The secondary palate forms by fusion of a pair of shelves that
originate on the inner side of the maxillary process (reviewed
by Ferguson, 1988). Upon meeting at the midline of the
oropharyngeal cavity, the shelves adhere to each other at their
medial edge epithelia (MEE) giving rise to the medial
epithelial seam (MES). The MES initially consists of a
multilayered epithelium that later becomes a single epithelial
layer. MES degeneration continues by fragmentation of the
adhered region forming epithelial islands (called epithelial
pearls) along the MES. These islands degenerate resulting in
fused palate shelves.
The MEE is composed of a basal columnar cell layer
covered by flat cells that constitute the periderm. During
shelf growth, MEE is histologically undistinguishable from
oral or nasal epithelium, but it acquires distinctive features
just prior to fusion. At the molecular level, the MEE region
appears defined by the expression of several genes such as
Tgfb3 (Fitzpatrick et al., 1990), Egfr (Brunet et al., 1993),
Tgfa (Citterio and Gaillard, 1994) and Fos (Yano et al.,
1996). It is believed that periderm cells shed before fusion
to allow intimate contact between shelves (Fitchett and Hay,
Epithelial-mesenchymal transformation (EMT) is considered
relevant for MES degeneration (Fitchett and Hay, 1989; Griffith
and Hay, 1992; Shuler et al., 1991). EMT stands for the
transdifferentiation of packed epithelial cells to more loose
mesenchymal cells, a process that involves basal lamina
degradation (and dramatic changes in the cytoskeleton), and
cell-cell and cell-extracellular matrix interactions (Boyer et al.,
1996). The migratory capacity of mesenchymal cells allows
them to move far from their site of origin. Once the EMT
process occurs, transdifferentiated cells can give rise to specific
cell types, just as neural crest cells do (Duband et al., 1995), or
can contribute to form structures such as the heart valves, which
are derived from endocardial cells (Markwald et al., 1975). In
the case of secondary palate, transdifferentiated mesenchymal
cells would not have a specific function. Several lines of
evidence suggest that EMT actually occurs during palate shelf
fusion (Fitchett and Hay, 1989; Griffith and Hay, 1992;
Martinez-Alvarez et al., 2000; Shuler et al., 1991; Shuler et al.,
1992). However, because only few cells have been detected to
During mammalian development, a pair of shelves fuses to
form the secondary palate, a process that requires the
adhesion of the medial edge epithelial tissue (MEE) of each
shelf and the degeneration of the resulting medial epithelial
seam (MES). It has been reported that epithelial-
mesenchymal transformation (EMT) occurs during shelf
fusion and is considered a fundamental process for MES
degeneration. We recently found that cell death is a
necessary process for shelf fusion. These findings
uncovered the relevance of cell death in MES degeneration;
however, they do not discard the participation of other
processes. In the present work, we focus on the evaluation
of the processes that could contribute to palate shelf fusion.
We tested EMT by traditional labeling of MEE cells with
a dye, by infection of MEE with an adenovirus carrying the
lacZ gene, and by fusing wild-type shelves with the ones
from EGFP-expressing mouse embryos. Fate of MEE
labeled cells was followed by culturing whole palates, or by
a novel slice culture system that allows individual cells
to be followed during the fusion process. Very few labeled
cells were found in the mesenchyme compartment, and
almost all were undergoing cell death. Inhibition of
metalloproteinases prevented basal lamina degradation
without affecting MES degeneration and MEE cell death.
Remarkably, independently of shelf fusion, activation of
cell death promoted the degradation of the basal lamina
underlying the MEE (‘cataptosis’). Finally, by specific
labeling of periderm cells (i.e. the superficial cells that
cover the basal epithelium), we observed that epithelial
triangles at oral and nasal ends of the epithelial seam do
not appear to result from MEE cell migration but rather
from periderm cell migration. Inhibition of migration or
removal of these periderm cells suggests that they have a
transient function controlling MEE cell adhesion and
survival, and ultimately die within the epithelial triangles.
We conclude that MES degeneration occurs almost
uniquely by cell death, and for the first time we show that
this process can activate basal lamina degradation during
a developmental process.
Key words: Morphogenesis, Apoptosis, Cell migration, Mouse
Death is the major fate of medial edge epithelial cells and the cause
of basal lamina degradation during palatogenesis
Rodrigo Cuervo and Luis Covarrubias*
Departament of Developmental Genetics and Molecular Physiology, Instituto de Biotecnología, Universidad Nacional Autónoma de
México, Cuernavaca, Morelos 62210, México
*Author for correspondence (e-mail: firstname.lastname@example.org)
Accepted 26 September 2003
Development 131, 15-24
Published by The Company of Biologists 2004
have transdifferentiated and because there is lack of quantitative
analyses, other mechanisms for MES degeneration should be
considered. Migration of MEE cells towards the nasal and oral
regions has also been proposed to participate in shelf fusion
(Carette and Ferguson, 1992). Cell death, which has been
known for many years to occur in the developing palate
(DeAngelis and Nalbandian, 1968; Farbman, 1968; Smiley and
Dixon, 1968), was until only recently implicated in MES
degeneration (Cuervo et al., 2002; Martinez-Alvarez et al.,
2000; Mori et al., 1994; Taniguchi et al., 1995). MES
degeneration could also result from a combination of cellular
mechanisms such as those described above.
The aim of the present work was to evaluate the relevance
of EMT, epithelial cell migration and cell death in palate shelf
fusion. Our results revealed a fundamental role of cell death in
MES degeneration, without a significant contribution from
EMT or basal MEE cell migration. However, we show data
indicating that the ordered migration of periderm cells out
from the basal MEE is necessary for normal shelf fusion.
