Gene Expression Signatures Identify Biologically Homogenous Subgroups of Grade 2 Meningiomas

Abstract

Introduction: Meningiomas are the most common brain tumor, with prevalence of approximately 3%. Histological grading has a major role in determining treatment choice and predicting outcome. While indolent grade 1 and aggressive grade 3 meningiomas exhibit relatively homogeneous clinical behavior, grade 2 meningiomas are far more heterogeneous, making outcome prediction challenging. We hypothesized two subgroups of grade 2 meningiomas which biologically resemble either World Health Organization (WHO) grade 1 or WHO grade 3. Our aim was to establish gene expression signatures that separate grade 2 meningiomas into two homogeneous subgroups: a more indolent subtype genetically resembling grade 1 and a more aggressive subtype resembling grade 3.

Methods: We carried out an observational meta-analysis on 212 meningiomas from six distinct studies retrieved from the open-access platform Gene Expression Omnibus. Microarray data was analyzed with systems-level gene co-expression network analysis. Fuzzy C-means clustering was employed to reclassify 34 of the 46 grade 2 meningiomas (74%) into a benign "grade 1-like" (13/46), and malignant "grade 3-like" (21/46) subgroup based on transcriptomic profiles. We verified shared biology between matching subgroups based on meta-gene expression and recurrence rates. These results were validated further using an independent RNA-seq dataset with 160 meningiomas, with similar results.

Results: Recurrence rates of "grade 1-like" and "grade 3- like" tumors were 0 and 75%, respectively, statistically similar to recurrence rates of grade 1 (17%) and 3 (85%). We also found overlapping biological processes of new subgroups with their adjacent grades 1 and 3.

Conclusion: These results underpin molecular signatures as complements to histological grading systems. They may help reshape prediction, follow-up planning, treatment decisions and recruitment protocols for future and ongoing clinical trials.

Keywords: bioinformatics; gene expression networks; marker discovery; meningioma; transcriptomics.

FIGURE 1

Gene expression signatures associated with meningioma grade. (A) Differential gene expression between grades 3 and 1 meningiomas reveal four upregulated and two downregulated genes in grade 3 tumors [| log₂(fold change)| ≥ 1.5, p < 0.0001] highlighted with blue and red dots, respectively. (B) Gene co-expression networks analysis. Dendrogram of genes based on the topological overlap map, with the 29 gene modules represented by colors in the bar below. Gray represents unclassified genes. (C) Plot of median module meta-gene expression differences between grades (“m” versus “n”). Only modules with significantly different expression between grades 1 and 3 are included (Mann–Whitney p < 0.05). Red indicates modules which are upregulated in grade (m), and darker shades indicate larger effect sizes. Notably, 11/15 modules are significantly different between grades 1 and 2 while 2/15 are also significantly different between grades 2 and 3. *p < 0.05 (Mann–Whitney).

FIGURE 2

Optimized soft clustering reveals two subgroups of grade 2 meningiomas. (A) Cost of multiple input configurations: “royal blue” (RB) and “tan” (T) modules in blue and optimized differentially expressed genes in gray. Top inset depicts shape of sigmoid function with varied alphas. (B) Summary graph of fuzzy C-means clustering best performing inputs (RB + T). The x-axis represents the probability of being in the grade-3 enriched cluster and y-axis represents the proportion of patients in each bin of 10%. Line graph component represents normalized frequency distribution of each histological grade (green = grade 1, black = grade 2, red = grade 3). Top jitter plot represents individual patients. Dark green and red bars above represent the 20 and 80% thresholding into grade 1-like and grade 3-like subgroups of grade 2 meningiomas. Recurrence rates are plotted on the right by grade (green, black, red) and subgroup (“grade 1-like” and “grade 3-like”). *Chi-square p < 0.05.

FIGURE 3

Validation of meningioma reclassification using RNA-seq data. Summary graph of fuzzy C-means clustering best performing inputs on the microarray data (modules RB + T). The x-axis represents the probability of being in the grade-3 enriched cluster and y-axis represents the proportion of patients in each bin of 10%. Line graph component represents normalized frequency distribution of each histological grade (green = grade 1, black = grade 2, red = grade 3). Top jitter plot represents individual patients. Dark green and red bars above represent the 20 and 80% thresholding into grade 1-like and grade 3-like subgroups of grade 2 meningiomas. Recurrence rates are plotted on the right by grade and subgroup (“grade 1-like” and “grade 3-like”). Notably, recurrence data is not available for the grade 3 meningiomas in this cohort.

FIGURE 4

Molecular identity of newly described grade 2 meningioma subgroups. (A) Scatter plot of median module meta-gene expression (unitless). Larger circles indicate Mann–Whitney p < 0.05. Colors correspond to previously identified modules in Figure 1B. ρ = Pearson coefficient, *p < 0.05. Note the positive correlation between the modules of grade 1 and “grade 1-like” and grade 3 and “grade 3-like” subtypes. (B) Scatter plots of genetic separation between grade 2 subtypes as histological grades. The x-axis represents the difference in median module expression between grades 3 and 1, while the y-axis represents the difference in median module expression between “grade 3-like” and “grade 1-like.” Large circles represent modules which are significantly different in both comparisons and empty circles indicate modules which are not significantly different in either. Of the remainder, 4/6 are significantly different between grades 3 and 1 only and 2/6 is significantly different between “grade 3-like” and “grade 1-like” (∼).