A1 astrocytes activated by 2-Chloroethanol via the ROS-induced p38 MAPK/NF-κB and AP-1 pathways modulate microglia polarization


 Background1,2-Dichloroethane (1,2-DCE) is a synthetic organic chemical that causes brain edema under subacute poisoning. Our previous studies indicated that the neuroinflammation could be induced due to activation of both astrocytes and microglia during the course of brain edema in 1,2-DCE intoxicated mice. However, the crosstalk between the two glial cells in 1,2-DCE-induced neuroinflammation is unclear. In the current studies, we hypothesized that astrocytes are the first responder to the effects of 1,2-DCE in the brain, as they adhere to the cerebral capillaries, and they are an essential component of the blood-brain barrier (BBB).MethodsWe used primary cultured rat astrocytes and microglia, as well as a highly aggressively proliferating immortalized (HAPI) microglia cell line to study the effects of astrocytes on microglia polarization following exposure to 2-CE.ResultsFindings from the present studies demonstrated that treatment of primary rat astrocytes with 2-chloroethanol (2-CE), the intermediate metabolite of 1,2-DCE in vivo, can stimulate the activation of A1 reactive astrocytes (A1s) through p38 mitogen-activated protein kinase (p38 MAPK)/ nuclear factor-κB (NF-κB) and activator protein-1 (AP-1) signaling pathways by the reactive oxygen species (ROS) produced during 2-CE metabolism. A1s activated by 2-CE can upregulate the expression of interleukin-1β (IL-1β), tumor necrosis factor-α (TNF-α), and inducible nitric oxide synthase (iNOS), and stimulate the M1 polarization of microglia through IL-1β and TNF-α released by 2-CE activated A1s. Microglia are less sensitive to 2-CE than astrocytes, since treatment of primary rat microglia with 30 mM 2-CE alone failed to activate them, though this dose of 2-CE can activate A1s and in turn stimulate M1 polarization of microglia through the factors released by A1s.ConclusionThe neuroinflammation induced by 1,2-DCE in the brain of mice is most probably triggered by the activation of astrocytes. The understanding of the multidimensional roles of reactive astrocytes may further the development of new treatment strategies in reducing neuroinflammation and brain edema following 1,2-DCE-induced toxic encephalopathy.

Astrocytes are the most abundant cells in the brain that are necessary for neuronal survival and function by maintenance of BBB integrity and extracellular homeostasis [4]. Accumulated evidences demonstrated that astrocytes can be activated under some pathological conditions, and then secrete a variety of proin ammatory factors that might induce neuroin ammatory responses, which degrades the tight junctions of BBB, thus leading to breakdown in BBB integrity and brain edema formation [5,6]. However, the functions of activated astrocytes under what contexts they may be helpful or harmful for the injury recovery remain unclear.
Recently, it was reported that neuroin ammation could induce two different types of reactive astrocytes, termed A1 and A2, respectively [7]. The A1 reactive astrocytes (A1s) are harmful since they highly upregulate many classical complement cascade genes shown to be destructive to neurons. In contrast, A2 reactive astrocytes (A2s) are protective due to up-regulating many neurotrophic factors [8,9]. Astrocytes, as they adhere to the cerebral capillaries, are an essential component of the BBB [4]. Therefore, astrocytes might be the rst exposing cells for the exogenous chemicals in the peripheral circulation from entering the brain. Because of high lipid solubility, 1,2-DCE can easily across the BBB. Thus, it is reasonable to assume that astrocytes might be both the rst targets of, as well as the responders to, 1,2-DCE.
Microglia are the primary immune effector cells and play an important role in the response of brain to both exogenous and endogenous insults [10]. As a myeloid-derived cell, activated microglia can polarize into M1 or M2 distinct macrophage subtypes [11]. The M1 phenotype is associated with prototypic in ammatory responses with increased release of proin ammatory cytokines. Adoption of the M2 phenotype promotes tissue repair through release of anti-in ammatory cytokines and neurotrophic factors [12]. Increasing lines of evidence indicate that the crosstalk between microglia and astrocyte play an important role in the neurodegenerative diseases and neurotoxicity of toxicants. A large number of experiments suggested that astrocytes are activated following the microglial activation. However, since astrocytes are the rst targets of 1,2-DCE in the brain, and a prominent contributor to several proin ammatory mediators, it is necessary to uncover the alteration in the microglial polarization following the activation of astrocytes. Up-to-date, the important interactions between astrocytes and microglia are virtually unknown in 1,2-DCE induced brain edema.
We previously reported that both astrocytes and microglia are activated during the course of brain edema formation in 1,2-DCE intoxicated mice [13]. In the current study, we postulated that 2-CE directly stimulate astrocytes through p38 MAPK signaling pathway via activation of NF-κB and AP-1, and that this could in turn enhance microglial polarization. To address this hypothesis, we utilized primary cultured rat astrocytes and microglia, as well as an immortalized microglia cell line to study the effects of astrocytes on microglia polarization following exposure to 2-CE. Our studies revealed that A1s activated by 2-CE are a potent inducer of M1 polarization in microglia, suggesting the crosstalk between astrocytes and microglia is critical for neuroin ammatory injury from 1,2-DCE poisoning.