Furthermore, in contrast to the activation of cell death by the
degradation of basal lamina (i.e. anoikis), we identified the
activation of basal lamina degradation as a consequence of
MEE cell death (‘cataptosis’).
Materials and methods
Animal handling and palate dissection
CD-1 and EGFP (Hadjantonakis et al., 1998) mouse strains were used
in this study. Pregnant females were sacrificed by cervical dislocation
between 13.5 and 14.5 days post coitus (d.p.c.). The day of detection
of vaginal plug was found at 0.5 d.p.c. Palate dissection was carried
out as previously described (Cuervo et al., 2002).
Whole palates were cultured on filters floating on serum-free medium
as previously described (Cuervo et al., 2002). We also developed a
palate slice culture system based on that reported by Knight et al.
(Knight et al., 1999). Initially shelves of whole palates were allowed
to contact for 3 hours and were then embedded in 5% low-melting
point agarose (SeaPlaque GTG, FMC Bioproducts, Rockland, ME) in
McCoy medium (Microlab, México). Slices (200 µm) were obtained
using a vibratome (Leica VT1000S, Wetzlar, Germany) and collected
in cold PBS (5.4 mM potassium chloride, 138 mM sodium chloride,
22 mM glucose, 2 mM sodium-potassium phosphate, pH 7.2). Slices
were placed at the bottom of a 35 mm petri dish and covered with a
layer of 1% low-melting point agarose and 2 ml of McCoy medium.
At the end of culture, live slices were washed with PBS, fixed with
4% paraformaldehyde and processed for TUNEL in wholemount
(Conlon et al., 1995). Cytochalasin D (6 µM; Sigma, St Louis, MO),
cycloheximide (20 µg/ml; Sigma, St Louis, MO), retinoic acid (20
µM; Sigma, St Louis, MO), staurosporin (20 µM Sigma, St Louis,
MO), BB3103 MMP inhibitor (10 µM; British Biotech, Oxford, UK),
or z-VAD (100 µM z-VAD; Biomol, Plymouth, PA) were added
directly to the culture medium. z-VAD anti-apoptotic activity has also
been tested in explant cultures of limbs undergoing interdigital
regression, and of developing spinal cords undergoing motoneuron
degeneration. Similarly, the inhibitory activity of BB3103 on
metalloproteinases has been tested by zymography using gelatin as
substrate, and by the ability to avoid the natural degradation of basal
lamina in explant cultures containing müllerian and wolfian ducts. All
reagents remained in the medium for the whole culture period.
Complete MEE labeling was obtained by submerging whole palates
in a 10 µM solution of 5-6-carboxy 2-7-dichlorofluorescein diacetate
succinimidyl ester (CCFSE; Molecular Probes, Eugene, OR) in PBS.
Samples were incubated at 37°C for 15 minutes in the dye solution
and washed twice with PBS before culturing. Selective labeling of
periderm cells with the same dye was attained by a short incubation
(30 seconds) of the palate explants in the same dye solution at room
temperature. Transfections with an adenovirus carrying the lacZ
reporter gene (Ad-lacZ) were carried out at 37°C for 1.5 hours as
previously described (Cuervo et al., 2002). Chimaeric palates were
formed using one shelf from a CD1 embryo and the other from an
EGFP embryo. Shelf fusion under this condition took about 36 hours.
In all cases described above palates were cultured for different time
periods, fixed with 4% paraformaldehyde, embedded in paraffin or
agarose, and sliced for histology, immunohistochemistry or cell death
Periderm cell removal
To remove superficial cells, palates were incubated in 0.25% trypsin
solution in Versene (Invitrogen, Grand Island, NY) at 4°C for 5
minutes. After incubation, palate medial edges were extensively
washed with the same solution until the thin periderm cell layer came
off. Samples were washed with PBS and then incubated in DMEM
containing 10% serum at room temperature for 10 minutes. Shelves
were separated from the rest of the tissue and cultured on filters. The
same treatment was given to control palates except that wash with
0.25% trypsin was not performed.
Histology, immunofluorescence and TUNEL procedures
Cell death detection by the TUNEL method was performed using
commercial kits (Roche, Mannheim, Germany) or a whole-mount
procedure as reported by Conlon et al. (Conlon et al., 1995).
Immunohistochemistry for laminin (Rabbit anti-laminin, Sigma, St.
Louis, MO) was performed according to standard protocols. Except
for the Ad-lacZ
labeling, labeled cells were detected by
epifluorescence (Eclipse TE 300, Nikon, Japan). Double detections
combined EGFP or fluorescein (CCFSE; in situ cell death detection
kit, fluorescein, Roche) with rhodamin (in situ cell death detection kit,
rhodamin, Roche) or Alexa-fluor 594 (goat anti-rabbit; Molecular
Probes, Eugene, OR). Photographs were taken with a digital camera
(CoolSnap, Roper Scientific Inc., Trenton, NJ).
Fate of MEE cells after palate shelf fusion
To study the shelf fusion process that is required for secondary
palate formation, we have cultured palate shelves integrated to
the nasal region of the head (Cuervo et al., 2002). In this culture
system, complete shelf fusion occurs within 24 hours (Fig. 1A,
T/L) whereas, when isolated shelves are used, fusion occurs
within periods of time of around 48 hours. Hence, any cellular
process participating in fusion must occur within the 24 hours
time window in our culture conditions. A frequently used
protocol to determine the fate of MEE cells consists on the
specific labeling of these cells with dyes that can not diffuse
trough the basal lamina. One such dye is the CCFSE, which,
once inside a cell, becomes modified and can not diffuse to
other cells. After MEE labeling with CCFSE, we forced the
contact between shelves and followed the fate of the labeled
cells at different time points during the fusion process.
