For this study, cores of tissue were obtained from the primary tumor and metastases from patients with metastatic renal cell carcinoma at the time of nephrectomy (n=6). Each of these cores (5-6 for the primary tumor and 1-3 for metastases) were subdivided into 3 mm sized fragments and were implanted into the chorioallantoic membrane of chick embryos at Day 9 of embryonic development. On Day 11, PDXs were treated topically with vehicle or sunitinib (10 uM final in dimethyl sulfoxide [DMSO]) every day until Day 17 of embryonic development. At the endpoint (Day 19), all PDXs were submitted to high-frequency ultrasound imaging to quantitate differences in tumor blood perfusion and tumor volume between treatment groups. Finally, after imaging, all PDXs were individually submitted to total exome sequencing.
This approach was applied to 1548 tumor regions from six renal cell carcinoma patients, achieving a 93.6% engraftment rate. Genomic DNA was profiled by whole-exome sequencing on tumor biopsies both before and after implementation and drug treatment. Data from the engrafted samples were aligned with both human and chicken genomes to remove contaminating sequences. Mutational profiling was performed to assess the intra-tumoral heterogeneity, resulting in the identification of 100 unique SNVs. No variants were identified with sufficient confidence in primary tumor regions. A recurrence analysis confirmed the genomic heterogeneity of spatially distinct tumors regions. The profile of mutated genes for each tumor sample demonstrated that only six genes were found to contain variants in multiple regions of the tumor. Variant allele frequency revealed the branched expansion of subclones and variants associated with sunitinib resistance. These subclones exhibited a branched expansion pattern, with two synonymous variants in PBRM1 identified as early mutational events.
Power concluded that this study established a rapid and efficient new method of surveying functional tumor heterogeneity in the context of drug resistance. Within 8-10 days of a drug challenge, various regions of the primary tumor and metastases can be determined to be resistant or sensitive to targeted therapy. Interestingly, no outright identifiable molecular signature correlated to drug resistance, however, this phenotype-based readout was superior to various prognostic scoring criteria systems. Upon validation, this PDX-based model may be used to predict de novo drug resistance in patients with metastatic renal cell carcinoma.
Presented by: Nicholas Power, MD, Western University, London, Ontario, Canada
Co-Authors: Matthew Lowerison2, Yaroslav Fedyshyn2, Karla Williams4, Ann Chambers1, James Lacefield1, Paul Boutros3, Hon Leong2.
1. Surgery, Western University, London, ON, Canada
Written By: Zachary Klaassen, MD, Urologic Oncology Fellow, University of Toronto, Princess Margaret Cancer Centre Twitter: @zklaassen_md at the 73rd Canadian Urological Association Annual Meeting - June 23 - 26, 2018 - Halifax, Nova Scotia