Development of a breast cancer spheroid model for use in drug discovery and cancer biology research

Abstract

The screening of lead compounds during cancer drug discovery still results in many hits that never reach the clinic. This is in part due to the distance between 2D cell culture, which is mostly used for screening assays and the complexity of an in vivo tumour setting. It is thus essential that screening setups are developed that better bridge the gap between in vitro and in vivo studies. Culturing cells in 3D has been shown to provide an architecture and gene expression profile that better resembles that found in tumours. A number of methods have been developed but many are laborious, require specialised equipment, or are prohibitively costly. In Southern Africa, we do not always have access to expensive machinery and reagents. Therefore, in this thesis, we aimed to set up a simple, easy to use, cost-effective 3D spheroid methodology for triple negative breast cancer that could be used in any laboratory with minimal reagents.

The non-metastatic triple negative breast cancer cell line, BT-20, was identified as the cell line most amenable to spheroid formation and measurement, compared to the metastatic cell lines MDA-MB-231 and MCF-7. Both these cell lines formed inconsistent, difficult to measure spheroids. A low attachment methodology was settled on to induce cell aggregation and spheroid formation. This proved to be an easy, low-cost, and reproducible method. Using the developed methodology, the efficacy of several novel compounds with anti-proliferative activity in 2D cancer cell line cultures were tested for their ability to affect cell survival in the spheroid model.

Our studies show that one non-sulfamoylated 2-methoxyestradiol derivative, EE-15-one, caused a loss of cells in the 2D cell survival assay while it had no effect on BT-20 spheroids. In contrast, two other 2-methoxyestradiol derivatives had similar effects in 2D and 3D. Another novel anticancer compound, STX1972, also had anti-proliferative capabilities in 2D which were lost in cells grown in 3D. As a result, our screening approach on a small number of samples was able to identify 2 of 5 compounds that had no effect on cancer cell growth in 3D. Further analysis of EE-15-one suggests that this compound does not inhibit cell cycle progression like the other derivatives tested but instead inhibits cell adhesion. Our data shows that integrin-based adhesion is replaced by cadherin-dependent cell-cell adhesion in spheroids. Changing the mode of adhesion correlates strongly with the efficacy of this compound suggesting that it is an inhibitor of integrin dependent cell adhesion. Further analysis show that the efficacy of this compound was not dependent on hypoxia, further strengthening the suggestion that it directly acts on cell adhesion.

It is commonly accepted that a tumour consists of a heterogeneous mix of subpopulations of cancer cells, each with altered genetic backgrounds. Furthermore, such subpopulations can affect each other changing behaviours of neighbouring cells. To replicate this scenario in vitro, we initiated the development of a co-culture spheroid system of different triple negative breast cancer cell lines to investigate how they behave within a spheroid. Our data shows that when a co-culture spheroid system of BT-20 and MDA-MB-231 cells is generated, the cells with migratory ability were able to migrate away from the spheroid onto rigid surfaces. In contrast, the non-metastatic BT-20 cells remained within the confines of the spheroid. This suggests that indeed different cell populations will continue to behave differently within a 3D cell culture setting.

In conclusion, we have developed a robust, cost-effective 3D culture system that has shown great potential for used in high throughput screening of novel anti-proliferative compounds. We have shown that these spheroids use different adhesive strategies than their counterparts kept in 2D. This suggests that these cells also change their cellular behaviour which is essential for better mimicking the in vivo tumour setting.