The morphological sequence of chick amniogenesis has been described (Hamburger and Hamilton,1951; Romanoff,1960; Davidson,1977) and there have been limited attempts to determine the mechanism(s) by which amniogenesis proceeds (Lillie,1903; Adamstone,1948; Davidson,1977; Overton,1989). However, the processes that drive the formation and eventual closure of the amnion are not understood. Early investigations suggested that amniogenesis involves tractional forces of the embryo (such as pushing of the embryo head into the amnion fold) and progressive differentiation of the amnion both anteriorly and posteriorly (Lillie,1903). Shore and Pickering (1889) described closure of the amnion as the closing of the lateral folds. Subsequently, Adamstone (1948), using cautery, argued that the formation of the amniotic folds is organized (in part) by groups of cells in the proamnion area and the primordium of the tail fold. Later experiments suggested tension and resistance forces at the leading edge cells of the amnion were involved, based on microfilament disruption experiments using Cytochalasin D (Overton,1989). Although amnion closure was prevented with Cytochalasin D, such microfilament disrupting agents lack discrimination in terms of cellular target (Cooper,1987), affecting not just the amnion but other microfilaments throughout the embryo. Miller et al. (1994) reported that the elevation of the lateral folds during amniogenesis correlated with the existence of domains of differential cell proliferation in the ectoderm epithelium and suggested that these patterns of proliferation possibly assist the folding and progression of the amnion.
Actin cables formed from accumulation of actin microfilaments in leading edge cells and contracting in a “purse string” mechanism have been demonstrated in embryonic wound closure (Martin and Lewis,1992; Brock et al.,1996; Lawson and England,1998; Bement et al.,1999; MacManus et al.,2006), some forms of adult wound closure (Danjo and Gipson,1998), and morphogenesis (Williams-Masson et al.,1997; Grose and Martin,1999; Jacinto et al.,2002). In addition to the formation of an actin cable, ultrastructural components such as lamellapodia and in particular, filopodia have been documented both in embryonic epithelial fusion (Martin and Wood,2002) and in the last phase of embryonic epithelial wound healing (Redd et al.,2004).
The exoenzyme C3 transferase is specific for Rho (Bishop and Hall,2000) and reacts to ADP-ribosylate Rho proteins making them biologically inactive (Ridley and Hall,1992). Rho regulates the assembly of actin stress fibres and focal adhesions in fibroblasts (Ridley and Hall,1992). Inhibition of Rho leads to the disruption of the actin stress fibres and focal adhesions, which means that various cellular mechanisms are unable to take place. Wound experiments have used scrape loaded C3 transferase to successfully block Rho and prevent the formation of a Rho mediated actin cable, therefore inhibiting wound closure (Brock et al.,1996; MacManus et al.,2006). However, cell permeable C3 transferase may be useful in this embryonic development study as cells cannot be scrape loaded via a cellular disruption method such as in wounding experiments.
The developing chick amnion is an avascular, noninnervated, free-floating, bi-layered membrane composed of an inner single cell epithelial layer and an outer mesenchymal layer (Romanoff,1960; Davidson,1977; Bellairs and Osmond,2005). The amnion is different from most other embryonic tissues as it does not have a cellular substratum. The lack of a substratum may influence the mechanism(s) involved in the formation of amniotic tissue and as a consequence differ from that reported in other embryonic tissue movements (England and Cowper,1977; Redd et al,2004; Woolner et al,2005). This study examines whether or not an actin cable is present during the development and closure of the chick amnion.
MATERIALS AND METHODS
Fertilized Ross eggs (Moy Park, Dungannon, Northern Ireland) were incubated at 37.5°C and windowed according to the method of Summerbell and Hornbruch (1981) on either Day 2 or Day 3 of incubation. Embryos were observed at Stage 10–18 according to the system of Hamburger and Hamilton (1951) using a Leica MZ6 (Wetzler, Germany) dissection microscope.
Scanning Electron Microscopy
To examine the surface morphology of the developing amnion, embryos at Stage 10–18 were prepared for scanning electron microscopy (SEM). Embryos were fixed in ovo with 3% glutaraldehyde in 0.1-M sodium cacodylate buffer (2–3 min), dissected out, and replaced in fresh fixative at 4°C for a minimum of 8 hr. The overlying chorion was removed from the extraembryonic areas after fixation as this was found to be technically easier than in ovo. Specimens were washed in 0.1-M sodium cacodylate buffer for 2 hr. Specimens were dehydrated through a graded series of alcohols, dried with hexamethyldisilazane, secured onto aluminium stubs, sputter coated with gold (10–15 nm) and viewed using a JEOL JSM 840A SEM (Tokyo, Japan). Representative micrographs were taken.
