C57BL/6 mice were housed and bred in the Comparative Medicine animal facilities of Tongji University School of Medicine (TUSM). All procedures were conducted according to protocols approved by the Institutional Animal Care and Use Committee (IACUC) of TUSM. Mouse primary microglia were isolated from whole brains of C57 mice at postnatal day 1 as previously described (9). Briefly, neonatal mouse brains were dissected out after removing peripheral blood vessels and washed with HBSS. Mouse brains were digested at 37°C for 30 min in 0.25% trypsin solution (Invitrogen) supplemented with 0.05% DNase I (Invitrogen). Digestion was stopped by FBS (Invitrogen). The tissue dissociates were centrifuged at 1500 rpm for 5 min at 4°C. Dissociated cells were re-suspended and cultured in DMEM with 10 ng/mL GM-CSF, 10% FBS, 50 U penicillin, and 50 mg/mL streptomycin at 37°C. The culture medium was replaced every three days. Microglia in the microglia-astrocytes mixed cultures were induced to detach by shaking and collected by 1500 rpm centrifugation for 5 min at 4°C.

The siRNA scrambled control and siRNAs for Igf2bp1 (sense sequence: 5’-GUCCCAAGGAGGAAGUAAATT-3’), Gbp11 (sense sequence: 5’-GGCCUUAUUUCAUUCUUUATT-3’), and Cp (sense sequence: 5’-GCCACCAAUUCAUGCAAAUTT-3’) were purchased from GenePharma (GenePharma Co., Ltd, Shanghai). Transfection was performed using the HiPerFect Transfection Reagent (Qiagen301705) according to the manufacturer’s instruction. Briefly, microglia were seeded at 1 × 10⁶ cells per well of a 6-well plate in 1500 µl of microglia culture medium one day before transfection. 3 μg siRNA was diluted in 1 ml culture medium without FBS. After 5 minutes, 12 µl HiPerFect Transfection Reagent was then add to the diluted siRNA. siRNA and transfection reagent were incubated for 10 min at room temperature to allow formation of transfection complexes. The complexes were added drop-wise onto the cells. After 8 h, siRNA-containing medium was replaced with microglia culture medium. Transfection efficiency was determined by examining the transcript levels of target genes 48 h post transfection via qRT-PCR. Specific primer sets were included in Supplementary Table 1 .

The messenger RNAs (mRNAs) were isolated from cell samples using Rneasy mini kit (Qiagen) according to the manufacturer’s instructions. Genomic DNA was removed using Dnase I digestion kit (Qiagen) and cDNA was synthesized using miScript II RT kit (Qiagen). Transcripts were amplified using specific primer sets ( Supplementary Table 1 ) and SYBR green PCR kit (Qiagen) with the ABI7500 (Applied Biosystems). Reactions were run in triplicates for each sample and no-template blanks were used as negative controls. Gapdh was used as internal control for value normalization.

Western blotting was performed as previously described (9). Cell samples were lysed in RIPA lysis and extraction buffer (ThermoFisher) containing a protease inhibitor cocktail (Sigma). Protein concentrations were determined using BCA Protein Assay Kit (Pierce). Proteins (20~30 mg) from cell lysates were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and electrophoretic transferred to polyvinylidene fluoride membranes (Millipore and Bio-Rad). Membranes were incubated with primary antibodies for TNFα (rabbit, Abcam, ab183218, 1:1000), IL1β (rabbit, Abcam, ab234437, 1:1000), CD68 (mouse, BD Abcam, ab125212, 1:1000), and β-actin (mouse, Abcam, ab8226, 1:10000), overnight at 4°C followed by a secondary anti-rabbit or anti-mouse antibody (Cell Signaling Technologies, 1:10000) incubation. Antigen-antibody complexes were visualized by Pierce ECL Western Blotting Substrate (ThermoFisher). Membrane images were acquired using CanonScan 9950F scanner and analyzed using ImageJ program.

Conditioned medium of primary mouse microglia with/without the perturbation of gene function was collected and the concentration of pro-inflammatory cytokines were measured with commercially available ELISA kits (Multi Sciences, EK282/4) according to manufacturer’s protocols. Briefly, microglia were seeded onto the 6-well at a density of 1 × 10⁶ cells/well. 48 h after siRNA transfection, culture medium was collected and centrifuged at 300 g for 10 min. Diluted standards and culture medium in triplicate were added to the corresponding wells and incubated at room temperature for 2 hours on a microplate shaker. Sample Diluent was used as blank control. 100 µl Diluted Streptavidin-HRP was added to all wells and incubate at room temperature for 45 minutes. 100 µl TMB Substrate Solution was then added to all wells and incubate at room temperature for 30 minutes. The enzyme reaction was stopped by quickly pipetting 100 µl of Stop Solution into each well. Absorbance of each well was read using spectrophotometer DV8200 (Drawell). TNFα concentrations were calculated according to the standard curve.

