Poster S19 - Mathematical Modeling of Chromosome Loss in 2D and 3D Spheroids: An Exploration of Luria-Delbruck Statistics

Saturday, October 26, 2024

10:00 AM – 11:00 AM EST

Location: Exhibit Hall E, F & G

Presenting Author(s)

Markus T. Sprenger (he/him/his)

Ph.D. student
University of Pennsylvania, United States

Co-Author(s)

AA

Alişya Anlaş, PhD

Postdoctoral Researcher
University of Pennsylvania
Philadelphia, Pennsylvania, United States

Primary Investigator(s)

Dennis Discher, PhD

Robert D. Bent Professor, Director Phys Sci Onc Ctr/Proj
Biophysical Eng'g Labs, University of Pennsylvania
Philadelphia, Pennsylvania, United States

Co-Author(s)

SP

Steven Phan

PhD Student
University of Pennsylvania, United States

Introduction: Chromosome gains and losses, a common feature of solid tumors, can be variable and heritable thereby contributing to genomic instability. Although chromosome losses are rare, they arise early during cancer progression, and play a crucial role in carcinogenesis. Chromosome loss parallels with the seminal work of Luria-Delbruck (LD), who applied fluctuation analysis to bacterial mutations, revealing them to be rare, random and heritable. Here we adapt the LD model to explore chromosome loss in 2D cell cultures as well as in spheroids aiming to approximate chromosome loss rates under altering conditions. By evaluating chromosome loss under various experimental conditions, we identify scenarios in which the chromosome loss event is more prevalent. Our findings offer a quantitative framework for investigating this critical aspect of tumorigenesis that contributes to chromosomal instability.

Materials and

Methods: H23 lung adenocarcinoma cells were engineered using CRISPR/Cas9 to edit constitutively active Lamin-B1 with GFP on one allele of Chromosome 5. Loss of GFP signal is therefore indicative of chromosome 5 loss. 2D and 3D spheroids of engineered H23 cells were sampled after various days and taken for flow cytometry analysis. Following from multiple samples, the percent GFP negative (indicating chromosome loss) and its variation was plotted in respects to the number of days in culture. Starting with LD's work a model was derived for both the large and small chromosome loss rate case, to model the GFPneg and its variation for 2D cultures and spheroids allowing for a measurement of the chromosome loss rate. In addition, H23 and a second A549 lung adenocarcinoma cell line, with similar to H23 Lamin-B1 editing with RFP, was treated with reversine and blebbistatin for 24 hours. The effects of each drug on chromosome loss rate were determined for 2D cultures using LD methodology and widefield fluorescent microscopy imaging. The 2D cultures, being imaged with widefield fluorescent microscopy, could be examined for the spatial relation of GFP and RFP negative cells and colonies of H23 GFPneg and A549 RFPneg cells could be identified.

Results, Conclusions, and Discussions: The LD model, originally developed to demonstrate that genetic mutations in bacteria are random and heritable, has been employed in our study of cancer cells to analyze chromosome loss in lung adenocarcinoma cancer cells. We employed sampling methods rather than whole population analysis to assess data from both 2D and 3D experiments. Our results reveal that although the 2D and 3D assays initially appear to follow different trends, their chromosome loss rates align with LD’s high or low mutation rate predictions. Therefore, our data support the use of fluctuation analysis for estimating chromosome loss rates in cancer cells. Further, the detection of colonies exhibiting chromosome loss provides evidence that chromosome loss is heritable. An increase of chromosome loss rate was seen for reversine, but not for blebbistatin treated cells indicating that the chromosome loss is environment dependent. The increase for chromosome loss rate with the addition of reversine was as high as 40 fold for A549 and 8 fold for H23. Notably, the transition from 2D to 3D cultures resulted in a substantial increase in chromosome loss rate, sufficient to alter the LD model’s assumption from rare to infrequent chromosome losses. These findings validate LD’s model on random mutations as applicable to chromosome loss in cancer cells and highlight the impact of microenvironment and treatment conditions on chromosomal instability.

Acknowledgements (Optional):