Intravenous antagomiR-494 lessens brain-infiltrating neutrophils by increasing HDAC2-mediated repression of multiple MMPs in experimental stroke

Background: Neutrophil infiltration and phenotypic transformation are believed to contribute to neuronal damage and clinical outcome in ischemic stroke. Emerging evidence suggests that HDAC2 is an epigenetic regulator of inflammatory cells. Here, we investigated whether miR-494 affects HDAC2-mediated neutrophil infiltration and phenotypic shift. Methods: The miR-494 levels in neutrophils from AIS patients were detected by real-time PCR. C57BL/6J mice were subjected to transient middle cerebral artery occlusion, and the N1/N2 neutrophil shift was examined. Cortical neurons were subjected to oxygen-glucose deprivation and stimulated with supernatant from differently treated neutrophils or were cocultured with neutrophils; neuronal injury was detected, and ChIP-Seq was performed to clarify which genes are the binding targets of HDAC2. Finally, a transwell assay was conducted to examine neutrophil migration. Results: Compared to the control subjects, AIS patients had increased neutrophil expression of miR-494, and in AIS patients, elevated miR-494 expression in neutrophils was a predictor of worse neurological outcomes. MiR-494 correlates with the upregulation of adhesion molecules in neutrophils of AIS patients. Systemically administered antagomiR-494 partly shifts neutrophils into the N2 phenotype in MCAO mice. AntagomiR-494-treated neutrophils exert a neuroprotective role in vitro. ChIP-seq revealed that HDAC2 targets multiple MMP genes in neutrophils of AIS patients. Further in vitro and in vivo experiments showed that antagomiR-494 repressed expression of MMP genes, including MMP7, MMP10, MMP13, and MMP16, to reduce the number of brain-infiltrating neutrophils by regulating HDAC2.


Introduction
Ischemic stroke is an acute and severe neurological disease that results in disability and death. Except for recanalization therapy, there are no effective pharmacotherapies available [1][2][3], which is due to two major reasons. First, single agent pharmacotherapy may be insufficient, and effective treatment of stroke requires targeting multiple complementary targets. Second, most therapeutics could not efficiently penetrate the brain. Therefore, a treatment that can regulate multiple pathological processes and alleviate brain injury via intravenous administration has the advantages of being easily translated into clinical use. The continuous inflammatory cascade reaction is a major factor in cerebral ischemia pathobiology and clinical outcomes [4]. Systemic immunomodulatory therapy based on intravenous cells is attracting increasing attention.
In particular, neutrophils are among the first circulating leukocytes to invade the ischemic tissue, which then attract other immune cells and exacerbate blood-brain barrier destruction and lesions by releasing a number of inflammatory factors, including metallomatrix proteases (MMPs) and reactive oxygen species [5,6].
Moreover, the population of infiltrated neutrophils in the brain is heterogeneous, and the peroxisome proliferator-activated receptor-γ agonist rosiglitazone or a tolllike receptor 4 deficiency was suggested to have the ability to reprogram neutrophils toward an anti-inflammatory (N2) neutrophil phenotype, which may assist in resolving inflammation and may provide neuroprotection in ischemic mice 5 [7][8][9]. However, the mechanisms involved in neutrophil transmigration and phenotypic shift in the ischemic brain are not fully elucidated. This makes it critical to find a molecule that can simultaneously target these two important neutrophilrelated pathological processes.
As a promising novel class of small regulatory molecules, microRNAs have been found to play pivotal roles in the inflammatory process associated with ischemic stroke [10], and accumulating evidence has indicated the value of miRNAs in diagnosis, prognosis and treatment of ischemic stroke [11,12]. According to the results of our previous research, miR-494 in lymphocytes was robustly upregulated and was related to lymphocyte polarization in ischemic stroke patients, and these findings were further confirmed by our animal experiments [13]. Thus, whether miR-494 has an effect on neutrophil phenotypic transformation and infiltration is not clear. Histone deacetylase 2 (HDAC2), a class I HDAC, is known to act as epigenetic and transcriptional regulators. HDAC2 was shown to be a target of miR-494 and plays a part in the T helper cell phenotype and microglia shift in ischemic stroke [13,14] and in MMP-9-mediated macrophages [15]; therefore, whether HDAC2 is implicated in neutrophil phenotypic transformation following ischemic stroke is not clear. The goal of our current research was to 1) evaluate the clinical significance of miR-494 changes in neutrophils of acute ischemic stroke (AIS) patients and 2) assess whether miR-494 was involved in the N1/N2 phenotypic transformation and infiltration of neutrophils by targeting HDAC2.

