Poldip2 mediates blood‐brain barrier disruption and cerebral edema by inducing AQP4 polarity loss in mouse bacterial meningitis model

Abstract Background Specific highly polarized aquaporin‐4 (AQP4) expression is reported to play a crucial role in blood‐brain barrier (BBB) integrity and brain water transport balance. The upregulation of polymerase δ‐interacting protein 2 (Poldip2) was involved in aggravating BBB disruption following ischemic stroke. This study aimed to investigate whether Poldip2‐mediated BBB disruption and cerebral edema formation in mouse bacterial meningitis (BM) model occur via induction of AQP4 polarity loss. Methods and Results Mouse BM model was induced by injecting mice with group B hemolytic streptococci via posterior cistern. Recombinant human Poldip2 (rh‐Poldip2) was administered intranasally at 1 hour after BM induction. Small interfering ribonucleic acid (siRNA) targeting Poldip2 was administered by intracerebroventricular (i.c.v) injection at 48 hours before BM induction. A specific inhibitor of matrix metalloproteinases (MMPs), UK383367, was administered intravenously at 0.5 hour before BM induction. Western blotting, immunofluorescence staining, quantitative real‐time PCR, neurobehavioral test, brain water content test, Evans blue (EB) permeability assay, transmission electron microscopy (TEM), and gelatin zymography were carried out. The results showed that Poldip2 was upregulated and AQP4 polarity was lost in mouse BM model. Both Poldip2 siRNA and UK383367 improved neurobehavioral outcomes, alleviated brain edema, preserved the integrity of BBB, and relieved the loss of AQP4 polarity in BM model. Rh‐Poldip2 upregulated the expression of MMPs and glial fibrillary acidic protein (GFAP) and downregulated the expression of β‐dystroglycan (β‐DG), zonula occludens‐1 (ZO‐1), occludin, and claudin‐5; whereas Poldip2 siRNA downregulated the expression of MMPs and GFAP, and upregulated β‐DG, ZO‐1, occludin, and claudin‐5. Similarly, UK383367 downregulated the expression of GFAP and upregulated the expression of β‐DG, ZO‐1, occludin, and claudin‐5. Conclusion Poldip2 inhibition alleviated brain edema and preserved the integrity of BBB partially by relieving the loss of AQP4 polarity via MMPs/β‐DG pathway.


| INTRODUC TI ON
Bacterial meningitis (BM) is a severe infectious disease with a high morbidity and mortality rate. 1,2 Despite the advances in antimicrobial therapy for BM, up to 50% of BM survivors develop long-term neurological sequelae. 3 One complication of BM is brain edema, which accounts for intracranial hypertension, potentially leading to brain ischemia, herniation, and even death. 4 Cerebral edema results from fluid accumulation in astrocytes due to energy failure or entry of plasma fluid into the brain extracellular space (ECS) caused by leaky blood-brain barrier (BBB). 5,6 The discovery of aquaporins (AQPs), a family of membrane-bound water channel proteins, contributes to deeper understanding of brain water transport. Aquaporin-4 (AQP4), the predominant aquaporin in the central nervous system, is expressed widely throughout the brain, particularly at brain-blood interfaces. AQP4 is abundantly expressed in astrocytic endfeet processes in direct contact with blood vessels. [7][8][9] The specific highly polarized expression of AQP4 was suggested to play critical roles in maintaining cerebral water balance and BBB integrity. [10][11][12][13] Previous studies have reported that AQP4 polarity was disturbed in various neurological diseases that cause brain edema and BBB disruption. [14][15][16][17] However, whether loss of AQP4 polarity mediates BBB disruption and cerebral edema formation in mouse BM model is still unknown.
The polarized distribution of AQP4 relies on dystrophin-glycoprotein complexes (DGC), a membrane-spanning group of proteins comprising syntrophin, dystroglycan (DG), utrophin, and dystrobrevins. [18][19][20] The proteins in DGC connect extracellular matrix (ECM) with cytoskeleton; AQP4 is tethered to syntrophin, a component of DGC; thus, AQP4 is anchored to astrocytic endfeet processes. [18][19][20] β-dystroglycan (β-DG), a membrane-spanning protein, is the central component of DGC and can be cleaved by matrix metalloproteinases (MMPs). 21,22 Therefore, we speculate that any stimuli in brain promoting MMPs activity will lead to loss of AQP4 polarity via β-DG degradation, which further affects the water transport balance in brain and the integrity of BBB.
Polymerase δ-interacting protein 2 (Poldip2) is a protein discovered originally as a binding partner of the p50 subunit of DNA polymerase δ. 23 Poldip2 has been shown to be involved in numerous cellular functions such as DNA repair, mitochondrial fusion, cytoskeletal remodeling, and cell proliferation and migration. 23 A recent study by Hernandes demonstrated that Poldip2 was upregulated following ischemic stroke. 24 However, changes in Poldip2 expression in BM animal models have not been reported. Furthermore, Hernandes's study reported that Poldip2 mediates the breakdown of BBB by MMPs activation following ischemic stroke, 24 and another study showed that loss of Poldip2 impaired MMPs activity in a hindlimb ischemic animal model. 25 Knowing that MMPs cleaved β-DG and β-DG is the central protein for AQP4 anchoring to the astrocytic endfeet processes, we induced mouse BM model by injecting mice with group B hemolytic streptococci via the posterior cistern and investigated whether Poldip2 affects AQP4 polarity via the MMPs/β-DG pathway in this model.

