Introduction

Stroke remains the leading cause of disability and the fifth leading cause of mortality in the USA. Ischemic stroke accounts for 80% of all strokes, and there is currently only one FDA-approved drug therapy, tissue plasminogen activator (tPA). tPA must be administered within a short time window of 3–4.5 h to be effective, which limits the proportion of patients eligible to receive this treatment (5.9–7.0%) [1, 2]. Therefore, there is an urgent need for the development of effective and safe therapeutics. Moreover, since stroke risk and stroke severity are modified by the age and biological sex of the patient, it is imperative for preclinical studies of potential stroke therapies to address both variables.

Epigenetic modifiers, such as small non-coding RNAs, have emerged as powerful candidates for several diseases including cancers and immune, infectious, cardiovascular, and neurodegenerative diseases [3,4,5,6]. MicroRNAs (miRNAs) are typically 18–25 nucleotides long and bind to complementary sequences in the 3′ UTR of multiple target mRNAs to regulate gene expression [7]. In the case of ischemic stroke, several studies have now shown that miRNA mimics or antagomirs can regulate acute and chronic stroke outcomes [8,9,10,11,12]. Previous work from our lab has shown that the efficacy of miRNA treatment may be restricted to specific age and sex groups. For example, anti-Let7f treatment after middle cerebral artery occlusion (MCAo) improved stroke outcomes in adult female rats but not older females or adult males [8, 13]. Similarly, miR-363-3p, identified by miRNA profiling of serum, modifies stroke outcomes only in females (adult and middle-aged) but not age-matched males [14]. In order to identify a more universally effective microRNA treatment, the present study used a novel approach by focusing on astrocytes.

Astrocytes are a crucial cell type for mediating brain energy homeostasis, providing neurotrophic support, and maintaining the blood–brain barrier [15]. Astrocytes have been shown to develop an “aging” phenotype, characterized by increased glial fibrillary acidic protein (GFAP) expression [16] and increased production of senescence-associated secretory phenotype (SASP) factors, such as interleukins (IL-1α [17], IL-6 [18, 19], IL-8 [20], IL-15 [21]) and matrix metalloproteases (MMP-1 [18], MMP-3, MMP-10 [20]). Astrocytes harvested from the ischemic cortex and striatum in middle-aged females show similar senescence-related changes, including reduced glutamate reuptake, decreased growth factor release, and increased chemokine release compared to adult females, all indicative of decreased neuroprotective capacity [22]. Furthermore, astrocytes from middle-aged females showed reduced trimethylation of histone H3K4, a transcriptional enhancer, as compared to adult females [23], indicating a global reduction in transcription potential, including the transcriptional start site for the miR-17a-92a cluster.

To assess the role of astrocytic miRNA in stroke outcomes, we compared miRNA profiles in astrocytes obtained from the ischemic forebrain of adult and middle-aged males and females. These groups were specifically selected because stroke outcomes are more severe in adult males and middle-aged males and females as compared to adult females, where stroke-induced infarction and sensory motor impairment is low. This approach confirmed the suppression of the miR-17–92 cluster observed in ChIP-seq analysis [23] and further identified miR-20a-3p as a uniquely age- and sex-regulated miRNA, whose expression was dramatically elevated in adult females as compared to all other groups. Bioinformatics identified mitochondrial and inflammation-associated genes as targets of miR-20a-3p, and select examples were confirmed in molecular and functional assays. A viral construct designed to conditionally elevated miR-20a-3p in astrocytes partially improved stroke outcomes. Subsequently, intravenous injection of fluorescently labeled miR-20a-3p after stroke was found to be preferentially taken up by neurons. A second viral construct was designed to conditionally express miR-20a-3p in neurons, resulting in robust neuroprotection. Finally, i.v. treatment of miR-20a-3p, a more translationally viable route of administration, was found to be neuroprotective in both middle-aged males and females. Together, these data indicate that while miR-20a-3p is produced by and profoundly alters the function of astrocytes after ischemia, stroke neuroprotection via this microRNA may depend on other cell types such as neurons receiving and utilizing miR-20a-3p.

