Organizational principles of integumentary Organs: Maximizing Variations for Effective Adaptation

Abstract

The integument serves as the interface between an organism and its environment. It primarily comprises ectoderm-derived epithelium and mesenchyme derived from various embryonic sources. These integumentary organs serve as a barrier defining the physiological boundary between the internal and exterior environments and fulfill diverse functions. How does the integument generate such a large diversity? Here, we attempt to decipher the organizational principles. We focus on amniotes and use appendage follicles as the primary examples. The integument begins as a simple planar sheet of coupled epithelial and mesenchymal cells, then becomes more complex through the following patterning processes. 1) De novo Turing periodic patterning process: This process converts the integument into multiple skin appendage units. 2) Adaptive patterning process: Dermal muscle, blood vessels, adipose tissue, and other components are assembled and organized around appendage follicles when present. 3) Cyclic renewal: Skin appendage follicles contain stem cells and their niches, enabling physiological molting and regeneration in the adult animal. 4) Spatial variations: Multiple appendage units allow modulation of shape, size, keratin types, and color patterns of feathers and hairs across the animal’s surface. 5) Temporal phenotypic plasticity: Cyclic renewal permits temporal transition of appendage phenotypes, i.e. regulatory patterning or integumentary metamorphosis, throughout an animal’s lifetime. The diversities in (4) and (5) can be generated epigenetically within the same animal. Over the evolutionary timescale, different species can modulate the number, size, and distributions of existing ectodermal organs in the context of micro-evolution, allowing effective adaptation to new climates as seen in the variation of hair length among mammals. Novel ectodermal organs can also emerge in the context of macro-evolution, enabling animals to explore new ecological niches, as seen in the emergence of feathers on dinosaurs. These principles demonstrate how multi-scale organ adaption in the amniotes can maximize diverse and flexible integumentary organ phenotypes, producing a vast repertoire for natural selection and thereby providing effective adaptation and evolutionary advantages.

Graphical Abstract

A conceptual diagram illustrating tissue patterning during the development and regeneration of a prototypical amniote skin. Complex tissue patterns are established through De novo patterning and multiple steps in adaptive patterning. Regulatory patterning refers to the modifications of appendage phenotypes in adults to adapt to environmental changes.

Regenerative patterning describes the reformation of tissue patterns after loss, as observed in wound regeneration or organoid morphogenesis of dissociated progenitor cells.

1. Introduction

“It is not the strongest nor the most intelligent of species that survives, but the one that is most adaptable to change.” – Charles Darwin

Natural selection and sexual selection remain the primary mechanisms driving evolution. For organisms to thrive in this process, they must demonstrate high adaptability. This adaptability is built upon the integration of multiple organ systems. However, among different organs, changes in ectodermal organs are the most visible. While the fundamental principles for generating variation to optimize adaptation likely apply to all organs, ectodermal derived organs form the external integuments and oral cavity, and are directly exposed to the changing external environment, where the need for adaptation is most critical. Moreover, the changes in these organs are visible and the system is experimentally manipulable. Therefore, we focus here on integumentary organs to uncover these underlying principles. Our goal is to understand how these organs, such as feathers, hair, teeth, glands, and many other integumentary organs, can be so diverse and adaptable, as these are defining features of the classes Aves and Mammalia. We found the complex integumentary organ phenotypes are achieved by combinatorial, multi-dimensional pattern formation, providing a vast repertoire of functional forms for selection.

1.1. Integumentary And Ectodermal Organ

The integument forms the interface between an organism and its environment. It serves as a physiological barrier between the internal and exterior environments. These integumentary organs are derived from the ectoderm (Fig. 1A). While oral cavity ectodermal organs such as teeth, taste buds, and salivary glands also originate from the ectoderm, they are not part of the external integument but are included here categorically as part of the integumentary organs to facilitate the discussion (Please see Box 1, Glossary). The integument provides the first line of defense against mechanical, biochemical, or microbial insults, thus protecting the animal (Chuong et al., 2002). They serve as sensors to detect and adapt to large-or small-scale changes in the external environment, to maintain temperature stability (e.g., hair, sweat glands, blubber) through metabolic activity. They form glands with specialized biochemical functions, secreting fluids that serve various purposes (e.g., milk, mucus, toxins). Integumentary organs can function as weapons for attack or defense (e.g., scales, horns, claws, teeth). Some evolved into a feeding apparati (e.g., tooth, beak), while others enable locomotion (e.g., bird feathers, wings of flying mammals, snake scales), allowing amniote animals to venture into diverse ecological niches. Additionally, the integument provides a platform for individuals to either conceal themselves or display flamboyant communication signals to intimidate predators or attract mates.

Figure 1. Organizational principles of ectodermal organs.

A schematic drawing of a conceptual amniote animal displaying various epithelial appendages, all resulting from epithelial-mesenchymal interactions. The epithelia can originate from either ectoderm (solid line) or endoderm (gray line) (Chuong, 1998; Wu et al., 2004b). B. The Tao of integuments: periodic patterning creates multiplicity, which allows for regional diversity, while stem cell-based follicles enable cyclic renewal and temporal transitions in appendage phenotypes. Together, these mechanisms equip the integument with the ability to generate variations and facilitate adaptation (Lai and Chuong, 2016).

Box 1: Glossary.

Amniotes:

Since reptilian, avian and mammalian embryos have amniotic or chorionic membranes these tetrapod vertebrates are called amniotes.

Ectodermal and integumentary organs:

In this review, we focus on the amniotes. In a triploblastic embryo, there are three germ layers: endoderm, mesoderm and ectoderm. Ectoderm gives rise to neural and surface ectoderms. In this special issue, the ectodermal organ indicates the use of the surface ectoderm and adjacent mesenchymal tissues to form an organ. Thus, categorically, ectodermal organs include teeth, mammary glands, hairs, feathers, etc. Integument refers to the protective layer that covers the outside of the living organism, encompassing the epidermis, dermis and associated structures such as blood vessels, nerves, etc. Integumentary organs are the diverse organs that develop on the integument across species that help animals interact with their environments. Surface ectodermal organs overlap with integumentary organs but they are not equivalent. Oral cavity ectodermal organs originate from the ectoderm. Strictly speaking, they line the oral cavity but are not part of the external integument. To facilitate our discussion of this topic, the oral cavity is included in this paper as part of the integumentary organs.

Epithelial and mesenchymal tissues:

Epithelial and mesenchymal tissues often work together to make an organ. They interact and influence each other to generate new cell fates and tissue organization. The epithelium is a continuous layer of epithelial cells supported by a basement membrane, separating it from the underlying connective tissue. Mesenchymal tissue refers to loosely organized fibroblast-like cells that form connective tissues. Both epithelia and mesenchyme are umbrella terms. The epithelium includes both ectodermal and endodermal derived epithelium. The mesenchyme can arise from paraxial mesoderm (somite), lateral plate mesoderm, neural crest-derived ectomesenchyme and epithelial-mesenchymal transition.

De novo, adaptive, regulatory and regenerative patterning:

Patterning is the process by which a group of similar cells within a tissue are organized into spatially or temporally distinct subsets, thus enhancing the functional efficiency and specialization of the tissue or organ.

De novo patterning refers to the emergence of patterns within a tissue without any pre-existing templates or guidance. A key example is Turing patterning, where periodic patterns arise through interactions between stochastic activators and inhibitors, following Turing principles.

Adaptive patterning involves the use of pre-existing signaling centers to guide further assembly or differentiation, enabling the formation of more complex tissue patterns.

Regulatory patterning describes the dynamic adjustment of tissue patterns in response to environmental changes or temporal shifts, enhancing adaptation. This process depends on the presence of stem cells and a modulable niche.

Regenerative patterning describes the restoration of tissue patterns after loss, as observed in wound induced hair neogenesis or self-organization of tissue patterns by dissociated progenitor cells

Thus, the animal integument exhibits a vast spectrum of variation and architectural principles, fulfilling numerous functions. Some invertebrates, notably cephalopods, display remarkably dynamic skin behaviors, including rapid color and texture changes for camouflage and communication (Shook et al., 2024). In vertebrates, fish integuments also demonstrate remarkable adaptations (Aman and Parichy, 2024). Given the extensive variations and differing environmental adaptations, a comprehensive review of all integumentary systems is beyond our scope. Therefore, in this review, we will focus on the integumentary organs present in amniotes. We depict some of the diverse ectodermal organs in a prototypical amniote animal (Chuong, 1998; Wu et al., 2004a). Different types of integumentary organs are distributed across the body surface with species-, spatial-, and temporal-specificity. During embryonic development, these ectodermal organs share the characteristic of being composed of epithelia and mesenchyme, arising from epithelial-mesenchymal interactions (Please see Box 1, Glossary). In adulthood, their homeostasis is maintained through the control of epidermal stem cell quiescence and activation. Some integumentary organs undergo cyclic renewal, while others utilize a continuous growth mode (e.g., claws, mouse incisors). With advancements in stem cell biology, we can now generate epithelial stem cells representing different stages and developmental lineages through reprogramming and culture conditions (Koehler and Hashino, 2014; Neumayer et al., 2023). Conditions supporting fibroblast lineages are more complex and how they modulate different ectodermal organ fates remain to be investigated (Plikus et al., 2021).

1.2. Organizational Principles That Enable The Generation Of Integumentary Organ Diversity

Can we extract some simple principles that govern these complexities? Given that birds exhibit the most distinct integumentary patterns, we use them as a primary model, comparing them with mammals and reptiles. In birds, this complexity is achieved by first partitioning the integument into multiple tracts (Chen et al., 2021; Dhouailly et al., 2004; Lucas and Stettenheim, 1972). Each tract becomes a morphogenetically competent field, undergoing periodic patterning to form a multiplicity of integumentary appendage primordia (Chuong et al., 2013). Each element becomes an integumentary appendage. This multiplicity allows each unit to function independently or in groups, creating new functionalities (Li et al., 2015). By housing stem cells within a protected niche (a follicle), each element can renew cyclically in response to molting or regenerate after injury (Wu et al., 2021). Appendage stem cells can respond to intrinsic signals (within the appendage follicle, such as growth factors) and/or extrinsic signals (outside the appendage follicles, such as temperature, hormonal changes, seasonal changes) to modulate the formation of appendage phenotypes at different life stages and in different body regions (Fig. 1B) (Lai and Chuong, 2016; Wu et al., 2021). For example, in birds, young chicks form downy feathers for warmth. In adulthood, they require feathers for flight and communication, leading to a transition to adult feather types, which can significantly alter their appearance from that of a chick. Evolution at the macro-level has led to dramatic changes in appendage types, further expanding integument functions and allowing amniote animals to adapt to different environments. The evolution of skin appendages has become a hallmark for new vertebrate classes: mammary glands for nurturing young define the mammalian class, while feathers enabling flight to define feathered dinosaurs and the entire avian class.

This review aims to decipher the organizing principles of ectodermal organs, identify signals controlling fate determination, tissue architecture, and homeostasis. This foundational knowledge will contribute to practical applications in regenerative medicine, such as guiding ectodermal progenitor stem cells to form specific ectodermal appendages of the desired type and size.

2. Tissue Patterning In The Developing Skin

The integument can range from being quite simple to forming highly complex architectures. Based on our developmental studies, this complexity is built step by step (Fig. 2A). A fundamental requirement is the coupling of epidermal and dermal cells to create the bi-composite structure of the integument. Deficiencies in this process can be observed in the Gibbin mutation syndrome (Collier et al., 2022). This planar integument can then bend, overlap, invaginate, or undergo other topological transformations to create different ectodermal organ architectures. The fundamental processes are de novo patterning, adaptive patterning, regulatory patterning and regenerative patterning. The first fundamental process involves “de novo patterning” of an ectodermal field (or region) into multiple units through periodic pattern formation. This process has been observed in dermal shark teeth (Cooper et al., 2023; Nicklin et al., 2024), fish scales (Harris et al., 2008), snake scales (Tzika et al., 2023), bird feathers (Jiang et al., 2004), mammal hairs (Sick et al., 2006), among others. The second process is “adaptive patterning”, through which new layers of blood vessels, nerves, muscle, adipose tissue, etc., are assembled using the pre-existing tissue structure as a template (Wu et al., 2019). These sequential steps enable the assembly of complex tissue patterns, which forms the basis of functional forms (Ramos et al., 2024). De novo and adaptive patterning are covered in this section. The third process is “regulative patterning” which refers to the temporal transition of appendage phenotypes that occur under physiological conditions and is discussed in section 5. Finally, the integument has the remarkable ability to regenerate and to restore patterns. We refer this as “regenerative patterning”. It includes wound induced hair neogenesis (Ito et al., 2007) and self-organization of skin organoids (Lei et al., 2017). We will not cover this topic here since it has recently been reviewed (Harn et al., 2023).

Figure 2. Tissue patterning in the developing skin.

