Pharmacological Inhibition of Receptor-Interacting Protein Kinase 2 (RIPK2) Elicits Neuroprotective Effects Following Experimental Ischemic Stroke

Abstract

Ischemic stroke induces a debilitating neurological insult, where inflammatory processes contribute greatly to the expansion and growth of the injury. Receptor-interacting protein kinase 2 (RIPK2) is most well-known for its role as the obligate kinase for NOD1/2 pattern recognition receptor signaling and is implicated in the pathology of various inflammatory conditions. Compared to a sham-operated control, ischemic stroke resulted in a dramatic increase in the active, phosphorylated form of RIPK2, indicating that RIPK2 may be implicated in the response to stroke injury. Here, we assessed the effects of pharmacological inhibition of RIPK2 to improve post-stroke outcomes in mice subjected to experimental ischemic stroke. We found that treatment at the onset of reperfusion with a RIPK2 inhibitor, which inhibits the phosphorylation and activation of RIPK2, resulted in marked improvements in post-stroke behavioral outcomes compared to the vehicle-administered group assessed 24 h after stroke. RIPK2 inhibitor-treated mice exhibited dramatic reductions in infarct volume, concurrent with reduced damage to the blood-brain barrier, as evidenced by reduced levels of active matrix metalloproteinase-9 (MMP-9) and leakage of blood-borne albumin in the ipsilateral cortex. To explore the protective mechanism of RIPK2 inhibition, we next pretreated mice with RIPK2 inhibitor or vehicle and examined transcriptomic alterations occurring in the ischemic brain 6 h after stroke. We observed a dramatic reduction in neuroinflammatory markers in the ipsilateral cortex of the inhibitor-treated group while also attaining a comprehensive view of the vast transcriptomic alterations occurring in the brain with inhibitor treatment through bulk RNA-sequencing of the injured cortex. Overall, we provide significant novel evidence that RIPK2 may represent a viable target for post-stroke pharmacotherapy and potentially other neuroinflammatory conditions.

Introduction

Ischemic stroke is a leading cause of death and disability around the world, with an estimated annual cost of $721 billion, as of 2022 (Feigin et al., 2022). An ischemic stroke occurs when blood flow to the brain is restricted due to the blockage of a cerebral artery. This results in the rapid death of neurons and other brain cells downstream of the cerebral artery as these cells succumb to the hypoxic and energy-deprived conditions. This, in turn, promotes a highly neuroinflammatory state within the brain parenchyma as glial cells, like microglia, respond to the damage caused by stroke injury by secreting cytokines, chemokines, and blood-brain barrier (BBB)-compromising matrix metalloproteinases (MMPs) (Candelario-Jalil et al., 2022). Peripheral immune cells are then drawn to the site of injury via chemotactic signals, aided by the breakdown of the BBB, where they further contribute to the neuroinflammatory state by releasing additional inflammatory factors (Iadecola et al., 2020). In this way, neuroinflammation contributes greatly to secondary brain injury after stroke. Modulation of neuroinflammatory pathways represents an attractive target to lessen the deleterious progression of stroke injury.

Various cell types utilize pattern recognition receptors (PRRs) to respond to molecular patterns and facilitate a rapid response to potentially dangerous stimuli (Li and Wu, 2021). In the context of stroke, PRRs like toll-like receptors (TLRs) and nucleotide oligomerization domain (NOD)-like receptors (NLRs) are used by immune cells and neurons to respond to damage-associated molecular patterns (DAMPs) released from dead/dying cells and promote NF-κB and MAPK pro-inflammatory pathway activation (Kumar, 2019; Santoni et al., 2015). Overactivation of these pathways promotes a highly neuroinflammatory environment with the consequence of direct damage to otherwise healthy neurons.

Receptor-interacting protein kinase 2 (RIPK2) is well characterized for being the obligate kinase for NOD1 and NOD2 NLR signaling. NOD1 and NOD2 are PRRs that are traditionally known for recognizing pathogen-associated molecular patterns (PAMPs) of bacterial origin. However, more recent evidence suggests that they may have a lesser-defined role in the recognition of DAMPs (Pei et al., 2021). Upon NOD1/NOD2 activation, RIPK2 autophosphorylates, initiating a signaling cascade that results in both NF-κB and MAPK pro-inflammatory pathway activation. Besides their role in the propagation of inflammatory signaling, NOD1/2 and RIPK2 have been shown to be involved in the cascade of mitochondrial endoplasmic reticulum (ER) stress (Keestra-Gounder et al., 2016), and overactivation of ER stress can shunt cells toward a more pro-apoptotic pathway (Han et al., 2021).

In addition to its association with and propagation of NOD1/2 signaling, RIPK2 has alternative roles that implicate it in the progression of stroke injury. For instance, RIPK2 was shown to associate with caspase-1 and contribute to the direct cell death of neurons under hypoxic conditions (Zhang et al., 2003). RIPK2 has also been implicated in the autophagy pathway, and its inhibition was shown to decrease the phosphorylation of the autophagy-initiating protein ULK1 in a mouse meningitis model (Gao et al., 2019; Wang et al., 2022). Excessive activation of the autophagy pathway in the brain after stroke is associated with worse outcomes in animal models of the disease (Mo et al., 2020; Shi et al., 2021). Previous work from our group found that global Ripk2 knockout mice and mice where Ripk2 is specifically deleted from microglia experienced dramatically reduced post-stroke deficits compared to their respective control animals (Larochelle et al., 2023).

