Responsive nanoparticles modulating microglia lactate transport alleviate M1-type polarization and neuroinflammation for brain injury therapy

Abstract

Abstract

Neuroinflammation mediated by microglia and excessive oxidative stress are key pathological processes driving the progression of brain injuries such as intracerebral hemorrhage (ICH) and traumatic brain injury (TBI). Modulating lactate transport has emerged as a promising approach to mitigate M1 microglia polarization and alleviate neuroinflammation. Liposomal nanoparticles provide a safe and efficient platform for drug delivery. Here, we developed reactive oxygen species (ROS) responsive MiRCM nanoparticles that co-deliver monocarboxylate transporter 1 (MCT1) inhibitor AR-C155858 and a ROS-scavenging PPS core, modified with CAQK and MG1 peptides for brain injury site and microglia targeting. MiRCM nanoparticles demonstrated ROS-induced degradation and controlled drug release, effectively protecting AR-C155858 from enzymatic degradation. In vitro and in vivo experiments exhibited that MiRCM nanoparticles selectively accumulated at injured brain regions and in M1-type microglia, where they scavenged ROS, inhibited lactate efflux, suppressed M1 microglia polarization and reduced inflammatory cytokine production in ICH and TBI models. Consequently, MiRCM treatment protected neurons, reduced astrocyte activation, preserved blood-brain barrier integrity, enhanced endogenous antioxidant enzyme activities, and alleviated neurological deficits in both ICH and TBI models. RNA sequencing further confirmed downregulation of inflammatory pathways. Moreover, biosafety evaluations revealed no significant histopathological or biochemical abnormalities in major organs, indicating good biocompatibility. In summary, this study offers a new MiRCM nanoparticle that effectively modulates lactate transport and eliminates ROS to suppress microglia M1 polarization and neuroinflammation, ultimately enhancing neural protection and functional recovery after ICH and TBI.

Graphical Abstract

Schematic of MiRCM nanoparticles for ICH and TBI treatment

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04355-9.

Introduction

Neuroinflammation is a critical contributor to the initiation and progression of brain injury-associated disorders. Within the central nervous system (CNS), microglia act as the primary immune cells, orchestrating inflammatory responses and neuronal recovery under pathological conditions such as intracerebral hemorrhage (ICH), traumatic brain injury (TBI), ischemic stroke, Alzheimer’s disease, and Parkinson’s disease [1–5]. Following brain injury, microglia rapidly respond to harmful stimuli, including pathogens and damage-associated molecular patterns (DAMPs), then engulf apoptotic cells, myelin fragments, and neurotoxic proteins such as amyloid-β [6, 7]. During this process, they exhibit dynamic morphological and functional alterations, transforming between pro-inflammatory (M1-like) and anti-inflammatory (M2-like) states, though recent transcriptomic evidence suggests more diverse profiles [8]. The M1 phenotype is characterized by the release of inflammatory factors such as inducible nitric oxide synthase (iNOS), reactive oxygen species (ROS), interleukin-1 (IL-1), interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α), which together amplify neuroinflammatory cascades and contribute to neuronal injury. Conversely, M2 microglia secrete anti-inflammatory cytokines, including interleukin-10 (IL-10) and arginase-1 (Arg-1), which support neuronal survival and facilitate tissue repair [9]. Besides, recent studies have also revealed that microglia function exhibits a distinct temporal pattern following brain injury. Previous studies have demonstrated that transient depletion of microglia during the first three days after TBI, followed by controlled microglial repopulation, improves both short- and long-term cognitive outcomes, through promoting IL-6 expression in hippocampal granule cells. In contrast, prolonged microglial depletion fails to provide the benefit, indicating that the temporal dynamics of microglial activation are critical for neurological recovery [10]. In the chronic phase, persistent microglia-mediated inflammation critically impairs neural restoration [11, 12]. In summary, microglia play a pivotal role in post-injury pathophysiological processes, including neuroinflammation, phenotypic polarization, neuronal death, cellular activation, blood-brain barrier (BBB) integrity, and neurogenesis after brain injury [1, 13, 14]. Currently, effective therapeutic options for brain injury remain limited. Therefore, developing a newly precise and targeted strategy to modulate microglial activity is essential for optimizing functional recovery following brain injury.

Lactate metabolism plays an essential role in maintaining normal cellular function by supporting energy production, regulating metabolism, and influencing protein modification [15–17]. As the final product of anaerobic glycolysis, lactate is produced from pyruvate by lactate dehydrogenase [18]. Hypoxia producing excessive lactate limits the entry of pyruvate into the mitochondrial tricarboxylic acid (TCA) cycle, reducing ATP synthesis and increasing ROS generation, ultimately exacerbating neuroinflammation [19–22]. In addition, inflammatory signaling also promotes nuclear translocation of the transcription factor hypoxia-inducible factor-1α (HIF-1α), which binds to glycolysis-related genes and induces “Warburg effect”, further enhancing lactate production [23]. Lactate modulates inflammation in a time-dependent manner: in the early phase, elevated lactate reflects enhanced anaerobic metabolism, contributing to acidosis and energy failure [18, 23]. Subsequently, accumulated lactate can feedback-inhibit glycolysis, promote pyruvate oxidation through the TCA cycle, restore energy homeostasis, initiate lactylation of Pyruvate kinase M2, and then promote macrophage polarization into the M2 type [18, 23, 24]. The elevated lactate/GPR81 pathway also inhibits either the AMPK/NF-κB or ARRB2/NLRP3 signaling pathway to promote resolution of the inflammatory response [18, 23]. Monocarboxylate transporters (MCTs), which mediate bidirectional lactate transport across cell membranes, are key regulators of the concentration of intracellular lactate [25]. Microglia express MCT 1 and MCT4, which mainly regulate the lactate exportation and importation respectively [26, 27]. Silencing MCT1 in microglia leads to intracellular lactate accumulation, which negative-feedback inhibits anaerobic metabolism and expression of Hif-1α to enhance mitochondrial oxidative phosphorylation and promote energy production [27]. This metabolic shift attenuates M1 polarization and alleviates neuroinflammation [27]. Therefore, targeting MCT1 to modulate microglia metabolic reprogramming represents a promising therapeutic strategy for mitigating inflammation and improving outcomes after brain injury.

Liposomal nanoparticles are spherical vesicles composed of one or more phospholipid layers that can encapsulate drugs, making them one of the most versatile nanocarriers in biomedical applications [28]. Owing to their biocompatibility, biodegradability, and structural similarity to biological membranes, liposomes offer a safe and efficient platform for drug delivery [29]. They can protect encapsulated drugs from enzymatic degradation, prolong circulation time in vivo, and enhance drug accumulation at target sites through passive or active targeting strategies [30, 31]. Furthermore, functionalization with targeting ligands, such as peptides, antibodies, or aptamers, enables site-specific delivery and minimizes off-target effects [29, 32, 33]. In this study, to efficiently deliver the MCT1 inhibitor, AR-C155858 [34], to microglia, we designed a novel nanoparticle system capable of protecting the drug from enzymatic degradation. Meanwhile, the hydrophobic core of the nanoparticles, composed of PPS, can react with and scavenge ROS [31, 35]. Moreover, the nanoparticles were engineered with cysteine-alanine-glutamine-lysine (CAQK) and CHHSSSARC (MG1) peptide modifications to achieve targeted delivery. The CAQK peptide recognizes the chondroitin sulfate proteoglycans in the brain injury site and facilitates nanoparticle penetration across the BBB [36, 37], while the MG1 peptide specifically directs the nanoparticles to M1-type microglia at the injury site [38, 39]. After accumulating in the injured brain tissue and M1 microglia, the PPS core of the nanoparticles responds to elevated ROS levels, simultaneously scavenging ROS and releasing the MCT1 inhibitor to exert neuroprotective effects.

In this study, we employed both ICH and TBI models to evaluate the therapeutic effects of the newly developed nanoparticles. ICH and TBI are representative brain injury disorders characterized by pronounced ROS and neuroinflammation, primarily mediated by activated microglia. Therefore, we designed novel liposomal nanoparticles incorporating MCT1 and PPS cores, which can effectively target M1-activated microglia and modulate their activation state. By targeting microglia to enhance ROS scavenging and suppress neuroinflammation through regulation of lactate transport, this strategy may represent a promising therapeutic approach for the treatment of brain injury-related disorders.

Methods

Synthesis and characterization of DSPE-PEG₂₀₀₀-MG1

DSPE-PEG₂₀₀₀-NHS (Ruixitech, China) was first dissolved in N, N-dimethylformamide (DMF; D1120000, Aladdin, China), followed by the addition of MG1 peptide (Ruixitech, China) and triethylamine (T103285, Aladdin, China) until complete dissolution was achieved. The reaction was maintained at room temperature for 12 h. The resulting mixture was then transferred into a dialysis bag (MWCO 1000 Da) and dialyzed against ultrapure water for 24 h to eliminate residual small peptides and other impurities. After dialysis, the purified product was collected and freeze-dried to obtain DSPE-PEG₂₀₀₀-MG1. The formation of the conjugate was confirmed through ¹H nuclear magnetic resonance spectroscopy (¹H-NMR, Fourier 80, Rruker, USA).