Primary cultures of astrocytes and microglia
Astrocytes and microglia were puri ed from the cerebral cortices of 1-2-day-old postnatal Wistar rats that were obtained from the Animal Experiment Center of China Medical University, which were cultured as previously described [3,14,15]. Brie y, the cerebral cortices of rat brains were taken immediately after decapitation, and carefully cleaned of meninges and blood vessels. They were enzymatically and mechanically dissociated to generate a single cell suspension. The cells were then seeded and incubated at 37℃ in a 5% CO 2 -humidi ed atmosphere in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 20% fetal bovine serum (FBS) and 1% penicillin-streptomycin. The medium was completely replaced every 3 d until the cells reached a con uence, which usually occurred after 12-14 d.
The microglia were isolated from con uent mixed glial cultures via the mild trypsinization [14,15]. After 0.08% trypsinization, the microglia were rmly attached to the bottom of the culture plates, and the supernatant containing a detached intact astrocyte layer was collected and reseeded on culture plates pre-coated with Poly-L-Lysine. The astrocytes were puri ed via four passages to remove oligodendrocytes and neurons. By contrast, the microglia were incubated for 24 h with 200 ng/mL macrophage colonystimulating factor (MCSF, Novoprotein Scienti c, Shanghai, China) to stimulate their proliferation [16,17]. Astrocyte and microglia purities were determined using glial brillary acidic protein (GFAP) and ionized calcium binding adapter molecule 1 (Iba1), respectively.
The highly aggressively proliferating immortalized (HAPI) microglia cell line purchased from Otwo Biotech (Shenzhen, China) was also used in this study. This cell line was obtained initially from mixed glial cultures prepared from 3-day-old rat brain and possessed the characteristics of microglia.

Treatment of rat astrocytes with 2-CE and preparation of astrocyte-conditioned medium (ACM)
First, astrocytes were treated with 2-CE at 7.5, 15, or 30 mM for 8 h or at 30 mM from 0 to 48 h to analyze changes in target proteins. Next, the cells were treated with NAC (5, 50, or 500 μM), FC (0.5, 1, or 2 μM), GI (1, 5, or 10 μM), SB202190 (1, 10, or 30 μM), PDTC (25 μM), or SR11302 (10 μM) for 1 h prior to the addition of 30 mM 2-CE. These inhibitors were dissolved in dimethyl sulfoxide (DMSO) and then diluted with DMEM containing 5% FBS. The nal DMSO concentration was 0.1% at most and did not affect the behavior of the cells. Moreover, the cells were transfected with siRNAs of CYP2E1, TNF-α or iNOS using Lipofectamine 3000 Transfection Reagent (Invitrogen, Carlsbad, USA) or infected with IL-1β shRNA according to the manufacturer's instructions. Astrocytes were incubated with transfection media for 6 h, thereafter the complete media were replaced and incubated for 24 h before 2-CE exposure. The cells were harvested after 8 h treatment with 30 mM 2-CE for subsequent experiments.
For preparation of ACM, the treated astrocytes were washed with D-Hank's solution, and then incubated with fresh medium for 6 h. Subsequently, the media were collected and ltered using 0.22 μm pore lters before addition into the cultures of primary microglia or HAPI cells. In this experiment, ACM from untreated astrocytes was used as the control. After 12 h incubation with ACM, the primary microglia and HAPI cells were collected for subsequent experiments.

qPCR
Total RNA in the cells was extracted using Trizol Reagent (Takara), and their concentrations were con rmed using a Nanophotometer N60 Touch (Implen, München, Germany). For the reverse transcriptase reactions, 500 ng RNA was used as a template by using the PrimeScript RT reagent kit (Takara). The expression levels of IL-1β, TNF-α, iNOS, CYP2E1, CD86, Arg-1, CD200R, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were quanti ed using SYBR Premix Ex Taq II (Takara) with the QuantStudio 6 Flex Real-time PCR System (Life Technologies, Carlsbad, USA). All the speci c primers used for the qPCR are listed in Table 1. The running program was carried out for 30 s at 95℃ followed by 40 cycles of 5 s at 95℃ and 34 s at 60℃. Triplicate PCR reactions were prepared for each sample.
Changes in gene expression (fold change) were calculated using the 2 −ΔΔCt method with normalization to GAPDH.

