Abstract AP19: SINGLE–CELL RNA–SEQUENCING OF PATIENT–DERIVED OVARIAN CANCER CELLS AND PATIENT–DERIVED XENOGRAFT MODELS

Abstract

Abstract BACKGROUND AND PURPOSE: Genetic heterogeneity is a hallmark of ovarian cancer (OvCa) biology and underlies treatment resistance. Macroscopic and or treatment resistant microscopic residual disease (MRD) after debulking surgery and chemotherapy are the source for disease recurrence and death in these patients. Current profiling methods are unable adequately reflect heterogeneity, and transcriptional programs associated with treatment resistance and MRD remain elusive. To elucidate transcriptional heterogeneity and potential mechanisms of drug-resistance, we isolated OvCa cells from patients with malignant ascites using flow-cytometry. We applied single-cell RNA-sequencing (sc-RNA-seq) to ascites-derived cells from patients with OvCa. We picked OvCa spheroids and profiled these separately. To investigate MRD, we used three PDX-models stably expressing mCherry, treated with carboplatin, and harvested tumor cells for sc-RNA-seq at three time points (pre-treatment, at time of MRD as determined by bio-luminescence imaging, and disease relapse). SUMMARY OF RESULTS: We successfully sequenced 770 single-cell transcriptomes from 6 individuals with treatment-resistant OvCa, including 3 patients with sequential samples. We mapped the landscape of chromosomal aberrations by inferred large-scale copy number variations (CNVs) at a single-cell level. We observed significant inter-tumor heterogeneity and started to deconstruct the genomic architecture of individual patients. Using experimentally validated gene sets, we determined the cell cycle state of individual cells and identified transcriptional programs related to the cell cycle as significant bias in publically available bulk RNA-sequencing data. Principal component analysis revealed the expression of a stem-ness signature, including CD133, ALDH1A and AXL, in a sub-set of non-cycling cells. An important driver of transcriptional heterogeneity common to patients included in this study was the expression of gene sets related to inflammatory pathways, such as the NFkB and JAK/STAT pathways. Hierarchical clustering of 42 spheroid profiles identified four major clusters, including a highly “inflamed” phenotype. Therapeutic inhibition of the STAT pathway abrogated the capacity of spheroid formation on an ultra-low attachment surface, indicating its importance for metastasis. We have successfully isolated thousands of individual cells for single-cell profiling from PDX models treated with carboplatin. These cells were collected at three time points, including at the MRD stage. We have successfully sequenced 100 cells from this collection and were able to generate whole-transcriptome data comparable to that of freshly isolated patient cells and thousands of single cells are currently undergoing sequencing. CONCLUSION: We have successfully applied single-cell RNA-sequencing to patient-derived ovarian cancer cells and PDX-models. Single-cell transcriptomes enabled inference of genomic information, genetic and transcriptional heterogeneity, cell cycle state, and programs related to stem-ness and inflammation, providing a unique and comprehensive perspective on ovarian cancer cell states. Ongoing profiling of carboplatin-resistant cells captured at the minimal residual disease stage in PDX-models will provide a unique opportunity to understand treatment resistance which ultimately leads to cancer recurrence. Citation Format: Benjamin Izar, Elizabeth Stover, Itay Tirosh, Asaf Rotem, Parin Shah, Chris Rodman, Sanjay Prakadan, Marc Wadsworth, Mei-Ju Su, Rachel Leeson, Sangeetha Palakurthi, Joyce Liu, Ursula Matulonis, Alex Shalek, Orit Rozenblatt-Rosen, Aviv Regev, Levi Garraway. SINGLE–CELL RNA–SEQUENCING OF PATIENT–DERIVED OVARIAN CANCER CELLS AND PATIENT–DERIVED XENOGRAFT MODELS [abstract]. In: Proceedings of the 11th Biennial Ovarian Cancer Research Symposium; Sep 12-13, 2016; Seattle, WA. Philadelphia (PA): AACR; Clin Cancer Res 2017;23(11 Suppl):Abstract nr AP19.