| Animals
All animal experiments were approved by the Animal Ethics Committee of Chongqing Medical University. A total of 259 male Kunming mice (weighing 20 ± 2 g) were used, of which 25 died during the intracisternal puncture process and were excluded from this study. All mice were housed in an automatically controlled light-temperature room with free access to food and water.

| Bacterial culture
Group B β-hemolytic streptococci strain was obtained from the microbiology laboratory of Jiulongpo People's Hospital, Chongqing.
Bacteria were inoculated on blood agar plates and incubated at 37°C for 16-24 hours in a 5% CO 2 incubator. The organisms were isolated at mid-logarithmic growth phase, washed with saline twice, and diluted to 10 8 colony-forming units (CFU)/L.

| Mouse bacterial meningitis (BM) model
The BM model was established by bacterial injection as previously described. 26 Briefly, the mice were anesthetized by intraperitoneal injection of 3.5% chloral hydrate (0.1 mL/10g), and then, 10 μL of cerebrospinal fluid (CSF) was withdrawn from intracisternal puncture followed by intracisternal injection of 10 μL β-hemolytic streptococcus suspension (10 8 CFU/L). For the sham animals, instead of bacterial suspension, 10 μL of sterile saline was injected.

| Study design
All mice were randomly distributed to the following five separate experiments.
Conclusion: Poldip2 inhibition alleviated brain edema and preserved the integrity of BBB partially by relieving the loss of AQP4 polarity via MMPs/β-DG pathway.

K E Y W O R D S
AQP4, bacterial meningitis, blood-brain barrier, cerebral edema, MMPs, Poldip2

| Experiment 1
To determine the time course of endogenous Poldip2 expression after BM induction, mice were randomly divided into seven groups: sham, 3 hours BM, 6 hours BM, 12 hours BM, 24 hours BM, 48 hours BM, and 72 hours BM (n = 6 per group). The brains of the mice were collected for Western blot analysis. In addition, double immunofluorescence staining was performed to test whether Poldip2 is expressed in astrocytes, neurons, microglia, neutrophils, and T cells at 12 hours after BM induction (n = 3 in sham and 12 hours BM group).

| Experiment 2
To choose the best Poldip2 siRNA from three candidate sequences (sequences A, B, and C, Figure 2) provided by OriGene (USA), quantitative real-time PCR and Western blot analysis were performed after BM induction. Mice were randomly divided into four groups: BM + scramble siRNA, BM + sequence A, BM + sequence B, and BM + sequence C (n = 6 per group).

