Observations on Closure of the Neuropores in the Chick Embryo GARY C . SCHOENWOLF Department of Anatomy, The Uniuersity of New Mexico, School of Medicine, Albuquerque, New Mexico 87131

ABSTRACT Neuropore closure was studied in chick embryos by light and electron microscopy. Surface ectoderm reflects over the crests of the neural folds a t all craniocaudal levels, merging with the neural ectoderm lining the neural groove. Apices of surface ectodermal cells have an essentially identical morphology prior to approximation of folds, both within the presumptive fusion sites and more laterally. Cells of these areas have slightly convex profiles exhibiting few cellular protrusions. Each neural fold contains a superficial half, composed of neural ectoderm covered by surface ectoderm, and a deep half consisting entirely of neural ectoderm. Initial contact between folds usually occurs near the junction between these halves in cranial regions, but is restricted primarily to surface ectoderm a t caudal levels. Subsequent fusion of folds at all levels involves both ectodermal layers. Cellular protrusions and small, morphologically unspecialized intercellular junctions often interconnect cells of apposed folds in areas undergoing fusion. The anterior neuropore closes a t stages 10-11, but fusion of folds in this region is not completed until stages 13-14. Fusion occurs dorsoventrally in this area and is more advanced internally than externally. Numerous pleomorphic inclusions and a few apparently necrotic cells are present in areas bordering the anterior neuropore. The posterior neuropore closes a t stages 12-13 and fusion is completed in this region during stages 13-14. The caudal end of the posterior neuropore closes dorsal to the developing tail bud. Several morphological features of this closure may a t least partially account for the high susceptibility to myeloschisis localized specifically a t caudal spinal cord levels. The neural tube originates by two totally different mechanisms in chick embryos: the entire brain and a large portion of the spinal cord form during primary neurulation; the most caudal spinal cord regions (i.e., part of the lumbosacral and all of the tail spinal cord) originate by secondary neurulation (Criley, '69; Schoenwolf, '77, '78a). Primary neurulation involves formation of an ectodermal thickening called the neural plate (composed of neural ectoderm). The lateral margins of this structure and the adjacent regions of surface ectoderm elevate and migrate toward the midline a s the paired neural folds, which ultimately meet to establish a hollow ectodermal cylinder, the neural tube. The neural folds first come into contact a t about the level of the future midbrain, during approximately Hamburger-Hamilton ('51) stage 8. A short neural tube is thereby estabAM. J. ANAT. (1979)155: 445-466.

lished which opens both cranially and caudally (fig. 1).The broad cranial opening bounded laterally by neural folds constitutes the anterior neuropore. The neural tube opens caudally into the neural groove, the depth of which gradually decreases a t more caudal levels. The posterior neuropore is an ambiguous structure. For the purposes of the present study, i t will be considered equivalent to the slit-like dorsal opening into the caudal part of the neural groove a t stages just prior to the beginning of tail bud formation (i.e., a t about stages 11 or 12). The phase of primary neurulation terminates as the posterior neuropore closes, and subsequent formation of the neural tube occurs by secondary neurulation. Part of the tail bud cavitates during this process, ultimately

+

Received Dec. 12, '78. Accepted Mar. 22, '79.

445

446

GARY C. SCHOENWOLF

ropores (anterior neuropore: Bancroft and Bellairs, '75; Waterman, '76; posterior neuropore: Waterman, '76; Schoenwolf, '78b). The paucity of studies on neuropore closure is surprising since failure of this process to occur has devastating effects on further development of the embryo. For example, dysraphic conditions, such as anencephaly and myeloschisis, result when neuropore closure fails to occur properly. Since these defects are often present in human neonates (frequency = approximately 20% of all congenital anomalies: McKeown and Record, '61) the sequence of events normally involved in neuropore closure deserves further study. The present investigation was thus undertaken, as part of a continuing study on neurulation, to determine exactly what morphological events are involved in closure of the anterior and posterior neuropores in the chick embryo. The results presented here complement and amplify the brief scanning electron microscopic observations on chick neuropore closure previously reported (Bancroft and Bellairs, '75; Schoenwolf, '78b). Fig. 1 SEM. Dorsal view of the cranial end of a s t a g e 4 embryo showing anterior neuropore (AN), surface ectoderm (SE), neural groove (NG), and neural folds (NF) which are in contact with one another over part of their length (asterisk). X 300.

creating a single lumen, and the surrounding tail bud cells form the neural tube walls. An area of transition exists within the future lumbosacral region, where the spinal cord forms by primary neurulation dorsally and by secondary neurulation ventrally (Criley , '69; Schoenwolf, '78b, '79). Closure of the caudal end of the posterior neuropore occurs within this overlapping region. One of the most interesting features of primary neurulation is the process of neural groove closure. This event recently has been studied ultrastructurally in a variety of vertebrates (chick: Gouda, '74; Portch and Barson, '74; Revel, '74; Bancroft and Bellairs, '75; Revel and Brown, '76; Santander and Cuadrado, '76; Schoenwolf, '78b; Silver and Kerns, '78; rodent: Waterman, '75, '76; amphibian: Tarin, '71; Rice and Moran, '77; Mak, '78). Many of these studies unfortunately concentrated only on the first regions of the neural tube established; more cranial and caudal levels have thus received considerably less a t tention. Furthermore, only three of these investigations examined closure of the neu-

