Vitexin reduces epilepsy after hypoxic ischemia in the neonatal brain via inhibition of NKCC1

Sprague-Dawley (SD) rats (postnatal day 7) were provided by the Animal Biosafety Level 3 Laboratory (ABSL-3, Wuhan University, China). All rats were housed at the standard laboratory animal facility (25 ± 2 °C, 12-h light/dark cycle) with access to food and water ad libitum in individual cages.

The rats (n = 128) were randomly assigned to four groups (n = 32 rats each): hypoxia-ischemia group (abbreviated: HI group), rats underwent the ligation of the right common carotid artery and had hypoxic injury, the details of which are discussed in the following experimental methods; sham-operated group (abbreviated: sham group), rats underwent the same operation, without ligation of the right common carotid artery nor the hypoxia treatment; HI+bumetanide group (abbreviated: HI+Bum group), bumetanide (0.5 mg/kg, intraperitoneal injection, i.p., Cat. sc-200727, Santa Cruz) was diluted with saline and administered 5 min after HI; bumetanide, an inhibitor of NKCC1, served as positive group in our study. HI+vitexin group (abbreviated: HI+Vit group), vitexin (45 mg/kg, intraperitoneal injection, i.p., Cat. E-0310, Tauto Biotech) was diluted with saline and administered 5 min after HI.

All procedures of the neonatal HI model in this study were based on the Rice-Vanucci study [47, 48]. Animal surgical procedures and experimental protocols were reviewed and approved by the Committee on the Ethics of Animal Experiments of Wuhan University (China). Briefly, P7 SD rats (13–19 g, equal males and females have been chosen for each group) were anesthetized by inhalation of isoflurane. The right common carotid artery (CCA) was exposed and ligated with 5-0 surgical silk. After ligation for 2 h, the rat pups were then placed in an airtight, transparent chamber at 37 °C and exposed to 8% O₂ in N₂ for 3 h to create hypoxic injury after brain ischemia. Thereafter, the littermates were returned to their mothers when they can move freely. Meanwhile, sham-operated rats were subjected to isolation and stringing of vessels without ligation and subsequent ischemia. The rats were sacrificed at 24 h after HI insult, and their ipsilateral hemispheres were collected for follow-up experiments.

Lines of rat brain microvascular endothelial cells (RBMECs) were purchased from BeNa Culture Collection (Cat. BNCC337880). RBMEC cultures were expanded and maintained in 89% basal medium, 10% fetal bovine serum (FBS), and 1% penicillin/streptomycin solution (P/S). They were then incubated in a humidified atmosphere containing 5% CO₂ at 37 °C.

OGD was conducted as described previously. Briefly, RBMECs were grown in complete growth media as monolayers in cell culture incubator (95% O₂ and 5% CO₂ at 37 °C). To initiate OGD in vitro, the cells at 4 DIV (days in vitro) were washed with 1× PBS three times, switched to OGD medium (serum- and glucose-free DMEM), and placed in a hypoxic/anoxic chamber (1% O₂, 5% CO₂, and 94% N₂ at 37 °C) to mimic OGD injury in incubator. Following the OGD carried out for 3.5 h, cells were removed from the anaerobic chamber, and the OGD medium in the cultures was then changed back to DMEM medium for an additional 24-h reperfusion under normal conditions. At the same time, control glucose-containing cultures remained in a regular incubator (5% CO₂ and 95% O₂). The supernatants and cell extracts were collected after OGD for the following experiments.

Membrane protein was extracted according to the manufacturer’s protocol from cultured RBMECs and the ischemic penumbra of the rat cortex with a Membrane and Cytosol Protein Extraction Kit (Cat. P0033, Beyotime, Shanghai, China). The concentrations were determined with a BCA protein assay kit (Cat. P0012, Beyotime, Shanghai, China). Samples were denatured for 10 min at 100 °C and frozen at − 20 °C before assay. Approximately 20 μl of the samples were separated by 8–12% SDS–PAGE gel and then transferred to polyvinylidene fluoride (PVDF) membranes. Subsequently, the membranes were blocked in 5% BSA for 1 h at room temperature following incubation with primary antibodies overnight at 4 °C. Dilutions for primary and secondary antibodies were listed in Table 1. Membranes were washed three times in TBST and specific binding was visualized by ECL reaction. The density of bands was detected using an imaging densitometer (Bio-Rad, Foster City, CA, USA), and the gray value of the bands was quantified using ImageJ Software (version 1.41).

