Bioinformatic Exploration of Circulating microRNAs Related to Functional Outcomes in Patients With Acute Ischemic Stroke: An Exploratory Prospective Study

Background Although epigenetic modifications have been expected to play an important role in neuroplasticity for stroke recovery, the role of dynamic microRNA (miRNA) regulation related to functional outcomes after ischemic stroke remains unclear. Therefore, the current study performed a comprehensive miRNA expression analysis in serum to identify specifically altered circulating miRNAs associated with different grades of functional outcomes in patients with acute ischemic stroke (AIS). Methods Twelve patients with AIS in the middle cerebral artery region were included in this study. Peripheral blood samples were collected from patients one or two days after hospitalization. Total RNA, including small RNAs, was extracted from 400 µL of serum, and comprehensive miRNA expression analysis was performed to identify specifically altered circulating miRNAs associated with different grades of functional outcomes. Functional outcomes were evaluated three months after stroke onset using the modified Rankin Scale (mRS), classified as favorable (mRS score of 0 or 1) or unfavorable (mRS score of 2 to 5). Differentially expressed miRNAs were analyzed using the DESeq2 package. Target genes of the miRNAs were explored using miRTargetLink 2.0. Results Acute miRNA expression dynamics were characterized by differences in the patients’ functional outcomes following ischemic stroke. The favorable outcome group exhibited significantly downregulated miRNAs, including hsa-miR-218-1, hsa-miR-218-2, hsa-miR-320e, hsa-miR-320d-1, hsa-miR-320d-2, hsa-miR-326, and hsa-miR-4429. In addition, 15 miRNAs, including hsa-miR-223, hsa-miR-18a, hsa-miR-411, and hsa-miR-128-1, were significantly upregulated in the favorable outcome group compared to the unfavorable outcome group. Interesting and strong validated networks between miRNAs and their target genes were identified. Conclusion This study identified specifically altered circulating miRNAs in serum associated with varying grades of functional outcomes in AIS patients and explored miRNA-target gene networks that might contribute to these outcomes. Although further studies are needed, this study highlights their potential role as biomarkers for predicting functional outcomes in patients with AIS.


Introduction
Ischemic strokes account for approximately 70% of all strokes [1].Occlusion or stenosis of the middle cerebral artery caused by atherothrombotic or cardioembolic events may lead to various residual disabilities.Although the development of acute revascularization treatments such as recombinant tissue plasminogen activator and mechanical thrombectomy has dramatically improved the prognosis of patients with major cerebral artery lesions [2], residual disabilities can often persist.Further advancements in treatment outcomes are essential to enhance patients' quality of life and minimize social security costs.
Rehabilitation is critical for improving daily functional independence and social participation among patients with confirmed brain infarction lesions and persistent functional impairment or disability, even after revascularization treatment [2].During the rehabilitation process, the degree of functional recovery can vary considerably between patients who respond to rehabilitative treatment and those who do not [3].Previous studies have suggested that factors such as the severity of neurological dysfunction, time elapsed from stroke onset, and the volume of the ischemic lesion can affect functional recovery [4].At the molecular level, Li et al. found that glial growth factor 2 promoted neural functional recovery after ischemic stroke [5].Although numerous molecules associated with functional recovery after ischemic stroke have been identified [5], the molecular dynamics related to the functional outcomes of ischemic stroke are not yet fully understood.
Epigenetic modifications have been recognized as important gene expression regulators and are expected to play a key role in neuroplasticity for stroke recovery [6].Among the various epigenetic modifications, microRNAs (miRNAs), which are short non-coding RNAs approximately 23 nucleotides in length, are considered important diagnostic or prognostic biomarkers because they are easily detectable and remain relatively stable in peripheral blood [7].A previous study compared miRNA expression between acute ischemic stroke (AIS) patients and healthy controls (i.e., those who did not experience a stroke) to examine its role as a diagnostic marker and investigate molecular changes following stroke onset [8].However, the role of dynamic miRNA regulation related to functional outcomes after ischemic stroke remains unclear due to the lack of studies that compare miRNA expression by different functional outcomes in AIS patients with similar initial stroke severity.Additionally, previous studies have been limited by the fact that they evaluated only selected miRNAs [8], and no study to date has conducted a comprehensive miRNA expression analysis in the context of functional outcomes following ischemic stroke.Therefore, the current study carried out a comprehensive miRNA expression analysis using next-generation sequencing in serum to identify specifically altered circulating miRNAs associated with different grades of functional outcomes in patients with AIS.

