of the cells against oxidative stress injury [11]. Pterostilbene significantly increased the distribution of Nrf2 in the nucleus in Aβ1-42-treated mouse brain in a dose-dependent manner [F(4, 70) = 13.069, p < 0.001, post hoc test, 20 mg/kg: p = 0.022, 40 mg/kg: p = 0.001, Fig. 6a, 6b]. Keap1 protein can bind Nrf2 to keep Nrf2 in the cytoplasm. Therefore, Keap1 can inhibit transcription of downstream genes by suppressing nuclear translocation of Nrf2 [12]. Pterostilbene reduced the protein levels of Keap1 and promoted Nrf2 translocation into the nucleus. The results also showed that pterostilbene dose-dependently initiated the transcription of downstream antioxidant stress genes, which was evidenced by increased HO-1 [F(4, 70) = 97.157, p < 0.001, post hoc test, 40 mg/kg: p < 0.001, Fig. 6c] and SOD mRNA content [F(4, 70) = 26.785, p < 0.001, post hoc test, 20 mg/kg: 0.007, 40 mg/kg: p < 0.001, Fig. 6d]. On the other hand, pterostilbene dose-dependently increased HO-1 expression [F(4, 70) = 62.158, p < 0.001, post hoc test, 20 mg/kg: p = 0.005, 40 mg/kg: p < 0.001, Fig. 6e] and enhanced the activity of SOD [F(4, 70) = 47.397, p < 0.001, post hoc test, 20 mg/kg: p = 0.009, 40 mg/kg: p < 0.002, Fig. 6f]. Pterostilbene decreased Keap1 expression in mice [F(4, 70) = 14.368, p < 0.001, post hoc test, 20 mg/kg: p = 0.001, 40 mg/kg: p < 0.003, Fig. 6g], thus indicating that pterostilbene could activate Nrf2 pathway by inhibiting Keap1 in vivo. 3.8 Pterostilbene activated the Nrf2 signaling pathway in SH-SY5Y cells To further explore whether pterostilbene reduced the content of ROS by activating Nrf2 pathway, Aβ1-42-stimulated SH-SY5Y cell model was established. Pterostilbene (10 μM) did not affect SH-SY5Y cell viability after treatment for 24 h [F(5, 48) = 11.333, p < 0.001, post hoc test, p = 0.260, Fig. 7a]. Aβ1-42 (3 μM) significantly reduced cell viability (p < 0.001) and increased ROS content after 24 h [F(5, 48) = 2.596, p = 0.037, post hoc test, p = 0.004, Fig. 7b]. Pterostilbene reduced cell damage (10 μM: p = 0.023) caused by Aβ1-42 and decreased ROS content (3 μM: p = 0.020, 10 μM: p = 0.014), showing that pterostilbene could protect neurons and inhibit ROS accumulation in SH-SY5Y cells. After treating SH-SY5Y cells for 6–24 h, pterostilbene (10 μM) significantly increased the protein levels of HO-1 [F(5, 48) = 12.242, p < 0.001, post hoc test, 6 h: p = 0.006, 12 h: p = 0.008, 24 h: p = 0.016, Fig. 7c]. Under co-stimulation with Aβ1-42, the treatment with pterostilbene for 6 h promoted the nuclear translocation of Nrf2 (Fig. 7d) and increased the protein content of HO-1 [F(3, 32) = 7.955, p < 0.001, post hoc test, versus control: p = 0.023, versus Aβ1-42: p = 0.019, Fig. 7e] and SOD enzyme activity [F(3, 32) = 5.290, p = 0.004, post hoc test, versus control: p < 0.001, versus Aβ1-42: p < 0.001, Fig. 7f]. Aβ1-42 promoted the generation of MDA in SH-SY5Y cells. Pterostilbene inhibited MDA production [F(3, 32) = 43.284, p < 0.001, post hoc test, versus control: p < 0.001, versus Aβ1-42: p < 0.001, Fig. 7g]. The Nrf2 inhibitor ML385 offset the ROS inhibitory effect of pterostilbene [F(4, 40) = 2.849, p = 0.036, post hoc test, p = 0.012] and inhibited the effect of pterostilbene on the increase in HO-1 protein content [F(4, 40) = 4.168, p = 0.006, post hoc test, p < 0.001] and SOD activity [F(4, 40) = 2.849, p = 0.036, post hoc test, p = 0.045, Fig. 8a–8c] suggesting that pterostilbene regulated ROS content through Nrf2 pathway. Pterostilbene protected SH-SY5Y cells from damage induced by Aβ1-42, and this effect of pterostilbene was blocked by ML385 [F(4, 40) = 10.729, p < 0.001, post hoc test, p = 0.002, Fig. 8d], thus indicating that pterostilbene could protect neuron damage by activating Nrf2. 