Loss of astrocytic Bmal1 promotes blood-brain barrier disruption and synaptic dysfunction during systemic inflammation

Abstract

Circadian rhythm disruption has been associated with the exaggerated inflammatory responses in peripheral tissues; however, its impact on neuroinflammation and blood-brain barrier (BBB) integrity remains unclear. Here, we identify the astrocytic circadian clock as a key regulator of BBB homeostasis during systemic inflammation. In a mouse model, circadian rhythm disruption for three weeks markedly increased BBB permeability in male mice, as evidenced by Evans blue leakage and myeloid cell infiltration into the brain parenchyma following lipopolysaccharide (LPS) challenge. Transcriptomic analyses using public datasets revealed that astrocytes exhibit the highest expression of core circadian clock genes among brain cell types. Accordingly, we generated tamoxifen-inducible, astrocyte-specific Bmal1-knockout (KO) mice. Deletion of Bmal1 in astrocytes significantly enhanced BBB leakage, astrogliosis and pericyte loss after LPS administration. Mechanistically, Bmal1-deficient astrocytes produced elevated levels of the chemokine CXCL5, which promoted CXCR2-dependent neutrophil recruitment into the brain. Pharmacological blockade of CXCR2 with SB225002 restored pericyte coverage and attenuated BBB disruption in astrocytic Bmal1 KO mice. Functionally, these mice exhibited impaired excitatory synaptic transmission following systemic inflammation, suggesting that astrocytic Bmal1 loss compromises neurovascular and synaptic integrity. Taken together, our findings demonstrate that astrocytic Bmal1 maintains BBB integrity and synaptic stability under inflammatory stress. This work also highlights astrocyte-intrinsic circadian regulation as a critical mechanism linking chemokine production to neurovascular vulnerability.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12974-026-03841-z.

Introduction

Circadian rhythm is an endogenous oscillation of physiological processes that follows a 24-hour cycle [1]. In mammals, circadian rhythm is maintained by cell-autonomous molecular clocks, organized as an autoregulatory transcription–translation feedback loop [2]. At its core, transcription factors Brain and Muscle Arnt-like 1 (BMAL1; also known as ARNTL) and Circadian Locomotor Output Cycles Kaput (CLOCK) form a heterodimer that binds to E-box elements to initiate transcription of clock-controlled genes, including Period (PER) and Cryptochrome (CRY) [3]. PER and CRY proteins in turn inhibit BMAL1/CLOCK activity, establishing a negative feedback loop that generates circadian rhythmicity. Disruption of circadian rhythms, whether by environmental or genetic factors, has been linked to impaired immune function and increased risk of inflammatory diseases [4]. Indeed, symptoms of inflammatory disorders such as atopic dermatitis and asthma fluctuate with time of day [5, 6]. As circadian regulatory mechanisms operate across diverse cell populations, including immune cells, glial cells, and neurons [7–9], elucidating their role in immune regulation may provide novel strategies for treating inflammatory diseases.

The blood–brain barrier (BBB) is a highly specialized interface that separates the brain parenchyma from the circulating blood [10]. Its architecture consists of endothelial cells connected by tight junctions, supported by pericytes and ensheathed by astrocytic end-feet. This multilayered structure ensures selective exchange between the brain and the bloodstream, thereby protecting brain parenchymal regions from peripheral toxins and pathogens. Severe systemic inflammation, as seen in infection or sepsis, can disrupt BBB integrity and lead to sepsis-associated encephalopathy [11]. When the BBB is compromised, peripheral immune cells and inflammatory mediators infiltrate the brain parenchyma, exacerbating neuroinflammation and neuronal injury [12]. Conversely, the restrictive nature of the BBB also limits effective drug delivery for central nervous system (CNS) disorders such as glioma, Alzheimer’s disease, and multiple sclerosis [13]. Thus, the BBB represents both a key contributor to neurological disease pathogenesis and a critical barrier to therapy.

Importantly, brain-resident cells also harbor intrinsic circadian clocks, which have been implicated in pathologies including cognitive decline and neurodegeneration [14]. Recent studies indicate that BBB permeability exhibits circadian oscillation over the 24-hour cycle [15]. However, it remains unclear whether these rhythmic changes are mediated by cell-autonomous circadian clocks or by other mechanisms. In this study, we sought to investigate whether circadian clocks regulate BBB permeability under systemic inflammatory conditions. Using a mouse model of lipopolysaccharide (LPS)-induced systemic inflammation [16], we hypothesized that BBB permeability would vary according to the time of day and circadian disruption. In particular, we aimed to define the role of cell-specific BMAL1 in modulating BBB integrity during peripheral inflammation.

Methods

Mice

C57BL/6 mice were purchased from Orient Bio. Bmal1^flox/flox (JAX #007668) and Aldh1l1^CreERT2 (JAX #029655) mice were obtained from The Jackson Laboratory. To generate astrocyte-specific Bmal1deficient mice, Bmal1^flox/flox mice were crossed with Aldh1l1^CreERT2 mice. All genetically modified mice were maintained on a C57BL/6 background, and genotypes were confirmed by polymerase chain reaction (PCR). Mice were housed in a specific pathogen-free facility at Yonsei University College of Medicine under controlled temperature and humidity, on a 12 h–12 h light-dark cycle (lights on at 8:00 a.m. and off at 8:00 p.m.) to ensure a stable circadian environment. Age matched male and female mice (7–12 weeks old) were used for the experiments. All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Yonsei University College of Medicine (2022 − 0121) or DGIST (DGIST-IACUC-25082201-0001). All experiments were conducted in accordance with the approved guidelines of the Institutional Ethical Committee.

Mice treatment

To induce astrocyte-specific Bmal1 deletion, Aldh1l1^CreERT2/+; Bmal1^flox/flox mice were administered with tamoxifen (75 mg/kg, intraperitoneally) dissolved in corn oil (Sigma-Aldrich, C8267) once daily for five consecutive days. Tamoxifen-treated Bmal1^flox/flox littermates were used as controls. For induction of systemic inflammation, mice received an intraperitoneal injection of lipopolysaccharide (LPS, 2.5, 4, or 10 mg/kg; Sigma-Aldrich, L3012) and were euthanized at 6–24 h after injection for analysis. Phosphate-buffered saline (PBS)-injected mice served as controls. Unless otherwise indicated, all LPS administrations were performed between ZT0 and ZT2. To inhibit neutrophil recruitment, mice were treated with the CXCR2 antagonist SB225002 (2 mg/kg, intraperitoneally, dissolved in a vehicle containing 2% DMSO, 2% Tween 80, and 20% PEG300) once daily for five consecutive days before LPS injection. For neutrophil depletion experiments, mice were administered either isotype control IgG or anti-Ly6G antibody (100 µg/mouse, intraperitoneally) 24 h prior to LPS challenge.