To visualize actin in the amnion, whole mounted embryos were stained with fluorescein isothiocyanate (FITC)-phalloidin (Sigma, Poole, UK) and examined by confocal scanning laser microscopy (CSLM). Embryos were fixed in ovo between Stages 10–18 in 4% paraformaldehyde in phosphate buffered solution (PBS). Embryos were dissected out and replaced in 4% paraformaldehyde/PBS at 4°C for a minimum of 30 min, and the chorion was removed. Embryos were washed with PBS for 1 hr with three changes at room temperature. Specimens were stained with FITC-phalloidin (0.05 mg/mL stock solution diluted 1:50 with PBS) for 1 hr in the dark at room temperature. Finally, the specimens were washed three times in PBS (in the dark, at room temperature) and mounted onto cavity slides using an aqueous anti-fade mountant and viewed under an MRC 600 CSLM (Bio-Rad Laboratories, Richmond, CA). Representative digital micrographs were recorded.
Effect of Cell-Permeable C3 Transferase
To investigate the effect of blocking Rho on chick amnion development between Stages 14–18, 10 μL of either 6.25, 25, 50, or 100μg/mL permeable C3 transferase (Universal Biologicals, Cambridge, UK) was applied to the surface of the amnion immediately after overlying chorion was removed in order to obtain a dose response of exoenzyme concentration after 2 hr incubation. A control group had 10 μL of distilled water added instead of the exoenzyme. Based on the pilot dose response effects of blocking Rho on chick amnion, development between Stages 14–18 was examined using 25 μg/mL permeable C3 transferase. Embryos were returned to incubator conditions for a period of 24 hr. At 24 hr, the embryos were observed and the status of the amnion recorded as either open, closed, or undetermined. Embryos that were dying/dead were noted. Surviving embryos were fixed in ovo with 4%PFA/PBS, dissected out and placed in fresh fixative. Representative digital micrographs were recorded. Embryos were subsequently stained for actin as previously described and representative digital micrographs were taken.
Transmission Electron Microscopy
To examine the ultrastructure of the leading edge of the developing amnion at Stage 13/14, embryos were prepared for transmission electron microscopy (TEM) using a modification of the protocol described by McDonald (1984). Embryos were fixed in ovo with 1% glutaraldehyde in 0.1-M sodium cacodylate buffer (2–3min), dissected out, and replaced in fresh fixative for a minimum of 8 hr. Specimens were rinsed 0.1-M sodium cacodylate buffer for 1 hr and treated in 0.8% potassium ferrocyanide in 0.1-M sodium cacodylate buffer for 15 min. Specimens were postfixed in 0.5% osmium tetroxide in 0.1-M sodium cacodylate buffer for 1 min and dehydrated through a series of alcohols. Specimens were processed in propylene oxide, propylene oxide/epoxy resin mixes (TAAB, Berkshire, UK), orientated in flat bed moulds and embedded in epoxy resin. Resin-embedded specimens were cured at 60°C for 48 hr. Ultrathin sections (70 nm) of the leading edge of the developing amnion (previously identified with 1 μm semi-sections of the embedded tissue stained with 1% Toluidine Blue) were cut using a Ultracut E ultramicrotome (Reichert-Jung, Austria), collected on copper grids and stained using uranyl acetate and lead citrate. Sections were viewed on a JEOL JEM-100CX II transmission electron microscope (Tokyo, Japan) and representative micrographs taken.
To visualize the leading edge of the amniotic folds, embryos at Stage 18 were stained in ovo with a saturated solution of Neutral Red. Before staining, the chorion was removed using watchmaker forceps (5 Inox). Approximately 10 μL of neutral red was dropped onto the surface of the embryo and left for 1 min. Excess neutral red stain was washed off with distilled water and removed with a micropipette. This process was repeated until the excess stain was removed. Embryos were observed using a Leica MZ6 dissection microscope and representative digital images recorded using an attached Nikon Coolpix digital camera (Tokyo, Japan).
