4.1 Introduction

For a time of such staggering importance that it eclipses birth or death, “gastrulation” certainly doesn’t have the same universal recognition. What is this process that commands such esteem, and how is it relevant to anatomy and physiology? Fundamentally, any study of vertebrate anatomy and physiology must consider vertebrate embryology, encompassing the processes that produce organismal shape and function—that is, the set of developmental events, including gastrulation, that an embryo undergoes as it proceeds from a unicellular fertilized zygote to a distinct, complete, multicellular organism. In addition, differences in form and function between related organisms necessarily involve evolutionary changes in these embryological processes. Thus, we will introduce a few key concepts of developmental biology and then examine how those concepts are illustrated in vertebrate biology and evolution.

4.2 Major Concepts of Developmental Biology

The progression of a fertilized zygote, the single-celled result of the fusion of egg and sperm, into a fully formed multicellular individual requires coordination between complex developmental processes that can differ in their specifics in different organisms. However, these mechanisms share fundamental similarities that operate across animal groups.

Differentiation

Although muscle cells and nerve cells, for example, appear obviously distinct in their shape and properties, vertebrate cells typically share the same full set of genetic information (that is, the genome of all cells is identical—with some interesting exceptions when it comes to cells of the immune system!). Therefore, distinctions between cell types must arise from differences in which portions of the genetic information are switched on or off (i.e., which genes are actively expressed and which are not). Differentiation is the process cells take on the shape and properties of their defined cell type, also referred to as their cell fate (Figure 4.1). Cell fates can be considered broadly (e.g., nervous system) or narrowly (e.g., pyramidal neurons). An important cell fate defying our typical expectations for what cells do is apoptosis (also known as programmed cell death), in which the fate of the cell is to self-destruct in a carefully controlled manner, helping to sculpt the shape of limbs and digits (Box 4.1—Cell Death in the Feet of Waterfowl) or to eliminate self-reactive cells in the immune system.

Determination

The fate of most cells is already established well before those cells actually take on the physical properties of their final cell fate. Determination is the set of mechanisms cells commit to a cell fate. Determination can be thought of as a progressive narrowing of the fates available to a given cell. Cells that can achieve any fate, of either the embryonic body or extraembryonic supporting tissues, are referred to as totipotent. Cells of the very early embryo are totipotent but typically lose this extreme flexibility after only a few cell divisions. Similarly, pluripotent cells have the ability to produce any embryonic cell type but not extraembryonic tissues (such as the placenta and amnion), which provide physical support and nutrient access for the developing embryo. Embryonic stem (ES) cells derived from mammalian embryos can be maintained in petri dishes in an undifferentiated state but, when returned to an embryo or stimulated with growth factors, can generate all types of embryonic cells, demonstrating their pluripotency (Figure 4.1). Determination is affected by the internal status of the cells, such as the activation state of specific genes, and by external conditions, which may include either nonbiotic environmental cues or communication between cells in an individual.

Induction

The cells, tissues, and organs of an organism must be carefully organized and regulated to ensure proper interaction and function. For example, the alignment of structures within the vertebrate eye is important for effective vision, and this critical alignment is achieved by a series of reciprocal interactions between cells that regulate the process (Figure 4.2). This specific case illustrates the more general concept of induction, by which one set of cells or tissue can trigger changes in the determination or differentiation of neighboring cells. The communication of these inductive signals can be via direct contact or by transmission of soluble signaling molecules, such as sonic hedgehog or decapentaplegic (many of these molecules were originally discovered in fruit flies or other model organisms, and their discoverers got to choose their odd, funny, or whimsical names).

Morphogenesis

Morphogenesis comprises the processes at the levels of cells and tissues that generate anatomically relevant shapes, including tubes, branches, and layers (Figure 4.2). These processes involve changes in cell growth, motility, adhesion, and division that together can create the emergent properties of tissues and organs. For example, if the cells lining the intestine failed to adhere to each other properly, the leaky gut structure would be significantly disruptive. Conversely, if cells on opposite sides of the intestine adhered to each other inappropriately, the tube would collapse and be blocked. The proper formation of a tube, whether in the nervous system, the intestines, or the lungs, requires specific cell movements and interactions.

Modularity

Modules, broadly considered, describe independent or distinguishable sets of parts that can interact with other modules to build more complex structures. Because organisms typically display distinct physiological, behavioral, or anatomical traits, and many of these traits can vary independently (or mostly independently) from one another, we can assert that organisms are modular in nature. This modularity, or the distinguishability and independence of complex traits, enhances evolutionary change by allowing variation of isolated traits without detrimental effects on other traits. Modularity can occur at multiple levels, including the morphological (e.g., different bones within the vertebrate limb have evolved independent size and shape changes, such as the elongated finger bones of bats or the shortened arm bones of whales and dolphins) or the developmental (e.g., shifting boundaries of Hox gene expression in nervous and skeletal system leading to differences in brain and axial skeleton development).

