Ginseng Extracts, GS-KG9 and GS-E3D, Prevent Blood–Brain Barrier Disruption and Thereby Inhibit Apoptotic Cell Death of Hippocampal Neurons in Streptozotocin-Induced Diabetic Rats

Type 1 diabetes mellitus is known to be linked to the impairment of blood–brain barrier (BBB) integrity following neuronal cell death. Here, we investigated whether GS-KG9 and GS-E3D, bioactive ginseng extracts from Korean ginseng (Panax ginseng Meyer), inhibit BBB disruption following neuronal death in the hippocampus in streptozotocin-induced diabetic rats showing type 1-like diabetes mellitus. GS-KG9 and GS-E3D (50, 150, or 300 mg/kg, twice a day for 4 weeks) administered orally showed antihyperglycemic activity in a dose-dependent manner and significantly attenuated the increase in BBB permeability and loss of tight junction proteins. GS-KG9 and GS-E3D also inhibited the expression and activation of matrix metalloproteinase-9 and the infiltration of macrophages into the brain parenchyma, especially into the hippocampal region. In addition, microglia and astrocyte activation in the hippocampus and the expression of proinflammatory mediators such as tnf-α, Il-1β, IL-6, cox-2, and inos were markedly alleviated in GS-KG9 and GS-E3D-treated group. Furthermore, apoptotic cell death of hippocampal neurons, especially in CA1 region, was significantly reduced in GS-KG9 and GS-E3D-treated groups as compared to vehicle control. These results suggest that GS-KG9 and GS-E3D effectively prevent apoptotic cell death of hippocampal neurons by inhibiting BBB disruption and may be a potential therapy for the treatment of diabetic patients.


Introduction
Diabetes mellitus, characterized by chronic hyperglycemia, is a metabolic disorder with semeiotic symptoms such as polyuria, polyphagia, and polydipsia along with low-grade inflammation [1]. Recent clinical and experimental studies suggest that prolonged hyperglycemic conditions influence a progressive impairment of neuronal function in the brain [2][3][4]. Diabetes mellitus has also been associated with increased severity of epileptic seizure [5] and risk of mortality following traumatic brain injury [6]. In particular, diabetes mellitus has been strongly linked to mild cognitive impairments [7,8] and is considered a predisposing factor for developing vascular dementia [9] and Alzheimer's disease [10]. Learning and memory are also known to be affected, especially in adults with poor glycemic control, longer duration of diabetes, or young age of onset [11,12]. 2% Evans blue dye (Sigma) solution in saline was administered i.p. Three hours later, the animals were perfused with PBS (Phosphate buffered saline) and the brain hippocampus region was removed and homogenized in a 50% trichloroacetic acid solution. After homogenization, samples were centrifuged at 10,000× g for 10 min, supernatants were collected and fluorescence was quantified at an excitation wavelength of 620 nm and an emission wavelength of 680 nm. The amount of dye in samples was determined as micrograms per gram of tissue from a standard curve plotted using known amounts of dye. For qualitative analysis, some animals were perfused with PBS and subsequently with 4% formaldehyde, as described above. The brains were then immersed in a 30% sucrose solution and brain tissue were cut into 20 µm thick sections with a cryostat. The fluorescence of Evans blue in the hippocampus was observed with a fluorescence microscope and the relative fluorescence intensity was determined by MetaMorph software (Molecular devices, Sunnyvale, CA, USA).
Tissue preparation. Rats were anesthetized with chloral hydrate (500 mg/kg, intraperitoneal injection) and perfused transcardially with phosphate buffer (100 mM, pH 7.4) followed by ice-cold 4% paraformaldehyde, they were and then decapitated. The brains were removed and post-fixed overnight in phosphate buffer (50 mM, pH 7.4) containing 4% paraformaldehyde. The brains were then immersed in a 30% sucrose solution (in 50 mM phosphate-buffered saline, PBS) and stored at 4 • C until sectioning. Frozen brains were sectioned along the coronal plane (30 µm) using a cryostat (Leica Microsystems AG, Wetzlar, Germany) and maintained in a storage solution at 4 • C. For molecular work, rats were perfused with PBS and bilateral hippocampal tissue samples were isolated and frozen at −80 • C.

Immunohistochemistry
Frozen sections were processed for immunohistochemistry with antibodies against OX-42 (1:100, Merck Millipore, Billerica, MA, USA) for staining microglia, GFAP (Glial fibrillary acidic protein) (1:5000, Merck Millipore) for staining astrocytes, and ED-1 (Bio-Rad, Hercules, CA, USA) for staining macrophages, as previously described [42]. The sections were incubated with primary antibodies, followed by biotin-conjugated secondary antibodies (Dako, Carpinteria, CA, USA). The ABC (Avidin-biotin complex)method was used to detect labeled cells using a Vectastain kit (Vector Labs, Burlingame, CA, USA). DAB (3, 3-diaminobenzidine)served as the substrate for peroxidase. For immunofluoresence staining, cy3-conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA, USA) were used. In addition, nuclei were labeled with DAPI (4 ,6-diamidino-2-phenylindole)according to the protocol of the manufacturer (ThermoFisher Scientific, Waltham, MA, USA). In all controls, reaction to the substrate was absent if the primary antibody was omitted or if the primary antibody was replaced by a non-immune, control antibody. Serial sections were also stained for histological analysis with Cresyl violet acetate. The counting of ED-1-positive cells was carried out by the investigators, who were blind to the experimental conditions. Every five sections throughout the hippocampus were selected and the ED-1-positive cells in the CA1 area (10 sections per animal) were counted and averaged.

