Regulation of IL-8 gene expression in gliomas by microRNA miR-93

Abstract

Background: Different strategies have been proposed to target neoangiogenesis in gliomas, besides those targeting Vascular Endothelial Growth Factor (VEGF). The chemokine Interleukin-8 (IL-8) has been shown to possess both tumorigenic and proangiogenic properties. Although different pathways of induction of IL-8 gene expression have been already elucidated, few data are available on its post-transcriptional regulation in gliomas.

Methods: Here we investigated the role of the microRNA miR-93 on the expression levels of IL-8 and other pro-inflammatory genes by RT-qPCR and Bio-Plex analysis. We used different disease model systems, including clinical samples from glioma patients and two glioma cell lines, U251 and T98G.

Results: IL-8 and VEGF transcripts are highly expressed in low and high grade gliomas in respect to reference healthy brain; miR-93 expression is also increased and inversely correlated with transcription of IL-8 and VEGF genes. Computational analysis showed the presence of miR-93 consensus sequences in the 3'UTR region of both VEGF and IL-8 mRNAs, predicting possible interaction with miR-93 and suggesting a potential regulatory role of this microRNA. In vitro transfection with pre-miR-93 and antagomiR-93 inversely modulated VEGF and IL-8 gene expression and protein release when the glioma cell line U251 was considered. Similar data were obtained on IL-8 gene regulation in the other glioma cell line analyzed, T98G. The effect of pre-miR-93 and antagomiR-93 in U251 cells has been extended to the secretion of a panel of cytokines, chemokines and growth factors, which consolidated the concept of a role of miR-93 in IL-8 and VEGF gene expression and evidenced a potential regulatory role also for MCP-1 and PDGF (also involved in angiogenesis).

Conclusion: In conclusion, our results suggest an increasing role of miR-93 in regulating the level of expression of several genes involved in the angiogenesis of gliomas.

Fig. 1

Expression of IL-8 and VEGF mRNA in glioblastoma. VEGF mRNA (a, c, e) and IL-8 mRNA (b, d, f) by mRNA in situ hybridization are shown in separate 5 μm serial tissue sections from glioblastoma specimens at different magnifications (a, b: x2; c, d: x10; e, f: x32) by peroxidase staining. Nuclei are counterstained with hematoxylin. Positive (GAPDH mRNA) and negative (DAPB mRNA) controls are reported (g, h: x20 magnification). Squared areas in panels A and B indicate the detail reported in panels c and d, respectively

Fig. 2

Expression of VEGF, IL-8 and miR-93 in Low-Grade Gliomas (LGGs) and High-Grade Gliomas (HGGs). VEGF mRNA (a) and IL-8 mRNA (b) levels relative to GAPDH were measured by RT-qPCR with TaqMan probes on RNAs isolated from FFPE sections of 6 LGG and 10 HGG and normalized to healthy brain reference RNA. Fold changes (FC) of expression over healthy brain reference RNA are reported. In the same LGGs and HGGs miR-93 was quantified (c) and normalized to healthy brain reference RNA. For panels a–c: dashed line: mean; solid line: median; grey box includes values from 5th to 95th centiles, vertical lines range from min to max values, excluding outliers which are represented by single dots. The data obtained in each glioma specimen are reported in the right side of the panels. d Relationship between VEGF mRNA and IL-8 mRNA in the same LGG (filled circles) and HGG (open circles) samples analyzed and reported in a–c. Regression straight line showing direct correlation was drawn by the least square method Sigmaplot. Inset reports the same graph expanded

Fig. 3

Correlations among the expression of miR-93, VEGF and IL-8 mRNAs in HGGs. Regression analysis between VEGF mRNA (a) and IL-8 mRNA (b) was performed as function of miR-93. Regression straight line showing inverse correlation was drawn by the least square method with Sigmaplot software. The graphical correlations among miR-93, VEGF and IL-8 mRNAs is represented in 3D plot (c). All data are reported as fold changes (FC) over healthy brain reference RNA

Fig. 4

Interactions of miR-93 with IL-8 mRNA and VEGF mRNA. Predicted secondary structure of the 3′UTR regions of VEGF mRNA (a) and IL-8 mRNA (b) based on the UCSC genome browser (http://genome.ucsc.edu), and of miR-93 by UNAFold Web Server (http://mfold.rna.albany.edu). Magnification is also shown of the central portion of 3′UTR IL-8 (b) and VEGF (a) mRNAs and pointing out the possible interaction between the 3′UTR target strands and the seed region of the lowest energy miR-93 potential stem loops

Fig. 5

Effects of the treatments of glioma U251 cells with pre-miR-93 and antagomiR-93. VEGF (a) and IL-8 released protein (b, c) were quantified by Bio-plex analysis. RNA was isolated from cultures after 48 h in vitro growth and analyzed by RT-qPCR. Internal RT-qPCR control were U6 snRNA and let-7c for miR-93, RPL13A and 18S for IL-8 mRNA. Data are in all cases reported in comparison to U251 cells treated with control scrambled sequences. Results represent the average ± S.D. of at least three independent experiments. * p < 0.05; ** p < 0.01

Fig. 6

Effects on IL-8 mRNA (a, c) and IL-8 protein (b, d) of the treatments of glioma U251 (a, b) and T98G (c, d) cells with pre-miR-93 and antagomiR-93. RNA was isolated from cultures after 48 h in vitro growth and analyzed by RT-qPCR. Internal RT-qPCR control was RPL13A. Released IL-8 protein was quantified by Bio-plex analysis. Data are in all cases reported in comparison to U251 and T98G cells treated with control sequences. Results represent the average ± S.D. of three independent experiments. * p < 0.05; ** p < 0.01

Fig. 7

Secretome profile of U251 cells. a, b Arrowed are protein exhibiting low level of secretion by U251 cells. c Changes in the protein profile after treating U251 cells with antagomiR-93 and pre-miR-93. The data generating this panel are shown in Table 1 and are originated by Bio-plex analysis

Fig. 8

a Proposed model for the control of the changes of IL-8 and miR-93 expression during the transition from normal brain tissues to LGG and to HGG. In the scheme is also indicated some effects (namely of VEGF, MCP-1, PDGF-B) associated to miR-93 down-regulation during LGG to HGG transition. b Proposed inhibitory effects of miR-93 within the microenvironment participating to glioma angiogenesis [59]