4.3 Fertilization and Early Development

Gametes and Fertilization

A convenient place to start in our consideration of vertebrate development is the formation of the zygote via fertilization of a haploid ovum (or egg cell) by a haploid spermatozoon (or sperm cell) to reconstitute a complete diploid genome. Fertilization can occur externally (outside of the reproductive tract) or internally (within the reproductive tract). Even animals with internal fertilization may then develop externally (for example, birds).

Both egg and sperm are gametes, haploid products of meiosis that carry just half of the genetic material of the parent. Gametes are cells of the germ line, which are the only cells of the organism to pass along their hereditary information. By contrast, the cells of the soma (also called somatic cells) form the structures of the organism but do not transmit information to subsequent generations. Thus, gametes represent a specialized cell fate within an organism, and developing sperm and eggs undergo specific differentiation events to generate cells that are, respectively, small, elongated, and motile or large, spherical, and nonmotile.

In different species of vertebrates, sperm shape (such as the hook-shaped heads of rodent sperm) and size (from 28 µm for a porcupine to 349 µm for a honey possum) can vary significantly. Sperm consist of three segments—head, midpiece, and flagellum (tail or terminal piece)—and the length of each component can vary independently of the others (Figure 4.3). Tail length appears to have an evolutionary relationship to fertilization mode (internal or external), but other variations in shape and size do not always have obvious evolutionary correlates.

Within a given species, the egg is always larger than the sperm cell, but across different groups, egg size also can vary widely. The egg contains all the nutrients and other components (such as RNAs, transcription factors that switch on or off RNA production, and signaling molecules that regulate physiological responses or transcription factor activity) for the initial developmental steps. Embryos that develop externally must have sufficient energy reserves and materials to complete development to a stage where the individual can obtain its own resources. Most of the resources are contained within the yolk, which consists of proteins and energy-rich lipid molecules, and the amount of yolk will affect how long an embryo can survive before reaching a feeding stage. Eggs with small amounts of yolk are termed microlecithal, whereas macrolecithal eggs have substantial quantities of yolk (Figure 4.4). By contrast, internally developing embryos are less dependent on yolk because they typically obtain nutrients from their parental host.

Fertilization requires the entry of sperm into the egg, fusion of the sperm and egg nuclei, and activation of cellular processes within the newly generated zygote. Sperm entry, sometimes envisioned as the culminating event of a single “victorious” sperm, may in many cases actually require the action of multiple sperm in order to penetrate the various external support cells and cell membranes of the egg. Eggs and sperm use various mechanisms to ensure that only sperm of the same species can undergo fusion with an egg, and evolutionary changes in the components of these systems can act as isolating mechanisms that may lead to speciation.

Sperm entry triggers several responses within the egg. First is the cortical reaction, which prevents entry of additional sperm. In addition, some eggs, including those of humans, do not finish their meiotic divisions until sperm entry stimulates completion. The specific location of sperm entry may also establish the embryonic head-to-tail (anterior-posterior) and back-to-belly (dorsal-ventral) axes by promoting the rearrangement of intracellular components.

Cleavage

Once the fertilization process is complete, the zygote undergoes a series of rapid cell divisions, known as cleavage, in which the divisions are not separated by any periods of cell growth. At the end of this process, the multicellular embryo, now called a blastula, will be approximately the same size as the zygote but now composed of many smaller cells called blastomeres. A small amount of growth and minor cell movements may occur at this stage, but the blastula is largely setting the stage for the major set of cell movements that will occur in the next developmental period.

The shape and organization of the blastula differ across vertebrate groups, ranging from spherical (amphibians and fishes, Figure 4.5A, B) to discoidal (a flattened, circular shape in birds, Figure 4.5E, F). The blastula often, but not always, has an interior cavity called the blastocoel. In eutherian mammals, the blastomeres form two distinct sets of cells—the outer layer of the trophectoderm, which will generate extraembryonic tissues like the placenta, and the internal cluster of the inner cell mass, from which the embryo proper will arise. Pluripotent embryonic stem cells can be derived from this inner cell mass. Overall, this multilayered embryo, referred to as a blastocyst (Figure 4.5C, D), will implant into the uterine wall and grow slowly into a cylindrical shape. The inner cell mass is itself divided into the hypoblast (primitive endoderm) and the epiblast, which form additional support structures, like the yolk sac, and the primary layers of the embryo, respectively.

Gastrulation

And so we finally arrive at that stage of promised importance, gastrulation, which involves extensive cell rearrangements that establish the three layers of all vertebrate embryos (for videos of these movements in Xenopus embryos, check out a cross-section view and this animation). These three layers, from outside to inside, are the ectoderm, mesoderm, and endoderm, and each will produce distinct tissue types in the mature animal. The three layers can be formed by different modes of cell movement, which largely depend on the type of blastula present at the start of gastrulation.

In amphibians and fish, gastrulation initiates at a small inward fold called the blastopore (Figure 4.6). The region located just dorsally of the blastopore is known as the Spemann-Mangold organizer, and transplanting this region into another embryo can trigger a second site of gastrulation and an entire second axis of development in the recipient embryo (Box 4.2—Ethel Browne, Hilde Mangold, and the Organizer Concept). Cells alongside the blastopore move inward in coherent sheets by the process of involution, spreading into the interior surface of the embryos. These cells will form the internal endoderm and mesoderm layers. The cells at the top of the blastula spread downward and around the outside of the embryo via epiboly (which is a bit like how caramel dripped onto the top of an apple might flow around the outside to cover it), forming the initial ectoderm tissue.