Nissl Staining
After being mounted onto gelatin-coated slides, tissue sections were stained with 0.5% Cresyl violet, dehydrated through graded alcohols (70%, 80%, 90%, and 100% × 2), placed in xylene, and covered with a coverslip after the addition of permount. The Nissl-stained cells in the CA1 area were counted by a researcher who was blinded to the experimental conditions. Every five sections throughout the hippocampus were processed for counting (10 sections per animal). The number of cells in CA1 was quantitatively expressed as a percentage compared to the normal control.

TUNEL Staining
To examine apoptotic cell death of hippocampal neurons, coronal sections including hippocampal regions were processed for terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate-biotin nick end labeling (TUNEL) staining using an Apoptag in situ kit (Merck Millipore). Investigators who were blind to the experimental conditions carried out all TUNEL analyses. Every five sections throughout the hippocampus were selected and the TUNEL-positive cells in the CA1 area (10 sections per animal) were counted and averaged.

Gelatin Zymography
The activity of matrix metalloproteinase (MMP)-2 and -9 at 28 d after starting the drug treatment was examined by gelatin zymography based on a previously described protocol, with some modifications [42]. Briefly, hippocampal tissue samples were weighed and homogenized in lysis buffer containing the following: 28 mM Tris-HCl, 22 mM Tris-base, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS (Sodium dodecylsulfate-polyacrylamide). The protein concentration of the homogenates was determined by the bicinchoninic acid method (BCA protein assay kit, Pierce, Rockford, IL, USA). After determination of protein concentration of the homogenates, equal amounts of protein (30 µg) were loaded on a Novex 10% zymogram gel (ThermoFisher Scientific) and separated by electrophoresis with 100 V (19 mA) at 4 • C for 6 h. The gel was then incubated with renaturing buffer (2.5% Triton X-100) at room temperature for 30 min to restore the gelatinolytic activity of the proteins. After incubation with developing buffer (50 mM Tris-HCl, pH 8.5, 0.2 M NaCl, 5 mM CaCl 2 , 0.02% Brii35) at 37 • C for 24 h, the gel was stained with 0.5% Coomassie blue for 60 min and then destained with 40% methanol containing 10% acetic acid until appropriate color contrast was achieved. Clear bands on the zymogram were indicative of gelatinase activity. Relative intensity of zymography (relative to sham or vehicle) was measured and analyzed by AlphaImager software (Alpha Innotech Corporation). Backgrounds were subtracted from the optical density measurements.

Statistical Analysis
All data are presented as the mean ± standard deviation (SD). Comparisons between STZ control and GS-KG9 or GS-E3D-treated groups were made by unpaired Student t test. Multiple comparisons between groups were performed by one-way ANOVA. Tukey's multiple comparison was used for post-hoc analysis. Statistical significance was accepted with p < 0.05. Statistical analyses were performed using SPSS 15.0 (SPSS Science, Chicago, IL, USA).

