Autocrine S100B in astrocytes promotes VEGF-dependent inammation and oxidative stress, and causes impairment of neuroprotection

Minimal hepatic encephalopathy (MHE) is strongly associated with neuroinammation. Nevertheless, the underlying mechanism of the induction of inammatory response in MHE astrocytes remains unclear. In this study, we further investigated the effect and mechanism of S100B, predominant isoform expressed and released from mature astrocytes, on MHE-like neuropathology in the MHE rat model. We discovered that S100B expressions and autocrine were signicantly increased in MHE rats and astrocytes isolated from MHE rats. Furthermore, we found that S100B stimulates VEGF expression via the interaction between TLR2 and RAGE in an autocrine manner. S100B-facilitated VEGF autocrine expression further led to a VEGFR2 and COX-2 interaction, which in turn induced the activation of NF (cid:0) B, eventually resulting in inammation and oxidative stress caused by MHE astrocytes. Compared to WT astrocytes, impairment of MHE astrocytes supported neuronal growth in co-culture. To sum up, comprehensive-understanding of the impact of S100B-overexpressed MHE astrocyte on MHE pathology may provide insights into the etiology of MHE. as ratios of relative light units of treatment This study determined an essential role of astrocyte dysfunction in MHE. Our data suggested a role of S100B in mediating astrocyte activity and MHE pathology via an autocrine manner. We found that S100B-induced pathological changes in astrocyte activity during MHE may accelerate inammation. In addition, we discovered that the VEGF-dependent inammation regulates the involvement of S100B in the pathogenesis of MHE. MHE astrocytes trigger neuronal impairment and halt the disease progression. We also found that MHE astrocytes maintained S100B overexpression in vivo and were functionally impaired in neuronal support, increasing VEGF expression. The autocrine mechanisms of S100B pro-inammatory action were associated with the binding of the VEGF and VEGFR and the interaction of VEGFR and COX-2. In conclusion, we suggested that targeting S100B may result in disease-modifying therapies and might be used as a pharmacological for MHE.

crystals of Cysin HBSS-Hepes (20mM)). Percoll gradient (Sigma) was used to remove myelin and lyse red blood cells. The remaining cells were plated in poly-D-lysine-coated 75 cm 2 culture asks at a concentration of 15 ×10 6 cells in 1% serum-containing DMEM/F12 medium for 72h. The medium was changed every 72h.

Primary cortical neurons culture
Primary cortical neuron cultures were prepared from the freshly dissected cerebral cortex from 1-day-old rat pups and dissociated to single-cell suspensions using papain digestion. Then, cells were placed in poly-L-lysine-precoated six-well plates, and co-cultured with adult primary astrocytes cultured in the insert Transwell dishes (Millipore) in Neurobasal ® Medium (1X) supplemented with 0.5 mM GlutaMAX™-I, B-27 ® for 24 h. The pore size of the membrane was 0.4 mm.

NF B -sensitive luciferase Assay
Astrocytes stably transfected with NF B reporter gene and VEGF siRNA were cultured on 12-well plates.
The cells were starved from serum overnight and then stimulated with S100B (2 μM) for 6h. The cells were lysed with 1×luciferase cell culture lysis reagent (Promega, Madison, WI). Supernatants from cell lysates were subjected to luciferase assay according to the manufacturer's instruction (Stratagene, LaJolla, CA). Brie y, 20 μl of supernatant were mixed with 100 μl of luciferase assay buffer (40mM Tricine, pH7.8, 0.5mM ATP, 10mM MgSO 4 , 0.5mM EDTA,10mM 1,4-dithiothreitol, 0.5mM coenzyme A, and 0.5mM luciferin), and the intensity of luminescence was immediately measured using a luminometer (Lumat LB9507; Berthold Technologies, OakRidge, TN). Samples were assayed in triplicate, and the luciferase activity was normalized based on protein concentrations of the cell lysate. The results are presented as ratios of relative light units of treatment groups to control groups.
Detection of COX-2 Activity COX-2 Activity Assay Kit (Abcam Inc., Boston, MA, USA) was used to detect COX-2 activity in the homogenate of MHE astrocytes with S100B treatment in preincubation of PTK787 following the manufacturer's instructions.
Real-time quantitative PCR (qPCR) Total RNA was extracted using TRIzol reagent (Invitrogen). Brie y, 1 μg of total RNA was reverse transcribed into cDNA using Superscript II reverse transcriptase (Invitrogen). Real-time PCR was performed with SYBR Green mix (1:20,000 SYBR Green) on an iCycler iQ real-time detection system (BioRad) according to the manufacturer's instructions. Each reaction was performed in duplicate or triplicate on at least three samples for each condition, in a volume of 25 μl. The mRNA expression was quanti ed using the relative 2 (−∆∆C(T)) method. Relative levels of mRNA were normalized to the values of GAPDH mRNA for each reaction. Each DNA sample was then separated by agarose gel. DNA fragments were subsequently visualized and photographed on a long wave UV lightbox. PCR primers are listed in Table 1.
Flow cytometry for evaluating cell apoptosis N2A cells were co-cultured with adult primary astrocytes culture in an insert Transwell dish for 48 h and harvested. Harvested N2A cells were washed in PBS, then with Binding Buffer (1 mM HEPES, 14 mM NaCl, 2.5 mM CaCl 2 in PBS). Cells were then incubated in Binding Buffer containing PE-conjugated Annexin V (eBioscience, San Diego, CA) for 10 min at room temperature. Dead cells were detected with DAPI solution (0.1 μg/ml). BD FACSCanto™ II ow cytometer equipped with FacsDiva program was used for analysis.