Results

Expression of MicroRNA from the miR-17–92 Cluster in Astrocytes from Adult and Middle-Aged Males and Females

Astrocytes were extracted at 48 h post stroke from adult and middle-aged females and males using positive selection for GLAST. Based on our previous data showing increased trimethylation of the miR-17–92 cluster in adult females, we analyzed sex and age differences in this cluster by qRT-PCR (normalized to U6). Members of this miRNA cluster showed a complex regulation due to age and sex. As shown in Fig. 1a, miR-17-5p, miR-18a-5p, miR-19a, miR-19b, miR-20a-5p, and miR-20a-3p displayed a significant age × sex effects, such that each of these miRNAs was elevated in the adult female as compared to the adult male and middle-aged male or female. In the case of miR-17-3p, there were a main effect of age and a main effect of sex, while in the case of miR-92a-3p, there was a main effect of age, where adult animals had a higher expression of these miRNAs as compared to middle-aged animals. The most remarkable age × sex regulation was seen in miR-20a-3p expression, which demonstrated a > 240,000-fold elevation in adult female as compared to middle-aged females. A separate analysis of the entire miRnome panel controlling for false discovery rate (FDR) showed that miR-20a-3p was the only astrocytic miRNA that was significantly elevated by age and sex.

Fig. 1
figure 1

MiRnome analysis of miR expression from the miR-17–92 cluster. A histogram depicting the mean (± SEM) expression of microRNA from the miR-17–92 cluster in adult and middle-aged males and females. Bar above each set indicates main and interaction effects from two-way ANOVAs. a Main effect of age. b Main effect of sex. c Interaction effect (age × sex). B MiR-20a-3p expression in serum, whole brain, and astrocyte. Pre-stroke and 2-day post-stroke serum, brain, and astrocyte samples were evaluated for miR-20a expression. Key: asterisk, significantly different from pre stroke; number sign, significantly different from adult females at that time point. p < 0.05. n = 6 in most groups, n = 3 in pre-stroke astrocytes and post stroke brain samples. Red line, adult females; black line, middle-aged females

Full size image

miR-20a-3p Expression in the Circulation, Brain, and Astrocytes in Adult and Middle-Aged Females

To determine if miR-20a-3p suppression is a tissue-specific response to stroke, miR-20a-3p expression was analyzed using qRT-PCR in serum, whole-brain homogenate, and astrocytes from adult and middle-aged females in sham animals and at 2 days post stroke. While basal expression of miR-20a-3p is similar at both ages, miR-20a-3p is altered in an age- and tissue-specific manner (Fig. 1b) during the early acute phase (2 days) of stroke. In circulation (serum), miR-20a-3p is significantly elevated at 2 days post stroke in both adult and middle-aged females (Fig. 1b(i)). In whole-brain homogenate (Fig. 1b(ii)) and in astrocytes (Fig. 1b(iii)), miR-20a-3p expression is preferentially elevated in adult females (red line) and profoundly suppressed in middle-aged females (black line) at 2 days post stroke.

Effect of MiR-20a-3p Treatment on Astrocyte Mitochondrial Function In Vitro

Bioinformatics analyses (using databases TargetScan and MiRWalk 2.0) indicate that miR-20a-3p regulates a large number of genes that are responsible for mediating mitochondrial function (Supplementary Fig. 1). In order to investigate specific mechanisms by which miR-20a-3p alters mitochondrial function, male and female human astrocytes were cultured and subjected to 6 h of oxygen glucose deprivation (OGD) (1% O2, 0 mM glucose) or normoxia (21% O2, 25 mM glucose) and treated with miR-20a-3p mimic, scrambled miR, or vehicle. Astrocyte cultures were confirmed by immunohistochemistry for GFAP (green) and co-labeled with nuclear dye DAPI (blue) (Supplementary Fig. 2a). Cultures that were incubated with 50 nM FAM-labeled miR-20a-3p mimic and MitoTracker Deep Red to stain living mitochondria for 6 h showed that the microRNA is readily taken up by astrocytes (Fig. 2a). In addition, qRT-PCR analysis indicates that female human astrocytes profoundly upregulate miR-20a-3p in response to ischemia, whereas the male astrocytes do not (Fig. 2b), recapitulating our in vivo observations.