Total RNA was extracted from the ischemic cerebral cortices (n = 6 per group) for the detection of NKCC1, IL-1β, IL-6, and TNF mRNA at 24 h after HI using Trizol Reagent (Invitrogen Life Technologies Corporation, USA) according to the manufacturer’s protocol. The quantity of total RNA was measured with a UV spectrophotometer (Biochrom Ltd., UK). Next, reverse transcription was performed using a cDNA synthesis kit (TaKaRa Biotechnology). Briefly, 2 μl of total RNA was combined with 4 μl of 5× Prime Script® Buffer. RNase Free ddH₂O was added to 20 μl, after which the mixture was heated at 37 °C for 15 min and then 85 °C for 5 s. Quantitative PCR was performed with SYBR-Green premix (Trans Gen Biotech) at the following conditions (denaturing at 95 °C for 10 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s) and detected by a real-time PCR system (Step One, Applied Biosystems). The expression of target genes was measured in triplicate and normalized to β-actin as an internal control. The ΔΔCt values of each group were analyzed, and mRNA expression levels were normalized to 2−ΔΔCt. Primers are listed in Table 2 (Sangon Biotech, Shanghai, Co., Ltd.).

The protein concentrations of IL-1β, IL-6, and TNF in each group of animals were quantified using an ELISA kit (purchased from 4A Biotech Co., Ltd) according to the manufacturer’s instruction. Absorbance at 450 nm was recorded and the concentration of the target protein was calculated according to the standard curve and normalized against the protein of the samples. Result was expressed as pg/mg protein.

Immunofluorescence staining was carried out to detect ZO-1 and CD31 expression in peri-ischemic brain tissue or ZO-1, NKCC1, and F-actin expression in RBMECs.

Rats in each group were deeply anesthetized with 10% chloral hydrate and transcardially perfused first with PBS and then fixed in 4% paraformaldehyde solution at room temperature, dehydrated, and embedded in paraffin at 24 h after HI. Post-fixation, the brains were removed and cryoprotected in 20% sucrose and 30% sucrose solutions and dehydrated by 30% sucrose for 72 h at 4 °C. Serial coronal sections (5-μm-thick with injury epicenter located centrally) prepared with cryotome (Leica, Wetzlar, Germany) were used for immunofluorescence labeling. Sections were incubated with a blocking solution (5% FBS) for 30 min at 37 °C. The tissue slices were then incubated overnight with the anti-ZO-1 and anti-CD31 antibodies. On the following day, the sections protected from light were washed and subsequently incubated with secondary antibodies Cy3-conjugated Goat Anti-Rabbit IgG (H+L) and Alexa Fluor® 488 Conjugates for 2 h at 37 °C. Dilutions for antibodies were listed in Table 3. Images were obtained using a confocal microscope (Leica-LCS-SP8-STED).

For the assessment of ZO-1, NKCC1, and F-actin expression in RBMECs, cells were fixed with methanol, washed with PBS-T, and incubated at 4 °C with anti-NKCC1, anti-F-actin, and anti-ZO-1 antibodies. Subsequently, cells were washed with PBS-T before incubation with mixtures of secondary antibodies: Alexa Fluor® 488 Conjugates, Cy3-conjugated Rabbit Anti-Goat IgG (H+L), and Cy3-conjugated Goat Anti-Rabbit IgG (H+L) diluted in blocking buffer for 2 h in the dark at room temperature. Dilutions for antibodies were also listed in Table 3. The cells were washed three times in PBS-T before they were mounted using DAPI. Finally, cellular co-localization was captured using confocal microscope (Leica-LCS-SP8-STED).

To evaluate the expression of NKCC1and the severity of HI-induced inflammation, the slices from paraffin-embedded tissues were subjected to immunohistochemical staining for NKCC1 and myeloperoxidase (MPO), respectively. MPO is a representative marker of neutrophils and is an important index for evaluating the severity of inflammation. It also reflects the extent of inflammation in brain tissue [49]. The sections were washed with PBS and blocked in 5% BSA for 2 h. Thereafter, the primary antibody goat anti-NKCC and rabbit anti-MPO polyclonal antibody were applied. After being rinsed with PBS, the sections were incubated with corresponding secondary antibodies, and nuclei were stained with DAPI. Images were obtained using a confocal laser-scanning microscope (Leica-LCS-SP8-STED).