Synthesis of the nanoparticles

To prepare MiRCM nanoparticles, PPS₁₂₀ (1.0 mg; Ruixitech, China), lecithin (0.1 mg; MB5129-1, MeilunBio, China), DSPE-PEG₂₀₀₀-CAQK (0.06 mg; Ruixitech, China), DSPE-PEG₂₀₀₀ (0.04 mg; D163619, Aladdin, China), DSPE-PEG₂₀₀₀-MG1 (0.06 mg; Ruixitech, China), and AR-C155858 (0.1 mg; HY-13248, MedChemExpress, USA) were dissolved in 200 µL of DMSO (D670381, Aladdin, China). The organic phase was then gradually added into 800 µL of PBS under gentle stirring. Following a 10-minute incubation, the mixture was continuously stirred at room temperature for approximately 2 h to facilitate nanoparticle self-assembly. Unencapsulated AR-C155858 and excess DMSO were removed via dialysis against PBS (10010023, Gibco, USA) for 12 h. The purified MiRCM nanoparticles were collected and stored at 4 °C until further use. Other nanoparticle formulations, including R, MiR, and MiRC, were prepared following the same procedure.

Characterization of the nanoparticles

The nanoparticles’ diameter and zeta potential were analyzed using a laser particle size analyzer (Zetasizer Nano Series, Malvern, UK). Their structural morphology was examined by transmission electron microscopy (TEM, Hitachi, Japan). The encapsulation efficiency of AR-C155858 was determined by high-performance liquid chromatography (HPLC; LC-20AD, Shimadzu, Japan) equipped with a C18 column (150 × 4.6 mm, 5 μm). The mobile phase consisted of 0.1% phosphoric acid water: acetonitrile (80:20) at a flow rate of 0.8 mL/min. The analysis was performed at 30 °C with an injection volume of 20 μL.

Drug release test

One milliliter of MiRCM nanoparticle suspension was placed into a dialysis bag and incubated in 100 mL of PBS or 0.5 µM H₂O₂ in PBS. The samples were maintained at 37 °C, and 200 µL samples were obtained at appointed time points. An equal volume of PBS was added after each test to maintain the same volume. The cumulative release of AR-C155858 was quantified by HPLC.

Cell culture and uptake of nanoparticles in vitro

BV2, HT22, Bend.3, and C8-D1A cell lines were obtained from the Cell Bank of the Chinese Academy of Sciences. Cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM; LD1111, Bioagrio, China) supplemented with 10% fetal bovine serum (FBS; S1356, Bioagrio, China) and 1% penicillin-streptomycin (XA4122, Bioagrio, China). The cells were incubated at 37 °C in a humidified incubator with 5% CO₂. Upon reaching roughly 80% confluence, cells were digested using trypsin and passaged into new dishes. For experiments, 2 × 10⁴ cells were plated in 96-well glass-bottom plates and allowed to adhere overnight. Experimental groups were treated with Hemin (60 µM, HY-19424, MedChemExpress, USA) and LPS (1 µg/mL; ST1470, Beyotime, China), followed by 5 µL of DiI (40726ES10, Yeasen, China)-labeled nanoparticles of various types. After an 8-hour incubation, cells were fixed in paraformaldehyde (PFA; G1101, Servicebio, China), and nanoparticles’ uptake was assessed using confocal fluorescence microscopy.

ROS detection

BV2 microglia were seeded into 96-well plates at a density of 2 × 10⁴ cells per well and treated with PBS, Hemin (60 µM), LPS (1 µg/mL) and 2 µL of various nanoparticles for 24 h. After incubation, the cells were replaced with serum-free DMEM containing DCFH-DA (10 µM; 50101ES01, Yeasen, China) and incubated for 30 min. Free dye was removed by washing with serum-free medium. Fluorescence signals were then visualized using a fluorescence microscope (excitation 488 nm, emission 519 nm), and images were quantitatively analyzed using ImageJ software (NIH, USA).

Cell co-culture in vitro

When BV2 microglia, HT22 neurons, Bend.3 endothelial cells, and C8-D1A astrocytes reached ~ 80% confluence, they were digested with trypsin. Subsequently, 1 × 10⁵ cells HT22, Bend.3, or C8-D1A cells were seeded into the basal chambers of 24-well transwell inserts, while 3 × 10⁴ BV2 cells were plated in the apical chambers (0.4 μm pore size; 3470, Coring, USA). The co-culture systems were maintained in DMEM supplemented with 10% FBS and 0.5% penicillin/streptomycin at 37 °C under 5% CO₂. To induce injury-like conditions, BV2 cells in the apical chamber were exposed to Hemin (60 µM) or LPS (1 µg/mL), with 1 µM AR-C155858 and different types of nanoparticles with the same amount of AR-C15585. After 24 h of treatment, cells were either fixed with PFA or collected for downstream analyses.

In vitro BBB model

To establish the in vitro BBB model, bEnd.3 endothelial cells were seeded onto the transwell inserts (3470, Coring, USA) and allowed to form a confluent monolayer. Cells were cultured in DMEM supplemented with 10% FBS for 7 days to promote barrier maturation. Barrier integrity was monitored by measuring transendothelial electrical resistance (TEER) and further confirmed by immunofluorescence analysis. Only monolayers exhibiting TEER values above 200 Ω·cm² were included in subsequent experiments. For morphological evaluation, the cells were fixed and permeabilized, followed by staining with FITC-phalloidin (Abcam, Cambridge, UK) to label F-actin and DAPI to visualize nuclei. Fluorescent images were captured using a fluorescence microscope.

For the permeability assay, microglia were first treated with Mi, R, MiR, or MiRM nanoparticles under Hemin (60 µM) or LPS (1 µg/mL) stimulation. After 24 h of treatment, the conditioned medium was collected and applied to the established in vitro BBB model. Following an additional 24 h incubation, 1 mM 40 kDa FITC-dextran was added to the upper chamber of the transwell inserts. Transwell inserts without bEnd.3 endothelial cells served as blank controls. After 3 h of incubation, medium from the lower chamber was collected, and fluorescence intensity was measured using a microplate reader (BioTek Synergy Neo2, Agilent, USA). The relative permeability ratio was calculated by normalizing fluorescence values to the blank control group.

Lactate levels test assays

Extracellular lactate and intracellular levels were quantified using a lactate assay kit (S0208S, Beyotime, China) according to the manufacturer’s instructions. Briefly, cells were seeded in 6-well plates and subjected to the indicated treatments. After 24 h treatment, the cell media were collected, and the cells were washed twice with ice-cold PBS. Subsequently, 200 µL of BeyoLysis™ Buffer A for Metabolic Assay was added to each well, and cells were lysed on ice for 10 min. Lysates were centrifuged at 12,000 × g for 5 min at 4 °C, and the supernatants were collected for analysis. A standard curve was generated using serial dilutions of a 100 mM lactate standard prepared in BeyoLysis™ Buffer A for Metabolic Assay. The WST-8 working solution (50 µL per reaction) was freshly prepared by mixing 44 µL Lactate Assay Buffer, 2 µL Enzyme Solution, 2 µL Substrate, and 2 µL WST-8 reagent. For the assay, 50 µL of either sample supernatant or standard solution was added to a 96-well plate. Wells containing only lysis buffer served as blank controls. Subsequently, 50 µL of WST-8 working solution was added to each well, followed by incubation at 37 °C for 30 min in the dark. Absorbance was measured at 450 nm using a microplate reader (BioTek Synergy Neo2, Agilent, USA). Lactate concentrations were calculated based on the standard curve.

Live/Dead assay

Cell viability was assessed using a Calcein AM/ PI double-staining kit (C2015S, Beyotime, China) according to the manufacturer’s instructions. Briefly, HT22 neurons were seeded in 24-well plates and co-cultured with BV2 microglia that were treated as indicated in the experimental design. After 24 h of treatment, the culture medium was removed and the adherent HT22 neurons were gently washed once with PBS. Then, 250 µL Calcein AM/PI working solution containing Calcein AM and PI was added to each well. Cells were incubated at 37 °C for 30 min in the dark. After incubation, fluorescence images were acquired using a fluorescence microscope. Live cells were identified by green fluorescence (Calcein AM, Ex/Em = 494/517 nm), and dead cells were identified by red fluorescence (PI, Ex/Em = 535/617 nm). When required, nuclei were counterstained with DAPI. All procedures were performed under light-protected conditions to prevent photobleaching.

Animals

All animal experiments were performed following the guidelines of Shanghai Jiao Tong University and approved by the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University. Male C57BL/6 mice (6–8 weeks old) were purchased from Shanghai Lingchang Biotechnology Co. (Shanghai, China) and maintained under specific pathogen-free conditions at 20–25 °C, 40–50% relative humidity, with a 12-h light/dark cycle, and free access to food and water. Mice were acclimated for at least five days prior to any experimental procedures.

