Neuroprotective and Neuro-rehabilitative Effects of Acute Purinergic Receptor P2X4 (P2X4R) Blockade after Ischemic Stroke

Abstract

Stroke remains a leading cause of disability in the United States. Despite recent advances, interventions to reduce damage and enhance recovery after stroke are lacking. P2X4R, a receptor for adenosine triphosphate (ATP), regulates activation of myeloid immune cells (infiltrating monocytes/macrophages and brain-resident microglia) after stroke injury. However, over-stimulation of P2X4Rs due to excessive ATP release from dying or damaged neuronal cells can contribute to ischemic injury. Therefore, we pharmacologically inhibited P2X4R to limit the over-stimulated myeloid cell immune response and improve both acute and chronic stroke recovery. We subjected 8–12-week-old male and female wild type mice to a 60 minute right middle cerebral artery occlusion (MCAo) followed by 3 or 30 days of reperfusion. We performed histological, RNA sequencing, behavioral (sensorimotor, anxiety, and depressive), and biochemical (Evans blue dye extravasation, western blot, quantitative PCR, and flow cytometry) analyses to determine the acute (3 days after MCAo) and chronic (30 days after MCAo) effects of P2X4R antagonist 5-BDBD (1 mg/kg P.O. daily × 3 days post 4 h of MCAo) treatment. 5-BDBD treatment significantly (p<0.05) reduced infarct volume, neurological deficit (ND) score, levels of cytokine interleukin-1 beta (IL-1β) and blood brain barrier (BBB) permeability in the 3-day group. Chronically, 5-BDBD treatment also conferred progressive recovery (p<0.05) of motor balance and coordination using a rotarod test, as well as reduced anxiety-like behavior over 30 days. Interestingly, depressive-type behavior was not observed in mice treated with 5-BDBD for 3 days. In addition, flow cytometric analysis revealed that 5-BDBD treatment decreased the total number of infiltrated leukocytes, and among those infiltrated leukocytes, pro-inflammatory cells of myeloid origin were specifically reduced. 5-BDBD treatment reduced the cell surface expression of P2X4R in flow cytometry-sorted monocytes and microglia without reducing the total P2X4R level in brain tissue. In summary, acute P2X4R inhibition protects against ischemic injury at both acute and chronic time-points after stroke. Reduced numbers of infiltrating pro-inflammatory myeloid cells, decreased surface P2X4R expression, and reduced BBB disruption are likely its mechanism of neuroprotection and neuro-rehabilitation.

1. Introduction

Ischemic stroke is one of the primary causes of death and long-term disability globally. Despite recent advances, therapeutic drug interventions to reduce damage and enhance recovery after stroke are lacking (Feigin et al., 2015). This therapeutic gap creates an urgent need to identify and validate new targets for improved stroke recovery. Studying the sequential pathological events after stroke is one of the primary means to identify new targets. After stroke, a burst of adenosine triphosphate (ATP) is released from dying brain cells, which activates P2X (ATP-gated cation channels) purinergic receptors (Huang et al., 2014; Verma et al., 2017) as well as, potentially, other purinergic receptors. In particular, P2X4R, which is highly expressed on immune cells (North and Jarvis, 2013), plays a predominant role in modulating the responses of myeloid cells (both central resident microglia and peripheral monocytes) and mediates their activation following stroke injury (Verma et al., 2017). An over-stimulation of P2X4R, due to release of ATP from dying or damaged neuronal cells, contributes to ischemic injury (Cheng et al., 2014; Rivera et al., 2016). We have previously shown that the absence of P2X4R in knockout mice provided acute neuroprotective effects but caused impaired long-term recovery (Verma et al., 2017). The reason for such paradoxical findings is not known but may be explained by P2X4R’s dual role of modulating acute ischemic brain damage and impacting long-term chronic recovery. P2X4R regulates at least two signaling cascades: one that leads to the formation and activation of inflammasome via plasma membrane receptors (Burnstock, 2008), and another via intracellular endo-lysosomal receptors that triggers synthesis and release of brain-derived neurotrophic factor (BDNF), which has neuro-restorative effects (Suurväli et al., 2017). Hence, we hypothesize that blocking P2X4R activity during the initial phase of stroke, where its pro-inflammatory action exacerbates stroke injury, is beneficial without compromising its positive restorative function during the delayed phase of stroke recovery. Therefore, temporal modulation of P2X4R activity holds promise for better outcomes after stroke.

To examine the effect of acute blockade of P2X4R on ischemic stroke recovery, we utilized an established specific inhibitor of P2X4R, 5-(3-bromophenyl)-1,3-dihydro-2H-benzofuro [3,2-e]-1,4-diazepin-2-one (5-BDBD). 5-BDBD is a potent (IC₅₀ ~ 0.5 μM) and specific P2X4R antagonist, which has no apparent effect on other CNS-dominant P2X receptors, e.g., P2X7R (Coddou et al., 2019). Peripherally, it shows very high affinity towards P2X4R but low affinity for P2X1R and P2X3R (Balázs et al., 2013), hence it is one of the most selective, commercially available P2X4R inhibitors. Here, we pharmacologically inhibited P2X4R with 5-BDBD during the acute stroke phase and studied functional recovery at both acute and chronic time points. We used histological, biochemical, molecular, and behavioral approaches to support our hypothesis that early P2X4R inhibition is neuroprotective and leads to benefits during chronic stroke recovery.

2. Material and Methods

2.1. Materials

5-BDBD (Cat. SML0450) was obtained from Sigma-Aldrich, St. Louis, MO. Flow cytometry fluorophores CD45-efluor450 (Cat 48045182), CD11b-APC-eFluor780 (Cat. 47011282), and Ly6C-PerCP-Cy5–5 (Cat. 45593282), as well as Chemiluminescent HRP substrate (Cat. 34580) and BCA Protein estimation kit (Cat. 23227) were procured from Thermo Fisher Scientific Inc., Rockford, IL. Fluorophore Ly6G-PE (Cat 501276U025) was from Tonbo Biosciences, San Diego, CA., IL.

