Ezetimibe Attenuates Oxidative Stress and Neuroinflammation via the AMPK/Nrf2/TXNIP Pathway after MCAO in Rats

Oxidative stress and neuroinflammation play essential roles in ischemic stroke-induced brain injury. Previous studies have reported that Ezetimibe (Eze) exerts antioxidative stress and anti-inflammatory properties in hepatocytes. In the present study, we investigated the effects of Eze on oxidative stress and neuroinflammation in a rat middle cerebral artery occlusion (MCAO) model. One hundred and ninety-eight male Sprague-Dawley rats were used. Animals assigned to MCAO were given either Eze or its control. To explore the downstream signaling of Eze, the following interventions were given: AMPK inhibitor dorsomorphin and nuclear factor erythroid 2-related factor 2 (Nrf2) siRNA. Intranasal administration of Eze, 1 h post-MCAO, further increased the endogenous p-AMPK expression, reducing brain infarction, neurologic deficits, neutrophil infiltration, microglia/macrophage activation, number of dihydroethidium- (DHE-) positive cells, and malonaldehyde (MDA) levels. Specifically, treatment with Eze increased the expression of p-AMPK, Nrf2, and HO-1; Romo-1, thioredoxin-interacting protein (TXNIP), NOD-like receptor protein 3 (NLRP3), Cleaved Caspase-1, and IL-1β were reduced. Dorsomorphin and Nrf2 siRNA reversed the protective effects of Eze. In summary, Eze decreases oxidative stress and subsequent neuroinflammation via activation of the AMPK/Nrf2/TXNIP pathway after MCAO in rats. Therefore, Eze may be a potential therapeutic approach for ischemic stroke patients.


Introduction
Stroke accounts for 10% of all deaths worldwide [1]. The pathophysiology of stroke is composed of complex sequelae of cellular processes: oxidative stress, apoptosis, blood-brain barrier disruption, and inflammation [2][3][4][5][6][7]. Although the majority of ischemic strokes occur from embolic arterial occlusion, oxidative stress and neuroinflammation play significant roles in transient ischemic stroke and the reperfusion process [8,9]. For example, neuroinflammatory responses to ischemic stroke are characterized by astrocyte activation and microglial resident, peripheral leukocyte infiltration, and proinflammatory mediator release. Moreover, infiltrated neutrophils and activated microglia produce free radicals and oxidants that damage the central nervous system tissue, leading to long-term disabilities and death in stroke patients [10]. Therefore, developing a protective strategy against oxidative stress and subsequent neuroinflammation may be an effective approach for the treatment of ischemic stroke patients.
Ezetimibe (Eze) is a new lipid-lowering agent that inhibits Niemann-Pick disease type C1-like 1-(NPC1L1-) dependent cholesterol absorption [11,12]; however, studies have shown Eze to exert pleiotropic effects independent of NPC1L1 [13][14][15]. For example, we have previously demonstrated that intranasal administration of Eze attenuated neuronal apoptosis through the activation of AMPK-dependent autophagy after MCAO in rats [16]. In a rat liver ischemia/reperfusion model, Eze therapeutically exerted antioxidation effects by modulating glutathione and glutathione peroxidase [14]. In an Alzheimer mouse model, researchers reported that treatment with Eze reduced the memory dysfunctions associated with dementia [17]. Of importance, a randomized and placebo-controlled clinical study reported that treatment with Eze prevented the progression of the deleterious symptoms associated with acute stroke [18]. Lastly, in hepatocyte mouse models, studies have shown that the anti-inflammatory effects of Eze were dependent on AMPK autophagic induction and NLRP3 inflammasome inhibition [19,20].
Therefore, in the current study, we assessed the hypothesis that intranasal administration of Eze may attenuate oxidative stress and neuroinflammation in a rat model of MCAO via the AMPK/Nrf2/TXNIP pathway.