A. A conceptual diagram illustrating tissue patterning during the development and regeneration of a prototypical amniote skin. Complex tissue patterns are established through De novo patterning and multiple steps in adaptive patterning. Regulatory patterning refers to the modifications of appendage phenotypes in adults to adapt to environmental changes. Regenerative patterning describes the reformation of tissue patterns after loss, as observed in wound regeneration or organoid morphogenesis of dissociated progenitor cells. B. Whole-mount in situ of β-Catenin in developing chicken embryo. C. Turing model and its molecular cellular mechanisms. D. In chickens, the reconstitution of skin explants shows feather bud formation simultaneously (in red circles), and feather density is dependent on dermal cell density (a-d). In vivo, under the influence of a global gradient, the formation process occurs in a propagatory manner (Inaba et al., 2019a). E, F. Roles of bio-electricity in feather bud morphogenesis. When short bud starts to elongate, it accompanied with an outward electric current and synchronized GCaMP6 activity (Li et al., 2018). External electric currents have effect on collective bud orientation which depends on the position of electrodes (Jiang et al., 2021). G. Assembly of dermal muscle network connect feather buds and is force dependent (Wu et al., 2019). H. The formation of the dermal vascular network depends on localized vasculogenesis within feather buds (Ou et al., 2024).

In amniote animals, integuments from different animals utilize similar processes to generate tissue patterns. In this review, we use the feather as the primary paradigm, but we also reference other integumentary organs when the principles are applicable.

2.1. De Novo Periodic Patterning

Periodic patterns refer to the repetition of basic elements with specific spacing across the integument surface (Ramos et al., 2024). These foundational patterns can be manifest as spots or stripes, upon which more complex designs are constructed. The repeating elements may include integumentary appendages or pigmentation patterns. Such periodic patterning is observed early in evolution among both invertebrates and vertebrates. Employing periodic patterning in biological architectures offers the following advantages.

First, generation of multiplicity.

Through periodic patterning, numerous integumentary organs form. For instance, instead of developing a single appendage, an organism can produce hundreds, creating a population of integumentary organs. This multiplicity allows for redundancy, where some organs can be dispensable, and others can undergo renewal cycles. Second, Modular Specialization. Each element functions as a module that can be modified for specialized purposes, providing a foundation for optimal adaptation. This modularity enables organisms to fine-tune specific body parts for particular functions, enhancing adaptability to diverse environments. In the following sections, we will explore the mechanisms underlying the formation of these periodic patterns, delving into the processes that contribute to the remarkable variations and adaptability of integumentary systems.

2.1.1. Turing Patterning Of Skin Appendages With Diffusible Morphogens

Diverse strategies employed across species to establish periodic patterns during development. For example, Drosophia larva segment formation is based on a hierarchical gene regulatory network involving gap genes and pair-rule genes (Patel, 1994). Vertebrate somite formation is based on a segmentation clock mechanism involving the Notch, Wnt and FGF pathway (Pourquié, 2022). Periodic formation of skin appendages is based on Turing reaction diffusion principles involving FGFs, Wnts, and BMPs (Jung et al., 1999; Sick et al., 2006). In amniotes, the pattern is most clearly demonstrated in the avian skin. The avian skin is first organized into multiple tracts or regions (Lucas and Stettenheim, 1972) which then form different skin appendage phenotypes. Each tract constitutes a morphogenetic field (Dhouailly et al., 2004). Within these fields, interactions between the dermis and epithelium generate appendage primordia (Chuong et al., 1996; Dhouailly, 2024, 1973). Turing activator / inhibitor principles has been shown to be a major mechanism for periodic pattern formation in the developing avian skin (Fig. 2C) (Jiang et al., 1999; Lin et al., 2009). In this patterning process, activators promote the formation of an element and stimulate the local production of both activators and inhibitors (Kondo and Miura, 2010; Turing, 1952).Inhibitors suppress the activators, thus limiting the expansion of the element and helping to create inter-element spacing. Initially, both activators and inhibitors are synthesized and secreted at the same location. Activators have a short diffusion range, while inhibitors have a long diffusion range. Consequently, elements are created where the activator-to-inhibitor ratio is highest, moving outward from the center. The long-range inhibitors prevent the formation of additional elements surrounding each bud, resulting in periodic arrangements of equidistantly spaced elements (Fig. 2C). In chicken dorsal skin, it has been shown that fibroblast growth factor 4 (FGF4) acts as an activator of feather bud formation, while bone morphogenetic protein 4 (BMP4) serves as an inhibitor(Jung et al., 1998). Further research revealed that the morphogenetic competent field expresses β-catenin all over the epithelium at E6, then gradually becomes restricted to the feather bud region at about E8 (Fig. 2B). While Shh is initially absent in the field, it later appears at the tip of each bud and has been used to visualize periodic bud formation (Jiang et al., 2004).

A variety of morphogen signals work together to guide proper skin appendage morphogenesis. Wnt-β-catenin signaling is key to the formation of an appendage (Widelitz, 2008). The roles of additional morphogens have been revealed. In the feather bud, Shh, enriched at the bud tip, is critical for elongation and growth (Ting‑Berreth and Chuong, 1996a). TGF-β is important for dermal condensation formation (Ting‑Berreth and Chuong, 1996b). The Notch-Delta pathway is crucial for establishing the anterior-posterior axis of the bud (Chen et al., 1997). Wnt7a, enriched in the posterior bud epithelium, can produce plateau-like buds when overexpressed throughout the bud (Widelitz et al., 1999).

Protein kinases also play an important role in establishing early feather bud development (Noveen et al., 1995). The cAMP response element binding protein (CREB) is initially expressed throughout the epithelium and mesenchyme. Beginning at H&H stage 29 phosphorylated CREB (P-CREB), corresponding to protein kinase A (PKA) activity is seen in the placode but not in inter-placode region. Later, P-CREB is expressed in the mesenchyme beneath the buds but not in the interbud regions between stages 33 and 36. Interestingly, protein kinase C (PKC) is first distributed throughout the early skin at H&H 30. Its expression levels then decrease beneath sites of developing feather germs. Functional tests show that PKA activators and PKC inhibitors can expand a feather bud domain by enhancing the size of the dermal condensation, while PKC activators and PKA inhibitors can expand interbud domains.

There are more examples of periodic patterning in feathers and other integumentary regions (Desmarquet‑Trin Dinh and Manceau, 2024). For example, the development of squamate footpad scales and avian reticulate scales involves similar epithelial appendages; however, their patterns differ. The patterning of reticulate scales aligns with the reaction-diffusion model and has been shown to involve the expression of Shh and β-catenin (Cooper et al., 2019), even though individual placodes are not clearly visible. Instead, the primary domain is divided into smaller units that give rise to individual scales. The footpad scales of squamates develop synchronously across the footpad surface, and their morphogenesis processes share common molecular signaling pathways with feathers (Cooper et al., 2019).

In developing mouse skin, Wnt and Dkk have been found to be the reaction diffusion pair during the periodic patterning of hair primordia (Sick et al., 2006). More molecular interactions involving Eda/Edar/NF-κB and Wnt/β-catenin signaling pathway between epithelium and dermis are also required (Zhang et al., 2009). While FGF 20 and β-catenin are required for dermal condensation specification, scRNA analyses revealed a transitory molecular state prior to the visible formation of dermal condensations (Gupta et al., 2019; Mok et al., 2019). Shh and Wnt signaling gradients also play significant roles in the formation of hair dermal condensations (Qu et al., 2022). Morphogens and the extracellular matrix (ECM) facilitate the elongation of hair primordia in a telescopic fashion (Morita et al., 2021). These findings suggest that while feathers and hairs result from convergent evolution, they are both formed by adopting parallel complex molecular cascades.

2.1.2. Epithelial-Mesenchymal Interactions And Appendage Morphogenesis

Once the appendage primordia are formed, they continue to develop into different phenotypes. Early tissue recombination experiments, including heterotopic and heterochronic recombinants, demonstrated that these phenotypes result from epithelial-mesenchymal interactions. The epithelium must be within a competent time window to respond to the inductive signals, and the dermis also exhibits transient inductive abilities (Dhouailly, 2024; Rawles, 1963). The most insightful classical experiment came from early cross-species epithelial-mesenchymal recombination experiments among developing mouse skin, chicken skin, and lizard skin. The results showed that the epithelia could receive inductive signals from the dermis of different species to produce primordia, but they can only produce the type of appendages intrinsic to their species (Dhouailly, 1973). Thus, the fundamental “make an appendage” signal is shared among amniotes, but specific phenotypes require more specific wiring for that particular appendage. The competence to make ectodermal appendages, may start as being multi-potent, but gradually becomes limited to specific appendage phenotypes.

A tissue recombination matrix using feather and scale epidermis/mesenchyme combinations illustrates this point (Hughes et al., 2011). When embryonic day 9 (E9) metatarsal epidermis is recombined with E7 dorsal skin, which has stronger feather-inducing ability, it can alter the scale fate to form feather-like appendages. However, when E11 or E13 metatarsal epidermis is recombined with E7 dorsal skin, a scale-like patterning with suppressed feather induction occurs, because the time window for epidermal competence has passed. Conversely, when E7 dorsal epidermis is recombined with E11 metatarsal mesenchyme, which has stronger scale-inducing ability, it can alter the feather fate of the dorsal skin and scale-like appendages are induced. However, if E9 or E12 metatarsal mesenchyme are used, they do not exhibit this inducing ability and shadowy weak appendages form. Thus, both epidermal bi-potent competence and dermal inductive abilities are restricted to specific time windows (Hughes et al., 2011). When heterotopic (tissues from different skin regions) and heterochronic (tissue from different embryonic ages) recombinations are used, the consequences will depend on their competence and inductive states. Successful induction occurs only when both states are present.

Within the mouth, several ectodermal organs arise from epithelial-mesenchymal interactions. When competent chicken oral mucosa is combined with feather dermis, it induces the formation of tooth follicle-like structures (Chen et al., 2000). Similarly, when chicken oral mucosa is recombined with mouse dental mesenchyme or when a segment of mouse neural crest is transplanted into a chicken embryo, tooth-like appendages form (Mitsiadis et al., 2003). However, when hair-inducing dorsal dermis from the mouth is recombined with mouse oral mucosa, it induces hair formation (Mäkelä and Mikkola, 2023). This difference can be explained by the developmental timing of the oral mucosa. In the case of chicken oral mucosa, the mucosa has already passed the bipotent phase but can still respond to the inducing signal by forming tooth-like appendages. Conversely, in the mouse oral mucosa, the tissue remains in a bipotent phase and is capable of switching its fate to form hairs.

Taste buds in the tongue are produced by a periodic patterning process (Barlow, 2024). BMP signaling and SHH signaling regulate the placement of taste buds and taste homeostasis in adult mice (Jung et al., 2004, 1999; Shechtman et al., 2023). In the taste bud progenitor epithelium, Sox 2, Pax 9 are important for the progenitor status, then Wnt / β-catenin signaling are important to promote differentiation of taste receptor cells, not lingual epithelium (Kist et al., 2014; Shechtman et al., 2023). This kind of shared progenitors for different ectodermal organs within the oral cavity can be traced back to the bony fish. In cichlids, the oral epithelium can form either teeth or taste buds. Wnt signaling guides tooth and taste bud density while downstream signaling directs the differentiation of pre-placode oral epithelium toward teeth (Shh) or taste buds (BMP) (Bloomquist et al., 2015).

2.2. Global Wave Enables Propagating Pattern Formation

The dorsal tract of the chicken has been extensively used as a model to study periodic patterning because it can be easily observed both in vivo and in vitro (Jiang et al., 2023; Riddell and Headon, 2025; Sengel, 1976). In the developing chicken skin, the first feather row forms along the longitudinal midline of the anterior-posterior axis at embryonic day 6.5 (E6.5). Subsequent rows of feathers are added bilaterally at the lateral edges, following a morphogenetic wave (Mayerson and Fallon, 1985). What constitutes the global wave that directs in vivo propagating feather bud patterning? Some molecular signals may induce cells to generate physical tension, which can facilitate Turing patterning. In chicken skin, a traveling wave of Ectodysplasin A (EDA) moves across the skin and induces mesenchymal cell proliferation, leading to increased cell density and FGF20-mediated mesenchymal cell aggregation (Fig. 2D) (Ho et al., 2019). Mechanical transduction mediated by cell aggregation has been implicated in the translocation of β-catenin into epithelial cell nuclei, which triggers the onset of feather follicle formation (Shyer et al., 2017). The travelling wave initiates at the midline and moves bilaterally creating a growth gradient with older buds along the midline and younger buds at the lateral edges of the tract.

2.2.1. Simultaneous Periodic Patterning

Is the propagation process essential for periodic patterning? Evidence from several lines of research suggests that propagation may not be essential. A reconstituted chicken skin explant with intact epithelium and dissociated dermal cells shows that periodic patterns can re-form simultaneously without bilateral propagation (Jiang et al., 1999). The experiment revealed that at low dermal cell density, no buds formed and as the ratio of mesenchymal to epidermal cells increased the size of the buds remained stable but the number of buds increased until a maximum bud density was achieved. Expressing a Turing inhibitor (BMP) caused the forming buds to be smaller. Expressing Noggin (a BMP antagonist) produced enlarged buds. These data demonstrate that feather bud size is regulated by Turing activators or inhibitors while bud number is determined by the ratio of mesenchymal to epithelial cells (Fig. 2D) (Jiang et al., 1999). Thus, Turing patterning represents local interactions and can occur without an oriented global wave (Inaba et al., 2019b).

