Neuroprotective Effects of Purpurin Against Ischemic Damage via Anti-inammatory and MAPK Pathway

Purpurin has various effects, including anti-inammatory effects, and can eciently cross the blood-brain barrier. In the present study, we investigated the effects of purpurin on oxidative stress in HT22 cells and ischemic damage in the hippocampal CA1 region of gerbils. Oxidative stress induced by H 2 O 2 was signicantly ameliorated by treatment with purpurin, based on changes in cell death, DNA fragmentation, formation of reactive oxygen species, and apoptosis (Bcl-2)/antiapoptosis (Bax)-related protein levels. In addition, treatment with purpurin signicantly reduced the phosphorylation of c-Jun N-terminal kinase (JNK), extracellular signal-regulated kinase 1/2 (ERK), and p38 signaling in HT22 cells. Transient forebrain ischemia in gerbils led to a signicant increase in locomotor activity 1 day after ischemia and signicant decrease in number of surviving cells in the CA1 region 4 days after ischemia. Administration of purpurin reduced the travel distance 1 day after ischemia and increased the number of NeuN-immunoreactive neurons in the hippocampal CA1 region of the dentate gyrus 4 days after ischemia. Purpurin treatment signicantly decreased microglial activation in the hippocampal CA1 region 4 days after ischemia and ameliorated the ischemia-induced increases in interleukin (IL)-1β, IL-6, and tumor necrosis factor (TNF)-α 6 h after ischemia. In addition, purpurin signicantly alleviated the ischemia-induced phosphorylation of JNK, ERK, and p38 in the hippocampus 1 day after ischemia. These results suggest that purpurin has neuroprotective potential to reduce inammatory processes and the phosphorylation of JNK, ERK, and p38 in the hippocampus. ± hoc


Introduction
Ischemic stroke is a life-threatening disease that affects approximately 15 million people worldwide annually [1]. Interruption of the blood ow into the brain causes a reduction in the supply of oxygen and glucose into the brain, resulting in damage to affected areas, including the hippocampus [2, 3].
Reperfusion of interrupted vessels into the brain enormously increases the blood supply to the brain, but glucose metabolism is impaired via the pyruvate dehydrogenase pathway in neurons and pyruvate carboxylase pathway in astrocytes [4]. Normally, oxygen radicals are generated from 0.2-2% of oxygen by the electron transport chain [5] and scavenged by antioxidants in the body [6,7]. However, ischemia/reperfusion signi cantly increases the formation of oxygen radicals, exceeding the scavenging capacity of antioxidant enzymes in neurons, and nally causing oxidative damage and propagating in ammatory damage in neurons after ischemia [8,9].
Many attempts have been made to prevent and reduce brain damage after ischemic damage using herbal extracts because of their high phenolic and avonoid contents [10,11]. Anthraquinones have a 9,10dioxoanthracene core substituted with phenolic hydroxyl and aliphatic groups in the two benzene rings. Anthraquinones are less highlighted, although they have various biological effects that inhibit the progression of diseases [12]. Purpurin, an anthraquinone, exhibits antioxidant, anti-in ammatory, and antifungal effects in in vitro assays [13,14] and anti-angiogenic effects in a zebra sh model [15]. In addition, purpurin inhibits monoamine oxidase and shows potential for drug development in depression [16,17]. Purpurin is able to cross the blood-brain barrier (BBB) assessed in human brain-like endothelial cells [18], which mimic the in vivo BBB [19].
However, no studies have examined the effects of purpurin against oxidative damage in HT22 cells and ischemic damage in the gerbil hippocampus. In the present study, we elucidated the effects of purpurin and its mechanisms based on H 2 O 2 -induced oxidative stress in HT22 cells and ischemia-induced neuronal damage in the gerbil hippocampal CA1 region.