Epithelial cell viability was not affected by CCFSE labeling
(data not shown). We previously reported some differences in
cell death activation within different MEE regions. Cell death
in the anterior MEE region is activated shortly after shelf
contact, whereas, cell death in the posterior MEE region is
Development 131 (1)Research article
17 PCD during palate shelf fusion
observed prior to shelf contact. It is therefore possible that
different mechanism control fusion in the anterior and posterior
MEE regions; thus, the fate of MEE cells was studied in both
regions. After 3 hours of contact, adhesion between shelves
was very strong and there were no signs of MEE cell death,
basal lamina degradation, migration or EMT (Fig. 1A, 3 hours
T/L and CCFSE). Three hours later, the MES started to disrupt
and many dying cells could be observed (Fig. 1A, 6 hours T/L
and CCFSE). Epithelial triangles were clearly identified at this
time, and stained cells were not detected outside the MES.
Twelve hours after contact, many dying cells and remnants of
basal lamina were still present within the MES (Fig. 1A, 12
hours T/L and CCFSE), but by 24 hours, all dead cells were
absent and the basal lamina was completely degraded (Fig. 1A,
24 hours T/L and CCFSE). Very few stained cells were
detected outside the MES after this period of time, and among
those, most of them were detected as dying cells (Table 1)
probably inside phagocytes.
An alternative strategy to follow the fate of developing cells
is by using viral vectors that carry a reporter gene. A recent
report shows that labeling of cells with a retroviral vector
carrying the lacZ gene allows the detection of many labeled
cells within the mesenchyme compartment after fusion
(Martinez-Alvarez et al., 2000). However, infection with a
retroviral vector requires induction of cell proliferation by
Fig. 1. Analysis of MEE cell fate using different
approaches. (A) Left lane (T/L) shows the cell death
and basal lamina degradation patterns during palate
shelf fusion in vitro (0-24 hours). Cell death detection
by TUNEL (green) and laminin immunohistochemistry
(red) were performed on the same slice. Laminin-
specific immunohistochemistry detected basal lamina
as well as blood vessels (V). Three hours after shelf
contact, MES and basal lamina were unaltered. By 6
hours, concomitant cell death and basal lamina
fragmentation indicate that MES degradation had
begun. By 12 hours, MES degradation was very
advanced, showing many dying cells within the
epithelial triangles and few dying in the MES within the
epithelial pearls (arrowhead) as indicated by the
surrounding basal lamina. No MEE cells (alive or dead)
and laminin were detected in the region surrounding the
MES at the end of culture (24 hours). MEE cells of
palates of equivalent stages were labeled with CCFSE
or Ad-lacZ (LacZ in figure) before contact and their
fate was analyzed after 3, 6, 12 and 24 hours in anterior
and posterior palatal regions. Epithelial pearls were
evident at 6 and 12 hours after contact (arrowheads)
using both labeling protocols. At the end of culture (24
hours) no labeled cells were detected in the MES. At no
time was the presence of labeled cells evident in the
mesenchyme compartment. By joining one wild-type
shelf with one from the EGFP mouse strain, we
produced chimeric palates (WT↔EGFP). At no time
were EGFP-positive cells detected in the wild-type
mesenchyme compartment of these chimeric palates.
Interestingly, chimeric palates showed intercalation of
MEE cells (asterisks), at the time abundant apoptotic
bodies were detected (arrows). After 24 hours of
culture, fusion was complete and no mesenchymal cell
migration was detected between halves. et, epithelial
triangles; mes, medial epithelial seam. (B) Although no
EGFP-positive cells were detected in the wild-type
mesenchyme compartment of chimaeric palates, EGFP-positive cells (green) were detected in the wild-type oral and nasal epithelia.
(C) CCFSE-labeled cells are not detected in the mesenchyme compartment when cell death is inhibited by z-VAD. Scale bar: 100 µm.
Table 1. Quantification of MEE labeled cells in the
mesenchyme compartment after fusion
94 (12 µm)
100 (12 µm)
60 (50 µm)
MEE cells of palates were labeled with CCFSE or Ad-lacZ before contact
and cultured for 24 hours. CCFSE labeled palates were also cultured in the
presence of z-VAD. Slices from those palates (three for each condition) were
produced and labeled cells in the mesenchyme compartment were searched
along anterior and posterior palatal regions. Cell death was also determined in
the same samples by the TUNEL technique. n.d., not determined.
serum. As serum could cause artifacts during fusion, we
preferred to use an adenovirus-based vector, also carrying the
lacZ gene, which does not require serum for infection. Cell
labeling with the adenoviral vector used here was more
extensive than that reported using the retroviral one. As with
the dye, most of the MEE infected cells were dying and none
was detected outside the MES (Fig. 1A, LacZ column; Table
We also followed the fate of MEE cells by forming
chimaeric palates between a wild-type shelf and a shelf from
a mouse embryo that expresses constitutively the GFP protein
(WT↔EGFP). In this situation, the cells that undergo EMT
would be detected only if they cross the MES. That is,
transdifferentiated cells that migrated within the same shelf
would not be detected. Of course, in principle, mesenchymal
cells present prior to shelf contact could also migrate between
shelves during or after fusion. No cells were detected to cross
the MES (Fig. 1A, WT↔EGFP), supporting the low frequency
of EMT occurrence and also the limited migratory ability of
mesenchymal cells from one shelf to another. Regardless of the
behavior of MEE and mesenchymal cells, nasal and oral
epithelial cells were observed to migrate between shelves (Fig.
1B). From these experiments, it was also interesting to observe
that as the double adhered epithelial layer turned into a single
epithelial layer, intercalation of epithelial cells from both
shelves (i.e. EGFP-positive and EGFP-negative cells) became
obvious (Fig. 1A, 6-12 hours/WT↔EGFP).