In vivo model of ICH

To establish ICH models, mice were fixed in a stereotaxic frame (RWD, China), and 0.075 U of collagenase IV (C4-28-100MG, Sigma, USA) in 0.4 µL PBS was injected into the left striatum following a modified protocol. Briefly, a 1-mm burr hole was made 2.2 mm lateral to the midline and 0.5 mm anterior to bregma. A microsyringe needle was lowered 3 mm into the striatum, and collagenase was infused over 90 s using a microinfusion pump (World Precision Instruments, USA). The needle remained in place for 5 min to prevent reflux, after which the hole was sealed with bone wax and the scalp sutured. Mice were recovered on a 37 °C heating pad. Sham-operated mice underwent the same procedure without collagenase injection. Treatment groups received PBS, free AR-C155858 (Mi), MiR, or MiRCM nanoparticles via tail vein injection at a dose equivalent to 4 mg/kg AR-C155858. The nanoparticles were administered for three consecutive days following model establishment, and subsequently once every three days.

In vivo model of CCI

For controlled cortical impact (CCI) models, male C57BL/6 mice (6–8 weeks old) were anesthetized with isoflurane and positioned in a stereotaxic frame (RWD, China). Under sterilized conditions, the skull was fully exposed and a craniectomy was performed without damaging the dura. A severe TBI model was established using a CCI device (Yihong Technology, China). A 3-mm impact tip was aligned perpendicular to the exposed cortical surface, and the following parameters were applied: the impact velocity was 3.5 m/s, the deformation depth was 2.5 mm, and the duration time was 400 ms. This reliably produced hippocampal and cortical damage without mortality. Mice were then recovered on a 37 °C heating pad. Sham animals underwent craniectomy only. Treatment groups received PBS, Mi, MiR, or MiRCM nanoparticles via tail vein injection at 4 mg/kg AR-C155858 equivalent. The nanoparticles were administered for three consecutive days following model establishment and subsequently once every three days.

Biodistribution analysis in vivo

DiD-labeled MiR, MiRC, or MiRCM nanoparticles were intravenously administered at 4 mg/kg AR-C155858 equivalent dose at 24 h after ICH or TBI injury. At 6 h post-injection, mice were perfused with saline, and major organs, including the brain, liver, heart, spleen, and lungs, were harvested for in vivo imaging using a small-animal imaging system (IVIS Spectrum, PerkinElmer, USA) to assess nanoparticle biodistribution (Ex = 640, Em = 680). Following imaging, brains were fixed in 4% PFA, sectioned, and subjected to immunofluorescence staining to further evaluate nanoparticle localization within brain tissue.

Cell immunofluorescence staining

BV2 microglia were seeded into 96-well plates at a density of 2 × 10⁴ cells per well and treated with Hemin (60 µM) or LPS (1 µg/mL), with 1 µM AR-C155858 and different types of nanoparticles with the same amount of AR-C155858. The co-culture and treatment procedures of HT22, C8, and Bend.3 cells were described above. After 24 h of treatment, cells were fixed for 30 min with 4% PFA at room temperature or -20℃ methanol and followed by three PBS washes. Samples were then permeabilized with 1% Triton X-100 in PBS and blocked with 5% bovine serum albumin in PBS for 1 h. Primary antibodies (Supplementary Table 1) were added and incubated overnight at 4°C. On the second day, samples were washed three times with PBS and incubated for 1 h at room temperature with Alexa Fluor 488-, 555-, or 647-conjugated secondary antibodies and 4’,6-diamidino-2-phenylindole (DAPI, 1:1000 dilution in PBS; 40728ES03, Yeasen, China). High-resolution images were obtained using either a confocal or standard fluorescence microscope.

Brain immunofluorescence staining

At specified time points, mice were deeply anesthetized with isoflurane (induction at 3–4% and maintenance at 1.5-2% in oxygen) using an inhalation anesthesia system, and then perfused transcardially with 0.9% saline followed by 4% PFA. Next, the collected brains were post-fixed in 4% PFA overnight and then cryoprotected in 30% sucrose for 48 h. Coronal brain sections  (20 μm) were prepared using a freezing microtome. Sections were permeabilized and blocked in 1% Triton X-100 (A110694-0100, Sangon, China) with 5% bovine serum albumin (BSA; ST023, Beyotime, China) for 1 h. Primary antibodies (Supplementary Table 1) were diluted in blocking solution and incubated overnight at 4 °C. After washing, sections were treated with Alexa Fluor 488-, 555-, or 647-conjugated secondary antibodies and counterstained with DAPI for 1 h at room temperature. Slides were mounted with coverslips, and high-resolution images were captured using confocal or conventional fluorescence microscopy.

RNA sequencing

Total RNA was isolated from mouse brain tissues collected three days after ICH or TBI using TRIzol^® Reagent (Magen, China) according to the manufacturer’s protocol. RNA concentration and purity were determined with a NanoDrop ND-2000 spectrophotometer (Thermo Fisher Scientific, USA), while RNA integrity was verified on an Agilent 4150 Bioanalyzer (Agilent Technologies, USA). Only samples meeting quality standards were selected for RNA sequencing library construction. Paired-end RNA libraries were generated using the VAHTS Universal V6 RNA-seq Library Prep Kit for Illumina (Vazyme, China) in accordance with the provided instructions. mRNA was isolated from 1 µg of total RNA using oligo(dT) magnetic beads, fragmented, and then reverse-transcribed into cDNA. Subsequently, second-strand synthesis, adapter ligation, and PCR amplification were then performed to generate the paired-end libraries. After purification (AMPure XP system) and quality assessment (Agilent 4150 Bioanalyzer), the libraries were sequenced on an Illumina NovaSeq 6000 platform to obtain 150 bp paired-end reads. The data generated from Illumina platform were used for bioinformatics analysis. Differential expression analysis was performed using the DESeq2 (http://bioconductor.org/packages/release/bioc/html/DESeq2.html), DEGs with log₂FC > 1 and P < 0.05 were considered to be significantly different expressed genes. The GO and KEGG enrichment analyses of differential genes can explain the functional enrichment of differential genes and clarify the differences between samples at the gene function level. We used dplyr software package for GO function enrichment and KEGG pathway enrichment analyses. When P < 0.05, it is considered that the GO or KEGG function is significantly enriched.

Evans blue assay

Three days after ICH or TBI induction, mice received an intravenous injection of Evans blue (2 mL/kg; A602025-0005, Sangon, China) via the tail vein. After dye circulating for 3 h, animals were perfused with saline to remove intravascular dye. Brains were then subsequently excised, photographed, and weighed. For ICH mice, brain slices were additionally collected for imaging analysis. The injured hemispheres were homogenized in 50% trichloroacetic acid. The homogenates were then centrifuged at 10,000 × g for 20 min at 4 °C. The supernatant was then carefully collected and diluted with 95% ethanol. The Evans blue content in supernatant was quantified measuring the fluorescence intensity at 620 nm. The corresponding Evans blue content was calculated by plotting a standard curve with known different concentration gradients.

Enzyme-linked immunosorbent assay (ELISA) test

For cells, BV2 microglia were seeded into 24-well plates at a density of 1 × 10⁵ cells per well and treated with Hemin (60 µM) or LPS (1 µg/mL), with 1 µM AR-C155858 and different types of nanoparticles with the same amount of AR-C155858. After 24 h treatment, cell culture supernatant from each treatment group was collected and centrifuged at 300 × g for 10 min to remove debris. For animals, on day 3 following ICH or TBI injury, the injured brain tissue was perfused with PBS, carefully collected, weighed, and homogenized with an equal ratio of PBS to ensure complete mixing. The homogenate was then centrifuged at 5000–6000 rpm for 5 min at 2–8 °C, and the supernatant was collected for further analysis. Then the levels of IL-1β, IL-6, TNF-α, and IL-12 were quantified using enzyme-linked immunosorbent assay kits (EK201B, EK206, EK282, EK2183, Multi Science Biological Technology, China) according to the manufacturers’ instructions. Standard sample solution dilutions were also prepared. The ELISA plate was prewashed with the supplied buffer. Then standard samples and tested supernatant from cells or tissues were added to individual wells, followed by incubation with detection antibodies for 1.5 h at 25 ± 3 °C under gentle agitation (100–300 rpm). After aspirating the liquid, the plate was washed six times. Then the plate was incubated with 100 µL streptavidin working solution for 30 min, washed again, and incubated with chromogenic substrate solution for 20 min. Then the reaction was terminated, and absorbance was read at 450 nm within 30 min using a microplate reader (BioTek Synergy Neo2 Agilent, USA). Sample concentrations were calculated from the standard curve.

Superoxide dismutase, malondialdehyde and glutathione peroxidase assays

On day 3 after ICH or TBI, the activities of superoxide dismutase (SOD), malondialdehyde (MDA), and glutathione peroxidase (GSH-Px) were measured. Mice were first perfused with saline, and the injured brains were harvested and carefully weighed. Each brain was then homogenized in nine volumes of cold saline (weight: volume, g: mL = 1:9) on ice, followed by centrifugation at 3,500 rpm for 10 min at 4 °C. Then the supernatant was collected for analysis, and protein concentration was determined prior to downstream assays. MDA, GSH-Px, and SOD levels were measured through assay kits (60745ES50, 60746ES50, 60741ES50, Yeasen, China) according to the manufacturers’ instructions.

