FTO Regulates Microglia-Induced Inammation by Stabilizing ADAM17 Expression After Experimental Traumatic Brain Injury

The neuroinammatory response mediated by microglial polarization plays an important role in the secondary nerve injury of traumatic brain injury (TBI). The post-transcriptional modication of n6-methyladenosine (m 6 A) is ubiquitous in the immune response of the central nervous system. The fat mass and obesity (FTO)-related protein can regulate the splicing process of pre-mRNA. However, after experimental traumatic brain injury (TBI), the role of FTO in microglial polarization and the subsequent neuroinammatory response is still unclear. TBI mice model was established by the Feeney weight-drop method. Neurological severity score, brain water content measurement and Nissl staining were used to detect the role of FTO in microglial polarization and the molecular mechanism of targeted RNA epigenetic modication. In vitro and in vivo experiments were conducted to evaluate microglial polarization and the neuroinammatory response by down-regulation of FTO expression. The level of m 6 A modication in M1 activated microglia was detected by qRT-PCR, m 6 A-MeRIP and m 6 A high-throughput sequencing. Fluorescent in situ hybridization combined with immunouorescence imaging were used to detect the epigenetic regulation of ADAM17 mediated by an FTO-m 6 A-dependent mechanism. Microglia-mediated neuroinammatory responses play important roles in secondary neurological injury after TBI. Regulation of the phenotype of microglia and inammatory factors is a potential strategy for the treatment of TBI. The fat mass and obesity (FTO)-related protein can regulate the splicing process of pre-mRNA. However, after TBI, the role of FTO in microglial polarization and the subsequent neuroinammatory response is still unclear. Here, we report that the expression of FTO was signicantly down-regulated in BV2 cells treated with lipopolysaccharide and mice with TBI. FTO can affect the transcription modication of ADAM17 and downstream TNF-α/NF-kB pathway related factors in microglia, which promote M1 polarization phenotype of microglia and the development of inammation after TBI. Taken together, our results reveal that FTO, as an important m 6 A demethylation enzyme, regulates microglia polarization and neuroinammation by stabilizing ADAM17 expression after TBI.

Introduction Traumatic brain injury (TBI) is a common serious disease, with high fatality and disability rates. TBI is associated with a primary injury that triggers a series of harmful in ammatory processes that further aggravate initial tissue damage and affect nerve cell repair [1][2][3]. Microglia are the innate immune cells of the central nervous system, which play an important role in neuroin ammation and neurological impairment after TBI [4]. The activation of microglia is closely related to their different polarization phenotypes. Speci cally, microglia activated by traumatic stress exhibit two polarized phenotypes, including the classic activated M1 phenotype and the alternative activated M2 phenotype, which are also called the pro-in ammatory phenotype and anti-in ammatory phenotype, respectively [5][6][7]. Our previous studies [8,9] con rmed that the polarization phenotype of microglia is closely related to the outcome of neuroin ammation. Blocking polarization of the M1 phenotype and adjusting the M1/M2 polarization can improve the prognosis of neuroin ammation after TBI and restore nerve function.
The eukaryotic RNA modi cation n6-methyladenosine (m 6 A) has recently been identi ed as a key posttranscriptional regulator of gene expression [17][18][19]. M 6 A mRNA modi cation is the most abundant type of gene modi cation, accounting for more than 60% of all post-transcriptional RNA modi cations [17].
Previous studies have reported that the modi cation of m 6 A mRNA mainly involves the modi cation of adenine sites, and which was precisely determined by writer, eraser, and reader proteins in the RNA omitted. Mice in the TBI+NADP group were injected with NADP (300 mg/kg/day; Sigma-Aldrich, St. Louis, MO, USA) 0.5 h after the surgery [33], those in the TBI+Vehicle group were given an equal volume of the vehicle dimethyl sulfoxide as a negative control.
Neurological impairment score Mice were subjected to exercise (muscular phenotype and abnormal action), sensation (visual, tactile, and balance), and re ex examinations and assigned a modi ed neurological severity score (mNSS). A score was recorded when the mice failed to complete the task or showed no corresponding re exes. The mNSS score ranged from 0 to 18 points, where a total score of 18 points indicated severe neurological de cits and a score of 0 indicated normal performance. Researchers, blinded to the experimental groups, measured the neurological function of mice at different time points.
Measurement of brain water content and blood brain barrier (BBB) permeability The wet weight-dry weight method was used to calculate the brain water content 8 . The animals were sacri ced after neurological assessment, and the cerebral cortex was excised at the edge of the bone window. Filter paper was used to remove excess blood and cerebrospinal uid. The wet weight was measured and the brains were dried in an oven for 24 h at 100 °C until a constant weight was achieved, at which point the dry weight was measured. The percentage of brain water content was calculated as (wet weight − dry weight)/wet weight × 100%.
BBB permeability was investigated by measuring the extravasation of Evans blue dye (2% in saline; 4 mL/kg; Sigma-Aldrich), which was injected intravenously 2 h prior to sacri ce on the third day after injury. Following sacri ce, the mice were transcardially perfused with PBS followed by PBS containing 4% paraformaldehyde. Each tissue sample was immediately weighed, homogenized in 1 mL of 50% trichloroacetic acid, and centrifuged. Then, the absorption of the supernatant was measured with a spectrophotometer (UV-1800 ENG 240V; Shimadzu Corpomiceion, Kyoto, Japan) at a wavelength of 620 nm. The quantity of Evans blue dye was calculated using a standard curve and expressed as µg/g of brain tissue.

