METTL3-Mediated m6A Modification of lncRNA-0949 Drives Microglial Inflammation in an vitro Model of Sepsis-Associated Encephalopathy Authors: Xing Zeng, Xiao-Juan Luo, Zhen-Ze Zhang, Jia-Li Lu, Yi-An Zhan Published: 2026 ABSTRACT IntroductionSepsis-associated encephalopathy (SAE) is a severe complication of sepsis with limited therapeutic options. Although neuroinflammation driven by microglial activation is central to SAE pathogenesis, the underlying epitranscriptional regulatory mechanisms remain poorly defined. Here, we investigated the role of the m6A methyltransferase METTL3 in regulating microglial inflammation using lipopolysaccharide (LPS)-stimulated HMO6 microglial cells as an in vitro SAE model.MethodsHMO6 cells were stimulated with LPS (1 μg/mL) for 0–24 h to establish an SAE model. METTL3 expression was assessed by Western blotting and immunofluorescence. MeRIP-qPCR was used to detect m6A deposition on lncRNA-0949. METTL3 was inhibited pharmacologically (3-DAA) or by siRNA knockdown. lncRNA-0949 stability was evaluated by actinomycin D chase assay. Luciferase reporters containing wild-type or m6A site-mutated lncRNA-0949 3′UTR were constructed to identify functional m6A sites. Wild-type and m6A site-mutated lncRNA-0949 overexpression vectors were employed to assess modification-dependent pro-inflammatory function. Cytokine mRNA (qRT-PCR) and protein (ELISA) levels were measured.ResultsLPS stimulation time-dependently increased oxidative stress (MDA), pro-inflammatory cytokines (MIP-2, IL-1β, TNF-α, IL-6), and METTL3 protein expression (2.99-fold at 24 h, P < 0.001). METTL3 catalyzed m6A deposition on lncRNA-0949, with enrichment reaching 21.3-fold at 24 h post-LPS (P < 0.001). METTL3 knockdown abolished this modification and reduced lncRNA-0949 stability, decreasing its half-life from ~30 h to 11 h (P < 0.001). The m6A site within the 3′UTR (RRACH motif) was essential for LPS-induced reporter activity (7.7-fold increase for WT vs. no response for Mut, P < 0.001). Overexpression of wild-type lncRNA-0949 amplified LPS-triggered cytokine release (eg, TNF-α increased by additional 33%, P < 0.001), whereas the m6A site mutant had no effect. Conversely, METTL3 knockdown attenuated LPS-induced inflammatory responses, with mRNA levels reduced by 35–58% and protein levels by 30–63% (all P < 0.001).ConclusionTogether, these findings define a METTL3–m6A–lncRNA-0949 regulatory axis that amplifies microglial inflammation in an in vitro SAE model. This study provides the first evidence that METTL3-driven m6A modification of lncRNA-0949 contributes to neuroinflammation, offering a new mechanistic perspective and highlighting the need for in vivo validation to assess therapeutic potential. Introduction Sepsis is a systemic inflammatory response caused by infection, which is the direct cause of death in critically ill patients in intensive care unit (ICU). In clinical practice, sepsis is mainly manifested by multiple organ dysfunction. Sepsis-associated encephalopathy is one of the common complications of sepsis, which is usually accompanied by acute neurological dysfunction, manifested as disturbance of consciousness, cognitive dysfunction, memory loss and other symptoms.1 Uncontrolled severe inflammation in the brain is the main cause of SAE.2 Microglia are the resident immune cells inherent in the central nervous system. When the CNS is injured or invaded by diseases, microglia will show complex responses and release a series of pro-inflammatory cytokines, thereby inducing neuroinflammatory responses.3–5 Therefore, understanding the mechanisms of how microglia are aberrantly activated is key to developing effective therapies for SAE. Although neuroinflammatory response is the pathological mechanism leading to the development of SAE, its post-transcriptional regulatory mechanisms, particularly at the RNA epigenetic level, remain unclear. Therefore, this study aims to explore the role of RNA epigenetic modifications, specifically m6A, in microglial inflammation during SAE. RNA epigenetic modifications such as m6A and m5C play an important role in regulating gene expression, cell function and inflammatory stress response.6 Among them, m6A (N-6 methyladenosine) is the most important modification in mammalian mRNA and lncRNA.7 In the process of regulating the dynamic modification of m6A, methyltransferases, demethylases and recognition enzymes cooperated. The methyltransferase complex (such as METTL3/METTL14 heterodimer), as the core writing enzyme of m6A modification, exerted global regulation on RNA metabolism by catalyzing the methylation of adenosine at N6 position.8 Studies have shown that METTL3 is the key enzyme that catalyzes the methylation of m6A in RNA, and it plays an important role in regulating the immune inflammatory response caused by macrophages. It can regulate the activation and function of immune cells by binding to the methylation sites of RNA.9,10 Recent studies have found that METTL3 can directly affect the “lifetime” of long non-coding RNA (lncRNA) with the help of m6A methylation, which opens a new door to understand the upstream regulation of lncRNA.11,12 Moreover, m6A modifications are recognized by reader proteins such as IGF2BP2, which can influence RNA stability and function.13 However, the role of METTL3 in microglia, the resident immune cells of the brain, remains poorly understood.14 Long non-coding RNA (lncRNA) is a kind of linear non-coding RNA, which plays an important role in gene regulation.15 In recent years, lncrnas have been found to be closely related to a variety of nervous system diseases and inflammatory responses, and their mechanisms involve transcriptional regulation, competing endogenous Rnas, and activation of related inflammasomes.16,17 LncRNA-0949, first identified in 2014,18 remains functionally uncharacterized. In our previous study, we observed its significant upregulation in LPS-treated iPSC-Exos co-incubated with HMO6 microglial cells (Supplementary Figure 1A and B)—a model where iPSC-Exos reduced microglial inflammatory cytokines (TNF-α, IL-1β, IL-6) upon LPS stimulation. This suggested a potential role for lncRNA-0949 in the anti-inflammatory effect of iPSC-Exos, prompting us to explore its function in microglial inflammation. Given that METTL3 can regulate lncRNA stability via m6A modification,19,20 we now ask whether METTL3 modifies lncRNA-0949 to regulate microglial inflammatory responses in the pathological context of SAE. This question is important because it links the chemical modification