Biocompatibility and systemic safety evaluation

On the seventh day after model establishment, blood samples and major organs (heart, liver, spleen, lungs, and kidneys) were collected from each experimental group to assess systemic biocompatibility. Serum biochemical markers, including alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), and creatinine (CREA), were analyzed to evaluate hepatic and renal function. The harvested organs were fixed and stained with hematoxylin and eosin (H&E) for histopathological examination under light microscopy.

Adhesive removal test

The adhesive removal test was employed to evaluate sensorimotor asymmetry and tactile sensitivity, as previously described [40]. Briefly, a 3 mm × 2 mm adhesive tape was applied to the contralateral forepaw, and the latency to completely remove the tape was recorded, with a maximum observation time of 120 s.

Foot fault test

The foot fault test was conducted to assess motor coordination and limb placement accuracy at the appointed time points. Mice were placed on a standard wire grid, and their movement was observed for 3 min. A foot fault was defined as a misstep where the paw slipped through the grid. Both the total number of foot faults and the total steps were recorded, and motor impairment was quantified as the percentage of foot faults relative to total steps.

Morris Water Maze (MWM)

Spatial learning and memory were evaluated using the Morris Water Maze test. Prior to surgery, mice underwent a period to familiarize themselves with the environment. Those failing to locate the platform on the pre-training day were excluded. During the subsequent 5-day learning phase, mice were placed into the pool from four randomized start positions and swam for 60 s to locate the hidden platform. Escape latency (the time to find the platform) was recorded as a measure of spatial learning. Those that failed to locate the platform were gently guided to it. 24 h after the learning phase, a 60-second probe trial was conducted with the platform removed to assess the memory ability. The proportion of time spent in the target quadrant and the number of platform site crossings were recorded and analyzed.

Statistical analysis

All statistical analyses were conducted using GraphPad Prism 8 (GraphPad Software LLC, USA). Data in figures were expressed as mean ± SD. Statistical significance was determined using one-way or two-way ANOVA, or Student’s t-test, as appropriate. The number of biological replicates or animals per group is provided in the corresponding figure legends. Significance levels were defined as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Results

Results 1. Preparation and characterization of nanoparticles

The hydrophobic polymer PPS₁₂₀ was first prepared using a reversible addition-fragmentation chain-transfer polymerization method. Under oxidative stress, ROS can oxidize the hydrophobic polysulfide units in PPS into hydrophilic polysulfoxide units, enabling PPS degradation to simultaneously consume ROS and thereby achieve ROS scavenging. Then, DSPE-PEG₂₀₀₀-CAQK and DSPE-PEG₂₀₀₀-MG1 were also synthesized. The synthesis process and chemical structures of DSPE-PEG₂₀₀₀-MG1 were verified via ¹H-NMR in Supplementary Fig. 1. Subsequently, PPS₁₂₀ and AR-C155858 were employed as the hydrophobic core of the nanoparticles. The nanoparticles’ shell was constructed using DSPE-PEG₂₀₀₀, DSPE-PEG₂₀₀₀-CAQK, DSPE-PEG₂₀₀₀-MG1, and lecithin. Through a one-step nanoprecipitation self-assembly process, brain injury targeted nanoparticles capable of selectively recognizing M1-polarized microglia were fabricated and designed for the delivery of MCT1 inhibitor. The average diameter of nanoparticles with only PPS core (R) was 94.09 ± 8.591 nm, that of nanoparticles encapsulated with MCT inhibitor (MiR) was 107.3 ± 6.717 nm, and that of nanoparticles with MCT inhibitor and CAQK/ MG1 peptides (MiRCM) was 105.9 ± 4.152 nm (Fig. 1A). TEM images revealed that the R, MiR, and MiRCM nanoparticles were uniformly spherical with good monodispersity (Fig. 1B). The zeta potentials of R, MiR, and MiRCM nanoparticles were -18.47 ± 1.686 mV, -18.8 ± 4.751 mV and −-20.6 ± 1.706 mV, respectively (Fig. 1C).

Fig. 1.

Structural and chemical characterization of nanoparticles. (A) Size distribution of R, MiR and MiRCM nanoparticles. (B) TME of R, MiR and MiRCM nanoparticles. (C) Zeta potentials of R, MiR and MiRCM nanoparticles. (D) Size distribution of R nanoparticles with H₂O₂ or PBS treated. (E) The drug release efficiency of AR-C155858 from MiRCM with H₂O₂ or PBS treated. Scale bars in (B) represent 100 nm

Next, we examined how ROS affected nanoparticles’ stability and the release profile of the MCT inhibitor. 0.5 mM H₂O₂ was added to the R particles for 12 h, after which their particle sizes were analyzed. The particle sizes of R particles increased when treated with H₂O₂ (Fig. 1D). These results confirm that the PPS-based nanoparticles are sensitive to ROS and undergo degradation upon oxidative stimulation. To further evaluate drug release behavior, the release of AR-C155858 from MiRCM nanoparticles was quantified by HPLC under conditions with or without H₂O₂. Consistent with their ROS-responsive design, MiRCM nanoparticles exhibited markedly accelerated drug release in the presence of H₂O₂, while release remained minimal under ROS-free conditions (Fig. 1E).

Results 2. Nanoparticles with CAQK and MG1 peptide modifications specifically target M1 microglia and brain injury sites

We next investigated whether the MG1 peptide could confer selective affinity of the nanoparticles toward M1-polarized microglia. So, BV2 microglia cells under different polarization conditions were incubated with DiI-labeled nanoparticles modified with MG1 peptide (RM) or their non-targeted counterparts (R). Previous studies have shown that Hemin and lipopolysaccharide (LPS) treatments mimic intracerebral hemorrhage (ICH) and traumatic brain injury (TBI) models, respectively, both driving microglia toward an M1 phenotype, whereas interleukin-4 (IL-4) promotes M2 polarization. So, BV2 cells were treated with Hemin, LPS, or IL-4, followed by exposure to either R or RM nanoparticles. As is shown in Fig. 2A, both R and RM nanoparticles were all internalized by BV2 cells under all conditions and the uptake efficiencies of R and RM were comparable in control group (Fig. 2B). Notably, Hemin and LPS stimulation markedly enhanced RM uptake but had no effect on R (Fig. 2A, B). However, IL-4 treatment did not increase uptake of RM nanoparticles (Fig. 2A, B). Subsequently, when BV2 cells were exposed to different concentrations of Hemin, RM uptake exhibited a dose-dependent increase (Fig. 2C, D). To further evaluate cell-type specificity, we assessed nanoparticle internalization in bEnd.3 endothelial cells, HT22 neurons, and C8-D1A astrocytes. These cell types showed limited uptake of either nanoparticle (Fig. 2E, F), and neither Hemin nor LPS pretreatment augmented their internalization efficiency (Fig. 2G, H). Collectively, these findings demonstrate that MG1 peptide modification effectively drives nanoparticle enrichment within M1-polarized microglia.

Fig. 2.

Efficient uptake of R and RM nanoparticles under different conditions in different cells. (A) Representative fluorescence images of DiI-labelled R and RM nanoparticles in BV2 microglia with Hemin (60 µM), LPS (1 µg/mL) or IL-4 (20 ng/mL) treated conditions. (B) Quantification of mean fluorescence intensity of DiI-labelled R and RM nanoparticles from (A). n = 6. (C) Representative fluorescence images of DiI-labelled RM nanoparticles in BV2 microglia with different concentrations of Hemin treated conditions. (D) Quantification of mean fluorescence intensity of DiI-labelled RM nanoparticles from (C). n = 6. (E) Representative fluorescence images of DiI-labelled RM nanoparticles in different cells under the control condition. (F) Quantification of mean fluorescence intensity of DiI-labelled RM nanoparticles from (E). n = 6. (G) Representative fluorescence images of DiI-labelled RM nanoparticles in different cells with Hemin and LPS treated conditions. (H) Quantification of mean fluorescence intensity of DiI-labelled RM nanoparticles from (G). n = 6. Data are shown as mean ± SD from six independent biological replicates (n = 6). Statistics in (B) were calculated using two-way ANOVA tests. Statistics in (D, F, H) were calculated using one-way ANOVA tests. NS: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Scale bars represent 50 μm