MATERIALS AND METHODS

Several dozen fertile White Leghorn eggs were incubated in a forced-draft incubator a t 38"Cuntil embryos reached stages 8-14 (Hamburger and Hamilton, '51). Eggs were then opened into finger bowls containing warm 0.9%saline, and the blastoderms were rapidly cut away from the yolk, washed in fresh saline, and immediately fixed and processed for light or electron microscopy. Processing for scanning electron microscopy (SEMI Blastoderms were immersed for two to three hours in one of three primary fixatives made up in 0.1 M cacodylate buffer (pH 7.2): 2% glutaraldehyde (Sabatini et al., '631, one-half strength Karnovsky's fixative (Karnovsky, ' 6 8 , or one-half strength Karnovsky's fixative containing 0.01% trinitrophenol (It0 and Karnovsky, '68). Essentially identical tissue preservation was obtained with all three primary fixatives, and no morphological differences were noted as a result of these variations in fixation procedure. Fixed embryos were then dissected free of surrounding membranes, pooled according to their stage, washed in several changes of buffer, postfixed with buffered 1%osmium tetroxide for one hour, dehydrated with a graded ethanol series, and critical-

NEUROPORE CLOSURE I N THE CHICK

447

one another in surface view (fig. 3). Cells of the surface ectoderm (both within the regions of the neural folds and more laterally) had slightly convex profiles exhibiting fine undulations. A few surface ectodermal cells contained small bleb-like projections. Cells containing such protrusions were located both within the areas of the neural folds and more laterally. It must be emphasized that prior to approximation of folds no obvious morphological differences were observed between those Processing for light microscopy (LM) surface ectodermal cells destined to undergo and transmission electron fusion during closure of the anterior neuromicroscopy ( T E M ) pore (i.e., the surface ectodermal cells of the Blastoderms were immersed for approx- neural folds) and more lateral surface ectoderimately 30 seconds in cold (4°C) cacodylate- ma1 cells. buffered (0.1 M a t pH 7.2) 2% glutaraldehyde The cells of the surface ectoderm and neural (plus 0.05% calcium chloride and 0.1 M su- ectoderm were morphologically distinct (fig. crose), while small blocks of tissue containing 31, although the transition between these two the desired regions of the embryos were dis- areas was gradual in surface view. The cells of sected out. These portions of the embryos were the neural ectoderm contained very irregular then fixed immediately on ice for one hour contours displaying numerous bleb-like prowith a cacodylate-buffered (0.1 M a t pH 7.2) trusions resembling those present a t far lower mixture containing a final concentration of frequency in the area of the surface ectoderm. 2%glutaraldehyde/l% osmium tetroxide (plus Other structures, which appeared to intercon0.05% calcium chloride). It had been deter- nect two neural ectodermal cells, were fremined previously that this fixative (modified quently observed (fig. 3: arrow). from Hasty and Hay, '77) provided better The neural folds came into contact, to close ultrastructural preservation of early chick the anterior neuropore, a t stages 10-11 (fig. 4). embryos than that obtained with a sequential A vertically oriented furrow marked the regimen of glutaraldehyde followed by osmi- former site of the anterior neuropore following um tetroxide (Schoenwolf, '79). Dissected re- this contact. Examination of both the internal gions were then dehydrated with ethanol, aspect of this furrow (as exposed in dissected transferred to propylene oxide, and embedded embryos) and its external aspect, revealed in Epon 812 (Luft, '61). Thick (1wm) and thin that fusion was slightly more advanced inter(,lo0nm) sections were then cut with diamond nally than externally (i.e., the internal porknives. Thick sections were stained with tion of the furrow is eliminated prior to disapmethylene blue/azure I1 (Richardson et al., pearance of its external part), and that fusion '60) and mounted on glass slides for examina- began a t the dorsal aspect of the furrow and tion by LM. Thin sections were collected on proceeded ventrad. By the time that the neucopper grids, stained with uranyl acetate and ral folds came into contact, to close the anterilead citrate (Reynolds, '631, and examined or neuropore, boundaries had become well with a Hitachi HS-7S TEM a t 50 kV. established between individual surface ectodermal cells, due to formation of numerous RESULTS microvilli and microplicae a t cellular periphClosure of the anterior neuropore eries (figs. 5, 6). The frequency with which SEM these structures occurred wasconsiderably reThe anterior neuropore at stage 8 consisted duced more centrally, however, especially in of a broad opening bounded laterally by neural the area immediately surrounding the central folds (figs. 1, 2). Surface ectodermal cells re- cilium, where such protrusions were usually flected over the crest of each fold and gradu- lacking. Scattered debris and occasional small, ally merged with cells of the neural ectoderm. Each cell of the neural and surface ectoderm spherical blebs were present within the furcontained a single cilium projecting from row between apposed neural folds in most emabout the center of its apical surface, and indi- bryos. Numerous cellular protrusions bridged vidual cells were not clearly delineated from the dorsal part of the cleft, indicating that

point-dried from liquid COz in a Samdri apparatus. Some embryos were transected with razor blades during dehydration to reveal their internal morphology and then criticalpoint-dried. All dried embryos were attached to aluminum stubs with double-stick Scotch tape and silver paint, coated with gold/palladium in a Hummer I sputter coater, and examined with an ETEC Autoscan SEM operated a t 10 kV.