Clinical subjects
The first-ever AIS patients were recruited between November 2017 and October 6 2018 and were approved for study participation by the Medical Ethics Committee of Xuanwu Hospital of Capital Medical University (trial registration no. ChiCTR-NCT-03577093). Written informed consent was obtained from all included patients or their relatives before enrolment. The inclusion criteria were as follows: (1) admitted within 6 hours of symptom onset, (2) brain MRI or repeated CT verified the diagnosis of AIS, and (3) blood samples were collected no more than 6 hours after stroke attack. Patients were excluded if they had previous stroke, malignant tumors, renal disease, acute myocardial infarction, severe heart failure, immune diseases or other neurological diseases. Finally, AIS patients were included in our study. The patients underwent standard neurological and general medical evaluations and assessments with the National Institute of Health Stroke Scale (NIHSS) score at admission, and ischemic stroke was diagnosed in accordance with the guidelines. Age-and sexmatched healthy controls were recruited from the Medical Examination Center. Some patients were examined at a follow-up visit one year later, and a modified Rankin Scale (mRS) score was used to investigate the patient's functional independence.

Separation of neutrophils
Venous blood samples from AIS patients and healthy controls were collected into K3EDTA tubes, and neutrophils were separated by a standard Ficoll-Paque Plus gradient method as previously reported [13]. Then, neutrophils were suspended in 0.5 ml of TRIzol and frozen at −80°C prior to the test.

Real-time reverse transcription PCR (RT-PCR)
Total RNA from neutrophils was isolated using oligo-d (T) primers and SuperScript III 7 /RNaseOUT Enzyme Mix (Invitrogen, Carlsbad, CA, USA). For miR and mRNA quantification, total RNA was reverse transcribed using the RNeasy Mini Kit (Qiagen, Gaithersburg, MD, USA) according to the manufacturer's protocol. MiRs and mRNA abundance was assessed by RT-PCR using All-in-One miR RT-PCR Reagent Kits.
Values were normalized to U6 or β-actin via the 2 − ΔΔCT method. RT-PCR was performed in triplicate. Primers are listed in Table 1.

ChIP sequencing
Neutrophils of AIS patients were cross-linked in 1% formaldehyde for 20 minutes at HiSeq 4000 using a HiSeq 3000/4000 SBS Kit (FC-410-1003, Illumina). Image and base calling, Solexa CHASTITY quality filter, human genome (UCSC hg19) alignment, peak detection, and peak annotation were performed, and the data were analyzed.
The ChIP samples from each sample served as negative inputs for peak calling. [16]. Animal groups were randomized and blinded to experimenters.

