Integument pattern formation involves genetic and epigenetic controls: feather arrays simulated by digital hormone models - PubMed (original) (raw)
Review
Integument pattern formation involves genetic and epigenetic controls: feather arrays simulated by digital hormone models
Ting-Xin Jiang et al. Int J Dev Biol. 2004.
Abstract
Pattern formation is a fundamental morphogenetic process. Models based on genetic and epigenetic control have been proposed but remain controversial. Here we use feather morphogenesis for further evaluation. Adhesion molecules and/or signaling molecules were first expressed homogenously in feather tracts (restrictive mode, appear earlier) or directly in bud or inter-bud regions ( de novo mode, appear later). They either activate or inhibit bud formation, but paradoxically colocalize in the bud. Using feather bud reconstitution, we showed that completely dissociated cells can reform periodic patterns without reference to previous positional codes. The patterning process has the characteristics of being self-organizing, dynamic and plastic. The final pattern is an equilibrium state reached by competition, and the number and size of buds can be altered based on cell number and activator/inhibitor ratio, respectively. We developed a Digital Hormone Model which consists of (1) competent cells without identity that move randomly in a space, (2) extracellular signaling hormones which diffuse by a reaction-diffusion mechanism and activate or inhibit cell adhesion, and (3) cells which respond with topological stochastic actions manifested as changes in cell adhesion. Based on probability, the results are cell clusters arranged in dots or stripes. Thus genetic control provides combinational molecular information which defines the properties of the cells but not the final pattern. Epigenetic control governs interactions among cells and their environment based on physical-chemical rules (such as those described in the Digital Hormone Model). Complex integument patterning is the sum of these two components of control and that is why integument patterns are usually similar but non-identical. These principles may be shared by other pattern formation processes such as barb ridge formation, fingerprints, pigmentation patterning, etc. The Digital Hormone Model can also be applied to swarming robot navigation, reaching intelligent automata and representing a self-re-configurable type of control rather than a follow-the-instruction type of control.
Figures
Fig. 1. Examples of integument patterns
(A) Feather stripes, patches and spots on feathers of a pheasant. There is regional variation in the color of the skin appendage markings. The stripes are brown on the tail feathers and white on the wing feathers. (B) Pigment stripes and spots on a cat. Stripes are present from the neck to the tail with a similar orientation. The stripes on the head have a different rostral-caudal orientation. Some stripes are complete and some are partial. In some regions, stripes break into dots. (C) Fingerprints from the same finger of identical twins. Although genetically identical, epigenetic events during fingerprint formation lead to subtle variations between twins. From Jain et al. (2002). (D) Skin ridges on dolphin integument. Muscles underlying the skin can respond to environmental pressures causing microvibrations which reduce water turbulence, enabling the skin to respond to its environment. The remarkable orientations of the lines imply that the environment has an effect. From Ridge and Carder (1993). (E) Asymmetric pigmentation patterns in pheasant tail feathers. Stripes are present on one side of the vane and hash marks (flecks) are present on the other side. Two tail feathers are placed side by side and show pigmentation patterns that are similar but non-identical. (F) Schematic diagram showing the human lines of Blaschko. These represent patterns of human skin diseases which may result from mutations which occur within a cell lineage during skin morphogenesis. The nature of these lines is cell lineage based. From Happle, (1985).
Fig. 2. Different integument patterns can be revealed during feather morphogenesis
(A) Embryonic chicken skin showing feather buds arranged in hexagonal patterns and groups of buds distributed in feather tracts. Buds were visualized by in situ staining for Shh. (B) Feather arrays on the quail. Note the different array pattern between chicken and quail, even though the two species are closely related. The black color is due to melanocytes. (C) β -catenin marks the initial appearance of feather primordia. During development, the β-catenin transcripts are first expressed in the whole tract field, before they become restricted to individual feather primordia. Each primordium is then surrounded by a region devoid of β -catenin expression (Widelitz et al., 2000). (D) Dermatome cell lineages. Quail somites were transplanted to the dorsal midline of chicken embryos. They can be seen to migrate out to populate certain regions of the dorsal lateral trunk. *This patch represents dermal cell lineages migrating out from specific somite regions. (E) Epithelial cell lineage. Epithelial precursor cells near the midline were transduced with LacZ as a lineage marker. The lineage of epithelial cells are distributed in horizontal lines. Note the similarity to the lines of Blaschko in Fig. 1F. From Chuong et al. (1998). (F) Schematic showing feather tracts. Tracts are shown as blue patches. The direction of progressive feather bud formation is shown by the arrow. (G) Modes of molecular expression. Genes are expressed with restrictive (left) or de novo (right) expression patterns. Both modes of expression can form bud or interbud expression patterns. (H) Examples of genes expressed with restrictive and de novo patterns. Expression patterns of L-CAM (Chuong and Edelman, 1985a), Eph A4 (Patel et al., 1999), Shh (Ting-Berreth et al., 1996a), L-fringe (Chen and Chuong, 2000), gremlin (Ohyama et al., 2001) and collagen I are shown.
