Cellular mechanisms of morphogenesis
"Morphogenesis" is a general term meaning the creation of form and it is used most frequently in the context of the creation of shape during animal development.
The range of animal forms is immense, both in terms of obvious external appearances and in terms of the internal anatomies of organs, tissues and cells; the vast number of different shapes can therefore make the analysis of morphogenetic mechanisms seem daunting at first (compare the hundreds of thousands of animal shapes, for example, with the modest number of differentiated cell types found in a complex animal, estimated at about 250 for humans). Fortunately, detailed analyses of morphogenetic events in a variety of animals suggests that most morphogenetic events are in fact driven by only ten basic cellular mechanisms, used in different combinations and sequences. These basic cellular mechanisms of morphogenesis are set out in the 'contents' box below. It should be emphasized that different organisms can add specialized mechanisms to this basic list, for the creation of unusual structures such as mammalian platelets, but most morphogenetic events in most animals are driven solely by the mechanisms of the basic list. The rest of this article will present brief explanations of each of these ten mechanisms, and examples in which each is used. Each depends on the presence of appropriate molecules in the cell, and these are made according to the ‘developmental programme’ encoded in the genome of the organism: the molecular bases of some mechanisms are well-understood, but others much less so.
Cell multiplication by mitosis, controlled directly by the internal mechanisms of the cell cycle and indirectly, often, by paracrine growth factor signals from other cells, typically acting on cyclin expression, is one of the main mechanisms for embryonic growth (embryos can also grow by cell expansion, and by making fluid-filled cavities). The direction of the mitotic spindle, and therefore the relative orientation of the daughter cells of the mitosis, may be tightly controlled. In most epithelia, for example, the spindle is directed at right angles to the axis of apico-basal polarity so that the daughters end up side-by-side in the epithelium: in the blastocyst epithelium of the early mammalian embryo, however, the spindle of some divisions lies along the apicobasal axis, resulting in some daughter cells being pushed out below the plane of the epithelium and giving rise to the inner cell mass (Figure 1).
Even where the mitotic spindle is directed at right angles to the axis of apico-basal polarity, it may also be constrained to lie at a precise orientation with respect to the axis of planar cell polarity. This means that the sheet of cells will tend to expand in one specific direction, and the effect is thought to be important, for example, in closure of the neural tube of mammals (Figure 2). Planar cell polarity is often encoded by the Wnt-JNK pathway.
Cell expansion, especially when it takes place in a specific direction, is an important mechanism in the morphogenesis of plants and is also important in some animals. In animals, it is common for cells in a particular part of an epithelium to elongate to become tall compared to their neighbours, in a process called ‘pallisading’ that usually occurs just before other morphogenetic mechanisms such as keystone deformation. The mechanism of this elongation used to be thought to rely on microtubule polymerization because it is blocked by the microtubule-polymerizing drug, colchicine. Recent experiments using other microtubule-polymerizing drugs (eg nocadazole) have cast doubt on this, however, and suspicion now falls on a combination of fluid influx that causes expansion and circumferential actin that restricts the expansion to be longitudinal only. Expansion can also affect just one part of a cell, to produce a long ‘process’ such as the acrosomal process of the sperm of the sea cucumbers (Thyone: figure 3), or microvilli in vertebrate epithelia. This is driven by rapid polymerization of actin filaments that are cross-linked into a bundle that forms the core of the cell process.
Elective cell death
Elective cell death, a term that includes apoptosis and other mechanisms for cells to die by their own actions, is responsible for eliminating cells from a developing body. This may be needed because cells of one type are present in excessive numbers for other types of cell with this they interact. More relevant to morphogenesis, elective cell death may be needed because the cells are needed for the process of development but not for final anatomy, in rather the way that scaffolding is needed for the construction of a building but not for the building itself. Elective cell death is used at a fine scale for eliminating small numbers of excessive cells in a developing tissue, and also at larger scales for eliminating entire structures, such as the presumptive uterus in male mammals, the presumptive vas deferens in female animals, the tail of a tadpole and the webs between the toes of a foot. The importance of elective cell death as a morphogenetic mechanism can be demonstrated by preventing it, for example with drugs that block the caspase cascade that controls apoptosis: if this is done during the development of chicken hindlimb development, the chick develops webbed feet rather like those of a duck. Elective cell death is typically controlled by paracrine signals from other cell types, that result, via control of cytochrome release from mitochondria, in activation of the caspase cascade.