Circadian rhythm disruption model

The experimental design is illustrated in Fig. 1A. C57BL/6 mice were housed in light-tight cages equipped with Leddy programable illumination control, with food and sterilized water provided ad libitum. Circadian rhythm disruption (CRD) was induced by advancing the light phase by 6 h every 2 days for three consecutive weeks, as previously described [17]. Control mice were housed under standard 12 h light/12 h dark (LD) conditions in identical light-tight cages for the same duration.

Fig. 1.

Circadian regulation of blood-brain barrier integrity in response to systemic inflammation. A Experimental scheme of light cycle manipulation. Male mice were subjected to either a standard 12 h: 12 h light-dark cycle (LD group) or a shifting light schedule with 6 h advances every 2 days (CRD group). Following 3 weeks of LD or CRD exposure, brains were collected 6 h after PBS or LPS (4 mg/kg) administration. B, C Representative brain images (B) and quantification (C) of Evans blue extravasation in male mice treated with PBS or LPS under LD or CRD conditions as shown in (A). (n = 5, PBS group; n = 6, LPS group) D-G Brain-infiltrating immune cells were analyzed by flow cytometry with the gating strategy (D) and quantification of CD45^hi CD11b^hi myeloid cells (E), Ly6G⁺ neutrophils (F), and Ly6C^hi monocytes (G) (n = 4, PBS group; n = 8, LPS group). H Experimental scheme of time-dependent BBB permeability. Male mice received LPS (2.5 mg/kg) at light onset (ZT0) or light offset (ZT12). I, J Representative brain images (I) and quantification (J) of Evans blue leakage showing time-of-day dependent BBB permeability after LPS injection at ZT0 or ZT12 (n = 9 per group). K-M Quantification of brain-infiltrating CD45^hi CD11b^hi myeloid cells (K), Ly6G⁺ neutrophils (L), and Ly6C^hi monocytes (M) (n = 10 per group). Asterisks indicate significant differences. *P < 0.05, **P < 0.01, ***P < 0.001

Reagents and antibodies

LPS (L3012), tamoxifen (T5648), sodium fluorescein (F6377), Evans blue (E2129), and trichloroacetic acid (T6399) were purchased from Sigma-Aldrich. CXCR2 antagonist SB225002 (S7651) was obtained from Selleckchem. Anti-mouse Ly6G (BE0075-1) and rat IgG2 isotype control (BE0089) were acquired from Bio X cell. Cy3-conjugated anti-rabbit (111-165-003), anti-rat (712-165-050) or anti-mouse (115-167-185) antibodies; Alexa Fluor 488-conjugated anti-rat (712-545-150) or anti-goat (205-545-108) antibodies; and fluorescein isothiocyanate (FITC)-conjugated anti-guinea pig (706-095-148) or anti-rabbit (111-095-046) antibodies were purchased from Jackson ImmunoResearch. Fluorochrome-conjugated monoclonal antibodies against CD45 (BioLegend, 103116), CD11b (Invitrogen, 12-0112-82), Ly6G (Invitrogen, 17-9668-82), Ly6C (Invitrogen, 53-5932-82), ACSA-2 (Miltenyi, 130-117-535), CD3 (BioLegend, 100204) and NK1.1 (BioLegend, 156506) were purchased from the indicated suppliers. For flow cytometric analysis, all antibodies were validated in our laboratory based on the manufacturer’s recommendations. For immunohistochemistry, antibody specificity was confirmed using controls stained with secondary antibodies alone.

Brain isolation and processing

Mice were anesthetized with isoflurane and transcardially perfused through the left ventricle with 20 ml of ice-cold PBS. Brains were carefully isolated, excluding the meninges, olfactory bulbs, brainstem, and cerebellum, and immediately processed for analysis. In some experiments, cortical and hippocampal regions were further dissected and processed separately.

BBB permeability assay

Blood–brain barrier (BBB) permeability was assessed by measuring extravascular leakage of Evans blue and sodium fluorescein into brain parenchyma. Mice were injected intravenously via the retro-orbital sinus with Evans blue (0.2 mg/kg) or sodium fluorescein (100 mg/kg) 1 h prior to euthanasia. Under deep isoflurane anesthesia, mice were perfused with PBS, and brains were collected. Brain tissues were homogenized in saline using a Dounce homogenizer, followed by addition of an equal volume of trichloroacetic acid (TCA). Samples were rotated for 1 h and centrifuged at 10,000 g for 30 min at 4 °C. The resulting supernatants were collected for fluorescence quantification. For Evans blue, supernatant was mixed with 95% ethanol; for sodium fluorescein, supernatant was mixed with 1.25 M NaOH in a clear 96-well plate. Fluorescence intensity was measured using a Varioskan Flash 3001 microplate fluorometer (Thermo Fisher Scientific) at 620/680 nm (excitation/emission) for Evans blue and 460/515 nm for sodium fluorescein. Concentrations were calculated based on standard curves for each dye and normalized to brain tissue weight, expressed as micrograms of dye per gram of tissue.

Flow cytometry

Harvested brain tissues were homogenized in RPMI-1640 medium containing DNase I and collagenase IV using a Dounce homogenizer. The homogenates were incubated at 37 °C for 1 h with gentle agitation and filtered through a 70 μm cell strainer. Myelin debris was removed by centrifugation on a 40% Percoll gradient. The resulting single-cell suspension was stained with fluorochrome-conjugated monoclonal antibodies against CD45, CD11b, Ly6C, and Ly6G. Cell viability was assessed using 4′,6-diamidino-2-phenylindole (DAPI) to exclude dead cells. Flow cytometric analysis was performed on a FACSVerse cytometer (BD Biosciences), and data were analyzed using FlowJo software (TreeStar).

Astrocyte isolation and culture

Primary astrocyte cultures were prepared from brains of postnatal day 0–3 (P0–P3) mouse pups; tail biopsies were collected for genotyping. Dissected brain tissues were gently triturated in ice-cold PBS using a fire-polished glass pipette, filtered through a 100 μm cell strainer, centrifuged (700 g, 3 min), and resuspended in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS), L-glutamine, and 1% penicillin–streptomycin. Cells were maintained at 37 °C in a humidified 5% CO₂ incubator, and the culture medium was replaced every 2–3 days. After 12–13 days in vitro, flasks were shaken at 300 rpm for 6 h to remove non-astrocytic cells, followed by mild trypsinization to detach adherent cells. The collected adherent fraction was used as enriched primary astrocytes. For astrocyte isolation from adult mouse brains, single-cell suspensions were labeled with ACSA-2 antibody, and ACSA-2 positive cells were gated and sorted using a FACS Aria Ⅱ (BD Biosciences).

