The reopening process of the esophagus in the normal chick and the crooked neck dwarf mutant

In an earlier publication (Allenspach & Hamilton, 1962) an account was given of chick esophageal development and histogenesis up to 10 days of incubation. The esophagus becomes completely occluded at 5 days (stage 26, Hamburger & Hamilton, 1951) and reopens rapidly at days (stages 33 –34). Detailed observations were made of the histological changes in the organs as correlated with the localization of ribonucleic acid and alkaline phosphatase.

The occlusion of the developing esophagus was apparently due to collapse and adhesion of roof and floor epithelia to form a solid epithelial bar and not to cell proliferation, as stated in Lillie’s Development of the Chick (Hamilton, 1952, pp. 384 –5). This process was the object of further investigation in which the role of proliferation in occlusion was determined by the use of colchicine. Results indicated that cell proliferation is not involved in the closing process (Allenspach, 1964).

The mechanisms which control the reopening process have also been partially elucidated. Reopening is initiated by the appearance of bilateral primary vesicles. Subsequent fusion of primary and secondary vesicles with concomitant desquamation of degenerating epithelial cells restores the lumen (Allenspach & Hamilton, 1962; Allenspach, 1964). During this period the esophageal epithelium changes from non-oriented cells in the occluded region to a stratified epithelium similar to that of the definitive organ. Reorganization is initiated around the most lateral vesicles. Another subtle mechanism apparently operates in the reopening process. Orientation of dorsal, epithelial cells suggested they were the object of vertical tension. By treating the esophagus with versene, a compound which disaggregates tissues, a cleft was induced between roof and floor. This suggests that (1) adhesions play a significant role during occlusion, and (2) forces exerted on the dorsal epithelium assist in separating the roof from the floor (Allenspach, 1964).

The present investigation was prompted by the report of Pun (1954) that the esophagus of the crooked neck dwarf mutant embryo is still largely occluded beyond 12 days. He suggested that affected embryos suffer from chronic starvation since they cannot swallow nutritious amniotic fluid that is essential for normal development (Pun, 1954). The object of this study is to approach an understanding of the controlling mechanisms of esophageal histogenesis by a comparison of mutant and normal embryos at critical stages of occlusion and reopening. The results obtained support the thesis that, even though vesiculation occurs, epithelial degeneration plays a major role in the reopening process which may itself be influenced by tensions not yet understood.

The material used in this study was collected at the University of Connecticut during the summers of 1962 and 1963. (The author is deeply indebted to Dr Louis J. Pierro and the Department of Animal Genetics, University of Connecticut, Storrs, Connecticut, for their co-operation in collecting and shipping the material used in this study.) Eggs from a cross of carriers heterozygous for the crooked neck dwarf mutant were incubated for days. Embryos were recovered, fixed and shipped in Bouin’s fixative and upon receipt were staged (Hamburger & Hamilton, 1951) and stored in 70% ethanol.

The nature of the study required recognition of mutants as early as possible. Embryos of 9 –10 days and older were recognized by edema, spindly legs, and abnormally short necks (Herrmann, Clark & Landauer, 1963) and were collected along with a few normal embryos. Since the gene involved is an autosomal recessive, earlier embryos were necessarily studied in greater numbers to ensure an accurate histological picture of esophageal development.

For this investigation the neck region of the embryo was removed, dehydrated, cleared in benzene, infiltrated in 56 –58 °C paraffin-bayberry wax mixture, and sectioned at 7 μ. With the exception of one series stained with toluidine blue, all were stained by a modified haematoxylin method.

The crooked neck dwarf mutation apparently had no effect on early esophageal organogenesis in embryos of stage 29 and younger. The occluded, vesiculated, and posterior open regions appear unchanged in all embryos studied.

Mutant embryos could be recognized from stage 30 onward by the appearance of deeply chromatophilic, degenerating mesenchymal cells outside the basement membrane as shown in Plate 1, fig. A. This degeneration results in deterioration of the submucosal region to a greater or lesser extent and occasionally the esophageal epithelium is partially destroyed (Plate 1, fig. A). Degradation always progresses from the submucosal layer through the basement membrane into the epithelial layer. More extensive deterioration is observed in the virtual destruction of the epithelial wall of the esophagus at stage 32 (Plate 1, fig. B). Lysed epithelial cells can be seen floating in the lumen.

