/. Embryol. exp. Morph. Vol. 46, pp. 37-52, 1978
Printed in Great Britain © Company of Biologists Limited 1978
37
Regional differences in mesenchymal
cell morphology and glycosaminoglycans in early
neural-fold stage rat embryos
By GILLIAN M. MORRISS 1 AND MICHAEL SOLURSH 2
From the Department of Human Anatomy, University of Oxford,
and Department of Zoology, University of Iowa
SUMMARY
Rat embryos with two to four pairs of somites (day 9 of gestation) were examined by
scanning electron microscopy and Alcian blue staining. The neural folds, which represent
only future brain region at this stage, form a pair of elongated hemispheres with a deep
neural groove between them. In transverse section the neural ectoderm is biconvex; the
cranial mesenchyme cells beneath them are widely separated by extracellular matrix (ECM)
and are joined to each other and to the ectodermal basement membrane by fine cytoplasmic
processes and strands of ECM material. In contrast, mesenchyme cells close to the primitive
streak are closely packed, having broad areas of surface contact and only small amounts of
ECM.
The nature and distribution of ECM, cell surface, and basement membrane glycosaminoglycans (GAG) were investigated by staining with Alcian blue at specific pH values in
combination with enzyme pretreatments, and at various critical electrolyte concentrations.
The results indicate that the GAG of the ectodermal basement membrane, mesenchymal
ECM and mesenchymal cell surfaces are in continuity with each other and consist
largely of hyaluronate and chondroitin sulphates.
Differences in morphology and histochemistry of neural fold and primitive streak regions
are discussed in relation to their possible morphogenetic significance.
INTRODUCTION
In a biochemical study of rat embryos from the pre-primitive streak stage
(day 7) to 10 somites (early day 10) we found that the synthesis of hyaluronate
(HA) was progressively increased (Solursh & Morriss, 1977). In contrast, levels
of sulphated glycosaminoglycans (GAG) remained low until the latest stages
examined. HA is a common component of the extracellular matrix, and has been
correlated with the formation of spaces around mesenchyme cells (Toole,
1973, 1976; Pratt, Larsen & Johnston, 1975; Solursh, 1976; Fisher & Solursh
1977). Its association with expanding intercellular spaces may be correlated
1
Author's address: Department of Human Anatomy, South Parks Road, Oxford OX1
3QX, U.K
2
Author's address: Department of Zoology, University of Iowa, Iowa City, la 52242,
U.S.A.
38
G. M. MORRISS AND M. SOLURSH
with the fact that it has a hydrodynamic volume which is 10 times larger than
the space occupied by the unhydrated polysaccharide chain (Ogston & Stanier,
1951).
Following [3H]glucosamine incorporation into GAG in day-9 embryos,
labelling of hyaluronate was higher in the neural fold than in the primitive
streak region (Solursh & Morriss, 1977). Mesenchymal cells underlying the
neural folds are widely separated by extracellular matrix (ECM), whereas
those of the primitive streak region have little ECM (Morriss & Solursh, 1978).
The present study was undertaken in order to make some correlation between
these regional differences in GAG synthesis and morphology during early
stages of cranial neural fold morphogenesis.
MATERIALS AND METHODS
Pregnant Sprague-Dawley or Wistar rats were killed on the afternoon of
day 9 of pregnancy (positive vaginal smear = day 0). The whole uterus was
removed and placed in Tyrode saline. Decidual swellings were freed from the
uterine wall, and embryos dissected from them. Embryos having 2-3 pairs of
somites, and in perfect condition, were selected for one of the following proced ures.
Scanning electron microscopy
Reichert's membrane was removed, and the embryo proper separated from
the extra-embryonic region. The embryo was fixed in 2-5 % cacodylate-buffered
glutaraldehyde (0-1 M, pH 7-2) for 1 h, and washed in buffer. The amnion was
then removed, and transverse cuts made by means of a cataract knife through
the highest point of the biconvex cranial neural folds, and/or the primitive
streak region. They were then dehydrated in graded acetones and dried in a
Polaron critical point drying apparatus, mounted on aluminium stubs with
double-sided Sellotape, and coated with gold in a sputter coater. They were
viewed in a Cambridge Stereoscan scanning electron microscope.