Enzyme-linked immunosorbent assay (ELISA)
The levels of mature IL-1β and TNF-α in culture media were assessed via ELISA kits (Elabscience Biotechnology, Wuhan, China) according to the manufacturer's protocol. During the co-culture of microglia with ACM, the ACM were sampled from astrocytes and detected via ELISA kits prior to being applied onto microglia.

Flow cytometry
For analysis of microglia polarization, the microglia were incubated at room temperature for 30 min with the uorescently labeled antibodies CD86-APC (1:200, BioLegend, San Diego, USA) and CD200R-FITC (1:200, BioLegend). After incubation, the cells were washed, centrifuged and then resuspended in PBS to a nal volume of 500 μL. Flow cytometry was performed on a FACSCanto apparatus (BD Biosciences, Bedford, USA). The obtained data were analyzed using FlowJo software (version 10.1, BD Biosciences).

Levels of reactive oxygen species (ROS) in the cells
The uorescent probe of 2,7-dichlorodihydro uorescein diacetate (DCFH-DA, Beyotime) was used for quantifying ROS levels in the cells, and measured using a multi-functional microplate reader (BioTek Instruments, Winooski, USA) at excitation and emission wavelengths of 488 nm and 535 nm, respectively.
To provide statistical data, the mean measured uorescence intensity for the control was set to 100%.
The ow cytometer was also used to uorescently quantify the cells. The uorescent images of ROS were visualized and acquired using a RVL-100 uorescence microscope camera system.

Quanti cation of NO production
The cumulated NO production was assessed via the assay kit of Griess reaction (Elabscience Biotechnology), which detects the level nitrite (NO 2 -), a stable metabolite of NO in the cells.

Statistical analysis
Data are presented as the mean ± standard deviation (SD) and analyzed using SPSS 20.0 (IBM, Armonk, USA). Four to six independent experiments using primary cells of different origins were performed. Student's t-test (two-tailed) was used for comparing the differences between two groups, while one-way ANOVA followed by Student-Newman-Keuls tests were applied for determining the signi cant differences among multiple groups. P < 0.05 was de ned as statistically signi cant.

Alteration in astrocyte phenotype affected by treatment with 2-CE
It has been reported that C3 is a characteristic and highly upregulated marker in A1s, it is not expressed in A2s [7,8]. In contrast, S100A10 is the marker that is upregulated in A2s [18,19]. As shown in Figures 1a and 1b, the expression levels of C3 protein in 15 and 30 mM 2-CE treated astrocytes increased signi cantly compared to that in the untreated astrocytes, and those in the 30 mM 2-CE treated group were signi cantly higher than those in the other 2-CE treated groups. However, the expression levels of S100A10 protein increased signi cantly in the 7.5 mM 2-CE treated group, whereas decreased signi cantly in the 30 mM 2-CE treated group as compared with the control group. Moreover, the protein levels of S100A10 in 15 and 30 mM 2-CE treated cells were signi cantly lower than in the 7.5 mM 2-CE treated cells. These results indicate that treatment of rat astrocytes with 2-CE could dose-dependently induce A1s and suppress A2s.
On the other hand, the changes in protein levels of IL-1β, TNF-α, and iNOS along with the doses and time of 2-CE treatment in the astrocytes were determined. The temporal pro les of IL-1β, TNF-α, and iNOS protein levels in the cells are illustrated in Figure 1c. The protein levels increased signi cantly, reached a peak and remained elevation above the control levels were at 4 h, 8 h and 12 h for IL-1β, at 2 h, 8 h and 12 h for TNF-α, and at 4 h, 12 h and 24 h for iNOS, following the treatment with 2-CE, respectively. Accordingly, the cells were treated with 7.5, 15, and 30 mM 2-CE for 8 h. As shown in Figure 1d, the IL-1β and TNF-α protein levels in the cells treated with 30 mM 2-CE increased signi cantly compared to the other groups, and the iNOS protein levels in both 15 and 30 mM 2-CE treated group were signi cantly higher than those in either 7.5 mM 2-CE treated group or control group. Moreover, the IL-1β mRNA levels in all 2-CE treated groups increased signi cantly compared to the control group, and those in the 30 mM 2-CE treated cells were signi cantly higher than in the other groups (Fig. 1e). The TNF-α and iNOS mRNA levels in the cells treated with 30 mM 2-CE were signi cantly higher than those in either 7.5 mM 2-CE treated group or control group, and those of iNOS in 30 mM 2-CE treated group were also signi cantly higher than those in 15 mM 2-CE treated group (Figs. 1f and 1g). Furthermore, the levels of mature IL-1β and TNF-α in the media from all 2-CE treated cultures increased signi cantly compared with those from the control culture ( Figs. 1h and 1i). The concentrations of nitrite, an indicator of cumulated nitric oxide (NO) released into the media, in the media from 15 and 30 mM 2-CE treated groups increased signi cantly and dose dependently (Fig. 1j). All these results indicate that treatment of primary rat astrocytes with 30 mM 2-CE for 8 h could dramatically upregulate the production and release of proin ammatory mediators, such as IL-1β, TNF-α, and iNOS due to inducing A1s.