448

GARY C . SCHOENWOLF

fusion was occurring actively in this area (fig. 6 ) . Such cellular processes were observed only in those embryos in which the apposed neural folds in the region of the anterior neuropore were already in contact. Thus, the cranial neural folds were never bridged by cellular protrusions a t stages prior to anterior neuropore closure ke., a t stages 8-91. Fusion of neural folds a t the cranial end of the embryo was completed a t stages 13-14. The most ventral part of the furrow was the last to be eliminated, since fusion occurred dorsoventrally a t the cranial end of the embryo (fig. 7). LM and TEM Neural folds in the region of the anterior neuropore were well defined by stage 8 (fig. 8). A distinct interface separated surface and neural ectoderm lateral to each fold, but these two ectodermal layers were less well separated more medially. Nevertheless, based on the arrangements of cells within the fold, it can be seen that its superficial (outer) half (portion above the bar in fig. 8 ) consisted of a core of neural ectodermal cells (denselypacked cells) capped by flattened cells of surface ectoderm (more loosely-packed cells), whereas its deep (inner) half (part below the bar in fig. 8) consisted solely of neural ectoderma1 cells. Initial contact between apposed neural folds in the anterior neuropore region usually occurred near the junction of the superficial and deep halves of each fold (ie., near the level of the left end of the bar in fig. 8 ) . Subsequent fusion of folds in cranial areas therefore involved cells of both surface and neural ectoderm. Examination of thick and thin sections confirmed SEM observations that the cells of the neural folds in cranial regions displayed few cellular protrusions prior to anterior neuropore closure. In many embryos the surfaces of the neural folds were remarkably smooth (fig. 91, while in others, occasional blebs extruded from one fold toward the contralateral one (fig. 10).Blebs in these regions usually appeared as empty vesicles resembling “glass bell-like” formations, as previously described (Klika and Jelinek, ’711, but some also contained cytoplasm, ribosomes, small vacuoles, and occasional mitochondria. They projected consistently from regions containing elaborate junctional complexes interconnecting adjacent ectodermal cells. Similar blebs were also present in regions of both surface and

neural ectoderm which did not undergo fusion during anterior neuropore closure. Closure of the anterior neuropore involves a progressive decrease in the space separating apposed neural folds, as revealed in embryos in which cranial neural folds were in contact (figs. 9,11,12).Very small, scattered intercellular junctions frequently formed across the midline between cells of the contiguous folds (fig. 11 and plate 3 inset: arrows). Unfortunately, it was impossible to ascertain exactly which junctional types were present between neural folds solely on the basis of conventional TEM of these unspecialized-appearing junctions. Small, scattered microvillus-like processes were occasionally present in the interface between apposed folds. Furthermore, section-planes frequently passed through cilia trapped within the interface (fig. 12). The space between folds was still relatively large in these areas. The fate of such imprisoned cilia is unknown; a considerable number of cilia must be trapped in this way since SEM revealed a frequency of one cilium per cell on the surfaces of the neural folds. The neural ectoderm in the region of the anterior neuropore often contained numerous darkly-stained particles in methylene blue/ azure II-stained thick sections (fig. 13). This massive accumulation of dark bodies within the neural ectoderm was present only in cranial regions; similar structures were observed only sporadically in the neural ectoderm a t more caudal levels. TEM revealed that most of these particles were inclusions contained within cells that otherwise appeared normal ultrastructurally (fig. 14). These inclusions varied in electron-density and contained particulate, membranous, and droplet-like components. They were usually a t least partially bounded by membranes. Other darkly-stained particles were entire cells apparently undergoing necrosis (fig. 15). Numerous organelles were evident in these electron-dense cells, including irregularly shaped nuclei, mitochondria, ribosomes, and endoplasmic reticulum. Closure of the posterior neuropore SEM The neural groove a t stages 8-11 consisted of a long, trough-like space bordered laterally by prominent neural folds and extending caudad from the developing neural tube. The caudal ends of the neural folds flared laterally in the region of the primitive knot and grad-

449

NEUROPORE CLOSURE IN THE CHICK

ually returned toward the midline as they merged with the short primitive ridges flanking the primitive groove (fig. 16). The entire diamond-shaped space bounded by the caudal neural folds and the primitive ridges constitutes the sinus rhomboidalis of older literature. The neural groove was closed throughout the cranial half of the embryo by stages 10-11. The neural folds flanking the neural groove in the caudal half of the embryo appeared similar to those present in cranial regions a t earlier stages. Each fold consisted of both neural and surface ectoderm, with the latter forming a cap covering the neural ectodermal cells a t the lip of each fold (fig. 17). An abrupt junction separated the surface ectodermal cells of the neural folds and the neural ectodermal cells lining the neural groove (fig. 17: arrow); the latter cells had much smaller apices than the former cells. The apical surfaces of the cells of the caudal neural folds each had a convex profile and contained a single, centrally located cilium; boundaries were well-established between cells (figs. 17, 18).Occasional bleb-like structures, as well as some scattered debris, were observed in most embryos. The neural folds generally came into contact along their entire length in caudal regions, to close the posterior neuropore a t stages 12-13. Infrequently, Le., in less than 10% of the embryos examined) the posterior neuropore was spanned by many filopodia-like processes a t these stages (fig. 19). These protrusions resembled slender threads where they bridged the neuropore, and cones where they connected to neural folds more laterally. Such processes were often associated with regions of torn plasmalemma, especially near the cranial end of the posterior neuropore (fig. 20). The primitive groove, which was initially continuous with the neural groove cranially (fig. 161, began to disappear as the posterior neuropore closed. Concomitantly, the cranial end of the primitive streak began to expand to form the tail bud. Elimination of the raphe marking the former site of the posterior neuropore was completed during stages 13-14 (fig. 21). With formation of the tail bud well under way a t this time, closure of this raphe did not occur a t the extreme caudal tip of the embryo, but rather slightly more cranially. LM and TEM The layers which composed the caudal neu-

