Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Development 122, 231-241 (1996) Printed in Great Britain © The Company of Biologists Limited 1996 DEV3253 231 The dorsal neural tube organizes the dermamyotome and induces axial myocytes in the avian embryo Martha S. Spence1, Joseph Yip2 and Carol A. Erickson1,* 1Section of Molecular and Cellular Biology, University of California, Davis, CA 95616, USA 2Department of Neurobiology, School of Medicine, University of Pittsburgh, Pittsburgh, PA 15261, USA *Author for correspondence (e-mail: [email protected]) SUMMARY Somites, like all axial structures, display dorsoventral polarity. The dorsal portion of the somite forms the dermamyotome, which gives rise to the dermis and axial musculature, whereas the ventromedial somite disperses to generate the sclerotome, which later comprises the vertebrae and intervertebral discs. Although the neural tube and notochord are known to regulate some aspects of this dorsoventral pattern, the precise tissues that initially specify the dermamyotome, and later the myotome from it, have been controversial. Indeed, dorsal and ventral neural tube, notochord, ectoderm and neural crest cells have all been proposed to influence dermamyotome formation or to regulate myocyte differentiation. In this report we describe a series of experimental manipulations in the chick embryo to show that dermamyotome formation is regulated by interactions with the dorsal neural tube. First, we demonstrate that when a neural tube is rotated 180° around its dorsoventral axis, a secondary dermamyotome is induced from what would normally have developed as sclerotome. Second, if we ablate the dorsal neural tube, dermamyotomes are absent in the majority of embryos. Third, if we graft pieces of dorsal neural tube into a ventral position between the notochord and ventral somite, a dermamyotome develops from the sclerotome that is proximate to the graft, and myocytes differentiate. In addition, we also show that myogenesis can be regulated by the dorsal neural tube because when pieces of dorsal neural tube and unsegmented paraxial mesoderm are combined in tissue culture, myocytes differentiate, whereas mesoderm cultures alone do not produce myocytes autonomously. In all of the experimental perturbations in vivo, the dorsal neural tube induced dorsal structures from the mesoderm in the presence of notochord and floorplate, which have been reported previously to induce sclerotome. Thus, we have demonstrated that in the context of the embryonic environment, a dorsalizing signal from the dorsal neural tube can compete with the diffusible ventralizing signal from the notochord. In contrast to dorsal neural tube, pieces of ventral neural tube, dorsal ectoderm or neural crest cells, all of which have been postulated to control dermamyotome formation or to induce myogenesis, either fail to do so or provoke only minimal inductive responses in any of our assays. However, complicating the issue, we find consistent with previous studies that following ablation of the entire neural tube, dermamyotome formation still proceeds adjacent to the dorsal ectoderm. Together these results suggest that, although dorsal ectoderm may be less potent than the dorsal neural tube in inducing dermamyotome, it does nonetheless possess some dermamyotomal-inducing activity. Based on our data and that of others, we propose a model for somite dorsoventral patterning in which competing diffusible signals from the dorsal neural tube and from the notochord/floorplate specify dermamyotome and sclerotome, respectively. In our model, the positioning of the dermamyotome dorsally is due to the absence or reduced levels of the notochord-derived ventralizing signals, as well as to the presence of dominant dorsalizing signals. These dorsal signals are possibly localized and amplified by binding to the basal lamina of the ectoderm, where they can signal the underlying somite, and may also be produced by the ectoderm as well. INTRODUCTION lack obvious morphological polarity (Bellairs, 1963; Christ et al., 1972; Mestres and Hinrichsen, 1976). Within a few hours, however, morphogenetic movements generate regional histological differences along the dorsal-ventral axis of each somite (Mestres and Hinrichsen, 1976; reviewed by Stern et al., 1988; Ordahl, 1993). Cells comprising the ventromedial portion of the somite undergo an epithelial-to-mesenchymal transforma- The paraxial mesoderm first arises by ingressing through the primitive streak to form a uniform band of tissue that flanks the neural tube. Although initially unsegmented, the paraxial mesoderm soon becomes partitioned in an anterior-to-posterior wave to form the somites, which are epithelial in nature and Key words: dermamyotome, myogenesis, neural tube, somites, avian embryo 232 M. S. Spence, J. Yip and C. A. Erickson tion to form the sclerotome, which lies lateral to the neural tube and notochord (Hay, 1968). In contrast, cells in the dorsal portion of the somite maintain an epithelial organization and constitute the transient dermamyotome. Cells at the cranial edge of the dermamyotome later give rise to the more ventrally situated myotome, while the remaining dermamyotome cells become dermatome (Kaehn et al., 1988; Tosney et al., 1994). Although the derivatives of the somites (i.e. dermatome, myotome and sclerotome) are positioned with obvious dorsalventral polarity, such regionalization is not intrinsic to either the unsegmented paraxial mesoderm or the most recently formed somites. This absence of regional specification at early stages is revealed in experiments in which 180° rotation of the paraxial mesoderm or early somites has no effect on the subsequent dorsal-ventral patterning (Aoyama and Asamoto, 1988). Similarly, when the dorsal half of an early somite is replaced with a ventral half, the grafted ventral tissue undergoes normal myogenesis (Christ et al., 1992). These studies suggest that dorsal-ventral pattern within each somite is established through interactions with surrounding axial structures during and subsequent to segmentation. Several axial tissues could potentially contribute to somite patterning. For example, the neural tube and notochord are known to play a role in the development of the sclerotome. Cocultures of somite with neural tube and notochord produce an abundance of cartilage, a sclerotome derivative (Lash et al., 1957; Lash, 1968; Kosher and Lash, 1975). Moreover, when a notochord is grafted ectopically to dorsal regions of the somite, the sclerotome expands at the expense of the dermamyotome (Pourquié et al., 1993; Bober et al., 1994; Goulding et al., 1994). Finally, early removal of the notochord enhances the development of the dermamyotome and results in the absence of sclerotome (van Stratten and Hekking, 1991; Rong et al., 1992; Goulding et al., 1993, 1994). Recent studies reveal that Sonic hedgehog is produced by the notochord and floor plate at the correct time to induce sclerotome (Johnson et al., 1994) and, furthermore, that heterologous cells expressing Sonic hedgehog can induce sclerotome formation, as assessed using molecular markers, both in culture assays (Fan and TessierLavigne, 1994) and in vivo (Johnson et al., 1994). Unlike the sclerotome, candidate molecules that regulate the formation of dermamyotome or the myotome from it have not been identified. Indeed, our