Oxygen-glucose deprivation-induced glial cell reactivity in the rat primary neuron-glia co-culture

Abstract

It has been demonstrated that in vivo brain ischemia induces activation and proliferation of astrocytes and microglia. However, the mechanism underlying the ischemia-induced activation and proliferation of these cells remains to be unclear. Oxygen-glucose deprivation (OGD), an in vitro ischemia mimic, has been extensively used to analyze the hypoxia response of various cell types. This study examined the OGD-induced changes in the expression level of astrocytes and microglia marker proteins and immunoreactivity for Ki-67, a marker protein for cell proliferation, using rat primary hippocampal neuron-glia co-culture (NGC) cells. Furthermore, OGD-induced changes in the expression of M1/M2 microglia phenotype-related genes were also examined. MTT assay indicated that 120 min of OGD decreased cell viability, and immunocytochemistry indicated that 120 min of OGD abolished most microtubule-associated protein 2 (MAP2)-immunopositive neurons. In contrast, glial fibrillary acidic protein (GFAP)-immunopositive astrocytes and ionized calcium-binding adapter protein-1 (Iba-1)-immunopositive microglia, and 2’,3’-cyclic nucleotide-3’-phosphodiesterase (CNPase)-immunopositive oligodendrocytes survived OGD. Western blot assays and double-immunofluorescent staining indicated that OGD increased the GFAP expression level and the Ki-67-immunopositive/GFAP-immunopositive cells’ ratio. Real-time PCR analysis showed that OGD altered M1 microglia phenotype-related genes. Specifically, OGD decreased the expression level of CD32 and interleukin-1β (IL-1β) genes and increased that of the inducible nitric oxide synthase (iNOS) gene. Therefore, applying OGD to NGC cells could serve as a useful in vitro tool to elucidate the molecular mechanisms underlying brain ischemia-induced changes in GFAP expression, astrocyte proliferation, and M1 microglia phenotype-related gene expression.

Ischemic neuronal death induces activation and proliferation of astrocytes and microglia. In the ischemic brain, astrocytes and microglia increase the expression levels of the glial fibrillary acidic protein (GFAP) and ionized calcium-binding adapter protein-1 (Iba-1), respectively [27]. In addition, astrocytes and microglia express Ki-67 or proliferating cell nuclear antigen protein [1, 7], which are cell proliferation marker proteins [5, 9]. Recently, the roles of M1 and M2 phenotypes, two distinct phenotypes of activated microglia in the ischemic brain, have been intensively investigated [32]. M1 and M2 microglia are characterized by cell surface antigen proteins, such as cluster of differentiation 32 (CD32) and CD206, respectively. M1 microglia express pro-inflammatory mediators, including the inducible nitric oxide synthase (iNOS), tumor necrosis factor-α (TNF-α), and interleukin-1β (IL-1β), and have harmful effects on neurons. On the other hand, M2 microglia express anti-inflammatory mediators, including interleukin-10 (IL-10) and transforming growth factor-β (TGF-β), and contribute to tissue repair [26]. The expression levels of M1/M2 phenotype-related genes, including CD32, CD206, iNOS, TNF-α, IL-1β, IL-10, and TGF-β, are changed in the ischemic brain [3, 6, 18]. Astrocytes and microglia contribute to brain inflammation and tissue repair in the ischemic brain through their activation and proliferation. Since brain inflammation deteriorates the prognosis of patients with brain infarction [21], it is considered a therapeutic target. Therefore, understanding the mechanisms underlying glial activation, proliferation, and phenotype shift may help develop novel treatments for the ischemic brain.

Animal models have been used for analyzing morphological and functional changes in glial cells in the ischemic brain. However, preparing brain ischemia models requires skilled surgical techniques. Therefore, a simple experimental system is required to analyze the molecular mechanisms underlying morphological and functional changes in postischemic glial cells. Oxygen-glucose deprivation (OGD), an in vitro ischemia mimic, has been extensively used for analyzing the hypoxia response of various cell types [10, 15, 29], and, in this study, the inexpensive Anaeropack system was used as a tool for analyzing such responses [8, 14]. Applying OGD to neuron-glia co-culture (NGC) cells has been extensively used for analyzing the mechanisms of ischemia-induced neuronal cell death or the neuroprotective effect of compounds against ischemia [17, 28, 31]. However, it has been used less for analyzing the changes in the activation and proliferation of glial cells after OGD-induced neuronal cell death.