| Experiment 3
The effect of Poldip2 siRNA (the best among three candidate sequences for interfering Poldip2 expression) on neurological function, brain water content, BBB permeability, astrocytic swelling, and AQP4 polarized distribution was evaluated by administering Poldip2 siRNA by intracerebroventricular (i.c.v) injection at 48 hours before BM induction. Neurological score assessment, brain water content test, EB permeability assay, transmission electron microscopy (TEM), and immunofluorescence staining experiments were performed after BM induction. Mice were randomly divided into four groups: sham, BM, BM + scramble siRNA, and BM + Poldip2 siRNA (n = 21 per group).

| Experiment 4
The downstream proteins of Poldip2 were explored, and the mechanisms underlying Poldip2 effect on BBB integrity were

| Experiment 5
To investigate whether MMPs were involved in Poldip2-mediated BBB disruption and loss of AQP4 polarity in mouse BM model, UK383367 (Selleck), a specific inhibitor of MMPs, was used to inhibit the activity of MMPs. Neurological function, brain water content, BBB permeability, and AQP4 polarity were evaluated by Loeffler's score assessment, brain water content test, EB permeability assay, and immunofluorescence staining experiments, respectively. The expression levels of β-DG, ZO-1, occludin, claudin-5, and GFAP were examined by Western blot. Mice were randomly divided into three groups: sham, BM + vehicle, and BM + UK383367 (n = 24 per group).

| In vivo RNAi
Intracerebroventricular (i.c.v) Poldip2 siRNA administration was performed as previously described. 27 Briefly, mice were anesthetized with chloral hydrate and fixed in a stereotactic frame in prone position. Next, the scalp was sterilized by isopropyl alcohol, and a 10-μL Hamilton microsyringe was inserted at 1.0 mm anterior and 1.5 mm lateral to the bregma, and 2.5 mm deep from the skull surface into the right lateral ventricle. At 48 hours before BM induction, 2 μL Poldip2 siRNA (100 pmol/μL; OriGene) or scramble siRNA (100 pmol/μL, OriGene) was injected slowly over a period of 5 min.
To prevent liquid reflux, the syringe was held in place for 10 minutes and then withdrawn slowly over a period of 5 minutes.

| Intranasal administration of rh-Poldip2
A total volume of 5 μL rh-Poldip2 (0.3 μg/mice) or vehicle (in sterile saline water) was administered into the bilateral nares. 1.25 μL rh-Poldip2 or vehicle was administered every 2 minutes into alternating nares. 28

| Neurological score assessment
Neurobehavioral functions were assessed by a blinded investigator using Loeffler's scoring method at 12 hours after BM induction, as described previously. 31 Briefly, the mice were placed on a horizontal floor and the neurobehavioral status of each mouse was recorded. The score criteria of Loeffler's method are as follows: score 1, no movement; score 2, no upright turning when positioned on the back; score 3, upright turning in more than 5 seconds; score 4, minimal ambulatory activity and upright turning in less than 5 seconds; and score 5, normal motor activity and upright turning in less than 5 seconds.

| Brain water content
Brain water content was measured at 12 hours after BM induction by dry-wet weight method as previously described. 15 Briefly, wet brain was weighed immediately after removal, dried in a vacuum oven at 120°C for 48 hours, and weighed again to determine the dry weight. Brain water content (%) was calculated as [(wet weight − dry weight)/wet weight] × 100%.

| Immunofluorescence
Immunofluorescence was performed as previously described. 15 Briefly, mice were deeply anesthetized and transcardially perfused with 30 mL 0.9% ice-cold saline followed by 4% paraformaldehyde at 12 hours after BM induction. The brains were removed and fixed in 4% paraformaldehyde overnight at 4°C and then placed in 30% sucrose until they sank. Then, the brains were embedded into optimal cutting temperature (OCT) compound and frozen, and 10-μmthick coronal slices were cut sequentially at − 20°C with a cryostat were visualized with a fluorescence microscope (IX73, Olympus), and immunoreactivity density was measured using Image-Pro Plus 6.0 software.