Nissl staining
The formaldehyde-xed specimens were embedded in para n, cut into 4 μm thick sections, depara nized with xylene, and rehydrated in a graded series of alcohol. After being treated with Nissl staining solution for 5 min, the damaged neurons were atrophied or contained vacuoles, while the cells of normal neurons were larger and fuller with larger nuclei. Five areas were randomly selected for microscopic examination by a researcher who was blinded to the to the experimental groups.
Cell culture and treatment

RNA isolation and RT-PCR
Total RNA from the tissues or the cultured samples was puri ed using TRIzol (Invitrogen, ThermoFisher Scienti c) and reverse transcribed using the ABI reverse transcriptase (ABI, ThermoFisher Scienti c), oligo (dT) primers, or speci c RT primers. Template (1 µL) was ampli ed by real-time PCR using the primers listed in Suppl. Table 4, Supporting Information (Integrated DNA Technologies). Each sample was run in triplicate in a 10 µl reaction with 100 nm forward and reverse primers, 2µl of SYBR Green mix (ABI, ThermoFisher Scienti c), and 10 ng cDNA. The PCR reactions were carried out using a STEP-ONE 96 realtime PCR system. GAPDH was used as an internal control for normalization. Ratios of mRNA levels from the treated groups or mRNA at levels different time points compared with the mRNA level of the normal control group were calculated using the ΔCt method (2 −ΔΔCt ). All data were normalized to GAPDH.

RNA m 6 A quanti cation
Total RNA was isolated with TRIzol (Invitrogen, ThermoFisher Scienti c) according to the manufacturer's instructions and RNA quality was measured by using a NanoDrop3000. The m 6 A RNA methylation quanti cation kit (Abcam, UK) was used to measure the m 6 A content of the RNA. Brie y, 200 ng of RNA was detected in each well. The capture antibody solution and detection antibody solution were then added to assay wells separately in a suitable diluted concentration following the manufacturer's instructions. The m 6 A levels were quanti ed calorimetrically by reading the OD 450 absorbance of each well and calculations were performed based on the standard curve.
RNA m 6 A sequence and m 6 A-RNA immunoprecipitation assay The chemically fragmented RNA (100 nucleotides) was incubated with the m 6 A antibody and immunoprecipitation was performed according to the standard protocol of the Magna methylated RNA immune-precipitation (MeRIP) m 6 A Kit (Merck Millipore, USA). Enrichment of m 6 A containing mRNA was analyzed by qRT-PCR using the primers listed in Suppl.

Immunohistochemical analysis
The formaldehyde-xed specimens were embedded in para n, cut into 4 μm thick sections, depara nized with xylene, and rehydrated in a graded series of alcohol. TAntigen retrieval was carried out by microwaving the sections in citric acid buffer. Sections were then incubated with an antibody against FTO (1:400, abcam, UK), washed, and then incubated with secondary antibody. The negative control was prepared without adding the primary antibody. Five randomly selected visual elds were analyzed as follows [3,8,37]: 0, no positive cells; 1, very few positive cells; 2, moderate number of positive cells; 3, many positive cells; and 4, the highest number of positive cells.

Immuno uorescence analysis
The formaldehyde-xed specimens were embedded in para n, cut into 4 μm-thick sections, depara nized with xylene, rehydrated in a graded series of alcohol, and then the antigen was retrieved as describe above. Sections were incubated overnight at 4 °C with antibodies against ionized calciumbinding adapter molecule-1 (Iba-1; 1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA), FTO (1:200; Santa Cruz Biotechnology), CD86, and CD206 (1:100; Boster Biotech, Wuhan, China). After washing, the sections were incubated with secondary antibodies for 1 h at room temperature, after which the cell nuclei were stained with 4',6-diamidino-2-phenylindole. Immuno-positive cells in ve randomly selected elds were counted under a microscope (Leica, Wetzlar, Germany) at 400× magni cation by investigators who were blinded to the experimental groups.