3.9 Pterostilbene facilitated the binding of p62 and Keap1 The autophagosome degradation substrate p62 competes with Nrf2 in binding with Keap1, and the binding can promote the degradation of Keap1. Thus, the content of free Nrf2 in the cytoplasm increases, and eventually the Nrf2 pathway is activated [13]. Co-IP experimental results (Fig. 8e) showed that pterostilbene enhanced the binding of p62 and Keap1 (Student t test: Fig. 8f, p = 0.006, Fig. 8g, p = 0.027, Fig. 8h, p = 0.002), suggesting that pterostilbene could activate the Nrf2 pathway by promoting the binding of p62 and Keap1. 4. Discussion The present study reported that pterostilbene alleviated Aβ1-42-induced cognitive dysfunction in mice. The underlying mechanism of action may be associated to the inhibition of oxidative stress probably through the activation of Nrf2 signaling pathway. One of the main pathological features of AD is senile plaques formed by large amounts of Aβ deposition. Aβ can cause oxidative stress, neuroinflammation, energy metabolism disorders, and ion disorders, which ultimately lead to degenerative lesions and neuron loss [28, 29]. The structural basis for development of cognitive dysfunction is that functional neurons are damaged, and the circuits formed between the neurons are destroyed [30, 31]. Aβ causes degenerative lesions and loss of neurons through various ways and ultimately results in cognitive dysfunction. Oxidative stress is one of these processes [32]. As a consequence of oxidative stress, per-oxidized proteins and lipids are found in the brains of AD patients [3]. In this study, Aβ1-42 caused oxidative stress and subsequently neuronal damage in mice or SH-SY5Y cells, consistent with previous studies [33]. Pterostilbene was reported to reduce Aβ toxicity and protect neurons [34]. Pterostilbene, a natural product with antioxidative activity, can alleviate oxidative stress in some central nervous system diseases [35, 36]. In this study, pterostilbene reduced ROS production and loss of mature neurons in the brains of Aβ1-42-treated mice. Pterostilbene reduced the ROS production caused by Aβ1-42 in SH-SY5Y cells and increased the survival rate of cells. This effect of pterostilbene was blocked by ML385, thus indicating that pterostilbene could play a neuroprotective role by activating Nrf2 pathway. Nrf2 is a key regulator on oxidative stress [9], consisting of seven Nrf2-ECH homology (Neh) domains [37, 38], among which Neh2 binds to the Kelch repeat domain of Keap1 and mediates the degradation of Nrf2 [39]. Keap1 is a major regulator of Nrf2. Owing to the antioxidative properties of Nrf2, inhibition of Nrf2 can cause neurodegenerative diseases, for example, Nrf2-specific inhibitor or Nrf2 gene knockout causes cognitive impairments [40]. This makes Nrf2 a potential target for treating AD. In this study, pterostilbene promoted the nuclear translocation of Nrf2 in the brain of Aβ1-42-treated mice and SH-SY5Y cells. Nrf2 specific inhibitor ML385 interferes with the binding of Nrf2 and DNA by binding the Neh1 domain [41]. In this study, ML385 blocked pterostilbene from reducing ROS in Aβ1-42-stimulated SH-SY5Y cells. This suggests that pterostilbene may reduce oxidative stress by activating Nrf2 pathway.