ral folds were identified easily in transverse sections. The surface ectoderm formed a thin covering overlying the neural ectodermal cells a t the crest of each fold, as in cranial regions (fig. 22). Some ectodermal cells immediately subjacent to the surface ectoderm appeared to be only marginally associated with the developing neural tube (fig. 22: arrows); these cells are presumed to belong to the neural crest since they occupied the characteristic location of crest cells. LM and TEM observations confirmed SEM data that the apical surfaces of the cells of the neural folds in caudal regions likewise contained few cellular protrusions (fig. 22). Small blebs t h a t were occasionally present in these areas emanated from regions of numerous junctional complexes, as in cranial regions, and they almost always appeared as empty structures which were frequently partially collapsed (fig. 23). The portions of the neural folds which made initial contact during posterior neuropore closure were much more restricted than in the region of the anterior neuropore. That is, the areas of the folds which made first contact across the midline were much broader in cranial regions than in caudal regions (cf. figs. 8, 24, 25). The initial contact usually occurred primarily between cells of the surface ectoderm in caudal areas (figs. 24, 25). The dorsalmost neural ectodermal cells (fig. 24: asterisk) subsequently establish the roof of the caudal spinal cord as they come into contact in the midline. Closure of the posterior neuropore a t its caudal end occurred dorsal to the developing tail bud (fig. 25). Cavitation of the latter (secondary neurulation) was already beginning in most embryos as closure occurred. The depth of the developing primary neural tube (i.e., its dorsoventral diameter) progressively decreased craniocaudally within this short overlap zone between primary and secondary neurulation. Thus, the caudal extremes of the neural folds fused in close proximity to (i.e., just above) the floor of the developing primary neural tube. DISCUSSION

Apical morphology o f ectodermal cells prior to approximation of folds Surface ectoderm reflects over the lips of the neural folds a t both cranial and caudal levels to merge with neural ectoderm lining the neural groove. The apical surfaces of the cells of the surface and neural ectoderm were

450

GARY C. SCHOENWOLF

morphologically distinct by stage 8, confirming previous observations (Backhouse, '74; Bancroft and Bellairs, '74, '75; Jacob et al., '74; Portch and Barson, '74; Schoenwolf, '78b). The apices of all surface ectodermal cells, prior to approximation of folds, have an essentially identical morphology whether they occupy the presumptive contact sites a t the crests of the folds or the more lateral prospective epidermal areas. The surface ectodermal cells of these regions have relatively smooth external profiles exhibiting few cellular protrusions. The cells of the neural folds in the future midbrain region (i.e., the level at which the neural tube is first established) have a morphology similar to that present a t more cranial and caudal levels (Silver and Kerns, '78). The cells of the neural folds a t all craniocaudal levels of the embryo therefore have an essentially identical topography. The surfaces of the neural folds in rodent embryos have a strikingly different morphology (Waterman,'75, '76). The crests of the folds in mouse and hamster embryos display elaborate protrusions prior to their approximation in the midline. These consist primarily of ruffles in mouse embryos and long filopodia and blebs in hamster embryos. The paucity of such protrusions in chick embryos strongly suggests that a t least in this organism they play a minimal role, if any, in guiding neural folds together. Delicate alterations in the surfaces of the neural fold, therefore, probably do not function in cellular recognition and adhesion during apposition of folds in chick embryos, as suggested for rodent embryos. Filopodia-like bridges were occasionally observed spanning the posterior neuropore in the present study. Similar structures have also been observed by others (Bancroft and Bellairs, '75; Bellairs and Bancroft, '75). These investigators suggested that such processes develop as extensions from cells of each fold and subsequently meet above the neural groove, to become firmly anchored together. It was further suggested that the slender bridges, thus formed, contract, actively towing the neural folds together. The available data does not support this view, however. The fact that the entire neural groove can apparently close in the complete absence of these processes in most embryos strongly suggests that these cellular bridges are not essential for normal closure of the neural groove. It is proposed, instead, that such bridging processes are probably artifacts since they are often associated with regions of torn plasmalemma. Artifacts

of this type presumably could be generated by separation of previously adhering folds during processing for microscopy. As the folds separated, patches of plasmalemma a t points of strong adhesion might be stretched into such configurations. If the bridging processes are indeed artifacts produced in this way, they could conceivably be formed a t several steps in processing (i.e., by stretching of the blastoderm during its removal from the yolk; or during the considerable shrinkage which occurs during fixation, dehydration, and critical-point-drying: Waterman, '74; Keller and Schoenwolf, '77). Further studies, perhaps utilizing serial thin sections, are needed to determine whether these processes are real or artifactual. In the present study, small thread-like to bleb-like structures were observed frequently on the apical surface of the neural ectoderm (fig. 3: arrow). Each of these processes apparently interconnects two neural ectodermal cells and probably represents the remains of a telophase bridge between daughter cells (Bancroft and Bellairs, '75).

