Short-Chain Fatty Acids Reduce Oligodendrocyte Precursor Cells Loss by Inhibiting the Activation of Astrocytes via the SGK1/IL-6 Signalling Pathway

Short-chain fatty acids (SCFAs) are known to be actively involved in neurological diseases, but their roles in hypoxic-ischaemic brain injury (HIBI) are unclear. In this study, a rat model of HIBI was established, and this study measured the changes in IL-6 and NOD-like receptor thermal protein domain associated protein 3 (NLRP3), in addition to proliferation and apoptosis indicators of oligodendrocyte precursor cells (OPCs). The mechanism of action of SCFA on astrocytes was also investigated. Astrocytes were subjected to hypoxia in vitro, and OPCs were treated with IL-6. The results showed that SCFAs significantly alleviated HIBI-induced activation of astrocytes and loss of OPCs. SCFA pretreatment (1) downregulated the expression of NLRP3, IL-6, CCL2, and IP-10; (2) had no effect on the proliferation of OPCs; (3) ameliorated the abnormal expression of Bax and Bcl-2; and (4) regulated IL-6 expression via the SGK1-related pathway in astrocytes. Our findings revealed that SCFAs alleviated the loss of OPCs by regulating astrocyte activation through the SGK1/IL-6 signalling pathway.


Introduction
Stroke is often associated with a poor neurological prognosis, and hypoxic-ischaemic brain injury (HIBI) is the main mechanism underlying this effect [1]. Axonal demyelination and neuroinflammation are important features of HIBI [2][3][4]. The adult brain contains a large population of oligodendrocyte progenitor cells (OPCs), which affect myelination and astrocyte-related neuroinflammation. Our previous study showed that HIBI results in axonal demyelination [3,[5][6][7], but the relationship between astrocytes and OPCs is unclear. Activation of astrocytes may be involved in OPC injury.
Recent studies have shown that activated astrocytes secrete inflammatory factors and chemokines that are involved in the regulation of neuroinflammation and that regulating astrocyte overactivation is a promising strategy for the treatment of HIBI [8,9]. The NOD-like receptor thermal protein domain associated protein 3 (NLRP3) inflammasome is one of the best studied and characterized inflammasomes due to its unique response to a wide diversity of stimuli. NLRP3 has been suggested to play a role in a variety of disorders, and its over-activation may lead to severe impairments (i.e., cell swelling, tissue damage, internal bleeding and respiratory disablement) [10,11]. These effects are mediated by caspase-1 activation and secretion of IL-1β, which ultimately triggers the pathophysiological changes of OPCs [3]. There is evidence that IL-1 leads to the increase of IL-6 in inflammatory states, and the latter cytokine may also be involved in NLRP3 mediated OPCS injury. However, the specific mechanism is not clear [12].
Recent studies have highlighted the important effects of natural products on brain diseases. Increasing evidence shows that short-chain fatty acids (SCFAs) are closely related to a variety of brain diseases. SCFAs, including acetate, propionate and butyrate, are produced by the metabolism of dietary fibre by intestinal flora. SCFAs can migrate from the intestine to the brain, play a variety of roles and regulate the function of the central nervous system [3,13]. However, the mechanism by which SCFAs regulate neuroinflammation and OPC loss remains unclear and may involve astrocyte activation. Serum and glucocorticoid-induced protein kinase 1 (SGK1) is a protein kinase that is involved in cell metabolism [14], but the role of SGK1 in astrocyte activation is not clear. Whether SCFAs increase the activation of astrocytes during HIBI through the regulation of the SGK1 signalling pathway is unclear. The present study aims to explore the molecular mechanism of astrocyte activation after treatment with SCFAs in the context of HIBI and to provide a scientific theoretical basis for the treatment of HIBI.

Animals and Study Design
The animals were purchased from Jihui Experimental Animal Breeding Co., Ltd. (Shanghai, China). A total of 100 male SD rats weighing 200-250 g were selected. Pentobarbital (30 mg/kg) was injected intraperitoneally, and anaesthesia induction time was 20 min. In this study, animals were divided into the following groups: (1) the sham operation group, (2) the bilateral common carotid artery occlusion (BCCAO) group (two vessel occlusion, 2VO; data collection was stopped if an animal died prematurely), and (3) the SCFA (3:1:1 acetate: propionate: butyrate, 500 mg/kg, intragastric administration for 7 d before BCCAO) + BCCAO group. The rats were fed under normal oxygen conditions for 1, 3 and 7 d before sacrifice. The animals were euthanized via cervical dislocation. The animal treatment and experimental protocols were approved by the Shanghai Municipal Commission for Animal Protection and Use.
To assess the effect of SCFAs (lentivirus injection via the caudal vein) on astrocyte activation via SGK1, rats were divided into the sham operation group, BCCAO group, BCCAO + SCFA group, BCCAO + SCFA + SGK1 overexpression group, BCCAO + SCFA + si-SGK1 group, and SCFA group. The rats were fed under normal oxygen conditions for 1 d and then sacrificed.

