/ . Embryo!, exp. Morph. Vol. 54, pp. 17-35, 1979
Printed in Great Britain © Company of Biologists Limited 1979
\"f
Effect of oxygen concentration on morphogenesis
of cranial neural folds and neural crest in cultured
rat embryos
By GILLTAN M. MORRISS 1 AND D. A. T. NEW 2
From the Department of Human Anatomy, Oxford, and
the Physiological Laboratory, Cambridge
SUMMARY
Rat embryos, 9^ days old, cultured with a 5 % or 10% O2 gas phase underwent normal or
near-normal cranial neurulation; however, culture at 20% or 40% O2 resulted in abnormal
morphogenesis of the cranial neural folds from the 9-somite stage onwards, and the brain
tube frequently failed to close. Normal morphogenesis was characterized by a narrowing
V-shaped profile, development of a slightly concave neuroepithelial surface, and formation
of a sharp mediad curvature of the most lateral region prior to midline apposition and fusion.
These morphogenetic events were related to cellular changes within the neuroepithelium,
namely cell death, onset of neural crest cell migration, and loss of apical microfilament
bundles from the most lateral cells. In 20% and 40% O2-cultured embryos, failure of
curvature of the neuroepithelium was associated with failure or retardation of the related
cellular changes; it may therefore have been due to the maintenance of an excessive
rigidity which opposed the forces involved in bringing about the final stage of brain-tube
formation.
Mitochondria in normal (low O2 and in vivo) embryos were of the anaerobic type, having
few cristae; in high O,-cultured embryos they were of the characteristic aerobic type, indicating an adaptation to the abnormal environment.
INTRODUCTION
Recent modifications in the technique of whole-rat-embryo culture have
enabled 9-^-day (presomite/early head-fold stage) embryos to undergo normal
morphogenesis during 48 h or more of in vitro development (New, Coppola &
Cockroft, 1916a, b). One of the essential conditions for normal development was
found to be a low (5 %) oxygen concentration in the gas phase during the first
24 h of culture. This period corresponds to the period of cranial neural-fold
morphogenesis; a 20% O2 gas phase was found to result in a high proportion
of embryos with misshapen or unclosed brain tubes. After the brain tube has
closed, i.e. after the first 28-30 h of culture, 20 % O2 supports normal morphogenesis and is the most suitable oxygen level during the subsequent period.
1
Author's address: Department of Human Anatomy, South Parks Road, Oxford OX1
3QX, U.K.
2
Author's address: Physiological Laboratory, Downing Street, Cambridge CB2 3EG, U.K.
18
G. M. MORRISS AND D. A. T. NEW
This study is an analysis of the differences in cranial neural-fold morphogenesis of embryos cultured in the presence of different concentrations of
oxygen. It was undertaken in order to discover the nature of the cellular events
associated with abnormal morphogenesis in rat embryos cultured at oxygen
levels which are excessively high only for this specific stage of in vitro
development.
MATERIALS AND METHODS
CFHB strain rat embryos were explanted in Hanks saline on the afternoon
of day 9 of gestation (day of positive vaginal smear = day 0) and Reichert's
membrane was opened. They were cultured at 38 °C in 30 ml cylindrical
bottles, each bottle containing three embryos, in 3 ml immediately-centrifuged,
heat-inactivated serum (Steele & New, 1974) containing 50 /*g/ml streptomycin.
Litters were evenly divided among the bottles to be gassed with different gas
mixtures. All embryos were at the late presomite stage prior to culture. The
bottles were rotated and gassed continuously throughout the period of culture
(New & Cockroft, 1979), with one of the following gas mixtures: 5 % O 3 /5 %
CO,/90%N 2 ; 1 0 % O 2 / 5 % C O 2 / 8 5 % N 2 ; 20 % O 2 /5 % CO 2 /75 % N 2 ; 40%
O 2 /5 % CO 2 /55 % N 2 . Cultures were terminated after 23-24 h, 28-31 h or
48 h. Embryos cultured for up to 31 h were washed in Hanks saline and fixed
immediately in 2-5% cacodylate-buffered glutaraldehyde (0-1 M, pH 7-2, with
2mM CaCl2). They were then washed in buffer, and the yolk sac and amnion were
removed. At this stage the somites were counted, and the cranial neural fold/
brain tube development was assessed in relation to somite number. They were
subsequently prepared either for scanning electron microscopy or for light
microscopy and transmission electron microscopy. Embryos cultured for 48 h
were examined without fixation, after removal of the membranes.
