Immigrant Cells in the Epidermis

Despite its homogeneous histological appearance, the epidermis is really a cellular mosaic, with contributions from cells derived not only from surface ectoderm, but also from other precursors, such as the neural crest or mesoderm. These cells play important specific roles in the function of the skin.

Early in the second month, melanoblasts derived from the neural crest migrate into the embryonic dermis; slightly later, they migrate into the epidermis. Although melanoblasts can be recognized early by staining with a monoclonal antibody (HMB-45, which reacts with a cytoplasmic antigen common to melanoblasts and melanomas [pigment cell tumors]), these cells do not begin to produce recognizable amounts of pigment until midpregnancy. This production occurs earlier in heavily pigmented individuals than in individuals with light complexions. The differentiation of melanoblasts into mature melanocytes involves the formation of pigment granules called melanosomes from premelanosomes.

The number of pigment cells in the skin does not differ greatly among the various races, but the melanocytes of dark-skinned individuals contain more pigment granules per cell. Albinism is a genetic trait characterized by the lack of pigmentation, but albinos typically contain normal numbers of melanocytes in their skin. The melanocytes of albinos are generally unable to express pigmentation because they lack the enzyme tyrosinase, which is involved in the conversion of the amino acid tyrosine to melanin.

Late in the first trimester, the epidermis is invaded by Langerhans’ cells, which arise from precursors in the bone marrow. These cells are peripheral components of the immune system and are involved in the presentation of antigens; they cooperate with T lymphocytes (white blood cells involved in cellular immune responses) in the skin to initiate cell-mediated responses against foreign antigens. Langerhans’ cells are present in low numbers (about 65 cells/mm² of epidermis) during the first two trimesters of pregnancy, but subsequently their numbers increase several-fold to 2% to 6% of the total number of epidermal cells in the adult.

A third cell type in the epidermis, the Merkel cell, appears in palmar and plantar epidermis at 8 to 12 weeks of gestation and is associated with free nerve terminals. Derived from the neural crest, these cells function as slow-adapting mechanoreceptors in the skin, but cytochemical evidence suggests that they may also function as neuroendocrine cells at some stage.

Epidermal Differentiation

Progression from a single-layered ectoderm to a stratified epithelium requires the activation of the transcription factor p63, possibly in response to signals from the underlying dermal mesenchyme. Subsequently, through the action of a microRNA (miR-203), p63 must be turned off for cells within the stratified epidermis to embark on their terminal differentiation program, which involves their leaving the cell cycle.

When the multilayered epidermis becomes established, a regular cellular organization and sequence of differentiation appear within it (Fig. 9.3). Stem cells* of the basal layer (stratum basale) divide and contribute daughter cells to the next layer, the stratum spinosum. The movement of epidermal cells away from the basal layer is preceded by a loss of adhesiveness to basal lamina components (e.g., fibronectin, laminin, and collagen types I and IV). These cellular properties can be explained by the loss of several integrins, which attach the basal cells to the underlying basal lamina. Cells of the stratum spinosum produce prominent bundles of keratin filaments, which converge on the patchlike desmosomes binding the cells to each other.

Keratohyalin granules, another marker of epidermal differentiation, begin to appear in the cytoplasm of the outer, postmitotic cells of the stratum spinosum and are prominent components of the stratum granulosum. Keratohyalin granules are composed of two types of protein aggregates—one histidine rich and one sulfur rich—closely associated with bundles of keratin filaments. Because of their high content of keratin, epidermal cells are given the generic name keratinocytes. As the keratinocytes move into the stratum granulosum, their nuclei begin to show characteristic signs of terminal differentiation, such as a flattened appearance, dense masses of nuclear chromatin, and early signs of breaking up of the nuclear membrane. In these cells, the bundles of keratin become more prominent, and the molecular weights of keratins that are synthesized are higher than in less mature keratinocytes.

As the cells move into the outer layer, the stratum corneum, they lose their nuclei and resemble flattened bags densely packed with keratin filaments. The cells of this layer are interconnected by the histidine-rich protein filaggrin, which is derived from one of the granular components of keratohyalin. Depending on the region of the body surface, the cells of the stratum corneum accumulate to form approximately 15 to 20 layers of dead cells. In postnatal life, whether through friction or the degradation of the desmosomes and filaggrin, these cells are eventually shed (e.g., about 1300 cells/cm²/hour in the human forearm) and commonly accumulate as house dust.

Biochemical work has correlated the expression of keratin proteins with specific stages of epidermal differentiation. Keratins K5, K14, and K15 are expressed in the basal layer, but as the stratum spinosum develops, the cells in that layer express K1 and K10. Loriccrin and filaggrin, an intracellular binding protein, appear as the stratum granulosum and the stratum corneum develop in the early fetus.

The proliferation of basal epidermal cells is under the control of a variety of growth factors. Some of these stimulate and others inhibit mitosis. Postnatally, keratinocytes commonly spend about 4 weeks in their passage from the basal layer of the epidermis to ultimate desquamation, but in some skin diseases, such as psoriasis, epidermal cell proliferation is poorly controlled, and keratinocytes may be shed within 1 week after their generation.

A prominent feature of the skin, particularly the thick skin of the palms and soles, is the presence of epidermal ridges and creases. On the tips of the digits, the ridges form loops and whorls in fingerprint patterns that are unique to the individual. These patterns form the basis for the science of dermatoglyphics, in which the patterns constitute the foundation for genetic analysis or criminal investigation.