Roles of ROS in 2-CE induced A1s
Our paper published recently reported that treatment of rat astrocytes with 2-CE could upregulate CYP2E1 expression and cause oxidative damage [20]. Accordingly, we rst examined whether upregulated expression of CYP2E1 was involved in 2-CE-induced A1s. The speci c siRNA was transfected into the rat astrocytes to knockdown CYP2E1 expression before treatment with 2-CE. As shown in Figures 2a and 2b, the expression levels of CYP2E1 protein and mRNA in the cells of CYP2E1 knockdown decreased signi cantly as compared to the control, and the negative control (NC) that was transfected with scrambled siRNA. Moreover, transfection of CYP2E1 siRNA almost completely abolished 2-CE-induced CYP2E1 overexpression (Figs. 2c and 2d), and reduced ROS production (Figs. 2e and 2f) and C3 protein upregulation, whereas enhanced the protein expression of S100A10 in astrocytes treated with 2-CE (Fig.  2g). Furthermore, we also measured the changes of C3 protein expression in the presence of NAC, a scavenger of ROS in 2-CE treated astrocytes. Pretreatment with NAC could reduce ROS production (Figs. 2h and 2i), and C3 protein expression (Figs. 2j and 2k) in the astrocytes treated with 2-CE, suggesting that ROS production might play an important role in 2-CE induced A1s.

A1s induced by 2-CE through p38 MAPK/NF-κB and AP-1 signaling pathways
In the previous study, we have found that pretreatment of 2-CE treated rat astrocytes with NAC could attenuate the phosphorylation of p38 MAPK in these cells, suggesting that ROS production might be involved in the activation of p38 MAPK signaling pathway. Accordingly, we further explore the role of p38 MAPK signaling pathway in 2-CE induced A1s. Pretreatment of 2-CE treated astrocytes with SB202190, a specific inhibitor of the p38 MAPK could dose-dependently reverse the protein expression of C3 (Fig. 3a), and reduce the expression of IL-1β, TNF-α, and iNOS at the transcriptional level and their secretion (Figs. 3b-h). These ndings imply that the p38 MAPK signaling pathway might be involved in 2-CE induced A1s.
It has been reported that both NF-κB and AP-1 are activated via p38 MAPK signaling pathway. Thus, we further to explore the contribution of NF-κB or AP-1 signaling pathway to 2-CE-induced A1s by pretreatment of 2-CE treated astrocytes with PDTC, an inhibitor of NF-κB [21] or SR11302, an inhibitor of AP-1 [22]. The alterations in expression of C3 protein, and expression and secretion of IL-1β, TNF-α, and iNOS in either PDTC or SR11302 pretreated cells following the 2-CE treatment were similar with those achieved by the inhibitor of p38 MAPK (Figs. 3i-p). These ndings suggest that treatment of rat astrocytes with 2-CE might induce A1s and upregulate the expression of proin ammatory mediators through p38 MAPK/ NF-κB and AP-1 signaling pathways.