understanding of how these tissues develop is further complicated by persistent controversy concerning which neighboring tissues exert inductive influences on the paraxial mesoderm. The neural tube has been demonstrated repeatedly to play a role in myogenesis (e.g. Watterson et al., 1954; Vivarelli and Cossu, 1986; Kenny-Mobbs and Thorogood, 1987; Christ et al., 1992; Rong et al., 1992; Borman and Yorde, 1994; Buffinger and Stockdale, 1994; Stern and Hauschka, 1995; Münsterberg and Lassar, 1995), and there is some evidence that a signal is produced specifically by the dorsal neural tube (Fan and Tessier-Lavigne, 1994). Yet several studies also indicate that dermamyotomeand myocyte-inducing molecules are produced by other tissues, including the surface ectoderm (Christ et al., 1972; Rong et al., 1992; Fan and Tessier-Lavigne, 1994; Kuratani et al., 1994), the notochord or ventral neural tube (Buffinger and Stockdale, 1994; Stern and Hauschka, 1995; Münsterberg and Lassar, 1995) and neural crest cells (Christ et al., 1992; Borman and Yorde, 1994). Finally, some evidence suggests that dorsal fate is a default pathway in the absence of a ventralizing signal from the neural tube and notochord (Pourquié et al., 1993), and therefore that dorsal-specifying signals may not be necessary at all. To characterize the cellular interactions that signal formation of the dermamyotome as well as specification of myocytes, we have employed the techniques of experimental embryology. Because most of our experiments attempt to identify and characterize various signaling centers in the context of the embryo, the environmental conditions and spatial organization of inducing and responding tissues are close to normal, in contrast with previously employed tissue culture assays where these signals may be lost or altered. Results from a series of manipulations including dorsoventral rotation of the neural tube, extirpation of the dorsal neural tube and grafting of putative inducing tissues ventrally adjacent to the sclerotome together support a model in which the dorsal neural tube induces the formation of the dermamyotome. In contrast, surface ectoderm, ventral neural tube and neural crest cells appear to have only modest, direct signaling roles, or do not participate in the induction of the dermamyotome. Furthermore, our study suggests that the signal emanating from the dorsal neural tube is a diffusible molecule that can compete, in the embryo, with the ventralizing signal from the notochord. MATERIALS AND METHODS Embryo culture White Leghorn chicken embryos were used for all in vivo manipulations (Avian Sciences Department, University of California-Davis or Western Scientific, Sacramento, CA). Some embryos were treated and maintained in ovo. On the day before the operation, a small hole was cut into the egg shell above the embryo, the window sealed with cellophane tape and the egg returned to the incubator. Just prior to any experimental manipulation, a drop of 0.02% neutral red in Locke’s saline was applied to the vitelline membrane to stain the embryo and reveal the vitelline membrane, which was subsequently slit with a tungsten needle to access the embryo. After surgery, the eggs were resealed and incubated at 38°C and 70% humidity until the embryos were fixed. Alternatively, some embryos were cultured in vitro after surgery. Embryos were cut from the blastoderm and transferred to a 100 mm Petri dish coated with a layer of 2% agar. The vitelline membrane was teased away, the experimental manipulation performed (see below) and the embryo cultured according to previously published methods (Erickson and Goins, 1995). Briefly, the embryo was oriented ventral side up, the saline removed so that the embryo was flat and stretched, and a filter paper ring (Whatman #1) centered over the embryo to keep the embryo fully spread. A Nuclepore filter (8 µm pore size, Costar Corp., Pleasanton, CA) was then layered over the embryo and ring, which served to support the embryo during later culture. The whole assemblage was then turned so that the embryo was dorsal side up and suspended over a well in an organ culture dish (Falcon #3037) containing 1.5 ml Liebowitz L-15 medium supplemented with 10% fetal calf serum, 2 mM glutamine and 1% penicillin-streptomycin (all from GIBCO, Grand Island, NY). The moat surrounding the well was filled with sterile double-distilled water and the embryo returned to the egg incubator at 38°C and 70% humidity. Neural tube rotations and ablations To discern whether the neural tube influenced the polarity of adjacent axial structures, the orientation of the neural tube was experimentally altered. Segments of neural tubes were excised with tungsten needles, Dorsal neural tube induces the dermamyotome leaving the notochord in place, as described previously (Yip, 1990). The neural tube was then rotated 180° around the dorsal-ventral axis and immediately repositioned in the embryo (Fig. 1). Embryos were incubated for another 9-26 hours and fixed. The dorsal neural tube was ablated from some embryos to determine if dermamyotome development was inhibited. The dorsal portion of the neural tube of stage 12-14 embryos was suctioned off using a mouth-controlled micropipette with a 0.1 mm tip diameter. Embryos were reincubated until they reached stage 16-18 and then fixed. Grafting experiments The ability of various embryonic tissues to induce an ectopic dermamyotome from the ventral somite was assessed by grafting these tissues into the space between the notochord and segmental plate. Embryos were cut from the blastoderm and transferred, ventral side up, to an agar-coated Petri dish filled with warm Locke’s saline. A small slit was made in the endoderm between the neural tube and paraxial mesoderm just posterior to the last-formed somite. The piece of donor tissue was lightly stained with neutral red and maneuvered into the slit with a tungsten needle. The embryos were then incubated dorsal side up for another 18 to 22 hours (until stage 16-18) and fixed. Embryonic tissues tested for their ability to induce the formation of a dermamyotome included dorsal, lateral and ventral neural tube, as well as ectoderm and neural crest cells. To obtain these tissues, trunks at the axial level of the unsegmented mesoderm of stage 1315 embryos were excised with tungsten needles and digested with Pancreatin (GIBCO) at 37°C until the tissues began to separate. The trunks were then transferred to cold Hank’s balanced salt solution (HBSS) and the various tissues separated by manual dissection. At this time, the neural tube was cut longitudinally into dorsal, ventral and lateral segments. The isolated tissues were incubated in F12 medium supplemented with 10% fetal calf serum, 2 mM glutamine, 3% chick embryo extract and 1% penicillin-streptomycin solution (complete medium) for 2 hours to allow the cells to re-express proteolytically digested surface proteins. Neural crest cells were obtained from 24-hour-old neural tube cultures, as described previously (e.g. Loring et al., 1981; Erickson and Goins, 1995). The tissues were washed in HBSS and stained with neutral red prior to grafting. In some cases, the tissues were labeled for 2 hours in a 50 µg/ml solution of wheat germ lectin conjugated with fluorescein (Sigma, St Louis, MO) to identify unequivocally the graft in tissue sections. 