The formation of epidermal ridges is closely associated with the earlier appearance of volar pads on the ventral surfaces of the fingers and toes (Fig. 9.4). Volar pads first form on the palms at about weeks, and by weeks, they have formed on the fingers. The volar pads begin to regress by about weeks, but while they are present, they set the stage for the formation of the epidermal ridges, which occurs between 11 and 17 weeks. Similar events in the foot occur approximately 1 week later than those in the hand.

The pattern of the epidermal ridges is correlated with the morphology of the volar pads when the ridges first form. If a volar pad is high and round, the epidermal ridges form a whorl; if the pad is low, an arch results. A pad of intermediate height results in a loop configuration of the digital epidermal ridges. The timing of ridge formation also seems to influence the morphology: Early formation of ridges is associated with whorls, and late formation is associated with arches. The primary basis for dermatoglyphic patterns is still not understood.

When the epidermal ridges first form, the tips of the digits are still smooth, and the fetal epidermis is covered with peridermal cells. Beneath the smooth surface, however, epidermal and dermal ridges begin to take shape (Fig. 9.5). Late in the fifth month of pregnancy, the epidermal ridges become recognizable features of the surface landscape.

Hair

Hairs are specialized epidermal derivatives that arise as the result of inductive stimuli from the dermis. There are many types of hairs, ranging from the coarse hairs of the eyelashes and eyebrows to the barely visible hairs on the abdomen and back. Regional differences in morphology and patterns of distribution are imposed on the epidermis by the underlying dermis.

Hair formation is first recognizable at about the twelfth week of pregnancy as regularly spaced epidermal placodes associated with small condensations of dermal cells called dermal papillae (Fig. 9.7). Under the continuing influence of a dermal papilla, the placode forms an epidermal downgrowth (hair germ), which over the next few weeks forms an early hair peg. In succeeding weeks, the epidermal peg overgrows the dermal papilla, and this process results in the shaping of an early hair follicle. At this stage, the hair follicle still does not protrude beyond the outer surface of the epidermis, but in the portion of the follicle that penetrates deeply into the dermis, two bulges presage the formation of sebaceous glands, which secrete an oily skin lubricant (sebum), and are the attachment site for the tiny arrector pili muscle. The arrector pili is a mesodermally derived smooth muscle that lifts the hair to a nearly vertical position in a cold environment. In many animals, this increases the insulation properties of the hair. The developing hair follicle induces the adjacent dermal mesoderm to form the smooth muscle cells of this muscle. As the developing hair matures, a small bulge below the sebaceous gland marks an aggregation of epidermal stem cells (Fig. 9.7D).

The formation of a hair involves a series of inductive interactions mediated by signals that are only partly understood. When a dense condensation of dermal cells has formed beneath the ectoderm (Fig. 9.8A), the first of two dermal inductions results in the thickening of ectoderm in very regularly arranged locations to form epidermal placodes (Fig. 9.8B). Fibroblast growth factor (FGF) and Wnt (mainly Wnt-11) signaling from the dermis, along with the inactivation of local bone morphogenetic proteins (BMPs), stimulates the activation of other Wnts in the ectoderm to form an epidermal placode. The response of the ectoderm is to produce other Wnts, acting through β-catenin intermediates, and Edar, the receptor for the signaling molecule ectodysplasin. In the areas where hairs will not develop (interfollicular areas), placode formation is inhibited by locally produced BMPs and by the inhibition of Wnts by Dickkopf. How the epidermal placodes are spaced in such a geometrically regular fashion is still not well understood.

The newly formed epidermal placodes become the inducing agent and stimulate the aggregation of mesenchymal cells beneath the placode to form the dermal papilla (Fig. 9.8C). Sonic hedgehog, produced by the epidermal placode, seems to be involved in this induction, but the identity of other signals is unknown. Next, the dermal papilla initiates the second dermal induction by stimulating downgrowth of the cells of the epidermal placode into the dermis (Fig. 9.8D). Epidermal downgrowth, which involves considerable epidermal cellular proliferation, is stimulated by expression of sonic hedgehog by the epidermal cells and the subsequent expression of cyclin D1, part of the cell cycling pathway. Later formation of a hair is structurally and biochemically an extremely complex process, which, among other things, involves the expression of a range of Hox genes in specific locations and at specific times along the length of each developing hair.

Once formed, an individual hair follows a regular cycle of growth and shedding (see Fig. 9.7). During anagen, the first phase in the cycle, the hair is actively growing (around 10 cm per year). This phase can last up to 5 to 6 years. Then it enters catagen, a phase lasting 1 or 2 weeks, during which the hair follicle regresses to only a fraction of its original length. The hair stops growing in the resting phase (telogen), which lasts 5 to 6 weeks, after which time the hair is shed (exogen). Adjacent hairs are frequently in different phases of the hair cycle.

Erupted hairs are first seen on the eyebrows shortly after 16 weeks. Within a couple of weeks, hairs cover the scalp. The eruption of hairs follows a cephalocaudal gradient over the body. During the later stages of hair formation, the hair bulb becomes infiltrated with melanocytes, which provide color to the hair. Starting around the fifth month, the epidermal cells of the hair shaft begin to undergo keratinization, forming firm granules of trichohyalin, which imparts hardness to the hair.