Effect of 2-CE induced A1s on microglia polarization
To further explore the impact of A1s on microglia, a co-culture system consisting of primary rat microglia and ACM that was from the cultures either treated or untreated with 2-CE was applied. The purity of the primary rat microglia was determined via Iba1 immunostaining (Fig. 4a). A dramatic change in the microglial morphology was observed due to the culture conditions. Microglia exposed to the ACM from 2-CE untreated astrocytes showed rami ed morphology with long branches, however, the morphology of microglia exposed to the ACM from 2-CE treated astrocytes, and exposed directly to 2-CE showed reduced number of branches, branch length, and branch complexity (Fig. 4b).
It is widely used that expression of CD86 is the marker of M1 polarization [23], and Arg-1 [24,25] or CD200R expression [26,27] is the marker of M2 polarization in the microglial phenotype. As shown in Figure 4c, more CD86-positive microglia and fewer Arg-1-positive microglia were observed in the culture exposed to 2-CE treated ACM compared to that exposed to 2-CE untreated ACM. The expression levels of CD86 mRNA and protein increased, conversely Arg-1 decreased in the microglia exposed to the ACM from 2-CE treated cells (Fig. 4d-f).
Similar to the results reported as before, the levels of mature IL-1β and TNF-α, and concentrations of NO in the ACM from the 2-CE treated cells were increased signi cantly (Figs. 4g-i). Surprisingly, although the microglia exposed directly to 30 mM 2-CE showed de-rami ed morphology, the expression levels of CD86 protein decreased, but Arg-1 increased in these cells (Figs. 4b-f), suggesting the stimulated effects of ACM prepared from 2-CE-treated astrocytes on M1 polarization of microglia is not due to the 2-CE. On the other hand, these results indicate that the primary rat microglia are not sensitive to the 2-CE effects as the primary rat astrocytes.

Factors released from the 2-CE induced A1s modulate M1 polarization of microglia
To elucidate the factors released from 2-CE-induced A1s modulate M1 polarization of microglia, the RNA interference to knock down the mRNA of IL-1β, TNF-α, or iNOS in A1s was performed before 2-CE treatment. As shown in Figures 5a-l, the speci c shRNA and siRNA effectively reduced the protein and mRNA levels of IL-1β, TNF-α and iNOS in the cells with or without 2-CE treatment. Accordingly, the levels of mature IL-1β, TNF-α, and concentrations of NO in the ACM prepared from shRNA infected or siRNA transfected astrocytes decreased signi cantly (Figs 5m-o).
It is well-known that the in vitro studies involving the primary microglia have been hindered by the inability to obtain high yields of puri ed cells due to their lack of proliferation in culture. Therefore, many in vitro studies were performed by using the immortalized microglia cell lines. In this study, an immortalized microglia cell line, HAPI, prepared initially from rat brain was used to further explore the factors in the ACM modulate the microglial polarization. As mentioned above, the HAPI microglial cells were co-cultured with the ACM that was prepared from the astrocytes transfected with or without the IL-1β shRNA, TNF-α siRNA, or iNOS siRNA prior to 2-CE treatment, which were termed as NC, 2-CE, siIL-1β, siTNF-α, and siiNOS group, respectively. The mRNA levels of CD86 decreased signi cantly, while the CD200R increased in the microglial cells of siIL-1β or siTNF-α group compared to the 2-CE group (Figs.  6a-d). The percentages of CD86-positive and CD200R-positive microglial cells examined using the ow cytometry in the NC, 2-CE, siIL-1β, siTNF-α and siNOS group were 19.0% and 2.85%, 31.7% and 2.24%, 22.7% and 2.85%, 20.8% and 2.44%, and 31.9% and 2.55%, respectively (Fig. 6g). The differences in percentage of CD86-positive cells between NC and 2-CE groups, and between either siIL-1β or siTNF-α and 2-CE groups were signi cantly (Fig. 6h). Our ndings from this part of study indicate that 2-CE induced A1s could stimulate M1 polarization of microglia through releasing IL-1β and TNF-α into the ACM. Although A1s could also suppress the M2 polarization of microglia, IL-1β and TNF-α in the ACM have no contribution to this effect (Fig. 6i). Furthermore, our results in this part of study suggest that knockdown of iNOS in A1s could not affect the microglial polarization (Figs. 6e-i).