FITC-phalloidin was used to label actin and determine the possible existence of an actin cable mechanism at different stages of amniogenesis and at closure of the amnion (n = 14). Staining for actin at the leading edge of the amnion head fold was positive at Stage 10 against the background staining of cortical actin (Fig. 1A). At Stage 13, positive staining for actin was evident in the deltoid region, the elongated leading edges and the nodule (Fig. 1B). The circumference of the amniotic aperture stained positive for actin at Stage 18, although several nodules along the perimeter of the amniotic aperture were apparent, which also stained positive for actin (Fig. 1C). Intense staining for actin was evident in the nodule and leading edge of the amnion in comparison to surrounding amniotic tissue (Figs. 1D,E).
Evidence for Rho-involvement in the actin cable mechanism during amniogenesis was investigated by inhibiting Rho with cell-permeable C3 transferase between Stages 14–18. A dose-dependent study showed that a concentration of 6.25 μg/mL did not appear to affect amniogenesis, and concentrations of 50 μg/mL and 100 μg/mL (n = 37) resulted in a high mortality rate. A baseline concentration of 25 μg/mL cell permeable C3 (n = 59) retarded the development of the amnion compared with one embryo from the control group (Table 1) where the amnion failed to close after 24 hr (n = 29). Embryos 24 hr after the addition of 25 μg/mL permeable C3 transferase revealed evidence that amnion progression was retarded, and the amniotic aperture remained patent.
Table 1. The effect of 25-μg/mL cell-permeable C3 transferase on chick amniogenesis at Stages 14–18 compared with control (distilled water, figures in brackets)
The surface morphology during the early development of the amnion was examined at various stages ranging from 10–18 of the embryo using SEM (n = 12). A crescent-shaped strip of elongated cells, the ectamnion, was seen above the head region in the ectoderm of the area pellucida at Stage 10 (Figs. 2A,B). At Stage 17, the amnion head fold had progressed posteriorly. A raised deltoid region consisting of flattened cells with indistinct boundaries (contrasting to the surrounding cobblestoned cells of the amnion) was apparent in the mid-line region. Elongated cells adjacent to either side at the leading edge of the amnion were seen either side of this deltoid structure (Fig. 2C, arrowheads). A nodule was apparent projecting from just below the deltoid area of cells; the nodule was composed of cells and cellular debris. Neither the deltoid-shaped area of cells nor nodule was apparent at Stage 10. Amnion closure was almost completed at Stage 18; elongated cells were arranged around the circumference edge of the dorsal opening of the amnion with the nodule and deltoid region still evident (Fig. 2D).
To establish the arrangement of actin filaments identified with FITC-phalloidin and to reveal the ultrastructural arrangement of the filaments within the leading edge cells and nodule of the developing amnion, tissue was examined using TEM (n = 6). The elongated cells at the circumferential edge of the amnion aperture revealed linearly arranged microfilaments (˜6 nm in diameter) running parallel to the long axis of the cells and concentrated in the cytoplasm subjacent to the plasma membrane of the leading lateral edge (Fig. 3A). This linearly arranged band of microfilaments runs throughout adjacent neighbouring cells around the perimeter of the amniotic aperture. Lamellipodia were absent from these perimarginal cells (Fig. 3B). Neutral red staining of the whole embryo at Stage 18 (n = 3) indicated numerous dead and dying cells in the raised deltoid-shaped area of ectamnion cells in the mid-line region of fusion and within the nodule of cells below this region. Staining was less intense around the leading edge of the closing amnion (Fig. 3C). Apoptotic cells, recognized by their condensed chromatin and shrunken appearance, were evident within the nodule of accumulated cells (Fig. 3D). A mesh of cytoplasmic actin filaments was dispersed throughout the nodule (Fig. 3E).
As early as Stage 10, an actin cable is seen in the amnion head fold, with cells elongated in alignment. The cells maintain this orientation at the circumference of the amnion aperture throughout amniogenesis suggesting a degree of circumferential tension. This actin cable appears to maintain spatiotemporal integrity throughout the development of the amnion until closure. This study has shown that the actin cable involved in amnion closure during amniogenesis is required to close the membrane over an area in excess of 2-mm diameter, and exceeds the largest wound size currently reported as healing via an actin cable (2-mm diameter, Danjo and Gipson,1998). Wound repair of the amnion membrane after surgical puncture in the chick (MacManus et al.,2006) demonstrated that the chick amnion retains the capability of forming an actin cable; using this contractile mechanism to heal quickly, without a scar, and re-establish the epithelial barrier avoiding persistent amniotic fluid leakage.
