SEARCH

SEARCH BY CITATION

Keywords:

  • chick embryo;
  • initial neurulation, actin;
  • cytochalasin D;
  • cell movements

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES
  8. Supporting Information

In a previous study, we have demonstrated that initial closure of the mesencephalic neural groove in the chick embryo is different from neurulation elsewhere. The neural groove invaginates, the walls appose and make contact in a ventrodorsal direction, and subsequently separate ventrally, forming an incipient neural tube lumen, which finally widens into a definitive lumen. In this study, a role for actin in the processes of this initial mesencephalic closure is studied. Based on rhodamine-phalloidin–stained sections, three distinct actin distribution patterns emerged, and time-lapse video microscopy revealed cytochalasin-D–reversible neurulation movements. We propose that actin is involved in formation and stabilization of the neural groove hinge point, in invagination of dorsal neuroepithelial cells into the neural groove, in the origin of the incipient lumen and the reinforcement of adhesion of the dorsal neural folds, and finally in the development of a wide lumen. Such a multifunctional effect of actin microfilaments within a narrow time window and at specific sites has not been reported yet. © 2002 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES
  8. Supporting Information

According to the traditional viewpoint, initial closure of the cephalic neural plate of the chick embryo involves bilateral elevation of the edges of the plate, flattening of the floor plate, subsequent convergence of the dorsal neural folds, followed by their adhesion and fusion. This process is clearly depicted by Nagele and coworkers (Nagele et al., 1989). Such a sequence suggests that the neural groove is wide almost from its inception. In a previous study (van Straaten et al., 1997), we proposed a new viewpoint for this initial closure: at the stage of 0 somites a neural groove forms by invagination, subsequently the walls of the neural groove appose and make contact. At the stages of 1–3 somites, the apposition/contact zone, which functions as a seal, shifts in dorsal direction, whereas the neural walls separate ventrally to form the incipient neural tube lumen. Finally, at the stage of 4 somites, the lumen expands in lateral direction, dorsally bordered by neural folds, and also opens in the rostral and caudal neuropores. By this mechanism, the neural walls and folds remain in contact from the beginning onward. Once initial closure is completed, further neurulation commences in rostral and caudal direction, according to the traditional viewpoint.

The underlying mechanism of this distinct mesencephalic closure is unknown. Many factors and processes are involved in neurulation movements; of these, we focus on apical constriction mediated by actin microfilaments for the following reasons: (1) Neuroepithelial cells possess abundant actin microfilaments, the majority of which is oriented circumferentially in the apical region of the cell (Baker and Schroeder, 1967; Sadler et al., 1982; Lee and Nagele, 1985; Morris Kay and Tuckett, 1985). It is proposed that these actin microfilaments constrict the apical ends of neuroepithelial cells by a “purse-string–like” mechanism, which generates a wedge-shape appearance of these cells and provides the bulk of driving forces for uplifting of the neural walls (Baker and Schroeder, 1967). (2) From morphologic and morphometrical studies, apical constriction was extrapolated to be an important driving force, especially in mesencephalic neurulation (Burnside, 1973; Nagele and Lee, 1980; Lee and Nagele, 1985). (3) Application of cytochalasin-D (Cyt-D), a depolymerizing agent for filamentous actin, reverses elevation and convergence of neural walls and folds, leading to neural tube closure defects, preferentially in the cephalic region as has been reported for hamster (Wiley, 1980), rat (Morriss Kay and Tuckett, 1985; Matsuda and Keino, 1994), mouse (Ybot Gonzalez and Copp, 1999), and chick (Schoenwolf et al., 1988). However, Schoenwolf et al. (1988) claimed that disturbance of only neural fold fusion could be ascribed to Cyt-D. (4) Several mice knockouts deficient in the putative regulators of cytoskeletal dynamics or in cytoskeletal proteins suffer from cephalic closure defects (Juriloff and Harris, 2000).

Thus we hypothesized that the distinct morphogenetic movements of initial mesencephalic neurulation are dependent on actin. To test this, we studied mesencephalic neurulation movements, determined actin distribution patterns and tested filamentous actin requirement with Cyt-D.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES
  8. Supporting Information

In this study, we present a model for the morphogenetic mechanism of initial mesencephalic closure and discuss, in a chronologic description, the prerequisite role of actin in this process. A summary is presented in Figure 3.