233 Co-culture studies To assay the ability of various embryonic tissues to induce myocytes from unspecified paraxial mesoderm, the trunk regions (encompassing the last 9 somites plus segmental plate) of stage 13-15 quail embryos (Coturnix coturnix japonica) were removed with tungsten needles and subsequently cut into 3-somite-length pieces (300 µm in length), including a piece comprising the entire segmental plate. The pieces were digested briefly with full-strength Pancreatin at 37°C until the tissues could be separated easily from each other using tungsten needles. For each axial level, the neural tube, notochord, ectoderm and mesoderm (either somites or segmental plate) were collected separately and stored in complete F12 medium until used. A piece of segmental plate approximately 300 µm in length was introduced into a small culture well (8-chamber culture slides, LabTek, Nunc, Naperville, IL) in complete F12 medium. A potential inducing tissue was also added to each well and the tissues maneuvered with a blunt tungsten needle until they touched. Cultures were incubated at 37°C in 5% CO2 for 4 days and the medium was replaced every other day. The presence of myocytes was determined by immunocytochemistry (see below). Because ectoderm does not readily spread on a planar substratum, we also employed collagen gels to immobilize and culture the tissues. Collagen solution was prepared as described previously (Tucker and Erickson, 1984) and 50 µl droplets gelled in 35 mm plastic Petri dishes. The mesoderm and potential inducers were pipetted onto this base and covered with a second layer of collagen so that the tissues were completely enveloped. Cultures were incubated for 4 days. Immunocytochemistry We used a variety of techniques to identify myocytes and neural crest cells, both in embryo sections and in culture. For most of our studies, we employed whole-mount labeling (see Erickson et al., 1992; Tosney et al., 1994 for details). A polyclonal antibody against the intermediate filament desmin (1:200, Biogenex, San Ramon, CA) was used to identify myocytes (Erickson et al., 1987; Kaehn et al., 1988; Tosney et al., 1994). In some cases, operated embryos were labeled with the HNK-1 antibody to visualize the distribution of neural crest cells. HNK-1-producing hybridoma cells were obtained from the American Type Culture Collection and culture supernatant produced in our facilities. After tissues were immunolabeled, they were embedded in Spurr’s resin, sectioned at 3 µm on a Sorvall MT-2 microtome using a diamond histoknife (Diatome) and viewed with a Leitz Diaplan microscope equipped with epifluorescence. The color micrographs in Somite Neural crest cells Ectoderm from dorsal surface Notochord Fig. 1. Diagrammatic representation of the neural tube rotation experiments. A segment of the neural tube from the level of somites 16 to 21 was excised, rotated dorsoventrally and replaced. The embryo was reincubated for another 9 to 26 hours. Note that these rotations were performed at an axial level where the dorsal-ventral axis of the somite was already apparent. Even so, a secondary dermamyotome develops from the sclerotome. 234 M. S. Spence, J. Yip and C. A. Erickson Fig. 4 were captured using a cooled CCD C72 MTI camera and NIH Image software and merged in Photoshop. Cultures were labeled with the monoclonal antibody 13F4 to identify myocytes (Rong et al., 1987). In our hands this antibody produced less background fluorescence in cultured tissue than the desmin antibody. Cultures were fixed in 4% paraformaldehyde for 20 minutes, washed in PBS, blocked in 0.1% BSA/PBS and incubated in 13F4 (1:100 dilution in 0.1% BSA/PBS; Developmental Studies Hybridoma Bank) for 2 hours. The cultures were washed with PBS (2×), blocked in 0.1% BSA/PBS and incubated in secondary antibody (goat anti-mouse conjugated with FITC or RITC, 1:100; Cappel) for one hour. The cultures were then washed, coverslipped and viewed as described above. In some cases, cultures were doubled-labeled with the HNK-1 antibody and the desmin antibody. RESULTS Nomenclature In this study, we use somite pairs as a convenient marker for identifying and reporting axial positions. Traditionally, the first formed somite pair is designated as ‘1’, the next most posterior pair as ‘2’, and so on. Our study focuses on the thoracic level, which encompasses somite pairs 20-27. Regardless of the stage of development, the relative maturity of each somite is consistent from one embryo to the next at the thoracic level when measured with reference to the most recently formed somite (Loring and Erickson, 1987; Ordahl, 1993; Tosney et al., 1994). Consequently, we also use a posterior-to-anterior staging scheme employed previously (see Loring and Erickson, 1987; Tosney et al., 1994), in which the most recently formed pair of somites is designated as ‘−1’, the next most anterior pair as ‘−2’, etc. Neural tube rotations Since there is reason to believe that the dorsoventral pattern of the somites is established through interactions with the neural tube (summarized in INTRODUCTION), we rotated the neural tube around its dorsoventral axis, leaving the notochord in its normal position, to test whether this could produce a corresponding inversion of paraxial mesoderm pattern. The neural tube was excised at the level of somites 16-21 in stage 17-20 chick embryos, rotated 180° and the embryos reincubated for 9-26 hours (Fig. 1). Prior to surgery, the ventral portion of the somites at the axial level of rotation had begun to disperse to form the sclerotome while the dermamyotome was morphologically distinct in the dorsal somite. After rotation of the neural tube and further development, however, the ventromedial margin of the somite was no longer mesenchymal, but rather was epithelial in character, resembling the dermamyotome that persisted dorsally (n=5; Fig. 2). This was particularly noteworthy since the sclerotome had already begun to disperse at the time of the operation. Proximity to the dorsal neural tube appears to be important for the induction of a second dermamyotome, because in one embryo in which the neural tube was oriented tangentially, the length of the ectopic dermamyotome that formed from each somite pair corresponded to the extent of the dorsal neural tube with which the somite was in contact (Fig. 2C). To examine whether the secondary dermamyotome could have developed from pieces of dorsal somite that may have adhered to the neural tube and been subsequently displaced ventrally, we fixed 6 embryos 7-10 hours after the rotation. No obvious clusters of cells could be detected adhering to the rotated neural tube, suggesting that the ectopic dermamyotome was not derived from dorsal somite. Dorsal neural tube ablations To test further the hypothesis that the dorsal neural tube induces dermamyotome formation, we ablated the dorsal portion of the neural tube, along with the narrow strip of ectoderm that adheres to it, in stage 12-14 embryos. Manipulated embryos were allowed to develop for an additional 1820 hours and subsequently analyzed for the presence of a dermamyotome at the level of ablation. When the dorsal neural tube was removed at the level of the unsegmented mesoderm (n=16), 87.5% of the embryos displayed either a complete absence (10 out of 16; 62.5%) or Fig. 2. Histological sections through embryos whose neural tubes (nt) were rotated and then fixed 