In the present study, we examined glial cell proliferation and the OGD-induced changes in the expression level of glial marker proteins and M1/M2 microglia phenotype marker genes using rat primary hippocampal NGC. We evaluated whether the application of OGD to NGC cells was a valuable tool to evaluate the mechanisms underlying glial activation, proliferation, and phenotype shift after ischemia-induced neuronal cell death.

MATERIALS AND METHODS

Preparation of rat primary hippocampal NGC

All procedures were carried out in accordance with the recommendations of the Guidelines for the Care and Use of Laboratory Animal Experiment Committee of Osaka Metropolitan University (approval no; 22-33). The pregnant Sprague-Dawley rats (Japan SLC, Inc., Hamamatsu, Japan) were kept under a 12 hr light/dark cycle, and food and water were provided ad libitum. The number of pregnant rats used in this study was 11, and the average number of fetuses used in a single experiment was 10.3. NGC cells were prepared from fetal rat hippocampus according to previously reported methods, with minor modifications [24]. Briefly, pregnant female rats were anesthetized with isoflurane, and embryos were taken out from their uterus at embryonic day 18. Following decapitation and brain extraction, the hemispheres were separated, and the hippocampus was dissected. After carefully removing the meninges, the tissue was digested with 2.5% trypsin. After enzymatic digestion, the tissue was washed and resuspended in a plating medium (PM). PM comprises neurobasal medium (Thermo Fisher Scientific K.K., Tokyo, Japan) supplemented with 2% B27 supplement (Thermo Fisher Scientific K.K.), 2 mM glutamine, 0.3% glucose, 37.5 mM NaCl, and 5% fetal bovine serum. Cells were plated on 24-well plates (1.3 × 10⁵ cells/cm²), 4-well slide chambers (Watson Bio Lab, San Diego, CA, USA) (1.3 × 10⁵ cells/cm²), and 35-mm dishes (1.4 × 10⁵ cells/cm²). They were incubated in a humidified incubator at 37°C and 5% CO₂. After two days, one-half to one-third of the total medium amount was replaced with serum-free medium (SFM). SFM is a neurobasal medium supplemented with 2% B27 supplement, 2 mM glutamine, 0.3% glucose, and 37.5 mM NaCl. The experiments were conducted 14 days after plating the cells on culture plates, dishes, or slide chambers.

OGD treatment

The OGD treatment was performed according to previously described methods, with minor modifications [14]. We used the Anaeropack system (Mitsubishi Gas Chemical Co., Tokyo, Japan) to produce a hypoxic atmosphere in vitro. Cells were washed four times with Earle’s Balanced Salt Solution (EBSS) containing mannitol (mannitol-EBSS) (116.24 mM NaCl, 5.37 mM KCl, 1.80 mM CaCl₂, 0.83 mM MgSO₄, 1.01 mM NaH₂PO₄, 26.19 mM NaHCO₃, and 5.56 mM mannitol) and cultured with mannitol-EBSS. The dishes, plates, or slide chambers were placed into an airtight Anaeropack container incubated at 37°C. For control (non-OGD) conditions, cells were cultured with EBSS containing 5.56 mM glucose in a humidified incubator at 37°C and 5% CO₂.

3-(4,5-di-methylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay

Cell viability was assessed using an MTT assay kit (Nakarai Tesque, Kyoto, Japan). The MTT solution was added and mixed with one-tenth of the medium, and cells were cultured in a 5% humidified CO₂ incubator at 37°C for 2 hr. Then, an equal amount of solubilization solution was added to the medium and mixed well to dissolve the cells. Cells were cultured overnight in a 5% humidified CO₂ incubator at 37°C. The next day, the precipitate was dissolved entirely, and the absorbance was measured with a microplate reader (Bio-Rad Laboratories, Inc., Hercules, CA, USA) at 570 nm with background subtraction at 630 nm.

