Activated microglia and astrocytes form a bidirectional regulatory loop to modulate neuroinflammation and glial scar formation. Astrocytes promote an anti-inflammatory milieu by secreting costimulatory molecules such as CD200 and interacting with microglia to activate anti-inflammatory cytokines (90). In preclinical models, C5aR1 on microglia integrates local and peripheral C5a signals, triggering a cascade of neuroinflammation amplification, neurotoxic astrocyte polarization and neutrophil recruitment that culminates in cerebral edema (51). In ICH, the C3-C3aR signaling axis, formed by astrocytic C3 and microglial C3aR, mediates the inhibitory effect of A1 astrocytes on the phagocytosis of myelin debris (91). The release of IL-15 from astrocytes triggers a shift in microglia toward a pro-inflammatory phenotype, creating a feedback loop that exacerbates neuroinflammation following ICH (92). TRPA1 in astrocytes regulates neuroinflammation and neurological deficits post-ICH, by suppressing the MAPK/NF-κB signaling pathway, driving a shift toward the neuroprotective A2 astrocytic phenotype and enhancing the proliferation of phagocytic microglia (93). Bidirectional regulation between microglia and astrocytes serves as a key mechanism modulating the inflammatory response after ICH, primarily through influencing cytokine profiles and cellular phenotypic transitions. The crosstalk between microglia and astrocytes represents a focal point and cutting-edge area in immunology research, with their bidirectional modulation of phenotypic and functional states warranting particular attention in the field of ICH.

Microglia serve as a bridge linking central and peripheral immunity after ICH. IL-10 modulates both microglial phagocytic function and the infiltration of monocyte-derived macrophages following ICH, with CD36 serving as a key phagocytic effector downstream of IL-10 signaling (94). Activin-A, derived from M2 microglia/macrophages, plays a crucial role in promoting oligodendrocyte differentiation and thereby fostering myelin regeneration, highlighting its potential as a therapeutic target for white matter injury following cerebral hemorrhage (95). In PHE, crosstalk between microglia-derived osteopontin and CD44-positive cells plays a critical role in regulating the local immune milieu, thereby influencing the pathological progression (96). Extracellular vesicles secreted by bone marrow-derived macrophages have the potential to boost microglial phagocytosis in the aftermath of ICH (97). Microglia act as an immune bridge, regulating phagocytic function and repair processes through specific molecules and responding to external signals such as extracellular vesicles. However, current findings largely remain as isolated mechanisms, and the mechanisms by which they integrate into a coordinated spatiotemporal network within the dynamically evolving ICH milieu remains unclear.

The communication between microglia and endothelial cells constitutes a critical axis within the neurovascular unit, fundamentally influencing the pathophysiology of ICH. This bidirectional crosstalk regulates central processes including BBB integrity, neuroinflammation and vascular repair, making it a pivotal focus for therapeutic intervention. Their functional interplay is facilitated by spatial proximity and involves both contact-dependent and soluble mediators. Activated microglia can directly influence BBB permeability through the release of inflammatory factors and proteases interacting with endothelial cells (67). In the context of ICH, endothelial cells lining the choroid plexus are prone to pyroptosis at the 24-h time point (98). ICH-induced pyroptosis in endothelial cells is accompanied by the release of pro-inflammatory cytokines IL-6 and IL-1β, which engage with microglia to activate the NF-κB signaling pathway and ultimately lead to increased transcriptional expression of Lcn2 and Msr1 (98). This initiates a cycle of endothelial damage activating microglia, whose response further exacerbates vascular dysfunction and inflammation. Hence, therapeutic strategies aimed at the microglia-endothelial axis, such as inhibiting key inflammatory pathways or protecting endothelial health, offer a promising avenue to alleviate SBI post-ICH.

Following ICH, the anti-inflammatory functions of microglia are mediated by diverse signaling axes that form complex networks intertwined with multiple biological processes. Elucidating these signaling pathways and their molecular foundations is critical for identifying promising therapeutic targets to mitigate neuropathological deficits in ICH. Among them, targeting intrinsic regulatory axes is particularly promising. This approach directly modulates the cell-autonomous mechanisms of microglia, including phagocytosis, polarization and inflammatory signaling, by acting on their intrinsic signaling pathways. Due to this cell-specific mode of action, it offers high target specificity with a potentially favorable systemic side-effect profile, representing the most promising direction for current clinical translation efforts in post-ICH neuroinflammation.