Neutrophil migration assay

Bone marrow cells were flushed from the femurs and tibias of adult mice. Neutrophils were isolated by negative selection using the MojoSort™ Mouse Neutrophil Isolation Kit (BioLegend, 480058). Enriched neutrophils (4 × 10⁵ cells) were seeded into the upper inserts of 24-well transwell plates (pore size = 8 μm). Conditioned medium was prepared from astrocyte cultures treated with 4-hydroxytamoxifen (4-OHT) to delete Bmal1, followed by LPS stimulation (250 ng/mL, 3 h). Supernatants were centrifuged to remove debris and used as conditioned media in the lower chamber. After 18 h of incubation at 37 °C, migrated neutrophils in the lower chamber were collected and counted using a hemocytometer.

Immunohistochemistry

Mice were deeply anesthetized with inhalation of isoflurane and immediately perfused, first with cold 1X PBS for 3 min and then with 4% paraformaldehyde (PFA) for 4 min. Brains were dissected, fixed in 4% PFA overnight and sliced into 40–50 μm-thick coronal sections using a vibratome (VT1200S; Leica) or Leica CM1860 cryostat. Brain sections were permeabilized with 0.2 or 0.5% Triton X-100 in PBS containing 5% bovine serum albumin and 5% horse serum for 1 h at room temperature (RT). For immunostaining, brain sections were incubated overnight at 4 °C with primary antibodies against GFAP (Invitrogen, PA1-10004, 1:2000 or Invitrogen; Cat# 13–0300, 1:1000), PECAM-1 (BD biosciences; Cat# 550274, 1:400), CD13 (R&D systems; Cat# AF2335-SP, 1:400), Claudin-5 (Invitrogen; Cat# 34-1600, 1:200), Iba1 (Wako, 019-19741, 1:500), VGLUT1 (Millipore; Cat# AB5905; 1:200), PSD-95 (JK016; 1:300) [18], VGAT (Synaptic Systems; Cat# 131 003; 1:500), and gephyrin (Synaptic Systems; Cat# 147 111; 1:500). Sections were washed with PBS and incubated with species-appropriate Cy3-, Alexa Fluor 488-, or FITC-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) for 2 h at room temperature. Nuclei were counterstained with DAPI, and sections were mounted with mounting medium (Invitrogen, P36934; H-1200; Vector Laboratories). Fluorescent images were acquired using confocal microscopes (LSM700, LSM780, and LSM800, Carl Zeiss) and processed with Zen Blue software (Zeiss). Morphological analysis of GFAP-positive astrocytes was performed using ImageJ. Astrocytic processes were skeletonized, and the maximum branch length and total number of branches were quantified. Synaptic puncta were quantified using MetaMorph software (Molecular Devices), and their density and average area were measured.

Quantification of mRNA

Total RNA was isolated from brain tissues using TRIzol reagent (Invitrogen, 15596018) according to the manufacturer’s instructions. Complementary DNA (cDNA) was synthesized from the extracted RNA using the PrimeScript RT Master Mix Kit (Takara, RR036A). Quantitative real-time PCR (qRT-PCR) was performed on a QuantStudio 3 Real-Time PCR System (Applied Biosystems, USA) using SYBR Green I Master Mix (ABclonal, RK21203). Relative mRNA expression levels were calculated using the comparative ΔΔCt method after normalization to Rn18s. Primers were as follows: 5’-GCT GGG ATT CAC CTC AAG AA-3’ and 5’-TGG GGA CAC CTT TTA GCA TC-3’ (Cxcl1); 5’-GTT TCT GGG GAG AGG GTG AG-3’ and 5’-TGT TCT ACT CTC CTC GGT GC-3’ (Cxcl2); 5’-GCC CTA CGG TGG AAG TCA TA-3’ and 5’-AGT GCA TTC CGC TTA GCT TT-3’ (Cxcl5); 5’-AGG TGT CCC AAA GAA GCT GT-3’ and 5’-ACA GAA GTG CTT GAG GTG GT-3’ (Ccl2); 5’-TTT CTT CTA TGC GCA GTT GG-3’ and 5’-GCA GTT TGG TGC CTA CTT CA-3’ (Cldn5); 5’-ACA GTC CAA TGG CCT ACT CC-3’ and 5’-ACT TCA GGC ACC AGA GGT GT-3’ (Ocln); 5’-CCA CCT CTG TCC AGC TCT TC-3’ and 5’-CAC CGG AGT GAT GGT TTT CT-3’ (Tjp1); 5’-TGG TGG AGC TAC GAT GCC AG-3’ and 5’-AGC TGT TGT TGT GTG TGC CG-3’ (Jam2); 5’-AGA AAG GTT GAA TCG CTG GA-3’ and 5’- CGG CGA TAG TCG TTA GCT TC -3’ (Gfap); 5’-TTC ACA CTG AAT GCC AGC TC-3’ and 5’-GTC TGC TGA GAC CCC TCT TG-3’ (Icam1); 5’-ATT TTC TGG GGC AGG AAG-3’ and 5’-ACG TCA GAA CAA CCG AAT CC-3’ (Vcam1); 5’-ATT GAG ACA GAC CCC AA CG-3’ and 5’-TTC TGG TTT TCT GGC AGC TT-3’ (Cdh5); 5’-CGC GGT TCT ATT TTG TTG GT-3’ and 5’-AGT CGG CAT CGT TTA TGG TC-3’ (Rn18s).

Enzyme-linked immunosorbent assay (ELISA)

Brain tissues were homogenized in buffer containing 10 mM Tris-Cl (pH 7.4), 150 mM NaCl, 2 mM EDTA, and protease inhibitors using a Dounce tissue grinder. Homogenates were centrifuged, and the resulting supernatants were collected for analysis. Total protein concentrations were determined using the Bradford assay. For primary astrocyte cultures, conditioned media were collected and centrifuged to remove cellular debris. The concentrations of CXCL1, CXCL2, and CXCL5 were quantified using commercially available ELISA DuoSet kits (R&D Systems) according to the manufacturer’s instructions.