Primary antibodies used in western blot against P2X4R (Cat. 135341AP), BDNF (Cat. 256991AP), Beta-Actin (Cat. HRP-60008), and secondary antibody Anti-mouse (Cat. SA00002–1) were from Protein Tech, Rosemont, IL. Primary anti- IBA1 (Cat. Ab5076) was procured from Abcam, Cambridge, MA. Secondary anti-rabbit (Cat. 7074S) was from Cell Signal Tech., Danvers, MA and anti-goat Alexa fluor 594 (Cat. A11058) was from Thermo Fisher Scientific Inc., Rockford, IL. Trizol (Cat.15596026) and TaqMan universal master mix reagent (Cat.444040) were from Ambion, Life Technologies, Camarillo, CA.

2.2. Mice

We used age- and weight-matched C57B/6 male and female (wild-type) WT mice generated in our breeding colony at UConn Health animal facility. Mice were fed standard chow diet and water ad libitum. Standard housing conditions were maintained at a controlled temperature with a 12-h light/dark cycle. All experiments were approved by the Institutional Animal Care and Use Committee of University of Connecticut Health and conducted in accordance with the U.S. National Institutes of Health Guidelines for the Care and Use of Laboratory Animals.

2.3. Experimental design and animals

Eight- to twelve-week-old mice were subjected to right middle cerebral artery occlusion (MCAo) for 60 min followed by 3 (acute) or 30 (chronic) days of reperfusion. A total of 174 mice were randomly divided into vehicle (0.5% methyl cellulose, oral gavage) and 5-BDBD (1 mg/kg PO daily × 3 days starting after 4 h of MCAo) groups. After stroke, mice in the acute cohort were sacrificed for various biochemical and histological analyses, while mice in the chronic survival cohort were subjected to neurobehavioral analysis at days 2, 7, 14, 21, and 28 using rotarod and open-field tests, and day 29 for tail suspension test. A detailed list of mice and their use in each group are given in Table 1. Mice from the chronic survival cohort were harvested for either immunohistochemistry (IHC) infarct/atrophy assessment after paraformaldehyde perfusion fixation or were flash-frozen for subsequent biochemical analysis. Similarly, mice from the acute survival group were either fixed or flash-frozen for various activities as listed in Table 1. We chose 3 and 30 days post-stroke as representative time points for acute outcomes (peak of post-stroke inflammation) and long-term chronic recovery (translationally relevant), respectively.

Table 1:

Number of mice utilized in each experiment

Groups	Acute (3 days post-stroke)					Long-Term (30 days post-stroke)
	IHC/ Infarct analysis	Western/ Protein	Flow Cytometry	Evans Blue Dye (EBD)	qPCR/ RNA Seq	Western/ Protein	IHC/ Atrophy/Behavior	qPCR/ RNA Seq
Vehicle	12	8	11	7	8	6	18	5
Excluded (died)	1	2	2	2	2	3	6	1
*5-BDBD	24	6	12	9	6	4	19	4
Excluded (died)	0	1	2	1	1	0	1
Total no. of mice	37	17	27	19	17	13	44	9

Note:

^*For 5-BDBD Group IHC/Infarct Analysis consists of n=3 (0.3 mg/kg); n=3 (3 mg/kg); n=18 (1 mg/kg). Other parameters of 5-BDBD were recorded at dose 1mg/kg.

2.4. In vitro efficacy of 5-BDBD

Human P2X4R-expressing cells (HEK 293) were trypsinized, counted, and seeded in black, clear-bottomed 96-well plates at a density of 50,000 cells per well and incubated overnight. The next day, media was removed from cell plates and 25 μl of assay buffer (1.11 mM CaCl₂, 0.43 mM MgCl^.6H₂O, 0.36 mM MgSO₄^.7H₂O, 4.98 mM KCl, 0.39 mM KH₂PO₄, 122 mM NaCl, 0.3 mM Na₂HPO₄, 4.86 mM D-glucose, 17.7 mM HEPES, pH 7.4) was added. Calcium 5 dye (Molecular Devices, San Jose, CA) was diluted in assay buffer and 10 μl was added to the wells and incubated at 37°C for 1 hour. Compound dilutions (including serial dilutions) were performed in 100% DMSO then transferred to intermediate dilutions for a very limited amount of time (<10 min) just before adding to the cell plate. The final concentration of DMSO in the assay was 0.5%. Various concentrations of 5-BDBD were added using a manual multichannel pipettor and incubated for 10 min at room temperature. The plates were then placed in the fluorescent imaging plate reader (FLIPR) and fluorescence monitored every 1.52 seconds. After baseline reading of standard agonist ATP (10 μl of 1 μM) for 20 seconds, ivermectin (100 nM) was added and the fluorescence was monitored for 5 min at excitation/emission: 488 nm/510–570 nm. Details of their HEK 293 stability and validation data is available on https://www.sbdrugdiscovery.com/p2x-receptors.asp which was the original source this cells and data. Briefly HEK cells were transfected with sequence-verified P2X4 expression plasmid using standard chemical transfection reagents and single cell clones were cultured under antibiotic selection. Surviving clones were screened using a FLIPR assay to identify clones showing correct pharmacology and optimal window over un transfected cells. The best clone was then characterized using manual patch clamp electrophysiology. Attached is a datasheet from the manual patch clamp assay. More details can be found on https://www.sbdrugdiscovery.com/datasheets/SB-HEK-P2X4%20Datasheet.pdf

2.5. Middle cerebral artery occlusion (MCAo)

We induced focal transient cerebral ischemia by a 60 min right MCAo under isoflurane anesthesia followed by reperfusion for either 3 or 30 days as described previously (Verma et al., 2017). Briefly, we proceeded with midline ventral neck incision and unilateral right MCAo using a 6.0 silicone rubber-coated monofilament (size 602145/602245; Doccol Corporation, Sharon, MA) placed 10–11 mm away from the bifurcation point of the internal carotid artery through an external carotid artery stump. Rectal temperatures were monitored and maintained at 37±0.5°C with the help of a heating pad. We used laser Doppler flowmetry (DRT 4/Moor Instruments Ltd, Devon, UK) to measure cerebral blood flow and to confirm occlusion [reduction to 15% of baseline cerebral blood flow; {452±72 (pre MCAo or baseline vrs. 69±17 (after occlusion) perfusion unit}] and reperfusion. All animals were fed wet mash for one week after surgery to ensure adequate nutrition for chronic endpoints, as animals have rearing deficits after stroke. In sham surgery mice, we performed identical surgeries except the suture was not advanced into the internal carotid artery.