Studies have demonstrated that rosiglitazone-mediated activation of PPAR-γ protects against BBB damage and ameliorates hemorrhage transformation, potentially by promoting anti-inflammatory polarization of microglia (104,105). Phagocytosis is critical for enhancing hematoma resolution, promoting healing after ICH injury. Zhao et al (106) found that PPAR-γ activators significantly upregulated PPAR-γ-regulated genes such as CD36 catalase, while downregulating pro-inflammatory genes and alleviating neuronal injury by enhancing microglial phagocytosis, whereas PPAR-γ gene knockdown or anti-CD36 antibody treatment significantly impaired this phagocytic function following ICH. PPAR-γ activation also enhances the phagocytic capacity of anti-inflammatory microglia via CD36 (52). Building on preclinical research, PPAR-γ activators have entered clinical applications. Simvastatin-activated PPAR-γ improves microglia-induced erythrocyte phagocytosis and exhibits neuroprotective effects in patients with ICH (107,108). These findings highlight PPAR-γ as a pivotal regulator of microglial phagocytosis and inflammation, with therapeutic activators offering promise for clinical translation in ICH.

Adenosine monophosphate-activated protein kinase (AMPK) regulates the balance between pro- and anti-inflammatory microglial subclasses by acting as a central sensor of brain injury (109,110). Adiponectin receptor 1 (AdipoR1), constitutively expressed in microglia, is activated by endogenous C1q/TNF-related protein 9 (CTRP9), which exhibits increased expression within the first 24 h after ICH. CTRP9 treatment upregulates AdipoR1 and phosphorylated (p-) AMPK protein levels while reducing inflammatory cytokines, thereby decreasing neuroinflammation by modulating the AdipoR1/AMPK/NF-κB signaling pathway (111). Following ICH, CTRP9-activated AdipoR1 further mitigates neuropathological damage and BBB dysfunction by engaging APPL1/AMPK/Nrf2 signaling, highlighting CTRP9 as a promising therapeutic agent for improving BBB integrity (112,113). Similarly, activation of melanocortin receptor 4 alleviates neuropathological deficits by regulating the AMPK signaling pathway (114). Future research should focus on developing brain-specific AMPK activators to enhance translational feasibility.

Inhibition of glycogen synthase kinase-3β (GSK-3β) confers neuroprotective effects in in vivo experiments (115-117). By enhancing microglial phagocytosis and promoting M2-phenotype microglial differentiation, GSK-3β inhibition significantly improves hematoma resolution and cognitive deficits in rats with ICH (118,119). Additionally, lithium chloride treatment downregulated GSK-3β activity, reducing mature oligodendrocytes death and upregulating brain-derived neurotrophic factor expression (120). Thus, GSK-3β is classified as a medium-priority target, requiring time-window stratified studies and dose-optimization research to confirm its clinical applicability.

As aforementioned, regulatory T lymphocytes suppress the microglia-induced inflammatory response and improve neurological deficits by inhibiting NF-κB activation via the JNK/ERK pathway (38,101,121). Hyperbaric oxygen therapy has been shown to reduce the expression levels of pro-inflammatory cytokines and p-JNK, proposing a link between JNK signaling and downregulation of M1 microglial immune activity (122,123). Current clinical management of ICH faces notable limitations, including concerns about treatment efficacy, surgical trauma, complications and drug side effects, with few standardized interventions available. Hyperbaric oxygen therapy (HOT) emerges as a promising alternative approach, although the specific mechanisms by which HOT modulates ICH pathophysiology require further investigation and validation. These findings highlight the need to explore non-pharmacological, pathway-targeted strategies such as HOT to address unmet clinical needs in ICH treatment.