Fate of MEE cells in a palate slice culture system
In order to follow the fate of MEE cells continuously during
MES degeneration, we established a slice culture system in
which cell migration can be studied in further detail. MEE cells
were labeled with CCFSE and shelves were put in contact as
described above. Three hours later, 200 µm slices were
obtained. At this time, MES appeared intact without signs of
degeneration (Fig. 2; 3 hours). Next, slices were cultured as
described in the Materials and methods. In slices cultured for
3 hours (i.e. 6 hours after contact), the MES thinned; 6 hours
later it seemed fragmented (Fig. 2; 6 hours and 12 hours,
respectively). At this later time, no labeled cells were found
outside the MES. Twenty-four hours after contact, the MES
could not be visualized by phase-contrast microscopy, but
fluorescent cells were still detected within the fusion region
(Fig. 2, 24 hours). The great majority of remaining cells at this
time was of dying cells (Fig. 2, TUNEL). Therefore, in this
culture system, MEE fusion takes place in the absence of
detectable EMT contribution.
Relationship between cell death and basal lamina
Despite the failure to detect EMT during MEE fusion in our in
vitro system, it is still possible that it takes place but it is
immediately followed by cell death. Under normal conditions,
cell death would preclude the detection of EMT as the cause
of MES degeneration. To address this issue, we blocked cell
death with an inhibitor of caspases, z-VAD, and searched for
labeled cells in the mesenchyme compartment. We previously
showed that this inhibitor is very effective for blocking cell
death in rugae and MES (Cuervo et al., 2002) (see also Fig. 4).
No such labeled cells were found in treated palates (Fig. 1C;
Table 1). These results indicate first, that failure to detect MEE
transdifferentiated cells was not because they died soon after
differentiation, and second, that EMT can not compensate for
MES degeneration in the absence of cell death.
In the previous experiments, we also assessed the integrity
of the basal lamina. To allow EMT, the basal lamina needs
to degrade by a mechanism involving metalloproteinases
(MMPs). It has actually been proposed that basal lamina
degradation is one of the earliest events of EMT (Lochter et
al., 1997; Song et al., 2000). In agreement with the rare
occurrence of EMT during shelf fusion, we did not detect basal
lamina degradation when cell death was blocked (Fig. 3, z-
VAD). Interestingly, MMP inhibition prevented basal lamina
degradation and blocked fusion without affecting the
occurrence of cell death (Fig. 3, MMI). Toxic effects, or other
different than preventing basal lamina degradation, by MMP
inhibitors were not observed in these experiments, given that
palate shelves looked histologically unaltered and cell death
was not detected in ectopic regions (Fig. 3 and data not shown).
Therefore, it appears that cell death is not induced by the basal
lamina degradation that should accompany EMT, but rather
basal lamina degradation appears to be a secondary event
activated by the dying cells (‘cataptosis’; see Discussion).
To obtain supporting evidence for the occurrence of
cataptosis, we treated untouched palate shelves with RA, a very
strong MEE cell death inducer. Under these conditions, RA-
induced cell death was also accompanied by basal lamina
degradation (Fig. 4, mee/RA). Basal lamina degradation was
not an independent event regulated by RA, as cell death
inhibition under this condition also inhibited basal lamina
degradation (Fig. 4, mee/z-VAD/RA). Interestingly, cataptosis
was restricted to the MEE region, as cell death in rugae
epithelium (natural or RA-induced) had no effect in the
integrity of basal lamina (Fig. 4, control and rugae/RA).
Finally, a general activation of cell death by staurosporin also
induced basal lamina degradation specifically in the MEE
region and preferentially in regions underlying apoptotic cells,
whereas basal lamina remained intact in the surrounding tissue
(Fig. 4/staurosporine). Therefore, these data suggest that as a
Development 131 (1)Research article
Fig. 2. Time-course analysis of MES degeneration in a live palate
slice. MEE cells of 14.5 dpc palate shelves were labeled with CCFSE
(bright signal) before contact. Three hours after contact, 200 µm
transverse slices were produced. Selected individual slices were
cultured, and micrographs of same slices were taken at 3, 6, 12 and
24 hours. At the end, cultured slices were processed for cell death
detection (TUNEL positive, red). Note the accumulation of labeled
cells in epithelial triangles at 3 and 6 hours of culture (blue arrows).
At 6 and 12 hours of culture, the fragmented MES was obvious. At
the end of culture (24 hours), the remaining labeled cells were
detected as dying cells and none was clearly detected in the
mesenchyme. These experiments were repeated more than three
times at least in triplicate for each condition. Scale bar: 100 µm.
19 PCD during palate shelf fusion
consequence of cell death activation in the MES, the MMPs
responsible for basal lamina degradation gets activated.
Migration of the periderm cells associated to the
As previously shown, most MEE cells die during fusion.
However, another component associated to the MEE is the
periderm that covers most of the shelf surface that first comes
into contact with the opposite shelf. In order to label periderm
cells preferentially, we stained palate shelves for a very short
period of time (i.e. 30 seconds). The selectivity of this staining
procedure can be clearly seen in slices of these preparations
(see, for example, Fig. 5, CCFSE/2 hours). To determine the
fate of periderm cells, they were followed during the fusion
process. In contrast to MEE cells that appeared to die in situ,
periderm cells migrated to the oral and nasal ends of the MES,
contributing to the formation of the epithelial triangles, where
most of them died (Fig. 5, 2-8 hours).
To gain insights on the role of periderm cells during fusion,
we inhibited cell migration with cytochalasin D, which blocks
actin polymerization. As expected, periderm cells did not
migrate, epithelial triangles did not form and complete
adhesion did not occur (Fig. 6A). Interestingly, cell death in
both basal MEE and the overlying periderm was not triggered
in the presence of cytochalasin D (Fig. 6B). Blockade of actin
polimerization-depolimerization does not appear to interfere in
general with the cell death execution process, as the drug did
not modify cell death in rugae (Fig. 6C). Cycloheximide also
inhibited migration and the epithelial triangles did not form;
however, in this case, cell death in rugae was affected as well
(data not shown). These data together suggest that migration
of periderm cells out of the MEE area is essential to initiate
the fusion process.
