A computational study of self-organization along the anterior-posterior axis in vertebrate embryos: somitogenesis and neuroepithelial rosette formation

Abstract

An incredibly elegant and extremely complex, yet seemingly effortless, coordination of behaviors takes place when clusters of cells work together to create an entire embryo from scratch. Starting from just a single cell, the cells divide, change cell state, move, adhere together, coordinate directions and react to their environment in an astonishingly tightly orchestrated way that ends with the creation of a brand new individual. Advances in understanding the components and processes that take place in embryonic development are made every day, however we still have a lot left to learn. One of the big questions in life that is still not fully understood is: How are babies made? With this thesis I wish to answer the same question seen from a different perspective: How do babies make themselves?

One of the first key processes that happen in the early stages of a vertebrate's life is the establishment of the head-to-tail axis, called the anterior-posterior axis. This thesis sets out to do a theoretical, mathematical and computational exploration of various aspects of the vital anterior-posterior development of mouse embryos. With a special focus on the self-organizing capabilities of embryonic cells, we study which mechanisms cells use to control their own behaviors within each individual cell, and how they coordinate their behaviors across the tissue to create marvelous patterns and shapes.

Recent advances in biological technologies now allow us to take a closer look at the cells while they are performing developmental processes. By removing the cells from their embryonic environment and culturing them in vitro, it is possible to study the cells' movement, signaling molecules, coordination and interactions with their environments. Amazingly, these experiments reveal the self-organizing properties of the cells and how they act without the influence of other embryonic tissues, geometric constraints, global gradients, signaling cues and more.

The first process I have studied takes place in the posterior (tail) side of the embryo and is called somitogenesis - the formation of body segment precursors (somites) which make the vertebrae, skeletal muscles and other structures found along the head to tail axis of the embryo. These segments are generated by coordinated genetic oscillations giving rise to waves along the head to tail axis of the embryo. With a reaction-diffusion model I found that a guided self-organizing process where cells are excited to oscillate by local interactions with their neighbors both propagate waves of signaling activity across the tail as seen in vivo, and also captures several key findings from in vitro experiments of tail explants and single cells. By developing a first ever simulation of a virtual explant of an embryonic mouse tail, I show that initial conditions and the excitable behavior, without the need for globally controlling gradients, are sufficient to generate the circular waves of genetic oscillations as observed with explants in vitro.

While somitogenesis is happening, on the anterior side of the embryo, neuroepithelial tissues begin to take form in the region of the embryo that will develop the head. Also in this case, in vitro experiments allow us to see, in part, how this may happen in a self-organizing manner. By creating aggregates of mouse embryonic stem cells called embryoid bodies, my colleagues in the laboratory at CABD have observed that under the right experimental conditions, epiblast stem cells differentiate into neuroepithelial cells that spontaneously self-organize into hollow spherical layers known as rosettes. These rosettes further re-organize into elongated tubular structures as the whole embryoid body elongates. These processes mimic some aspects of the morphogenesis seen during neurulation in the embryo. Using an agent-based simulation with cell movement, cell shape changes and cellular polarity, I discovered that cellular polarity and apical constriction is sufficient to model the local organization that forms global neural rosettes and tubes in virtual embryoid bodies.

Overall, the goal of this thesis is to improve our understanding of these fundamental processes in order to better understand how life is made.