Ex vivo electrophysiological recordings

Male mice were anesthetized with isoflurane and decapitated to remove the brain. The brains were quickly transferred in an ice-cold slicing solution (3.3 mM KCl, 1.3 mM NaH₂PO₄, 26 mM NaHCO₃, 11 mM D-glucose, 211 mM sucrose, 0.5 mM CaCl₂ and 10 mM MgCl₂). Brain slices (300 μm) were prepared using a vibrating-knife microtome (VT1000S; Leica Microsystems). Brain slices were stabilized for at 37 °C for at least 1 h in an oxygenated (95% O₂/5% CO₂) artificial cerebrospinal fluid (124 mM NaCl, 3.3 mM KCl, 1.3 mM NaH₂PO₄, 26 mM NaHCO₃, 11 mM D-glucose, 2 mM CaCl₂ and 1 mM MgCl₂). Slices were transferred to the recording chamber and perfused in an oxygenated (95% O₂/5% CO₂) artificial cerebrospinal fluid solution. Patch electrodes (3–8 MΩ) were prepared with capillary glass (Warner Instruments, USA) using a micropipette puller (P-97, Sutter Instrument, CA). Glass pipettes (2–5 MΩ) were filled with an intracellular solution containing (in mM): 145 mM CsCl, 5 mM NaCl, 10 mM HEPES, 10 mM EGTA, 4 mM Mg-ATP, and 0.3 mM Na-GTP, adjusted to pH 7.3 with CsOH and was used for the measurement of miniature inhibitory postsynaptic currents (mIPSCs) and miniature excitatory postsynaptic currents (mEPSCs). The osmolarity of the internal solution was 280–290 mOsm. Cells were voltage-clamped at − 70 mV. The mIPSCs were isolated by blocking Na⁺ channels, NMDARs, and AMPARs with bath-applied tetrodotoxin (TTX, 1 µM), D-2-amino-5-phosphonopentanoate (D-AP5, 50 µM) and cyanquixaline (CNQX, 10 µM), respectively. The mEPSCs were isolated by blocking Na⁺ channels and GABA_A receptors by bath-applying TTX (1 µM) and picrotoxin (PTX, 50 µM), respectively. All recordings were performed using a Multiclamp 700B amplifier and a Digidata 1550B digitizer (Molecular Devices). Only cells with access resistance (Ra) in the range of 5 MΩ < Ra < 30 MΩ were analyzed.

Data analysis

Publicly available transcriptomic datasets were obtained from the NCBI Gene Expression Omnibus (GEO), including human brain single-nucleus RNA-seq data (GSE163577; healthy donor subset) and mouse brain single-cell RNA-seq data (GSE263094; WT PBS subset). Bioinformatic analyses were performed using R 4.4.1 within RStudio (R Foundation for Statistical Computing, Vienna, Austria). Data preprocessing followed the Seurat workflow using Seurat v4.3.0. Low-quality and outlier cells were filtered based on the number of detected genes and the proportion of mitochondrial transcripts. Subsequent normalization, scaling, dimensionality reduction, and Louvain clustering were conducted using Seurat built-in functions with dataset-specific parameters. Differential gene expression was assessed to annotate clusters, and cell-type identities were assigned based on canonical marker genes reported in the literature.

Statistical analysis

All data are presented as mean ± standard error of the mean (s.e.m.). The number of biological replicates (n) is indicated in each figure legend. Statistical analyses were performed using R 4.4.1 or Prism 10 (GraphPad Software). The normality of data distributions was evaluated using the D’Agostino & Pearson test. Comparisons between two groups were made using unpaired two-tailed Student’s t-tests or Mann-Whitney U test. For multiple group comparisons, one-way or two-way analysis of variance (ANOVA) followed by Tukey’s post hoc test or Kruskal-Wallis test was used. A P ≤ 0.05 was considered statistically significant.

Results

Systemic inflammation-induced BBB disruption is dependent on circadian rhythm

To investigate whether BBB permeability is regulated by circadian rhythms, we employed a circadian disruption mouse model. Under normal conditions (12 h light/ 12 h dark, termed LD), mice were maintained on a standard light-dark cycle. In contrast, under circadian rhythm disruption (CRD) conditions, the light phase was advanced by 6 h every 2 days for three weeks (Fig. 1A). To induce mild BBB disruption, mice received an intraperitoneal (i.p.) injection of lipopolysaccharide (LPS, 4 mg/kg), and BBB permeability was assessed by quantifying extravascular Evans blue leakage. Evans blue was administered intravenously 1 h before sacrifice, and brain parenchyma was analyzed 6 h after LPS injection (Supplementary Fig. S1A). LPS treatment significantly increased Evans blue extravasation into the brain (Supplementary Figs. S1B and C). Notably, LPS-induced BBB permeability was further enhanced under CRD than ND conditions in male mice (Figs. 1B and C). By contrast, in female mice, LPS (4 mg/kg) did not induce significant BBB disruption under either LD or CRD conditions (Supplementary Figs. S1D and E).

We next examined the infiltration of circulating immune cells into the brain following peripheral inflammation (Fig. 1D). LPS injection elicited a robust infiltration of myeloid cells (CD45^hiCD11b^hi) into the brain parenchyma of male mice, and this response was significantly greater under CRD compared with LD conditions (Fig. 1E). Consistently, LPS-induced infiltration of neutrophils and monocytes was increased under CRD relative to LD condition (Figs. 1 F, G), although the increase in monocytes did not reach statistical significance. In contrast to myeloid cells, LPS administration did not induce significant infiltration of lymphocytes, including NK cells or T cells, into the brain parenchyma (Supplementary Figs. S1F and G).

We then investigated whether the time of day of LPS stimulation influenced BBB permeability. To this end, LPS injections were administered at zeitgeber time (ZT) 0 or ZT 12 (Fig. 1H). LPS treatment at ZT 12 induced greater Evans blue leakage (Figs. 1I and J) and myeloid cells infiltration (Figs. 1K-M) in male mice, whereas no significant effects were observed in female mice (Supplementary Fig. S1H and I). Together, these findings indicate that peripheral inflammation-induced BBB disruption is regulated in a circadian rhythm-dependent manner.

Deletion of Bmal1 in astrocytes enhances peripheral inflammation-induced BBB disruption

The above results linking circadian rhythm to BBB permeability suggest that intrinsic circadian clocks in peripheral immune cells or brain-resident cells may influence BBB disruption. To assess the contribution of brain-resident circadian clocks to BBB integrity, we examined the expression of circadian clock genes in brain cell types using Gene Expression Omnibus (GEO) datasets. Two independent transcriptomic analyses [19, 20] revealed that astrocytes displayed the highest core clock genes (CCG) scores among brain-resident cells (Figs. 2 A, B; Supplementary Figs. S2A, B). Based on these findings, we generated astrocyte-specific Bmal1-deficient mice, given that Bmal1 is a key regulator of the molecular clock machinery [21].

Fig. 2.