ELISA assay
Enzyme-linked immunosorbent assay (ELISA) kits (R&D Systems, Minneapolis, MN, USA) were used to measure extracellular S100B/VEGF/TNFα levels in the culture medium of MHE astrocytes according to the manufacturer's recommendations. Levels of these cytokines were then analyzed using a Thermo Fisher Multiskan MCC plate reader (Waltham, MA, USA) by the spectrophotometrical method.
Densitometrical quanti cations were indicated as fold change relative to control as fold of proteins to βactin levels.

Fluorescent staining
Four-micron frozen cerebral cortex sections on slides or cells cultured on glass coverslips were xed with 4% paraformaldehyde for 30 min and blocked with PBS containing 3% BSA and 0.2% Triton-x 100 for 1 h.

Assessment of NADPH oxidase activity
Photon emission from the chromogenic substrate lucigenin as a function of acceptance of electron/O2generated by the NADPH oxidase complex was measured at speci c intervals in a Bioteck luminometer (Bioteck, USA). NADPH oxidase assay buffer containing 250 mM HEPES (pH7.4), 120 mM NaCl, 5.9 mM KCl, 1.2 mM MgSO 4 (7H 2 O), 1.75 mM CaCl 2 (2H 2 O), 11 mM glucose, 0.5mMEDTA, 100 μM NADH and 5μM lucigenin was used. The data were converted to relative light units/min/mg of protein. The lucigenin activity of MHE astrocytes homogenate from the control group was arbitrarily set at 100%.

ROS Detection Assay
MHE astrocytes were tested for reactive oxygen species generation using a ROS-ID Total ROS Detection Kit (Enzo Life Sciences) following manufacturer's protocol. Cells were treated with ROS detection reagent and incubated for 4h at 37°C in the dark. Following incubation, ROS level was analyzed via uorescence microscopy. Obtained data were expressed as a percentage.
MTT assay N2A cell viability was evaluated by MTT assay. Brie y, cells in 96-well plates were treated using a medium containing 0.5 mg/ml MTT at 37 °C for 3 h and then incubated using dimethylsulfoxide. The light absorption at 490 nm was measured by quantitative colorimetric assay with a microplate reader (TECAN).

Statistical analysis
All of the data were expressed as mean±SD. Data comparisons were analyzed using either Student's t-test (two-group comparison) or one-way analysis of variance (ANOVA) followed by Dunnett's post hoc multiple comparison test (more than two groups). P < 0.05 or P < 0.01 was considered as statistical signi cance.