Materials and Methods
2.1. Animals. All experiments were approved by the Institutional Animal Care and Use Committee of Loma Linda University in accordance with the NIH Guide for the Care and Use of Laboratory Animals (NIH Publications No. 8023, revised 1978) and the ARRIVE2009 Guidelines for Reporting Animal Research [27]. A total of 198 adult male Sprague-Dawley rats (260-280 g) were obtained from the Experimental Animal Center of Loma Linda University. Rats were housed in a controlled humidity and temperature room with a 12 h light/dark cycle and free access to water and food.

MCAO
Model. The transient MCAO model was used in male Sprague-Dawley rats as previously described [28]. Briefly, anesthesia was induced intraperitoneally with a mixture of ketamine (80 mg/kg, K2573; Sigma-Aldrich, St. Louis, MO, USA) and xylazine (10 mg/kg, X1126; Sigma-Aldrich, St. Louis, MO, USA). Next, atropine was administered (0.1 mg/kg) subcutaneously. The depth of anesthesia was checked by pinch-paw reflex. The right common carotid artery (CCA), internal carotid artery (ICA), and external carotid artery (ECA) were surgically exposed. The ECA was ligated, and a 4-0 nylon suture with a silicon tip was then inserted through the ECA stump into the ICA, occluding the MCA, approximately 18 to 22 mm from the insertion point. After 2 h of MCAO, the suture was removed to begin reperfusion. Sham rats underwent the same protocol without occlusion of the MCA.

Experimental
Design. Animals were divided into groups for three experimental studies in a randomized fashion by generating random numbers using Excel, and experiments were performed in a blinded manner (Figure 1): the experimental groups and sample size are listed in Table 1
To explore the effects of Eze treatment on neutrophil infiltration and microglia/macrophage activation at 24 h after MCAO, 18 rats were randomly divided into sham, MCAO +vehicle, and MCAO+Eze 500 μg/kg groups (n = 6 per group). Immunofluorescence staining of myeloperoxidase (MPO) and Iba-1 was performed, and quantitative analyses of MPO and Iba-1-positive cells were counted in the ischemic penumbra at 24 h after MCAO. The expression of MPO and Iba-1 among the three groups was measured by western blot at 24 h after MCAO.
To explore the effects of Eze treatment on oxidative stress at 24 h after MCAO, another 18 rats from sham, MCAO +vehicle, and MCAO+Eze 500 μg/kg were used to measure malonaldehyde (MDA) levels (n = 6 per group). Dihydroethidium (DHE) staining was performed, and the number of DHE-positive cells was counted in the ischemic penumbra at 24 h after MCAO (shared with the immunofluorescencestained samples). The expression of Romo-1 among the three groups was measured by western blot at 24 h after MCAO.

Groups
Additionally, to evaluate the effects of AMPK inhibition via dorsomorphin, 24 rats were randomly divided into four groups: naive+DMSO, naive+dorsomorphin, naive +Eze+DMSO, and naive+Eze+dorsomorphin. The expression of phosphorylated AMPK was evaluated by western blot.

Intranasal Administration of Eze.
Intranasal administration was performed as previously described [29]. Briefly, rats were treated 1 h after MCAO with DMSO or Eze (250 μg/kg, 500 μg/kg, and 1 mg/kg dissolved in 10%DMSO, purity ≥ 98%, SML-1629, Sigma-Aldrich, St. Louis, MO, USA) in a supine position under 2% isoflurane anesthesia. A total volume of 25 μl was delivered into the bilateral nares, alternating one naris at a time, every 5 min over a period of 20 min.