Patient recruitment and blood sample collection
This study included patients diagnosed with AIS and admitted for treatment to the Department of Neurosurgery, JA Onomichi General Hospital.The inclusion criteria were as follows: (1) ischemic stroke (atherothrombotic or cardioembolic) in the middle cerebral artery region for the first time; (2) Diffusion-Weighted Imaging-Alberta Stroke Program Early Computed Tomography Scores (DWI-ASPECTS) ranging from 5 to 10 (established ischemic lesion after acute revascularization treatment); and (3) age at the time of consent ranging between 20 and 85 years.The exclusion criteria were as follows: (1) Functional Ambulation Categories (FAC) of 3 or less prior to stroke onset or (2) inability to conduct standard neurosurgical treatment or rehabilitation due to serious medical complications during hospitalization.All treatments, including acute revascularization and rehabilitation, were performed in accordance with the Japan Stroke Society Guideline 2021 for the Treatment of Stroke [2].After acute management at JA Onomichi General Hospital, the patients were transferred to a convalescent rehabilitation hospital as appropriate for continued rehabilitation.
Peripheral blood samples were collected in serum tubes from patients one or two days after hospitalization.The samples were then allowed to clot at room temperature for at least 30 minutes and, after centrifugation at 3500 rpm for 15 minutes at 4°C, were aliquoted into microtubes and frozen at −20°C until required for analysis.
This study was approved by the ethical committees of Hiroshima Cosmopolitan University (approval number: 2020008) and JA Onomichi General Hospital (approval number: OJH-202192).Written informed consent was obtained from all patients, and the study was conducted in accordance with the Declaration of Helsinki.

Ischemic lesion evaluation and functional assessment
Evaluation of established ischemic lesions was carried out by an expert vascular neurosurgeon (M.A.) blinded to miRNA expression, using the DWI-ASPECTS.Patients who underwent acute revascularization therapy to minimize ischemic lesions were evaluated with the score after completion of treatment.Imaging data was acquired using a 3-T MRI scanner (GE Healthcare Japan).
Functional stroke severity upon acute rehabilitation admission was evaluated by a nurse blinded to the patient's DWI-ASPECTS score and miRNA expression, using the National Institutes of Health Stroke Scale (NIHSS).Functional outcome evaluation was conducted by an expert vascular neurosurgeon (M.A.) blinded to miRNA expression during the patient's follow-up appointment at JA Onomichi General Hospital three months after stroke onset, using the modified Rankin Scale (mRS).The mRS at three months has generally been used as the functional outcome following AIS [9,10].The functional outcomes were classified as favorable (mRS score of 0 or 1) or unfavorable (mRS score of 2 to 5), in accordance with previous studies [9,10].
The MicroRNA Isolation Kit (BioChain) was used to extract and purify total RNA, including small RNAs, from 400 µL of serum, in accordance with the manufacturer's protocol.The quality and quantity of the extracted small RNAs were assessed using NanoDrop (Thermo Fisher Scientific, Waltham, Massachusetts), the QuantiFluor RNA System (Promega, Madison, Wisconsin), and the Agilent Small RNA kit of the BioAnalyzer 2100 System (Agilent Technologies, Santa Clara, California), as per the manufacturer's instructions.RNA integrity number ≥ 7 and A260/A280 and A260/230 ratios ≥ 1.8 were set as the quality check criteria.Based on miRNA concentration, 10 of the 12 samples were selected for preparation of a small RNA library for nextgeneration sequencing analysis, and the remaining 2 samples were excluded from further analysis.

