DEVELOPMENTAL DYNAMICS 207:30%318 (1996) Neural Tube Closure in the Chick Embryo Is Multiphasic HENNY W.M. VAN STRAATEN, HENNIE C.J.P. JANSSEN, MARIAN C.E. PEETERS, ANDREW J. COPP, AND JOWAN W.M. HEKKING Department AnatornylErnbryology, University of Limburg, Maastricht, Netherlands (H.W.M.V.S.,H.C.J.P.J., M.CE P., J.W.M.H.);Division of Cell and Molecular Biology, Institute of Child Health, University of London, London W C l N LEH, United Kingdom (A.J.C.) ABSTRACT Progression of neurulation in the chick embryo has not been well documented. To provide a detailed description, chick embryos were stained in ovo after the least manipulation possible to avoid distortion of the neural plate and folds. This allowed a morphological and morphometric description of the process of neurulation in relatively undisturbed chick embryos. Neurulation comprises several specific phases with distinct closure patterns and closure rates. The first closure event occurs, de novo, in the future mesencephalon at the 4-6 somite stage (sst 4-6). Soon afterwards, at sst 8-7, de novo closure is seen at the rhombocervical level in the form of multisite contacts of the neural folds. These contacts occur in register with the somites, suggesting that the somites may play a role in forcing elevation and apposition of the neural folds. The mesencephalic and rhombocervical closure events define an intervening rhombencephalic neuropore, which is present for a brief period before it closes. The remaining pear-shaped posterior neuropore (PNP) narrows and displaces caudally, but its length remains constant in embryos with seven to ten somites, indicating that the caudal extension of the rhombocervical closure point and elongation of the caudal neural plate are keeping pace with each other. From sst 10 onward, the tapered cranial portion of the PNP closes fast in a zipper-like manner, and, subsequently, the wide caudal portion of the PNP closes rapidly as a result of the parallel alignment of its folds, with numerous button-like temporary contact points. A role for convergent extension in this closure event is suggested. The final remnant of the PNP closes at sst 18. Thus, as in mammals, chick neurulation involves multisite closure and probably results from several different development mechanisms at varying levels of the body axis. 0 1996 WtIey-Lisa, Ine. Key words: Chick, Embryo,Neurulation, Closure rate, Neuropore, Morphology, Morphometry INTRODUCTION Neurulation, the process of neural tube formation, involves the concerted action of a variety of intrinsic 0 1996 WILEY-LISS, INC. and extrinsic forces, resulting in the complex shaping and folding of the neural plate and closure of the neural tube. In view of the complexity of the process, it is not surprising that the underlying mechanisms are only partially understood. Work with amphibian and chick embryos has defined the principal morphogenetic events that comprise neurulation (for review, see Karfunkel, 1974; Schoenwolf, 1982, 1994; Gordon, 1985; Copp et al., 1990; Schoenwolf and Smith, 1990; Jacobson, 1991). The neural plate extends longitudinally, narrows transversely, and thickens apicobasally during the process of shaping. This elongation appears to result mainly from cell rearrangement and directed cell proliferation. Simultaneously, the lateral borders of the neural plate elevate, and the arising neural folds converge to the midline. The mechanisms of elevation and convergence have been suggested to include apical constriction of neuroepithelial cells, the formation of medial and dorsolateral hinge points, and extrinsic factors, such as expansion of underlying mesoderm and the medial extension of the surface ectoderm. Finally, the neural folds appose, adhere, and fuse in the midline, processes that are very poorly understood in mechanistic terms. Relatively little attention has been paid to the sequential events of neural tube closure along the body axis in the chick embryo. Closure has been described as proceeding in an orderly sequence from the initial closure site in both rostra1 and caudal directions, in a zipper-like manner (Portch and Barson, 1974; Bancroft and Bellairs, 1975; Silver and Kerns, 1978; Schoenwolf, 1982, 1985). However, a recent detailed study using the high-definition microscope has identified closure occurring as separate de novo events a t mesencephalic and rhombencephalic levels (Jaskoll et al., 1991). Indeed, multisite closure was suggested to occur in the chick in an analogous manner to that described in mammals, in which up to five different closure sites have been identified (for review, see Golden and Chernoff, 1993). High-magnification observations by Jaskoll et al. (1991) suggest that neural tube closure may be a more subtle process than previously recognized. ARer initial Received March 11, 1996;accepted June 7, 1996. Address reprint requestdcorrespondence to H.W.M. Van Straaten, Department AnatomylEmbryology,University of Lirnburg, P.O.Box 616,6200 MD Maastricht, Netherlands. 