GS-KG9 and GS-E3D Inhibit Hyperglycemia in STZ-Induced Diabetic Rats
To induce hyperglycemia, we injected STZ (60 mg/kg) into rats intraperitoneally (i.p.), which resulted in diabetic syndromes verified by the presence of polydipsia, polyuria, hyperglycemia, and weight loss in the diabetic animals [44]. Hyperglycemia was confirmed by measuring the blood glucose level in a tail vein and then the animals showing blood glucose levels over 250 mg/dL were selected and randomly divided into each experimental group (Figure 1). At 4 days after STZ injection, the mean blood glucose levels were significantly higher in STZ-induced diabetic rats than in normal rats (STZ, 364.4 ± 24.5 mg/dL vs. normal, 112.0 ± 5.5 mg/dL). We first determined whether GS-KG9 and GS-E3D reduce glucose level in STZ-induced diabetic rats. As shown in Figure 2A, in STZ-injected diabetic rats, the blood glucose level was increased and maintained at a high level during the experiments (7 d, 337.3 ± 55.2; 14 d, 324.3 ± 32.2; 21 d, 345.9 ± 39.9; 28 d, 342.4 ± 37.8 mg/dL). However, both GS-KG9 and GS-E3D (150 and 300 mg/kg) significantly reduced the level of blood glucose at 28 d after drug treatment in a dose-dependent manner. Our results also showed that a dose of 300 mg/kg of the drugs was an optimal dose to achieve an anti-hyperglycemic effect in STZ-induced diabetic rats ( Figure 2A) and thus this dosage was used in these experiments. Especially, GS-KG9 or GS-E3D (300 mg/kg) treatment significantly decreased the blood glucose level from 14 d after drug treatment as compared to the STZ control group (at 28 d: STZ, 342.4 ± 37.8; STZ+GS-KG9, 240.4 ± 32.1; STZ+GS-E3D, 238.3 ± 43.0 mg/dL) ( Figure 2B). Evident body weight loss ( Figure 2C), polyphagia ( Figure 2D), and polydipsia ( Figure 2E) were observed in STZ-induced diabetic rats, whereas both body weight loss and polydipsia in STZ-treated rats were significantly alleviated by GS-E3D treatment from 20 d and 28 d after administration, respectively ( Figure 2C,D). Meanwhile, the body weight and polydipsia in GS-KG9-treated groups were slightly higher than in the STZ-treated control group, although not to a level of significance ( Figure 2C,D). In addition, polyphagia was not significantly changed by GS-KG9 or GS-E3D treatment ( Figure 2E). In addition, any significant change in body weight and toxicity signs (such as piloerection, alteration in the locomotor activity, or diarrhea) was not observed in either of the experimental groups or normal rats during the experiment (data not shown). These results indicate that GS-KG9 and GS-E3D reduce glucose levels and improve hyperglycemia-associated symptoms in STZ-induced diabetic rats.
Nutrients 2020, 12, x FOR PEER REVIEW 6 of 19 weight and toxicity signs (such as piloerection, alteration in the locomotor activity, or diarrhea) was not observed in either of the experimental groups or normal rats during the experiment (data not shown). These results indicate that GS-KG9 and GS-E3D reduce glucose levels and improve hyperglycemia-associated symptoms in STZ-induced diabetic rats.  It is known that insulin causes a reduction in blood glucose by facilitating the uptake and storage of glucose. In addition, the blood glucose level reduction can also depend on a downregulation of carbohydrate hydrolyzing enzymes such as α-amylase and α-glucosidase in digestive tract. αamylase is the enzyme that hydrolyzes starch to maltose and consequentially higher postprandial hyperglycemia, while α-glucosidase is responsible for the breakdown of oligo-and/or disaccharides weight and toxicity signs (such as piloerection, alteration in the locomotor activity, or diarrhea) was not observed in either of the experimental groups or normal rats during the experiment (data not shown). These results indicate that GS-KG9 and GS-E3D reduce glucose levels and improve hyperglycemia-associated symptoms in STZ-induced diabetic rats.  It is known that insulin causes a reduction in blood glucose by facilitating the uptake and storage of glucose. In addition, the blood glucose level reduction can also depend on a downregulation of carbohydrate hydrolyzing enzymes such as α-amylase and α-glucosidase in digestive tract. αamylase is the enzyme that hydrolyzes starch to maltose and consequentially higher postprandial hyperglycemia, while α-glucosidase is responsible for the breakdown of oligo-and/or disaccharides It is known that insulin causes a reduction in blood glucose by facilitating the uptake and storage of glucose. In addition, the blood glucose level reduction can also depend on a downregulation of carbohydrate hydrolyzing enzymes such as α-amylase and α-glucosidase in digestive tract. α-amylase is the enzyme that hydrolyzes starch to maltose and consequentially higher postprandial hyperglycemia, while α-glucosidase is responsible for the breakdown of oligo-and/or disaccharides to monosaccharides [45]. The inhibition of these enzymes leads to a decrease in blood glucose level, because monosaccharides are a form of carbohydrate which is absorbed through the mucosal border in the small intestine [46]. It has been demonstrated that Panax ginseng extracts ameliorate hyperglycemia in STZ-induced diabetic animal models by increasing serum insulin levels [47,48]. In addition, a recent report showed that Korean red ginseng extract inhibits α-glucosidase and α-amylase activities and decreases glucose uptake and transport rate in human colon cell lines [48]. Although the level of insulin and the activity of carbohydrate hydrolyzing enzymes was not examined in the present study, we cannot rule out the possibility that GS-KG9 and GS-E3D may be involved in the downregulation of serum insulin levels or inhibition of glucose production and glucose absorption in the digestive tract, thereby reducing the blood glucose level in STZ-induced diabetic rats. We will seek to elucidate the precise mechanism underlying GS-KG9 and GS-E3D-mediated inhibition of blood glucose level in a future study.