Scanning electron microscopy (SEM)
Embryos were either prepared whole, or after transverse cuts through the
head region had been made (Fig. 11). They were dehydrated in graded acetones,
critical-point-dried, mounted on aluminium stubs with double-sided Sellotape,
coated with gold in a sputter coater, and viewed in a Cambridge Stereoscan
scanning electron microscope. Forty embryos of equivalent stages explanted
directly from day-10 pregnant rats were prepared by the same method for
comparison.
Light microscopy (LM) and transmission electron microscopy (TEM)
Embryos were post-fixed in cacodylate-buffered osmium tetroxide, washed,
and dehydrated to 70 % alcohol. They were photographed whole for reference
during subsequent sectioning. Dehydration was then completed, and the
embryos were embedded individually in Spurr resin. They were orientated at
this stage, and during cutting, as indicated in Fig. 11. Sections, 1 fim thick, were
Oxygen and cranial neurulation
19
mounted on. glass slides and stained with toluidine blue for light microscopy.
Ultrathin sections were stained with uranyl acetate and lead citrate for electron
microscopy. Five embryos of 10-14 somites explanted directly from day-10
pregnant rats were used for comparison.
RESULTS
All cultured embryos were assessed as having normal cranial neural-fold
shape in relation to somite number at eight somites, in comparison with embryos
which had developed to this stage in vivo. During the 8- to 16-somite stage in
vitro, each pair of somites was added, on average, in just less than 2 h. The
appearance of embryos examined with the dissecting microscope immediately
after culture was as follows.
5% O2-cultured embryos: 18 embryos were cultured for 23-31 h, and the
somite number ranged from 8 to 16. Sixteen embryos were assessed as having
normal cranial neural-fold shape; one embryo with 9 somites and one with
12 somites had cranial neural folds similar to those of in vivo embryos with 8
and 11 somites respectively.
10% O2-cultured embryos: 6 embryos were cultured for 23-30 h. Three
embryos with 8-10 somites had normal cranial neural-fold shape, while three
embryos with 10-13 somites showed an open region which was slightly too
broad for the somite number. Seven embryos which were cultured for 48 h
(22-25 somites) had normal, closed brain tubes.
20% O2-cuItured embryos: 24 embryos were cultured for 23-31 h, and the
somite number ranged from 8 to 17. Five embryos with 8 or 9 somites were
assessed as normal with respect to cranial neural-fold shape; in four embryos
with 10 somites the open region was too broad; 11 embryos with 11-14 somites
showed a broad open area (cf. narrowing and closing at this stage in vivo); two
embryos with 16 somites had normal, closed brain tubes; two embryos with
17 somites had large open areas. Six embryos were cultured for 48 h (21-26
somites): four of these showed normal, closed brain tubes, while two showed
persistent open areas.
40% O2-cultured embryos: 18 embryos were cultured for 23-31 h, and the
somite number ranged from 8 to 16. Three embryos with 8-9 somites were
assessed as normal; two embryos with 9-10 somites had cranial neural folds
more broadly separated than in vivo embryos of this stage; in 13 embryos
with 11-16 somites the cranial neural folds remained widely separated, and
the anterior forebrain region was not apposed prior to 12 somites (cf. 10-11
somites in vivo). Six embryos were cultured for 48 h (22-25 somites); only one
of these showed a closed brain tube.
20
G. M. MORRISS AND D. A. T. NEW
Scanning electron microscopy of whole embryos
Closure of the cranial neural tube in in vivo rat embryos was found to follow
a closely similar pattern to that of the hamster and mouse (Waterman, 1976).