310 VAN STRAATEN ET AL. Fig. 1. A: Visualization of a chick embryo with seven somites after staining with toluidine blue via the vitelline membrane. 6: The shape and size of the neuropores of the embryo in dorsal view following careful removal of the vitelline membrane. Structures that were in contact with the vitelline membrane (e.g., the neural folds) are stained most intensely, whereas deeper structures like floor plate area (f), the primitive streak (ps), and the contacts between the neural folds are not stained. Somites 1 and 7 are indicated by arrows. Bulgings of the neural folds or fold contacts are seen in register with the somite pairs. A rhombencephalic neuropore (RNP) is present; its neural folds and the adjacent ectodem are stained only lightly due to their deep location. Caudally, a pearshaped posterior neuropore (PNP) is present. Note that few morphological differences exist between the embryo before (A) and after (6) removal of the vitelline membrane. The length is reduced by 4%, and the RNP is slightly widened. Thus, the preparation method used is relatively free from attifacts. ANP: anterior neuropore; H: Hensen’s node. Scale bar = 250 pm. using scanning electron microscopy. Using the highdefinition microscope we recently found (Van Straaten et al., 1993b) that, during PNP closure, numerous button-like contacts arise, not only during fusion, but even as early as stages of apposition and adhesion of the neural folds. This finding suggests that a zipper-like model of neural tube closure may be an oversimplification. In order to shed more light on this morphogenetic process, the present study was undertaken to document the sequence of de novo closure points, the changes in shape, size, and position of the neuropores, and the presence and extension of the various types of neural fold contact. Care was taken to avoid manipulating the embryo during preparation because our previous study had shown that artefacts of PNP length and width and reopening of de novo adhesion sites can easily be introduced. The detailed description reveals that closure in the chick embryo is multiphasic and based on distinct closure patterns at different levels of the body axis. Fig. 1 (legend in facing column). contact between the mesencephalic folds, apposition occurred as an imperfect “zipping up,” in which several nonfused sites were interspersed with fused sites (Jaskoll et al., 1991).Bancroft and Bellairs (1975) were the first to mention such focal fusion sites, which they observed during closure of the posterior neuropore (PNP) RESULTS A total of 117 embryos between somite stages 4 and 18 (sst 4-18) were used to gather data on several morphological and morphometric parameters of neurulation in the chick. The embryos were studied in the dorsal view with the least possible manipulation (Fig. 1)in order to preserve the natural morphology and sequence of neurulation events. A representative series of embryos is shown in Figure 2, and the morphometric data on neuraxis elongation, neural fold contacts, and neuropore size are presented in Figures 3-6. The morphometric data, combined with the drawings of the changing shape of neuropores and neural folds, are summarized in a general schematic representation of neural tube closure (Fig. 7). This drawing reveals a multiphasic pattern of neurulation in the chick embryo. Mesencephalic Closure At the earliest stage studied (sst 4), the mesencephalic neural folds are in contact over a short distance. 311 MULTIPHASIC NEURAL TUBE CLOSURE length (pm) 7000 0 6500 0 6000 5500 5000 4500 4000 3500 0 0 3000 0 2500 2000 2 4 6 8 10 12 development (somites) 14 16 18 20 Fig. 3 Elongation of the neuraxis during neurulation in the chick embryo. Neuraxial length was measured between the anterior extremity of the embryo and the caudal extremity oi the PNP, as defined in Figure 8. The neuraxis elongates progressively, but its data are described better by a polynomial distribution than by a linear regression. The polynomial indicates a relatively enhanced rate of elongation between sornite stages (sst) 7 and 11 and delay between sst 11 and 14. Fig. 2 Changes in the shape of the neuropores during neurulation. Chick embryos were stained with toluidine blue via the vitelline membrane, which was removed subsequently for clarity of observation in this figure. The numbers of somites are indicated. Arrowheads indicate the caudal extremity of the neuraxis, as defined in Figure 8. A,B: Mesencephalic contact (rn) progresses, and the PNP is narrowed at the rhombocervical level (r). C,D The PNP is pear-shaped with tapered (t) and wide (w) portions. The PNP narrows progressively but retains its length. E: The tapered portion of the PNP is almost closed. F: The wide portion of the PNP is closing, and its neural folds are oriented in parallel. Scale bar = 400 pm. Anterior to this closure site is