Gastrulation in birds and mammals starts at a structure called the primitive streak, which starts forming at the posterior pole and then extends anteriorly and will reflect the main body axis of the developing vertebrate. Cells converge at the primitive streak and then move inward at a condensation of cells, called Hensen’s node in birds, which has equivalent organizing properties to the Spemann organizer in amphibians. Instead of the coherent sheets of cells (epithelia) during involution, cells from the discoidal bird blastula move individually via ingression (Figure 4.7). The loose aggregates of cells (mesenchyme) that have moved internally will again form mesoderm and endoderm, whereas the cells that remained external to the primitive streak will contribute to the ectoderm. Mechanisms of epithelial-mesenchymal transition (EMT), critical here for proper development, can be disrupted in cancer cells, allowing them to detach from their originating tissue and move elsewhere in the body to form metastases. In the mammalian egg cylinder blastula, the primitive streak movements also involve convergent extension, wherein cells move toward and then past each other, displacing into adjacent layers and extending the length of the overall axis.

Regardless of the specific cell movements that produce the layers, the fates generated by the layers are remarkably similar (Figure 4.8). The endoderm forms the lining of the digestive tract, as well as structures evolutionarily derived from it, such as the lungs. The ectoderm forms the obviously externally facing structures, such as skin (epidermis), and also the nervous system, which administers the animal’s interface with the outside world. Sensory structures, like eyes, nose, ears, and lateral line organs, are derived from this layer, often via interactions between epidermal and neural ectoderm cells. Finally, mesoderm derivatives include the notochord, muscles, bone, heart, kidneys, and blood.

4.4 Neurulation

Once the three cell layers have been established, the work of building the distinct structures of the body can begin in earnest. Of critical importance because of its role in regulating the function of many other systems is the early development of the nervous system. Following directly from the dramatic tissue reorganization of gastrulation, the process of neurulation folds up portions of the ectoderm to generate the neural tube, which will form the developmental basis for the brain and spinal cord. Because the specifics of gastrulation differ between vertebrates, so too must the details of neurulation. Despite these differences, the outcome is a generalized nervous system with similar major structures.

The nervous system arises from the ectoderm via inductive interactions with the underlying mesoderm tissue. A crucial inductive signal comes from the notochord, which is the defining taxonomic characteristic of chordates, a condensed rod of tissue that runs along the anterior-posterior axis, providing both structural support and inductive instructions. The notochord signals to the overlying ectoderm cells, causing them to elongate and to form the thickened neural plate. The various changes in cell shape ultimately cause the edge of the neural plate to rise up, generating the neural folds with the lower neural groove in the region in between (Figure 4.9). The neural folds move toward each other before meeting at the dorsal midline of the embryo, where they fuse to form the hollow neural tube. The neural tube becomes separate from the ectoderm above it, which will become epidermal, and the cells of the neural crest will derive from the neural fold regions near the conjunction of the neural and nonneural ectoderm. This process occurs toward the anterior end of most vertebrate embryos and is referred to as primary neurulation. Toward the posterior of vertebrate embryos (and all along the axis in some, such as teleost fishes), secondary neurulation forms the neural tube from a solid rod of cells, similar to the notochord. Small openings form in this medullary cord, and the progressive fusion of these small cavities forms a hollow tube via the process of cavitation.

The proper formation of the neural tube is critical, and defects in the process can lead to severe defects in the embryo. Failure of primary neurulation at the anterior may result in anencephaly, where the brain and other head structures fail to develop entirely. Improper secondary neurulation at the posterior end of the neural tube may contribute to spina bifida, the severity of which depends on how much of the neural tube is defective. Collectively, neural tube defects are among the most common during human development, appearing in roughly 1 in 500 live births and additionally contributing to some significant proportion of spontaneous miscarriages.

4.5 Early Mesoderm Development and Somite Formation

Nearly simultaneously with the process of neurulation in the ectoderm, the vertebrate mesoderm begins to form primordia for the various mesodermal tissues and organs. Within the mesoderm, three main regions emerge on each side of the dorsal midline. The paraxial (“beside the axis”) mesoderm lies alongside the neural tube, whereas the lateral plate mesoderm forms a slightly flattened region farthest from the midline, with the intermediate mesoderm sandwiched between the two other regions (Figure 4.10).

As the lateral plate mesoderm splits into two layers, the definitive coelom, or main body cavity, will form and expand in the space between the two layers. The inner, or visceral, layer will support and surround the internal organs, such as the gut, heart, and lungs. The outer, or somatic, layer will contribute to the muscles of the body wall. From the internal layer, the splanchnic mesoderm forms the muscles surrounding the gut, as well as the arteries, veins, heart muscle, and blood cells of the circulatory system. The intermediate mesoderm generates the kidneys as well as other portions of the excretory and reproductive systems.