Results
Astrocytes from MHE rats overexpress S100B In astrocytes, S100B, a cytokine marker of astroglial activation in brain disorders, induces proin ammatory mediators' secretion (Hu and Van Eldik, 1996). Thus, we sought to determine whether S100B was elevated in astrocytes in neurological diseases MHE. We examined the expression of S100B using IB analysis and found that the expression of S100B was signi cantly increased in cortexes of MHE rats ( Figure 1A and 1B). A maximal increase in S100B mRNA transcription was observed in cortexes (qPCR, Figure 1C and 1D), suggesting overexpression of S100B in the MHE brain.
To further investigate astrocytes' role in cortexes of MHE rats, we established adult astrocytes culture from MHE rats and WT rats. The astrocyte cell culture purity was analyzed by immuno uorescence staining with an astrocyte-speci c marker (GFAP) and other cell-type-speci c markers for neurons (MAP2), microglia (IBA1), endothelial cells (CD31), and oligodendrocytes (O4). We found that 99% of the cells were stained with GFAP ( Figure 1E). We further evaluated our cell culture using RT-PCR by targeting these speci c cell markers. We found that cultured astrocytes had an enriched expression of astrocytespeci c gene GFAP and low expression of other cell-type-speci c genes compared to the whole brain ( Figure 1F). Altogether, these experiments showed that this culture protocol generates a highly enriched astrocyte culture. Then, S100B production and secretion were assessed using IB analysis, qPCR, and ELISA assay. IB analysis and qPCR assay showed that MHE astrocytes showed a signi cantly higher S100B protein ( Figure 1G and 1H) and mRNA expression ( Figure 1I and 1J) compared to WT astrocytes. Moreover, an ELISA assay suggested that the level of S100B released from MHE astrocytes into the cell culture medium was obviously elevated compared to WT astrocytes ( Figure 1K). Thus, we deduced there was an overload of S100B in MHE astrocytes.
Furthermore, to demonstrate the speci city of S100B in in ammation and oxidative stress, we inhibited S100B using an anti-S100B antibody in media and siRNA-induced-silencing in MHE astrocytes. We rst tested the e ciency of S100B siRNA transfection into MHE astrocytes. As determined by qPCR, MHE astrocytes showed weak S100B expression after S100B siRNA transfection ( Figure 1L) compared to control siRNA transfection, indicating e cient transfection. Then, we measured the in ammatory factors expression and secretion in MHE astrocytes. We found that the level of TNFα in the cell culture medium of MHE astrocytes was signi cantly increased compared to WT astrocytes. After knocking down the S100B in MHE astrocytes, the TNFα secretion was reduced or even inhibited ( Figure 1M), indicating the induction of in ammation in MHE astrocytes by S100B.
MHE astrocytes are associated with S100B and VEGF production Astrocyte-derived VEGF, as a pro-in ammatory cytokine (Byzova et al., 2000), has a signi cant role in lesion pathogenesis of CNS disease characterized by in ammation (Baker et al., 1995;Weis and Cheresh, 2005). In this study, we investigated whether autocrine pro-in ammatory S100B from astrocytes affects the VEGF production and secretion. First of all, we examined the association of S100B and VEGF in MHE astrocytes in vivo and in vitro. Immunostaining of the MHE rat's brain revealed that S100B was colocalized with VEGFa in astrocytes ( Figure 2A). Furthermore, confocal images from MHE astrocytes cultured in vitro showed colocalization of S100B with VEGF-A in the cytoplasm ( Figure 2B).
Next, we addressed whether MHE astrocytes-derived S100B enhances VEGF production in an autocrine manner using S100B siRNA transfected-MHE astrocytes. IB and qPCR analysis showed signi cant elevation in VEGF protein ( Figure 2C and 2D) and mRNA expressions ( Figure 2E and 2F) in MHE astrocytes; however, these levels were decreased in cells transfected with S100B siRNA. Similar results were obtained from the ELISA assay; the cell culture medium of MHE astrocytes transfected with the control vector showed a signi cant increase in VEGF level, while S100B-de cient MHE astrocytes displayed decreased VEGF secretion ( Figure 2G). Hence, our data suggested that autocrine S100B facilitated VEGF overproduction by MHE astrocytes.
Next, we assessed whether VEGF autocrine accumulation impacted the S100B expression in MHE astrocytes using VEGF siRNA transfection into MHE astrocytes. VEGF siRNA obviously decreased VEGF mRNA in MHE astrocytes by qPCR assay ( Figure 2H), indicating successful transfection of VEGF siRNA. More importantly, IB analysis for S100B expression showed no change in S100B protein level after VEGF siRNA transfection in MHE astrocytes ( Figure 2I and 2J). Additionally, the ELISA assay also showed that the level of S100B in the cell culture medium of MHE astrocytes was not altered after AVAB treatment ( Figure 2K), indicating that changes in VEGF do not affect the S100B overload. S100B-enhances VEGF production through a TLR2 and RAGE interaction TLRs are involved in the release of in ammatory cytokines (Giuseppe et al., 2007). A previous study suggested that S100B protein binds to RAGE and, in turn, activates TLRs (Tian et al., 2007). We estimated that autocrine S100B might trigger TLR activation via stimulation of RAGE, leading to VEGF production in MHE astrocytes. Thus, we tested the binding between RAGE and S100B in vitro using a coimmunoprecipitation experiment. The results showed that RAGE coimmunoprecipitated with S100B in MHE astrocytes ( Figure 3A). We also validated this result by performing a reciprocal experiment using anti-rage, and found that RAGE and S100B physically interacted in MHE astrocytes ( Figure 3A) Next, we analyzed the association of RAGE with TLR in MHE rats in vivo using immunostaining experiments. Confocal images showed the co-localization of TLR2 with RAGE combined with GFAP in astrocytes in cortexes in MHE rats ( Figure 3B), whereas the co-localization of TLR4 with RAGE was not found in astrocytes of cortexes ( Figure 3C). Furthermore, we tested the effect of S100B on the interaction between RAGE and TLR2 or TLR4 in MHE astrocytes in vitro using co-immunoprecipitation experiments. The results showed that RAGE was coimmunoprecipitated with TLR2 but no TLR4 in MHE astrocytes. We also validated this result by performing a reciprocal experiment using anti-TLR2. RAGE and TLR2 physically interacted, while RAGE failed to be coimmunoprecipitated with TLR2 after ASAB treatment ( Figure 3D). Thus, we deduced that S100B elicits the RAGE and TLR2 but no RAGE and TLR4 interaction.
To assess whether S100B stimulates the VEGF expression through RAGE/TLR2 signaling, we incubated MHE astrocytes with RAGE inhibitor FPS-ZM1 and TLR2 inhibitor C29 preincubated with various concentration of S100B. Quantitative analysis by IB showed an increase in S100B-dependent VEGF production in a dose-response manner. To evaluate the contribution of RAGE and TLR2 interaction to S100B activity, we used RAGE inhibitor FPS-ZM1 and TLR2 inhibitor C29 in a competition assay. We discovered that FPS-ZM1 or C29 signi cantly reduced the effect of a high dose of S100B on VEGF expression in MHE astrocytes ( Figure 3E and 3F). Similar results were found by ELISA assay; FPS-ZM1 or C29 dramatically decreased the high dose of S100B-induced VEGF secretion in MHE astrocytes ( Figure  3G). Hence, we deduced that RAGE and TLR2 are required for S100B-mediated VEGF production.