the anterior neuropore (ANP). Caudally, a large posterior neuropore (PW) is present, which is narrowed a t the level of the somites (Fig. 2A). The mesencephalic contact extends in both rostra1 and caudal directions and, a t this stage, comprises apposition and adhesion but not fusion of the neural folds (for definitions of these terms, see Experimental Procedures). Fusion apparently lags behind, because it could be detected for the first time only in sst 6 embryos (Fig. 4). The length of the ANP gradually decreases to zero as the prosencephalic neural tube is formed (Table 1, Fig. 5). The closure point of the PNP progresses in a caudal direction (Fig. 7) at a rate (relative to somite 1) of about 200 pdsomite stage between sst 4 and 6. Rhombocervical Closure Multisite de novo closures at the rhombocervical level occur a t sst 6-7. Preceding closure, the neural folds show local bulges, which are in register with the somites (Fig. 2B)and which result in a number of separate contact sites that are first visible at the level of somites 3-4(‘%utton-like”closure; Fig. 1B). In between, 312 VAN STRAATEN ET AL. length (pm) 7000 length (pm) 3500 ' I 0 r 6000 3000 I 5000 2500 - .~ . . oPNP I - PNP, wide portion 1 +RNP oANP ~ 0 4000 0 2000 0 0 3000 1500 2000 0 1000 O \ 1000 + 0 0 0 0 a a \&a 500 0 0 2000 2 4 6 8 10 12 14 16 18 20 development (somites) 1000 0 2 4 6 8 10 12 14 16 18 20 development (somites) Fig. 4 Length of neural fold contacts during neurulation in the chick embryo. Data from apposition and adhesion were taken together, because distinction between them was not always clear. When a specific contact was scattered throughout an embryo, the values were added. Regression lines are drawn for clarity only. The average length of apposition and adhesion shows a roughly biphasic pattern, with relatively high values at sst 7-8 and 13-14; at sst 9-10, both high and low values are seen. Fusion is seen at sst 6 for the first time, and the length of fused neural tube increases from that stage onward, although it lags behind that of apposition and adhesion by about three somite stages. A marked threefold increase in fusion occurs between sst 8 and 11. The rate of increase slows up to sst 14 but increases afterwards. At sst 18, the neural folds are in contact over their full length, but not all contacts are transformed into fusion yet. Each data point represents a single embryo, with the exception of the solid circles, which represent data combined from 7 embryos (at sst 4), from 11 embryos (at sst 5),and from 13 embryos (at sst 6). Fig. 5 Reduction in length of neuropores during neurulation in the chick embryo. Data for the three major neuropores, ANP, RNP, and PNP, are shown. The ANP and PNP are present from sst 4 onward. The ANP exhibits a gradual reduction of its length and no longer seen after sst 9. The RNP is present briefly between sst 6 and 8. The PNP is the most pronounced and long-lived neuropore. The length of the PNP during sst 6-7 also includes the lengths of the several small intermediate neuropores. The length of the wide portion of the PNP is plotted separately. Two major length reductions occur during development, at sst 6-7 and at sst 10-14, whereas, between sst 7 and 10, no reduction in PNP length is seen. Variation between embryos at a single somite stage is limited. This is especially evident from the close correlation between PNP length and sst during the rhombocervicalclosure and the closure between sst 10 and 14. Each data point represents a single embryo and is included in the regression lines (with the exception of the solid circles). has a rhomboid shape, and a pear-shaped PNP (Fig. lB), which has distinguishable tapered and wide portions (see Fig. 8). The RNP is transiently present for 1.5 somite stages on average, predominantly at sst 6-7 (Table 1, Fig. 5). The separation of the original PNP explains its dramatic length reduction from 2,500 to 1,500 pm between sst 6 and 7 (Fig.5). small neuropores arise temporarily. The rhombocervi- Narrowing and Caudal Shifting of the PNP During sst 7-10, the width of the PNP (measured at cal closure occurs fast: caudal progression of the PNP closure point now occurs a t a rate of 1,200 pdsomite its wide portion) shrinks steadily (Fig. 61,but its length stage between sst 6 and 7. This abrupt increase in con- remains constant, at about 800 pm for the wide portion tact is predominantly based on apposition and adhesion and about 700 pm for the tapered portion (Figs. 2C,D, (Fig. 4),whereas fusion again lags behind, because a 5). This means that the elongation rate of the neuraxis marked increase in fusion length (from 1,000 to 3,000 and the rate of caudal progression of the PNP closure pm) does not occur until sat 8-11 (Fig. 4). The rhomb- point both must be equal at about 300 pm/somite stage. ocervical closure results in separation of the original Because the rate of somite gain is only 140 pdsomite PNP into a