Although we don’t think of vertebrates as obviously segmented (unlike, say, arthropods or annelids), we can observe within the paraxial mesoderm a clear indication of vertebrate segmentation in the form of somites. Somites are paired blocks of tissue that arise alongside the notochord and will form the muscles and bones of the trunk (Figure 4.11).

Humans, mice, and birds all have ~50 somites, whereas snakes can have many hundreds. Much of the process of somite formation has been described in chick embryos because of the ease of observation and manipulation (including transplantations, dye injections, and gene expression labeling). Between Hensen’s node (the site of active gastrulation movements) and the most recently formed somite is the presomitic mesoderm, from which the next pair of somites will arise. New somite pairs are formed every 90 minutes within the chick embryo; the highly regular interval is determined by a complex interaction of signals within the somites and presomitic mesoderm described as the “clock-and-wavefront” model. Waves of expression of the transcription factor c-hairy1 and components of the Delta-Notch signaling pathway, including the amusingly named Lunatic Fringe protein, start at the tail end of presomitic mesoderm and move anteriorly until they reach the site of the next somite, where they persist through the visible formation of the condensing somite.

4.6 Nervous System Formation

The nervous system in adult vertebrates consists of both central and peripheral components, most of which are ultimately derived from the tissue divisions formed during neurulation. In general terms, the central nervous system arises from the neural tube, whereas the peripheral nervous system arises from the neural crest. In addition, various sensory systems, such as the nose and ear, will develop from neurogenic placodes, ectodermal thickenings lateral to the forming neural tube (Figure 4.12).

The anterior portion of the closed neural tube will develop into the brain, which is divided during development into the forebrain (prosencephalon), midbrain (mesencephalon), and hindbrain (rhombencephalon; Figure 4.13).

The forebrain and midbrain are unsegmented, but the hindbrain is divided into eight segmented regions (r1–r8) called rhombomeres (Figure 4.12). Cells within one rhombomere remain separate and distinct from cells of neighboring rhombomeres, and different cranial nerves arise in each rhombomere before extending into the distinct pharyngeal arches (also known as branchial arches or branchiomeres) from which the facial skeleton will arise. All developing vertebrates display branchial arches, and gill openings do form between pairs of branchial arches in embryonic fishes (both cartilaginous and teleost). These openings are the source of the inaccurate statement that humans have gills during development, but the gill openings either are transitory or never open at all in amniotes and never resemble anything close to functional gills.

The closed neural tube also gives rise to different populations of neurons across its dorsal-ventral axis. This dorsal-ventral pattern is established by complex, competing interactions between concentration gradients of the sonic hedgehog (Shh) signal released from the ventral notochord and bone morphogenetic protein (BMP) signals induced by epidermal ectoderm in the dorsal neural tube (Figure 4.14). The signals stimulate transcription factor responses in cells, based on the high-to-low concentration of each signal, resulting in different gene expressions in different cells based on where they are physically located within the neural tube. Cells closest to the notochord form the floor plate, which serves as a ventral organizing structure influencing neuronal patterning and migration (in part by its own production and release of Shh). Spinal motor neurons and interneurons develop from ventral regions near the floor plate, with different subtypes depending on local concentration of Shh. By contrast, commissural neurons, which send their axons across the neuronal midline in order to connect distant regions within the neural tube, arise dorsally.

The sensory neurons of the peripheral nervous system develop from neural crest cells outside of the neural tube but can make connections in the dorsal region of the dorsal neural tube. Neural crest derivatives also include pigment-forming cells, facial nerves, sensory and autonomic ganglia, and some head skeletal structures.

The proper function of the entire nervous system is dependent on the formation of appropriate connections between neurons in both the central and peripheral nervous systems. These specialized connections within the nervous system are called synapses, which occur where neurons come into close proximity to each other or target muscle cells but are still physically separated by a short distance, called the synaptic cleft (Figure 4.15). Chemical, rather than electrical, signals are passed across the synapse, and so proper patterning of neuronal connections is absolutely essential for transmission of neuronal messages throughout the body.

As neurons develop, they extend long, thin axons that will eventually form one-half of a synaptic connection. Axons lengthen toward their targets by the dynamic extension and retraction of cytoplasmic filaments from the growth cone region, often appearing to be searching their environment for the correct path. This search behavior occurs by the growth cone responding to chemically attractive or repulsive signals, including netrins, semaphorins, cadherins, and ephrins. The signals may be diffusible long-range molecules or contact-dependent short-range molecules, but in both cases they will interact with receptors in the growth cones to influence the extension and retraction movements.

The other major source of sensory structures is the neurogenic placodes, which are initially formed in the head region but may migrate to trunk locations (Figure 4.12). Examples include the nasal placode, which will form the odorant-responsive cells of the olfactory epithelium; the otic placode, which will form the fluid-filled ducts of the inner ear; and several lateral line placodes, from which the tactile and electroreceptors of the lateral line systems of fishes and amphibians are derived.