| Western blotting
Mice were anesthetized, and the brains were quickly removed and stored in − 80°C freezer. Western blotting was performed as previously described. 32 Briefly, mice brains were homogenized in radioimmunoprecipitation assay (RIPA) lysis buffer (Beyotime, P0013B) with protease inhibitor cocktail and centrifuged at 14 000 g for 20 minutes at 4°C. The supernatant was collected, and total protein content was determined using a Bicinchoninic Acid (BCA) Protein Assay Kit and analyzed by Quantity One software (version 4.6.2).

| Quantitative real-time PCR
Total RNA was isolated using the RNAiso Plus reagent (TaKaRa) according to the manufacturer's protocol. 33 RNA concentration was measured using a NanoDrop 1000 spectrophotometer (Thermo). The total RNA was reverse-transcribed to cDNA using

| Transmission electron microscopy (TEM)
Mice were anesthetized and transcardially perfused with ice-cold 0.9% NaCl followed by fixing solution containing 4% paraformaldehyde and 1% glutaraldehyde. The brains were removed and cut into 1 mm × 1 mm × 1 mm blocks, followed by fixing in 2.5% glutaraldehyde overnight at 4°C and postfixing in 2% osmium tetroxide for 2 hours at 4°C. Next, the blocks were dehydrated in alcohol, passed through propylene oxide, embedded in Epon-Araldite, and cut into ultrathin sections (100 nm). The prepared sections were stained with uranyl acetate and lead citrate and examined using TEM (H-7500; Hitachi).

| Evans blue (EB) permeability assay
For analysis of BBB permeability, mice were injected with 2% Evans blue (3 mL/kg) (Aladdin, China) intravenously. 34 After three hours, mice were anesthetized with 3.5% chloral hydrate and transcardially perfused with 0.1 mol/L cold phosphate-buffered saline (PBS, pH 7.4). Then, the brains were removed immediately and stored at − 80°C until use. The brains were homogenized in PBS, sonicated, and centrifuged for 30 minutes at 12 000 g and 4°C.
The supernatant was collected and equal volume of trichloroacetic acid (TCA) was added, followed by incubation at 4°C for 24 hours.
After centrifugation (12 000 g, 4°C, 30 minutes), Evans blue absorbance was measured at 610 nm using a spectrophotometer (Thermo Fisher Scientific) and interpreted based on the standard curve.

| Gelatin zymography
The process of protein extraction and total protein content detection was same as that of Western blotting. 35 Protein samples (30 μg proteins/well) were resolved using 10% SDS-polyacrylamide gel containing 0.1% gelatin at 100 V and 50 A for 2 hours. After that, the gel was washed twice with 2.5% Triton X-100 and incubated in incubation buffer (50 mmol/L Tris-HCl, 5mmol/L CaCl 2 , 1 μmol/L ZnCl 2 , 0.02% NaN 3 ) for 20 hours at 37°C. Next, the gel was stained with Coomassie Brilliant Blue (Beyotime, P0017) and photographed.
Band intensity was measured using Quantity One software (version 4.6.2).

| Statistical analysis
Statistical analyses were performed using GraphPad Prism 6 (GraphPad software). Data are expressed as mean ± standard error of mean (SEM). KS normality test was used to assess the normality of data distribution. Statistical difference between groups was analyzed using one-way analysis of variance (ANOVA) followed by multiple comparisons using Tukey's post hoc test. P values were two-sided, and P < .05 was considered statistically significant.

| Time course and spatial expression of endogenous Poldip2 in mouse BM model
Endogenous Poldip2 expression increased significantly in a timedependent manner as compared to the sham control group (P < .05), reaching a peak value at 12 hours after BM induction ( Figure 1A).

| In vivo knockdown of Poldip2 by Poldip2 siRNA
Three candidate Poldip2 siRNA sequences (A, B, and C; Figure 2A) were administered by i.c.v injection 48 hours before BM induction.
Quantitative real-time PCR analysis showed that Poldip2 mRNA expression decreased significantly in groups A and C as compared to that in the scramble group at 12 hours after BM induction (P < .05), while sequence B treatment had no significant effect on Poldip2 mRNA expression as compared to the scramble group (P > .05, Figure 2B). Western blotting revealed that Poldip2 protein expression decreased significantly in groups A and C (P < .05), but no significant difference was observed in group B (P > .05, Figure 2C,D) when compared to the scramble group. Considering that group C showed slightly lower Poldip2 expression than group A, though not significant, sequence C siRNA was selected for the further experiments.