Transient middle cerebral artery occlusion (tMCAO) surgery and groups
The MCAO model was induced through intraluminal occlusion of the right middle cerebral artery as published previously [17]. Briefly, mice were anesthetized with isoflurane, and their body temperature was maintained at 37.0 ± 0.5°C with a heat lamp during the surgery. A silicone rubber-coated nylon filament with a 0.19 mm diameter silicon tip (Catalog number: 701956PK5Re; Doccol Corporation, Sharon, MA, USA) was inserted into the right middle cerebral artery for 45 minutes to obstruct the flow of blood, and then the suture was removed to allow for reperfusion. Local cerebral blood flow was monitored using a transcranial laser Doppler (LDF, PeriFlux System 5000; Perimed, Sweden) to verify the occlusion of the middle cerebral artery. Sham-operated mice underwent a similar procedure without suture insertion. All attempts were made to minimize animal suffering and the number of animals used.
Mice were randomly grouped into three groups with a lottery drawing box: (1) sham group: mice accepted similar procedures without inserting the suture; (2) MCAO + 9 control group: mice underwent MCAO surgery and were injected with negative control miRNA; (3) MCAO + antagomiR-494 group: mice underwent MCAO surgery and were injected with antagomiR-494. AntagomiR-494 and negative control miRNA (100 µM, GenePharma, Suzhou, China) were administered to mice via right intracerebroventricular injection over 20 minutes, and the needle was held in situ for 10 minutes before it was slowly withdrawn. The sequence of the negative control miRNA was: 5'-CAGUACUUUUGUGUAGUACAA-3', and the sequence of antagomiR-494 was 5'-GAGGUUUCCCGUGUAUGUUUCA-3'. Three to four mice in each group were used for flow cytometry analysis (bone marrow and blood, respectively), and six mice in each group were used for immunofluorescence staining at 3 days poststroke.

Immunofluorescence staining
Brain tissues for immunofluorescence analysis were obtained 3 days after MCAO and fixed in 4% paraformaldehyde for at least 48 hours. Mouse brains were transferred to a 30% sucrose solution and allowed to dehydrate. Then, 20-mm-thick coronal brain slices were acquired on a cryostat vibratome and incubated with a primary antibody against Ly-6G (1:200, Santa Cruz Biotechnology, CA, USA) at 4°C overnight followed by incubation with a donkey anti-rat antibody conjugated to Alexa 488 (1:200, Jackson ImmunoResearch) for an hour at room temperature. Images were then captured using a fluorescence microscope (Olympus, Japan), and the number of Ly-6G-positive cells in the infarction border cortex was counted and analyzed with ImageJ software.

Flow cytometry analysis
For flow cytometry analysis, blood cell lysis buffer (eBioscience, San Diego, CA, USA) was added to peripheral blood and bone marrow cells. Then, the cells were centrifuged and suspended in 2% fetal bovine serum (FBS) in PBS. Next, cells were stained with Ly-6G-PE, CD11b-PerCP/PECy5.5, CD16-FITC and CD206-APC (BD Biosciences, New Jersey, USA). After filtration through a 40 μm nylon cell strainer, samples were detected on a BD Accuri™ C6 for fluorescence flow cytometry analysis.

Cortical primary neuron cultures and oxygen-glucose deprivation
Cortices from the brains of C57BL/6J mice of embryonic day 16-18 were dissected and minced into small pieces using a sterile scalpel and digested for 3 minutes using trypsin (Sigma-Aldrich). Then, trypsin was discarded, and fresh buffer with DNAzyme was added, followed by filtration through a 70-μm mesh cell. Single cell suspensions were suspended in Dulbecco's modified Eagle medium (DMEM) containing 10% FBS and 100 U/ml penicillin-streptomycin and seeded into culture plates precoated with poly-D-lysine (0.1 mg/mL) at a 6*10 5 cells/mL density. Six hours later, neurons medium was replaced with Neurobasal/B27/Lglutamine/Gentamicin medium. Five days later, 80-90% confluent neurons were used for oxygen-glucose deprivation (OGD) treatment for 2 hours. After OGD, neurons were stimulated with supernatant collected from HL-60 cells or cocultured with transfected HL-60 cells. Twenty-four hours later, neurons in 48-well plates were fixed and assessed by immunofluorescence staining and a high-content screening assay for neurite outgrowth.