Primary Culture of Astrocytes and Oligodendrocytes (OLs)
Cells were cultured as described in the literature [3]. Briefly, mixed cells were isolated from 1-day-old SD rats. The cells were mixed for 10 d and then rotated at 180 RPM and 36.5 °C for 1 h to remove OPCs and microglia. The cells were maintained under normal conditions for 24 h and then subjected to different treatments. In this study, astrocytes with a purity of more than 90% were cultured. The purified OPCs were cultured in OPC medium (OPCM, proliferation medium) (ScienCell Research Laboratories, USA, Cat. No. 1601) for 1 d in 5% carbon dioxide (CO 2 ) and 95% air at 37 °C.

Treatment of OPCs and Astrocytes
The effect of SCFAs on OPCs under oxygen-glucose deprivation (OGD) conditions were explored. In the OGD study, glucose-deprived DMEM (GD-DMEM) was incubated in 1% oxygen 94% N 2 /5% CO 2 in a tri-gas incubator (Astec, Japan) overnight to obtain OGD-DMEM. The next day, the cells were rinsed with GD-DMEM rather than OPCM and incubated with OPCs in OGD-DMEM for 1 d [15].
The cells were divided into the following groups: the control group, OGD group, and OGD + 40 nmol/L SCFA group.
To determine the effect of SCFAs on NLRP3, IL-6, CCL2, and IP-10 levels in astrocytes, astrocytes were divided into the following groups: the control group, the OGD group, the OGD + 40 nmol/L SCFA group, and the 40 nmol/L SCFA group. In the OGD study, GD-DMEM was incubated in 1% oxygen 94% N2/5% CO2 in a tri-gas incubator (Astec, Japan) overnight to generate OGD-DMEM. The next day, astrocytes were rinsed with GD-DMEM and then incubated in OGD-DMEM for 1 h [1].
To assess the effect of SCFAs on SGK1 levels in astrocytes, cells were divided into the following groups: the control group, OGD group, OGD + 40 nmol/L SCFA group, OGD + 40 nmol/L SCFA + SGK1 group, and OGD + 40 nmol/L SCFA + si-SGK1 group (the transfection method is described in Ref. [6]).

In Vivo Lentivirus Transfection
A U6-sh-SGK1-CMV-GFP lentiviral vector, which was used to silence SGK1 expression and a U6-overexpression-SGK1-CMV-GFP lentiviral vector, which was used to increase SGK1 expression, were constructed by GeneChem (Shanghai, China). The titres were 1 × 10 9 TU/ml, and 10 µl lentivirus was injected into rats in the appropriate groups via the caudal vein.

Plasmid Transfection
The SGK1 overexpression gene or sh-SGK1 was introduced into the pEGFP vector to silence the SGK1 gene after OGD exposure. Plasmids containing 4 µg DNA were dissolved in 240 µL of serum-free Opti-MEM. A total of 10 µL Lipofectamine 2000 (Life, 11,668,027) was dissolved in 240 µL serum-free Opti-MEM mixed with the plasmid solution in a tube for 20 min at room temperature. Astrocytes were cultured in a six-well plate and incubated under OGD conditions for 1 h. Western blot analysis was then carried out as described below.