Effects of the treatment with uorocitrate and GIBH-130 (piperazinyl pyrimidine) on A1s, and their roles in modulation of M1 polarization of microglia
Both uorocitrate (FC) and GIBH-130 (GI) are the inhibitors of neuroin ammation reported recently [28,29]. Since FC is preferentially taken up by the astrocytes and can inhibit the tricarboxylic acid cycle via targeting aconitase, it is used to explore the roles of activated astrocytes under various pathological conditions. GI is a novel anti-neuroin ammatory agent that was identi ed through microglia-based phenotypic screenings, however, its inhibitory effects on astrocytic activation is unclear. In the part of study, we rst study the effects of either FC or GI on 2-CE induced A1s, and then further explore the alteration in M1 polarization of microglia through co-culturing microglia with the ACM prepared from the 2-CE treated astrocytes.
Our results revealed that pretreatment of rat astrocytes with either FC or GI could inhibit 2-CE-induced A1s by suppressing the expression levels of C3 protein, the speci c marker of A1s (Figs. 7a-c). Moreover, pretreatment of 2-CE treated rat astrocytes with FC or GI could reduce the protein expression of IL-1β, TNF-α, and iNOS in the cells, and their excreted levels in the culture media (Figs. 7d-k). Furthermore, pretreatment of 2-CE-exposed rat astrocytes with FC or GI could also reduce the activation of p38 MAPK/ NF-κB and AP-1 signaling pathways by inhibiting the phosphorylation of p38 MAPK, p65, IκB, and c-Jun proteins (Figs. 7l and 7m) On the other hand, the primary rat microglia were examined following co-culture with the ACM prepared from 2-CE exposed astrocytes pretreated with either FC or GI. The results showed that the M1 polarization of microglia was suppressed in the FC or GI group compared to the 2-CE group, since both the proportion of CD86-positive microglia and expression levels of CD86 in the two groups decreased signi cantly. In contrast, the suppressed M2 polarization of microglia in the 2-CE group were relieved in the GI group as the percentages of Arg-1-positive microglia and expression levels of Arg-1 increased signi cantly compared to the 2-CE group. However, no signi cant alteration in the percentages of Arg-1-positive microglia and expression levels of Arg-1 were observed between the 2-CE and FC groups (Fig. 8a-e). Meanwhile, the levels of IL-1β and TNF-α, and the concentrations of NO in the ACM from FC or GI pretreated cells decreased signi cantly compared to the 2-CE treated cells (Figs. 8f-h). These results demonstrated that the inhibition of A1s achieved by pretreatment with FC or GI could also suppress the M1 polarization of microglia through the mediators released into the ACM.
The results in this part of study provided the evidences concerning the inhibitory effects of FC and GI on 2-CE induced A1s, and indicated that IL-1β, TNF-α, and NO in the ACM might modulate the effects of A1s on M1 polarization of microglia, which are agreement with the ndings as mentioned above.