To effectively deliver the MCT inhibitor to M1 microglia at brain injury site, we next evaluated the in vivo targeting capability of different nanoparticle types in animal models. According to previous studies [31, 36], CAQK is a tetrapeptide that can target chondroitin sulfate proteoglycans at the injury site following TBI, delivering drugs to the brain injury site. So, the brain targeted ability of nanoparticles modified with CAQK and MG1 peptides was evaluated. DiD-labeled MiR, MiRC (nanoparticles with MCT inhibitor and CAQK peptides), and MiRCM, along with PBS as a control, were intravenously administered to mice subjected to 24 h after induction of ICH or TBI. Six hours post-injection, a small animal in vivo imaging system was performed to assess nanoparticle distribution within the brain. MiR, which lacked any targeting modification, exhibited minimal accumulation in the lesion sites of both ICH and TBI models (Fig. 3A-D). In contrast, MiRC decorated with the CAQK targeting peptide demonstrated substantial enrichment at injury regions, and MiRCM, modified with both CAQK and MG1 peptides, displayed a broader distribution within these lesions compared to MiRC nanoparticles (Fig. 3A-D). Next, biodistribution analysis across major organs revealed that all nanoparticle types—MiR, MiRC, and MiRCM—predominantly accumulated in the liver and kidneys, with no significant differences among formulations (Fig. 3E and F, Supplement Fig. 2A and B). We also quantified the targeting efficiency of nanoparticles to the brain by normalizing organ-specific fluorescence intensity to the total detected fluorescence. As shown in Supplementary Fig. 2C, approximately 20% of the total fluorescence signal was detected in the ICH brain, whereas the remaining were mainly distributed in the liver and kidneys. To further validate the targeting specificity at the tissue and cellular levels, we performed immunofluorescence staining. Compared to MiR, MiRC exhibited more efficient localization to the lesion sites in ICH and TBI brains (Supplementary Fig. 2D-F). Moreover, MiRCM showed significantly higher colocalization with CD68⁺/IBA1⁺ microglia than MiRC, indicating a preferential accumulation within M1-polarized microglia in the damaged regions (Fig. 3G, H). Finally, analysis of serum drug concentrations revealed that free AR-C155858 was rapidly cleared from the circulation. In contrast, AR-C155858 encapsulated within MiR nanoparticles exhibited a markedly slower clearance rate, suggesting that nanoparticle encapsulation protected the drug from elimination in serum and thereby prolonged its systemic retention (Supplementary Fig. 2G). Moreover, surface modification with the CAQK or MG1 targeting peptides also exhibited a slower clearance rate of AR-C155858 in serum (Supplementary Fig. 2G). Collectively, these findings demonstrate that MiRCM nanoparticles effectively target both lesion areas and M1 microglia in ICH and TBI models, enabling efficient delivery of the MCT1 inhibitor AR-C155858 to microglia.

Fig. 3.

Targeted distribution of MiR, MiRC and MiRCM nanoparticles in ICH and TBI mouse models. (A-B) Representative in vivo living fluorescence images of brains from ICH and TBI mice at 6 h after intravenous injection of DiD-labeled MiR, MiRC, or MiRCM nanoparticles. (C-D) Quantitative analysis of brain fluorescence intensity corresponding to (A) and (B), respectively. n = 3 mice. (E) Representative in vivo living fluorescence images of major organs harvested from ICH mice at 6 h after intravenous injection with DiD-labeled MiR, MiRC and MiRCM nanoparticles. (F) Quantitative analysis of major organs fluorescence intensity corresponding to (E). n = 3 mice. (G-H) Representative immunofluorescence images showing CD68 (green), IBA1 (red), and DiD-labeled nanoparticles (purple) in the perilesional brain regions of Sham, ICH and TBI mice. DiD-labeled MiRC or MiRCM nanoparticles were intravenously administered 24 h after injury, and brain tissues were collected 6 h post-injection. Data are shown as mean ± SD from three independent biological replicates (n = 3). Statistics were calculated using one-way ANOVA tests. NS: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Scale bars represent 50 μm in (G, H)

Results 3. MiRM nanoparticles exert antioxidant effects and inhibit M1 microglial polarization in vitro

Given that the nanoparticles incorporate PPS which has intrinsic antioxidant activity, we next assessed their capacity to scavenge ROS. Intracellular ROS levels in BV2 microglia were quantified using a DCFH-DA assay. Hemin and LPS markedly elevated ROS levels, whereas treatment with R nanoparticles significantly attenuated this increase (Fig. 4A, B). Notably, RM nanoparticles, which are decorated with MG1 targeting peptide, achieved a greater reduction of ROS compared to R, indicating that MG1 modification further enhances antioxidant activity (Fig. 4A, B). We then examined the impact of different MCT inhibitor-loaded nanoparticles on M1-polarized BV2 cells. As AR-C155858 effectively inhibits MCT1 activity, which further hinders the efflux of lactate, we next examined the concentration of lactate in BV2 microglia culture medium. As shown in Supplementary Fig. 3A, B, Hemin and LPS increased the concentration of lactate in the cell medium. Free MCT1 inhibitor (AR-C155858), R, MiR and MiRM nanoparticles all reduced the concentration of lactate in the medium, and the MiRM had the most potent effect. Mechanistically, Mi and MiR significantly increased intracellular lactate levels compared with Hemin/LPS and Hemin/LPS + R groups respectively, rather than reducing intracellular lactate content (Supplementary Fig. 3C, D). This finding indicates that the observed decrease in extracellular lactate in the culture medium is attribute to inhibition of lactate efflux, rather than suppression of lactate production. In contrast, the Hemin/LPS + R group exhibited reduced intracellular lactate levels compared with the Hemin/LPS-treated group, suggesting that R decreases intracellular lactate production by ROS (Supplementary Fig. 3C, D), thereby lowering extracellular lactate accumulation. Besides, immunofluorescence staining for iNOS and CD68 confirmed that Hemin and LPS robustly increased expression of these M1 markers (Fig. 4C-F). While free MCT1 inhibitor, R, and MiR all reduced iNOS and CD68 expression, more pronounced suppression was observed in the MiRM-treated group. (Fig. 4C-F). Consistent with these findings, ELISA analysis revealed that Hemin and LPS stimulation markedly elevated secretion of pro-inflammatory cytokines, including IL-1β, IL-6, TNF-α, and IL-12 (Fig. 4G, H). Treatment with free MCT inhibitor, R, MiR, or MiRM mitigated this effect, with MiRM again producing the most pronounced inhibition (Fig. 4G, H). These results indicate that MG1-modified nanoparticles effectively suppress lactate efflux, M1 polarization, and inflammatory cytokine release in microglia.

Fig. 4.

Nanoparticles scavenge ROS, inhibit M1 microglia polarization, and suppress pro-inflammatory cytokine production in vitro. (A) Representative ROS fluorescence images of BV2 microglia treated with R and RM nanoparticles under Hemin (60 µM) or LPS (1 µg/mL) stimulation. (B) Quantification of mean fluorescence intensity of ROS corresponding to (A). n = 3. (C) Representative fluorescence images of iNOS and CD68 in BV2 microglia treated with Mi, R, MiR and MiRM nanoparticles under Hemin (60 µM) stimulation. (D) Quantification of mean fluorescence intensity of iNOS and CD68 in BV2 microglia corresponding to (C). n = 3. (E) Representative fluorescence images of iNOS and CD68 in BV2 microglia treated with Mi, R, MiR and MiRM nanoparticles under LPS (1 µg/mL) stimulation. (F) Quantification of mean fluorescence intensity of iNOS and CD68 in BV2 microglia corresponding to (E). n = 3. (G) ELISA results of inflammatory cytokines of microglia supernatants treated with Mi, R, MiR and MiRM nanoparticles under Hemin (60 µM) stimulation. n = 3. (H) ELISA results of inflammatory cytokines of microglia supernatants treated with Mi, R, MiR and MiRM nanoparticles under LPS (1 µg/mL) stimulation. n = 3. Data are shown as mean ± SD from three independent biological replicates (n = 3). Statistics were calculated using one-way ANOVA tests. NS: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Scale bars represent 50 μm in (A, C, E)

Results 4. MiRM nanoparticles inhibit neuronal death, BBB disruption, and astrocyte activation in vitro

To partly mimic the cellular microenvironment of ICH and TBI and explore the nanoparticles’ effect on others cells through microglia, we established a transwell-based co-culture system in which BV2 microglia were seeded in the upper chamber and neurons, endothelial cells, or astrocytes in the lower chamber. Following Hemin and LPS treatment, cells were exposed to free MCT inhibitor, R, MiR, or MiRM. Live/dead assays showed that Hemin and LPS promoted neuronal death, whereas all treatments attenuated this effect, with MiRM providing the strongest neuroprotection (Fig. 5A-D). Immunofluorescence analysis further revealed that Hemin and LPS suppressed expression of the tight junction protein ZO-1 in Bend.3 endothelial cells, indicating BBB disruption. This suppression was reversed by all treatment groups, again with MiRM exhibiting superior efficacy (Fig. 5E-H). To further examine the BBB integrity, the in vitro BBB model was established using a transwell (Supplement Fig. 4). The Bend.3 cells were seeded in the upper chamber and formed a fully confluent monolayer on day 7 (Supplement Fig. 4A). F-actin was stained with FITC-phalloidin to visualize filamentous actin and cell morphology, demonstrating proper cellular alignment and tight junction formation (Supplement Fig. 4A). In addition, the transendothelial electrical resistance (TEER) reached 200 Ω·cm² on day 7, indicating the formation of a functional BBB (Supplement Fig. 4B). We then evaluated the effects of nanoparticles on BBB integrity using a co-culture system with conditioned medium, as illustrated in Supplementary Fig. 4C. BBB permeability was assessed by measuring FITC-dextran penetration and TEER values. As shown in Supplementary Fig. 4D, Hemin and LPS significantly increased BBB permeability, whereas all nanoparticle treatments attenuated this effect, with MiRM showing the most pronounced protective effect. Consistently, TEER measurements demonstrated a similar trend (Supplementary Fig. 4E). Together, these results further demonstrate that MiRM effectively preserves BBB integrity under stimulated conditions. Finally, astrocytes in the co-culture system exhibited increased expression of the A1-reactive astrocyte marker C3 after LPS and Hemin exposure, which was markedly reduced upon the nanoparticle treatment. Similarly, the MiRM showed the highest efficiency (Fig. 5I-L).