2.6. Cresyl violet staining for infarct volume and tissue atrophy analysis

At the end of the experiment, mice were anaesthetized using an overdose of avertin (250 mg/kg intraperitoneally [i.p.]). Following blood collection, mice underwent transcardiac perfusion with cold phosphate buffered saline (PBS) followed by 4% paraformaldehyde. Brains were fixed overnight in 4% paraformaldehyde and then transferred to cryoprotectant (30% sucrose in 1x PBS) for 72 h prior to sectioning. Brains were then sliced into 30 μm free-floating sections using a freezing microtome. A total of eight slices covering ischemic territory were mounted and stained with cresyl violet to assess infarct volume (3 days after stroke) and tissue atrophy (30 days after stroke) as described earlier (Verma et al., 2017). An investigator blinded to the experimental cohort conducted the data analyses.

2.7. Quantitative PCR (qPCR)

Trizol (Ambion, Life Technologies Camarillo, CA) method was utilized for total RNA isolation from right hemisphere (perilesional cortex) of vehicle-treated and drug-treated WT mice. Reverse transcription (RT) and qPCR was carried out in accordance with manufacturer instructions - TaqMan RNA reverse transcription kit and TaqMan universal master mix reagent (Ambion, Life Technologies, Camarillo, CA).

2.8. RNA-Seq

Details can be found in supplementary files

2.9. Behavioral analysis

Details can be found in supplementary files

2.10. Flow cytometry analysis of resident microglia and infiltrating monocytes

For sorting microglia and monocytes using flow cytometry, we collected and prepared tissue samples as described before (Verma et al., 2017). Briefly, all the cells isolated from the interphase 70%/30% Percoll gradient were washed and blocked with mouse Fc Block (clone 93, eBioscience, Thermo Fisher, Waltham, MA) before staining for 30 min in dark and cold conditions with primary antibody-conjugated fluorophores: CD45-eF450 (clone 2D1, eBioscience), CD11b-APCeF780 (clone M1/70, eBioscience), Ly6G-AF700 (clone 1A8, eBioscience), Ly6C-PerCP-Cy5.5 (clone HK1.4, BioLegend, San Diego, CA), CX3CR1 APC (Clone SA011F11, BioLegend), and P2X4R-FITC (AFR-024-F, Alomone Labs, Jerusalem, Israel). Cells with marker profile CD45^int CD11b⁺Ly6C⁻CX3XR1+ were identified as resident microglia, while those with CD45^hiCD11b⁺Ly6C⁺Ly6G⁻ were categorized as bone marrow-derived monocytes and CD45^hiCD11b⁺Ly6C⁺Ly6G⁺ as bone marrow-derived neutrophils. For live/dead cell discrimination, a fixable viability dye, carboxylic acid succinimidyl ester (CASE-AF350, Invitrogen, Carlsbad, CA), was used at a dilution of 1:300 from a working stock of 0.3 mg/mL. LSR II flow cytometer (BD Biosciences, Billerica, MA) enabled with FACsDIVA 6.0 (BD Biosciences, Billerica, MA) and FlowJo (Treestar Inc., Ashland, OR) software was used for data acquisition. The gating for each antibody being used was determined based on fluorescence minus one control.

2.11. ImageStream data acquisition and analysis

Flow-sorted microglia and monocytes from mouse brain were fixed with 2% paraformaldehyde and images of individuals from each sample (1000 cells/sample) were acquired on the ImageStream®X MARK II (Amnis Corp., Seattle, WA). P2X4R-FITC and CD11b-APCeF780 cell-surface expression levels were analyzed using the ImageStream Data Exploration and Analysis Software (IDEAS, Amnis) as described previously (Lapilla et al., 2011). Ratio of P2X4R/CD11b surface expression was used on Y axis to compare P2X4R surface expression between the groups and cells type. CD11b was used a marker of myeloid cells.

2.15. Western blot analysis

Mice were euthanized 3 or 30 days after stroke surgery with an overdose of avertin (250 mg/kg, i.p.). Following collection of blood from the right ventricle, mice underwent transcardiac perfusion using cold PBS supplemented with 5 mM EDTA. The brain was removed and perilesional cortical tissue from the right (ischemic) hemisphere was isolated and homogenized as described previously (Verma et al., 2016). Protein concentration was determined by bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific Inc., Rockford, IL). 20 mg of protein was loaded into each well of 4–15% polyacrylamide gels (Bio-Rad, Hercules, California) and transferred to PVDF membranes (Bio-Rad, Hercules, California). Membranes were incubated with primary antibodies overnight [P2X4R (1:500); BDNF (1:1000)] followed by incubation in HRP-linked secondary antibody. Chemiluminescent HRP substrate was used for developing the blot and β-Actin (1: 3000) was probed as loading control.

2.16. Immunohistochemistry

Immunostaining was used to visualize differences in the quantity and morphology of microglia/macrophages between 5-BDBD and vehicle-treated groups. Additional brain slices (30 μm) that were sectioned on a freezing microtome for infarct and atrophy analyses were mounted on the glass side. Antigen retrieval was done using citrate buffer (pH 6.0) and sections were incubated with blocking buffer followed by incubation overnight with primary antibodies [TMEM119 1:500; Ionized calcium binding adaptor molecule 1 (IBA1) 1:1000; from Abcam (Cambridge, MA) and P2X4R (1:150, Alomone Labs, Jerusalem, Israel)]. Three coronal brain sections per mouse (n = 4 per group) were taken 0.45 and 0.98 mm from bregma, stained, and visualized for quantification at 20x magnification at the junction of core and penumbra regions. A blinded observer quantified IBA1, TMEM119, and P2X4R-positive alone or co-labelled cells using Image J software (NIH). DAPI staining was used to determine the number of nuclei and to assess gross cell morphology. The average numbers of cells visualized from 3 separate regions at the junction core and penumbra were recorded for each mouse.