Immunocytochemistry

The cells were washed with 0.01 M phosphate-buffered saline (PBS) (pH 7.4) and fixed with 4% paraformaldehyde-0.1 M PB (pH 7.4) for 30 min at room temperature (RT). Then the cells were treated with 0.3% H₂O₂-methanol for 20 min at RT, incubated with 3% normal goat serum (Immune BioScience, Miami, FL, USA) or 3% normal donkey serum (Immune Bio Science) for 30 min at 32°C, and incubated overnight at 4°C with the primary antibody solution containing specific microtubule-associated protein 2 (MAP2) (1:1,000; #8707; Cell Signaling Technology, MA, USA), GFAP (1:20,000; MAB360; Merk Millipore, MA, USA), Iba-1 (1:1,500; ab5076; Abcam, Cambridge, UK), 2′,3′-cyclic nucleotide 3′-phosphodiesterase (CNPase) (1:1,800; MAB326; Merk Millipore), or Ki-67 (1:2,500; NB110-89717; Novus Biologicals, LLC., CO, USA) antibodies. Cells were then incubated with the secondary antibody solution containing DyLight 488-labeled goat anti-rabbit IgG (1:600; AB_2338046; Jackson ImmunoResearch Laboratory Inc., West Grove, PA, USA) and DyLight 549-labeled goat anti-mouse IgG (1:2,000; 115-505-266; Jackson ImmunoResearch Laboratory Inc.), or DyLight 488-labeled donkey anti-rabbit IgG (1:600; AB_2313584; Jackson ImmunoResearch Laboratory Inc.) and DyLight 594-labeled donkey anti-goat IgG (1:2,000; ab96933; Abcam) for 1 hr at 32°C. Cell nuclei were stained with Hoechst 33258 (Dojindo, Tokyo, Japan). After cells mounted, immunoreactivity was observed under fluorescent microscopy (Olympus IX71).

Western blot analysis

Cells were lysed in RIPA buffer (10 mM Tris-HCl, 400 mM NaCl, 1 mM EDTA, 0.1% Sodium dodecyl sulfate, 1% TritonX-100, Protease Inhibitor Cocktail [Nakalai Tesque]). Then, protein lysates were centrifuged at 12,000 g for 20 min, and the supernatants were used as samples. Protein concentrations in samples were measured using a Protein Assay Lowry kit (Nakarai Tesque). Aliquots containing 5 to 30 μg of protein were submitted to 10% or 12% SDS-PAGE and transferred onto a PVDF membrane. Membranes were blocked with 3% non-fat milk in TBS-T solution [10 mM Tris-HCl (pH 7.4), 0.15 M NaCl, and 0.05% Tween 20] for 1 hr at RT and incubated overnight at 4°C with primary antibody solution containing specific GFAP (1:10,000; MAB360; Merk Millipore), Iba-1 (1:1,500; ab5076; Abcam), or β-actin (1:3,000; #3700; Cell Signaling Technology) antibodies.

Membranes were washed three times with TBS-T for 10 min each and incubated for 1 hr at RT with the secondary anti-mouse IgG horseradish peroxidase (HRP)-linked antibody (#7076; Cell Signaling Technology) or anti-goat IgG HRP-linked antibody (SA0001-4; Proteintech, Tokyo, Japan). After washing, the membranes were processed with 20× LumiGLO^® Reagent and 20x Peroxide (#7003; Cell Signaling Technology).

Real-time qPCR

mRNA expression levels were assessed using real-time qPCR. Total RNA was isolated from the cell culture using TRIzol (Bioline Ltd., London, UK) according to the manufacturer’s instructions. Reverse transcription of 1 μg of total RNA was performed using M-MLV reverse transcriptase (Promega, Madison, WI, USA) and an oligo (dT) 18 primer at 42°C for 50 min. The expression of CD32, CD206, iNOS, TNF-α, and IL-1β was quantified using Thunderbird SYBR qPCR MIX (Toyobo, Osaka, Japan). The primers used in this study are listed in Table 1. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an endogenous control to normalize gene expressions. The PCR analysis was repeated three times for the experimental group, and the mean values were used for statistical analysis.

Table 1. List of primers used in the present study.