Evidence in support of the idea that an actin cable mechanism drives amnion development was the inhibition of amniogenesis by the Rho blocking agent cell-permeable C3 transferase. Amniogenesis inhibition occurred more consistently when the blocking agent was administered early in amnion development (approximately Stage 14/15). Earlier stages could not be used in this study due to the technical difficulty of removing the overlying chorion. The embryonic environment is already bathed in fluids such as amniotic fluid and albumin, which will affect the concentration/amount of permeable C3 reaching the cellular edges of the developing amnion. This may explain the closure of some amnions treated with 25 μg/mL of cell-permeable C3 transferase, but addition at later stages had little effect on the closure. Alternately, contractile forces may play a greater role in amniogenesis in the early development stages of the amnion. Amniogenesis might continue after the addition of the blocker as it becomes exhausted during ADP-ribosylation of Rho proteins, thus more stress fibres are able to form at the advancing edge of the developing amnion. Aepfelbacher et al. (1997) showed that that cellular microinjection of non-cell permeable C3 transferase to block human umbilical vein endothelial cell migration in an in vitro wound repair experiment was less effective at longer times post-wounding perhaps due to the shorter half-life of C3 in the cytosol. This suggests C3 becomes less effective the longer it remains in the cell.
Previous studies (Davidson,1977; Overton,1989), did not describe the surface morphology of the amnion at early developmental stages. The distinct crescent-shaped arrangement of elongated cells in the area pellucida noted at Stage 10 suggests the beginning of actin cable formation. As the amniogenesis progresses posteriorly, it is also apparent that the amnion does not contact the underlying ectodermal surface of the embryo. Amniogenesis was usually completed by Stage 18 although intra-embryo temporal variation of amniogenesis has been reported (Davidson,1977). The cellular morphology of amniogenesis seen with the SEM mirrored the spatial pattern of actin staining seen with the FITC-phalloidin labelling throughout the developing stages.
The perimarginal cells of the closing amnion did not possess lamellipodia, and filopodia were poorly developed suggesting that lamellipodial crawling is not involved in closure during amniogenesis. Closure of adult epithelial wound healing is orchestrated via ultrastructural mechanisms such as lamellipodial crawling (Jacinto et al,2001), a process which requires a tissue substratum (Dipasquale,1975; Farooqui and Fenteaney,2004). Fusion of embryonic tissue, such as dorsal closure in the Drosophila embryo, is also thought to require a tissue substratum, the amnioserosa (Kiehart et al,2000). Another study (Redd et al,2004) demonstrated that an actin cable alone is not involved in embryonic epithelial wound closure of Drosophila, but that lamellipodia and filopodia assist cells in crawling over the tissue substratum. It is possible that such ultrastructural protrusions are absent from the leading edge of the amnion perhaps because of the absence of a substratum.
Overton (1989) examined chick amniogenesis from Stage 14 through to Stage 18 and showed apoptotic cells within the region of fusion, and suggested that this region exerts tension, which could bring together the amnion folds. In this study, the nodule of accumulated cells (or globular projection), which lies immediately below the deltoid region of fusion, absent at Stage 10 but noted at later stages, was also found to have a high density of apoptotic cells. Apoptosis is a necessary function during embryogenesis, helping to shape the structure of tissues and remove cells that are no longer necessary; it is conceivable that apoptosis plays a role in amniogenesis. A mesh of actin filaments was observed within the nodule and studies have shown that apoptotic cells are removed from simple epithelial by contraction of actin rings around these cells (Rosenblatt et al,2001), supporting the hypothesis that certain actin filaments within the nodule are acting as an elimination mechanism of apoptotic cells. It could be argued that this is the fate of some of the perimarginal cells as the amniotic aperture is shrinking in size and these leading edge cells are therefore diminishing in number.
Although this study demonstrates the existence of an actin cable mechanism during the development of the amnion similar to that seen in embryonic wound repair, it is possible that amniogenesis does not proceed, nor is reliant upon, this one mechanism alone. It may operate in conjunction with proliferation (Miller et al,1994), apoptosis (Rosenblatt et al,2001) or embryonic movement (Lillie,1903).
The authors would like to thank Professor Forbes Davidson for permission to retain a copy of his PhD thesis.