Circadian disruption in astrocytes exacerbates BBB breakdown in response to peripheral inflammation. A, B Human brain single-nucleus RNA sequencing was reanalyzed using publicly available datasets (GSE163577). A A heatmap presents the proportion of brain cell populations expressing core clock genes (CCG). B A violin plot displays the CCG score calculated from the total normalized expression of core clock genes. VSMC, vascular smooth muscle cells; OPCs, oligodendrocyte precursor cells. C Experimental scheme of BBB permeability assay. All male mice were treated with tamoxifen and one week later Control (Bmal1^flox/flox; referred to as CTRL) and astrocyte-specific Bmal1 knockout (Aldh1l1^CreERT2;Bmal1^flox/flox; referred to as AS-B1KO) mice were sacrificed 6 h after PBS or LPS (4 mg/kg) injection. D, E Representative brain images (D) and quantification of Evans blue (E) or Sodium fluorescein (NaFl) F extravasation in mice treated as shown in (C). (E, n = 5, CTRL-PBS; n = 6, AS-B1KO-PBS; n = 7, LPS group: F, n = 6 per group). G Representative flow cytometry analysis depicting the population of neutrophils and monocytes. H-J Quantification of CD45hi CD11bhi myeloid cells (H), Ly6G+ neutrophils (I), and Ly6Chi monocytes (J) in the brain of mice treated as shown in (C). (n = 7, CTRL-PBS; n = 6, AS-B1KO-PBS; n = 13, LPS group) K, L Representative brain images (K) and quantification of Evans blue leakage (L) in CTRL and AS-B1KO mice following LPS treatment at ZT0 or ZT12 (n = 8, CTRL-LPS at ZT0; n = 7, KO-LPS at ZT0; n = 8, LPS at ZT12 group). Asterisks indicate significant differences. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. not significant

By crossing Aldh1l1-CreERT2 mice with Bmal1^fl/fl mice, we generated Aldh1l1^CreERT2/+; Bmal1^fl/fl mice (Supplementary Fig. S2C), in which Bmal1 deletion was efficiently induced in ACSA2⁺ astrocytes following tamoxifen treatment (Supplementary Figs. S2D–G). Ablation of astrocytic Bmal1 markedly increased extravascular leakage of Evans blue and sodium fluorescein (NaFl), another commonly used tracer of BBB permeability, at 6 h post LPS administration (Figs. 2 C–F). Consistently, peripheral inflammation–induced infiltration of myeloid cells into the brain parenchyma was significantly elevated in astrocyte-specific Bmal1 knockout (AS-B1 KO) male mice compared with control (CTRL) littermates (Figs. 2G–J). Notably, the increased Evans blue leakage in astrocytic Bmal1 deficiency was observed following LPS stimulation at ZT 0, but not at ZT12 (Figs. 2K and L), indicating that the time-of-day-dependent effect of LPS stimulation is diminished in the absence of astrocytic Bmal1.

We next examined whether enhanced BBB disruption in Bmal1-deficient mice is observed in females. As described above, LPS (4 mg/kg) injection did not induce significant BBB permeability in female mice under either genotype (Supplementary Figs. S3A and B). In contrast, administration of a higher LPS dose (10 mg/kg) elicited substantial Evans blue and NaFl leakage, which was further exacerbated in astrocytic Bmal1-deficient female mice (Supplementary Figs. S3A and B). Consistent with this finding, BBB permeability was increased following LPS treatment at ZT 12 compared with ZT 0 in female mice under higher LPS stimulation (Supplementary Figs. S3C and D). Similarly, enhanced BBB disruption under CRD conditions was also observed in female mice at higher LPS doses (Supplementary Figs. S3E and F). Although we did not assess circadian effects on myeloid cells’ infiltration in female mice, these findings suggest that female mice are relatively more resistant to peripheral inflammation–induced BBB disruption, while the circadian-dependent increase in BBB permeability is conserved across sexes. Based on this difference in LPS sensitivity, we used male mice for subsequent experiments examining the circadian regulation of BBB integrity.

Deletion of Bmal1 in astrocytes induces astrogliosis and pericyte degeneration

We next examined whether Bmal1 deficiency affects the abundance of BBB-composing cells in response to peripheral inflammation. At 24 h after LPS injection, infiltration of myeloid cells remained elevated in the whole brain (Figs. 3A-D), hippocampal region (Figs. 3E and F), and cortical region (Supplementary Fig. S4A and B) of astrocytic Bmal1 KO mice compared with control mice. Unlike neutrophils, the Bmal1-dependent increase in monocyte infiltration did not reach statistical significance at 24 h post LPS treatment (Fig. 3D).

Fig. 3.

Astrocytic Bmal1 deficiency increases immune cell infiltration and astrogliosis. A Brains were collected 24 h after PBS or LPS (4 mg/kg) injection in CTRL and AS-B1KO mice. B-D Quantification of brain-infiltrating myeloid cells (B), Ly6G+ neutrophils (C), and Ly6Chi monocytes (D) in the brain of mice treated as in (A). (n = 6, CTRL group; n = 7, AS-B1KO group) E, F Quantification of brain-infiltrating myeloid cells (E) and Ly6G+ neutrophils (F) in the hippocampal region of mice brain treated as in (A). (n = 9, CTRL group; n = 10, AS-B1KO group) G Representative immunofluorescence images of the hippocampal CA1 regions of LPS-injected CTRL and AS-B1KO mice. The brain coronal sections were stained with anti-GFAP (green) and anti-Iba1 (red). DAPI represents nuclei signal (blue). H Quantification of relative GFAP+ and IBA1+ areas per field of view. (n = 7, CTRL group; n = 6, AS-B1KO group) I, J Representative immunofluorescence images (I) of cortical region of LPS-injected CTRL and AS-B1KO mice as stained by anti-CD13 (cyan) and anti-PECAM-1 (magenta) antibodies and quantification of PECAM-1+ endothelial cells (J) in the images. (n = 7, CTRL group; n = 6, AS-B1KO group) K Quantification of Cldn5 mRNA levels in the brain tissue from LPS-injected CTRL and AS-B1KO mice. (n = 6 per group) L Representative immunofluorescence images of cortical region of LPS-injected CTRL and AS-B1KO mice as stained by anti-Claudin-5 antibody. M Quantification of Claudin-5-stained areas in the images. (n = 7, CTRL group; n = 6, AS-B1KO group) N Quantification of CD13 + pericyte areas with PECAM-1 + endothelial cells in the cortical regions of LPS-injected CTRL and AS-B1KO mice as shown in (I). (n = 7, CTRL group; n = 6, AS-B1KO group) Scale bars, 50 μm. Asterisks indicate significant differences. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. not significant

Within the hippocampus, the relative area of GFAP-positive astrocytes was increased in Bmal1 KO mice, whereas the area of Iba1-positive microglia was unchanged between genotypes (Figs. 3G-H; Supplementary Fig. S5A). To further characterize astrocyte activation, we analyzed the morphology of GFAP-positive astrocytes. LPS-primed Bmal1-deficient astrocytes exhibited a modest increase in branch length compared with controls (Supplementary Fig. S5B and C), while branch number remained unchanged (Supplementary Fig. S5D). Notably, astrocytic Bmal1 ablation also led to increased mRNA expression of Gfap, a canonical marker of astrocyte activation, in response to LPS stimulation (Supplementary Fig. S5E).

