New Brain Tumor Entities Emerge from Molecular Classification of CNS-PNETs

Abstract

Primitive neuroectodermal tumors of the central nervous system (CNS-PNETs) are highly aggressive, poorly differentiated embryonal tumors occurring predominantly in young children but also affecting adolescents and adults. Herein, we demonstrate that a significant proportion of institutionally diagnosed CNS-PNETs display molecular profiles indistinguishable from those of various other well-defined CNS tumor entities, facilitating diagnosis and appropriate therapy for patients with these tumors. From the remaining fraction of CNS-PNETs, we identify four new CNS tumor entities, each associated with a recurrent genetic alteration and distinct histopathological and clinical features. These new molecular entities, designated "CNS neuroblastoma with FOXR2 activation (CNS NB-FOXR2)," "CNS Ewing sarcoma family tumor with CIC alteration (CNS EFT-CIC)," "CNS high-grade neuroepithelial tumor with MN1 alteration (CNS HGNET-MN1)," and "CNS high-grade neuroepithelial tumor with BCOR alteration (CNS HGNET-BCOR)," will enable meaningful clinical trials and the development of therapeutic strategies for patients affected by poorly differentiated CNS tumors.

Figure 1. Molecular Classification of CNS-PNETs by DNA Methylation Profiling

(A) Unsupervised clustering of DNA methylation patterns of 323 CNS-PNET samples alongside 211 reference samples representing CNS tumors of known histology and molecular subtype using the 10,000 most variably methylated probes. Molecular diagnostic reference tumors or CNS-PNETs (inner circle) and gene expression subgroup assignment (outer circle) are depicted by colored bars as indicated. DNA methylation clusters are highlighted by colors as indicated. Grey bars indicate samples unclassifiable by gene expression analyses. (B) Two dimensional representation of pairwise sample correlations using the 10,000 most variably methylated probes by t-Distributed Stochastic Neighbor Embedding (tSNE) dimensionality reduction. The same samples as in (A) are used (n = 534). Reference samples are colored according to their molecular reference entity. CNS-PNET samples are colored in black. Lines connect each sample to the centroid of its respective molecular CNS tumor entity. (C) Re-classification of 323 CNS-PNETs into known molecular reference entities and four new CNS tumor entities by molecular profiling. Entities correspond to DNA methylation clusters and are represented by colors as indicated. See also Figure S1 and Table S1.

Figure 2. Molecular and Clinical Characteristics of Re-Classified CNS-PNET Groups

(A-D) Molecular characteristics of CNS-PNETs from ETMR (A), AT/RT (B), HGG_IDH, HGG_K27, and HGG_G34 (C), and HGG_MYCN (D) DNA methylation clusters. Detection and frequency of characteristic molecular alterations in each group are indicated. Representative copy-number profiles in (A), (B), and (D) depict genomic gains (green dots) and losses (red dots) on individual chromosomes as indicated. FISH and IHC images in (A), (B), and (D) show representative tumor samples. (E-H) Tumor location and age at diagnosis from ETMR (E), AT/RT (F), HGG_IDH, HGG_K27, and HGG_G34 (G), and HGG_MYCN (H) DNA methylation clusters. Black bars in age plots indicate the median. Numbers in brackets indicate group size with available data. See also Figure S2 and Tables S1 and S2.

Figure 3. Identification of New CNS Tumor Entities Across Histologies

(A) Unsupervised clustering of DNA methylation patterns of 77 CNS-PNET samples alongside 159 reference samples and 59 additional samples representing CNS tumors of varying histology using the 10,000 most variably methylated probes. Molecular subgroup assignment by DNA methylation (inner circle) or gene expression patterns (middle circle) correspond to subgroup labels. Original tumor histology (outer circle) is depicted for tumors from new molecular CNS tumor entities by colored bars as indicated. (B) Composition of four new CNS tumor entities by histological diagnosis. Tumor histology is represented by colors as indicated. (C-F) Clinical patient information for four novel CNS tumor entities CNS NB-FOXR2 (C), CNS EFT-CIC (D), CNS HGNET-MN1 (E), and CNS HGNET-BCOR (F). For each entity, tumor location (left panel), age at diagnosis (middle panel), and gender distribution (right panel) are shown. Numbers in brackets indicate group size with available data. See also Figure S3 and Table S3.

Figure 4. Histopathological Patterns of New CNS Tumor Entities

(A-C) The CNS NB-FOXR2 entity was characterized by uniform round embryonal cells with minimal cytological pleomorphism. Nuclear palisades and neurocytic differentiation were frequently encountered. (D-F) CNS EFT-CIC tumors were composed of small monotonous cells. The tumor architecture was variable and included fascicular and alveolar growth. Select examples demonstrated a spindle cell phenotype. (G-I) CNS HGNET-MN1 tumors were composed of monotonous neuroepithelial cells with oval forms. Pseudopapillary architecture and dense stromal hyalinization was often encountered. (J-L) The CNS HGNET-BCOR entity was characterized by oval to elongated cells. Perivascular anuclear zones were often present and glial fibrillary processes were typical. Scale bars represent 50 μm. See also Figure S4 and Table S2.

Figure 5. Recurrent Molecular Alterations in the CNS NB-FOXR2 Entity

(A) Schematic representation depicting chromosomal location, wild-type RNA transcripts, and exon structures resulting from an exemplary genetic alteration affecting the FOXR2 gene. (B) Frequency of FOXR2 re-arrangements identified by RNA/DNA sequencing or copy-number data. (C) Gene expression levels of FOXR2 in various CNS tumor entities. See also Figure S5 and Table S4.

Figure 6. Recurrent Molecular Alterations in CNS EFT-CIC, CNS HGNET-MN1 and CNS HGNET-BCOR Entities

(A-I) Schematic representation, frequency, and transcriptomic effects of recurrent molecular alterations found in tumors from the CNS EFT-CIC (A-C), CNS HGNET-MN1 (D-F), and CNS HGNET-BCOR (G-I) entities. Schematics in (A), (D), and (G) depict chromosomal location, wild-type RNA transcripts, and exon structures resulting from recurrent alterations. Frequencies of the respective events detected by different methods are depicted in panels (B), (E), and (H). Gene expression levels of NUTM1, BEND2, and BCOR across various CNS tumor entities are displayed in panels (C), (F), and (I). See also Figure S6 and Table S4.

Figure 7. Transcriptional Profiling of New CNS Tumor Entities

(A) Heatmap representing the expression levels of the ten most significantly differentially up-regulated genes comparing one new CNS tumor entity vs. the three others. Each column represents one sample, each lane represents one gene. Gene expression levels are represented by a color scale as indicated. (B) Individually selected marker genes specifically up-regulated in one of the new CNS tumor entities compared with other CNS tumor entities as indicated. See also Figure S7 and Table S5 and S6.