(I-346) Migration Based Sorting of Breast Cancer Cells Reveal Bi-Directional Relationship with Epithelial to Mesenchymal Phenotypes

As of 2023, the average risk of a woman in the US developing breast cancer is 13%. The chances of their survival greatly decrease once metastasis occurs as cancer cells migrate from the primary tumor through the stroma, ultimately forming secondary sites. A previous study from our lab has shown that cancer cells can be sorted into subpopulations based on their migratory capabilities. In this study there was a positive correlation between highly migratory cells and EMT phenotype. This led us to investigate how differences in migration can be associated with EMT. Literature has shown that with increased EMT, there is increased cell migration, but little is known if sorting by migratory behavior can influence EMT phenotype.  The aim of this study is to focus on the role of migratory sorting on EMT phenotype and its crosstalk with migration. Utilizing our migration-based subpopulations we hypothesized that altering a weakly migratory cell’s mesenchymal properties may lead to a change in its ability to migrate and become more similar to their highly migratory counterpart. This was tested by utilizing 3D collagen gels, and treatment with TGFb to induce EMT. 

Parental MDA-MB-231 cells were sorted using consecutive transwell assays, resulting in highly and weakly migratory cells (Fig. 1A). RNA sequencing (RNAseq) was conducted at Vanderbilt Technologies for Advanced Genomics Core. Cell images for aspect ratios were taken on a 10x magnification on a Zeiss Axio Observer Z1 Inverted Microscope and quantified using ImageJ processing. For3D migration assays cells were seeded in collagen gels with a concentration of 3mg/mL 24 hours prior to the 12-hour timelapse. For the TGFb trials the same format was followed as previously described with the addition of TGFb 24 hours before seeding.

Using our sorting method (Fig. 1A) that utilizes a series of transwell assays, we sorted a highly and weakly migratory subpopulation. To confirm their migratory differences, we put them in 3D collagen gels and assessed their migration over 12 hours. The highly migratory subpopulation exhibited a higher mean velocity during trials than the weakly migratory population (Fig. 1B). Meaning that the highly migratory subpopulations migrated farther during that time. The two subpopulations were examined for genotypic and phenotypic characteristics. Utilizing RNAseq, we quantified the average number of mesenchymal genes divided by the average number of epithelial genes to create an EMT score (Fig. 1C), and individual cell aspect ratios were taken (Fig. 1D). The highly migratory subpopulation displayed a higher EMT score and mean aspect ratio than their counterparts. This means there are elevated levels of EMT in the highly migratory subpopulation. To promote EMT in the weakly migratory subpopulation TGFb was added at a concentration of 5 ng/mL. Aspect ratios were quantified 24 hours post treatment. The treated weakly migratory cells displayed a higher mean aspect ratio than the water-treated control group (Fig. 1E). When the migratory capabilities of the TGFb treated group and water treated group were compared, the TGFb group had a higher mean velocity in 3D collagen (Fig. 1F). Ultimately, our work indicates that migration-based sorting also indirectly sorts for EMT phenotype in breast cancer cells, suggesting that there is a bi-directional relationship between migration and EMT state.