Materials And Methods
Cell preparation and determination of cellular toxicity in HT22 cells Murine hippocampal HT22 cells were obtained from ATCC (Manassas, VA, USA) and cultured in Dulbecco's modi ed Eagle's medium as described in previous studies [20,21]. Purpurin was dissolved in 200-mM dimethyl sulfoxide (DMSO) and various concentrations of purpurin (1-200 μM) were added to HT22 cells for 60 min. The cells were then harvested to observe the cellular toxicity of purpurin in HT22 cells. Cellular toxicity was assessed by measuring the uorescence of formazan produced using the WST-1 assay kit (Sigma, St. Louis, MO, USA) and a Fluoroskan ELISA plate reader (Labsystems Multiskan MCC/340, Helsinki, Finland) as described in previous studies [20,21].
Measurements of reactive oxygen species, DNA fragmentation, and cell viability in HT22 cells Cells were exposed to 25-μM purpurin or DMSO immediately after treatment with 1-mM H 2 O 2 to induce oxidative stress. For reactive oxygen species (ROS) formation, 20-μM 2',7'-dichloro uorescein diacetate (DCF-DA) was added to HT22 cells 10 min after H 2 O 2 treatment to induce the formation of DCF, which has strong uorescence. Cells were harvested 30 min after DCF-DA treatment. DNA fragmentation was validated using terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining as described in previous studies [20,21]. Brie y, cells were harvested 3 h after H 2 O 2 treatment, and DNA fragmentation was visualized using a TUNEL staining kit (Sigma). Microphotographs from DCF-DA and TUNEL staining were taken using a confocal uorescence microscope (LSM 510 META NLO; Zeiss GmbH, Jena, Germany), and the uorescence intensity was measured using a Fluoroskan ELISA plate reader (Labsystems Multiskan MCC/340). Cell death was assessed using a WST-1 assay 5 h after H 2 O 2 treatment, and formazan uorescence was measured using a Fluoroskan ELISA plate reader.
In addition, we did not use DMSO as a vehicle for in vivo studies because it shows neuroprotective effects against ischemic damage [23].

Spontaneous motor activity
Motor activity was monitored 1 day after ischemia for 60 min because hyperactivity was induced due to cellular damage in the hippocampal CA1 region [24]. Traveling activity was recorded using a digital camera system (Basler 106200, Ahrensburg, Germany), and the travel distance and duration of immobile/mobile phases were analyzed using Ethovision XT14 (Wageningen, Netherlands).

Measurements of pro-in ammatory cytokines
To elucidate the mechanisms of purpurin's effects against ischemic damage, animals (n = 5 in each group) were euthanized with 75 mg/kg alfaxalone and 10 mg/kg xylazine 6 h after ischemia/reperfusion, when pro-in ammatory cytokine levels were signi cantly increased [26,27]. In brief, interleukin (IL)-1β, IL-6, and tumor necrosis factor (TNF)-α levels were measured based on comparisons with linear calibration curves generated using IL-1β, IL-6, and TNF-α standard solutions.

Western blot analysis in gerbil hippocampus
To elucidate the MAPK pathway in gerbil hippocampus after ischemia, animals were sacri ced 24 h after ischemia. Hippocampi were obtained from the brain and homogenized. Thereafter, cells were lysed with ice-cold radioimmunoprecipitation assay buffer (Thermo Scienti c, IL, USA), and western blotting for mitogen-activated protein kinases (MAPKs) was performed described above.

Statistical analysis
Data are presented as mean with the standard deviation, and differences in means were compared and statistically analyzed using one-way or two-way analysis of variance (ANOVA) followed by Bonferroni's post-hoc test using GraphPad Prism 5.01 software (GraphPad Software, Inc., La Jolla, CA, USA).

Results
Neuroprotective effects of purpurin against oxidative stress in HT2 cells First, we validated the toxicity of purpurin in HT22 cells to determine the effective, but non-toxic, concentration of purpurin. Purpurin treatment for 60 min showed no toxic effects at a concentration 25 μM, and higher concentrations of purpurin decreased cell viability in a concentration-dependent manner (Fig. 1A).