Fig. 2
figure 2

Effect of miR20a-3p mimic on mitochondrial dynamics. a Primary human astrocytes stained with MitoTracker Deep Red and incubated with FAM-labeled miR-20a-3p for 6 h. b qRT-PCR expression of miR-20a-30 in primary human astrocytes. c Calcein assay on female and male human astrocytes. Histogram depicting the mean (± SEM) percent of viable cells relative to normoxic conditions. d FRAP analysis of female and male human astrocytes. The recovery was normalized to the level of bleaching in the cells (bleaching value was set at 1). e Seahorse XFe96 Mito Stress Test on human astrocytes. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001

Full size image

To determine the effect of miR-20a-3p on astrocyte survival after OGD, cells were incubated with Calcein-AM dye (Life Technologies, CA) after OGD treatment. MiR-20a-3p treatment significantly increased cell viability in both male and female cells compared to scrambled miR treatment, but with different degrees of efficacy. In females, cell viability in the miR-20a-3p treatment group was virtually similar to that in normoxic controls, while cell viability in males that received miR-20a-3p was greater than that in cultures that received the scrambled oligo but was far lower than that in normoxic controls. Moreover, the female astrocytes demonstrated a greater degree of viability as compared to the male astrocytes after exposure to OGD (p = 0.002, Fig. 2c).

Two functional assays were used to explore the effect of miR-20a-3p treatment on astrocytic mitochondrial function. The first technique used was fluorescence recovery after photobleaching (FRAP), in which human astrocyte cultures were labeled with MitoTracker Deep Red and a small area of the cell was photobleached with a high intensity 405-nm laser and the subsequent fluorescent recovery was quantified using ImageJ. MiR-20a-3p treatment significantly increased recovery in both OGD and normoxic female groups immediately after bleaching (< 4 s), while the miR-20a-3p—treated male cells only exhibited significantly increased recovery in the OGD group after 18 s (Fig. 2d). Another functional assay used to assess mitochondrial dynamics is the Mito Stress Test using the Seahorse 96XFe Analyzer. This assay involves the serial injection of various drugs (oligomycin, FCCP, and rotenone + antimycin A) that target components of the electron transport chain. In normoxic conditions, female astrocytes show no difference between the miR-20a-3p—treated group and either of the control groups; however, the miR-20a-3p—treated male astrocytes show reduced oxygen consumption rate relative to both vehicle and scrambled oligo–treated cells (Fig. 2e). Moreover, miR-20a-3p treatment significantly reduces oxygen consumption rates in the male and female astrocytes in OGD conditions, though the male cells exhibit significantly more variability than the female cells (Fig. 2e). Increased oxygen consumption could be indicative of cell senescence induced by oxidative damage [24], and these data suggest that miR-20a-3p may reduce this senescent feature in an ischemic environment.

Astrocyte-Specific Enhancement of MiR-20a-3p Improves Select Stroke Outcomes in Middle-Aged Females

To assess the neuroprotective potential of astrocyte-derived miR-20a-3p in the context of stroke, a viral construct (rAAV5-TetOn-GFAP-miR-20a-3p-mCherry, SignaGen) under the control of a TetOn system with a GFAP promoter and conjugated with a mCherry reporter (Supplementary Figure 3a) and a control vector (Supplementary Figure 3b) was synthesized and injected into the left hemisphere of middle-aged female rat brains. Figure 3c demonstrates that the recombinant adeno-associated virus (rAAV) is activated by doxycycline and primarily co-localizes with GFAP. Two doses (2.5 × 1011 VP/ml; 2.5 × 1010 VP/ml) of the rAAV were used in pilot studies to determine the most effective dose to assess the effect of astrocyte-specific miR-20a-3p expression after stroke. Striatal injections of the high dose (2.5 × 1011 VP/ml) did not result in any adverse effects on body weight or spleen weight and were analyzed extensively for stroke outcomes (Fig. 3a–e). The Kaplan–Meier survival analysis (Fig. 3a) showed that the striatal injection of the rAAV containing miR-20a-3p reduced stroke-induced mortality (p = 0.0372). Infarct volume (Fig. 3b, c) assessed at 5 days after MCAo in animals that received the control vector was greater (53.42%) than that in animals that received the miR-20a-3p—containing vector (40.39%), although this difference did not meet the criteria for statistical significance (p = 0.1117). Sensory motor performance assessed by the adhesive removal test showed that both groups had increased latency at 2 days post stroke on the contralesional limb, while at 5 days post stroke, the miR-20a-3p group showed improved recovery and significantly reduced latency as compared to the control vector group (Fig. 3d). The vibrissae-evoked forelimb placement task did not demonstrate any improvement in the percentage of correct responses post MCAo when the same-side vibrissae were stimulated nor when the cross-midline vibrissae were stimulated (Supp. Fig. 4A).