of RNA with the function of lncRNA and helps us to understand how neuroinflammation is regulated. The core research hypothesis is that the expression of METTL3 is increased by inflammatory mediators under SAE conditions. The up-regulated METTL3 enhanced the stability of lncRNA-0949 via m6A modification, which ultimately aggravated neuroinflammation. To test this hypothesis, we constructed an in vitro model of LPS-activated human microglia (HMO6) to mimic the SAE environment by (1) confirming whether METTL3 is responsible for the m6A modification and stability of lncRNA-0949; (2) To investigate whether m6A modification is necessary for the pro-inflammatory function of lncRNA-0949; (3) to verify the central role of “METTL3-m6A-lncRNA-0949” in driving the inflammatory response of microglia. We aimed to uncover a novel epigenetic regulatory pathway in SAE and provide new targets for the pathogenesis and potential treatment of the sepsis-associated brain. Materials and Methods Cell Culture and Processing HMO6 cells were purchased from Shanghai Enzyme-linked Biotechnology (cat# C1024) and cultured in DMEM (Hyclone, cat# SH30243.02) supplemented with 10% fetal bovine serum (FBS, Gibco) and 1% penicillin-streptomycin (PS, Hyclone, cat# SV30010) at 37°C in a 5% CO2 humidified incubator. The culture medium was refreshed every 2–3 days. For LPS treatment, cells were seeded in 6-well plates at a density of 1×105 cells per well, allowed to adhere overnight, and then stimulated with 1 μg/mL lipopolysaccharide (LPS, from Escherichia coli O111:B4, Sigma, cat# L5293) dissolved in PBS for 0, 6, 12, or 24 h in DMEM (Hyclone, cat# SH30243.02) containing 1% FBS (Gibco). Control cells received an equivalent volume of PBS. All experiments were performed with at least three independent biological replicates. MDA MDA levels were measured using a commercial kit (S0131S; Beyotime, China) per manufacturer’s instructions. Cells were lysed in PBS by sonication, centrifuged, and supernatants incubated with TBA working solution at 100°C for 15 min. After cooling and centrifugation, absorbance was read at 532 nm. MDA concentrations were calculated via a standard curve and normalized to protein content. RNA m6A Quantification Total m6A levels were measured using the EpiQuik m6A RNA Methylation Quantification Kit (Epigentek, cat# P-9005-48; distributed by Wuhan AmyJet Scientific, China) per manufacturer’s instructions. Briefly, 200 ng of poly(A)-purified RNA was incubated with capture antibody, detection antibody, and enhancer solutions, and absorbance was read at 450 nm. m6A percentage was calculated using the kit’s formula. RT-qPCR Total RNA was extracted from HMO6 cells using TRIzol (Invitrogen). RT-qPCR was performed with SYBR Green Master Mix (Thermo Fisher Scientific) on a QuantStudio 5 system. Relative mRNA expression was calculated by the 2−ΔΔCt21 method with GAPDH as internal control. Primers were designed and synthesized by GenePharma (Shanghai, China). Primer sequences are listed in Table 1.Table 1Primer Sequences of Each GeneGenePrimer Sequence (5′→3′)TNF-αForward: CAGGGGCCACCACGCTCTTCReverse: CTTGGGGCAGGGGCTCTTGAIL-6Forward: ATGAACTCCTTCTCCACAAGCGCReverse: GAAGAGCCCTCAGGCTGGACTGIL-1βForward: TCCCTTCATCTTTGAAGAAGAReverse: GAGGCCCCAAGGCCACAGGMIP-2Forward: TCCAAGAAAGGGCGAAATAAGGReverse: TGCAGCTCTATCTGAATGTCTGTGAPDHForward: TGTGGATGGCCCCTCTGGAAReverse: TGACCTTGCCCACAGCCTTG Table 1Primer Sequences of Each GeneGenePrimer Sequence (5′→3′)TNF-αForward: CAGGGGCCACCACGCTCTTCReverse: CTTGGGGCAGGGGCTCTTGAIL-6Forward: ATGAACTCCTTCTCCACAAGCGCReverse: GAAGAGCCCTCAGGCTGGACTGIL-1βForward: TCCCTTCATCTTTGAAGAAGAReverse: GAGGCCCCAAGGCCACAGGMIP-2Forward: TCCAAGAAAGGGCGAAATAAGGReverse: TGCAGCTCTATCTGAATGTCTGTGAPDHForward: TGTGGATGGCCCCTCTGGAAReverse: TGACCTTGCCCACAGCCTTG Primer Sequences of Each Gene Me-RIP MeRIP was performed using a commercial kit (Junli Biology, GK-4043). Briefly, 100 μg of RNA was fragmented to ~200 nt (70°C, 6 min). Ten percent was saved as input; the remaining RNA was immunoprecipitated with anti-m6A antibody (Abcam, ab151230) at 4°C for 3 h. After washing, bound RNA was eluted and analyzed by RT-qPCR with lncRNA-0949-specific primers (Table 1). m6A enrichment was calculated as %input following the kit manual. Immunofluorescence HMO6 cells on coverslips were fixed, permeabilized, blocked, and incubated with rabbit anti-METTL3 (1:400; 10761-1-AP; Proteintech) and rabbit anti-IGF2BP2 (1:400; 11601-1-AP; Proteintech) at 4°C overnight. After washing, cells were incubated with Alexa Fluor 488 goat anti-rabbit IgG (1:1000; SA00013-2; Proteintech) or Alexa Fluor 555 goat anti-rabbit IgG (1:50; SA00013-4; Proteintech) for 90 min at room temperature. Nuclei were counterstained with DAPI. Fluorescence images were captured using a fluorescence microscope. All experiments were performed with three independent replicates. siRNA Transfection The siRNA targeting human METTL3 (sense strand: 5′-GGAGAUCCUAGAGCUAUUA-3′; antisense strand: 5′-CCUCUAGGAUCUCGAUAAU-3′) was designed based on a previously validated sequence22 and synthesized by GenePharma (Shanghai, China). HMO6 cells were seeded in culture plates and transfected upon reaching 70–80% confluence. According to the manufacturer’s instructions, cells were transfected with either siMETTL3 (experimental group) or negative control siRNA (control group) using siRNA-Mate plus transfection reagent (GenePharma, Cat. No. G04002). Cells were harvested 48–72 hours post-transfection for subsequent experiments. Western Blot Cells were lysed in RIPA buffer (Roche). Protein concentration was determined by BCA assay (Beyotime). Equal protein (50 μg) was resolved by 12% SDS-PAGE, transferred to PVDF membranes, blocked with 5% skim milk, and incubated with primary antibodies (anti-METTL3 ab195352 1:1000, anti-IGF2BP2 ab128175 1:1500, anti-GAPDH ab9485 1:5000; Abcam) overnight at 4°C, followed by HRP-conjugated goat anti-rabbit secondary antibody (ab6721 1:6000; Abcam) for 1 h. Bands were visualized by ECL (Beyotime) and quantified using Quantity One (Bio-Rad) with GAPDH as loading control. ELISA Cell culture supernatants were collected and centrifuged. Levels of TNF-α (88–7346-22), IL-1β (88–7261-22), IL-6 (88–7066-22), and MIP-2 (900-K120) were measured using human-specific ELISA kits (Thermo Scientific, USA) according to the manufacturer’s protocols. Briefly, samples were added to pre-coated plates, incubated with detection antibody, and developed with TMB substrate. Absorbance was read at 450 nm, and cytokine concentrations were calculated