Fluorescent in situ hybridization (FISH) combined with immuno uorescence imaging
The Cy3-labeled probes of ADAM17 mRNA were designed and synthesized by Sangon Inc. (Shanghai), while mouse monoclonal antibody to FTO and goat anti-mouse IgG H&L were obtained from abcam (Alexa Fluor® 488, abcam, UK). FISH combined with immuno uorescence experiments were performed according to the manufacturer's instructions. Bv2 cells were seeded in a 24-well plate on chamber cover slips and treated as described above. After the cells had reached 60%-70% con uency, they were xed with 4% paraformaldehyde for 30 min at room temperature, permeabilized with pre-cooled 0.5% Triton-X-100 for 5 min at 4 °C, washed three times with PBS, and prehybridized for 30 min at 37 °C with 200 μL pre-hybridization buffer. mRNA ADAM17 FISH Probe Mix Storage solution (2.5 μL, 20 μM; mRNA FISH Probe Mix) and 100 μL hybridization buffer were added and the cells were incubated overnight at 37°C in a humidi ed chamber in the dark. The cells were then washed three times for 5 min each with 4× SSC and 2× SSC for 5 min and 1 × SSC for 5 min at 42 °C, followed by a 5 min wash with PBS at room temperature in the dark. Finally, glass coverslips were sealed with an anti-quenching adhesive containing DAPI and images were acquired on an IX51 inverted microscope (Olympus, Japan).

Western Blotting Analysis
Samples, including brain tissues and BV2 cells, were prepared by using the nuclear and cytoplasmic proteins puri cation assay kit (KeyGEN Biotech, China), with modi ed RIPA lysis buffer (50mM Tris-HCl pH 7.4, 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS). The protein concentrations of the nuclear and cytosolic lysates, respectively, were determined with a BCA kit (

Statistical analysis
Data were analyzed using SPSS v.18.0 software (SPSS Inc., Chicago, IL, USA). All experiments were performed in triplicate unless otherwise noted, and the results are expressed as the mean ± SD. The unpaired Student's t test was used for comparison between groups. Multiple-group comparisons were assessed by one-way ANOVA and post hoc multiple comparisons were performed using Student-Newman-Keuls tests. P< 0.05 was considered statistically signi cant.

The level of m 6 A modi cation is increased in M1 activated microglia
In order to explore whether m 6 A methylation was related to microglial polarization, we rst investigated BV2 cells without any treatment (M0 phenotype group) and lipopolysaccharide (LPS)-stimulated BV2 cells (M1 phenotype group) to determine the abundance of m 6 A and m 6 A writers/erasers in the two groups. m 6 A RNA methylation quantitative experiments revealed that, compared with the M0 phenotype, the level of m 6 A modi cation in the M1 phenotype was signi cantly higher ( Fig. 1a). RT-PCR results showed that the expressions of hnRNP, YTHDF1, and YTHDF2 increased in the M1 phenotype group compared with the M0 phenotype group, although the difference was not statistically signi cant. By contrast, FTO, WTAP, METTL3, and METTL14 decreased in the M1 phenotype group, with the FTO level showing a signi cant reduction (P < 0.01), and which was consistent with the dynamic change of the m 6 A modi cation level (Fig. 1b). The expression of FTO protein was lower in the M1 phenotype group, but with no signi cant change in the expression of ALKBH5 (Fig. 1c). Furthermore, we used MeRIP-Seq to analyze the transcriptome-wide mRNA modi cations in the M0 and M1 phenotype groups. Speci cally, the clean reads of the two groups were compared with the mouse genome to obtain positional information of the reference genome (details in Suppl. Table 1). In the M0 and M1-phenotype groups, the transcripts of 4828 and 5767 genes were enriched with high-con dence m 6 A peaks, respectively. Motif enrichment analysis revealed that the m 6 A peaks identi ed above shared a common sequence element [U]GGAC[U]A (Fig. 1d) and that the CDS and 5'UTR, together with the 3'UTR, harbored the largest fraction of peaks (Fig. 1e, f). Furthermore, the results showed that the m 6 A modi cation level in the M1 phenotype group was signi cantly higher than that in the M0 phenotype group. Additionally, multiple m 6 A peaks in the M1 phenotype group were enriched in transcripts of the phosphoinositide 3-kinase (PI3K)/Akt and NF-κB signaling pathways compared with the M0 phenotype group (Fig. 1g, h). These data indicated the potential role of m 6 A in activating M1 microglia during the initial stages of in ammation.