26 hours later. (A,B) Phase contrast and fluorescence micrographs of an HNK-1-labeled embryo. HNK-1 immunoreactivity identifies early migratory neural crest cells. The notochord (n) was left in its normal position, although a small piece remains attached to the former ventral surface of the neural tube. A dermamyotome (dm) developed in a normal position. Also an ectopic dermamyotomelike epithelium developed from what would have been sclerotome (arrowheads). The vesicle attached to the dorsal neural tube is probably a remnant of ectoderm (e). (C) Fluorescence micrograph of a section taken through a region where the neural tube is tangential with respect to the embryonic dorsal-ventral axis. The size of the ectopic dermamyotome (limits marked by arrowheads) is proportional to the extent of dorsal neural tube in direct contact with the somite, suggesting that close contact is necessary for signaling to occur. s, sensory ganglion. Scale bar, 50 µm. Dorsal neural tube induces the dermamyotome 235 Fig. 3. Sections through 6 embryos whose dorsal neural tubes were ablated at the level of the segmental plate. The degree to which dermamyotome formation was affected is variable. (A,B) In most cases, there was a complete absence of the dermamyotome. Note that there are no neural crest cells in these sections (as assayed by HNK-1 immunoreactivity), indicating that the dorsal neural tube has been removed and has not regenerated. (C,D) In a smaller number of instances, remnants of the dermamyotome develop (arrowheads). Generally these are found in lateral positions in the embryo, as seen in C. Rarely, small epithelial vesicles are observed (D). (E) Although the dorsal portion of the neural tube has been successfully ablated in this embryo, as evidenced by the lack of neural crest cells, normal dermamyotomes are found. Possibly enough neural tube with inducing ability remains to stimulate dermamyotome development. (F) In this instance, where the entire neural tube was removed, one large dermamyotome forms dorsally. nt, neural tube; n, notochord; dm, dermamyotome. Scale bar, 50 µm. a significant loss (4 out of 16; 25%) of the dermamyotome at the level of ablation (Fig. 3A-D). The remaining embryos (2 out of 16; 12.5%) revealed no apparent effect on dermamyotome morphology (Fig. 3E). These results suggest that the dorsal neural tube, or perhaps the narrow strip of ectoderm covering it, specifies dorsal somite structures. The appearance of part or all of the dermamyotome in some cases may be due to the variability in the amount of neural epithelium that was removed from each embryo. Alternatively, several studies show that the lateral dermamyotome develops independently of the cues that regulate formation of the medial dermamyotome (Bober et al., 1994; Pourquié et al., 1995), which may explain why when remnants of dermamyotome are found, they are positioned laterally (e.g. Fig. 2C). Curiously, when the entire neural tube was removed in one individual, leaving an intact notochord in place, a fused somite with a dorsally positioned dermamyotome resulted (Fig. 3F). The paraxial mesoderm apparently responds to a dorsalizing cue associated with the dorsal neural tube quite early and irreversibly, because when the dorsal neural tube is ablated at progressively more anterior levels (i.e. embryologically older regions), dermamyotome formation is less affected. Specifically, when the dorsal neural tube is ablated at the more anterior level of somites −1 to −3 (n=7), only 43% (3 out of 7) of the embryos exhibit a loss of the dermamyotome. There is no loss of the dermamyotome when the neural tube is removed from the axial level of somites −4 to −6 (n=5). Embryonic tissues grafted medial to the ventral somite The previous two experiments demonstrated a role for dorsal axial tissues in the induction of the dermamyotome. However, these could not rule out the possibility that dermamyotomeinducing signals are produced either by the medial ectoderm adhering to the neural tube or by neural crest cells emigrating from the dorsal neural tube. To test potential roles for these tissues and further investigate the influence of the neural tube, we used two additional experimental strategies. First, pieces of isolated embryonic tissues were grafted ventrally into the space between the notochord and paraxial mesoderm at the axial level where somites have not yet formed to determine what embryonic tissues are capable of stimulating ectopic dermamyotome formation. The presence of a dermamyotome was assayed by morphological criteria. We also used an antibody against the intermediate filament protein desmin in order to immunocytochemically detect myocytes. Out of 37 embryos in which pieces of dorsal neural tube were grafted ventrally (Fig. 4A,B), 49% displayed either the presence of a morphologically distinct dermamyotome (6 out of 37) or desmin-positive cells proximal to the graft (12 out of 37). In the remainder of the embryos that showed no obvious ectopic dermamyotome, the graft was quite distant from the somite, suggesting that the grafted tissue must be close to the mesoderm to exert an effect. Pieces of lateral neural tube also induced either an organized epithelium in the ventral somite or desmin-positive cells (Fig. 4C,D), although the response was less robust. 236 M. S. Spence, J. Yip and C. A. Erickson Fig. 4. Pieces of isolated embryonic tissues were labeled with fluoresceinconjugated lectin (green labeling), grafted between the notochord and somite, and the presence of myocytes assessed by immunoreactivity with an antibody to desmin (red labeling). (A,C,E) Fluorescence images; (B,D,F) corresponding phase images. (A,B) A piece of dorsal neural tube was grafted ventrally and the embryo fixed 24 hours later. Myocytes are distributed ventrally (arrowheads) proximate to the grafted tissue. In a phase micrograph of the same section (B), note also the epithelial arrangement of the somitic cells (indicated by arrowheads) associated with the grafted piece of dorsal neural tube. Normal myotome structure and position is observed on the contralateral side. (C,D) A piece of lateral neural tube was grafted between the notochord and somite. A few myocytes arise in contact with the graft (arrowheads), but these cells are not organized into an epithelium. (E,F) A piece of ventral neural tube was grafted into a ventral position but, although it is in intimate contact with the somite, no myocytes have differentiated. Two small areas of rhodamine fluorescence (arrowheads) dorsal to the graft are blood vessels. Scale bar, 50 µm. In contrast, when ventral neural tube pieces were grafted in an identical fashion (n=14), the ventral somite cells were not organized in an epithelium in any of the embryos, and only 7% of the embryos (1 out of 14) displayed desmin-positive cells next to the graft (Fig. 4E,F). This difference between the dorsal and ventral grafts is statistically significant (Chi-square analysis, P=0.05). When pieces of ectoderm were grafted adjacent to the notochord (n=10), 90% of these embryos displayed either a vague epithelial arrangement of somite cells (5 out of 10; Fig. 5A) or a few desmin-positive cells (4 out of 10). The response was not nearly as dramatic as that generated by dorsal neural tube pieces, however (Fig. 5B). Finally, we tested the role of neural crest cells in dermamyotome formation by grafting