Fig. 3. A model for feather periodic patterning involving reaction-diffusion and competitive equilibrium
(A) Through experimentation, some molecules are found to enhance feather formation (named activators) and some suppress feather formation (named inhibitors). Exemplary activators (FGF 4) and inhibitor (BMP 4) are shown. PMA increases the size of the interbud domain (inhibitor of bud formation) whereas Forskolin decreases the size of the interbud domain (activator of bud formation). (B) Data show that both activators and inhibitors of bud formation are located in the bud region, not in the bud and interbud, respectively. The activators induce both activators and inhibitors, while the inhibitors suppress the activators. Some growth factors promote or suppress interbud formation. Examples include Shh (Ting-Berreth et al., 1996a), FGF (Widelitz et al., 1996; Song et al., 1996; Jung et al., 1998), PKA (Noveen et al., 1995b), follistatin (Patel et al., 1999), BMP (Jung et al., 1998), Delta-1 (Crowe et al., 1998), retinoic acid (Chuong et al., 1992), EGF (Atit et al., 2003), PKC (Noveen et al., 1995b), Wnt 7a (Widelitz et al., 1997). These results favor the involvement of a reaction diffusion mechanism (Turing, 1952; Nagorcka and Mooney, 1985; Moore et al., 1998; Jiang et al., 1999).
Fig. 4. A novel reconstitution assay which allows the study of periodic pattern formation from the ground state
(A) Embryonic day 6 chicken skin was dissected. (B) Epidermis (shown) was separated from the mesenchyme. (C) Mesenchyme was dissociated into single cells. (D) Mesenchymal cells were recombined with epithelium and plated. In 3 days, they self-organize into many feather buds simultaneously (Jiang et al., 1999). (E) Using the reconstitution model, we tested the relationship using a fixed size of epidermis while increasing the number of competent mesenchymal cells. Logically, either the number of buds or the size of the feather primordia could increase. We found that, using skin from a certain region, the size of the feather primordia is constant. At low density, buds did not form. At higher cell density, feather primordia started to appear randomly. The density of feather primordia gradually increased until they reached the highest packing density, which yields the hexagonal patterning (Jiang et al., 1999). Our results show that it increased the number of feather buds. We propose that competent cells are first distributed homogeneously in the field. These cells adhere randomly and this adhesion is reversible. When these small unstable aggregates surpass a threshold density, they become stable dermal condensations. (F) The size of each dermal condensation is dependent on the ratio of activator molecules (noggin, FGF, Shh, etc) to inhibitor molecules (BMPs). A higher activator to inhibitor ratio allows the formation of larger sized feather buds, while a higher inhibitor to activator ratio favors the formation of the interbud region (Jiang et al., 1999). (G) Schematic diagram showing that increasing mesenchymal cell number increases feather number at a constant size. The size of feather buds is influenced by activators (Noggin) and inhibitors (BMP).
Fig. 5. Cell adhesion molecules act as mediators of patterning
(A) Immunostaining for NCAM in reconstituted explants or in dissociated mesenchymal cells (N-E) after 4, 10, 18 and 24 h. in culture. Scale bar, 200 μm. (B) Pseudocolor and high power views. (C) NCAM was expressed at moderate levels throughout the cultures and was upregulated in cell aggregates by 18 h. (arrow 1), but was expressed at basal levels in lower density aggregates (arrow 2) and not expressed where cells remained aggregate-free (arrow 3). (D) Schematic drawing shows that cells are initially attached to their substrates, but through extracellular matrix molecules, cell-cell adhesion becomes greater than cell-substrate adhesion leading to cell condensation formation. (E) Hence, during feather formation, global events help to form the tract field. Random unstable cell aggregates begin to form in the mesenchyme. Local events now begin to dominate the feather forming processes. Stable aggregates form and signal to the overlying epithelium to form a placode. This induces alterations of molecular expression. Through a reaction - diffusion mechanism, these molecular expression patterns are intensified into periodic patterns which shape and consolidate the forming feather primordia.