In the previous experiments, it is possible that cell death was
Fig. 3. Relationship between cell death and basal lamina degradation.
Shelves of developing palates were put in contact and cultured in the
presence of either a caspase inhibitor (z-VAD) or a metalloproteinase
inhibitor (MMI). At the end of culture (24 hours), cell death (green)
and laminin (red) were detected in the same palate slice by TUNEL
and specific immunohistochemistry, respectively. After 12 hours in
culture, control palates show an advanced MEE cell death and basal
lamina degradation. At the end of culture, MEE and basal lamina
completely disappeared. However, z-VAD treatment inhibited cell
death but basal lamina remained intact. Application of 10 µM
BB3103 (MMI) did not alter the apoptotic fate of MEE cells but, as
expected, inhibited basal lamina degradation (arrowhead). These
experiments were repeated more than three times with at least a
triplicate for each condition. v, blood vessel. Scale bar: 50 µm.
Fig. 4. Activation of basal lamina degradation by MEE cell death
stimuli. Individual palate shelves were cultured without contact in
the presence of retinoic acid (RA; a MEE cell death activator) or
staurosporine (a broad-spectrum cell death activator). At the end of
culture (10 hours), cell death (green) and laminin (red) were detected
in the same palate slice by TUNEL and specific
immunohistochemistry, respectively. RA induced extensive cell death
in the MEE and rugae (r) of isolated shelves, but basal lamina
degraded only in the MEE apoptotic region (arrows). Treatment with
z-VAD blocked RA-induced cell death and basal lamina remained
intact. Generalized induction of epithelial cell death with
staurosporine activated the degradation of the basal lamina
underlying the dying MEE cells (arrowheads), whereas basal lamina
underlying the MEE adjacent epithelium (left from the arrow) was
unaffected. These experiments were repeated more than three times
at least in triplicate for each condition. Scale bar: 100 µm.
not activated in the MEE because periderm cells act as a barrier
for intimate contact between shelves, a conceivable
requirement for cell death activation and fusion (Cuervo et al.,
2002). To test this possibility, we removed this cell layer by
controlled trypsin treatment (see Materials and methods).
Palate shelves lacking periderm cells adhered, activated cell
death and fused. However, epithelial triangles did not form,
resulting in a thinner secondary palate (Fig. 7A). Interestingly,
viability of MEE cells in isolated shelves markedly decreased
when periderm cells were removed (Fig. 7B). In conclusion,
periderm cells do not appear to be necessary for the fusion
process itself, but they need to migrate out of the MES to allow
contact and cell death activation.
Our studies indicate that the primary and major fate of MEE
cells of secondary palate shelves is death, a requirement for
MES degeneration and fusion. We did not find evidence of
EMT, and migration could be detected only for the periderm
cells that overlie the MEE. Furthermore, we found, for the first
time, that basal lamina degrades as a consequence of cell death,
emphasizing the relevance of this latter process in shelf fusion.
Several years ago, Fitchett and Hay (Fitchett and Hay, 1989)
presented the first evidence to suggest that EMT causes MES
degradation. Subsequently, other authors presented additional
evidences supporting this role of EMT (Kaartinen et al., 1997;
Martinez-Alvarez et al., 2000; Shuler et al., 1991; Shuler et al.,
1992). Currently, this idea prevails, in some cases neglecting the
participation of cell death (Young et al., 2000). In some of
our experiments, we used experimental strategies similar to
those used in previous reports, but using an improved palate
culture system in which shelf fusion occurs within a similar
time window as observed in vivo. Under these conditions,
we were unable to detect any obvious participation of EMT.
The few labeled cells found around the MES at the end of
culture were undergoing cell death (Table 1). Furthermore,
when cell death was prevented, EMT was still not detected,
indicating that EMT does not compensate for the inability
to eliminate MEE by cell death. Because it is possible that
individual transdifferentiated cells escaped to our detection
methods, we analyzed each MEE cell within a 200 µm MES
region using a novel slice culture system. Again, we failed
to detect any evidence of EMT and found, instead, that most
MEE cells were dying.
Why the discrepancy between our data and those previous
reports? With some exceptions (Sun et al., 1998), very few
MEE cells have been reported to undergo EMT, and in most
reports quantitative analyses are lacking. Moreover, in none of
these reports, has it been determined whether the assumed
transdifferentiated cells are the dying cells or whether they are
phagocytes containing dying cells instead. These drawbacks
make it difficult to estimate the contribution of EMT to MES
degeneration. It is also possible that palate shelf fusion does
not occur equally along the rostrocaudal axis, which might
explain the discrepancies if the studies were performed at
various points along this axis. We considered this possibility
and, thus, studied MES degeneration along the complete
length of the rostrocaudal axis. Another possible explanation
for the contrasting results obtained in our experiments is the
use of an improved culture system. Dissected shelves are
usually put together and then cultured in the presence of serum
to allow fusion. We have found that our culture system allows
a more precise contact between shelves and an efficient fusion
in the absence of serum. Shuler et al. (Shuler et al., 1992) also
labeled MEE in vivo with DiI and showed that clumps of
labeled cells remain around the fusion line. It is possible that
artifactual staining occurred in those experiments (DiI can
easily precipitate), as individual cells could not be visualized
and DiI membrane incorporation would not ensure the transfer
of the dye to other more internal cells. Despite these
apparently conflicting results, we propose that cell death is the
major contributor to MES degeneration, even considering a
low occurrence of EMT.