100 μg of total RNA was extracted and purified using RiboMinus™ Eukaryote Kit v2 (A15020, Invitrogen) to deplete the ribosomal RNA from the total RNA. Next, RNA Fragmentation Reagents (AM8740, Invitrogen) were used to shear the RNA into approximately 100-nt fragments. Approximately 1/10 of the fragmented RNA was saved as the input control for further RNA sequencing. The remaining were incubated with an anti-m6A antibody (202,203, Synaptic Systems) for one hour at 4°C, and then mixed with prewashed Pierce™ Protein A/G Magnetic Beads (88,803, Thermo Scientific) in immunoprecipitation buffer at 4°C overnight. The m6A antibody was digested with proteinase K digestion buffer and the methylated RNA was purified for MeRIP-seq. Paired-end 2 × 150 bp sequencing was performed on the Illumina Novaseq 6000 platform of LC-BIO Bio-tech. (Hangzhou, China). Mapped reads of immunoprecipitation and input libraries were provided for R package exomePeak. HOMER was used for de novo and known motif finding, followed by localization of the motif with respect to peak summit by Perl scripts. Then, StringTie was used to measure expression levels for all mRNAs from input libraries by calculating fragments per kilobase of transcript per million fragments mapped (FPKM). Differential expression analysis was performed using the DESeq2 R package. P-values and q-values were adjusted using the Benjamini and Hochberg’s approach for controlling the false discovery rate. Genes with q-value < 0.05 were considered as differentially expressed. Differentially expressed genes (DEGs) with over 2 fold changes were mapped to Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analysis. GO and KEGG enrichment analysis was performed using The Database for Annotation, Visualization and Integrated Discovery (DAVID) (http://david.ncifcrf.gov/).

To detect objective RNA stability, primary microglia were seeded in 12-well plates and treated with 5 μg/mL actinomycin D (Med-ChemExpress) and then collected at the indicated time points. Total RNA was extracted by RNeasy mini kit (Qiagen) according to the manufacturer’s instructions. qRT-PCR assay was used to determine the levels of mRNAs. Specific primer sets were included in Supplementary Table 1 . Gapdh was used as internal control for value normalization. The mRNA half-live times were estimated according to the linear regression analysis.

All results are the means of at least three independent experiments ± s.d. Data from two groups were evaluated statistically by two-tailed unpaired student t test. Data from multiple groups were evaluated statistically by either one-way ANOVA followed by Tukey’s post hoc test or two-way ANOVA followed by Sidak’s multiple comparisons test according to the experiment design. Significance was considered when p < 0.05.

To examine the changes of m6A profiles in microglial activation, we utilized LPS to stimulate the inflammatory responses of microglia in vitro (9). Mouse primary microglia were treated with 100 ng/ml LPS for 3 h and the inflammatory responses of microglia were tested at 48 h. Western blotting results found enhanced expression of pro-inflammatory proteins TNFα, IL1β, and CD68 in LPS-treated microglia ( Supplementary Figure 1A ). The expression of transcripts corresponding to pro-inflammatory genes TNF, Il1b, and Nos2 was similarly increased in LPS-treated microglia ( Supplementary Figure 1B ). The release of TNFα from LPS-treated microglia was also promoted, confirming the successful activation of microglia via LPS stimulation ( Supplementary Figure 1C ).

The m6A signature of pro-inflammatory microglia was then determined by MeRIP-seq approach. MeRIP-seq analysis identified 12952 and 14552 m6A sites in control and LPS-stimulated microglia, respectively, and among them, 10026 m6A sites were shared by both groups ( Figure 1A ). MeRIP-seq analysis showed that m6A peaks both in control and LPS-stimulated microglia were primarily enriched in the coding sequence and 3’ untranslated regions (UTR) regions ( Figure 1B ). Furthermore, there were 3421 up-regulated m6A peaks and 4504 down-regulated m6A peaks in LPS-stimulated microglia, which distributed in all chromosomes ( Figures 1C, D ). The sequence logo showed distinct enriched m6A motifs in control and LPS-stimulated microglia ( Figures 1E, F ). By analyzing the localization of all m6A sites in mRNAs, 2391 and 1731 mRNAs were identified with down-regulated and up-regulated m6A modifications, respectively ( Figure 1G ). The majority of mRNAs had one m6A peak (about 55%) ( Figure 1H ). Meanwhile, there were over 3% of mRNAs had over 5 m6A peaks (126/4123) ( Figure 1H ).

