The CEBPB–AP-1 (JunB/Fos) Axis Drives Neuroinflammation and Microglial Dysfunction Via TNF Signaling in Ischemic Stroke Authors: Kun Liang, Shuangshuang Lu, Weihao Shi, Weijian Fan, Yijun Huang, Jianwei Chen, Lei Zhu Published: 2026 ABSTRACT Ischemic stroke triggers a strong neuroinflammatory response, with microglia playing dual roles in both exacerbating tissue damage and promoting repair. The molecular mechanisms regulating microglial activation and polarization remain inadequately defined. In this study, we demonstrate that the transcription factor CEBPB is a key upstream regulator of the AP-1 (JunB-Fos) complex, driving pro-inflammatory transcriptional programs in microglia following transient focal cerebral ischemia (tFCI). Transcriptomic profiling of microglia from tFCI mice revealed significant gene expression changes, particularly enrichment in inflammatory, immune, and cytokine production pathways, and identified Cebpb, Junb, Fos, and Tnf as key hub genes. Mechanistically, CEBPB is upregulated in post-ischemic microglia and directly binds to the Fos promoter to activate AP-1, inducing downstream inflammatory mediators, including IL-1β and TNF-α. Inhibition of AP-1 with the selective inhibitor T-5224 shifted microglia from a pro-inflammatory to an anti-inflammatory phenotype in vitro, decreasing pro-inflammatory cytokine production and enhancing anti-inflammatory mediator release. In vivo, T-5224 treatment in tFCI mice suppressed neuroinflammation, reduced neuronal apoptosis in the striatum and cortex, promoted a restorative microglial phenotype, and improved long-term sensorimotor and cognitive function. These findings establish the CEBPB/AP-1 axis as a critical driver of neuroinflammation in ischemic stroke and highlight it as a promising therapeutic target.Supplementary InformationThe online version contains supplementary material available at 10.1007/s10753-026-02525-3. Introduction Ischemic stroke is a leading cause of long-term disability and mortality worldwide, driven by a complex cascade of pathophysiological events that include excitotoxicity, oxidative stress, and a robust neuroinflammatory response [1, 2]. The post-ischemic inflammatory milieu is shaped by the rapid activation of resident brain immune cells, with microglia emerging as central orchestrators of both tissue damage and repair [3, 4]. Within minutes of ischemic insult, microglia undergo dramatic morphological and transcriptional changes, transitioning from a ramified, homeostatic state to an activated phenotype that migrates to the infarct core and penumbra [5]. This activation is not a binary switch but a spectrum of functional states, broadly categorized into pro-inflammatory (M1-like) and anti-inflammatory (M2-like) phenotypes [6, 7]. Pro-inflammatory microglia release cytokines such as TNF-α, IL-1β, and IL-6, exacerbating blood-brain barrier breakdown and neuronal apoptosis, while anti-inflammatory microglia produce IL-10 and TGF-β, promoting debris clearance, angiogenesis, and neurogenesis [8–10]. Despite the well-documented dual role of microglia in stroke, the molecular mechanisms that govern their phenotypic polarization and functional plasticity remain incompletely understood. CCAAT/enhancer-binding protein beta (CEBPB) is a transcription factor that plays a pivotal role in regulating immune and inflammatory gene expression [11]. In the central nervous system, CEBPB has been implicated in the pathogenesis of various neurodegenerative and neuroinflammatory disorders, including Alzheimer’s disease, multiple sclerosis, and traumatic brain injury [12, 13]. Recent studies have shown that CEBPB is upregulated in activated microglia and astrocytes following brain injury, where it drives the expression of pro-inflammatory mediators [14, 15]. However, the specific function of CEBPB in modulating microglial heterogeneity and the post-ischemic inflammatory response in stroke has not been fully elucidated. Activator protein-1 (AP-1) is a dimeric transcription factor complex, most commonly composed of Jun (e.g., JunB) and Fos (e.g., c-Fos) family members [16], that regulates gene expression in response to a wide range of extracellular stimuli, including cytokines, growth factors, and stress [17]. In the context of brain injury, AP-1 has been shown to be rapidly activated in neurons and glia, where it contributes to the regulation of cell survival, death, and inflammation [18, 19]. While AP-1 is known to be a downstream effector of multiple signaling pathways in the brain, its upstream regulators and its role in mediating microglial polarization after ischemic stroke remain poorly characterized. Here, we use a mouse model of transient focal cerebral ischemia (tFCI) to investigate the molecular mechanisms underlying microglial activation and polarization [20, 21]. We demonstrate that CEBPB acts as a critical upstream regulator of the AP-1 (JunB/c-Fos) complex, driving a pro-inflammatory transcriptional program in microglia after ischemia. Mechanistically, we show that CEBPB is upregulated in post-ischemic microglia and directly binds to the promoter of Fos to activate AP-1, thereby inducing the expression of downstream inflammatory mediators. Inhibition of AP-1 with the selective inhibitor T-5224 shifts microglia from a pro-inflammatory to an anti-inflammatory phenotype, reduces neuroinflammation, and improves functional recovery in tFCI mice. Our findings establish the CEBPB/AP-1 axis as a key regulator of microglial function in ischemic stroke and provide a promising therapeutic target for the treatment of this devastating disease. Methods Animals Adult male C57BL/6J mice (10–12 weeks of age, 22.7–26.8 g) were housed in a specific pathogen-free (SPF) environment with controlled temperature (22 ± 2 °C), humidity, and a 12-hour light/dark cycle. Standard chow and sterile water were provided ad libitum. Mice were randomly assigned to experimental groups using a computer-generated randomization protocol. Female mice were excluded due to the neuroprotective effects of ovarian hormones, particularly estrogen, which modulate neuroinflammation [22]. All procedures, including surgery, drug administration, behavioral assessments, and outcome quantification, were conducted by investigators blinded to the treatment groups to minimize bias. Animal experiments were approved by the Institutional Animal Care and Use Committee and conducted according to the NIH Guide for the Care and Use of Laboratory Animals. Transient Focal Cerebral Ischemia Model Transient focal cerebral ischemia (tFCI) was induced by the intraluminal filament method, resulting in middle cerebral artery occlusion. Mice were anesthetized with 5% isoflurane for induction and 2% for maintenance, administered in oxygen-air mixtures. A feedback-controlled heating pad maintained a constant body temperature of 37 °C to prevent hypothermia. A midline cervical incision exposed the common carotid artery (CCA), external carotid artery (ECA), and internal carotid artery (ICA). The silicone-coated monofilament (RWD Life Science, model MSMC21B120PK50) was inserted into the ICA and advanced until mild resistance was encountered. Regional cerebral blood flow (rCBF) was monitored in real-time using a laser Doppler flowmeter to confirm successful occlusion, defined as a reduction in rCBF by over 80% (Figure S1). After 60 min of ischemia, the filament was withdrawn to allow reperfusion. Sham-operated mice underwent identical procedures without filament insertion. Postoperative survival rates ranged from 80 to 90%, and only mice with stable vital signs were included in experiments. Cell Culture and Oxygen-Glucose Deprivation/Reoxygenation (OGD/R) Modeling BV2 microglial cells were cultured in high-glucose DMEM supplemented with 10% heat-inactivated fetal bovine serum (FBS) and 1% penicillin-streptomycin. Cells were maintained in a humidified incubator at 37 °C with 5% CO₂ and passaged every 2–3 days. For OGD/R, cells were washed with phosphate-buffered saline (PBS), incubated in glucose-free, serum-reduced DMEM (2% FBS), and placed in a hypoxia chamber (95% N₂, 5% CO₂) for 2 h. Reoxygenation was initiated by returning cells to normoxic conditions with complete culture medium for 24 h. Control cells were handled under normoxic conditions with glucose-containing