Actin Is Involved in Neural Furrow Formation

The initial development of the neural groove, at the stage of 0 somites, is visualized in our movie 1 (see http://jws-edci.interscience.wiley.com:8998/jpages/1058-8388/suppmat/224_1/v224.html to view the 7 AVI movies). The initial groove in the future mesencephalic region, formed by two small, apposed neural walls, extends in rostral and caudal directions during development. Hensen's node and the primitive streak move caudally and the head fold develops. Figure 1A shows that actin microfilaments are already present and concentrated at the apical side of the neuroepithelial cells. Figure 1D shows that actin is rather evenly distributed throughout the cephalic neuroepithelium, with some enhancement in the median groove, the medial hinge point (MHP). This local actin enhancement has been reported before (Burnside, 1973; Nagele and Lee, 1980; Lee and Nagele, 1985). Upon application of Cyt-D, the neural groove flattens (movie 2, Fig. 2A,B), which strengthens the suggestion of the above-mentioned authors, that actin is required for the MHP cells to become wedge-shaped, resulting in furrowing of the neural plate (Fig. 3A). Our finding contradicts another report that showed that the MHP furrowing in the chick is not sensitive to Cyt-D (Schoenwolf et al., 1988) but confirms data showing that formation of this furrow is also inhibited by prevention of calcium ion transport, where calcium ions are proposed to play a role in actin microfilament contraction (Ferreira and Hilfer, 1993). Therefore, we propose that actin is involved in formation of the MHP.

thumbnail image

Figure 1. Presence and distribution of actin in neuroepithelial cells of the chick embryo during mesencephalic closure, as presented in the cephalic region of three representative embryos. A–C: Presence and distribution of actin-coupled rhodamine at the apical side of neuroepithelial cells in plastic sections at three stages of mesencephalic closure. In B, adhesion between the incipient neural folds is released due to the histotechnical procedure. D–F: Three distinct actin distribution patterns in the cephalic neural plate. Data are reconstructed from serial sections. The plate is drawn unfolded, flattened, and in dorsal view. Rostral is up. Gray levels represent extinction of actin-coupled rhodamine; darkest gray level correspond to highest extinction. Square size: 25 × 25 μm2. The sections in A–C are from the indicated levels, which correspond to the centers of the mesencephalic closure.

Download figure to PowerPoint

thumbnail image

Figure 2. Effect of cytochalasin-D (Cyt-D) on initial neurulation. A: Control chick embryo. Arrows indicate the partly invaginated neural groove in the mesencephalic region. B: After 1.5 hr of exposure to 1.6 μg/ml Cyt-D, the head fold is not developed, the neural groove has flattened, and the median furrow (arrow) is only weakly present. C: Control chick embryo. D: After 1 hr of exposure to 3.2 μg/ml Cyt-D. The rostral (RNP) and caudal neuropore (CNP) are diverged, the mesencephalic neural tube is flattened, but the mesencephalic neural folds (MNF) are still apposed. HN, Hensen's node; HF, head fold; NP, neural plate. Scale bars = 50 μm in A,B, 100 μm in C,D.

Download figure to PowerPoint

thumbnail image

Figure 3. Proposed mechanism of initial mesencephalic closure in the chick embryo, as explained by changes in cell shape due to contraction of apical actin microfilaments. Green, neuroepithelium; orange, ectoderm; light green, individual cell; red, localization and relative amount of actin; blue, contact zone. Grey arrows indicate proposed movements.

Download figure to PowerPoint

Actin Forces Surface Cells to Form the Neural Groove by a Roll-In Movement

A striking phenomenon is evident from movie 1 as well: the sheet of rostral surface cells migrates in rostromedial direction, contributing to the extension of the head fold, but, more importantly, part of the cells disappear into the neural groove. In movie 3, these roll-in movements are seen in more detail. The invaginating cells must be neuroepithelial cells, as appears from Figure 1A, where many of these high-columnar cells are still at the dorsal surface (compare with Fig. 1C). Thus the arise of the neural groove is likely due to a roll-in movement of surface cells. Therefore, we prefer to indicate this process as “invagination of the neural groove” rather than the traditional “elevation of neural walls.”