By inhibiting RIPK2’s autophosphorylation event, RIPK2’s signaling activity can be ablated (Tigno-Aranjuez et al., 2010). Inhibition of RIPK2 has proven effective in a variety of inflammatory disease models; in particular, inflammatory bowel disease (Hollenbach et al., 2004), Crohn’s disease and ulcerative colitis (Haile et al., 2016; Hollenbach et al., 2005), multiple sclerosis (Nachbur et al., 2015), arthritis (Rosenzweig et al., 2011), and intracerebral hemorrhage (Wang et al., 2020). Multiple inhibitors for RIPK2 have been developed, and currently, there is a focused effort to create novel inhibitors for RIPK2 with enhanced potency and selectivity, as well as other molecules that block RIPK2 activity, such as the development of RIPK2-specific proteolysis-targeting chimeras (PROTACs) (Bondeson et al., 2015; Mares et al., 2020; Miah et al., 2021). Herein, we utilize an inhibitor of RIPK2 that was proven to be highly efficacious at doses ranging from 1-15 mg/kg of body weight in mice, while also avoiding typical off-target effects that were common in previous iterations of RIPK2 inhibitors (Salla et al., 2018).

In our present study, we subjected adult male mice to 45 minutes of transient middle cerebral artery occlusion (tMCAO) and administered doses of either RIPK2 inhibitor or vehicle control upon reperfusion. We hypothesized that inhibition of RIPK2 will improve animal outcomes after ischemic stroke primarily by dampening neuroinflammatory processes, thus preserving functional brain tissue from the sequelae of stroke injury. We assessed the effects of RIPK2 inhibition on infarct volume and post-stroke behavioral deficits, dampening neuroinflammation, and preservation of the BBB in the acute phase of stroke injury. Finally, to delve into molecular mechanisms underlying the effects of RIPK2 inhibition in stroke, we performed bulk RNAseq in the ischemic cerebral cortex of vehicle-administered and inhibitor-pretreated animals to investigate how RIPK2 blockade impacts stroke-induced transcriptional programs during the early stages of ischemic brain injury.

Methods

Animals

Three-to-six-month-old male C57BL/6J male mice (Jackson Laboratory) were used for this study. Mice were housed in a specific pathogen-free facility with a 12-hour light/12-hour dark cycle and ad libitum access to food and water. All animal experiment procedures were conducted following the NIH Guide for the Care and Use of Laboratory Animals. All procedures were approved by the University of Florida Institutional Animal Care and Use Committee (animal protocol numbers 201907934 and 202200000201).

Power analysis, sample size calculation, blinding, and randomization

The number of animals for each experiment was calculated based on an a priori power analysis using the G*Power v.3.1.9.7 software. To estimate the required sample size, we utilized variances from our preliminary studies in the transient MCAO model in male mice to calculate the Cohen effect size (d) by comparing two independent groups using α=0.05 and β (type II error) =0.1 with a power of 90%. This study was powered with the expectation that we would detect a difference in infarct size of at least 25% between our vehicle- and inhibitor-treated animals, which is a biologically meaningful result (Schlattmann and Dirnagl, 2010). We calculated that a sample size of n=10 with an effect size of d= 2.102 would likely yield a statistically significant result. After adjusting for a potential 15% attrition rate, 11 mice per group were needed for the experiments described in this study. We calculated a sample size of n=5 for the molecular biology analyses with an effect size of d=2.476. All mice were coded and randomly allocated to experimental groups using the randomization tool developed by GraphPad Prism (http://www.graphpad.com/quickcalcs/randomize1.cfm). All experiments and analyses were performed by investigators blinded to animal treatment groups. The specific number of animals for each analysis is stated in the figure legends.

Induction of transient middle cerebral artery occlusion

Adult mice were subjected to transient focal cerebral ischemia induced by intraluminal occlusion of the right middle cerebral artery (MCA) for 45 min, as previously described by our group (Yang et al., 2020). Briefly, mice were anesthetized with 3% isoflurane, and surgical levels of anesthesia were maintained by inhalation of 1.5-2% isoflurane in medical-grade oxygen. Body temperature was maintained at 37°C during surgery using a heating pad and temperature regulator with a rectal probe. The right common carotid artery (CCA), external carotid artery (ECA), and internal carotid artery (ICA) were exposed via a midline vertical incision in the anterior neck. A 12-mm-long 6-0 silicone-coated nylon filament (Doccol, Cat#602123) was advanced gently into the ICA approximately 9-10 mm from the carotid bifurcation until mild resistance was felt, and cerebral blood flow (CBF) was reduced by at least 75% of the baseline value, as assessed by laser Doppler flowmetry. After 45 min of MCA occlusion, the filament was gently retracted to allow reperfusion (confirmed by laser Doppler flowmetry). The skin was closed, anesthesia was discontinued, and the mice were allowed to recover in a temperature-controlled chamber. Sham-operated animals received the same surgical procedures except for the MCA occlusion. Animals showing complete loss of spontaneous activity for a prolonged time (>2 h), subarachnoid hemorrhage upon tissue harvesting, or lack of neurological deficits after 45 min of stroke were excluded from the analysis.

Preparation and Administration of Pharmacological Reagents

Dilutions of the potent RIPK2 inhibitor 3-benzamido-4-methyl-N-[3-(1-methyl-1H-imidazol-2yl)phenyl] benzamide (Salla et al., 2018) (referred to throughout the paper simply as RIPK2 inhibitor) were prepared by first dissolving the drug (Sigma Aldrich, Cat. #: SML2554) in dimethyl sulfoxide (DMSO) to create a stock solution at 7.5 mg/mL. This stock solution was further dissolved into a final solution containing 2% of the DMSO stock in 10% Captisol (Cat#NC1238671, Carbosynth LLC). Animals were then administered a single dose of 3 mg/kg of body weight of RIPK2 inhibitor or vehicle control via i.p. injection either immediately upon reperfusion or 15 min prior to the start of MCA occlusion. Animals received the vehicle or inhibitor at 20 mL/kg, corresponding to a total volume of solution ranging from 500-700μL. Since DMSO is at 2% in the final solution, mice received 0.4 mL/kg of DMSO, corresponding to a total DMSO volume ranging from 10-14μL depending on the animal’s body weight.

