Ethics statement

The current study followed the guidelines established by the China Regulations for the Administration of Affairs Concerning Experimental Animals (1988) and Regulations for the Administration of Affairs Concerning Experimental Animals in Hubei Province (2005). All procedures and handling techniques were approved by The Scientific Ethic Committee of Huazhong Agricultural University (Animal Welfare Assurance No. HZAUMO-2023–0289 and No. HZAUMO-2025–0156). All efforts were made to ensure the ethical treatment of all experimental animals in this study and to minimize their suffering.

Cell culture and bacterial strains

Human brain microvascular endothelial cells (hBMEC) were purchased from ScienCell (Carlsbad, CA, USA). hBMEC were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum and penicillin-streptomycin (100 U/mL). The human astrocytoma cell line U251 and HEK293T (ATCC CRL-3216) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum. All the cells were cultured in 37°C incubators under 5% CO₂ until reaching monolayer confluence. The meningitic E. coli strain PCN033, a highly virulent isolate, was initially isolated from cerebrospinal fluid of porcine in China in 2006 [20].

Animals and infection

Specific-pathogen-free (SPF) C57BL/6 mice were purchased from the Laboratory Animal Services Centre of Huazhong Agricultural University (Hubei, China). Vascular CK2β-specific knockout C57BL/6J mice (CK2βVas^-/-) used in this study were obtained by crossing CK2βflox/flox mice (Cat No. S-CKO-01917, Cyagen Biosciences Inc, Jiangsu, China) with Cdh5-Cre mice (Cat No. 001023, Cyagen Biosciences Inc, Jiangsu, China) [21]. The CK2βflox/flox mice were obtained through genetic editing technology, with a loxP site inserted into the intron 2 and intron 3 of the Csnk2b gene on the 17th chromosome of the C57BL/6J mouse fertilized egg. When Cre enzyme is present, the Csnk2b gene can be knocked out by deleting the exon 3. The Cdh5-Cre mice were obtained by inserting the Cre element downstream of the promoter of the Cdh5 gene, which encodes the vascular endothelial cadherin, in the fertilized eggs of C57BL/6N mice. Cdh5-Cre-negative CK2β^flox/flox littermates were used as the control group. These mice can specifically express the Cre enzyme in the blood vessels and the conditional knockout was induced by estrogen analogues (tamoxifen). Tamoxifen (MCE, HY-13757A) was dissolved in a mixture of ethanol and corn oil (Solarbio, 1:9 by volume) at a concentration of 25 mg/mL concentration. It was administered through intraperitoneal injection at 2 weeks of age, with a dose of 40 mg/kg body weight every day for a duration of 7 days. Male and female mice 3 w of age were used for experiments.

The meningitic E. coli infection model was established as follows. Mice in the infection group were challenged with PCN033 via intravenous tail injection at a density of 1 × 10⁷ CFU. Mice in the control group were injected with isometric PBS. Then, the mice were euthanized with 30% CO₂ inhalation in a dark chamber at 2 hpi and 12 hpi (Figs 1A–1G; S1A–S1F, S1H–S1K)/ at 2 hpi only (Figs 3G, 3H; 6A, 6D–6G; and S5E–S5H), and their brains were collected for mRNA/protein sample isolation and IF assays [22].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Meningitic E. coli-stimulated BMEC induces astrocyte reaction.

(A) Representative immunofluorescence of GFAP (green) and S100β (red) in the hippocampal region of control, 2 hpi and 12 hpi group. Scale bar, 500 μm and 50 μm. (B) Number of S100β and GFAP double-positive cells was quantified in 1 mm2 of hippocampal region per section (n = 3 mice per group. For each mouse, four random zones per section were analyzed. The median value from these four fields in all three mice was used for statistical analysis). (C) Quantification of the Evans blue extravasation in the mouse brain (n = 3 mice per group). (D) Representative immunofluorescence of Iba1 (red) and the skeletonized images in the hippocampal region of control, 2 hpi and 12 hpi group. Scale bar = 50 μm. (E-G) Total process length, branches number and endpoints number of the top 10 Iba-1 positive microglia in branches number in 4 fixed regions within the hippocampal region of control, 2 hpi and 12 hpi group (n = 3 mice per group. The median value from these four regions in all three mice was used for statistical analysis). (H) Schematic diagram of the BMEC-astrocyte coculture model. (I) The mouse primary brain microvascular endothelial cells in upper chamber were challenged for 2 h with 10 MOI meningitic E. coli or left untreated; mRNA fold changes of astrocytic pan-reactive markers, LPS-specific reactive markers and MCAO-specific reactive markers in mouse primary astrocytes. n = 3. (J) mRNA fold changes of pro-inflammatory factors in mouse primary astrocytes at 2 hpi. n = 3. Data are shown as mean ± SEM. One-way ANOVA (B, C, E, F, G) followed by Tukey’s multiple comparison and Two-way ANOVA (I, J) followed by Bonferroni’s multiple comparison tests; *p < 0.033, ** p < 0.002, *** p < 0.001, ns, not significant.

https://doi.org/10.1371/journal.ppat.1013464.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Differential secretome profiling of BMEC in response to meningitic E. coli infection.

(A) Volcano plot comparing DEPs from the hBMEC culture supernatants of the control/infected groups, n = 3. (B) Heatmap visualization shows unsupervised clustering of differentially expressed proteins (DEPs) from the hBMEC culture supernatants of the control/infected groups. The expression profiles are displayed with three biologically independent samples in each group. Blue represents lower expression, and red represents higher expression. (C) KEGG analysis of the top 10 enriched pathways of the secreted DEPs. (D) Fold changes and p-values of 14 randomly selected DEPs. (E) Western blot verification of the expression of the randomly selected DEPs in the cell culture supernatant and lysate of fresh samples; ACTN4 (fold change of 1.01 in secretome result) here was used as a loading control for the supernatant sample, and fold changes of DEPs (presented as infection/control ratios under the bands of infected group) in culture supernatant were analyzed by Image J. n = 3. Data are shown as mean ± SEM.

https://doi.org/10.1371/journal.ppat.1013464.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. CK2 induces astrocyte reaction in vitro and in vivo.

