Epigenetic and genomic profiling of chordoid meningioma: implications for clinical management

Abstract

Chordoid meningioma is a morphological variant of meningioma designated as WHO grade 2. However, the recurrence rates varied widely in different case series, and to date, a unifying molecular genetic signature has not been identified. Among 1897 meningiomas resected at our institution, we identified 12 primary chordoid meningiomas from 12 patients. Histologically, all 12 cases had predominant (> 50%) chordoid morphology. Ten were otherwise grade 1, and two were also atypical. We performed DNA global methylation profile, copy number variation analysis, and targeted next-generation sequencing on 11 chordoid meningiomas, and compared to those of 51 non-chordoid, mostly high grade meningiomas. The chordoid meningiomas demonstrated a unique methylation profile in tSNE, UMAP, and hierarchical heatmap clustering analyses of the most differentially methylated CpGs. The most common copy number variation in chordoid meningioma was loss of 1p (7/11, 64%). Three chordoid meningiomas had 2p loss, which was significantly higher than the non-chordoid control cohort (27% vs 7.2%, p = 0.035). 22q loss was only seen in the two cases with additional atypical histological features. Chordoid meningiomas were enriched in mutations in chromatin remodeling genes EP400 (8/11,73%) KMT2C (4/11, 36%) and KMT2D (4/11, 36%), and showed low or absent NF2, TERT, SMO, and AKT1 mutations. Prognosis wise, only one case recurred. This case had atypical histology and high-grade molecular features including truncating NF2 mutation, 1p, 8p, 10, 14, 22q loss, and homozygous deletion of CDKN2A/B. Progression free survival of chordoid, otherwise grade 1 meningioma was comparable to non-chordoid WHO grade 1 meningioma (p = 0.75), and significantly better than chordoid WHO grade 2 meningioma (p = 0.019). Conclusion: the chordoid histology alone may not justify a universal WHO grade 2 designation. Screening for additional atypical histological or molecular genetic features is recommended.

Keywords: Atypical meningioma; Brain tumor; Chordoid meningioma; Global methylation profile; Prognosis; WHO grade.

Fig. 1

Chordoid meningioma histology. A–D Chordoid meningiomas with otherwise benign histology, H&E. A Case 4, tumor cells arranged as individual cells. B Patient 1, tumor cells arranged as single file cords. C,D Patient 6, tumor cells arranged as small nests. Asterisk indicates solid area without mucin. E–H Patient 11, Chordoid meningioma with atypical histology, H&E. E Chordoid morphology dominated the tumor. F Extensive brain invasion. G Mitosis (arrow) and prominent nucleoli. H Geographic necrosis

Fig. 2

Kaplan–Meier survival analyses of progression free survival and overall survival of chordoid meningiomas compared to non-chordoid meningiomas

Fig. 3

Unsupervised clustering of epigenetic markers from chordoid meningioma samples reveals distinct subtypes. DNA methylation from 62 patient samples was analyzed using the EPIC array; the resulting methylation matrix was then passed through two commonly used dimensionality reduction algorithms, tSNE (left) and UMAP (right). Samples are annotated by histological classification. ATM: atypical meningioma. ANA: anaplastic meningioma. CHOM: chordoid meningioma. MNG: meningothelial and transitional meningiomas

Fig. 4

Differentially Methylated Probes (DMPs) analysis of a heatmap of the top 5000 most variable methylated CpGs. Columns represent samples and rows methylation sites; individual cells are color coded by their methylation value. Hierarchical clustering was performed on both axes of the heatmap. The distance of each methylation site from a transcription start site is annotated to the right of the heatmap. Differential methylation highlights the distinct signature of chordoid meningiomas. Meningioma classification is annotated beneath the heatmap: ANM- Anaplastic meningioma, ATM- Atypical meningioma, CHOM- Chordoid meningioma, MNG- meningothelial and transitional

Fig. 5

Genomic profiles of chordoid meningiomas. Visualization of recurrent genomic alteration events in Meningiomas by OncoPrint. The samples are clustered by Meningioma type. Alteration percentage per gene is shown on the left side