Figure 1 indicates the different regions of neural ectoderm which are evident
in an 8-somite rat embryo. The neural tube is closed at the level of somites
4-8, and open anteriorly and posteriorly. Initial apposition within the anterior
Fig. 1. 8-somite embryo, SEM. Regions of neural ectoderm: (1) anterior forebrain
region; (2) broad cranial neural fold (BCNF) region; (3) lower hindbrain (myelencephalic) region; (4) closed spinal (cervical) neural tube; (5) open spinal neural
groove, with primitive streak posteriorly, c: 'crumpled' region of surface ectoderm;
fa: forebrain apposition point; pos: preotic sulcus.
FIGURES
2-6
Fig. 2. Early 9-somite embryo, in vivo: open BCNF area forms a narrower V shape
than at 8 somites. Arrows indicate forebrain apposition movement.
Fig. 3. Late 9-somite embryo, in vivo: more anterior view than Figs 2 and 4-6, to show
the two sides of the forebrain apposition point (arrows) and anterior forebrain
region approaching each other medially. The neural epithelium has developed a slight
concavity, and the V shape has narrowed further.
Fig. 4. 10-somite embryo, in vivo: spindle-shaped opening of BCNF region, with
fused forebrain apposition point; further narrowing of the V shape and increased
concavity of the surface.
Fig. 5. 11-somite embryo, in vivo: narrower spindle-shaped opening, narrower
V shape of neural epithelial surfaces.
Fig. 6. 12-somite embryo, in vivo: further narrowing of spindle-shaped opening,
with apposition anteriorly and posteriorly.
Oxygen and cranial neurulation
21
22
G. M. MORRISS AND D. A. T. NEW
100 fxm
Fig. 7. 9-somite embryo, 20% O2-culture: forebrain apposition region normal,
posterior (lower) BCNF region excessively broad.
Fig. 8. 13-somite embryo, 20% O2-culture: the kite-shaped opening is characteristic
of 11-somite and older embryos cultured at this O2 level (compare with the small
opening in a normal 13-somite embryo, Fig. 14).
Fig. 9. 17-somite embryo, 20% O2-culture: opening similar to that of Fig. 8.
Fig. 10. 13-somite embryo, 40% O2-culture: much broader opening than seen in
20% O2-cultured embryos.
(cranial) region occurs at a point between the anterior forebrain and the region
of broad cranial neural folds (posterior forebrain to otic region of hindbrain).
This apposition was observed to occur at the late 9-somite to early 10-somite
stage (Fig. 3). Apposition of the whole anterior forebrain region occurred at
Oxygen and cranial neurulation
23
10-11 somites; since closure of this region was either normal or only slightly
retarded in all cultured embryos, it will not be described further. The region that
was frequently abnormal in shape/somite-number relationship in embryos
cultured with a high oxygen gas phase was the region of broad cranial neural
folds (BCNF). Figures 2-10 show the shape of this area in 9- to 12-somite
embryos in vivo and in 9- to 17-somite cultured embryos.
Closure of the BCNF region of in vivo embryos showed a clear relation to
somite number. In 8-somite and early 9-somite embryos the neural epithelium
formed a broad V shape in profile, the surface on each side being at first flat
(Figs 2, 12). Shortly before apposition was made at the forebrain apposition
point, the surface just posterior (caudal) to it became slightly concave (Fig. 3).
Following neural-fold fusion at this point, the concave area spread caudally
(10 somites, Fig. 4). The open area was spindle-shaped at this stage, and the
V shape of the neural epithelium was broadest at its mid-point. From 10 to
12 somites (Figs 4-6) the V shape narrowed throughout the whole length of
the open area simultaneously, and the lateral edges of the neural ectodermal
surfaces developed a sharp mediad curvature (in addition to the slight concavity
of the whole epithelium) and thus approached each other in the midline, so
that at 12 somites the formerly spindle-shaped opening became a narrow slit
(Fig. 6); at 13 somites only a small opening remained (Fig. 14) and all 14-somite
embryos examined had completed closure of the brain tube.
SEM of whole cultured embryos confirmed the observations of 5 % and
10% O2-cultured embryos made immediately after culture, and also confirmed
that embryos cultured at the higher oxygen concentrations were normal at the
8-somite stage. However, all five 9-somite embryos cultured at 20% or 40% O2
were found to have a broader than normal BCNF region (compare Figs 7 and 2).