With the exception of the ‘degenerate cell’ marker in the mutant mesenchyme, histological examination of all embryos within stages 30 –33 demonstrated normal esophageal organogenesis. Briefly stated, the closing process is brought about by fusion of roof and floor epithelia. The vesiculated region contains many mitotic figures adjacent to the numerous vesicles, an indistinct basement membrane limiting the dorsal epithelium, and prominent nucleoli. Rearrangement of the epithelium from an unorganized to a pseudostratified type occurs around the primary vesicles. The posterior open region, with pseudostratified columnar epithelium, is limited by a distinct basement membrane. As the lumen increases in diameter the epithelium is reduced to a low stratified type characterized by few mitoses and small nucleoli.

It is noteworthy that there is little secretion exuding from the roof of the posterior open esophagus in the mutant embryos (Plate 1, fig. C). Small amounts of basophilic secretion appear at the free surface of the organ, but the lumen is never full of mucus, as it is in normal embryos.

Up to stage 34, development of the mutant esophagus is apparently normal; closing by fusion and vesiculation occur without change. However, significant differences in the reopening process are observed.

Since major histological changes occur at this stage, a detailed histological study along the cephalo-caudal axis of the mutant esophagus is of interest.

A representative section of the broad, occluded epithelial bar just behind the pharynx is shown in Plate 1, fig. D. Vesicles appear posteriorly, but the intervesicular epithelium shows no degeneration (Plate 1, fig. E). As the broad esophagus narrows, the middle segment becomes vesiculated (Plate 2, fig. A). The two primary, bilateral vesicles exist with several secondary vesicles in the medial epithelial bridge which unites the roof and floor, but no degeneration occurs within the organ (see Plate 1, fig. F, or Plate 2, fig. A). However, the mutant esophagus does contain, with some variability, degenerating cells, always in the posterior part of the vesiculated region (Plate 2, fig. I). The dearth of degenerative cells contrasts sharply with the massive desquamation observed in the normal embryo where comparable regions are already reopened (Plate 2, figs. E, F), and where full-scale degeneration has proceeded anteriorly into the broad epithelial bar just posterior to the tracheal bifurcation (Plate 2, figs. C, D). The most extensive degeneration observed through stage 38 is observed in an embryo of stage 34 (Plate 1, fig. G).

The vesiculated region extends posteriorly by the persistence of a thin epithelial bridge connecting the roof and floor in mutant embryos (Plate 1, fig. H). There is no evidence of pycnosis, and the large nucleoli reflect a period of syn thesis rather than degradation (Plate 2, fig. H). Also, the epithelial bridges are coated with a generous supply of extracellular matrix. Projecting from the floor in the reopened region is a pinnacle of living epithelial cells, a remnant of the median cellular connexion (Plate 1, fig. I). This phenomenon prevails in the open region even at later stages.

Perhaps the most curious phenomenon in the vesiculated region is the cleft observed infrequently in some cytoplasmic bridges (Plate 1, fig. F). The cleft partially severs, for a short distance, the connexion between roof and floor. Cells adjacent to the break are living, with intact nuclei and prominent nucleoli, even though neighbouring cell membranes may be ruptured as a result of the separation. Epithelial cells dorsal to the cleft are vertically oriented along an indistinct, basement membrane. The indistinct, dorsal basement membrane, best observed in Plate 3, fig. I, and also directly above the cleft in fig. J, same plate, is characteristic of the region and the stage of development.

While a broad epithelial bar exists, it is not lengthy, for it is rapidly penetrated by vesicles. In the mutant it becomes multi-vesiculated and resembles the normal esophagus at stage 34 (see Plate 1, fig. E).

Frequent mitoses are observed in the unorganized intervesicular epithelium, and large, basophilic nucleoli are common. There was no evidence of pycnosis.

The extent to which epithelial cells degenerate is shown more posteriorly in the vesiculated region (Plate 2, fig. B). Only a small population of cells has degenerated (Plate 2, fig. I). Occasionally free-floating masses as well as projecting clusters of living cells are observed in the vesicles. That this embryo is a mutant is confirmed by the presence of deeply basophilic, degenerating mesenchyme cells below the basement membrane.

