Spemann (1, 2) was the first to employ the term "chimera" and to consider the great potential for surgically created chimeric embryos in the analysis of developmental mechanisms. The chimera method has frequently involved imaginative experimental procedures by which cells of one species are grafted into another. Any animal thus composed of different cell populations that derive from more than one fertilized egg should be considered as a chimera. This type of animal can currently be constructed in amphibians, birds, and mammals.
Many fundamental concepts of embryology have been at least partly formulated on the basis of results of cell or tissue transplants between two different embryos, usually separate species of amphibians. The most spectacular transplantation experiments, published by Spemann and Mangold in 1924 (3), demonstrated the organizing power of the dorsal lip of the blastopore during gastrulation by interspecific transplantations of this area. More recently, a model of lens induction was developed by using chimeric eyes (4). The technical advantages of producing amphibian chimeras are straightforward, owing to the independence of the embryos from their parents. They are easily accessible and receptive to foreign tissue, even across species barriers. All these qualities are not shared by higher vertebrates, such as birds and mammals. Nonetheless, bird embryos have several advantages over other vertebrate embryos, making certain interesting approaches feasible. The greatest advantage is continual accessibility within the egg throughout the developmental period. Another is the ease with which the various rudiments can be delineated, and thus removed and replaced, with extreme precision. An avian chimera obtained by combining quail and chick cells has been the most successful method, having provided a continual source of new data about developmental mechanisms for almost 30 years (5, 6). With the advent of the quail-chick nuclear marker, which is particularly simple to employ, easy to identify, and endowed with great resolving power, avian chimeras have been used to study the ontogeny of the nervous system, the development of the hematopoietic and immune systems, and the formation of muscles and skeleton. Quail heterochromatin in the nucleus is concentrated around the nucleolus. This creates a large, deeply staining mass that is easily distinguishable from the diffuse heterochromatin of chick cells. Moreover, there are some antigens that are quail-specific and not detectable in chick cells. These phenomena allow individual quail cells to be readily distinguished, even when most of the cell population is chick.
Although the avian embryo is a practical model, perfectly suitable for tissue graft experiments after the incubated egg is opened, it is difficult to undertake this type of investigation in the mammalian fetus in utero. Nonetheless, it has become routine to remove postimplanted mammalian embryos from the uterus, manipulate them, and return them to a foster mother for further development. Thus, chimeric mice are the result of two or more early-cleavage (usually 4- or 8-cell) embryos that have been artificially aggregated to form a composite embryo. Since each cell is able to produce any component of the body, the construction of the chimeric mouse has very important consequences for the study of mammalian ontogeny. A very powerful application of this technique is the transfer of genes into every cell of the mouse embryo. During mouse development, there is a stage when only two cell types are present: outer cells, which will form the fetal portion of the placenta, and inner cells, which will give rise to the embryo itself. These inner cells are known as embryonic stem cells because each in isolation can generate all the cells of the embryo (7, 8). These cells can be grown in culture, where they are treated to incorporate new DNA. The new embryonic stem cells can then be injected into another early-stage mouse embryo, resulting in a chimeric mouse. Mice that derive from these animals are transgenic mice. Our understanding of regulatory mechanisms in mammalian development is improving increasingly rapidly as a result of the construction of these genetically modified mice. A combination of the tools of developmental genetics with those of embryology should lead to real advances in the study of such mechanisms. In this field, our group has pioneered the grafting of embryonic tissues from transgenic mice into the chick embryo (9, 10). Owing to these interspecies grafting experiments, it is possible to monitor factors that regulate the expression of a particular gene in vivo. The value of this technique is greatly increased when a reporter gene is used to follow the changes in gene expression of the grafted cells. The possibility of conducting grafts until late stages of in vivo development allows the behavior of wild and mutant mouse cells to be observed at any developmental stage and location.