2.17. Evans blue dye (EBD) blood brain barrier (BBB) permeability assay

EBD (4 ml/kg body weight) was prepared as a 1% solution in 0.9% saline and injected as a single bolus (i.v.) 2 h prior to sacrifice in stroke mice. Thereafter, mice were euthanized by overdose of avertin (250 mg/kg, i.p.) and were transcardially perfused with 0.9% saline to wash out the remnant dye and brains were collected. Right hemispheres were isolated followed by extraction of EBD in 50% TCA at 1:3 volume ratios. Isolated brain regions were homogenized with a Dounce homogenizer and the extracts were centrifuged at 10,000 × g for 20 min to remove precipitates. Supernatant was collected and 30 μl was added per well of a 96-well plate following addition of 90 μl of 95% ethanol. Proper mixing was assured by pipetting and absorbance was read at 620 nm. A standard curve (1–200 μg) was used to calculate unknown concentrations of EBD uptake and data expressed as μg/mg tissue.

2.18. ELISA for IL-1β and TNF-α expression in plasma and tissue

Blood samples from both vehicle- and 5-BDBD-treated groups were collected in 40 mM

EDTA before sacrificing mice. Plasma was collected by centrifugation (10,000 × g, 10 min, 4°C) and stored at −80°C until further use. Similarly, a small fraction of whole cell tissue lysate from biochemical analysis was separated for tissue analysis of IL-1β and TNF-α using a Mouse ELISA ready-set-go kit (eBioscience).

2.19. Statistics

Data from each experiment were represented as mean ± SD and significance was determined using Student’s t-test for comparing the experimental groups between vehicle and drug (5-BDBD) treatment at days 3 and 30. If there were more than two groups, the data were analyzed by one-way or two-way analysis of variance (ANOVA) with a Bonferroni post-hoc test with repeated measure for correction of multiple comparisons (GraphPad Prism Software Inc., San Diego, CA). Owing to the ordinal nature of ND scores, the Mann-Whitney U test was used for ND score analysis. For survival curve analysis we used log-rank (Mantel-cox) analysis to determine statistical significance between the groups. The probability value of p<0.05 was considered statistically significant. An investigator blinded to the experimental groups carried out the behavioral experiments and final data analyses of all the experiment. Experimenters were not blinded for biochemical experiments.

3. Results

In this study, we analyzed the acute and chronic neuroprotective and neuro-rehabilitative effects of short-term 5-BDBD treatment after MCAo-induced stroke. Our experimental design for pharmacological treatment in stroke, and post-stroke behavioral, histological, and biochemical analyses is presented in Figure 1.

Figure 1. Study design of 5-BDBD treatment in chronic stroke.

D=day; ND=neurological deficit score; OFT=open-field test; BBB=blood brain barrier; WB=western blot; IHC=immunohistochemistry; TST=tail suspension test.

3.1. 5-BDBD is a potent inhibitor of P2X4R

We used the HEK293 cell line stably expressing P2X4R to calculate IC₅₀ of 5-BDBD, which was found to be 3.8 × 10⁻⁶ M (Fig. 2).

Figure 2. Determining IC50 of 5-BDBD.

Plot indicates the effect of 5-BDBD at various concentrations (10 nM to 30 μM; log10 [nM] = −4.52 to −8) after P2X4R agonist ATP (1 μM) induced [Ca²⁺]i response. Calculated IC50 was 3.8 × 10⁻⁶ M.

3.2. 5-BDBD treatments confer acute neuroprotection

We optimized the dose of P2X4R inhibitor 5-BDBD based on its effects on brain infarct volume. We saw a significant reduction (Supplementary Fig. 1) in infarct volume both at 1 and 3 mg/kg dose (P.O. daily dose for 3 days after stroke). We chose a dose of 1 mg/kg of 5-BDBD for all other experiments. WT mice treated with 1 mg/kg/day 5-BDBD showed significant reductions not only in cortical and hemispheric infarct volumes (and a reduction trend in striatum) but also neurological deficit (ND) score at 3 days after stroke (Fig. 3). 5-BDBD treatment did not reduce mortality at 3 days after stroke but significantly reduced mortality by day 30 after stroke (Supplementary Fig. 2). We did not see any change in tissue atrophy after 5-BDBD treatment at 30 days post-stroke (Supplementary Fig. 3). These data are consistent with our prior finding, that loss of P2X4R confers acute neuroprotection after stroke (Verma et al., 2017).

Figure 3. 5-BDBD treatment reduces infarct volume and neurological deficit during the acute stage of stroke.

Representative cresyl violet stained section showing infarct area (dotted line) in upper panel. Three-day treatment with 5-BDBD (1 mg/kg p.o./day in 1% DMSO and 0.5% methyl cellulose suspension) significantly (#p<0.05; vs Veh; Student’s t-test) reduced total hemispheric and cortical infarct volume (n=8–12/group) (middle panel) and neurological deficit (ND) score after stroke (lower panel). Middle panel graph is mean ± SD (*p<0.05; vs. Veh; Mann-Whitney U test) and lower panel is a box and whisker plot with median (horizontal red line within box), upper/lower quartiles (top/bottom of box), and maximum/minimum (top/bottom bars) values.

3.3. RNA-Seq analysis of gene expression profiles after 5-BDBD treatments

We used RNA-Seq analysis to determine the changes in expression levels of a total of 52,637 mRNAs from perilesional cortexes of 5-BDBD-treated mice after 3 days of stroke. We found 15 differentially expressed transcripts (DETs) in this treatment group using a threshold of >1.5-fold change in expression and p adjusted value <0.05 (Table 2). The number of up-regulated DETs was greater than the number of down-regulated DETs, 10 vs 5. Among the most down-regulated genes, Zfp708, Zfp 260, and Mmp9 are indirectly or directly related to extracellular matrix (ECM) receptor pathways implicated in BBB integrity (Dong et al., 2009; Han et al., 2016), suggesting a potential role of 5-BDBD treatment on regulating BBB permeability. Other DETs affected by 5-BDBD are listed in Table 2. Briefly, COX6A2 affects ROS production and is an important component of the electron transport chain (Quintens et al., 2013), Fam196a (also termed INSYN2A) is a constituent present at the inhibitory synapses involved in protein-protein interactions (Uezu et al., 2016), fgf16 is involved in enhancing cell survival (Maddaluno et al., 2017), and Tmprss6 is involved in iron homeostasis and is downregulated by inflammation (Meynard et al., 2013).