Fig. 3
figure 3

Effect of rAAV-TetOn-GFAP-miR-20a-3p on stroke outcome in middle-aged females. A Kaplan–Meier survival plot shows significantly greater mortality in animals that received the control vector as compared to the miR-20a-3p–containing vector. B Representative images of TTC-stained sections obtained 5 days post MCAo. C Histogram depicting mean (± SEM) infarct volume. D Histogram of mean (± SEM) latency to remove the adhesive tape pre MCAo, 2 days post MCAo, and 5 days post MCAo. N = 12 (miR-20a-3p) and n = 11 (scrambled oligo) pre MCAo. *p ≤ 0.05, **p ≤ 0.01

Full size image

FAM-Labeled miR-20a-3p Localizes to Neurons Rather Than Astrocytes After MCAo

We next examined the efficacy of intravenously injected miR-20a-3p. First, a FAM-labeled miR-20a-3p mimic was injected via tail vein 4 h after MCAo or sham surgery. The FAM-labeled mimic was chosen in order to visualize the localization of the miR-20a-3p mimic after intravenous treatment. FAM-labeled miR-20a-3p (green) preferentially co-localized with NeuN+ cells rather than GFAP+ cells in the animals that received MCAo (Fig. 4a, b). The proportion of FAM-labeled miR-20a-3p expressed in cells (co-labeled with dapi) that co-localized with NeuN+ cells compared to GFAP+ cells confirmed that neurons preferentially uptake the FAM-labeled miR (Fig. 4e). Interestingly, in animals that received the sham surgery, this preferential uptake by neurons is not observed (Fig. 4c–e). These data indicate that neurons may uptake this miRNA more readily under ischemic conditions.

Fig. 4
figure 4

Cellular localization of i.v. miR-20a-3p. Photomicrographs of coronal sections from animals injected with FAM-labeled miR-20a-3p after MCAo or sham surgery probed for glial and neuronal markers and counterstained for DAPI (blue). A MCAo: GFAP (red). B MCAo: NeuN (red). C Sham: GFAP (red). D Sham: NeuN (red). Arrows indicate FAM-labeled miR-20a-3p (green). E Quantification of FAM-miR-20a-3p co-localization. **p ≤ 0.01

Full size image

Effect of MiR-20a-3p on Neuronal Mitochondrial Function In Vitro

In light of the cellular localization of i.v. FAM-miR-20a-3p injections, we decided to investigate the effects of miR-20a-3p on neuronal mitochondrial functions. Primary human neurons were cultured in identical conditions to the previously described astrocytes. Neuronal identify was confirmed via immunohistochemistry for NeuN (red) and co-labeled with the nuclear stain DAPI (Supplementary Fig. 2b). OGD was reduced to 30 min to account for greater cell death of the neurons relative to astrocytes in ischemic conditions. Confocal images indicate that the neurons uptake the FAM-labeled miR-20a-3p mimic after 30 min of OGD (Fig. 5a). qPCR analysis was performed to assess the expression of miR-20a-3p in normoxic and OGD conditions, and no differences in the expression of this microRNA were observed in either sex (Fig. 5b).

Fig. 5
figure 5

Effect of miR20a-3p mimic on neuronal mitochondrial dynamics. a Primary human neurons stained with MitoTracker Deep Red and incubated with FAM-labeled miR-20a-3p for 30 min. b qRT-PCR expression of miR-20a-30 in primary human neurons. c Calcein assay on female and male human neurons. Histogram depicting the mean (± SEM) percent of viable cells relative to normoxic conditions. d FRAP analysis of female and male human neurons. The recovery was normalized to the level of bleaching in the cells (bleaching value was set at 1). e Seahorse XFe96 Mito Stress Test on human neurons. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001

Full size image

A cell viability assay, FRAP analysis, and the Seahorse Mito Stress Test assay were performed to assess the effect of miR-20a-3p on neuronal mitochondrial function. The cell viability assay did not show any improvement of neuronal survival after 30 min of OGD, though the female neurons did demonstrate a greater degree of viability as compared to the male neurons (p = 0.0007, Fig. 5c); however, key functional differences were elucidated in the FRAP and Seahorse assays. In all cases, neurons treated with miR-20a-3p significantly improved fluorescent recovery after photobleaching (Fig. 5d), similar to what was observed in the astrocytes. The Seahorse assay, however, demonstrated differences between neurons and astrocytes regarding the effect of miR-20a-3p on oxygen consumption. In normoxic conditions, male neurons treated with miR-20a-3p showed significantly elevated levels of oxygen consumption rate relative to control groups (Fig. 5e). The female neurons treated with miR-20a-3p exhibited greater oxygen consumption relative to vehicle, but there was not a significant difference between miR-20a-3p-treated and scrambled miR–treated groups. OGD reduced the rate of oxygen consumption in both sexes, though the difference between the miR-20a-3p–treated groups and the vehicle groups is not significant (Fig. 5e).