rhombencephalic neuropore (RNP), which stage, the PNP as a whole is seen to be displaced in a MULTIPHASIC NEURAL TUBE CLOSURE width (pm) 700 600 1 I ! ca 0 0 0 $ 0 0 I 100 0 temporary button-like closure sites evident, mostly adhesion points. A small aperture remains. The caudal progression of the PNP closure point is rapid between sst 11 and 13, with a maximum rate of 480 pdsomite stage. This coincides with an increase in apposition and adhesion (between sst 10 and 141, whereas an increase of neural fold fusion occurs later, between sst 14 and 16 (Fig. 4). The rapid reduction in PNP length from sst 10 onward is shown dramatically in Figure 5, but the increase of closure rate (Fig. 7) during this period appears to be less dramatic: from 300 to 480 pdsec. This is due to a simultaneous decrease in the rate of neuraxial elongation (Fig. 3). I 4 500 200 313 -I b 1 1 O 0 & 0 O Q ) @ o 0 Closure of the PNP Remnant From sst 14 onward, the small remnant of the PNP further narrows and shortens, concluding in its closure. This opening appears to be localized mostly a t the caudal extremity of the PNP, because, from this aperture onward, reopening was successful in the cranial direction but failed in the caudal direction. Apposition and adhesion decrease, indicating that the apposed walls of the previously tapered and wide portions of the PNP are becoming fused. After sst 18, the PNP is invariably closed. I 2 4 6 8 10 12 14 16 18 20 development (somites) Fig. 6 Reduction in width of the PNP during neurulation in the chick embryo. The width was measured at the wide portion of the PNP. Note that reduction of width of the PNP, as measured in the dorsal view, results both from narrowing due to elevation and convergence and from reduction in width of the neural plate. A relatively uniform rate of width reduction is seen up to sst 12 followed by a slow rate of reduction until the PNP finally closes. DISCUSSION Possible Distortion of the Neural Plate During Embryo Preparation Descriptions of neurulation in the chick embryo have been based mostly on studies in which the embryo was manipulated to a greater or lesser extent, which could caudal direction as development progresses. The most have resulted in distortion of the neural plate (Van recently formed somite is passed by the PNP closure Straaten et al., 1993b). In the present study, we took point at sst 8 (Fig. 7). The reduced rate of progression great care to avoid manipulating the embryos in order of the PNP closure point (300 vs. 1,200 pdsomite stage to prevent artifactual distortion. It is possible that at sst 7-10 vs. sst 6-7) is reflected in a reduction of opening the egg and immersion of embryos in saline apposition and adhesion lengths until sst 11 (Fig. 4). can introduce artefacts, but our experience has shown Button-like adhesion and fusion points, as observed a t that even pushing on the vitelline membrane does not the rhombocervical closure, are present in register deform the embryo; therefore, we assume that distorwith the somites during caudal progression of the PNP tion is unlikely a t this stage of preparation. The viclosure point. telline membrane probably acts as a protective cover for the embryo against mechanical forces. On several Closure of the PNP occasions, removal of the vitelline membrane did cause A second major phase of reduction in PNP length distortion of the embryo, but these artefacts could be (from 1,500to 165 pm) occurs between sst 10 and 14 recognized and the embryos omitted from the analysis. (Fig. 5). While the PNP narrows, the neural folds of the We assume that the description of the chick embryo more rostra1 tapered portion appose (Fig. 2D) and sub- given in this paper reflects naturally occurring events sequently close over their full length between sst 10 of neural plate morphogenesis during neurulation. and 12 (Fig. 2E). This causes the length of the PNP to reduce from 1,500 to 800 p m (Fig. 5 ) ; relatively few Multiphasic Neural Tube Closure We found a multiphasic pattern of closure of the PNP button-like contact points are seen during this closure event. The wide portion of the PNP does not change in with varying morphology of closure and varying rates length until sst 12 (Fig. 5 ) , although its folds continue of neural tube closure progression a t specific locations to approach the midline and become oriented almost along the body axis. This nonuniform pattern of neuparallel t o each other. Between sst 12 and 14, the folds rulation contrasts with the traditional view of a smooth in this region close rapidly (Fig. 2F) with numerous progression of closure from the mesencephalic region 314 VAN STRAATEN ET AL. 0 1 neural tube: ....... open c3closed - -- apposition adhesion fusion somites 18 Fig. 7 Schematic representation of the entire neurulation process in the chick embryo. The drawing is based on morphometrical data on neuraxis elongation, neural fold contacts, and neuropore sizes