1. Amphibian Chimera
Using amphibians, Spemann and Mangold (3) improved our understanding of the specification of the nervous system by transplanting dorsal blastopore lip tissue from an early gastrula into the ventral ectoderm of another gastrula. They used differently pigmented embryos from two species of newt: darkly pigmented Triturus taeniatus and nonpigmented Triturus cristatus. On the basis of color, it was easy to distinguish host and donor tissues. The dorsal blastopore lip tissue from early Triturus taeniatus gastrula, once transplanted into an early Triturus cristatus gastrula in the region, would normally become ventral epidermis. In fact, the donor tissue did not become belly skin but invaginated and formed a secondary embryo, face to face with its host. The more recent use of nuclear markers has allowed Spemann’s results to be confirmed (11). Such chimeras elegantly demonstrate the organizing power of the dorsal lip of the blastopore in amphibian gastrula, since whole secondary embryos formed under the influence of the transplanted tissue.
Considerable advances have also taken place in the field of differentiation and organogenesis through the use of amphibian chimeras. The more common examples are those involving the interaction of epithelia with adjacent mesenchyme. After being separated, embryonic epithelium and mesenchyme can be recombined in different ways (12). In a classic experiment, Spemann and Schotte (13) transplanted flank ectoderm from an early frog gastrula into the region of newt gastrula destined to become part of the mouth. Similarly, the presumptive flank ectodermal tissue of newt gastrula was placed into the presumptive oral regions of frog embryos. The structures of the mouth region differ greatly between salamander and frog larvae. The Triturus salamander larva has club-shaped balancers beneath its mouth, whereas frog tadpoles produce mucus-secreting glands and suckers. Frog tadpoles also have a horny jaw without teeth, whereas the salamander has a set of calcareous teeth in its jaw. The larvae resulting from the transplants were chimeras. The salamander larvae had froglike mouths, and the frog tadpoles had salamander teeth and balancers. In other words, the mesodermal cells instructed the ectoderm to make a mouth, but the ectoderm responded by making the only mouth it "knew" how to make. Thus, instructions sent by mesenchymal tissue can cross species barriers, although the response of the epithelium is species-specific. Thus, organ-type specificity is usually controlled by the mesenchyme within a species, but species specificity is usually controlled by the responding epithelium.
The cells that form the lens are derived from a region of head ectoderm in contact with optic vesicles of the anterior neural plate. Servetnick and Grainger (14) removed animal cap ectoderm from various gastrula stages and then transplanted them into the presumptive lens region of neural-plate-stage embryos. Ectoderm from early gastrula showed little or no competence to form lenses, but ectoderm from slightly later stages was able to. By the end of gastrulation, this ability to respond to the neural plate signal had been lost. This competence was found to be inherent within the ectoderm itself and not induced by other surrounding tissues. These observations showed that only mid- to late-gastrula ectoderm is able to respond to signals from the anterior neural plate. It has recently been demonstrated that the transcription factor Pax6 may play a role in the determination processes of eye tissue (see Pax Genes).
2. Avian Chimera
When portions of quail embryo are grafted into a similar region of chick embryo, the cells become integrated into the host and participate in the construction of the appropriate organs. This grafting is done while the embryo is still inside the egg, and the chick that hatches is a "chimera," having a portion of its body composed of quail cells (15). The quail is usually chosen as a donor because it is easier to identify quail cells among chick cells than the reverse.
The ontogeny of the peripheral nervous system is one of the fields in which the use of avian chimeras is the most fruitful. This system arises almost entirely from the neural crest, a transient structure that develops at the neural tube apex. To follow the fate of chick neural tube-bearing neural crests, a segment is replaced by a homologous quail fragment. This method has permitted the diverse range of neural crest potentialities to be well-described, and the list of neural crest derivatives well-defined (6).