medium. Magnetic Bead-Based Microglial Isolation Microglia were isolated from the brains of sham and tFCI mice using magnetic-activated cell sorting (MACS). Mice were anesthetized with isoflurane and perfused transcardially with ice-cold sterile PBS. The ipsilateral brain hemisphere was dissected and mechanically minced, followed by enzymatic digestion with collagenase IV and DNase I at 37 °C for 30 min. The tissue was passed through a 70-µm cell strainer, and myelin debris was removed using a 30% Percoll gradient. CD11b⁺ microglia were isolated using CD11b MicroBeads and MACS columns. The purity of isolated microglia was confirmed by immunostaining for Iba1 (> 95%) and trypan blue viability staining (≥ 90%). Drug Treatments In vitro, BV2 microglial cells and primary microglia were divided into control, OGD/R, and OGD/R + T-5224 groups. T-5224 (HY-12270, MedChemExpress), an AP-1 inhibitor, was dissolved in DMSO and added to the culture medium (10 µM) 1 h before OGD. For in vivo experiments, T-5224 was administered intraperitoneally at 30 mg/kg immediately after reperfusion and daily for 3 days. The dose and administration route were based on previous studies to ensure effective inhibition of AP-1 activity without toxicity [23–25]. siRNA-Mediated CEBPB Knockdown in Microglial Cells BV2 cells were seeded at 5 × 10⁵ cells/well in 6-well plates and transfected with siRNA targeting Cebpb (si-Cebpb) or non-targeting control (si-NC) usingLipofectamine 3000 (Invitrogen). After 48 h, transfection efficiency was verified by qPCR. Cells were then subjected to OGD/R or other functional assays. Chromatin Immunoprecipitation (ChIP)-qPCR ChIP-qPCR was used to verify CEBPB binding to the Fos promoter in BV2 microglial cells. Cells were cross-linked with 1% formaldehyde, lysed, and chromatin was sheared into 200–500 bp fragments. Immunoprecipitation was performed with anti-CEBPB antibody or normal rabbit IgG (control). The chromatin was washed, eluted, and reverse-cross-linked at 65 °C overnight. DNA was purified, and qPCR was performed using primers for the Fos promoter region. Enrichment of CEBPB at the Fos promoter was calculated as the fold change compared to the IgG control, normalized to the input. Western Blotting Microglial cells were lysed in RIPA buffer with protease and phosphatase inhibitors. Protein concentration was determined by BCA assay, and 30 µg of protein was separated by SDS-PAGE and transferred to PVDF membranes. Membranes were blocked and incubated overnight with primary antibodies against iNOS (CST, 13120, 1:1000), COX‑2 (CST, 12282, 1:1000), TNF‑α (CST, 3707, 1:1000), and β‑actin (loading control). After washing, membranes were incubated with HRP-conjugated secondary antibodies and visualized using chemiluminescence. Densitometric analysis was performed using ImageJ, with target protein expression normalized to β-actin. Perfusion, Tissue Processing, and Cryosectioning Mice were perfused transcardially with ice-cold PBS followed by 4% paraformaldehyde (PFA) for fixation. Brains were post-fixed in 4% PFA at 4 °C for 24 h, cryoprotected in sucrose solutions (10%, 20%, 30%), embedded in OCT compound, and stored at −80 °C. Coronal brain Sect. (25 μm) were cut using a cryostat and stored in antifreeze solution at −20 °C. Immunofluorescence on Brain Cryosections Brain cryosections were thawed at room temperature, washed three times with PBS to remove antifreeze solution, and permeabilized with 0.3% Triton X-100 in PBS for 15 min. Non‑specific binding was blocked with 5% normal goat serum in PBS for 1 h at room temperature. Sections were then incubated overnight at 4 °C with primary antibodies against Iba1 (Abcam, AB178846, 1:1000), iNOS (CST, 13120, 1:500), CD16 (BD Biosciences, 553142, 1:300), Arg1 (Santa Cruz, SC‑271430, 1:300), CEBPB (Abcam, ab32358, 1:500), AP‑1 (Proteintech, 10486‑1‑AP, 1:500), and NeuN (Abcam, ab177487, 1:1000). After three washes with PBS, sections were incubated with species‑matched Alexa Fluor‑conjugated secondary antibodies (Invitrogen) for 1 h at room temperature in the dark. Nuclei were counterstained with DAPI for 5 min. Sections were mounted with anti‑fade mounting medium (Thermo Fisher Scientific) and imaged using a slide scanner or confocal microscope with identical acquisition settings across all groups to ensure quantitative consistency. Image quantification was performed using ImageJ by investigators blinded to treatment groups, with predefined thresholds for positive staining and region‑of‑interest (ROI) analysis focused on the cortex and striatum. TUNEL Staining Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining was used to detect apoptotic cells in frozen brain sections, combined with NeuN immunofluorescence to identify neurons. The One-Step TUNEL Apoptosis Detection Kit (green fluorescence) from Beyotime Biotechnology (Cat. No. C1086) was used. Brain sections were fixed with immunostaining fixative, washed twice with PBS to remove residual fixative, and permeabilized with PBS containing 0.5% Triton X-100. Non-specific binding sites were blocked with blocking solution, and sections were incubated overnight at 4 °C with primary anti-NeuN antibody (Abcam, Cat. No. ab177487) diluted 1:1000. After washing with PBS, sections were incubated with fluorescence-conjugated secondary antibody at room temperature, washed again with PBS, and incubated with TUNEL detection solution at 37 °C for 60 min in the dark, following the kit instructions. After three washes with PBS, sections were mounted with anti-fluorescence quenching mounting medium. TUNEL-NeuN double-positive cells (apoptotic neurons) were counted and statistically analyzed to assess neuronal apoptosis. Primary Microglia Isolation and Culture Primary microglia (pMG) were isolated from neonatal C57BL/6J mouse pups (P0–P3). Brains were dissected on ice, and the meninges were removed. The brain tissue was minced and digested with 0.25% trypsin-EDTA for 15 min at 37 °C. The resulting single-cell suspension was filtered through a 70-µm cell strainer. Cells were plated in T75 flasks with DMEM supplemented with 10% FBS, 1% penicillin-streptomycin, and 20 ng/mL recombinant mouse M-CSF (PeproTech) to promote microglial proliferation. Cultures were maintained at 37 °C with 5% CO₂, with media changes every 3 days. After 10–14 days, confluent mixed glial cultures were shaken mildly (200 rpm, 30 min, 37 °C) to detach microglia. Isolated microglia were collected by centrifugation (300 × g, 5 min), replated in 6-well or 24-well plates, and allowed to adhere overnight before experimental use. Microglial purity was confirmed by Iba1 immunostaining. Cell Immunofluorescence BV2 microglial cells or primary microglia were seeded on glass coverslips in 24-well plates and cultured to 70–80% confluence. After treatment (OGD/R, drug treatment, or siRNA transfection), cells were fixed with 4% PFA for 15 min, permeabilized with 0.3% Triton X-100 for 10 min, and blocked with 5% normal goat serum for 1 h. Cells were incubated overnight at 4 °C with primary antibodies against CD16, iNOS, Arg1, or Iba1. After washing with PBS, cells were incubated with fluorescent secondary antibodies for 1 h at room temperature, followed by DAPI staining for 5 min. Coverslips were mounted on slides with anti-fade mounting medium and imaged using fluorescence or confocal microscopy. Quantification included mean fluorescence intensity (MFI) and the percentage of marker-positive cells per field, analyzed using ImageJ software with consistent settings across all groups. RNA-Seq Library Preparation, Sequencing, and Analysis Total RNA was extracted from magnetically isolated microglia using TRIzol reagent (Invitrogen). RNA integrity was assessed using an Agilent 2100 Bioanalyzer, and only samples with RNA Integrity Number (RIN) ≥ 8.0 were used. RNA-seq libraries were prepared with the NEBNext Ultra II RNA Library Prep Kit (New England Biolabs), using poly(A) enrichment for mRNA. Paired-end sequencing (150 bp) was performed on an Illumina NovaSeq 6000. Raw reads were quality-controlled with FastQC, adapter sequences and low-quality reads (Phred score < 20) were trimmed using Trimmomatic, and clean reads were aligned to the mouse reference genome (mm10) using STAR. Gene expression counts were quantified with featureCounts, and differential