Neuron-Specific Enhancement of MiR-20a-3p is Neuroprotective in Middle-Aged Female Rats After Stroke

A second viral construct (rAAV5-TetOn-NSE-miR-20a-3p mCherry, SignaGen) was created to assess the effect of conditional neuronal miR-20a-3p expression. rAAV striatal injections and endothelin-1 (ET-1) surgeries were performed as indicated previously, and all procedures were identical to the previous rAAV experiment. However, unlike the GFAP vector, the neuron-specific enolase (NSE) vector showed significant reduction of infarct volume in animals that received the rAAV vector (Fig. 6b, c, p = 0.0087) and more significant improvement in the sensory motor tasks at both 2 days and 5 days post stroke (Fig. 6d, p < 0.0001). While post-stroke mortality in the rAAV-NSE-miR-20a-3p was almost half that of the group receiving the control vector, the Kaplan–Meier survival analysis only indicated a statistical trend in survival (Fig. 6a, p = 0.0805).

Fig. 6
figure 6

Effect of rAAV-TetOn-NSE-miR-20a-3p on stroke outcome in middle-aged females. a Kaplan–Meier survival plot shows a trend toward greater mortality in animals that received the control vector as compared to the miR-20a-3p—containing vector. b Representative images of TTC-stained sections obtained 5 days post MCAo. c Histogram depicting mean (± SEM) infarct volume. d Histogram of mean (± SEM) latency to remove the adhesive tape pre MCAo, 2 days post MCAo, and 5 days post MCAo. N = 10 (miR-20a-3p) and n = 9 (scrambled oligo) pre MCAo. *p ≤ 0.01, ****p ≤ 0.0001

Full size image

MiR-20a-3p Treatment Significantly Reduced Infarct Volume in Middle-Aged Female and Male Rats

Although the rAAV experiments provide key insight into the cellular mechanism of neuroprotection of miR-20a-3p, i.v. treatment is much less invasive and therefore more translationally viable, so we next assessed the effect of i.v. miR-20a-3p on acute stroke outcomes. Animals were subjected to MCAo and injected via the tail vein with miR-20a-3p mimic either concurrent (immediate treatment (IT)) with the onset of ischemia, 4 h after ischemic onset (delayed treatment (DT)), or 24 h after the onset of ischemia (very delayed treatment (VDT)). As shown in Fig. 7a, triphenyl tetrazolium chloride (TTC)-stained images indicate that the cortico-striatal infarct is significantly reduced in middle-aged females in animals that received miR-20-3p 4 h after MCAo (DT) as compared to the age-matched counterparts that received i.v. injections of the scrambled oligos. Delayed miR-20a-3p treatment decreased infarct volume by 60% and improved performance on the adhesive removal test when measured 5 days post stroke (Fig. 7a). On the other hand, i.v. injections of miR-20a-3p–delivered concurrent with the onset of ischemia (IT) or 24 h after MCAo (VDT) did not show any decrease in infarct size or reduction in sensory motor deficits (Fig. 7b, c, Supp. Fig. 4c, d).

Fig. 7
figure 7

Effect of intravenous miR-20a-3p mimics on stroke outcomes in middle-aged females at IT, DT, and VDT: A(i)–C(i) TTC-stained coronal sections from middle-aged females treated with scrambled oligo or miR-20a-3p. a(ii)–c(ii) Histogram depicts average infarct volume (± SEM) normalized to the volume of the non-ischemic hemisphere. a(iii)–c(iii) Sensory motor performance on the adhesive removal test was evaluated before and after stroke. Histograms depict mean (± SEM) latency in seconds to remove the tape. DT, N = 6 (control or treatment); IT, N = 7 (control) and N = 8 (treatment); VDT, N = 8 (control) and N = 8 (treatment). *p ≤ 0.05, **p ≤ 0.01

Full size image

Middle-aged males (a group that also sustains severe stroke outcomes) showed significantly reduced infarct following delayed miR-20a-3p mimic treatment (Supplementary Fig. 3A, B; p < 0.05) and significantly improved performance on the adhesive removal test, which is impaired after MCAo (Supplementary Fig. 3C). Collectively, these data indicate that, similar to middle-aged female rats, i.v. miR-20a-3p mimic to middle-aged male rats improved stroke outcomes.