Alcian blue histochemistry
Reichert's membrane was left intact. Embryos were fixed in Carnoy's fluid,
dehydrated, cleared, and embedded in paraffin wax.* Sections were cut at
5 /-cm thickness, orientated transversely to the long axis of the 'egg cylinder'.
They were mounted three to four per slide.
* Some embryos were fixed in Carnoy's fluid containing lOOmg/ml cetylpyridinium
chloride (CPC), since leaching of GAG into the fixative has been shown to occur (Szirami,
1963). However, there was no difference in the staining intensity and distribution with and
without CPC, although the texture of the ECM was altered. Comparison of glutaraldehydefixed embryos with and without CPC by electron microscopy showed severe membrane
damage in the former, with membranous vesicles in the intercellular spaces. Ruggeri et al.
(1975) used Alcian blue staining successfully at the ultrastructural level without CPC. Use
of the detergent was therefore discontinued.
Mesenchyme, GAG, and neural folds in rat embryos
39
Enzyme histochemistry. Prior to staining, sections were incubated for 3 h at
37 °C in one of the following solutions: (a) 0-1 M phosphate buffer, pH 5-6;
(b) buffer containing 50 TRU/ml Streptomyces hyaluronidase (Calbiochem);
(c) buffer containing 400 units/ml purified, protease-free testicular hyaluronidase
(Worthington, HSEP). Following incubation, the slides were rinsed in buffer
and stained in Alcian blue 8GX (Gurr) at pH 2-5 or pH 1, dehydrated, cleared
and mounted.
Only sulphated GAG stain at pH 1, whereas at pH 2-5 sulphated and nonsulphated GAG and other polyanions stain (Lev & Spicer, 1964). Streptomyces
hyaluronidase is specific for hyaluronate (Ohya & Kaneko, 1970); testicular
hyaluronidase degrades hyaluronate, chondroitin, and chondroitin sulphates
A and C.
Alcian blue staining at critical electrolyte concentrations. Staining solutions
were prepared according to Scott & Dorling (1965), and the staining procedure
of Ruggeri, Dell'Orbo & Quacci (1975) was followed. Solutions contained
0-0 M, 0-1 M, 0-3 M, 0-65 M, and 0-8 M-MgCl2.
Scott & Dorling (1965) observed that hyaluronate standard on filter paper
does not stain in the presence of MgCl2 concentrations greater than 0-1 M,
but stains weakly at 0-1 M and intensely at lower concentrations; chondroitin
sulphate stains weakly in the presence of 0-65 M-MgCla; heparan, and presumably heparan sulphate, stains weakly in 0-8 M-MgCla. (The critical electrolyte concentration was 30-50 % lower on filter paper than in tissues.)
Scott & Stockwell (1967) have emphasized that the CEC method does not
permit unique identifications in the absence of other chemical data. The utilization of this technique in the present study must therefore be regarded as
complementary to the use of enzyme histochemistry, and to our previously
reported biochemical data (Solursh & Morriss, 1977).
RESULTS
Scanning electron microscopy (Figs. 1-4).
Figure 1 shows a 3-somite embryo viewed from the ectodermal surface.
It has the anteroposterior concave curvature characteristic of rodent embryos
of this stage of development, but is otherwise similar to early neurula stages of
other mammalian species, e.g. pig, human (Patten, 1964). Figure 4 shows a
2-somite embryo with the neural folds and primitive streak cut transversely.
Anteriorly, the cranial neural folds form a pair of elongated hemispheres which
are biconvex in section and separated by a deep neural groove (Figure 3).
This region of the neural ectoderm is bounded posteriorly by a transverse groove,
the preotic sulcus, which marks the position of the 3rd rhombomere of the
hindbrain (Adelmann, 1925) (Fig. 1). Posterior to it is the otic region, and an
area of shallow neural groove between the somite pairs. Hensen's node could
not be detected in surface view. The epiblast of the primitive streak region
40
G. M. MORRISS AND M. SOLURSH
100 u
Fig. .1. Three-somite embryo, side view. The yolk sac and amnion have been
removed and a transverse cut made across the primitive streak region (left). Cranial
neural folds either side of the neural groove are biconvex in form; the transverse
groove (arrowed) is the preotic sulcus. Neural ectoderm and 'undifferentiated'
epiblast are indistinguishable from each other in surface view, but are clearly
delimited from the presumptive surface ectoderm (s) laterally, c, Cardiogenic region.