Autocrine S100B-mediated VEGF triggers VEGFR2 and COX-2 interaction The roles of VEGF receptors (Saban et al., 2007) and COX-2 (Harris et al., 2002) are associated with in ammatory processes. VEGFR2 participates in VEGF-mediated COX induction (Smith et al., 1998). Therefore, we categorized the contribution of VEGFR2 or COX-2 to S100B-mediated VEGF activity. First, we used mouse anti-COX-2 and anti-GFAP antibodies combined with VEGFR1/2 to test the co-localization of VEGFR1/2 and COX-2 in vivo. Confocal images con rmed the presence of VEGFR2, but not VEGFR1, colocalized with COX-2 in the cortex astrocytes of MHE rats ( Figure 4A and 4B). Next, we assessed the effect of S100B on the interaction between VEGFR and COX-2 in MHE astrocytes in vitro using coimmunoprecipitation analysis. As shown in Figure 4C, substantial COX-2 was coimmunoprecipitated with abundant VEGFR2 but not with VEGFR1. We also validated this result by performing a reciprocal experiment using anti-VEGFR2. Our result indicated that VEGFR2 and COX2 physically interacted in MHE astrocytes, whereas ASAB addition induced no interaction of VEGFR2 with COX2, thus suggesting that S100B facilitates the interaction of VEGFR2 with COX-2.
Next, to investigate whether VEGFR2/COX-2 signaling was implicated in S100B-mediated autocrine VEGF, we treated MHE astrocytes with AVAB or VEGF siRNA transfection in the presence of S100B. qPCR analysis showed that AVAB decreased the VEGFR2 mRNA expression in MHE astrocytes ( Figure 4D and 4E). Moreover, the knockdown of VEGF and AVAB attenuated the effect of S100B and decreased the expression of COX-2 ( Figure 4F-H). Hence, our data indicated that VEGF is required for S100B-mediated VEGFR2/COX-2 signaling. S100B-mediated activation of NF B depends on autocrine VEGF by astrocyte Nuclear factor kappa-light-chain-enhancer of activated B cells (NF B), as a central transcription factor, has a major role in pro-in ammatory response (Uwe, 2003). Studies have found that S100B can induce nuclear translocation and transcriptional activity of NF B (p65) (Arumugam et al., 2004;Hofmann et al., 1999a). However, the exact mechanism of S100B-mediated NF B activation in an autocrine action still remained unclear. Thus, we examined the contribution of S100B autocrine to activation of NF B in a competition assay. Immunoblots of cytoplasmic and nuclear NF B protein showed that nuclear NF B content was signi cantly increased in MHE astrocytes. Moreover, the addition of S100B substantially enhanced NF B nuclear translocation. Then, we evaluated the contribution of VEGF to S100B activity. Silencing of VEGF by VEGF siRNA transfection signi cantly diminished S100B-mediated NF B nuclear translocation in MHE astrocytes, indicating the involvement of VEGF in the S100B-response NF B nuclear translocation ( Figure 5A-C).
Next, we used a luciferase-based reporter gene construct to test an NF B-sensitive gene transcription. NF B-dependent transcription indeed was activated in MHE astrocytes, and the enhancement of transcriptional response was observed after treating cells with exogenous S100B. However, knockdown of VEGF effectively blocks the S100B-mediated activation of the NF B-dependent transcription ( Figure   5D), indicating the role of VEGF in the S100B-response NF B activation.
To assess whether S100B affects NF B activation through COX-2, we treated MHE astrocytes with S100B together with VEGFR2 inhibitor PTK787 or COX-2 inhibitor celecoxib. Immunostaining con rmed that NF B was largely localized to the nucleus in S100B-treated MHE astrocytes. After AVAB, PTK787, or celecoxib treatment, NF B was hardly detected in the cell nucleus ( Figure 5E), suggesting an involvement of VEGFR2 and COX-2 in S100B-dependent NF B nuclear translocation. We then examined the levels of pro-in ammatory factors using qPCR analysis. S100B induced an excessive TNFα level in MHE astrocytes, while AVAB addition signi cantly diminished S100B-induced TNFα production ( Figure 5F and 5G), indicating the responsibility of VEGF for the S100B-induced in ammatory action. S100B-mediated oxidative stress depends on autocrine VEGF by astrocytes Elevated oxidative stress (Khaper and Singal, 2001) and in ammation, (Aukrust et al., 1999) which have been associated with the severity of disease, have been reported in a variety of pathophysiological conditions. The linkage between increased oxidative stress and acute-phase in ammation is already well understood (Danielski et al., 2003). Like in ammation, S100B is implicated in oxidative stress. In this study, we transfected MHE astrocytes with S100B siRNA to determine the contribution of S100B in oxidative stress. ROS production assay showed that ROS level was obviously increased in MHE astrocytes. However, the ROS production was signi cantly lowered with the knockdown of S100B ( Figure  6A). Moreover, IB analysis of NADPH oxidase expression showed signi cant increases in NOX1/NOX2 level in MHE astrocytes, while the knockdown of S100B signi cantly reduced the protein expression ( Figure 6B and 6C). In addition, analysis of NADPH oxidase activity displayed an obvious increase in MHE astrocytes, while ASAB treatment led to decreased activity ( Figure 6D), indicating the induction of in ammation in MHE astrocytes by S100B.
The binding of VEGF to VEGFR2 results in a burst of ROS (Obrosova et al., 2006). To examine whether S100B autocrine enhances the oxidative stress through VEGF autocrine in MHE astrocytes, we preincubated MHE astrocytes with S100B and then transfected cells with VEGF siRNA. The addition of S100B further ampli ed ROS production, while the silencing of VEGF lowered the levels of ROS in S100Btreated MHE astrocytes, indicating the responsibility of VEGF for the S100B-induced in ammatory action ( Figure 7A). Moreover, the knockdown of VEGF signi cantly diminished the action of S100B, indicating the involvement of VEGF in the S100B-response NOX1 and NOX2 regulation ( Figure 7B and 7C).
NF-kB up-regulates NADPH oxidase activity, leading to an increase in ROS levels (Bonizzi et al., 1999). COX2 contributes to oxidative stress (Dey et al., 2004) and regulates NADPHox-driven ROS production (Sonia et al., 2013). To assess the impact of S100B on the oxidative stress depending on COX-2 and NF B, we treated MHE astrocytes with S100B together with NFKB inhibitor PDTC or COX-2 inhibitor celeCOXib. NADPH oxidase activity was signi cantly increased in MHE astrocytes; increased NADPH oxidase activity was further activated after treatment of exogenous S100B. However, PTK787, celeCOXib, PDTC, or AVAB addition diminished the effect of S100B ( Figure 7D), indicating the involvement of COX-2/nfkb signaling in S100B-response oxidative stress.