expression was analyzed with DESeq2. Genes with an adjusted p-value < 0.05 and absolute log2(fold change) > 1 were considered differentially expressed (DEGs). Functional enrichment analyses, including Gene Ontology (GO) biological process, KEGG pathway, and Gene Set Enrichment Analysis (GSEA), were conducted to identify enriched biological processes and pathways. Quantitative Real-Time PCR (qPCR) Total RNA from cells or brain tissue was extracted using TRIzol reagent. cDNA was synthesized from 1 µg RNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). qPCR was performed with SYBR Green Master Mix (Applied Biosystems) on a StepOnePlus Real-Time PCR System. Gene-specific primers (Sangon Biotech) were used to amplify target genes. Reaction conditions were: 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. Gapdh was used as a housekeeping gene. Primer sequences are provided in Table S1. Relative gene expression was calculated using the 2^−ΔΔCt method. Enzyme-Linked Immunosorbent Assay (ELISA) Cytokine levels were measured in both cell culture supernatants and brain tissue lysates. For supernatants, microglia were treated as indicated, and the medium was collected, centrifuged at 1000 × g for 10 min, and stored at − 80 °C. For brain lysates, mice were sacrificed 3 days after tFCI; the ipsilateral hemisphere was homogenized in ice‑cold RIPA buffer containing protease inhibitors, centrifuged at 12,000 × g for 15 min at 4 °C, and the supernatant was collected. TNF‑α, IL‑1β, IL‑6, IL‑10, and TGF‑β in cell supernatants, as well as TNF‑α, IL‑1β, CCL2, CCL3, IL‑4, IL‑10, and IL‑13 in brain tissue lysates, were quantified using commercial ELISA kits (TNF‑α, CCL2, CCL3: Elabscience, Cat. Nos. E‑EL‑M3063, E‑EL‑M3001, E‑EL‑M0007; IL‑1β, IL‑6, IL‑10, IL‑4, IL‑13: mlbio, Cat. Nos. ml106733, ml063159, ml037873, ml064310, ml106729). Absorbance was read on a microplate reader (Bio‑Rad), and cytokine concentrations were calculated from standard curves. Behavioral Assessments Garcia score The Garcia score ranges from 0 to 18, with higher scores indicating better function. It evaluates six functional domains: spontaneous activity, limb symmetry, climbing ability, body proprioception, vibrissae response, and lateral turning. Each domain is scored on a 0–3 scale (0 = severe impairment, 3 = normal function). Mice were acclimated to the testing environment, and assessments were performed at predefined postoperative time points. Adhesive Removal Test and Rotarod Test For the adhesive removal test, a 3 mm × 3 mm adhesive tape was applied to the forepaw (contralateral to the infarct side). Latency to touch (time to first contact with the tape) and latency to remove (time to completely peel off the tape) were recorded, with a maximum cutoff time of 120 s. Each mouse was tested three times per session, and the average values were analyzed. In the rotarod test, mice were trained on an accelerating rotarod (4–40 rpm) over 5 min for 3 consecutive days before tFCI. Post-tFCI, mice were tested every 7 days up to 35 days, and the latency to fall was recorded. Each mouse was tested three times per session, and the average latency was used for analysis. Novel Object Recognition (NOR) Test The NOR test assessed recognition memory. Mice were habituated to an open-field arena (40 cm × 40 cm × 30 cm) for 10 min per day for 2 consecutive days. On the training day, two identical objects (Object A) were placed in the arena, and mice were allowed to explore for 10 min. After a 1-hour retention interval, one object was replaced with a novel object (Object B), and mice explored the arena for another 10 min. Exploration time for each object was recorded by a blinded observer using predefined criteria. Morris Water Maze (MWM) Test The MWM test evaluated spatial learning and memory in a circular pool (120 cm diameter, 50 cm depth) filled with water (22 ± 1 °C) and made opaque with non-toxic white paint. A hidden escape platform (10 cm diameter) was submerged 1 cm below the water surface in a fixed location (target quadrant). During the acquisition phase (days 29–33 post-tFCI), mice were trained to find the platform with four trials per day. Escape latency and swim paths were recorded using video tracking (EthoVision XT, Noldus). On day 34 post-tFCI, a probe trial was conducted with the platform removed. Mice swam freely for 60 s, and the time spent in the target quadrant, the number of platform crossings, and swimming speed were recorded. Statistical Analysis All statistical analyses were performed using GraphPad Prism 9. Data are presented as mean ± standard deviation (SD). Comparisons between two groups were made using an unpaired two-tailed Student’s t-test (normal distribution) or the Mann-Whitney U test (non-normal distribution). For multiple groups, one-way or two-way ANOVA was used, followed by Tukey’s post hoc test. RNA-seq data were adjusted for false discovery rate (FDR). A p-value < 0.05 was considered statistically significant. Results Transcriptomic Analysis Reveals Activation of Inflammatory Pathways in Microglia Following Transient Focal Cerebral Ischemia RNA sequencing was performed on microglia isolated from mice subjected to tFCI or sham surgery to assess the molecular changes induced by ischemic stroke (Fig. 1A; Figures S2−4). Principal component analysis (PCA) and hierarchical clustering displayed a clear distinction in gene expression profiles between the two groups, confirming the robustness of the experimental model and the quality of the RNA-seq data (Fig. 1B, C). Differential expression analysis of microglial transcriptomes from tFCI mice versus sham controls uncovered widespread transcriptional dysregulation (Fig. 1D and E). Volcano plot visualization revealed 1,148 significantly upregulated genes and 2,382 significantly downregulated genes in the tFCI group (|log₂FC| > 1, FDR < 0.05; Tables S2 and S3).Fig. 1Transcriptomic profiling of microglia after transient focal cerebral ischemia (tFCI) reveals dysregulation of inflammatory and immune-related pathways. (A) Schematic of the experimental workflow. (B) Principal component analysis (PCA) of RNA-seq data showing distinct separation between the sham and tFCI groups. (C) Hierarchical clustering of samples based on gene expression profiles, demonstrating clear separation between the three biological replicates of sham and tFCI groups. (D) Volcano plot of differentially expressed genes (DEGs) between tFCI and sham groups. Genes with significant upregulation (red), downregulation (blue), or no significant change (gray) are indicated. (E) Heatmap of DEGs, illustrating the distinct gene expression patterns between sham and tFCI groups. (F-I) Gene Set Enrichment Analysis (GSEA) showing significant positive enrichment of gene sets related to inflammatory and immune functions in tFCI microglia. (J-M) GSEA showing significant negative enrichment of gene sets in tFCI microglia Transcriptomic profiling of microglia after transient focal cerebral ischemia (tFCI) reveals dysregulation of inflammatory and immune-related pathways. (A) Schematic of the experimental workflow. (B) Principal component analysis (PCA) of RNA-seq data showing distinct separation between the sham and tFCI groups. (C) Hierarchical clustering of samples based on gene expression profiles, demonstrating clear separation between the three biological replicates of sham and tFCI groups. (D) Volcano plot of differentially expressed genes (DEGs) between tFCI and sham groups. Genes with significant upregulation (red), downregulation (blue), or no significant change (gray) are indicated. (E) Heatmap of DEGs, illustrating the distinct gene expression patterns between sham and tFCI groups. (F-I) Gene Set Enrichment Analysis (GSEA) showing significant positive enrichment of gene sets related to inflammatory and immune functions in tFCI microglia. (J-M) GSEA showing significant negative enrichment of gene sets in tFCI microglia Gene Set Enrichment Analysis (GSEA) showed that genes involved in inflammatory and immune-related processes were notably enriched in tFCI microglia. These pathways included the positive regulation of cytokine production, inflammatory response, inflammatory response to antigenic stimulus, and myeloid leukocyte-mediated immunity (Fig. 1F–I). In contrast, gene sets that were negatively