Background
Ischemic stroke triggers complex, multipart cascade of events leading to irreversible brain damage and dysfunction of the neurovascular network in the ischemic core [
1]. Hypoxic but still viable peri-infarct area surrounding the core region remains to be a subject of intensive investigation, with the focus on neuroprotective and pro-regenerative treatments for preservation of salvageable brain tissue. Post-ischemic inflammatory response is an integral part of both brain damage and recovery, yet its function remains controversial. Lower levels of pro-inflammatory cytokines and higher expression of anti-inflammatory cytokines are associated with lower infarct size and a better clinical outcome [
2,
3]. On the other hand, number of pro-inflammatory cytokines is implicated in neuroprotection and post-stroke plasticity and, thus, in facilitation of the restorative processes and tissue remodeling [
4,
5]. Original elevation of cytokine expression during acute inflammation starts as early as several hours and is associated with recruitment of different types of cells into the ischemic area, including neutrophils, lymphocytes, monocytes, and activation of resident microglia, astrocytes, and endothelial cells [
3]. Activation of microglia and astrocytes leading to additional release of pro-inflammatory factors can last for up to several weeks after stroke [
5].
Ischemia is associated with significant changes in the molecular profile of the neurovascular elements [
6,
7]. Short messenger RNA (mRNA)-interfering molecules microRNAs (miRNAs) are recently identified as essential modulators of inflammatory response in the brain [
8,
9]. Multifunctional miRNA miR-155 is among the miRNAs with expression profiles significantly affected by cerebral ischemia [
10‐
12]. miR-155 has been implicated in regulating various physiological and pathological processes including immunity and inflammation [
13,
14]. Recent investigations identified miR-155 as a potent regulator of the neuroinflammatory response in Japanese encephalitis and Alzheimer’s disease [
15,
16]. Microglia-mediated immune response associated with several pathological states is regulated by miR-155 [
17‐
19]. Specific inhibition of this miRNA is accompanied by reduced inflammation, in several pro-inflammatory conditions [
20,
21]. miR-155 mediates the inflammatory response through negative targeting of its direct target proteins including Src homology 2-containing inositol phosphatase 1 (SHIP-1) [
15,
22], suppressor of cytokine signaling molecules SOCS-1 and SOCS-6 [
19,
23,
24], and transcription factor CCAAT/enhancer binding protein beta (C/EBP-β) [
25]. All these miR-155-targeted factors control the inflammatory response via suppression of cytokine transcription and signaling.
In our previous study, we found that specific in vivo inhibition of miR-155 after experimental mouse ischemia supports brain microvasculature in the peri-infarct area, reduces brain tissue damage, and improves the animal functional recovery. In addition, miR-155 inhibition after distal middle cerebral artery occlusion (dMCAO) resulted in alterations of several cytokine/chemokine gene expression [
26], prompting our interest to possible role of miR-155 in post-ischemic inflammation. miR-155 is expressed in hematopoietic cells (including B cells, T cells, monocytes, and granulocytes), endothelial cells, microglia, and astrocytes [
17,
23,
27]. All these cell types are able to express and secrete pro- and anti-inflammatory molecules and, thus, actively participate in the inflammation process. Therefore, we proposed that systemic in vivo inhibition of miR-155 could significantly affect the post-stroke inflammatory response. While cytokine expression has been mostly studied during the acute inflammation, few studies have examined the temporal profile of pro- and anti-inflammatory molecules during subacute phase of stroke. The present investigation focuses on the long-lasting effect of miR-155 inhibition on the inflammatory response.
To our knowledge, this is a first report describing the effect of intravenous anti-miRNA injections on the time course of cytokine expression and cellular inflammatory response following mouse stroke.
Methods
Animal groups
All institutional and national guidelines for the care and use of laboratory animals were followed during the experiments. C57BL/6 male mice (2-month old, Jackson Laboratories;
https://www.jax.org/strain/000664) were used in our studies. Experimental groups included sham-operated mice, mice subjected to dMCAO and specific miR-155 inhibitor injections (inhibitor group), and mice subjected to dMCAO and control (scrambled) miRNA inhibitor injections (control group).