Intracerebroventricular Injection.
Intracerebroventricular (i.c.v.) administration was performed as previously described [28]. Briefly, rats were placed in a stereotaxic apparatus under 2.5% isoflurane anesthesia. A scalp incision was made along the midline, and a 1 mm burr hole was drilled into the skull. The stereotactic i.c.v. injection site was relative to the bregma: anteroposterior 1 mm, right lateral 1.5 mm, and depth 3.5 mm. The AMPK-specific inhibitor, dorsomorphin (0.1 μmol, purity ≥ 98%, P5499, Sigma-Aldrich, St. Louis, MO, USA), was dissolved in 20% DMSO in PBS, and 10 μl was delivered into the ipsilateral ventricle with a Hamilton syringe (Microliter 701; Hamilton Company, USA) 30 min before MCAO [30]. The same volume of DMSO was used as a negative control. Nrf2 siRNA (SR508224; OriGene, Rockville, MD, USA) and scrambled siRNA (SR30004; OriGene) were prepared at 500 pmol in RNAse free suspension buffer and administered (5 μl of the siRNAs) 48 h before MCAO [31]. Lastly, the burr hole was sealed with bone wax, and the dissection was sutured.
2.6. Neurobehavioral Function Assessment. Neurobehavioral function was assessed with the modified Garcia and beam walking tests by an independent, blinded researcher at 24 h after MCAO, as previously described [32]. To understand the effect of neuronal lesions on sensorimotor areas, the modified Garcia test was used to measure hemiplegia, motor performance deficits, and abnormal postures [33]. The modified Garcia scoring system consisted of 6 tests covering spontaneous activity, symmetry in the movement of four limbs, forepaw outstretching, climbing, body proprioception, and response to vibrissae touch, with a maximum score of 18, higher scores indicating better performance. In addition, to better asses cortical motor injury, the beam walking test was used to measure dysfunction in memory, motivation, attention, somatomotor, and locomotor functions [34]. The beam walking test was performed with a 0-5-point scale as previously described [35].

Cerebral Infarction Volume Assessment.
Under deep anesthesia, animals were perfused with cold PBS (0.1 M, pH 7.4) as previously described [36]. Brains were removed and coronally sliced into 2 mm thick sections. Brain slices were incubated in 2% 2,3,5 triphenyltetrazolium chloride (TTC, Sigma-Aldrich, St. Louis, MO, USA) for 15 min at 37°C. The infarcted brain tissue appeared white, whereas the noninfarcted region appeared red. The infarct and total hemispheric areas of each slice were measured using ImageJ (ImageJ 1.5; NIH, Bethesda, MD, USA). The area of each slice was calculated using the following formula: ððarea of contralateral − area of noninf arcted ipsilateral tissueÞ/2 * ðarea of contralateralÞÞ * 100%. The area was calculated for each slice, and the average was taken to represent the percentage of infarcted area for that animal [37,38].

Immunofluorescence
Staining. Twenty-four hours after MCAO, under deep anesthesia, rats were perfused with icecold PBS and then 10% formalin. The brains were removed and fixed in formalin and then dehydrated with 30% sucrose. Next, brain samples were snap-frozen and cut into 10 μm thick coronal sections using a cryostat (LM3050S; Leica Microsystems, Bannockburn, Germany). Immunofluorescence staining was performed as previously described [39]. Briefly, brain samples were incubated overnight at 4°C with primary antibodies including anti-Iba-1 (1 : 100, Abcam, ab5076) and anti-MPO (1 : 500, Abcam, ab65871). The sections were then incubated with the appropriate fluorescence-conjugated secondary antibodies (1 : 200, Jackson ImmunoResearch) for 1 h at room temperature and then visualized with a fluorescence microscope (DMi8, Leica Microsystems, Germany). 2.9.2. DHE Staining. Dihydroethidium (DHE) staining was performed as previously described [40]. Briefly, 10 μm thick frozen brain sections were incubated with 2 μmol/l fluorescent dye DHE (D1168, Thermo Fisher Scientific, Waltham, MA, USA) at 37°C for 30 min in a humidified chamber and protected from light. The DHE-positive cells were observed under a fluorescence microscope (DMi8, Leica Microsystems, Germany), and the positive cells were counted by using ImageJ software (ImageJ 1.5; NIH, Bethesda, MD, USA).