Fig. 5.

Nanoparticles protect neurons from death, preserve endothelial function, and inhibit A1 astrocyte transformation in vitro. (A) Representative fluorescence images of live (green) and dead (red) neurons after 24 h co-cultured with microglia treated with Mi, R, MiR and MiRM nanoparticles under Hemin (60 µM) stimulation. (B) Quantification of dead/live cell ratios corresponding to (A). n = 3. (C) Representative fluorescence images of live (green) and dead (red) neurons after 24 h co-cultured with microglia treated with Mi, R, MiR and MiRM nanoparticles under LPS (1 µg/mL) stimulation. (D) Quantification of dead/live cell ratios corresponding to (C). n = 3. (E) Representative fluorescence images of ZO-1 in endothelial cells after 24 h co-cultured with microglia treated with Mi, R, MiR and MiRM nanoparticles under Hemin (60 µM) stimulation. (F) Quantification of the mean fluorescence intensity of ZO-1 corresponding to (E). n = 3. (G) Representative fluorescence images of ZO-1 in endothelial cells after 24 h co-cultured with microglia treated with Mi, R, MiR and MiRM nanoparticles under LPS (1 µg/mL) stimulation. (H) Quantification of the mean fluorescence intensity of ZO-1 corresponding to (G). n = 3. (I) Representative fluorescence images of C3 in astrocytes after 24 h co-cultured with microglia treated with Mi, R, MiR and MiRM nanoparticles under Hemin (60 µM) stimulation. (J) Quantification of the mean fluorescence intensity of C3 corresponding to (I). n = 3. (K) Representative fluorescence images of C3 in astrocytes cells after 24 h co-cultured with microglia treated with Mi, R, MiR and MiRM nanoparticles under LPS (1 µg/mL) stimulation. (L) Quantification of the mean fluorescence intensity of C3 corresponding to (K). n = 3. Data are shown as mean ± SD from three independent biological replicates (n = 3). Statistics were calculated using one-way ANOVA tests. NS: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Scale bars represent 50 μm in (A, C, E, G, I, K)

Collectively, these findings demonstrate that MiRM nanoparticles not only mitigate microglial M1 polarization and pro-inflammatory signaling but also protect against neuronal death, BBB disruption, and astrocyte activation in in vitro ICH and TBI cell models.

Results 5. MiRCM nanoparticles exert neuroprotective and antioxidant effects in vivo

We next evaluated the therapeutic efficacy of the nanoparticles in vivo using a collagenase-induced ICH model and a controlled cortical impact (CCI) model of TBI. Mice received intravenous injections of free MCT inhibitor, R, MiR, or MiRCM for three consecutive days following model establishment. Brain pathology was assessed on day 3. Because the nanoparticles mainly target microglia, we first evaluated the nanoparticle effect on M1 polarized microglia. As shown in Fig. 6A, B and Supplementary Fig. 5A, B, immunofluorescence analysis revealed a marked increase in M1-polarized microglia in the peri-hematomal region of ICH brains and in the peri-injury site of TBI brains, which was significantly attenuated by all treatments, with MiRCM exhibiting the most pronounced effect. Subsequently, the structure of neurons were examined through Nissl staining. As is shown in Fig. 6C, D and Supplementary Fig. 5C, D, the Nissl staining demonstrated substantial neuronal loss in ICH and TBI brains, which was effectively mitigated by nanoparticle administration, again with MiRCM showing the strongest neuroprotection. Similarly, astrocytic activation around the lesion site was strongly induced by ICH and TBI but was reversed by all treatment groups, with MiRCM exerting superior inhibitory effects (Fig. 6E, F and Supplementary Fig. 5E, F). To assess BBB integrity, we next performed immunostaining for ZO-1 and Evans blue extravasation assays. ICH and TBI disrupted ZO-1 expression of vessels and increased Evans blue leakage around the lesion site, indicating severe BBB breakdown. Treatment with Mi, R, MiR, or MiRCM restored ZO-1 levels and reduced vascular permeability, with MiRCM providing the most significant protection against BBB disruption (Fig. 6G-J and Supplementary Fig. 5G-J). Given the intrinsic antioxidant properties of the nanoparticles, we next evaluated their impact on endogenous antioxidant defenses in the brain. Levels of SOD, GSH-Px, and MDA were measured as key markers of oxidative stress. SOD eliminates superoxide radicals, GSH-Px interrupts lipid peroxidation chains, and MDA reflects the extent of oxidative damage. Remarkably, beyond direct ROS scavenging, MiRCM treatment significantly enhanced antioxidant capacity, as evidenced by increased SOD and GSH-Px activity and reduced MDA accumulation, indicating robust protection against oxidative injury (Fig. 6K-M and Supplementary Fig. 5K-M). Finally, we assessed inflammatory cytokine levels in ICH and TBI brains. ICH and TBI markedly elevated IL-1β, IL-6, TNF-α, and IL-12 expression, which was substantially reduced by all treatments, with MiRCM achieving the more pronounced suppression (Fig. 6N and Supplementary Fig. 5N). Collectively, these findings demonstrate that MiRCM nanoparticles effectively attenuate M1 microglial activation, neuronal loss, astrocytic reactivity, and BBB disruption, while enhancing antioxidant defenses and suppressing neuroinflammation in ICH and TBI models.

Fig. 6.

MiRCM nanoparticles inhibit M1 microglia polarization, protect neurons, attenuate astrocyte activation, preserve BBB integrity, and suppress pro-inflammatory responses of ICH mice in vivo. (A) Representative immunofluorescence images of IBA (red) and CD68 (green) at the peri-hematoma region in sham and ICH mice treated intravenously with PBS, Mi, R, MiR, or MiRCM at day 3 post-ICH. (B) Quantification of the percentage of IBA1/CD68 double positive cells corresponding to (A). n = 5 mice. (C) Representative Nissl brain staining at the peri-hematoma region in sham and ICH mice treated intravenously with PBS, Mi, R, MiR, or MiRCM at day 3 post-ICH. (D) Quantification of viable neurons corresponding to (C). n = 5 mice. (E) Representative immunofluorescence images of GFAP (green) at the peri-hematoma region in sham and ICH mice treated intravenously with PBS, Mi, R, MiR, or MiRCM at day 3 post-ICH. (F) Quantification of mean fluorescence intensity of GFAP positive cells corresponding to (E). n = 5 mice. (G) Representative immunofluorescence images of CD31 (green) and ZO-1 (red) at the peri-hematoma region in sham and ICH mice treated intravenously with PBS, Mi, R, MiR, or MiRCM at day 3 post-ICH. (H) Quantification of the colocalization coefficient of ZO-1/CD31 corresponding to (G). n = 5 mice. (I) Representative Evans Blue leakage images of ICH mice treated intravenously with PBS, Mi, R, MiR, or MiRCM at day 3 post-ICH. (J) Quantification of Evans Blue leakage corresponding to (I). n = 6 mice. (K-M) Quantification of SOD (K), MDA (L) and GSH-Px (M) of the hematoma brain in sham and ICH mice treated intravenously with PBS, Mi, R, MiR, or MiRCM at day 3 post-ICH. n = 6 mice. (N) ELISA results of inflammatory cytokines of hematoma brain in sham and ICH mice treated intravenously with PBS, Mi, R, MiR, or MiRCM at day 3 post-ICH. n = 6 mice. Data are shown as mean ± SD from five or six independent biological replicates (n = 5-6). Statistics were calculated using one-way ANOVA tests. NS: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Scale bars represent 50 μm in (A, C, E, G)

Results 6. RNA-sequencing analysis of gene expression changes in response to MiRCM treatment

We next performed transcriptomic profiling of brain tissues between Sham + PBS, ICH + PBS and ICH+MiRCM treated group to evaluate gene expression changes under different conditions. As shown in Fig. 7A, Supplementary Fig. 6A and Supplementary Table 2, the ICH + PBS group exhibited 1930 upregulated and 543 downregulated genes compared with the Sham + PBS group. The GO pathway enrichment analyses of these differentially expressed genes between ICH + PBS and Sham + PBS group revealed that upregulated genes in the ICH + PBS group were predominantly associated with pathways related to immune response, defense response and external stimulus (Supplementary Fig. 6C, Supplementary Table 3), while the downregulated genes in the ICH + PBS group were predominantly associated with nervous system development, neurogenesis, generation of neuron and cell junction organization (Supplementary Fig. 6D, Supplementary Table 4). And KEGG analysis indicated that upregulated genes in the ICH + PBS group were mainly enriched in NOD-like receptor signaling pathway, NF-kappa B signaling pathway, TNF signaling pathway, Toll-like receptor signaling pathway (Supplementary Fig. 6E, Supplementary Table 5), while the downregulated genes were mainly enriched in neuroactive ligand-receptor interaction, cell adhesin molecules, tight junction, axon guidance, compared to Sham + PBS group (Supplementary Fig. 6F, Supplementary Table 6). These results indicate that ICH significantly disrupts neuronal and axonal integrity, impairs normal intercellular connections, and triggers oxidative stress and neuroinflammation.