Primer	Direction	Sequence
GAPDH	Forward	5′-ACATCAAATGGGGTGATGCT-3′
	Reverse	5′-GGATGCAGGGATGATGTTCT-3′
CD32	Forward	5′-AATGGCTACTCCCACCACTG-3′
	Reverse	5′-GCACCGGTATTCTTCACTGT-3′
CD206	Forward	5′-CAAGGAAGGTTGGCATTTGT-3′
	Reverse	5′-CAAAGGAACGTGTGCTCTGA-3′
TNF-α	Forward	5′-ATGGGCTCCCTCTCATCAGT-3′
	Reverse	5′-AAATGGCAAATCGGCTGACG-3′
iNOS	Forward	5′-GCATTCAGATCCCGAAACGC-3′
	Reverse	5′-GCCCTCGAAGGTGAGTTGAA-3′
IL-1β	Forward	5′-TTGGGCAGAGAATCCGCAAT-3′
	Reverse	5′-GCTGTCCTCAAGGCTTCCAT-3′

CD32, cluster of differentiation 32; CD206, cluster of differentiation 206; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; iNOS, inducible nitric oxide synthase; IL-1β, interleukin-1β; TNF-α,tumor necrosis factor-α.

Statistical analysis

Statistical analysis was performed using SPSS statistics. The data are expressed as the mean ± SD. Levene’s t-test or the Mann–Whitney U-test was used to analyze statistical differences between two groups. Data were confirmed to present a normal distribution and equal variance using the Kolmogorov-Smirnov test and Levene’s test, respectively. If the data passed the Kolmogorov-Smirnov test and Levene’s test, the statistical analysis of the data was performed using Leven’s t-test, a parametric test. If the data did not pass the Kolmogorov-Smirnov test, statistical analysis was performed using the Mann–Whitney U-test. Statistical analysis of data from experiments containing multiple groups was performed using the Kruskal-Wallis test, followed by the Dunn-Bonferroni test. A P-value inferior to 0.05 was considered significant. All figures indicate significance as follows: *, P<0.05; **, P<0.01; and ***, P<0.001.

RESULTS

Characteristics of hippocampal NGC cells

We characterized the hippocampal NGC cells by immunocytochemical staining. In our NGC cells, MAP2-immunopositive, GFAP-immunopositive, Iba-1-immunopositive, and CNPase-immunopositive cells made for 29.0 ± 5.2%, 40.2 ± 7.2%, 12.9 ± 6.6%, and 3.5 ± 1.8%, respectively. The remaining 14.4 ± 1.6% of cells were immunonegative for MAP2, GFAP, Iba-1, and CNPase (Fig. 1).

Fig. 1.

Representative image of the immunoreactivity for microtubule-associated protein 2 (MAP2), glial fibrillary acidic protein (GFAP), ionized calcium-binding adapter protein-1 (Iba-1), and 2′,3′-cyclic nucleotide-3′-phosphodiesterase (CNPase) of untreated cells.