Western Blot Analysis.
After TTC staining at 24 h after MCAO, brain slices were separated into the contralateral and ipsilateral hemispheres, flash frozen in liquid nitrogen, and then stored at −80°C freezer. Western blot was performed as previously described [41]. Nuclear proteins were extracted from tissue homogenates using a nuclear extraction kit (ab113474, Abcam, Cambridge, MA, USA) according to 4 Oxidative Appropriate secondary antibodies (1 : 4000, Santa Cruz Biotechnology) were selected for the incubated membrane the following day for 1 h at room temperature. Immunoblots were then visualized with an ECL Plus chemiluminescence reagent kit (RPN3243; Amersham Bioscience, Bensenville, IL, USA) and quantified with optical methods using the ImageJ software (ImageJ 1.5; NIH, Bethesda, MD, USA). The results were normalized using β-actin or Lamin B1 as an internal control.
2.11. Statistical Analysis. All data were expressed as the mean and standard deviation (mean ± SD). Statistical analysis was performed with GraphPad Prism 6 software (La Jolla, CA, USA). Before analysis, the Shapiro-Wilk test was used to test normality. For parametric data, one-way ANOVA with post hoc Tukey test was used to test for differences among groups. p < 0:05 was considered statistically significant.

Mortality and Exclusion.
Of the 198 total animals used, 144 were subjected to MCAO, and the overall mortality was 16.7% (24/144). No significant difference was observed in mortality between the MCAO groups (p > 0:05). No rats died in the sham and naive groups. Six animals were excluded from this study due to no infarction volume after MCAO (Table 1).

Time-Course of Endogenous p-AMPK after MCAO.
The expression of endogenous p-AMPK in the ipsilateral/right cerebral hemispheres after MCAO was assessed by western blot. As shown in Figure 2(a), the expression of p-AMPK increased at 6 h, reaching its peak at 24 h, and decreased by 72 h after MCAO compared to the sham group (p < 0:05). After Eze treatment, the endogenous p-AMPK expression further increased at 6, 12, 24, and 72 h after MCAO compared to the sham group (p < 0:05, Figure 2(b)).

Eze Treatment Reduced Brain Infraction and
Ameliorated Neurobehavioral Deficiency at 24 h after MCAO. In the vehicle group, infarction volume was significantly increased, while Garcia scores were significantly decreased compared to the sham group (p < 0:05, Figure 3). Intranasally administered Eze 500 μg/kg and Eze 1 mg/kg significantly reduced the infarction volume and improved neurological outcomes at 24 h after MCAO compared to the vehicle group (p < 0:05, Figures 3(a)-3(c)). With no additional therapeutic effects observed with Eze 1 mg/kg treatments, we used Eze 500 μg/kg for the subsequent studies. There were no significant differences in beam walking scores between the MCAO groups ( Figure 3(d)).

Eze Treatment Inhibited Neutrophil Infiltration and Microglia/Macrophage Activation at 24 h after MCAO.
MPO levels were used to assess neutrophil infiltration [8]. Iba-1 levels were used to evaluate microglia/macrophage activation in the brain tissue [42]. Immunofluorescence staining and western blot were used to evaluate whether the anti-inflammatory effects of Eze were caused by a reduction of neutrophil infiltration or microglia/macrophage activation in the ischemic penumbra at 24 h after MCAO. The immunofluorescence staining results showed that Eze treatment significantly decreased the number of MPO and Iba-1positive cells in the ischemic penumbra compared to the   Figures 5(a)-5(c)). In the vehicle group, Romo-1, a marker of oxidative stress [43], was increased compared to that in the sham group (p < 0:05, Figure 5(d)), while treatment with Eze significantly reduced the expression of Romo-1 at 24 h after MCAO (p < 0:05, Figure 5(d)). Nrf2 is a transcription factor involved in the endogenous antioxidant stress system [26]. Our results showed no significant difference in total-Nrf2; however, the nuclear-Nrf2 expression increased in the vehicle group compared to the sham group (Figures 5(e) and 5(f)). Treatment with Eze significantly increased total-Nrf2 expression and further increased the expression of nuclear-Nrf2 at 24 h after MCAO compared to the vehicle group (p < 0:05, Figures 5(e) and 5(f)).