ROS formation was visualized by the formation of DCF uorescence after H 2 O 2 treatment of HT22 cells.
In the control group, DCF uorescence was faintly detected, but in the DMSO-treated group, some cells showed strong DCF uorescence, although no statistically signi cant difference in DCF uorescence was detected between the control and DMSO-treated groups. In the DMSO and H 2 O 2 -treated (H 2 O 2 +DMSO) group, numerous DCF uorescent cells were found, and the uorescence intensity was signi cantly higher (511.3%) than that in the control group. In the purpurin and H 2 O 2 -treated (H 2 O 2 +Purpurin) group, a few DCF uorescent cells were found, and uorescence intensity was signi cantly lower than that in the H 2 O 2 +DMSO group (Fig. 1B).
DNA fragmentation was observed using TUNEL staining after H 2 O 2 treatment of HT22 cells. In the control and DMSO groups, few TUNEL-positive cells were detectable among the HT22 cells and the TUNEL uorescence intensity was low. In the H 2 O 2 +DMSO group, many TUNEL-positive cells were observed among HT22 cells, and the uorescence intensity was signi cantly increased to 490.3% of that in the control group. In the H 2 O 2 +Purpurin group, few TUNEL-positive cells were found, and the uorescence intensity was signi cantly lower than that in the H 2 O 2 +DMSO group at 201.5% of the intensity in the control group (Fig. 1C).
Cell viability was measured using formazan uorescence from tetrazolium salts after H 2 O 2 treatment in HT22 cells. In the DMSO group, the cell viability was similar to that of the control group, but cell viability in the H 2 O 2 +DMSO group was signi cantly lower after H 2 O 2 treatment at 56.6% of that of the control group. In the H 2 O 2 +Purpurin group, cell viability was signi cantly increased compared to that in the H 2 O 2 +DMSO group, and cell viability in this group was at 78.3% of that of the control group (Fig. 1D).
Neuroprotective mechanisms of purpurin against oxidative stress in HT2 cells Bax and Bcl-2 protein levels were measured using western blotting after H 2 O 2 treatment of HT22 cells. In the DMSO group, Bax and Bcl-2 protein levels did not show any signi cant changes relative to those in the control group. However, in the H 2 O 2 +DMSO group, Bax protein levels were signi cantly higher at 469.7% of those in the control group, while Bcl-2 levels were dramatically lower at 24.2% of those in the control group. In the H 2 O 2 +Purpurin group, changes in Bax and Bcl-2 levels were ameliorated compared to those in the H 2 O 2 +DMSO group, respectively, and they were 336.5% and 55.9% of those in the control group, respectively ( Fig. 2A).
JNK, ERK, p38 proteins and their phosphorylated forms (p-JNK, p-ERK1/2, and p-p38) were assessed using western blotting after H 2 O 2 treatment of HT22 cells, and the ratio of phosphorylated and naïve forms were analyzed. In the DMSO group, the p-JNK/JNK, p-ERK/ERK, and p-p38/p38 ratios were similar to those in the control group. In the H 2 O 2 +DMSO group, the p-JNK/JNK, p-ERK/ERK, and p-p38/p38 ratios were signi cantly higher at 381.8%, 472.3%, and 176.4% of those in the control group, respectively. In the H 2 O 2 +Purpurin group, the ratio of p-JNK/JNK, p-ERK/ERK, and p-p38/p38 was signi cantly lower than those in the H 2 O 2 +DMSO group at 222.2%, 283.5%, and 109.7% of the ratios in the control group, respectively (Fig. 2B).