Distal middle cerebral artery procedure
A distal (direct) middle cerebral artery occlusion (dMCAO) was utilized as an experimental model of cortical ischemia [
28]. dMCAO was performed on 2-month-old male C57BL/6 mice, as described in [
26]. The mice were anesthetized using isoflurane gas (induction dosage 2–3 %; maintenance dosage 1.5–2 %) and a mixture of O
2:N
2O gases in the ratio 2:1, delivered during the surgery. The MCA was exposed via transtemporal approach [
28,
29]. A small burr hole (located 1 mm rostral to the fusion of zygoma and squamosal bone and 3 mm ventral to the parietal bone) was made on the left side of the skull surface, and the MCA was coagulated with low-heat electrocautery (Bovie Medical). In sham-operated animals, the MCA was exposed but not coagulated.
miRNA inhibitor injections
Injections of specific anti-miR-155 miRCURY LNA™ (Product#4101082-001,
https://www.exiqon.com/microrna-knockdown-probes) inhibitor or control inhibitor (scrambled oligonucleotide, Product#199006) from Exiqon Company were initiated at 48 h after dMCAO and performed for three consecutive days. Oligonucleotides were introduced via mouse lateral tail vein; the dose was 10 mg/kg in saline, total injected volume 100 μl [
26].
Mouse cytokine and chemokine PCR array
Mouse cytokine and chemokine PCR array (Qiagen, Cat# PAMM-150Z,
http://www.sabiosciences.com/rt_pcr_product/HTML/PAMM-150A.html) was utilized to evaluate the expression of genes encoding major pro- and anti-inflammatory cytokines and chemokines in the RNA samples. At 7 days after dMCAO, three brains per experimental group (inhibitor and control) were used to generate separate sample triplicates for the analyses. Brain cortices were dissected and stored in RNAlater solution (Ambion). Total RNA was isolated using mirVana miRNA (AB/Ambion) isolation kit, from the hemispheres ipsilateral to dMCAO injury. All measurements and data quantification were performed by Qiagen Company Sample & Assay Technologies team. The obtained raw data were analyzed using RT2 Profiler software (Qiagen). Only the genes with consistent expression levels (within the triplicate samples) were picked up for statistical analysis. The fold changes of gene expressions (inhibitor vs control) were calculated, and the transcripts that showed ≥2-fold change in expression (either up- or downregulated) were retained. At the final step, statistical significance (
p value <0.05) and reliability of the results was automatically evaluated. The raw data are deposited in the Open Science Framework general data repository, link:
https://osf.io/3zhc4/?view_only=0826f6e687884b90ab774328c2746ae1.
Cytokine protein expression analysis
At 48 h and 7, 14, and 21 days after dMCAO, six brains per experimental group (sham, inhibitor, and control) were used to generate separate samples. Brain cortices from both ipsi- and contralateral (to dMCAO damage) hemispheres were dissected on ice and rapidly frozen. Lesioned and intact hemispheres were analyzed separately. Brain tissue was homogenized in tissue extraction buffer (Life Tech/Invitrogen Cat# FNN0071, 5 ml per 1 g of brain tissue) with the addition of protease inhibitor cocktail (Sigma). The samples were centrifuged at 10,000 rpm for 5 min, and supernatant was collected and kept on ice. Protein concentration was determined for each sample, using DC protein assay kit from BioRad. Brain tissue samples were normalized for total protein content and diluted at 1:10 in assay buffer. Expression levels of CCL12 and CXCL3 were detected using Mouse CCL12/MCP5 PicoKine™ (Boster Biological Technology, Cat# EK1128) and Mouse CXCL3 PicoKine™ ELISA Kits (Boster Biological Technology, Cat# EK1364), according to manufacturer’s recommendations. Other cytokine protein expression was detected using Mouse Cytokine Magnetic 20-Plex Panel Kit (Life Tech/Invitrogen, Cat# LMC0006M,
https://www.thermofisher.com/order/catalog/product/LMC0006M), according to the manufacturer’s recommendations. The Panel is designed for the quantitative determination of FGF-basic, GM-CSF, IFN-γ, IL-1α, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-10, IL-12 (p40/p70), IL-13, IL-17, IP-10, KC, MCP-1, MIG, MIP-1α, TNF-α, and VEGF expression. The measurements were done using Luminex xMAP-100 system, at the UNM Center of Molecular Discovery. Cytokine concentrations were calculated automatically, using Specialized Luminex system software. For quantification, only cytokines with consistent expression throughout the samples were retained. Two-way ANOVA followed by Tukey’s multiple comparison test was used for final statistical analysis. The raw data are deposited in the Open Science Framework general data repository, link:
https://osf.io/dz5ue/?view_only=4f2c586e7562432595d894b86154b97e.