100 ng/mL recombinant mCK2 was added to the cell culture supernatant of mouse primary astrocytes for 2 h, the mRNA fold change in the expression of (A) astrocytic pan-reactive markers, LPS-specific reactive markers in mouse primary astrocytes, and (B) pro-inflammatory factors in mouse primary astrocytes, n = 3. (C-D) Protein levels and fold changes of GFAP in mouse primary astrocytes treated with 0, 50, 100 and 200 ng/mL mCK2 for 2 h, n = 3. (E-F) Immunofluorescence and quantification of C3, CXCL10 and STEAP4 (all in red) in mouse primary astrocytes after 200 ng/mL mCK2 treatment. Scale bar, 50 μm. (G) Schematic illustration of mice pretreated with CX4945 (2.5 mg/kg) or vehicle (DMSO) for 2 h via intraventricular injection, followed by intravenous tail injection of meningitic E. coli (1 × 10⁷ CFU) for 2 h. (H) mRNA fold changes of reactive astrocyte markers and pro-inflammatory factors in the hippocampus of mice. n = 6. Data are shown as mean ± SEM. Two-way ANOVA (A, B, F) followed by Bonferroni’s multiple comparison test and One-way ANOVA (D, H) followed by Tukey’s multiple comparison; *p < 0.033, **p < 0.002, *** p < 0.001.

https://doi.org/10.1371/journal.ppat.1013464.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Extracellular CK2 induces astrocyte reaction via the NF-κB signaling pathway.

(A) mRNA fold change of Gfap in mouse primary astrocytes pretreated with 10 μM BAY 11-7082 (NF-κB inhibitor), Ruxolitinib (JAK/STAT inhibitor), SP600125 (MAPK-JNK inhibitor), Wortmannin (PI3K-Akt inhibitor), U0126 (MAPK-ERK inhibitor) or DMSO (vehicle) for 2 h before mCK2 (100 ng/mL) stimulation. n = 3. Protein levels of phosphorylated p65 and p65 and fold change analysis in mouse primary astrocytes treated with recombinant CK2 in (B-C) a time and (D-E) a concentration gradient. The CK2 reconstitution buffer was used as the vehicle. n = 3. (F) mRNA fold changes of reactive astrocyte markers and pro-inflammatory factors in mouse primary astrocytes pretreated with 10 μM BAY 11-7082 or isometric DMSO for 2 h before 100 ng/mL mCK2 stimulation for 2 h. n = 3. (G) mRNA fold changes of pro-inflammatory factors in U251 pretreated with 10 μM BAY 11-7082 or isometric DMSO for 2 h before 100 ng/mL CK2 stimulation for 2 h. n = 3. (H) Schematic illustration of mice treated with 1.5 mg/kg BAY 11-7082 (or vehicle, DMSO) and 0.3 μg/kg mCK2 simultaneously for 3 h. (I) mRNA fold change of reactive astrocyte markers and pro-inflammatory factors in the hippocampus of the mice. n = 6. Data are shown as mean ± SEM. One-way ANOVA (A, I) followed by Tukey’s multiple comparison and Two-way ANOVA (C, E, F, G) followed by Bonferroni’s multiple comparison test. *p < 0.033, **p < 0.002, ***p < 0.001, ns, not significant.

https://doi.org/10.1371/journal.ppat.1013464.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Extracellular CK2 interacts with myosin 9 in astrocytes.

(A) Analysis of endogenous astrocytic proteins associated with extracellular CK2 by immunoprecipitation and LC-MS/MS. A total of 1 μg/mL recombinant CK2 consisting of His-tagged CK2α & GST-tagged CK2β was added into the cell culture supernatant of U251 cultured in 10 cm dishes for 3 h. The cells were subsequently lysed and centrifuged, the clarified cell supernatant was immunoprecipitated with a mouse monoclonal antibody for His and GST tags (2 μg per 500 μg of cell lysates), and mouse IgG was used as a negative control. Specific bands in lane 3 in the vehicle group and CK2 group (a and b) were then identified by LC-MS/MS. Relevant information on the top-ranked protein myosin 9 is shown. (B) U251 cells were incubated with 1 μg/mL recombinant CK2 or His-tagged/GST-tagged CDT (as negative controls) for 2 h. Then, the cell lysates were immunoprecipitated with an anti-myosin 9 antibody and immunoblotted with an anti-His antibody or anti-GST antibody, respectively. (C) Immunostaining of His-tagged CK2α (red), GST-tagged CK2β (red) and endogenous myosin 9 (green) in U251 after stimulation with 1 μg/mL CK2 for 2 h. Scale bar, 10 μm. (D) HEK293T cells were transfected with expression vectors encoding GFP or CK2β truncation mutants. The cell lysates were immunoprecipitated with an anti-GFP antibody and immunoblotted with an anti-GFP antibody or anti-myosin 9 antibody. (E) Protein levels of phosphorylated myosin 9, myosin 9, phosphorylated p65 and p65 in mouse primary astrocytes at 30 min after 100 ng/mL CK2α, CK2β and CK2 treatment. n = 3. (F) Protein levels of myosin 9, phosphorylated p65 and p65 in WT and myosin 9-KO U251 cells after 100 ng/mL CK2 treatment for 1 h. CK2 reconstitution buffer was used as the vehicle. n = 3. (H) mRNA fold change of pro-inflammatory genes in WT/ myosin 9-KO U251 cells stimulated with 100 ng/mL CK2 for 2 h. n = 3. (G) Protein levels of phosphorylated myosin 9 and myosin 9 in mouse primary astrocytes pretreated with BAY 11-7082 (NF-κB inhibitor) or vehicle (DMSO) following 100 ng/mL CK2 treatment for 10 min. n = 3. (I) Schematic diagram of the hBMEC (WT) cocultured with U251 (WT) or U251 (myosin 9-KO). (J) mRNA fold change of pro-inflammatory genes in WT/myosin 9-KO U251 within the coculture model under infection conditions. n = 3. Data are shown as mean ± SEM. Two-way ANOVA (H, J) followed by Bonferroni’s multiple comparison test. ***p < 0.001, ns, not significant.

https://doi.org/10.1371/journal.ppat.1013464.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Loss of CK2β in the BMEC attenuates astrocyte reaction induced by the meningitic E. coli infection.

(A) Protein levels of CK2α and CK2β in supernatants and lysates of WT and CK2β-knockdown hBMEC after meningitic E. coli infection (10 MOI for 2 h). n = 3. (B) Schematic diagram of the hBMEC (WT) or hBMEC (CK2β-KO) cocultured with U251 cells. (C) mRNA fold change of pro-inflammatory genes in U251 cells cocultured with WT/CK2β-KO hBMEC under infection condition. n = 3. Immunofluorescence of (D) STEAP4, (E) CXCL10, and (F) C3 (all in red) with GFAP (green) in the hippocampus of WT and CK2βVas^-/- mice post 2 h infection. Scale bar, 20 μm. n = 3. (G) mRNA fold change of reactive astrocyte markers and pro-inflammatory factors in hippocampus of WT and CK2βVas^-/- mice challenged with meningitic E. coli (1 × 10⁷ CFU). n = 6. (H) Survival rates of WT and CK2βVas^-/- mice were monitored for 25 h after meningitic E. coli infection. n = 7. Data are shown as mean ± SEM. One-way ANOVA (G) followed by Tukey’s multiple comparison and Two-way ANOVA (C) followed by Bonferroni’s multiple comparison test. Log-rank (Mantel-Cox) test (H). ** p < 0.002, *** p < 0.001, ns, not significant.