The density of endothelial cells, assessed by PECAM1 staining, was not altered in the cortical regions of Bmal1-deficient mice (Figs. 3I and J). Consistently, ablation of Bmal1 did not affect the expression of tight junction proteins, including Claudin-5, at either mRNA or protein level (Figs. 3K-M and Supplementary Figs. S6A-C). Moreover, Bmal1 deficiency did not alter the expression of major adhesion molecules in brain extracts (Supplementary Figs. S6D-F). In contrast, the population of CD13-positive pericytes associated with endothelial cells was significantly reduced in the cortical regions of Bmal1 KO mice compared with controls (Figs. 3I and N). These findings indicate that astrocytic BMAL1 contributes to BBB homeostasis during peripheral inflammation by regulating astrocyte activation and maintaining pericyte integrity.

Astrocytic Bmal1 deficiency enhances LPS-induced CXCL5 chemokine production

We next investigated how astrocytic Bmal1 influences neutrophil infiltration into the brain parenchyma during peripheral inflammation. To analyze astrocyte-derived chemokines, astrocytes were isolated from mixed glial cultures prepared from neonatal mouse brains (Supplementary Fig. S7A). Bmal1 expression was then deleted in Aldh1l1-positive astrocytes by treatment with 4-hydroxytamoxifen (4-OHT) (Supplementary Fig. S7B and C). Control and Bmal1-deficient astrocytes were stimulated with LPS, and their culture supernatants (conditioned medium) were subsequently cocultured with bone marrow-derived neutrophils in a transwell co-culture assay (Fig. 4A and Supplementary Fig. S7D).

Fig. 4.

Astrocyte-specific Bmal1 deletion drives neutrophil migration through enhanced chemokine secretion following LPS challenge. A Schematic illustration of the neutrophil migration assay. The lower chamber contained conditioned media (CM) collected after LPS treatment (250 ng/mL) from control or Bmal1-deficient astrocyte cultures. Bone marrow-derived neutrophils were seeded into the upper insert. B Quantification of migrated neutrophils into the lower chamber in (A). (n = 6 per group) C Quantification of CXCL1, CXCL2, CXCL5 in the conditioned media collected from PBS- or LPS (250 ng/mL)-treated CTRL or AS-B1KO astrocytes. (n = 6, PBS group; n = 8, LPS group) D Experimental scheme for the analysis of brain tissue or ACSA2⁺ astrocytes. LPS, 4 mg/kg. E Quantification of chemokine protein levels in the brain tissues as assessed by ELISA. (n = 5 per group) F Quantification of chemokine mRNA levels in the sorted ACSA2 + astrocytes. (n = 6 per group) G Quantification of CXCL5 protein levels in the brain homogenates from AS-B1KO mice following LPS injection at ZT0 or ZT12 (n = 6 per group). Asterisks indicate significant differences. *P < 0.05, **P < 0.01, n.s. not significant

Neutrophil migration from the transwell insert to the lower chamber was then quantified. Conditioned medium from Bmal1-deficient astrocytes induced significantly greater neutrophil migration compared with that from control astrocytes (Fig. 4B). Given that brain-infiltrating neutrophils mainly express the chemokine receptor CXCR2 [20], we then analyzed CXCL chemokines in the astrocyte-conditioned medium. LPS treatment induced robust production of CXCL1 and CXCL2 by astrocytes, but levels were comparable between control and Bmal1-deficient cells (Fig. 4C). In contrast, CXCL5 production was significantly elevated in Bmal1-deficient astrocytes relative to controls (Fig. 4C), suggesting that BMAL1 negatively regulates CXCL5 expression in astrocytes.

To validate this finding in vivo, we examined CXCL chemokine expression in whole brain extracts from control and astrocytic Bmal1 KO mice 6 h after PBS or LPS injection (Fig. 4D). LPS-induced CXCL1 and CXCL2 levels were comparable between genotypes, whereas CXCL5 levels were significantly elevated in Bmal1-deficient mice (Fig. 4E). Consistent with these results, analysis of sorted astrocytes from control and Bmal1 KO mice confirmed that CXCL5 was selectively upregulated in the absence of Bmal1, while other chemokines remained unchanged (Fig. 4F). Notably, the increased CXCL5 production in Bmal1-deficient mice was not influenced by the time of LPS administration (Fig. 4G), consistent with the loss of time-of-day-dependent regulation of BBB permeability (Fig. 2L). Together, these findings demonstrate that astrocytic Bmal1 deficiency selectively enhances CXCL5 production upon LPS stimulation, thereby promoting CXCL5-dependent neutrophil chemotaxis into the brain.

CXCR2 blockade mitigates neutrophil-driven BBB disruption in astrocytic Bmal1-deficient mice

We next tested whether enhanced neutrophil chemotaxis contributes to the increased BBB permeability in astrocytic Bmal1 KO mice. To inhibit neutrophil infiltration into the brain, mice were treated with CXCR2 antagonist SB225002 (Fig. 5A). Consistent with earlier results (Fig. 2), astrocytic Bmal1 KO mice displayed increased infiltration of myeloid cells and enhanced recruitment of neutrophils into the brain following LPS injection compared with control mice (Figs. 5B and C). Treatment with SB225002 markedly reduced the accumulation of these cells in the brain parenchyma of Bmal1 KO mice (Figs. 5B and C). Alongside reduced immune cell extravasation, SB225002 treatment also significantly decreased Evans blue leakage in astrocytic Bmal1 KO mice (Figs. 5D and E), indicating a critical role for neutrophil recruitment in the exacerbated BBB disruption associated with Bmal1 deficiency.

Fig. 5.