Infarct Volume Quantification

2,3,5-triphenyltetrazolium chloride (TTC) staining was used to measure brain infarction at 24 h after stroke, as previously described (Yang et al., 2021). Mice were euthanized following tMCAO and were perfused transcardially with ice-cold saline. Brains were harvested, placed in a slicing matrix (Zivic Instruments, Pittsburgh, PA), and sliced into six 1-mm-thick coronal sections, approximately corresponding to bregma level 2 mm (slice 1) and ending at level -3 mm (slice 6). The fourth section starting from the rostral side was dissected into ipsilateral and contralateral cerebral cortices and subcortices. These tissues were immediately frozen in liquid nitrogen for storage at -80°C for later processing. The remaining sections were stained with 2% TTC in phosphate-buffered saline (PBS, pH 7.4) at room temperature for 30 min, then were fixed with 4% paraformaldehyde (PFA) in PBS, pH 7.4. The stained sections were laid rostral-side down and scanned at 600 dpi using an HP Scanjet 8300 scanner (Palo Alto, CA) and saved as a JPEG file. The caudal side of the 3^rd section was scanned and served as the representative image of the 4^th section, as it corresponds to the rostral side of the 4^th slice. The infarction volume was calculated by integrating the lesion areas of all six brain slices.

Protein extraction from brain for Western blotting and ELISA

At 24 h after tMCAO, brains were dissected after transcardiac perfusion with ice-cold saline. The ipsilateral and contralateral cerebral cortex and subcortex of the brain were collected and immediately frozen in liquid nitrogen and saved at -80° for further processing. Tissues were homogenized in radioimmunoprecipitation (RIPA) lysis buffer consisting of 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5 mM EDTA, 1 mM EGTA, 1% NP-40, 0.5% sodium deoxycholate and 0.1% SDS plus protease and phosphatase inhibitor cocktails (Cat#78430 and Cat#78428, respectively, Thermo Fischer Scientific; Rockford, IL), and total protein concentration was determined using the Pierce^™ BCA assay kit (Cat#23277, Thermo Scientific, Rockford, IL). An equal amount of protein was separated using 4-20% polyacrylamide gradient gels (BioRad, Hercules, CA) and transferred to nitrocellulose membranes. Membranes were blocked with 5% milk and were incubated overnight at 4°C with rabbit anti-RIPK2 (1:1000; Cat#4142S, Cell Signaling Technology, Danvers. MA or 1:500: Cat#GTX28428, GeneTex, Irvine, CA), rabbit antibody recognizing phosphorylated RIPK2 at residue Thr474 (1:500; Cat#19510, QED Bioscience, San Diego, CA), rat anti-β-Actin (1:5000; Cat#664802, BioLegend; San Diego, CA), rabbit anti-MMP-9 (1:500; Cat#sc-6841-R, Santa Cruz Biotechnology; Dallas, TX). After primary antibody incubation, the membranes were washed 3 times with TBST, and incubated with the following secondary antibodies for 1h at room temperature: goat anti-rabbit IRDye 800CW (1:30000; Li-Cor; Lincoln, NE), goat anti-rat IRDye 680LT (1:40000; Li-Cor; Lincoln, NE). Membranes were visualized, and densitometric analysis was performed using the Odyssey infrared scanner and Image Studio 2.0 software (Li-Cor). Unedited immunoblots are presented in the Supplemental Material data file.

Brain albumin levels were measured using a commercially available ELISA kit (Cat#E90-134, Bethyl Laboratories Inc, Montgomery, TX) according to the manufacturer’s instructions. A total of 3 μg of protein from the ipsilateral and contralateral cortices of the mouse brain extracted as stated above was used. All samples were assayed in duplicate and optical absorbance was measured at 450 nm with a Synergy^™ HT Multi-Mode Plate Reader (BioTek Instruments; Winooski, VT).

RNA extraction from brain and quantitative real-time PCR

Total RNA from 50 μL of the total cortical tissue homogenate was isolated using a modified acid guanidinium thiocyanate-phenol-chloroform extraction method (Chomczynski and Sacchi, 2006; Yang et al., 2021). RNA concentration and purity were determined by a Take3 Micro-Volume Plate Reader (BioTek Instruments; Winooski, VT). Quantitative real-time PCR was performed in a total reaction volume of 10 μL using Luna^® Universal One-Step RT-qPCR Kit (Cat#E3005, New England BioLabs; Ipswich, MA) following the manufacturer’s protocol. Reactions were performed in 96-well plates run in a BioRad CFX96 Touch Real-Time instrument. Each reaction was performed in duplicate or triplicate, and the relative expression value for each target gene was calculated using the ^ΔΔCt method after normalization to the housekeeping gene Ywhaz. The primer sequences used in the reactions can be found in Table 1.

Table 1.