Simultaneous periodic patterning can also occur in vivo. Interestingly, simultaneously formed buds are also found in the embryos of flightless birds (ostriches, emus) (Bailleul et al., 2019; Ho et al., 2019). Ostriches form distinct tracts, while the integument of emus has a lower mesenchymal cell density and fails to form tracts (Ho et al., 2019). These flightless birds do not exhibit proliferation waves, and their feather bud arrangements are less precise. Various avian species display different feather patterns, which appear to be based on species-specific pre-patterning of the integument (Bailleul et al., 2019). Comparing feather patterning across multiple species (emus, ostriches, chickens, Japanese quail, zebra finches, and penguins) revealed that the configuration of cells expressing β-catenin within feather primordia can pre-pattern the integument (Curantz et al., 2022). Subsequently, feather buds form either sequentially or simultaneously across the integument in a patterning wave. However, differences in mesenchymal cell morphological orientation (anisotropy) in different species correlate with varying inter-primordia spacing. These results suggest that patterning the species-specific self-organizing feather bud array relies on cell anisotropy. Perturbing the integumenťs tension by inhibiting actin polymerization or stretching the skin modulates mesenchymal cell polarity and subsequent patterning fidelity. The regularly spaced primordia became displaced when tension was reduced in the integument, suggesting that cell polarity is crucial for proper feather primordia patterning (Curantz et al., 2022). This demonstrates that in early feather primordia, cellular polarity generates tension that regulates species-specific feather patterns across the body. Hence, the physical forces from cellular tension contribute to more precise feather patterning.

2.2.2. Channel And Bioelectric Activities

However, on average the diffusion range of morphogens is relatively short (around 200 μm), and the morphogenetic field is usually large, suggesting that additional mechanisms may be involved in Turing periodic patterning. This was found to be the case in mutant zebrafish pigment stripe formation suggesting a Turing-like patterning mechanism without diffusion. In this scenario, cell adhesion and repulsion could act as activators and inhibitors for melanophore stripes (Hamada et al., 2014). Moreover, gap junction Cx40 mutants also result in disrupted pigment patterns (Watanabe et al., 2016). Back to the feather model, the involvement of gap junctions in periodic patterning was demonstrated when gap junction activities were inhibited using 18 α-glycyrrhetinic acid (AGA) or carbenoxolone during the short feather bud stage. Surprisingly, this led to waves of ectopic bud emergence. The first wave occurred in the interbud space, away from existing buds, while later waves appeared near the base of the existing buds (Tseng et al., 2024). These findings suggest that a long-distance inhibitor is mediated by gap junctions. The first wave appears because overall inhibitor activity is reduced, while the second wave emerges as each growing bud also secretes some activators around its base. Thus, while the Turing principle is in action, the threshold for bud formation can vary based on the dynamic morphogen landscape in the developing skin (Tseng et al., 2024).

Recent work shows more cross talk between biophysical events and biochemical signaling. In addition to tissue mechanics, bioelectricity produced through ion channel activity has been proposed to play an important role in morphogenesis (Levin, 2014). An unexpected finding is that when a pulsed exogenous electric field was applied across the feather explant, the orientation of the feather buds collectively reoriented toward the cathode. Epithelium-mesenchymal recombination experiments showed that the “memory” of prior electric stimulation was retained by the epithelium. Perturbing calcium channels by adding inhibitors to the medium caused buds to grow in random directions (Fig. 2E-F). These findings suggest that bioelectricity influences feather bud orientation (Jiang et al., 2021). Measuring endogenous electric currents with a vibrating probe revealed that current flowed into the anterior, posterior, and inter-primordia regions at the early placode stage. However, as the buds developed morphological asymmetry at the long bud stage, the current flowed into the posterior bud but out of the anterior bud. This was mediated by calcium ion channels (voltage-gated and calcium release-activated calcium channels) that spread through the mesenchymal cell population via gap junctions. As the feathers elongate, mesenchymal cells expressing Shh displayed synchronized calcium oscillations, directing the dermal cells to grow distally in a coordinated fashion (Li et al., 2018). Thus, the distal elongation of feather buds, and therefore feather orientation, involves biochemical-bioelectric signaling, with Shh, gap junctions, and calcium activity playing essential roles. These works provide clues for future research to elucidate the role of channel activities in skin morphogenesis, which may reveal a faster communication system that spans a longer distance than diffusible morphogens.

2.3. Adaptive Patterning: Stepwise Assembly To Build Complex Architectures

In addition to forming unique integumentary appendages, these appendage primordia become new organization centers for the assembly of other tissue components (Fig. 2A). Muscles enable individual feathers to change their angles to improve aerodynamics by controlling the morphology of the wing (Hieronymus, 2016). Blood vessels penetrate the feather follicle pulp to provide nutrients during growth phase (Ou et al., 2024) while nerves provide electrical stimulation. Then nerves and blood vessels grow within the follicle pulp and are replaced in each cycle ((Yu et al., 2004). How do skin appendages guide formation of these complex architectures in development? Using the developing chicken skin model, we show two examples of this assembly.

2.3.1. Assembly Of The Dermal Muscle Network

In birds, feather follicles are connected by a dermal muscle network composed of four smooth muscles—two in the front and two in the back (Homberger and de Silva, 2003; Lucas and Stettenheim, 1972). This raises the question: Does this network form through a top-down or bottom-up patterning mechanism? In the top-down scenario, each follicle might secrete morphogens that attract smooth muscle cells to connect directly with the follicle. Alternatively, in the bottom-up scenario, smooth muscle cells might initially be dispersed throughout the area and subsequently stabilize between the feather follicles. Smooth muscle cells were found to migrate radially from each feather follicle. When fibers extending from neighboring buds make contact, the muscle fibers stabilize, forming a network pattern. This process depends on tension forces, as the network fails to form when tension is suppressed (Fig. 2G) (Wu et al., 2019). Additionally, following muscle fiber ablation, neighboring smooth muscle cells were able to rearrange and rebuild the connections, suggesting an equilibrium between free and assembled smooth muscle cells (Wu et al., 2019). The findings indicate that the network is built through a bottom-up self-assembly process rather than a top-down signaling center mechanism.

2.3.2. Assembly Of The Vasculature Network

During development, feather buds initially lack blood vessels. Subsequently, a vascular network connects each bud to a central vascular tree. This raises the question: Does the central vasculature sprout new branches to each bud via angiogenesis, or does vasculogenesis occur within each feather bud, followed by anastomosis with the central vascular tree? Using Tie1 transgenic quail, which express Tie1 in the endothelium, we demonstrated that Tie1-positive cells emerge from the bud dermis and then connect with the central vascular tree. Transcriptome analysis further revealed that Tie1-positive cells appear in the bud first and exhibit characteristics of dermal cells (Fig. 2G) (Ou et al., 2024). Therefore, in developing chicken skin, the vascular network is constructed through vasculogenesis within the target organ (feather buds), followed by connection to the central vasculature, rather than through angiogenetic sprouting from the central vascular tree.

3. Homeostasis In The Adult: Cyclic Renewal And Other Stem Cell-Based Strategies

In the adult animal, the skin undergoes wear and tear, and continuous renewal maintains the same skin thickness through the process of homoeostasis. Skin cell birth, differentiation, and loss are sensed and kept in equilibrium. Additionally, differentiated cells can be reprogrammed to replace stem cells when needed (Gadre et al., 2024). Ectodermal appendages also employ various strategies to maintain their homeostasis. Some, like claws, undergo continuous growth. Others, such as hairs, feathers, and reptile teeth, follow a cycle of renewal and shedding. Certain appendages, like the mammary gland, do not undergo cyclic renewal but can regrow after fulfilling their physiological functions. Here we will compare three cyclic renewable ectodermal organs, followed by the description of several non-cyclic renewable ectodermal organs.

3.1. Cyclically Renewable Ectodermal Appendage Follicles

Cyclic renewal allows species to physiologically reset their ectodermal appendages, enabling changes that occur regionally, seasonally, or during different life stages (macro-environmental signals). These appendages share a fundamental follicle architecture, where stem cells are housed in a protected niche (Fig. 3A, B). This design allows for the shedding of differentiated appendages after they have served their purpose and become worn out. It also allows the mesenchymal niche to serve as an intermediate to receive the macro-environmental signals from outside the follicle.

Figure 3. Homeostasis in the adult: Cyclic renewal and other stem cell-based strategies.

A. Schematic three-dimensional view of feather follicles (Lucas and Stettenheim, 1972). B. Schematic drawings showing feather follicle and cells within (Lin et al., 2013). C. The structure and regeneration cycle of a feather follicle. D. The structure and regeneration cycle of a Tooth follicle. E. The structure and regeneration cycle of a hair follicle. F. Population regenerative behavior among hair follicle stem cells in mouse skin. G. Population feather length control define wing shape for different flight modes. Photos of falcon and tern are by Violet Shen.

3.1.1. Feather Cycling

Feathers also have follicular structures. Their epithelial stem cells are protected within the collar bulge located in the follicle, in close proximity to the dermal papilla for activation (Yue et al., 2005). These stem cells cycle through phases of initiation, growth, regression, rest, and loss (Fig. 3C). In feathers, the dermal papilla undergoes cyclic growth and shrinkage. As the rest stage approaches, dermal papilla stem cells descend to reside in the apical dermal papilla, where they retain dermal progenitor cells for the next feather cycle (Wu et al., 2021). Upon activation, the dermal papilla proliferates and generates peripheral and central pulp. The peripheral pulp is adjacent to the epidermal cylinder and has inductive capabilities. It interacts with epidermal and melanocyte progenitors to generate distinct patterns (Lin et al., 2013; Yu et al., 2002). The central pulp contains new blood vessels, nerves, extracellular matrix (ECM), and provides nourishment. Some dermal stem cells remain quiescent, reserved for the next feather cycle. Wnt-β-catenin signaling activates new stem cells in each cycle to induce a new round of feather cycling , and this appears to be regulated by dermal papilla DKK2 and Frzb (Chu et al., 2014).

3.1.2. Hair Cycling

Hair follicles in the skin undergo continuous cycles of regeneration (anagen), regression (catagen), and rest (telogen) throughout life, a process known as hair cycling (Fig. 3D) (Paus and Cotsarelis, 1999; Saxena et al., 2019). Hair follicle stem cells (HFSCs) reside within the bulge located on the side of each hair follicle, beneath the sebaceous gland (Kostic et al., 2017). The dermal papilla, situated at the base of the hair follicle, acts as a dermal signaling center, modulating the quiescence or activation of stem cells (microenvironmental signals). The interaction between HFSCs and the dermal papilla is crucial for cyclic hair renewal. Upon leaving the niche, HFSCs migrate downward to form transient amplifying cells that proliferate in the hair matrix region, and differentiating toward the distal to form the inner root sheath, hair fiber, and hair medulla (Hsu et al., 2014). The length of the hair fiber is primarily determined by the duration of the anagen phase. Intrinsic and extrinsic factors within each hair’s dermal niche also determine the type of hair that forms (Guard, Awl, Auchene, or Zigzag) (Driskell et al., 2009; Takeo et al., 2023). Wnt-β-catenin signaling between the dermal papilla and HFSCs are key to controlling hair follicle cycling and interactions between these cell compartments control the phenotypes of individual hairs.

3.1.3. Tooth Cycling

While tooth can be replaced in fish (Square et al., 2021) and sharks (Nicklin et al., 2024), most mammals, including humans, experience only one cycle of tooth renewal. While mouse incisors grow continuously, they do not undergo cyclic renewal. In contrast, reptile teeth exhibit a robust cycling ability (Henriquez and Richman, 2024; Razmadze et al., 2024). For example, alligators can replace their teeth repetitively throughout their lives. This capacity arises from the presence of a complex tooth family unit at each tooth position, consisting of a functional tooth, a replacement tooth, and a dental lamina. The dental lamina is located on the lingual side of the tooth family unit. In the mandible, the dental lamina forms a continuous structure connecting every tooth family unit. Stem cells are housed in a bulge of the dental lamina adjacent to the replacement tooth. When the functional tooth is lost, the replacement tooth is stimulated to develop into a new functional tooth, while the dental lamina bulge initiates the formation of a new replacement tooth (Fig. 3E). A new dental lamina bulge, containing stem cells, then branches off the lingual side of the developing replacement tooth, ensuring the continued renewal of the tooth family unit (Wu et al., 2013). Although these tooth units do not form a single, well-organized follicle, the dermal niche surrounding the dental lamina stem cells regulates their interactions. Molecular analysis reveals that the Wnt/β-catenin pathway plays a key role in this regulation (Tsai et al., 2016; Wu et al., 2013).