The neural epithelial surface of 9- and 10-somite embryos cultured at 20 % O2
frequently failed to develop the slightly concave shape observed in normal
embryos, except close to the points of forebrain apposition and closed hindbrain neural tube. Subsequent stages showed that failure of the two neural
folds to approach each other and become concave persisted in this region, and
the lateral edges failed to develop a mediad curvature, so that 12- and 13somite embryos had a large kite-shaped opening (Fig. 8) that was still present
in several embryos at stages that should have closed brain tubes (14 somites
or more, Fig. 9). Embryos cultured with a 40% O2 gas phase showed an
exaggeration of this effect (Fig. 10).
SEM of cut cranial neural folds
Cuts made in fixed embryos tend to bring about a natural fracture, so that
individual cells remain whole, although they may occasionally be slightly
displaced. All embryos were cut along a line beginning at the junction of mandibular arch and anterior forebrain (Fig. 11). Figures 12-14 show normal morphology of the region of closure in 8-, 10- and 13-somite embryos. At 8 somites
24
G. M. MORRISS AND D. A. T. NEW
the BCNF region has a wide V-shaped profile; at 10 somites the V shape is
narrower, and a mediad curvature is present at the extreme lateral edges. At
the 8-somite stage the neural epithehum is thinner close to the midline than
elsewhere, giving the appearance of a 'hinge' region in relation to the subsequent
narrowing of the V shape. The curved lateral region of neuroepithelium of a
10-somite embryo is shown in greater detail in Fig. 16; the surface curvature
is reflected in an even greater curvature of the basal surface of the neural
Fig. 11. Diagram to show (a) line of cut for SEM (Figs 12-17); (b) line of cut for
LM and TEM (Figs 18-27); arrows indicate the direction of' pull' which would result
from elongation or curvature of the head (see discussion and Fig. 28). The notochord
is indicated by a broken line. (Drawn from a photograph of a 13-somite 20% O2culture embryo fixed prior to embedding.)
epithelium, brought about partly by a progressive decrease in height (length) of
the cells towards the lateral edge, and partly by the predominantly basal
distribution of the nuclei. The presence of pro-neural crest cells with leading
lamellae indicated loss of the basement membrane in this region. The neural
epithelium of embryos cultured with a 20 % O2 or 40 % O2 gas phase was
normal at 8 somites, but the broad V shape persisted at later stages and the
FIGURES
12-15
Figs 12-15. SEM of cut surface, mid-BCNF level.
Fig. 12. 7/8-somite embryo, in vivo. Wide V shape with flat neuroepithelial surface
except the near mid-line, where the epithelium is thinner, n: notochord.
Fig. 13. Late 10-somite embryo in vitro, 5% O2 (normal). The neural epithelium
has thickened, and part of the surface has become concave.
Fig. 14.13-somite embryo, in vivo. Arrow: final open area on outer surface. Double
arrows: neural crest cells. Dorsal neuroepithelial cells (fusion region) are flat or
cuboidal, cf. elongated cells of the lateral walls of the neural tube.
Fig. 15. 13-somite embryo in vitro, 20% O2. The neural epithelium has thickened
normally, but there is no surface concavity, and no narrowing of the V shape.
Oxygen and cranial neurulation
25
26
G. M. MORRISS AND D. A. T. NEW
Fig. 16. Lateral neural ectoderm edge, 5% O2, 10 somites. Cells with basally
placed nuclei (arrow) and cells with leading lamellae (double arrows) appear to be
pro-neural crest.
Fig. 17. Lateral neural ectoderm edge, 20% O2, 13 somites. Random arrangement
of nuclei and smooth basal surface of neural epithelium suggests little or no neural
crest cell migration, an, apical surface of neural ectoderm; cm, cranial mesenchyme.
lateral region showed a lesser concave curvature than normal (Fig. 15). A
curvature of the basal surface, related to a gradation in cell height, was present,
but there were no features indicative of neural-crest-cell migration. Although
the V shape was similar to that of 8-somite embryos, the neural epithelium was
thicker, i.e. the cells had elongated as in normal embryos of this somite
number.