Products of the fetal sebaceous glands accumulate on the surface of the skin as vernix caseosa. This substance may serve as a protective coating for the epidermis, which is continually exposed to amniotic fluid.

The first fetal hairs are very fine in texture and are close together. Known as lanugo, they are most prominent during the seventh and eighth months. Lanugo hairs are typically shed just before birth and are replaced by coarser definitive hairs, which arise from newly formed follicles.

The pattern of epidermal appendages such as hairs has been shown experimentally to relate to patterns generated in the dermis. Other studies have compared patterns of scalp hairs between normal embryos and embryos with cranial malformations (Fig. 9.9) and have shown a correlation between whorls and the direction of hair growth and the tension on the epidermis at the time of formation of the hair follicles.

Nails

Toward the end of the third month, epidermal thickenings (primary nail field) on the dorsal surfaces of the digits mark the beginnings of nail development. Cells from the primary nail field expand proximally to undercut the adjacent epidermis (Fig. 9.10). Proliferation of cells in the proximal part of the nail field results in the formation of a proximal matrix, which gives rise to the nail plate that grows distally to cover the nail bed. The nail plate itself consists of highly keratinized epidermal cells. A thin epidermal layer, the eponychium, initially covers the entire nail plate, but it eventually degenerates, except for a thin persisting rim along the proximal end of the nail. The thickened epidermis underlying the distalmost part of the nail is called the hyponychium, and it marks the border between dorsal and ventral skin. Outgrowing fingernails reach the ends of the digits by about 32 weeks, whereas in toenails, this does not occur until 36 weeks.

Mammary Glands

As with many glandular structures, the mammary glands arise as epithelial (in this case, ectodermal) downgrowths into mesenchyme in response to inductive influences by the mesenchyme. The first morphological evidence of mammary gland development is the appearance of two bands of ectodermal thickenings called milk lines (part of the wolffian ridge [see p. 111]) running along the ventrolateral body walls in embryos of both genders at about 6 weeks (Fig. 9.11A). They are marked by the expression of various Wnts within the ectodermal cells. The thickened ectoderm of the milk lines undergoes fragmentation, and remaining areas form the primordia of the mammary glands. The craniocaudal level and the extent along the milk lines at which mammary tissue develops vary among species. Comparing the location of mammary tissue in cows (caudal), humans (in the pectoral region), and dogs (along the length of the milk line) shows the wide variation in location and number of mammary glands. In humans, supernumerary mammary tissue or nipples can be found anywhere along the length of the original milk lines (Fig. 9.11B). Individual mammary placodes form from aggregation and proliferation of ectodermal cells of the milk line under the inductive influence of the signaling molecule neuregulin-3. Their dorsoventral location is marked by the expression of the transcription factor Tbx-3.

Mammary ductal epithelial downgrowths (Fig. 9.12) are associated with two types of mesoderm: fibroblastic and fatty. The early epithelial downgrowth secretes parathyroid hormone–related hormone, which increases the sensitivity of the underlying mesenchymal cells to BMP-4. BMP-4 signals within the underlying mesenchyme have two principal effects (see Fig. 9.12B). First, they stimulate further downgrowth of the mammary epithelial bud. Second, they stimulate the expression of the transcription factor Msx-2, which inhibits the formation of hair follicles in the region of the nipple. Experimental evidence suggests that inductive interactions with the fatty component of the connective tissue are responsible for the characteristic shaping of the mammary duct system. As with many developing glandular structures, the inductive message seems to be mediated to a great extent by the extracellular matrix of the connective tissue.

Although the mesoderm controls the branching pattern of the ductal epithelium, the functional properties of the mammary ducts are intrinsic to the epithelial component. An experiment in which mouse mammary ectoderm was combined with salivary gland mesenchyme illustrates this point. The mammary ducts developed a branching pattern characteristic of salivary gland epithelium, but despite this, the mammary duct cells produced one of the milk proteins, α-lactalbumin.

In keeping with their role as secondary sexual characteristics, mammary glands are extremely responsive to the hormonal environment. This has been shown by experiments conducted on mice. In contrast to the continued downgrowth of ductal epithelium in female mice, the mammary ducts in male mice respond to the presence of testosterone by undergoing a rapid involution. Female mammary ducts react similarly if they are exposed to testosterone. Further analysis has shown that the effect of testosterone is mediated through the mammary mesenchyme, rather than acting directly on the ductal epithelium. Conversely, if male mammary ducts are allowed to develop in the absence of testosterone, they assume a female morphology. Mammary duct development in male humans does not differ from that in female humans until puberty. Even then, a rudimentary duct system remains, which is why men can develop gynecomastia or breast cancer in later life.

The role of the mesoderm and testosterone receptors is well illustrated in experiments involving mice with a genetic mutant, androgen insensitivity syndrome. This is the counterpart of a human condition called the testicular feminization syndrome, in which genetic male individuals lack testosterone receptors. Despite having high circulating levels of testosterone, these individuals develop female phenotypes, including typical female breast development (Fig. 9.13A), because without receptors, the tissue cannot respond to the testosterone.