Primer sequences for real-time PCR analysis

Gene	Primer Forward Sequence	Primer Reverse Sequence	Accession Number
Il1β	GACCTGTTCTTTGAAGTTGACG	CTCTTGTTGATGTGCTGCTG	NM_008361.4
Il6	AGCCAGAGTCCTTCAGAGA	TCCTTAGCCACTCCTTCTGT	NM_031168.2
Cxcl1	CCAAACCGAAGTCATAGCCA	GTGCCATCAGAGCAGTCT	NM_008176.3
Ccl2	CATCCACGTGTTGGCTCA	AACTACAGCTTCTTTGGGACA	NM_011333.3
Ywhaz	TGTTCTAGCCTGTTTCCCCG	ACGATGACGTCAAACGCTTC	NM_001356569.1
Lcn2	ATGTCACCTCCATCCTGGTCAG	GCCACTTGCACATTGTAGCTCTG	NM_008491.1
Icam1	AAACCAGACCCTGGAACTGCAC	GCCTGGCATTTCAGAGTCTGCT	NM_010493.3

Gene abbreviations: Il1β, interleukin 1 beta; Il6, interleukin 6; Cxcl1, C-X-C motif chemokine ligand 1; Ccl2, chemokine C-C motif ligand 2; Ywhaz, tyrosine 3-monooxygenase/tryptophan 5-monoogygenase activation protein zeta; Lcn2, Lipocalin 2; Icam1, intercellular adhesion molecule 1.

Bulk RNA-sequencing and Analysis

Poly(A) RNA sequencing library was prepared following Illumina’s TruSeq-stranded-mRNA sample preparation protocol (Illumina, Inc, San Diego, CA). RNA integrity was checked with Agilent Technologies 2100 Bioanalyzer (Agilent Technologies, Inc, Santa Clara, CA). Poly(A) tail-containing mRNAs were purified using oligo-(dT) magnetic beads with two rounds of purification. After purification, poly(A) RNA was fragmented using a divalent cation buffer at elevated temperatures. The DNA library was constructed. Quality control analysis and quantification of the sequencing library were performed using Agilent Technologies 2100 Bioanalyzer High Sensitivity DNA Chip (Cat# 5067-4626). Paired-ended sequencing was performed on Illumina’s NovaSeq 6000 sequencing system.

Transcripts Assembly

Firstly, Cutadapt (Martin, 2011) and Perl scripts made in house were used to remove the reads that contained adaptor contamination, low quality bases, and undetermined bases. Then, sequence quality was verified using FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). We used HISAT2 (Kim et al., 2015) to map reads to the genome of ftp://ftp.ensembl.org/pub/release-101/fasta/mus_musculus/dna/. The mapped reads of each sample were assembled using StringTie (Pertea et al., 2015). Then, all transcriptomes were merged to reconstruct a comprehensive transcriptome using Perl scripts and gffcompare. After the final transcriptome was generated, StringTie and ballgown (http://www.bioconductor.org/packages/release/bioc/html/ballgown.html) were used to estimate the expression levels of all transcripts.

Different expression analysis of mRNAs

StringTie (Pertea et al., 2015) was used to perform expression levels for mRNAs by calculating Fragments per Kilobase of transcript per Million (FPKM). mRNAs differential expression analysis was performed by R package DESeq2 (Love et al., 2014) between two different groups (and by R package edgeR (Robinson et al., 2010) between two samples). The mRNAs with the false discovery rate (FDR) parameter below 0.05 and absolute fold change ≥ 2 were considered differentially expressed mRNAs.

Behavioral Tests

Animals were subjected to several behavioral tests by investigators blinded to treatment groups. Tests performed include neurological deficit score, open field locomotor activity test, weight grip test, and vertical grid to assess the neurological performance of mice at various time points after stroke.

Neurological Deficit Scores

The neurological deficits scoring assessment evaluates overall neurological abnormalities along with multiple physical deficits in animal studies of stroke (Bieber et al., 2019; Schaar et al., 2010). At 24 h after stroke, a neurological deficit score (NDS) was determined for each animal according to six different parameters: body symmetry, gait, circling behavior, front limb symmetry, compulsory circling, and climbing, as previously described (Glushakov et al., 2013; Liu et al., 2019a). Each test was scored independently by two trained investigators using a 4-point scoring system (0, no deficits; 4 severe deficits). The average score of each mouse from two investigators was used for statistical analysis by nonparametric tests.

Open field

The open field test is a reliable behavioral test to assess locomotor behaviors in mice. Within one week prior to stroke (baseline) and at 24 h post-stroke, the spontaneous locomotor activity of mice was measured in an open field paradigm using automated video tracking software (Anymaze software; Stoelting, Wood Dale, IL) as previously described (DeMars et al., 2019). Mice were individually placed in an open field chamber (40 × 40 × 40 cm) with grey sidewalls and were allowed to freely explore for 10 min. The total distance traveled was used as the indices of motor/exploratory behavior of each animal. The open field arena was thoroughly cleaned with 70% ethanol between tests.

Vertical grid test

The vertical grid test is a sensitive test intended to assess animals' neuromuscular strength and motor coordination after stroke (Kim et al., 2010; Tillerson et al., 2002). The vertical grid is an open frame apparatus (55 cm H × 8 cm W ×5 cm L) with a wire mesh (0.8 cm × 0.8 cm aperture) on the backside. The grid is placed in a cage filled with soft bedding material. Within one week prior to surgery and at 24 h post-tMCAO induction, each mouse was placed to the highest point on the grid facing downward and was allowed to descend the grid into the cage. A blinded investigator recorded the time required for the animal to descend. Animals were subjected to 3 trials with intervals of 30 s between each trial. The average of the three trials constitutes the animal’s score on the test. Animals that failed to descend the grid within 60 s or were unable to maintain a firm grip of the grid and fell were assigned the maximum score of 60 s for that trial.