Contact, adhesion, and fusion of neural folds The neural folds, in all regions examined, consisted of two morphologically distinct halves as early as stage 8: a superficial (outer) portion composed of a core of neural ectoderma1 cells capped by flattened cells of surface ectoderm; and a deep (inner) part composed entirely of neural ectodermal cells. The two halves of each fold can not be resolved by light microscopy of paraffin sections (for example see DiVirgilio et al., '67: figs. 1, l a , 3, 41, but these halves are seen easily in plastic sections examined by both light and electron microscopy (figs. 8, 9, 22, 24, 25; Bancroft and Bellairs, '75: fig. 22; Santander and Cuadrado, '76: fig. 101, as well as in cross-fractured embryos examined by SEM (fig. 17; Revel and Brown, '76: fig. 4d). The neural folds in cranial regions are much broader in a superficial-to-deep plane than are folds in caudal areas. Furthermore, the exact areas of the folds which initially come into apposition during closure of the neural groove differ in cranial and caudal regions. The specific areas of the folds which make first contact also vary at different body levels in mouse embryos (Geelen and Langman, '77; Sadler, '78). The apical surfaces of the cells of the neural folds of chick embryos have been reported to undergo considerable alteration following

NEUROPORE CLOSURE I N THE CHICK

their initial apposition (Gouda, '74; Revel, '74; Bancroft and Bellairs, '75; Revel and Brown, '76; Santander and Cuadrado, '76; Schoenwolf, '78b; Silver and Kerns, '78) and these observations are confirmed here. Small, membranous ruffles may occasionally form in these areas even slightly before folds make contact, as suggested by examination of dissected embryos (Revel and Brown, '76: fig. 4d). But it is also possible that the small gap observed between folds in these embryos was artificially produced during the fracturing process. Formation of cellular protrusions following apposition of folds results in a "healing over" of the raphe demarcating the former site of the opening into the neural groove. Careful examination of the pattern of formation of such cellular protrusions has revealed that fusion of folds throughout the craniocaudal extent of the embryo does not occur in a "zipper-like" fashion; instead, fusion is discontinuous, both superficial-to-deep and cranial-to-caudal (Silver and Kerns, '78). Furthermore, examination of protrusions bridging the fusing neural folds in the region of the anterior neuropore in the present study has shown that fusion a t the cranial extreme of the embryo occurs dorsoventrally. Once neural folds come into contact, what forces hold these structures together prior to initiation of fusion? Several investigators have recently suggested that neural fold adhesion is mediated by the presence of extracellular material (ECM) a t the crests of the folds (chick: Lee et al., '76; Silver and Kerns, '78; rodent: Sadler, '78; amphibian: Moran and Rice, '75; Rice and Moran, '77; Mak, '78). Apparent interference with carbohydrate-rich ECM a t these regions in chick embryos prevents neural groove closure, resulting in gross brain and spinal cord defects (Lee et al., '76). In the present study evidence is presented which suggests that intercellular junctions may also be involved in adhesion of folds, since these structures were often present between apposed folds. Similar junctions have been observed between neural folds by others (Santander and Cuadrado, '76) as well as between contiguous endocardia1 cushions during formation of the septum intermedium of the developing heart (Hay and Low, '72). The junctions observed in the present study may actually be sites of intercellular communication, however, rather than loci of adhesion. These structures can not be classified specifically by conventional TEM alone, and it is, therefore, impossible to predict how they may

451

function. Further studies using lanthanum tracer techniques and freeze-fracture would be required to determine exactly which types of junctions form between apposed neural folds.

Possible factors involved in lumbosacral myeloschisis Lumbosacral myeloschisis frequently occurs following various experimental manipulations in chick embryos (see Criley, '67, for a review). For example, the simple removal of vitelline membrane in ovo results in open clefts in the lumbosacral spinal cord in about 70% of the cases (Criley, '66). Furthermore, defects of the spinal cord and its coverings most frequently occur a t caudal levels in human embryos (i.e., a t low thoracic, lumbar, and sacral regions: Fisher et al., '52; Barson, '70). The posterior neuropore closes within the future lumbosacral region in the chick embryo (Criley, '69), and certainly within caudal levels in the human embryo, although the exact area of closure is unknown. It is, therefore, likely that the high incidence of myeloschisis in caudal regions of both of these embryos is due to faulty closure of the posterior neuropore. Several morphological features of normal posterior neuropore closure reported here might account for the greater susceptibility of this structure to failure to close: (1)the neural groove is very shallow a t its caudal end. Closure of the posterior neuropore, therefore, occurs just dorsal to the floor of a very slender primary neural tube. (2) The neural folds are much broader in a superficial-to-deepplane in cranial regions than in caudal regions. The area of the folds which first come into apposition is, thus, relatively small a t caudal levels. (3) Closure of the posterior neuropore occurs within the region of the developing tail bud. The complex morphogenetic changes occurring as the tail bud forms in this area (Schoenwolf, '79) may actually make closure of the posterior neuropore less likely to occur. Since the exact etiology of caudal myeloschisis remains unknown, experimental studies are needed to determine which of these morphological factors, if any, may predispose to failure of closure of the posterior neuropore. Inclusions i n the region of the anterior neuropore A massive accumulation of darkly-stained particles was observed in the present study in the neural ectoderm bordering the anterior