The open region is partially bisected by a mid-ventral projection of living epithelial cells (cf. Plate 1, fig. I). This is a remnant of the median, cellular bridge observed in the vesiculated region. More posteriorly in the open region the principal changes include further development of the longitudinal folds and the rearrangement of epithelium into a low stratified type. In addition, numerous vesicles are observed within the thin wall of stratified epithelium.

Occasionally a short occluded zone is observed in the esophagus of mutant embryos; however, the anterior region is typically represented in Plate 3, fig. A. The organ is compartmentalized by extremely long, thin epithelial bridges which are coated with considerable extracellular matrix. More posteriorly the organ is greatly restricted (Plate 3, fig. B) and characterized by nuclei with pro-minent nucleoli and mitotic figures adjacent to the vesicles. Degenerate, desquamated masses of epithelial cells are lacking.

Of questionable significance is the presence of an epithelial cleft in one stage 36 mutant embryo, similar to that described earlier (cf. Plate 1, fig. F).

The open region and the small vesicles within the stratified epithelium of the stellate-shaped esophagus have been described earlier.

Mutant embryos of stages 37 –38 do not exhibit a distinct region of occlusion, but the broad anterior esophagus has a reduced diameter by virtue of a median connexion between roof and floor (Plate 3, fig. G). The median epithelial bridge contains small vesicles. Some intervesicular epithelial cells located farthest from the basement membrane are highly vacuolated, but their nuclei never show indications of pycnosis (Plate 3, fig. G, inset).

The vesiculated region more posteriorly takes on the character shown in Plate 3, fig. D. The large, bilateral vesicles never fuse with medial vesicles, and the latter enlarge to fashion thin, cellular bridges between roof and floor. The majority of the cellular strands are living (Plate 3, fig. K(a)); occasionally degenerate cells are observed in lateral epithelial connexions while medial cellular strands persist (Plate 3, fig. K(b)). Degeneration is never extensive enough to effect reopening since the organ is still occluded in 18 –19-day embryos (Pun, 1954). A representative section showing well-developed longitudinal folds in the posterior esophagus at stage 38 is seen in Plate 3, fig. H. Note that the median part of the organ, including one longitudinal fold, is occluded. Small vesicles occur in the medial, occluded epithelium and are particularly abundant within the epithelium of the longitudinal folds. Typically the esophagus is reopened at this time (cf. Plate 3, fig. C).

The open region remains unchanged, characterized by remnants of cellular connexions protruding from the floor, groups of free-floating cells in the lumen, and numerous vesicles within the stratified epithelium (Plate 3, fig. F).

Our observations support those reported on older material by Pun (1954) but give details on earlier stages. Studies on 12- and 18-day mutant embryos revealed that the size of the esophageal lumen was slightly smaller and, in some cases, completely obliterated while the normal esophagus was open along its length. Cellular strands ‘suggesting closure or anastomosis between all folds of the esophagus’ were observed connecting walls of longitudinal grooves (Pun, 1954, p. 107).

Esophageal histogenesis in the mutant embryo resembles the normal pattern up to stage 33. Beyond this stage, however, significant differences are noted; perhaps the most important is the failure of the esophagus in the mutant embryo to reopen at the normal time.

Vesiculation and degeneration are important in reopening the normal, occluded organ. However, that reopening is not necessarily caused by vesiculation alone is shown by observations in the mutant embryo where vesiculation occurs extensively in the occluded epithelium through stage 38 to produce a broad, much restricted, multi-vesicular organ (Plate 3, figs. G, H), yet the roof and floor fail to separate and reopen the organ. This evidence precludes reopening of the esophagus by vesiculation alone and points to epithelial degeneration as an essential mechanism for removing connexions between roof and floor.

One curious phenomenon observed is the appearance of apparently natural clefts in the epithelial bridge between roof and floor in mutant embryos of stages 34 and older (Plate 2, fig. F; Plate 3, fig. J). In an earlier work, treatment of the occluded, embryonic esophagus with versene induced separation of roof and floor into clefts similar to those observed in mutants (Plate 2, fig. G). Versene-induced clefts were noted at stages before epithelial degeneration commences, namely stages 33 –34 (Allenspach, 1964). The interpretation was offered that a force exerted on the esophageal epithelium must exceed the cellular adhesiveness to effect reopening. The clefts noted in mutant material occur at stage 34 and later and always in regions where epithelial cells appear to be subject to tensions, as noted by their specific orientation near a subjacent dorsal, indistinct basement membrane (Plate 3, fig. J). These clefts may indicate that an intrinsic force exists in the mesenchyme and plays a role, along with degeneration, in the separation of roof and floor. The possibility of these clefts being artifacts is not justified since fixation of other tissues was excellent.