and also incorporates morphological data with respect to the shape of neuropores and neural folds. Neuropore outlines are indicated by the thick lines. Apposition, adhesion, and fusion of the neural folds are based on an average impression of their location. Progression of the ANP and PNP closure points are indicated by the curved border line between heavy and light shading. This drawing reveals a multiphasic pattern of neural tube closure with distinct phases, as seen at the mesencephalic level at sst 4, at the rhornbocervical level (sst 6-7), and for the PNP in two steps (sst 10-14) and for its remnant (sst 14-18). onward (Portch and Barson, 1974; Bancroft and Bellairs, 1975; Schoenwolf, 1979, 1982). Even the detailed longitudinal study of Schoenwolf (1985) did not report the enhanced rate of final PNP closure, although Figure 9 of that study does indicate a marked rate of PNP closure between Hamburger and Hamilton (HH) stages 10 and 11(Hamburger and Hamilton, 1951). Jaskoll et al. (1991) described independent closure events at the mesencephalic and rhombocervical levels that appeared to result from different morphogenetic mechanisms. These findings have been confirmed and extended in the present study. In the following sections, we discuss the mechanisms that may be involved pre- dominantly in each phase of the neurulation process in the chick embryo. Possible Role for the Somites in Rhombocervical Closure The pattern of the rhombocervical closure suggests that the somites may be involved in closure of the neural folds. Progression of neurulation appears enhanced at the future rhombocervical level, as indicated by the local narrowing of the PNP, and subsequent buttonlike contact points between the neural folds are in register with the somites. Both observations suggest that somite expansion could be aiding in the dorsomedial 315 MULTIPHASIC NEURAL TUBE CLOSURE TABLE 1. Presence of the Anterior Neuropore (ANP) and the Rhombencephalic Neuropore (RNP) With Increasing Somite Numbers During Development Presence of ANPI no. of embryos Percentage 4 616 100 5 6 10110 11112 7 516 119 100 92 83 Somites 8 9 016 10 016 closure point 11 0 0 - t n z n neural fold TI f! P 1 t notochord n z n aa, Hensen's node 3 caudal neuraxis I . Prlm"'''n .I, streirn 1LIV-z I Bff u u Fig. 8 Drawing of the PNP illustrating the definition of its length from the cbsure point to the caudal extremity of the neuraxis and the arrangement of its wide and tapered portions. movement of the neural walls and folds. Indeed, Schroeder (1970) proposed that expansion of paraxial mesoderm could play a role in neural plate morphogenesis in amphibian embryos. Several observations support this idea for rhombocervical closure in the chick. First, the enhanced rhombocervical narrowing of the PNP and subsequent closure initiates at the level of somites 3-4, that is, midway along the row of 6-7 somites. This may indicate that the combined action of several pairs of somites is necessary to force the neural walls to elevate sufficiently for contact to be initiated. Second, there is considerable supportive evidence to validate the existence of the button-like, segmental closure pattern. This has been described in detail by Jas- Presence of RNP/ no. of embryos 016 017 4115 618 1/14 017 016 Percentage 0 0 27 75 7 0 0 koll et al. (1991) and mentioned by other authors (Gouda, 1974; Bancroft and Bellairs, 1975; Nagele and Lee, 1987; Nagele et al., 1989). Furthermore, at later stages of development, the neural tube exhibits evidence of morphological segments. These periodic undulations do not match a specific spatiotemporal pattern of neuroepithelial proliferation and differentiation and are regarded as being the result of mechanical moulding of the neuroepithelium by the somites (Lim et al., 1991). Corresponding cell lineage restrictions within the neural tube have been suggested to be imposed secondarily by the somites (Stern et al., 1991). The present study indicates that the periodic bulging of the neural tube originates before neural tube closure and supports the idea that this periodicity is imposed by the somites. Axial Curvature M a y Affect Closure of the Rhombencephalic Neuropore The independent closure events at mesencephalic and rhombocervical levels lead to the formation of the RNP. On many occasions, we observed that removal of the vitelline membrane resulted in simultaneous upward lifting of the head and widening of the RNP (see Fig. 11, suggesting that RNP closure delay may be aided by progressive dorsal flexion of the axis. This is similar to the mechanism proposed for the role of axial curvature in closure of the caudal neural tube in both chick and mouse embryos (Brook et al., 1991; Van Straaten et al., 1993a; Peeters et al., 1996). Closure of the PNP is Dependent on Convergent Extension Following the rapid rhombocervical closure, the PNP closure point gradually progresses caudally beyond the last somite