When Cyt-D is applied, the cell movements come to an immediate halt and even reverse (movie 2), and at slightly later stages, cells of the neural walls migrate laterally (after a transient apposition of the walls) and the neural groove decreases in depth (movie 4). Thus, invagination of the neural groove appears to be reversed by Cyt-D.

What could drive the dorsal neuroepithelial cells to move into the neural groove? We can think of three mechanisms, as follows:

  • 1
    The ectoderm pushes neuroepithelial cells medially and subsequently into the groove. Studies on this subject have lead to the proposal that expansion of the ectoderm by cell reshaping and by oriented mitoses is the main driving force in elevation and convergence (Schoenwolf and Alvarez, 1991; Smith and Schoenwolf, 1991; Sauseso et al., 1997). However, as our movies 1 and 3 show, the dorsal cell sheet (both neuroepithelial and ectodermal cells) migrates as a whole. In case of intrinsic ectoderm expansion, lateral cells would not migrate medially or at a low rate. Therefore, expansion of the ectoderm does not seem to provide a major contribution to neural groove formation.
  • 2
    The neural groove is pulled into the depth by underlying structures. This mechanism is rather unlikely, as the neuroepithelium itself seems to comprise most of the force-generating cellular mass (see Fig. 1A). On the other hand, convergent extension of the neural plate/notochord has been proposed to evoke elevating forces by a buckling mechanism (Keller et al., 1992; Jacobson and Moury, 1995), but this mechanism would especially affect the mid-embryonic part of the neural groove not the extreme cephalic part.
  • 3
    The roll-in mechanism resides in the neural groove itself. We propose that apical contraction of actin filaments is especially involved in those neuroepithelial cells that are at the position to roll inward the neural groove. Such contraction will change their inverted wedge shape (base medially directed) into a columnar shape (Fig. 3B), with two consequences: (A) the new columnar cells become part of the neural walls (thus the groove deepens), and (B) the dorsal sheet of neuroepithelial and ectodermal cells will be pulled medially. This mechanism will work only when the neural walls remain aligned in parallel, which is possible if the MHP forms a frozen hinge, and/or the walls remain connected at their contact zone. The former is likely evoked by the relative abundance of actin microfilaments, stabilizing the hinge, while the latter does occur, as is discussed below. Thus actin is proposed to affect two different processes in groove invagination: a roll-in movement of cells and MHP stabilization. A comparable mechanism has been described in Xenopus: superficial neuroepithelial cells contract apically, roll the neural plate into a trough and appear to pull the superficial epidermal cell sheet medially (Davidson and Keller, 1999).

Actin Initiates the Incipient Lumen and Reinforces Wall Contact

In movie 5, the development from a 1-somite into a 4-somite embryo is visualized. The neural walls of the initial closure site extend their apposition in the rostral and caudal direction and finally the rostral and caudal neuropores arise.

At these stages actin is predominantly present in a zone halfway up the neural walls (Fig. 1B,E). We propose that increased apical constriction in those cells changes the shape of the neural wall cells from columnar into wedge-shaped (Fig. 3C), resulting in transformation of the flat neural wall into a concave one, with two consequences: initiation of an incipient lumen and strengthening of contact between the dorsal neural walls. Application of Cyt-D between the stages of 3 and 5 somites invariably results in a fast (within 20 min) and profound widening of both neuropores (movie 6) as has been reported by others, but the mesencephalic neural folds remain apposed (although in some cases torn apart by the widening neuropores). The only effect is flattening of the mesencephalic neural tube (Fig. 2C,D), indicating that the walls have lost their rigidity.

We suspected that adhesion between the apposed neural walls prevented their divergence after Cyt-D application. In movie 7, the neural folds are mechanically released with a glass needle approximately 15 min after Cyt-D application. The released folds appeared very floppy. Once directed laterally, they have a similar appearance as those of the rostral and caudal neuropores (movie 7). When this experiment was performed in controls, the walls apposed immediately after their mechanical release. This experiment indicates that actin is involved in the formation of the concave walls of the incipient lumen.