Neuroprotective effects of purpurin against ischemic damage in gerbils
The neuroprotective effects of purpurin were validated using locomotor behavior 1 day after ischemia. In the vehicle-treated ischemic group, the time in the mobile and immobile phases was signi cantly changed to 115.6% and 57.3% of those in the control group, respectively. The traveled distance in the vehicle-treated ischemic group was signi cantly longer than that in the control group (292.9% of that in the control group). In the 1 or 3 mg/kg purpurin-treated ischemic groups, the time spent in the mobile and non-mobile phases was similar to those in the vehicle-treated group, but in the 6 mg/kg purpurin-treated group, they did not show signi cant differences compared to those in the vehicle-treated group control or vehicle-treated ischemic group. Similarly, the traveled distance was signi cantly longer in 1 or 3 mg/kg purpurin-treated ischemic groups than in the control group. However, in the 6 mg/kg purpurin-treated group, the traveled distance was signi cantly less than that in the vehicle-or 1 mg/kg purpurin-treated ischemic groups (181.9% of that in the control group) (Fig. 3A).
The neuroprotective effects of purpurin were con rmed using immunohistochemical staining for NeuN in the hippocampus 4 days after ischemia. In the control group, abundant NeuN-immunoreactive cells were found in the hippocampus. In the vehicle-treated ischemic group, a few NeuN-immunoreactive cells were detected in the hippocampal CA1 region (5.1% of control), whereas in other regions, NeuNimmunoreactive cells were similar levels were seen as in the control group. In the 1 or 3 mg/kg purpurintreated groups, NeuN-immunoreactive neurons were similarly observed in the hippocampal CA1 region compared to vehicle-treated group (7.5% and 9.9% of control). In the 6 mg/kg purpurin-treated ischemic group, many NeuN-immunoreactive cells were found in the CA1 region, and the number of NeuNimmunoreactive neurons was signi cantly higher (60.2% of control) than that in the vehicle-treated ischemic group (Fig. 3B).
Neuroprotective mechanisms of purpurin's effects against ischemic damage in gerbils The neuroprotective mechanisms of 6 mg/kg purpurin were evaluated in terms of anti-in ammatory responses in the hippocampus using an ELISA assay for IL-1β, IL-6, and TNF-α 6 h after ischemia. In the vehicle-treated ischemic group, IL-1β, IL-6, and TNF-α levels were signi cantly higher at 529.6%, 312.4%, and 1255.0% of those in the control group, respectively. In the purpurin-treated ischemic group, IL-1β, IL-6, and TNF-α levels were signi cantly lower than those in vehicle-treated ischemic group and were 203.2%, 178.2%, and 626.1% of those in the control group (Fig. 4A).
Microglia were visualized using immunohistochemical staining for Iba-1 4 days after ischemia. In the control group, Iba-1 immunoreactive microglia had a small cell body and thin processes. In the vehicletreated ischemic group, Iba-1 immunoreactive microglia in the stratum pyramidale had a round cell body, but they had a hypertrophied cell body and thick processes in the stratum oriens and radiatum. In this group, Iba-1 immunoreactivity was signi cantly increased to 711.7% of that in the control group. In the purpurin-treated ischemic group, Iba-1 immunoreactive microglia had a large cell body and lessdeveloped processes compared to those in the vehicle-treated ischemic group. In this group, Iba-1 immunoreactivity was signi cantly than that in the vehicle-treated ischemic group and was 459.9% of that in the control group (Fig. 4B).
MAPKs and their phosphorylated forms were validated using western blotting 1 day after ischemia in gerbil hippocampus and the ratio of phosphorylated and naïve forms were analyzed. In the vehicletreated ischemic group, the ratios of p-JNK/JNK, p-ERK/ERK, and p-p38/p38 were signi cantly increased to 221.8%, 692.4%, and 223.9% of control group, respectively although naïve forms of MAPKs showed similar levels compared to respective control group. In the purpurin-treated ischemic group, the ratios of p-JNK/JNK, p-ERK/ERK, and p-p38/p38 were signi cantly lowered to 129.9%, 406.0%, and 124.2% of those in the control group compared to respective vehicle-treated ischemic group (Fig. 5).