In vitro recombination experiments on mice with androgen insensitivity have been instrumental in understanding the role of the mesoderm in mediating the effects of testosterone on mammary duct development (Fig. 9.13B). If mutant mammary ectoderm is combined with normal mesoderm in the presence of testosterone, the mammary ducts regress, but normal ectoderm combined with mutant mesoderm continues to form normal mammary ducts despite being exposed to high levels of testosterone. This shows that the genetic defect in testicular feminization is expressed in the mesoderm.

The postnatal development of female mammary gland tissue is also highly responsive to its hormonal environment. The simple mammary duct system that was laid down in the embryo remains in an infantile condition until it is exposed to the changing hormonal environment at the onset of puberty (Fig. 9.14A). Increasing levels of circulating estrogens, acting on a base of growth hormone and insulinlike growth factor activity, stimulate the proliferation of the mammary ducts and enlargement of the pad of fatty tissue that underlie it (Fig. 9.14B). As is the case with testosterone effects, estrogen effects on the epithelium of the mammary ducts are mediated via paracrine influences from the mammary connective tissue stroma, which contains the estrogen receptors.

The next major change in the complete cycle of mammary tissue development occurs during pregnancy, although minor cyclical changes in mammary tissue are detectable in each menstrual cycle. During pregnancy, increased amounts of progesterone, along with prolactin and placental lactogen, stimulate the development of secretory alveoli at the ends of the branched ducts (Fig. 9.14C). With continuing development of the alveoli, the epithelial cells build up increased numbers of the cytoplasmic organelles, such as rough endoplasmic reticulum and the Golgi apparatus, which are involved in protein synthesis and secretion.

Lactation involves numerous reciprocal influences between the mammary glands and the brain; these are summarized in Figure 9.14D. Stimulated by prolactin secretion from the anterior pituitary, the alveolar cells synthesize milk proteins (casein and α-lactalbumin) and lipids. In a rapid response to the suckling stimulus, the ejection of milk is triggered by the release of oxytocin by the posterior portion of the pituitary. Oxytocin causes the contraction of myoepithelial cells, which surround the alveoli. Suckling also causes an inhibition of the release of luteinizing hormone–releasing hormone by the hypothalamus that results in the inhibition of ovulation and a natural form of birth control.

With cessation of nursing, reduced prolactin secretion and the inhibitory effects of nonejected milk in the mammary alveoli result in the cessation of milk production. The mammary alveoli regress, and the duct system of the mammary gland returns to the nonpregnant state (Fig. 9.14E).

Clinical Correlation 9.1 summarizes several types of anomalies that affect the integumentary system.

Skull

The skull is a composite structure consisting of two major subdivisions: the neurocranium, which surrounds the brain; and the viscerocranium, which surrounds the oral cavity, pharynx, and upper respiratory passages. Each of these subdivisions consists of two components, one in which the individual bones are first represented by cartilaginous models and are subsequently replaced by bone through endochondral ossification and another in which bone arises directly through the ossification of mesenchyme.

The phylogenetic and ontogenetic foundation of the skull is represented by the chondrocranium, which is the cartilaginous base of the neurocranium (Fig. 9.26A). The fundamental pattern of the chondrocranium has been remarkably preserved in the course of phylogeny. It is initially represented by several sets of paired cartilages. One group (parachordals, hypophyseal cartilages, and trabeculae cranii) is closely related to midline structures. Caudal to the parachordal cartilages are four occipital sclerotomes. Along with the parachordal cartilages, the occipital sclerotomes, which are homologous with precursors of the vertebrae, fuse to form the base of the occipital bone. More laterally, the chondrocranium is represented by pairs of cartilage that are associated with epithelial primordia of the sense organs (olfactory organ, eyes, and auditory organ). Molecular signals from the preoral gut endoderm are required for chondrification of the chondrocranium rostral to the tip of the notochord, which is of neural crest origin, whereas caudal to the tip of the notochord, notochordal signals promote chondrification of the mesodermally derived posterior chondrocranium.

The precursors of the chondrocranium all fuse to form a continuous cartilaginous structure that extends from the future foramen magnum to the interorbital area. This structure must elongate to keep pace with the growing embryo. It does so by first forming numerous primary ossification centers along its length. Then the cartilage between ossification centers develops two mirror-image growth plates, similar to those at the ends of long bones, but with cartilage rather than a joint cavity between them (therefore, they are called synchondroses). Under the influence of Indian hedgehog, the growth plates cause the primary ossification centers to elongate, resulting in the overall elongation of the basicranium.

The individual primordial elements of the chondrocranium undergo several patterns of growth and fusion to form the structurally complex bones of the basicranium (the occipital, sphenoid, and temporal bones and much of the deep bony support of the nasal cavity) (Fig. 9.26B). In addition, some of these bones (e.g., the occipital and temporal bones) incorporate membranous components during their development, so in their final form, they are truly composite structures (see Fig. 9.26D). Other components of the neurocranium, such as the parietal and frontal bones, are purely membranous bones (Box 9.1).

Virtually all the bones of the neurocranium arise as the result of an inductive influence of an epithelial structure on the neighboring mesenchyme. These interactions are typically mediated by growth factors and the extracellular matrix. Immunocytochemical studies have shown the transient appearance of type II collagen (the principal collagenous component of cartilage) at the sites and times during which the interactions leading to the formation of the chondrocranium occur. In addition to type II collagen, a cartilage-specific proteoglycan accumulates in areas of induction of chondrocranial elements. There is increasing evidence that epithelial elements in the head not only induce the skeleton, but also control its morphogenesis. This situation contrasts with morphogenetic control of the appendicular skeleton, which is determined by the mesoderm rather than the ectoderm of the limb bud.