Effect of DT MiR-20a-3p Treatment on Gene Targets Post MCAo

To assess whether i.v. treatment of miR-20a-3p affected astrocyte gene expression, astrocytes were isolated from the ischemic hemisphere for qRT-PCR analysis. In view of the improved stroke outcomes in the DT group, the analysis was only performed in this group. Several matrix metalloproteinases (MMPs) and mitochondrial gene targets of miR-20a-3p demonstrated altered regulation in the ischemic hemisphere of the animals that received DT as compared to the animal that received scrambled treatment. Gene expression of MMP-2, MMP-9, and MMP-14 was elevated in the ischemic hemisphere of animals that received scrambled oligo treatment, while this elevation was mitigated in the ischemic hemisphere of animals receiving miR-20a-3p (Fig. 8a–c, *p < 0.05). MMP activity was assessed by gelatin zymography. Protein was isolated from the ischemic hemisphere of animals subjected to MCAo that received either miR-20a-3p mimic or scrambled oligo (DT). Lytic activity was noted in the region between 50 and 70 kDa (Fig. 8e), which contains MMP-2 and MMP-14. This region was quantified for densitometry by ImageJ and revealed a significant reduction in MMP activity in the miR-20a-3p—treated animals (Fig. 8f). MMPs have been implicated in the blood–brain barrier (BBB) and cerebral microvasculature damage as well as hemorrhagic bleeds [25]. BBB damage was assessed by ELISA for serum GFAP, which is a surrogate measure of BBB leakiness. As shown in Fig. 8g, there was a trend (p = 0.12) toward decreased GFAP levels in the miR-20a-3p—treated groups. Hemorrhagic foci were inspected in TTC-stained sections (shown in Fig. 8h) by a code-blinded observer and were found to be present in all (100%) scrambled oligo-treated middle-aged female rats and in one-third of the miR-20a-3p—treated rats (chi-square; p < 0.0143), generally in the vicinity of striatal vessels. Collectively, these data suggest that miR-20a-3p treatment 4 h after MCAo may protect the brain against MMP-mediated BBB damage and subsequent hemorrhagic transformation.

Fig. 8
figure 8

Effect of intravenous miR-20a-3p on markers blood–brain barrier permeability. a–d qPCR analysis of MMPs from the ischemic and non-ischemic hemispheres of scrambled and miR-20a-3p–treated animals. e Representative image of gelatin Zymogram depicting MMP activity from protein lysates from the ischemic hemisphere. f Histogram showing the mean (±) of MMP activity from gelatin zymography. g ELISA analysis of GFAP expression in serum. h Representative TTC images of brain slices with hemorrhagic foci from females treated with miR-20a-3p mimic and scrambled oligo. Foci are indicated by white arrows. N = 6 (miR-20a-3p) and n = 6 (scrambled oligo). #p < 0.1, *p ≤ 0.05

Full size image

IL-17A is a pro-inflammatory cytokine, which is a predicted target of miR-20a-3p and a therapeutic target for anti-inflammatory drugs to improve recovery post stroke [26]. Similar to the MMPs, IL-17A was elevated in the ischemic hemisphere of animals that received scrambled oligo treatment, while this elevation was mitigated in the ischemic hemisphere of animals receiving miR-20a-3p (Fig. 9a). This was further confirmed by ELISA assay, showing that IL-17A expression was significantly reduced in the ischemic hemisphere of the miR-20a-3p–treated middle-aged females as compared to the scrambled controls (Fig. 9b, p ≤ 0.05). Circulating (serum) levels of IL-17A were not different in the two groups (data not shown), suggesting that miR-20a-3p either has a brain-specific effect for this cytokine or regulates the blood–brain barrier to prevent extravasation of IL-17A or IL-17A–producing cells.