formed until it is flanked by the presomitic mesoderm. Clearly, closure factors other than the somites must become gradually more important in achieving closure. These could include convergent extension, apical constriction, and ectodermal expansion. The process of convergent extension transforms the initially short and broad caudal neural plate into an elongated and slender structure. In amphibia, this reshaping is based mainly on directed cell rearrangement along the midline of the neural plate probably driven by changes in the notoplate (Jacobson and Gardon, 1976; Jacobson, 1978, 1991; Keller et al., 1985; 316 VAN STRAATEN ET AL. Jacobson et al., 1986; Keller and Tibbetts, 1989). In the chick embryo, a marked eightfold elongation and a twofold narrowing occurs throughout HH stages 4-1 1, which appear to involve not only cell rearrangement but also changes in cell shape and cell number (Schoenwolf, 1986, 1994; Schoenwolf and Alvarez, 1989; Schoenwolf and Sheard, 1989). Elongation of the neural plate appears to be especially associated with the phase of neural tube closure in the amphibian and chick embryo (Jacobson and Gordon, 1976; Jacobson, 1984; Schoenwolf and Alvarez, 1989). Thus, the rate of elongation was seven times higher in the portion of the PNP cranial to Hensen’s node than in the closed neural tube of the chick (Jacobson, 1981). The enhanced neuraxial elongation between sst 7 and 11 in the present study indicates that convergent extension is occurring and may play an important role in the mechanism of PNP closure between sst 7 and 11. Rapid, Two-step Closure of the PNP From sst 10 onward, the rate of caudal progression of the PNP closure point accelerates to 480 pdsomite stage, and the length of the PNP rapidly diminishes, resulting in its final closure (except for a small remnant). This new phase of neurulation seems to be unique in its sudden enhancement, but it is the likely continuation of the gradual reduction in width of the PNP, which, at this time, becomes sufficiently narrowed to allow fast closure. Following rhombocervical closure, the PNP exhibits a pear-like shape, with cranial tapered and caudal wide portions. The pear shape can be explained by assuming that movements of convergent extension (and, thus, narrowing of the neural plate) progress in a craniocaudal direction along the body axis that are active in the tapered portion and that are becoming active in the wide portion of the PNP (Fig. 8).This peculiar shape of the PNP coincides with a two-step closure. Due to integral narrowing of the PNP, its tapered portion is the first to close. We hardly observed button-like closure points during closure of this PNP portion, which suggests that the folds close in a craniocaudal, zipper-like manner rather than simultaneously over their full length. Narrowing continues in the wide portion of the PNP, which transforms into a slit. From sst 12 onward, this slit appears to close almost instantaneously over its entire length (except for the caudal region). Numerous button-like contacts were observed during closure in this region, suggesting a fundamentally different method of closure from that of the tapered region of the PNP. The slit-like appearance of the PNP has also been noted by others (Portch and Barson, 1974; Schoenwolf, 1979).Convergent extension may assist in the creation of this shape: elongation of the midline has been postulated to generate transverse buckling tension, which forces the neural walls to elevate and the neural folds to converge (Jacobson, 1978). We suggest that the enhanced elongation preceding and during the final phase of PNP closure causes transverse buckling of the Fig. 9 Scanning electron micrograph of the PNP of an sst 11 chick embryo. The PNP is slit like, with the neural folds extending dorsally and undergoing convergence. The wide portion (w) of the PNP is open, whereas the tapered portion (t) shows a suture line that indicates apposition or adhesion contact between the neural folds. More cranially, the folds are fused (f). The lengthof both portions of the PNP amountedto 750 pm, as measured in dorsal view, although, in this oblique scanning electron micrograph view, the PNP seems shortened. Scale bar = 100 pm. neural folds and results in the slit-like appearance of the PNP (Fig. 91, leading to rapid completion of neural tube closure. EXPERIMENTAL PROCEDURES Preparation, Manipulation, and Measurement of Embryos Eggs of White Leghorn chicks were incubated at 37°C and 55% humidity in a roller incubator (Poly- 317 MULTIPHASIC NEURAL TUBE CLOSURE hatch; Brinsea Products, Sandford, United Kingdom) for 35-50 hr to obtain embryos ranging from HH stages 7 to 15. After cracking the egg shell very gently, the contents were floated into a bowl with excess warm saline (Locke’s solution: 154 mM NaC1, 6 mM KC1, 2 mM CaC12, 10 mM D-glucose), and the embryo was viewed with a WILD M5 dissecting microscope. By using side