452

GARY C. SCHOENWOLF

neuropore. Similar structures were observed in this region many years ago in a variety of vertebrate embryos (see Glucksmann, ’51, for a review), as well as more recently in t h e mouse embryo (Geelen and Langman, ’77). Since previous investigators used LM exclusively, it could not be ascertained whether these darkly-stained bodies were located intracellularly or extracellularly. I t has been shown here, however, t h a t t h e ultrastructural equivalent of these particles is mainly electron-dense inclusions, but also some apparently dying cells. What role these probably necrotic cells may play in anterior neuropore closure is unknown. Even more enigmatic is the specific origin of the numerous dense inclusions found in this area. I t seems likely that these inclusions might form in a t least four possible ways, but experimental evidence for any of these mechanisms is wanting: (1) phagocytosis and assimilation of yolk granules; (2) phagocytosis of extracellular materials; (3) phagocytosis of degenerating cellular remains; and (4) early manifestations of autolysis Le., autophagic vacuole formation). The third possibility seems a likely origin for many of these inclusions since the remainder of each inclusion-containing cell appears normal ultrastructurally, and since some neural ectodermal cells in the anterior neuropore region are apparently necrotic (when degeneration of neural ectodermal cells is induced in mouse embryos with fluorodeoxyuridine, nuclear condensation precedes fragmentation of the cell into membrane-bound bodies; adjacent, apparently unaffected neural ectoderma1 cells engulf degenerating cellular fragments, forming inclusions similar t o those described here: Langman and Cardell, ’78: cf. inclusion in their figs. 19, 20 and those in fig. 14). ACKNOWLEDGMENTS

I wish to express my appreciation to Doctor Robert E. Waterman for encouragement and constructive criticism during t h e course of this investigation, to Ms. Helen Eason and Ms. Judi DeLongo for expert technical assistance, and to Ms. Anita Kimbrell for typing the manuscript. Supported by USPHS Grant No. 1-F32-NS06055-01 from the National Institutes of Health. LITERATURE CITED Backhouse, M. 1974 Observations on the development of

the early chick embryo. SEMi1974, IIT Research Institute, Chicago, Illinois, pp. 525-532. Bancroft, M., and R. Bellairs 1974 The onset of differentiation in the epiblast of the chick blastoderm (SEM and TEM). Cell Tiss. Res., 155: 399-418. 1975 Differentiation of the neural plate and neural tube in the young chick embryo. Anat. Embryol., 147: 309-335. Barson, A. J. 1970 Spina bifida: The significance of the level and extent of the defect to the morphogenesis. Develop. Med. Child. Neurol., 12: 129-144. Bellairs, R., and M. Bancroft 1975 Midbodies and beaded threads. Am. J. Anat., 143: 393-398. Criley, B. B. 1966 Development of the chick embryo i n ouo following complete removal of t h e vitelline membrane over the blastodisc. Amer. Zool., 6: 608-609. 1967 Analysis of the Embryonic Sources and Mechanisms of Development of Posterior Levels of Chick Neural Tubes. Ph.D. thesis, University of Illinois, Champaign/Urbana, Illinois. 1969 Analysis of the embryonic sources and mechanisms of development of posterior levels of chick neural tubes. J. Morph., 128: 465-501. DiVirgilio, G., N. Lavenda and J . L. Worden 1967 Sequence of events in neural tube closure and the formation of neural crest in the chick embryo. Acta Anat., 68: 127-146. Fisher, R. G., A. Uihlein and H. M. Keith 1952 Spina bifida and cranium bifidum: Study of 530 cases. Proc. Staff Meet. Mayo Clin., 27: 33-38. Geelen, J . A. G., and J. Langman 1977 Closureof theneural tube in the cephalic region of the mouse embryo. Anat. Rec., 189: 625-640. Glucksmann, A. 1951 Cell deaths in normal vertebrate ontogeny. Biol. Rev., 26: 59-86. Gouda, J. G. 1974 Closure of the neural tube in relation to the developing somites in the chick embryo (Gallus galEus domesticus). J. Anat., 118: 360-361. Hamburger, V., and H. L. Hamilton 1951 A series of normal stages in the development of the chick embryo. J. Morph., 88: 49-92. Hasty, D. L., and E. D. Hay 1977 Freeze-fracture studies of the developing cell surface. Formation of particle-free membrane blebs during glutaraldehyde fixation. J. Cell Biol., 75: 234a. Hay, D. A., and F. N. Low 1972 The fusion of dorsal and ventral endocardial cushions in the embryonic chick heart: A study in fine structure. Am. J. Anat., 133: 1-24. Ito, S., and M. J. Karnovsky 1968 Formaldehyde-glutaraldehyde fixatives containing trinitro compounds. J. Cell Biol., 39: 168a-169a. Jacob, H. J . , B. Christ, M. Jacob and G. V. Bijvank 1974 Scanning electron microscope @EM) studies on the epiblast of young chick embryos. Z. Anat. Entwickl: Gesch., 143: 205-214. Karnovsky, M. J. 1965 A formaldehyde-glutaraldehydefixative of high osmolality for use in electron microscopy. J. Cell Biol., 27: 137a-138a. Keller, R. E., and G. C. Schoenwolf 1977 An SEM study of cellular morphology, contact, and arrangement, as related to gastrulation in Xenopus laeuis. Wilhelm Roux’s Archives of Develop. Biol., 182: 165.186. Klika, E., and R. Jelinek 1971 “Glass bell-like’’formations as modified cellular junctions in the developing neural tube and endoderm of the chick embryo. Folia Morphol. (Praha), 19: 137-145. Langman, J., and E. L. Cardell 1978 Ultrastructural observations on FudR-induced cell death and subsequent elimination of cell debris. Teratology, 17: 229-270.