from standard curves. DB Experiment RNA samples from control and LPS-treated cells (0,6,12,24 h) were serially diluted (1000–250 ng/μL), denatured (70°C, 5 min), and spotted (2 μL per dot) onto Hybond-N+ membrane (Cytiva, cat# RPN303B) in duplicate for m6A detection and methylene blue staining. After UV-crosslinking, membranes were blocked and incubated with anti-m6A antibody (Abcam, cat# ab151230, 1:1000), followed by HRP-conjugated secondary antibody. Signals were detected by ECL (Beyotime, cat# P0018AS) and quantified by densitometry. Relative m6A enrichment was calculated as fold change vs. the corresponding control at each time point. Methylene blue (Sigma Aldrich, cat# M9140) staining served as loading control. All experiments were performed with three independent biological replicates. RNA Stability Assay Cells transduced with METTL3 shRNA or control shRNA (referred to as siMETTL3 and siCtrl) were either left untreated or stimulated with LPS for 24 h. After actinomycin D (5 μg/mL) treatment, total RNA was collected at 0, 6, 12, and 24 h. lncRNA-0949 levels were quantified by RT-qPCR (GAPDH as internal control, 2−ΔΔCt method), and half-life was calculated from decay curves. Luciferase Reporter Assay The wild-type (WT) lncRNA-0949 3′UTR and its m6A site-mutated (mut, GGACA→GGCCA) sequence were cloned into the pmirGLO vector (Wuhan MiaoLing Biotechnology). HMO6 cells were co-transfected with the indicated reporter constructs and siRNAs, then stimulated with LPS. Luciferase activities were measured using the Dual-Luciferase Reporter Assay System (Beyotime). Firefly luciferase activity was normalized to Renilla luciferase activity.23 Statistical Methods All data are presented as mean ± standard error of the mean (SEM) from at least three independent biological replicates, each performed with three technical replicates. Statistical analyses were conducted using GraphPad Prism (version 9.0). Normality was assessed using the Shapiro–Wilk test, and homogeneity of variances was verified by the Brown–Forsythe test; all datasets met the assumptions for parametric testing. For comparisons between two groups, an unpaired two-tailed Student’s t-test was used. For multiple group comparisons, one-way or two-way analysis of variance (ANOVA) was applied as appropriate, followed by Tukey’s or Šidák’s post hoc test for multiple comparisons. Statistical significance was defined as ***P < 0.001. Results LPS Time-Dependently Activates Microglia and Upregulates the Core Methyltransferase METTL3 To establish an LPS-stimulated human microglial (HMO6) model recapitulating key aspects of sepsis-associated encephalopathy (SAE), we challenged cells with lipopolysaccharide (LPS; 1 μg/mL) and assessed the pathogenic cascade over 0, 6, 12, and 24 h. Neuroinflammatory activation was confirmed by a significant rise in the oxidative stress marker malondialdehyde (MDA). LPS stimulation significantly increased MDA levels in a time-dependent manner (Figure 1A). Compared to control cells, MDA levels in LPS-treated cells were elevated at 6 h (0.64 ± 0.12 vs. 0.29 ± 0.12, 2.2-fold, P = 0.01), 12 h (1.59 ± 0.12 vs. 0.25 ± 0.12, 6.3-fold, P < 0.001), and 24 h (2.23 ± 0.12 vs. 0.29 ± 0.12, 7.7-fold, P < 0.001). Transcript levels of the pro-inflammatory mediators MIP-2, IL-1β, TNF-α, and IL-6 increased significantly in a time-dependent manner following LPS stimulation (P < 0.001 for both time and treatment effects; Figure 1B), a pattern corroborated at the protein level by ELISA (eg, TNF-α reached 305.6 pg/mL at 24 h vs. 26.5 pg/mL in controls, P < 0.001; Figure 1C). This shift to a pro-inflammatory phenotype prompted an investigation of upstream epitranscriptomic regulation. Given our prior transcriptomic data implicating lncRNA-0949 in SAE-relevant contexts (Supplementary Figure 1A and B), we analyzed the core N6-methyladenosine (m6A) methyltransferase METTL3. Western blotting revealed progressive accumulation of METTL3 protein throughout LPS stimulation (2.99 ± 0.19-fold increase at 24 h vs. 0 h, P < 0.001; Figure 1D), a finding localized to the cell via intensified immunofluorescence signal at 24 h (Figure 1E). Collectively, these data position the sustained upregulation of METTL3 as a concomitant feature of evolving neuroinflammation in this cellular model. Figure 1LPS activates microglia and upregulates METTL3 protein. HMO6 cells were stimulated with LPS (1 μg/mL) for 0, 6, 12, and 24 h, with untreated cells serving as controls. (A) Intracellular MDA levels were determined using the TBA method. (B) The mRNA expression of pro-inflammatory cytokines (MIP-2, IL-1β, TNF-α, IL-6) was quantified by RT-qPCR. (C) The secretion of the corresponding cytokines into the supernatant was measured by ELISA. (D) METTL3 protein expression was analyzed via Western blot, with representative blots and densitometric quantification shown. (E) Immunofluorescence staining for METTL3 (red) was performed; nuclei were stained with DAPI (blue). Representative images at 24 h are presented. Scale bar: 5 μm. Data represent the mean ± SD from three independent experiments (n = 3). Statistical significance was determined by two-way ANOVA followed by Šídák’s multiple comparisons test. Asterisks denote statistical significance (*P < 0.05, **P < 0.01, ***P < 0.001).The image A showing a bar chart with legend Control and LPS. X-axis label: Time (hours). Ticks: 0, 6, 12, 24. Y-axis label: MDA level (nmol/mgprot). Range: 0 to 3. Bars: at 0 hours both near 0.3 to 0.4; at 6 hours Control near 0.3 and LPS near 0.7 with asterisk; at 12 hours Control near 0.3 and LPS near 1.6 with three asterisks; at 24 hours Control near 0.3 and LPS near 2.4 with three asterisks. The image B showing four bar charts with legend Control and LPS. All x-axes: Time (hours) with ticks 0, 6, 12, 24. Top-left y-axis label: Relative mRNA expression MIP2 versus GAPDH, range 0 to 8. Values: 0 hours both near 1; 6 hours Control near 1 and LPS near 2 with asterisk; 12 hours Control near 1 and LPS near 4 with three asterisks; 24 hours Control near 1 and LPS near 5 to 6 with three asterisks. Top-right y-axis label: Relative mRNA expression TNF-alpha versus GAPDH, range 0 to 8. Values: 0 hours both near 1; 6 hours LPS near 2 with asterisk; 12 hours LPS near 4 with three asterisks; 24 hours LPS near 6 to 7 with three asterisks; Control near 1 at all times. Bottom-left y-axis label: Relative mRNA expression IL-1beta versus GAPDH, range 0 to 5. Values: 0 hours near 1; 6 hours LPS near 2 with two asterisks; 12 hours LPS near 3.5 to 4 with three asterisks; 