The goals of this research project are to: 1) elucidate the role of the hypoxia-inducible transcription factor (HIF-1a) in the behavior of two different tumor models, 2) model the transport of oxygen and nutrients in these tumor models using a sandwich culture system designed to mimic the tumor microenvironment, and 3) translate the in vitro experimental results into an explanation for the in vivo behavior of tumors containing cells lacking HIF-1a.

ES cells with inactivated HIF-1a genes (HIF-1a^-/-) and wild type controls (HIF-1a^+/+) were used for all experiments. The HIF-1a^-/- and HIF-1a^+/+ cells were injected into a dorsal skin window of a mouse to develop an ES-cell derived tumor, and the tumor was visualized using intravital fluorescence microscopy. For in vitro testing, co-cultures of the two cell types were grown in a sandwich culture system consisting of the cells grown on a glass slide with another glass slide placed on top. The combination of diffusion and consumption creates a gradient of oxygen and nutrients such that the center of the culture system experiences hypoxic and hypoglycemic conditions.

In vivo, the embryonic stem cells lacking HIF-1a showed a preference for regions of the tumor distant from vasculature, where oxygen and nutrient concentrations are lowest. Likewise, in co-cultures of the two cell types, the cells formed clumps with the HIF-1a^-/- cells primarily in the center. By growing the co-cultures under conditions promoting differentiation, a pseudo-monolayer of cells was obtained consisting of clumps of predominantly HIF-1a^-/- cells with interspersed monolayers of predominantly HIF-1a^+/+ cells. When these co-cultures were placed in the sandwich culture system, the interspersed monolayer of HIF-1a^+/+ cells disappeared in the central hypoxic region. Currently, studies are underway to determine whether the preference for hypoxic regions of the HIF-1a^-/- cells is due to apoptosis, migration, or another mechanism.