MHE astrocytes have impaired ability to promote neuronal growth
Given the extensive dependency of neurons on astrocytic support (growth and viability) (Müller and Seifert, 1982), we co-cultured N2A cells with MHE astrocytes to address the effect of the MHE astrocytes on the neuronal viability using MTT assay. MHE astrocytes resulted in obvious N2A cell loss after 18h and 24h of co-culturing ( Figure 8A). Moreover, we evaluated the apoptotic effect of MHE astrocytes on N2A cells in vitro by ow cytometry. Flow cytometric analysis of apoptosis showed that the percentage of apoptotic neurons was signi cantly higher after co-culturing N2A cells with MHE astrocytes than with WT astrocytes ( Figure 8B and 8C). Moreover, similar results were observed in vivo. Brie y, we isolated neurons from P1-P2 pups with co-cultured of MHE astrocytes to assess primary cortical neuronal apoptosis using the TUNEL study. Primary cortical neurons showed the obvious increase in TUNEL positive cells (apoptosis) in co-cultures with MHE astrocytes as compared to co-cultures with MHE astrocytes ( Figure 8D).
We also characterized whether MHE astrocytes were impaired in promoting neurite growth length of the longest neurite per primary cortical neuron. Length of the longest neurite per neuron, evident by β-tubulin staining, was measured in at least 100 neurons per coverslip. We discovered that neurons had longer neurites in contact with WT astrocytes. MHE astrocytes promoted the growth of shorter neurites by cocultures as compared to WT astrocytes culture ( Figure 8E and 8F). These data indicated that activated astrocytes impair neuronal survival in MHE.