Intrinsic regulatory targets with high druggability and existing clinical repurposing potential should be prioritized for subsequent translational studies. AMPK and GSK-3β, while mechanistically promising, require specificity enhancement and clinical feasibility validation to advance toward clinical application.

Intercepting detrimental extracellular signals targeting microglial pattern recognition receptors (PRRs) is a key upstream intervention strategy. This approach aims to block the initial 'trigger' of microglial activation, thereby inhibiting downstream inflammatory cascades. However, PRRs are often widely expressed in myeloid cells, leading to potential systemic immune suppression, which poses a major translational issue for this class of targets.

Toll-like receptor 4 (TLR4), a central player in innate immune responses, has been recognized as a key pattern recognition receptor (124). TLR4 deficiency attenuated perihematomal inflammation by reducing the infiltration of pro-inflammatory microglia in in vivo ICH experiments (125,126). By contrast, TLR4 activation impairs microglial phagocytosis of erythrocytes, delaying CD36-mediated hematoma resolution and exacerbating neurological deficits in ICH (127,128). Additionally, TLR4-driven microglial autophagy has been linked to exacerbated neuroinflammation in ICH mouse models (129). Collectively, these studies highlight the dual role of TLR4 in modulating both inflammatory recruitment and hematoma clearance after ICH. Targeting TLR4 signaling represents a promising therapeutic strategy to balance anti-inflammatory effects with enhanced phagocytic function, positioning it as a potential candidate for future ICH interventions.

DNA damage initiates the innate immune response, with cGAS, a key DNA sensor, detecting disease-associated damaged DNA to activate its downstream effector, STING. This cascade results in the upregulation of type I interferon production and phosphorylation of interferon regulatory factor 3 (130). cGAS acts as a critical regulator of inflammation and autophagy, with upregulated cGAS in striatal brain damage mediating inflammation via increased expression of pro-inflammatory genes (Ccl5 and Cxcl10) and activating autophagy through the major initiators LC3A and LC3B (131). In ischemic stroke models, tPA therapy enhances neutrophil extracellular trap expression in plasma and brain tissue. DNase I treatment or peptidyl-arginine deiminase 4 deficiency, both of which inhibit neutrophil extracellular trap formation, reverse tPA-induced cGAS upregulation. By contrast, in tPA-treated mice, cGAMP (a cGAS-derived second messenger) suppresses DNase I-induced anti-hemorrhagic effects by dampening STING and type I interferon signaling (132). Jiang et al (133) further demonstrated that cGAS knockdown furthers M2 microglial polarization and attenuates the inflammatory response by disrupting the cGAS-STING axis in stroke mice. Shi et al (134) extended these findings, demonstrating that a cGAS inhibitor delivered via immunosuppressive nanoparticles reduces microglial inflammation and enhances anti-inflammatory polarization in post-stroke rats. IronQ pretreatment confers mesenchymal stem cells (MSCs) with a synergistic effect to alleviate neuroinflammation and improve neurological function, achieved by inhibiting the cGAS-STING signaling pathway (135).

C-type lectin-like receptors, a family of transmembrane PRRs predominantly expressed in myeloid cells, play a critical role in innate immunity (136). Dysregulation of C-type lectin-like receptors following tissue injury can drive excessive production of inflammatory mediators and exacerbate inflammatory progression. Among C-type lectin-like receptors, microglial macrophage-inducible C-type lectin (Mincle), widely expressed on macrophages, binds nuclear spliceosome-associated protein 130 released from necrotic cells, potently enhancing the inflammatory response (137,138). In alcohol-induced liver injury models, Mincle activation via its downstream effector spleen tyrosine kinase (Syk) upregulates the expression levels of pro-inflammatory genes (139), while in Crohn's disease, the Mincle/Syk axis exacerbates intestinal mucosal inflammation through macrophage pyroptosis (140). Inhibiting this pathway attenuates inflammatory responses, highlighting its therapeutic potential. Preclinical studies across multiple neurological disorders have demonstrated that blocking the Mincle/Syk signaling axis confers neuroprotection (141). For example, interventions including the Syk inhibitor BAY61-3606, acupuncture and MSC transplantation suppress Mincle/Syk-mediated microglial activation, reducing neuroinflammation in in vivo stroke models (138,141-145). Collectively, these findings establish the Mincle/Syk pathway as a promising therapeutic target for ICH, offering a rationale for developing novel immunomodulatory strategies to mitigate post-ICH neuroinflammation and promote neurological recovery.