illumination, the number of somite pairs could be visualized and subsequently counted. The egg white above the embryo was removed. A droplet of stain (1%toluidine blue in 1%borax diluted 10 times with saline) was floated over the vitelline membrane, and, within 2 min, areas of the embryo contacting the vitelline membrane (specifically the neural folds) were sufficiently stained. A photograph was taken. Several morphometrical parameters were determined on the embryo in dorsal view: length of the neuraxis, length of all neuropores, and width of the PNP, by using a scaled eyepiece graticule at a magnification of x 25. Data were depicted on a scale drawing of each embryo. The anterior extremity of the neuraxis was defined by the anterior end of the embryo. The caudal extremity of the neuraxis, the length of the PNP and of its portions, and the most caudal point of neural fold closure (the PNP closure point) were defined as indicated in Figure 8. To expose the embryo, saline was injected gently underneath the vitelline membrane, and the membrane was torn away over an area that exceeded only slightly the size of the embryo. A second photograph was taken. The length of the neuraxis and of the PNP were measured again and were compared with the previous data. Embryos in which these lengths differed by more than 5% (less than 10%of the total studied) were excluded from the next part of the study. The locations and types of neural fold contacts were determined by opening the neural tube step by step with two tungsten needles; the smallest step was about 50 pm. Fusion was noted when the folds could not be separated undamaged, adhesion was noted when the folds could be separated with adhesion bridges temporarily present, and apposition was noted when folds could be separated without any adhesive contacts (Van Straaten et al., 1993b). The positions and lengths of these contacts were depicted on the drawing of each embryo. Morphometric Analysis and Reconstruction of the Neurulation Process A total of 117 embryos were used between sst 4 and 18. It was not possible to measure all parameters in every embryo, so the data presented in Table 1and in Figures 3-6 are based on different numbers of embryos for each parameter. In 70 of the embryos, the distance between the anterior extremity of the embryo and the first somite and the craniocaudal length of the somites were determined. Although individual somites change in size during development, their average length ap- peared uniform throughout the stages in this study and was assumed to be 140 Fm at all stages. The morphometrical data were plotted graphically. For the data on neuraxis elongation, a polynomial function was found to fit the data better than a linear regression function (Fig. 3). When computing progression rates of the PNP closure, partial regression lines were constructed. A reconstruction of the entire neural tube closure process was performed by using the position of the first somite as a reference (Fig. 7). The lines indicating the anterior and posterior extremities of the embryo were deduced from the distance between the anterior extremity and the first somite and from the polynomial relationship in Figure 3. The length of the neuropores and the width of the PNP were drawn according to their linear regression formulae. The PNP closure points at each stage were connected and drawn as a continuous, curved line; its irregularities at sst 6-8 as well as the shape of the neuropores and of the neural folds were deduced from the photographs. The lengths of neural fold contacts were based on Figure 4, and their locations were based on the drawings, The position of the PNP closure point a t a given somite stage was calculated as follows: [length of the neuraxis] - [length (anterior extremity to somite l ) ] [length of the PNP]. The rate of progression of the closure point, relative t~ somite 1,was calculated subsequently for several phases of neurulation. ACKNOWLEDGMENTS We acknowledge the technical assistance of Paul van Dijk. This work was supported in part by a grant from the Netherlands Organization for Scientific Research (431241). REFERENCES Bancroft, M., and Bellairs, R. (1975) Differentiation of the neural plate and neural tube in the young chick embryo. Anat. Embryol. 147:309-335. Brook, F.A., Shum, A.S.W., Van Straaten, H.W.M., and Copp, A.J. (1991) Curvature of the caudal region is responsible for failure of neural tube closure in the curly tail (d)mouse embryo. Development 113:671-678. Copp, A.J., Brook, F.A., Estibeiro, J.P., Shum, A.S.W., and Cockroft, D.L. (1990)The embryonic development of mammalian neural tube defects. Progr. Neurobiol. 35:363-403. Golden,J.A., and Chernoff, G.F. (1993)Intermittent pattern of neural tube closure in 2 strains of mice. Teratology 47:73-80. Gordon, R. (1986)A review of the theories of vertebrate neurulation and their relationship to the mechanics of neural tube birth defect. Rev. J. Embryol. Exp. Morphol. 