MicroRNA Analysis of Human Stroke Brain Tissue Resected during Decompressive Craniectomy/Stroke-Ectomy Surgery

Abstract

Background: Signaling pathways mediated by microRNAs (miRNAs) have been identified as one of the mechanisms that regulate stroke progression and recovery. Recent investigations using stroke patient blood and cerebrospinal fluid (CSF) demonstrated disease-specific alterations in miRNA expression. In this study, for the first time, we investigated miRNA expression signatures in freshly removed human stroke brain tissue.

Methods: Human brain samples were obtained during craniectomy and brain tissue resection in severe stroke patients with life-threatening brain swelling. The tissue samples were subjected to histopathological and immunofluorescence microscopy evaluation, next generation miRNA sequencing (NGS), and bioinformatic analysis.

Results: miRNA NGS analysis detected 34 miRNAs with significantly aberrant expression in stroke tissue, as compared to non-stroke samples. Of these miRNAs, 19 were previously identified in stroke patient blood and CSF, while dysregulation of 15 miRNAs was newly detected in this study. miRNA direct target gene analysis and bioinformatics approach demonstrated a strong association of the identified miRNAs with stroke-related biological processes and signaling pathways.

Conclusions: Dysregulated miRNAs detected in our study could be regarded as potential candidates for biomarkers and/or targets for therapeutic intervention. The results described herein further our understanding of the molecular basis of stroke and provide valuable information for the future functional studies in the experimental models of stroke.

Keywords: malignant hemispheric infarction; microRNA; next generation microRNA sequencing; stroke.

Figure 2

Morphology of human stroke brain tissue. Morphological evaluation of the temporal lobe tissue, HE staining. (A,B), Control non-stroke samples; (C,D), Brain tissue collected at 48 h after stroke onset. Black arrows: eosinophilic (red) neurons with darkly stained pyknotic nuclei. Blue arrows: necrotic “ghost” neurons with an absent nuclear detail. Stars: peri-neuronal and peri-vascular space widening associated with edema. Neutrophils (arrowheads) were detected in the vicinity of large vessels and capillaries (E,F), tissue collected at 60 h after stroke onset), as well as in the brain parenchyma (G,H), tissue collected at 28 h after stroke onset). (E) and higher magnification in (F): Margination and pavementing of neutrophils along the blood vessel wall; some neutrophils migrated out and are detected in the blood vessel vicinity. (G) and higher magnification in (H): Neutrophils are detected at a considerable distance from the blood vessels, which demonstrates the process of neutrophil invasion into the damaged brain tissue. BV- blood vessel. Bars: (A,C): 50 μm; (B,D,E–H): 20 μm.

Figure 1

Representative pre-operative and post-operative images from the same stroke patient. (A): Axial computed tomography (CT) scan at one day post stroke demonstrating left sided middle cerebral artery infarct with loss of grey-white matter differentiation in a large territory. The white dotted line points out the approximate infarct border. (B): Post-operative CT scout image demonstrating the large craniotomy defect (dotted line). (C): Post-operative axial CT scan demonstrating completed infarct, bony decompression, and the region of partial stroke resection in the temporal lobe (arrow).

Figure 3

Immunofluorescence analysis of human stroke brain tissue IF staining of human stroke brain tissue collected at 48 h after stroke onset. A, B: Pan-neuronal antibody staining (red) visualizing small unipolar and bipolar (A) and spindle-shaped (B) neurons. (C): GFAP (green) immunostaining for astrocytes. (D): Iba-1 (red) immunostaining for microglia. Bars: (A,C,D): 20 μm, and (B): 50 μm.

Figure 4

Visualization of miRNA expression in stroke and control tissue. (A): Volcano plot showing the magnitude of the difference in expression values of the samples in Control and Stroke groups. The graph is constructed by plotting the −log₁₀(FDR corrected p-value) on the y-axis, and the log₂(fold-change) on the x-axis from the group of 172 miRNAs with impaired expression in stroke patient vs. control samples. Grey circles represent samples with FDR p-value > 0.05, green circles represent miRNAs that have a fold-change > 1.5 and FDR p-value > 0.05, blue circles represent miRNAs that have an FRD p-value < 0.05 and a fold-change > 1.5, and red circles represent miRNAs that have an FDR p-value < 0.05 and fold-change >2.0. (B): Principal component analysis (PCA) plot for stroke and control samples. The PCA was performed on 34 miRNAs identified to be differentially expressed between stroke and non-stroke control samples using the miRNAs that have the largest coefficient of variation based on TMM normalized counts. Each asterisk represents a sample. Red asterisk–stroke samples from patients S1–S5; black–control non-stroke samples C1–C3. (C): A heat map demonstrating expression levels of 50 microRNAs in three different control samples. Each column represents one miRNA and each row represents one sample. The color represents the difference of the count value to the row mean. N = 5 stroke and 3 control samples.

Figure 5

Pathway enrichment analysis for newly detected miRNAs. Diagram showing selected groups of GO terms (with the corresponding FDR p-value) significantly associated with the identified miRNAs (depicted in the box). The annotations describe molecular functions and signaling pathways, as well as general and brain-specific biological processes. Abbreviations: BMP- bone morphogenetic protein; MAPK—mitogen-activated protein kinase; PDGF—platelet-derived growth factor; RIG-I—retinoic acid-inducible gene receptor I; TGF-beta—transforming growth factor beta.

Figure 6

miR-155 target gene analysis. (A) A list of miR-155 direct target genes significantly downregulated in stroke tissue, compared to control samples. Blue—genes identified by PCR Array, green—additional genes identified by whole transcriptome sequencing. (B) A portion of a heatmap (after non-hierarchical clustering) displays a group of target genes with significantly lower (green) expression in the stroke samples. The heat map was generated based on delta Ct values. Genes with higher expression levels are shown in red, whereas genes with lower expression levels are shown in green. Genes with average expression levels are shown in black. (C) Scatter plot compares the normalized gene expression between the stroke and control samples. The central line indicates unchanged gene expression. Red circles identify genes significantly upregulated in stroke tissue, blue—genes significantly downregulated in stroke group. The graph plots the log10 of normalized gene expression levels in control samples (x-axis) versus stroke samples (y-axis). N = 3 samples per control and stroke groups.