Table 2.

Downregulated (Red) and upregulated (Green) mRNA transcripts in the brain tissue of 5-BDBD- treated mice.

S. No.	N	log2Fold Change	P value	Adjusted P value
1	Zfp708	−19.9	1.32E-09	4.09E-05
2	Lox	−4.57	6.31E-07	0.006514
3	Mmp9	−3.5	1.98E-05	0.04517
4	Enpp1	−2.93	9.62E-06	0.033202
6	Zfp260	−2.55	5.62E-07	0.006514
7	Sar1a	−1.49	1.71E-05	0.04517
8	Tceal3	1.097	2.19E-06	0.013593
9	Cox6a2	1.373	1.12E-05	0.034773
10	Fam196a	1.707	2.04E-05	0.04517
11	Rprd1b	2.266	9.65E-06	0.033202
12	Gm45623	2.472	1.87E-05	0.04517
13	Gm42763	5.759	2.37E-05	0.048929
14	Fgf16	6.371	6.66E-06	0.033202
15	Tmprss6	7.146	1.03E-06	0.007974
16	Mir8114	9.57	7.76E-06	0.033202

3.4. 5-BDBD reduces BBB permeability of EBD

BBB integrity was investigated in sham surgery and stroke (vehicle- and 5-BDBD-treated) WT mice at 3 days after MCAo using the EBD method. Control sham mice did not show any EBD extravasation, while vehicle-treated mice showed apparent EBD extravasation extending from striatum to cortical regions. 5-BDBD treatment significantly reduced EBD extravasation, which was confined to the striatal region (Fig. 4).

Figure 4. 5-BDBD treatment reduces EBD extravasation after stroke.

EBD was injected as a single bolus (i.v.) 2 h prior to sacrifice as 2% solution at 4 ml/kg body weight. A) Representative images showing EBD extravasation in brain parenchyma. B) 5-BDBD-treated group showed a significant reduction in EBD accumulation in brain parenchyma compared with vehicle-treated group, as measured by spectrophotometer. Data are mean ± SD μg/gm of tissue weight (n=7–9). *p<0.05; 5-BDBD vs. Vehicle; Student’s t-test.

3.6. 5-BDBD treatment reduces total leukocytes and myeloid cell infiltration

Quantitative analysis of infiltrated leukocytes by flow cytometry revealed that 5-BDBD treatment significantly reduced infiltrated total leukocytes including both neutrophils and monocytes (myeloid cells) into the brain 3 days after stroke compared with vehicle (Fig. 5A–D; Gating strategy shown in Supplementary Fig. 4). Additionally, 5-BDBD treatment showed a trend toward increased microglia (Fig 5D). Among the infiltrated leukocytes, the number of Ly6G^+hi neutrophils were significantly reduced in 5-BDBD-treated mice (Fig 5E). 5-BDBD also significantly reduced the total number of infiltrated monocytes, however, no difference was found in the levels of Ly6C^+hi monocytes between 5-BDBD and vehicle groups (Fig. 5E). These data suggest that 5-BDBD confers its neuroprotective effects by reducing infiltration of total leukocytes including myeloid cells. These results are consistent with a functional correlation between preserved BBB and decreased infiltration of leukocytes (Jickling et al., 2015).

Figure 5. 5-BDBD treatment reduces total infilterated leukocytes, Ly6G+ neutrophils, and Ly6C+ monocytes in the ischemic hemisphere of the brain at 3 days after stroke.

Representative flow plot with the gating strategy used to identify A) brain-resident microglia (CD45^int CD11b+ Ly6C−) and infiltrated CD45^hi CD11b+ myeloid cells (For detail see supplementary fig 4). B) Ly6G^hi and Ly6G^low subpopulation of (CD45^hiCD11b+Ly6G+) neutrophils. C) Ly6C^hi and Ly6C^low subpopulations of CD45^hiCD11b+Ly6C+Ly6G− monocytes^. D) 5-BDBD treatment significantly reduced (*p<0.05 vs Veh; n=6/group, student’s t-test) total infiltrated leukocytes (left panel) as well as number of infiltrating Ly6G+ neutrophils and Ly6C+ monocytes (right panel) in the ischemic hemisphere of the brain. E) Upon further separation of Ly6G+ neutrophils and Ly6C+ monocytes between high and low populations of their respective markers, 5-BDBD treatment significantly reduced the mature Ly6G hi neutrophil population (*p<0.05 vs Veh; n=7/group; t-test). Ratio of Ly6C hi vs low was not significantly different in the ischemic hemisphere of the brain despite their overall low counts. Data are mean ± SD.

3.7. Effect of 5-BDBD on BDNF mRNA and protein levels

We next studied changes in BDNF mRNA expression following acute blockade of P2X4R at both 3 and 30 days after stroke. 5-BDBD treatment increased BDNF mRNA levels at day 3 (2.2-fold) and day 30 (6-fold) (Supplementary Fig. 5), which was unexpected, especially at an early time point (day 3), as P2X4R blockade is thought to reduce BDNF synthesis (Trang et al., 2009). To confirm if these effects were restricted to mRNA level or also translated to protein level, we measured BDNF protein expression. Analysis of BDNF protein forms showed significantly reduced pro BDNF precursor (proBDNF) and notable trend in reduction of mature BDNF (mBDNF) protein expression at day 3 after 5-BDBD treatment (Fig. 6A). Interestingly, at day 30 after stroke, total expression of proBDNF and mBDNF proteins (Fig. 6B) and ratio of mature/proBDNF protein (Fig. 6C) were significantly increased with 5-BDBD treatment vs vehicle. These data suggest that acute blockade of P2X4R by 5-BDBD may transiently block BDNF protein synthesis, but during chronic recovery when P2X4R is no longer inhibited, BDNF synthesis increases due to stimulation of P2X4R by endogenous ATP

Figure 6. 5-BDBD treatment show differential response on BDNF levels at 3 days and at 30 days after stroke.