GS-KG9 and GS-E3D Inhibit the Increase in BBB Permeability in the Hippocampal Regions of STZ-Induced Diabetic Rats
It is known that all forms of diabetes are characterized by chronic hyperglycemia, resulting in the development of a number of microvascular and macrovascular pathologies. Diabetes is also associated with changes in brain microvasculature, leading to dysfunction and ultimately BBB disruption [17,[49][50][51]. Since GS-KG9 and GS-E3D caused a significant reduction in blood glucose levels in STZ-induced diabetic rats (in Figure 2), we next investigated whether STZ-induced hyperglycemia would lead to microvascular changes and whether GS-KG9 and GS-E3D inhibit BBB disruption in the brain parenchyma, especially in the hippocampal region. To examine the change in BBB permeability, we injected Evans blue dye into the STZ-treated rats (i.p.) administered for four weeks with GS-KG9 and GS-E3D. As a result, we found leakage of Evans blue dye into the brain parenchyma of STZ-induced diabetic rats, especially into the hippocampal region, but not in normal rats, which means that STZ-induced hyperglycemia resulted in BBB disruption. Furthermore, GS-KG9 or GS-E3D treatment significantly reduced the amount of Evans blue dye extravasation when compared with the STZ control (STZ, 20.3 ± 2.7; STZ+GS-KG9, 9.8 ± 2.1; STZ+GS-E3D, 7.3 ± 2.3 mg/dL) ( Figure 3A). Consistent with these results, the fluorescence microscopy images of hippocampal CA1 regions of STZ-induced diabetic rat brains showed the presence of Evans blue dye (bright red) in the brain parenchyma around blood vessels, demonstrating the leakage of the dye across the BBB. However, the intensity of Evans blue was reduced in GS-KG9 or GS-E3D-treated groups ( Figure 3B,C). These results indicate that GS-KG9 and GS-E3D inhibit BBB disruption by STZ-induced hyperglycemia.

GS-KG9 and GS-E3D Inhibit the Expression and Activity of MMP-9 in STZ-Induced Diabetic Rats
It has been known that hyperglycemia increases matrix metalloproteinase (MMP) activity and thereby results in the breakdown of the tight junction following the increase in BBB permeability [14]. MMPs are a family of zinc-dependent proteolytic enzymes that degrade components of the extracellular matrix in various pathophysiological conditions. The excessive proteolytic activity of MMPs such as MMP-2 and MMP-9 results in BBB disruption after CNS injury such as stroke and spinal cord injury [42,52,53]. Recently, Aggarwal et al. [54] reported that MMP-9 activity is significantly elevated in STZ-induced diabetic rat brains and the inhibition of MMP-9 activity leads to the restoration of BBB integrity and improves learning and memory in STZ-induced diabetic rats. Since GS-KG9 and GS-E3D treatment reduced BBB disruption in the hippocampus of STZ-induced diabetic rats, we examined whether these drugs would inhibit the expression and activity of MMP-2 and/or MMP-9. As shown in Figure 4A, the levels of mmp-9 mRNA increased in STZ-injected rats compared with normal rats. Furthermore, mmp-9 mRNA expression was significantly inhibited by GS-KG9 or GS-E3D administration compared with STZ control group ( Figure 4A,B). Using gelatin zymography, GS-KG9 and GS-E3D also significantly inhibited the increase in MMP-9 activity (active MMP-9 band) as compared with the STZ control ( Figure 4C,D) (active MMP-9; STZ, 8.7 ± 0.6; GS-KG9, 6.1 ± 0.7; GS-KG9, 6.1 ± 0.7). No significant change was observed in either mRNA expression or gelatinase activity of MMP-2 in diabetic rats, as reported [54]. Our data thus indicate that GS-KG9 and GS-E3D reduce both mmp-9 mRNA expression and MMP-9 activation in chronic hyperglycemic diabetic rats.  Figure 3B). The bottom panels are higher magnification images of the box area in the upper panels. Arrows indicate Evans blue extravasation. Red, fluorescence signal; blue, DAPI (4',6-diamidino-2-phenylindole)stained signal. Note that the fluorescence intensity of Evans blue in the hippocampal radiatum layer (Rad) was significantly reduced in GS-KG9 and GS-E3D (300 mg/kg) treated rats when compared to that in STZ control. Scale bars, 100 μm.