89(Suppl.):229-255. Gouda, J.G. (1974) Closure of the neural tube in relation to the developing somites in the chick embryo. J. Anat. 118:360-361. Hamburger, V., and Hamilton, H.G. (1951) A series of normal stages in the development of the chick embryo. J . Morphol. 88:49-92. Jacobson, A.G. (1978) Some forces that shape the nervous system. %on 6:13-21. Jacobson, A.G. (1981)Morphogenesisof the neural plate and tube. In: “Morphogenesisand Pattern Formation,” Connelly, T.G., Brinkley, L.L., and Carlson, B.M. (eds). New York Raven Press, pp. 233-263. Jacobson, A.G. (1984) Further evidence that formation of the neural 318 VAN STRAATEN ET AL. tube requires elongation of the nervous system. J. Exp. Zool. 230: 23-28. Jacobson, A.G. (1991)Experimental analyses of the shaping of the neural plate and tube. Am. Zool. 31:628-643. Jacobson, A.G., and Gordon, R. (1976)Changes in the shape of the developing vertebrate nervous system analyzed experimentally, mathematically, and by computer simulation. J. Exp. Zool. 197: 191-246. Jacobson, A.G., Oster, G.F., Odell, G.M., and Cheng, L.Y. (1986)Neurulation and the cortical tractor model for epithelial folding. J. Embryol. Exp. Morphol. 96:19-49. Jaskoll, T., Greenberg, G., and Melnick, M. (1991)Neural tube and neural crest-A new view with time-lapse high definition photomicroscopy. Am. J. Med. Genet. 41:333-345. Karfunkel, P. (1974)The mechanism of neural tube formation. Review. Int. Rev. Cytol. 38:245-271. Keller, R.E., and Tibbetts, P. (1989)Mediolateral cell intercalation in the dorsal, axial mesoderm of Xenopus laeuis. Dev. Biol. 131:539549. Keller, R.E., Danilchick, M., Gimlich, R., and Shih, J. (1985)The function and mechanism of convergent extension during gastrulation of Xenopus lueuis. J . Embryol. Exp. Morphol. 89(Suppl.):185209. Lim, T.M., Jaques, K.F., Stern, C.D., and Keynes, R.J. (1991)A n evaluation of myelomeres and segmentation of the chick embryo spinal cord. Development 113:227-238. Nagele, R.G., and Lee, H.Y. (1987)Studies on the mechanisms of neurulation in the chick Morphometric analysis of the relationship between regional variations in cell shape and sites of motive force generation. J. Exp. Zool. 244:197-205. Nagele, R.G., Bush, K.T., Kosciuk, M.C., Hunter, E.T., Steinberg, A.B., and Lee, H.Y. (1989)Intrinsic and extrinsic factors collaborate to generate driving forces for neural tube formation in the chick-A study using morphometry and computerized 3-dimensional reconstruction. Dev. Brain Res. 50101-111. Peeters, M.C.E., Shum, A.S.W., Hekking, J.W.M., Copp, A.J., and Van Straaten, H.W.M. (1996)Relationship between altered axial curvature and neural tube closure in normal and mutant (curly tail) mouse embryos. Anat. Embryol. 193:123-130. Portch, P.A., and Barson, A.J. (1974)Scanning electron microscopy of neurulation in the chick. J . Anat. 117:341-350. Schoenwolf, G.C. (1979)Observations on closure of the neuropores in the chick embryo. Am. J. Anat. 155:445-466. Schoenwolf, G.C. (1982)On the morphogenesis of the early rudiments of the developing central nervous system. Scan. Elec. Micmsc. 1289-308. Schoenwolf, G.C. (1985)Shaping and bending of the avian neuroepithelium: Morphometric analysis. Dev. Biol. 109:127-139. Schoenwolf, G.C. (1994)Formation and patterning of the avian neuraxis: One dozen hypotheses. In: “Neural Tube Defects,” Bock, G., and Marsh, J. (eds). CIBA Foundation Symposium 181.Chichester: John Wiley and Sons, pp. 25-50. Schoenwolf, G.C., and Alvarez, IS. (1989)Roles of neuroepithelial cell rearrangement and division in shaping of the avian neural plate. Development 106427-439. Schoenwolf, G.C., and Sheard, P. (1989)Shaping and bending of the avian neural plate as analysed with a fluorescent-histochemical marker. Development 105:17-25. Schoenwolf, G.C., and Smith, J.L. (1990)Mechanisms of neurulation-traditional viewpoint and recent advances. Development 109 243-270. Schroeder, T.E. (1970)Neurulation in Xenopus Zaeuis. An analysis and model based upon light and electron microscopy. J. Embryol. Exp. Morphol. 23~427-462. Silver, M.H., and Kerns, J.M. (1978)Ultrastructure of neural fold fusion in chick embryos. Scan. Elec. Microsc. Ik209-215. Stern, C.D. Jaques, K.F., k a s e r , S.E., and Keynes, R.J. (1991)Segmental lineage restrictions in the chick embryo spinal cord depend on the adjacent somites. Development 113:239-244. Van Straaten, H.W.M., Hekking, J.W.M., Consten, C., and Copp, A. (1993a)Intrinsic and extrinsic factors in the mechanism of neurulation: Effect of curvature of the body axis on closure of the posterior neumpore. Development 117:1163-1 172. Van Straaten, H.W.M., Jaakoll, T., Rousseau, A.M.J., Terwindt-Rouwenhorst, E.A.W., Greenberg, G., Shankar, K., and Melnick, M. (1993b)Raphe of the posterior neural tube in the chick embryo-its closure and reopening as studied in living embryos with a High Definition light microscope. Dev. Dynamics 19865-76.