Representative western blots are shown in the top panel. Effects of vehicle or 5-BDBD on pro and mature BDNF at (A) 3 days and (B) 30 days after stroke. C) Effects of vehicle and 5-BDBD treatment on pro/mature BDNF ratio 3 and 30 days after stroke. (D) P2X4R expression at 3 and 30 days after stroke were not different; Data are Mean ± SD (n=4–8 mice/group/time point). *p<0.05 for 5-BDBD vs vehicle by Student’s t-test. β-actin was used as internal control to normalize the data.

3.8. 5-BDBD treatment does not affect total P2X4R expression but reduces cell surface expression of P2X4R

Treatment with 5-BDBD caused no significant change in total P2X4R protein expression measured by western blot in peri-infarct brain cortex tissue at either 3 or 30 days after stroke (Fig. 6D). However, immunostaining of brain sections showed that 5-BDBD treatment significantly reduced P2X4R-expressing IBA1 positive cells in the perilesional cortex (penumbra tissue) at 3 days after stroke (Fig. 7). 5-BDBD treatment also increased the number of microglial cells in the perilesional tissue as indicated by increased number of TMEM119 +ve cells (Fig 8) probably due to reduced microglial death. While TMEM119 is expressed only in brain resident microglia and stains only resting microglial cells (van Wageningen et al., 2019), the IBA1 is expressed in all the myeloid cells (i.e. Both microglia and monocyte/macrophage) and imaging analysis of IBA1 positive cells can be used to distinguish highly activated, moderately activated or resting myeloid cells, specifically microglia which show changes in morphology during stress response. Amoeboid or round morphology without process indicates a highly activated state of microglia (Fig. 8 inset) (shorter processes, less arborization, and larger soma indicate intermediate activation of microglia, and small soma with elongated processes indicate non-activated or regular cell type (Coddou et al., 2019; Verma et al., 2017). We found that 5-BDBD treatment reduces microglia activation as suggested by the microglial phenotype in the perilesional cortex (Fig 8). Given that both infiltrating monocytes and highly activated microglia show morphologically indistinguishable phenotypes (i.e. both are round cells and express IBA1), we used flow cytometry analysis to sort microglia and infiltrating monocytes at 3 days after stroke as shown in fig 5 and subjected them to ImageStream analysis. We found that: 1) Stroke increases cell surface expression of P2X4R; 2) cell surface expression of P2X4R was reduced after 5-BDBD treatment in both microglia and monocytes; and 3) microglia express significantly less cell surface expression of P2X4R as compared to monocytes (Fig. 9). Overall, these data suggest that 5-BDBD decreases microglia/monocyte activity via reducing P2X4R activity but not its expression level.

Figure 7. P2X4R-expressing IBA1+ myeloid cells are significantly reduced near the core/penumbra junction after 5-BDBD treatment.

Images (top) and quantification (bottom) of P2X4R expression (green) in lba-1 positive cells (red) from immunostaining of brain sections 3 days after stroke. 20x magnification. DAPI (blue) = 4′,6-diamidino-2-phenylindole (nuclear stain). Scale bar 20 μm. Data are Mean ± SD (n=3 mice/group). *p<0.05, 5-BDBD vs. Vehicle; Student’s t-test.

Figure 8. 5-BDBD decreases microglia/monocyte activation after ischemic stroke.

A) 5-BDBD-treated mice showed intermediate activation of TMEM119+ and/or IBA1+ microglia (arrows), based on shorter processes, less arborization, and larger soma, as compared to Vehicle-treated group, which showed an amoeboid or round morphology characteristic of a highly activated state (arrows). White dotted line box in CV stained brain section shows one of the areas of penumbra tissue used in imaging. TMEM119 (green; marker for brain-resident microglia); IBA1 (red; microglia/monocyte activation marker); DAPI (blue) = 4′,6-diamidino-2-phenylindole (nuclear stain). B) A quantitative analysis suggests no change in total IBA1+ myeloid cells in the penumbra region, which might be due to reduced microglia survival and increased monocyte infiltration or vice versa in both vehicle and 5-BDBD treated group respectively. TMEM119+ microglia were increased in penumbra region of the 5-BDBD-treated group suggesting that microglial cell death is reduced in this group. (10x; scale bar 50 μm). Data are Mean ± SD (n=3mice/group). *p<0.05, 5-BDBD vs. Vehicle; Student’s t-test.

Figure 9: Treatment with 5-BDBD reduces P2X4R surface expression on both brain resident microglia and infiltrating monocytes.

Flow-sorted microglia and monocytes cells from Fig 5 were fixed (2% paraformaldehyde) and analyzed by an ImageStream cytometer. Using the IDEAS software package, all doublets or aggregated cells were excluded by gating on single cells. A) P2X4R and CD11b+ were focused and selected. Images of single cells (~300 cells/sample, N=3 per group) of the bright field and fluorescent channel (P2X4R-FITC and CD11b-APC-eFluor780) were acquired, and overlays were generated. Magnification=60x. B) Ratio of median fluorescence intensity (MFI) of P2X4R/CD11b was significantly reduced after 5-BDBD treatment in both microglia and monocytes. Further, a separate comparison of microglia vs. monocytes P2X4R MFI suggests reduced expression of P2X4R in microglia irrespective of treatment. Data are Mean ± SD; *p<0.05, 5-BDBD vs. Vehicle; #p<0.05, microglia vs. monocytes; Student’s t-test.

3.9. 5-BDBD reduces interleukin-1β (IL-1β) in brain tissue and plasma

Elevated levels of pro-inflammatory cytokines are established markers of a neuroinflammatory response following stroke. Short-term treatment with 5-BDBD resulted in decreased expression and release of IL-1β, a pro-inflammatory cytokine, as assessed by ELISA at 3 days after stroke in both brain tissue and plasma (Fig. 10A and B). However, we did not see changes in IL-1β levels at 30 days in either brain tissue or plasma, suggesting that short-term treatment with 5-BDBD does not cause long-lasting inhibition of IL-1β release or synthesis. Unlike IL-1β, we did not observe any differences in the levels of TNF-α in plasma or brain tissue either at 3 or 30 days after stroke between the 5-BDBD and vehicle groups (Fig. 10C and D), suggesting that the effect of 5-BDBD treatment is specific for IL-1β release and mediated by P2X4R.