Elements of the membranous neurocranium (the paired parietal and frontal bones and the interparietal part of the occipital bone) arise as flat, platelike aggregations of bony spicules (trabeculae) from mesenchyme that has been induced by specific parts of the developing brain. These bones remain separate structures during fetal development, and even at birth, they are separated by connective tissue sutures. Intersections between sutures where more than two bones meet are occupied by broader areas of connective tissue called fontanelles. The most prominent fontanelles are the anterior fontanelle, located at the intersection of the two frontal and two parietal bones, and the posterior fontanelle, located at the intersection of the parietal bones and the single occipital bone (Fig. 9.27).

During normal development, some sutures between cranial bones close; others remain open. Which of the sutures close and which remain open depend on an intricate interplay among several molecules. Ubiquitously expressed BMP in the embryonic cranium stimulates widespread bone formation, but the BMP antagonist, noggin, is expressed in all embryonic sutures. Under the influence of locally expressed FGF-2, noggin is downregulated in the sutures that fuse, thus allowing BMP-mediated bone formation to cement the two adjacent skull bones. Conversely, the absence of local FGF-2 allows noggin to repress BMP in the sutures that are destined not to fuse.

Similar to the neurocranium, the viscerocranium consists of two divisions: a cartilaginous viscerocranium and a membranous viscerocranium. In contrast to much of the neurocranium, the bones of the viscerocranium originate largely from neural crest–derived mesenchyme. Phylogenetically, the viscerocranium is related to the skeleton of the branchial arches (gill arches). Each branchial arch (more commonly called a pharyngeal arch in humans) is supported by a cartilaginous rod, which gives rise to numerous definitive skeletal elements characteristic of that arch (see Box 9.1). (Details of the organization and derivatives of the noncranial pharyngeal arch cartilages are discussed in Chapter 14 [see Fig. 14.36].)

The membranous viscerocranium consists of a series of bones associated with the upper and lower jaws and the region of the ear (see Fig. 9.26D). These arise in association with the first arch cartilage (Meckel’s cartilage) and take over some of the functions originally subserved by Meckel’s cartilage and many new functions, such as sound transmission in the middle ear. Clinical Correlation 9.3 discusses conditions resulting from skull deformities.

Clinical Correlation 9.3   Conditions Resulting from Skull Deformities

Many conditions are recognizable by gross deformities of the skull. Although many of these are true congenital malformations, others fall into the class of deformities that can be attributed to mechanical stress during intrauterine life or childbirth. Some malformations of the skull are secondary to disturbances in development of the brain. In this category are conditions such as the following: acrania and anencephaly (see Fig. 8.4), which are associated with severe malformations of the brain; microcephaly (see Fig. 9.9), in which the size of the cranial vault accommodates to a very small brain; and hydrocephaly (see Fig. 11.38), in which an extremely enlarged cranial vault represents the response of the skeleton of the head to an excessive buildup of cerebrospinal fluid and a great expansion of the brain.

One family of cranial malformations called craniosynostosis results from premature closure of certain sutures between major membrane bones of the neurocranium. Craniosynostosis is a feature of more than 100 human genetic syndromes and is seen in 1 in 3000 live births. Many of these syndromes are the result of gain-of-function mutants of fibroblast growth factor receptors. One type, the Boston variant, is a dominant gain-of-function mutant of the homeobox-containing gene Msx2 expressed in the mesenchymal tissue of the early sutures and the underlying neural tissue. Exactly how this mutation is translated into premature suture closure is unknown. Premature closure of the sagittal suture between the two parietal bones produces a long, keel-shaped skull referred to as scaphocephaly (Fig. 9.29).

Oxycephaly, or turret skull, is the result of premature fusion of the coronal suture, which is located between the frontal and parietal bones. A dominant genetic condition, Crouzon’s syndrome, has a gross appearance similar to that of oxycephaly, but the malformation of the cranial vault is typically accompanied by malformations of the face, teeth, and ears and occasional malformations in other parts of the body (Fig. 9.30).

Determination and Differentiation of Skeletal Muscle

The mature skeletal muscle fiber is a complex multinucleated cell that is specialized for contraction. Precursors of most muscle lineages (myogenic cells) have been traced to the myotome of the somite (see Fig. 6.10). Although these cells look like the mesenchymal cells that can give rise to many other cell types in the embryo, they have undergone a restriction event committing them to the muscle-forming line. Committed myogenic cells pass through several additional mitotic divisions before completing a terminal mitotic division and becoming postmitotic myoblasts.

Proliferating myogenic cells are kept in the cell cycle through the action of growth factors, such as FGF and transforming growth factor-β. With the accumulation of myogenic regulatory factors (see next section), myogenic cells upregulate the synthesis of the cell cycle protein p21, which irreversibly removes them from the cell cycle. Under the influence of other growth factors, such as insulinlike growth factor, the postmitotic myoblasts begin to transcribe the mRNAs for the major contractile proteins actin and myosin, but the major event in the life cycle of a postmitotic myoblast is its fusion with other similar cells into a multinucleated myotube (Fig. 9.32). The fusion of myoblasts is a precise process involving their lining up and adhering by calcium (Ca⁺⁺)–mediated recognition mechanisms, involving molecules such as M-cadherin, and the ultimate union of their plasma membranes.