https://doi.org/10.1371/journal.ppat.1013464.g006

For stereotaxic injection, mice were anesthetized with isoflurane (1.5%) in pure oxygen using continuous flow via a nose cone to maintain a surgical plane of anesthesia. Sedation was assessed via absence of both the pedal (toe pinch) and tail pinch reflexes, and anesthesia was maintained at a constant rate. After confirming proper sedation, the fur from the top of the scalp to the base of the neck was shaved and mice were fixed on a stereotaxic apparatus with ear bars. After the skull was exposed via a small incision, a small hole was drilled for injection based on coordinates with the bregma. For in vivo compound or protein delivery, microinjection was performed at the following stereotaxic coordinates of the bilateral ventricle: anteroposterior (AP), -0.8 mm; mediolateral (ML), ± 1.7 mm; and dorsoventral (DV), -2.5 mm from bregma. The injection was performed via a 10 μL gas phase (tip) microinjector (Shanghai Gaoge Industry and Trade Co., Ltd) over 10 min; the needle was not removed until 10 min after the infusion. All the compounds and proteins were injected at a total volume of 10 μL [23]. After completing injections into both hemispheres, the incision was closed using a stainless steel wound clip and the mice were placed back on the warming pad until they woke up and moved around normally.

Reagents, antibodies and inhibitors

The NF-κB inhibitor BAY11–7082, the JAK/STAT inhibitor Ruxolitinib, the JNK inhibitor SP600125, the PI3K-Akt inhibitor Wortmannin, the ERK1/2 inhibitor U0126, and the CK2 inhibitor Silmitasertib (CX-4945) were purchased from MedChem Express (Monmouth, USA). Recombinant human casein kinase II (CK2) was purchased from Sigma Aldrich (218701, St. Louis, USA). The nucleic acid dye, 4’-6-diamidino-2-phenylindole (DAPI), was obtained from Solarbio (Beijing, China). All the antibody information was placed in S5 Table.

Evaluation of BBB permeability by Evans blue

The mice were anesthetized according to the method described in “Animal and infection”. Evans blue (2% wt/vol in PBS, 2 mL/kg) was administered through the tail vein 0.5 h before mice were euthanized. Whereafter, mice were perfused through the left ventricle with 10 mL PBS to remove intravascular albumin-EB. Then half of the brain was weighed and homogenized in 1 mL of PBS, and 0.5 mL homogenization was mixed with 1.5 mL of 50% trichloroacetic acid, followed by vortex mixing for 2 min. After centrifugation at 10,000 × g for 30 min at 4°C, the supernatant was collected. 30 µL of supernatant was mixed with 90 µL 95% ethanol in a clear 96-well plate. The fluorescence intensity was measured at 620 nm excitation/680 nm emission on a microplate fluorescence reader (SPARK 10M, TECAN) and compared against a standard curve (serial dilutions of the stock dye solution in the concentration range of 0, 0.25, 0.5, 1, 5, 20 μg/mL), as described previously [24]. The Evans blue content was calculated and expressed as per gram of brain tissue.

Preparation of mouse primary astrocyte and endothelial cell cultures

Primary cultures of astrocytes were prepared from neonatal mice, as described previously [25]. Newborn mice (1-2d postnatal) were killed by decapitation, and the brains were removed. After the meninges were removed from the entire brain, the neocortical regions were excised and collected. The tissues were dissociated with 0.25% trypsin at 37°C, and the reaction was terminated with DMEM supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. The cell culture medium was refreshed 24 h later and changed every 3 days. Experiments were performed on confluent 14-day-old cultures.

Endothelial cells were isolated from adult mice of 6–8-week-old as described previously [26]. The brains were cut along the longitudinal fissure to separate the two hemispheres. The meninges and choroid plexuses were removed by rolling the brain tissues on P8 grade filter paper. Minced brain pieces were forced through sterile needles in an 18G, 20G and 22G needles order. After centrifugation at 800 × g for 8 min at 4°C, the precipitate of the homogenate was digested and pelleted. The myelin was separated by resuspending the pellet in DMEM containing 20% BSA and centrifuging at 4˚C, 1,000 × g for 20 min. In order to separate microvessel from brain cellular contaminants, the pellet was resuspended in 1 ml of ice-cold DMEM and overlay on top of a cold 33% continuous isotonic Percoll gradient and then centrifuged at 4˚C, 1,000 × g for 20 min. The microvessel layer was collected, washed and seeded at a dilution corresponding to one brain per well on 6-well plates previously coated for 4 h with type IV collagen. Microvessel were cultured in high glucose DMEM supplemented with 20% FBS, 1% penicillin/streptomycin, 1 ng/mL basic fibroblast growth factor (bFGF) and 100 μg/mL heparin. The culture medium was replaced every other day using fresh ECs culture medium containing only 4 μg/mL puromycin.

BMEC-astrocytes transwell coculture and infection assays

The in vitro BMEC-astrocyte coculture model was constructed via a 0.4 μm pore size transwell chamber following previously reported methods [27]. In brief, mouse primary endothelial cells or hBMEC were pre-seeded in the upper chamber of transwell inserts (Corning Cellgro, Manassas, VA, USA) at 8 × 10³ cells per well, and mouse primary astrocytes or U251 cells were synchronously seeded in a 24-well plate. After 12 h of growth, the transwell inserts were transferred to the 24-well plates, and the coculture medium was changed to DMEM supplemented with 1 ng/mL bFGF (for primary cells) or RPMI 1640 medium (for cell lines). Meningitic E. coli infection of coculture model was performed as follows. The confluent cells were starved in serum-free medium for 12 h. The meningitic E. coli cultures were resuspended, diluted in the same serum-free medium and added into the transwell inserts at the indicated time points (10 MOI in 100μL final volume). Afterwards, the cells in the lower chamber were washed twice with pre-chilled PBS and collected for the subsequent RNA isolation assay.