CXCR2 blockade attenuates systemic inflammation-induced BBB permeability in astrocytic Bmal1 knockout mice. A Schematic illustration of the experimental procedure. CTRL and AS-B1KO mice received vehicle or CXCR2 antagonist SB225002 (2 mg/kg) once daily for 5 days, followed by LPS (4 mg/kg) treatment. B, C Quantification of brain-infiltrating CD45hi CD11bhi myeloid cells (B) and Ly6G+ neutrophils (C) at 6 h post LPS injection. (n = 6 per group) D, E Representative brain images (D) and quantification of Evans blue extravasation (E) in mice at 6 h post LPS injection. (n = 5 per group) F Quantification of GFAP+ area per field of view in the hippocampal CA1 regions 24 h after LPS administration. (n = 5 per group). G Representative immunofluorescence images of cortical regions at 24 h post LPS injection, as stained by anti-CD13 (cyan) and anti-PECAM-1 (magenta) antibodies. Scale bars, 50 μm. H Quantification of PECAM-1+ endothelial cells and CD13+ pericytes coverage in (G). (n = 5 per group) I Schematic illustration of neutrophil depletion experiments. Control and astrocytic Bmal1 KO mice received an intraperitoneal injection of anti-Ly6G antibody or isotype control IgG, followed by LPS treatment. J, K Representative brain images (J) and quantification (K) of Evans blue leakage in control and astrocytic Bmal1 KO mice following IgG + LPS or anti-Ly6G + LPS at 6 h post-LPS injection. (n = 6, CTRL + IgG; n = 6, AS-B1KO + IgG; n = 5, AS-B1KO + anti-Ly6G) Asterisks indicate significant differences. *P < 0.05, **P < 0.01, ***P < 0.001, n.s. not significant

We then examined the effects of CXCR2 blockade on BBB-associated cells. As shown in Fig. 3H, Bmal1 deficiency increased the relative area of GFAP+ astrocytes in the hippocampus of LPS-treated mice. Notably, inhibition of neutrophil infiltration by SB225002 did not reduce this increased astrogliosis in Bmal1 KO mice (Fig. 5F and Supplementary Fig. S8), suggesting that astrocyte activation is not directly dependent on neutrophil recruitment. In contrast, while SB225002 treatment did not alter endothelial cell intensity, it significantly restored pericyte coverage of blood vessels in cortical regions, which was reduced in astrocyte Bmal1 KO mice (Figs. 5G and H). To further validate the contribution of infiltrating neutrophils to BBB disruption in this model, we depleted neutrophils using an anti-Ly6G antibody prior to LPS stimulation (Fig. 5I). Neutrophil depletion significantly attenuated LPS-induced Evans blue leakage in astrocytic Bmal1 KO mice (Figs. 5J and K). Together, these results indicate that enhanced neutrophil infiltration via CXCL5-CXCR2 signaling plays a critical role in BBB disruption during peripheral inflammation in astrocytic Bmal1-deficient mice.

Astrocytic Bmal1 deficiency selectively impairs excitatory synaptic transmission after LPS administration

To explore the physiological consequences of increased BBB permeability in astrocytic Bmal1-deficient mice during peripheral inflammation, we examined synaptic alterations in the hippocampus. To assess excitatory synaptic transmission, we recorded miniature excitatory postsynaptic currents (mEPSCs) from CA1 pyramidal neurons five days post LPS injection (Fig. 6A). Astrocytic Bmal1 KO mice exhibited significantly reduced mEPSC frequency and amplitude compared with control mice (Figs. 6B-D), indicating impaired excitatory synaptic transmission under inflammatory challenge in the absence of astrocytic Bmal1.

Fig. 6.

Excitatory synaptic deficits in astrocytic Bmal1-deficient mice after LPS administration. A Schematic illustration of the experimental procedure. B Representative traces of miniature excitatory postsynaptic currents (mEPSCs) recorded from CA1 pyramidal neurons in CTRL and AS-B1 KO treated with PBS or LPS. C, D Summary graphs showing the amplitude (C) and frequency (D) of mEPSCs. Bar graphs show mean ± SEMs (**P < 0.01; ANOVA with a non-parametric Kruskal–Wallis test; ‘n’ denotes the total number of neurons analyzed as follows; CTRL + PBS, n = 23 cells from 6 mice; AS-B1 KO + PBS, n = 31 cells from 9 mice; CTRL + LPS, n = 37 cells from 8 mice; AS-B1 KO + LPS, n = 31 cells from 9 mice). E Representative images showing colocalization of VGLUT1 (green) and PSD-95 (red) puncta in the stratum oriens (SO) and stratum radiatum (SR) of the hippocampal CA1 region. Scale bar: 10 μm. F Quantification of the density (left) and size (right) of VGLUT1+ and PSD-95+ synaptic puncta. Data are means ±SEM (*P < 0.05; Mann-Whitney U test; ‘n’ denotes the number of images analyzed; n = 10–15 tissues from 4–5 mice per group)

To further explore potential structural correlates of these functional deficits, we performed immunostaining for the vesicular glutamate transporter 1 (VGLUT1), a presynaptic marker, and postsynaptic density protein 95 (PSD-95), a postsynaptic marker, and analyzed their colocalization within the stratum oriens (SO) and stratum radiatum (SR) of the CA1 region. Consistent with the electrophysiological deficits, astrocytic Bmal1 KO mice showed a significant reduction in the density of VGLUT1/PSD-95 colocalized puncta, accompanied by a modest but non-significant decrease in puncta size (Figs. 6E and F), indicating a loss of functional excitatory synapses in the hippocampus. Importantly, a similar significant reduction in VGLUT1/PSD-95 colocalized puncta density was also observed in the cortex of astrocytic Bmal1 KO mice under the same inflammatory conditions (Supplementary Fig. S9), demonstrating that excitatory synaptic deficits are not restricted to the hippocampus.

On the contrary, recordings of miniature inhibitory postsynaptic currents (mIPSCs) from CA1 pyramidal neurons revealed no significant differences in either frequency or amplitude between astrocytic Bmal1 KO and control mice (Supplementary Figs. S10A-D). Consistent with these electrophysiological findings, immunostaining for vesicular GABA transporter (VGAT) and gephyrin, markers of pre- and postsynaptic inhibitory specializations, respectively, showed comparable densities and sizes of VGAT/gephyrin colocalized puncta within the SO, stratum pyramidale (SP), and SR of the CA1 region (Supplementary Figs. S10E and F). These results indicate that astrocytic Bmal1 deficiency selectively affects excitatory, but not inhibitory, synapses under mild inflammatory conditions.

Discussion

Circadian rhythm has been implicated in the regulation of inflammatory responses in peripheral tissues [22], as reflected by diurnal fluctuations in the severity of inflammatory phenotypes [5, 6]. Disruption of circadian clock genes in immune cells has been shown to alter inflammatory responses [23], indicating that cell-intrinsic circadian clocks are crucial determinants of both the sensitivity and magnitude of inflammation. However, in contrast to peripheral inflammation, the mechanisms by which circadian rhythms regulate neuroinflammation remain poorly understood. In this context, our findings provide new insight into how BBB integrity during peripheral inflammation can be modulated by the astrocytic circadian clock.