OGD-induced changes in cell viability and immunostaining

Cells were exposed to OGD for different periods (60, 120, and 180 min) to evaluate its effect on cell death. Cell viability was evaluated by the MTT assay as an indicator of cell death (Fig. 2A). Cell viability was already reduced immediately (0 hr) after 120 min and 180 min of OGD treatment compared to control conditions (non-OGD group). The reduction of cell viability was also observed at 24 hr after 120 min and 180 min of OGD. However, cell viability at 24 hr after 120 min or 180 min of OGD was not significantly different from that at 0 hr after 120 min or 180 min of OGD. On the other hand, OGD for 60 min did not decrease cell viability but increased it at 24 hr of reoxygenation compared to the non-OGD group. We performed immunocytochemistry to observe the morphology of surviving cells after 120 min and 180 min of OGD. Figure 2B presents immunostainings for MAP2, GFAP, Iba-1, and CNPase in NGC subjected to 120 min of OGD. After 120 min of OGD, most MAP2-immunopositive neurons were lost, while many GFAP-immunopositive astrocytes survived. Although no remarkable changes were observed in GFAP immunostaining after OGD treatment at 24 hr of reoxygenation, GFAP immunostaining was a little stronger in OGD-treated NGC cells than in control cells after 48 hr of reoxygenation. Iba-1-immunopositive microglia and CNPase-immunopositive oligodendrocytes also survived after 120 min of OGD. No significant changes were observed in Iba-1 and CNPase immunostainings after OGD treatment. However, oligodendrocytes decreased their processes after 120 min of OGD. MAP2-immunopositive neurons, GFAP-immunopositive astrocytes, and CNPase-immunopositive oligodendrocytes were markedly decreased by 180 min of OGD (Fig. 2C). OGD for 180 min reduced glial processes and enlarged the cell body in surviving GFAP-immunopositive astrocytes. In contrast, Iba-1-immunopositive microglia survived (Fig. 2C). No significant changes were observed in Iba-1 immunostainings after OGD treatment. These results indicate that neurons died after 120 min of OGD, but glial cells survived. Control conditions (non-OGD) did not cause cell morphological changes after 120 min or 180 min of OGD. Constituent cell types in NGC at 72 and 96 hr after OGD treatment were almost the same as those at 48 hr after 120 min or 180 min of OGD. However, prolonged reoxygenation after OGD induced changes in the morphology of GFAP-immunopositive astrocytes. Specifically, cell processes and cell body of GFAP-immunopositive astrocytes tended to be enlarged at 72 and 96 hr after 120 min and 180 min of OGD as compared with 48 hr after 120 min and 180 min of OGD (Supplementary Fig. 1). Therefore, we selected 120 min of OGD as an optimal condition for analyzing ischemia-induced changes in the glial cells in vitro.

Fig. 2.

Effect of oxygen-glucose deprivation (OGD) on neuron-glia co-culture (NGC) cells. (A) Cells were subjected to OGD for 60, 120, and 180 min, then reoxygenation for 0 or 24 hr. Cell survival rate was assayed by 3-(4,5-di-methylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Data are presented as the mean percentage of MTT reduction ± SD from duplicate experiments (n=5), assuming the absorbance of control cells (non-OGD) was 100%. Small circles represent data points. Data were tested using the Kruskal-Wallis test, followed by the Dunn-Bonferroni test. *P<0.05, **P<0.01. Representative image of the immunoreactivity for microtubule-associated protein 2 (MAP2), glial fibrillary acidic protein (GFAP), ionized calcium-binding adapter protein-1 (Iba-1), and 2′,3′-cyclic nucleotide-3′-phosphodiesterase (CNPase) in cells subjected to OGD for 120 min (B) or 180 min (C) followed by reoxygenation for 24 or 48 hr.

Western blot analysis of OGD-induced GFAP expression changes

Brain ischemia was reported to increase the expression levels of GFAP [27]. Therefore, western blot was performed to evaluate quantitatively whether OGD increase GFAP expression levels. At 48 hr of reoxygenation after 120 min of OGD, band intensity of GFAP was stronger than that in control NGC. The OGD-induced increase in the band intensity of GFAP was also observed at 96 hr of reoxygenation after 120 min OGD. On the other hand, the band intensity of β-actin tended to be weaker in NGC subjected to OGD than in control NGC. The ratio of GFAP band intensity to β-actin band intensity was higher at 48 hr of reoxygenation after 120 min of OGD than that in control NGC (t=−20.978, df=8, P<0.001) (Fig. 3A). The ratio of GFAP band intensity to β-actin band intensity was also higher at 96 hr of reoxygenation after 120 min of OGD than that in control NGC (t=−4.816, df=8, P<0.001) (Fig. 3B). Since brain ischemia also increases the expression levels of Iba-1 [27], we evaluated Iba-1 expression levels after OGD by western blot. Unlike GFAP, OGD did not affect Iba-1 expression (Fig. 3A and 3B).

Fig. 3.

Western blot analysis of glial fibrillary acidic protein (GFAP) expression level at 48 hr (A) and 96 hr (B) of reoxygenation after 120 min of oxygen-glucose deprivation (OGD). Data are expressed as the mean of densitometric values of bands obtained from control samples ± SD. Values are expressed as fold change of band density from control samples. Small circles represent data points. n=5. Data was tested using Levene’s t-test. ***P<0.001, **P<0.01. OGD for 120 min increased the protein expression level of GFAP.