Light microscopy
Light microscopy of plastic-embedded embryos confirmed the SEM observations, and revealed some additional features. Observation of whole embryos
prior to embedding showed that from the 8- to 14-somite stages the longitudinal
axis of the cranial region (notochord and neural folds) of embryos cultured at
all O2 levels (and in vivo) became more curved, i.e. the cranial flexure became
more pronounced. (The method of observation did not enable fine comparisons
to be made between the degree of curvature in different embryos.) The shape
of the cranial flexure in a 13-somite, 20% O2-cultured embryo is illustrated in.
Fig. 11, which also indicates that increase in the flexure would be accompanied
by an elongation of the open area at the level of the neural/surface ectodermal
junction.
Serial sections showed that in normal embryos, thickening of the neural
epithelium was correlated not only with increasing somite number, but also
with the degree of narrowing of the V shape at different levels of the open area
within the same embryo. In embryos cultured with 20 % or 40 % O2 the epithe-
Oxygen and cranial neurulation
27
lium thickened with increasing somite number even though the V shape was
excessively broad (e.g. Fig. 19).
Examination of the lateral edge of the neural ectoderm at higher power
(Figs 20-23) indicated that in normal embryos development of the sharp
mediad curvature was accompanied by neural crest cell migration and cell
death. Loss of cells was correlated with the appearance of a cell-free space
between the most lateral neural ectoderm cells and the basal surface of the
surface ectoderm (Fig. 20); this space was taken up as the surface ectoderm was
(apparently) drawn medially with the curving neural epithelium (Fig. 21). Those
neural epithelial cells which remained in the most lateral region lacked the
apical micronlament/desmosome band which characterized the major part
of the epithelium (see below). This band formed a continuous line which was
clearly visible at all magnifications; cells which contributed to this region had
narrow necks and surface projections. In contrast, the lateral edge cells had
broad, flat apical surfaces, and were a simple cuboidal to columnar shape
(graded from the extreme edge).
Embryos cultured with 20 % and 40 % O2 showed late and reduced neural
crest migration (Figs 22 and 23). Loss of neural crest cells was not accompanied
by loss of the apical micronlament/desmosome bands in the remaining cells,
so that a cell-free space was not formed between the lateral edge of the neural
epithelium and the basal surface of the surface ectoderm. Necrotic cells were
extremely rare in any part of the neural epithelium.
The surface ectoderm adjacent (lateral) to the open BCNF region was very
uneven in both normal and abnormal embryos (Figs 18, 19). This corresponded
to the 'crumpled' area observed with SEM, and flattened out on closure of the
neural tube. It probably represents 'slack', which is taken up as the lateral
edges of the neural folds move towards the midline.
Transmission electron microscopy
Dense microfilament bundles oriented parallel to the surface were observed
in association with both horizontally and vertically orientated desmosomes in
the apical region of neural epithelium of embryos from all sources and at all
stages examined (Fig. 24). The distribution of the micronlament/desmosome
lattice has been described above. Microtubules, oriented along the long axis of
the cells, were a prominent feature of neural epithelial cells (Fig. 24). Mitochondrial morphology differed in embryos cultured at low and high O2 levels
(Figs 25-27). The mitochondria of all cells (i.e. neural ectoderm, surface
ectoderm, mesoderm, and neural crest-derived mesenchyme) of 5 % O2-cultured
and in vivo embryos were rounded in section and had very few or no cristae.
In contrast, the mitochondria of 20 % and 40 % O2-cultured embryos showed
the morphology typical of aerobic cells, having an ovoid shape in section, and
containing a large number of parallel cristae. Both types were closely associated
with rough endoplasmic reticulum.
3
EMB
54
28
G. M. MORRISS AND D. A. T. NEW
Oxygen and cranial neurulation
29
DISCUSSION
Closure of the cranial neural tube (brain tube) in rat embryos involves
(a) contact at the forebrain apposition point; (b) closure of the anterior forebrain region (anterior neuropore); (c) closure of the spindle-shaped opening
posterior to the forebrain apposition point. This study has not contributed to
an understanding of the mechanism involved in (b), which will therefore not be
discussed.