In relation to its origin, neural crest not only participates in the formation of spinal and cranial sensory ganglia, sympathetic and parasympathetic ganglia and plexuses, and Schwann cells of the peripheral nerves, but it also gives rise to endocrine and paraendocrine cells (calcitonin-producing cells, carotid body type-I cells, adrenomedullary cells) and pigment cells. The construction of an avian chimera has demonstrated the contribution of the cephalic neural crest to mesectodermal derivatives. For example, nasal and maxillary processes are built up partly by crest cells of mesencephalic origin. The mesencephalic crest cells also form the cephalic skeleton, the upper and lower jaws, the palate, and the tongue. The rhombencephalic crest participates in formation of the pre-optic region and the hyoid arch skeleton. In addition to cartilage and bone, the cephalic neural crest takes part in other derivatives of the head and neck, such as dermis and connective tissue.
Brain development has also been characterized by the use of the quail-chick chimera technique (16). Quail-chick brain chimeras can hatch and survive without showing impaired movement or locomotion, which indicates that functional synapses have been established between host and donor neurones, as well as between donor neurones and host muscles. Moreover, the normal behavior of the chimeras demonstrates that proper neuronal connections develop in the brain, which means that quail axons recognize local signals for growth and directionality in the chick environment, as they do in normal development. An interesting application of the chimeras is illustrated by chicks with transplanted quail mesencephalic-diencephalic brain areas, which then exhibit quail vocalization (17). Another important issue is the elucidation of when and where groups of cells in the brain make commitments to particular development pathways. Genes with spatially restricted expression are of particular interest, since they may indicate the existence of committed groups of cells that are important in pattern formation, but are not discernible on the basis of morphology. An example is provided by the zinc-finger gene Krox-20, which has restricted domains of expression in the early neural plate. Krox-20 is expressed in the early neural epithelium, first in one stripe, and then in two stripes in the hindbrain. Subsequently, Krox-20 is expressed in two alternating segments (rhombomeres) in the hindbrain. The generation of stripes of Krox-20 expression in the early neural plate suggests that rhombomeric precursor cells are committed prior to the morphological appearance of the rhombomere.
Hox homeotic gene expression is seen along the dorsal axis, from the anterior boundary through to the tail. Direct evidence for hindbrain plasticity comes from quail-chick chimera experiments showing that anterior-to-posterior transpositions can reprogram Hox expression and induce a transformation in cell fate (18). The quail-chick system has also shown that environmental cues play a significant role in maintaining the Hox code in the neuroepithelium.
Chimeric experiments involving the muscles of the body, limbs, and skeleton have also been carried out in birds. For development, somites are the primitive metameric structures of the vertebrate body from which arise the vertebrae that surround the spinal cord, the muscles and connective tissue holding the vertebrae, the dermal layer of the skin of the back, and the limb musculature. All these data were obtained by constructing a chimeric embryo after interspecific grafts of somites between quail and chick embryos (19, 20). The somites appear as pairs of epithelial spheres that bud off from the unsegmented paraxial mesoderm in a craniocaudal direction. They become polarized into a ventral mesenchymal compartment, the sclerotome, which yields the dorsal skeleton, and a dorsal epithelial component, the dermomyotome, from which striated muscle and dermis arise. By constructing a quail-chick chimera after interspecific exchanges of medial or lateral halves of newly-formed somites, Ordhal and Le Douarin (21) determined that, regardless of the somites, cells farthest from the neural tube (ie, lateral) migrate to form the body wall and limb musculature.
The interspecies graft method in birds has contributed considerably to understanding the control mechanisms of somite patterning. This method demonstrates that specification of the somite is accomplished by the interaction of different tissues that form its environment. In fact, the newly formed somite is composed mostly of unspecified cells, and the determination of somite compartments toward the different lineages is regulated by environmental cues. The ventral-medial portion of the somite is induced to become the sclerotome by factors, especially Sonic hedgehog protein, that are secreted from the notochord and neural tube floor plate. If portions of the notochord, the source of Sonic hedgehog, are transplanted next to other regions of the somite, those regions also become sclerotome cells. These sclerotome cells express a new transcription factor, Pax1, which activates cartilage-specific genes and is necessary for the formation of vertebrae (22). In similar ways, the myotome is induced by two distinct signals. The muscles surrounding the body axis, which arise from the medial portion of the somite, are induced by factors from the dorsal neural tube, probably members of the Wnt family (23, 24). The muscles derived from the lateral portion of the somite, which form the musculature of the limbs and body wall, are probably induced through the combination of Wnt proteins from the epidermis with bone morphogenetic protein-4 (BMP4) protein from the lateral plate mesoderm (25). These factors cause the myotome cells to express particular transcription factors, MyoD and Myf5, that activate muscle-specific genes. The dermatome differentiates in response to another factor secreted by the neural tube, neurotrophin 3 (NT-3) (26).