Immunohistochemistry and immunofluorescence

For immunohistochemistry, anesthetized mice (according to the method described in “Animal and infection”) were perfused with PBS transcardially, followed by 4% paraformaldehyde. Then, the mouse brain tissues were fixed with 4% paraformaldehyde at 4°C for 24 h. A series of gradient alcohols, anisole and xylene were used for the tissue dehydration before paraffin embedding: 75% ethanol for 5 min, 85% ethanol for 5 min, 90% ethanol for 5 min, 95% ethanol for 5 min, absolute ethanol twice for 5 min each, anisole for 5 min, xylene twice for 5 min each. Fixed brains were sliced into 5 μm thickness coronal sections using a Leica RM2125 RTS microtome. The sections were placed in a 37°C warm water bath until it is completely flat and then were mounted by being placed in 37 °C heater for 2 h. Sections were deparaffinized by xylene twice for 5 min each and were hydrated with 100% ethanol twice for 5 min each, 95% ethanol twice for 5 min each, and 70% ethanol twice for 5 min each and then rinsed in distilled water. Then, the paraffin sections were heated in citrate buffer (pH 6.0) at 60°C for 10 min. After natural cooling, sections were placed in PBS and washed 3 times. Then, sections were processed with 0.1% Triton X for 5 mins, followed by 5% bovine serum albumin blocking for 1 h at room temperature. After staining with primary antibody (diluted in the blocking buffer) overnight at 4°C. After washing off with PBS, the samples were then incubated with secondary antibodies for 1 h at room temperature. After that, the samples were covered with DAPI (diluted in the blocking buffer) in the dark. For immunofluorescence for cultured cells, cells were washed with PBS at 4°C and fixed with 4% paraformaldehyde for 30 min at room temperature, followed by incubation in blocking and permeabilization buffer containing PBS with 0.1% Triton X-100 and 5% normal bovine serum.

Primary and secondary antibodies are listed in S5 Table. Finally, the cells or sections were stained with the indicated antibodies and DAPI. Images were taken with an LSM 880 confocal microscope (Carl Zeiss AG, Oberkochen, Germany). For quantification, the intensities of the immunopositive cells were measured via Fiji ImageJ software. For S1I Fig, Pearson’s correlation coefficient was computed using Fiji ImageJ software with the Trainable Weka Segmentation and ROI manager to assess the degree of colocalization between albumin and CD31, providing a semi-quantitative indicator of dynamic changes of BBB permeability over time. For Fig 1E–1G, morphological analysis (skeletonization, total process length measurement, and branch number/endpoints counting) was performed by plugins AnalyzeSkeleton (2D/3D) of Fiji ImageJ software according to the previous studies [28].

Sample preparation, data acquisition, and analysis of the BMEC secretome

The mass spectrometry data of the BMEC secretome have been deposited in the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD021427. Primary hBMEC were divided into control and infected groups and were starved in a serum-free medium (1:1 mixture of Ham’s F-12 and M-199) for 16 h. The medium of the control group was replaced with the fresh serum-free medium when the other group was infected with PCN033 for 3 h (MOI = 10) at 37°C. Cell culture supernatants from the infected and control groups were obtained and centrifuged at 4,000 × g for 15 min at 4°C. The insoluble debris was removed, and the supernatant was centrifuged at 15,000 × g for 2 h at 4°C. The protein concentration in the supernatant was measured via a BCA protein assay kit (Beyotime, China) and then reduced and alkylated with 10 mM Tris-(2-carboxyethyl) phosphine and 40 mM chloroacetamide at 95°C for 5 min. Alkylated proteins were diluted fivefold with 50 mM triethylammonium bicarbonate and digested with Lys-C at a 1:100 (wt/wt) enzyme-to-protein ratio for 3 h at 37°C. Trypsin was added at a final 1:50 (wt/wt) enzyme-to-protein ratio overnight digestion at 37°C. The digested peptides were acidified with trifluoroacetic acid to a final concentration of 0.5% trifluoroacetic acid, and 250 μL of ethyl acetate was added to 250 μL of the digested solution. The mixture was shaken for 2 min and then centrifuged at 20,000 × g for 2 min to obtain aqueous and organic phases. The aqueous phase was collected and desalted via a 100 mg Sep-Pak C18 column. The in-house-constructed immobilized metal ion affinity chromatography (IMAC) tip was made by capping the end with a 20 μm polypropylene frit disk. Peptides were dissolved in 4 μL of 0.3% formic acid with 3% (vol/vol) acetonitrile and injected into an EasynLC 1000. Peptides were separated on a 45-cm in-house packed column (360 μm OD × 75 μm inner diameter) containing C18 resin (2.2 μm, 100 Å) with a 30-cm column heater, and the temperature was set at 50°C. The mobile phase buffer consisted of 0.1% formic acid in ultrapure water (buffer A) with an elution buffer of 0.1% formic acid in 80% (vol/vol) acetonitrile (buffer B), which was run with a linear 45- or 60-min gradient of 6–30% buffer B at a flow rate of 250 nL/min. The EasynLC 1000 was coupled online with a hybrid high-resolution LTQ-Orbitrap Velos Pro mass spectrometer. The mass spectrometer was operated in data-dependent mode, in which a full-scan MS (from m/z 350–1,500 with a resolution of 30,000 at m/z 400) was used, followed by MS/MS on the 10 most intense ions (normalized collision energy, 30%; automatic gain control 3E4; maximum injection time, 100 ms). For the raw MS/MS data, the intensities of the peptides were extracted via the MaxQuant search engine with Label-Free Quantitation (LFQ), and the analysis was processed via Andromeda. The tandem mass spectra results were subsequently matched with the Human UniProt database. The relative protein abundances were presented as the infection/control ratios. The statistical significance of the difference between the control and infection group was analyzed by t-test followed by Bonferroni’s multiple comparisons. Proteins whose expression changed more than 2 times or less than 0.5 times and whose p-value was less than 0.05 were regarded as differentially expressed proteins. After removing the proteins that were marked by UniProt for their nuclear localization, SignalP 6.0 (https://services.healthtech.dtu.dk/services/SignalP-6.0/) was used to predict whether the remaining proteins carried signal peptides, and SecretomeP 2.0 (https://services.healthtech.dtu.dk/services/SecretomeP-2.0/) was used to predict non-classical secretory proteins (NN-SCORE > 0.6).

RNA extraction and qRT-PCR

The total RNA from cultured cells or tissue isolated with TRIzol reagent was reverse transcribed via HiScript II Q RT SuperMix reverse transcriptase (Takara, Japan). Quantitative real-time PCR (qRT-PCR) was performed on a QuantStudio 3 real-time PCR system (Thermo) with 2 × M5 HiPer SYBR Premix EsTaq (Mei5bio, Beijing, China) according to the manufacturer’s instructions. The primers used for qRT-PCR are shown in S1 Table. Each sample’s abundance of individual mRNA transcripts was assayed in triplicate and normalized to the GAPDH (for the human sample) or β-actin (for the mouse sample) mRNA level via the 2^-ΔΔCTmethod.