Myotubes are intensively involved in mRNA and protein synthesis. In addition to forming actin and myosin, myotubes synthesize a wide variety of other proteins, including the regulatory proteins of muscle contraction—troponin and tropomyosin. These proteins assemble into myofibrils, which are precisely arranged aggregates of functional contractile units called sarcomeres. As the myotubes fill with myofibrils, their nuclei, which had been arranged in regular central chains, migrate to the periphery of the myotube. At this stage, the myotube is considered to have differentiated into a muscle fiber, the final stage in the differentiation of the skeletal muscle cell.

The development of a muscle fiber is not complete, however, with the peripheral migration of the nuclei of the myotube. The nuclei (myonuclei) of a multinucleated muscle fiber are no longer able to proliferate, but the muscle fiber must continue to grow in proportion to the rapid growth of the fetus and then the infant. Muscle fiber growth is accomplished by means of a population of myogenic cells, called satellite cells, which take up positions between the muscle fiber and the basal lamina in which each muscle fiber encases itself (see Fig. 9.32). Operating under a poorly understood control mechanism, possibly involving the Delta/Notch signaling system, satellite cells divide slowly during the growth of an individual. Some of the daughter cells fuse with the muscle fiber so that it contains an adequate number of nuclei to direct the continuing synthesis of contractile proteins required by the muscle fiber. After muscle fiber damage, satellite cells proliferate and fuse to form regenerating muscle fibers.

A typical muscle is not composed of homogeneous muscle fibers. Instead, usually several types of muscle fibers are distinguished by their contractile properties and morphology and by their possession of different isoforms of the contractile proteins. For the purposes of this text, muscle fibers are considered to be either fast or slow.

Muscle Transcription Factors

Myogenesis begins with a restriction event that channels a population of mesenchymal cells into a lineage of committed myogenic cells. The molecular basis for this commitment is the action of members of families of myogenic regulatory factors, which, acting as master genetic regulators, turn on muscle-specific genes in the premuscle mesenchymal cells.

The first-discovered family of myogenic regulatory factors is a group of four basic helix-loop-helix transcription factors, sometimes called the MyoD family (Fig. 9.33). Another regulatory factor, called muscle enhancer factor-2 (MEF-2), works in concert with the MyoD family, but all these myogenic regulatory factors are capable of converting nonmuscle cells (e.g., fibroblasts, adipocytes, chondrocytes, retinal pigment cells) to cells expressing the full range of muscle proteins.

As with many helix-loop-helix proteins, myogenic regulatory proteins of the MyoD family form dimers and bind to a specific DNA sequence (CANNTG), called the E box, in the enhancer region of muscle-specific genes. The myogenic specificity of these proteins is encoded in the basic region (see Fig. 9.33).

The regulatory activities of MyoD and other members of that family are themselves regulated by other regulatory proteins, which can modify their activities (Fig. 9.34). Many cells contain a transcriptional activator designated E12. When a molecule of E12 forms a heterodimer with a molecule of MyoD, the complex binds more tightly to the muscle-enhancer region of DNA than does a pure MyoD dimer. This increases the efficiency of transcription of the muscle genes. A transcriptional inhibitor called Id (inhibitor of DNA binding) can form a heterodimer with a molecule of MyoD. Id contains a loop-helix-loop region, but no basic region, which is the DNA-binding part of the molecule. The Id molecule has a greater binding affinity for a MyoD molecule than another molecule of MyoD and can displace one of the units of a MyoD dimer, thus resulting in more Id-MyoD heterodimers. These bind poorly to DNA and often fail to activate muscle-specific genes.

During muscle development, the myogenic regulatory factors of the MyoD family are expressed in a regular sequence (Fig. 9.35). In mice, the events leading to muscle formation begin in the somite, where Pax-3 and Myf-5, working through apparently separate pathways, activate MyoD and cause certain cells of the dermomyotome to become committed to forming muscle. With increased levels of MyoD, the mononuclear cells withdraw from the mitotic cycle and begin to fuse into myotubes. At this stage, myogenin is expressed. Finally, in maturing myotubes, Myf-6 (formerly called MRF-4) is expressed.

In knockout mice, the absence of a single myogenic regulatory factor (e.g., myf-5, MyoD) alone does not prevent the formation of skeletal muscle (although there may be other minor observable defects), but when myf-5 and MyoD are knocked out simultaneously, muscle fails to form. Another very instructive double knockout of Pax-3 and myf-5 produces mice that are totally lacking in muscles of the trunk and limbs, but the head musculature remains intact. This research shows that in the earliest stages of determination, different regulatory pathways are followed by muscle-forming cells of the head and trunk (see Fig. 9.41).

Muscle growth is negatively controlled as well. Myostatin, a member of the transforming growth factor-β family of signaling molecules, arrests muscle growth when a muscle has attained its normal size. In the absence of myostatin function, animals develop a grossly hypertrophic musculature. Breeds of “double-muscled” cattle are known to have mutations of the myostatin gene.

Because each of the regulators, activators, and inhibitors is itself a protein, their formation is subject to similar positive and negative controls. The complex examples of the regulation of the first steps in myogenesis give some idea of the multiple levels of the control of gene expression and the stages of cytodifferentiation in mammals. Although the molecular aspects of myogenesis are better understood than are the stages underlying the differentiation of most cell types, it is likely that similar sets of interlinked regulatory mechanisms operate in their differentiation of other cells.