Contact at the forebrain apposition point followed a rapid narrowing of the
V shape at this point. This process took 3-4 h (8-somite to early 10-somite
stage). It was associated with a progressive increase in curvature of the underlying anterior part of the notochord, i.e. with the development of a pronounced
cranial flexure. The cranial flexure of the rat (and mouse and hamster, Waterman, 1976) at this stage is greater than that of some other mammalian embryos
(e.g. human, Corliss, 1976). 'Precocious' development of a forebrain apposition
point divides the open cranial neural folds into two regions, which appear to
close independently. Only the posterior (BCNF) open area was severely affected
by culture at high oxygen levels.
FIGURES
18-23
Figs 18-23. Plastic-embedded embryos, 1 /*m sections, stained toluidine blue.
Orientation as shown in Fig. 11.
Fig. 18. 10-somite embryo, 5% O2: apposed anterior forebrain neural folds; in the
BCNF region, concave curvature of the neural epithelium has just begun, and neural
crest cells («c) have begun to migrate; the surface ectoderm adjacent to the BCNF
region has a crumpled appearance (c). n, Notochord, which is curved and therefore
sectioned at two levels.
Fig. 19. 12-somite embryo, 40% O2: the anterior forebrain neural folds are not
apposed: the BCNF region has the 10-somite shape, but has thickened normally;
few neural crest cells («c?) can be identified (dense areas (b) are blood cells); the
surface ectoderm has retained its crumpled appearance.
Figs 20-23. Lateral edge of neural epithelium, BCNF region. Arrows delimit the
extent of the microfilament-free region, or indicate the junction between microfilament containing cells and surface ectoderm. Dark-stained cells are necrotic. Apical
microfilament bundles with desmosomes (m) show as a dark line close to the surface.
Fig. 20. 10-somite embryo, 5%O 2 : emigration of neural crest cells from the
microfilament-free region has left a cell-free space between neural and surface
ectoderms.
Fig. 21. 11-somite embryo, 5% O2: the microfilament-free region of neural epithelium has moved medially, so that the apical surface of this region is at right angles to
the main microfilament-containing neural epithelium; the surface ectoderm and basal
neural ectoderm are closely apposed, i.e. the cell-free space has disappeared.
Fig. 22. 12-somite embryo, 20% O2: the lateral neural epithelium has no microfilament-free region; neural crest cells have begun to migrate but there are no
dead cells.
Fig. 23. 12-somite embryo, 40% O2: possibly one microfilament-free cell; little or
no neural crest cell migration; no dead cells in the lateral edge region.
3-2
30
G. M. MORRISS AND D. A. T. NEW
Fig. 24.12-somite embryo, 20 % O2: TEM of the apical surface of the neural epithelium to show the microfilament bundle (m/)/desmosome arrangement which was
observed in all embryos at this stage, mt, microtubules.
Oxygen and cranial neurulation
Figs 25-27. Mitochondria in cultured embryos: all are from mesoderm cells, but are
representative of mitochondria from all types of cell.
Fig. 25. 5% O2-cultured embryo: rounded mitochondria with few cristae.
Fig. 26. 20 % O2-cultured embryo: elongated mitochondria with many cristae.
Fig. 27.40 % O2-cultured embryo: mitochondria similar to those of 20 % O2-cultured
embryos.
31
32
G. M. MORRISS AND D. A. T. NEW
Comparison of embryos cultured with a 20 % or 40 % O2 gas phase with those
cultured at 5 % O2 or developing in vivo showed the following differences in
morphogenesis of the neural epithelium of the BCNF region: the epithelium
thickened normally but showed less narrowing of the V shape and less development of a concave surface; the lateral edges failed to develop a sharp mediad
curvature. Failure of curvature of the lateral edges was associated with failure
of neural crest cell migration, failure of normal cell death, and retention of
apical microfilament bundles. These cellular events are therefore considered to
be associated with normal cranial neural tube closure in rat embryos, and to
be dependent on a low O2 environment.