FTO regulates microglial M1 polarization
Since FTO was down-regulated in the M1 phenotype group, we further veri ed the role of FTO in the polarization and activation of microglia. FTO siRNAs and plasmids were utilized to knock down and overexpress FTO in BV2 cells, respectively to explore the function of FTO in regulating the microglial polarization. The level of FTO mRNA was signi cantly reduced after siRNA treatment (siFTO group), and upregulated after pcDNA3.1-Flag-FTO transfection (oeFTO group) (Fig. 2a, b). As shown in Fig. 2c, under LPS stimulation, the M1 phenotype (CD11b+/CD86+) in the siFTO group was higher (P < 0.05) compared with the negative control and oeFTO groups. Furthermore, immuno uorescence and ELISA were used to detect alterations in the levels of in ammatory cytokines following both FTO knockdown and overexpression. The expression of anti-in ammatory factors (IL-10 and TGF-β1) was found to be the opposite ( Fig. 2d-f). These data suggest that the m 6 A demethylase FTO inhibits M1 microglial polarization during the in ammatory response.
3. ADAM17 is the downstream target of FTO-mediated m 6 A modi cation in microglia Next, we investigated the molecular mechanism of FTO in regulating microglial polarization and identi ed downstream transcription targets. Compared with the control group, under LPS stimulation, 918 genes were altered, of which 532 genes were up-regulated and 386 genes were down-regulated (Fig. 3a).
Using standard GEO2R analysis and quantile normalization, we chose 88 genes with signi cant changes in the two signaling pathways associated with the initial in ammatory response (Suppl. Table 2). In order to characterize the potential targets involved in the in ammatory response process of BV2 cells regulated by m 6 A, we identi ed 88 in ammatory response process genes with key functions. We overlapped the related genes with 19 genes from the 918 m 6 A regulatory genes in the initial in ammatory response (> two-fold change of m 6 A, Fig. 3b), including TNF-α/NF-kB pathway genes (Fig. 3c). The TNF-α/NF-kB pathway is critical to the polarization of the M1 microglial phenotype [8,38]. The activity of TNF-α is closely associated with the extracellular domain cleaved by ADAM17 [39,40]. According to our analysis, PI3K/Akt and TNF-α/NF-kB signaling pathway genes were enriched in the M1 phenotype group compared with the M0 phenotype group, with several m 6 A motifs that were enriched in mRNA transcripts including HMGB1, RELA, ADAM17, and TNF-α ( Fig. 3d and Suppl. Table 3). The results indicated that ADAM17/TNF-α/NF-kB pathway genes were not affected by FTO-mediated m 6 A modi cations. Moreover, compared with the siFTO group, both the mRNA and protein levels of ADAM17, TNF-α, and NF-κB decreased (Fig. 3e, f). Based on these data, we concluded that the down-regulation of FTO promoted the expression of ADAM17 protein at the translation level by maintaining the stability of the ADAM17 mRNA transcript.

Epigenetic regulation of ADAM17 is mediated by an FTO-m 6 A-dependent mechanism
To verify the hypothesis that FTO promotes ADAM17 protein expression by enhancing the stability of the ADAM17 mRNA transcript, we used real time-PCR to detect the expression of the ADAM17 precursor and mature mRNA. As expected, compared with oeFTO group, both the ADAM17 precursor and mature mRNA in the siFTO group were signi cantly increased (Fig. 4a). Since the mRNA level depends on its transcription and stability, we carried out half-life detection by using actinomycin D treatment and found that the ADAM17 precursor mRNA levels were similar in the FTO overexpression and siFTO groups; however, the mature ADAM17 mRNA levels in the siFTO group were signi cantly increased compared with those in the FTO overexpression group (Fig. 4b). Consistent with the above results, western blot analysis showed that when cells in the groups were treated with the protein translation inhibitor cycloheximide (CHX), the half-life of ADAM17 protein in cells of the siFTO group was longer than that in cells of the FTO overexpression group (Fig. 4c). The results suggest that FTO-m 6 A-modi cation of ADAM17 not only increases the translation level of ADAM17 protein by enhancing the stability of mature ADAM17 mRNA, but also enhances the stability of ADAM17 protein.
In order to prove that the m 6 A modi cation site of ADAM17 mRNA was directly demethylated by FTO, we prepared multiple fragments of the ADAM17 transcript through in vitro mRNA transcription experiments, including the 5'UTR (1-200) CDS (201-2484), and 3' UTR (2489-4451) regions (Fig. 4d). We used sitedirected mutagenesis to replace the adenosine base in the m 6 A consensus sequence of ADAM17 with thymine, thereby eliminating the three potential m 6 A sites in the CDS and 3'UTR regions (RRACH) (Fig.  4e). As a result, the mutant 03 of ADAM17 had markedly decreased luciferase activity compared with wild type ADAM17 when normalized to the Renilla data (Fig. 4f). These data demonstrated that FTOmediated m 6 A demethylation accelerated the degradation of ADAM17 mRNA, while silencing of FTO enhanced the stability of ADAM17 mRNA, which in turn increased its protein expression and stability in an m 6 A demethylase-dependent manner.