GS-KG9 and GS-E3D inhibit the expression and activity of MMP-9 in STZ-induced diabetic rats.
It has been known that hyperglycemia increases matrix metalloproteinase (MMP) activity and thereby results in the breakdown of the tight junction following the increase in BBB permeability [14]. MMPs are a family of zinc-dependent proteolytic enzymes that degrade components of the extracellular matrix in various pathophysiological conditions. The excessive proteolytic activity of MMPs such as MMP-2 and MMP-9 results in BBB disruption after CNS injury such as stroke and spinal cord injury [42,52,53]. Recently, Aggarwal et al. [54] reported that MMP-9 activity is significantly elevated in STZ-induced diabetic rat brains and the inhibition of MMP-9 activity leads to the restoration of BBB integrity and improves learning and memory in STZ-induced diabetic rats. Since GS-KG9 and GS-E3D treatment reduced BBB disruption in the hippocampus of STZ-induced diabetic rats, we examined whether these drugs would inhibit the expression and activity of MMP-2 and/or MMP-9. As shown in Figure 4A, the levels of mmp-9 mRNA increased in STZ-injected rats compared with normal rats. Furthermore, mmp-9 mRNA expression was significantly inhibited by GS-KG9 or GS-E3D administration compared with STZ control group ( Figure 4A,B). Using gelatin zymography, GS-KG9 and GS-E3D also significantly inhibited the increase in MMP-9 activity (active MMP-9 band) as compared with the STZ control ( Figure 4C,D) (active MMP-9; STZ, 8.7 ± 0.6; GS-  Figure 3B). The bottom panels are higher magnification images of the box area in the upper panels. Arrows indicate Evans blue extravasation. Red, fluorescence signal; blue, DAPI (4 ,6-diamidino-2-phenylindole)stained signal. Note that the fluorescence intensity of Evans blue in the hippocampal radiatum layer (Rad) was significantly reduced in GS-KG9 and GS-E3D (300 mg/kg) treated rats when compared to that in STZ control. Scale bars, 100 µm.
Nutrients 2020, 12, x FOR PEER REVIEW 9 of 19 KG9, 6.1 ± 0.7; GS-KG9, 6.1 ± 0.7). No significant change was observed in either mRNA expression or gelatinase activity of MMP-2 in diabetic rats, as reported [54]. Our data thus indicate that GS-KG9 and GS-E3D reduce both mmp-9 mRNA expression and MMP-9 activation in chronic hyperglycemic diabetic rats. Since we showed that GS-KG9 and GS-E3D inhibited BBB disruption in the hippocampus of STZ-induced diabetic rats (Figure 3), we next examined whether these drugs would also inhibit the loss of tight junction proteins by Western blot. The tight junction in the endothelial cells of capillary

GS-KG9 and GS-E3D Alleviate the Loss of Tight Junction Proteins in the Hippocampus of STZ-Induced Diabetic Rats
Since we showed that GS-KG9 and GS-E3D inhibited BBB disruption in the hippocampus of STZ-induced diabetic rats (Figure 3), we next examined whether these drugs would also inhibit the loss of tight junction proteins by Western blot. The tight junction in the endothelial cells of capillary blood vessels is essential for BBB integrity in the brain [55]. To determine whether STZ-induced hyperpermeability of BBB was due to tight junction alterations, the expression of the tight junction-associated proteins such as ZO-1, occludin, and claudin-5 in the hippocampus was examined at 28 d after drug treatment. As shown in Figure 5A, the levels of ZO-1, occludin, and claudin-5 were markedly decreased in STZ-induced diabetic rats, indicating that hyperglycemia leads to a decrease in BBB integrity by resulting in the loss of tight junction proteins. However, GS-KG9 or GS-E3D treatment significantly decreased the levels of ZO-1, occludin, and claudin-5 as compared with STZ control (Figure 5A,B). These data indicate that GS-KG9 and GS-E3D preserve tight junction integrity by inhibiting the degradation of tight junction molecules and thereby preventing BBB disruption in STZ-induced diabetic rats.

GS-KG9 and GS-E3D inhibit macrophage infiltration in the hippocampus of STZ-induced diabetic rats.
It is known that blood cell infiltration following BBB disruption initiates inflammatory responses, leading to the secondary injury cascade by producing inflammatory mediators. Therefore, we examined the effect of GS-KG9 and GS-E3D on blood cell infiltration by immunofluorescence staining with a macrophage/monocyte cell marker, ED-1 antibody. Immunofluorescence staining showed that ED-1 positive cells were observed in the pyramidal layer and radiatum layer of CA1 at 28 d after drug treatment ( Figure 6A, STZ), whereas no ED-1 positive cells were observed in the normal brain ( Figure 6A, normal). Furthermore, the number of ED-1 positive cells was significantly decreased in GS-KG9 and GS-E3D-treated groups as compared with the STZ control (STZ, 58 ± 5.8; GS-KG9, 21.8 ± 3.3; GS-E3D, 15.8 ± 3.9) (Figure 6A,B). These results indicate that GS-KG9 and GS-E3D inhibit macrophage infiltration in STZ-induced diabetic rats by preserving BBB integrity.

GS-KG9 and GS-E3D Inhibit Macrophage Infiltration in the Hippocampus of STZ-Induced Diabetic Rats
It is known that blood cell infiltration following BBB disruption initiates inflammatory responses, leading to the secondary injury cascade by producing inflammatory mediators. Therefore, we examined the effect of GS-KG9 and GS-E3D on blood cell infiltration by immunofluorescence staining with a macrophage/monocyte cell marker, ED-1 antibody. Immunofluorescence staining showed that ED-1 positive cells were observed in the pyramidal layer and radiatum layer of CA1 at 28 d after drug treatment ( Figure 6A, STZ), whereas no ED-1 positive cells were observed in the normal brain ( Figure 6A, normal). Furthermore, the number of ED-1 positive cells was significantly decreased in GS-KG9 and GS-E3D-treated groups as compared with the STZ control (STZ, 58 ± 5.8; GS-KG9, 21.8 ± 3.3; GS-E3D, 15.8 ± 3.9) (Figure 6A,B). These results indicate that GS-KG9 and GS-E3D inhibit macrophage infiltration in STZ-induced diabetic rats by preserving BBB integrity.