Neural tube formation has been considered by many authors to be the result
of contraction of apical microfilament bundles (Baker & Schroeder, 1967;
Burnside, 1971; Freeman, 1972) and this hypothesis has received recent experimental support (Messier & Seguin, 1978). However, there is no evidence in the
present study to suggest that deficient curvature of the BCNF region of embryos
cultured in high O2 was related to deficient contraction of apical microfilament
bundles. Although it is not possible to assess their relative contraction forces,
they were observed by TEM to be well organized in all embryos, showing an
abnormality at the higher O2 levels only in extending to the extreme lateral
edges of the neural epithelium of 10- to 14-somite embryos. We suggest that
their presence may be essential for control of shape (e.g. surface area) and
maintenance of cohesion of the neural epithelium, but that they do not actually
bring about neural tube formation.
Jacobson & Gordon (1976) have demonstrated that in the newt Taricha
torosa neurulation is associated with notochordal elongation as well as microfilament contraction. Jacobson (1978) considers that notochordal elongation
generates stretch forces in the neural ectoderm which result in the formation
of neural folds and neural tube. However, the newt embryo is a non-growing
system, in which the whole neural plate undergoes neurulation simultaneously.
The rat embryo, like all amniotes, is continuously growing, and different
anteroposterior levels of the neural tube are formed at different times. Cranial
neurulation in the rat embryo (and probably in all mammalian embryos) is
further complicated by the fact that a sequence of shape changes, probably
involving a sequence of different mechanisms occurs (Waterman, 1976; Morriss
& Solursh, 1978a, b). Nevertheless, the process of BCNF morphogenesis in.
9- to 13-somite embryos described in this paper has some features which are
compatible with the theory of Jacobson & Gordon (1976). Narrowing of the
spindle-shaped BCNF opening from the 10- to 12-somite stage occurs simultaneously throughout its length before fusion of the neural folds in this region
begins. The change in shape of the opening could be explained as resulting from a
longitudinal stretch force, although this would have to act at surface ectoderm
level as well as at the level of the notochord. The observed increase in cranial
flexure during the period of closure may have this effect; the postulated stretch
Oxygen and cranial neurulation
(A)
(B)
Fig. 28. Diagrams of the neural ectoderm of the broad cranial neural fold (BCNF)
closure region in (A) a normal 11 -somite embryo, and (B) an .11- to 14-somite embryo
cultured with a 20% or 40% O2 gas phase. The orientation is as in Fig. 11. The
lower line on each represents the notochord. The upper arrows indicate the expected
elongation of this region which would be associated with the observed increase in
cranial flexure at this time. The cross-sections are based on observations of relevant
embryos (e.g. Figs 4-6, 8, 13-15, 21-23). The diagrams illustrate the suggestion
made in the discussion, that in normal embryos elongation of the open area may
contribute to closure through bringing about the passive medial movement of
flexible lateral edges, whereas in 20%- and 40% O2-cultured embryos the failure
of mediad movement of the lateral edges maybe due to their relative inflexibility (see
text for details of cellular specializations in the lateral region).
33
34
G. M. MORRISS AND D. A. T. NEW
force is indicated by arrows in Fig. 11, and in more detail in Fig. 28 A. The
cross-sectional shape of the closed areas immediately rostral and caudal
(anterior and posterior) to the opening includes a dorsal horizontal element
which is continuous with the thinned, microfilament-free lateral edges of the
BCNF region; a longitudinal pull would therefore involve passive mediad
movement of the lateral edges, but only if they were sufficiently flexible to allow
bending to occur (compare Fig. 28 A and B). We suggest that the loss of microfilaments, the emigration of neural crest cells, and the death of some cells in the
lateral region may provide the necessary flexibility. Although the observations
presented here are compatible with the hypothesis that closure of the spindleshaped BCNF opening depends on a longitudinal stretch force, no direct
evidence is at present available that such forces exist in mammalian embryos,
and further investigation is needed.
Failure of closure of the BCNF region in conditions of high O2 may be
explained in relation to the above hypothesis either as a deficiency of the
longitudinal stretch force, or as an increase in resistance of the neural epithelium
to deformation, or both. The observations presented here do not contribute
information concerning the possibility of deficiency of a longitudinal stretch
force, but they are compatible with the suggestion of an increase in resistance
of the neural epithelium. Evidence in support of the possibility that the epithelium may be abnormally rigid is provided by the general lack of dead celis
within it, and by the absence or late appearance of the lateral specializations
which were considered above in relation to the development of flexibility here.
