Characterization of Dysregulated miRNA in Peripheral Blood Mononuclear Cells from Ischemic Stroke Patients

Abstract

Epigenetic modification may play an important role in pathophysiology of ischemic stroke (IS) risk. MicroRNAs (miRNAs), which constitute one of the modes of epigenetic regulation, have been shown to be associated with a number of clinical disorders including IS. The purpose of this study was to investigate the miRNA profile in the peripheral blood mononuclear cells (PBMCs) of IS patients and compare it with stroke-free controls. Blood samples were obtained from 19 healthy age-gender-race matched individuals who served as controls to 20 IS patients. miRNA microarray analysis with RNA from PBMCs was performed and significantly dysregulated miRNAs common among IS patients were identified. We identified 117 miRNAs with linear fold values of at least ±1.5, of which, 29 were significantly altered (p value <0.05). Ingenuity Pathway Analysis (IPA) indicated a role for the dysregulated miRNAs in conditions relevant to IS (e.g., organismal injury and abnormalities, hematological disease and immunological disease). Pro-inflammatory genes like STAT3, interleukin (IL) 12A, and IL12B were some of the highly predicted targets for the dysregulated miRNAs. Notably, we further identified three common and significantly upregulated miRNAs (hsa-miR-4656, -432, -503) and one downregulated miRNA (hsa-miR-874) among all IS patients. Molecular interactive network analysis revealed that the commonly dysregulated miRNAs share several targets with roles relevant to IS. Altogether, we report dysregulation of miRNAs in IS PBMCs and provide evidence for their involvement in the immune system alteration during IS pathophysiology.

Keywords: Inflammation; Ischemic stroke; PBMCs; miRNA.

Fig. 1

miRNAs are dysregulated in IS PBMCs. a Heat map showing the expression level of all 1009 miRNAs (number of probes used to detect miRNAs during the array) in the controls (n=5) and IS samples (n=5) used for the microarray analysis with RNA isolated from PBMCs. b Scatter plot of the linear fold-change values of all 1009 miRNAs analyzed in the microarray. c Volcano plot showing the expression levels of the miRNAs with fold change above the cutoff value (±1.5) and p<0.05 (X-axis: Log2 fold-change; Y-axis: -Log10 of p-values. On the X-axis, 0.585 is ~1.5 linear fold-change and 1.0 is ~2 linear fold-change. On the Y-axis, the line denotes the p-value 0.05).

Fig. 2

Identification of molecular networks using IPA. a The most significant network by IPA: ‘Organismal injury and abnormalities, Reproductive system disease, Cancer’. To enhance the reliability of the results, only the miRNAs with a p-value <0.05 were selected thereby obtaining 29 miRNAs. b Heat map showing the expression levels of only the 29 significantly dysregulated miRNAs. (C: control, P: IS patients). c Top network obtained from IPA, with only the 29 significantly dysregulated miRNAs, under the category: ‘Cancer, organismal injury and abnormalities’ and ‘reproductive system diseases’. In the network, the following genes were manually added based on information for miRNA-gene interaction available on www.microrna.org (SMAD3, TGFB2, TGFB3, MKL2, ETS1, TGFBR1, IL12A, IL12B).

Fig. 2

Identification of molecular networks using IPA. a The most significant network by IPA: ‘Organismal injury and abnormalities, Reproductive system disease, Cancer’. To enhance the reliability of the results, only the miRNAs with a p-value <0.05 were selected thereby obtaining 29 miRNAs. b Heat map showing the expression levels of only the 29 significantly dysregulated miRNAs. (C: control, P: IS patients). c Top network obtained from IPA, with only the 29 significantly dysregulated miRNAs, under the category: ‘Cancer, organismal injury and abnormalities’ and ‘reproductive system diseases’. In the network, the following genes were manually added based on information for miRNA-gene interaction available on www.microrna.org (SMAD3, TGFB2, TGFB3, MKL2, ETS1, TGFBR1, IL12A, IL12B).

Fig. 3

Identification of the dysregulated miRNAs common in all the IS samples included for the miRNA microarray. Linear fold-change of all the miRNAs was calculated for individual IS patients. A threshold of ±1.5 was set and the miRNAs with expression level beyond the threshold was analyzed by Venn to get the common miRNAs seen in IS patients. a Venn diagram showing miRNAs (5) that are commonly up-regulated (hsa-miRs-4656, −432, −487b, −503, −409-3p) in the 5 IS patients. b Venn diagram showing miRNAs that are commonly downregulated (hsa-miR-874-3p) in the 5 IS patients compared to controls. The targets of four miRNAs (hsa-miRs-4656, −432, −503 referred to as select miRNAs and −487b) were analyzed by Cytoscape to generate the molecular interaction network. c Molecular interaction network of the select miRNAs, hsa-miR487b and their target genes.

Fig. 4

Validation of select miRNAs and target genes by qRT-PCR. a Linear fold-change values of the commonly dysregulated miRNAs after microarray. b Relative abundance (RA) after qRT-PCR validation of representative dysregulated miRNAs. We randomly selected the miRNAs for validation from dysregulated miRNAs of which miRNAs, hsa-miRs-4656, −432 and −503, were among the commonly up-regulated miRNAs in all the IS patients Greater than 90% of the IS patients had higher expression for hsa-miRs-4656, −432 and −503 upon validation by qRT-PCR. The hsa-miR-320a was included here to further confirm the reliability of our microarray result. c) Result from qRT-PCR validation of 8 target genes of significantly dysregulated miRNAs. The values represent relative abundance of the transcripts in IS PBMCs (Control n=19, IS n=20). 18S rRNA was used as an internal control.

Fig. 5

MiR-432-5p directly interacts with the UTRs of human TGFB3, CELSR2 and ITM2C. a Nucleotide sequence of hsa-miR-432-5p and the UTR of the genes predicted to interact with miR-432-5p. The sequence details, miRSVR and PhastCons scores were obtained from www.microrna.org. b The UTRs were cloned into pmiRGLO vector and co-transfected with the hsa-miR-432-5p mimics into THP1 cells for luciferase activity measurement. The bar graph shows the relative luciferase activity 48h after transfection with the different recombinant vectors compared to control and mutant. Becuase the seed sequence complementary bases was same for all the three UTRs, we used only one mutant which had all 7 nucleotides of the seed sequence changed.