24 hours LPS near 3.5 with three asterisks; Control near 1. Bottom-right y-axis label: Relative mRNA expression IL-6 versus GAPDH, range 0 to 8. Values: 0 hours near 1; 6 hours LPS near 4 with three asterisks; 12 hours LPS near 6 to 7 with three asterisks; 24 hours LPS near 6 with three asterisks; Control near 1. The image C showing four bar charts with legend Control and LPS. All x-axes: Time (hours) with ticks 0, 6, 12, 24. Top-left y-axis label: MIP2 content (pg/mL), range 0 to 600. Values: 0 hours near 20; 6 hours LPS near 320 with three asterisks; 12 hours LPS near 420 with three asterisks; 24 hours LPS near 500 with three asterisks; Control near 20. Top-right y-axis label: TNF-alpha content (pg/mL), range 0 to 400. Values: 0 hours near 20; 6 hours LPS near 200 with three asterisks; 12 hours LPS near 300 with three asterisks; 24 hours LPS near 320 with three asterisks; Control near 20. Bottom-left y-axis label: IL-1beta content (pg/mL), range 0 to 400. Values: 0 hours near 20; 6 hours LPS near 200 with three asterisks; 12 hours LPS near 280 with three asterisks; 24 hours LPS near 280 with three asterisks; Control near 20. Bottom-right y-axis label: IL-6 content (pg/mL), range 0 to 1000. Values: 0 hours near 20; 6 hours LPS near 320 with three asterisks; 12 hours LPS near 520 with three asterisks; 24 hours LPS near 820 with three asterisks; Control near 20.A set of 9 bar charts showing time course effects of lipopolysaccharide versus control on markers.Figure 1continued.Western blot and immunofluorescence analysis of METTL3 and IGF2BP2 expression over time with LPS treatment.The image D shows Western blot analysis of METTL3, IGF2BP2 and GAPDH expression at 0, 6, 12 and 24 hours under control and LPS conditions. Relative protein expression graphs for METTL3 and IGF2BP2 versus GAPDH are shown, indicating significant differences at various time points. The image E displays immunofluorescence analysis of DAPI, METTL3, IGF2BP2 and merged images at 0, 6, 12 and 24 hours under control and LPS conditions. DAPI stains nuclei, METTL3 and IGF2BP2 are shown separately and the merge combines all signals to visualize colocalization. The analysis highlights changes in protein localization and expression over time with LPS treatment. LPS activates microglia and upregulates METTL3 protein. HMO6 cells were stimulated with LPS (1 μg/mL) for 0, 6, 12, and 24 h, with untreated cells serving as controls. (A) Intracellular MDA levels were determined using the TBA method. (B) The mRNA expression of pro-inflammatory cytokines (MIP-2, IL-1β, TNF-α, IL-6) was quantified by RT-qPCR. (C) The secretion of the corresponding cytokines into the supernatant was measured by ELISA. (D) METTL3 protein expression was analyzed via Western blot, with representative blots and densitometric quantification shown. (E) Immunofluorescence staining for METTL3 (red) was performed; nuclei were stained with DAPI (blue). Representative images at 24 h are presented. Scale bar: 5 μm. Data represent the mean ± SD from three independent experiments (n = 3). Statistical significance was determined by two-way ANOVA followed by Šídák’s multiple comparisons test. Asterisks denote statistical significance (*P < 0.05, **P < 0.01, ***P < 0.001). continued. LPS Induces Specific m6A Hypermethylation of lncRNA-0949 Through METTL3 Having established LPS-driven METTL3 upregulation, we next asked whether this activity directly targets lncRNA-0949. MeRIP–qPCR using an anti-m6A antibody was performed on RNA isolated from HMO6 cells at 0, 6, 12, and 24 h after LPS treatment. LPS stimulation led to a time-dependent increase in m6A enrichment on lncRNA-0949. Compared to control cells at the same time point, m6A levels in LPS-treated cells were significantly elevated at 6 h (0.85 ± 0.14 vs. 0.25 ± 0.14, 3.4-fold, P < 0.001), 12 h (3.42 ± 0.14 vs. 0.25 ± 0.14, 13.5-fold, P < 0.001), and 24 h (5.37 ± 0.14 vs. 0.25 ± 0.14, 21.3-fold, P < 0.001; Figure 2A). A concomitant screen of m6A regulatory factors identified parallel upregulation of METTL3 mRNA (6.1-fold increase at 24 h vs. 0 h, Supplementary Figure 2A), pinpointing a coherent regulatory axis. Figure 2LPS induces METTL3-mediated m6A hypermethylation of lncRNA-0949. MeRIP-qPCR analysis of m6A enrichment on lncRNA-0949 in HMO6 cells at 0, 6, 12, and 24 h. (A) Time course of m6A levels in control and LPS-treated cells. (B) Effect of the m6A methyltransferase inhibitor 3-DAA on LPS-induced m6A modification. Cells were pretreated with 3-DAA prior to LPS stimulation. (C) Effect of METTL3 knockdown on LPS-induced m6A modification. Cells were transfected with control siRNA (siNC) or METTL3 siRNA (siMETTL3) and then left untreated or stimulated with LPS. Data are presented as mean ± SD from three independent biological replicates (n = 3). For (A), statistical significance was determined by two-way ANOVA followed by Šídák’s multiple comparisons test. For (B) and (C), statistical significance was determined by two-way ANOVA followed by Tukey’s post hoc test. Asterisks denote statistical significance (***P < 0.001).The images A, B and C depict grouped bar charts titled Me-RIP, illustrating Relative RNA expression (methylated lncRNA-0949) over time (0, 6, 12, 24 hours). The y-axis ranges from 0 to 10. Image A compares Control and LPS: Control remains around 1, while LPS increases from 1 at 0 hours to 4.5 at 6 hours, 6.8 at 12 hours and 7.2 at 24 hours, with significant differences indicated by brackets and three asterisks. Image B includes Control, LPS, Control plus 3-DAA and LPS plus 3-DAA: At 0 hours, all groups are about 1. At 6 hours, LPS rises to 4.5, while others stay at 1; similar patterns occur at 12 and 24 hours, with LPS reaching 4.8 and 8.5, respectively. Image C features Control, LPS, Control plus siMETTL3 and LPS plus siMETTL3: At 0 hours, all groups are about 1. At 6 hours, LPS is 4.5, while others remain at 1; at 12 and 24 hours, LPS is 4.8 and 7.8, respectively, with significant differences marked by brackets and three asterisks.A grouped bar chart set of three plots showing Me-RIP relative RNA expression over time. LPS induces METTL3-mediated m6A hypermethylation of lncRNA-0949. MeRIP-qPCR analysis of m6A enrichment on lncRNA-0949 in HMO6 cells at 0, 6, 12, and 24 h. (A) Time course of m6A levels in control and LPS-treated cells. (B) Effect of the m6A methyltransferase inhibitor 3-DAA on LPS-induced m6A modification. Cells were pretreated with 3-DAA prior to LPS stimulation. (C) Effect of METTL3 knockdown on LPS-induced m6A modification. Cells were transfected