Lateral edges which are abnormally thick, and with an abnormally high ratio
of cells: extracellular matrix, would provide an abnormally high resistance to
the tendency of a longitudinal stretch force to pull them together in the midline.
Figure 28 B illustrates the postulated stretch force (arrows) in relation to
observations made on whole, cut and sectioned embryos of 11—14 somites from
20% and 40% O2 cultures.
Mitochondrial morphology showed a clear relationship to the gas phase,
having few cristae in 5 % O2-cultured and in vivo embryos, and being of the
characteristic aerobic type in 20 % and 40 % O2-cultured embryos. It is interesting
to note that Shepard, Tanimura & Robkin (1970) demonstrated a high rate of
glycolysis in day-10 rat embryos (equivalent to 24 h after the start of culture in
this study), with relatively little activity of the Krebs cycle and electron transport
system. It would seem that during neurulation the rat embryo is adapted for
anaerobic respiration, possibly because the blood circulation of the embryo
and yolk sac is not established until neurulation is almost complete, while
growth during this period is considerable. When the O2 level is increased,
mitochondrial structure is modified accordingly, but other cellular changes
occur which result in disturbance of morphogenetic mechanisms.
We wish to thank M. Barker, W. Mouel, B. Archer and P. Blundell for assistance, and the
MRC forfinancialsupport.
Oxygen and cranial neurulation
35
REFERENCES
P. C. & SCHROEDER, T. E. (1967). Cytoplasmic filaments and morphogenetic movement in the amphibian neural tube. Devi Biol. 15, 432-456.
BURNSIDE, B. (1971). Microtubules and microfilaments in newt neurulation. Devi Biol. 26,
416-441.
CORLISS, C. E. (1976). Patten's Human Embryology. New York: McGraw-Hill.
FREEMAN, B. G. (1972). Surface modifications of neural epithelial cells during formation of
the neural tube in the rat embryo. /. Embryol. exp. Morph. 28, 437-448.
JACOBSON, A. G. (1978). Some forces that shape the nervous system. Zoon 6, 13-21.
JACOBSON, A. G. & GORDON, R. (1976). Changes in the shape of the developing vertebrate
nervous system analysed experimentally, mathematically and by computer simulation.
/. exp. Zool. 197, 191-246.
MESSIER, P.-E. &SEGUIN, C. (1978). The effects of high hydrostatic pressure on microfilaments
and microtubules in Xenopus laevis. J. Embryol. exp. Morph. 44, 281-295.
MoRRiss, G. M. & SOLURSH, M. (1978o). Regional differences in mesenchymal cell morphology and glycosaminoglycans in early neural-fold stage rat embryos. /. Embryol. exp.
Morph. 46, 37-52.
MORRISS, G. M. & SOLURSH, M. (19786). The role of primary mesenchyme in normal and
abnormal morphogenesis of mammalian neural folds. Zoon 6, 33-38.
NEW, D. A. T. & COCKROFT, D. L. (1979). A rotating bottle culture method with continuous
replacement of the gas phase. Experientia 35, 138-139.
NEW, D. A. T., COPPOLA, P. T. & COCKROFT, D. L. (1976a). Improved development of
head-fold rat embryos in culture resulting from low oxygen and modifications of the
culture serum. /. Reprod. Fert. 48, 219-222.
NEW, D. A. T., COPPOLA, P. T. & COCKROFT, D. L. (19766). Comparison of growth in vitro
and in vivo of post-implantation rat embryos. J. Embryol. exp. Morph. 36, 133-144.
SHEPARD, T. H., TANIMURA, T. & ROBKIN, M. A. (1970). Energy metabolism in early mammalian embryos. Devi Biol. Suppl. 4, 42-58.
STEELE, C. E. & NEW, D. A. T. (1974). Serum variants causing the formation of double
hearts and other abnormalities in explanted rat embryos. J. Embryol. exp. Morph. 31,
707-719.
WATERMAN, R. E. (1976). Topographical changes along the neural fold associated with
neurulation in the hamster and mouse. Am. J. Anat. 146, 151-172.
BAKER,
{Received 23 April 1979, revised 25 June 1979)