Fig. 6. Diagram of self-organizing models for pattern formation
(A) Homogenously distributed cells through random interactions form unstable aggregates. This forms random variations which are amplified above a threshold at which the patterns become set. Distinct patterns are formed by competition between intrinsic factors (properties of the membranes and extracellular matrix), concentrations of activators and inhibitors and the size of the primordial field. (B) Digital hormone model. Cells (black dots) can move within the grid. They secrete activators (red) and inhibitors (green) that influence neighboring cells which fall within their sphere of influence. Activators and inhibitors cancel each other out in the space between the red and green regions. In the lower portion of the figure, two cells are interacting through their hormones. If the activator/inhibitor ratio is high (lower left figure) feather formation is highly favored. If the activator/inhibitor ratio is low (lower middle figure) feather formation is suppressed. If the activator/inhibitor ratio is balanced (lower right figure) feather buds and interbuds will form.
Fig. 7. Simulation of feather patterning by the Digital Hormone Model
(A) Simulations of pattern formation with increasing mesenchymal cell densities (10%, 25%, 50% of grids filled) in a fixed field size indicate that higher cell density favors more aggregates of similar size. The process was captured at step 0, 50, 500 and 1000 to see the dynamic flux of the cell movements. (B) The presence of hormone activators and inhibitors upon pattern formation. Starting with a uniformly dispersed cell population, inhibitors blocked the formation of aggregates. Activators led to the formation of aggregates. As the ratio of activator to inhibitor increased, the size of the cell aggregates increased. The dynamic cell sorting was captured at step 0, 50, 500 and 1000. (C) To test the influence of field shape on pattern formation we started with a field shaped like a plus sign. This caused the virtual cells to form stripes.
Fig. 8. Further patterning of feather buds and feather filaments
Several factors influence the final shape of each feather. (A) As feathers elongate, a proximal-distal axis starts to develop and molecular differences are seen along this axis. Localized growth zones (LoGZ) are located at the tip of the feather during early development, but become localized near the feather base later. (A′) Tangential section of early buds show that the homogenous feather buds become heterogeneous following the development of the anterior -posterior axis. Examples shown include tenascin (Jiang and Chuong, 1992); Wnt 7a (Widelitz et al., 1997); Eph A4 (Patel et al., 1999), Notch (Chen et al., 1997) and Follistatin (Patel et al., 1999). (B) Cross sections of different levels of feather filament are shown schematically which also represent different developmental stages (distal is more mature). Stage 0: initiation of the stratified epithelial cylinder; the basal layer is beginning to form. Stage 1: barb ridge formation initates, the basal layer is well formed. Stage 2: barbule plates and marginal plates begin to form. Stage 3: axial plates are forming and barbule plates are well formed. Stage 4: keratinization is complete; barbs have separated. The numbers and sizes of each structure depicted are schematic and do not reflect the actual values (from Chuong and Edelman, 1985b). (C) Wholemount view of the elongated feather buds. The barb ridges alternate with the marginal plate which is highlighted by staining with Shh. Cell lineage was traced by injecting replication-defective spleen necrosis virus directing the expression of β-galactosidase at E10 and analyzing expression at E18. The results indicate that it is not one clone - one marginal or barb plate. Instead each individual barb is polyclonal, containing some β-galactosidase positive and some β-galactosidase negative cells. ap, axial plate; br, barb ridge; fe, feather epithelium; mp, marginal plate; pe, pulp epithelium.
Fig. 9. Schematic diagram depicting the genotype, the epigenetic events at different organization levels and the resultant phenotypes
While genetics (genome, transcriptome, proteome) provides the basic molecular composition, probability events and variations occur at the cell interaction level. Cell fate decisions are based on basic cellular interactions (proliferation, adhesion, migration, death and differentiation). They are ruled by physiochemical phenomena (stochastic randomness, competitive equilibrium, self-organization). These factors have been modeled by reaction diffusion, cellular automata and digital hormones.
References
- Artavanis-Tsakonas S, Matsuno K, Fortini ME. Notch signling. Science. 1995;268:225–268. - PubMed
- Asai R, Taguchi E, Kume Y, Saito M, Kondo S. Zebrafish leopard gene as a component of the putative reaction-diffusion system. Mech Dev. 1999;89:87–92. - PubMed
- Atit R, Conlon RA, Niswander L. EGF signaling patterns the feather array by promoting the interbud fate. Dev Cell. 2003;4:231–40. - PubMed
- Bereiter-Hahn J, Matoltsy AG, Richards KS, editors. Biology of the Integument 2 Vertebrates. Berlin: Springer Verlag; 1986.
- Botchkarev VA, Botchkareva NV, Sharov AA, Funa K, Huber O, Gilchrest BA. Modulation of BMP signaling by noggin is required for induction of the secondary (nontylotrich) hair follicles. J Invest Dermatol. 2002;118:3–10. - PubMed