clusters of neural crest cells isolated from 24-hour-old cultures between the somite and notochord. In 12 embryos, none showed any evidence of dermamyotome formation proximate to the graft (data not shown). Co-culture studies To further distinguish between the relative roles of the dorsal, ventral and lateral neural tube in the induction of myogenesis, we co-cultured portions of the neural tubes with paraxial mesoderm in vitro. Such studies allowed us to better control the size of tissue pieces used and quantitate the inductive response. Dissected neural tubes approximately 300 µm in length were taken from the axial level of the segmental plate, and cut longitudinally into 3/4 dorsal and 1/4 ventral pieces, or alternatively, into 3/4 ventral and 1/4 dorsal pieces (Fig. 6A). These neural tube pieces were then cultured individually with a piece of segmental plate (approximately 300 µm in length). The extent of myogenesis was assessed by quantifying the number of 13F4-positive cells per culture. Because the density of overlapping cells precluded accurate counting, we categorized the cultures as: <10 13F4-positive cells; 10≥100 cells; and >100 cells (Fig. 7). Control cultures of paraxial mesoderm alone produced no myocytes (n=29 cultures). The results of coculture treatments are summarized in Fig. 6B. The greatest number of 13F4-immunoreactive cells was found in the 1/4 dorsal neural tube cultures (n=39 cultures; over 90% of all cultures contained more than 10 myocytes). Fewer immunoreactive cells were found in cultures that contained either 3/4 dorsal or 3/4 ventral neural tube pieces. The smallest number of immunoreactive cultures resulted from 1/4 ventral neural tube explants (n=56 cultures; 55% of these Dorsal neural tube induces the dermamyotome 237 Fig. 5. (A) When pieces of ectoderm (e) are grafted ventrally, cells that should have been mesenchymal, as on the contralateral side, are organized into a thin epithelium (indicated by arrowheads). This is not as robust a response as seen when pieces of dorsal neural tube (dnt) are similarly grafted (B). Note in B the epithelial organization of the ventral somite (arrowheads) despite its proximity to the notochord (n). The grafted tissue is not acting as a physical barrier between the somite and notochord and thereby blocking ventralizing signals. nt, neural tube. Scale bar, 50 µm. contained no myocytes), demonstrating that the dorsal neural tube has the greatest ability to induce myocyte formation under these culture conditions. Neural crest cells derived from 24-hour-old cultures and cocultured with paraxial mesoderm (n=11) resulted in no myocytes in 55% (6 out of 11) of the cultures, or a moderate level (10≥100 myocytes) of myogenesis in 45% (5 out of 11) of the cultures. We found it difficult to establish cultures of dorsal ectoderm on planar substrata and therefore could not assess their ability to induce myogenesis using this assay. However, we repeated some of these experiments by culturing tissues in collagen gels, where adhesion to the substratum was not essential and where the tissues could be forced to stay together. Under these conditions, two or three pieces of ectoderm co-cultured with a piece of unsegmented mesoderm (n=14) produced no response in 43% (6 out of 14), a minimal response in 43% (6 out of 14), and a substantial response in 14% (2 out of 14) of the cultures. A Co-culture of dorsal neural tube and paraxial mesoderm in collagen gels still yielded a dramatic response (>100 myocytes/culture) in all cultures (n=4), as expected from the planar culture assay. We also assessed the relative ability of progressively older neural tubes to induce myogenesis (Fig. 8). Neural tubes from 4 different axial levels – unsegmented mesoderm (n=11), somite level −1 to −3 (n=11), −4 to −6 (n=10), and −7 to −9 (n=12) – were cut in half longitudinally and the dorsal halves co-cultured with segmental plate. This experiment did not reveal statistically significant differences in the inductive effects of dorsal neural tube from different axial levels. Stern and Hauschka (1995) also reported that rostral and caudal neural tubes are equivalent in their capacity to promote myogenesis. This result is not surprising in light of the neural tube rotation experiments described earlier, in which all levels of the rotated piece could induce a secondary dermamyotome. Our culture system also allowed us to evaluate at what axial B Neural Tube Pieces + SP 3/4 D 1/4 V + + Percentage of Cultures 75 < 10 myocytes 10 ≥ 100 myocytes > 100 myocytes 31 32 50 27 20 19 18 3/4 V 11 14 25 7 1/4 D 24 6 4 + + 0 1/4 V 3/4 V 3/4 D 1/4 D Neural Tube Pieces Fig. 6. (A) Diagrammatic summary of co-culture experiments in which paraxial mesoderm derived from the axial level of the segmental plate (SP; indicated by arrow) was combined with dorsal or ventral pieces of neural tube for 4 days and the differentiation of myocytes assessed by immunoreactivity with the 13F4 monoclonal antibody, which is a marker for myocytes. Cultures were characterized as: containing <10 myocytes; 10≥100 myocytes or >100 myocytes. (B) Percentages of cultures in each category are summarized in the bar graph. Total number of cultures for each category are indicated at the top of each bar. 238 M. S. Spence, J. Yip and C. A. Erickson Fig. 7. Examples of cultures immunolabeled with the 13F4 antibody, which reveals myocytes, and rated as >100 myocytes (A), 10≥100 myocytes (B) and <10 myocytes (C). Note in A how far the myocytes are distributed from the piece of dorsal neural tube. i, putative inducer. Scale bar, 100 µm. level myocytes were specified. As noted above, unsegmented mesoderm cultured alone produced no myocytes, as measured by 13F4 immunoreactivity (n=29). In contrast, somites −1 to −3 cultured alone (n=15) produced myocytes in 60% of the cultures and somites −4 to −6 (n=10) and somites −7 to −9 (n=12) produced myocytes 100% of the time. These results are identical to other studies that analyze myogenesis in stage 13 embryos using culture assays (Stern and Hauschka, 1995) or in vivo experimental manipulations (Aoyama and Asamoto, 1988). Thus, myocyte specification occurs within a few hours of somite segmentation. DISCUSSION Our results, derived from both in vivo and in vitro analysis, suggest that factors from the dorsal neural tube induce dermamyotome formation and myocyte differentiation, that these factors are diffusible over a short range and compete with ventralizing factors, and that specification of dorsal fate occurs shortly after somite segmentation. Dorsal neural tube is an inducer Many different embryonic tissues have been proposed to specify dorsal somite structures, but our results indicate that the dorsal neural tube is the most potent of these. When the neural tube is rotated so that the dorsal surface abuts the ventral somite, sclerotome dispersion is inhibited locally and an epithelial structure resembling the dermamyotome forms. Also, when the dorsal neural tube is ablated at a posterior axial level, no dermamyotome develops in the majority of cases. Finally, when pieces of dorsal neural tube are co-cultured with paraxial mesoderm either in tissue culture or in the embryo, an ectopic dermamyotome forms and myocytes differentiate from the unspecified mesoderm. Since no dermamyotome forms or muscle cells differentiate in the absence of the dorsal neural tube in