4.7 Hox Genes and Evolution

Some major questions we have been avoiding up to now are, How are the limbs correctly placed along the body, and how do the head and tail form at the proper ends of the animal? Narrowing further to the developing brain, how are the fore-, mid-, and hindbrain regions organized in the appropriate order? To a large extent, the answer to these questions of regional specification involves the action of the Hox gene family.

Hox genes were originally identified in the fruit fly but have since been identified in virtually all animals and share functions in anterior-posterior specification in all bilaterally symmetrical animals, including vertebrates (Figure 4.16). The Hox genes all encode transcription factors that can switch other genes on or off, thereby regulating cell fates. Curiously, the Hox genes form clusters of closely spaced genes, and a gene’s position within the cluster parallels its expression along the anterior-posterior axis. Although fruit flies possess only a single Hox cluster with eight genes, vertebrates have at least three clusters (in jawless cyclostome fishes) up to a maximum of eight clusters (in teleost fishes), with all tetrapods possessing four clusters. The cluster duplications in teleosts have been hypothesized as a trigger facilitating the adaptive radiation of morphological diversity within this largest vertebrate group. The clusters in vertebrates have both overlapping and separable functions, with major roles in the development of the mesodermally derived axial skeleton and ectodermal derivatives in the nervous system.

One of the most visible aspects of Hox function in the nervous system is in the patterning of the r1–r8 rhombomeres (Figure 4.17). Expression boundaries for genes of the Hoxa and Hoxb clusters correspond to the boundaries between rhombomeres, and experimental manipulation of Hox expression can change the fate of the cranial nerves emanating from the altered rhombomeres. In the mesoderm, misexpression or loss of Hox6 and Hox10 paralogs can transform cervical and lumbar vertebrae into thoracic-like vertebrae with ectopic rib structures.

Despite the largely similar complements of Hox genes across vertebrates, changes in precisely where, when, and in what combinations they are expressed may account for some of the observable variation in anterior-posterior morphology. These differences in expression pattern and resulting morphology are a good example of developmental modularity. For example, the pectoral girdle where bird forelimbs articulate always forms at the boundary between cervical and thoracic vertebrae, but different birds have different numbers of cervical vertebrae and therefore different forelimb positions. Zebrafinch, chicken, and ostrich embryos develop limbs at different axial positions. Zebrafinches, which have the fewest cervical vertebrae, express Hoxb4 in the smallest domain and for the shortest duration, whereas ostrich embryos (with the most cervical vertebrae) express Hoxb4 for longer and in a larger region than either zebrafinches or chickens. Furthermore, changing of Hox expression on just one side of the body in chicken embryos can shift the limb position with respect to the unmanipulated side, suggesting a causal role for Hox expression in limb positioning. Another example comes in the secondary loss of limbs in snakes. Snakes lack forelimbs entirely, though some groups of snakes develop rudimentary hindlimb buds during development, which in some cases are retained as vestigial adult structures. Hox expression associated with thoracic development in other groups is expanded in python embryos, suggesting that the snake axial skeleton takes on a riblike thoracic identity, eliminating the domains from which limbs would arise, preventing forelimb development entirely and minimizing the hindlimb development that can occur.

4.8 Limb Development

The diversity of functions and morphologies for limb structures within the vertebrates makes limb development a fascinating system for the study of pattern formation, development, and evolution. We have already seen how changes in Hox expression are correlated with evolutionary changes in limb position along the body in tetrapods. We will consider in this section major mechanisms for patterning within the limbs and look at an example of how variations in limb developmental mechanisms contribute to evolutionary variation.

Most vertebrate limbs have differences along three discrete developmental axes: proximal-distal (from closest to furthest from the trunk), anterior-posterior (digit 1 [e.g., thumb] to digit 5 [e.g., little finger]), and dorsal-ventral (from top to bottom [e.g., paw pads on a dog or cat]). The specific mechanisms for each axis are distinct, but all follow the familiar model of inductive signaling interactions and changes in gene expression that produce different outcomes in other body regions. And although we will focus on one axis at a time, they are highly integrated in the ways they interact with one another to coordinate the final shape of the vertebrate limb.

Proximal-Distal

The close-to-far differentiation of limbs is easily observed in tetrapods, which have a general skeletal pattern of “one bone, two bones, many bones” (corresponding to the zeugopod [humerus], stylopod [radius/ulna], and autopod [wrist/digits] regions) that itself is presaged in limbs of transitional Devonian-era tetrapods like Tiktaalik (Figure 4.18).

Initial development of the limb bud (Figure 4.19) occurs as an outgrowth of the somatic mesoderm, covered by a thin layer of ectoderm. A raised portion of the ectoderm, referred to as the apical ectodermal ridge (AER), is essential for continued outgrowth of the limb bud. Removal of the AER early results in truncation of the entire limb, but removal at progressively later stages allows formation of progressively more distal limb structures. Replacement of the AER with a bead releasing the growth factor FGF (fibroblast growth factor) can restore limb development, indicating a key role for this signaling molecule in the process of proximal-distal patterning.