3. Mammalian Chimera
The laboratory mouse is by far the most popular mammal used to construct chimeras for developmental studies. Mice offer a large variety of genetically well-characterized strains that provide scope for using genetic markers. Moreover, they breed throughout the year and thus continually supply embryos. The most convenient time for manipulation of a mammalian embryo is before its implantation. During the first three to four days of gestation, mouse embryos have not yet formed an attachment in the mother’s reproductive tract and can thus be explanted, manipulated, and then retransplanted into another mouse to continue their development. Chimeric mice are generally formed by the fusion of two early embryos that organize to produce a single mouse with two distinct cell populations.
A skeletal muscle cell (myotube) is an elongated cell containing many nuclei. It has been widely debated whether this cell is derived from the fusion of several mononucleated precursor cells (myoblasts) or from a single cell that undergoes nuclear division without cell division. Evidence for the fusion process has been provided by mouse chimeric constructions. Mintz and Baker (27) fused mouse embryos from two strains that produce different types of the dimeric enzyme isocitrate dehydrogenase: one strain makes the A subunit and the other B. If myotubes are formed from one cell whose nuclei divide without cytokinesis, the dimeric enzyme will be purely AA or BB. If, however, myotubes are formed by fusion between cells, some might code for the B subunit and others for A, in which case the molecules of the enzyme will be hybrid (AB). Electrophoresis can separate these three types (see Isozyme, Isoenzyme). The presence of the hybrid AB enzyme in extracts of skeletal muscle tissue confirms the fusion model.
The study of mutations that impair development or function has long been recognized as a valuable means of elucidating the normal role of genes in such processes. Genetically chimeric animals consisting of mixtures of mutant and wild type cells can be of considerable value for studying the genes that are primarily implicated in developmental mechanisms. Methods of manipulating mouse embryos and transferring genes into every embryonic cell are now standard procedures. In recent years, an extraordinary increase in the use of these methods has led to numerous exciting prospects.
The most widely used method to produce transgenic mice is to inject cloned DNA into a pronucleus of a fertilized egg. Although this method has been used primarily for studies of gene expression, an unexpected benefit is the frequent integration of the injected DNA into genes, causing insertional mutations. Experimental infection of embryos by retroviruses has also been developed as a method for gene transfer. As in the case of spontaneous infection, this approach has resulted in insertional mutations. An additional approach to producing transgenic mice involves genetic modification of tissue culture lines of embryonic stem cells, followed by their re-incorporation into growing embryos. By applying somatic cell genetic techniques to these cells while in culture, both random and selected insertional mutations have been produced. Another method for assaying the function of a gene is to eliminate it from the genome of the whole organism. Very recent studies, using strategies to inactivate genes by homologous recombination in embryonic stem cells, with the subsequent generation of germ-line chimeras, have made this scenario possible and thereby initiated a new era in mammalian developmental biology.
The power of this technology can be illustrated by the int-1 gene, which is expressed during central nervous system development in a temporally and spatially restricted fashion. Following inactivation of the gene by homologous recombination in embryonic stem cells, germ-line chimeras were obtained and bred to derive homozygous int-1 -deficient mice. These homozygous mice were not viable, because specific regions of the brain were missing. This experiment indicates that int-1 is an essential gene for early development and suggests that its expression might determine the fate of a cell. The Hox a-3 gene has been found to control segment-specific gene expression in Drosophila, but what is its role in mammals? Chisaka and Capecchi (28) used gene targeting to determine the function of Hox a-3 in the development of the mouse. Homozygous mutants of Hox a-3 were found to have severe anomalies in development of the hindbrain and in mesectodermal neural crest derivatives and to lack thyroid, parathyroid, and thymus glands. These studies clearly open up new avenues for the study of mammalian development at the molecular level and will certainly be instrumental in unraveling the molecular networks responsible for the functioning of a multicellular organism.