with control siRNA (siNC) or METTL3 siRNA (siMETTL3) and then left untreated or stimulated with LPS. Data are presented as mean ± SD from three independent biological replicates (n = 3). For (A), statistical significance was determined by two-way ANOVA followed by Šídák’s multiple comparisons test. For (B) and (C), statistical significance was determined by two-way ANOVA followed by Tukey’s post hoc test. Asterisks denote statistical significance (***P < 0.001). To establish METTL3 as the necessary writer, we inhibited its catalytic function pharmacologically with 3-deazaadenosine (3-DAA) or knocked down the transcript with specific siRNA. The efficiency of METTL3 knockdown was confirmed by the complete loss of LPS-induced m6A enrichment on lncRNA-0949 (Figure 2C), which serves as a functional readout of effective silencing. Both interventions significantly attenuated the LPS-induced m6A enrichment on lncRNA-0949 at 24 h (P < 0.001 vs. LPS alone for both 3-DAA and siMETTL3; Figure 2B and C), a reversal corroborated by dot blot analysis (Supplementary Figure 2B). Thus, in LPS-stimulated HMO6 cells, inflammatory signalling co-opts the METTL3 methyltransferase to selectively deposit m6A marks on lncRNA-0949. METTL3 Regulates the RNA Stability of lncRNA-0949 Through Its Specific m6A Site We first assessed whether METTL3-mediated m6A deposition affects lncRNA-0949 stability. Following actinomycin D treatment to block transcription, the decay of lncRNA-0949 was monitored over 0, 6, 12, and 24 h by qRT-PCR with GAPDH as an internal control (the Ct values of GAPDH remained stable across all time points, variation < 0.5 cycles). HMO6 cells were divided into four groups: control (siNC, untreated), LPS (siNC + LPS), control+siMETTL3 (siMETTL3, untreated), and LPS+siMETTL3 (siMETTL3 + LPS). METTL3 knockdown significantly accelerated the decay of lncRNA-0949 compared to the scrambled siRNA control. The estimated half-life of lncRNA-0949 was reduced from approximately 30 hours in control cells to 11 hours in METTL3-deficient cells (P < 0.001; Figure 3A). Figure 3METTL3 regulates the RNA stability of lncRNA-0949 through its specific m6A sites. HMO6 cells were transfected with control siRNA (siNC) or METTL3 siRNA (siMETTL3) and then left untreated or stimulated with LPS, generating four experimental groups: control (siNC, untreated), LPS (siNC + LPS), control+siMETTL3 (siMETTL3, untreated), and LPS+siMETTL3. (A) RNA stability assay. After LPS stimulation, actinomycin D (5 μg/mL) was added, and the relative residual levels of lncRNA-0949 were detected by qPCR at 0, 6, 12, and 24 h (the 0 h value was set as 100%). (B) Luciferase reporter assay. Cells were co-transfected with reporter vectors carrying wild-type (WT) or m6A site-mutated (Mut; GGACA→GGCCA) lncRNA-0949 3′UTR sequences, along with siNC or siMETTL3 as indicated. Luciferase activity was measured at 0, 6, 12, and 24 h post-treatment. Data are presented as mean ± SD from three independent biological replicates (n = 3). Statistical significance was determined by two-way ANOVA followed by Tukey’s post hoc test. Asterisks denote statistical significance (***P < 0.001).A showing a line graph with legend entries Control, LPS, Control plus siMETTL3 and LPS plus siMETTL3. The x-axis label is plus Actinomycin D Time (hours), with ticks at 0, 6, 12 and 24 hours. The y-axis label is Relative RNA expression lncRNA0949 versus GAPDH, ranging from 0.0 to 1.5. Approximate plotted means: at 0 hours all groups near 1.0. At 6 hours: Control about 0.85, LPS about 0.90, Control plus siMETTL3 about 0.80, LPS plus siMETTL3 about 0.78. At 12 hours: Control about 0.55, LPS about 0.75, Control plus siMETTL3 about 0.45, LPS plus siMETTL3 about 0.40. At 24 hours: Control about 0.30, LPS about 0.60, Control plus siMETTL3 about 0.22, LPS plus siMETTL3 about 0.18. Error bars are shown at each time point. B showing two circular plasmid maps labeled Reformed luciferase reporter WT. Left map includes labels CMV Promoter, Firefly Luciferase, Renilla Luciferase and lncRNA0949 WT. Right map includes labels CMV Promoter, Firefly Luciferase and lncRNA0949 Mut. Below B showing a grouped bar chart. The x-axis categories are lncRNA0949 WT and lncRNA0949 Mut. The y-axis label is Relative luciferase activity (normalized to control), ranging from 0 to 10. For lncRNA0949 WT: Control about 1, LPS about 8, Control plus siMETTL3 about 1, LPS plus siMETTL3 about 1. For lncRNA0949 Mut: all four groups about 1. Brackets with triple asterisks appear above multiple comparisons.A line graph and a bar graph showing lncRNA0949 stability and luciferase activity across treatments. METTL3 regulates the RNA stability of lncRNA-0949 through its specific m6A sites. HMO6 cells were transfected with control siRNA (siNC) or METTL3 siRNA (siMETTL3) and then left untreated or stimulated with LPS, generating four experimental groups: control (siNC, untreated), LPS (siNC + LPS), control+siMETTL3 (siMETTL3, untreated), and LPS+siMETTL3. (A) RNA stability assay. After LPS stimulation, actinomycin D (5 μg/mL) was added, and the relative residual levels of lncRNA-0949 were detected by qPCR at 0, 6, 12, and 24 h (the 0 h value was set as 100%). (B) Luciferase reporter assay. Cells were co-transfected with reporter vectors carrying wild-type (WT) or m6A site-mutated (Mut; GGACA→GGCCA) lncRNA-0949 3′UTR sequences, along with siNC or siMETTL3 as indicated. Luciferase activity was measured at 0, 6, 12, and 24 h post-treatment. Data are presented as mean ± SD from three independent biological replicates (n = 3). Statistical significance was determined by two-way ANOVA followed by Tukey’s post hoc test. Asterisks denote statistical significance (***P < 0.001). To pinpoint the responsible cis-element, we engineered luciferase reporters bearing the lncRNA-0949 3′UTR, either wild-type (WT) or with a mutated core RRACH motif (mut), using the pmirGLO dual-luciferase vector, which expresses both firefly and Renilla luciferases from a single construct. HMO6 cells were transfected with each reporter along with siNC or siMETTL3 as indicated. Firefly luciferase activity was normalized to Renilla activity to control for transfection efficiency, and then expressed as fold change relative to the untreated control group. LPS stimulation significantly increased the activity of the WT reporter (approximately 7.7-fold vs. control, P < 0.001), while this increase was completely abolished by METTL3 knockdown (P < 0.001; Figure 3B). In contrast, the Mut reporter showed no response to LPS or METTL3 