An in vitro CRISPR screen of cell-free DNA identifies apoptosis as the primary mediator of cell-free DNA release

ABTRACT

Clinical circulating cell-free DNA (cfDNA) testing is now routine, however test accuracy remains limited. By understanding the life-cycle of cfDNA, we might identify opportunities to increase test performance. Here, we profile cfDNA release across a 24-cell line panel and utilize a cell-free CRISPR screen (cfCRISPR) to identify mediators of cfDNA release. Our panel outlines two distinct groups of cell lines: one which releases cfDNA fragmented similarly to clinical samples and purported as characteristic of apoptosis, and another which releases larger fragments associated with vesicular or necrotic DNA. Our cfCRISPR screens reveal that genes mediating cfDNA release are primarily involved with apoptosis, but also identify other subsets of genes such as RNA binding proteins as potential regulators of cfDNA release. We observe that both groups of cells lines identified primarily produce cfDNA through apoptosis. These results establish the utility of cfCRISPR, genetically validate apoptosis as a major mediator of DNA release in vitro, and implicate ways to improve cfDNA assays.

NOTCH1 PEST domain variants are responsive to standard of care treatments despite distinct transformative properties in a breast cancer model

Activating variants in the PEST region of NOTCH1 have been associated with aggressive phenotypes in human cancers, including triple-negative breast cancer (TNBC). Previous studies suggested that PEST domain variants in TNBC patients resulted in increased cell proliferation, invasiveness, and decreased overall survival. In this study, we assess the phenotypic transformation of activating NOTCH1 variants and their response to standard of care therapies. AAV-mediated gene targeting was used to isogenically incorporate 3 NOTCH1 variants, including a novel TNBC frameshift variant, in two non-tumorigenic breast epithelial cell lines, MCF10A and hTERT-IMEC. Two different variants at the NOTCH1 A2241 site (A2441fs and A2441T) both demonstrated increased transformative properties when compared to a non-transformative PEST domain variant (S2523L). These phenotypic changes include proliferation, migration, anchorage-independent growth, and MAPK pathway activation. In contrast to previous studies, activating NOTCH1 variants did not display sensitivity to a gamma secretase inhibitor (GSI) or resistance to chemotherapies. This study demonstrates distinct transformative phenotypes are specific to a given variant within NOTCH1 and these phenotypes do not correlate with sensitivities or resistance to chemotherapies or GSIs. Although previous studies have suggested NOTCH1 variants may be prognostic for TNBC, our study does not demonstrate prognostic ability of these variants and suggests further characterization would be required for clinical applications.

Hierarchical tumor heterogeneity mediated by cell contact between distinct genetic subclones

Intratumor heterogeneity is an important mediator of poor outcomes in many cancers, including breast cancer. Genetic subclones frequently contribute to this heterogeneity; however, their growth dynamics and interactions remain poorly understood. PIK3CA and HER2 alterations are known to coexist in breast and other cancers. Herein, we present data that describe the ability of oncogenic PIK3CA mutant cells to induce the proliferation of quiescent HER2 mutant cells through a cell contact-mediated mechanism. Interestingly, the HER2 cells proliferated to become the major subclone over PIK3CA counterparts both in vitro and in vivo. Furthermore, this phenotype was observed in both hormone receptor-positive and -negative cell lines, and was dependent on the expression of fibronectin from mutant PIK3CA cells. Analysis of human tumors demonstrated similar HER2:PIK3CA clonal dynamics and fibronectin expression. Our study provides insight into nonrandom subclonal architecture of heterogenous tumors, which may aid the understanding of tumor evolution and inform future strategies for personalized medicine.