One model, called the “progress zone model,” for how this FGF signal might be translated into distinct limb elements suggested that the length of time a cell resided in the population of rapidly dividing mesoderm near the AER (and therefore, responding to the FGF signal) would establish its proximal-distal (P-D) fate. However, more recent evidence suggests that a combination of a proximal retinoic acid signal and the distal FGF signal from the AER may establish P-D fate early, followed by outgrowth of the limb bud tissue to form the different P-D regions.

A reciprocal signaling interaction from the mesoderm to the ectoderm is also critical, as removing the mesodermal progress zone stops limb development, and transplanting the leg progress zone into a developing wing bud in chickens will cause the bud to produce leg structures.

Anterior-Posterior

The proliferation of acronyms continues as we consider the anterior-posterior (A-P) patterning of the limbs. A small region of mesodermal tissue at the posterior margin of the limb bud, called the zone of polarizing activity (ZPA), releases the sonic hedgehog (Shh) signal (Figure 4.19). Either transplanting a second ZPA into the anterior of the limb bud or adding Shh can cause a mirror-image duplication of the digits. Genetic mutations associated with polydactyly in cats, mice, and human families cause Shh to be switched on inappropriately at the anterior margin of the limb bud, replicating the digit duplications observed by experimental manipulation.

Positive feedback between the AER and ZPA is critical to maintain proper growth of the limb. The BMP proteins we saw earlier during nervous system patterning also act here, signaling to inhibit FGF in the AER. However, Shh from the ZPA blocks the action of the BMPs, allowing continued FGF expression; that FGF, in turn, enhances expression of Shh. This type of feedback loop helps coordinate proper formation of the limb. In fact, in both pythons and dolphins, hindlimb buds begin to develop, but one or more aspects of the AER-ZPA signaling loop appear to be defective, leading to adult animals that completely lack hindlimbs.

Dorsal-Ventral

The AER forms at the border between dorsal and ventral ectoderm in the developing limb bud, so the establishment of the dorsal and ventral compartments is also critical for proximal-distal patterning and outgrowth of the limb. The ventral ectoderm expresses a gene called en1, which is induced by signaling from the neighboring lateral somatopleuric mesoderm. This gene is not activated in the dorsal ectoderm, and the populations of expressing and nonexpressing cells are present and separate even before the limb bud itself is visible. Beyond positioning the AER, en1 also acts in later portions of dorsal-ventral patterning by regulating the expression of later target genes that are directly involved in aspects of limb determination.

Evolutionary Modification of Limbs

Vertebrate limbs are highly variable—the range of forelimb structures, for example, includes the “arm wings” of birds and the “finger wings” of bats, the paddle-like flippers of whales and dolphins, and the shovel-like paws of moles, as well as our own finely articulated grasping hands. Given the shared similarity in the underlying bones and the underlying developmental axes, the vertebrate limb provides an intriguing model for examining how changes in development can generate differences in final limb structures.

Some of the most stunning limb modifications in mammals are the dramatic elongation of the fingers and expansion of interdigital membranes that allow powered flight in Chiroptera (bats). Comparing development in bats and mice has been fruitful in identifying processes that vary in the two groups and likely contribute to the formation of the specialized bat wing. Importantly, limb buds in both groups initiate their development virtually identically, and the gene sequences of the limb development factors in both groups are either nearly or absolutely identical. These observations further support a model in which relatively subtle changes during development, often in when, where, and at what level a gene is active, are capable of producing substantial differences in final outcomes.

For example, levels of the BMPs, which we have seen in multiple contexts, are higher in developing bat limbs than developing mouse limbs (Figure 4.20). Growing limbs from either animal in artificial culture media with added BMP also causes greater limb length, adding a causative effect to the correlation observed for expression. Similarly, the gene Prx1 has a higher expression level in bat limbs than mouse limbs. The DNA regulatory sequences that control this gene expression have differences between bats and mice, and substituting the bat sequence into the mouse genome causes those mice to develop longer limbs, suggesting that the DNA differences may have contributed to bat forelimb elongation. Beyond the elongated bones of the fingers, another notable modification in bats is the retention of the membranous webbing between the digits. Our good friend BMP makes another appearance here, though in this instance, BMP causes the interdigital cells to undergo apoptotic cell death, and it is the novel presence of the Gremlin BMP-inhibitor protein in bats that allows the webbing to survive.

4.9 Summary

Vertebrate development is a highly dynamic process involving dramatic cell movements and careful coordination between regions via signaling pathways that are used repeatedly across different contexts, ultimately resulting in cells, tissues, and organs with very different physical properties from one another. This chapter has emphasized the shared developmental processes that generate the different basic regions and structures of vertebrates, and we will see throughout the remainder of the book how those physical properties affect overall structure and function of vertebrate anatomical systems.