Discussion
Purpurin, an alizarin-type anthraquinone, has free radical scavenging activity [14,[28][29][30] and antioxidant effects against Trp-P-2 carcinogen by reducing DNA adducts in the liver [31]. In the present study, we investigated the role of purpurin against oxidative stress induced by H 2 O 2 in HT22 cells and against ischemic damage in gerbils. First, we screened the toxicity of purpurin in HT22 cells to determine the optimal concentration without toxicity in HT22 cells. We observed that 25-μM purpurin was the optimal concentration with minimal toxicity in HT22 cells. The optimal concentration may differ depending on the cell type. In 3T3-L1 adipose cells, 50-and 100-μM purpurin had positive effects [30].
Oxidative stress was induced by treatment with H 2 O 2 , which increases ROS formation and decreases cell viability in a concentration-dependent manner in HT22 cells [32].  [14]. In addition, purpurin reduces hTau accumulation in an in vitro culture system [18].
Next, we examined the protein levels of Bax and Bcl-2, which are the main components of the apoptosis and anti-apoptosis pathways, respectively, because high levels of ROS lead to mitochondrial membrane damage and release of pro-apoptotic proteins such as Bax [33]. Treatment with H 2 O 2 signi cantly increased Bax levels and decreased Bcl-2 levels in HT22 cells, consistent with previous studies [32,34]. Incubation with purpurin signi cantly ameliorated the changes in Bax and Bcl-2 induced by H 2 O 2 treatment in HT22 cells. We also observed the phosphorylation of MAPKs, including JNK, ERK, and p38, because MAPKs play important roles in ROS-induced cell death and H 2 O 2 signi cantly increased the expression of p-ERK 1/2, p-JNK, and p-p38 in HT22 cells [32]. Treatment with H 2 O 2 signi cantly increased the p-JNK/JNK, p-ERK/ERK, and p-p38/p38 ratios in HT22 cells, and incubation with purpurin signi cantly mitigated the increase in the ratio.
In the present study, we also investigated the effects of purpurin against ischemic damage following oral treatment with 6 mg/kg purpurin because purpurin is able to cross the blood brain barrier [18,35]. In addition, purpurin caused no signi cant changes in physiological or blood chemistry variables in an acute oral toxicity study [36]. We observed the locomotor activity 1 day after ischemia because the locomotor test is a predictive measure for assessing neuronal damage in the hippocampus [37,38].
Transient forebrain ischemia signi cantly increased the travel distance and time in the mobile phase, indicating hyperactivity in gerbils 1 day after ischemia. Purpurin treatment signi cantly reduced the travel distance and time in the mobile phase. In addition, we con rmed that 6 mg/kg, not 1 or 3 mg/kg, purpurin treatment ameliorated the ischemia-induced reduction in NeuN-immunoreactive neurons in the hippocampal CA1 region. This result suggests that purpurin has the potential to reduce neuronal death induced by ischemia.
To elucidate the possible role of 6 mg/kg purpurin against ischemia, we observed the morphology of microglia and pro-in ammatory cytokines in the hippocampus because a recent study showed the antiin ammatory roles of purpurin in RAW 264.7 murine macrophage cells [14]. The animals were sacri ced 6 h after ischemia to measure IL-1β, IL-6, and TNF-α levels in the hippocampus because these levels are signi cantly increased in the early period of ischemia [26, 27, 39]. In addition, the IL-1 receptor antagonist showed neuroprotective effects against ischemic damage in rats [40]. In the vehicle-treated group, IL-1β, IL-6, and TNF-α levels were signi cantly increased 6 h after ischemia/reperfusion compared to those in the control group. In the purpurin-treated ischemic group, IL-1β, IL-6, and TNF-α levels were dramatically lower in the hippocampal homogenates. This result suggests that purpurin treatment signi cantly reduces the release of pro-in ammatory cytokines in the hippocampus 6 h after ischemia. In addition, we con rmed microglial activation based on microglial morphology in the hippocampus 4 days after ischemia. In the vehicle-treated group, Iba-1-immunoreactive microglia had hypertrophied cell body and thickened processes (activated microglia), and the phagocytic form (round cell body without processes) of microglia were also found in the stratum pyramidale of the CA1 region 4 days after ischemia/reperfusion. This result was consistent with previous studies showing that ischemia induced microglial activation and morphological changes in the hippocampus [41,42]. Treatment with purpurin reduced the phagocytic form of microglia in the stratum pyramidale, and overall Iba-1 immunoreactivity was signi cantly decreased in the hippocampal CA1 region compared to that in the vehicle-treated group. A molecular docking study demonstrated that purpurin had a strong inhibitory effect on the nucleotide-binding domain leucine-rich repeat and pyrin domain containing receptor 3, which is one of the main contributors to neuroin ammation [43].
In the present study, we also observed the ischemia signi cantly increased the ratios of p-JNK/JNK, p-ERK/ERK, and p-p38/p38 in the gerbil hippocampus 1 day after ischemia result. This result is consistent with in vitro study in HT22 cells that oxidative stress induced by H 2 O 2 treatment signi cantly increased the phosphorylation of MAPKs. In addition, several studies demonstrate the increases of MAPK phosphorylation in the hippocampus after ischemia [22,44,45] and treatment with JNK blocker ameliorates the neuronal death induced by ischemia [46]. In addition, the close relationship has been reported between the cytokine-related in ammation and MAPKs [47,48]. In the present study, we observed the purpurin treatment signi cantly decreased the activation of MAPK pathway in the hippocampus after ischemia.