Development 131 (1)Research article
Fig. 5. Analysis of periderm cell fate. Periderm cells were
labeled with CCFSE as described in the Materials and methods,
and palates were cultured for 8 hours. Samples were analyzed
every 2 hours. At the beginning of culture (2 hours), periderm
cells were confined to the MES middle line between the two
basal MEE; cell death was not detected at this time. As fusion
proceeded (4, 6 and 8 hours), accumulation of labeled cells
occurred at the apex of the MES, constituting a large proportion
of epithelial triangle cells. Most labeled cells died within the
epithelial triangles (yellow cells; see also Fig. 5). As noted here,
basal MEE cells appeared to die in situ within the epithelial
pearls (arrows at 8 hours; compare with Fig. 1). Arrowheads
indicate autofluorescent erythrocytes. These experiments were
repeated more than three times at least in triplicate for each
condition. Scale bar: 100 µm.
21 PCD during palate shelf fusion
MEE cell migration has also been proposed as a mechanism
for MES degeneration (Carette and Ferguson, 1992). However,
those studies did not take into account the migration of the
periderm cells that overlay the MEE. In our study, we stained
periderm cells preferentially and demonstrated that soon after
contact they migrate toward the oral and nasal cavities and
form the epithelial triangles. In keeping with this observation,
when migration was blocked or periderm cells were selectively
eliminated, the epithelial triangles did not form. Shelf fusion
did take place in the absence of periderm cells, but it resulted
in a thinner palate. It has been considered that periderm cells
shed before contact (Fitchett and Hay, 1989); our data indicate,
however, that epithelial triangles result from periderm cell
migration, a process that appears to be necessary for proper
fusion. Furthermore, a relevant finding of the present work was
that periderm cell migration plays role in activation of cell
death of both periderm and basal MEE cells. Periderm cells are
Fig. 6. Effect of inhibition of periderm cell migration on cell death
and fusion. After periderm cell labeling (green), palate shelves were
put in contact and cultured in the presence or absence of 6 µM
cytochalasin D for 10 hours. (A) When cytochalasin D was included
in the medium, palate morphology showed the lack of epithelial
triangles (et) and weak shelf adhesion. To detect cell death, palates
were either stained in whole-mount with Acridine Orange (C; bright
spots) or slices processed for the TUNEL technique (B; red).
(B) Periderm cells of control palates died within epithelial triangles
(yellow; see also Fig. 3), whereas those from cytochalasin D-treated
palates did not reach the oral and nasal closures and did not die
(green cells; arrowheads). (C) Specific reduction in cell death was
observed in the mes of cytochalasin D-treated palates with a
minimum effect in rugae (r). These experiments were repeated more
than three times at least in triplicate for each condition. Scale bars: in
A, 100 µm for A; in C, 500 µm for C.
Fig. 7. Effect of periderm cell removal on basal MEE cell viability
and shelf fusion. Periderm cells were removed by washing the MEE
region after controlled trypsin digestion performed on 14.5 dpc
palate shelves before contact. Control palates were also treated with
trypsin but washing was not performed. (A) Treated shelves were put
in contact and fusion was analyzed 24 hours later by standard
Hematoxylin-Eosin staining (HE). Although MES degenerated,
proper fusion between ‘denuded’ palate shelves did not occur. Note
the absence of epithelial triangles and marked reduction in MES
thickness when compared with a control sample (brackets).
(B) Isolated halves were cultured for 10 hours and cell death
analyzed with the TUNEL technique. More dying MEE cells were
detected in ‘denuded’ palates (i.e. without periderm cells). These
experiments were repeated more than three times at least in triplicate
for each condition. Scale bar: 100 µm.
likely to produce the filopodia and to be the source of the
proteoglycans required for shelf adhesion (Gato et al., 2002;
Taya et al., 1999). We propose that periderm cell migration is
relevant for the efficient shelf fusion regulating adhesion and
cell death activation.
Anoikis is a term given to the process of cell death induced
by the lack of contact with the extracellular matrix. The basal
lamina has been considered to be an essential survival factor
for epithelial cells in vitro and in vivo (Coucouvanis and
Martin, 1995; Ruoslahti and Reed, 1994). For example, it has
been shown that during mammary gland involution or Müller
duct regression, disrupting the underlying extracellular matrix
induces epithelial cell death (Pullan et al., 1996; Roberts et al.,
2002). This prompted us to assess whether the trigger for MEE
cell death activation was basal lamina degradation. Blocking
basal lamina degradation by inhibiting MMP activity, however,
had no effect on cell death. These results contrast with two
recent reports (Blavier et al., 2001; Brown et al., 2002) showing
partial or no MEE degeneration in the presence of the same
MMP inhibitor used here. In our experiments, we demonstrated
[with high reproducibility even at the low inhibitor dose (10
µM)] that the basal lamina was intact. In the aforementioned
reports, the integrity of basal lamina was not demonstrated, and
hence, their results can be interpreted as incomplete MEE
degeneration. Our observations in the presence of the MMP
inhibitor do not imply that basal lamina has no survival activity
on MEE cells, but suggests that MEE cell death is not triggered
by basal lamina degradation. On the contrary, we found that
cell death activates basal lamina degradation (see Fig. 4). To
our knowledge, this is an unprecedented finding that gives a
new function to the process of cell death. We propose the term
‘cataptosis’ (a Greek word meaning downfall) to describe this
phenomenon. Cataptosis may occur in different developmental
process involving tissue regression, in order to coordinate cell
degeneration with extracellular matrix degradation. In the
palate, the basal lamina degradation activity is restricted to the
dying MEE cells, suggesting that specific factors give them this
property (see below). This conclusion is also in opposition to
the participation of EMT, as inhibition of basal lamina
degradation can block EMT (Song et al., 2000), and MMPs can
directly induce EMT (Lochter et al., 1997).