To examine the association of m6A methylation with gene expression in activated microglia, we then carried out RNA-seq using control and LPS-treated microglia. 455 and 278 mRNAs were significantly up- and down-regulated in LPS-treated microglia, respectively, versus controls ( Figures 1I, J ). Afterwards, four-quadrant plots were utilized to divide mRNAs with over two folds m6A methylation and expression levels between control and LPS-treated microglia ( Figure 1K ). Among them, 2106 hypermethylated mRNAs that were significantly up-regulated (2043) or down-regulated (63) in expression, and 3751 hypomethylated m6A peaks in mRNAs that were significantly up-regulated (17) or down-regulated (3734) in expression ( Figure 1K ). We next drew venn diagram using miRNAs with significantly different m6A methylation and expression levels (over two folds, p<0.05). Over 80% of up-regulated mRNAs were hypermethylated ( Figure 1L ), and about 75% of down-regulated mRNAs were hypomethylated ( Figure 1O ). In contrast, only 1 down-regulated mRNA was hypermethylated ( Figure 1O ), and no overlapping mRNAs were identified among up-regulated and hypomethylated mRNAs ( Figure 1L ). We next examined the potential roles of mRNAs with altered m6A methylation and expression patterns via GO and KEGG analyses. GO biological process analysis revealed that the identified 225 mRNAs with up-regulated m6A peaks and expression levels were enriched in immune response-related GO terms, suggesting the strong association of m6A-hypermethylated mRNAs with the pro-inflammatory function of microglia ( Figure 1M ). Similarly, KEGG analysis indicated that the mRNAs with up-regulated m6A peaks and expression levels were enriched in inflammatory pathways including TNFβ, Toll-like receptor, and Nf-κb signaling ( Figure 1N ). In contrast, both GO and KEGG analyses revealed that the mRNAs with down-regulated m6A peaks and expression levels were involved in cell cycle and metabolic regulation ( Figures 1P, Q ). Hence, the MeRIP-seq and RNA-seq results demonstrated a positive correlation between the m6A methylation and expression of pro-inflammatory genes in LPS-stimulated microglia, implying the involvement of m6A modification in the regulation of microglial activation.

To identify the key factor that mediates the LPS-induced m6A modifications, we first screened the expression patterns of currently known m6A writers, erasers, and readers in control and LPS-treated microglia using RNA-seq data ( Figure 2A ). Among all tested m6A regulators, Igf2bp1, a recently identified m6A reader (10, 11), was demonstrated as the most significantly regulated one. The RNA-seq data was confirmed by qRT-PCR assay, in which the expression levels of Igf2bp1 transcripts exhibited the largest fold changes in LPS-treated microglia versus controls ( Figure 2B ). Furthermore, we determined the effects of LPS stimulation on Igf2bp1 expression by treating microglia with different doses of LPS. qRT-PCR results revealed an elevation of Igf2bp1 transcript levels with the increase of LPS concentration till 500 ng/ml ( Figure 2C ). Western blotting results suggested that the expression of Igf2bp1 proteins was promoted by increasing LPS concentration from 0 ng/ml to 100 ng/ml ( Figure 2D ). No significant difference was observed in the levels of Igf2bp1 proteins when microglia were treated with LPS of 100 ng/ml or higher doses ( Figure 2D ). In addition, the expression levels of both Igf2bp1 mRNAs and Igf2bp1 proteins were also elevated by LPS treated in a time-dependent manner ( Figures 2E, F ).

Next, we examined the expression levels of Igf2bp1 in mouse brains under inflammatory conditions. Neuroinflammation was stimulated by intraperitoneal injection of LPS for 3 d (one injection per day, dose: 1 mg/kg). qRT-PCR analysis demonstrated significant increase of Igf2bp1 transcripts in mouse hippocampal tissues 3 days post LPS administration ( Figure 2G ). Immunohistochemical analysis further revealed a significant increase of the proportions of Igf2bp1⁺cells in Iba1⁺ activated microglia in mouse hippocampal tissues ( Figures 2H, I ). Hence, both in vitro and in vivo studies indicated that the expression of Igf2bp1 in activated microglia was significantly up-regulated, implying a role of Igf2bp1 in LPS-induced m6A modifications in microglia.