(Fig. 2) consists of a columnar epithelium continuous with the neural plate,
whereas the presumptive surface ectoderm lying lateral to the epiblast and the
neural plate forms a cuboidal epithelium (Fig. 3). Bottle-shaped cells can be
seen in the streak itself (Fig. 2).
The endoderm cells form a thin squamous epithelium. Mesoderm cells
(primary mesenchyme) are very close-packed in the primitive streak region,
and consequently somewhat spherical in shape (Fig. 2). In the presumptive
surface ectoderm region their intercellular spaces are slightly greater. Under
the cranial neural folds the mesoderm cells are stellate in shape, with numerous
F I G U R E S 2-4
Figs. 2-4. Two-somite embryo. The primitive streak region has been cut as in Fig. 1.
In addition, a transverse cut has been made through the cranial neural folds.
Fig. 2. Primitive streak region. Characteristic bottle-shaped cells can be seen in the
streak, a, Amniotic cavity.
Fig. 3. Cephalic region. Neural ectoderm forms a convex curve lateral to the notochord. In the neural groove, the two apical neuroectodermal surfaces are in close
apposition. Primary mesenchyme cells are more widely separated than those of the
primitive streak region. Endoderm shows an intucking, the lateral fold of the
developing foregut.
Fig. 4. Whole specimen, for reference.
Mesenchyme, GAG, and neural folds in rat embryos
41
42
G. M. MORRISS AND M. SOLURSH
processes joining the cells to each other, to the overlying neuroectodermal
basement membrane and to the underlying endoderm. The intercellular spaces
are largest in this region (Fig. 3).
Aldan blue staining (Figs. 5-16, Tables 1 and 2).
Staining intensities of cell surfaces and extracellular matrix in the cranial
neural fold and primitive streak regions were scored as strong, moderate, weak
or absent. The results are interpreted on the basis of enzyme sensitivities (Table
1) or from the specificities of different critical electrolyte concentrations (Table 2)
according to the criteria described in the Materials and Methods section.
Cytoplasmic and nuclear staining have been included in the tables to indicate
the intensity of the reaction of RNA, DNA, and other intracellular polyanions.
This staining provided a degree of 'background', and care was taken to make
comparisons between specific features on different slides.
Cell surfaces stained at pH 2-5 (Figs. 5 and 8 and Table 1), but the stained
material (with the exception of the apical endodermal and cranial mesenchymal
surfaces, which stain much more strongly than other regions) was not sensitive
to hyaluronidase treatment and could be any polyanionic material in addition
to GAG. The material staining at pH 1, when only sulphated GAG are expected
to stain, was largely hyaluronidase-sensitive (Figs. 9 and 10). The enzymesensitive sulphated GAG is likely to be chondroitin sulphate, while the enzymeresistant material associated with the apical endodermal and cranial mesenchymal surfaces is probably heparan sulphate. These two conclusions are further
supported by the decreased staining when the MgCl2 concentration is increased
from 0-3 to 0-65 M, and the further decrease in the apical endodermal and
cranial mesenchymal surfaces when the M'gCl2 concentration is increased
from 0-65 to 0-8 M (Figs. 13-15 and Table 2). The staining associated with the
cranial mesenchymal cell surfaces is similar to that observed in the mesenchymal
matrix, suggesting continuity of material.
Mesenchymal extracellular matrix. This stained more intensely in the cranial
region than in the primitive streak region at pH 1 and pH 2-5 (Figs. 5, 8 and 9).
Since the mesenchyme cells in the primitive streak region (Figs. 8 and 10) were
closely packed together, there is little extracellular matrix available for staining,
and apposed cell surfaces were unstained. Assessment of the staining intensities
in this region are therefore based on a few small spaces for each section. The
staining reactions in the primitive streak region are qualitatively similar to
those of the cranial mesenchyme, except that in the primitive streak region no
stained material could be detected after testicular hyaluronidase treatment
(Table 1).
