LAUSR

Activated platelet derived growth factor receptor alpha (PDGFRα) is essential for the maintenance of oligodendrocyte progenitors. Differential signaling results from phosphorylation of specific tyrosine residues in the kinase insert (KI) domain present in most RTKs in variable length. Our goal is to identify the molecular contexts in which PDGFRA activation is an essential driver for maintenance of the malignant phenotype in glioblastoma. RNA sequencing was conducted on 13 glioblastoma mouse orthotopic PDXs (n=3/line). PDGFRA mRNA overexpression was observed for 3 glioblastomas: one carrying PDGFRA extrachromosomal (ecDNA) amplification (HF3253) and two non-amplified (Tukey p=0.41, 0.19). Corresponding PDGFRα KI activation (pY754) was observed only for HF3253. To assess activation as an adaptation to microenvironmental pressures, pY754 was quantified by reverse phase protein array on 4 growth conditions (normal, -growth factors, two serum concentrations) for 8 neurospheres. PDGFRɑ activation was highest in growth factor-depleted conditions in the absence of exogenous ligand (Tukey p=0.02). Taking advantage of genomic heterogeneity in HF3253, we isolated cellular lineages with divergent frequencies of PDGFRA ecDNA. PDGFRA copy number (TaqMan) decreased longitudinally. PDX tumors from higher passage had slower tumor growth due to strong selection for initial low frequency PDGFRA ecDNA-amplified clones. IHC showed that the tumors had comparably high total and activated PDGFRα to low passage tumors. Further exploiting intra-tumoral heterogeneity, we have isolated single cell clones. Tumor growth was reduced in 2 ecDNA(-) tumors compared to parental ecDNA(+) (log-rank test p= 3e-3, 1e-3). In contrast to parental, ecDNA(-) tumors demonstrated diffuse tumor morphology and weak PDGFRα activation. The staining index (mean intensity x area) for the ecDNA(-) clones was significantly lower than the ecDNA(+) (1-wayANOVA p=2e-16). We established dependency on PDGFRα signaling in a patient-derived glioblastoma model. Our data demonstrates that detection of ecDNA-amplified PDGFRA has the potential to be a predictive biomarker of future PDGFRɑ-targeted therapies.