Box 4.1—Cell Death in the Feet of Waterfowl

One common adaptation in waterbirds is the presence of partial or full webbing in the foot to aid in water-based locomotion, as we might commonly envision in ducks. But birds not adapted to water do not typically have this feature, and the lack of webbing appears to be the ancestral character state for birds. How did the webbing get retained in particular lineages of birds, and how do we think of this difference in terms of cell fates? In other vertebrates without webbing, interdigital cells undergo programmed cell death, and so well-separated fingers and toes are partially sculpted not just by growth and bone condensation but also by the self-destruction of the cells in between digits. In birds with webbed feet, the fate of those cells is changed—they must survive instead of dying. Across birds with various levels of webbing, we see one common feature that may help cells opt for this new fate of surviving instead of dying. During development, cells in the feet of birds with webbing express a gene called Gremlin1 that is not switched on in the feet of birds without webbing, such as quail and chickens. The activation of this new factor appears to be enough for the cell to choose to live instead of to self-destruct.

Box 4.2—Ethel Browne, Hilde Mangold, and the Organizer Concept

The history of the organizer concept in developmental biology demonstrates how circumstances and personalities can puncture the myth of scientific objectivity that often accompanies tales of scientific discovery. As we will see, the conventional narrative of the organizer importantly leaves out critical individuals and experiments.

In the standard narrative, Hans Spemann, a prominent developmental biologist at the University of Freiberg in Germany, was awarded a Nobel Prize in 1935 for the discovery of the organizer, the tiny chunk of tissue (subsequently named after him) that promotes formation of a duplicate axis when transplanted from one embryo into another. In 1918, Spemann had published experiments describing transplantations from early and late gastrula stages, demonstrating that early cells were not yet determined in their fates but that during gastrulation cells did commit to particular fates. These experiments set the stage for the organizer experiments, but Spemann himself actually did not perform those experiments for which the Nobel Prize was awarded.

Instead, his graduate student Hilde Mangold used transplantation techniques originally developed by Spemann (including using a tiny glass needle to carefully carve cells from a developing embryo and a loop of human hair to gently move the embryos and transplants from dish to dish) to transfer the dorsal blastopore lip region between pigmented and unpigmented newts to be able to distinguish native from transplanted tissue. Over two years of experiments, she completed over 200 transplantations, but only results from six embryos were ever published to demonstrate the stunning outcome of a newt embryo with two distinct and properly patterned axes, involving both transplanted and host tissue (Figure 4.22).

The first successful embryo was added as a postscript to a 1921 publication of Spemann that also introduced and defined the “organizer” terminology. The full set of experiments was submitted in 1923 and published in 1924, with Spemann’s name listed first. Viktor Hamburger, a graduate student who joined Spemann’s laboratory at the same time as Mangold and who himself later became one of the most remarkable developmental biologists of the 20th century, described in his memoirs Mangold’s reaction to that arrangement: “[Mangold] was not happy that Spemann had added his name to her thesis publication, while Holtfreter and I and all the rest of us saw ourselves proudly in print as sole authors. Moreover, Spemann had insisted on having his name precede hers! But Spemann was perfectly right in claiming precedence, while she apparently did not fully realize the significance of her results.” The basis on which Hamburger makes these claims is not obvious, and clearly Spemann had not insisted on primary authorship (or any form of coauthorship) for the work of any of his other graduate students, who were all men. Tragically, the question of whether Mangold should have shared in the Nobel Prize never arose because Mangold died from severe burns suffered from an oven fire in 1924, and the Nobel committee does not award posthumous prizes.

But the trail of omissions in the organizer story appears to have yet another link to consider. Before Mangold began the newt experiments, Spemann had asked her to replicate the 18th-century experiments of French naturalist Arthur Tremblay on inverting the Hydra freshwater cnidarian to investigate the interchangeability of inner and outer layers of cells. Ultimately, neither Mangold nor Spemann could repeat this technically challenging experiment (the Hydra would inevitably uncurl itself, undoing the inversion), so Spemann eventually let Mangold shift to her work on the newts. But if all the other students in Spemann’s lab were focused on newts, why might he assign a student to instead pursue work in a completely different (and nonvertebrate!) system?

An intriguing possibility is the potential influence of a 1909 paper by Ethel Browne on induction in Hydra. Browne transplanted unpigmented hypostome tissue (from the mouth region of the Hydra polyp) to the flank of a pigmented host individual. She then observed that the unpigmented tissue induced the host tissue to form a completely new secondary axis. The similarities between Browne’s and Mangold’s experiments are striking: the use of pigmented and unpigmented tissues; the formation of a new axis following transplantation; and the reorganization of host tissue, not just growth of the transplant, to generate the new axis.

Browne did not use the term organizer in her publication, but her experiments clearly describe a phenomenon closely resembling the organizer from the Spemann and Mangold paper. Curiously, that paper does not cite Browne’s work (nor do any other of Spemann’s published papers). Did Spemann perhaps not know Browne’s work with Hydra? Notably, developmental biologist Howard Lenhoff described in 1991 a reprint of Browne’s work in a collection of Spemann’s papers—a reprint with only the single phrase “induced the formation of” underlined and an exclamation point in the margin. While it cannot be definitively established that it was Spemann who read and marked the paper, its presence in Spemann’s papers coupled with his directive that Mangold start her thesis work with Hydra suggests he likely had some familiarity with the topic.