Double Immunofluorescence
Samples from each group were incubated with antibodies against IL-6 (1:100) and GFAP (1:100) and IL-6R (1:100). To assess the apoptosis and proliferation of OPCs, cells were incubated with anti-cleaved caspase-3 (1:100) and anti-BrdU (1:100) antibodies as described previously [6]. A total of eight randomly selected microscopic fields (each with area of 4.5 mm 2 ) from each section were analyzed at 40 × Cells with an NG2-positive (green) body and a cleaved Caspase-3 (red)-or BrdU (red)-positive nucleus were counted as apoptotic or proliferating OPCs, and the number of apoptotic or proliferating OPCs was divided by the total number of NG2positive cells. Cells that emitted green fluorescence only were counted as nonapoptotic or nonproliferative OPCs. The total number and percentage of apoptotic and proliferative OPCs were calculated and averaged. Cells were incubated with an anti-NG2 or anti-PDGFR-α (1:100) antibody overnight at 4 °C followed by secondary antibody and observed under a fluoroscope. The cells were then incubated with an Alexa Fluor 555-conjugated goat anti-rabbit secondary antibody (1:100, Life, Cat. No. A-21428) and FITC-conjugated goat anti-mouse secondary antibody (1:100 Chemicon International, Cat. No. AP130F) for 1 h. Finally, the cells were counterstained with DAPI and examined under a fluorescence microscope (Olympus System Microscope Model BX53, Olympus Company Pte, Tokyo, Japan).

Electron Microscopy
The animals were perfused with 0.1 M phosphate buffer containing a mixture of 2% paraformaldehyde and 3% glutaraldehyde. Then, the brains were removed, and coronal sections (approximately 1 mm thick) were cut. The sections were then fixed in 1% osmium tetroxide for 2 h, dehydrated, and subsequently processed. Ultrathin sections were cut and observed under an electron microscope (FEI Corporation, Hillsboro, OR).

Statistical Analysis
All data were evaluated using SPSS 13.0. The data are expressed as the mean ± standard deviation. Univariate data with homogeneity of variance were compared between four groups by analysis of variance (ANOVA). In addition, Welch's ANOVA was used for analysis. If the data had homogeneity of variance, the least significant difference (LSD) method was used for multiple comparisons. P < 0.05 was considered statistically significant.

IL-6 and NLRP3 Expression Levels In Vivo
In the corpus callosum (CC), IL-6 expression levels were measured (Fig. 1A-J). IL-6 immunoreactivity was higher in the BCCAO groups ( Fig. 1D-F) than in the control group (Fig. 1A-C). However, SCFAs decreased the expression of IL-6 ( Fig. 1G-I). The number of positive cells is shown in K. The immunoreactive bands of NLRP3 and IL-6, which appeared at approximately 62 and 23 kDa, respectively (Fig. 1L), increased significantly in optical density at 1 day after BCCAO when compared with the controls. This effect was significantly inhibited in the presence of SCFAs (Fig. 1)M and N.

OPC Apoptosis in the CC
After BCCAO, an increase in cleaved caspase-3 levels was observed ( Fig. 2A-J). Cleaved caspase-3 levels in the CC were increased in the BCCAO group (Fig. 2D-F) compared to the control group (2A-2C) 1 d after surgery. These changes were reversed by SCFA treatment (Fig. 2G-J). The observed trend in Bax immunoreactivity was identical to that revealed by double immunofluorescence (Fig. 2K-L), and the trend in the expression of bcl-2 was opposite that in the expression of Bax (Fig. 2K-M).

OPC Proliferation in the Cc
The number of PDGFR-α + OPCs was significantly decreased in the BCCAO group (Fig. 3D-F) compared with the control group ( Fig. 3A-C). However, SCFAs had no effect on the number of BrdU + OPCs (Fig. 3G-J). The observed trends in the protein expression of PCNA and Ki-67 were identical those revealed by double immunofluorescence, and SCFAs had no effect on the levels of these proteins (Fig. 3K-M). The experimental data showed that SCFAs had no effect on the proliferation of OPCs. SCFAs may affect the number of OPCs in other ways.

SCFAs Affect the Expression of SGK1 and GFAP and OPC Apoptosis
SGK1 immunoreactivity in the CC was decreased in the BCCAO group (Fig. 4B) compared with the control group (Fig. 4A). These changes were reversed by SCFA treatment (Fig. 4C). The observed trends in the expression of GFAP and SGK1 were identical to those revealed by double immunofluorescence, and SCFAs had no effect on GFAP expression (Fig. 4D-F). We used electron microscopy to observe nuclear fragmentation in OPCs. Nuclear breakage occurred in the BCCAO group compared with the control group (Fig.4 G-I). SCFAs decreased SGK1 levels and reduced OPC loss, and in vitro experiments showed that SCFAs alone had no effect on OPC apoptosis (Fig. 4J-L). These results suggest that SCFAs affect the loss of OPCs in other ways and that astrocyte-derived IL-6 may be involved in this effect.