NEUROPORE CLOSURE IN THE CHICK Lee, H.-Y., J. B. Sheffield, R. G. Nagele, Jr. and G. W. Kalmus 1976 The role of extracellular material in chick neurulation. I. Effects of concanavalin A. J. Exp. Zool., 198: 261-266. Luft, J. H. 1961 Improvements in epoxy resin embedding methods. J. Biophys. Biochem. Cytol., 9: 409-415. Mak, L. L. 1978 Ultrastructural studies of amphibian neural fold fusion. Develop. Biol., 65: 435-446. McKeown, T., and R. G. Record 1961 Malformations in a population observed for five years after birth. In: Ciba Foundation Symposium on Congenital Malformations. G. E. W. Wolstenholme and C. M. OConnor, eds. Little, Brown and Co.,Boston, pp. 2-16. Moran, D., and R. W. Rice 1975 An ultrastructural examination of the role of cell membrane surface coat material during neurulation. J. Cell Biol., 64: 172-181. Portch, P. A., and A. J. Barson 1974 Scanningelectron microscopy of neurulation in the chick. J. Anat., 117: 341-350. Revel, J:P. 1974 Scanning electron microscope studies of cell surface morphology and labeling, in situ and in vitro. SEM/1974, IIT Research Institute, Chicago, Illinois, pp. 541-548. Revel, J:P., and S. S. Brown 1976 Cell junctions in development, with particular reference to the neural tube. Cold Springs Harbor Symp. Quant. Biol., 40: 443-455. Reynolds, E. S. 1963 The use of lead citrate at high pH as a n electron-opaque stain in electron microscopy. J. Cell Biol., 17: 208-212. Rice, R. W., and D. J. Moran 1977 A scanning electron microscopic and X-ray microanalytic study of cell surface material during amphibian neurulation. J. Exp. Zool., 201: 471-478. Richardson, K. C., L. J a r e t t and E. H. Finke 1960 Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technol., 35: 313-323. Sabatini, D. D., K. Bensch and R. J. Barrnett 1963 Cytochemistry and electron microscopy. The preservation of

453

cellular ultrastructure and enzymatic activity by aldehyde fixation. J. Cell Biol., 17: 19-58. Sadler, T. W. 1978 Distribution of surface coat material on fusing neural folds of mouse embryos during neurulation. Anat. Rec., 191: 345-350. Santander, R. G., and G. M. Cuadrado 1976 Ultrastructure of t h e neural canal closure in the chicken embryo. Acta Anat., 95: 368-383. Schoenwolf, G. C. 1977 Tail (end) bud contributions to the posterior region of t h e chick embryo. J. Exp. Zool., 201: 227-246. 1978a Effects of complete tail bud extirpation on early development of the posterior region of the chick embryo. Anat. Rec., 192: 289-296. 1978b An SEM study of posterior spinal cord development in the chick embryo. SEM/1978Nol. 11, Scanning Electron Microscopy, Incorporated, Chicago, Illinois, pp. 739-746. 1979 Histological and ultrastructural observations of tail bud formation in the chick embryo. Anat. Rec., 193: 131-148. Silver, M. H., and J. M. Kerns 1978 Ultrastructure of neural fold fusion in chick embryos. SEM/1978Nol. 11, Scanning Electron Microscopy, Incorporated, Chicago, Illinois, pp. 209-215. Tarin, D. 1971 Scanning electron microscopical studies of the embryonic surface during gastrulation and neurulation in Xenopus laeuis. J. Anat., 109: 535-547. Waterman, R. E. 1974 Scanning electron microscopic techniques applied to embryonic and fetal tissues of vertebrates. In: Principles and Techniques of Scanning Electron Microscopy. M. A. Hayat, ed. Van Nostrand Reinhold, Co., New York, Vol. 2, Chap. 8, pp. 93-110. 1975 SEM observations of surface alterations associated with neural tube closure in the mouse and hamster. Anat. Rec., 183: 95-98. 1976 Topographical changes along the neural fold associated with neurulation In the hamster and mouse. Am. J. Anat., 146: 151-172.

Abbreuiations B, Bleb-like structures BB, Basal bodies C, Cilium CV, Coated vesicle I, Inclusion JC, Junctional complexes M, Mitochondrion N, Nucleus NC, Notochord NE, Neural ectcderm

NF, Neural fold NG, Neural groove OV, Optic vesicle PG, Primitive groove PN, Posterior neuropore SE, Surface ectoderm TB, Tail bud V, Ventral surface of head VA, Vacuole

PLATE 1 EXPLANATION O F FIGURES

2

SEM. Cranial view of t h e anterior neuropore of t h e stage-8 embryo shown in figure 1. X 500.

3 Enlargement of central portion of figure 2. Several bleb-like protrusions are visible on the apical surface of the neural ectoderm. Arrow indicates a structure which apparently interconnects two adjacent ectodermal cells. X 2,400. 4

SEM. Cranial view of the head of a stage-11 embryo. The former site of the anterior neuropore is indicated by a vertical, midline furrow lying between the prominent optic vesicles. Lower arrow indicates area enlarged in figure 5; upper arrow indicates area enlarged in figure 6. X 150.

