Posttranscriptional deregulation of signaling pathways in meningioma subtypes by differential expression of miRNAs

Abstract

Background: Micro (mi)RNAs are key regulators of gene expression and offer themselves as biomarkers for cancer development and progression. Meningioma is one of the most frequent primary intracranial tumors. As of yet, there are limited data on the role of miRNAs in meningioma of different histological subtypes and the affected signaling pathways.

Methods: In this study, we compared expression of 1205 miRNAs in different meningioma grades and histological subtypes using microarrays and independently validated deregulation of selected miRNAs with quantitative real-time PCR. Clinical utility of a subset of miRNAs as biomarkers for World Health Organization (WHO) grade II meningioma based on quantitative real-time data was tested. Potential targets of deregulated miRNAs were discovered with an in silico analysis.

Results: We identified 13 miRNAs deregulated between different subtypes of benign meningiomas, and 52 miRNAs deregulated in anaplastic meningioma compared with benign meningiomas. Known and putative target genes of deregulated miRNAs include genes involved in epithelial-to-mesenchymal transition for benign meningiomas, and Wnt, transforming growth factor-β, and vascular endothelial growth factor signaling for higher-grade meningiomas. Furthermore, a 4-miRNA signature (miR-222, -34a*, -136, and -497) shows promise as a biomarker differentiating WHO grade II from grade I meningiomas with an area under the curve of 0.75.

Conclusions: Our data provide novel insights into the contribution of miRNAs to the phenotypic spectrum in benign meningiomas. By deregulating translation of genes belonging to signaling pathways known to be important for meningioma genesis and progression, miRNAs provide a second in line amplification of growth promoting cellular signals. MiRNAs as biomarkers for diagnosis of aggressive meningiomas might prove useful and should be explored further in a prospective manner.

Keywords: histological subtypes; meningioma; miRNA; qRT-PCR.

Fig. 1.

Hierarchical clustering of the array samples based on expression of the 57 miRNAs with significant expression differences among WHO grades or histological subtypes. Cluster A contains 6 grade II (orange), 10 grade III (red), and only 3 grade I meningiomas, and cluster B consists of the remaining 30 grade I, 4 grade II, and 2 grade III meningiomas. Meningothelial, transitional, fibroblastic, grade II, and grade III meningiomas are indicated with blue, green, yellow, orange, and red bars, respectively.

Fig. 2.

(A) Significantly deregulated miRNA in WHO grade I meningiomas. Arrows: number of significantly deregulated miRNAs for the comparisons of meningothelial vs fibroblastic, meningothelial vs transitional, and transitional vs fibroblastic meningiomas. Green: downregulated in meningothelial meningioma. Red: upregulated in meningothelial meningioma. (B) Significantly deregulated miRNAs in WHO grade III meningioma compared with grade I meningioma with respect to the histological subtypes. Green: downregulated in anaplastic meningioma. Red: upregulated in anaplastic meningioma. Underlined: miRNAs located on chromosome 14q.

Fig. 3.

Quantitative RT-PCR validation of 6 meningioma-deregulated miRNAs in validation set (n = 95). Mean ΔCT of 6 meningioma-deregulated miRNAs in meningothelial (white bar), fibroblastic (light grey bar), and transitional (dark grey bar) subtypes, WHO grade II (heavy dark grey bar) and grade III meningiomas (black bar). Significant expression differences are marked with square brackets above. Horizontal bar represents all grade I meningiomas combined, eg, for miR-195 significant expression difference between combined grade I meningiomas compared with grade III meningiomas. Significant expression differences between grade I histological subtypes separately compared with either grade II or III meningiomas are not displayed for reason of clarity, but are included in Supplementary Table S6. Higher mean ΔCT values indicate lower expression of miRNA.

Fig. 4.

Receiver operator characteristics (ROC) analysis. ROC curves for SVM-based prediction models for differentiating WHO grade I from grade II meningiomas using expression of miR-34a* and -136, separately, and the 497/34a*/136/222 combined signature in training (left panels) and test set (right panels). AUC value for each analysis is given.