Western blot analysis

The cells and homogenized tissues were lysed in ice-cold cell lysis buffer or RIPA buffer (Beyotime, Shanghai, China), and then the samples enriched with 1 × loading buffer were heated to 100°C for 10 min. The protein samples were separated via SDS-PAGE and transferred to polyvinylidene difluoride membranes. The membranes were blocked with 5% bovine serum albumin in TBST for 2.5 h at room temperature and incubated with specific primary antibodies diluted in the blocking buffer overnight at 4°C. The membranes were then probed with the corresponding peroxidase-conjugated secondary antibody for 1.5 h at room temperature. Then the membranes were washed with TBST three times and developed with Clarity Western ECL Substrate (Bio-Rad), the results were visualized via the ChemiDoc XRS1 imaging system (Bio-Rad Laboratories). Primary and secondary antibodies used in western blot are listed in S5 Table. The quantitative analysis of bands was assessed using ImageJ, and β-actin (as well as ACTN4 in the cell culture supernatant protein sample in Figs 2E and 6A) was used as the internal reference protein. Uncropped versions of the blots are shown in S1 Data.

Recombinant CK2 expression and purification

Human CK2α/CK2β, as well as mouse CK2α/CK2β (mCK2α/mCK2β), were expressed in a prokaryotic expression system as described below. Human CSNK2A1 and CSNK2Bwere amplified from hBMEC cDNA, and mouse Csnk2a1 and Csnk2b were amplified from the cDNA of a mouse brain microvascular endothelial cell line (bEND.3, ATCC CRL-2299), the primer used in the amplification are listed in Supplementary S1 Table. His-tagged CK2α and mCK2α, as well as GST-tagged CK2β and mCK2β, were then separately overexpressed in E. coli BL21 (DE3) after being subcloned and inserted into pET-28a and pGEX-6p-1. The soluble fraction of the bacterial lysates was purified with Ni-NTA His-Tag Purification Agarose (MCE) or Glutathione Sepharose (GE) according to the manufacturer’s instructions. Edulcoration and concentration of the purified proteins were performed with 10 kDa Ultra Centrifugal Filters (Millipore, UFC901024). The endotoxin of the purified proteins was removed by Pierce High-Capacity Endotoxin Removal Resin (Thermo, No.88276). Finally, the purified proteins were mixed with stock buffer (20 mM MOPS, pH 7.4, 1 mM EDTA, 0.1 mM Triton X-100, 0.3% 2-mercaptoethanol, and 50% glycerol) for storage at −80°C. The α and β subunits were reconstituted into the holoenzyme in vitro as described previously [29]. Equimolar quantities of each subunit (0.1 nmoL) were incubated for 1 h at 30°C in 200 μL of reconstitution buffer (50 mM Tris-HCl pH 7.0, 150 mM NaCl, 0.1% BSA).

The enzyme activity of α subunits and holoenzymes is reflected by ATP consumption. Briefly, astrocytes cultured in 10 cm dishes were lysed with reconstitution buffer, and the lysate was collected and centrifuged at 10,000 × g for 3 min. For each of the three parallel wells in a 96-well opaque plate, 50 μL of supernatant was incubated with 50 ng of CK2α, mCK2α, CK2 or mCK2 for 10 min at 37 °C. BSA was used as a negative control. The ATP content was subsequently detected via an Enhanced ATP Assay Kit according to the manufacturer’s instructions (Beyotime, Shanghai, China).

The cytotoxicity of CK2 subunits and holoenzymes was assessed by measuring cell viability with a Cell Counting Kit-8 according to the manufacturer’s instructions (Beyotime, Shanghai, China). Astrocytes were seeded and grown in 96-well plates until they reached monolayer confluence. mCK2α, CK2α, mCK2β, CK2β, mCK2 and CK2 (400 ng/mL) were added to the plates for 1 h and five parallel wells were used for each treatment.

Immunoprecipitation and LC-MS/MS analysis

After CK2 incubation, cellular samples were obtained via lysis buffer for western blot and IP and were subsequently centrifuged at 4°C for 15 min at 15,000 × g to remove insoluble cell debris. The clarified cell lysates (500 μg) were incubated with 2 μg indicated primary antibodies overnight at 4°C, and 25 μL protein A/G agarose (Beyotime, China) was then added for 2 h of incubation at room temperature. After five washes with PBST, the immunoprecipitates were resuspended in 1 × loading buffer and boiled for 5 min. Then, the protein samples were separated by SDS-PAGE. The differential protein bands from the GST-tag antibody-immunoprecipitated samples were identified via LC-MS/MS, performed using a nanoLC-LTQ-Orbitrap XL mass spectrometer (Thermo, San Jose, USA) coupled with an Eksigent nano-LC 1D plus HPLC system at APTBIO (Shanghai, China). Tryptic peptides were fully enzymatically digested and ionized via nano electrospray ionization. The data were analyzed via a full-scan mass spectrometer (300–1800 m/z). Finally, Proteome Discoverer (version 1.4.0.288, Thermo Scientific) was used to analyze the MS data.

Gene transfection, RNAi, and generation of knockout cell lines

The transfection of siRNA or DNA was performed via jetPRIME transfection reagent (Polyplus, France) according to the manufacturer’s protocol. Human CSNK2B siRNA and negative control were synthesized by JTS Scientific, and the knockdown efficiency was determined at 48 h post-transfection via western blot analysis. The sgRNA sequences targeting human CSNK2B and the sgRNA targeting human MYH9 were predicted via the online CRISPR/Cas9 design tool (https://www.genscript.com.cn/gRNA-design-tool.html) and inserted into individually cloned into a pYSY-SpCas9-sgRNA-Puro plasmid. The recombinant sgRNA expression plasmids were cotransfected into cells for 24 h. Then, the positive-transfected cells were selected by fresh medium containing 1 μg/mL puromycin for another 24 h. Surviving cells were transferred into 96-well plates with limiting dilution and incubated until a single-cell clone formed. Positive single-cell clones were validated via DNA sequencing and immunoblotting analyses.

Statistical analysis

The results are expressed as mean ± SEM, for datasets that do not follow a normal distribution, median was used for statistical analysis, and all the statistical analyses are described in the figure legends. Brown-Forsythe test was used to conduct a homogeneity of variance test on the data. In all vivo experiments, we prepared two sections of each mouse and took two photos of each section. Two randomly selected photos from different brain slices of each mouse were used for immunohistochemistry analysis. Especially, 4 regions of interest (S1G Fig) per mouse were analyzed in Figs 1B, 1E–1G and S1I Data in Figs 1B, 1E–1G and S1I were presented in the form of significance of the differences between each group was analyzed by one-way ANOVA followed by Tukey’s multiple comparison. In all the figures of subculture cells, “n = 3” represents three separate wells of distinct cell cultures. In the case of using primary cells, “n = 3” refers to three independent cultures obtained from the same cell isolation. For normally distributed datasets, data was analyzed by one-way ANOVA followed by Tukey’s multiple comparison, two-way ANOVA followed by Bonferroni’s multiple comparison test or Log-rank (Mantel-Cox) test (Fig 6H). The statistical analyses of all data were performed with GraphPad Prism, version 8.0 (GraphPad Software Inc., La Jolla, CA, USA). p < 0.033 (*) was considered statistically significant, and p < 0.002 (**), as well as p < 0.001 (***), were considered extremely significant.