Inhibition of ADAM17 blocks M1 microglial polarization driven by FTO-m 6 A -modi cations
To further verify the role of ADAM17 as a downstream target gene of FTO in microglial polarization, we explored the co-localization of FTO and ADAM17 mRNA in BV2 cells. Over-expression of FTO resulted in the degradation of ADAM17 RNA. Over-expression of FTO showed a strong signal of red uorescence (FTO+) and a relatively weak signal of green uorescence (ADAM17+), while down-regulation of FTO showed the opposite trend. In agreement with the results of luciferase activity and mRNA half-life results, the Cy3-labeled ADAM17 mRNA was signi cantly more localized to the FTO protein in the LPS-treated group compared with the untreated group, as detected by AlexaFluor488 uorescence. We found that BV2 of oeFTO group possessed the strongest yellow signal with a signi cant positive correlation coe cient of Rr value both in scatter and line pro le analyze responding to LPS (Fig. 5a, b). Given that the silencing of FTO promoted the expression of ADAM17 in the polarization process of BV2 cells, we next inhibited ADAM17 expression to examine whether M1 polarization could be blocked. As we expected, despite LPS stimulation, inhibition of ADAM17 reduced the FTO-mediated M1 polarization in BV2 cells treated with TAPI-1. To further examine the ADAM17-mediated effects on M1 polarization, we detected a number of genes associated M1/M2 polarization (Fig. 5c, d) and tested a series of pro-in ammatory and antiin ammatory factors by ELISA (Fig. 5e). Taken together, we concluded that reduced ADAM17 expression resulted in increased FTO expression to promote M1 polarization of BV2 cells.
6. FTO is closely related to brain injury after TBI A TBI model was used to verify whether FTO had an effect on microglial polarization during the initial in ammatory process caused by brain injury in vivo (Fig. 6a). Western blot results showed that FTO levels decreased signi cantly at days 1, 3, and after TBI, with the most signi cant decline on day 3 (P < 0.05), and gradually increased near baseline levels on day 14 after injury (Fig. 6b). Immunohistochemistry also showed that FTO in the cortical injury area was signi cantly reduced at day 3 after in the TBI group (Fig.  6c). The results of immuno uorescence double staining showed that FTO was mainly expressed in microglia (Fig. 6d). Over-expression of FTO induced by NADP has been veri ed [33], and our results were in agreement Fig. 6e.
Modi ed neurological severity scoring (mNSS), brain water, and Evans blue dye content were used to evaluate post-TBI neurological function after over-expression of FTO. The results showed that, compared with the TBI group, the mNSS of the TBI+NADP group was remarkably improved at day 3 after TBI (Fig.   6f). Compared with the TBI group, the brain water content and the penetration rate of Evans blue in the TBI+NADP group were signi cantly reduced compared with the TBI group ( Fig. 6g-i).
Concurrently, compared with the TBI group, the neuronal apoptosis rate in the TBI+NADP group was signi cantly lower at day 3 after TBI compared with the TBI group (Fig. 7a-c). The number of M1 microglia (CD86+/Iba+) in the TBI+NADP group was lower, whereas the number of M2 microglia (CD206, Arg-1) was signi cantly increased, compared with that in the TBI group. (Fig. 7d, e). Finally, the production of proin ammatory cytokines decreased remarkably in the TBI+NADP group (Fig. 7f). These data demonstrate that over-expression of FTO inhibited microglia-induced in ammation and improved neurological function after TBI.

FTO regulates neuroin ammation after TBI by targeting ADAM17 in microglia
Western blot results showed that NADP intervention highly inhibited ADAM1 , TNF-alpha, and NF-κB p65 expression (Fig. 8a). Subsequently, exosomes in the supernatant of ADAM17-BV2 overexpression group were collected for identi cation, concentration, and quanti cation (Suppl. Fig. 1a-c). Based on the experimental results of the TBI group and TBI+NADP group (Fig. 6-7), we selected the intervention at 3 days after TBI (Fig. 8b, Schematic Diagram of Modeling). Nissl staining and Evans blue dye were used to evaluate the therapeutic effect of exo-oeADAM17 on neurological function after FTO overexpression.
Compared with the TBI+NADP group, the apoptosis rate of neurons in the TBI+NADP+exo-oeADAM17 group was signi cantly increased (Fig. 8c). Meanwhile, the penetration of the corresponding Evans blue dye in the TBI+NADP+exo-oeADAM17 group was signi cantly increased (Fig. 8d). Compared with the TBI+NADP group, the TBI+NADP+exo-oeADAM17 group had a higher level of the M1 microglial biomarkers CD86 and iNOS (Fig. 8e), and the production of pro-in ammatory cytokines was remarkably increased (Fig. 8f). These data indicate that the over-expression of FTO contributed to the inhibition of the in ammation induced by microglia and improved the nerve function after brain trauma.

Discussion
The main ndings of this study include: 1) In BV2 cells and mice with TBI, the expression of FTO in the LPS treatment group was signi cantly down-regulated. The down-regulation of FTO expression increased the m 6 A level in M1 microglia in the entire transcriptome. 2) After FTO interference, the M1/M0 phenotype detection experiments revealed the BV2 cells shifted from the M0 to M1 the phenotype as the population of CD11b + /CD86 + and secretion of pro-in ammatory cytokines increased.
3) The m 6 A peaks localized to the ADAM17 and TNF-α genes increased, especially in the 3'UTR and 5'UTR regions of the ADAM17 gene.