all our experimental conditions, our results further suggest that a dorsalizing signal is required to specify the dermamyotome, rather than dorsal structures simply arising by default. Although we demonstrate that the dorsal neural tube can induce dermamyotome formation and myocyte differentiation, and is also the most potent inducer in all of our assays, we have not ruled out the possibility that redundant or overlapping cues emanate from other tissues. For example, ectoderm has been proposed to signal dermamyotome formation directly (Christ et al., 1972; Fan and TessierLavigne, 1994; Kuratani et al., 1994). In support of a role for ectoderm is the observation that in the absence of the entire neural tube, but in the presence of a notochord, the somites fuse in the midline and one large overarching dermamyotome develops (our results; Christ et al., 1972; Bober et al., 1994). Thus, in the absence of dorsal neural tube cues, a normal dermamyotome develops in association with the ectoderm and suggests that redundant cues are produced by ectoderm and neural tube. If the ectoderm does produce such a cue, it is difficult to explain why dermamyotomes do not develop when the dorsal neural tube is ablated but the ectoderm over the somites remains intact. Moreover when pieces of ectoderm are grafted ventrally in the embryo adjacent to the ventral somite, only a very minimal response is observed (see also Kenny-Mobbs and Thorogood, 1987; Fan and TessierLavigne, 1994) compared with the dramatic response to the dorsal neural tube. Given that a putative inducing factor produced by the dorsal neural tube appears to be diffusible (Fan and Tessier-Lavigne, 1994; see below), another explanation for the weak inductive activity displayed by the ectoderm is that such a factor could diffuse from the dorsal neural tube and bind to the extracellular matrix associated with the basal lamina of the ectoderm. Another possibility that our study cannot rule out is that the dorsal neural tube may signal the overlying ectoderm to acquire dermamyotome-inducing ability. There is precedent for such a mechanism being involved in ventral somite specification, since Sonic hedgehog produced by the notochord induces formation of the neural tube floor plate, which in turn produces more Sonic hedgehog and is apparently responsible for both stimulating sclerotome formation as well as repressing dermamyotome formation (Johnson et al., 1994; Fan and Tessier-Lavigne, 1994). Thus when the dorsal neural tube was ablated in our studies, it is conceivable that we may have interfered with signaling to the ectoderm. Nevertheless, one additional observation argues against this possibility. Bober et al. (1994) report that, when one lateral half of a neural tube is removed, the medial portion of the dermamyotome on the depleted side does not form, even though the other half of the neural tube is still intact, including its dorsal component. The remaining neural tube could presumably still signal the overlying ectoderm just as effectively. Thus we suggest that if Dorsal neural tube induces the dermamyotome 239 Dorsal neural tube from different axial levels co-cultured with segmental plate (SP) DNT SP DNT -7 to -9 75% positive (n=12) 40% positive (n=10) SP -4 to -6 -1 to -3 DNT 64% positive (n=11) Segmental Plate SP DNT SP the ectoderm plays a role in dermamyotome development, it is likely to be redundant. Neural crest cells have also been proposed to induce the formation of the myotome as well as stimulate myogenesis, but we view this possibility to be unlikely. First, neural crest cells grafted ventrally adjacent to the sclerotome fail to induce dermamyotome formation. Similarly in co-culture studies, few or no myocytes arise in the presence of pure populations of neural crest cells. The moderate number of myocytes observed in some cultures can be explained by the fact that purified crest cells must be harvested from neural tube cultures, which may contaminate them with a dorsal neural tube-derived inducing factor. Finally, myotome formation occurs concomitant with the invasion of neural crest cells into the somite at the thoracic level (e.g. Loring and Erickson, 1987; Tosney et al., 1994) but, at more posterior levels, crest cells invade the somite many hours prior to the appearance of the myotome (our unpublished results). Thus, there is no strict correlation between contact with neural crest cells and the appearance of the myotome. There are several reports suggesting that the notochord may induce dermamyotome formation and muscle cell differentiation. First, as already discussed, when the entire neural tube is ablated but the notochord in left in place (Christ et al.,1972; Bober et al., 1994; and this study), a single large dermamyotome spans the space where the neural tube would be normally. Second, the notochord apparently has a myocyteinductive capability in the absence of the neural tube in the embryo. For example, Rong and co-workers (1992) removed the neural tube and notochord along the entire embryonic axis and then replaced these axial structures with 2 or 3 exogenous notochords. Under these circumstances muscle cells differentiate. Finally, when paraxial mesoderm is co-cultured with notochord, muscle cells differentiate (Buffinger and Stockdale, 1994; Stern and Hauschka, 1995; Münsterberg and Lassar, 1995; our unreported results). It is difficult to reconcile these observations with the fact that grafting a notochord into the 72% positive (n=11) Fig. 8. Summary of co-culture experiments in which the myocyte-inducing ability of the dorsal neural tube (DNT) from increasingly anterior axial levels was assessed. The dorsal half of the neural tube was isolated from four different axial levels and then combined with pieces of paraxial mesoderm isolated from the level of the segmental plate (SP). There is no significant difference in the inducing ability of the neural tube as it ages (i.e. from the more anterior axial levels). dorsal portion of a somite represses dermamyotome formation (Pourquié et al., 1993; Goulding et al., 1994) and that notochord-derived Sonic hedgehog competes with the dorsal signal to turn off dermamyotome-specific gene expression (Johnson et al., 1994; Fan and Tessier-Lavigne, 1994). In addition, when the notochord is removed, resulting in the dorsalization of the ventral neural tube (van Straaten and Hekking, 1991; Goulding et al., 1993), dermatome-specific gene expression and myotome formation is extended ventrally, again supporting the model that the notochord and ventral neural tube suppress the development of the dermamyotome. A definitive explanation of these conflicting results must await the identification of the inducing factors. However, one possibility is that the neural tube reciprocally signals the notochord and that, in the absence of neural tube influence (as would be the case in Rong et al., 1992; Buffinger and Stockdale, 1994; Stern and Hauschka, 1995), the notochord no longer produces factors that repress dermamyotome formation, thereby revealing myogenic factors that it also generates. The observation that ectopically expressed Sonic hedgehog alone in the dorsal somite results in the expansion of the myotome rather than its repression is evidence of a notochord-derived myogenic inducer (Johnson et al., 1994). The dorsal neural tube product is diffusible It is likely that