Weight grip test

The weight grip test was performed with minor modifications from a previous study (Deacon, 2013), as we have previously reported (Yang et al., 2023), to assess the muscular strength of the forepaws. Five different weights (weight-1: 16.2 g, weight-2: 30.4 g, weight-3: 44.6 g, weight-4: 58.2 g, and weight-5: 71.4 g) were prepared by attaching a metal mesh to stainless steel lines. The animals were suspended from the middle/base of the tail and allowed to grasp the first weight (weight-1). A timer starts when the mouse successfully grips the weight using its forepaws, then the animal is lifted until the steel links are completely lifted from the bench. The mouse must hold the weight for 3 s for the test to be successful. If the mouse was able to hold weight-1 for 3 seconds, the investigator proceeded to the next weight in sequential order. If the mouse were to drop a weight in less than 3 seconds, the test would conclude, and no further weights would be attempted. The mice were permitted three tries to successfully hold the weight for 3 seconds. A final score was tallied as a sum of the point of each weight that the mouse holds, multiplied by the number of seconds that weight was held. For example, a mouse that held up to weight-5 for 1 second is assigned a score of 1×3 + 2×3 + 3×3 + 4×3 + 5×1 = 35. This test was conducted 1 week prior to surgery (baseline) and at 24 h post-stroke.

Statistical analysis

Prism software (GraphPad v.8) was used to perform the statistical analyses. An independent unpaired Student’s t-test (parametric) or Mann-Whitney test (nonparametric) was performed to compare two groups. Two-way ANOVA followed by Šídák’s post hoc test was used for multiple comparisons. Values were expressed as mean ± SEM, and a p-value of less than 0.05 was considered statistically significant. The number of animals per group is clearly stated in the figure legends. Animals were assigned nondescript coded IDs upon enrollment in the study. All studies were performed by investigators blinded to the treatment of the animals in analyses.

Results

RIPK2 Inhibitor Reduces Stroke-induced RIPK2 Phosphorylation

We first sought to determine whether the active, phosphorylated form of RIPK2 can be detected in the ischemic brain. We compared the levels of phosphorylated-RIPK2 (p-RIPK2) at residue Thr474 in the ipsilateral cortex (CXI) of mice subjected mice to 45 min of tMCAO or sham surgery and found that stroke elicits a dramatic increase in the levels of p-RIPK2 in the brain 24 h after injury (Fig.1 A-B). We found no detectable difference between sham and stroke in the total levels of RIPK2 protein in the CXI. However, there was a significant difference in the ratio of p-RIPK2 to RIPK2 between groups (Fig.1 C-D). Thus, for the first time, we demonstrated that p-RIPK2 increases in the ischemic brain, which led us to hypothesize that this phosphorylation event can be blocked by utilizing a kinase inhibitor selective for RIPK2.

Fig.1. RIPK2 inhibitor dramatically reduces stroke-induced phosphorylated-RIPK2 (p-RIPK2) in the brain.

A Western blot for p-RIPK2 and RIPK2 in the ipsilateral cortex (CXI) of sham and stroke mice after 24 h of reperfusion. B Quantification of p-RIPK2 depicted in (A). C Quantification of RIPK2 depicted in (A). D Quantification of the ratio between p-RIPK2 and RIPK2 detected in the blot in panel (A). E Chemical structure of the RIPK2 Inhibitor used throughout this study. F Western blot for p-RIPK2 and RIPK2 in the cerebral cortex 24 h after ischemic stroke. G Quantification of p-RIPK2 depicted in (F). H Quantification of RIPK2 depicted in (F) showing no difference between levels of unphosphorylated RIPK2 between vehicle- and inhibitor-treated animals after stroke. I Quantification of the ratio between p-RIPK2 and RIPK2 detected in the WB in (F). n=5-6/group. Differences determined by Student’s t test. * P<0.05, ** P<0.01, **** P<0.0001.

To test whether the RIPK2 inhibitor was inhibiting RIPK2’s autophosphorylation, we subjected mice to 45 min of tMCAO and administered 3 mg/kg of RIPK2 inhibitor or vehicle upon reperfusion. We found that treatment with the inhibitor, whose chemical structure is depicted in Fig.1 E, dramatically reduced the levels of p-RIPK2 in the brain after stroke compared to vehicle-administered animals (Fig.1 F-G). Brain lysate from mice deficient for the Ripk2 allele were used as a negative control for the presence of both phosphorylated- and unposphorylated-RIPK2 (Fig.1 F). Importantly, we show that RIPK2 inhibition does not reduce the total levels of RIPK2 between inhibitor- and vehicle-administered animals (Fig.1 H); however, we do show that there is a clear difference in the ratio of p-RIPK2 to RIPK2, indicating that the inhibitor is, in fact, reducing the phosphorylated form of RIPK2 (Fig.1 I). This data provides evidence that the RIPK2 inhibitor reduces the active, phosphorylated form of RIPK2 in vivo in ischemic stroke, demonstrating target engagement at the dose of 3 mg/kg delivered intraperitoneally.

Improved Behavioral Outcomes in Acute Stroke Injury with RIPK2 Inhibition

Mice were subjected to 45 min of tMCAO then administered doses of 3mg/kg of RIPK2 inhibitor or vehicle at the onset of reperfusion. To attain a comprehensive understanding of the potential neuroprotective effect of RIPK2 inhibition, we subjected vehicle- and inhibitor-treated mice to a battery of behavioral tests 24 h after the onset of reperfusion following experimental ischemic stroke. Mice were subjected to the open field test, vertical grid test, weight grip test, and neurological deficit score assessment. We observed a powerful effect on the preservation of motor and neurological function following stroke in mice treated with the RIPK2 inhibitor across our different behavioral paradigms. RIPK2 inhibitor-treated mice traveled greater distances in the open field chamber compared to vehicle controls (Fig.2 B-C), with clear separation in their distances traveled during each individual minute of the ten-minute open field test (Fig.2 D). Inhibitor-treated animals displayed a significant preservation in both their ability to descend a vertical grid (Fig.2 E) and their ability to grip weights of increasing mass (Fig.2 F) 24 h after ischemic stroke. Finally, we observed clear differences in the neurological functioning of these animals by NDS scoring, where inhibitor-treated animals scored significantly lower in both the total NDS score (Fig.2 G) as well as scoring lower on each of the six individual parameters that comprise the assessment (Fig.2 H). Upon examination of this data in its totality, it is clear that inhibition of RIPK2 exerts profound neuroprotective effects that are reflected in improved behavioral outcomes during the acute phase of stroke injury.