OGD-induced changes in immunostainings for Ki-67

We performed immunocytochemical staining for Ki-67 to examine whether OGD induced astrocyte and microglia proliferation. In this study, Ki-67 immunoreactivity was detected in GFAP-immunopositive astrocytes (Fig. 4) but not in Iba-1-immunopositive microglia (Supplemental Fig. 2). The Ki-67-immunopositive/GFAP-immunopositive astrocyte ratio significantly increased in OGD-treated NGC cells compared to control NGC cells at 72 hr of reoxygenation (t=−8.704, df=9.630, P<0.001) (Fig. 4A). In contrast, the OGD treatment did not affect the Ki-67-immunopositive/GFAP-immunopositive astrocyte ratio in NGC cells at 24 hr of reoxygenation (t=−0.17, df=10, P=0.987) (Fig. 4B). In this study, Ki-67 immunoreactivity was also detected in GFAP-immunonegative or Iba-1-immunonegative cells (Supplementary Fig. 2).

Fig. 4.

Immunocytochemistry analysis for Ki-67 in astrocytes. The number of Ki67- or glial fibrillary acidic protein (GFAP)-immunopositive cells was counted in four wells in a slide chamber. The percentage of Ki-67-immunopositive cells in GFAP-immunopositive cells was calculated from Ki-67- or GFAP-immunopositive cells detected in four fields (0.33 × 0.44 mm² per field) at 72 hr (A) and 24 hr (B) after 120 min of oxygen-glucose deprivation (OGD). The data are expressed as mean ± SD and were tested using Levene’s t-test. Small circles represent data points. n=6. ***P<0.001. Arrows indicate Ki-67 immunoreactivity in GFAP-immunopositive cells. OGD for 120 min increased the Ki67-immunopositive/GFAP-immunopositive astrocyte ratio.

Real-time PCR analysis

The OGD-induced expression levels of CD32, an M1 phenotype marker, and CD206, an M2 phenotype marker, were examined by real-time PCR (Fig. 5A and 5B). At 48 hr of reoxygenation, the expression level of CD32 was significantly lower in OGD-treated NGC cells than in control cells (t=7.342, df=7, P<0.01). The same difference was observed after 96 hr of reoxygenation (t=4.132, df=8, P=0.03). In contrast, no significant differences were observed in CD206 expression between OGD-treated and control NGC cells. Since OGD changed the expression level of CD32, the expression level of M1-related genes, including iNOS, IL-1β, and TNF-α, was examined (Fig. 5A and 5B). The expression level of iNOS was significantly higher in OGD-treated NGC cells compared to control cells at 48 hr (P=0.016) and 96 hr of reoxygenation (P=0.036). The expression level of IL-1β was significantly lower in OGD-treated NGC cells compared to control cells at 96 hr of reoxygenation (P=0.007), but no differences were observed at 48 hr of reoxygenation (P=0.337) (Fig. 5A and 5B). No significant difference was observed in TNF-α expression level between OGD-treated and control NGC cells (Fig. 5A and 5B).

Fig. 5.

Real-time qPCR analysis of the expression level of cluster of differentiation (CD)32, CD206, inducible nitric oxide synthase (iNOS), interleukin (IL)-1β, and tumor necrosis factor (TNF)-α at 48 hr (A) and 96 hr (B) after 120 min of oxygen-glucose deprivation (OGD). The data are expressed as the mean ± SD. Small circles represent data points. n=3–5. The results of CD32 and IL-1β expression levels at 48 and 96 hr after OGD treatment were tested using Levene’s t-test, and those of iNOS and CD206 at 48 and 96 hr after OGD treatment were tested using the Mann–Whitney U-test. *P<0.05, **P<0.01. OGD for 120 min affected the expression level of CD32, iNOS, and IL-1β.