Figure 10. 5-BDBD reduces inflammatory cytokine IL-1β in plasma and brain tissue of mice 3 days post-stroke.

IL-1β and TNF-α were measured in plasma (pg/ml) and brain tissue (pg/mg protein) using ELISA. IL-1β levels in plasma (A) and brain tissue (B) at 3 and 30-days post-stroke. TNF-α levels in plasma (C) and brain tissue (D) at 3- and 30-days post-stroke. Veh=vehicle. Data are Mean ± SD (n=7–10 mice/group). *p<0.05; 5-BDBD vs. Veh; Student’s t-test.

3.10. 5-BDBD treatment shows sustained improvement in sensorimotor deficits after stroke

Following ischemic stroke, both vehicle- and 5-BDBD-treated groups did not show any differences in OFT total exploratory behavior at any time point, suggesting no effect of 5-BDBD on gross locomotion activity (Fig.11A). However, we found that acute treatment with 5-BDBD caused less anxiety-like behavior (Fig. 11B) and better motor coordination and balance over 30 days (Fig. 11C). We did not see any effect of 5-BDBD treatment on behavior in the absence of stroke, i.e., sham mice treated with vehicle or 5-BDBD had no difference in any behavioral activities. These data suggest that: 1) Stroke-induced activation of P2X4R is required for the action of 5-BDBD and 2) acute inhibition of P2X4R lays the foundation for sustained improvement in functional recovery during the late recovery phase of stroke perhaps due to reduced tissue damage acutely after stroke.

Figure 11. 5-BDBD improve motor-coordination, sociability and anxiety-like behavior in post-stroke mice.

Sham surgery (SH) or stroke (ST) mice were treated with acute 5-BDBD (1 mg/kg × daily for 3 days) or vehicle and subjected to various behavioral tests. I) Repeated measure task A) Total locomotion activity over time in the open field test (OFT), was not different between the vehicle or 5-BDBD treatment group[ B) 5-BDBD treatment significantly reduced anxiety-like behavior in the mice after stroke (two-way repeated-measures ANOVA with greisser greenhouse correction, F (5, 174) = 29.62, interaction; F (5, 174) = 24.98, time; F (1, 174) = 182.1, Treatment and C) motor balance coordination (rotarod test; 4–40 rpm, 5 mins) was improved progressively after 5-BDBD treatment in mice after stroke. (Two-way repeated-measures ANOVA with greisser greenhouse correction, F (5, 174) = 14.17, p = 0.37, interaction; F (5, 174) = 22.47, p = 0.365, time; F (1, 174) = 96.16, Treatment. (Data are mean ± SD. n=3 SH mice per treatment 15–16 ST mice per treatment. Sham mice were not used in statistical analysis II) One time measured tasks such as sociability and tail suspension test (TST) D) 5-BDBD treatment does induce depression-like behavior analyzed by TST measured at Day29 (n=15–16/group) and E) 5-BDBD treatment induce sociability behavior (sociability index or percent time spent in the right chamber) measured at day 28 after stroke (Data are mean ± SD. *P<0.05 vs Vehicle). No shams were used in these tasks.

3.11. 5-BDBD treatment did not cause depression-like behavior

WT mice treated with 5-BDBD did not show depression-like behavior as assessed by a tail suspension test (TST) at 30 days (Fig. 11D), suggesting early P2X4R inhibition is critical and can provide subsequent long-term chronic recovery without predisposing mice to depression-like behavior. Unlike global P2X4R KO mice where we saw depression-like behavior and reduced BDNF levels at 30 days after stroke (Verma et al., 2017), short-term P2X4R inhibition in WT mice did not cause depression-like behavior, consistent with the concept that the restoration of BDNF during chronic recovery may have prevented depression-like behavior.

3.12. 5-BDBD treatment increases sociability in mice

Given that P2X4R is implicated in socio-communicative disturbances, we measured if acute blockade of P2X4R affects social behavior during chronic recovery after stroke. We used a sociability test to measure social approach behavior in mice after 5-BDBD or vehicle treatment. During baseline or the habituation period in both groups, we did not see any side preference (right versus left chamber) in the absence of a stranger mouse. Mice treated with 5-BDBD showed higher sociability at 30 days after stroke, compared with vehicle alone (Fig. 11E), suggesting that acute P2X4R inhibition enhanced social interaction probably by reducing anxiety-like behavior or sensorimotor deficits.

4. Discussion

In this study, we tested whether acute pharmacological blockade of P2X4R confers neuroprotection and neuro-rehabilitation after ischemic stroke. We previously showed that both global- and myeloid-specific P2X4R KO provide acute neuroprotection after ischemic stroke in mice. However, despite acute neuroprotection and swift recovery during an early phase of stroke, P2X4R KO mice exhibited depression-like behavior and worse functional recovery in the chronic phase after stroke (Verma et al., 2017). These paradoxical findings prompted us to dissect the acute and chronic effects of P2X4R activation after stroke.

During acute stroke injury, dead or damaged brain cells release supraphysiologic amounts of extracellular ATP that cause cell surface translocation (Qureshi et al., 2007) and excessive activation of P2X4R (Suurväli et al., 2017). Excessive acute activation of P2X4R leads to the release of pro-inflammatory cytokines such as IL-1β (Liang et al., 2016), which later participates in brain injury after stroke. On the other hand, the physiological activation of microglial P2X4Rs contributes to several cellular functions such as synthesis and release of BDNF and maintenance of long term potentiation (LTP), which are necessary for maintaining synaptic plasticity as well as post-stroke rehabilitation and recovery (Trang et al., 2009; Verma et al., 2016). These findings suggest that permanent absence of P2X4R, such as in KO animals, might be detrimental due to a lack of trophic support during chronic recovery. Therefore, we postulate that inhibition of P2X4R during the acute stage of stroke provides early ischemic protection after stroke by preventing acute inflammation and thus neuroprotection, which might be sufficient for both acute and chronic recovery.