the inducing signals derived from the dorsal neural tube are diffusible, although the distance over which they can diffuse is difficult to discern from our current observations. We noted from our co-cultures on planar substrata that the dorsal neural tube had to be placed in close proximity to the mesoderm (but not necessarily touching) to see strong induction but, in this assay, we were not able to quantitate precisely this distance. Stern and Hauschka (1995) also noted the need for close positioning. Under our culture conditions, a diffusible factor might fall to very low levels a short distance from the dorsal neural tube. In fact, co-cultures embedded in 240 M. S. Spence, J. Yip and C. A. Erickson collagen gels (Fan and Tessier-Lavigne, 1994), where the diffusing molecule may be concentrated or even bound to the gel matrix, suggest effects of over 100 µm. We also noted that, although the inducing and responding tissues needed to be close, myocytes appeared widely throughout the paraxial mesoderm (>100 µm from the inducing source) and not just in the contact zone (see Fig. 8A), suggesting that this factor can diffuse through and perhaps be concentrated in the tissue environment of the paraxial mesoderm. We did not rule out directly that an inducing signal in our system is propagated from one myocyte to the next. Such propagation seems unlikely, however, because we might have observed similar numbers of myocytes arising in our cultures of somites at different ages but, instead, we found that greater numbers of myocytes differentiated from progressively older somites. Our in vivo studies also cannot reveal how far such a factor may diffuse since tests of induction were always done in competition with the host notochord or floorplate, which complicates the interpretation. Nevertheless myocytes often appeared >50 µm from the closest edge of the dorsal neural tube fragments. Any further tests of how far such an inducer could act must clearly be carried out in the embryo since rates and distances of diffusion will depend upon the ability to move through and bind to the extracellular matrix or through other tissues. Our studies, as well as those of others, suggest that dorsal neural tube-derived factors can compete with ventralizing signals. Fan and Tessier-Lavigne (1994) elegantly demonstrated such competition when they co-cultured dorsal neural tube and Sonic hedgehog-producing heterologous cells with unspecified mouse presomitic mesoderm. Similarly in our study, when 3/4 dorsal neural tubes, which are likely to contain ventralizing signals, are cultured with paraxial mesoderm, fewer myocytes differentiate compared to 1/4 dorsal neural tubes, again suggesting antagonizing signals. We have been able to demonstrate directly that such a competition occurs in the embryo as well. In two different experimental paradigms – either when the neural tube was rotated 180° around its dorsalventral axis or when pieces of dorsal neural tube were grafted ventrally – host notochord was always left in its normal position. Under these circumstances, dorsal structures (either a dermamyotome or myocytes) differentiated in response to a dorsalizing cue despite their proximity to the host notochord. This competition is not due simply to the grafted tissue physically blocking a ventralizing signal from the notochord, since the dorsalizing effect of the grafted dorsal neural tube was noted even when the notochord had unimpeded simultaneous contact with the somite (Fig. 5B). Timing of specification Several recent studies where progressively older somites were placed in culture suggest that, within only a few hours after separation from the segmental plate, the myocytes are already determined and will later differentiate, even in the absence of the neural tube (Vivarelli and Cossu, 1986; Kenny-Mobbs and Thorogood, 1987; Buffinger and Stockdale, 1994; Stern and Hauschka, 1995). Our tissue culture data are consistent with these observations; although myocytes do not differentiate from unsegmented paraxial mesoderm, older somites give rise to progressively more myocytes. The dorsal neural tube ablation studies extended these observations to the embryonic environment. A complete dermamyotome rarely forms when the dorsal neural tube is ablated at the level of the segmental plate, suggesting that specification has not yet occurred. In contrast, dorsal neural tube ablation at the level of the three last-formed somites results in dermamyotome development in 50% of the cases, while somites −4 to −6 apparently no longer need the influence of the dorsal neural tube since ablation has no effect on dermamyotome differentiation. Aoyama and Asamoto (1988) also showed that myocytes are specified by somite −3 in stage 13 embryos by rotating a somite around its dorsoventral axis, which then produced an inverted dorsoventral pattern. Interestingly, the somites remain responsive to signaling from the neural tube after segmentation (see also Stern and Hauschka, 1995) and the dorsal neural tube itself retains inducing properties at more anterior axial levels, presumably many hours after they are needed. Such precision in identifying the timing of specification should be useful in elucidating the molecular basis for the various steps in the specification, determination and differentiation of myocytes and dermal precursors. Model for dorsal-ventral pattern Our data, along with a wide array of previously published studies, suggest the following model for the establishment of dorsal-ventral pattern in the somite. Competing diffusible signals from the dorsal neural tube and notochord/floorplate specify dermamyotome and sclerotome, respectively. The dorsal neural tube signal is proposed to diffuse broadly but in ventral regions is out-competed by Sonic hedgehog and perhaps other ventral signals. Thus, the spatial positioning of the dermamyotome dorsally is due to the absence or low concentration of notochord-derived signals, as well as to the presence of dorsal signals. It is possible that these latter signals are localized and amplified by binding to the ectoderm or that the ectoderm also produces dorsalizing factors. An alternative possibility is that the dorsal neural tube signals only the medial edge of the somite, and this dermamyotomespecifying signal is propagated laterally in the plane of the somitic epithelium. This signal, if propagated ventrally, would be inhibited by ventralizing signals from the floorplate and notochord. Identification of the dorsal signaling molecules awaits further experimentation, but particularly attractive candidates are members of the Wnt family, which are diffusible signaling factors produced in the neural tube that act over both a short and long range (Parr et al., 1993). Finally, it is likely that several dorsalizing signals are necessary, including those that maintain the epithelial structure of the dermamyotome, as well as additional factors that may dictate specific cell fate. We thank Charles Ordahl and David McClay for insightful comments concerning somitogenesis and signaling mechanisms and Mark Reedy, David Parichy and Tuan Duong for careful scrutiny of the manuscript. Two anonymous reviewers contributed substantially to the tightened focus and balance of the Discussion, for which we thank them. This research was supported by NIH grants DE 05630 to C. A. E. and NS 23916 to J. Y. REFERENCES Aoyama, H. and Asamoto, K. (1988). Determination of somite cells: Dorsal neural tube induces the dermamyotome independence of cell differentiation and morphogenesis. Development 104, 15-28. Basler, K., Edlund, E., Jessell, T. M. and Yamada, T. (1993). Control of cell pattern in the neural tube: Regulation of cell differentiation by dorsalin-1, a novel TGFß family member. Cell 73, 687-702. Bellairs, R. (1963). The development of the somites in the chick embryo. J. Embryol. Exp. Morph. 