For visualization of cytoskeleton F-actin, the cultured RBMECs were processed for direct confocal imaging with FITC-conjugated phalloidin (Cat. P5282, Sigma), which specifically combines with F-actin. Monolayers were rinsed with PBS solution, fixed with 4% paraformaldehyde for 30 min at 4 °C, and permeabilized with 0.3% Triton X-100 for 30 min. The cells were incubated with FITC-phalloidin (5 μg/ml, Sigma) for 1 h at room temperature in the dark. Cells were then counterstained with DAPI for nuclear labelling. Visualization was performed with a Leica-LCS-SP8-STED confocal microscope (Leica Microsystems).

N-(ethoxycarbonylmethyl)-6-methoxyquinolinium bromide (MQAE), a chloride-sensitive fluorescent indicator inversely related to intracellular chloride ion concentration, was used to detect [Cl⁻] _i as previously described [50]. This dye detects the ion via diffusion-limited collisional quenching. The cultured RBMECs were then incubated with 10 mM MQAE (Cat. E3101, Invitrogen) in a Kreb HEPES-buffered isotonic solution [DMEM, 0.1% BSA, 10 mM 4-(2-hydroxyethyl)-1 piperazine-ethanesulfonic acid (HEPES), pH 7.5] for 1 h at 37 °C in the dark. Subsequently, cells were washed with DMEM three times. Fluorescence was excited every 60 s at 340 nm, and emission fluorescence at 460 nm was recorded. Images were collected and analyzed with the Image-Pro Plus 6.0 image-processing software.

The number of NKCC1⁺ RBMECs was analyzed by flow cytometry. Single-cell suspensions were harvested at 24 h after OGD and flushed with PBS. The suspensions were then stained with fluorescently labeled antibodies, anti-NKCC1 antibody, and FITC rabbit anti-goat antibodies. The staining was performed according to the manufacturer’s instructions. Flow cytometric analysis was performed using Flow Jo (version 9.2; Tree Star Inc.).

For the recording of the electrical activity of the brain, electrodes were implanted in rats (n = 15 per group) on P28. The rats were anesthetized with 10% chloral hydrate (3 ml/kg) and then fixed into the stereotaxic apparatus. The electrodes were bipolar twisted silver steel and embedded in the skull with dental cement. These electrodes were implanted into the bilateral hippocampal CA3 (3.5 mm posterior to bregma, 3.5 mm lateral, 3.5 mm ventral to the dura mater), according to the coordinates derived from the atlas of Paxinos and Watson. Spontaneous EEG seizures in the dentate gyrus were recorded in freely moving animals after 3 days of recovery, which were then individually placed in a cage and connected to a neurophysiology workstation (AD Instruments Lab Chart 8) through a flexible cable that prevents twisting [51]. The frequency and mean duration of these spontaneous EEG seizures during an EEG recording session were examined for 2 h. The EEG signals were digitized with Lab Chart software (AD Instruments). Seizure severity was classified into five levels by Racine’s scale [52]: I, facial movement; II, head nodding; III, unilateral forelimb clonus; IV, bilateral forelimb clonus; V, tonic clonic seizure, rearing, and failing. The rats in which seizure severity reached a level III were identified as grand mal seizure disorder. Seizures were also identified by consistent changes in the power of the fast Fourier transform of EEG, including changes in the frequency of activity during the course of the event. These criteria have been used successfully by experts in the field [53].

Statistical differences between groups were analyzed with either an unpaired t test or one-way analysis of variance (ANOVA) where appropriate. Post hoc analysis was performed with the Newman-Keuls multiple-comparison test. Differences were considered statistically significant at a critical value of *P < 0.05. All values are presented as the mean ± standard error of the mean (SEM).