Expression of IL-6R in OPCs In Vivo
IL-6R was localized to OPCs and was colocalized with NG2 ( Fig. 5A-J). IL-6R immunoreactivity in OPCs was increased in the BCCAO group (Fig. 5D-F) compared with the control group ( Fig. 5A-C). SCFAs had no effect on IL-6Ra levels ( Fig. 5G-I). The number of IL-6Ra-positive cells is shown in Fig. 5K. The observed trend in the IL-6R protein level was identical to that revealed by double immunofluorescence (Fig. 5L-M).

IL-6 Regulates OPC Apoptosis In Vitro
Double immunofluorescence showed that very few OPCs were double labeled with NG2 and caspase-3 at 1 day after treatment with IL-6 or IL-6+IL-6Ra and the matching controls ( Fig. 6A-I). There were more NG2-labeled OPCs that were caspase-3 immunoreactive at 1 d aftertreatment with IL-6 ( Fig. 6D-F) in comparison with the control (Fig.6A-C). However, IL-6Ra significantly reversed the caspase-3 approximately 23KD and 30KD, respectively (Figs. 6K), decreased (Bcl-2) or increased (Bax) significantly in optical density at 1 d after after treatment with IL-6 as compared with the controls (Figs. 6L, M). However, IL-6Ra significantly improve the above phenomena (Figs. 6L, M). These data demonstrated that the apoptosis of OPCs was significantly activated after treatment with IL-6 as compared with the controls.

SCFAs Can Inhibit IL-6 Expression In Vitro
Double immunofluorescence showed that IL-6 was increased at 1 h after treatment with OGD ( Fig. 7D-F) in comparison with the control (Fig. 7A-C). However, SCFAs significantly reversed the IL-6 immunoreactivity in astrocytes as treated with OGD ( Fig. 7G-I). The immunoreactive bands of NLRP3 and IL-6 protein levels that appeared at approximately 62 kDa,23 kDa, respectively (Fig. 7J), increased significantly in optical density at 1 h after after treatment with OGD as compared with the controls. However, SCFAs significantly improve the above phenomena (Fig. 7K, l). These data demonstrated that the expression of IL-6 and NLRP3 was significantly inhibited after treatment with SCFAs as compared with the OGD.

SCFAs Regulate CCL2 and IP-10 Levels in Astrocytes
CCL2 expression was upregulated in the OGD group compared with the control group at 1 h. These changes were reversed by SCFA treatment (Fig. 8G-I). The levels of CCL2 and IP-10 were increased in the OGD group compared with the control group at 1 h (Fig. 8J-L). SCFAs alleviated the overexpression of the CCL2 and IP-10 proteins. The above results show that SCFAs regulate astrocyte activation in vitro; however, the mechanism is unclear.

SCFAs Regulate IL-6 Expression via SGK1 In Vivo and In Vitro
We hypothesized that SCFAs regulate IL-6 expression via the SGK1 signalling pathway. SGK1 expression was markedly downregulated in the OGD group compared with the control group at 1 h. SCFAs increased SGK1 protein levels, and this effect was weakened by treatment with SGK1 siRNA. When SGK1 was overexpressed, the opposite result was obtained (Fig. 9A-B). The protein level of IL-6 was markedly increased in the OGD group compared with the control group. Overexpression of the SGK1 gene alleviated the overexpression of the IL-6 and CCL2 proteins. si-SGK1 transfected had the opposite effect (Fig. 9C-E). To further verify the regulatory effect of SGK1 on IL-6 expression, in vivo experiments were  performed. The results showed that the effect of SCFAs was weakened after transfection of si-SGK1. However, SGK1 overexpression had the opposite effect ( Fig. 9F and  G). Therefore, the in vivo and in vitro data confirmed the effectiveness of SCFAs in reducing astrocyte activation via the SGK1/IL-6 signalling pathway.