4) FTO may affect the transcription modi cation of ADAM17 and the expression of downstream factors
associated with the TNF-α/NF-kB pathway. 5) Inhibition of ADAM17 blocked the M1-phenotypic transformation of microglia caused by FTO-m 6 A-modi cation. In short, our study found that FTO-related m 6 A modi cation regulated the activation of microglia and neuroin ammatory response. FTO regulated the in ammatory response induced by microglia by stabilizing the expression of ADAM17, and may be considered as a new potential target for the treatment of brain injury (Fig. 9).
Our previous studies con rmed that the neuroin ammatory response, mediated by microglial polarization, plays an important role in the secondary nerve injury after TBI [3,8,9]. Exploring the molecular mechanisms of microglial polarization regulation is particularly important for improving neurological function after TBI. The results of the present study revealed that m 6 A methylation was related to microglial polarization. Here, we found that FTO down-regulated the translation and expression of ADAM17 gene by inhibiting the 3'UTR and 5'UTR m 6 A modi cation levels of ADAM17 mRNA in microglia, which affected the expression of downstream factors associated with the TNF-α/NF-kB pathway, thereby inhibiting M1 microglial polarization. Down-regulation of FTO led to abnormally high expression of ADAM17 in microglia, which promoted in ammation after TBI (Fig. 9).
M 6 A modi cation is one of the most common ways to modify mRNA in eukaryotic cells [21,24]. As the rst identi ed m 6 A demethylase, FTO was found to be involved in the regulation of dopamine signal transduction in the midbrain of mice [27,41]. The midbrain includes areas associated with the formation of learning and memory, as well as adult neurogenesis [41]. FTO also has an important contribution to immune in ammation [42,43]. However, the role of FTO in TBI is not fully understood. More importantly, the role of FTO in the immunophenotypic transition of microglia remains unclear. This study has found that FTO regulates the expression of ADAM17 in microglia by inhibiting m 6 A modi cation, which in turn blocked the immunophenotypic transformation of microglia, thereby affecting the early in ammatory response after TBI. Although we have revealed the epigenetic regulation of FTO in microglia, another m 6 A methyltransferase, METTL3, was previously found to promote LPS-induced microglial in ammation by activating the TRAF6/NF-κB signal pathway [24,44]. METTL3 also utilizes m 6 A to up-regulate the expression of TRAF6, and promote the expression of in ammatory cytokines and proteins related to the M1 phenotype [45][46][47]. The YTHDF proteins, known as m 6 A-binding proteins, may also in uence the stability of methylated RNA to regulate transcription [37,48]. In this study, we systematically analyzed the key regulatory factors of m 6 A modi cation that were related to M1 microglial polarization after TBI and in an LPS-induced microglial cell line. We found that the expression of YTHDF1 and YTHDF2 increased in the M1 phenotype group compared with that of the M0 phenotype group but that the difference was not statistically signi cant. FTO and METTL3 were decreased in the M1 phenotype group compared with the M0 phenotype group and, in particular, the FTO level decreased signi cantly. The down-regulation of FTO expression was closely related to the polarization of microglia after TBI. Compared with the M0 normal group, the overall level of m 6 A in M1 microglia induced by LPS or after TBI was signi cantly increased.
Further studies con rmed that the changes in methylation levels were related to the low expression of FTO and clari ed the role of FTO in the regulation of microglial M1 polarization during the early stages of TBI.
ADAMs are a family of metalloendopeptidases belonging to the zinc dependent superfamily of enzymes, which are involved in a variety of biological processes [10]. ADAM protein sheddase activity mediates the separation of the extracellular domain of membrane-anchored receptors that can cleave a variety of substrates, including growth factors (all the EGFR ligands), cytokines (e.g., pro-TNF-α), cytokine receptors (e.g., IL-6R, TNF-R, and TGF-βRs), ErbB ligands (e.g., TGF-α and TGF-β), and amyloid precursor protein [11,12,49,50]. In particular, TNF-α, TGF-α, and amphiregulin are cleaved by ADAM17 sheddase activity [11,12]. TNF-α is a type II transmembrane protein expressed on the cell surface in a membrane-bound form.
After cleavage by ADAM17, soluble TNF-α binds to the TNF-α R to activate the NF-κB-related signaling pathway, which in turn initiates and regulates the in ammation cascade [14,15]. According to our data, TNF-α/NF-kB signaling pathway genes were enriched in LPS-stimulated BV2 cells (M1 phenotype) and that the m 6 A motifs were enriched in ADAM17 and TNF-α mRNA transcripts, but not in ADAM10. Our data indicated that the ADAM17/TNF-α/NF-kB pathway genes are affected by FTO-mediated m 6 A modi cation and are critical to M1 microglial polarization.
In order to con rm that ADAM17 is the main downstream target of FTO-mediated regulation of M1 microglial polarization, we screened differentially modi ed gene transcripts through m 6 A-sequencing.
Through mRNA half-life and protein half-life experiments, the regulatory mechanism of FTO on ADAM17 was analyzed at both the molecular and protein levels. The results showed that FTO participates in the transcriptional regulation of ADAM17 by affecting the stability of ADAM17 mRNA and increasing the translational speed of ADAM17 protein. Double luciferase mutation further veri ed that siFTO mainly regulated the 3'UTR of ADAM17, suggesting that ADAM17 is the main downstream target of FTO. Finally, we con rmed that ADAM17 is the key gene target of FTO-mediated regulation of M1 microglial polarization, by showing that FTO overexpression speci cally inhibited ADAM17 both in vitro and in vivo. Targeted overexpression of ADAM17 signi cantly counteracted the anti-in ammatory effects of M1 microglial polarization induced by FTO overexpression. In the future, research on FTO knockout mice should be carried out to study the mechanism of FTO-mediated inhibition of ADAM17 and regulation of microglial-induced in ammation.

