Abstract 3900: Integrating single-molecule localization microscopy and computational pipelines to elucidate ecDNA dynamics

Abstract Extrachromosomal DNA (ecDNA) is prevalent across various cancer types, notably Glioblastoma Multiforme (GBM), where it facilitates the rapid acquisition of high oncogene copy numbers by cancer cells. These ecDNAs exhibit self-replication and are randomly distributed to daughter cells during mitosis, resulting in uneven segregation. This leads to variations in oncogene copy numbers among subclones, contributing to significant intratumoral heterogeneity. Traditional sequence-based analyses are insufficient for addressing spatial heterogeneity. To overcome this limitation, we propose utilizing sequential OligoSTORM imaging combined with single-molecule localization microscopy (SMLM), an innovative hybridization technique, and a computational pipeline. Our hybridization method significantly enhances imaging speed: whereas imaging approximately 40 loci with typical sequential OligoSTORM requires 11 days to obtain data from about 20 cells, this method enables imaging of 104 loci within the same time frame. Additionally, we developed a computational pipeline to complement this technique. Together, these tools enable 3D imaging of chromosomes with a high resolution of 20 kb. Our hybridization approach allows imaging at both 100-kb and 20-kb resolution steps, organized hierarchically in a Matryoshka-like structure. Our pipeline incorporates advanced computational algorithms, including iterative closest point (ICP) and space-partitioning data structures, to precisely align each 20-kb sub-step to its corresponding 100-kb step. Using this methodology, we target the EGFR gene locus with high resolution. To validate our approach, we are applying it to a non-cancerous microglial cell line (HMC3) as a control and a patient-derived glioblastoma stem-like cell (GSC) line as our GBM model. The GSC line, rich in ecDNA content with an average of 18 ecDNAs per cell, demonstrates pronounced tumor heterogeneity, including significant EGFR amplifications among subclones. Through this methodology, we can elucidate the structural organization of ecDNAs containing oncogenes and their regulatory elements with fine precision. Furthermore, our approach seamlessly integrates these structural insights with matched single-cell genomic data. This combined strategy offers a comprehensive understanding of ecDNA dynamics in cancer, shedding light on tumor heterogeneity at both spatial and genomic levels and paving the way for novel insights into oncogene-driven cancer biology. Citation Format: Mesut Unal, Jessica Alsing, Gabrielle S. Dewson, Guy Nir, Kadir C. Akdemir. Integrating single-molecule localization microscopy and computational pipelines to elucidate ecDNA dynamics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3900.