Molecular heterogeneity and CXorf67 alterations in posterior fossa group A (PFA) ependymomas

Abstract

Of nine ependymoma molecular groups detected by DNA methylation profiling, the posterior fossa type A (PFA) is most prevalent. We used DNA methylation profiling to look for further molecular heterogeneity among 675 PFA ependymomas. Two major subgroups, PFA-1 and PFA-2, and nine minor subtypes were discovered. Transcriptome profiling suggested a distinct histogenesis for PFA-1 and PFA-2, but their clinical parameters were similar. In contrast, PFA subtypes differed with respect to age at diagnosis, gender ratio, outcome, and frequencies of genetic alterations. One subtype, PFA-1c, was enriched for 1q gain and had a relatively poor outcome, while patients with PFA-2c ependymomas showed an overall survival at 5 years of > 90%. Unlike other ependymomas, PFA-2c tumors express high levels of OTX2, a potential biomarker for this ependymoma subtype with a good prognosis. We also discovered recurrent mutations among PFA ependymomas. H3 K27M mutations were present in 4.2%, occurring only in PFA-1 tumors, and missense mutations in an uncharacterized gene, CXorf67, were found in 9.4% of PFA ependymomas, but not in other groups. We detected high levels of wildtype or mutant CXorf67 expression in all PFA subtypes except PFA-1f, which is enriched for H3 K27M mutations. PFA ependymomas are characterized by lack of H3 K27 trimethylation (H3 K27-me3), and we tested the hypothesis that CXorf67 binds to PRC2 and can modulate levels of H3 K27-me3. Immunoprecipitation/mass spectrometry detected EZH2, SUZ12, and EED, core components of the PRC2 complex, bound to CXorf67 in the Daoy cell line, which shows high levels of CXorf67 and no expression of H3 K27-me3. Enforced reduction of CXorf67 in Daoy cells restored H3 K27-me3 levels, while enforced expression of CXorf67 in HEK293T and neural stem cells reduced H3 K27-me3 levels. Our data suggest that heterogeneity among PFA ependymomas could have clinicopathologic utility and that CXorf67 may have a functional role in these tumors.

Keywords: CXorf67; DNA methylation profiling; Ependymoma; H3 K27-trimethylation; H3 K27M; Molecular heterogeneity; PRC2.

Figures

Figure 1.. PFA ependymomas comprise two subgroups and nine subtypes.

(A) Heat map representation of an unsupervised consensus hierarchical clustering of DNA methylation profiles from 675 PFA ependymomas using the 5000 most differentially methylated probes across all tumors. Each row represents a probe, and each column represents a sample. The level of DNA methylation (beta-value) is represented by a red-blue color scale, as depicted in the key (lower right). For each sample, subtype association, institutional origin, gender, and patient age (years) are provided in four rows below the heat map. Dendrograms related to the two subgroups, PFA-1 and PFA-2, are depicted in red and blue, respectively. The dotted line transecting the dendrogram represents the cut-off for nine subtypes as established through an analysis of the cumulative distribution function (see Figure S2). (B) TSNE plot of DNA methylation array data from 675 PFA ependymomas. Samples are colored according to their respective consensus cluster affiliation in Figure 1A and as shown in the key at lower right.

Figure 2.. Differential expression of genes involved at brain stem sites during CNS embryogenesis in PFA-1 and PFA-2 ependymomas.

(A) Unsupervised clustering analysis of gene expression profiles generated by transcriptome sequencing on a subset of PFA ependymomas (n=28) and using the top 100 most differentially expressed genes. Consensus matrix heat maps and a principal components analysis (PCA) plot were used to demonstrate stable cluster number and two subgroups, PFA-1 and PFA-2 (PC1 & PC2 = standard deviation / dispersion ellipses). (B) Volcano plot of transcriptome sequencing data showing genes overexpressed in PFA-1 (red) and PFA-2 ependymomas (blue). Many genes overexpressed in PFA-1 tumors, relative to PFA-2 tumors, belong to the HOX family of genes that encode transcription factors involved in CNS patterning. Relatively overexpressed patterning genes in the two subgroups have actions at different levels of the brain stem during development. PFA-1 HOX genes are active in the caudal brain stem (red), while PFA-2 genes are active more rostrally, at the midbrain-hindbrain boundary (blue), illustrated diagrammatically (C), or in preparations from the Allen Brain Atlas of mouse embryogenesis (Image credit: Allen Institute), where expression levels of HOXA4 (D) and EN2 (E) are detected by in situ hybridization (arrows).

Figure 3.. Clinical and genetic heterogeneity among nine subtypes of PFA ependymoma.

Nine subtypes of PFA ependymoma displaying variability in: (A) age at diagnosis (years), (B) gender ratio, (C) ratio of pathological grade, (D) progression-free survival (PFS), and (E) overall survival (OS). (F) Elevated OTX2 expression in PFA-2c ependymomas demonstrated by gene expression profiling (2x Affymetrix u133 v2 array probes) and immunohistochemistry (G). Immunonegative PFA ependymoma of another subtype (H). Varying frequencies across subtypes of the four commonest chromosome arm copy number alterations found in PFA ependymomas; 1q gain (I), 6q loss (J), 10q loss (K), 22q loss (L). Scale bars (G/H) = 100μm.

Figure 4.. Mutations of CXorf67 and H3 K27M in PFA ependymomas.