Western blot analysis
Five to six brains per experimental group were collected at 7, 14, and 21 days after dMCAO and used to generate separate samples. Brain cortices from ipsi- and contralateral (to dMCAO damage) hemispheres were dissected on ice and rapidly frozen. For tissue lysate preparation, brain tissue was homogenized in tissue extraction buffer (Life Tech/Invitrogen Cat# FNN0071, 5 ml per 1 g of brain tissue) with the addition of protease inhibitor cocktail (Sigma). The samples were centrifuged at 10,000 rpm for 5 min, and supernatant was collected and kept on ice. Protein concentration was determined for each sample, using DC protein assay kit from BioRad. The proteins were separated on 4–20 % gradient Criterion precast gels (Bio-Rad). A broad range molecular weight calibration marker from 10,000 to 250,000 MW (Bio-Rad) was used as a standard. Janus kinase (JAK)/signal transducer and activator of transcription (STAT) signaling pathway analysis was performed using phospho-STAT antibody sampler kit (Cell Signaling Technology Cat#9914, RRID:AB_330385). Other antibodies used were as follows: mouse monoclonal anti-STAT-3 (Cell Signaling Technology Cat# 9139, RRID:AB_331757); rabbit polyclonal anti-SOCS-1 (Cell Signaling Technology Cat# 3950S, RRID:AB_2192983); anti-SHIP-1 (Cell Signaling Technology Cat# 2728, RRID:AB_2126244); anti-C/EBP-β (Cell Signaling Technology Cat# 3087, RRID:AB_2078052); rabbit polyclonal anti-SOCS-6 (Santa Cruz Biotechnology Cat# sc-5608, RRID:AB_661195); rabbit polyclonal anti-Iba-1 (Wako Cat# 019-19741, RRID:AB_839504); rat anti-mouse CD68 (AbD Serotec Cat# MCA1957, RRID:AB_322219); goat polyclonal anti-CD206 (R and D Systems Cat# AF2535, RRID:AB_2063012); and anti-CD45 (R and D Systems Cat# AF114, RRID:AB_442146). Loading was confirmed by comparing actin immunoreactivity across the lanes, using mouse monoclonal anti-actin (Sigma-Aldrich Cat# A2228, RRID:AB_476697). Horseradish peroxidase-labeled secondary antibodies were from Cell Signaling and Amersham Biosciences. The density of the protein bands was determined using ImageJ software (NIH Image, RRID:SCR_003073), normalized by actin expression and quantified using Microsoft Excel software.