5 Enlargement of area indicated by lower arrow in figure 4. Scattered debris lies within the furrow. x 2,000.

6 Enlargement of area indicated by upper arrow in figure 4. Cellular protrusions (arrows) extend across the furrow beneath contiguous neural folds. x 1,500.

7 SEM. Cranial view of the head of a stage-13 embryo. The contiguous neural folds have not yet completely fused and a shallow furrow (asterisk) is still present between them. x 140.

454

NEUROPORE CLOSURE IN THE CHICK Gary C . Schoenwolf

PLATE 1

455

PLATE 2 EXPLANATION OF FIGURES

8 LM. Transverse section. Neural fold in the region of the anterior neuropore in a stage-8 embryo. Arrowheads indicate boundary between surface and neural ectodermal limbs of neural fold. The superficial half of the fold is located above the level of the bar; its deep half lies below this level. x 480. 9 TEM. Transverse section. Neural folds are in contact, closing the anterior neuropore in this stage-11 embryo. Arrows indicate expanse of area enlarged in figure 11. Asterisks indicate spaces separating neural and surface ectoderm. x 7,000. 10 TEM. Transverse section. Enlargement of the surface of a neural fold in the re-

gion of the anterior neuropore in a stage-8 embryo. Small protrusions from the apices of the cells of t he neural fold contain either cytoplasm and some organelles or appear as empty vesicles. x 13,200.

456

NEUROPORE CLOSURE I N THE CHICK Gary C. Schoenwolf

PLATE 2

457

PLATE 3 EXPLANATION OF FIGURES

Inset

TEM. Transverse section showing closely apposed neural folds in the region of the anterior neuropore in a stage-11 embryo. Asterisks indicate wide extracellular spaces interspersed between intercellular junctions (arrows) interconnecting cells of t h e adjacent folds. x 92,800.

11 Enlargement of area indicated by arrows in figure 9. Arrow indicates a developing intercellular junction interconnecting t h e closely apposed neural folds. The basal bodies are located toward the ventral aspect of the section. x 62,000. 12 Enlargement of apposed neural folds in the anterior neuropore region from the same embryo as that shown in figure 9. Micrograph from area just ventral to field of view shown in figure 9. Note cilium occupying interspace between approaching neural folds. The base of the cilium is located toward the dorsal aspect of the section. x 62,000.

458

NEUROPORE CLOSURE IN THE CHICK Gary C. Schoenwolf

PLATE 3

459

PLATE 4 EXPLANATION OF FIGURES

13 LM. Transverse section through region of anterior neuropore in a stage-8 embryo. Numerous darkly-stained particles are visible within the neural ectcderm. Arrow indicates debris on surface of neural fold. x 120. 14 TEM. Transverse section. Stage 11. Several neural ectodermal cells bordering the closed anterior neuropore contain pleomorphic inclusions. x 13,200.

15 TEM. Transverse section. Stage 8. Region of the anterior neuropore. A neural ectodermal cell, presumed t o be dying, contains organelles and vacuoles. x 17,200.

460

NEUROPORE CLOSURE IN THE CHICK Gary C. Schoenwolf

PLATE 4

461

PLATE 5 EXPLANATION OF FIGURES

16 SEM. Dorsal view of the caudal half of a stage-11 embryo showing neural folds, neural groove, and primitive groove. Arrow indicates area enlarged in figure 18. X 175. 17 SEM. Transverse fracture through the neural groove of a stage-11 embryo showing closely apposed neural folds. Each neural fold is composed of surface and neural ectoderm. Arrow indicates boundary between these two areas. x 1,400. 18 Enlargement of the closely apposed neural folds in area indicated by arrow In figure 16. x 2,000. 19 SEM. Dorsal view of the caudal end of a stage-12 embryo showing thin cellular processes bridging the posterior neuropore. Arrow indicates area enlarged in figure 20. X 270.

20 Enlargement of area indicated by arrow in figure 19. An area of disrupted plasmalemma (arrow) is present a t t h e base of a cellular protrusion. x 3,000. 21 SEM. Dorsal view of the caudal end of a stage-13 embryo. The contiguous neural folds are not yet completely fused and a shallow furrow (asterisk) is still present between them. The tail bud is forming a t the caudal end of the embryo. x 120.

462

NEUROPORE CLOSURE IN THE CHICK Gary C. Schoenwolf

PI.ATE 5

463

PLATE 6 EXPLANATION OF FIGURES

22

TEM. Transverse section of caudal neural fold in a stage-10 embryo. Asterisk indicates space separating neural and surface ectoderm. Arrows indicate neural ectodermal cells presumed to be part of the neural crest. x 63,000.

23 TEM. Transverse section. Partially collapsed bleb in region of caudal neural folds in a stage.10 embryo. X 25,000. 24

LM. Transverse section. The surface ectodermal components of the neural folds have come into contact to close the posterior neuropore in this stage-12 embryo. Asterisk indicates approximate extent of the neural ectodermal components of the folds, which will soon make contact in the midline. x 90.

25 LM. Transverse section. Stage 13. The caudal end of the posterior neuropore has just closed dorsal to the cranial end of the tail bud. Cavitation (asterisk) is under way in the latter. x 90.

465