depletion (all P > 0.05 vs. control; Figure 3B). These results indicate that, in this LPS-stimulated microglial model, the METTL3-dependent m6A modification at the 3′UTR is essential for the LPS-induced increase in lncRNA-0949 activity, establishing that METTL3 regulates lncRNA-0949 abundance primarily via this specific m6A site. The Pro-Inflammatory Function of lncRNA-0949 Strictly Depends on Its m6A Modification Status To determine the functional consequence of this epitranscriptomic mark, we tested the pro-inflammatory activity of lncRNA-0949 in its m6A-modified versus unmodified states. Cells were transfected with wild-type (OE-WT) or m6A site-mutated (OE-Mut) lncRNA-0949 overexpression vectors and then left untreated or stimulated with LPS. Overexpression of wild-type lncRNA-0949 amplified LPS-triggered release of pro-inflammatory cytokines at 24 h compared to the LPS alone group (eg, TNF-α increased by an additional 33%, P < 0.001; Figure 4A and B) and elevated oxidative stress (Supplementary Figure 3A). In contrast, overexpression of the m6A site mutant produced no such enhancement (P > 0.05 vs. LPS alone), demonstrating that the RNA’s pro-inflammatory function in this context requires its specific modification, not simply its overexpression (Figure 4A and B, Supplementary Figure 3A). Figure 4The pro-inflammatory function of lncRNA-0949 strictly depends on its m6A modification. (A and B) Effects of overexpressing wild-type or m6A site-mutated lncRNA-0949 on LPS-induced inflammatory responses at 0, 6, 12, and 24 h. Cells were transfected with wild-type (OE-WT) or m6A site-mutated (OE-mut) lncRNA-0949 overexpression vectors and then left untreated or stimulated with LPS. The following six groups are shown: untreated control (Control), OE-WT without LPS (Control+OE-lncRNA0949-WT), OE-mut without LPS (Control+OE-lncRNA0949-mut), LPS stimulation alone (LPS), OE-WT with LPS (LPS+OE-lncRNA0949-WT), and OE-mut with LPS (LPS+OE-lncRNA0949-mut). (A) mRNA expression of pro-inflammatory factors (MIP-2, IL-1β, TNF-α, and IL-6) measured by RT-qPCR. (B) Protein concentrations of corresponding cytokines in cell supernatants measured by ELISA. (C and D) Effects of METTL3 knockdown on LPS-induced inflammatory responses at 0, 6, 12, and 24 h. HMO6 cells were transfected with control siRNA (siNC) or METTL3 siRNA (siMETTL3) and then left untreated or stimulated with LPS, generating four experimental groups: control (siNC, untreated), LPS (siNC + LPS), control+siMETTL3 (siMETTL3, untreated), and LPS+siMETTL3 (siMETTL3 + LPS). (C) mRNA expression of pro-inflammatory factors (MIP-2, IL-1β, TNF-α, and IL-6) measured by RT-qPCR. (D) Protein concentrations of corresponding cytokines measured by ELISA. Data are presented as mean ± SD from three independent biological replicates (n = 3). Statistical significance was determined by two-way ANOVA followed by Tukey’s post hoc test. Asterisks denote statistical significance (*P < 0.05, **P < 0.01, ***P < 0.001).The image showing four grouped bar charts labeled A. Top-left chart: y-axis label Relative messenger ribonucleic acid expression MIP2 versus GAPDH, unit not shown; y-axis range 0 to 15. x-axis label Time (hours); x-axis values 0, 6, 12, 24. Legend entries: Control; LPS; Control plus OE lncRNA0949 WT; LPS plus OE lncRNA0949 WT; Control plus OE lncRNA0949 mut; LPS plus OE lncRNA0949 mut. Bars at 0 hours are near 0 to 1 for all groups. At 6, 12, 24 hours, LPS plus OE lncRNA0949 WT is highest (about 4, about 7, about 9 to 10). LPS plus OE lncRNA0949 mut is next (about 2, about 5, about 6). LPS is about 1, about 3 to 4, about 6 to 7. Control and both Control plus OE groups stay near 0 to 1. Multiple brackets show asterisk, two asterisks and three asterisks. Top-right chart: y-axis label Relative messenger ribonucleic acid expression TNF alpha versus GAPDH, unit not shown; y-axis range 0 to 15. x-axis label Time (hours); x-axis values 0, 6, 12, 24. Same legend. At 0 hours all groups near 0 to 1. At 6, 12, 24 hours, LPS plus OE lncRNA0949 WT is highest (about 4, about 6 to 7, about 10 to 11). LPS plus OE lncRNA0949 mut is about 2, about 4 to 5, about 6. LPS is about 2, about 4, about 7. Control and both Control plus OE groups near 0 to 1. Brackets with two asterisks and three asterisks. Bottom-left chart: y-axis label Relative messenger ribonucleic acid expression IL1 beta versus GAPDH, unit not shown; y-axis range 0 to 15. x-axis label Time (hours); x-axis values 0, 6, 12, 24. Same legend. At 0 hours all groups near 0 to 1. At 6, 12, 24 hours, LPS plus OE lncRNA0949 WT is highest (about 4, about 7, about 12 to 13). LPS plus OE lncRNA0949 mut is about 2, about 5, about 7. LPS is about 2, about 5, about 8. Control and both Control plus OE groups near 0 to 1. Brackets with two asterisks and three asterisks. Bottom-right chart: y-axis label Relative messenger ribonucleic acid expression IL6 versus GAPDH, unit not shown; y-axis range 0 to 10. x-axis label Time (hours); x-axis values 0, 6, 12, 24. Same legend. At 0 hours all groups near 0 to 1. At 6, 12, 24 hours, LPS plus OE lncRNA0949 WT is highest (about 3, about 5, about 9). LPS plus OE lncRNA0949 mut is about 2, about 4, about 5. LPS is about 2, about 3, about 5. Control and both Control plus OE groups near 0 to 1. Brackets with three asterisks.Four grouped bar charts showing relative messenger ribonucleic acid expression over time across treatments.Figure 4continued.A four-plot grouped bar chart showing cytokine concentrations over time across six treatment groups.A composite figure with four sub-images. The image A showing a grouped bar chart of MIP2 Concentration (pg/mL) versus Time (hours). X-axis ticks: 0, 6, 12, 24. Y-axis range: 0 to 800. Legend groups: Control; LPS; Control+OE-lncRNA0949-WT; LPS+OE-lncRNA0949-WT; Control+OE-lncRNA0949-mut; LPS+OE-lncRNA0949-mut. Bars at 0 hours are near baseline for all groups. At 6 hours, LPS is about 150, LPS+OE-lncRNA0949-WT about 250, LPS+OE-lncRNA0949-mut about 180; other groups near baseline. At 12 hours, LPS about 420, LPS+OE-lncRNA0949-WT about 560, LPS+OE-lncRNA0949-mut about 450; others near baseline. At 24 hours, LPS about 520, LPS+OE-lncRNA0949-WT about 600, LPS+OE-lncRNA0949-mut about 480; others near baseline. Multiple bracketed comparisons marked with asterisks, including single, double and triple asterisks. The image B showing a grouped bar chart of TNF-alpha Concentration (pg/mL) versus Time (hours). X-axis ticks: 0, 6, 12, 24. Y-axis range: 0 to 600. Same six legend groups. At 0 hours, all groups near baseline. At 