A large portion of the material that stains at pH 2-5 is sensitive to treatment
with either Streptomyces hyaluronidase or testicular hyaluronidase (Table 1,
Figs. 6 and 7). These results suggest the presence of hyaluronate and possibly
also chondroitin and chondroitin sulphate. Hyaluronate was identified by
Mesenchyme, GAG, and neural folds in rat embryos
43
Table 1
pH 2-5
Location of stain
Cell surfaces
Apical neural ectoderm
Apical non-neural
ectoderm
Apical endoderm
Cranial mesenchyme
PS region mesenchyme*
Mesenchymal ECM
Cranial neural folds
PS region*
Basement membranes
Medial neural ectoderm
Other ectoderm
Cells
Cytoplasm
Nuclei
pH 1
Buffer
only
Streptomyces
HAase
Testicular
HAase
Buffer
only
Testicular
HAase
+
±
+
±
±
±
±
±
—
—
+ 4++
(±)
++
±
(±)
+
±
(±)
++
+
(±)
±
±
—
H- +
+
±
(±)
±
(±)
+
(±)
±
-
++
+
±
±
+
±
+
±
±
—
+
+
+
+
+
+
—
—
—
—
* Very few spaces. No staining of closely apposed surfaces.
Key to tables: + + strong; + moderate; ± weak; (±) weak staining of a few small areas
only; — no staining.
Table 2
Critical electrolyte concentration
Location of stain
Cell surfaces
Apical neural ectoderm
Apical non-neural ectoderm
Apical endoderm
Cranial mesenchyme
PS region mesenchyme*
Mesenchymal ECM
Cranial neural folds
PS region*
Basement membranes
Medial neural ectoderm
Other ectoderm
Cells
Cytoplasm
Nuclei
00M
01M
0-3 M
0-65 M
+
+
+
+
(±)
+
+
+
+
(±)
—
—
±
+
(±)
—
—
±
±
—
++
+
++
+
+
±
±
-
±
-
++
4-
++
+
++
+
+
±
+
±f
++
++
+
++
+
±
-
-
* See Table 1.
t Lateral cranial neural ectoderm only.
08 M
—
—
±
±
—
44
G. M. MORRISS AND M. SOLURSH
Mesenchyme, GAG, and neural folds in rat embryos
45
comparing Alcian blue staining intensities between the following pairs of
preparations: (a) pH 1 and pH 2-5; (b) pH 2-5 following preincubation with
either buffer alone or buffer containing Streptomyces hyaluronidase; (c) critical
electrolyte concentration 0-0 M and higher molarities. These comparisons
showed that significant quantities of HA are present in the cranial mesenchymal
ECM and that HA is also present in the small amount of ECM in the streak
region. Some staining reaction was also obtained in the cranial region at pH 1,
and at the higher critical electrolyte concentrations (0-3-0-8 M) showing that
sulphated GAG were also present. These were identified separately or in groups
as follows: comparison of staining at pH 2-5 following digestion with the two
types of hyaluronidase indicates the presence of chondroitin sulphates A and C,
as does a comparison between staining at 0-3 and 0-65 M; staining at pH 1
following testicular hyaluronidase, and staining at 0-65 and 0-8 M in the CEC
series, indicate some heparin/heparan sulphates. [Keratan sulphate would also
stain at this CEC, but in our biochemical study (Solursh & Morriss, 1977)
we demonstrated the presence of heparan sulphate, whereas we were unable to
detect keratan sulphate.] It is possible that some of the material staining at pH 1
was hyaluronate of exceptionally high molecular weight (Derby & Pintar,
1978).
Basement membrane. In rat embryos of this age there is only one basement
membrane, that of the ectoderm. It showed a clear regional difference in staining
intensity, being very heavily stained in the medial region of the cranial neural
folds (Figs. 5, 9 and 11-14); the lateral extent of this region of heavy staining
was rather variable in different specimens, but never extended outwards more
than two-thirds of the arc of each neural fold. The basement membrane is
continuous between the neural folds of the two sides. Comparison of the staining
FIGURES
5-10
Figs. 5-16. Three-somite embryos (Carnoy-fixed and paraffin-embedded) stained
with Alcian blue. Sections 5-7 and 9-15 show part of neural ectoderm (approximately
one-third width of neural folds from midline) and underlying mesenchyme. Figs.
8 and 16 show the region immediately lateral to the primitive streak. For details, see
Materials and Methods and Results sections.
Fig. 5. pH 2-5, incubated with buffer only. Heavy staining of basement membrane,
mesenchymal cell surfaces, and ECM material.
