By Lina Sorg
Cell migration is crucial in all multicellular species and contributes to development, homeostasis, and disease progression. While researchers often study and characterize the movement of individual cells, the collective motion of cell clusters through heterogeneous environments is less clear. During a minisymposium at the 2018 SIAM Annual Meeting, currently taking place in Portland, Ore., Bradford E. Peercy of the University of Maryland, Baltimore County, examines clustered cell migration in the egg chamber of Drosophila melanogaster, the common fruit fly. “Drosophila as a biological system provides a beautiful set of homologous genes, and has a nice correlation to what humans experience in their genes,” Peercy said. “We have the tools to model the genetic pieces that go into the mechanisms of cell migration in Drosophila.”
Multicellular egg chambers develop in the fly’s abdomen. Small somatic cells surround the outside of the chamber, while large germline (nurse) cells are inside, with an oocyte cell located at the end. During egg development, different cell types coalesce into a group and traverse among the germline cells in the center of the chamber. Polar cells at the very tip of the chamber secrete a chemical signal that triggers neighboring epithelial cells to begin moving towards the oocyte. Intercellular signaling involves the JAK-STAT signaling pathway, a chain of interactions between cell proteins. Peercy examines two processes that occur in the fruit fly egg chamber: (i) secretion-induced symmetric epithelial cell activation, which transpires prior to cell clustering, and (ii) chemoattractant distribution, which directs clustered cell migration.
Drosophila egg chambers offer an excellent viewing point for modeling the progression of cell migration behaviors. “How we model this system depends on the questions we’re asking,” Peercy said. “Researchers have focused on questions spanning initial activation of cell clustering to full cluster migration.” Possible queries include the following:
“When it comes to intracellular space, we consider a well-mixed volume and typically work with ordinary differential equations,” Peercy said. Meanwhile, the extracellular space around the cells has a significant impact on migration. For this area, he considers a continuum three-dimensional space to implement reaction-diffusion partial differential equations. However, the space’s thinness can yield two-dimensional surface-like approximations. Boundary conditions—due to actions occurring on the cells’ surface—are either no-flux or mixed. Peercy determines that an agent-based framework best handles cluster migration. “The discrete nature of the cells dissuaded us from using a continuum approximation for the cellular density,” he said.
Traditionally, animations of Drosophila cell motion display a lateral perspective of clustering cells coming together and making their way through the egg chamber. Peercy challenges this by flipping the conventional football shape on its side. “If you look at it from a different angle, what is the activation pattern of the surrounding cells going to be?” he asked. He initially hypothesizes that STAT signaling around the polar cells is radially symmetric. Upon extracting and examining the egg chambers, Peercy and his colleagues see that while some of the cells do exhibit a radially-symmetric pattern, even more show complete asymmetry and reveal existing gaps. “How can this activation pattern be different in so many ways?” he queried.
Peercy’s new, second hypothesis claims that the underlying geometry is responsible for asymmetric border cell activation. The germline cells form a cleft that could alter the distribution of Upd, so he creates a reaction-diffusion model whose source is at the polar cells. This leads him to ask whether one can use a model to capture the cluster migration and even rotation observed in experiments via only adhesive, repulsive, chemotactic, and small stochastic forces. “If there is adhesion and you’re sensitive to the chemoattractant, you can use that to push off and move up the gradient like a rock climber climbing up a cliff,” Peercy said. “The cluster cells have the ability to see the chemoattractant and the ability to move.”
The rotation of the cluster is also of interest. While Peercy is able to capture some of the rotation, he cannot capture it all. “There’s probably some additional chemical signaling going on that we need to capture to fully represent this mechanism,” he said. “In the end, what we find is that you get more of an exponential approach because the chemoattractant is going to have that exponential shape.”
Ultimately, Peercy proves that the extracellular domain has significant effects on the cell fate and time course of migration. However, he acknowledges that activating the cluster yields an accurate pattern but an incorrect number. The extracellular space in the simulation is also much too big, as Peercy and his team were limited by the constraints of MatLab. The team requires something akin to the 500 radius to inter-germline cell ratio, which is more realistic. Further work is required to tackle that computation.