ABSTRACT Increased oncogene expression mediated by focal amplifications is a common mechanism for oncogene activation in human cancers. Two major mechanisms leading to oncogene amplification have been described: chromosomal amplification and non-chromosomal amplification. The latter mechanism is characterized by the presence of multiple copies of circular DNAs that are thought to originate following the fragmentation and subsequent circularization of pieces of chromosomes. These “extrachromosomal circular DNAs” (ecDNAs) have long been known as “double minutes” for their appearance in metaphase spreads and by the lack of centromeric sequences. In the past few years renewed interest in this class of cancer-associated chromosomal rearrangements has been fueled by technological advances and by the realization that, due to their random segregation at mitosis, ecDNAs can accelerate tumor evolution, mediate drug resistance, and generally promote a more aggressive phenotype. Despite substantial progress, however, several key questions regarding the biology of ecDNAs, their dynamics during the early stages of tumor formation, and their contribution to tumor initiation and progression, remain unanswered. This is in part due to the lack of effective means to engineer and track ecDNAs in normal cells and in model organisms. Our group has extensive expertise in the generation and characterization of germline and somatic mouse models of human cancers, and we have pioneered the use of somatic genome editing to engineer chromosomal rearrangements in mice. In this grant application, supported by strong preliminary data, we describe a novel general strategy to model ecDNAs in cells and in mice. We have already generated three new genetically engineered mouse strains in which the formation of ecDNAs containing the oncogenes most commonly amplified in human cancers can be induced in a temporally and spatially controlled manner. Using a similar strategy, we have also generated cell lines in which formation of specific ecDNAs can be induced and tracked non-invasively using fluorescent reporters and selectable markers. We propose to use these innovative tools and reagents to address the following key questions: 1) Can oncogenic ecDNAs initiate tumor formation and/or accelerate tumor progression and metastasis in vivo? 2) How do oncogenic ecDNAs respond to changes in intracellular and extracellular environment? 3) Are there mechanisms preventing ecDNA formation and propagation in primary cells? 4) Is the presence of ecDNAs in cancer cells associated with unique therapeutically actionable vulnerabilities? 5) Can ecDNAs be transmitted horizontally between cells?