Discussion
This study determined an essential role of astrocyte dysfunction in MHE. Our data suggested a role of S100B in mediating astrocyte activity and MHE pathology via an autocrine manner. We found that S100B-induced pathological changes in astrocyte activity during MHE may accelerate in ammation. In addition, we discovered that the VEGF-dependent in ammation regulates the involvement of S100B in the pathogenesis of MHE. MHE astrocytes trigger neuronal impairment and halt the disease progression. We also found that MHE astrocytes maintained S100B overexpression in vivo and were functionally impaired in neuronal support, increasing VEGF expression. The autocrine mechanisms of S100B pro-in ammatory action were associated with the binding of the VEGF and VEGFR and the interaction of VEGFR and COX-2.
In conclusion, we suggested that targeting S100B may result in disease-modifying therapies and might be used as a pharmacological approach for MHE.
The protein S100B is abundant in the nervous system and is mainly expressed in astrocytes (Donato, 2001). S100B is secreted by astrocytes to the extracellular space (Van Eldik and Zimmer, 1987), thereby affecting astrocytes in an autocrine manner and neurons in a paracrine manner (Scotto et al., 1998).
Once released by astrocytes into the extracellular brain space, S100B exerts a dual effect on brain cells, acting as a neurotrophic factor at low (i.e., nanomolar) concentrations and as a toxic factor at high (i.e., micromolar) concentrations (Donato, 2001). Micromolar S100B levels turn astrocytes into a proin ammatory neurodegenerative phenotype. Moreover, when activated, astrocytes can become an important source of in ammatory cytokines (Ronaldson and Davis, 2012). In vitro studies on astrocytes have suggested that excessive production of S100B has a pro-in ammatory effect leading to the secretion of tumor necrosis factor-alpha (TNFα) and subsequent enhanced in ammation (Hu and Van Eldik, 1996). Increases of brain S100B amount partly mimic pathological brain conditions in Down's syndrome, Alzheimer's disease, and other neuronal diseases (Van Eldik and Wainwright, 2003), (Gri n et al., 1989). The constitutive overexpression of S100B might cause chronic neuronal damage (Reeves et al., 1994); thus, S100B has been considered a speci c marker of brain tissue damage (Araque et al., 2001). Moreover, transgenic mice overexpressing human S100B exhibits impaired hippocampal LTP and spatial learning (Gerlai et al., 1995). However, studies have suggested that on the accumulation in the brain tissue and astrocytes from MHE rats, S100B may be a pathogenic factor to cause MHE pathology. Low S100B and RAGE were found in the brain of normal rats, and primary adult normal rats cultured astrocytes, while cortices and primary adult MHE rats cultured astrocytes showed excessive S100B and high TNFα expression. Knockdown of S100B in primary adult MHE rats' cultured astrocytes induced low TNFα levels, suggesting that the autocrine function of extracellular S100B affected astrocytes to stimulate the in ammatory cytokines production and secretion.
Vascular endothelial growth factors (VEGFs), as a pro-in ammatory cytokine, are implicated in generating clinical de cits of many chronic in ammatory diseases (Baker et al., 1995;Sophie et al., 2005), (Byzova et al., 2000). Notably, astrocytes stimulate basal levels of VEGF secretion under normal physiological conditions (Chow et al., 2001). Reactive astrocytes result in VEGF-A overproduction in CNS in ammatory disease (Proescholdt et al., 2002b;Weis and Cheresh, 2005). Recent studies have found functional autocrine VEGF/VEGF receptor loops in many types of cells, such as hematopoietic stem cells (HSC) (Gerber et al., 2002), human leukemia cells (Santos and Dias, 2004), osteosarcoma cells (Ohba et al., 2014), and malignant cells (Lee et al., 2007;Lichtenberger et al., 2010). VEGF and its receptors (VEGF receptor-1, -2, and -3 (VEGFR1-3)), can be expressed in astrocytes (Kaur et al., 2010). The role of VEGF and its receptors are more related to in ammation (Saban et al., 2007). The endogenous secreted S100B from astrocytes appears to be su cient to induce VEGF secretion from astrocytes in an autocrinedependent manner, while the elevation of VEGF does not affect S100B secretion. Autocrine extracellular S100B interacts with RAGE, thus affecting the VEGF production and secretion from astrocytes. We showed that S100B essentially elicited the in ammatory reaction due to VEGF mediation in an autocrinedependent manner in MHE astrocytes.
The cell surface receptor for S100B on astrocytes mediating the autocrine effects has not been identi ed so far. Receptor for AGE (RAGE) has been reported to be a multiligand member of the immunoglobulin superfamily of receptors (Neeper et al., 1992) that engages with various ligands, including members of the S100 family S100B, and is implicated in the expression of the receptor itself (Hofmann et al., 1999b). (Huttunen et al., 1999). RAGE recruitment in astrocytes (Adami et al., 2004b), (Alejandro et al., 2014 has been identi ed as a responsible receptor mediating autocrine S100B effects (Ranuncolo et al., 2002). RAGE recruitment by ligand has been found to be directly involved in facilitating the resolution of cellular pro-in ammatory responses (Shin et al., 2010), resulting from the activation of the pro-in ammatory transcription factor nucleus factor-kappa-light-chain enhancer of activated B cells (NFκB) (Yan et al., 1996), leading to the expression of TNFα (G et al., 2002). Elevated RAGE expression was reported to be related to many in ammatory conditions or tissue injury (Chavakis et al., 2004), (Stewart et al., 1998). In vivo and in vitro results showed that RAGE expression in MHE astrocytes was decreased by knockdown of S100B. We also found that S100B was colocalized with RAGE in the cortex of MHE rats, while in astrocytic S100B has been shown to interact with RAGE to induce the pro-in ammatory mediators. It was observed that autocrine S100B discretely interacted with RAGE in the cytoplasm in MHE astrocytes. Thus, we believe that astrocytes expressing RAGE are a target of S100B. In MHE, astrocytes S100B's ability have been shown to activate astrocytes and induce the RAGE engagement through an autocrine mechanism. Thus, the S100B's ability to activate astrocytes via RAGE engagement and upregulate RAGE expression in astrocytes might signi cantly contribute to neuroin ammation.