Identification of differentially expressed miRNAs and exploration of their target genes
Differentially expressed miRNAs (DEMs) were analyzed using the DESeq2 package (version 1.20.0) to identify miRNAs that were uniquely altered in the favorable outcome group compared to the unfavorable outcome group.The thresholds for significant DEMs were set at |log2FC| >1.5 and P value <0.05.Significant DEMs with log2FC >1.5 or log2FC <−1.5 were categorized as upregulated or downregulated DEMs, respectively, in accordance with a previous study [7].We used Shinyheatmap (http://shinyheatmap.com/) to display the heatmap of DEMs.Target genes of the miRNAs that have been reported to be associated with functional recovery after stroke in previous studies were explored using miRTargetLink 2.0 (https://ccbcompute.cs.uni-saarland.de/mirtargetlink2).We selected "strong validated only" when presenting the interaction between miRNAs and target genes using miRTargetLink 2.0.

Functional enrichment analysis
Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) (https://david.ncifcrf.gov/)to elucidate the biological implications and functions of the significantly downregulated or upregulated DEMs.Statistical significance was set at P value < 0.05.We applied the Benjamini-Hochberg procedure to adjust P values.GO enrichment analysis included three domains: biological process, cellular component, and molecular function.We used Metascape (https://metascape.org/) to display the bar graph of enriched terms.

Statistical analysis
All statistical analyses were performed using JSTAT (Sato, Japan) and R (version 3.6.1,R Foundation, Vienna, Austria) software.The Mann-Whitney U test was used for the comparison of age, DWI-ASPECTS, and NIHSS upon rehabilitation admission, while Fisher's exact test was used to compare sex, ischemic stroke type, and acute revascularization treatment between groups.A P-value <0.05 was considered statistically significant.

Patient demographics
Of the 12 serum samples collected from patients with AIS, 2 were excluded from the sequencing analysis due to low miRNA concentrations.Of the 10 patients, 4 were classified as the unfavorable outcome group (2 patients with mRS 3 and 2 patients with mRS 2) and 6 were classified as the favorable outcome group (3 patients with mRS 0 and 3 patients with mRS 1) according to the mRS at 3 months after stroke onset.The mean age of the patients did not significantly differ between the groups (unfavorable outcome group: 77.75±7.93years; favorable outcome group: 76.83±4.83years; Table 1).All patients in the unfavorable outcome group were female, while 4 out of 6 patients in the favorable outcome group were female.There were no significant differences in ischemic stroke type between the groups (unfavorable outcome group: 2 patients with atherothrombotic and 2 with cardioembolic; favorable outcome group: 1 patient with atherothrombotic and 5 with cardioembolic) or in the acute revascularization treatment.Established ischemic lesions (DWI-ASPECTS scores; unfavorable outcome group: 7.25±1.71;favorable outcome group: 8.33±1.86)and functional stroke severity upon acute rehabilitation admission (NIHSS scores; unfavorable outcome group: 9.50±2.38;favorable outcome group: 8.67±6.62)also did not differ significantly between the groups.

Identification of differentially expressed miRNAs and experimentally validated target genes
Sequencing analysis confirmed the expression of 997 miRNAs out of a total of 1,879 miRNAs registered in the miRBase.The miRNA expression profiles differed significantly between the unfavorable and favorable outcome groups.Comparison of miRNA expression between groups showed that 35 miRNAs, including hsa-miR-218-1, hsa-miR-218-2, hsa-miR-320e, hsa-miR-320d-1, hsa-miR-320d-2, hsa-miR-326, and hsa-miR-4429, were significantly downregulated in the favorable outcome group compared to the unfavorable outcome group (Figure 1 and Table 2).Additionally, 15 miRNAs, including hsa-miR-223, hsa-miR-18a, hsa-miR-411, and hsa-miR-128-1, were significantly upregulated in the favorable outcome group compared to the unfavorable outcome group (Figure 1 and Table 3).Target genes of downregulated or upregulated DEMs that have been reported to be associated with functional recovery after stroke in previous studies were explored.The strongly validated networks between miRNAs and their target genes are represented in Figures 2, 3.

