Unsupervised Consensus Clustering of Spinal Ependymoma Gene Expression Signatures

SUPPLEMENTARY FIGURE LEGENDS:

Supplementary Figure 1:

Unsupervised consensus clustering of spinal ependymoma gene expression signatures

(A)  Unsupervised consensus hierarchical clustering of 35 spinal ependymoma gene expression signatures from a rank 2 to 4 classification.

(B)  Unsupervised consensus Non-Negative Matrix Factorization of 35 spinal ependymoma gene expression signatures from a rank 2 to 4 classification.

Supplementary Figure 2:

Significant somatic copy number alterations in spinal ependymoma

(A)  Significant focal amplifications in spinal ependymoma detected by GISTIC2. Highlighted in red are amplifications encompassing the entire gene.

(B)  Significant focal deletions in spinal ependymoma detected by GISTIC2. Highlighted in blue are deletions encompassing all or portions of a gene.

(C)  Examination of focal amplifications encompassing C15ORF54, with gains indicated in red, neutral copy number in white, and genomic loss in blue. Grade II (orange) and myxopapillary (purple) sample classification is indicated in the top panel.

Supplementary Figure 3:

HIF-1α mRNA transcript and HIF-1α transcriptional targets are up-regulated in spinal MPE

(A)  Boxplot of HIF-1α expression in spinal Grade II and Myxopapillary ependymoma. Statistical comparison between the two groups was assessed using a two-sided wilcoxon-rank sum test.

(B)  Single-sample gene set enrichment analysis demonstrating significant enrichment of the HIF-1α pathway and its transcriptional targets by gene expression profiling. Statistical comparison between the two groups was assessed using a two-sided wilcoxon-rank sum test.

Supplementary Figure 4:

Western blot and functional assays supporting a Warburg phenotype in myxopapillary ependymoma

(A)  Western blot analysis of numerous metabolic associated proteins including HIF1a, HK2, HK1, LDHA, PDK1, PDK2, pPDHE1a, and PDHE1a in normal adult and fetal spine, Grade II (n=6) and Myxopapillary (n=6) ependymoma.

(B)  Linear quantification of protein expression shown in Figure S4A for LDHA protein.

(C)  Linear quantification of protein expression shown in Figure S4A for the PDK2 protein.

(D)  Lactate measurement in normal adult and fetal spine, Grade II (n=6) and Myxopapillary (n=6) ependymoma.

Supplementary Figure 5:

Evaluation of TCA cycle activity in spinal myxopapillary ependymoma through phosphorylation of pyruvate dehydrogenase E1A.

(A)  Demonstration of phosphorylation of PDHE1A relative to total PDHE1A as determined by western blot analysis.

(B)  Linear quantification of pPDHE1a levels relative to total PDHE1a demonstrated in Figure S3C.

Supplementary Figure 6:

Glycolysis pathway in spinal myxopapillary ependymoma and putative targets for therapy

A schematic of the glycolytic pathway in myxopapillary ependymoma illustrating deregulation of proteins and metabolites and possible targets for therapeutic intervention.

SUPPLEMENTARY TABLE LEGENDS:

Supplementary Table 1:

Clinical and demographic details for spinal ependymomas analyzed in this study including molecular subgroup, histological grade, gender, and age.

Supplementary Table 2:

Antibodies used for western blot analysis in this study and their associated concentrations and manufacturers.

Supplementary Table 3:

Differentially expressed genes as calculated by a signal to noise ration between spinal myxopapillary ependymoma and Grade II spinal tumors.

Supplementary Table 4:

Statistically significant amplifications detected by GISTIC2 in spinal ependymoma. Copy number data was analyzed using a circular binary segmentation algorithm, median centering, then analysis using GISTIC2.

Supplementary Table 5:

Statistically significant deletions detected by GISTIC2 in spinal ependymoma. Copy number data was analyzed using a circular binary segmentation algorithm, median centering, then analysis using GISTIC2.

Supplementary Table 6:

Enriched pathways identified by Gene Set Enrichment Analysis in Grade II spinal ependymoma using datasets from MSIGDB Broad Pathway database (C2 and C5).

Supplementary Table 7:

Enriched pathways identified by Gene Set Enrichment Analysis in spinal myxopapillary ependymoma using datasets from MSIGDB Broad Pathway database (C2 and C5).