By combining with ARE, Nrf2 initiates the transcription of approximately 250 genes encoding a variety of proteases involved in antioxidant metabolism, iron metabolism, and xenobiotic detoxification [42, 43]. Among them, HO-1 and SOD are two main antioxidant stress genes whose transcriptional events are initiated by Nrf2 [10, 11]. Studies showed that HO-1 was able to resist oxidative stress [44]. Activation of HO-1 could reduce Aβ1-42 toxicity [45, 46]. SOD catalyzes the disproportionation of superoxide anion radicals to form oxygen and hydrogen peroxide, thereby maintaining redox balance. In this study, pterostilbene increased HO-1 expression and SOD activity. This may be part of the mechanism of pterostilbene scavenging ROS. Pterostilbene relieved oxidative stress injury by activating HO-1 expression [35] and SOD expression [47], and our results were consistent with previous studies.
One of the regulators to mediate Nrf2 activation is p62, mainly through the combination of p62 and Keap1 [13]. Keap1 has four domains containing a double glycine repeat (DGR or Kelch repeat) domain and a carboxy-terminal region (CTR) combined with residues 346–359 in p62 [48]. The DGR is also a domain that binds Nrf2 [12], therefore, p62 and Nrf2 compete for binding Keap1. Simply inhibiting autophagy increases the content of p62, contributing to the increased binding of p62 and Keap1 [48] and Nrf2 dissociation in the cytoplasm. Physiologically, Keap1 retains Nrf2 in the cytoplasm and inhibits Nrf2 from entering the nucleus. While, in this study, pterostilbene significantly reduced the protein content of Keap1, which may be the reason why Nrf2 dissociated from Keap1 and entered the nucleus. One of the reasons for the decrease of Keap1 content may be the ubiquitination degradation of Keap1 by p62 [48]. We found that pterostilbene treatment increased the binding of p62 and Keap1. Although the specific mechanism is not clear, this may be one of the reasons why pterostilbene reduces Keap1 content.
In summary, we found for the first time that pterostilbene could alleviate Aβ1-42-induced cognitive dysfunction. Pterostilbene reduced neuronal damage and inhibited oxidative stress in our AD-like model. The mechanism of action may be explained by the fact that pterostilbene activates Nrf2 pathway by increasing the binding of Keap1 and p62. However, it is not clear whether the activation effect of pterostilbene on Nrf2 pathway depends on its direct target action on p62. In our future study, p62 overexpression or knockdown experiments will be carried out to obtain further understanding of the action mechanism and neuroprotective effects of pterostilbene.

References

[1] J. M. Long, D. M. Holtzman, Cell. 2019, 179, 312.
[2] B. De Strooper, E. Karran, Cell. 2016, 164, 603.
[3] T. Jiang, Q. Sun, S. Chen, Prog Neurobiol. 2016, 147, 1.
[4] D. A. Butterfield, D. Boyd-Kimball, J Alzheimers Dis. 2018, 62, 1345.
[5] R. N. Martins, C. G. Harper, G. B. Stokes, C. L. Masters, J Neurochem. 1986, 46, 1042.
[6] H. J. Lee, J. M. Ryu, Y. H. Jung, S. J. Lee, J. Y. Kim, S. H. Lee, I. K. Hwang, J. K. Seong, H. J. Han, Sci Rep. 2016, 6, 36746.
[7] J. Navarro-Yepes, L. Zavala-Flores, A. Anandhan, F. Wang, M. Skotak, N. Chandra, M. Li, A. Pappa, D. Martinez-Fong, L. M. Del Razo, B. Quintanilla-Vega, R. Franco, Pharmacol Ther. 2014, 142, 206.
[8] F. Sesti, S. Liu, S. Q. Cai, Trends Cell Biol. 2010, 20, 45.
[9] M. Yamamoto, T. W. Kensler, H. Motohashi, Physiol Rev. 2018, 98, 1169.
[10] Maruyama, J. Mimura, N. Harada, K. Itoh, Nucleic Acids Res. 2013, 41, 5223. [11] E. Y. Park, H. M. Rho, Mol Cell Biochem. 2002, 240, 47.
[12] P. Canning, F. J. Sorrell, A. N. Bullock, Free Radic Biol Med. 2015, 88, 101.
