Algorithm to quantify spheres derived from cancer cell lines.

ANA BELÉN PEÑAHERRERA PAZMIÑO,; RAMIRO ISA-JARA, ; SILVIA GÓMEZ, ; ELSA HINCAPIÉ, ; DENISE BELGOROSKY,; EDUARDO IMANOL AGÜERO, ; ANA MARIA EIJAN; MAXIMILIANO PÉREZ, ; BETIANA LERNER

Congreso; Reunion Anual de Sociedades de Biociencias; 2023

Introduction: Sphere formation assay is accepted as a method for selecting and enriching cancer stem cells (CSCs). They play a crucial role in chemoresistance and cancer recurrence. Thus, there is intense research that aims at a better understanding of CSCs behavior. In our laboratory we study the growth of CSCs in microdevices to develop predicting chemotherapy assays in cancer. Counting spheres cultured in different devices is laborious and operator dependent. Objective: the aim of this work is to develop a computational program that identifies, counts and measures spheres automatically.Methods: The software named Spheres Interface has been developed using Phyton. It is a graphical user interface (GUI) mainly based on PySimpleGUI and SKIMAGE libraries to detect regions with spheres. Using an inverted microscope, 10X images of U251 human glioblastoma cell line were acquired. Then, the algorithm was trained to calculate the area in square micrometers, and it quantifies the number of spheres in each image. This algorithm has been evaluated with images that present different environments such as uniform and non-uniform background.Results: It has been observed that the performance of the algorithm is better when images present uniform background. When images have non-uniform background, artifacts are interpreted as spheres. A number of 100 images acquired from U251 cells has been processed with the algorithm and the sphere number reported has been compared with the manual count. The mean difference between the two methods and the standard deviation of the difference were estimated (mean± SD: 2±4, ns by t-student test). Also, the algorithm allows to number the spheres, which enables the monitoring of individual spheres in the time course.Conclusion: This algorithm has allowed to count spheres derived from cell lines. Although it needs to be improved, it can be a useful tool for automated CSC quantification from cancer cell lines and primary culture cells.