GS-KG9 and GS-E3D inhibit microglia and astrocyte activation in the hippocampus of STZ-induced diabetic rats.
A recent report showed that astrocyte and microglia are activated in the hippocampus of diabetic rats and involved in cognitive impairment [56]. Since GS-KG9 and GS-E3D inhibited BBB disruption and macrophage infiltration in the hippocampus of diabetic rats ( Figure 3; Figure 6), we postulated that GS-KG9 and GS-E3D would inhibit microglia and astrocyte activation in the hippocampus of STZ-induced diabetic rats. As shown in Figure 7A, immunostaining with OX-42 antibody, a microglia marker, revealed that microglia morphology displayed a small soma bearing thin-branched or ramified processes in normal rats, indicating a resting state ( Figure 7A, normal). However, OX-42positive cells displayed a significantly-activated morphology, demonstrating cell body hypertrophy and retraction of cytoplasmic processes in STZ-injected diabetic rats ( Figure 7A, STZ), whereas GS-KG9 and GS-E3D treatment markedly reduced the number of activated microglia ( Figure 7A,  STZ+KG9 and STZ+E3D). In addition, immunofluorescence staining with GFAP, an astrocyte marker, showed that astrocytes were also activated morphologically in the hippocampal areas of STZ-treated diabetic rats ( Figure 7B, STZ) as compared with normal controls. However, STZ-induced astrocyte activation was markedly attenuated in GS-KG9 and GS-E3D-treated groups ( Figure 7B, STZ+ KG9 and STZ+E3D). These results indicate that GS-KG9 and GS-E3D inhibit microglia and astrocyte activation in the hippocampus of STZ-induced diabetic rats.

GS-KG9 and GS-E3D Inhibit Microglia and Astrocyte Activation in the Hippocampus of STZ-Induced Diabetic Rats
A recent report showed that astrocyte and microglia are activated in the hippocampus of diabetic rats and involved in cognitive impairment [56]. Since GS-KG9 and GS-E3D inhibited BBB disruption and macrophage infiltration in the hippocampus of diabetic rats (Figures 3 and 6), we postulated that GS-KG9 and GS-E3D would inhibit microglia and astrocyte activation in the hippocampus of STZ-induced diabetic rats. As shown in Figure 7A, immunostaining with OX-42 antibody, a microglia marker, revealed that microglia morphology displayed a small soma bearing thin-branched or ramified processes in normal rats, indicating a resting state ( Figure 7A, normal). However, OX-42-positive cells displayed a significantly-activated morphology, demonstrating cell body hypertrophy and retraction of cytoplasmic processes in STZ-injected diabetic rats ( Figure 7A, STZ), whereas GS-KG9 and GS-E3D treatment markedly reduced the number of activated microglia ( Figure 7A, STZ+KG9 and STZ+E3D). In addition, immunofluorescence staining with GFAP, an astrocyte marker, showed that astrocytes were also activated morphologically in the hippocampal areas of STZ-treated diabetic rats ( Figure 7B, STZ) as compared with normal controls. However, STZ-induced astrocyte activation was markedly attenuated in GS-KG9 and GS-E3D-treated groups ( Figure 7B, STZ+ KG9 and STZ+E3D). These results indicate that GS-KG9 and GS-E3D inhibit microglia and astrocyte activation in the hippocampus of STZ-induced diabetic rats.

GS-KG9 and GS-E3D inhibit the expression of pro-inflammatory cytokines and mediators in STZinduced diabetic rats.
It is known that the activation of microglia and macrophages is associated with the production of pro-inflammatory cytokines and mediators. Astrocytes are also active participants in propagating and regulating neuroinflammation [57]. Since our data showed that GS-KG9 and GS-E3D reduced macrophage infiltration and inhibited microglia and astrocyte activation after STZ injection, we expected that GS-KG9 and GS-E3D would inhibit the expression of proinflammatory cytokines and mediators in the hippocampus of diabetic rats. As shown in Figure 8A, the mRNA expression of proinflammatory cytokines such as tnf-α, il-1β, il-6 and mediators including cox-2 and inos was markedly increased in STZ-induced diabetic rats as compared with normal controls. Furthermore, their expression was significantly decreased by GS-KG9 and GS-E3D treatment at 28 d after drug treatment ( Figure 8A,B). These results suggest that the reduction of neuroinflammation in the hippocampus of GS-KG9 and GS-E3D-treated rats may be mediated in part by inhibiting hyperglycemia-induced glial activation in the hippocampus. However, several reports showed that some inflammatory factors such as IL-6 and Cox-2 are expressed in neurons under the pathological condition [58,59]. Thus, we will examine the effect of GS-KG9 and GS-E3D on the production of proinflammatory factors in neuronal cells.