The insight gained from studies on earlier developmental processes helps in the interpretation of observations reported on 18 –19-day mutant embryos (Pun, 1954). Pun discovered that the entrance to the esophagus was occluded in 18 –19-day mutant embryos. Some of the longitudinal grooves were completely closed while others were united by anastomosing, epithelial strands which greatly reduced the lumen (Pun, 1954). The persisting occlusion of longitudinal grooves (Plate 3, fig. H) and cellular strands (Plate 3, fig. F) is due to failure of epithelial cells to degenerate at 8 days. The lack of occluded longitudinal grooves and cellular strands is evident in the normal esophagus (Plate 3, fig. E). The net result is the same : chronic starvation due to failure of nutrients to pass from amnion to stomach.

Very little secretion was observed in the mutant esophagus, resembling the situation in the crooked neck dwarf turkey embryo where the epithelium is apparently non-secretory (Asmundson & Pun, 1956). Considerable mucus is secreted by normal epithelial cells, but it is believed that the secretion, acidic mucopolysaccharides by nature, is not responsible for the reopening process (Allenspach & Hamilton, 1962). On the other hand, degenerate epithelial cells are sloughed into the lumen, and their by-products may be responsible for autolysis of other epithelial cells. The role, if any, of secretory products in esophageal reopening remains unclear.

Since the crooked neck dwarf gene is lethal, the question arises as to which abnormality of the pleiotropic pattern is the primary cause of death. Several organ systems are known to be involved. That mutant embryos suffer from chronic starvation because of their inability to swallow amniotic fluid has been advanced by Pun (1954). However, a similar mutant in the turkey results in harmful effects before esophageal atresia prevents starvation (Asmundson & Pun, 1956). Recent results of experimental work on normal and abnormal muscle have led Herrmann et al. (1963) to declare the ‘starvation hypothesis’ an unlikely cause of abnormal development. Observations reported here and others (Hadorn, 1961; Allenspach, unpublished) support this contention. Abnormal developmental processes of the lethal embryo can be recognized before any effects of starvation occur.

Esophageal organogenesis and more specifically the reopening process may be controlled by a minimal number of mechanisms : cell degeneration, vesiculation and intercellular tensions. To understand these controlling mechanisms further in highly organized tissues requires, first of all, an understanding of cell contacts and intracellular aspects of cell degradation. The fact that some epithelial cells in the esophagus degenerate, while adjacent cells do not, causes speculation as to what happens at the molecular level within these cells. These problems are currently under investigation.

1. On a étudié le développement de l’œsophage chez le mutant nain à cou tordu, entre jours et 12 jours d’incubation (stades 26 –38).

2. L’histogenèse œsophageienne du mutant ressemble à celle du développement normal jusqu’au stade 33.

3. Le vésiculisation seule ne provoque pas la réouverture de l’œsophage qui dépend de la participation de la dégénérescence de l’épithélium pour achever ce processus.

4. Les recherches sur la dégénérescence épithéliale du mutant font ressortir son gradient caudo-céphalique dans le processus de réouverture.

5. L’apparition de fissures dans l’épithélium intervésiculaire suggère que la tension joue un rôle dans le processus de réouverture.

6. Les segments occlus et les cordons épithéliaux persistants del œsophage embryonnaire du mutant du 18e jour, résultent de défectuosités critiques du développement au 8e jour de l’incubation (stade 34).

7. Le mutant nain à cou tordu présente un marqueur histogénétique sous la forme de ‘cellules mésenchymateuses dégénérées’, dès le stade 30.

This investigation was supported by a special grant from Albright College. The technical assistance of Linda J. Doerr and Russell Reidinger is hereby acknowledged.

Special thanks are due to Mr Forest Moyer, Chief Microphotographer, Reading Hospital, Reading, Pennsylvania, whose generous time and valuable criticisms are deeply appreciated.

Most of the figures are photographs of cross-sections of the esophagus in the crooked neck dwarf embryo. Several other photographs are included for comparative purposes. Unless otherwise stated, sections were stained with iron haematoxylin. The magnification is indicated by a scale-line on each photograph.