Fig.2: Inhibition of RIPK2 upon reperfusion elicits profound protective effects on neurobehavioral outcomes.

A Schematic of experimental design. B Representative open field track plots and heat maps comparing vehicle- and RIPK2 inhibitor-treated mice at baseline and 24 h after ischemic stroke. C Quantification of open field data depicted in (A). D Distance traveled during each individual minute of the open field test. E Time required for animals to descend the vertical grid. F Weight grip scores for both groups obtained from the weight grip test. G Total NDS scores of treatment groups 24 h after stroke. H Scores of the six individual parameters of the NDS assessment. n=11/group. (B-D) Differences determined by two-way ANOVA, Šídák’s multiple comparisons test. (E-G) Differences determined by Mann-Whitney test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001

Reduced Infarct Volume 24 h after stroke with RIPK2 Inhibition

We next wanted to assess the neuroprotective effect of pharmacological inhibition of RIPK2 by measuring infarct volume 24 h after the onset of stroke. Mice were sacrificed 24 h post-stroke and their brains were sectioned and stained with TTC to observe the effect on acute infarction. We found that mice RIPK2 inhibitor-treated mice had profoundly decreased infarct volumes compared to vehicle-administered control animals as determined by TTC staining (Fig.3). TTC staining showed a consistent reduction in infarction (Fig.3 A), with profound effects observed at the levels of the cortex and total hemisphere. In contrast, infarct at the level of the subcortex showed no significant difference (Fig.3 B). By measuring the infarct volume at each 1 mm-thick slice, we show that infarct is reduced across all levels of the cortex and total hemisphere, while there are trends for decreasing infarct at various slice numbers in the subcortex (Fig.3 C). Taken together, this data reveals a powerful neuroprotective effect of RIPK2 inhibition in reducing infarct size during the acute stage of stroke injury.

Fig.3: Inhibition of RIPK2 dramatically reduces infarct volume 24 h after ischemic stroke.

A Representative TTC staining of vehicle- and inhibitor-treated animals 24 h after experimental ischemic stroke. B Quantification of infarct volumes at the levels of the subcortex, cortex, and total hemisphere. C Quantification of the infarct area at the level of each individual slice at the level of the subcortex, cortex, and total hemisphere. n=11/group. (B) Differences determined by Student’s t test. (C) Differences determined by two-way ANOVA, Šídák’s multiple comparisons test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001

Reduced Quantity of BBB-compromising MMP-9 and Albumin Leakage

Since we determined the neuroprotective effect of RIPK2 inhibition in the acute phase of stroke injury, we wanted to further probe the mechanism by which this protective effect is mediated. To do so, we examined the levels of MMP-9 in the ipsilateral cortex of vehicle- and inhibitor-treated mice 24 h after stroke and examined the leakage of albumin into the ipsilateral cortex as a way of measuring damage to the BBB. We found that levels of active MMP-9 were dramatically lower in the inhibitor-treated mice relative to the vehicle-administered controls (Fig.4 A-B). Using an ELISA to detect albumin, we observed a dramatic reduction in extravasated albumin levels in the ipsilateral cortex of mice treated with the RIPK2 inhibitor relative to the vehicle group (Fig.4 C). This data indicates that RIPK2 inhibition reduces the levels of the BBB-damaging enzyme MMP-9 and decreases cortical levels of albumin, both of which are indicative of greater preservation of the BBB in the inhibitor-treated mice compared to the vehicle controls.

Fig.4: Inhibition of RIPK2 reduces active MMP-9 in the ipsilateral cortex 24 h after ischemic stroke.

A Western blot of the ipsilateral cortex of vehicle- and inhibitor-treated mice 24 h after stroke depicting the levels of active- and inactive-MMP-9. B Quantification of WB shown in (A). C ELISA for the detection of albumin in the ipsilateral (CXI) and contralateral (CXC) cortices 24 h after stroke. (A-B) n=5-6/group. Differences determined by Student’s t test. (C) n=10/group. Differences determined by two-way ANOVA, Šídák’s multiple comparisons test. *P<0.05, **P<0.01

RIPK2 Inhibition Reduces Stroke-induced Inflammatory Gene Transcription

To further explore the mechanism by which the RIPK2 inhibitor exerts its observed protective effects in stroke injury, we performed RT-qPCR on the ipsilateral cortex of mice treated with either our vehicle or RIPK2 inhibitor to assess the expression of various neuroinflammation-related genes (Fig.5). In this experiment, mice were pretreated with either RIPK2 inhibitor or vehicle 15 min prior to stroke onset and sacrificed after 6 h. We found that RIPK2 inhibition dramatically reduced markers of neuroinflammation, such as Il1β¸ Cxcl1, Ccl2, Icam1, and Lcn2 compared to the vehicle controls, and Il6 showed a strong trend toward being significantly decreased (Fig.5 B). This data provides mechanistic evidence for the manner in which RIPK2 inhibition exerts a protective effect in stroke injury; namely, by reducing neuroinflammation in the brain during the acute stage of injury immediately following stroke induction.