BBB permeability has long been recognized as a major barrier to the delivery of therapeutic agents in neurological diseases [24]. Accordingly, the discovery of circadian rhythm–dependent fluctuations in BBB permeability has generated growing interest in chronotherapeutic strategies [25]. Indeed, recent studies have demonstrated that BBB integrity varies with the time of day [26]. Beyond its pharmacological implications, selective BBB permeability is also essential for maintaining immune privilege in the CNS by restricting the entry of circulating immune cells and inflammatory mediators. Thus, elucidating the mechanisms underlying circadian regulation of BBB integrity has broad implications for both BBB physiology and pathophysiology.

To date, cell-intrinsic circadian regulation of BBB function has been primarily described in endothelial cells and pericytes [27, 28]. Our analysis of publicly available transcriptomic datasets revealed that astrocytes exhibit particularly high expression of core clock genes, including Bmal1. However, whether astrocytic circadian clocks influence BBB permeability has not been clearly clarified, despite the well-known role of astrocytes in maintaining BBB structure and function. Previous studies of astrocytic Bmal1 have largely focused on intracellular processes such as endolysosomal function and autophagy [29, 30]. Indeed, deletion of Bmal1 in astrocytes has been shown to ameliorate tau and α-synuclein pathology by enhancing autophagic clearance [30]. In this context, it will be important to determine whether Bmal1-dependent alterations in astrocytic phagocytosis or vesicular trafficking contribute to BBB stability. In addition, whether astrocytic Bmal1 deficiency indirectly affects endothelial cell function warrants further investigation, as this may influence BBB integrity.

In our study, astrocyte-specific Bmal1 KO male mice exhibited increased BBB vulnerability at a mild LPS dose, whereas KO female mice showed comparable effects only at a higher dose. One possible explanation is that the lower LPS dose was insufficient to elicit BBB disruption in females, consistent with prior reports indicating that BBB integrity is generally more robust in females than in males [31]. Supporting this notion, estrogen has been reported to enhance BBB integrity in ovariectomy models [32, 33]. Additionally, sex-dependent differences in peripheral immune activation may partially contribute to the observed variation in BBB permeability [32, 34]. As most previous studies on BBB regulation have focused on male animals [27, 35, 36], it remains unclear whether circadian modulation of BBB permeability differs between genotypes. Notably, astrocyte-specific Bmal1 deletion has been reported to induce sex-dependent metabolic phenotypes [37], suggesting that BMAL1 function itself may be sexually dimorphic. In line with this, our data indicate that female mice are relatively more resistant to LPS-induced BBB disruption, yet still exhibit circadian-dependent increases in BBB permeability under higher inflammatory conditions. Nevertheless, hormonal fluctuations may contribute to variability in BBB responses, and future studies incorporating sex-specific molecular profiling will be required to clarify these mechanisms.

Mechanistically, our data indicate that increased CXCL5 production in astrocytic Bmal1-deficient mice promotes enhanced neutrophil recruitment into brain vessels and parenchyma, contributing to pericyte impairment and BBB disruption (Fig. 5B-E and H). In contrast, pharmacological inhibition of CXCR2 signaling, which effectively reduces neutrophil infiltration [38], significantly attenuated BBB permeability in astrocytic Bmal1 loss. Given that neutrophils can disrupt the BBB through the release of matrix metalloproteinases, as shown in our recent study [20], enhanced neutrophil recruitment is likely a key driver of BBB breakdown in this model. Notably, LPS-induced neutrophil recruitment itself exhibits circadian dependence (Fig. 1F and L), suggesting that both CXCL5 production and downstream neutrophil infiltration contribute to circadian modulation of BBB permeability in astrocytic Bmal1 deficiency.

Beyond BBB regulation, our electrophysiological data further indicate that astrocytic circadian disruption compromises neuronal communication, likely as a downstream consequence of BBB dysfunction. The observed reductions in mEPSC frequency and amplitude in astrocytic Bmal1-deficient mice suggest that increased BBB permeability and subsequent neuroinflammatory stress weaken excitatory synaptic transmission. While we cannot exclude the possibility that astrocytic Bmal1 loss directly affects neuron–astrocyte interactions, such as glutamate uptake, gliotransmitter release, or metabolic coupling, these mechanisms remain to be investigated. Given that BMAL1 regulates diverse astrocytic functions, including redox homeostasis, energy metabolism, and cytokine production [39, 40], future studies should delineate how astrocytic circadian disruption links BBB instability to synaptic vulnerability at the molecular and circuit levels. Collectively, our findings suggest that astrocytic Bmal1 coordinates neurovascular and synaptic stability under inflammatory conditions, thereby linking circadian regulation to neurovascular homeostasis.

Limitations

Previous studies have shown that astrocytic circadian disruption alters behavioral rhythms and cognitive function [41, 42]. Those studies, which used Glast-CreERT2 mice targeting broader neuronal and astrocytic populations and allowed longer post-tamoxifen intervals, may have captured cumulative synaptic and circuit-level remodeling. In contrast, our study used younger mice with shorter intervals between tamoxifen induction and analysis, which may have limited the detection of phenotypes requiring longer-term reorganization.

In addition, although astrocytes play a pivotal role in hypothalamic–pituitary hormone signaling [43], endocrine profiles were not assessed in this study. Astrocyte-specific Bmal1 deficiency has also been reported to alter systemic metabolism [42] and energy expenditure [44], which may indirectly influence BBB stability. Therefore, it will be important to investigate the systemic effects of astrocytic Bmal1 loss, as well as the potential systemic contributions of CXCR2 blockade. Furthermore, the contribution of circadian-independent functions of astrocytic Bmal1, including endocrine and metabolic alterations, to BBB vulnerability remains to be determined. Finally, we observed astrocyte-intrinsic activation in astrocytic Bmal1-deficient mice under LPS stimulation (Supplementary Figs. S5B-E), which was independent of neutrophil recruitment. These findings suggest that altered astrocyte activation may also contribute to BBB disruption through neutrophil-independent mechanisms. Future studies will be required to delineate these pathways.

Authors’ contributions

C.L., J.W.U., and J.-W.Y. conceived and designed the entire project. C.L., Y.L., W.C.J., I.H., H.B., D.-W.S., and H.J. performed the experiments. C.L., J.W.U., and J.-W.Y. wrote the manuscript. All authors read and approved the final version of the manuscript.

Funding

This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government (RS-2023-00207834, RS-2025-18362970 to J.-W.Y, RS-2023-NR076948 to J.W.U) and by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (RS-2024-00405260 to J.-W.Y).

Data availability

All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials. Single-nucleus or single-cell RNA-Seq datasets used in this study are available through the National Center for Biotechnology Information Gene Expression Omnibus. (GSE163577 and GSE263094).

Declarations

Ethics approval and consent to participate

All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Yonsei University College of Medicine (2022 − 0121) or DGIST (DGIST-IACUC-25082201-0001). All experiments were conducted in accordance with the approved guidelines of the Institutional Ethical Committee.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.