Hyaluronic Acid-Based Multilayer Films Regulate Hypoxic Multicellular Aggregation of Pancreatic Cancer Cells with Distinct Cancer Stem-Cell-like Properties.

In vitro spherical cancer models have been widely used in cancer stem cell (CSC) research, and the ability of CSCs to form multicellular colonies is recognized as a morphological marker. However, although several spherical/colony models share a common three-dimensional (3D) conformation, each model displays its own intrinsic properties. Thus, the CSC phenotypes with distinct multicellular aggregate morphologies must be defined and clarified. Here, a novel 3D model was designed to regulate the type of pancreatic CSC colonies that form using niche mimetic hyaluronic acid (HA)-based multilayer nanofilms and hypoxia. The multicellular aggregate morphology, CSC phenotypes, CSC-related marker expression, cell cycle, invasion, and drug resistance were determined. On the basis of the results of a cell morphology analysis, colonies formed on multilayer nanofilms in response to both normoxia and hypoxia, but round and island-type colonies, were investigated. Immunostaining results revealed a significantly higher expression of stem cell markers, such as OCT4, CXCR4, and CD44v6, in colonies that formed on multilayer nanofilms. These colonies also expressed higher levels of E-cadherin, hypoxia-inducible factor-1α, and vimentin, particularly the round-type colonies that formed on HA-based multilayer nanofilms, [poly(allylamine) (PAH)/HA]3, indicating that these colonies exhibit hybrid and metastable epithelial/mesenchymal phenotypes. Moreover, the cell cycle and invasion tests revealed that most of the cells in colonies growing on multilayer nanofilms showed a quiescent, slow cycling phenotype but displayed higher invasion after induction. Furthermore, a hypoxic environment strongly influences the drug resistance. This study describes a useful tool to investigate the diverse phenotypes of pancreatic CSC colonies and to study their regulatory factors that may benefit CSC research.