To date, autophagy exhibits dualistic functions, making it challenging to assess whether its effects after ICH are predominantly harmful or beneficial. Excessive autophagy exacerbates endoplasmic reticulum stress (ERS)-induced brain injury as early as 6 h following ICH. By contrast, at day 7 post-ICH in rats, autophagy strengthens ERS protective functions by clearing cellular debris (146). Tan et al (147) demonstrated that enhanced autophagy attenuates oxidative stress injury following ICH by upregulating antioxidant protein expression, while autophagy inhibitors reverse this neuroprotective effect. Previous studies further indicate that autophagy positively regulates the inflammatory response in the context of ICH (148,149).

The levels of ILs during ICH progression are modulated by microglial functions. Xu et al (150) demonstrated that IL-4-loaded nanoparticles activating the IL-4/STAT6 axis promoted hematoma resolution and functional recovery in mice with ICH. By contrast, IL-15, a pro-inflammatory cytokine, regulates homeostasis and microglial immunoreactivity following CNS inflammatory events (92). Yu et al (151) showed that an IL-17A-neutralizing antibody attenuates microglial activation and blocks ICH-mediated pro-inflammatory cytokines. Shi et al (152) further revealed that IL-17A enhances microglial autophagy and neuroinflammation. Administration of an IL-17A-neutralizing antibody significantly reduced brain edema and improved neurological outcomes in mice with ICH, while genetic inhibition of autophagy-related genes ATG5 and ATG7 suppressed microglial autophagy and inflammation (153). Intraventricular injection of IL-33 alleviated white matter and neuronal injury by promoting microglial M2 polarization following ICH (154). Additionally, inhibiting the NF-κB signaling pathway through diverse interventions, including microRNAs (miRNAs or miRs), GATA-binding protein 4 and traditional Chinese medicines, can reduce neuroinflammation in ICH therapy (155-159). These findings highlight the bidirectional role of ILs and NF-κB in regulating microglial phenotypes and underscore their therapeutic potential in modulating neuroinflammation after ICH.

Intercepting extracellular signals via PRRs represents a rational upstream therapeutic strategy, yet its translational potential is constrained by insufficient target specificity and limited validation in ICH-specific models. Therefore, future research should focus on developing microglia-specific PRR modulators and conduct in-depth validation in clinically relevant ICH models. In summary, although extensive preclinical studies have made notable progress at the mechanistic level, successfully translating these findings into clinical applications requires further systematic investigation of promising interventions (Fig. 2).

Exosomes and miRNAs represent emerging precision therapeutic tools for modulating microglia after ICH, offering advantages such as low immunogenicity, high target specificity and the ability to regulate multiple pathways simultaneously, thereby effectively addressing the spatiotemporal complexity of microglial dynamics post-ICH. To translate these promising tools into precise therapies, a rational framework is required. This involves identifying key molecular targets within microglial dynamic continuum, categorizing them based on their pathological function, and employing engineered exosomes for spatiotemporally controlled delivery. Current evidence points to three strategic categories of targets: (i) Inflammatory initiators (for example, TLR4), the suppression of which quenches upstream danger signaling; (ii) polarization nodes (for example, CKIP-1 and C/EBP-α), the modulation of which actively reprogrammes microglia toward a reparative phenotype; and (iii) Cellular stress regulators (for example, mTOR), the intervention of which maintains microglial homeostasis. The convergence of exosome biology and miRNA targetology offers a unique path to address these targets in a combined and specific manner.