Fig. 7.

Bulk RNA-sequencing analysis of gene expression changes of ICH mice following MiRCM treatment. (A) Heatmap of RNA sequencing data showing gene expression profiles of hematoma brain in Sham + PBS, ICH + PBS, and ICH+MiRCM groups. n = 3 biological replicates per group. (B) Volcano plot showing differentially expressed genes of hematoma brain in ICH mice treated with PBS and MiRCM. n = 3 biological replicates per group. (C) GO enrichment analysis of down-regulated and up-regulated genes of hematoma brain in ICH mice treated with MiRCM versus PBS, ranked by p-value. The top 10 enriched Biological Processes pathways are shown. (D) KEGG enrichment analysis of down-regulated genes of hematoma brain in ICH mice treated with MiRCM versus PBS, ranked by p-value. The top 20 enriched KEGG pathways are shown. (E) GSEA analysis highlighting downregulation of the “interleukin 4 and interleukin 13 signaling” pathway in ICH+MiRCM versus ICH + PBS groups. (F) Heatmap corresponding to (E) showing the differentially expressed genes of hematoma brain in ICH mice treated with PBS and MiRCM

We next analyzed the differentially expressed genes between the ICH + MiRCM group and the ICH + PBS group to evaluate the effects of MiRCM nanoparticles. As shown in the heatmap and (Fig. 7A, B), the ICH+MiRCM group exhibited 298 upregulated and 662 downregulated genes compared with the ICH + PBS group (Supplementary Table 7). GO pathway enrichment analyses of these differentially expressed genes revealed that downregulated genes in the ICH+MiRCM group were predominantly associated with pathways related to defense response, response to external stimulus, inflammatory response, response to stress, and immune response (Fig. 7C, Supplementary Table 8). And KEGG analysis indicated that downregulated genes were mainly enriched in the IL-17 signaling pathway, TNF signaling pathway, and NF-κB signaling pathway (Fig. 7D, Supplementary Table 9). Further GSEA of the downregulated genes demonstrated significant suppression of the interleukin-4 and interleukin-13 signaling pathway, which included multiple inflammation-related genes such as Il1b, Ccl12, Ccl2, Il6, and Tnf (Fig. 7E, F). The transcriptomic analysis also revealed that several genes associated with oxidative stress amplification, inflammatory propagation, and tight junction were significantly decreased following MiRCM treatment, such as Tnf, Il6, Ripk3, Duox2, Mmp9, Alox5, Lcn2, Cd36, and Il18rap, that are well-recognized contributors to ROS generation, lipid peroxidation, neuroinflammatory cascades, and BBB integrity (Supplementary Table 7). Tnf and Il6 activate NF-κB and STAT3 pathways, promoting secondary injury [41, 42], while Ripk3 mediates necroptotic cell death, further exacerbating tissue damage [43]. Duox2 directly generates ROS [44], and Mmp9 compromises BBB integrity by degrading tight junctions, collagen and laminin [45]. Lcn2 also promotes acute inflammation and oxidative stress through iron accumulation [46]. Collectively, these findings suggest that MiRCM nanoparticles administration attenuates immune and inflammatory responses as well as oxidative stress signaling in the brain after ICH, thereby facilitating neurological recovery.

For the ICH+MiRCM and Sham + PBS groups, the ICH+MiRCM group exhibited 1212 upregulated and 485 downregulated genes compared with the Sham + PBS group (Supplementary Fig. 6B, Supplementary Table 10). And the GO pathway enrichment analyses of these differentially expressed genes revealed that upregulated genes in the ICH+MiRCM group were also predominantly associated with pathways related to immune response, defense response and external stimulus (Supplementary Fig. 6G, Supplementary Table 11), while the downregulated genes in the ICH+MiRCM group were predominantly associated with axon ensheathment, ensheathment of neuron, nervous system development, oligodendrocyte differentiation, myelination (Supplementary Fig. 6H, Supplementary Table 12). And KEGG analysis indicated that upregulated genes in the ICH+MiRCM group were mainly enriched in cytokine-cytokine receptor interaction, complement and coagulation cascades, NOD-like receptor signaling pathway, NF-kappa B signaling pathway, TNF signaling pathway (Supplementary Fig. 6I, Supplementary Table 13), while the downregulated genes were mainly enriched in cAMP signaling pathway, tight junction, PI3K-AKT signaling, ECM-receptor interactions, neuroactive ligand-receptor interactions (Supplementary Fig. 6J, Supplementary Table 14), compared to Sham + PBS group. These results indicate that although MiRCM effectively suppresses inflammation- and oxidative stress-related signaling pathways, a certain degree of structural and functional damage persists compared with the Sham + PBS group.

Results 7. MiRCM nanoparticles improve motor, sensory, and cognitive function in ICH and TBI mice in vivo

Improving the prognosis of ICH and TBI is the ultimate therapeutic goal. Previous studies have shown that ICH and TBI can not only cause motor and sensory impairment but also impair learning and memory in mice [40, 47]. The impairment is related to the damage of hippocampal neurons caused by ROS and neuroinflammation generated after ICH and TBI. These results above have proved that the MiRCM could scavenge ROS and inhibit MCT expression by targeting M1 microglia, thereby modulating inflammation and suppressing oxidative stress in the mouse brain during the acute phase. This modulation exerted a protective effect on brain tissue. Therefore, long-term administration of the nanoparticles should also exert a beneficial effect on neurological recovery. Therefore, we further investigated whether nanoparticle treatment improves motor, sensory, and cognitive function in ICH and TBI mice during the chronic phase. A schematic timeline of the animal experimental procedures is shown in Fig. 8A. Mi, R, MiR, MiRM, and MiRCM were injected via the tail vein for three consecutive days following model establishment, and subsequently once every three days. Then, we used the Morris Water Maze (MWM) test to assess learning and memory in mice from days 15 to 20 after ICH and TBI. It was found that ICH significantly impaired the mice’s ability to locate the hidden platform, as reflected by increased escape latency and path length, but not swimming speed (Fig. 8B and C, Supplementary Fig. 7A and B). Tail vein injection of Mi, R, MiR, MiRM, and MiRCM significantly reduced the latency time and path length, but not swimming speed, with MiRCM showing the best effect (Fig. 8B and C, Supplementary Fig. 7A and B), indicating that MiRCM can significantly improve the mice’s learning ability. Furthermore, we removed the platform on the 20st day after ICH and evaluated the memory ability of the mice in different groups, by measuring the time the mice spent in the target quadrant and the number of times they crossed the platform area. The results showed that ICH significantly reduced the target quadrant time and the number of crossing times during the memory phase, indicating that ICH impairs the mice’s memory ability (Fig. 8B, C). But R, MiR, MiRM, and MiRCM treatments significantly increased target quadrant time and crossing numbers, with MiRCM showing the best effect (Fig. 8B, C). These results exhibited that MiRCM can significantly improve learning and memory of ICH mice. Furthermore, we similarly found that MiRCM improved learning and memory abilities in TBI mice (Fig. 8G and H, Supplementary Fig. 7C and D). Next, the effects of these nanoparticles on motor and sensory function in mice with ICH and TBI were further investigated. The mNSS score is commonly used to assess neurological impairment, with higher scores indicating more severe neurological impairment. The results showed that injection of R, MiR, MiRM, and MiRCM significantly reduced mNSS scores in mice with ICH and TBI on days 3, 7, and 14, indicating improved neurological function (Fig. 8D, I). We further evaluated the motor and sensory function of mice using the adhesion removal test and foot fault test on days 1, 3, 7, 14, and 28. As shown in Fig. 8E, F, J and K, ICH and TBI significantly prolonged the time required to remove adhesive tapes and increased the frequency of missteps. Injection of R, MiR, MiRM, and MiRCM significantly reduced both removal time and frequency of missteps, with MiRCM producing the most improvement. This suggests that MiRCM can significantly improve the motor and sensory function of mice with ICH and TBI. In summary, MiRCM nanoparticles can improve both short-term and long-term motor, sensory, and learning and memory function in ICH and TBI mice.

Fig. 8.