Fig.5: Pre-treatment with RIPK2 inhibitor shows significant reduction in inflammatory gene transcription 6 h post-stroke.

A Schematic of experimental design. B RT-qPCR for inflammatory markers Il1β, Il6, Ccl2, Cxcl1, Lcn2, and Icam1 in the ipsilateral (CXI) and contralateral (CXC) cortices. Significant differences were observed with expression of Il1β, Cxcl1, Icam1, Ccl2, and Lcn2, while ll6 showed a strong trend toward significance. n=7-10/group. Differences determined by two-way ANOVA, Šídák’s multiple comparisons test. P values displayed on the graph.

RNA-sequencing Reveals Transcriptional Alterations in the Ipsilateral Cortex

In order to broaden our understanding of pathways that may be affected by the RIPK2 inhibitor treatment in the context of stroke injury, we performed bulk RNA-sequencing of the ipsilateral cortex in mice that received either RIPK2 inhibitor or vehicle pretreatment 6 h after ischemic stroke induction. We observed dramatic alterations to the transcriptional landscape in the injured area between our two treatment groups. Compared to the vehicle-administered group, we observed a total of 254 genes that were downregulated and 305 genes that were upregulated in the ipsilateral cortex of the RIPK2 inhibitor-treated group, evidencing the vast changes that occur within the injured brain after stroke when RIPK2 is inhibited. The genes most significantly up- and downregulated are depicted in Fig.6 A. Utilizing this data, we computed the significantly altered GO term pathways that are regulated by RIPK2 inhibitor treatment, both upregulated (Fig.6 B) and downregulated (Fig.6 C). With these data, we provide ample evidence that inhibition of RIPK2 elicits wide transcriptional alterations that serve to protect the brain after ischemic stroke.

Fig.6: Bulk RNA-sequencing of the ipsilateral cortex reveals that pre-treatment with RIPK2 inhibitor results in a markedly altered transcriptional landscape 6 h post-stroke compared to vehicle control.

A Heat map of significantly different genes expressed in the ipsilateral cortex of both treatment groups. B Significantly different GO Terms that were downregulated in the RIPK2 Inhibitor-treated group compared to the vehicle-administered group. C Significantly different GO Terms that were upregulated in the RIPK2 Inhibitor-treated group compared to the vehicle-administered group. n=11/group. Genes with a false discovery rate (FDR) below 0.05 and absolute fold change of ≥ 2 were considered differentially expressed. Differentially expressed genes were then subjected to enrichment analysis of GO functions using geneontology.org.

Discussion

Effective post-stroke pharmacotherapies have the potential to drastically reduce the neurological insult that develops in the timeframe immediately following ischemic stroke and thereby dramatically improve patient outcomes and long-term recoveries. Unfortunately, achieving reperfusion with recombinant tissue plasminogen activator (rtPA) remains the only pharmacological post-stroke therapy. Even if reperfusion is achieved, there remains a need to develop neuroprotective strategies to protect the brain against the damaging effects of ischemia and reperfusion (Chamorro et al., 2016). Unfortunately, very few therapeutics have shown efficacy in reducing the debilitating effects of stroke in patient populations, likely since many of these therapies modulate only a single detrimental pathway within the ischemic brain (Paul and Candelario-Jalil, 2021). As such, there remains a pressing need to discover and characterize novel druggable targets that may prove suitable for post-stroke pharmacotherapy.

To date, pharmacological inhibition of RIPK2 has never been employed as a mode for the treatment of experimental ischemic stroke. Recently, we have shown that genetic deletion of Ripk2, both globally and specifically in microglia, improves post-stroke outcomes in mouse models of the disease (Larochelle et al., 2023). Herein we explored the feasibility of inhibiting RIPK2 in vivo using a small molecule inhibitor to improve outcomes in mice after experimental ischemic stroke. In this way, we intended to replicate our previous observations in genetic models via a separate modality and assess the therapeutic relevance of RIPK2 pharmacological inhibition for post-stroke treatment.

We first determined that the RIPK2 inhibitor, in fact, blocked the kinase activity of RIPK2 after stroke by treating the animals with the RIPK2 inhibitor and examining the levels of p-RIPK2 relative to the vehicle controls. Satisfied with the observed reduction of p-RIPK2 and engagement of our target, we next sought to observe its effects in an in vivo model of ischemic stroke. We initially determined its powerful effects on reducing infarct volume and improving acute behavioral outcomes during the acute stage of stroke injury. Of note, inhibition of RIPK2 led to dramatic preservation of spontaneous locomotor activity, preservation of grip strength, and reduced neurological deficits. Grip strength, as assessed by both the vertical grid and weight grip tests, is a critical metric in post-stroke recovery and is a clinical correlate with positive post-stroke outcomes (Bohannon, 2019; Carson, 2018; Ekstrand et al., 2016).

Treatment with the RIPK2 inhibitor resulted in profound neuroprotective effects, evidenced by both dramatically preserved behavioral functioning and a significant reduction in infarct volume 24 h after stroke. We also found that the RIPK2 inhibitor exerted its neuroprotective effect by preserving BBB integrity, reducing the levels of active MMP-9 in the brain after stroke, and reducing the leakage of albumin into the brain after stroke. Of note is the fact that RIPK2 inhibitor treatment was delivered as a single dose at the time of reperfusion and no additional doses were given thereafter. This makes inhibition of RIPK2 an attractive modality for post-stroke therapy for its simplicity and potential to block many nodes of the ischemic cascade during the early phase of the injury, positively impacting stroke outcomes.