DISCUSSION

This study used rat primary hippocampal NGC cells to examine OGD-induced changes in GFAP and Iba-1 expression levels, Ki-67 immunoreactivity, and M1/M2 microglia phenotype-related genes. The main results indicated that: 1) OGD increased the GFAP protein expression; 2) OGD increased the Ki-67-immunopositive/GFAP-immunopositive cells’ ratio; 3) OGD affected the expression of M1 microglia phenotype-related genes, including CD32, iNOS, and IL-1β. These OGD-induced results correspond to those induced after brain ischemia in vivo [3, 6, 7, 18, 27]. Therefore, applying OGD to NGC cells could help elucidate the molecular mechanisms underlying ischemia-induced changes in GFAP expression, astrocyte proliferation, and M1 microglia phenotype-related gene expressions.

GFAP is an intermediate filament protein in astrocytes, and brain ischemia increases GFAP expression levels [27]. However, the functional significance of that increase remains unclear. GFAP deficiency increases ischemia-induced neuronal death but does not affect the morphology of astrocytes in the mouse hippocampus [23]. Although the mechanism by which GFAP deficiency increases neuronal susceptibility to ischemia has not been fully elucidated, it is likely related to ischemia-induced hippocampal long-term potentiation [23]. Applying OGD to NGC cells from GFAP knockout animals will enable more detailed studies at the cellular level to elucidate the role of GFAP in brain ischemia. In the present study, the OGD treatment increased GFAP expression level. Several molecules contributing to the regulation of GFAP expression have been reported. However, the molecular mechanisms leading to the ischemia-induced increase in GFAP expression levels remain unclear. Applying OGD to NGC cells may be a useful tool to identify the molecules contributing to the regulation of ischemia-induced increase in the GFAP expression.

In the western blot, the intensity of the β-actin band tended to decrease in the OGD group compared to non-OGD group. In our preliminary experiments, GAPDH was also used instead of β-actin as an internal standard. We found that the band density of GAPDH also tended to decrease in the OGD-treated cell group (data not shown). Although it is unclear why OGD decreases in β-actin band intensity, the following two factors may be involved in the decrease in β-actin band intensity in OGD group. One is OGD-indued neuronal death. The expression level of β-actin is probably higher in neurons than in astrocytes. The other is OGD-induced increase in the expression levels of astrocyte proteins. Studies using a cerebral ischemia model show increased expression levels of various proteins, including GFAP, S100, and heme oxygenase1 (HO-1), in astrocytes [11, 30]. Although this study showed that the OGD-induced increase in the expression level of GFAP in cultured cells, it is likely that the expression levels of several kinds of proteins also increase in astrocytes after OGD. OGD-induced increase in the expression levels of several kinds of proteins in astrocytes may reduce the relative amount of β-actin in the protein abundance of the sample for western blot. A combination of OGD-induced neuronal loss and OGD-induced increase in the expression level of proteins in astrocyte induced after OGD may have decreased β-actin band intensity in OGD group.

OGD increased the Ki-67-immunopositive/GFAP-immunopositive astrocyte ratio, suggesting that OGD promotes the proliferation of GFAP-immunopositive astrocytes. In this study, OGD increased the Ki-67-immunopositive/GFAP-immunopositive astrocytes’ ratio at 72 hr but not 24 hr of reoxygenation. On the other hand, OGD increased the expression level of GFAP at 48 and 96 hr of reoxygenation. According to these results, the proliferation of GFAP-immunopositive astrocytes may partly contribute to the OGD-induced increase in the GFAP expression level at 96 hr of reoxygenation. In contrast, the contribution of the proliferation of GFAP-immunopositive astrocytes to OGD-induced increase of GFAP expression level at 48 hr of reoxygenation may be weak. Astrocytes generally do not proliferate in the healthy adult brain but slightly in the ischemic brain [7, 22]. In an infarcted brain, proliferated astrocytes contribute to form a glial scar. Glial scar is thought to function as a barrier to prevent the spread of the injured area by confining the ischemia-induced inflammatory response within the injured area. On the other hand, the glial scar is also a barrier to prevent neurite regeneration after ischemia [22]. Therefore, glial scar attracts attention as a therapeutic target for improving the prognosis of stroke patients, and many researchers are vigorously trying to elucidate the mechanism of glial cell proliferation. Applying OGD to NGC cells may help elucidate the mechanism of glial scar formation.