To determine the profile of NKCC1, we analyzed the protein and mRNA expression of NKCC1 at 24 h following HI. We induced the HI model in P7 neonatal rats, and samples were extracted from the ipsilateral cerebral cortex at 24 h after HI (Fig. 1a). Our results showed an upregulation of the protein and mRNA expression of NKCC1 in the peri-ischemic brain tissue. The optical density of the immune-reactive bands of NKCC1 protein levels that appeared at approximately 170 kDa were significantly increased at 24 h following HI compared with that in the sham group (Fig. 1c). However, after treatment with vitexin, the structure of which is depicted in Fig. 1b, the optical density was decreased significantly when compared with the optical density in ischemic rats (Fig. 1d, *P < 0.05). NKCC1 mRNA expression was significantly increased in the peri-ischemic brain tissue at 24 h following HI rats in comparison with the expression in the sham-operated rats. However, after treatment with bumetanide, which acted as a positive control, or vitexin, the level of NKCC1 mRNA expression in the peri-ischemic brain tissue was significantly decreased when compared with that in the ischemic rats (Fig. 1e, *P < 0.05). Moreover, we performed immunohistochemistry of coronal sections from the ipsilateral cortex to detect NKCC1⁺ cells using anti-NKCC1 antibody (Fig. 1f,  gf, ***P < 0.001). NKCC1⁺ cells were rarely observed in the sham-operated rats, whereas they were more abundant and more extensively distributed in the HI group, particularly in the ischemic penumbra (Fig. 1f, g). After treatment with bumetanide or vitexin, NKCC1⁺ cells in the ipsilateral penumbra were dramatically reduced compared with those in the HI group (Fig. 1f, g). In addition, NKCC1 is located predominantly in the luminal membrane of BBB endothelial cells in situ [54, 55].

Our previous studies have shown that vitexin has a protective effect on the BBB in HI neonatal brain injury [47]. To further investigate the mechanisms underlying HI-induced BBB disruption, we adopted an in vitro BBB model composed of a monolayer of RBMECs and subjected this model to the ischemia-like insult OGD for 3.5 h [47]. Subsequently, the expression of NKCC1 in RBMECs was confirmed by confocal imaging (Fig. 2a).

To further confirm NKCC1 expression in RBMECs under different conditions, flow cytometry was also used. As shown in Fig. 2b, the number of NKCC1-positive cells increased after OGD compared to that in controls. The cell quantities in both the bumetanide (100 mM; 24 h) and vitexin (100 mM; 24 h) groups were decreased compared to those in the OGD group. In addition, we performed MQAE to analyze the concentration of Cl⁻ in RBMECs (Fig. 2c). This was studied by preloading the cells with the Cl⁻ fluorescent indicator, MQAE, with subsequent analysis of the changes in fluorescence that occurred with changes in the intracellular Cl⁻ levels, which were inversely related to the intracellular chloride ion concentration ([Cl⁻] _i ). Compared with the fluorescence intensity in the controls, the fluorescence intensity of RBMECs was significantly decreased in the OGD group, indicating that [Cl⁻] _i was enhanced following OGD (Fig. 2d). Fluorescence was dramatically decreased with treatment with bumetanide or vitexin. Taken together, these results suggested that vitexin effectively suppresses intracellular [Cl⁻] _i , which indirectly represents the expression of NKCC1.

In addition to its conventional function as an ion transporter, NKCC1 also modulates cell migration ability by interacting with the cytoskeleton and acting as an anchor that transduces contractile forces [27, 28]. We therefore examined the NKCC1-positive cells co-stained with the F-actin antibody with confocal microscopy. Our results showed that NKCC1 was localized to F-actin in cultured RBMECs at 24 h following HI (Fig. 3a–c). We also investigated the effects of vitexin on cytoskeletal proteins in RBMECs after OGD (Fig. 3d, e). Representative confocal images of RBMECs showed F-actin cytoskeleton staining with FITC-phalloidin. In the control group, phalloidin staining in RBMECs demonstrated no obvious stress fiber formation, which appeared as large bundles of actin filaments. Instead, the cells showed an apparent increase in stress fiber formation in the OGD group. Treatment of OGD-exposed cells with bumetanide (100 mM; 24 h) or vitexin (100 mM; 24 h) resulted in a decrease in F-actin stress fiber formation (Fig. 3b).

Consistent with the above findings, western blot analysis of lysates from cultured RBMECs also proved that the significant increase in F-actin levels after OGD could be rescued by bumetanide (100 mM; 24 h) or vitexin (100 mM; 24 h) treatment (Fig. 3c). These results further validated our hypothesis that vitexin inhibits OGD-induced stress fiber formation in RBMECs through the NKCC1/F-actin pathway.

Robust actin polymerization and stress fiber formation in RBMECs lead to cell contraction and redistribution/disassembly of tight-junction proteins (TJs) [48]. Confocal imaging demonstrated that OGD weakened ZO-1 expression at extracellular cell–cell contact sites (Fig. 4a). In the control group, the subcellular location of ZO-1 was presented continuously at the RBMEC membrane, outlining the points of cell–cell contact and presumably the TJs. When cells were exposed to OGD, ZO-1 showed a discontinuous and diffuse pattern of staining at regions of cell–cell contact (Fig. 4a). Compared with the OGD group, those treated with vitexin exhibited more areas where the membrane-bound location of ZO-1 was maintained (Fig. 4a).