Discussion
Astrocytes and microglia are the important glial cells in the brain, and play the crucial roles in the biological responses to the external stimuli. Especially in the neuroin ammation, the crosstalk between the astrocytes and microglia can determine the magnitude and duration of the in ammatory responses. Increasing evidence indicates that the sustained in ammatory responses may result in the increased production of in ammatory mediators including NO, TNF-α, and IL-1β that can directly induce the pathological changes in the brain or amplify the local in ammatory response [30][31][32]. Although activated microglia play the dominated roles in neuroin ammation, astrocytes can also produce and secrete the proin ammatory mediators such as NO, TNF-α, and IL-1β, while they were activated by the external stimuli. Therefore, it is necessary to understand the regulatory effects of activated astrocytes on activation of microglia, which are critical with respect to a wide spectrum of neurological diseases.
Based on the knowledge that astrocytes might be the rst target cells and responders to 1,2-DCE in the brain, we have done a lot of in vivo and in vitro studies on the responses of astrocytes to either 1,2-DCE or 2-CE [3,13]. Although our cumulated evidences have disclosed that the neuroin ammation is involved in 1,2-DCE induced brain edema, the molecular mechanisms underlying 1,2-DCE induced activation of astrocytes and microglia in the brain have not been well demonstrated.
In the current study, we rst investigated the effects of 2-CE on the phenotype of reactive astrocytes. To our knowledge, this is the rst study to reveal the role of 2-CE in A1/A2 reactive astrocytic alteration. The present results indicated that treatment of the primary rat astrocytes with the 7.5 to 30 mM 2-CE can dose-dependently stimulate A1s, and 7.5 mM 2-CE may dominantly promote the A2s. Consequently, the expression levels of IL-1β, TNF-α, and iNOS in the 30 mM 2-CE treated astrocytes increased dramatically. However, the mRNA levels of IL-1β in the cells, as well as the levels of IL-1β and TNF-α in the media in 7.5 mM 2-CE treated group also increased markedly.
It is well-known that astrocytes undergo a dramatic transformation called "reactive astrocytes" and upregulate many genes after brain injury and disease. Liddelow et al. (2017) rstly reported that two different types of reactive astrocytes, termed "A1s" and "A2s", can be stimulated by lipopolysaccharide (LPS)-induced neuroin ammation. Many proin ammatory mediators are highly up-regulated in the A1s, which may stimulate the in ammatory response, so A1s are destructive to the recovery and repair in the brain. In contrast, many neurotrophic factors and anti-in ammatory mediators are up-regulated in the A2s, thus they are neuroprotective, and promote brain healing after injury. Accordingly, the present data suggested that treatment with 30 mM 2-CE could induce in ammation and cause damage in the brain.
Next, we investigated the ways by which 2-CE stimulates the A1s. In our previous studies, we have found that ROS produced through CYP2E1 mediated metabolism of 1,2-DCE or 2-CE can activate the p38 MAPK/ NF-κB and AP-1 signaling pathway, and upregulate the expression of MMP-9 that may disrupt the BBB integrity, and then enhance BBB permeability and brain edema formation. Therefore, we investigated whether this signaling pathway play a role in 2-CE induced A1s.
CYP2E1 is the main enzyme that is implicated in 1,2-DCE metabolism in vivo [33]. Although its expression in the brain is much lower than in the liver, the enzyme is highly concentrated and inducible in the cortical astrocytes [34][35][36][37]. It has been reported that CYP2E1 can convert 1,2-DCE into 2-CE, chloroacetaldehyde and chloroacetic acid [38][39][40][41]. Because of the high oxidase activity, CYP2E1-mediated metabolism of 1,2-DCE and its intermediates can lead to the generation of ROS, as a consequence induce oxidative damage in the brain [20,33] The present results revealed that CYP2E1 knockdown reversed 2-CE-induced augmentation of ROS production and A1s induction, suggesting that ROS production is involved in A1 activation. Using the ROS scavenger NAC, we further determined that ROS induce A1s. Overall, our data demonstrates for the rst time that 2-CE induces A1s via ROS that was overproduced by CYP2E1-mediated metabolism of 2-CE.
Although we have reported that p38 MAPK/ NF-κB and AP-1 signaling pathways are involved in the upregulation of MMP-9 in 2-CE treated astrocytes, whether these signaling pathways also contribute to 2-CE-induced A1s and upregulation of TNF-α, IL-1β, and iNOS in astrocytes are unclear. This report is the rst to provide strong evidence of the involvement of the p38 MAPK/ NF-κB and AP-1 pathways in 2-CEactivated A1s. Our data indicate that as A1s are activated, p38, p65, IκBα and c-Jun phosphorylation increases. To further explore the roles of these pathways in A1s activation, astrocytes were pretreated with the p38 inhibitor SB202190, the NF-κB inhibitor PDTC, and the AP-1 inhibitor SR11302. The results clearly demonstrate that p38 MAPK, NF-κB, and AP-1 regulate A1s activation and also provide evidence of these pathways in upregulating TNF-α, IL-1β, and iNOS.
Analysis on protein and gene expression of CD86 and Arg-1, the typical marker of M1 and M2 polarization revealed marked differences in microglia treated directly with 2-CE, as well as co-cultured with the ACM prepared from 2-CE treated and untreated astrocytes. Treatment with ACM prepared from 2-CE-stimulated astrocytes signi cantly increased and decreased, respectively, microglial expression of CD86 and Arg-1. By contrast, the treatment of microglia with 2-CE slightly increased Arg-1 expression but decreased CD86 expression. Thus, our ndings suggested that treatment of primary rat microglia with 30 mM 2-CE alone was insu cient to induce their M1 polarization, possibly because the production of ROS is less than in the astrocytes due to few expressed CYP2E1 in the microglia [34]. Furthermore, our ndings demonstrated that the most possibilities of M1 polarization in microglia are activated through the mediator released by 2-CE stimulated astrocytes in the brain.
Since the aforementioned results suggested that the ACM prepared from 2-CE treated astrocytes could activate microglia, it is essential to determine the factors that play these roles in the ACM. For this aim, the gene silence target at IL-1β, TNF-α, and iNOS was performed, and the ndings indicated that the IL-1β and TNF-α released by A1s into the ACM might contribute to microglial activation. However, the iNOS gene silence in 2-CE treated astrocytes failed to affect the microglial M1 polarization although the concentrations of NO determined in the ACM increased due to upregulated expression of iNOS. The reason is most probably due to the short half-life of NO in the culture media, where it may disappear rapidly due to oxidation and be formed nitrite [42][43][44]. Thus, the NO concentrations evaluated by nitrite just re ect the cumulated NO production, which is not the indicator of NO with the bioactivity. Therefore, study on co-culture of microglia with ACM can't be used to evaluate the effects of upregulated iNOS in astrocytes on microglia. For this purpose, a co-culture model by Transwell that mimics the crosstalk between cells should be used in our further study.
IL-1β, TNF-α, and NO are the dominated in ammatory mediators that may directly induce neuronal apoptosis or amplify the local in ammatory response. Although most studies reported that they were released by microglia activation and can activate A1s, these factors may also be produced and released by A1s. In current study, our ndings for the rst time demonstrated that these factors might rst produced and released by 2-CE treated astrocytes, and then stimulated the microglial activation. . It was also demonstrated that inhibiting reactive astrocytes with FC immediately before reperfusion may protect against ischemic stroke-induced memory impairment [28].
GI, the pyridazine-3-carboxamide derivative, is a novel anti-neuroin ammatory agent that has been approved by China Food and Drug Administration for clinical trials against AD. The antineuroin ammatory e cacy was identi ed via the microglia-based phenotypic screenings in AD animal models [29]. It has been reported that pretreatment with GI can selectively suppress the production of NO, TNF-α and IL-1β through modulating MAPK signal transduction pathway in LPS-activated microglia.
However, the effects of GI on astrocytic activation is unclear, and further insight into its molecular mechanisms of action is necessary.
The present results revealed that pretreatment of the primary rat astrocytes with either FC or GI could suppress A1s and activation of p38 MAPK/NF-κB and AP-1 signaling pathways in 2-CE treated astrocytes. Moreover, FC treatment can alleviate the activation of both A1s and A2s, in contrast, GI attenuated A1s, but not A2s in 2-CE treated astrocytes. Furthermore, co-culture with the ACM prepared from either FC or GI pretreated astrocytes attenuated the microglial M1 polarization, however, GI pretreated ACM might promote microglia towards the M2 state.