4. Mammalian Avian Chimera
In chimeras, the analysis of cell lineages depends on the ability to distinguish between cells of different origins. Markers are required to trace the developmental fate of cells and to recognize them at all ages in chimeric embryos. In mammals, genetic manipulations appear to be valuable means of elucidating the normal role of genes and identifying altered cells. Furthermore, the environment of the chimera allows these cells to survive and to display their phenotype. In chimeras, identification of all the descendants of genetically modified cells has thus far provided a reliable and sensitive means of measuring alterations in embryogenesis. In conjunction with an in situ marker, this may help in determining the cell type and the nature of the functions affected by the particular mutant. Extremely sensitive labeling is possible with transgenic mouse lines since cells that integrate foreign DNA, such as a reporter gene encoding for Escherichia coli b-galactosidase, are detectable by simple histochemical revelation, which can be used to differentiate grafted cells from host cells. Currently,the major means of exploring this genetic tool is by in vitro experiments in which mutant-induced tissue is cultured with wild-type tissue. The culture system is efficient in tissue development for only limited periods, however, and the media used may affect the tissue outcome. In vivo micromanipulation remains the most powerful means of studying the fates of cells, their origins, and the cell-cell interactions that govern development. Since implantation of mouse embryo is particularly unsuitable for in vivo manipulation, chick embryo is used as the site for developing mouse embryonic cells. Studies in such chimeras provide considerable information on the lineage and the differentiation of mouse embryonic cells, since in ovo grafted mouse cells are accessible throughout chick host embryogenesis. Moreover, these results are indicative of the information to be obtained through genetic manipulations in conjunction with embryonic tissue transplantation. It has been clearly demonstrated that in ovo transplantation provides a suitable environment for the development of mammalian cells and that the information supplied by this environment is capable of promoting mouse cell differentiation.
Chimeras were prepared by transplanting somites from nine-day postcoitum mouse embryos into two-day-old chick embryos at different axial levels. Mouse somitic cells then differentiated in ovo into dermis, cartilage, and skeletal muscle, as they normally do in the course of development, and were able to migrate into chick host limb. To trace the behavior of somitic myogenic stem cells more closely, somites arising from mice bearing a transgene of the desmin gene linked to a reporter gene coding for E. coli b-galactosidase were grafted in ovo. Interestingly, the transgene was rapidly expressed in myotomal muscles derived from implants. In the limb muscle mass, positive cells were found several days after implantation. This method facilitates investigation of the mechanisms of mammalian development, allowing the normal fate of implanted mouse cells to be studied and providing suitable conditions for identification of descendants of genetically modified cells.
Chimeras were also prepared by transplanting fragments of neural primordium mouse embryos into chick embryos at different axial levels. Mouse neuroepithelial cells differentiated and organized to form the different cellular compartments normally constituting the central nervous system. The graft entered into the development of the peripheral nervous system through migration of neural crest cells associated with mouse neuroepithelium. Depending on the graft level, mouse crest cells participated in the formation of various derivatives, such as head components, sensory ganglia, orthosympathetic ganglionic chain, nerves, and neuroendocrine glands. Knock-out mice with the tenascin genes inactivated, which express LacZ instead of tenascin and show no tenascin production (29), were used specifically to follow Schwann cells lining nerves derived from the implant. Together with the previous results on somite development, this study shows that the chick embryo constitutes a privileged environment, facilitating access to the developmental potentials of normal or defective mammalian cells. It allows study of the histogenesis and precise timing of formation of a known structure, as well as the implications of a given gene at all equivalent mammalian embryonic stages.