enriched in tFCI microglia were associated with the IL-17 signaling pathway, RNA methyltransferase activity, positive regulation of amine transport, and BMP binding (Fig. 1J–M). These results indicate that microglia undergo significant transcriptional reprogramming following ischemia, with activation of pro-inflammatory and immune-related pathways and suppression of specific cellular processes. Functional Enrichment Analysis Highlights Key Pathways and Hub Genes in Post-Ischemic Microglia KEGG pathway and Gene Ontology (GO) term enrichment analyses were conducted to further explore the functional implications of the transcriptional alterations observed in tFCI microglia. KEGG analysis of upregulated genes revealed significant enrichment in pathways related to inflammation, immunity, and cell proliferation, including the TNF signaling pathway, phagosome, efferocytosis, apoptosis, DNA replication, and cell cycle. Conversely, GO analysis of downregulated genes revealed enrichment in processes linked to transcriptional regulation and nucleic acid metabolism (Fig. 2A-B; Tables S4 and S5).Fig. 2Functional enrichment and validation of key inflammatory and immune-related genes in microglia following tFCI. (A) KEGG pathway enrichment analysis of genes significantly upregulated in tFCI microglia compared to sham controls. The top 15 pathways are displayed. (B) GO biological process enrichment analysis of genes significantly downregulated in tFCI microglia compared to sham controls. The top 15 terms are displayed. (C-E) Heatmaps showing the expression of genes involved in the TNF signaling pathway, Phagosome, and Efferocytosis in sham and tFCI microglia. (F) Protein-protein interaction (PPI) network of key genes from the TNF signaling pathway, Phagosome, and Efferocytosis, highlighting the central hub genes Cebpb, Junb, Fos, Ptgs2, and Tnf. (G-J) Genome browser tracks (bigWig visualization) showing the transcriptional activation of Cebpb, Fos, Ptgs2, and Tnf in tFCI microglia compared to sham microglia. (K) Quantitative real-time PCR (qPCR) validation of selected hub genes in isolated microglia from sham and tFCI mice. N = 6 per group. Data are presented as mean ± SD. **p < 0.01, ***p < 0.001 Functional enrichment and validation of key inflammatory and immune-related genes in microglia following tFCI. (A) KEGG pathway enrichment analysis of genes significantly upregulated in tFCI microglia compared to sham controls. The top 15 pathways are displayed. (B) GO biological process enrichment analysis of genes significantly downregulated in tFCI microglia compared to sham controls. The top 15 terms are displayed. (C-E) Heatmaps showing the expression of genes involved in the TNF signaling pathway, Phagosome, and Efferocytosis in sham and tFCI microglia. (F) Protein-protein interaction (PPI) network of key genes from the TNF signaling pathway, Phagosome, and Efferocytosis, highlighting the central hub genes Cebpb, Junb, Fos, Ptgs2, and Tnf. (G-J) Genome browser tracks (bigWig visualization) showing the transcriptional activation of Cebpb, Fos, Ptgs2, and Tnf in tFCI microglia compared to sham microglia. (K) Quantitative real-time PCR (qPCR) validation of selected hub genes in isolated microglia from sham and tFCI mice. N = 6 per group. Data are presented as mean ± SD. **p < 0.01, ***p < 0.001 Focusing on the expression patterns of genes involved in key inflammatory pathways, heatmaps confirmed a robust upregulation of genes associated with the TNF signaling pathway, phagosome, and efferocytosis in tFCI microglia compared to sham controls (Fig. 2C–E). Protein-protein interaction (PPI) network analysis identified a central regulatory hub involving Cebpb, Junb, Fos, Ptgs2, and Tnf, suggesting that these genes may play pivotal roles in mediating the microglial response to ischemia (Fig. 2F). BigWig visualization of RNA-seq data further confirmed increased transcriptional activity of these hub genes, including Cebpb, Fos, Ptgs2, and Tnf, in tFCI microglia (Fig. 2G–J). Finally, qPCR validation in isolated microglia corroborated the significant upregulation of these inflammatory genes (Cebpb, Junb, Fos, Ptgs2, Il1b, Ccl2, and Tnf) following tFCI, reinforcing their role in the post-ischemic inflammatory response (Fig. 2K). CEBPB Upregulation Drives AP-1-Mediated Inflammation in Microglia After Ischemia To validate our transcriptomic findings and further investigate the regulatory relationship between CEBPB and AP-1 (Junb/Fos) in post-ischemic microglia, we performed both in vivo and in vitro experiments. In the striatum (STR) of tFCI mice at 3 days post-ischemia, Iba1 + microglia exhibited markedly elevated CEBPB and AP-1 expression compared with sham controls (Fig. 3A, C). Similarly, in primary mouse microglia (pMG) subjected to OGD/R, we observed a robust upregulation of CEBPB and AP-1, recapitulating the in vivo findings (Fig. 3B, D).Fig. 3CEBPB acts as an upstream regulator of AP-1 (Junb/Fos) to drive the inflammatory response in microglia after cerebral ischemia. (A, C) Representative immunofluorescence images and quantification of CEBPB and AP-1 co-labeled with Iba1 in the striatum (STR) of mice 3 days after sham operation or tFCI. N ≥ 20 per group. (B, D) Representative immunofluorescence images and quantification of CEBPB and AP-1 co-labeled with Iba1 in primary mouse microglia (pMG) under control conditions or after OGD/R. N ≥ 20 per group. (E) qPCR analysis of the mRNA expression of Cebpb, Junb, Fos, Ptgs2, Il1b, Tnf, and Il6 in microglia under control conditions, OGD/R + si-NC, or OGD/R + si-Cebpb. N = 5 per group. (F) Western blot analysis of iNOS, COX-2, and TNF-α protein levels in microglia under OGD/R + si-NC or OGD/R + si-Cebpb conditions. β-actin was used as a loading control. (G) Quantification of the protein expression levels. N = 3 per group. (H) ChIP-qPCR analysis of CEBPB binding to its target promoter in microglia under Control or OGD/R conditions. Data are presented as the percentage of input DNA recovered after immunoprecipitation with anti-CEBPB antibody or non-specific IgG control. The IgG control confirms the specificity of the CEBPB ChIP signal. n = 3 independent experiments. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001 CEBPB acts as an upstream regulator of AP-1 (Junb/Fos) to drive the inflammatory response in microglia after cerebral ischemia. (A, C) Representative immunofluorescence images and quantification of CEBPB and AP-1 co-labeled with Iba1 in the striatum (STR) of mice 3 days after sham operation or tFCI. N ≥ 20 per group. (B, D) Representative immunofluorescence images and quantification of CEBPB and AP-1 co-labeled with Iba1 in primary mouse microglia (pMG) under control conditions or after OGD/R. N ≥ 20 per group. (E) qPCR analysis of the mRNA expression of Cebpb, Junb, Fos, Ptgs2, Il1b, Tnf, and Il6 in microglia under control conditions, OGD/R + si-NC, or OGD/R + si-Cebpb. N = 5 per group. (F) Western blot analysis of iNOS, COX-2, and TNF-α protein levels in microglia under OGD/R + si-NC or OGD/R + si-Cebpb conditions. β-actin was used as a loading control. (G) Quantification of the protein expression levels. N = 3 per group. (H) ChIP-qPCR analysis of CEBPB binding to its target promoter in microglia under Control or OGD/R conditions. Data are presented as the percentage of input DNA recovered after immunoprecipitation with anti-CEBPB antibody or non-specific IgG control. The IgG control confirms the specificity of the CEBPB ChIP signal. n = 3 independent experiments. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001 To assess whether CEBPB functions upstream of AP-1, we knocked down Cebpb expression in microglia using siRNA prior to OGD/R. qPCR analysis showed that while OGD/R significantly induced the expression of Cebpb, Junb, Fos, and downstream inflammatory genes (Ptgs2, Il1b, Tnf, Il6), this induction was markedly attenuated by Cebpb knockdown (Fig. 3E; Figures S5). Notably, CEBPB exists in two functionally distinct isoforms (LAP and LIP), and our primers were specifically designed to amplify the pro-inflammatory LAP isoform. Consistent with the mRNA data, protein levels of pro-inflammatory mediators, including iNOS, COX-2, and TNF-α, were significantly reduced in Cebpb-deficient microglia after OGD/R (Fig. 3F, G). Mechanistically, chromatin immunoprecipitation (ChIP)-qPCR in OGD/R-treated microglia demonstrated that CEBPB directly binds to the promoter region of Fos (Fig. 3H). The binding efficiency was significantly higher in OGD/R microglia compared to controls, indicating that ischemia enhances CEBPB’s transcriptional activity on the Fos promoter. These results establish CEBPB as a critical upstream regulator of AP-1 activity, driving the pro-inflammatory transcriptional program in microglia following cerebral ischemia. Inhibition of the CEBPB/AP-1 Axis Mitigates Pro-Inflammatory Activation of Microglia After OGD/R Building on our findings that CEBPB drives AP-1-mediated inflammation, we investigated the functional consequences of AP-1 inhibition in microglia. Primary mouse microglia were subjected to OGD/R in the presence or absence of T-5224, a selective AP-1 inhibitor. Immunofluorescence staining revealed that compared with control conditions, OGD/R markedly increased the expression of the pro-inflammatory markers CD16 and iNOS. Treatment with the AP-1 inhibitor T-5224 reversed this microglial phenotype: relative to OGD/R alone, T-5224 reduced the proportion of CD16 + and iNOS+ cells and elevated the anti-inflammatory marker Arg1 (Fig. 4A, B, H; Figure S6A–B).Fig. 4Inhibition of AP-1 with T-5224 shifts microglia toward an anti-inflammatory phenotype and reduces pro-inflammatory cytokine production. (A, B) Representative immunofluorescence images of primary mouse microglia under control, OGD/R, or OGD/R + T-5224 (AP-1 inhibitor) conditions, co-labeled with Iba1 and iNOS (M1-like marker) or Arg1 (M2-like marker). (C–G) ELISA quantification of pro-inflammatory cytokines (TNF-α, IL-1β, IL-6) and anti-inflammatory cytokines (IL-10, TGF-β) in the culture supernatants of primary microglia under control, OGD/R, or OGD/R + T-5224 conditions. N = 4 per group. (H) Quantification of the percentage of iNOS + and Arg1 + cells among Iba1 + primary microglia under control, OGD/R, or OGD/R + T-5224 conditions. N = 4 per group. (I) qPCR analysis of Tnf, Il1b, and Il10 mRNA expression in BV2 microglial cells under control, OGD/R, or OGD/R + T-5224 conditions. N = 4 per group. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001 Inhibition of AP-1 with T-5224 shifts microglia toward an anti-inflammatory phenotype and reduces pro-inflammatory cytokine production. (A, B) Representative immunofluorescence images of primary mouse microglia under control, OGD/R, or OGD/R + T-5224 (AP-1 inhibitor) conditions, co-labeled with Iba1 and iNOS (M1-like marker) or Arg1 (M2-like marker). (C–G) ELISA quantification of pro-inflammatory cytokines (TNF-α, IL-1β, IL-6) and anti-inflammatory cytokines (IL-10, TGF-β) in the culture supernatants of primary microglia under control, OGD/R, or OGD/R + T-5224 conditions. N = 4 per group. (H) Quantification of the percentage of iNOS + and Arg1 + cells among Iba1 + primary microglia under control, OGD/R, or OGD/R + T-5224 conditions. N = 4 per group. (I) qPCR analysis of Tnf, Il1b, and Il10 mRNA expression in BV2 microglial cells under control, OGD/R, or OGD/R + T-5224 conditions. N = 4 per group. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001 Consistent with the phenotypic shift, OGD/R induced a robust release of pro-inflammatory cytokines (TNF-α, IL-1β, IL-6) into the culture supernatant, which was significantly attenuated by T-5224 treatment. Conversely, the production of anti-inflammatory cytokines (IL-10, TGF-β) was further enhanced by AP-1 inhibition (Fig. 4C–G). These findings were recapitulated in BV2 microglial cells, where T-5224 treatment significantly reduced OGD/R-induced upregulation of Tnf and Il1b mRNA, while further increasing Il10 expression (Fig. 4I). To confirm the functional role of TNF-α as a critical downstream effector of the CEBPB/AP-1 axis, rescue experiments were performed using a TNF-α neutralizing antibody in microglia under OGD/R conditions. Treatment with T-5224 significantly attenuated the upregulation of both Tnf-α and Il1b. Notably, the addition of TNF-α neutralizing antibody did not result in further reduction in cytokine expression, suggesting no additive effect (Figure S7). These data confirm that the anti-inflammatory action of T-5224 is primarily mediated through the inhibition of TNF-α production downstream of AP-1 activation. This solidifies TNF-α as a key mediator in the pro-inflammatory response triggered by the CEBPB/AP-1 axis in ischemic microglia. Collectively, these results demonstrate that the CEBPB/AP-1 axis is a critical regulator of microglial inflammatory polarization under ischemic stress, and its pharmacological inhibition effectively mitigates the pro-inflammatory response. AP-1 Inhibition Mitigates Post-Ischemic Neuroinflammation and Promotes a Restorative Microglial Phenotype In Vivo To extend our in vitro findings to a clinically relevant model the effects of AP-1 inhibition on neuroinflammation and microglial polarization were assessed in mice subjected to tFCI. Mice were treated with T-5224 or vehicle control immediately after reperfusion. A dose-response and delayed intervention study was conducted to determine the optimal therapeutic dose and time window for T-5224 treatment in vivo. First the anti-inflammatory efficacy of T-5224 was evaluated at three different doses (10 30 and 60 mg/kg). Treatment with T-5224 reduced the expression of pro-inflammatory cytokines TNF-α and IL-1β in ischemic tissue in a dose-dependent manner. The 30 mg/kg dose achieved maximal efficacy significantly lowering cytokine expression while no additional benefit was observed at the 60 mg/kg dose (Figure S8A-B). Based on these results the 30 mg/kg dose was selected for subsequent studies. Next the therapeutic time window was assessed by administering T-5224 (30 mg/kg) at 0 3 or 6 h after reperfusion.Functional recovery was evaluated 3 days post-ischemia using the adhesive-removal test. Mice treated at 0–3 h showed significant and comparable improvements in sensorimotor function, whereas those treated at 6 h exhibited significantly worse outcomes, though still showing improvement compared to vehicle controls (Figure S8C). These findings indicate that early intervention within the first 3 h post-reperfusion is critical for optimal therapeutic benefit. ELISA analysis of brain tissue lysates revealed a significant increase in pro-inflammatory cytokines (TNF-α, IL-1β, CCL2, CCL3) and a moderate increase in anti-inflammatory cytokines (IL-4, IL-10, IL-13) following tFCI. T-5224 treatment significantly reduced pro-inflammatory cytokine expression while enhancing anti-inflammatory cytokine production (Fig. 5A, B). Immunofluorescence analysis of the striatum and cortex showed that tFCI induced a shift in microglial polarization, characterized by an increase in pro-inflammatory CD16 + cells and a decrease in anti-inflammatory Arg1 + cells. Inhibition of AP-1 with T-5224 significantly reversed this shift, decreasing CD16 + microglia and increasing Arg1 + microglia in both brain regions (Fig. 5C–J). Notably, T-5224 treatment had no effect on microglial polarization in sham-operated mice, confirming its specificity to the post-ischemic environment.Fig. 5Inhibition of AP-1 with T-5224 suppresses neuroinflammation and shifts microglial polarization toward an anti-inflammatory phenotype in tFCI model. (A, B) ELISA quantification of pro-inflammatory cytokines (TNF-α, IL-1β, CCL2, CCL3) and anti-inflammatory cytokines (IL-4, IL-10, IL-13) in brain tissue lysates from mice subjected to tFCI, treated with either vehicle (Veh) or T-5224. N = 4 per group. (C, D, G, H) Representative immunofluorescence images of microglia in the (C, D) striatum (STR) and (G, H) cortex (CTX) of mice from the indicated treatment groups, co-labeled with Iba1 and CD16 or Arg1. (E, F, I, J) Quantification of the percentage of (E, I) CD16 + and (F, J) Arg1 + cells among Iba1 + microglia in the striatum and cortex, respectively. N = 6 per group. Data are presented as mean ± SD. ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001 Inhibition of AP-1 with T-5224 suppresses neuroinflammation and shifts microglial polarization toward an anti-inflammatory phenotype in tFCI model. (A, B) ELISA quantification of pro-inflammatory cytokines (TNF-α, IL-1β, CCL2, CCL3) and anti-inflammatory cytokines (IL-4, IL-10, IL-13) in brain tissue lysates from mice subjected to tFCI, treated with either vehicle (Veh) or T-5224. N = 4 per group. (C, D, G, H) Representative immunofluorescence images of microglia in the (C, D) striatum (STR) and (G, H) cortex (CTX) of mice from the indicated treatment groups, co-labeled with Iba1 and CD16 or Arg1. (E, F, I, J) Quantification of the percentage of (E, I) CD16 + and (F, J) Arg1 + cells among Iba1 + microglia in the striatum and cortex, respectively. N = 6 per group. Data are presented as mean ± SD. ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001 To assess the effect of T-5224 on myelin debris clearance, immunofluorescence staining for Iba1 and degraded myelin debris (dMBP) was performed in the striatum of tFCI mice. T-5224 treatment enhanced the engulfment of myelin debris by microglia, as evidenced by a higher percentage of Iba1+/dMBP+ cells (Figure S9). These results indicate that T-5224 not only shifts microglia toward an anti-inflammatory phenotype but also promotes the clearance of myelin debris, crucial for tissue repair after ischemic stroke. These findings confirm that the CEBPB/AP-1 axis is a critical driver of neuroinflammation in vivo, and its inhibition promotes a restorative microglial phenotype, thereby mitigating the detrimental effects of cerebral ischemia. AP-1 Inhibition Attenuates Post-Ischemic Neuronal Apoptosis via a Microglia-Dependent Paracrine Mechanism Having confirmed that pharmacological inhibition of AP-1 alleviates microglial pro-inflammatory activation and drives a reparative phenotypic switch, we further explored its in vivo impact on neuronal survival after tFCI. TUNEL and NeuN double immunofluorescence staining showed negligible neuronal apoptosis in sham mice with or without T-5224 intervention. By contrast, tFCI markedly elevated the number of TUNEL⁺NeuN⁺ apoptotic neurons in both the striatum and cortex of vehicle-treated mice. Notably, T-5224 administration significantly diminished neuronal apoptosis across both brain regions (Figs. 6A–E).Fig. 6T-5224-mediated AP-1 inhibition attenuates neuronal apoptosis in the striatum and cortex after tFCI and confers neuroprotection through a microglia-dependent mechanism in an OGD/R co-culture model. (A) Representative low-magnification image of a mouse brain coronal section at 3 days post-surgery, showing the cortex (CTX) and striatum (STR) regions analyzed for neuronal apoptosis. (B, C) Representative high-magnification immunofluorescence images of the striatum and cortex from mice in the indicated treatment groups, co-labeled with TUNEL and NeuN. (D, E) Quantification of neuronal apoptosis in the (D) striatum and (E) cortex, presented as the percentage of TUNEL+NeuN+ cells among total NeuN+ neurons (left) and the density of TUNEL+NeuN+ cells per mm² (right). (F-G) Flow cytometric detection and quantification of neuronal apoptosis in the OGD/R neuron-microglia co-culture system. N = 3–6 per group. Data are presented as mean ± SD. ns = not significant, ***p < 0.001 T-5224-mediated AP-1 inhibition attenuates neuronal apoptosis in the striatum and cortex after tFCI and confers neuroprotection through a microglia-dependent mechanism in an OGD/R co-culture model. (A) Representative low-magnification image of a mouse brain coronal section at 3 days post-surgery, showing the cortex (CTX) and striatum (STR) regions analyzed for neuronal apoptosis. (B, C) Representative high-magnification immunofluorescence images of the striatum and cortex from mice in the indicated treatment groups, co-labeled with TUNEL and NeuN. (D, E) Quantification of neuronal apoptosis in the (D) striatum and (E) cortex, presented as the percentage of TUNEL+NeuN+ cells among total NeuN+ neurons (left) and the density of TUNEL+NeuN+ cells per mm² (right). (F-G) Flow cytometric detection and quantification of neuronal apoptosis in the OGD/R neuron-microglia co-culture system. N = 3–6 per group. Data are presented as mean ± SD. ns = not significant, ***p < 0.001 To clarify whether the neuroprotective efficacy of T-5224 originates from direct neuronal protection or indirect microglial regulation, we constructed an OGD/R-induced neuron–microglia co-culture system (Figure S10). In pure neuronal cultures, OGD/R robustly triggered neuronal apoptosis, whereas T-5224 monotherapy failed to reverse this injury, indicating no direct protective effect on neurons. Of note, co-culture with OGD/R-activated microglia further aggravated neuronal apoptosis; however, T-5224 treatment remarkably rescued neurons from microglia-mediated apoptotic injury in the co-culture context (Figs. 6F–G). Collectively, these results demonstrate that T-5224-mediated neuroprotection is fully reliant on suppressing microglial pro-inflammatory activation, rather than acting directly on neurons. Our findings underscore that blockade of the CEBPB/AP-1 axis concurrently mitigates post-stroke neuroinflammation and reduces neuronal apoptosis, supporting this pathway as a promising therapeutic target for ischemic stroke intervention. AP-1 Inhibition Enhances Functional Recovery After tFCI by Improving Sensorimotor Function and Cognitive Performance To assess the translational relevance of our findings, we evaluated the effect of AP-1 inhibition on long-term functional outcomes in tFCI mice. Behavioral tests were performed to assess sensorimotor and cognitive recovery in animals treated with T-5224 or vehicle control. Sensorimotor function was evaluated using the Garcia score, adhesive-removal test, and rotarod test. tFCI + Veh mice exhibited severe neurological deficits, as indicated by lower Garcia scores, prolonged time to remove the adhesive patch, and reduced latency to fall from the rotarod. Treatment with T-5224 significantly improved all these measures, with recovery approaching that of sham-operated mice (Fig. 7A–D).Fig. 7Inhibition of AP-1 with T-5224 improves sensorimotor and cognitive recovery following tFCI in mice. (A) Garcia score assessment of neurological function over 35 days post-surgery. N = 10 per group. (B, C) Adhesive-removal test measuring (B) time to touch and (C) time to remove an adhesive sticker from the forepaw, indicating somatosensory and motor function. N = 10 per group. (D) Rotarod test measuring motor coordination and balance, presented as latency to fall from an accelerating rotarod. N = 10 per group. (E) Schematic of the novel object recognition (NOR) test. (F) NOR recognition index, calculated as the time spent exploring the novel object divided by the total exploration time. N = 10 per group. (G) Escape latency during the acquisition phase of the Morris water maze (MWM) test, measuring spatial learning. N = 10 per group. (H) Escape latency on the probe trial day of the MWM test. (I) Swimming speed during the MWM test, indicating no significant differences in motor ability between groups. (J) Representative swim paths during the MWM probe trial for each treatment group. (K) Time spent in the target quadrant during the MWM probe trial, indicating spatial memory retention. (L) Frequency of crossing the former platform location during the MWM probe trial. N = 10 per group. Data are presented as mean ± SD. ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001 Inhibition of AP-1 with T-5224 improves sensorimotor and cognitive recovery following tFCI in mice. (A) Garcia score assessment of neurological function over 35 days post-surgery. N = 10 per group. (B, C) Adhesive-removal test measuring (B) time to touch and (C) time to remove an adhesive sticker from the forepaw, indicating somatosensory and motor function. N = 10 per group. (D) Rotarod test measuring motor coordination and balance, presented as latency to fall from an accelerating rotarod. N = 10 per group. (E) Schematic of the novel object recognition (NOR) test. (F) NOR recognition index, calculated as the time spent exploring the novel object divided by the total exploration time. N = 10 per group. (G) Escape latency during the acquisition phase of the Morris water maze (MWM) test, measuring spatial learning. N = 10 per group. (H) Escape latency on the probe trial day of the MWM test. (I) Swimming speed during the MWM test, indicating no significant differences in motor ability between groups. (J) Representative swim paths during the MWM probe trial for each treatment group. (K) Time spent in the target quadrant during the MWM probe trial, indicating spatial memory retention. (L) Frequency of crossing the former platform location during the MWM probe trial. N = 10 per group. Data are presented as mean ± SD. ns = not significant, *p < 0.05, **p < 0.01, ***p < 0.001 Cognitive function was assessed using the novel object recognition (NOR) test and Morris water maze (MWM) test. tFCI + Veh mice showed significant impairment in recognition memory in the NOR test, while T-5224 treatment restored memory to near-sham levels (Fig. 7E, F). In the MWM test, tFCI + Veh mice displayed significantly longer escape latencies, indicating impaired spatial learning. T-5224 treatment significantly reduced these latencies (Fig. 7G, H). Importantly, swimming speed was unaffected across groups, ruling out motor impairments as a confounding factor (Fig. 7I). In the probe trial, tFCI + Veh mice spent less time in the target quadrant and crossed the platform location fewer times, indicating impaired spatial memory retention. T-5224 treatment significantly improved both measures of spatial memory (Fig. 7J–L). These results demonstrate that inhibition of the CEBPB/AP-1 axis not only reduces neuroinflammation and neuronal apoptosis but also promotes substantial improvements in sensorimotor and cognitive recovery after ischemic stroke. Discussion This study systematically investigates the molecular mechanisms underlying microglial activation and polarization in ischemic stroke, identifying the CEBPB/AP-1 (Junb-Fos) axis as a critical driver of post-ischemic neuroinflammation and a promising therapeutic target. Our key findings demonstrate that CEBPB is upregulated in microglia following cerebral ischemia and directly binds to the Fos promoter to activate the AP-1 complex. This activation induces the expression of pro-inflammatory mediators and promotes a pro-inflammatory microglial phenotype. Inhibition of AP-1 with the selective inhibitor T-5224 reverses microglial polarization toward an anti-inflammatory phenotype, suppresses neuroinflammation, reduces neuronal apoptosis, and improves long-term sensorimotor and cognitive recovery in tFCI mice. The novelty of this study lies in establishing a direct regulatory link between CEBPB and AP-1 in post-ischemic microglia, unveiling a new transcriptional cascade that modulates microglial function and neuroinflammatory responses, and translating these mechanistic insights into a potential therapeutic strategy for ischemic stroke. Transcription factors orchestrate cellular responses to pathological stimuli, and our study positions the CEBPB/AP-1 axis as a central regulator of microglial activation and neuroinflammation following ischemic stroke. We explore these findings within the broader context of transcription factor cooperativity, emphasizing the cell-type-specific roles of CEBPB and AP-1 in central nervous system (CNS) pathology and discussing the therapeutic implications of targeting this axis. Pioneer factors, such as CEBPB, modulate chromatin accessibility and establish competence for gene regulation, with their functional synergy with partner transcription factors being increasingly recognized as a core mechanism in cellular programming. A high-throughput ChIP-ISO assay revealed that FOXA1’s genomic binding is profoundly influenced by co-binding transcription factors like AP-1 and CEBPB, with FOXA1 and AP-1 exhibiting binding cooperativity in vitro [26]. This finding strengthens the potential for crosstalk between CEBPB and AP-1 in regulating transcriptional programs and provides a mechanistic basis for the direct regulatory link observed between these two factors in post-ischemic microglia. The concept of transcription factor cooperativity extends beyond pioneer factor relationships. CEBPB and AP-1 have been shown to act in concert across diverse biological contexts. For example, during early adipogenesis, CEBPB and AP-1 cooperatively bind to a “hotspot” region in the Pparg2 promoter, with TEAD4 reinforcing the occupancy of both factors—a critical interaction for inducing Pparg2, a master regulator of adipocyte differentiation. In a similar context, studies on the murine surfactant-associated protein B (Sftpb) gene have shown that JUN, a key AP-1 component, inhibits promoter activity by heterodimerizing with CEBPB, thereby tethering JUN to the Sftpb promoter via the CEBP consensus element [27, 28]. These findings confirm that CEBPB and AP-1 family members can form functional heterodimers to regulate transcription, providing a precedent for the regulatory relationship identified in microglia. The relevance of CEBPB-AP-1 cooperation extends to pathological conditions, including hypoxia and cancer. In hypoxia-induced retinopathy, GPR91 activation via ERK1/2 signaling upregulates both CEBPB and AP-1 (c-Fos), which directly bind to the VEGF promoter to induce its expression [29]. In glaucoma-related retinal ganglion cell death, endothelin-1 treatment upregulates both c-Jun and CEBPB, interacting to form a positive feed-forward loop promoting endothelin receptor expression and apoptosis [30, 31]. These studies underscore the biological plausibility of CEBPB-AP-1 cooperation and reinforce the relevance of our findings in ischemic microglia. However, the regulatory relationship between CEBPB and AP-1 in cerebral ischemia appears to be cell-type specific, as evidenced by contrasting reports in the literature. A study on neuronal apoptosis in cerebral ischemia/reperfusion (I/R) injury found reduced Jun expression in I/R-injured brains, with its restoration alleviating neuronal damage by activating SOCS1, which promotes CEBPB degradation [32]. While this suggests an inverse regulatory relationship between AP-1 and CEBPB in neurons, our study demonstrates a direct activating role of CEBPB on AP-1 in microglia. This divergence underscores the critical importance of cell-type-specific regulatory networks in ischemic stroke pathophysiology and highlights the need for therapeutic strategies that account for these distinct roles in various CNS cell populations. CEBPB plays a central role in microglial function, particularly in neuroinflammation and neurodegenerative disease. For instance, in Alzheimer’s disease (AD), the ubiquitin ligase COP1 suppresses neuroinflammation by degrading CEBPB in microglia, while COP1 deficiency leads to CEBPB accumulation, promoting pro-inflammatory gene expression and neurotoxicity, effects mitigated by partial Cebpb deletion [11]. This supports CEBPB’s role in pathogenic microglial activation, aligning with our findings in ischemic stroke. The essential nature of CEBPB for microglial identity was demonstrated by a study showing that CEBPB, among six factors, suffices to differentiate human iPSCs into functional microglia-like cells [33, 34]. In CNS injury models, CEBPB drives microglial pyroptosis and neuroinflammation; for example, a spinal cord injury study found that microglial CEBPB upregulation activates Fcgr1 transcription and promotes pyroptosis, with CEBPB knockdown reducing pro-inflammatory cytokines and improving neurological recovery [35]. These findings align with our discovery that CEBPB contributes to microglia-mediated neuronal damage. The therapeutic potential of targeting CEBPB in stroke is further supported by studies showing that reprogramming microglia to an anti-inflammatory phenotype by modulating CEBPB-regulated NLRP3 inflammasome activation improves recovery after stroke [36]. Additional studies have shown that inhibiting CEBPB suppresses microglial inflammation; for example, isoliquiritigenin alleviates neuropathic pain by inhibiting spinal microglial activation via the ERK/CEBPB pathway, with CEBPB knockdown also reducing microglial activation [37]. However, CEBPB function is context-dependent; in depression models, pentraxin 3 drives microglial M2 polarization via the CREB/CEBPB axis [38], contrasting with our findings and reflecting the pleiotropic roles of CEBPB in microglial polarization. Developmental programming may also influence CEBPB-driven inflammation, as postnatal LPS exposure attenuates adult cortical microglial pro-inflammatory gene responses, including CEBPB, in a region- and sex-dependent manner [39]. Furthermore, CEBPB’s subcellular