Conclusions
Here the evidence from our in vitro studies for the rst time demonstrate that treatment of rat astrocytes with 2-CE stimulate activation of A1s through ROS mediated p38 MAPK/NF-κB and AP-1 signaling pathways, which enhance the production and secretion of IL-1β, TNF-α, and NO; next, treatment of rat astrocytes with 2-CE can stimulate M1 polarization in microglia through IL-1β and TNF-α released by 2-CE treated astrocytes; nally, microglial activation is less sensitive to 2-CE than astrocytes, suggesting that neuroin ammation induced by microglial activation is triggered by astrocytic activation in the brain of      Immunostaining of primary rat microglia with the M1 marker cluster of differentiation (CD) 86 and M2 marker arginase-1 (Arg-1). (d) Western blot analysis for primary rat microglial CD86 and Arg-1. (e, f) qPCR for CD86 and Arg-1 mRNA. (g-i) ELISA was performed to detect IL-1β and TNF-α levels in ACM. The concentrations of NO in ACM were measured using the Griess reaction. (* p < 0.05 vs. control ACM, # p < 0.05 vs. 30 mM 2-CE ACM; mean ± SD; one-way ANOVA followed by SNK tests or Student's t-test).

Figure 5
IL-1β, TNF-α, and iNOS knockdown in astrocytes are achieved via RNA interference. Successful knockdown of IL-1β, TNF-α, and iNOS protein (a, b, c) and mRNA (d, e, f) in astrocytes as assessed via western blotting and qPCR. IL-1β, TNF-α, and iNOS protein (g, h, i) and mRNA (j, k, l) expression levels in astrocytes following 2-CE treatment as detected by western blotting and qPCR. (m, n, o) IL-1β, TNF-α, and NO concentrations in ACM were detected via ELISA or the Griess reaction. (* p < 0.05 vs. control or vs. NC, # p < 0.05 vs. 30 mM 2-CE; mean ± SD; one-way ANOVA followed by SNK tests).  and TNF-α levels in culture media were quanti ed using ELISA. NO concentrations in the culture media were measured using the Griess reaction. (h) Effect of GI on IL-1β, TNF-α, and iNOS protein levels. (i-k) IL-1β and TNF-α levels in culture media were detected using ELISA, while the Griess reaction was applied to measure NO levels. (l, m) p38, p65, IκBα, and c-Jun phosphorylation were detected using western blotting.