Histogenesis of Muscle

Muscle as a tissue consists not only of muscle fibers, but also of connective tissue, blood vessels, and nerves. Even the muscle fibers themselves are not homogeneous, but can be separated into functionally and biochemically different types.

As muscles first form, the myoblasts are intermingled with future connective tissue mesenchyme. The role of the connective tissue in the morphogenesis of a muscle is discussed in the next section. Capillary sprouts grow into the forming muscle for nourishment, and motor nerve fibers enter shortly after the first myoblasts begin forming myotubes.

At one time, it was thought that all myoblasts were essentially identical, and that their different characteristics (e.g., fast or slow) were imposed on them by their motor innervation. Research has shown, however, that in birds and several species of mammals, distinct populations of fast and slow muscle cells exist as early as the myoblast stage, well before nerve fibers reach the developing muscles.

Not only are there fast and slow myoblasts, but also there are early and late cellular isoforms of myoblasts, which have different requirements for serum factors and nerve interactions in their differentiation. When the earliest myoblasts fuse into myotubes, they give rise to primary myotubes, which form the initial basis for an embryonic muscle. The differentiation of primary myotubes occurs before motor nerve axons have entered the newly forming muscle. Subsequently, smaller secondary myotubes arising from late myoblasts form alongside the primary myotubes (Fig. 9.36). By the time secondary myotubes form, early motor axons are present in the muscles, and there is evidence that the presence of nerves is required for the formation of secondary myotubes. A primary muscle fiber and its associated secondary muscle fibers are initially contained within a common basal lamina and are electrically coupled. These muscle fibers actively synthesize a wide variety of contractile proteins.

Early in their life history, embryonic muscle fibers are innervated by motoneurons. Although it has long been assumed that fast and slow motoneurons impose their own functional characteristics on the developing muscle fibers, it now seems that they may select muscle fibers of a compatible type through information contained on their cell surfaces. Initially, a motor nerve may terminate on both fast and slow muscle fibers, but ultimately, inappropriate connections are broken, so fast nerve fibers innervate only fast muscle fibers, and slow nerves innervate only slow muscle fibers.

The phenotypes of muscle fibers depend on the nature of the specific proteins that compose their contractile apparatus. There are qualitative differences in many of the contractile proteins between fast and slow muscle fibers, and within each type of muscle fiber is a succession of isoforms of major proteins during embryonic development. (The isoform transitions of myosin in a developing muscle fiber are used as an example.)

The myosin molecule is complex, consisting of two heavy chains and a series of four light chains (LCs) (Fig. 9.37). Mature fast fibers have one LC1, two LC2, and one LC3 subunits; slow muscle myosin contains two LC1 and two LC2 subunits. In addition, there are fast and slow forms (MHC_f and MHC_s) of the myosin heavy chain (MHC) subunits. The myosin molecules possess adenosine triphosphatase activity, and differences in this activity account partly for differences in the speed of contraction between fast and slow muscle fibers.

The myosin molecule undergoes a succession of isoform transitions during development. From the fetal period to maturity, a series of three developmental isoforms of the MHC (embryonic [MHC_emb], neonatal [MHC_neo], and adult fast [MHC_f]) pass through a fast muscle fiber. (Developmental changes in the LC and MLC subunits are summarized in Figure 9.37.) Other contractile proteins of muscle fibers (e.g., actin, troponin) pass through similar isoform transitions. After injury to muscle in an adult, the regenerating muscle fibers undergo sets of cellular and molecular isoform transitions that closely recapitulate the transitions occurring in normal ontogenesis.

The phenotype of muscle fibers is not irreversibly fixed. Even postnatal muscle fibers possess a remarkable degree of plasticity. These fibers respond to exercise by undergoing hypertrophy or becoming more resistant to fatigue. They also adapt to inactivity or denervation by becoming atrophic. All these changes are accompanied by various changes in gene expression. Many other types of cells can also modify their phenotypes in response to changes in the environment, but the molecular changes are not always as striking as those seen in muscle fibers.

Morphogenesis of Muscle

At a higher level of organization, muscle development involves the formation of anatomically identifiable muscles. The overall form of a muscle is determined principally by its connective tissue framework rather than the myoblasts themselves. Experiments have shown that myogenic cells from somites are essentially interchangeable. Myogenic cells from somites that would normally form muscles of the trunk can participate in the formation of normal leg muscles. In contrast, the cells of the connective tissue component of the muscles seem to be imprinted with the morphogenetic blueprint.

Muscles of the Trunk and Limbs

Quail/chick grafting experiments have clearly shown that the major groups of skeletal muscles in the trunk and limbs arise from myogenic precursors located in the somites. In the thorax and abdomen, the intrinsic muscles of the back (the epaxial muscles) are derived from cells arising in the dorsal myotomal lip, whereas ventrolateral muscles (hypaxial muscles) arise from epithelially organized ventral buds of the somites. Tendons of the epaxial muscles arise from the syndetome layer within the somites (see Box 6.1), whereas tendons of the limb and hypaxial musculature arise from lateral plate mesoderm. In the limb regions, myogenic cells migrate from the epithelium of the ventrolateral dermomyotome early during development. More cranial myogenic cells originating in similar regions of the occipital somites migrate into the developing tongue and diaphragm. At the lumbar levels, precursors of the abdominal muscles also move out of the epithelium of ventrolateral somitic buds.