MiRCM nanoparticles enhanced the motor, sensor and cognitive function of mice. (A) Schematic timeline of the animal experimental procedures. (B) Representative swim path traces of sham and ICH mice treated intravenously with PBS, Mi, R, MiR, or MiRCM during the MWM learning phase (hidden platform) and memory phase (probe trial). (C) Escape latency of ICH mice during the learning phase of MWM (left). Time spent in the target quadrant (middle) and number of platform crossings (right) of ICH mice during the MWM memory phase. n = 6 mice. (D) mNSS scores of sham and ICH mice treated intravenously with PBS, Mi, R, MiR, or MiRCM. (E-F) Results of the adhesive removal test (E) and foot fault test (F) of sham and ICH mice treated intravenously with PBS, Mi, R, MiR, or MiRCM. n = 6 mice. (G) Representative swim path traces of sham and TBI mice treated intravenously with PBS, Mi, R, MiR, or MiRCM during the MWM learning phase (hidden platform) and memory phase (probe trial). (H) Escape latency of TBI mice during the learning phase of MWM (left). Time spent in the target quadrant (middle) and number of platform crossings (right) of TBI mice during the MWM memory phase. n = 6 mice. (I) mNSS scores of sham and TBI mice treated intravenously with PBS, Mi, R, MiR, or MiRCM. (J-K) Results of the adhesive removal test (J) and foot fault test (K) of sham and TBI mice treated with PBS, Mi, R, MiR, or MiRCM. n = 6 mice. Data are shown as mean ± SD from six independent biological replicates (n = 6). Statistics in (C, D, E, F, H, I, J, K) were calculated using two-way ANOVA tests. Statistics in (C, H) were calculated using one-way ANOVA tests. NS: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Results 8. The nanoparticles are safe for the mice

Finally, we evaluated the biosafety of the nanoparticles. Major organs were collected for H&E staining, and blood samples were analyzed to assess liver and kidney function on day 7 following ICH injury. As shown in Supplementary Fig. 8A, H&E staining revealed no noticeable morphological abnormalities in the brains of ICH mice treated with different nanoparticles. In addition, serum biochemical analyses showed no significant changes in AST, ALT, BUN and CREA levels across different treatment groups (Supplementary Fig. 8B). These results indicate that the nanoparticles are biocompatible and exhibit no apparent toxicity to major organs or liver and kidney function.

Discussion

ICH and TBI are severe neurological disorders that threaten human health and quality of life, and they are two typical forms of brain injury. Previous studies have shown that both ICH and TBI trigger intense inflammatory responses and excessive production of ROS after injury, which in turn exacerbate secondary neuronal damage and worsen clinical outcomes [47–50]. Microglia, as the primary immune cells of the central nervous system, play a key role in initiating and amplifying these inflammatory processes and ROS generation [48, 51]. Notably, ROS and inflammation are tightly interconnected: ROS promotes microglia-mediated neuroinflammation [21], while neuroinflammation further enhances ROS production, forming a vicious cycle [52]. Therefore, targeting microglia-mediated neuroinflammation and generated ROS represents a promising therapeutic strategy for ICH and TBI.

Notably, lactate metabolism plays a critical role in hypoxic conditions. In the brains of ICH and TBI model mice, cells experience severe hypoxia [53, 54], where anaerobic glycolysis becomes the predominant pathway for energy production compared to the tricarboxylic acid (TCA) cycle [55]. However, reduced oxygen availability and increased glycolysis lead to excessive ROS generation and exacerbated neuroinflammatory responses [56, 57]. Recent studies have shown that microglia mainly express MCT1 and MCT4 [26, 27]. Modulation of lactate levels through MCT inhibition can suppress anaerobic glycolysis, thereby reducing M1 microglial polarization and pro-inflammatory cytokine production [27, 58]. Besides, a recent study has used liposomal nanoparticles to deliver agents such as IL-10, PLX5622, and Danshensu to modulate microglia [59–61]. Liposomal nanoparticles thus exhibit excellent biocompatibility and biodegradability, making them ideal for drug delivery. Based on these, we designed a novel liposome nanoparticle that encapsulates the MCT1 inhibitor AR-C155858 and incorporates a ROS-responsive PPS core. In addition, the nanoparticle surface was modified with CAQK and MG1 targeting peptides to facilitate delivery to the injury site and selective uptake by microglia. These nanoparticles protect AR-C155858 from degradation and efficiently deliver it to microglia at brain injury sites. In response to ROS, the PPS core is consumed and AR-C155858 is released to inhibit MCT1.

In this study, we found that the newly developed MiRCM nanoparticles, with favorable biocompatibility, could effectively release the MCT inhibitor AR-C155858 upon exposure to H₂O₂, while protecting it from enzymatic degradation in mouse serum. Furthermore, compared with nanoparticles lacking the CAQK and MG1 target peptides, MiRCM with target peptides exhibited significantly enhanced targeting efficiency toward injured brain regions and M1-type microglia, both in vitro and in vivo, in ICH and TBI models. Owing to the ROS-responsive PPS structure, MiRCM nanoparticles efficiently scavenged ROS, thereby reducing oxidative stress. In addition to their intrinsic antioxidant capacity, MiRCM enabled effective intracellular delivery of AR-C155858 into microglia, inhibiting lactate transport into the extracellular medium rather than lactate production itself. Through this dual mechanism, MiRCM markedly suppressed M1-type microglial polarization and attenuated neuroinflammation in both in vitro and in vivo. Moreover, MiRCM enhanced endogenous antioxidant defenses by increasing the activity of intracellular antioxidant enzymes, including SOD and GSH-Px, while reducing MDA accumulation. Considering that pro-inflammatory cytokines, ROS-induced oxidative stress, and subsequent secondary injury contribute to neuronal death, astrocyte activation, and BBB disruption, we further demonstrated that MiRCM nanoparticles protect neurons from apoptosis, limit excessive astrocyte activation, and preserve BBB integrity. To further elucidate the underlying molecular mechanisms, bulk RNA-sequencing analysis in ICH model revealed that MiRCM nanoparticles downregulate pro-inflammatory signaling and oxidative stress pathways, including key genes such as Il1a, Il1b, Il6, Tnf and Ripk3, Duox2, Mmp9, Alox5, Lcn2, Cd36, Il18rap, indicating the efficiency of MiRCM to target microglia through ROS scavenging and lactate transport modulation. Finally, behavioral assessments confirmed that MiRCM treatment significantly improved motor, sensory, and cognitive functions in ICH- and TBI-induced mice. Collectively, these findings indicate that MiRCM nanoparticles confer comprehensive neuroprotection by modulating M1 microglia lactate transportation and ROS scavenging to suppress neuroinflammation and oxidative stress, thereby attenuating disease progression in both injury models.

However, several limitations of the present study should be acknowledged. First, the therapeutic efficacy of MiRCM nanoparticles was evaluated only in mouse models. Further studies in large-animal models and, ultimately, clinical investigations are necessary to assess their safety, efficacy, and translational potential. Second, although our data suggest that MiRCM exerts protective effects through lactate transportation, the precise molecular mechanisms underlying these regulatory processes require further elucidation through gene modulation. Third, while we focused on acute and subacute outcomes (with the longest follow-up extending to 28 days), longer-term behavioral and histopathological assessments will be conducted to determine the durability and sustained therapeutic benefits of MiRCM treatment.

Conclusion

In summary, this study develops MiRCM nanoparticles that synergistically combine a lactate transport inhibitor (MCT1 inhibitor AR-C155858) with ROS-scavenging PPS for the treatment of ICH and TBI. Owing to their CAQK and MG1 target peptides, MiRCM nanoparticles efficiently accumulate at brain injury sites and in M1-polarized microglia after ICH or TBI, where they scavenge excessive ROS by PPS and simultaneously release AR-C155858 to inhibit M1 microglia-mediated neuroinflammation by modulating lactate transportation. This dual action disrupts the vicious cycle of oxidative stress and neuroinflammation, thereby inhibiting A1 astrocyte activation, preserving neuronal viability and structure, BBB integrity, and promoting long-term neurological recovery in ICH and TBI mice (Fig. 9). Collectively, these findings highlight the therapeutic potential of MiRCM nanoparticles for the ICH and TBI, and provide novel insights into the clinical management of brain injuries.

Fig. 9.

Schematic of MiRCM nanoparticles for ICH and TBI treatment

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Y. M., G. C., P. X. and B. W. contributed equally to this work. L.B., Y.S., and Y. Z. conceived the study and designed the experiments. Y.M., G.C., P.X. and B.W. carried out the experiments. T.T. and Y. Z. assisted with data analysis and analyzed the results. Y.L., Q.Z., and Q.S. discussed the results. Y.M., G.C., T.L. and Z.W. wrote the manuscript. All authors reviewed and approved the final version of the manuscript.

Funding

This study was supported by grants from the Medical Scientific Research Projects of the Jiangsu Provincial Health Commission​​ (MQ2024025 to Y. Z.), and the National Natural Science Foundation of China (82471318 to L.B., 82171283 to Y.S., 82501642 to Y.M. and 82501560 to Y.L.).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

All animal experiments were performed following the guidelines of Shanghai Jiao Tong University and approved by the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University (A2024431-001).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.