[13] T. Jiang, B. Harder, M. Rojo de la Vega, P. K. Wong, E. Chapman, D. D. Zhang, Free Radic Biol Med. 2015, 88, 199.
[14] A. M. Rimando, W. Kalt, J. B. Magee, J. Dewey, J. R. Ballington, J Agric Food Chem. 2004, 52, 4713.
[15] D. M. Kambiranda, S. M. Basha, S. J. Stringer, J. O. Obuya, J. J. Snowden, Molecules. 2019, 24.
[16] S. M. Poulose, N. Thangthaeng, M. G. Miller, B. Shukitt-Hale, Neurochem Int. 2015, 89, 227.
[17] Y. S. Chiou, M. L. Tsai, K. Nagabhushanam, Y. J. Wang, C. H. Wu, C. T. Ho, M. H. Pan, J Agric Food Chem. 2011, 59, 2725.
[18] N. Suh, S. Paul, X. Hao, B. Simi, H. Xiao, A. M. Rimando, B. S. Reddy, Clin Cancer Res. 2007, 13, 350.
[19] M. Benlloch, E. Obrador, S. L. Valles, M. L. Rodriguez, J. A. Sirerol, J. Alcacer, J. A. Pellicer, R. Salvador, C. Cerda, G. T. Saez, J. M. Estrela, Antioxid Redox Signal. 2016, 24, 974.
[20] G. Bahn, J. S. Park, U. J. Yun, Y. J. Lee, Y. Choi, J. S. Park, S. H. Baek, B. Y. Choi, Y. S. Cho, H. K. Kim, J. Han, J. H. Sul, S. H. Baik, J. Lim, N. Wakabayashi, S.H. Bae, J. W. Han, T. V. Arumugam, M. P. Mattson, D. G. Jo, Proc Natl Acad Sci USA. 2019, 116, 12516.
[21] M. A. Kostylev, M. D. Tuttle, S. Lee, L. E. Klein, H. Takahashi, T. O. Cox, E. C. Gunther, K. W. Zilm, S. M. Strittmatter, Mol Cell. 2018, 72, 426.
[22] C. P. Figueiredo, J. R. Clarke, J. H. Ledo, F. C. Ribeiro, C. V. Costa, H. M. Melo, A. P. Mota-Sales, L. M. Saraiva, W. L. Klein, A. Sebollela, F. G. De Felice, S. T, J Neurosci. 2013, 33, 9626.
[23] P. S. Frost, F. Barros-Aragao, R. T. da Silva, A. Venancio, I. Matias, E. S. N. M. Lyra, G. C. Kincheski, P. M. Pimentel-Coelho, F. G. De Felice, F. C. A. Gomes, S. T. Ferreira, C. P. Figueiredo, J. R, Cell Death Dis. 2019, 10, 323.
[24] A. L. Dinel, C. Andre, A. Aubert, G. Ferreira, S. Laye, N. Castanon, PLoS One. 2011, 6, e24325.
[25] S. J. Cohen, R. W. Stackman, Jr., Behav Brain Res. 2015, 285, 105.
[26] C. V. Vorhees, M. T. Williams, Nat Protoc. 2006, 1, 848.
[27] D. H. Kim, J. M. Kim, S. J. Park, S. Lee, C. Y. Shin, J. H. Cheong, J. H. Ryu, Neuropsychopharmacology. 2012, 37, 1234.
[28] S. J. Lee, E. Nam, H. J. Lee, M. G. Savelieff, M. H. Lim, Chem Soc Rev. 2017, 46, 310.
[29] D. H. Small, S. S. Mok, J. C. Bornstein, Nat Rev Neurosci. 2001, 2, 595.
[30] J. J. Palop, L. Mucke, Nat Rev Neurosci. 2016, 17, 777.
[31] A. L. Barth, A. Ray, Neuron. 2019, 104, 37.
[32] J. W. Lustbader, M. Cirilli, C. Lin, H. W. Xu, K. Takuma, N. Wang, C. Caspersen, X. Chen, S. Pollak, M. Chaney, F. Trinchese, S. Liu, F. Gunn-Moore, L. F. Lue, D. G. Walker, P. Kuppusamy, Z. L. Zewier, O. Arancio, D. Stern, S. S. Yan, H. Wu, Science. 2004, 304, 448.