GS-KG9 and GS-E3D Inhibit the Expression of Pro-Inflammatory Cytokines and Mediators in STZ-Induced Diabetic Rats
It is known that the activation of microglia and macrophages is associated with the production of pro-inflammatory cytokines and mediators. Astrocytes are also active participants in propagating and regulating neuroinflammation [57]. Since our data showed that GS-KG9 and GS-E3D reduced macrophage infiltration and inhibited microglia and astrocyte activation after STZ injection, we expected that GS-KG9 and GS-E3D would inhibit the expression of proinflammatory cytokines and mediators in the hippocampus of diabetic rats. As shown in Figure 8A, the mRNA expression of proinflammatory cytokines such as tnf-α, il-1β, il-6 and mediators including cox-2 and inos was markedly increased in STZ-induced diabetic rats as compared with normal controls. Furthermore, their expression was significantly decreased by GS-KG9 and GS-E3D treatment at 28 d after drug treatment ( Figure 8A,B). These results suggest that the reduction of neuroinflammation in the hippocampus of GS-KG9 and GS-E3D-treated rats may be mediated in part by inhibiting hyperglycemia-induced glial activation in the hippocampus. However, several reports showed that some inflammatory factors such as IL-6 and Cox-2 are expressed in neurons under the pathological condition [58,59]. Thus, we will examine the effect of GS-KG9 and GS-E3D on the production of proinflammatory factors in neuronal cells.

GS-KG9 and GS-E3D inhibit apoptotic cell death of hippocampal neurons in STZ-induced diabetic rats.
There is growing evidence to support the notion that diabetes has adverse effects on the brain, especially on the hippocampus, which is particularly susceptible to neuronal injury and behavior disorder [60][61][62]. Since we showed the inhibitory effect of GS-KG9 and GS-E3D on BBB disruption and neuroinflammation after STZ injection, we postulated that GS-KG9 and GS-E3D would inhibit hippocampal cell damage such as necrosis or apoptosis by STZ-induced hyperglycemia. To examine the cell damage of hippocampal neurons, Nissl staining was performed and hippocampal neurons were counted. As a result, the numbers of pyramidal neurons in hippocampal CA1 region at 32 d after STZ injection were diminished as compared with the normal control group, whereas the loss of hippocampal neurons by STZ injection was attenuated in GS-KG9 and GS-E3D-treated groups ( Figure 9A). Quantitative analysis revealed that the number of viable pyramidal neurons in CA1 was decreased to 50% of those of the normal group by STZ, whereas GS-KG9 and GS-E3D treatment significantly attenuated this loss (STZ, 51.4 ± 7.9; GS-KG9, 75.6 ± 6.2; GS-E3D, 78.9 ± 4.7) ( Figure 9B). To determine whether the loss of hippocampal neurons in diabetic rats was mediated by apoptotic cell death, TUNEL-staining was performed. As shown in Figure 9C, TUNEL-positive cells were mainly observed in hippocampal CA1 region in STZ-injected rats. However, when compared with the STZ-injected group, TUNEL-positive cells were reduced in GS-KG9 and GS-E3D-treated groups. Quantitative analysis revealed that the number of TUNEL-positive cells in the hippocampal CA1 region was significantly lower in the GS-KG9 and GS-E3D-treated groups than in the STZ-injected control group (STZ, 38 ± 7.9; STZ+GS-KG9, 10.5 ± 3.2; STZ+GS-E3D, 9.8 ± 4.1) ( Figure 9C,D). These data imply that GS-KG9 and GS-E3D prevent the loss of hippocampal neurons by apoptotic cell death in STZ-induced diabetic rats.