To test this hypothesis, we used a specific inhibitor of P2X4R, 5-BDBD, during the acute phase after stroke and analyzed both acute and chronic outcomes. Consistent with our previous observation in P2X4R KO mice (Verma et al., 2017), 5-BDBD-treated mice (for 3 days post-stroke) showed acute neuroprotection and reduced inflammatory cytokine IL-1β synthesis or release. Intriguingly, acute blockade of P2X4R also conferred sustained chronic recovery in neurobehavioral endpoints such as improved motor coordination and less anxiety-like behavior. Further, acute blockade of P2X4R also did not result in depression-like behavior during the chronic phase after stroke, indicating another benefit of acute P2X4R blockade, i.e., preservation of a salutary effect of P2X4R during the chronic recovery phase.

To understand these beneficial effects of acute P2X4R blockade, we first determined the location and expression intensity of this receptor. We had previously shown that following an episode of acute stroke, P2X4R activity and expression is increased in myeloid cells including both infiltrating monocytes/macrophages and brain-resident microglia (Verma et al., 2017). Infiltrating leukocytes such as myeloid cells (including neutrophils and monocytes) exacerbate ischemic injury after stroke (Kim et al., 2016) and inhibition of leukocyte infiltration reduces stroke injury (Schilling et al., 2009). It is well established that diminished acute activation of both peripheral (monocytes and neutrophils) or central myeloid cells (microglia) reduce stroke injury (Dimitrijevic et al., 2007). 5-BDBD treatment decreased the infiltration of peripheral myeloid cells, which contributed to recovery by reducing P2X4R activation in both infiltrated monocytes/macrophages. Reduced activation of resident microglia specifically in the perilesional cortex after 5-BDBD treatment suggests that acute inhibition of microglial P2X4R activation is also beneficial for stroke recovery, which is consistent with our previous observation in P2X4R KO mice (Verma et al., 2017).

This striking effect of 5-BDBD may be due to direct inhibition of P2X4R on monocytes/macrophages, leading to their reduced activation and migration to the brain, indirectly due to a decrease in BBB permeability, or both. During stroke, chemokine CCL2 has been shown to regulate leukocyte infiltration into the brain and its absence reduces infarct area after stroke (Schilling et al., 2009). P2X4R is a downstream target of CCL2, where CCL2 helps translocate P2X4R to the cell surface to participate in ischemic injury (Toyomitsu et al., 2012). It is conceivable that inhibition of extracellular P2X4R would reduce CCL2 activity via a feedback mechanism, which might reduce leukocyte infiltration independent of reduced BBB permeability. Our imaging data on flow-sorted monocytes and microglia, where we saw reduced cell-surface expression of P2X4R after 5-BDBD treatment, further strengthens our observation that reduced cell-surface expression of P2X4R decreases myeloid cell activation. Our EBD extravasation data suggest that 5-BDBD treatment reduced BBB permeability. These observations were supplemented by RNA-Seq data of 5-BDBD-treated mice that showed reduced expression of matrix metallopeptidase 9 (Mmp9), a zinc-metalloproteinase family member that degrades extracellular matrix and promotes influx of infiltrating leukocytes (Dong et al., 2009; Ramos-Fernandez et al., 2011; Xu et al., 2015; Yuan et al., 2013). Similarly, ZFP708 and ZFP 260 are the member of zinc finger proteins. These proteins are known to regulate MMP9 activity (Khrenova et al., 2014) to reduce BBB permeability. Hypoxic conditions such as stroke are known to induce the extracellular matrix protein lysyl oxidase (LOX) which further promotes MMP2/9 expression (Liu et al., 2014). Reduction of Lox in parallel to Mmp9 further supports our data that 5-BDBD reduces ischemic injury by reducing BBB permeability.

While P2X4R over-activation can induce acute inflammation, its stimulation at physiologic levels is required for the synthesis and release of BDNF (Trang et al., 2009), especially during the chronic neuro-rehabilitory phase. 5-BDBD treatment did not change mBDNF protein expression acutely during ischemic injury. On the other hand, 5-BDBD treatment increased BDNF protein expression in the longer term during chronic recovery. This may be due to the rescue of penumbra tissue by the anti-inflammatory effect from acute P2X4R inhibition. Finally, the continued presence of endogenous P2X4R in WT animals permits its salutary function during the chronic phase, a phenomenon that is absent in global P2X4R KO mice.

Here we also acknowledge a potential limitation of our work as we did not directly determine the level of 5-BDBD in brain tissue. However, our EBD extravasation data and other observations, such as reduced P2X4R expression on IBA1 positive microglia and monocytes near perilesional tissue, reduced cell-surface expression of P2X4R on flow-sorted microglia, and reduced IL-1β levels in brain tissue, suggest that 5-BDBD likely has crossed the BBB to produce its effects on P2X4R. The data provided functional indirect evidence for an effect on the resident brain microglia by 5-BDBD that had crossed into the brain. Overall, our data suggest that 5-BDBD blocked P2X4R activity on peripheral myeloid cells thus reducing their activation and infiltration, resulting in decreased ischemic injury. 5-BDBD likely also reduced activation of brain-resident microglia after stroke. Overall, 5-BDBD treatment during the acute phase of stroke showed considerable neuroprotective and neuro-rehabilitory effects. Acute pharmacological blockade of P2X4R during ischemic stroke represents a new approach and warrants further investigation.

5. Summary and Conclusion

Acute P2X4R inhibition by 5-BDBD protects against ischemic injury at both acute and chronic time points after stroke. Reduced numbers of infiltrating proinflammatory myeloid cells, decreased amounts of pro-inflammatory cytokines, and a better preserved BBB as well as a possible P2X4R blockade of the brain microglia by 5-BDBD contribute to the mechanism of neuroprotection after stroke.

Highlights:

1. Acute blockade of P2X4R provide neuroprotection and neuro-rehabilitation.

2. 5-BDBD reduce infiltration of peripheral immune cells after stroke.

3. 5-BDBD treatment diminish BBB permeability and myeloid cell activation.