Immunohistochemistry and fluorescence microscopy
The animals were perfused with PBS and 4 % paraformaldehyde (PFA); brains were removed and fixed in PFA and subsequently cryoprotected in 30 % sucrose. The brains were sectioned along the rostral-caudal axis; 16 μm coronal sections were mounted and subjected to immunostaining. Briefly, the slices were post-fixed with 4 % PFA and quenched with 50 mM NH4Cl. For antigen retrieval, the mixture of 1 % antigen unmasking solution (Vector Laboratories Cat# H-3301, RRID:AB_2336227) and 10 % of 1 M sodium citrate in PBST (PBS containing 0.05 % Triton X-100) was boiled, and tissue slices were immersed in the mixture for 15 min. Ten percent normal goat serum containing 0.05 % Triton-X was used as a blocking buffer. The slides were subsequently incubated with primary antibodies, at 4 °C overnight. Following antibodies were used for the immunostaining: rabbit polyclonal anti-Iba-1 (Wako Cat# 019-19741, RRID:AB_839504); mouse monoclonal anti-GFAP (BD Biosciences Cat# 556328, RRID:AB_396366); Cy-3-conjugated anti-NeuN (Millipore Cat# ABN78C3, RRID:AB_11204707); goat polyclonal anti-CD206 (R and D Systems Cat# AF2535, RRID:AB_2063012); anti-CD45 (R and D Systems Cat# AF114, RRID:AB_442146); and Alexa Fluor 488-conjugated rat anti-mouse CD68 (AbD Serotec Cat# MCA1957A488, RRID:AB_324822). FITC-, rhodamine-, and Cy-5-conjugated secondary antibodies (1:200 concentration, 2 h at RT) were from Jackson Immunoresearch. DAPI staining was used to visualize nuclei. The incubations were performed in the humidity chamber.
Quantification of the immunofluorescence staining intensity
Mouse brain coronal sections co-immunostained for Iba-1/CD206/DAPI or Iba-1/CD45/CD68 were imaged using single-scan and tile-scan imaging on Zeiss LSM510-META and Zeiss LSM-800 Airyscan confocal microscopes. Tile-scan (10 × 15 tiles) imaging performed with ×40 and ×60 magnification allowed us to acquire high-resolution scans of the entire infarct core and peri-infarct areas in the injured hemisphere, as well as to capture a corresponding area in the contralateral hemisphere. Occasionally, Z-stack imaging was used to localize and confirm the intracellular labeling. Three to four coronal sections from the rostral part of the brain (mostly affected by dMCAO) per mouse (
N = 5–6 per experimental group/per time point) were imaged and quantified. ImageJ software (NIH Image, RRID:SCR_003073) and a modified fluorescence intensity quantification method described in [
30,
31] were used for quantification. The brain area of the same size throughout all images (945 × 387 pixels, corresponding to ~500 × 200 μm), including the edge of the infarct (glial scar) and part of the peri-infarct region, was selected and analyzed. The following formula was used for calculations: Corrected total fluorescence of the area = total fluorescence intensity of the selected area − (selected area × mean fluorescence of background readings).
Electron microscopy
Mice were perfused with 0.5 % glutaraldehyde and 2.5 % paraformaldehyde in 0.1 M Sorensen’s buffer. Cortical tissue (1 × 1 mm) samples were fixed in 2 % osmium, dehydrated, embedded in Araldite, thin-sectioned to 80–90 nm, and stained with 4 % uranyl acetate in methanol followed by lead citrate as described in [
26,
32]. Images were acquired using Hitachi H7500 Transmission Electron Microscope equipped with AMT X60 camera. Three to four animals per experimental group (control and inhibitor) were used for analysis; high magnification (×800–15,000) images were taken both in the lesioned and intact hemispheres.
Statistical analysis
PCR array data were analyzed using RT2 Profiler software (Qiagen). The p values were calculated based on a Student’s t test of the replicate 2^(−Delta Ct) values for each gene in the control and inhibitor groups. Statistical analysis for all other obtained data was performed using Prism and GraphPad Prism v.6.05, or R software. Two-way ANOVA followed by Tukey’s multiple comparison test was used for cytokine protein expression analysis. Student’s t test was performed for the Western blot, immunofluorescence microscopy, and electron microscopy data analyses.
