Effects of cell packing on chemoattractant distribution within a tissue

Diffusible signals provide critical information to cells in biological systems, often in a concentration-dependent manner. In animal development, such signals can determine different cell fates or guide motile cells to their proper locations. It is well-known that migrating cells respond to graded chemoattractant cues by moving toward areas of higher concentrations. However, it is not clear how cell-dense animal tissues impact the distribution of chemoattractants in three dimensions. We leverage the simple architecture of the Drosophila egg chamber to explore this idea. In this context, sixteen large germline cells are packed together, enveloped by a somatic epithelium. A small set of epithelial cells, the border cells, form a motile cell cluster and respond to guidance signals by moving across the egg chamber during oogenesis. We created a geometrically-realistic model of the egg chamber and determined the distribution of the chemoattractants through that domain using a reaction-diffusion system. We used this information to determine reasonable biophysical parameters of chemoattractant that would facilitate gradient formation in the appropriate developmental time, and to explore the effects of different secretion locations in the egg chamber. Our model revealed several interesting features: The chemoattractant is more concentrated and the gradient sets up more quickly in a cell-packed space, and cell packing creates dips in the concentration and changes in gradient along the migratory path. We simulated migration with our calculated chemoattractant gradient and compared it to that with a constant gradient. We found that with our calculated gradient migration was slower initially than in the constant gradient, which could be due to the exponential nature of the gradient or other variation in signal due to the heterogeneous domain. Given the many situations in which cell migration occurs in complex spatio-temporal environments, including development, immune response, and cancer metastasis, we believe modeling chemoattractant distribution in heterogeneous domains is widely relevant.