Fig. 6. pH 2-5, Streptomyceshya.\won\dase. Weak staining of neuroectodermal basement membrane, mesenchymal cell surfaces and ECM.
Fig. 7. pH 2-5, testicular hyaluronidase. Overall reduction of the staining reaction.
Fig. 8. pH 2-5, buffer only (same section as Fig. 5). Cell surface and extracellular
staining is weak and sporadic, except on the apical ectodermal surface. Spaces
between mesoderm cells may be partly artifactual.
Fig. 9. pH 1, buffer only. Moderate staining of neuroectodermal basement membrane; moderate but sporadic staining of mesenchymal cell surface and ECM
material.
Fig. 10. pH 1, testicular hyaluronidase. Weak staining of neuroectodermal basement membrane, mesenchymal cell surfaces, and ECM material.
4
E M B 46
46
G. M. MORRISS AND M. SOLURSH
••'•;^.j5f^
•^
, • . / -
•*••••
.»,.:
•
v'>.
« . 4'
. " > ^<f2.-i
10/im
v
* ^ ^
(
Mesenchyme, GAG, and neural folds in rat embryos
47
reactions of this region with the rest of the basement membrane suggests a
quantitative rather than a qualitative difference.
In all regions the basement membrane stained more intensely at pH 2-5
than at pH 1. The sensitivity of this material staining at pH 2-5 to hyaluronidase
treatment suggests the presence of hyaluronate (Figs. 6 and 7). The presence of
testicular hyaluronidase-sensitive material that stains at pH 1 demonstrates
the presence of chondroitin sulphate (Figs. 9 and 10); in addition, the medial
part of the neural fold basement membrane contains some enzyme-resistant
sulphated GAG. The decreased staining observed in the presence of MgCla at
concentrations between 0-3 and 0-65 M (Figs. 13 and 14) suggests that chondroitin sulphate is the major GAG present in the basement membrane. Some
material stains faintly in the presence of 0-8 M-MgCl2 (Fig. 15). The effect of
Streptomyces hyaluronidase (Fig. 6) on the stainable material in the basement
membrane suggests that hyaluronate may be closely associated with chondroitin
sulphate here, as in the mesenchymal ECM.
DISCUSSION
Morphological and histochemical differences between the
neural fold and primitive streak regions
The observations presented here confirm that in 2-3-somite rat embryos
the cranial mesenchyme cells are widely separated by extracellular matrix
(ECM), and show that following fixation, this ECM contains Alcian bluepositive strands and granules. The strands stretch from one cell to another, or
from mesenchyme cells to the neuroectodermal basement membrane. The
cells make contact with each other and with the basement membrane by fine
cytoplasmic processes, so that the three-dimensional structure of the mesenchyme
under each neural fold is a dome-shaped meshwork of cytoplasmic and matrix
strands alternating with the bodies of the cells.
FIGURES
11-16
Fig. 11. Critical electrolyte concentration (CEC) 0 0 M. Strong staining of medial
neuroectodermal basement membrane (extreme left), and extracellular matrix
material; moderate staining of mesenchymal cell surfaces.
Fig. 12. CEC 01 M. Lighter intracellular staining; cell surface, basement membrane and ECM staining similar to 00M.
Fig. 13. CEC 03 M. Moderate but uneven staining of mesenchymal cell surfaces and
ECM; strong staining of neuroectodermal basement membrane.
Fig. 14. CEC 0-65 M. Occasional weak staining of mesenchymal cell surfaces and
ECM; moderate staining of neuroectodermal basement membrane.
Fig. 15. CEC 08 M. Some weak staining of mesenchymal cell surfaces and ECM;
moderate staining of neuroectodermal basement membrane.
Fig. 16. CEC 0-3 M, primitive streak region. Staining of epiblast basement membrane only.
4-2
48
G. M. MORRISS AND M. SOLURSH
In contrast, mesenchyme of the primitive streak region contains relatively
little ECM. The lesser staining intensity of the ECM and cell surface of this
region, compared to that of the neural fold region, appears to be due to the
smaller amount of ECM rather than to qualitative differences in the ECM and
cell surface material.