Collagen IV and laminin, the most abundant components of
basal lamina, are likely to be the major MMPs substrates
during basal lamina degradation. MMPs are found
extracellularly and also bound to the plasma membrane (MT-
MMPs) (Birkedal-Hansen, 1995). The MMP activity could
be regulated at different levels. MMP gene expression is
characteristic during tissue remodeling but postranslational
regulation is crucial for enzymatic activity. With the exception
of MT-MMPs, MMPs are synthesized as inactive proenzymes
that need to be processed by other enzymes, such as plasmin
or other MMPs, to become active (Nagase, 1997).
Furthermore, MMP activity can be negatively regulated by
direct binding of proteins such as members of the tissue
inhibitor of metalloproteinases (TIMP) family (Gomez et al.,
1997). Among the several MMPs and their inhibitors described
to date, MT1-Mmp (Mmp14 – Mouse Genome Informatics),
Mmp2, Mmp3, Mmp9, Mmp13, Timp1 and Timp2 are expressed
in the developing palate (Blavier et al., 2001; Morris-Wiman
et al., 2000). MT1-Mmp, Mmp13 and Timp2 are specifically
expressed in the MEE at the time fusion occurs (Blavier et al.,
2001). These genes could be the ones that provide the MEE
with the distinct ability to activate cataptosis. MMP2 and
MMP13 can digest collagen IV (Knäuper et al., 1997),
suggesting a role for these enzymes in basal lamina
degradation, although their role in vivo has not been
demonstrated. TIMP2 could prevent extracellular matrix
degradation; it has been shown, however, that it is also relevant
for the efficient MMP2 activation in vivo (Wang et al., 2000).
MT1-MMP, conversely, could be the initiator of a cascade of
MMPs involved in cataptosis. Recently, it was reported that
MT1-MMP and MMP2 translocate from the cytoplasm to the
plasma membrane in endothelial cells upon stimulation of
apoptosis (Levkau et al., 2002). Furthermore, MT1-MMP can
activate proMMP13 (Knäuper et al., 2002). However, as Mt1-
Mmp null mutants do not have an obvious palate phenotype
(Holmbeck et al., 1999), the latter hypothesis would imply that
redundant mechanisms are activated during MES cataptosis.
In summary, secondary palate shelf fusion process can be
described as follows (Fig. 8). Initially, shelves approach each
other until contact is made probably between filopodia from
Development 131 (1)Research article
Fig. 8. Schematic representation of palate shelf fusion. (A) Initially,
shelves approach each other at the time the periderm cells (yellow
cells) overlying the basal MEE cells (white cells) emit filopodia.
(B) First contact and adhesion occurs between periderm cells;
proteoglycans appear to be important at this stage. Adhesion
becomes stronger as periderm cells move up and down (arrows) the
MES (bracket) forming the epithelial triangles (et). (C) Basal MEE
cells of each shelf intercalate (convergent extension) resulting in a
single epithelial layer. (D) MES breaks up and epithelial pearls (ep)
form; periderm and MEE cells start to die within epithelial triangles
and epithelial pearls, respectively (red cells). (E) MES, which is
composed of periderm and basal MEE cells, essentially degenerates
by cell death; dying cells activate basal lamina degradation
(cataptosis; broken orange line). (F) Fusion is complete without a
major mesenchymal cell movement across the midline; some oral
and nasal epithelial cells do move across the middle line (double-
headed arrows). Pink cells represent mesenchymal cells. Orange
lines represent basal lamina.
23 PCD during palate shelf fusion
periderm cells and with the help of proteoglycans. More
intimate MEE contact proceeds through a process that is
accompanied by the migration of periderm cells to the oral
and nasal ends. Progressive adhesion appears to be controlled
by periderm cell migration. If this periderm cell migration is
not in place, a thinner palate would form. Periderm cells in
the epithelial triangles could also be important for sealing the
ends of the MES. Chimaeric palates CD1↔EGFP clearly
reveal the intercalation between MEE cells from each shelf
as recently reported (Tudela et al., 2002). This process, which
results in a single epithelial sheath, represent a classical
convergent extension phenomenon (Wallingford et al., 2002),
and cause MES growth in both anteroposterior and oronasal
axes. Strong adhesion between shelves probably originates
from this MEE cell intercalation, and we propose that it is up
to this stage that cell death is activated. Finally, dying cells
actively promote the activation of MMPs such as MT1-MMP
and MMP13 that cause basal lamina degradation. Basal
lamina initially breaks down in fragments generating the
epithelial pearls that later degenerate to produce the fused
Two key molecules have been identified to be relevant in the
fusion process. Retinoic acid appears to be essential for the
control of cell death (Cuervo et al., 2002), and consequently
for basal lamina degradation. TGFβ3, however, has been
proposed to be involved in the control of EMT (Kaartinen et
al., 1997; Sun et al., 1998). In the heart, strong evidence
supports a role of TGFβ3 in EMT (Ramsdell and Markwald,
1997). Tgfb3 knock-out mice display cleft palate, nevertheless,
the specific TGFβ3 function that is responsible for this
phenotype remains to be elucidated. Tgfb3–/–shelves lack the
characteristic MEE associated filopodia normally observed
prior to shelf contact, and also show a significant decrease in
proteoglycans on the MEE surface (Gato et al., 2002; Taya et
al., 1999). These data suggest that TGFβ3 is crucial for the
initial contact and during adhesion between shelves. TGFβ3
could also play a role in the control of periderm cell migration.
It is interesting to note that palates treated with cytochalasin
D are a phenocopy of those from Tgfb3–/–embryos (see Fig.
