Integrated genomic and functional analysis of patient-derived sarcoma cell lines validates their use for candidate treatment discovery including knockdown of a novel BCOR fusion gene.

Abstract

e23572 Background: Malignant peripheral nerve sheath tumors (MPNST) are rare, biologically-aggressive soft tissue sarcomas of neural origin, characterized by high risk of recurrence, metastases, treatment resistance and lethality. Preclinical models accurately reflecting underlying genetics are needed for exploring the tumor’s biology and developing novel therapeutic approaches. Using genomic, biochemical and bioinformatic approaches we generated and deeply characterized a patient-derived MPNST cell-line model (PDCL) and functionally validated identified therapeutic targets including a novel BCOR fusion gene. Methods: Under IRB consent, two geographically distinct metastatic tumor samples were obtained from a right peritoneal nodule (PN) and the pelvis (P) of a 63-year-old woman, with recurrent uterine MPNST to generate PDCL. Whole exome (WES) and RNA sequencing was performed on all samples. Data and bioinformatic analysis was performed using the QIAGEN software suite (CLC Genomic Workbench, Ingenuity Variant and Pathway Analysis, OmicSoft-ArrayStudio). Results: To define the degree of shared genomic diversity, WES and RNAseq of tumors and PDCL was performed. WES of metastatic samples identified both shared (n = 39) and novel (98 and 116, respectively) cancer driver variants, which were also present in the paired PDCL (38% and 39%, respectively). Furthermore, somatic copy number alterations (SCNAs) were also shared between tissue and paired PDCL. These SCNAs are present in the MPNST TCGA dataset suggesting broad biologic significance. Finally, a novel and expressed BCOR-CREBRF fusion was present in all samples. With the potential to identify and validate candidate therapies, bioinformatic analysis was used to identify shared canonical signaling pathways (Ephrin receptor signaling) and upstream regulators (HBEGF, VEGFA, and ERBB2) between tumors and PDCL. Based upon this analysis, the inhibitory effects of a targeted list of candidate drugs was performed: Dasatinib, Pazopanib, PD153035, and Lapatinib all selectively resulted in tumor cell death. Moreover, siRNA-mediated BCOR-CREBRF knockdown significantly inhibited PDCL growth. Conclusions: We demonstrate that MPNST PDCL capture the genomic diversity inherent in their founder tumor tissues and potentially other MPNST. Moreover, using a bioinformatic approach for therapeutic discovery, demonstrate the use of these PDCL for testing treatment candidates for possible use in MPNST patients.