Eze Treatment Attenuated Oxidative Stress and
Neuroinflammation via Activation of the AMPK/Nrf2/TXNIP Pathway after MCAO in Rats. In the naive rat, the endogenous p-AMPK expression was significantly increased in the ipsilateral cortex after Eze treatment. Inhibition of AMPK with dorsomorphin significantly decreased the expression of p-AMPK in both naive and Eze-treated animals (p < 0:05, Figure S1).

Discussion
In the present study, we demonstrated that Eze attenuated oxidative stress and neuroinflammation after MCAO. We made the following novel observations: (1) Eze further increased the endogenous p-AMPK expression after MCAO; (2) intranasal administration of Eze significantly reduced the infarction volume and improved neurological outcomes after MCAO; (3) Eze treatment inhibited neutrophil infiltration, microglia/macrophage activation, and oxidative stressassociated injuries in the ischemic penumbra regions after MCAO; (4) the antioxidative stress and anti-inflammatory effects of Eze were facilitated through the increased expression of p-AMPK, Nrf2, and HO-1, while Romo-1, TXNIP, NLRP3, Cleaved Caspase-1, and IL-1β were reduced following MCAO; and (5) pretreatment with dorsomorphin and Nrf2 siRNA reversed the beneficial effects of Eze on brain infarction, neurobehavioral function, and inflammatory protein expression. Taken together, our findings suggest that Eze attenuated oxidative stress and neuroinflammatory sequelae of MCAO via activation of the AMPK/Nrf2/TXNIP signaling pathway (Figure 8).
Accumulating scientific evidence suggests that neuroinflammation and oxidative stress are the main pathological processes responsible for the impairment of neurological function in MCAO [26,29,44,45]. Clinically, Eze, a NPC1L1 inhibitor, is mainly used as a treatment for hypercholesterolemia; however, in addition to its lipid-lowering activity, several studies have reported that Eze may attenuate ischemic-related oxidative stress and inflammation [14,19]. For example, in a rat liver ischemia/reperfusion model, Eze was reported to attenuate oxidative radicals, modulate NO production, and increase eNOS activity [13]. Independent of cholesterol regulation, other studies have shown that treatment with Eze improved renal injury outcomes in nondiabetic chronic kidney disease patients with dyslipidemia, which may be explained by the asymmetric dimethylargininelowering and antioxidative effects of Eze [46]. In a clinical study evaluating the neurological deterioration after embolic stroke resulting from atrial fibrillation in older patients, Lappegard et al. demonstrated that anti-inflammatory therapy with Eze may ameliorate the deterioration of neurocognitive function and loss of volume in cerebral areas [47]. Consistent with these findings, we demonstrated that  Oxidative Medicine and Cellular Longevity treatment with Eze reduced brain infarction, neutrophil infiltration, microglia/macrophage activation, MDA levels, and DHE-positive cell numbers. Specifically, Eze treatment reduced the protein expression of Romo-1 (oxidative stress marker) and IL-1β (inflammatory marker). Congruently, neurological outcomes were improved after Eze administration at 24 h after MCAO. Given that studies have reported Eze treatment to increase AMPK phosphorylation, novel and unique mechanistic endpoints are being investigated around AMPK regulation [16,19,20]. According to literature, Eze increases the oxygen consumption rate (OCR) and decreases the amount of ATP [48], which causes an elevation of ADP/ATP ratio that subsequently activates AMPK [20]. Therefore, Eze maintains cellular energy by regulating ATP consumption and generation via phosphorylation of AMPK [19]. Confirming this mechanistic link, our results showed that the endogenous p-AMPK expression was acutely increased after MCAO, and it was further increased after Eze treatment. To confirm that Eze regulates AMPK, a specific inhibitor of AMPK was administered with Eze. This intervention reversed the neuroprotective effects of Eze, increasing the infraction volume and neurobehavioral deficits; reducing the expression of p-AMPK, Nrf-2, and HO-1; and upregulating the expression of TXNIP, NLRP3, Cleaved Caspase-1, and IL-1β. This mechanistic study confirms a link between Eze and AMPK.