Like hairs and feathers, lizards, snakes and geckos teeth also are derived from epithelial (placode) mesenchymal interactions (Henriquez and Richman, 2024; Razmadze et al., 2024). Tooth formation requires the odontogenic mesenchyme. Combining dental epithelia with mesenchyme from another organ site generates the skin appendage from which the mesenchyme was derived. Geckos grow 40 functional teeth which are replaced by successional teeth, each underlain by a successional tooth and a dental lamina. Despite differences in follicle design, these three ectodermal appendages share the remarkable ability for cyclic renewal that is dependent on Wnt-β-catenin signaling, and it is a result of convergent evolution. A more detailed comparison of these three types of cycling appendages in a table can be seen in Wu et al., 2024 (Wu et al., 2024).

3.2. Non-Cyclic Renewable Ectodermal Appendages

Here we compare three examples of non-cyclic renewal ectodermal appendages: the mammary gland, representing a cyclic renewable ectodermal organ; the sweat gland, representing an ectodermal appendage that maintains a consistent size; and nails or claws, representing continuously growing appendages. Recent works have led to new understanding into their homeostasis regulation at the chromatin level, and interested readers are referred to those reviews (Botchkarev, 2017; Branch et al., 2024).

3.2.1. Mammary Glands

Although mammary glands do not form a follicular architecture, they exhibit remarkable adaptability, expanding in response to pregnancy and involuting after nursing is complete. In mice, an immature ductal structure, formed through epithelial-mesenchymal interactions, is present at birth. During puberty, branching morphogenesis is triggered by estrogen, growth hormone, and IGF1, leading to the formation of a ductal tree within the fat pad. Pregnancy further induces alveolar development, enabling milk secretion to nourish the young. Once the newborn is weaned, the gland undergoes involution, returning to its pre-pregnancy state and preparing for another growth phase in the next pregnancy (Macias and Hinck, 2012).

3.2.2. Sweat Glands

Sweat glands play a crucial role in thermoregulation by secreting moisture onto the skin's surface when external temperatures rise. They also respond to the fight-or-flight reaction by producing moisture (Aldea et al., 2023). Although sweat glands do not form dermal condensations, they do develop an epithelial placode (Dingwall et al., 2023). The expression of Engrailed 1 (En1) and EDA is essential for sweat gland development. In conditional En1 knockout mice, where En1 is deleted in basal epithelial cells, there is a significant reduction in sweat gland numbers, accompanied by an increase in hair follicle production (Dingwall et al., 2023). RNA-seq analysis revealed that Dkk4, a Wnt signaling inhibitor, is highly expressed in developing sweat glands but not in hair follicles. In En1 knockout mice, reduced Dkk4 expression in basal cells allows hair follicles to form. Additionally, S100a4 is upregulated in the dermis beneath developing sweat glands in an En1-dependent manner. When S100a4-expressing dermal cells were selectively ablated using a tamoxifen-inducible diphtheria toxin, En1 expression persisted, but sweat gland development was significantly impaired (Dingwall et al., 2023). This study identified an En1-dependent dermal niche critical for sweat gland development.

3.2.3. Nails Or Claws

Newborn mice and humans have the remarkable ability to regenerate the digit tip after injury (Johnson and Lehoczky, 2022; Lee et al., 2024). It was shown that Msx1 induction of BMP4 was an essential part of the regenerative response in fetal mice (Han et al., 2003). Interestingly, the blastema that forms after amputating the terminal phalanx expresses BMP4 which aids in phalanx regeneration (Han et al., 2008). K15, a stem cell marker, is expressed in the label-retaining putative stem cell population within nails. Studies in BMP knockout mice have shown that BMP signaling is crucial for the proper development and differentiation of nails. These nail stem cells have dual fates and can become epidermis or nails, depending on the need (Leung et al., 2014). During nail regeneration, the formation of a blastema—a mass of proliferating cells—occurs first, requiring the activation of the canonical Wnt-β-catenin pathway (Takeo et al., 2013). Lgr6, a marker found in nail stem cells, is also present in the blastema, indicating its role in digit tip regeneration (Lehoczky and Tabin, 2015). Gene expression during this process is regulated by histone acetylation status. Conditional deletion of histone deacetylase 1 and 2 in the epidermis leads to the loss of stem cells and causes nail dystrophies, including the formation of supernumerary hyperpigmented nails (Hughes et al., 2014).

3.3. Emergence Of Collective Regenerative Behavior In A Population

Skin appendages were thought to cycle in response to intrinsic factors but several examples have been found in which multiunit skin appendages can cycle in response to extrafollicular macro-environmental factors. For example, the ability of telogen hair follicles to initiate a new hair cycle is regulated in part by factors released from the underlying adipose tissue. This mechanism causes hairs to regenerate in a progressive wave (Plikus et al., 2008).). In a second example, hair follicle injury induced immune activation can lead to a minor or major regeneration response, and it depends on whether a quorum sensing threshold is reached (Fig. 3F) (C.‑C. Chen et al., 2015).

3.3.1. Regenerative Hair Wave

As discussed in section 3.1.1, hairs cycle through anagen, catagen, telogen and exogen stages. However, signals stored in adipose tissue within the extrafollicular environment can determine whether HFSCs are competent or refractory to activation. Interestingly, the amount of adipose tissue varies during hair cycling indicating that hair follicles and adipose tissue have a mutual regulatory relationship. The hair cycle refractory state is produced by the synthesis and storage of BMPs (BMP2 and 4) within the mature intrafollicular adipose tissue (Plikus et al., 2008). Adipose tissue-derived BMPs decrease at the late telogen, priming the HFSC environment to become competent to regenerate. Thus, BMP levels partition telogen into refractory telogen and competent telogen phases (Fig. 3F, upper panel). Depending on the amount of Wnt signaling, anagen can be partitioned into propagatory or autonomous anagen phases. HFSCs progress through hair cycling stages by continuous integration of intrinsic cell state and extrinsic environmental input. A two dimensional cellular automaton model with different thresholds was developed to predict the regenerative pattern of a hair follicle population. The model can describe the more typical wave propagation pattern seen in dorsal mouse skin as well as the fractal like complex patterns observed in rabbit skin in which multiple stem cells are activated in one hair follicle (Plikus et al., 2011).

Further molecular studies show more molecular interactions between hair follicles and the dermal adipose tissue. Immature preadipocytes in the dermis can secrete follistatin and platelet-derived growth factor (PDGF) to promote HFSCs activation (Festa et al., 2011). Dermal white adipose tissue also produces hepatocyte growth factor (HGF) that enhances human hair growth and pigmentation. Hence, although stem cells are housed in separate compartments of individual hairs, they can communicate with one another through environmental factors that help to determine whether to initiate a new hair cycle.

3.3.2. Activation Of Hair Follicles in a quorum sensing fashion

Environmental factors also play a role in coordinating hair regeneration following injury. Plucking 200 hairs at different follicle densities and in different shapes elicit different regenerative responses. Only the plucked hairs regenerate below a threshold level. However above the threshold level, both plucked and unplucked hairs regenerate amounting to a 5-fold increase in the number of regenerated hairs (Fig. 3F, lower panel) (C.‑C. Chen et al., 2015). This collective response to injury is based on a quorum sensing mechanism regulated through an immune cascade. Here, CCL2 recruits TNF-α-secreting macrophages which elicits an NF-κB - Wnt signaling pathway to activate hair follicle stem cells (C.‑C. Chen et al., 2015). This quorum sensing pathway enables the skin to ignore subthreshold injury levels but to induce a large-scale regenerative response when the signals are above threshold levels. Hence the immune system is another extrafollicular factor that helps to regulate hair cycling.

3.3.3. Control Of Relative Feather Length In A Follicle Population

Adjacent flight feathers on avian wings have tapered differences in their lengths which shape the wing for aerodynamic flight (Fig. 3G) (Li et al., 2017). This morphological characteristic is likely due to interactions between tissues and their environments or to interactions between neighboring repeated units. It is unknown if individual feather lengths is regulated autonomously or through an exogenous wave. Like feathers, the length of alligator teeth grow to be either long or short, depending on their position within the jaw. The long teeth in the upper jaw will meet the short teeth region in the lower jaw and thus enhance biting. How this is collectively controlled remains unknown. It has been postulated that alligator tooth length can be regulated by lateral inhibition among adjacent teeth or by the innate growth potential of each individual tooth (Osborn, 1998).

4. Regional Specificity Optimizes the Usage of the Integumentary Surface

Different body regions display different appendage phenotypes (e.g., trunk, wing, leg, etc.), thus optimizing the function of the integument (Fig. 4A). We try to elucidate how the form, color, and keratin types of skin appendages are controlled and how that particular phenotypes are limited in different skin regions.

Figure 4. Feather branches provide functional forms for endothermy, flight and display.

A. Regional specificity of feather forms: birds have flight, contour, downy, and tail feather types (Lucas and Stettenheim, 1972). B. A schematic drawing of rachis morphogenesis with representative molecular control and morphological transition. C. A schematic drawing of barb morphogenesis. D. Birds with different flight modes show different bio-architectures of the rachis of their flight feathers. On the right is a feather trapped in amber dated 99 million years ago. Cross sections show ancient rachis shape (Chang et al., 2019; Wang et al., 2020).

4.1. Feather Branches Provide Functional Forms For Endothermy, Flight, And Display

Feathers are integumentary organs that exhibit remarkable regional specificity and undergo significant temporal changes (Fig. 4A) (Chen et al., 2021). During embryogenesis and at hatching, birds are covered by downy feathers. These radially symmetric, plumulaceous (fluffy) feathers provide essential warmth to young chicks. As birds grow, a primary transition occurs during the early juvenile stage, where the downy feathers are replaced by bilaterally symmetric contour feathers. Later, at sexual maturity, a secondary transition leads to the replacement of these feathers with region-specific types, including bilaterally asymmetric feathers in the wings and tail, which are crucial for flight, and longer feathers in the hackles and saddles, which are used to attract mates (Fig. 4A, right panel). In the following sections, we will elaborate on the differences in the barb branches, main shaft (rachis), and the formation of symmetric and asymmetric feather forms.

4.1.1. Feather Barbs And Feather Shape

The feather rachis branches into structures known as barbs, which can develop barbules at their tips. Barbs are formed at the base of the feather follicle through the invagination of the epithelium into the underlying mesenchyme. Marginal plates adjacent to the barb ridges later undergo apoptosis, releasing the barbs to create lightweight feathers. In plumulaceous feathers, the barbules on proximal and distal barbs have the same shape, and the barbs are separated from each other. These feathers are primarily used for thermoregulation and do not form a feather vane. In contrast, pennaceous (vane-shaped) feathers require the adjacent barbs to organize into an aerodynamic surface. In these feathers, the distal barbules form hooklets that link to the proximal barbules of adjacent barbs, similar to how Velcro works, creating a planar surface in contour and flight feathers (Chang et al., 2019). Contour feathers feature a combination of proximal plumulaceous regions, which maintain body temperature, and distal pennaceous regions, which aid in aerodynamics. Flight feathers, on the other hand, have medial-lateral asymmetric pennaceous branches, which enhance lift and enable flight.

The phenotypes of feathers are modulated by the dermal papilla. Removing the anterior dermal papilla reduces the size of the distal pennaceous vane, while removing the posterior dermal papilla diminishes the proximal plumulaceous region. This suggests that different regions of the dermal papilla control distinct feather branch phenotypes. Transcriptomic analyses reveal that Wnt agonists and antagonists are expressed contralaterally along the anterior-posterior axis, forming a Wnt gradient. Inhibiting Wnt signaling in the dermal papilla of a pennaceous feather alters cell adhesion properties and disrupts barbule cell formation. Conversely, increasing Wnt2b expression in the anterior regions of plumulaceous feathers promotes hooklet formation. These findings demonstrate that the shape of hooklets is regulated by Wnt2b secreted from the dermal papilla (Chang et al., 2019).

Feather shape enables birds to adapt to diverse environments. For instance, the Taiwan blue magpie exhibits feathers of varying lengths and different symmetric versus asymmetric configurations (Fig. 4A). The shape of the vane is controlled by factors that fine-tune the core branching modules (Fig. 4C). Feather branches are regulated along the proximal-distal axis by the relative activity levels of FGFs and Sprouty (Yue et al., 2012). FGF and Notch signaling mediate basal cell filopodia interactions during this process (Cheng et al., 2018). Cross-sections of differently shaped feathers show that symmetry or asymmetry is determined by the relative position of rachis and the barb generative zone. If positioned directly opposite the rachis, the feather becomes symmetric. If offset to one side, barbs on one side of the feather will be longer than those on the contralateral side. Increasing GREM1 expression expands the barb generative zone and can cause branching of the rachis (Yu et al., 2002). Mesenchymal GDF10 establishes the position and topology of the rachis, while GREM1 interacts with an anterior-posterior Wnt gradient to set the position of the barb generative zone, determining whether a feather is bilaterally symmetric or asymmetric. A retinoic acid gradient, regulated by CRAPBP1 and CYP26B1, influences feather width (Li et al., 2017). Additionally, the angle at which barbs tilt from the rachis is regulated by planar cell polarity (Lin and Yue, 2018). It has been suggested that interactions between core branching modules and morpho-modulatory models can produce the diverse array of feather forms observed (Chuong et al., 2013).

