Development of silica-based physical confinement models to isolate and study dormancy-prone cancer cells

Abstract

Metastatic cancers account for the majority of cancer-related deaths. Some metastatic tumors arise after a latent, disease-free period. The latency is attributed to cancer cells being in a dormant state that is eventually overcome, leading to metastatic progression. The ability to isolate dormant cancer cells to study and develop treatments to prevent relapse has remained an elusive goal. In this dissertation, a novel process to isolate and study dormancy-prone cells is presented. The process involves immobilizing cancer cells within a highly porous silica-poly(ethylene glycol) gel that physically confines cells. Two separate gelation models are presented. In the first model (SPEG), a distinct viability response was observed in which MCF-7, a dormancy-prone cell line, survived physical confinement significantly better than dormancy-resistant cell lines (MDA-MB-231, MDA-MB-468). Surviving MCF-7 cells were demonstrated to be in a reversible cell cycle arrested state akin to clinically observed single-cell dormancy. It was also found that tumor cells from breast and ovarian cancers that survived physical confinement were in a cell cycle arrested state. The second model, MSPEG, was developed as an improved system that allows efficient and viable cell extraction. In the MSPEG model, cells are first coated individually in a thin layer of agarose using flow-focusing microfluidic devices before encapsulation in a silica-PEG gel for protection during the extraction process. The microfluidic system conditions such as microfluidic device dimensions, flow rates, agarose concentration, oil, and surfactants were optimized to produce individually coated cells with high viability at high throughput levels. The agarose coating could be degraded to recover cells or in situ while in silica to awaken dormant cells. The silica-PEG composition was also re-engineered for better disintegration and facile silica separation from cells by modifying the molecular weight and type of PEG used and introducing iron oxide nanoparticles stabilized with fumed silica, respectively. The MSPEG model was evaluated as a clinically relevant dormancy model by examining the protein expression of p38 and ERK, the RNA expression of CDK2, cyclin D1, and cyclin E1. Additionally, we confirmed the cell cycle arrest observed was reversible by examining Ki-67 expression, senescence-associated factors, and proliferation of cells before and after physical confinement. The two models presented in this thesis can therefore be used to isolate and study dormancy-prone cells.