Self-organization in embryonic development: myth and reality

"Self-organization" has become a watchword in developmental biology, characterizing observations in which embryonic or induced stem cells of animals replicate morphological steps and outcomes seen in intact embryos. While the term was introduced in the 18th century by the philosopher Immanuel Kant to describe the goal-directed properties of living systems, it came into modern use for non-living materials in which complex forms and patterns emerge through dynamical, energy-expending physical processes. What is the relationship among these uses of the term? While multicellular forms arose dozens of times from single-celled organisms, only some of these undergo development, and not all developmental processes are self-organizing. The evolution of the animals (metazoans) from unicellular holozoans was accompanied by the addition of novel gene products which mediated the constitution the resulting cell clusters as liquid-, liquid crystal-, and solid-like materials with versatile morphogenetic propensities. Such materials variously exhibited multilayering, lumen formation and elongation, echoing the self-organizing properties of nonliving matter, "generic" based on such parallels, though with biologically based subunit properties and modes of interaction. These effects provided evolutionary templates for embryonic forms and morphological motifs of metazoan lineages. Embryos and organ primordia of present-day animal species continue to generate forms that resemble the outcomes of these physical effects. Their development, however, employs overdetermined, highly evolved mechanisms that are often disconnected from their originating processes. Using the examples of gastrulation, somitogenesis, and limb skeletal development, I provide instances of, and a conceptual framework for understanding, the relationships between transparently physical and evolved types of developmental self-organization.