6 hours, LPS about 200, LPS+OE-lncRNA0949-WT about 280, LPS+OE-lncRNA0949-mut about 160; others near baseline. At 12 hours, LPS about 380, LPS+OE-lncRNA0949-WT about 420, LPS+OE-lncRNA0949-mut about 320; others near baseline. At 24 hours, LPS about 500, LPS+OE-lncRNA0949-WT about 420, LPS+OE-lncRNA0949-mut about 300; others near baseline. Brackets with asterisks appear above timepoints. The image C showing a grouped bar chart of IL-1beta Concentration (pg/mL) versus Time (hours). X-axis ticks: 0, 6, 12, 24. Y-axis range: 0 to 800. Same six legend groups. At 0 hours, all groups near baseline. At 6 hours, LPS about 200, LPS+OE-lncRNA0949-WT about 280, LPS+OE-lncRNA0949-mut about 180; others near baseline. At 12 hours, LPS about 400, LPS+OE-lncRNA0949-WT about 480, LPS+OE-lncRNA0949-mut about 380; others near baseline. At 24 hours, LPS about 500, LPS+OE-lncRNA0949-WT about 650, LPS+OE-lncRNA0949-mut about 450; others near baseline. Bracketed significance annotations include single, double and triple asterisks. The image D showing a grouped bar chart of IL-6 Concentration (pg/mL) versus Time (hours). X-axis ticks: 0, 6, 12, 24. Y-axis range: 0 to 1500. Same six legend groups. At 0 hours, all groups near baseline. At 6 hours, LPS about 450, LPS+OE-lncRNA0949-WT about 520, LPS+OE-lncRNA0949-mut about 380; others near baseline. At 12 hours, LPS about 520, LPS+OE-lncRNA0949-WT about 900, LPS+OE-lncRNA0949-mut about 500; others near baseline. At 24 hours, LPS about 800, LPS+OE-lncRNA0949-WT about 1000, LPS+OE-lncRNA0949-mut about 850; others near baseline. Multiple brackets with asterisks appear above groups at 6, 12 and 24 hours.Figure 4continued.A set of 8 grouped bar charts of MIP2, TNF-alpha, IL-1beta, IL-6 under Control, LPS, plus siMETTL3.{“Error”: “The request requires exact plotted values for each group at each timepoint, but the bar heights and point values are not legible enough to extract reliably from the provided image. Previously stated numeric content is limited to axis labels, units and ranges, which are preserved below.”} Shared elements: Eight grouped bar charts labeled C and D. Legend groups: Control, LPS, Control plus siMETTL3, LPS plus siMETTL3. X-axis on all charts: Time (hours) with ticks 0, 6, 12, 24. C1: Y-axis label: Relative mRNA expression MIP2 versus GAPDH. Y-axis range 0 to 6. Pattern: LPS increases over time with highest bars at 12 and 24; LPS plus siMETTL3 is lower than LPS at 12 and 24; Control and Control plus siMETTL3 remain near baseline across time. Asterisks mark multiple comparisons. C2: Y-axis label: Relative mRNA expression TNF-alpha versus GAPDH. Y-axis range 0 to 8. Pattern: LPS rises strongly at 6, 12, 24 with highest at 24; LPS plus siMETTL3 is reduced versus LPS at later times; Control groups remain near baseline. Asterisks shown. C3: Y-axis label: Relative mRNA expression IL-1beta versus GAPDH. Y-axis range 0 to 6. Pattern: LPS peaks around 12 then remains elevated at 24; LPS plus siMETTL3 is lower than LPS at 12 and 24; Control groups stay low. Asterisks shown. C4: Y-axis label: Relative mRNA expression IL-6 versus GAPDH. Y-axis range 0 to 6. Pattern: LPS increases at 6 and peaks around 12 and 24; LPS plus siMETTL3 is lower than LPS at 12 and 24; Control groups remain low. Asterisks shown. D1: Y-axis label: MIP2 Concentration (pg/mL). Y-axis range 0 to 600. Pattern: LPS increases from 6 to 24 with highest near 24; LPS plus siMETTL3 is lower than LPS at 12 and 24; Control groups remain low. Asterisks shown. D2: Y-axis label: TNF-alpha Concentration (pg/mL). Y-axis range 0 to 400. Pattern: LPS increases at 6 and remains elevated at 12 and 24; LPS plus siMETTL3 is reduced versus LPS at later times; Control groups remain low. Asterisks shown. D3: Y-axis label: IL-1beta Concentration (pg/mL). Y-axis range 0 to 400. Pattern: LPS increases at 6 and peaks around 12 and 24; LPS plus siMETTL3 is lower than LPS at 12 and 24; Control groups remain low. Asterisks shown. D4: Y-axis label: IL-6 Concentration (pg/mL). Y-axis range 0 to 1500. Pattern: LPS increases strongly at 6 and continues higher at 12 and 24; LPS plus siMETTL3 is lower than LPS at 12 and 24; Control groups remain near baseline. Asterisks shown. The pro-inflammatory function of lncRNA-0949 strictly depends on its m6A modification. (A and B) Effects of overexpressing wild-type or m6A site-mutated lncRNA-0949 on LPS-induced inflammatory responses at 0, 6, 12, and 24 h. Cells were transfected with wild-type (OE-WT) or m6A site-mutated (OE-mut) lncRNA-0949 overexpression vectors and then left untreated or stimulated with LPS. The following six groups are shown: untreated control (Control), OE-WT without LPS (Control+OE-lncRNA0949-WT), OE-mut without LPS (Control+OE-lncRNA0949-mut), LPS stimulation alone (LPS), OE-WT with LPS (LPS+OE-lncRNA0949-WT), and OE-mut with LPS (LPS+OE-lncRNA0949-mut). (A) mRNA expression of pro-inflammatory factors (MIP-2, IL-1β, TNF-α, and IL-6) measured by RT-qPCR. (B) Protein concentrations of corresponding cytokines in cell supernatants measured by ELISA. (C and D) Effects of METTL3 knockdown on LPS-induced inflammatory responses at 0, 6, 12, and 24 h. HMO6 cells were transfected with control siRNA (siNC) or METTL3 siRNA (siMETTL3) and then left untreated or stimulated with LPS, generating four experimental groups: control (siNC, untreated), LPS (siNC + LPS), control+siMETTL3 (siMETTL3, untreated), and LPS+siMETTL3 (siMETTL3 + LPS). (C) mRNA expression of pro-inflammatory factors (MIP-2, IL-1β, TNF-α, and IL-6) measured by RT-qPCR. (D) Protein concentrations of corresponding cytokines measured by ELISA. Data are presented as mean ± SD from three independent biological replicates (n = 3). Statistical significance was determined by two-way ANOVA followed by Tukey’s post hoc test. Asterisks denote statistical significance (*P < 0.05, **P < 0.01, ***P < 0.001). continued. continued. To position lncRNA-0949 as a functional effector within the METTL3–m6A pathway, we intervened upstream. METTL3 knockdown significantly attenuated LPS-induced inflammatory responses at 24 h, with mRNA levels reduced by approximately 35–58% (Figure 4C) and protein levels reduced by approximately 30–63% (Figure 4D) compared to the LPS + siNC group (all P < 0.001) and associated oxidative stress (Supplementary Figure 3A). Collectively, these findings delineate a METTL3–m6A–lncRNA-0949 regulatory axis in LPS-stimulated HMO6 microglia: inflammatory signalling induces METTL3, which stabilizes lncRNA-0949 via site-specific m6A deposition. The m6A-modified