Brain ischemia upregulates Iba-1 and Ki-67 expression in microglia [7, 27]. In our study, OGD did not upregulate Iba-1 and Ki67 expression. The reason remains unclear. Cultured microglia proliferate and increase Iba-1 expression levels [12, 19, 25]. Therefore, the reason why OGD cannot induce microglial proliferation and Iba-1 expression remains to be investigated. Sliced hippocampal culture is also used as an in vitro ischemia model tool. In sliced hippocampal culture, OGD induces microglial proliferation but not astrocytes proliferation [33]. It may be necessary to select an appropriate in vitro ischemia model for the analysis of glial cell proliferation.

In this study, Ki-67 immunoreactivity was also detected in GFAP-immunonegative or Iba-1-immunonegative cells. NGC cells contained oligodendrocytes in addition to astrocytes and microglia as glia cells. Therefore, Ki-67-positive/GFAP-negative cells and Ki-67-positive/Iba-1-negative cells may be oligodendrocytes. Alternatively, they could be undifferentiated cells contained in cultured cells. Further studies are needed to identify cells displaying Ki-67-positive/GFAP-negative or Ki-67-positive/Iba-1-negative immunoreactivity.

In this study, CNPase antibody was used to identify oligodendrocytes. CNPase is recognized as an enzyme that contributes to the formation of myelin. It has been reported that CNPase is expressed in both oligodendrocyte precursor cells and mature oligodendrocytes [13]. However, we did not perform immunostaining with marker proteins for oligodendrocyte precursor cells and mature oligodendrocytes in the present study. On the other hand, oligodendrocyte precursor cells distribute in the adult rat and human brain [13, 16]. Therefore, we believe that our cell culture mimics oligodendrocyte distribution of rat or human brain.

Changes in the expression levels of M1 microglia phenotype-related genes after OGD were not uniform. This result suggests that OGD modifies specific gene expression level. Specifically, OGD increases only iNOS gene expression level among M1 microglia phenotype-related genes including CD32, iNOS, IL-1β and, TNF-α. Brain ischemia increases the expression level of iNOS [3]. Although the functional significance of iNOS in brain ischemia has not been completely elucidated, NO derived from iNOS has cytotoxic effect [2, 4]. Therefore, control of the iNOS expression may be helpful for improving the ischemia-induced neurological deficit. However, the molecular mechanism of ischemia-induced up-regulation of iNOS is not completely understood. The application of OGD to NGC may help to develop a method to control ischemia-induced the expression of iNOS. This study showed for the first time that M1 microglia phenotype-related genes are altered in NGC subjected to OGD. We believe that our results will serve as a basis for further investigation of the molecular mechanisms that cause ischemia-induced changes in the expression levels of M1 phenotype-related genes.

The heterogeneous changes in the expression levels of M1-related genes after OGD may be due to different cell types expressing those genes. Cells other than microglia may also contribute to OGD-induced iNOS gene expression. In brain ischemia animal models, the iNOS expression level increased in microglia and astrocytes [3]. Further studies are needed to clarify the cell types which induce iNOS gene expression after OGD.

OGD did not affect the expression level of CD206. In this study, the analysis of OGD-induced changes in CD206 expression were limited to 48 and 96 hr after reoxygenation. Therefore, we cannot conclude that OGD did not affect the expression level of CD206, and further studies are needed to clarify this point.

The cerebral cortex subjected to infarction is divided into two regions: the core and penumbra region [20]. In the core region, necrotic cell death caused by ATP depletion is dominant. Therefore, cell death in the core region is irreversible. On the other hand, cell death progresses slowly in the penumbra region. It is thought that cell death in the penumbra region is caused by disruption of the intracellular homeostasis balance and salvageable with appropriate treatment. In this study, cell viability was evaluated by MTT at 0 and 24 hr after 120 min of OGD. Compared to the control non-OGD group, cell viability was already reduced at 0 hr after OGD. On the other hand, cell viability at 24 hr after 120 min OGD was not significantly different from that at 0 hr after OGD. Judging from these results of the MTT assay, we believe that OGD-induced cell death is mainly caused by ATP depletion. Therefore, the OGD-induced glial reactivity may correspond to the glial reactivity in the core region of stroke model animals.

CONFLICT OF INTEREST

The authors declare no conflict of interest associated with this manuscript.