RAGE receptor has been involved in intracellular signal transduction
Toll-like receptors (TLRs), a family of type 1 transmembrane proteins (Mousa et al., 2004), have a critical role in the in ammatory response. Convergence and ampli cation of RAGE and TLR signaling sustain and amplify in ammatory responses (Beijnum et al., 2008;Masakiyo et al., 2011). Previous studies have found that S100B protein is also involved in the TLR signaling pathway by interaction with RAGE (Giuseppe et al., 2007), (Tian et al., 2007). S100B-exposed astrocytes showed a pro-in ammatory phenotype with the expression of pro-in ammatory TLR2 (Selinfreund et al., 1990;Villarreal et al., 2014).
Previous studies have shown that high levels of TLR4 and RAGE are associated with neuroin ammation (Koprnová et al., 2001). Several studies have suggested that RAGE/TLR2 physical association is involved in S100B-induced cellular effects (Adami et al., 2004a). Emerging evidence suggests that S100B is su cient to mediate TLR2 upregulation but not TLR4 expression in MHE astrocytes. Thus, RAGE/TLR2 might be a complex receptor mediating autocrine S100B effects. We also found that S100B stimulates the overproduction of pro-in ammatory mediator VEGF via TLR2 and RAGE interaction in MHE astrocytes. Based on these ndings, we suggest that S100B acts in an autocrine manner by mediating the expression and secretion of potent inducers of in ammation through a RAGE and TLR2 interaction.
High S100B induced interaction between TLR2 and RAGE to promote the production of VEGF.
Cyclo-oxygenase 2 (COX-2) is an immediate-early response gene that exerts its key pro-in ammatory role by promoting an in ammatory process (Luisa, 2004), (Minghetti, 2004). Nonetheless, COX-2 expression in the brain has been associated with pro-in ammatory activities that are thought to be instrumental in the neurodegenerative processes of several acute and chronic diseases (Harris et al., 2002). S100B can upregulate the pro-in ammatory enzyme COX-2 expression via RAGE engagement (Bianchi et al., 2010)-8(Shanmugam et al., 2003. Speci cally, VEGF-dependent COX-2 induction is mediated through the VEGFR2 (Smith et al., 1998). In this study, we demonstrated that S100B mediates VEGF-A molecules' expression and stimulates VEGFR1 and VEGFR2 overexpression in an autocrine fashion in MHE astrocytes. S100B-induced upregulation of COX-2 has shown to be dependent on VEGF engagement. VEGF binding results in VEGFR2 no VEGFR1 interaction with COX-2. S100B can trigger COX-2 activity via VEGFa-mediated VEGFR2. These protein interactions are essential for mediating intracellular in ammatory signaling.
NFκB is a tightly regulated transcription factor that induces several genes, including TNFα (Hiscott et al., 1993). S100B can stimulate NF B transcriptional activity (Adami et al., 2004a), depending on RAGE (Arumugam et al., 2004). The inhibition of NF B results in signi cant suppression of S100B-induced TNFα expression (Li et al., 2010). RAGE is involved in cellular pro-in ammatory responses (Hofmann et al., 1999b) resulting from the activation of the NFκB (Yan et al., 1994), including the expression of TNFα (Howes et al., 2015). In the present study, we found that VEGF acts upstream of the nuclear translocation of NF B (p65) and enhances NF B (p65) transcriptional activity in the S100B-mediated in ammatory signaling pathway. Under stimulation by its ligands S100B and accumulation of VEGF, RAGE activates COX-2 signaling cascade, thereby evoking unrestrained NF B nuclear translocation NF B-transcriptional activity. COX-2 appears to act upstream of the nuclear translocation of NF B (p65) and enhances NF B (p65) transcriptional activity. Speci cally, S100B-dependent modulation of VEGF expression binds VEGFR2 to enhance the nuclear translocation and transcriptional activity of NF B (p65). S100B-induced RAGE via TLR2 ligation functions in an autocrine manner to elicit the downstream VEGF overproduction in the setting of interaction of VEGFR with COX-2, leading to the enhancement of COX-2 activity.
Existing studies have shown that cell-cell communication between primary neurons and astrocytes is crucial for the development, repair, and metabolism of neuronal systems (Kidambi et al., 2010). Available evidence indicates that astrocytes exert the capability to synthesize and release soluble paracrine factors to surrounding neurons to promote neuron growth and survival (Gomes et al., 2001). One of the most important functions of astrocytes is the regulation of neurite growth. Astrocytes promote neurite growth by providing various diffusible and nondiffusible proteins (Tomaselli et al., 1988). Our data further demonstrated astrocytic impairment in neuronal support at the functional level; MHE astrocytes induced less neurite growth when co-cultured with neonatal neurons compared to WT astrocytes. Primary neurons co-cultured randomly with MHE astrocytes showed an increased sensitivity to the toxicity as compared to pure neuronal cultures. Thus, MHE astrocytes could exert detrimental effects on neuronal growth. Higher TNFα expression by MHE astrocytes might account for the neuronal growth defects found in our coculture experiments. As the in ammatory response from astrocytes might cause marked changes in neuronal survival and growth, detailed analysis of the composition of factors secreted from WT and MHE astrocytes is needed to gain a better understanding of mechanisms through which astrocytes affect neuronal survival in MHE.