11, 697-714. Bober, E., Brand-Saberi, B., Ebensperger, C., Wilting, J., Balling, R., Paterson, B. M., Arnold, H.-H. and Christ, B. (1994). Initial steps of myogenesis in somites are independent of influence from axial structures. Development 120, 3073-3082. Borman, W. H. and Yorde, D. E. (1994). Barrier inhibition of a temporal neuraxial influence on early chick somitic myogenesis. Dev. Dynamics 200, 68-78. Buffinger, N. and Stockdale, F. E. (1994). Myogenic specification in somites: induction by axial structures. Development 120, 1443-1452. Christ, B., Brand-Saberi, B., Grim, M. and Wilting, J. (1992). Local signaling in dermomyotomal cell type specification. Anat. Embryol. 186, 505-510. Christ, B., Jacob, H. J. and Jacob, M. (1972). Experimentelle Untersuchungen zur Somitenentstehung beim Hühnerembryo. Z. Ant. Entwickl. Gesch. 138, 82-97. Erickson, C. A. and Goins, T. L. (1995). Avian neural crest cells can migrate in the dorsolateral path only if they are specified as melanocytes. Development 121, 915-924. Erickson, C. A., Duong, T. D. and Tosney, K. W. (1992). Descriptive and experimental analysis of the dispersion of neural crest cells along the dorsolateral path and their entry into ectoderm in the chick embryo. Dev. Biol. 151, 251-272. Erickson, C. A., Tucker, R. P. and Edwards, B. F. (1987). Changes in the distribution of intermediate-filament types in Japanese quail during morphogenesis. Differentiation 34, 88-97. Fan, C.-M. and Tessier-Lavigne, M. (1994). Patterning of mammalian somites by surface ectoderm and notochord: evidence for sclerotome induction by a hedgehog homolog. Cell 79, 1175-1186. Goulding, M. D., Lumsden, A. and Gruss, P. (1993). Signals from the notochord and floor plate regulate the region-specific expression of two Pax genes in the developing spinal cord. Development 117, 1001-1016. Goulding, M., Lumsden, A. and Paquette, A. J. (1994). Regulation of Pax-3 expression in the dermamyotome and its role in muscle development. Development 120, 957-971. Hay, E. D. (1968). Organization and fine structure of epithelium and mesenchyme in the developing chick embryo. In Epithelial-Mesenchymal Interactions (eds. R. Fleischmajer and R. E. Billingham). pp 31-55. Baltimore: Williams and Wilkins. Johnson, R. L., Laufer, E., Riddle, R. D. and Tabin, C. (1994). Ectopic expression of Sonic hedgehog alters dorsal-ventral patterning of somites. Cell 79, 1165-1173. Kaehn, K., Jacob, H. J., Christ, B., Hinrichsen, K. and Poelmann, R. E. (1988). The onset of myotome formation in the chick. Anat. Embryol. 177, 191-201. Kenny-Mobbs, T. and Thorogood, P. (1987). Autonomy of differentiation in avian brachial somites and the influence of adjacent tissues. Development 100, 449-462. Kosher, R. A. and Lash, J. W. (1975). Notochordal stimulation of in vitro somite chondrogenesis before and after enzymatic removal of perinotochordal materials. Dev. Biol. 42, 362-378. Kuratani, S., Martin, J. F., Wawersik, S., Lilly, B., Eichele, G. and Olson, E. N. (1994). The expression pattern of the chick homeobox gene gMHox suggests a role in patterning of the limbs and face and in compartmentalization of somites. Dev. Biol. 161, 357-369. 241 Lash, J. W. (1968). Somitic mesenchyme and its response to cartilage induction. In Epithelial-Mesenchymal Interactions (eds. R. Fleischmajer and R. E. Billingham). pp. 165-172. Baltimore: Williams and Wilkins. Lash, J., Holtzer, S. and Holtzer, H. (1957). An experimental analysis of the development of the spinal column. Exp. Cell Res. 13, 292-203. Loring, J. F. and Erickson, C. A. (1987). Neural crest cell migratory pathways in the trunk of the chick embryo. Dev. Biol. 121, 220-236. Loring, J., Glimelius, R., Erickson, C. and Weston, J. A. (1981). Analysis of developmentally homogeneous neural crest cell populations in vitro. I. Formation, morphology and differentiative behavior. Dev. Biol. 82, 86-94. Mestres, P. von and Hinrichsen, K. (1976). Zur Histogenese des Somiten beim Hühnchen. J. Embryol. Exp. Morph. 36, 669-683. Münsterberg, A.E. and Lassar, A.B. (1995). Combinatorial signals from the neural tube, floor plate and notochord induce myogenic bHLH gene expression in the somite. Development 121, 651-660. Ordahl, C. P. (1993). Myogenic lineages within the developing somite. In Molecular Basis of Morphogenesis, 51st Annual Symposium of the Society of Developmental Biology (ed. M. Bernfield). pp. 165-176. New York: Wiley-Liss. Ordahl, C. P. and Le Douarin, N. M. (1992). Two myogenic lineages within the developing somite. Development 114, 339-353. Parr, B. A., Shea, M. J., Vassileva, G. and McMahon, A. P. (1993). Mouse Wnt genes exhibit discrete domains of expression in the early embryonic CNS and limb buds. Development 119, 247-261. Pourquié, O., Coltey, M., Teillet, M.-A., Ordahl, C. and Le Douarin, N. M. (1993). Control of dorsoventral patterning of somitic derivatives by notochord and floor plate. Proc. Natl. Acad. Sci. USA 90, 5242-5246. Pourquié, O., Coltey, M., Bréant, C. and Le Douarin, N. M. (1995). Control of somite patterning by signals from the lateral plate. Proc. Natl. Acad. Sci. USA 92, 3219-3223. Rong, P.-M., Teillet, M.-A., Ziller, C. and Le Douarin, N. M. (1992). The neural tube/notochord complex is necessary for vertebral but not limb and body wall striated muscle differentiation. Development 115, 657-672. Rong, P.-M., Ziller, C. Pena-Melion, A. and Le Douarin, N. M. (1987). A monoclonal antibody specific for avian early myogenic cells and differentiated muscle. Dev. Biol. 122, 338-353. Stern, C. D., Fraser, S. E., Keynes, R. J. and Primmett, D. R. N. (1988). A cell lineage analysis of segmentation in the chick embryo. Development 104 Supplement 231-244. Stern, H. M. and Hauschka, S. D. (1995). Neural tube and notochord promote in vitro myogenesis in single somite explants. Dev. Biol. 167, 87-103. Tosney, K. W., Dehnbostel, D. B. and Erickson, C. A. (1994). Neural crest cells prefer the myotome’s basal lamina over the sclerotome as a substratum. Dev. Biol. 163, 389-406. Tucker, R. P. and Erickson, C. A. (1984). Morphology and behavior of quail neural crest cells in artificial three-dimensional extracellular matrices. Dev. Biol. 104, 390-405. van Straaten, H. W. M. and Hekking, J. W. M. (1991). Development of floor plate, neurons and axonal outgrowth pattern in the early spinal cord of the notochord-deficient chick embryo. Anat. Embryol. 184, 55-63. Vivarelli, E. and Cossu, G. (1986). Neural control of early myogenic differentiation in cultures of mouse somites. Dev. Biol. 117, 319-325. Watterson, R. L., Fowler, I. and Fowler, B. J. (1954). The role of the neural tube and notochord in development of the axial skeleton of the chick embryo. Am. J. Anat. 95, 337-397. Yip, J. W. (1990). Identification of location and timing of guidance cues in sympathetic preganglionic axons of the chick. J. Neurosci. 10, 2476-2484. (Accepted 27 September 1995)