(A) Upper: CXorf67 missense SNVs (upper) detected in 22/234 (9.4%) PFA ependymomas. The red bar marks a hotspot between codons 71 and 122. Lower: In silico analysis of CXorf67 protein. Based on IUPred [10], combined disorder score by DisMeta [19], and the meta-predictor PONDR-FIT [54], CXorf67 is predicted to be largely disordered. The SNV hotspot is located within a small ordered region that is predicted by ANCHOR to be a protein-protein interaction domain [11]. The NCPR (linear net charge per residue) plot shows a net negative charge in the mutation hotspot. (B) Frequencies of CXorf67 mutation are listed for various tumors (source data are: SJ PCGP, TCGA, TARGET and dbGaP study phs000754). (C) Frequencies of CXorf67 mutation and different H3 gene mutations across PFA ependymoma subtypes.

Figure 5.. Expression of CXorf67 in PFA ependymomas.

(A) Affymetrix u133v2 array and transcriptome sequencing (RNA-seq) data demonstrating significantly higher CXorf67 expression in PFA than in PFB or supratentorial (ST) ependymomas (Affymetrix u133 data, p=3×10–29 Welch ANOVA; RNA-seq data, p=4×10–9, Welch ANOVA). (B) CXorf67 expression at the RNA level in different types and molecular groups of CNS tumor (data derived from DKFZ cohorts). (C) CXorf67 expression at the RNA level (Affymetrix u133v2 array data) across PFA subtypes and in PFB or ST ependymomas. PFA-1f tumors express CXorf67 at the same level as PFB and ST tumors, significantly different (p=7×10–18, Welch ANOVA) from other PFA tumors. (D) Strong CXorf67 immunoreactivity in a PFA-1c tumor. (E) Negligible CXorf67 expression in a PFA-1f ependymoma that has an H3 K27M mutation. (F) Almost identical expression levels of CXorf67 (Affymetrix u133v2 array data) in PFA ependymomas with or without a CXorf67 mutation (p=0.51 Welch t-test). (G) Heat map of methylation at CpG sites across the CXorf67 gene demonstrating relative hypomethylation (blue) in the promoter region (red bar) for PFA ependymomas, but relative hypermethylation (yellow) at the same sites for PFB / ST ependymomas. (H) Significantly higher CXorf67 promoter region methylation in H3-mutant tumors than in H3-wildtype tumors – average beta for 6 probes (p=3×10–13 Welch t-test). Scale bars (D/E) = 100μm.

Figure 6.. CXorf67 binds core elements of PRC2.

(A) A 2-D SAINT plot from data generated by immunoprecipitation (IP) / mass spectrometry (MS) and using an anti-CXorf67 antibody against protein lysates from two cell lines, Daoy and U2-OS, which express high levels of CXorf67. (B) A 3-D SAINT plot from data generated by IP / MS and using anti-CXorf67, anti-EZH2, and anti-SUZ12 antibodies against protein lysates of nuclear extracts from Daoy (see also Table S5). Both SAINT plots show that CXorf67 binds to EZH2, SUZ12, and EED, the three core elements of PRC2, as well as to other proteins related to the PRC2 complex. (C) Composite image of immunoblotting lysates used in IP / MS to generate the 3-D SAINT plot. Anti-CXorf67, anti-EZH2, and anti-SUZ12 antibodies were used to pull down and then detect proteins in the nuclear fraction of Daoy cells. (D) Immunoblot showing that the PFA ependymoma cell line, EPD210FH, expresses CXorf67, but not H3 K27-me3, in contrast to the RELA fusion-positive ependymoma cell line, EP1NS. (E) IP with an anti-CXorf67 antibody on lysates from EPD210FH demonstrating interactions between CXorf67 and EZH2 and SUZ12.

Figure 7.. Expression of CXorf67 reduces H3 K27-me3 in HEK293 and NSCs and eliminating CXorf67 restores H3 K27-me3 in Daoy cells.

(A) Reduced H3 K27-me3 in HEK293 cells infected with a lentiviral vector expressing wildtype or mutant CXorf67. The population of infected HEK293 cells was enriched by flow sorting, and dual immunofluorescence with antibodies to CXorf67 (green) and H3 K27-me3 (red) demonstrates a reciprocal relationship between levels of these two proteins (DAPI – blue). (B) Immunoblotting of cell lysates from the experimental conditions in Figure 7A. A reduction of H3 K27-me3 follows expression of wildtype or mutant CXorf67. An increase in H3 K27-acetylation (H3 K27-ac) is evident following CXorf67 expression. (C) Immunoblotting showing similar results in neural stem cells infected with wildtype or mutant CXorf67. A reduction of H3 K27-me3 follows expression of wildtype or mutant CXorf67 and a slight increase in H3 K27-ac is evident following CXorf67 expression. (D) Daoy cells carrying vector with or without CRISPR guide 1 or 2. With an empty vector, Daoy cells express CXorf67, but not H3 K27-me3. Introduction of either CRISPR guide into some cells produces combined lack of CXorf67 expression and elevated H3 K27-me3 expression (arrows). (E) In Daoy cells flow-sorted to contain only those infected with vector, immunoblotting confirms knock-out of CXorf67 in those containing either CRISPR guide. Concomitantly, H3 K27-me3 expression is restored, while H3 K27-ac is reduced. HEK293 cells are included as a control (composite image from blots run in parallel). Elimination of CXorf67 expression reduced the growth of Daoy cells by more than half, as illustrated by a change in relative luminescence (F).