High-content screening assay
Neurons in 48-well plates were fixed with 4% paraformaldehyde and then blocked with 5% donkey serum and 0.3% Triton X-100 in PBS for 1 hour at room temperature. Then, the cells were incubated with primary antibodies against synaptophysin (1:150, catalog 32127, Abcam) overnight at 4°C. Then, the neurons were incubated with a donkey anti-rabbit antibody conjugated to Alexa 594 (1:200, Jackson ImmunoResearch), followed by counterstaining with DAPI. Total outgrowth, total processes, mean processes, total branches and total cell body areas were evaluated under a phase-contrast microscope using a 10× objective of ImageXpress Micro XLS System (ImageXpress Micro XLS System, Molecular Devices, USA). Sixteen areas in each well were randomly selected. Neurite length was measured using MetaXpress (RRID: SCR_016654). In this multiparameter assay, the software can segment the neuronal cell body and compute fluorescently labeled neurites with respect to their total outgrowth, branches, processes, cell body area and mean process.

Statistics
Data are presented as the mean±SD. Independent sample t-test was used for comparisons between two groups, while one-way ANOVA with Tukey's post hoc test was applied for comparisons between multiple groups. Correlation analysis was conducted using the Spearman correlation test. Receiver operator characteristic (ROC) curve analysis was adopted to calculate the predictive power of the sensitivity and specificity for the diagnosis of ischemic stroke, and the area under the ROC curve (AUC) was calculated. Differences were regarded as significant if

Elevated miR-494 in neutrophils of AIS patients predicts worse neurological outcomes
First, we assessed whether miR-494 was abnormally expressed in neutrophils of AIS patients, and RT-PCR analysis revealed that miR-494 was obviously elevated in AIS patients compared to that in healthy controls ( Fig. 1A; p<0.05 year after symptom onset (Fig. 1C, D; p<0.05). These findings indicate that increased miR-494 expression in neutrophils of AIS patients is related to serious neurological deficits and predicts worse neurological outcomes.

MiR-494 correlates with the upregulation of adhesion molecules in neutrophils of AIS patients
To further characterize the phenotypic shift of peripheral neutrophils of AIS patients, we analyzed the relevance between expression of N1 and N2 markers of 14 neutrophils and neutrophil miR-494 expression. First, the RT-PCR results demonstrated that both the CD16 and CD206 levels in neutrophils were significantly decreased in AIS patients compared with those in healthy controls ( Fig. 2A, B; p<0.05). However, our correlation analysis showed no obvious linear correlation between the CD16 level and miR-494 expression in neutrophils, nor a correlation between the CD206 level and miR-494 expression in neutrophils (Fig. 2C, D).
Given that our preceding bioinformatics analysis indicated that miR-494 target genes are involved in the focal adhesion molecule pathway [13], we investigated the expression of the classical adhesion molecules CD11b and MMP9 in the peripheral neutrophils of AIS patients and analyzed their relationship with the miR-494 levels. Consistent with bioinformatics analysis, both CD11b and MMP9 were prominently increased in neutrophils of AIS patients and were negatively associated with miR-494 expression in neutrophils (Fig. 2E-H; p<0.05). These results demonstrate that miR-494 may target the regulation of neutrophil infiltration after ischemic stroke.

Intravenous antagomiR-494 partly shifts neutrophils into the N2 phenotype in MCAO mice
To further explore the immunomodulatory effect of miR-494 on neutrophils in vivo, we implemented an MCAO mouse model. Flow cytometry analysis revealed that the ratio of both N1 and N2 neutrophils in the peripheral blood was increased (Fig. 3A-C) and that the ratio of N2 neutrophils was decreased in the bone marrow of ischemic mice (Fig. 3D-F); in addition, antagomiR-494 reversed the reduction in N2 neutrophils in the bone marrow remarkably (Fig. 3F; p<0.05). However, the ratio of N2 neutrophils in peripheral blood in MCAO+antagomiR-494 mice was not obviously changed compared with that in MCAO mice (Fig. 3A, C). Noticeably, the alteration of N1 neutrophils in both peripheral blood and bone marrow was not evident after antagomiR-494 treatment (Fig. 3A, B, D, E). These results suggested that ischemia led to distinct changes in N2 neutrophils and that miR-494 could induce an apparent increase in N2 neutrophils in bone marrow after MCAO.