These observations of quantitative differences in ECM of primitive streak
and cranial neural fold regions correlate well with our earlier observation of a
differential rate of labelling of HA/unit protein between the two (1-75:1,
neural plate:primitive streak) (Solursh & Morriss, 1977). In addition, the
proportion of GAG which was hyaluronate was slightly higher in the cranial
region (87 %, cf. 80 % in the streak region). Thus an increase in the rate of HA
synthesis may be related temporally and spatially to the formation of biconvex
cranial neural folds, and histologically to the accumulation of large amounts of
ECM in the mesenchyme under the neuroectodermal epithelium.
The histochemical observations described here provide further information
concerning the organization of the ECM. In the cranial neural fold region,
the Alcian blue staining at pH 2-5 was always associated with strands of matrix
material. There was a clear spatial association between HA and sulphated
GAG in this region. This association, and the sensitivity of sulphated GAG
to Streptomyces hyaluronidase, suggest that some or all of the HA of the cranial
mesenchymal ECM may be present in association with proteoglycan. Even
though some of the staining at pH 1 may be due to exceptionally high molecular
weight HA (Derby & Pintar, 1978), the criteria used here confirm the presence
of sulphated GAG in the ECM. Since the staining of the primitive streak region
was weak, and confined to such a small amount of ECM, it is not possible to
comment on its likely composition, except that it appears to contain some
sulphated GAG in addition to HA. These histochemical results agree with our
previous biochemical results (Solursh & Morriss, 1977), as well as with biochemical (Solursh, 1976) and histochemical (Fisher & Solursh, 1977) results
obtained in the chick. Hyaluronate is the predominant GAG, with smaller
amounts of chondroitin sulphate and heparan sulphate. HA has also been
identified as the predominant ECM GAG in neurula stages of the skate embryo
(McConnachie & Ford, 1966).
The functional significance of the present observations will now be discussed.
(a)
The relationship between cranial mesenchyme and the neuroectodermal
basement membrane
Stained matrix strands provided continuity between the basement membrane
and the mesenchymal cell surfaces, suggesting a functional relationship. Even
with the finer fixation and preservation of structure provided by scanning
electron microscopy, mesenchyme cells were rarely observed to be in broad
contact with the basement membrane. It is not possible, with the techniques
used in this study, to distinguish between strands of stained ECM material
Mesenchyme, GAG, and neural folds in rat embryos
49
and fine cytoplasmic strands covered with stained material. High voltage
electron microscopy has shown that both are present, and that both provide
continuity between the mesenchymal cell bodies and the basement membrane
(Morriss & Solursh, 1978). This organization would allow functional interaction to take place between the mesenchyme and the neural ectoderm, and yet
enable the different growth and cell movement patterns of the two tissues to
proceed without creating stresses in each other.
(b) Distribution of stainable material in the basement membrane
Restriction of strong basement membrane staining to the more medial area
of neural ectoderm is remarkable. (This observation was made separately by
both authors in different laboratories). Heavy basement membrane staining has
been observed in another region of convex curvative in an epithelium: Bernfield, Banerjee & Cohn (1972) identified GAG in convex regions of salivary
epithelium basement membrane during lobule formation. The presence of GAG
in the basement membrane was associated with the appearance of basal microfilaments in the overlying cells, suggesting a direct relationship between basement membrane GAG composition and formation of the convex shape by the
epithelial cells. If the basement membrane staining observed in the present
study is similarly associated with the convex curvative of the overlying neural
epithelium, it is curious that it is not distributed more evenly under the neural
folds. Cohn, Banerjee & Bernfield (1977) have recently suggested that the
salivary gland basement membrane contains complexes of HA and proteoglycan; this organization is likely to exist in the neuroectodermal basement
membrane also, since staining was markedly reduced following Streptomyces
hyaluronidase incubation. Similar effects of Streptomyces hyaluronidase on
Alcian blue staining were also observed in the ectodermal basement membrane
in the early chick embryo (Fisher & Solursh, 1977). Cohn et al. (1977) found
hyaluronate and chondroitin sulphate in salivary gland basement membrane but
did not measure heparan sulphate. On the other hand, Hay & Meier (1974)
found chondroitin sulphate and heparan sulphate in several embryonic basement membranes. They did not measure non-sulphated GAG. The present
results suggest that hyaluronate, chondroitin sulphate and heparan sulphate
might be commonly found in embryonic basement membranes, in transit or as a
constituent.