Instead of a narrative throughline from Ethel Browne to Hilde Mangold to the organizer concept of Hans Spemann’s Nobel Prize, we instead typically hear a much-truncated version that emphasizes solitary discovery and obscures the experimental and conceptual complexity contributed by two often overlooked women researchers. When considering the history of scientific knowledge, it’s critical that we endeavor to look beyond narrow, simplified narratives to better appreciate the full spectrum of contributors, who may have deliberately or inadvertently been left out of those narratives.

4.10 References

1. Cohn, Martin, and Cheryll Tickle. “Developmental basis of limblessness and axial patterning in snakes.” Nature 399 (1999): 474–9. https://doi.org/10.1038/20944.
2. Cooper, Lisa Noelle, and Karen E. Sears. “How to grow a bat wing.” In Bat Evolution, Ecology, and Conservation, edited by Adams, Rick A. and Scott C. Pedersen, 3–20. New York, NY: Springer New York, 2013.
3. Crow, Karen D., Peter F. Stadler, Vincent J. Lynch, Chris Amemiya, and Günter P. Wagner. “The ‘fish-specific’ Hox cluster duplication is coincident with the origin of Teleosts.” Molecular Biology and Evolution 23, no. 1 (2005): 121–136. https://doi.org/10.1093/molbev/msj020.
4. Fassler, Peter E., and Klaus Sander. “Hilde Mangold (1898–1924) and Spemann’s Organizer: Achievement and tragedy.” Roux’s Archives of Developmental Biology 205, no. 7–8 (May 1, 1996): 323–332. https://doi.org/10.1007/BF00377212.
5. Hamburger, Viktor. “Hilde Mangold, co-discoverer of the Organizer.” Journal of the History of Biology 17, no. 1 (1984): 1–11. http://www.jstor.org/stable/4330877.
6. Hopwood, Nick. “‘Not birth, marriage or death, but gastrulation’: The life of a quotation in biology.” British Journal for the History of Science 55, no. 1 (2022): 1–26. https://doi.org/10.1017/S0007087421000790.
7. Jaeger, Johannes, and Nick Monk. “Dynamical modules in metabolism, cell and developmental biology.” Interface Focus 11, no. 3 (2021): 20210011. https://doi.org/10.1098/rsfs.2021.0011.
8. Kahrl, Ariel F., Rhonda R. Snook, and John L. Fitzpatrick. “Fertilization mode differentially impacts the evolution of vertebrate sperm components.” Nature Communications 13, no. 1 (2022): 6809. https://doi.org/10.1038/s41467-022-34609-7.
9. Lenhoff, Howard M. “Ethel Browne, Hans Spemann, and the discovery of the Organizer phenomenon.” Biological Bulletin 181, no. 1 (1991): 72–80. https://doi.org/10.2307/1542490.
10. “Lewis Wolpert Obituary.” The Guardian, January 29, 2021. https://www.theguardian.com/science/2021/jan/29/lewis-wolpert-obituary.
11. Miao, Yuchuan, and Olivier Pourquié. “Cellular and molecular control of vertebrate somitogenesis.” Nature Reviews Molecular Cell Biology 25, no. 7 (February 28, 2024). https://doi.org/10.1038/s41580-024-00709-z.
12. Moreau, Chloe, Paolo Caldarelli, Didier Rocancourt, Julian Roussel, Nicolas Denans, Olivier Pourquie, and Jerome Gros. “Timed collinear activation of Hox genes during gastrulation controls the avian forelimb position.” Current Biology 29, no. 1 (January 7, 2019): 35–50.e4. https://doi.org/10.1016/j.cub.2018.11.009.
13. Parker, Hugo J., Marianne E. Bronner, and Robb Krumlauf. The Vertebrate Hox Gene Regulatory Network for Hindbrain Segmentation: Evolution and Diversification. Vol. 38. Wiley, 2016. https://doi.org/10.1002/bies.201600010.
14. Prakash, Seppan, Elumalai Prithiviraj, Sekar Suresh, Nagella Venkata Lakshmi, Mohanraj Karthik Ganesh, Murugesan Anuradha, Lakshmanan Ganesh, and Premavathy Dinesh. “Morphological diversity of sperm: A mini review.” Iranian Journal of Reproductive Medicine 12, no. 4 (April 1, 2014): 239–242.
15. Serrano Nájera, Guillermo, and Cornelis J. Weijer. “The evolution of gastrulation morphologies.” Development (Cambridge, England) 150, no. 7 (April 1, 2023): dev200885. https://doi.org/10.1242/dev.200885.
16. Tokita, Masayoshi, Hiroya Matsushita, and Yuya Asakura. “Developmental mechanisms underlying webbed foot morphological diversity in waterbirds.” Scientific Reports 10, no. 1 (2020): 8028. https://doi.org/10.1038/s41598-020-64786-8.
17. Verheyden, Jamie M., and Xin Sun. “Embryology meets molecular biology: Deciphering the apical ectodermal ridge.” Developmental Biology 429, no. 2 (2017): 387–390. https://doi.org/10.1016/j.ydbio.2017.01.017.