AntagomiR-494-treated neutrophils exert a neuroprotective role in vitro
To explicate whether antagomiR-494-treated neutrophils could directly or indirectly influence hypoxia-induced neuronal injury, we treated neurons subjected to OGD with conditioned medium from differently treated HL-60 cells (Fig. 4A) or cocultured neurons with differently treated HL-60 cells (Fig. 4B) for 24 hours. High-content analysis for the neurite marker synaptophysin showed that there was a substantial increase in axonal injury in OGD-treated neurons, and treating HL-60 cell supernatant with antagomiR-494 partly mitigated axonal injury, as indicated by an increased number of neuron processes ( Fig. 4A; p<0.05). However, antagomiR-494treated HL-60 cells seemingly did not show any protective effect on the hypoxiaassociated neurotoxic response when cocultured with neurons (Fig. 4B). Thus, we demonstrated that antagomiR-494 might exert a slightly neuroprotective effect on ischemic neurons by shifting neutrophils toward a beneficial N2 phenotype.

HDAC2 targets multiple MMP genes in neutrophils of AIS patients
Previously, we showed that miR-494 regulates the Th1/Th2 shift by directly targeting HDAC2. In light of this, we assessed expression of HDAC2 in neutrophils and found that expression was obviously downregulated in AIS patients compared to that in healthy controls; these results are contrary to those of miR-494 expression 16 (Fig. 5A). To identify HDAC2 target genes in neutrophils of AIS patients, we performed ChIP-seq and revealed that HDAC2 bound to various DNA-binding sites of several MMPs. Among them, HDAC2 was found to bind to the upstream region of the MMP21 DNA and MMP25-AS1 genes, the intergenic region of the MMP16, MMP13 and MMP10 genes, and the intron region of the MMP7 gene. Noticeably, HDAC2 could bind to both the upstream and intergenic regions of the MMP27 gene at different loci (Fig. 5B, C, D). These data indicate that downregulated HDAC2 potentially regulates the expression of multiple MMP genes in neutrophils of AIS patients.

AntagomiR-494 repressed MMP expression and brain-infiltrating neutrophils by regulating HDAC2
To verify whether miR-494 affects MMPs by targeting HDAC2, we transfected HL-60 cells with antagomiR-494 or HDAC2-specific siRNA and examined the mRNA levels of Immunofluorescence staining demonstrated that ischemia led to an evident increase in neutrophils in the ischemic brain; however, antagomiR-494 administration through the tail vein largely reversed this situation (Fig. 6). These results illustrate that antagomiR-494 could reduce the expression of MMPs and neutrophil infiltration in ischemic brains by regulating HDAC2.