(c) ECM of the primitive streak region mesenchyme
The identification of hyaluronate in the primitive streak region of rat embryos
in this and our previous study (Solursh & Morriss, 1977) suggests a correlation
with cell migration. Toole (1972) has suggested that one of the functions of HA
may be the physical separation of cells, thus inhibiting cyto-differentiation and
facilitating migration. The degree of contact between migrating primary
mesenchyme cells observed here suggests that separation of cells is not essential
50
G. M. MORRISS AND M. SOLURSH
for either of these functions. Ebendal (1977, personal communication) has
added hyaluronate to chick heart fibroblasts in vitro, and found no effect on
the rate or direction of cell migration. Furthermore, a hyaluronate-containing
substratum inhibits the spreading of limb, somite, and neural crest cells (Fisher
& Solursh, in preparation).
(d) ECM of the cranial mesenchyme
In contrast to the primitive streak region, large spaces separate the cranial
mesenchyme cells. These cells have presumably reached the end of their migration pathway. The difference in the amount of intercellular space in the two
regions suggests that although cell migration requires very little ECM in these
embryos, an important morphogenetic function is served by the mesenchymal
ECM of the cranial region. During the presomite to 6-somite period of development the cranial neural ectoderm changes shape from flat neural plate to
biconvex folds while the volume of mesenchyme (cells and matrix) beneath it
increases (Morriss & Solursh, 1978). Studies on vitamin A-induced exencephaly
have provided evidence that the increasing number of mesenchyme cells is an
essential component of neural fold elevation (Marin-Padilla, 1966; Morriss,
1972). The architecture of the cranial mesenchymal ECM is a three-dimensional
meshwork around the cells, so that cells and matrix together form an elongated
dome under each neural fold.
If, as the hypervitaminosis-A studies suggest, the mesenchyme cells have a
support function in neural fold elevation, their integral ECM must also be a
component of this morphogenetic process. In a study of migration of primary
mesenchyme cells in vitro, the three germ layers of late presomite and early
somite stage rat embryos were separated in order to isolate the mesenchyme
(Morriss, 1975); the dome shape of the cranial mesenchyme was maintained
after separation, suggesting that its shape was not entirely dependent on the
surrounding epithelia. Similarly Nakamura & Manasek (1977) have shown that
cardiac jelly dissected from embryonic heart will maintain its shape. These two
observations suggest that hyaluronate/proteoglycan extracellular matrices
may have a skeletal function in relation to adjacent epithelia.
Conclusion
The observations presented in this paper show a correlation between regional
differences in amounts of mesenchymal ECM and shape of the overlying
epithelium. This correlation involves an anteroposterior difference in the
amount of ECM, which is therefore also related to the age of the primary
mesenchyme cells and their position in the migration pathway. We suggest that
the increase in mesenchyme cells, and the accumulation around them of a
hyaluronate-rich extracellular matrix, are essential components of cranial
neural fold morphogenesis. While this mechanism may exist to a certain
degree in all vertebrate embryos, the shape of the neural ectoderm during the
Mesenchyme, GAG, and neural folds in rat embryos
51
first stage of cranial neurulation in the rat and other mammalian embryos
suggests that mechanical support provided by the underlying mesenchyme may
have increased in importance concomitantly with the evolution of the larger
brain of the adult.
This study was supported by an MRC project grant to G.M. M. and N1H grant no. HD
05505 to M.S. We wish to thank K. Thurley, W. Mouel, M. Barker and B. Crutch for
assistance, and B. Archer for photography.
REFERENCES
H. B. (1925). The development of the neural folds and cranial ganglia in the rat.
J. comp. New. 39, 19-172.
BERNFIELD, M. R., BANERJEE, S. D. & COHN, R. H. (1972). Dependence of salivary epithelial
morphology and branching morphogenesis upon acid mucopolysaccharide protein
(proteoglycan) at the epithelial surface. /. Cell Biol. 52, 647-689.
COHN, R. H., BANERJEE, S. D. & BERNFIELD, M. R. (1977). Basal lamina of embryonic salivary
epithelia. Nature of glycosaminoglycan and organization of extracellular materials.
/. Cell Biol. 73, 464-478.