[33] J. H. Nam, K. W. Park, E. S. Park, Y. B. Lee, H. G. Lee, H. H. Baik, Y. S. Kim, S. Maeng, J. Park, B. K. Jin, Antioxid Redox Signal. 2012, 16, 1369.
[34] Z. Fu, J. Yang, Y. Wei, J. Li, Food Funct. 2016, 7, 1014.
[35] Y. Yang, J. Wang, Y. Li, C. Fan, S. Jiang, L. Zhao, S. Di, Z. Xin, B. Wang, G. Wu, X. Li, Z. Li, X. Gao, Y. Dong, Y. Qu, Mol Neurobiol. 2016, 53, 2339.
[36] H. Liu, X. Wu, J. Luo, X. Wang, H. Guo, D. Feng, L. Zhao, H. Bai, M. Song, X. Liu, W. Guo, X. Li, L. Yue, B. Wang, Y. Qu, Front Immunol. 2019, 10, 2408.
[37] P. Nioi, T. Nguyen, P. J. Sherratt, C. B. Pickett, Mol Cell Biol. 2005, 25, 10895.
[38] Y. Katoh, K. Itoh, E. Yoshida, M. Miyagishi, A. Fukamizu, M. Yamamoto, Genes Cells. 2001, 6, 857.
[39] L. E. Tebay, H. Robertson, S. T. Durant, S. R. Vitale, T. M. Penning, A. T. Dinkova-Kostova, J. D. Hayes, Free Radic Biol Med. 2015, 88, 108.
[40] A. I. Rojo, M. Pajares, P. Rada, A. Nunez, A. J. Nevado-Holgado, R. Killik, F. Van Leuven, E. Ribe, S. Lovestone, M. Yamamoto, A. Cuadrado, Redox Biol. 2017, 13, 444.
[41] A. Singh, S. Venkannagari, K. H. Oh, Y. Q. Zhang, J. M. Rohde, L. Liu, S. Nimmagadda, K. Sudini, K. R. Brimacombe, S. Gajghate, J. Ma, A. Wang, X. Xu, S. A. Shahane, M. Xia, J. Woo, G. A. Mensah, Z. Wang, M. Ferrer, E. Gabrielson, Z. Li, F. Rastinejad, M. Shen, M. B. Boxer, S. Biswal, ACS Chem Biol. 2016, 11, 3214.
[42] K. Itoh, T. Chiba, S. Takahashi, T. Ishii, K. Igarashi, Y. Katoh, T. Oyake, N. Hayashi, K. Satoh, I. Hatayama, M. Yamamoto, Y. Nabeshima, Biochem Biophys Res Commun. 1997, 236, 313.
[43] J. D. Hayes, A. T. Dinkova-Kostova, Trends Biochem Sci. 2014, 39, 199.
[44] Y. Wang, C. Yang, N. A. H. Elsheikh, C. Li, F. Yang, G. Wang, L. Li, Aging (Albany NY). 2019, 11, 5535.
[45] N. Hettiarachchi, M. Dallas, M. Al-Owais, H. Griffiths, N. Hooper, J. Scragg, J. Boyle, C. Peers, Cell Death Dis. 2014, 5, e1569.
[46] N. T. Hettiarachchi, J. P. Boyle, M. L. Dallas, M. M. Al-Owais, J. L. Scragg, C. Peers, Cell Death Dis. 2017, 8, e2884.
[47] E. Bhakkiyalakshmi, D. Shalini, T. V. Sekar, P. Rajaguru, R. Paulmurugan, K. M. Ramkumar, Br J Pharmacol. 2014, 171, 1747.
[48] M. Komatsu, H. Kurokawa, S. Waguri, K. Taguchi, A. Kobayashi, Y. Ichimura, Y. S. Sou, I. Ueno, A. Sakamoto, K. I. Tong, M. Kim, Y. Nishito, S. Iemura, T. Natsume, T. Ueno, E. Kominami, H. Motohashi, K. Tanaka, M. Yamamoto, Nat