Nrf2 has traditionally been involved in upregulating antioxidant systems to reduce oxidative stress in the brain [49]. Under oxidative stress conditions, Nrf2 dissociates from Kelch-like ECH-associated protein 1 (Keap1) and translocates  Oxidative Medicine and Cellular Longevity into the nucleus to bind to antioxidant response elements (ARE), which activates downstream antioxidant defense enzymes, such as HO-1 [50]. NLRP3 is an inflammasome protein complex located in the cell that binds to pro-Caspase-1 to cause neuronal apoptosis and inflammation in ischemic injuries [51]. Activation of AMPK and downstream inhibition of both NLRP3 inflammasomes and IL-1β mediates the anti-inflammatory effects of Eze [19]. Therefore, inhibition of NLRP3 inflammasomes via the AMPK pathway is neuroprotective in ischemic stroke [45]. Of importance, TXNIP, a redox-regulated protein, can bind to and activate the NLRP3 inflammasome in response to the oxidative stress associated with stroke [23], and when TXNIP is inhibited, Nrf2 acts as a negative regulator of the NLRP3 inflammasome [26]. Our results showed that Eze significantly enhanced the phosphorylation of AMPK; increased the expression of Nrf2 and HO-1; and subsequently decreased the expression of TXNIP, NLRP3, Cleaved Caspase-1, and IL-1β. Similarly,  when knocking down the endogenous Nrf2 with Nrf2 siRNA, the protective effects of Eze were reversed: the intervention increased infarction volumes and neurobehavioral deficits, reducing the activation of Nrf2 and HO-1, with an associated increase in the levels of TXNIP, NLRP3, Cleaved Caspase-1, and IL-1β. Taken together, we can conclude that Eze exerts antioxidative and anti-inflammatory effects via activation of the AMPK/Nrf2/TXNIP signaling pathway.
There were some limitations in our study. First, due to the limited nature of this pilot study, we only evaluated a one-time window target for Eze treatment after MCAO. Second, our results do not fully exclude the possibility of alternative pathways that modulate the inflammasome pathway; thus, further research will need to investigate the relationship between other inflammasome activators (e.g., NF-κB [52]) and the proposed pathway of Eze to fully exclude or incorporate alternative pathways. Finally, we have previously reported that Eze attenuates neuronal apoptosis via AMPK-induced autophagy [16]. In addition to these downstream targets, in the present study, Eze also decreased oxidative stress and subsequent neuroinflammation via activation of the AMPK/Nrf2/TXNIP pathway. Therefore, Eze may exert these effects by increasing the oxygen consumption rate, which then decreases the amount of ATP and phosphorylates AMPK, activating other downstream targets such as autophagy. However, since lowering cholesterol may also contribute to AMPK activation [16], NPC1L1 may have played a role in our proposed pathway. In sum, further research is needed to better understand the pleiotropic effects of Eze.

Conclusion
In summary, our findings suggest that intranasal administration of Eze reduced brain infarction, oxidative stress, and neuroinflammation, while neurological outcomes were improved after transient MCAO in rats. Mechanistically, the neuroprotective effects of Eze were mediated through activation of the AMPK/Nrf2/TXNIP pathway. This research supports the continued investigation of Eze as a potential therapy for the treatment of patients with ischemic stroke.

Data Availability
All data are available upon request.

Conflicts of Interest
The authors declare that they have no conflict of interest.

Authors' Contributions
Jing Yu and Wen-na Wang contributed equally to this manuscript.

12
Oxidative Medicine and Cellular Longevity