Conclusion
Above all, accumulating evidence demonstrated that astrocytic S100B triggers the secretion of VEGF in an autocrine manner, which then regulates the secretion of in ammatory cytokines in an autocrine manner. VEGF has a vital role in the perpetuation of S100B-induced potent in ammatory response found in MHE astrocytes. Our results suggest that the S100B-enhancing effects of VEGF occur due to interaction between RAGE and TLR2 in MHE astrocytes. The present study indicated that the autocrinedependent effect of VEGF might, in part, mediate the interaction of VEGFR2 with COX-2, thereby promoting the activation of NF B and leading to in ammatory response and oxidative stress in animals' models of MHE disease (Figure 8). In conclusion, S100B may be a potential pathological factor for MHE, achieving the in ammatory effect via VEGF mediation in astrocytes, which in turn irritates the impairment of neuronal survival.

Declarations
Ethics approval and consent to participate The study was approved by the Ethics Committee of the First A liated Hospital of Wenzhou Medical University.

Consent for publication
Not applicable

Data availability statement
The data that support the ndings of this study are available from the corresponding author upon reasonable request.

Competing interests
The authors have declared no con ict of interest.

Funding
This study was supported by Basic Scienti c Research Projects of Wenzhou city (Y20180076) Natural Science Foundation of Zhejiang province (LY21H030012) and Natural Science Foundation of China (81671042, 81300308).

Authors' Contributions
Saidan Ding supervised the entire project, designed the research and analysed the data and critically revised the manuscript. Shuya Feng and Baihui Chen conceived and designed the experiments, performed the research interpreted, and analysed the data, and wrote the paper. Xuebao Wang conceived and designed the experiments, interpreted and analysed the data, and supervised all the experimental procedure. Leping Liu and He Yu performed the research and analysed the data. All authors read and approved the nal manuscript.   Autocrine S100B induces VEGF production in MHE astrocytes. (A) Immunostaining of free-oating coronal sections from MHE rats using antibodies against VEGF-A (red), S100B (blue), GFAP (green). (B) Immunostaining of MHE astrocytes culture using antibodies against VEGF-A (red), S100B (blue), GFAP (green), S100B. (C, D) Analysis for VEGF-A mRNA of MHE astrocytes culture transfected with S100B siRNA by qPCR. (E, F) Immunoblot analysis and densitometry of lysates from MHE astrocytes culture transfected with S100B siRNA using anti-VEGF-A and β-actin antibodies. (G) ELISA assay for VEGF level of supernatants from MHE astrocytes culture treated with anti-S100B antibody (ASAB  Autocrine S100B by astrocytes facilitates TLR2 and RAGE interaction to elicit VEGF production. (A) Coimmunoprecipitation analysis for the association between RAGE and S100B of lysates of MHE astrocytes culture. (B) Immunostaining of free-oating coronal sections from MHE rats using antibodies against RAGE (red), TLR2 (blue), GFAP (green). (C) Immunostaining of free-oating coronal sections from MHE rats using antibodies against RAGE (red), TLR4 (blue), GFAP (green). (D) Co-immunoprecipitation analysis for the association between RAGE and TLR2/4 of lysates from MHE astrocytes culture treated with ASAB. (E, F) Immunoblot analysis and densitometry of lysates from MHE astrocytes culture treated with FPS-ZM1 or C29 in the preincubation of S100B (0.1 or 2μM) using anti-VEGF-A and β-actin antibodies. (G) ELISA assay for VEGF level of supernatants from MHE astrocytes culture treated with FPS-ZM1 or C29 in the preincubation of S100B (0.1 or 2μM). Data are shown as mean± SD. *P <0.05, **P <0.01. n.s., not signi cant. scale bar, 25 μm. MRGD, merged image.  VEGF mediation is required for autocrine S100B-induced in ammation by astrocytes. (A-C) Immunoblot analysis and densitometry of nuclear/cytosolic fraction lysate from MHE astrocytes culture with VEGF siRNA transfection in the presence of 2μM S100B using anti-NF B and lamin B1/β-actin antibodies. (D) Assay for NF B transcriptional activation of MHE astrocytes culture with VEGF siRNA transfection in the presence of 2μM S100B via a luciferase reporter construct. (E) Immunostaining of MHE astrocytes culture treated with AVAB, PTK787, or celecoxib in the presence of 2μM S100B using antibodies against NF B (green). (F, G) Analysis for TNFα mRNA of MHE astrocytes culture treated with AVAB in the presence of 2μM S100B by qPCR. Data are shown as mean± SD. *P <0.05, **P <0.01. n.s., not signi cant. scale bar, 25 μm. MRGD, merged image.

Figure 6
Autocrine S100B stimulates oxidative stress by astrocytes. (A) Assessment of ROS of lysate from MHE astrocytes culture transfected with S100B siRNA. (B, C) Immunoblot analysis and densitometry of lysates from MHE astrocytes culture transfected with S100B siRNA using anti-NOX1/NOX2 and β-actin antibodies. (D) Assessment of NADPH oxidase activity of lysate from MHE astrocytes culture treated with ASAB. Data are shown as mean± SD. n.s., not signi cant. *P <0.05, **P <0.01. Immunoblot analysis and densitometry of lysate from MHE astrocytes culture with VEGF siRNA transfection in the presence of 2μM S100B using anti-NOX1/NOX2 and β-actin antibodies. (D) Assessment of NADPH oxidase activity of lysate from MHE astrocytes culture treated with PTK787, Page 35/36 celecoxib, PDTC, or AVAB in pre-incubation of 2μM S100B. Data are shown as mean± SD. *P <0.05, **P <0.01. S100B-VEGF autocrine signaling pathway. Astrocytes-derived S100B initiates its autocrine effects by binding to RAGE; activated RAGE interacts with TLR2, which stimulates VEGF expression and secretion.
Autocrine VEGF leads to the interaction of VEGFR2 and COX-2, and activation of COX-2. Once activated, COX-2 recruits NF B to translocate into the nucleus, which triggers in ammatory cytokines production.