Given their minimal risks of immunogenicity and tumorigenicity, exosomes secreted by human umbilical cord mesenchymal stem cells (hUCMSC-exo) represent a promising alternative to cell-based therapies. Evidence indicates that hUCMSC-exo administration can alleviate astrocyte inflammation and mitochondrial damage associated with ICH by inhibiting the TLR4/NF-κB signaling pathway (160). Exosomal PTEN-induced kinase 1 from hUCMSCs attenuates neurological deficits, dysregulated microglial M1/M2 polarization and inflammatory responses after ICH in mice (161). Exosomes derived from human adipose-derived mesenchymal stem cells (hADSCs-Exo) were successfully internalized by microglia cells and exerted robust anti-inflammatory actions by suppressing the exerted robust anti-inflammatory actions of inflammatory factors and promoting M1 to M2 transition (162). Exosomes, which are nanoscale vesicles measuring 30-150 nm in diameter, encapsulate a diverse cargo encompassing mRNA, miRNA, proteins and growth factors (163,164). Beyond serving as delivery vehicles, the specific contents of exosomes, especially miRNAs, are pivotal for their immunomodulatory effects. Consequently, identifying key miRNAs and elucidating their mechanisms in microglial polarization has become a major research frontier in developing RNA-based therapeutics for ICH.

Increasing genetic and epigenetic evidence highlights the critical role of miRNAs in modulating gene expression and microglial polarization following ICH (165). For example, let-7a modulates microglial M2 polarization by targeting CKIP-1 at day 3 post-ICH in mice. In the aforementioned study, let-7a overexpression reduced CKIP-1 protein levels, promoted M2 polarization characterized by increased expression of IL-10 and Arg-1, and alleviated neuroinflammation. By contrast, let-7a inhibition upregulated CKIP-1, driving M1 polarization (165). Similarly, miR-7 overexpression suppresses TLR4 protein expression, attenuating microglial inflammation in both ICH rats and lipoprotein-stimulated microglial inflammation models (166). Further research has demonstrated that targeting TLR4 through the inhibition of the Prx1/TLR4/NF-κB signaling axis at day 3 post-ICH provides neuroprotection against ICH-induced brain injury, presenting a promising anti-neuroinflammatory strategy (155). In stroke in vivo and in vitro experiments, miR-182-5p and miR-27a modulate inflammation by targeting TLR4, and their overexpression downregulates TLR4 protein levels while reducing the release of pro-inflammatory factors (167,168).

Previous evidence identified that integrin subunit β8, a direct target of miR-222, can be negatively regulated by miR-222 to alleviate microglial inflammation and apoptosis (169). Previously, miR-132 was shown to inhibit the cholinergic inflammatory response by targeting acetylcholinesterase. Zhang et al (170) reported that lentivirus-mediated miR-132 overexpression in the right caudate nucleus 14 days before autologous blood-induced ICH suppressed pro-inflammatory microglial activation, alleviated BBB dysfunction and reduced neuronal reduction at day 3 post-ICH. miRNAs also act as key mediators of autophagy-induced microglial inflammation, post-transcriptionally suppressing gene expression and function (171). It has been demonstrated that miR-144 enhances hemoglobin-mediated microglial autophagic inflammation by directly targeting the 3' untranslated regions (UTRs) of mTOR (171,172). Similarly, miR-124 promotes microglial M2 polarization and alleviates inflammation by targeting the 3'-UTR of C/EBP-α. In ICH mice, miR-124 mimic administration significantly reduced neurological deficits, brain water content (BWC) and C/EBP-α expression at day 3, compared with miR-124 inhibitor treatment. Consistent negative regulation of C/EBP-α by miR-124 was also observed in erythrocyte lysate-stimulated microglia transduced with miR-124 mimics or inhibitors (173).

Exosomes and miRNAs represent more than just a list of promising entities; they form the core of an emerging precision immunomodulation paradigm for ICH. Moving beyond the generic suppression of neuroinflammation, this paradigm is defined by the rational selection of targets across the microglial functional spectrum (for example, TLR4, CKIP-1 and mTOR) and their precise manipulation via engineered exosomes loaded with specific miRNA combinations. The primary translational challenge no longer lies solely in target identification, but in optimizing these intelligent delivery systems to ensure their stability, targeting efficiency, and controlled cargo release, and in validating their synergistic efficacy in advanced disease models. This approach heralds a shift from broad anti-inflammatory strategies toward context-sensitive immune remodeling tailored to the spatiotemporal dynamics of ICH.