We provide evidence that this beneficial effect may be mediated by a significant early reduction in neuroinflammatory gene transcription, as we observed in our PCR results at the acute 6 h mark post-stroke. Using extremely high throughput bulk RNA-sequencing of the ipsilateral cortex, we were able to identify a stark transcriptomic contrast between the RIPK2 inhibitor-treated and vehicle-administered animals, indicating that inhibition of RIPK2 induces a profound effect on the response of various cell types in the brain after stroke. Of particular interest was the ability of the RIPK2 inhibitor to downregulate pathways associated with endocytosis and the generation of new neurons, perhaps a reflection of reduced damage to brain cells following reperfusion. Inhibitor treatment also resulted in a downregulation in pathways associated with RNA transcription, perhaps a reflection of decreased NF-κB and p38-mediated pro-inflammatory gene transcription in response to stroke injury, as RIPK2 is known to mediate the activation of these two pathways. These exciting results provide additional evidence that RIPK2 plays a pathological role in the initial propagation of stroke injury and more clearly identify RIPK2 as a therapeutic target for post-stroke pharmacotherapy.

In recent years, ischemic stroke has become more closely associated with perturbations of the gut; specifically, gut permeability and the trafficking and leakage of gut-derived pathogens and their products, as well as gut-derived immune cells into the systemic circulation (Ahnstedt et al., 2020; Benakis et al., 2016; Brea et al., 2021; Crapser et al., 2016; El-Hakim et al., 2021; Kurita et al., 2020; Lee et al., 2020; Liu et al., 2019b; Stanley et al., 2016; Wen et al., 2019). We hypothesize that this phenomenon may contribute to the pro-inflammatory priming of peripheral immune cells, which, expressing the receptors NOD1 and NOD2, are thereby utilizing RIPK2 to respond to an increased presence of peripheral bacterial products and, in doing so, exacerbate cerebral injury upon their infiltration into the brain after stroke. Increased activity of RIPK2 has been observed in inflammatory bowel disease by several groups, including a recent publication which suggests that hyperactivation of RIPK2 in both colon sections and leukocytes from Crohn’s disease and ulcerative colitis patients correlates with loss of expression of the tumor suppressor, Ras associated family protein 1A (RASSF1A) (Salla et al., 2023). Further experimental work is required to strengthen this hypothesis.

Future efforts will focus on administering RIPK2 inhibitor treatment throughout the duration of a long-term post-stroke survival experiment to assess whether progressive inhibition of RIPK2 produces a sustained protective effect following ischemic stroke. To improve the therapeutic relevance of RIPK2 inhibition as a therapy for post-stroke treatment, future studies will also include time-course experiments to assess the efficacy of RIPK2 inhibition when the inhibitor is delivered at escalating time points after the insult (i.e. 2 h, 4 h, 6 h post-stroke) and establish a therapeutic time window. This would then help us to understand the time dependence of RIPK2 inhibition in exerting its beneficial effects after stroke and establish a clinically relevant time window for therapeutic intervention with RIPK2 inhibitors.

A limitation of this current study is that the effect of RIPK2 inhibition in ischemic stroke was only assessed in young male mice. Future experiments will need to assess the effect of RIPK2 inhibition in animals of both biological sexes and aged animals to strengthen its therapeutic relevance as a post-stroke pharmacotherapy. This limitation is mitigated by the fact that we determined that aged female mice deficient for the Ripk2 allele are protected from stroke injury, thus providing evidence that the beneficial effect of reducing RIPK2 signaling post-stroke is not a sex-dependent effect (Larochelle et al., 2023). This study is also limited in that we only assessed the protective effect of RIPK2 inhibition in the acute phase of stroke injury. As mentioned previously, we intend to perform longitudinal behavioral experiments assessing the effects of RIPK2 inhibition on long-term recovery after stroke.

Herein, we provide transcriptional evidence that RIPK2 inhibition reduces the early pro-inflammatory response to stroke injury both through a reduction of inflammatory markers and a plethora of other pathways, which were revealed through bulk RNA sequencing of the injured cortex. Due to the limitations of the methods we employed, we are unable to extrapolate these findings to claim the relative contribution of different cell types responsible for such signaling. In future studies, we intend to employ flow cytometric cell sorting to assess post-stroke immune cell infiltration into the brain with RIPK2 inhibitor vs. vehicle control to determine the extent to which RIPK2 inhibition is reducing immune cell trafficking after stroke.

In conclusion, this study is the first to assess the effect of RIPK2 inhibition in ischemic stroke. This work is significant because it identifies a powerful effect of RIPK2 inhibition in improving animals’ outcomes during the acute stage of stroke injury and strengthens the evidence that RIPK2 is a viable target for post-stroke pharmacotherapy. We hope this work will advance the exploration of RIPK2 and its multifaceted roles in inflammatory/neuroinflammatory conditions. Moreover, we hope that in identifying novel pathways in the progression of stroke injury, we may lead to improvement in therapeutic interventions for the millions of individuals around the globe who suffer from the debilitating effects of ischemic stroke.

Abbreviations

IBD

inflammatory bowel disease

EAE

experimental autoimmune encephalomyelitis

MS

multiple sclerosis

CNS

central nervous system

MMP

matrix metalloproteinase

NOD

nucleotide oligomerization domain

PRR

pattern recognition receptor

DAMP

damage-associated molecular pattern

PAMP

pathogen-associated molecular pattern

ER

endoplasmic reticulum

rtPA

recombinant tissue plasminogen activator

TLR

Toll-like receptor

NLR

NOD-like receptor

RASSF1A

Ras associated family protein 1A

Data availability

The original data supporting this study's findings are available from the corresponding author upon reasonable request.