FIGURE 1: The heatmap of differentially expressed miRNAs between favorable outcome group and unfavorable outcome group
35 miRNAs from hsa-miR-4516 to hsa-miR-3180-5 were identified as downregulated miRNAs and 15 miRNAs from hsa-miR-548j to hsa-miR-409 were identified as upregulated miRNAs in favorable outcome group.

Functional enrichment analysis
Figure 4A shows the bar graph of enriched terms for 35 downregulated miRNAs, and Figure 4B shows those for 15 upregulated miRNAs via Metascape.Enrichment analysis was performed by annotating the categories "GOTERM" and "KEGG_PATHWAY" via DAVID to understand the biological implications of the 35 downregulated (Table 4) and 15 upregulated DEMs (Table 5).Both downregulated and upregulated DEMs were commonly enriched in "gene silencing by miRNA," "miRNA-mediated inhibition of translation," "mRNA cleavage involved in gene silencing by miRNA," "RISC complex," "mRNA binding involved in posttranscriptional gene silencing," and "mRNA 3'-UTR binding" in the GOTERM category and "MicroRNAs in cancer" in the KEGG_PATHWAY category.In the GOTERM category, downregulated DEMs were specifically enriched in "positive regulation of stem cell proliferation," "long-term synaptic potentiation," "cellular response to glucose stimulus," and "extracellular vesicle," while upregulated DEMs were specifically enriched in "negative regulation of epithelial to mesenchymal transition" and "extracellular space."