transcript in turn contributes to a pro-inflammatory program, mirroring features of the dysregulated microglial activation associated with SAE. Discussion Our results define a pathogenic epitranscriptional circuit that sustains neuroinflammation in an in vitro model of sepsis-associated encephalopathy (SAE).24,25 Specifically, we establish that inflammatory signalling upregulates the m6A methyltransferase METTL3 in microglia. METTL3 then catalyses m6A deposition on the long non-coding RNA lncRNA-0949, thereby stabilizing it and driving a sustained pro-inflammatory response (Supplementary Figure 2C). This pathway—LPS → METTL3 → m6A–lncRNA-0949 → inflammation—forges a causal chain linking pathogen sensing, RNA modification, non-coding RNA stability and inflammatory phenotype (Supplementary Figure 2D–F),26–28 offering a novel molecular framework for understanding SAE pathology. Our data indicate that METTL3 is a key regulatory molecule in this pathway. In microglia, METTL3 expression increased in a time-dependent manner after LPS stimulation (Figure 1D and E), consistent with its role as an early response gene downstream of TLR4 signaling in immune cells.29,30 The functional importance of METTL3 was confirmed by gene knockout experiments, which showed significant reductions in inflammatory cytokine release and oxidative stress markers upon METTL3 depletion (Figure 4C and D, Supplementary Figure 3A). These findings position METTL3 as a molecular switch linking innate immune recognition to downstream inflammatory transcriptome reprogramming.27,31 Previous studies have focused more on the role of METTL3 in maintaining microglia homeostasis,14,32 while our work is the first to explicitly link its pathological hyperactivation to the neuroinflammatory mechanism of SAE, thus revealing its potential as a therapeutic target for intervention in neuroinflammatory conditions such as SAE. This study provides a mechanistic explanation for how lncRNA-0949 is regulated. It was found that lncRNA-0949 acts as a direct downstream target of METTL3. Specifically, METTL3 was shown to add an m6A modification at a particular site within the 3′UTR of lncRNA-0949 (Figure 2). This modification significantly increased the stability of the lncRNA-0949 transcript (Figure 3A). This stabilizing effect is in line with the known function of m6A “reader” proteins like YTHDF and IGF2BPs.19,33–35 Conclusive genetic evidence was obtained to confirm the functional requirement of this m6A mark. When wild-type lncRNA-0949 was overexpressed, inflammatory responses were intensified. In contrast, overexpression of a mutant form of lncRNA-0949, in which the m6A site was disrupted, produced no such effect (Figure 4A and B). Therefore, the m6A modification is not merely associated with the lncRNA’s function but is a direct cause of it. This chemical modification forms the essential basis for the pro-inflammatory activity of lncRNA-0949. In this way, a direct epitranscriptional regulatory mechanism is established.36,37 Notably, both METTL3 and the m6A reader IGF2BP2—which stabilizes modified RNAs—were co-upregulated under inflammatory stress (Supplementary Figure 2E).27,38 While direct evidence for IGF2BP2 binding to m6A-modified lncRNA-0949 is currently lacking, their co-upregulation is consistent with a potential functional link. Given their respective catalytic and stabilizing functions34,38–40 and the prevalence of feedback in signalling networks,41–43 we posit a testable model: that METTL3 and IGF2BP2 may form a positive feedback loop (METTL3 → m6A–lncRNA-0949 → IGF2BP2 stabilization → sustained inflammation → METTL3 upregulation). We emphasize that this speculative model requires direct experimental validation, which would provide a novel mechanistic explanation for the self-sustaining nature of pathological neuroinflammation. While this work defines a core epitranscriptional pathway linking METTL3 to lncRNA-0949 in microglial inflammation, several important questions remain. First, the precise mechanism by which m6A-modified lncRNA-0949 executes its pro-inflammatory program requires resolution. Although our data demonstrate that the m6A site is essential for lncRNA-0949 function, the molecular mode of action—whether it functions as a competing endogenous RNA (ceRNA) sequestering regulatory microRNAs (eg, miR-146a, miR-155),44–46 or as a scaffold assembling components of signalling hubs such as NF-κB or the NLRP3 inflammasome47,48—remains speculative. RNA pulldown coupled with mass spectrometry could identify its interacting protein partners, while luciferase reporters bearing predicted miRNA binding sites could test the ceRNA hypothesis. Second, while METTL3 knockdown and lncRNA-0949 overexpression experiments collectively implicate lncRNA-0949 as a downstream effector of METTL3, direct genetic manipulation—specifically, lncRNA-0949 knockdown and rescue with WT versus mutant constructs—would further solidify causality. Third, although our in vitro model establishes cell-autonomous function in microglia, the operational logic and necessity of this axis within the integrated neural circuitry of an intact brain—involving cross-talk with neurons and astrocytes—demands validation in SAE animal models.49,50 Finally, translation to human pathology will require profiling METTL3 activity and m6A-modified lncRNA-0949 in patient samples, where correlation with clinical severity could validate their pathogenic role and evaluate their potential as biomarkers or therapeutic guides.51–54 In summary, this work delineates a pathogenic epitranscriptional circuit—the METTL3–m6A–lncRNA-0949 axis—that drives and sustains neuroinflammation in an in vitro model of SAE, providing a mechanistic framework with potential relevance to the human condition. Conclusion Our research delineates a pathogenic mechanism in an in vitro model of sepsis-associated encephalopathy (SAE), in which inflammatory signalling upregulates the m6A methyltransferase METTL3 to catalyse site-specific modification of the long non-coding RNA lncRNA-0949, whose pro-inflammatory function we identify for the first time, thereby stabilizing it and propagating inflammation. By establishing a complete causal chain from environmental stimulus → epitranscriptomic writer → this functionally uncharacterized lncRNA → inflammatory phenotype, our findings bridge discrete levels of regulation and nominate METTL3-mediated m6A deposition as a potential therapeutic target for mitigating aberrant microglial activation in neuroinflammatory conditions such as SAE.