Crosstalk with BMEC facilitates the inflammatory response of astrocytes during meningitic E. coli infection

As two major cell types involved in maintaining homeostasis within the CNS, astrocytes and microglia also play an indispensable role in neuroinflammation [30,31]. The immunoreactivity of glial fibrillary acidic protein (GFAP) and S100β, markers of astrogliosis, were predominantly elevated in the hippocampal and internal capsule regions as the duration of infection increased (S1A Fig). While GFAP immunoreactivity levels in both regions were significantly upregulated at 12 h post infection (hpi), unlike in the internal capsule, GFAP levels in the hippocampal region were already significantly enhanced as early as 2 hpi (S1B, S1C Fig). The number of S100-positive cells also increased during the infection (S1D–S1F Fig). More importantly, the number of S100β and GFAP double-positive cells significantly increased at 2 hpi and further rose at 12 hpi in the hippocampal region (Fig 1A and 1B). To ascertain whether the observed astrogliosis could be attributed to the stimulation of factors such as pathogens and virulence proteins leading to peripheral blood bacterial-induced disruption of the BBB, we investigated the dynamic changes in BBB integrity following meningitic E. coli infection. Mice were intravenously injected with Evans blue 0.5 h prior to each analysis time point. Compared with the control group, no significant difference in BBB permeability was observed at 2 hpi. However, severe BBB disruption was detected at 12 hpi (Fig 1C). To quantitatively evaluate the extent of albumin extravasation, we utilized Pearson’s correlation coefficient to measure the colocalization of albumin with the brain microvessel marker CD31. A higher Pearson’s coefficient in the control and 2 hpi groups indicates strong colocalization, suggesting that albumin predominantly remains within the microvessel. Conversely, a lower Pearson’s coefficient in the 12 hpi group reflects weaker colocalization, indicating the extravasation of albumin from microvessel. (S1H, S1I Fig). Microglia are widely recognized as key instigators of astrogliosis. To investigate their potential involvement in astrogliosis during the early stage of infection, we performed a detailed morphological analysis of microglia in four fixed regions within the hippocampal region [32]. Compared to the control group, microglia at 12 hpi exhibited a significant reduction in total process length, branch number, and endpoints number, indicative of heightened microglial activation. In contrast, no substantial changes were observed in microglia at 2 hpi (Fig 1D–1G).

Given the critical role of brain microvascular endothelial cells (BMEC) in the blood-brain barrier (BBB) during the progression of E. coli meningitis, as well as the close interaction between BMEC and astrocytes, we investigated whether BMEC contributes to the astrocyte response at the early stage of this process. To explore this hypothesis, an in vitro coculture system was established to mimic the cell crosstalk between BMEC and astrocytes (Fig 1H). In this model, primary mouse brain microvascular endothelial cells were cultured in transwell inserts, while primary mouse astrocytes were maintained in culture plates. Both cell types shared the same culture medium but remained physically separated. The endothelial cells within the transwell inserts were infected with meningitic E. coli PCN033 (MOI = 10) for 2 h. Subsequently, the transcriptional profiles of genes previously identified as pan-reactive markers of astrocytic activation (Lcn2, Aspg, Osmr, S1pr3, Gfap, Cxcl10 and Steap4), LPS (Lipopolysaccharide)-specific reactive markers (H2-T23, H2-D1, Srgn, Psmb8, Amigo2, C3 and Iigp1), and MCAO (Middle cerebral artery occluding)-specific reactive markers (Slc10a6, CD109, Tm4sf1, S100a10, Emp1, Sphk1, and Clcf1) were analyzed [33]. In comparison to the control group, the results revealed a significant increase in mRNA levels for all 7 pan-reactive markers and 7 LPS-specific reactive markers, while no significant changes were observed in the mRNA levels of MCAO-specific reactive markers. This suggests that astrocytes in the lower chamber may undergo an LPS-specific reaction (Fig 1I). Additionally, we detected transcriptional upregulation of seven pro-inflammatory-related genes (Tnf-α, Il-1β, Il-6, Mip2, Ccl2, Cxcl3) in mouse primary astrocytes, which was further validated in hBMEC-U251 coculture model (Figs 1J and S1L). Based on these findings, we propose that the crosstalk between BMEC and astrocytes plays a critical role in astrocyte reactions and neuroinflammation during meningitic E. coli infection.

Differential secretion proteomics of hBMEC in response to infection

To further investigate potential BMEC regulatory molecules involved in astrocyte reaction, we collected culture supernatants from meningitic E. coli-infected as well as uninfected hBMEC for label-free quantitative proteomics analysis. A total of 140 proteins were identified as significant differential expression (fold change >2, p-value <0.05) among the 1680 common proteins identified from the cell culture supernatant. Among these proteins, 76 were upregulated and 64 were downregulated in the infection samples compared to the control samples (Fig 2A). The relative abundance of differentially expressed proteins (DEPs) in three biological replicates for each group is presented in a heatmap (Fig 2B). Furthermore, the DEPs were screened using the secreted protein analysis software SignalP and SecretomeP, resulting in the identification of a total of 61 possible secreted proteins (non-classical secretory proteins and classical secretory proteins carrying signal peptides, S2 Table). A Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis demonstrated that these secreted DEPs were predominantly associated with meningitic E. coli infection of the BBB (e.g., ECM-receptor interaction, focal adhesion, and adherens junction) and neuroinflammation (e.g., the PI3K-Akt signaling pathway, the Ras signaling pathway, pathogenic E. coli infection, and the NF-κB signaling pathway) (Fig 2C, S3 Table). Subsequently, 14 proteins were randomly selected from the 61 secreted DEPs for western blot validation. The selected DEPs included 12 significantly upregulated proteins and 2 significantly downregulated proteins. The fold change of these DEPs in the quantitative proteomics analysis, along with the associated p-values, are presented in Fig 2D. In comparison to the internal reference protein ACTN4, the expression of ANXA6 and the other 11 proteins significantly increased in the culture supernatant of hBMEC, and the expression of AGRN and TGFBI decreased, which aligns with the findings of the quantitative proteomics analysis (Figs 2E and S2). Our research demonstrated that meningitic E. coli infection of hBMEC resulted in significant alterations in the secretion proteomic profiles of these cells.