Discussion
The present study demonstrates that miR-494 plays a crucial role in mediating both brain infiltration of neutrophils and N1/N2 neutrophil programming after ischemic stroke. Here, we show that elevated neutrophil miR-494 levels in AIS patients were correlated with worse clinical neurological outcomes. Moreover, antagomiR-494 partly shifted neutrophils into the N2 phenotype in vivo and mediated certain neuroprotective effects in vitro. More importantly, we discovered that miR-494 levels correlated with the upregulation oftheadhesion molecules CD11b and MMP9 in neutrophils of AIS patients. Further, HDAC2 binds to multiple MMP genes in neutrophils of AIS patients, and antagomiR-494 represses MMP expression and brain-infiltrating neutrophils by regulating HDAC2. Overall, our study provides the first evidence of neutrophil reprogramming in neuroinflammation following ischemic stroke by miR-494-induced modulation of HDAC2.
Combined with our previous studies, miR-494 was found to be increased in peripheral neutrophils as well as plasma and lymphocytes [13] of AIS patients and reflects both neurological deficits at admission and long-term neurological outcome scores. As there is no ideal biomarker to predict the long-term prognosis of AIS, we speculate that miR-494 in the peripheral blood may provide an alternative choice.
After stroke, the population of neutrophils is heterogeneous, including pro-and antiinflammatory neutrophils (N1/N2). Our clinical data demonstrated that both N1 and N2 neutrophils were decreased in the blood of AIS patients but did not show a linear correlation with miR-494 levels. In MCAO mice, ischemia led to an obvious increase in N2 neutrophils in the blood and a decrease in N2 neutrophils in the bone marrow, and these results are consistent with those of other studies [7]; antagomiR-494 treatment via the tail vein partly reduced the number of N1 neutrophils in the peripheral blood and markedly elevated the number of N2 neutrophils in the bone marrow. In accordance with these results, an in vitro study showed that antagomiR-494-treated neutrophils or the supernatant could reverse OGD-induced primary neuronal damage. Taken together, it is highly likely that systemically administered antagomiR-494 confers a neuroprotective effect by mediating the N2 shift and Th2 shift [13] as well as by mitigating brain-infiltrated neutrophils after ischemic stroke. Our laboratory has already proven that miR-494 can directly bind to the 3' noncoding region of HDAC2 and inhibit HDAC2 expression by a luciferase reporter gene assay. Thus, we examined the expression of HDAC2 in neutrophils and found that HDAC2 was remarkably decreased in AIS patients compared with that in healthy controls, and these results are in direct contrast to the change in miR-494 expression. Therefore, we further applied ChIP sequencing to study the regulation of HDAC2 involved in the expression of genes in neutrophils in patients with AIS.  25], and that MMP10 could be regulated by HDAC3 and HDAC7 [25,26]. For the first time, we found that HDAC2 could potentially modulate the MMP10, MMP16, MMP21, MMP25-AS1, and MMP27 genes in neutrophils of AIS patients.
Then, through our in vitro experiments, we further verified that prescribing neutrophils with antagomiR-494 could decrease expression of MMP7, MMP10, MMP13 and MMP16. Among them, MMP7 was reported to be involved in neutrophil migration and activation [27,28]. Moreover, MMP13 deficiency was involved in a reduced infarction area, reduced hemorrhagic events and improved neurological outcome in the acute phase of ischemic stroke [29], and it was found to be related to neutrophil infiltration in bladder cancer [30]. Moreover, MMP10 was significantly increased in the brains of mice subjected to ischemic stroke [31] and was associated with macrophage-mediated neuroinflammation and atherosclerosis [32]. MMP16 was found to be associated with the migration of HUVECs [33], with no reports about its effects on neutrophil migration. Here, we found that MMPs may be involved in miR-494-mediated neutrophil migration in ischemic stroke. However, only MMP7, MMP10 and MMP13 could be upregulated after miR-494 plus HDAC2 siRNA treatment but could not be upregulated with antagomiR-494 treatment. Therefore, although antagomiR-494 could regulate MMPs by targeting HDAC2, it might not be the only 20 modulatory pathway. The administration of HDAC2-specific siRNA did not completely reverse the downregulation of MMPs. Hence, our results suggested that miR-494 could regulate the expression of MMPs partly by targeting HDAC2.
However, some flaws exist in our study. Our study found that not all downregulated MMPs in antagomiR-494-treated neutrophils could be reversed by HDAC2 siRNA administration. As discussed above, multiple MMPs could be regulated by the HDAC family, including HDAC1, HDAC2, HDAC3, and HDAC7. Our previous study suggested that HDAC3 was a direct target of miR-494 [34]. It is possible that miR-494 could regulate these MMPs by targeting HDAC3 or other members of the HDAC family.  AntagomiR-494 repressed MMP expression and brain-infiltrating neutrophils by regulating HD Figure 7 Schematic diagram of the proposed molecular mechanism. Our study demonstrates that anta