GS-KG9 and GS-E3D Inhibit Apoptotic Cell Death of Hippocampal Neurons in STZ-Induced Diabetic Rats
There is growing evidence to support the notion that diabetes has adverse effects on the brain, especially on the hippocampus, which is particularly susceptible to neuronal injury and behavior disorder [60][61][62]. Since we showed the inhibitory effect of GS-KG9 and GS-E3D on BBB disruption and neuroinflammation after STZ injection, we postulated that GS-KG9 and GS-E3D would inhibit hippocampal cell damage such as necrosis or apoptosis by STZ-induced hyperglycemia. To examine the cell damage of hippocampal neurons, Nissl staining was performed and hippocampal neurons were counted. As a result, the numbers of pyramidal neurons in hippocampal CA1 region at 32 d after STZ injection were diminished as compared with the normal control group, whereas the loss of hippocampal neurons by STZ injection was attenuated in GS-KG9 and GS-E3D-treated groups ( Figure 9A). Quantitative analysis revealed that the number of viable pyramidal neurons in CA1 was decreased to 50% of those of the normal group by STZ, whereas GS-KG9 and GS-E3D treatment significantly attenuated this loss (STZ, 51.4 ± 7.9; GS-KG9, 75.6 ± 6.2; GS-E3D, 78.9 ± 4.7) ( Figure 9B). To determine whether the loss of hippocampal neurons in diabetic rats was mediated by apoptotic cell death, TUNEL-staining was performed. As shown in Figure 9C, TUNEL-positive cells were mainly observed in hippocampal CA1 region in STZ-injected rats. However, when compared with the STZ-injected group, TUNEL-positive cells were reduced in GS-KG9 and GS-E3D-treated groups. Quantitative analysis revealed that the number of TUNEL-positive cells in the hippocampal CA1 region was significantly lower in the GS-KG9 and GS-E3D-treated groups than in the STZ-injected control group (STZ, 38 ± 7.9; STZ+GS-KG9, 10.5 ± 3.2; STZ+GS-E3D, 9.8 ± 4.1) (Figure 9C,D). These data imply that GS-KG9 and GS-E3D prevent the loss of hippocampal neurons by apoptotic cell death in STZ-induced diabetic rats.
Neuroinflammation in the hippocampus has been reported as one of the main mechanisms involved in the pathogenesis of hyperglycemia-induced neurodegeneration, which eventually leads to anxiety and depression [8]. Moreover, proinflammatory factors such as IL-6, IL-1β, and TNF-α are known to play a critical role in diabetes-causing anxiety and depression [62][63][64]. Recently, it was reported that anti-inflammatory drugs such as minocycline, taurine, and eugenol reduce blood glucose levels in STZ-induced diabetic rats and inhibit neuropathy and brain damage caused by hyperglycemia [65][66][67][68]. Thus, based on these reports, our results suggest that the neuroprotective effect of GS-KG9 and GS-E3D might be mediated by attenuating BBB disruption, thereby preventing neuroinflammation in the hippocampus in STZ-induced chronic diabetic rats. On the other hand, several studies suggest that oxidative stress is involved in the development of diabetic neurotoxicity and that antioxidant therapy can prevent or reverse hyperglycemia-induced nerve dysfunctions [69][70][71][72]. Recent reports showed that GS-E3D alleviates diabetes-related renal dysfunction by inhibiting the production of 8-hydroxy-2 -deoxyguanosine, a DNA oxidation marker [73]. In addition, E3D and KG9 also showed antioxidant effects in human hair dermal papilla cells proliferation and D-galactosamine-induced liver damage animal model, respectively [74,75]. These evidences taken together suggest that the neuroprotective effect of GS-KG9 and GS-E3D in STZ-induced diabetic rats may be mediated in part by inhibiting oxidative stress. Thus, we will investigate the effect of GS-KG9 and GS-E3D on oxidative stress in a future study. In addition, we will determine the underlying mechanism of the neuroprotective effect of GS-KG9 and GS-E3D in diabetic rats and the effect of these drugs on cognition, anxiety, and depression by chronic hyperglycemia. Neuroinflammation in the hippocampus has been reported as one of the main mechanisms involved in the pathogenesis of hyperglycemia-induced neurodegeneration, which eventually leads to anxiety and depression [8]. Moreover, proinflammatory factors such as IL-6, IL-1β, and TNF-α are known to play a critical role in diabetes-causing anxiety and depression [62][63][64]. Recently, it was reported that anti-inflammatory drugs such as minocycline, taurine, and eugenol reduce blood glucose levels in STZ-induced diabetic rats and inhibit neuropathy and brain damage caused by hyperglycemia [65][66][67][68]. Thus, based on these reports, our results suggest that the neuroprotective effect of GS-KG9 and GS-E3D might be mediated by attenuating BBB disruption, thereby preventing neuroinflammation in the hippocampus in STZ-induced chronic diabetic rats. On the other hand, several studies suggest that oxidative stress is involved in the development of diabetic neurotoxicity and that antioxidant therapy can prevent or reverse hyperglycemia-induced nerve dysfunctions [69][70][71][72]. Recent reports showed that GS-E3D alleviates diabetes-related renal dysfunction by inhibiting the production of 8-hydroxy-2'-deoxyguanosine, a DNA oxidation marker [73]. In addition, E3D and KG9 also showed antioxidant effects in human hair dermal papilla cells proliferation and Dgalactosamine-induced liver damage animal model, respectively [74,75]. These evidences taken

Conclusions
In conclusion, this study provides evidence that bioactive extracts from Korean red and white ginseng (GS-KG9 and GS-E3D) attenuate hyperglycemia and BBB disruption by inhibiting MMP-9 activation in STZ-induced diabetic rats. Furthermore, GS-KG9 and GS-E3D treatment reduced macrophage infiltration and glial activation in the hippocampus, leading to the alleviation of the apoptotic cell death of hippocampal neurons. As a possible mechanism of GS-KG9 and GS-E3D to prevent BBB damage, thereby alleviating neurodegeneration, the inhibitory effect of MMP-9 expression and activation by these drugs can be considered. Thus, we will further investigate the epigenetic regulation by this compound for MMP-9 expression in a future study. Taken together, our results suggest that GS-KG9 and GS-E3D could be used for diabetic patients as a health food and/or therapeutic treatment for preserving the BBB integrity of the brain.