DNA-based copy number analysis confirms genomic evolution of PDX models

We previously reported the genomic evolution of the copy number (CN) landscapes of patient-derived xenografts (PDXs) during their engraftment and passaging1. Woo et al. argue that the CN profiles of PDXs are highly conserved, and that the main conclusions of our paper are invalid due to our use of expression-based CN profiles2. Here, we reassess genomic evolution of PDXs using the DNA-based CN profiles reported by Woo et al. We find that the degree of genomic evolution in the DNA-based dataset of Woo et al. is similar to that which we had previously reported. While the overall Pearson’s correlation of CN profiles between primary tumors (PTs) and their derived PDXs is high (as reported in our original paper as well), a median of ~10% of the genome is differentially altered between PTs and PDXs across cohorts (range, 0% to 73% across all models). In 24% of the matched PT-PDX samples, over a quarter of the genome is differentially affected by CN alterations. Moreover, in matched analyses of PTs and their derived PDXs at multiple passages, later-passage PDXs are significantly less similar to their parental PTs than earlier-passage PDXs, indicative of genomic divergence. We conclude that genomic evolution of PDX models during model generation and propagation should not be dismissed, and that the phenotypic consequences of this evolution ought to be assessed in order to optimize the application of these valuable cancer models.

EPEN-42. MOLECULAR PROFILING REVEALS DISTINCT SUBGROUPS OF PEDIATRIC SPINAL EPENDYMOMA

Paediatric spinal ependymomas are important, albeit uncommon, malignant central nervous system tumours. Unlike adults, children with these tumours are likely to experience a more aggressive disease course, with a higher rate of local failure and a higher rate of metastases. The clinical and molecular factors underlying these differences remain poorly characterized. We analyzed spinal ependymoma (SEPN) tumour samples from 27 paediatric patients (female: 11, male: 15; age range: 4–18 years) using genome-wide DNA methylation profiling, copy-number analysis, as well as transcriptome profiling. Using DNA methylation profiles, two distinct unsupervised consensus-clustering approaches, hierarchical clustering and non-negative matrix factorization reliably identified two subgroups. These subgroups were designated as Myxopapillary ependymomas (SP-MPE) and spinal ependymomas (SP-EPN) based on the online Classifier tool (MNP2.0). The genome-wide copy-number analysis showed differences in numbers and pattern of copy-number alterations between these groups. The gain of chromosome 20 (39%) followed by loss of chromosomes 6 (28%), 10 (28%), and 13 (28%) were detected in the SP-MPE group, whereas loss of chromosome 22 was frequent (60%) in the SP-EPN group. Transcriptomics analysis showed that genes associated with oxidative phosphorylation, TCA cycle components, electron transport, and Interferon-gamma production characterize the SP-MPE group whereas potassium ion import and regulation of astrocyte differentiation characterize the SP-EPN group. Western blot analysis validated the increased protein expression of oxidative phosphorylation complexes in SP-MPE. With this study, we provide a foundation for further molecular characterization of pediatric SEPN subgroups. Our results suggest that mitochondrial oxidative phosphorylation may drive the regulation of energy metabolism of SP-MPE tumours.