4.1.2. Feather Rachis (Main Shaft) And Feather Rigidity

The position of the feather rachis is at the anterior end of the feather filament and is determined by a Wnt3a gradient (Yue et al., 2006). The rachis, derived from the supra-basal epidermis, differentiate into a cortex and vacuolated medulla, become a bi-composite material that is both strong and lightweight (Sullivan et al., 2019). During rachis formation, the timing and distribution of molecular expression are important. K17 is expressed in the cortex and β-catenin and desmoglein 1 are expressed in both the cortex and medulla. K75 is enriched in the medulla. As the medulla expands, the cortex extends to encase it, and the medulla ultimately becomes vacuolated. In frizzled chickens, where the K75 molecule is mutated, the medulla is significantly reduced and each feather shaft is defective, resulting in a curved feather phenotype(Fig. 6C) (Chen Siang Ng et al., 2012). The formation of rachis configurations is guided by signaling through the BMP and TGF-β pathways. Ectopic expression of BMP4 thickens the dorsal cortex and reduces medulla cell size, while suppression of TGF-β via Ski inhibits medulla formation and decreases the expression of keratins typically found in the cortex (Fig. 4B) (Chang et al., 2019). In an individual bird, the rachis architecture is different at different life stages and in different body regions. The rachis can have a simple rod-like structure in downy feathers, a complex structure with cortical ridges and medulla organization in flight feathers, and an intermediate structure in contour feathers. These structures control the rigidity of individual feathers (Chang et al., 2019).

Figure 6. Diverse keratins offer remarkable variations for intricate appendage architectures.

A. A schematic drawing of expression patterns of α-and β-keratin genes in different skin appendages (Wu et al., 2015). B. Genomic organization of keratin clusters (Liang et al., 2020). C. The chicken frizzle feather with mutated KRT75 (Chen Siang Ng et al., 2012). D. In situ hybridization to show representative keratin expressions in different body regions (Liang et al., 2020). E. Evolutionary origin and diversification of epidermal differentiation complex which made barrier proteins (Strasser et al., 2014). β-keratins, also named CBP (corneous β-proteins), are derived from EDC members.

The rachis serves as the backbone of each feather. It is composed of an outer cortex and inner porous medulla. The diverse bio-architecture of the rachis in flight feathers among eagles, sparrows, chickens, and flightless birds like emus and ostriches is noteworthy. The varying morphologies of the rachis in flight feathers enable birds to adapt to different flight modes (Fig. 4D). For example, 1) Eagles use their feathers to catch updrafts and soar through the sky. These feathers have a large medulla cell pore size, making them light weight and flexible. 2) Sparrows rely on rapid wing flapping for flight. Their medulla cells also have a large pore size but their low body mass enables them to fly by flapping their wings. 3) Chickens exhibit short bursts of flight and need a more powerful rachis. In the chicken flight feather, the ventral region of the dorsal cortex folds into the medulla to form cortical ridges that strengthen the architecture. They have a small medulla cell pore size. 4) Ostriches belong to flightless birds. The bird has a large body mass and the rachis of its flight feathers exhibits a small pore size in its medulla. This structural characteristic contributes to the feather's strength and flexibility, enabling it to withstand the aerodynamic forces encountered during flight by birds with different flight modes (Chang et al., 2019).

4.1.3. Regional Differences In The Mammalian Skin

In the mammal, there are also regional differences. One major difference is in the palm and foot sole. For example, mouse plantar skin has foot pads with few hairs, but many sweat glands. Transgenic mice overexpressing noggin from a K14 promoter exhibit the conversion of foot pad sweat glands into hairs (Plikus et al., 2004). Genetic deletion of Dkk2 promoted the formation of fully functional hair follicles in the plantar skin (Song et al., 2018). In humans, the palm also does not have hair but expresses many sweat glands. Humans form fingerprints by making ridges on the fingertips (Glover et al., 2023). The ridges are patterned by a Turing reaction-diffusion mechanism using Wnt, EDAR and BMP pathways as seen in hair formation. The synthesis and subsequent spread of signaling waves is determined by the initiation site microenvironment and the complex contours of the digit. The distribution of these signaling molecules induces proliferative bands of epithelial cells to form under the differentiated suprabasal layer. This system enables ridges to form with remarkable patterning diversity.

4.2. Complex Color Patterns On The Integument Surface Convey Communication Signals

The integument serves as a canvas for animals to display vivid color patterns that either attract mates during mating season or help the animal blend in with surroundings to avoid the attention of predators. These color patterns can result from varying pigmentation levels or structural colors across different regions of the body. They may appear as stripes, spots, rosettes, or blotches, with the potential to change temporally and spatially, serving diverse functional purposes. In birds, these color patterns can be distributed across the body in what is termed “macro-patterning” (Fig. 5A) (Inaba and Chuong, 2020) which may align with specific body regions, such as dorsal, ventral, or femoral areas, or even the tail. Notably, these patterns are not always consistent with feather tracts. Color differences can also manifest within individual integumentary appendages in a phenomenon known as “micro-patterning” (Fig. 5B) (Inaba and Chuong, 2020). This can produce solid colors, spots, or stripes of varying widths and spacings. These differences are often observed between the medial and lateral, or proximal and distal regions of a feather vane.

Figure 5. Complex color patterns on the integument surface coney communication signals.

A. Macro- (across the body) color pattern in different body regions. Barbet photo by Violet Shen. Peacock photo by CM Chuong. B. Periodic stripe formation in Japanese quail and periodic expression of Agouti genes (Inaba et al., 2019b). Quail photo by Zhou Yu. C. Representative pigment patterns within a feather. D. Structure and biochemical basis of yellow feather pigmentation in budgerigars (Cooke et al., 2017).

4.2.1. Across-The-Body Macro-Color Patterning

In many cases, these patterns do not correspond to known anatomical structures; instead, they form color boundaries that run across skin domains. However, in some instances, macro-color patterns align with specific organs, such as the eyes, ears, or wings (Fig. 5A). The mechanisms underlying these patterns have long intrigued researchers. These patterns may arise from interactions between melanocytes, fibroblasts, and mesenchymal cells. Melanocytes, derived from the neural crest, produce melanosomes that are transferred to keratinocytes, while adjacent fibroblasts regulate melanocyte differentiation. Possible mechanisms may be based on autonomous patterning by melanocytes or fibroblasts. For example, many young birds and mammals exhibit longitudinal stripes that may provide camouflage in their environment (Fig. 5B). In different quail and pheasant species, the number of black and yellow stripes on the dorsal skin varies, with the yellow stripes resulting from pheomelanin production influenced by agouti-expressing fibroblasts. Remarkably, when somites from these birds were transplanted during early embryogenesis, the resulting pattern adhered to the original species of the somite, suggesting that dermatome cells in somites possess intrinsic information for agouti expression patterning (Haupaix et al., 2018). However, the translation of this somite-derived information into large-scale periodic yellow stripes remains to be studied.

Another study revealed the autonomous ability of neural crest-derived melanocytes to form periodic black stripes. When Japanese quail neural crest cells were transplanted into host chick embryos with different endogenous patterns or into non-pigmented hosts, the resulting pigmentation pattern mirrored that of the Japanese quail. Further, the suppression of connexin40 (cx40) with MITF-specific promoters in melanocytes altered pigmentation patterns, suggesting that early melanocytes may form a network in which pigment patterning requires communication via gap junction channels (Fig. 5B) (Inaba et al., 2019b). How melanocyte and fibroblast interactions mediate patterning remains to be worked out.

The above is just an example. There are more cases of across-the-body macro-color patterns, and their underlying mechanisms await further exploration. Furthermore, adult birds can display different color patterns with seasonal dependence and sexual dimorphsms. These will be discussed further in the section on secondary feather transitions.

4.2.2. Within-An-Appendage Micro-Color Patterning

Different pigment patterns can also form within a single hair or feather. It is more obvious in feathers than hairs because the larger feather vane surface enhances visibility. This patterning mechanism originates within the feather follicle, where melanocyte progenitor stem cells play a major role. These progenitor cells are arranged in a horizontal ring around the proximal collar bulge of the follicle. During the growth phase, they continuously send melanocytes upward into the epithelial cylinder of the feather as it grows. The different pigment patterns form by modulating the presence, arrangement, or differentiation of these melanocytes. The decision of melanocyte progenitors to differentiate or migrate is influenced by their interactions with surrounding fibroblasts in the peripheral pulp and feather filament keratinocytes. These interactions determine whether the melanocytes will differentiate and move into the feather filament, contributing to the complex pigmentation patterns observed in feathers (Lin et al., 2013). When we see a particular color, it may not be produced by the same mechanism. For instance, both the absence of melanocytes or the suppression of melanocyte differentiation can create white regions.

Similar to macro-color patterning, micro-color patterning can arise from melanocyte-autonomous mechanisms or interactions with fibroblasts or epithelial cells (Fig. 5C). In the Barred Plymouth Rock chicken, for example, horizontal black and white stripes across the feather vane result from a mutation in CDKN2A, leading to premature melanocyte differentiation and depletion of melanocytes (Schwochow Thalmann et al., 2017). Developmental studies show that white stripe regions lack melanocytes, prompting melanocyte stem cells to send out a new wave of melanocytes, which form the next black stripe. So the white zone in this case is due to the absence of melanocytes (Lin et al., 2013). Yet, white spots on the feather vane of Guineafowls show MITF positive melanocyte prognitors. They exhibit white color because of suppressed melanocyte differentiation. These white spots face agouti-positive peripheral pulp dermal fibroblasts. In Silver Laced Wyandotte chickens, pigmentation at the feather edges is influenced by sexually dimorphic agouti expression in the peripheral pulp that affects pigmentation patterns (Lin et al., 2013).

Genetic studies provide further insights. For instance, the melanotic mutation causes eumelanin (black pigmentation) to expand into regions normally pigmented with pheomelanin (red or brown color) toward the outer edge of feathers. This mutation reduces cx40 (GJA5, a gap junction protein) expression levels. Normally, cx40 is transiently expressed in the melanoblasts and epithelial cells in the collar and branch-forming regions. Reduced gap junction activity may decrease cell communication, altering MC1R or agouti patterning and leading to modified melanocyte distribution and periodic pigmentation patterns (Li et al., 2021a). This finding parallels the observation that longitudinal black stripe formation across the body depends on gap junctions (Inaba et al., 2019b) and is reminiscent of studies showing that reduced gap junction activity leads to the emergence of new feather buds (Tseng et al., 2024). Gap junctions may facilitate the transport of small molecules, such as calcium and cAMP, between cells or mediate electric coupling of the action potential. How these molecules work together to generate periodic patterning is another unresolved question.

4.2.3. Structural Colors And Combinatorial Coloring

Some species exhibit vibrant structural colors alongside pigmentation. For instance, parrot feathers are brightly pigmented with psittacofulvins, which produce yellow to red colors (McGraw and Nogare, 2005). The budgerigar (Melopsittacus undulatus) can display white/blue or yellow/green colors in the same bird. The blue coloration is due to structural color, while yellow pigments are synthesized by MuPKS, a polyketide synthase (Fig. 5D). Genome-wide association mapping identified MuPKS, originally expressed in liver and other visceral organs in many birds. However, in budgerigars, MuPKS is expressed in feather keratinocytes, specifically in the axial plate epithelia. These axial plate cells deposit yellow pigments on the barbs, which disintegrate as the feather matures. When this happens to a white feather, the feather appears yellow; when it occurs in a blue feather, the feather turns green as the colors combine (Cooke et al., 2017).

4.2.4. Mammalian Color Patterns

Mammals also exhibit remarkable across-the-body color patterns. For example, members of the feline family can form black or brown stripes, spots, and intermediate patterns on their body surfaces. Pigmentation patterning in mammals is thought to be controlled in two steps. First, the skin color pattern is pre-patterned during development by the distribution of Transmembrane aminopeptidase Q (Taqpep). Later in embryonic development, the secretion of Endothelin-3 (Edn3) from mesenchymal cells promotes melanocyte proliferation. After birth, regions lacking Taqpep with high levels of Edn3 form dark pigments, while regions with Taqpep, which have low levels of Edn3, form yellow stripes (Kaelin et al., 2012). How are these patterns established in developing skin? Further study showed that periodic patterned epidermal ridges are established prior to the initiation of hair formation. These thin and thick regions form yellow and black areas, respectively. Transcriptome analysis revealed that thick regions express higher levels of Dkk4, a Wnt signaling antagonist. Further analysis indicated that nuclear Wnt signaling and Edar, a downstream target, are upregulated in thickened skin regions (Kaelin et al., 2021). This suggests that Turing patterning may be involved in setting up these transient epidermal ridges during embryogenesis, which are later translated into overt pigment patterns in the adult animals.