| Effect of endogenous Poldip2 knockdown on neurological status, brain edema, and BBB disruption in mouse BM model
In the BM group, at 12 h after BM induction, Loeffler's neurological score significantly decreased (P < .05, Figure 3A) and brain water content (P < .05, Figure 3B) and EB extravasation (P < .05, Figure 3C) increased as compared to the sham group. Whereas, in the Poldip2 siRNA group, Loeffler's neurological score significantly improved (P < .05, Figure 3A) and brain water content (P < .05, Figure 3B) and EB extravasation (P < .05, Figure 3C) decreased as compared to the scramble siRNA group. TEM demonstrated obvious astrocytic swelling and tight junctions appeared as diffuse, disorganized structures in BM (12 hours after BM induction) and scramble siRNA groups, while the Poldip2 siRNA group showed reduction in astrocytic swelling as well as less disorganized tight junctions ( Figure 3D).

| Effect of endogenous Poldip2 knockdown on AQP4 polarity loss in mouse BM model
Early endosome antigen 1 (EEA1) is a marker protein of early endosome. The colocalization of AQP4 with EEA1 indicates that the polarized localization of AQP4 at astrocyte membrane is lost and AQP4 is internalized into the cytoplasm containing early endosome.
Double immunofluorescence staining of AQP4 and EEA1 showed that the number of AQP4-positive cells and the ratio of AQP4 coexpressed with EEA1 to total AQP4 increased significantly in the cerebral cortex of BM group at 12 hours after BM induction as compared to the sham group (P < .05, Figure 4A-C), but decreased in the Poldip2 siRNA group as compared to the scramble siRNA group (P < .05, Figure 4A-C). Furthermore, double immunofluorescence staining of AQP4 and GFAP showed that AQP4 colocalized with astrocytic cytoplasm in the BM group at 12 hours after BM induction while AQP4 colocalized with astrocytic endfeet in the sham group ( Figure 4D). The activity of MMP9 and MMP2 in the scramble siRNA group was higher than that in the sham group (P < .05). Poldip2 siRNA downregulated the activity of MMP9 and MMP2 as compared to the scramble siRNA group (P < .05, Figure 6).

| Effect of MMPs inhibitor on neurological deficits, AQP4 polarization, brain edema, and BBB permeability after BM induction
In the vehicle group at 12 hours after BM induction, Loeffler's neurological score significantly decreased (P < .05, Figure 7A) and brain water content (P < .05, Figure 7B) and EB extravasation (P < .05, Figure 7C) increased as compared to the sham group. Whereas, in the UK383367 group, Loeffler's neurological score increased (P < .05, Figure 7A) and brain water content (P < .05, Figure 7B) and EB extravasation (P < .05, Figure 7C) decreased as compared to the vehicle group.
Double immunofluorescence staining of AQP4 and EEA1 showed that the number of AQP4-positive cells and the ratio of AQP4 coexpressed with EEA1 to total AQP4 increased significantly in the vehicle group at 12 hours after BM induction as compared to the sham group (P < .05, Figure 8) but decreased in the UK383367 group as compared to the vehicle group (P < .05, Figure 8).
Western blotting showed that the expression levels of β-DG, ZO-1, occludin, and claudin-5 decreased and GFAP expression level increased in the BM induction group as compared to the sham group (P < .05), while the expression levels of β-DG, ZO-1, occludin, and claudin-5 increased and GFAP expression level decreased in the UK383367 group as compared to the vehicle group (P < .05, Figure 9).