Fig. 9
figure 9

Effect of intravenous miR-20a-3p on predicted mitochondrial targets. a qPCR and b ELISA analysis of IL-17A expression from the ischemic and non-ischemic hemispheres of scrambled and miR-20a-3p–treated animals. c–e qPCR analysis of predicted mitochondrial genes from the ischemic and non-ischemic hemispheres of scrambled and miR-20a-3p–treated animals. f Protein from the ischemic hemisphere of scrambled and miR-20a-3p–treated animals probed for Drp1 by western blot and total protein stain. g Histogram of the mean (± SEM) of Drp1 normalized to total protein. N = 6 (miR-20a-3p) and n = 6 (scrambled oligo). *p ≤ 0.05, **p ≤ 0.01

Full size image

Mitochondrial gene targets including Uqcfs1, Ndufa1, and Atp5s were reduced in the ischemic hemisphere of the scrambled oligo–treated animals, while there were no differences in the expression of these genes in the two hemispheres of the miR-20a-3p–treated animals (Fig. 9c–e), indicating that miR-20a-3p treatment preserved the expression of mitochondrial genes. Quantitative western blot analysis was performed on protein isolated from the ischemic hemisphere of animals subjected to MCAo that received DT of either miR-20a-3p mimic or scrambled oligo. The blot was probed with a Drp1 (mitochondrial fission protein) antibody (Fig. 9f) and normalized to the LiCOR Total Protein Stain. MiR-20a-3p–treated animals demonstrated significantly less expression of Drp1 as compared to scrambled oligo–treated animals (Fig. 9g).

Discussion

The present study used well-known age and sex differences in stroke outcomes to identify microRNA with neuroprotective potential. We observed that the miR-17–92 cluster, which has been shown to improve neural plasticity after stroke in male rats/mice [Protein Analysis

Immunohistochemistry

Immunofluorescence for NeuN, ALDH1L1, and GFAP were performed on 30-µm brain sections mounted on glass slides. The sections were incubated with a blocking buffer (5% bovine serum albumin, 0.1% Triton X-100 in PBS, pH 7.4) for 1 h at room temperature. Sections were then incubated overnight at 4 °C with primary antibody (NeuN: anti-mouse [EMD Millipore], 1:250 µl; GFAP: anti-rabbit [Sigma-Aldrich], 1:3000 µl; Iba1: anti-rabbit [Wako Chemicals], 1:500 µl). Secondary antibodies (Alexa Fluor 488 and 594 anti-rabbit and anti-mouse) were then used at a 1:500 µl dilution at room temperature for 1 h. The sections were then washed thrice with PBS and then cover slipped with mounting media containing the nuclear dye DAPI (Fluoroshield, Abcam). Sections were visualized and imaged using an FV12-IX83 confocal microscope.

Protein Extraction

Cell proteins from the ischemic hemisphere (cortex and striatum) from animals terminated at 48 h were harvested and lysed in RIPA lysis buffer (Thermo Scientific, Grand Island, NY) and centrifuged at 20,000 rpm for 30 min. Supernatant was collected and stored at − 20 °C until further analysis. Protein concentrations were determined using the BCA protein assay kit (Pierce, Rockford, IL).

Western Blots

Protein extract (30 µg) from the ischemic hemisphere was loaded into 4–12% Novex gels and run at 60 V for 30 min and then at 100 V for 90 min. The protein was then transferred onto a PDVF membrane and probed with an antibody for Drp1 (anti-mouse, Abcam, 1:1000 µl) and conjugated to a fluorescent secondary antibody (LiCOR goat anti-mouse IRDye 680RD), normalizing to total protein (LiCOR Total Protein Stain).

Zymography

Protein extract (25 µg) from the ischemic hemisphere of middle-aged female rats post MCAo was loaded into a 10% Novex Zymogram Plus (gelatin) gel and run at 60 V for 30 min and then at 100 V for 90 min in a non-reducing sample buffer. The gel was then washed and incubated for 24 h at 37 °C. The gel was then stained with Coomassie blue for 30 min and then destained until clear bands could be seen.

IL-17a Assay

IL-17A expression was measured using the Rat IL-17A Platinum ELISA (Thermo Scientific, Grand Island, NY) according to the manufacturer’s instructions. Briefly, standards, controls, and aliquots of serum and protein lysates from ischemic cortex and striatum samples were loaded into a 96-well plate pre-coated with antibodies specific for IL-17A and followed by the addition of 100 µl of biotin-conjugated anti-rat IL-17A antibody and incubated at room temperature for 2 h on an orbital microplate shaker at 400 rpm for 30 s. With intervening washes, plates were sequentially incubated with 100 µl of streptavidin-HRP for 2 h, and 100 µl of TMB substrate solution for 30 min. The color reaction was stopped by an equal volume of stop solution and read at 450 nm in a microplate reader (Tecan, Switzerland). Standard curves were established from optical densities of wells containing known dilutions of the standard (1.6–100 pg/ml), and sample measurements were interpolated from standard curves. The experiments were performed in duplicates.