Discussion
In our previous study on the in vivo inhibition of miR-155 after mouse dMCAO, we detected that the vascular support achieved by anti-miR-155 injections at 7 days after the experimental stroke played a critical role in the prevention of delayed neuronal loss in the peri-infarct area at the later stages. Our previous analysis showed significant (34 %) reduction of infarct size in miR-155 inhibitor-injected animals at 21 days after dMCAO. Reduced brain injury and preservation of brain tissue reflected in efficient functional recovery of inhibitor-injected animals: bilateral asymmetry/adhesive removal and gait/locomotion tests demonstrated that the mice from the inhibitor group regained their sensorimotor deficits faster than controls [
26]. Based on these data, we assumed that vascular support achieved by miR-155 inhibition played a critical role in the prevention of neuronal loss in the peri-infarct area and, thus, in the improvement of functional outcome after stroke. We proposed that increased blood flow and improved vascular integrity could be mainly attributed to miR-155 inhibition-induced stabilization of tight junction (TJ) protein ZO1, via the upregulation of miR-155 target protein Rheb. Here, we hypothesized that, in addition to this possible mechanism, downregulation of pro-inflammatory miR-155 could improve stroke outcome by significantly influencing a post-stroke inflammation. At 7 days after dMCAO, there was a decreased mRNA and protein expression of pro-inflammatory cytokines CCL12 (also known as monocyte chemotactic protein MCP-5) and CXCL3 in the inhibitor group. Both cytokines are known to trigger vascular inflammation via activation of monocyte adhesion and migration through the vascular endothelium, as well as atherosclerotic plaque formation [
48,
49]. Seven days time point was also accompanied by significant upregulation of miR-155-targeted proteins. miR-155 direct targets SOCS-1, SOCS-6, and SHIP-1 inhibit cytokine signaling, using multiple suppression mechanisms [
50,
51]. Another miR-155 target C/EBP-β promotes the expression of anti-inflammatory cytokines (including IL-10) and contributes to injury repair and neuroprotection [
52,
53]. In the inhibitor samples from 7 days after dMCAO, upregulation of SOCS-1 was accompanied by reduced STAT-3 phosphorylation. Cytokine signaling is mediated via essential JAK/STAT pathway, in which STAT phosphorylation is an indicator of cytokine signaling activation [
54]. SHIP-1 upregulation corresponds with dephosphorylation of its direct target protein Akt, demonstrated earlier [
26]. Together, these data indicate that at 7 days post-dMCAO, miR-155 inhibition potentially suppresses STAT-3- and PI-3K-mediated cytokine signaling via its direct targets SOCS-1 and SHIP-1 activity. We speculate that miR-155 inhibition-associated upregulation of SOCS-1 and SHIP-1 could suppress number of different cytokines, without affecting their mRNA or protein levels. Altered cytokine expression and signaling (as compared to control group) at 7 days after dMCAO was also accompanied by alterations in perivascular M/M phenotype. Electron microscopy imaging revealed that, in contrast to inhibitors, peri-vascular microglia in the control group was characterized by high phagocytic activity. Engulfment of blood vessels by phagocytic microglia results in the degradation of blood vessels and the active breakdown of the BBB. It is suggested that phagocytic microglia actively participate in the transfer of stroke-induced injury in healthy neighboring tissue by the disassembly of blood vessels and the resulting decrease in blood flow in the ischemic penumbra [
44]. Thus, present EM data are in good correlation with our previous findings that at 7 days post-dMCAO, 30 % in control animals had disrupted tight junctions, as opposed to only 9 % in the inhibitor group. In addition, at the same time point after dMCAO, cerebral blood flow in the peri-infarct area of the inhibitor animals was significantly higher than in controls [
26]. Immunofluorescence analysis of the brain sections from 7 days after dMCAO also confirmed decreased expression of leukocyte/ macrophage marker CD45 and active phagocytosis marker CD68 in the anti-miR-155-injected group. Taken together all these findings, we concluded that miR-155-induced suppressed cytokine signaling at 7 days, accompanied by decreased M/M phagocytic activity, could contribute to preservation of TJs observed at 7 days after dMCAO.