Alteration in TJ-associated proteins such as ZO-1 has been reported to contribute to the loss of BBB function in many CNS diseases and disorders [56]. Next, ZO-1 expression in the brain tissue from animals of each group was detected by immunofluorescence staining. ZO-1 expression was greatly reduced in CD31-positive capillaries in the ischemic penumbra region, which is indicative of a damaged BBB (Fig. 4b). However, after treatment with bumetanide or vitexin, a certain degree of rescue of ZO-1 expression was observed, suggesting that BBB destruction after HI was attenuated. Similarly, western blotting and q-PCR also proved that the significant reduction in ZO-1 expression after HI could be rescued by vitexin treatment (Fig. 4c, d). Based on these experiments, we concluded that vitexin protected ischemic brain damage through increasing ZO-1 expression, which consequently improved the integrity of the BBB and protected vasogenic edema.

Mounting evidence suggests that inflammation is a key contributor to the severity of CNS hypoxia-ischemia injury. In rats subjected to 3.5 h of HI, their ipsilateral hemispheres displayed an inflammatory response as shown by increased expression of hallmark cytokines such as IL-1β, IL-6, and TNF. We further evaluated changes in the mRNA and protein levels of pro-inflammatory cytokine at 24 h after HI insult by q-PCR and ELISA. We demonstrated that HI caused a significant increasement (***P < 0.001) in the secretion of IL-1β, IL-6, and TNF when compared to sham treatment. However, treatment with vitexin (45 mg/kg) reduced IL-β, IL-6, and TNF mRNA at 24 h after HI (Fig. 5a–c, *P < 0.05, ***P < 0.001). Correspondingly, a significant decrease of protein levels of IL-β, IL-6, and TNF had also been detected (Fig. 5d–f). In sum, these results illustrate that vitexin is a potent suppressor of HI-induced inflammation. To investigate the effect of vitexin on neutrophil infiltration into ipsilateral hemispheres at 24 h after HI, we performed immunohistochemistry of coronal sections from the ipsilateral cortex to detect MPO⁺ cells using anti-MPO antibodies. MPO activity is an indicator of inflammation and is used to evaluate neutrophil accumulation [57]. Neutrophils were rarely observed in the sham-operated rats, whereas they were more abundant and more extensively distributed in the HI group, particularly in the ischemic penumbra (Fig. 5g, h). After treatment with bumetanide or vitexin, the number of MPO⁺ cells in the ipsilateral penumbra was dramatically reduced compared with that in the HI group (Fig. 5g, h). Collectively, these findings strongly suggest that vitexin can reduce the infiltration of neutrophils to a large extent under HI conditions.

Based on a previous study, we provided evidence that hypoxia-ischemia induced neonatal seizure which is thought to contribute to abnormal brain activity [51]. We therefore performed EEG monitoring to examine whether vitexin treatment in the early phase after HI has a long-term effect on HI rats. Rats were sorted into a sham group, HI group, or vitexin treatment group at P30 (Fig. 1a). The P7 rats were treated with HI and were then injected with vitexin (45 mg/kg, 5 min after HI). Electrodes were implanted in rats (n = 15 per group) on P28, and after 3 days of recovery, spontaneous EEG seizures were recorded in freely moving animals (Fig. 6a). During the EEG recording session, no spontaneous EEG seizures were observed in rats in the sham group. Rats in the HI+Vit group, which were given an intraperitoneal injection of vitexin, exhibited less intense spontaneous EEG seizures than rats in the HI group. Vitexin-treated animals also exhibited a marked reduction in the mean duration of spontaneous EEG seizures from 38.22 s/seizures (min. 14.0 s/seizures; max. 61.0 s/seizures) in the HI group to 7.75 s/seizures (min. 5.0 s/seizures; max. 21.0 s/seizures) in the HI+Vit group (Fig. 6b). Also, vitexin treatment significantly decreased the frequency of spontaneous EEG seizures in rats in the HI+Vit group (0.62 (min. 0; max. 2)) compared with the frequency in rats in the HI group (1.72 (min. 1; max. 3); Fig. 6b; *P < 0.05). Taken together, our results demonstrated that treatment with vitexin during a vulnerable period was able to decrease pentylenetetrazol (PTZ)-induced seizure in rats after HI.