To examine the effects of Igf2bp1 in the inflammatory responses of microglia, we performed loss-of-function (LOF) approaches in vitro. Microglia were pre-treated with 100 ng/ml LPS for 3 h and transfected with either Igf2bp1 siRNA or scrambled control for 48 h. The transfection efficiency was determined by demonstrating a ~60% reduction of Igf2bp1 transcripts levels ( Figure 3A ) and over 70% decline of Igf2bp1 protein levels ( Figure 3B ) in LOF group versus LPS controls. qRT-PCR analysis further revealed that Igf2bp1 LOF abrogated the overproduction of transcripts corresponding to TNF, Il1b, and Nos2 induced by LPS treatment in microglia ( Figure 3C ). Similarly, Western blotting results demonstrated that Igf2bp1 LOF significantly reduced the LPS-induced excessive expression of pro-inflammatory factors TNFα, IL1β, and CD68 ( Figure 3D ). ELISA assay also showed a significant decrease of the concentration of TNFα in the conditioned medium in Igf2bp2 LOF group, compared with LPS group ( Figure 3E ). Hence, our results suggested Igf2bp1 as a key regulator for the inflammatory responses of microglia.

Afterwards, we examined the roles of Igf2bp1 on LPS-induced m6A modification in microglia. MeRIP-seq analysis identified 3321 hypermethylated and 2917 hypomethylated m6A peaks in Igf2bp1 LOF group versus LPS group ( Figure 3F ). The differentially methylated m6A peaks distributed into all chromosomes except Y chromosome ( Figure 3G ). There were 2384 hypermethylated mRNAs and 2084 hypomethylated ones, and among them, over 80% mRNAs had one hypermethylated peak ( Figures 3H, I ). Moreover, there were 235 mRNAs that were hypermethylated with LPS stimulation and hypomethylated post Igf2bp1 LOF ( Figure 3J ). The GO and KEGG analyses revealed that those overlapping mRNAs were strongly associated with the immune responses of microglia ( Figures 3K, L ). The bioinformatical analyses indicated that Igf2bp1 LOF reversed the hypermethylation of inflammation-related mRNAs. Besides, there were 372 mRNAs that were hypomethylated with LPS stimulation and hypermethylated post Igf2bp1 LOF ( Figure 3M ). GO and KEGG analyses further demonstrated that Igf2bp1 LOF erased the LPS-induced hypomethylation of cell cycle and metabolic regulation-related mRNAs ( Figures 3N, O ). Therefore, our MeRIP-seq results showed that Igf2bp1 played an important role in the LPS-induced alteration of m6A modifications.

Due to the importance of Igf2bp1 in microglial activation and inflammation-related m6A modification, we hypothesized that Igf2bp1 regulates the inflammatory responses of microglia via modifying mRNA m6A signatures. To test our premise, we first carried out RNA-seq analysis to determine the transcript expression patterns of LPS-stimulated microglia with/without Igf2bp1 LOF ( Figure 4A ). RNA-seq analysis identified 77 up-regulated and 73 down-regulated genes in Igf2bp1 LOF group versus control LPS group ( Figure 4B ). Furthermore, from the 235 genes that were hypermethylated with LPS stimulation and hypomethylated post Igf2bp1 LOF, Ceruloplasmin (Cp) and Guanylate-binding protein 11 (Gbp11) were identified with significantly reduced expression levels post Igf2bp1 LOF ( Figure 4C ). The high-throughput analysis results were confirmed by qRT-PCR which showed significant reduction of the expression levels of both Gbp11 and Cp transcripts in Igf2bp1 LOF groups versus LPS controls ( Figure 4D ). To explore whether Igf2bp1 had impact on the stability of Gbp11 and Cp mRNAs, we treated microglia with actinomycin D in indicated time points before total RNA was acquired. The results showed a significantly shortened half-life of both Gbp11 and Cp mRNAs in Igf2bp1 deficient microglia versus control LPS group ( Figure 4E ).

To examine whether Gbp11 and Cp are potential downstream factors of Igf2bp1, we carried out Gbp11 and Cp LOF by transfecting LPS-stimulated microglia with either scrambled siRNA control, Gbp11 siRNA, or Cp siRNA. qRT-PCR results demonstrated over 50% knockdown of Gbp11 and more than 80% knockdown of Cp transcripts in LOF groups versus LPS controls, suggesting efficient transfection ( Figure 4F ). qRT-PCR analysis further showed significant decline of the expression levels of transcripts corresponding to TNF, Il1b, and Nos2 in both Gbp11 and Cp LOF groups versus LPS controls ( Figure 4G ). Western blotting results also demonstrated that both Gbp11 and Cp LOF significantly reduced the LPS-induced up-regulation of TNFα, IL1β, and CD68 protein expressions ( Figure 4H ). Moreover, ELISA assay suggested that both Gbp11 and Cp LOF erased the LPS-induced excessive release of TNFα from microglia ( Figure 4I ). Hence, our results suggested Gbp11 and Cp as pro-inflammatory proteins, implicating that Igf2bp1 mediates microglial activation via stabilizing Gbp11 and Cp mRNAs.