In the inhibitor group, expression of several important cytokines lasted for as long as 14 days after dMCAO (in contrast to a sharp decline after 7 days, detected in the control samples). Sustained increase in expression of IL-10 in the inhibitor-injected animals is in agreement with recent investigations demonstrating suppression of this cytokine by miR-155 [
55]. This major anti-inflammatory cytokine was shown to trigger anti-inflammatory response beneficial for stroke outcome [
41,
56]. Significantly higher (as compared to controls) expression of five other cytokines was also detected in the inhibitor group at 14 days after dMCAO. These cytokines exhibiting context-dependent dual action are implicated in having a significant impact on neuroprotection and overall stroke outcome. Il-4 (mostly regarded as anti-inflammatory cytokine) and IL-5 were found to play a beneficial role in brain repair, modulate microglial response, and suppress post-stroke inflammation [
57‐
59]. IL-6, apart from its well-known pro-inflammatory function, can also exhibit neurotrophic and regenerative features following cerebral ischemia [
60,
61]. MIP-1α, a member of the CC chemokine subfamily, is upregulated at early acute stages of cerebral ischemia and may have a role in promoting inflammatory and/or repair processes in ischemic brain [
62]; at later stages, this chemokine can also serve as a chemoattractant for stem cell migration after ischemic injury [
63]. The late (6 days after stroke) peak of IL-17 was detected in other studies, proposing a possible role of this cytokine in neovascularization after MCAO [
40]. High levels of IL-10 at 14 days after dMCAO correlated with and could be induced by sustained elevation of C/EBP-β detected at 7 and 14 days after dMCAO. In addition, increased IL-10 levels at 14 days could be associated with the reduced suppression from SOCS-1 and SHIP-1 [
55]. STAT-3 activation at this time point could be associated with (a) reduced suppression from SOCS-1 and (b) sustained upregulation of IL-10. IL-10/STAT-3 pathway activation triggers IL-10-mediated anti-inflammatory response (AIR): upon IL-10-induced activation, STAT-3 stimulates the expression of AIR factors, which specifically suppress pro-inflammatory cytokine signaling [
33,
64]. Based on the literature and our previous studies, miR-155 inhibition could potentiate IL-10/STAT-3-mediated AIR, which could significantly contribute to improved post-stroke recovery reported in our previous study [
26].
Since miR-155 is implicated in the regulation of macrophage differentiation and polarization toward anti-inflammatory phenotype, we expected that miR-155 inhibition would lead to an increased expression of the anti-inflammatory phenotype marker CD206. Instead, we detected altered expression of CD45 and CD68 markers. Increased CD45 expression in the inhibitor group (as compared to controls) at 14 days after stroke may be associated with prolonged upregulation of IL-10, since IL-10 was found to activate CD45 protein tyrosine phosphatase [
65]. CD45 is regarded as a negative regulator of pro-inflammatory microglia activation associated with MAPK signaling propagation and neuronal death [
66]. According to the literature, Iba-1/CD45-positive macrophages expressing active phagocytosis marker CD68 facilitate brain recovery process following LPS-induced brain injury [
67] and stroke [
68] and mediate neuroprotection in Alzheimer’s disease [
69]. The upregulation of CD68 and, thus, phagocytic activity at 14 days after dMCAO could facilitate removal of debris and dead tissue, as well as promote revascularization, neuroprotection, neurogenesis, and overall recovery [
70]. Based on the literature, ameboid Iba-1-positive M/Ms populating glial scar area are represented by macrophages, that infiltrated the site of the injury at early post-stroke stages, and resident microglia, that adopted macrophage-like morphology in response to ischemia. Following cerebral ischemia, initially, pro-inflammatory M/Ms undergo differentiation at later stages after stroke and acquire anti-inflammatory pro-regenerative features, as described in [
71]. Based on our data, we propose that miR-155 inhibition at 48 h after stroke results in suppression of early transient harmful actions of the activated M/Ms at 7 days, followed by enhancement of their protective and reparative actions at 14 days after dMCAO. In the present study, we did not aim to phenotypically distinguish between the resident microglia and monocyte-derived infiltrated macrophages, or different types of infiltrated leukocytes. This issue will be addressed in our upcoming studies involving Cx3cr1
GFP/+ /Ccr2
RFP/+ transgenic mice; this approach will minimize the artifacts associated with cell isolation/sorting and, thus, assure a truly reliable separation of two different cell types comprising the M/M population.
Acknowledgements
We would like to thank Dr. Kiran Bhaskar for sharing his expertise in microglia biology. Transmission EM imaging was performed in UNM HSC Electron Microscopy Facility. Confocal images were generated in the UNM Cancer Center Fluorescence Microscopy Facility.