Cell migration is critical to the development of most animals, being responsible for the dispersal of cells from one place to others (eg cells originating in the neural crest that disperse to form melanocytes, some neural tissue and much of the face, or primordial germ cells that migrate to the developing gonads) and also being responsible for the extension of neural axons and dendrites to their targets. Cell migration is driven mainly by organization of a motile leading edge (lamellipodium) in one part of the cell, this leading edge being based on polymerization of actin, controlled by ARP2/3 complexes that are in turn regulated, via proteins such as WASP, by small GTPases such as Rac and cdc42. The activity of these proteins can be regulated by signal transduction cascades controlled ultimately by extracellular molecules. This allows cells to be guided by gradients of attractive or repulsive extracellular molecules, in a process called chemotaxis. The migration of cells can also be controlled by differences in substrate adhesion, generally choosing more adhesive surfaces when presented with a choice, for biophysical reasons. Adhesion to substrates depends on the extracellular matrix molecules within the substrate and the repertoire of matrix receptors (for example, integrins) expressed by the cells themselves. Different cells express different combinations of integrins, so can respond in a cell type-specific manner to the same extracellular matrix cues of a substrate. Guidance by substrate is called haptotaxis.
Cell aggregation or condensation
Aggregation and condensation of previously loose-packed mesenchymal cells is a common feature of vertebrate development and is the prelude to the formation of bones, hair follicles, teeth and many other structures. Whether cells condense or remain separated depends on two things; (a) whether cell-cell or cell-matrix interactions are most energetically favourable and (b) whether the cells have the opportunity to meet. Cells can make cell-cell interactions more energetically favourably by expressing specific cell-cell adhesion molecules: an example is provided by limb bud mesenchyme cells, which express the homophilic cell-cell adhesion molecule, N-cadherin, as a means of condensing together prior to making cartilage. This condensation is prevented if an experimenter uses antibodies to block N-cadherin-mediated adhesion. Cells can also be brought together, creating the opportunity for condensation, by elimination of the extracellular matrix that separates them. An example is provided in the formation of hair follicles, in which CD44 is expressed to digest the hyaluronic acid that separates cells. Experimental injection of bacterial hyaluronidases is sufficient to promote condensation in this system.
Cell fusion is used to create syncytial cells, such as skeletal muscles that form by myoblast fusion, and it is also used to make very fine tubes (when a process of a cell extends out and wraps back on itself to make a cylindrical vessel).
Mesenchymal to epithelial transition (MET) is used to create new epithelia from condensations of mesenchymal cells. It is a relatively rare event in vertebrate development, but is critical to the formation of many epithelia of the urogenital system. EMT is achieved by the expression of ‘epithelial’ adhesion molecules such as E-cadherin, desmogleins, occludins etc, basement membrane and its integrin receptors, and other systems concerned with apico-basal polarity: it is accompanied by loss of mesenchymal integrin receptors interstital matrix and of cell motility.
Keystone deformation is an important mechanism for the morphogenesis of epithelial (and endothelial) sheets and tubes. The term refers to the constriction of one end of a cell, for example the apical pole, so that the cell takes on a cross-section rather like the keystone of an arch, with lateral surfaces that slope rather than stand at right angles to the basal surface. Because the lateral surfaces of epithelial cells are connected to each other, if a group of cells undergoes keystone deformation, the epithelium itself will be forced to bend (Figure 4). This mechanism is though to be responsible for bending of epithelia (eg in the folds of the colon), the initiation of epithelial invagination or evagination, which may then make use of cell multiplication or convergent extension to elongate the resulting tubule, and for the initiation of new branches during branching morphogenesis. Keystone deformation can also be used to drive the closure of holes in the plane of an epithelium.