Modulating downstream effector molecules targets the terminal pathological processes of microglial activation, aiming to block the final 'damage cascade' rather than regulating microglial function directly. This strategy is characterized by clear mechanisms of action but may lack specificity due to the widespread involvement of effector molecules in normal physiological processes.

Matrix metalloproteinases (MMPs), a universal superfamily of structurally related zinc-dependent endopeptidases, are upregulated following ICH and contribute to extracellular matrix degradation. The role of MMPs in ECM disruption driven by inflammation serves as a pathological basis for stroke, as evidenced in numerous studies (174). Wells et al (175) investigated MMP function in ICH mice and found that MMP-12 was the most significantly elevated isoform, with experiments revealing that MMP-12 knockout mice exhibited substantial forelimb motor recovery compared with WT mice, while WT mice showed increased recruitment of Iba1⁺ cells with macrophage morphology to the injury site, indicating that MMP-12 exacerbates SBI after ICH (175). Subsequent studies demonstrated that stem cell therapy or minocycline administration following ICH reduced perihematomal MMP-12 expression and attenuated microglia-infiltrated inflammatory responses (176,177), positioning MMP-induced microglial activation as a potential therapeutic target. For example, minocycline promotes the microglial phenotypic shift from pro-inflammatory M1 to anti-inflammatory M2 (178). However, while minocycline therapy effectively mitigates the early elevation of MMP-12 and TNF-α, its therapeutic efficacy diminishes by 1-week post-ICH. Additional studies have revealed that stem cell therapy notably reduces microglial infiltration and MMP-12 expression in perihematomal regions after ICH (177,179).

Increasing evidence indicates that ferrous iron released from hemolysis represents a key pathogenic factor in hematoma following ICH (180). Iron toxicity-mediated microglial activation and pro-inflammatory response are major contributors to brain injury after ICH. Deferoxamine, an iron chelator capable of crossing the BBB to bind iron, has been identified to moderately modulate microglial activation following ICH when administered intraperitoneally to decrease iron accumulation (181-183). As a microglial activation inhibitor, minocycline can also reduce iron levels to prevent neuronal death after ICH (184,185). Similarly, VK-28, a brain-permeable iron chelator, promotes microglial M2 phenotype polarization, reduces BWC, alleviates white matter injury and improves neurobehavioral deficits after ICH (186,187). Observational studies highlight the complexity of regulatory systems governing microglial function, underscoring the critical need to decipher phenotypic and genotypic variations to develop promising therapeutic strategies for ICH.

Modulating downstream effector molecules offers mechanism-specific intervention for post-ICH injury. Iron chelation is a high-priority strategy due to its ICH-specific mechanisms and existing clinical drug availability, while MMP-12 is limited by narrow intervention windows and poor specificity. For downstream targets, defining precise intervention timelines (aligned with pathological progression) is critical for balancing therapeutic efficacy and safety. Therefore, advancing these downstream strategies into the clinic critically depends on translational research that rigorously defines their optimal therapeutic windows in complex, time-course models (Fig. 3).

The core academic contribution of this review lies in the establishment of an innovative conceptual framework for the hierarchical targeted regulation of microglial neuroinflammation, which systematically categorizes regulatory strategies into four dimensions: Targeting the intrinsic regulatory axes of microglia, intercepting detrimental extracellular signals, implementing precise interventions with exosomes and miRNAs, and regulating downstream effector molecules, thus achieving a logical upgrade from mechanistic description to translation-oriented research. Its unique research perspective is reflected in taking translational feasibility, targeting specificity and pathological dynamic adaptability as the core criteria to clarify the regulatory priority and clinical translational potential of each signaling axis; it combines with the time dependence of pathological progression in ICH to distinguish the intervention windows and applicable scenarios of different regulatory strategies, and meanwhile focuses on the key challenges in translational medicine to analyze the clinical translational bottlenecks and optimization directions of various strategies. In summary, the present review systematically integrates the latest regulatory mechanisms of microglial neuroinflammation following ICH.