localization determines its functional outcomes; a depression study found that circDYM inhibits LPS-induced microglial apoptosis by sequestering CEBPB in the cytoplasm, where nuclear CEBPB would otherwise downregulate ZC3H4 and promote autophagy and apoptosis [40]. AP-1 is similarly established as a key mediator of microglial inflammation in various CNS pathologies. Studies on the enhancer RNA AANCR, induced by AP-1 activation via ATM/ERK signaling, reinforce AP-1’s role in CNS inflammation [41], reinforcing AP-1’s role in CNS inflammation. The therapeutic potential of AP-1 inhibition is demonstrated by multiple studies: in Parkinson’s disease, pazopanib suppresses LPS-induced microglial activation and pro-inflammatory cytokine expression via the MEK4-JNK-AP-1 pathway, protecting dopaminergic neurons [42]; in radiation injury, the PPARδ agonist L-165,041 inhibits the PKCα/MEK1/2/ERK1/2/AP-1 pathway, reducing reactive oxygen species and pro-inflammatory cytokines [43]; sulforaphane reduces LPS-induced pro-inflammatory mediators via JNK/AP-1/NF-κB inhibition [44]; pristimerin inhibits LPS-induced pro-inflammatory cytokines and microglial migration by inhibiting the IRAK1/TRAF6/TAK1-MKK7/JNK/AP-1 pathway [45]; kalopanaxsaponin A inhibits LPS-induced JNK phosphorylation and AP-1 DNA-binding activity [46]; and fucoxanthin blocks the MAPKs/AP-1 pathway [44]. AP-1 also cooperates with other transcriptional regulators in microglia; TonEBP acts as a transcriptional co-factor for AP-1 and NF-κB, promoting LPS-induced microglial activation and neuronal death [47], and noradrenaline synergistically increases IL-1β production via β2 adrenergic receptor-mediated AP-1 activation, with AP-1/NF-κB cross-coupling amplifying this effect [48]. The role of AP-1 extends to microglial-astrocyte crosstalk; fluorocitrate and GIBH-130 suppress reactive astrocytes via p38 MAPK/AP-1 inhibition, reducing pro-inflammatory microglial polarization [49]. Studies also link AP-1 to specific microglial responses: peptidoglycan-induced microglial IL-6 expression is AP-1-dependent via the TLR2/JNK/c-Jun pathway [50], and S100B/RAGE signaling co-activates AP-1 and NF-κB to upregulate COX-2, IL-1β, and TNF-α [51]. Notably, the carbazole compound LCY-2-CHO inhibits LPS/IFNγ-induced NO production by abolishing both CEBP (including CEBPB) and AP-1 DNA-binding activity [52], directly linking these factors in microglial inflammation and providing a pharmacological precedent for our findings. While the roles of CEBPB and AP-1 in microglial inflammation are well-established, the direct regulatory relationship between these two factors in ischemic stroke has remained unclear. Our study addresses this gap by demonstrating that CEBPB acts as a direct upstream regulator of AP-1 in post-ischemic microglia. Additionally, we identify a bidirectional regulatory loop, wherein CEBPB/AP-1 activation induces the expression of TNF-α, which, in turn, amplifies AP-1 signaling. This self-perpetuating inflammatory mechanism can be interrupted by AP-1 inhibition, underscoring the central role of the CEBPB/AP-1-TNF axis in post-ischemic neuroinflammation. Given the established roles of CEBPB and AP-1 in a variety of CNS pathologies, targeting this axis holds promise for broader therapeutic applications in neurodegenerative and neuroinflammatory diseases. This model aligns with previous studies demonstrating that maximal expression of certain inflammatory genes requires concurrent activation of both CEBP and AP-1 family members [52], and that these factors can physically interact to form heterodimeric complexes with distinct DNA-binding specificities [28]. Given the pivotal roles of CEBPB and AP-1 in a range of CNS disorders, targeting this axis could have far-reaching applications in treating neurodegenerative and neuroinflammatory diseases. Our findings encourage further exploration of upstream signals that activate CEBPB following ischemic insults, as well as the specific CEBPB isoforms involved in microglial activation. Moreover, investigating the potential for combination therapies targeting both CEBPB and AP-1, or their downstream effectors, could enhance neuroprotection and lead to more effective treatments. Additionally, considering the cell-type-specific differences in CEBPB-AP-1 regulation—highlighted by comparing our microglial findings with neuronal studies [32, 53–55]—will be essential in developing targeted therapeutic strategies that account for these distinct cellular responses. In the search for novel therapeutic targets for cerebrovascular diseases, similar to how miR-322‑5p inhibition, ECHS1 restoration, or HACE exosome blockade each ameliorates specific neuropathologies [56–58], our study underscores the significance of converging mechanisms as actionable entry points for ischemic stroke therapy. The therapeutic relevance of targeting AP-1 in inflammatory conditions is further supported by studies in pain models, which provide insights into the mechanisms by which AP-1 inhibition exerts anti-inflammatory effects. For instance, intrathecal administration of T-5224 alleviates Freund’s complete adjuvant- or IL-6-mediated inflammatory pain, associated with decreased expression of c-Fos, GFAP, Iba-1, and IL-6 in the spinal cord, demonstrating that AP-1 functions as a critical node in integrating glial activation and inflammatory signaling [59]. The specificity and CNS penetration of T-5224 have been validated by studies showing that it blocks spexin-induced upregulation of downstream target genes, confirming its utility as a pharmacological tool for probing AP-1 function in vivo [60]. Additionally, T-5224-mediated AP-1 inhibition suppresses the CCL2/CCR2 chemokine axis, identifying a downstream mechanism linking AP-1 activation to immune cell recruitment, which provides additional targets for therapeutic intervention [61]. Despite the strengths of our study, several limitations should be acknowledged. First, our experiments were primarily conducted in male C57BL/6J mice, which may limit the generalizability of our findings to female mice or other animal models, given the well-documented sex differences in stroke pathophysiology and neuroinflammatory responses. Second, while our study focuses on the CEBPB/AP-1 axis in microglia, this transcriptional cascade may also be active in other CNS cell types (e.g., astrocytes, neurons), potentially influencing post-ischemic outcomes. Future studies should examine the cell-type-specific roles of CEBPB/AP-1 in ischemic stroke. Third, we used a single dose and administration route of T-5224, and further research is needed to optimize dosing, timing, and the route of administration to maximize therapeutic efficacy and minimize potential side effects. Finally, our study did not assess the long-term effects of CEBPB/AP-1 inhibition on neurogenesis, angiogenesis, and other repair processes critical for long-term functional recovery. Further investigations into these areas will be necessary to fully understand the therapeutic potential of targeting the CEBPB/AP-1 axis in ischemic stroke. Conclusions The CEBPB/AP-1 (Junb-Fos) axis is a critical transcriptional regulator of microglial activation and polarization in ischemic stroke. CEBPB upregulation in post-ischemic microglia activates AP-1 by binding to the Fos promoter, driving pro-inflammatory responses and neuronal damage. Selective AP-1 inhibition with T-5224 reverses microglial pro-inflammatory polarization, suppresses neuroinflammation, and enhances long-term functional recovery. These findings position the CEBPB/AP-1 axis as a promising therapeutic target for ischemic stroke, advancing our understanding of microglial plasticity and neuroinflammation in cerebral ischemia. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 (PDF 1.36 MB) Supplementary Material 1 (PDF 1.36 MB) Supplementary Material 2 (PDF 998 KB) Supplementary Material 2 (PDF 998 KB) Supplementary Material 3 (XLSX 9.93 KB) Supplementary Material 3 (XLSX 9.93 KB) Supplementary Material 4 (XLSX 75.6 KB) Supplementary Material 4 (XLSX 75.6 KB) Supplementary Material 5 (XLSX 151 KB) Supplementary Material 5 (XLSX 151 KB) Supplementary Material 6 (XLSX 21.0 KB) Supplementary Material 6 (XLSX 21.0 KB) Supplementary Material 7 (XLSX 83.6 KB) Supplementary Material 7 (XLSX 83.6 KB)