Keystone deformation has traditionally thought to be driven by myosin-mediated tension in the actin network that runs between adhearens junctions (this tension can be driven by the small GTPase Rho, which acts via ROCK and myosin light chain kinase). Recent observations that blocking actin-myosin activity with drugs does not block keystone deformation in a number of systems, including neural tube formation, has cast doubt on this idea and genetic and cell biological data now point to the importance of Shroom proteins, which may cause cell-cell junctions around one end of of a cell to become less separated and this to reduce the circumference of that end, resulting in a keystone shape. At present (2008), much uncertainty remains.
Convergent extension is a mechanism that allows a structure to become long and thin without any net increase in cell volume or number, and it seems to be particularly important in late development of some invertebrates. In Drosophila germ band extension, the only system to be studied in detail, cells achieve convergent extension movements by treating their boundaries differently: boundaries that lie within the plane of the epithelium but perpendicular to the axis of elongation shorten, possibly under the action of actin-myosin contraction under the the control of a myosin regulatory chain (encoded by the gene Spaghetti squash) that is located only at these boundaries. The consequence of this can be best understood with reference to figure 5, in which the shortening of the boundary between cells C and H brings together the cells G and I and separates C and H along the axis of tissue elongation: repeated across the tissue, this process makes a short, broad epithelium into a long, thin one. The cells ‘know’ which boundaries are which by reference to the planar cell polarity system in the tissue.
Epithelial-to-mesenchymal transition (EMT) is the means by which the first mesenchyme is formed in vertebrate development (at gastrulation) and it is also used at other stages of development, for example the creation of neural crest from neurepithelium. EMT is achieved by cessation of expression of ‘epithelial’ cell-cell adhesion molecules such as E-cadherin and desmocollins, and increased expression of ‘mesenchymal’ cell-matrix adhesion molecules (integrins etc), and activation of motility. The change of adhesion and the active motility causes cells to leave their old epithelium and to migrate away as mesenchyme. This large-scale switch of gene expression can be controlled by transcription factors such as Slug and Snail.
- Davies J. (2005) Mechanisms of Morphogenesis. Academic Press/ Elsevier. This book considers cellular mechanisms of morphogenesis in detail, and contains numerous references to the primary literature. Each main topic is considered first at an overview level, suitable for early-year undergraduates, and then in more advanced chapters suitable for finals and postgraduate students.
- Gilbert S. (2006) Developmental Biology, 8th edn. Sinauer, Sunderland, Massachusetts. This book is a superb introduction to developmental biology and provides much valuable context for morphogenesis.
- Gilbert S. & Raunio A. (1997) Embryology: constructing the organism ppIX-X. In: pp. IX-X. Sinauer.This book provides much interesting information, at least at the descriptive level, about morphogenesis of organisms that are not the usual models for developmental biology.
- Alberts B., Bray D., Lewis J., Raff M., Roberts K. & Watson J.D. (1994) Molecular Biology of the Cell, 3rd edn. Garland, London.This provides an excellent text for cell biology in general. It is also the source of the count of differentiated cell types discussed at the top of this article.
- Gierer A. (1981a) Generation of biological patterns and form: Some physical, mathematical, and logical aspects. Prog. Biophys. Molec. Biol. 37, 1-47. Gierer has made theoretical contributions using shell theory (as used by architects). The treatment of the bending of cell sheets - the crucial point - starts on page 34 of this article (PDF at http://www.eb.tuebingen.mpg.de/departments/former-departments/h-meinhardt/Old%20Paper%20PDF/Generation%20of%20biological%20patterns.pdf)