Discussion
The current study carried out a comprehensive miRNA expression analysis using miRNA sequencing to explore specifically altered circulating miRNAs by functional outcomes in patients with AIS.Additionally, miRNA-target gene networks that might be associated with functional outcomes were explored.The present findings suggest that acute miRNA expression dynamics are characterized by differences in the patient's functional outcomes following an ischemic stroke.
Liang et al. investigated the role of miR-320 in cerebral ischemia/reperfusion (I/R) injury using a mouse model and oxygen-glucose deprivation (OGD)-treated PC12 cells [13].Interestingly, miR-320 was found to enhance cell apoptosis in the cortical infarcted peripheral zone and increase brain infarction or edema volume in the mouse model, while miR-320 inhibitors decreased the risk of these negative events [13].The miRTargetLink 2.0 analysis showed that hsa-miR-326 had networks with FGF1, SP1, and GLI1.FGF1 and SP1 are known as effective neuroprotective factors, including their role in reducing brain edema [14,15].Additionally, Xiao et al. reported that GLI1 is related to axon guidance and neurogenesis in acute ischemic stroke model mice [16].Although the role of miR-326 in ischemic brain injury remains unclear, this study identified the hsa-miR-326-FGF1 network, hsa-miR-326-SP1 network, and hsa-miR-326-GLI1 network, which may be associated with functional outcomes after ischemic stroke.The current study identified specifically downregulated miRNAs with dynamics that are consistent with the biological roles reported in previous studies, suggesting that they could be involved in functional recovery after ischemic brain pathogenesis and, consequently, contribute to better functional outcomes.
The favorable outcome group also exhibited 15 miRNAs, including hsa-miR-223, hsa-miR-18a, hsa-miR-411, and hsa-miR-128-1, that were significantly upregulated, and these DEMs were significantly enriched in "negative regulation of epithelial to mesenchymal transition" and "extracellular space."Interestingly, all the upregulated miRNAs confirmed to be associated with functional recovery from stroke based on the findings of previous studies [17][18][19][20] had target genes related to functional recovery.Harraz et al. found that overexpression of miR-223 lowered GluR2 and NR2B levels, inhibited NMDA-induced calcium influx in hippocampal neurons, and protected the brain from neuronal cell death after transient global ischemia and excitotoxic injury [17].The miRTargetLink 2.0 analysis showed that hsa-miR-223-3p had networks with FOXO1, NLRP3, HSP90B1, STAT3, PARP1, RHOB, and SEMA3A.Inhibition of FOXO1, NLRP3, HSP90, or STAT3 has been reported to promote functional recovery after brain injury [21][22][23][24].Additionally, inhibition of PARP1 has neuroprotective effects, such as reducing infarct volume and brain swelling [25], and RHOB impacts neuronal death in transient ischemic model mice [26].Moreover, Hira et al. reported that inhibition of SEMA3A promotes axonal elongation [27].Zhang et al. demonstrated that miR-18a is one of the more enriched miRNAs, compared to other members of the miR-17-92 cluster, in the distal axons of sympathetic neurons, implying that it may regulate genes associated with axonal outgrowth [18].Regarding the hsa-miR-18a-target gene network, the hsa-miR-18a-3p-CBX7 network and hsa-18a-5p-PTEN network were identified.Interestingly, inhibition of PTEN or CBX7 has been related to the improvement of motor function or cognitive function in cerebral ischemic model rats, respectively [28,29].
Gong et al. conducted a study investigating the role of miR-411 using OGD-treated PC12 cells and a spinal cord injury rat model [19].They found that a miR-411 mimic could dramatically reduce the increased percentage of apoptotic cells caused by OGD and significantly elevate neurological function in the spinal cord injury model rat.The miRTargetLink 2.0 analysis showed that hsa-miR-411-5p had a network with SPRY4, and siRNA-mediated downregulation of SPRY2/4 diminished ischemic brain injury [30].Karam et al. highlighted the possibility of miR-128 promoting the maturation of neural precursor cells that were already committed to the neural fate [20].It has also been suggested that the miR-128/nonsense-mediated mRNA decay circuit could directly impact synaptic plasticity, thus playing a significant role in processes such as cognition and memory [20].Regarding the hsa-miR-128-1-5p-target gene network, the hsa-miR-128-1-5p-E2F3 network and hsa-miR-128-1-5p-WNT3A network were identified.It has been reported that downregulation of E2F3 decreased neuronal cell death [31].Additionally, Ma et al. suggested that the downregulation of WNT3A was associated with functional recovery in ischemic brain injury model rats [32].The specific upregulated miRNA dynamics observed in the favorable outcome group could potentially reflect the biological processes related to functional outcomes following ischemic stroke, suggesting their potential role as biomarkers for the prediction of functional outcomes in patients.

FIGURE 2 :
FIGURE 2: Downregulated miRNA-target gene networkInteractions between downregulated miRNAs and strong validated target genes are displayed.

FIGURE 3 :
FIGURE 3: Upregulated miRNA-target gene networkInteractions between upregulated miRNAs and strong validated target genes are displayed.

FIGURE 4 :
FIGURE 4: Bar graph of enriched terms via the Metascape Bar graph of enriched terms of 35 downregulated miRNAs (A) and 15 upregulated miRNAs (B).

TABLE 1 : Baseline characteristics of subjects
DWI-ASPECTS, Diffusion-Weighted Imaging-Alberta Stroke Program Early Computed Tomography Score; DWI, diffusion-weighted imaging; MT, mechanical thrombectomy; NIHSS, National Institutes of Health Stroke Scale; SD, standard deviation; t-PA, tissue plasminogen activator.*The Mann-Whitney U test was applied.**The Fisher's Exact test was applied.

TABLE 4 : Enrichment analysis for the downregulated miRNAs in the favorable outcome group compared to the unfavorable outcome group BP
, biological process; CC, cellular component; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function.

TABLE 5 : Enrichment analysis for the upregulated miRNAs in the favorable outcome group compared to the unfavorable outcome group BP
, biological process; CC, cellular component; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function.