Another interesting example is the longitudinal stripes observed in some flying rodents. The dorsal pigment pattern of African striped mice (Rhabdomys pumilio) features dark longitudinal stripes separated by light stripes. RNA-seq analysis identified arista-like 3 (Alx3) as an epithelial gene upregulated in the light stripes. Overexpressing Alx3 significantly reduced Mitf levels, potentially by forming a complex with Pax3 to lower Mitf transcription (Mallarino et al., 2016). Additionally, a gradient of Soluble Frizzled 2 in the developing dorsal skin of embryos has been observed. The model suggests that a global gradient and local Turing patterning may together explain the emergence of these longitudinal pigmented stripes (Johnson et al., 2023).

Other species may use different molecules to display color patterns. For example, horses with the Dun phenotype are light in color with a prominent dark dorsal stripe. The expression of TBX3 in Dun horses produces this dun coloration by asymmetrically diluting pigmentation, resulting in a lighter overall coat color. TBX3 inhibits pigment synthesis in a subpopulation of hair cortex epithelial cells by modulating KITLG, a key factor in melanocyte migration. In contrast, non-dun horses exhibit a deeper coat color with only a weak dorsal stripe. This difference arises from mutations that reduce TBX3 expression, leading to increased KITLG levels. Interestingly, no significant changes in melanocytic pathways were observed in the dark dorsal stripe of non-dun horses, suggesting that the stripe's pigmentation is regulated independently of the broader coat color mechanisms (Imsland et al., 2016).

Mammalian skin appendages can also exhibit within-a-hair color patterning. For example, the agouti in a single mouse hair can be present only distally. Hedgehog spines display periodic striking black and white segment patterns. These examples suggest that there may be multiple mechanisms at play to achieve similar color patterns. In principle, complexity patterns can result from a combination of a global gradient, originating from the body midline or specific organs, and local self-organizing patterning mechanisms (Hiscock and Megason, 2015). Such combinatorial mechanisms offer numerous possibilities for generating diverse color patterns at different scales: within a single feather, across the entire body with regional specificity, and during an animal's lifetime through appendage molting and transition.

4.3. Diverse Keratins Offer Remarkable Variations For Intricate Appendage Architectures

Keratins (Krts) constitute the largest subgroup of intermediate filaments. They are a major component of the skin and integumentary appendages. Keratins can be subdivided into α-Krts and β-Krts based on their secondary structures (Wang et al., 2016). α-keratins are present in all vertebrates, while β-keratins are found only in birds and reptiles, and also are named corneous β-proteins (CBP) (Holthaus et al., 2018). Each keratin type includes a family of members that evolved through gene duplication, followed by changes in their gene sequences and regulation, leading to diverse temporal and spatial expression patterns. α-keratins are composed of pairs of acidic (Type I) and basic (Type II) subunits. β-keratins (CBP) were found in feathered dinosaurs that lived over 200 million years ago, and the evolution of keratins enabled improved keratin interactions necessary for flight in birds. Keratin variation offers a range of physical characteristics which play a crucial role in the various functions of the skin and ectodermal organs, such as providing mechanical support (e.g., bird beaks) and thermal insulation (e.g., furs) (Holthaus et al., 2018; Lazarus et al., 2021). In chicken feathers, a dramatic example is the mutated keratin 75, in which the rachis medulla is poorly developed and the rigidity of the rachis is weakened, leading to the frizzled feather appearance (Fig. 6C) (C S Ng et al., 2012).

4.3.1. Region-Specific Expression Of Keratin Genes

The avian Krt gene family consists of more than 30 α-Krt and more than 140 β-Krt genes. The combination of these many α- and β-keratin genes contributes to the enormous morphological and structural variations of avian skin appendages, with β-keratins conferring intricate complexity in building intra-appendage architectures (Fig. 6A) (Ng et al., 2014; Wu et al., 2015). Sequencing of the bird genome focusing on β-keratins revealed that a group of β-keratins is clustered within the epidermal differentiation complex (EDC) on chromosome 25 (Chr25) (Fig. 6E) (Strasser et al., 2014). β-keratins on chromosome 25 are organized into five subclusters expressed in specific structures such as the claw, feather, feather-like structures, scale, and keratinocytes (Wu et al., 2015). ChIP-seq analysis with H3K27ac and H3K4me1 markers revealed that these β-keratin clusters are independently regulated by simple enhancers (Fig. 6B) (Liang et al., 2020). Additionally, the entire gene cluster containing 48 β-keratins on Chromosome 27 (Chr27), encodes feather keratins, yet their expression within different regions of a single feather are regulated by chromatin loops (Liang et al., 2020). These loops are different in different feathers and scales, and are proposed to contribute to the large keratin reservoirs required to build complex feather architectures. How are these regulated at the epigenetic level?

4.3.2. Epigenetic Regulation Of The Keratin Gene Clusters Across The Body And Within-A-Feather Keratin Differences

The ability to regulate keratin expression across the body and also within a feather, expand the variations of feather shapes, textures and functions. Using developing chicken skin as a model, multi-omic approaches showed that β-keratin gene clusters employ two distinct epigenetic strategies to organize their chromatin for regional expression (Liang et al., 2020). For across-the-body keratin differences, a single enhancer drives the co-expression of all subclustered keratins (Fig. 6B, D). For within-feather keratin differences, differential intra-cluster chromatin looping elaborates the expression of different sets of keratins (Fig. 6B, D). Three categories of identified factors (competence factors, regional specifiers, and chromatin organizers) have been proposed to explain how these loops are established. Analyses suggest regional specification occurs through competence factors (e.g., AP1) that make chromatin accessible, regional specifiers (e.g., Zic1) that target specific genome regions, and chromatin regulators (e.g., CTCF and SATBs) that configure looping. This demonstrates how epigenetic regulators control chromatin structures modulate skin appendage variations at both macro and micro levels. This clarifies how chromatin interactions are involved in epidermal lineage commitment and provides new insights into the mechanisms that achieve skin heterogeneity.

Moreover, Special AT-rich Sequence-Binding Protein 2 (SATB2), a potential genome organizer, mediates the Krt gene clusters in different developing skin regions. Nuclear matrix-associated SATB proteins have been shown to serve as scaffolds to help form tissue-specific chromatin architectures (Cai et al., 2003). Misexpression of SATB2 in feathers, beaks, and claws causes epidermal differentiation abnormalities and cluster-wise misexpression of Krt and EDC genes (Kohwi et al., 2021; Lin et al., 2022, 2021). Hence chromatin loops, requiring nuclear matrix-associated proteins like SATB2, are essential for the proper expression of clustered β-Krt genes during development. Other factors also affect the epigenetic regulation of keratin expression. SATB2 can regulate subcluster switching in both α- and β-keratins. Over expressing SATB2 induces appendage dysplasia (Lin et al., 2022) and suppresses the epidermal differentiation cysteine-rich protein (EDCRP) subcluster in feathers and the epidermal differentiation proteins containing cysteine histidine motifs (EDCH) subcluster in beaks (Lin et al., 2021). This finding indicates that the avian EDC regulates regional and temporal expression at the gene/gene cluster level.

5. Regulatory Patterning: Integumentary Metamorphosis Enables Adaptability

Physiological cyclic regeneration occurs throughout the life of each bird. This cyclic renewal allows appendage phenotypes to change following each molt, thereby expanding the variation of skin appendages. Hair or feather stem cells respond to modulatory signals influenced by multiple environmental factors, including body region, age, sex, and seasons. These can be seen in human beings in which the facial hair distribution changes from childhood to young adults to the elderly (Fig. 7A). These factors can modulate differently shaped skin appendages in various regions and times, helping individuals adapt to their specialized niches (C.‑C. Chen et al., 2016; Hsu and Fuchs, 2022).

Figure 7. Temporal transition: Regulatory patterning enable a high level of adaptability.

A. Regional difference and temporal difference of skin appendages in human and chicken (Chuong et al., 2012). B. Primary and Secondary feather transitions in birds. 1^st row: Juvenile frigate birds show white and fluffy feathers. In the adult, the feathers transition to powerful flight and contour feathers, including red breast sacs that are used to attract a mate. 2nd row: feathers from the chicken (Chen et al., 2024; Widelitz et al., 2019). Frigate birds photo by CM Chuong. Darwin images from Wikipedia website.

5.1. Temporal Transition Of Feather Phenotypes Over A Lifetime

Feathers exhibit different characteristics depending on the body region, life stage, and season, responding to factors such as temperature, diurnal cycles, and hormones. Each follicle cycle is influenced by local signals from within the feather follicle as well as external signals. These signals can produce dramatic changes in feather morphology, pigmentation patterns, and appendage rigidity, optimizing adaptation at different times during an individual’s life. The appearance of baby-to-adult skin appendages represents an example of organ-level metamorphosis (Chuong et al., 2013) and is driven by the varying needs throughout different life stages. This process can be termed "integumentary metamorphosis."

5.1.1. Primary Feather Transition

Young chicks, who stay within their nest and are cared for by their parents, exhibit radially symmetric feathers that cover most of their bodies, providing them with a warm coat of downy feathers. During the primary transition, the natal down is replaced with juvenile feathers that have a bilaterally symmetric vane (Fig. 7B, left panel) (Chen et al., 2024). These feathers also develop a rachis and a strong feather sheath, enabling the chicks to begin flying and leave the nest. This transition is marked by the up-regulation of the canonical Wnt signaling pathway, which is involved in forming the rachis and promoting bilateral symmetry. The reorganization of extracellular matrix proteins fosters peripheral pulp development, leading to feather branch formation. The stem cell niche necessary for cyclic regeneration is formed, in part, by the expression of smooth muscle actin. Additionally, Sox14 in the juvenile feather sheath induces several “scale keratins,” which are stiff keratin subtypes originally identified in scales, providing feathers with additional strength (Chen et al., 2024).

5.1.2. Secondary Feather Transition

The juvenile feathers are next replaced by adult feathers during the secondary transition. While some feathers retain downy branches toward their base to continue providing warmth, they form feather vane in the distal feathers for communication and flight. Another major change is the unique color patterns produced in the adult birds which can be used for species recognition or mating choice. These are vividly shown in frigate birds. (Fig. 7B, right panel). An elegant study on the complex color patterns of estrildid finches revealed that despite their different color patterns, the basic domain patterns are conserved and similar. These domains are then filled with different colors depending on the species. Some transcription factors for these domains have been identified. The results suggest that pigmentation colors and patterns can be regulated by local transcription factors that interact with pigment synthesis pathways (Hidalgo et al., 2022). Understanding how these varied pathways can interact to produce complex color patterns and how visual cognition in the avian brain interacts with these patterns will enhance our understanding of the mechanism and function of secondary feather transition.

One of the most dramatic changes is the production of sexually dimorphic feathers observed in many, but not all, avian species (Fig. 7A) (Prum, 2017; Widelitz et al., 2019). How are sexual dimorphic feathers produced? Injecting male White Leghorn chickens with exogenous estradiol significantly shortened their saddle feathers compared to controls. Injecting females with testosterone had only a minor effect on feather length. However, blocking androgen conversion to estradiol using Femara dramatically increased feather length. These data demonstrate that estradiol dictates the length of female feathers. Hormones also influenced the ratio of pennaceous to plumulaceous barbs. The pennaceous-to-plumulaceous barb ratio is higher in males than in females. Exogenous estradiol increased the percentage of plumulaceous branches in male feathers, while exogenous androgen had no effect on the ratio in females. Transcriptomic data showed that male feathers up-regulated Sprouty expression, while females increased SMAD7 expression in the periodic patterning pathway of epithelium and mesenchyme. The pigmentation pathway was up-regulated in the male collar epithelium, where melanocyte stem cells reside, while Agouti, which suppresses pigmentation, was up-regulated in females (Widelitz et al., 2019). These experiments demonstrate that estradiol guides the female feather and pigmentation patterns, while a lack of estradiol guides the male patterns. Yet the molecular interface between sex hormones and feather morphogenesis remains to be elucidated.

5.1.3. Altricial And Precocious Birds.

The evolution of altricial mode is an important process in both avian and mammalian development, provide baby birds with longer time period to enhance parent-baby bonding: more time for the baby to develop social relationship and more opportunity for the baby to learn (Mota‑Rojas et al., 2023). Based on this behavior, birds can be classified into altricial (e.g., finch) and precocial (e.g., chicken). The hatchlings of altricial birds are almost naked, and require parents to keep them warm and fed, whereas those of precocial birds are covered with natal down and can survive even without parents. One major altricial trait is the delay of feather development in newborn chicks. Altricial traits have evolved independently several times in birds, thus altricial traits in finches, parrots, and pigeons are the results of convergent evolution (Chen et al., 2019). How the timing of feather development is delayed hetero-chronically is not known. In Zebra finch, FGF 16 is shown to be one of the molecules involved in this signaling cascade. Over-expression of FGF16 can suppress SHH and induce FGF 10, resulting in reduced feather development (C.‑K. Chen et al., 2016). While the whole signaling cascade controlling altricial remains to be worked out, this study provides a clue to the regulatory divergence in natal down formation between precocial and altricial birds.