DERBY, M. A. & PINTAR, J. E. (1978). The histochemical specificity of Streptomyces hyaluronidase and chondroitinase ABC. /. Histochem. Cytochem. (submitted).
FISHER, M. & SOLURSH, M. (1977). Glycosaminoglycan localization and role in maintenance
of tissue spaces in the early chick embryo. /. Embryol. exp. Morph. 42, 195-207.
HAY, E. D. & MEIER, S. (1974). Glycosaminoglycan synthesis by embryonic inductors neural tube, notochord and lens. /. Cell Biol. 62, 889-898.
LEV, R. & SPICER, S. S. (1964). Specific staining of sulphate groups with Alcian blue at low
pH. J. Histochem. Cytochem. 12, 309.
MARTN-PADILLA, M. (1966). Mesodermal alterations induced by hypervitaminosis A. /.
Embryol. exp. Morph. 15, 251-269.
MCCONNACHIE, P. R. & FORD, P. (1966). Acid mucopolysaccharides in the development of
the Pacific great skate, Raja binoculata. J. Embryol. exp. Morph. 16, 17-24.
MORRISS, G. M. (1972). Morphogenesis of the malformations induced in rat embryos by
maternal hypervitaminosis A. /. Anat. 113, 241-250.
MORRISS, G. M. (1975). Abnormal cell migration as a possible factor in the genesis of vitamin
A-induced craniofacial anomalies. In New Approaches to the Evaluation of Abnormal
Mammalian Embryonic Development (ed. D. Neubert). Berlin: Geo. Thieme Yerlag.
MORRISS, G. M. & SOLURSH, M. (1978). The role of primary mesenchyme in normal and
abnormal morphogenesis of mammalian neural folds. Zoon. 6. (In the Press.)
NAKAMURA, A. & MANASEK, F. J. (1977). The intrinsic shape of the chick embryonic cardiac
jelly. Anat. Rec. 187, 663 (Abstract).
OGSTON, A. G. & STANIER, J. E. (1951). The dimensions of the particle of hyaluronic acid
complex in synovial fluid. Biochem. J. 49, 585-590.
OHYA, T. & KANEKO, Y. (1970). Novel hyaluronidase from Streptomyces. Biochim. Biophys.
Acta. 198, 607-609.
PATTEN, B. M. (1964). Foundations of Embryology, 2nd ed. McGraw-Hill.
PRATT, R. M., LARSEN, M. A. & JOHNSTON, M. C. (1975). Migration of cranial neural crest
cells in a cell-free hyaluronate-rich matrix. Devi Biol. 44, 298-305.
RUGGERI, A., DELL'ORBO, C. & QUACCI, D. (1975). Electron microscopic visualization of
proteoglycans with Alcian Blue. Histochem. J. 7, 187-197.
SCOTT, J. E. & DORLING, J. (1965). Differential staining of acid glycosaminoglycans (mucopolysaccharides) by alcian blue in salt solutions. Histochemie 5, 221-233.
SCOTT, J. E. & STOCKWELL, R. A. (1967). On the use and abuse of the critical electrolyte
concentration approach to the localization of tissue polyanions. /. Histochem. Cytochem.
15, 111-113.
ADELMANN,
52
G. M. MORRISS AND M. SOLURSH
M. (1976). Glycosaminoglycan synthesis in the chick gastrula. Devi Biol. 50,
525-530.
SOLURSH, M. & MORRISS, G. M. (1977). Glycosaminoglycan synthesis in rat embryos during
the formation of the primary mesenchyme and neural folds. Devi Biol. 57, 75-86.
SZIRMAI, J. A. (1963). Quantitative approaches in the histochemistry of mucopolysaccharides.
/. Histochem. Cytochem 11, 24-34.
TOOLE, B. P. (1972). Hyaluronate turnover during chondrogenesis in the developing chick
limb and axial skeleton. Devi Biol. 29, 321-329.
TOOLE, B. P. (1973). Hyaluronate and hyaluronidase in morphogenesis and differentiation,
Amer. Zool. 13, 1062-1065.
TOOLE, B. P. (1976). Morphogenetic role of glycosaminoglycans (acid mucopolysaccharides)
in brain and other tissues. In Neuronal Recognition (ed. S. Barondes), pp. 275-329. London:
Chapman and Hall.
SOLURSH,
{Received 21 October 1977, revised 9 March 1978)