Early specification of the future hypaxial musculature within the epithelial somite is initially regulated by dorsalizing (possibly a member of the Wnt family) and lateralizing (BMP-4) signals from the ectoderm and lateral plate mesoderm. This process activates two early transcription factors (Six [sine oculis] and Eya [eyes absent]), which leads to a more intense expression of Pax3 and the expression of Lbx1, a homeobox gene that is exclusively expressed in the lateral dermomyotomal lips. Lbx1 may prevent the premature differentiation of the hypaxial musculature. It is highly likely that the prune belly syndrome (Fig. 9.38), which is characterized by the absence of the abdominal musculature, will be found to be caused by a molecular deficiency in this population of myogenic cells.

More recent experiments have shown different cellular behavior in areas of the myotomes adjacent to limb and nonlimb regions. In thoracic segments, cells of the dermatome surround the lateral edges of the myotome; this is followed by an increase in the number of myotubes formed in the myotome and the penetration of the muscle primordia into the body wall. In contrast, at the levels of the limb buds, dermatome cells die before surrounding the early myotubes that form in the myotome. These myotubes neither increase significantly in number nor move out from the myotomes to form separate muscle primordia.

After their origin from the somites, the muscle primordia of the trunk and abdomen become organized into well-defined groups and layers (Fig. 9.39). (Morphogenesis of the limb muscles is discussed in Chapter 10.) The results of numerous experiments have shown fundamental differences in cellular properties between the cellular precursors of limb muscles and axial muscles. These differences are summarized in Table 9.3.

Table 9.3

Differences between Cellular Precursors of Axial and Limb Muscles

Axial Muscles	Limb Muscles
Usual location in medial half of somite	Location in lateral half of somite
Differentiation largely in situ	Migration into limb buds before differentiation
Initial differentiation into mononucleate myocytes	Initial differentiation into multinucleate myotubes
Myogenic determination factors (Myf-5, MyoD) expressed at or before the onset of myotome formation	Expression of myogenic determination genes delayed until limb muscle masses begin to coalesce
Differentiation appears to be strongly influenced by neural tube and notochord	Migration and differentiation little influenced by axial structures

Muscles of the Head and Cervical Region

The skeletal muscle of the head and neck is mesodermal in origin. Quail/chick grafting experiments have shown that the paraxial mesoderm, specifically the somitomeres, constitutes the main source of the cranial musculature. The cells that make up some of the extraocular muscles arise from the prechordal plate of the early embryo.

Myogenesis in the head differs significantly from that in the trunk (Fig. 9.40). Much of the cranial musculature, especially that associated with mastication, arises from cranial unsegmented paraxial mesoderm, equivalent to the somites. Other craniofacial muscles, especially those in the lower jaw and neck, arise from lateral splanchnic mesoderm, as does cardiac muscle. In the early postgastrulation period, the lateral splanchnic mesoderm (sometimes called the cardiocraniofacial morphogenetic field), associated with the future pharynx and probably responding to the same inductive signals from the pharyngeal endoderm, gives rise to both the lower cranial musculature and the secondary heart field. Early in the determination process, both these types of muscle are under the control of transcription factors (e.g., Isl-1, Tbx-1, and Nkx 2.5) that are different from those that control the early development of the trunk musculature. Different types of musculature develop under different sets of early controls before entering similar pathways of differentiation (Fig. 9.41).

As with muscles in the trunk, muscles in the head and neck arise by the movement of myogenic cells away from the paraxial mesoderm through mesenchyme (either neural crest–derived or mesodermal mesenchyme) on their way to their final destination. Morphogenesis of muscles in the cranial region is determined by information inherent in the neural crest–derived connective tissues that ensheathe the muscles. There is no early level specificity in the paraxial myogenic cells; this has been determined by grafting somites or somitomeres from one craniocaudal level to another. In these cases, the myogenic cells that leave the grafted structures form muscles normal for the region into which they migrate, rather than muscles appropriate for the level of origin of the grafted somites.

Certain muscles of the head, in particular muscles of the tongue, arise from the occipital somites in the manner of trunk muscles and undergo extensive migration into the enlarging head. Their more caudal level of origin is evidenced by the innervation by the hypoglossal nerve (cranial nerve XII), which, according to some comparative anatomists, is a series of highly modified spinal nerves. Similar to the myogenic cells of the limb, precursor cells of the tongue musculature express Pax-3 while they are migrating into the head. Despite their final location in the head, these muscles are subjected to the same types of early molecular regulation of myogenesis as trunk muscles.

Anomalies of Skeletal Muscles

Variations and anomalies of skeletal muscles are common. Some, such as the absence of portions of the pectoralis major muscle, are associated with malformations of other structures. Further discussion of anomalies of specific muscles requires a level of anatomical knowledge beyond that assumed for this text.

Muscular dystrophy is a family of genetic diseases characterized by the repeated degeneration and regeneration of various groups of muscles during postnatal life. In Duchenne’s muscular dystrophy, which occurs in young boys, a membrane-associated protein called dystrophin is lacking from the muscle fibers. Its absence makes the muscle fibers more susceptible to damage when they are physically stressed.