Conclusions
In this study, we provide convincing evidence that FTO, as a key m 6 A demethylase, is down-regulated in a TBI mouse model and that low expression of FTO induces an increase in the level of methylation that is signi cantly associated with microglial polarization. M 6 A modi cation of FTO at the 3'UTR site of ADAM17 causes the degradation of ADAM17. In the TBI model, the low expression of FTO resulted in the up-regulation of ADAM17 expression and the M1 polarization of microglia M1, which was accelerated through epigenetic modi cation. In summary, our results indicate that FTO regulates the in ammatory response by inducing M1 microglial polarization and stabilizing the expression of ADAM17. FTO is expected to become a new target for the treatment of craniocerebral injury.

Declarations
Ethics approval and consent to participate Human subjects or samples were not used in this study. All animal experiments were approved by the Ethics Committee of the Second A liated Hospital of Fujian Medical University.

Consent for publication
Consent for publication is not applicable for this manuscript.

Availability of data and materials
All the datasets and materials supporting the conclusions of this article are presented in the manuscript.

Competing interests
The authors declare that they have no competing interests.   cells. Compared with the FTO overexpression group, de ciency of FTO in BV2 cells caused an upregulation of the downstream effector molecules of the ADAM17/TNF-α/NF-kB pathway, including ADAM17, NF-κB p65, IL-6, IL-1β, and TNF-α (P < 0.05). d m 6 A RNA-sequencing pro le of blank and LPStreated BV2 cells. By comparing m 6 A-sequencing with FTO overexpression BV2 cells, ADAM17, TNF-α and NF-κB mRNA in siFTO cells was mainly concentrated in the CDS and 3'UTR regions. e, f Compared with siFTO BV2 cells, the levels of ADAM17, TNF-α, and Nf-κB mRNA in FTO overexpression BV2 cells decreased (e), as did the protein levels (f). Values are expressed as mean ± SD from at least 3 independent experiments and the dots represent the value of each experiment. N.S. , nonsigni cant, P > 0.05; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 4
The epigenetic regulation of ADAM17 was by an FTO-m 6 A-dependent mechanism. a Detection of ADAM17 precursor and mature transcripts by qPCR. Compared with FTO overexpression BV2 cells, the precursor mRNA and mature mRNA of ADAM17 in siFTO BV2 cells were signi cantly enhanced. b Detection of the half-life of ADAM17 in FTO overexpression and siFTO BV2 cells pretreated with actinomycin-D for 90 min and analyzed for precursor or mature ADAM17 mRNA at 0, 2, 4, 6, and 8 h. The results revealed that in FTO overexpression and siFTO BV2 cells, precursor ADAM17 mRNA showed no signi cant difference, however, compared with the FTO overexpression BV2 cells, mature ADAM17 mRNA in the siFTO BV2 cells was signi cantly increased (P < 0.05). c The FTO overexpression and siFTO BV2 cells were pretreated with CHX for 90 min, and western blot analysis was applied to examine the expression of ADAM17. The results showed that compared with the cells treated with CHX, ADAM17 protein in the siFTO BV2 cells had a longer half-life (P < 0.05). d Schematic representation of positions of the m 6 A motifs within ADAM17 mRNA. e Schematic representation of mutated (GGAC to GGTC) 3'UTR pmirGLO vector to investigate the role of m 6 A in ADAM17 expression. f BV2 cells transfected with pmirGLO-3′UTR or pmirGLO-3′ UTR-Mut1/2 reporter plasmids. The mutant 03 of ADAM17 had markedly decreased luciferase activity compared with wild type BV2 cells, normalized to the Renilla data. Values are expressed as mean ± SD from at least 3 independent experiments and the dots represent the value of each experiment. N.S. , nonsigni cant, P > 0.05; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 5
Inhibition of ADAM17 in vitro can block microglial polarization after FTO-m 6 A modi cation. a, b In the FTO overexpression (oeFTO) and FTO knockdown (siFTO) BV2 cells treated with or without LPS, in situ hybridization results indicated that ADAM17 co-localized with FTO (a). Scale bars = 20 μm. We found that BV2 cells of oeFTO group possessed the strongest yellow signal with a signi cant positive correlation coe cient of Rr value both in scatter and line pro le analyze responding to LPS (b). c Suppression of ADAM17 alleviated reduced M1 (CD86 + /Iba-1 + ) microglial polarization in TAPI-1 treated BV2 cells mediated by FTO, despite LPS treatment. Scale bars = 50 μm. d M1 phenotype biomarkers expression in the NC, siFTO, and siFTO+TAPI-1 groups. Compared with the siFTO group, the siFTO+TAPI-1 group had signi cantly reduced expression of CD86, ADAM17, and iNOS, and increased the expression of CD206 and Arg-1 (P < 0.05). e ELISA results showed that the siFTO+TAPI-1 group had signi cantly reduced expression of IL-1β, TNF-α, and IL-6 and increased expression of TGF-β1, compared with the siFTO group (P < 0.05). Values are expressed as mean ± SD from at least 3 independent experiments and the dots represent the value of each experiment. N.S. , nonsigni cant, P > 0.05; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 6