| D ISCUSS I ON
Poldip2 is expressed extensively in the central nervous system and peripheral tissues, including brain, thyroid, heart, blood vessels, lungs, and kidneys. 24,[36][37][38][39][40] Many studies have reported the role of Poldip2 in peripheral tissues, 23,36-40 but its role in the brain has not been well studied and is limited to very few investigations. 23,24,41 In the present study, we observed that Poldip2 is expressed extensively in various types of cells in the mouse brain, suggesting the need to showing the effect of different Poldip2 siRNA sequences on Poldip2 protein expression. Data are presented as mean ± SEM, n = 6 per group. *P < .05 vs. scramble siRNA group brain edema, characterized by BBB disruption. AQP4 polarization on the astrocyte endfeet is the structural basis for its water transport function; therefore, loss of AQP4 polarity will disrupt BBB integrity and mediate brain edema formation.
AQP4 expression level was greatly increased in mouse BM model, which is consistent with other studies. 44,45 For water transportation function, AQP4 expression at brain-blood interfaces is essential, suggesting that a change in the location of AQP4 expression is important rather than a change in AQP4 expression level. The ratio of AQP4 coexpressed with EEA1 to total AQP4 increased in mouse BM model and decreased after Poldip2 siRNA injection, revealing that AQP4 polarization was lost in BM and Poldip2 inhibition could reverse the loss of AQP4 polarity in BM brain. However, the underlying mechanism of AQP4 polarity loss mediated by Poldip2 remains to be explored.
It has been demonstrated that Poldip2 is responsible for the up- B-E, Quantitative analysis of pro-MMP2, active MMP2, pro-MMP9, and active MMP9 activity levels. Data are presented as mean ± SEM, n = 6 per group. *P < .05 vs. sham group; & P < .05 vs. vehicle group; # P < .05 vs. scramble siRNA group F I G U R E 7 Effect of the MMPs inhibitor UK383367 on neurological functions, brain water content, and BBB permeability at 12 hours after BM induction. A, Effect of UK383367 on neurological function. Loeffler's neurological score decreased significantly in the vehicle group as compared to the sham group and increased significantly in the UK383367 group as compared to the vehicle group. B, Effect of UK383367 on brain water content. The brain water content increased in the vehicle group as compared to the sham group and decreased significantly in the UK383367 group as compared to the vehicle group. C, Effect of UK383367 on BBB permeability. EB extravasation increased in the vehicle group as compared to the sham group and decreased significantly in the UK383367 group as compared to the vehicle group. Data are presented as mean ± SEM, n = 6 per group. *P < .05 vs. sham group, & P < .05 vs. vehicle group and play important roles in BBB disruption and brain edema formation in mouse BM model.
In conclusion, our study showed for the first time that endogenous Poldip2 expression was increased in the brain of mouse BM model, and Poldip2 inhibition could alleviate brain edema and BBB disruption by reversing AQP4 polarity loss via MMPs/β-DG pathway. The mechanism of AQP4 polarity loss being induced by Poldip2 in mouse bacterial meningitis model is summarized in graphical abstract.
Moreover, Poldip2 was expressed not only in astrocytes but also in microglia and neurons, which suggests that Poldip2 may play multiple roles in inflammation, neuronal apoptosis, and so on. However, our study focused only on the role of Poldip2 in BBB integrity and brain edema in mouse BM model; its effects F I G U R E 8 Effect of the MMPs inhibitor UK383367 on AQP4 polarized expression. A, Double immunofluorescence images showing colocalization of AQP4 with EEA1 in the sham, BM + vehicle, and BM + UK383367 groups. The nuclei were stained by DAPI (blue). Top panel indicates the location of immunofluorescence staining (small black box). Scale bar = 25 μm. B-C, Quantification of AQP4-positive cells and ratio of AQP4 coexpressed with EEA1 to total AQP4 in different groups. Data are presented as mean ± SEM, n = 6 per group. *P < .05 vs. sham group, & P < .05 vs. vehicle group on inflammation and neuronal apoptosis need to be elucidated in further studies.

ACK N OWLED G M ENTS
We thank the microbiology laboratory, Jiulongpo People's Hospital, Chongqing, for providing the strain of group B β-hemolytic streptococcus, and Dr Ge Xu from the Institute of Life Sciences, Chongqing Medical University, for providing experimental assistance.

CO N FLI C T S O F I NTE R E S T
The authors declare that they have no conflicts of interest.

S U P P L E M E N TA RY DATA
The original images of western blots can be found at Appendix S1.