In Vitro Studies

Cell Culture and OGD

Human astrocytes and neurons from male and female donors of 18–20 gestational weeks of age were purchased from ScienCell Research Laboratories, grown in “Neural Growth Media (NGM)”, consisting of Neurobasal media with 2% B-27 supplement, 2% heat-inactivated gelded horse serum, 1% GlutaMAX, 1% penicillin, 0.1% ascorbic acid, 0.05% ampicillin, and 0.05% kanamycin, and plated in T-25 or T-75 cell culture-treated flasks. For experiments, cells were plated in poly-d-lysine—coated 6- or 96-well plates or in glass-bottom culture dishes in densities appropriate for the assay, which were ascertained via cell titration. Cultures were grown in normoxic conditions (5% CO2 and 21% O2, 37 °C) until confluent. Astrocytes were then subject to OGD (1% O2, 95% N2, and 5% CO2 in glucose-free DMEM, 37 °C) for 6 h with miR-20a-3p mimic (50 nM), scrambled miRNA (50 nM), or vehicle (PBS). Neurons were subjected to the same OGD conditions for 30 min. Culture media was collected, and cells were used for assays or processed for RNA isolation.

Calcein Assay

Cells were seeded at a density of 2 × 104 in a 96-well plate and subjected to OGD and treatment conditions for 6 h. Cell viability was determined using the Calcein-AM dye (Life Technologies, CA). After OGD, cells were incubated with Calcein-AM (2.5 µm) in Calcein-AM buffer for 20 min at 37 °C, and fluorescence was measured on a plate reader (Tecan, Switzerland) with excitation/emission set at 480 nm and 530 nm, respectively.

FRAP Analysis

Cells were seeded at a density of 1 × 105 into 2-ml glass-bottom culture dishes, cultured for 2 days, and then subjected to OGD and treatment conditions for 6 h. Culture media was collected, and cells were washed twice with gas-free recording buffer. Cells were then incubated with 120 nM MitoTracker Deep Red in gas-free recording buffer (154 mM NaCl, 5 mM KCl, 2 mM CaCl2·H2O, 0.5 mM MgCl2·H2O, 5 mM d-glucose, 10 mM HEPES) for 1 h. After incubation, cells were washed twice with gas-free recording buffer and imaged using confocal microscopy (FV12-IX83). Individual cells with MitoTracker Deep Red labeling were identified, and a region of interest containing a dense area of mitochondria close to the soma was selected. Pre-activation was recorded for 30 s, cells were bleached with high-intensity laser (405 nm, 80% power of total laser output) for 3 s, and subsequent activity was recorded for 90 s post bleach. Fluorescence intensity over time was plotted, and cells that did not achieve at least 75% bleaching were excluded from analysis.

Mitochondrial Respiration

Cells were seeded at a density of 4 × 104 in 96-well Seahorse microplates 15 h before OGD. After OGD, cells were washed twice with Seahorse XF DMEM medium (pH 7.4) and incubated for 45 min in a CO2-free incubator. The Mito Stress Test was then performed according to the manufacturer’s instructions. Briefly, this includes taking basal oxygen consumption measurements and then oxygen consumption measurements on the Seahorse XFe96 Analyzer after serial injections of oligomycin, FCCP, and rotenone/antimycin A.

Statistical Analysis

Power analysis for group sizes was computed based on effect sizes seen in previous data and pilot studies. In order to achieve a power of 0.9 (1 − β) and type 1 error rate (α = 0.05), the minimum sample size is 5. For these studies, group sizes ranged from 5 to 10. For comparisons between adult and middle-aged males and females on miR-20a-3p expression, a two-way ANOVA with post hoc comparisons was performed. For behavioral tests, a two-way ANOVA coded for repeated measure was used for each group, comparing the values obtained pre and post stroke. For all other tests, an unpaired t test was performed. Group differences were considered significant at p < 0.05 in each case. All in vitro assays were conducted with 3–5 replicate runs, and each run consisted of 3–12 technical replicates. Statistical analysis was performed using unpaired t tests for the Calcein assay, multiple t tests (one per time point) for FRAP assay, and 2-way ANOVA with post hoc comparisons for Seahorse assay. All statistical analyses were performed using GraphPad Prism software (v. 9.0).