Conclusions
The current ndings suggest that purpurin may be a strong neuroprotective agent to reduce oxidative  reader. Data are expressed as mean value ± standard deviation and were analyzed using one-way ANOVA followed by Bonferroni's post hoc test (ap < 0.05, signi cantly different from the control group; bp < 0.05, signi cantly different from the DMSO group; cp < 0.05, signi cantly different from the H2O2+DMSO group).

Figure 2
Mechanisms of purpurin' effects against oxidative damage in HT22 cells. (A) Protein levels related to cell death and survival were measured after H2O2-induced oxidative stress in HT22 cells using western blot analysis for Bax and Bcl-2, respectively. Protein levels of Bax and Bcl-2 were calibrated to the β-actin level. (B) Cell signaling pathway related to MAPKs were validated using western blot analysis for JNK, ERK, p38, and their phosphorylated forms. Protein levels were converted into p-JNK/JNK, p-ERK/ERK, and p-p38/p38 ratios in each group. Data are expressed as mean value ± standard deviation and were analyzed using one-way ANOVA followed by Bonferroni's post hoc test (ap < 0.05, signi cantly different from the control group; bp < 0.05, signi cantly different from the DMSO group; cp < 0.05, signi cantly different from the H2O2+DMSO group). Effect of purpurin against ischemic damage in gerbils. (A) Traveled distance and cumulative duration was measured in gerbils 1 day after ischemia in sham-operated (control), ischemia-induced vehicletreated (vehicle), and ischemia-induced purpurin-treated (purpurin) groups (n = 10 per group). (B) Mature neurons are visualized to show the surviving neurons after ischemic damage in the control, vehicle, and purpurin groups using NeuN immunohistochemical staining. SO, stratum oriens; SP, stratum pyramidale; SR, stratum radiatum. Scale bar = 50 μm. The number of NeuN-immunoreactive neurons is shown as a percentile value vs. control group (n = 10 per group). (A and B) Data are expressed as mean ± standard deviation and were analyzed using one-way ANOVA followed by Bonferroni's post hoc test (ap < 0.05, signi cantly different from the control group; bp < 0.05, signi cantly different from the vehicle group). and purpurin groups (n = 5 per group). (B) Microglia were visualized to show the morphological changes after ischemia in the CA1 region of the control, vehicle, and purpurin groups with Iba-1 immunohistochemical staining. Scale bar = 50 μm. Optical density was measured and expressed as a percentage of the value vs. control group (n = 5 per group). Data are expressed as mean value ± standard deviation and were analyzed using one-way ANOVA followed by Bonferroni's post hoc test (ap < 0.05, signi cantly different from the control group; bp < 0.05, signi cantly different from the vehicle group). Figure 5