5.2. Puberty, Aging, And Other Macro-Environmental Factors

Throughout life, integumentary appendage phenotypes can change physiologically at different ages (Chuong et al., 2012). The human fetus surface is covered by lanugo hair, which turns into vellus hairs after birth. Hair in some regions then become terminal hairs, such as those in eyebrows, scalps, etc. During the development of sexual maturity terminal hairs develop in regions such as the face (moustache, beard) and pubis (Fig. 7A). Transition between hair filament phenotypes at different ages are made possible through cyclic renewal. In some elder individuals, terminal hairs can become vellus hairs and androgenetic alopecia can develop. Extrafollicular environmental factors such as the circadian clock (Plikus et al., 2013), seasonal variations, or hormonal cycles associated with puberty or pregnancy (Randall, 2007; Schneider et al., 2009) can significantly impact mammalian hair cycling and phenotype specification. These are particularly apparent in mammals that live in the wild, where changes beyond hairs can be seen in other integumentary appendages, such as horns.

5.2.1. Aging

Aging can lead to hair greying (O’Sullivan et al., 2021; Rosenberg et al., 2021) and/or hair follicle miniaturization and subsequent loss (Matsumura et al., 2016). In aged skin, bulge HFSCs are still present but their numbers are diminished in the hair germ (Garza et al., 2011; Lei and Chuong, 2016). In aging mouse dorsal skin, the expression of activators (e.g., Wnts and Follistatin) decreases, while the expression of inhibitors (e.g., Bmps, Dkks, Sfrps) increases (Chen et al., 2014), slowing hair cycle wave progression. Extrafollicular expression of Dkk1 and Sfrp4 inhibits the Wnt signaling pathway during the anagen phase in aged mice. To test whether the change is intrinsic to the hair follicle or can be affected by the environment, a patch of older mouse skin was transplanted to the backs of young mice. Interestingly, hairs from older skin enter anagen near the transplant border (about 200 μm zone) where there are interactions with the adjacent younger skin, but hairs in the center of the transplanted skin stay in telogen. Low BMP2 and higher follistatin was detected in the old skin after transplantation. This experiment demonstrates that the young skin environment could partially rescue hair cycling in the older skin, possibly through molecular diffusion (Chen et al., 2014).

While HFSCs try to maintain their lineage identity, the ECM environment for HFSCs is gradually altered during skin aging (Ge et al., 2020). External conditions can modulate the speed of this aging process. Among the ECM, Collagen XVII was identified as a key supporter for HFSCs, and accumulated DNA damage from the environment can lead to degradation of collagen XVII during aging. HFSCs lose their normal molecular signatures and undergo precocious epidermal differentiation (Matsumura et al., 2016). The depleted HFSCs result in miniaturized hair follicles. These aging-related phenotypes can be mimicked by experimentally depleting Collagen XVII and can be prevented by maintaining Collagen XVII in HFSCs. Decreased Collagen XVII synthesis during aging leads to a reduced number of melanocyte stem cells, resulting in lower tyrosinase activity, decreased melanin production, and eventually hair greying (Tanimura et al., 2011).

5.2.2. Stress

Stress can affect various tissues differently and has long been considered a contributor to hair greying and hair loss. Stress-induced hair greying occurs through the loss of melanocyte stem cells, which is shown to be triggered by the activation of the norepinephrine receptor β2 adrenergic receptor (Adrb2). Normally, norepinephrine is secreted from sympathetic nerve axon termini located very close to the hair stem cell bulge, which contains melanocyte stem cells. Injecting norepinephrine into the dermis without stress also resulted in grey hairs. So how do melanocyte stem cells respond to sympathetic nerve stimulation? Stress dramatically increased melanocyte stem cell activation, as indicated by the upregulation of proliferation (Cdk2) and differentiation markers (Mitf, Tyr, Tyrp1, Oca2, and Pmel), leading to melanocyte stem cell depletion. However, blocking melanocyte proliferation maintained melanocyte stem cells within their niche, even under stress (Zhang et al., 2020). Human cortisol is upregulated in response to stress and is often referred to as the "stress hormone." In mice, the equivalent hormone is corticosterone, which is produced by the adrenal glands and also increases under chronic stress. In a mouse model of chronic stress, researchers found that elevated levels of corticosterone act on the dermal papillae, suppressing the expression of Gas6. This suppression keeps the hair follicle stem cells in a quiescent state, thereby prolonging the hair follicles' resting phase (Choi et al., 2021).

The mammalian hair follicle macro-environment consists of adipose tissue, nerves, extracellular matrix, and hormones. Each can regulate the health and number of hair follicle or melanocyte stem cells. Stress and aging can induce the stem cells to proliferate and differentiate, leading to stem cell depletion, which causes hair greying or alopecia. Interestingly, it seems that a younger macro-environment can rescue these consequences (Koester et al., 2021). These findings hold great promise for future treatments for these common conditions.

6. Evo-Devo Of The Integumentary Organs

The integument is located at the interface of the organism and its environment. Through the evolutionary time scale, integuments are often modified to provide new functions that enable animals to adapt to changing environments or to populate new habitats. Successful adaptations that survive natural selection are passed on to future generations. We have identified two general strategies used in amniotes to provide diversity. The cyclic renewal of skin appendages in birds and mammals supports changes in the number, size, shape, rigidity, color, and distribution of hairs or feathers in each cyle by modulating the stem cell niche (C.‑C. Chen et al., 2016). We call this micro-evolution of ectodermal organs. On the other hand, there are numerous skin appendages on the body. Hence, animals can afford to try new integumentary appendage phenotypes in different specialized regions of the body without compromising its survival. This is most obviously seen in the evolution of feathers in feathered dinosaurs. We call this macro-evolution of ectodermal organs.

6.1. Micro-evolution.

6.1.1. Feather phenotypes

The successful evolution of feathers depends on the novel appearance of two integumentary appendage features: 1) follicles allow feathers to undergo molting and physiological regeneration and 2) feathers branch to generate different temporo-spatial forms of feather filaments. We will discuss the origin of the follicle in the macro-evolution section. For the evolution of feather branch forms, fossils from the Jehol Biota in China, Germany, Russia, and Canada revealed the existence of feathered dinosaurs during the Jurassic and Cretaceous periods. These fossils exhibit a large spectrum of protofeathers with unusual characteristics, reflecting the evolutionary experimentation of different functional feather forms (Fig. 8A, B) (Xu et al. 2014; Benton et al. 2019; Dhouailly et al. 2019; Ksepka 2020; Xu 2020; Prum and Brush 2002). In the early evolutionary stage, these are called filament appendages or protofeathers. For example, Tianyulong had monofilamentous feathers, Sinosauropteryx had radially branched feathers, and Dilong had bilaterally branched feathers. Scales were found on the legs of therapods and ornithischians (a clade of mainly herbivorous dinosaurs) (Fig. 8D, compare to modern scale in lizard in 8E). Presumably, feathers at this time were needed for thermal regulation and communication. Early regional differences can be observed in a non-avian theropod (Xu et al., 2009). More apparent variation such as large symmetric feathers on the wing can be seen in Caudipteryx. Crown feather differences can be seen in Mesozoic birds such Longirostravis (Fig. 8F) (Hou et al., 2004). Asymmetric flight feathers were seen on the legs of Microraptors Sapeornis and wings of Archaeopteryx. The transition of flight feathers to the wings is still not completely understood although this would have required modification of the forelimb shape, musculature, rachis, barb and barbule forms to form individual feather vanes, and collective flight feather morphogenesis to improve aerodynamics (Xu et al., 2014).

Figure. 8. Micro- and macro-evolution of feathers.

A, B, Schematic representative of filamentous appendage, protofeathers and feathers in feathered dinosaurs and Mesozoic birds (Ksepka, 2020). C. Major molecular events in feather development which may also occurred in evolution (Liang et al., 2020; Li et al., 2017; Wu et al., 2018; Xu et al., 2014). D. Scale like protuberances (Plikus, 2024) in early amniotic skin dated 286 million years ago (Mooney et al., 2024). E. Green iguana scale. F. The Mesozoic bird, Longirostravis, shown with a toothed beak and the specialization of crown feathers (Hou et al., 2004). G. The chicken comb is converted into feathers in Polish breeds (Photo by Ping Wu). H. Definition of feathers (Wu et al., 2018).

We proposed that different molecular circuits act during different stages of feather development and regeneration (C.‑F. Chen et al., 2015; Li et al., 2017). Molecular circuits of feather branches involved in vane formation, medio-lateral asymmetry, and remarkable heterogeneous rigidity have been discovered (discussed in section 4.1.1 and Fig. 8C). Looking at extreme feather morphologies in extant birds also can provide insight into their role in adapting birds to their environments. The bi-composite configuration of the rachidial cortex and medulla again comes into play. The rachis of Emus is composed of concentric cortex and medulla. For example, the Ostrich rachis has an ill-defined border between the cortex and medulla suggesting they are unable to provide strength. Hummingbirds evolved a way to have downward strokes when the wing moves forward as well as backward. Their rachis has a stiff dorsal and ventral cortex with a thin lateral cortex which offers flexibility. Penguins swim in water and their flight and tail feathers predominantly have a stiff cortex while the contour feathers can trap air in their medulla to regulate temperature (Chang et al., 2019).

Both the barbs and rachis did not achieve their current architecture in one step (Xu, 2020). Cretaceous feathers embedded in amber about 100 million years ago have preserved 3-dimesional structures and provide new insight into barbule and rachis evolution. The contour feathers in amber exhibit bilaterally symmetric barbs with very long barbules without hooklets. The feathers form a primitive vane by overlapping the proximal and distal barbules from the adjacent barbs (Chang et al., 2019). It suggests that formation of the Velcro like mechanism connecting barbules evolved later. The primitive feather vane may work as a plane made of overlapping flight feathers.

The complex rachis architecture also evolved gradually. Feather fossils embedded in amber 99 million years ago show a dominant rachis structure without a medulla. Yet, the morphology of these ancient and extinct rachises presents complex cortexes, flattened and with strange ridge shapes, reflecting an early evolutionary progression toward having a strong but light structure to support the increasing complexity of barb branches (Fig. 4D, right panel; (Carroll et al., 2019; Wang et al., 2020). This suggests that, prior to the evolutionary development of the medulla, feathers employed alternative strategies to achieve a strong yet lightweight rachis.

6.1.2. Hair Phenotypes

Humans differ from other primates in many ways, but perhaps the most striking feature is the evolutionary degeneration of our body and facial hair (Fig. 7A). In humans, there is a much larger facial area that exhibits vellus hairs. The facial traits of early humans may have evolved to enhance cognition and communication. The human face plays a crucial role in social interaction, emotion, and communication, suggesting that changes in facial skin were influenced by social contexts. In addition, the evolution of facial musculature has been linked to conveying information relevant for social interactions. The connection of these facial muscles may parallel the adaptive patterning of the dermal muscle network described in section 2.3.1 (Wu et al., 2019) and warrants further investigation. Hypertrichosis, also known as the ‘werewolf syndrome’, is a rare condition with excessive hair growth all over the body and the face, representing a potential atavistic mutation — the reappearance of an ancient evolutionary trait that is normally suppressed. Genomic studies found hypertrichosis patients had chromosome deletions, insertions, inversions, or duplications, instead of point mutations in a specific gene (Sun et al., 2009; Zhu et al., 2011).

Regional changes in hair length seems to have occurred often during evolution. This may be because there are many potential pathways that can lead to long hair phenotypes. African elephants have short hairs. When they migrated toward cold, subarctic habitats, they began to exhibit a dense hair coat with exceptional length as seen in wooly mammoths to help them adapt to the cold. Some of them exhibit tusks with a flat-shaped tip, which is likely to help them dig plants for food from the frozen land (Roca et al., 2009). This is an example of micro-evolution of tooth shapes.

The long hair phenotype also can occur only in specific body regions. This is seen in the belly hairs of the yak. In several feline species, the long hair on the tips of their ears, known as ear tufts, helps enhance their hearing. For this, a clue was obtained from the study of the Koala mutant mouse, that has a hairy ear and muzzle. The Koala mutant shows a chromosome inversion that led to changes of the epigenetic landscape, and eventually enhanced hair length by changing Hoxc gene cluster expression (Yu et al., 2018).

6.1.3. Beak Phenotypes

The need to feed efficiently in different environments has always been a pressure for evolutionary modification. Darwin’s classical work analyzing Galapagos finch beak shapes inspired his concepts of evolution. The Galápagos Islands present either dry or wet landscapes which necessitated modifications of the original finch beak shape. Their beaks evolved to be 1) small and symmetric, 2) broad and deep, or 3) long and pointed. Morphogenesis of these diverse beak shapes relies on the timing of bone morphogenetic protein (BMP) expression (Abzhanov et al., 2004). A related developmental study showed the growth zone differences between chicken and duck beaks (Wu et al., 2006, 2004b). Higher BMP4 levels were seen in the duck beaks compared to chicken beaks. Expressing exogenous BMP4 increased beak length, width and depth. Expressing noggin, a BMP antagonist, reduced beak size. These studies suggest that BMP4 levels can be tuned to modify beak growth zones. What could be upstream of BMP4? By sequencing 120 different individuals of Galapagos finches, a genetic study identified the ALX1 locus as a strong contributor to beak shape variation (Lamichhaney et al., 2015). Furthermore, whole-genome data from 3955 individual Galapagos finches revealed six loci responsible for the variation in beak size caused by natural selection during a drought. These studies may provide clues for us to understand the bridges between environment factors and the actuators of beak shapes (BMP4).