FTO is closely related to brain injury after TBI. a Experimental scheme and a schematic brain section after TBI. The molecular biology study was performed on post-injury days 1, 3, 7, and 14. Areas in red refer to lesion sites. b The FTO level decreased signi cantly at days 1, 3 and 7 after TBI. In particular, the FTO level dropped most signi cantly at day 3 after TBI, after which the FTO level increased slowly to a near normal level at day 14 after TBI (P < 0.05). c Immunohistochemistry also showed that FTO expression in the cortical injury area was signi cantly reduced at day 3 after TBI the TBI group. Scale bars = 50 μm. d Double immuno uorescence staining showed that FTO was mainly expressed in microglia (Iba-1+). Representative photomicrographs of immuno uorescence double staining are shown.
e Over-expression of FTO was induced by NADP. Compared with the TBI group, NADP intervention signi cantly increased FTO expression. f Compared with the TBI group, the nerve function score of the TBI+NADP group was signi cantly improved at day 3 after TBI (P < 0.05). g The water content of brain tissue in the TBI+NADP group was signi cantly reduced at day 3 after TBI (P < 0.05). h The TBI+NADP group had signi cantly less extravasation of Evans Blue dye than the TBI group (P < 0.05). i Representative photos of Evans blue dye extravasation in the experimental groups at 3 days after the TBI. Values are expressed as mean ± SD from at least 3 independent experiments and the dots represent the value of each experiment. N.S. , nonsigni cant, P > 0.05; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 7
Enhancement of FTO can reduce neuronal apoptosis and inhibit neuroin ammation after TBI in vivo. a, b The percentage of apoptotic cells was higher in the TBI group than in the sham group (P < 0.05). Compared with the TBI group, the neuronal apoptosis rate in the TBI+NADP group was signi cantly lower at day 3 after TBI (P < 0.05). Representative photomicrographs of Nissl-stained neurons are shown. The arrows indicate apoptotic neurons. Scale bars = 50 μm. c Western blot analyses revealed that TBI resulted in the up-regulation of apoptotic factors in the injured cortex at day 3 after TBI. Compared with the TBI group, the levels of cleaved caspase-3 and Bax in the TBI+NADP group were decreased, and the antiapoptotic factor Bcl-2 was increased (P < 0.05). d Enhancement of FTO redcued M1 (CD86+/Iba-1+) microglial polarization. Representative photomicrographs of CD16-positive microglia are shown. Scale bars = 50 μm. e Compared with the TBI group, the TBI+NADP group had signi cantly reduced expression of ADAM17, CD86, and iNOS, and increased the expression of CD206 and Arg-1 (P < 0.05). f ELISA results showed that the TBI+NADP group had signi cantly reduced the expression of TNF-α, IL-1β, IL-6 and IFN-γ, compared with the TBI group (P < 0.05). Values are expressed as mean ± SD from at least 3 independent experiments and the dots represent the value of each experiment. N.S. , nonsigni cant, P > 0.05; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 8
FTO regulates neuroin ammation in vivo by targeting ADAM17 in microglia after TBI. a Western Blot demonstrated that NADP intervention highly inhibited ADAM17, TNF-alpha, and NF-κB p65 at the protein level. b Exosomes in the supernatant of ADAM17 overexpression cells were collected for identi cation, concentration, and quanti cation. The exosomes were subsequently injected into the damaged cortex. c The apoptosis rate of neurons in the TBI+NADP+exo-oeADAM17 group was signi cantly higher than that in the TBI+NADP group at day 3 after TBI (P < 0.05). Representative photomicrographs of the Nisslstained neurons are shown. The arrows indicate apoptotic neurons (P < 0.05). Scale bars = 50 μm. d The brain water content of the TBI+NADP+exo-oeADAM17 group increased signi cantly at day 3 after TBI (P < 0.05). e Compared with the TBI+NADP group, the TBI+NADP+exo-oeADAM17 group had signi cantly increased expression of ADAM17, CD86, and iNOS, and decreased expression of CD206 and Arg-1 (P < 0.05). f ELISA results showed that the TBI+NADP+exo-oeADAM17 group had signi cantly increased expression of TNF-α, IL-1 β, IL-6 and IFN-γ (P < 0.05). Values are expressed as mean ± SD from at least 3 independent experiments and the dots represent the value of each experiment. N.S. , nonsigni cant, P > 0.05; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 9
Schematic illustrating the possible mechanisms of FTO in microglial polarization and the neuroin ammatory response after TBI.
As illustrated, FTO, as an important m 6 A demethylation enzyme, can affect the transcriptional modi cation of A disintegrin and metalloproteinase 17 (ADAM17). The TNF-α/NF-kB pathway is critical to the polarization of the M1 microglial phenotype, which is regulated by the cleavage function of ADAM17. Down-regulation of FTO expression causes abnormally high expression of ADAM17 and downstream TNF-α/NF-kB pathway related factors in microglia, which promote the development of in ammation in the early pro-in ammatory process after TBI.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.