Actin Is Involved in the Widening of the Incipient Lumen

From the stage of 4 somites onward, the incipient lumen is wide (Fig. 1C). Actin appears predominantly present in the ventral half of the neural tube (Fig. 1C,F) and the extinction is higher than at any previous stage. We propose that actin changes the slit-like incipient lumen into a round definitive lumen by transforming all ventrolateral neural wall cells into wedge-shaped cells (Fig. 3D).

In conclusion, based on the distinct cell movements, the specific actin distribution patterns, and the effects of Cyt-D, we propose that actin is an important factor in initial mesencephalic closure and that several processes of this closure can be explained by sequential actions of actin at specific sites.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES
  8. Supporting Information

Fertilized White Leghorn chicken eggs were incubated for approximately 30 hr. Embryos were prepared with great care to avoid mechanical disturbance of neurulation movements. Only stage 6–8 embryos (0–5 somites) were further processed according to one of the following procedures.

Actin Staining and Quantification

Embryos were fixed with 4% formaldehyde, permeabilized with acetone at −20°C and incubated with rhodamine-conjugated phalloidin (Molecular Probes, Inc., Eugene, OR) in PBS (25 U/ml) for 1 hr in the dark. Next, the embryos were rinsed in PBS, dehydrated in an acetone series, and embedded in Technovit 8100 (Kulzer and Co, GmbH, Wehrheim, Germany). Serial transverse sections of 5 μm were cut, and 1 of 5 sections were collected. Rhodamine was visualized on a Leica DM RXA fluorescence microscope by using a 25× immersion objective (N.A. 0.75). Fluorescent images of each section were acquired, integrated by a COHU camera and digitized. The integration time of the camera was adjusted such that a linear relation between the extinction of rhodamine (which is a linear measure for the amount of actin) and gray levels of the digitized image occurred.

In each section, the extinction and distribution of rhodamine-coupled actin was determined with a Leica Quantimet 500 Image Analysis System. In short, per section, a one-pixel line was computed along maximum extinction points at the apical edge of the neural plate. This line was divided in stretches of 25 μm and per stretch the average extinction was determined. Data from all sections were put into a two-dimensional matrix, smoothed by one round of a 3 × 3 moving average and plotted as gray intensities onto an unfolded dorsal projection drawing of the neural plate. Each data point, thus, represented a neural plate surface area of 25 × 25 μm2 (25-μm section interdistance and 25-μm line stretch). Twenty-one embryos were processed according to this procedure.

Time-Lapse Video Microscopy

For time-lapse images, the embryo was placed in a Petri dish with the lid on, placed on a translucent thermo-stage and viewed with a Leica MZFLIII stereomicroscope, equipped which a 3-CCD camera. The embryo was obliquely transilluminated. The vitelline membrane, which was always on top of the embryo to prevent drying, did not interfere with imaging. Images were grabbed at fixed time intervals (mostly 2 min). After adaptation of 15–30 min, 40 μl of either Locke's saline, 0.5% dimethyl sulfoxide (DMSO) (Sigma-Aldrich, Inc) in saline, or Cytochalasin-D (Sigma-Aldrich, Inc) in 0.5% DMSO in saline, with final concentrations of 1.6 μg/ml at 0 and 1 somite stages and 3.2 μg/ml at 2–5 somite stages (concentrations from Schoenwolf et al., 1988, and empirically adjusted for age), were applied. Images were grabbed for an additional 2 hr; embryos appeared to remain viable for at least 6 hr. Movies were made from 150 embryos.

Scanning Electron Microscopy

For scanning electron microscopy, embryos were fixed in 4% buffered formaldehyde, dehydrated in ethanol, critical-point-dried by using liquid CO2, attached to aluminium stubs with silver paint, coated with gold, and viewed at 10 kV with a scanning electron microscope (Philips 505).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES
  8. Supporting Information

The authors thank Dr. Wout Lamers and Dr. Madeleine Brouns for the discussions as well as for their critical comments on the manuscript. We also thank Antone G. Jacobson for the several long and fruitful discussions about neurulation mechanisms.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES
  8. Supporting Information

The Supplementary Materials referred to in this article can be found at http://jws-edci.interscience.wiley.com:8998/jpages/1058-8388/suppmat/224_1/v224.html

FilenameFormatSizeDescription
movies.zip32383KSupporting Information file movies.zip

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.