CK2 contributes to the astrocyte inflammatory response

Since NF-κB represents a pivotal inflammatory signaling pathway and both CK2α and CK2β are enriched within this pathway, we selected to pursue subsequent validation of the biological function of CK2. First, we successfully expressed recombinant human and mouse CK2 proteins. Furthermore, recombinant human and mouse CK2 were verified to be enzymatically active and non-cytotoxic (S3A and S3B Fig). mouse primary astrocytes were subsequently treated with mouse recombinant CK2 (mCK2) at a concentration of 100 ng/mL for 2 h. Real-time PCR results demonstrated that mCK2 stimulation significantly upregulated the expression of astrocytic pan-reactive markers (Aspg, Osmr, S1pr3, Gfap, Cxcl10 and Steap4), LPS-specific reactive markers (H2-T23, H2-D1, Srgn, Psmb8, Amigo2, C3 and Iigp1) and pro-inflammatory factors (TNF-α, IL-1β, IL-6, MIP2, CCL2, CXCL3, and CXCL10) (Fig 3A and 3B). The western blot results demonstrated that CK2 protein at concentrations of 50, 100, and 200 ng/mL for 2 h stimulation of mouse primary astrocytes dose-dependently induced a significant upregulation of GFAP, which is an important marker molecule for astrogliosis (Fig 3C and 3D). Furthermore, we investigated the expression of C3, CXCL10, and STEAP4 via immunofluorescence following the administration of recombinant mouse CK2 at a concentration of 200 ng/mL to mouse primary astrocytes for 2 h. The results revealed that recombinant mouse CK2 stimulation significantly induced the expression of C3, CXCL10, and STEAP4 in mouse primary astrocytes (Fig 3E and 3F). The role of CK2 in the meningitic E. coli infection-induced astrocyte reaction was further analyzed in a mouse model. The mice were administered CX4945 (2.5 mg/kg), an inhibitor of CK2 activity, via intravenous injection 2 h before infection (Fig 3G). At 2 hpi, the brain hippocampal tissues of the mice were harvested to analyze the expression of astrocyte pan-activation markers, LPS-specific reactive markers, and pro-inflammatory cytokines via real-time PCR. As shown in Fig 3H, the meningitic E. coli challenge resulted in significant upregulation of astrocyte pan-activation markers (Gfap, Steap4 and Cxcl10), LPS-specific reactive markers (C3, Iigp1 and Amigo2), and pro-inflammatory cytokines (TNF-α, IL-1β and IL-6). The above in vivo and in vitro data collectively indicate that CK2 plays a significant role in the induction of the astrocyte reaction.

CK2 induces astrocyte reaction by activating the NF-κB signaling pathway

We subsequently sought to elucidate the molecular mechanisms by which CK2 induces the astrocyte reaction, thereby promoting neuroinflammation. We selected several classical signaling pathways with established roles in regulating astrocyte reaction and Gfap transcription, including the JAK/STAT, PI3K-Akt, ERK1/2, JNK, and NF-κB pathways, for testing. Mouse primary astrocytes were pretreated with 10 μM of the JAK/STAT-specific inhibitor Ruxolitinib, 10 μM of the PI3K-Akt-specific inhibitor Wortmannin, 10 μM of the ERK1/2-specific inhibitor U0126, 10 μM of the JNK-specific inhibitor SP600125 and 10 μM of the NF-κB-specific inhibitor BAY 11–7082, and the results demonstrated that the transcriptional upregulation of Gfap induced by CK2 was significantly suppressed by BAY 11–7082, Ruxolitinib, Wortmannin and SP600125. Among them, the level of Gfap in the BAY 11–7082 treatment group decreased the most significantly. (Fig 4A). The activation of the NF-κB pathway was subsequently investigated following the treatment of mouse primary astrocytes with mouse recombinant CK2. The results of the western blot analysis demonstrated that CK2 stimulation dose-dependently and time-dependently promoted the phosphorylation of p65, which led to the activation of the NF-κB signaling pathway (Fig 4B–4E). Subsequent investigations demonstrated that the inhibition of the NF-κB signaling pathway led to a notable reduction in the inflammatory response induced by mCK2 in mouse primary astrocytes. As illustrated in Fig 4F, the mCK2-induced upregulation of astrocyte pan-activation markers (Gfap, Cxcl10 and Steap4), LPS-specific reactive markers (C3, Iigp1 and Amigo2), and pro-inflammatory cytokines (Tnf-α and Il-6) in mouse primary astrocytes were markedly suppressed by the NF-κB-specific inhibitor BAY 11–7082. And the CK2-induced pro-inflammatory cytokines upregulation in U251 (TNF-α, IL-6, CXCL3 and CXCL10) were also blocked by BAY 11–7082 treatment (Fig 4G). The role of the NF-κB signaling pathway in CK2-induced neuroinflammation was further analyzed in a mouse model (Fig 4H). Mice were administered BAY 11–7082 (1.5 mg/kg) and a total of 0.3 μg/kg of mCK2 via intravenous injection for 2 h, and the brain hippocampal tissues from the mice were harvested to perform the real-time PCR analysis, and the results were in accordance with those of the in vitro analysis (Fig 4I). The inhibition of the NF-κB signaling pathway resulted in a significant reduction of astrocyte pan-activation markers (Gfap, Cxcl10 and Steap4), LPS-specific reactive markers (C3, Iigp1 and Amigo2), and pro-inflammatory cytokines (Tnf-α, Il-1β and Il-6) transcription induced by CK2. The in vitro and in vivo findings indicate that CK2 induces neuroinflammation by activating NF-κB signaling.

Myosin 9 mediates the CK2-induced astrocyte reaction

To identify the endogenous proteins that interact with extracellular CK2, U251 cells were incubated with 1 μg/mL recombinant CK2 (His-tagged CK2α and GST-tagged CK2β) for 3 h, after which the cell lysate supernatants were immunoprecipitated with mouse IgG, mouse-derived His-tag antibody, and mouse-derived GST tag antibody, respectively. By SDS-PAGE we detected a distinct protein band at approximately 180 kDa in the upstream lane of the GST-tag antibody incubated lysate from the CK2 protein-treated cells compared with the CK2 reconstitution buffer-treated cells. We next employed LC-MS/MS to analyze the protein composition of this specific band. By further analysis of the average peptide coverage and number of unique peptides, we identified myosin 9 as a target protein for CK2 (Fig 5A, S4 Table). Immunoblotting was used to analyze the co-precipitation of endogenous myosin 9 and CK2. CDT, which has been reported as a cell-binding exotoxin, was used as a His-/GST-labeled recombinant protein control [34]. As shown in Fig 5B, we found that endogenous myosin 9 significantly interacts with the CK2β subunit of CK2. The co-localization of extracellular CK2 and myosin 9 was also confirmed by immunofluorescence experiments, in which the recruitment of myosin 9 was observed in CK2-treated cells. We found that CK2β rather than CK2α co-localized with myosin 9, which was consistent with the Co-IP results (Fig 5C). The above results implied that exogenous CK2β can recruit and bind to myosin 9 in astrocytes. Next, the truncated forms of CSNK2B were cloned and inserted into pEGFP-C1 with an N-terminal GFP tag based on the reported domain architecture of CK2β [35] (S4A Fig). The truncated constructs of CK2β were then overexpressed in HEK293T cells, and the cells were lysed and subjected to an immunoprecipitation assay with an anti-myosin 9 antibody. The results revealed that CK2β_91–215 interacted with myosin 9 (Fig 5D). Overall, we have shown that the β-subunit of CK2 recruits myosin 9 and interacts with myosin 9 through its amino acids 91–215.