6.2. Macro-Evolution

The generation of novel ectodermal organs is called Macro-evolution. A highly successful evolutionary novel ectodermal organ can help produce a new branch of animals that can then expand their ecospace and produce further diversification. This can be seen in Aves which, among other traits, are built on the successful macro-evolution of feathers and beaks. This also occurred in mammals, which built their success on the macro-evolution of hairs and mammary glands. Before this, in the evolution of early amniotes, the first essential major eventswas the formation of a water-proof barrier that allowed animals to live on the land. This was built on the evolution of the epidermal differentiation complex (EDC) (Strasser et al., 2014). There are also other macro-evolutionary ectodermal organs that do not have such profound impacts. Combs and wattles appear in some avian species but they are not widespread. Horns are successful adaptations for ruminants and rhinos as defense weapons, but they are not present on all mammals. Baleen is made of hair keratins for filter feeding, and a transition of oral mucosa from a tooth fate (Peredo et al., 2018). While it is highly successful, its presence is limited to Mysticeti.

6.2.1. Early Amniotic Skin: Scale-Like Protuberances

Recently, in a Paleozoic cave system dating back 286 million years, well-preserved early amniote fossilized skins were discovered. These fossils exhibit a non-overlapping, pebbled, tuberculated epidermis, forming scale-like protuberances (Mooney et al., 2024; Strasser et al., 2014). This may represent an early attempt to strengthen localized regions for mechanical support or water resistance. A unique characteristic is the scattered arrangement and large spacing between the "scales," which contrasts with the tightly packed scales of modern reptiles that lack interbud spacing (Fig. 8D, E) (Plikus, 2024). One possible explanation can be based on available dermal cells. As observed in reconstituted chicken skin explants, when dermal cell density is low, some appendage primordia form randomly, however once dermal cells reach the bud-forming threshold appendages become evenly spaced with maximum packing density throughout the region (Fig. 2D). It can be speculated that early amniote skin underwent evolutionary processes aligning with periodic patterning principles to achieve effective adaptation.

6.2.2. Feather In Feathered Dinosaurs And Birds

Since we cannot trace back to study the evolution of novel molecular mechanisms, we rely on molecular perturbations in existent avians to study gene function. Molecular perturbation with retinoic acid (RA) (Dhouailly et al., 1980), Wnt/β-catenin (Widelitz et al., 2000), Notch/Delta pathway activation (Crowe and Niswander, 1998), BMP pathway suppression or Shh pathway (Cooper and Milinkovitch, 2023) activation can convert chicken scales to feathers. Scale to feather changes are also present in Silkie chicken and pigeon mutants. In chickens, the application of growth factors or there inhibitors generated several intermediate conversions from scutate scales to feathers (Wu et al., 2018). A real feather is defined to have five criteria, including a localized growth zone (LoGZ), invagination, branching, feather β-keratin, and dermal papilla. Different phenotypes may represent various combinations of these five criteria. Phenotypes produced in Sox18, β-catenin, and RA-treated samples meet all criteria necessary to be considered real feathers. The spectrum of these “intermediate morphotypes” suggests that the five key morphogenetic events can be uncoupled and specific criteria can be induced by specific molecular perturbations (Wu et al., 2018). Given the finding, that some features of filament integuments may evolve in different reptiles (Xu, 2020), it is likely that successful extant feathers evolved by integrating these distinct features, not by sequentially events. Eventually, the most successful combination lead to today’s feathers, and the less successful morphotypes gradually became extinct. These major features define extant feathers today (Fig. 8H).

Interestingly, two scale-feather converters, Spry2 and β-catenin, identified from birds can also induce an ectopic LoGZ and grow feather bud-like elongated appendages from embryonic alligator skin, but they do not go on to form follicles or branches, because they do not have those molecular circuits (Wu et al., 2018).

6.2.3. Beak Versus Jaw

The evolutionary novelty for the origin of beak formation is 1) the elongation of the jaw, 2) formation of the horny sheath. In extant birds, beak formation is often accompanied by a loss of teeth. The molecular basis for the transition from toothed dinosaurs to beaked birds is largely unknown. Ontogenetic vestigialization of alveoli in two theropod dinosaur lineages may represent transitional phenotypes in beak evolution (Wang et al., 2017). BMP4 may be involved in peramorphic expansions of keratinized epithelia and linked with progenetic truncation of odontogenesis in beaked theropod lineages. Mesozoic birds still have teeth, and the teeth undergo cycling just like alligators (Fig. 8F) (Wu et al., 2021). So, the emergence of new ectodermal organs likely requires a transitional period over millions of years and does not occur abruptly.

6.2.4. Glands Versus Hairs

Earlier experimental work has shown that retinoic acid can convert hairs to glandular structures (Hardy et al., 1990). This gives the hint that ectodermal organs can alter their fates when exposed to strong modulators. This is demonstrated experimentally in the mouse. As ectodermal organs, both hair follicles and sweat glands are developed from morphologically similar epithelial placodes during embryogenesis when Wnt signaling instructs progenitors within the epithelial sheet. Higher Shh but lower BMPs and FGFs in the murine dorsal back skin mesenchyme guide hair follicle formation, whereas lower Shh and Noggin but higher BMP in the mouse foot mesenchyme permit sweat gland production (Lu et al., 2016). Interestingly, hair follicles and sweat glands coexist in many regions of human skin, although they are developed at different embryonic stages (i.e, in human skin, hair follicles develop at week 10 on the head and sweat glands emerge at week 20). An adaptive human Ectodysplasin receptor varian, EDARV370A, has been identified as the key player for increased scalp hair thickness and the number of active eccrine glands in the Han Chinese (Kamberov et al., 2013). Moreover, increasing Engrailed 1 expression from the human-specific enhancer promotes the production of eccrine glands, contributing to the evolution of humans’ signature thermoregulatory capabilities (Aldea et al., 2021).

6.2.5. Comb Versus Feather

Phenotypic variations, in feather forms or colors, are often seen in the crown, or scalp, of many birds. The earliest known modification of the crown feather is seen in Longirostravis (Fig. 8F) (Hou et al., 2004). The White Leghorn chicken exhibits a large comb juxtaposed with short cranial feathers on its scalp, whereas the Polish chicken features many long crest feathers surrounded by shorter cranial feathers (Fig. 8G). This provides the opportunity to find out the molecular mechanism involved in comb – feather transition. Notably, the pronounced crest feathers of the Polish chicken dominates much of the scalp area, significantly reducing the size of its comb and relocating it to an anterior position near the beak. This transition from comb to crest feathers in Polish chickens has been found to correlate with elevated levels of HoxC8 expression (Wang et al., 2012). Recently, it was reported that a 195 bp mutant duplication within the intron of HoxC10 contributes to the crest phenotype observed in Polish chickens (Li et al., 2021b). Further work showed there is co-linear expression of HoxC gene cluster members in the midline dermis from scalp to the tail. The HoxC-containing topologically associating domain (TAD) is normally closed in the scalp but open in the dorsal and tail regions, allowing multiple long-distance contacts. In the scalp of the Polish chicken variant, the 195 bp duplication opens the TAD, resulting in high HoxC expression (Hsieh Li et al., n.d.). The work opens up frontiers to study the control of region-specific epigenetic profiles and how HoxC may regulate downstream morphogens to make different integumentary organ phenotypes.

6.2.6. External Ear In Mammals

External ears do not exisit in reptiles or birds. External ears evolved from cells originally used for the first and second pharyngeal arches (Passos‑Bueno et al., 2009). Auricles are shaped into different morphologies to help mammals in different environments gather and focus acoustic waves which offers great advantages for animal survival. Two recent papers advance our knowledge about the macro-evolution of external ears. One paper shows the auricle cartilage represents a new forms of cartilage, made of lipid vacuoles (Ramos et al., 2025). Another paper showed that the gill gene regulatory program in the fish was repurposed to form the external ear (Thiruppathy et al., 2025). Thus, while the exact evolutionary origin of the external ear remains to be investigated, these studies beautifully illustratrate that the epithelial tissue can interact with mesenchymal tissue derived from the somite, mesodermal or cranial neural crest cells to form a new integumentary organ.

7. Summary And Future Directions

Here we examine the variations of the amniote integument, including the oral cavity lining. It becomes clear how varied and adaptable these structures can be. They exhibit highly functional physical forms and biochemical compositions. These integumentary phenotypes are body region-specific, allowing animals to optimize different body parts for specific uses, making the integumentary surface highly adaptable. Furthermore, animals have different needs at various life stages, and integument metamorphosis—the temporal transition of integumentary phenotypes—adds an additional layer of adaptability.

In this review, we aim to decipher the principles that underlie complex pattern formation in the integumental organs of amniotes. First, periodic patterning segments the integument into multiple units, giving it the power of multiplicity. Through global or local cues, spatial specificity develops, enabling the integument surface to serve multiple functions. The formation of stem cell-based follicles allows for temporal renewal and the transition to new phenotypes at different stages of life. By combining Turing-based periodic patterning and stem cell-based follicle development, the four-dimensional patterning mechanisms of the integument provide a vast repertoire of diverse appendage phenotypes for selection to act upon. This enables individuals or species to adapt to diverse environments and venture into new ecological niches.

To take a very brief historical overview, research into the mechanism of integumentary organ formation began with pioneering tissue recombination experiments in the 1960’s. These studies provide insights into the distinct contributions of epithelial and mesenchymal tissues (Dhouailly, 1973; Dhouailly et al., 2004; Sengel, 1976). Starting in the 1990’s, our understanding expanded to include the roles of morphogens in skin morphogenesis, particularly in chickens (Jung et al., 1998; Li et al., 2017) and mice (Millar, 2002), and more recently reptiles (Tzika et al., 2023). The identification of follicle stem cells in mouse hairs (Cotsarelis et al., 1990) and avian feathers (Yue et al., 2005) has further elucidation new understanding in the molecular control of stem cell cyclic renewal (Fuchs et al., 2001) and phenotypic specification (Li et al., 2017) in adult animals. In parallel, since the 1990’s, 2000’s, a series of fossilized feathered dinosaurs has inspired more thinking into integument evolution (Dhouailly et al., 2019) and into the evolution of feather forms (Prum and Brush, 2002; Xu, 2020; Xu et al., 2014).

Exploring future directions in integumentary research, several areas merit increased attention: 1) Epigenetic Basis of Temporal Transition and Regional Specification. While the influence of morphogens on appendage phenotypes has been established, understanding their epigenetic regulation, including epigenetic memory, remains crucial. Advancements in omics technologies position us for significant breakthroughs in this frontier. 2) Crosstalk Between Biophysical Inputs and Biochemical Signaling in Morphogenesis: Historically, the focus has been on biochemical signaling's role in morphogenesis. Emerging research underscores the importance of biophysical inputs, including tissue mechanics and bioelectricity, in modulating biochemical pathways during skin development and wound healing. Investigating these interactions can deepen our comprehension of morphogenetic processes. 3) Integration of Integumentary Organs Within the Organism. Beyond neural and hormonal mediation, recent findings highlight the intricate regulation of integumentary stem cells by macro-environmental niches and adaptive immune signals. From point 1 to 3, these layers of controls offer novel therapeutic potentials for conditions from alopecia to advancements in regenerative medicine. 4) Applying Integumentary Principles to Evolutionary Studies. The cyclic renewal characteristic of integumentary organs facilitates continuous optimization and adaptation throughout an individual's lifespan. The multiplicity and dispensability of these organs allow species to experiment with new skin appendages when exploring novel ecological niches. Large-scale genomic sequencing enhances our capacity to understand these evolutionary adaptations. Focusing on these areas promises to advance our knowledge of integumentary systems, fostering innovations in both basic science and clinical applications.

Returning to the fundamental question of variation and adaptation posed at the beginning of this article, integumentary organs stand out due to their unique bio-architectures. By studying these visible and experimentally manipulable phenotypes, we can reveal the principles that drive the generation of diversity. Other organs, such as those in the digestive system with their biochemical pathways or the brain with its neural circuitry, employ different strategies to evolve adaptability. However, we believe that fundamental principles such as multiplicity, spatial specificity, and temporal transitions are applicable, in various ways, to different organ systems. Using ectodermal organs as a blueprint, we hope the concepts identified here will provide new insights into how variation is generated within individuals and species, fostering robust adaptation and evolution.

Highlights:

1. Amniote integuments are complex yet adaptable throughout their lifetime

2. In development, tissue patterns are set via de novo periodic and adaptive patterning

3. Hair, feather and tooth undergo cyclic renewal because of stem cell niche topology

4. Regional specification is set by local epigenetic profiles for effective function

5. Temporal transition of adult appendage phenotypes allow adaption to environments