Discussion
Excessive inflammatory factor release is a sign of astrocyte activation [16]. SCFAs are known to be actively involved in neurological diseases, but their roles in HIBI are unclear [10]. The results showed that NLRP3 and IL-6 expression  -3 (B, E, H, red). The co-localized expression of NG2 and caspase-3 can be seen in panels C, F, I. Bar graph in J shows the number of caspase-3/NG2-positive OPCs/mm2 at 1 d after the IL-6 administration or IL-6+IL-6R atreatment when compared with the corresponding control. Note that IL-6Ra attenuates the apop-tosis of OPCs and IL-6Ra may reverse the course. Panel K shows Bcl-2 (30 kDa), Bax (23 kDa), and â-actin (42 kDa) immunoreactive bands. Bar graphs in L and M show significant decreases or increases in the optical density of Bcl-2 or Bax at 1 d after IL-6 administration when compared with the corresponding controls. The IL-6Ra antagonist may attenuate the increment was upregulated in rats after BCCAO and that activated astrocytes were the source of IL-6. Our results showed that the expression levels of the IL-6R were increased by BCCAO, and IL-6R on OPCs and OPCs may be the target cells of IL-6. OPCs are NG2-or PDGFR-α-positive bipolar cells that sequentially differentiate into OLs [17][18][19]. OLs are myelinating cells that are present in the central nervous system throughout adulthood. OPCs differentiate into OLs, and loss of OPCs results in neurotransmission dysfunction [20][21][22]. The results showed that cleaved caspase-3 and bax were significantly increased and bcl-2 levels were decreased after HIBI, suggesting that OPC apoptosis occurs after HIBI and that SCFAs can ameliorate apoptosis. Our results showed that SCFAs did not affect the proliferation of OPCs at 1 d. GFAP is a marker of astrocytes. SCFAs did not affect GFAP levels. SCFAs may affect the function of astrocytes, but the mechanism is unknown. SGK1 is a member of the AGC subfamily of protein kinases and is involved in cell metabolism [22][23][24][25]. Our results showed that SCFAs increased the expression of SGK1, indicating that they may attenuate astrocyte overactivation via SGK1 and that IL-6 may be involved in this process. To verify these findings, in vitro experiments were carried out. The results preliminarily showed that SCFAs could regulate IL-6 and NLRP3 expression in vitro. NG2-positive cells are thought to be mainly OPCs [26], and apoptosis of OPCs is recognized as the main mechanism of OPC loss [27]. The in vitro results showed that the number of NG2-/cleaved caspase-3-positive cells was increased after IL-6 treatment. IL-6Ra prevented OPC loss, indicating that IL-6 can promote apoptosis of OPCs. Controlling the secretion of IL-6 is the key to reducing the loss of OPCs. Overexpression of chemokines is an indicator of astrocyte activation [28][29][30]. The results showed that SCFAs inhibited the overexpression of cytokines. We hypothesized that SCFAs inhibit astrocyte activation by promoting the SGK1 signalling pathway. This study found that SGK1 expression levels were decreased after OGD exposure and that SCFAs increased SGK1 levels after OGD exposure. IL-6 and CCL2 expression levels increased after OGD exposure, whereas SCFAs increased IL-6 and CCL2 expression. In addition, when siRNA-SGK1 was added, this effect was ameliorated, indicating that SCFAs regulate the expression of IL-6 and CCL2 through SGK1. However, when SGK1 was overexpressed, the effect of SCFAs was strengthened, suggesting that SCFAs inhibit IL-6 expression via SGK1. To further determine the underlying mechanism, we performed in vivo validation experiments and obtained the same results. Therefore, SCFAs alleviate the activation of astrocytes by regulating the SGK1/IL-6 signalling pathway.
In summary, activated astrocytes in the CC produce excess IL-6, which induces neuroinflammation and OPC loss. However, SCFAs decrease IL-6 levels. Further analysis revealed that SCFAs reduce IL-6 levels by regulating the SGK1 signalling pathway in astrocytes. Further research on this process will facilitate the development of effective therapeutic strategies for relieving neuroinflammation.

Data Availability
The data can be provided if necessary.

Conflict of interest
The authors declare that they have no competing interests.
Ethical Approval This study was approved by the Laboratory Animal Ethical and Welfare Committee of Xin Hua Hospital Affiliated with Shanghai Jiao Tong University School of Medicine (XHEC-F-2018-038).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.

Fig. 9
SCFAs regulates IL-6 via SGK1 in astrocyte. A Western blot of protein expression levels of SGK1 in astrocytes at 1 h (42 kDa). B Significantly downregulated protein expression level of SGK1 at after OGD. C IL-6 and CCL2 levels in astrocytes at 1 h. D-E) Significantly upregulated protein level of IL-6 and CCL2 at 1 h after treatment with OGD compared to control group. overexpression-SGK1 can improve IL-6 and CCL2 proteins. The effect weakens after adding si-SGK1. (G) Significantly downregulated protein expression level of IL-6 after treatment with SGK1 gene overexpression after OGD + SCFAs compared with that of the OGD + SCFAs. N = 5 experiments were performed in independent preparations. *P < 0.05