Based on the evidence that exogenous CK2 interacts with myosin 9, we investigated the potential involvement of the CK2-myosin 9 interaction in regulating the inflammatory response of astrocytes. Western blot analysis demonstrated that treatment of astrocytes with 100 ng/mL CK2 resulted in a significant time-dependent increase in the phosphorylation of myosin 9 and p65 of the mouse primary astrocytes. Furthermore, our observations indicated that CK2-induced phosphorylation of myosin 9 preceded that of p65 (S4B–S4C Fig). In contrast, treatment of 100 ng/mL α or β subunit alone was ineffective at inducing myosin 9 and p65 phosphorylation, respectively (Figs 5E and S4D). To investigate the regulatory relationship between myosin 9 and p65, we employed the CRISPR-Cas9 technique to delete myosin 9 in U251 cells. As shown in S5A Fig, myosin 9 was effectively knocked out in U251 cells. Subsequently, wild-type and myosin 9-KO U251 cells were treated with CK2, and the results of a western blot analysis indicated that myosin 9 deletion resulted in partial inhibition of p65 phosphorylation (Figs 5F and S4E). The real-time PCR results also revealed that U251 cells lacking myosin 9 presented significantly lower some of the CK2-induced pro-inflammatory factors, including TNF-α, IL-1β and CXCL10 (Fig 5H). Moreover, our findings indicate that CK2-induced phosphorylation of myosin 9 remains unaltered by pretreatment with the NF-κB-specific inhibitor BAY 11–7082 (Figs 5G and S4F). Additionally, myosin 9-KO U251 cells were utilized in the coculture model, and real-time PCR results indicated that myosin 9 depletion also reduced the transcription of meningitic E. coli-induced pro-inflammatory factors (Fig 5I and 5J). These results demonstrate that CK2 promotes the astrocyte response by regulating myosin 9 translocation and triggering downstream NF-κB activation.

CK2 deficiency in BMEC attenuates neuroinflammation induced by meningitic E. coli

It has been demonstrated that the complete absence of both CK2α and CK2β in mice is embryonically lethal because of the essential role of CK2 in development. Even conditional KO mice targeting CK2β in the central nervous system died shortly after birth [36,37]. Considering the potential impact of the loss of the catalytic subunit α in BMEC on the cell cycle, we investigated the effect of BMEC-derived CK2 on the astrocyte reaction via the knockdown of the β subunit, which is essential for the export of the CK2 holoenzyme [38]. As shown in S5B Fig, the CK2β protein was effectively knocked out in hBMEC. To ascertain whether CK2β plays a role in regulating CK2 extracellular export, the supernatants and lysates of wild-type and CK2β-knockdown hBMEC cell culture were collected. Western blot analysis revealed the absence of extracellular CK2α in the supernatant of CK2β-deficient hBMEC, even in the presence of meningitic E. coli infection (Fig 6A). Additionally, CK2β-deficient hBMEC were utilized in the coculture model, and real-time PCR results indicated that CK2β depletion also reduced some of the meningitic E. coli-induced pro-inflammatory factors transcription, including TNF-α, IL-8, MIP2, CXCL3 and CXCL10 (Fig 6B and 6C). To further elucidate the function of endothelial CK2β in astrocyte reaction in response to meningitic E. coli, a vascular-specific knockout (CK2βVas^-/-) mouse model was generated through Cre/LoxP-mediated genome engineering and knockout identification was validated by PCR (S5C Fig). CK2βVas^-/- and wild-type mice were treated with tamoxifen for 7 days. Subsequently, isolated cerebrovascular and whole-brain tissues were examined by western blot for CK2β expression. The results demonstrated that CK2β expression was undetectable in cerebrovascular samples derived from CK2βVas^-/- mice, whereas CK2β expression was detected in the CK2βVas^-/- mice whole-brain tissue. In contrast, CK2β expression was detected in both cerebrovascular samples and whole-brain tissue from wild-type mice (S5D Fig). Moreover, immunofluorescence was used to determine the expression of CK2β in the cerebral vessels of CK2βVas^-/- mice. The cerebrovascular was labeled with CD31, and no evidence of CK2β expression was detected in the cerebrovascular of CK2βVas^-/- mice (S5E Fig). The above results demonstrate the successful generation of mice with conditional deficiency for CK2β in vascular tissue. To further determine the effects of CK2β on astrocyte reaction and neuroinflammation induced by meningitic E. coli, immunofluorescence was performed to examine the expression of CXCL10, STEAP4, C3 and GFAP in infected wild-type and CK2βVas^-/- mice. It was observed that meningitic E. coli infection of wild-type mice resulted in a notable increase in GFAP fluorescence intensity. The expression of CXCL10, STEAP4, and C3 were also significantly upregulated in brain tissue. Conversely, infection-induced upregulation of CXCL10, STEAP4, C3 and GFAP was significantly reduced in CK2βVas-/- mice (Figs 6D–6F, S5F–S5H). Furthermore, we subsequently assessed its ability to regulate astrocyte pan-activation markers, LPS-specific reactive markers, and pro-inflammatory cytokines transcription in response to in vivo infection via real-time PCR. We found that Gfap, Cxcl10, Steap4, C3, Iigp1, Amigo2, TNF-α, IL-1β and IL-6 levels in the hippocampus were significantly augmented in bacterium-infected wild-type mice; however, the upregulation of these inflammatory molecules was significantly reduced in bacterium-infected CK2βVas^-/- mice (Fig 6G). Therefore, our results revealed that loss of CK2β in vascular endothelial cells significantly affects astrocyte reaction in the early stage of E. coli meningitis. Considering the deleterious effect of CK2β on the exacerbation of neuroinflammation in mice infected with meningitic E. coli, we investigated whether CK2β knockout could confer protection against meningitic E. coli-induced mortality in mice. As evidenced by the results, all mice in the uninfected wild-type and CK2βVas^-/- groups survived throughout the observation period. In contrast, the mortality rate among wild-type mice infected with bacteria was 100%. Nevertheless, CK2β knockout afforded protection from mortality in 42% of the CK2βVas^-/- mice with bacterial infection (Fig 6H). These results indicate that CK2β is indispensable for the transportation and extracellular localization of CK2. Deprivation of CK2β in BMEC prevents meningitic E. coli-induced astrocyte reaction and neuroinflammation, thereby affording the host a certain degree of protection.