6). Palates in both conditions show poor shelf adhesion,
reduced cell death and lack of epithelial triangles (Martinez-
Alvarez et al., 2000; Taya et al., 1999). Inhibition of actin
polymerization would block cell motility, resulting in absence
of filopodia and cell migration. In our experiments, it is
unlikely that defective contact and adhesion, owing to lack of
filopodia, is causing the unfused palate phenotype, because
shelf contact was forced. Therefore, TGFβ3 might have a
migration-promoting activity on periderm cells, as it appears
to occur on endocardial cells transformed to mesenchyme
(Ramsdell and Markwald, 1997). Despite the proapoptotic
activity that has been demonstrated for TGFβ3 (Nguyen and
Pollard, 2000; Opperman et al., 2000; Dunker et al., 2002),
reduced MEE cell death in palates from Tgfb3–/–mice
(Martinez-Alvarez et al., 2000) could result from defective
adhesion or periderm cell migration. Additionally, TGFβ3
could influence the MMP activity initiated by cell death,
because it is known that it regulates Mmp13 expression in the
palate (Blavier et al., 2001). Regulation in the opposite
direction, however, could also occur: MMP13 could process
the TGFβ3 precursor and in this way regulate its activity
(Sternlicht and Werb, 2001; Yu and Stamenkovic, 2000).
Further experiments are needed to determine the significance
of this positive regulatory loop in basal lamina degradation.
We are grateful to Elizabeth Mata, Sergio González, Barbara
Mondragón and Concepción Valencia for assistance in mice care and
reproduction; Dr L. McCrae (British Biotech, Oxford, UK) and Dr
Andras Nagy (Samuel Lunenfeld Research Institute, Toronto, Canada)
for the gift of BB3103 inhibitor and the EGFP mouse strain,
respectively; Dr Horacio Merchant and Dr Rosana Sánchez for
continuous valuable support; and Dr Alfredo Varela-Echavarría and
Dr Katia Del Río-Tsonis for the careful reading of the manuscript.
This work was funded by the Consejo Nacional de Ciencia y
Tecnología (grants 31730-N and 39930-Q, and doctoral fellowships
to R.C.) and the Dirección General de Apoyo al Personal Académico
Birkedal-Hansen, H. (1995). Proteolytic remodeling of extracellular matrix.
Curr. Opin. Cell Biol 7, 728-735.
Blavier, L., Lazaryev, A., Groffen, J., Heisterkamp, N., DeClerck, Y. A.
and Kaartinen, V. (2001). TGF-beta 3-induced palatogenesis requires
matrix metalloproteinases. Mol. Biol. Cell 12, 1457-1466.
Boyer, B., Valles, A. M. and Thiery, J. P. (1996). Model systems of
epithelium-mesenchyme transitions. Acta Anatomica 156, 227-239.
Brown, N. L., Yarram, S. J., Mansell, J. P. and Sandy, J. R. (2002). Matrix
metalloproteinases have a role in palatogenesis. J. Dent. Res. 81, 826-830.
Brunet, C. L., Sharpe, P. M. and Ferguson, M. W. (1993). The distribution
of epidermal growth factor binding sites in the developing mouse palate. Int.
J. Dev. Biol. 37, 451-458.
Carette, M. J. and Ferguson, M. W. (1992). The fate of medial edge epithelial
cells during palatal fusion in vitro: an analysis by DiI labelling and confocal
microscopy. Development 114, 379-388.
Citterio, H. L. and Gaillard, D. A. (1994). Expression of transforming growth
factor alpha (TGF alpha), epidermal growth factor receptor (EGF-R) and
cell proliferation during human palatogenesis: an immunohistochemical
study. Int. J. Dev. Biol. 38, 499-505.
Conlon, R. A., Reaume, A. G. and Rossant, J. (1995). Notch 1 is required
for the coordinate segmentation of somites. Development 121, 1533-1545.
Coucouvanis, E. and Martin, G. R. (1995). Signals for death and survival:
a two-step mechanism for cavitation in the vertebrate embryo. Cell 83, 279-
Cuervo, R., Valencia, C., Chandraratna, R. A. and Covarrubias, L. (2002).
Programmed cell death is required for palate shelf fusion and is regulated
by retinoic acid. Dev. Biol. 245, 145-156.
DeAngelis, V. and Nalbandian, J. (1968). Ultrastructure of mouse and rat
palatal processes prior to and during secondary palate formation. Arch. Oral
Biol. 13, 601-608.
Duband, J. L., Monier, F., Delannet, M. and Newgreen, D. (1995).
Epithelium-mesenchyme transition during neural crest development. Acta
Anat. (Basel) 154, 63-78.
Dunker, N., Schmitt, K. and Krieglstein, K. (2002). TGF-beta is required
for programmed cell death in interdigital webs of the developing mouse
limb. Mech. Dev. 113, 111-120.
Farbman, A. I. (1968). Electron microscope study of palate fusion in mouse
embryos. Dev. Biol. 18, 93-116.
Ferguson, M. W. (1988). Palate development. Development Suppl. 41-60.
Fitchett, J. E. and Hay, E. D. (1989). Medial edge epithelium transforms to
mesenchyme after embryonic palatal shelves fuse. Dev. Biol. 131, 455-474.
Fitzpatrick, D. R., Denhez, F., Kondaiah, P. and Akhurst, R. J. (1990).
Differential expression of TGF beta isoforms in murine palatogenesis.
Development 109, 585-595.
Gato, A., Martinez, M. L., Tudela, C., Alonso, I., Moro, J. A., Formoso,
M. A., Ferguson, M. W. and Martinez-Alvarez, C. (2002). TGF-beta(3)-
induced chondroitin sulphate proteoglycan mediates palatal shelf adhesion.
Dev. Biol. 250, 393-405.
Gomez, D. E., Alonso, D. F., Yoshiji, H. and Thorgeirsson, U. P. (1997).
Tissue inhibitors of metalloproteinases: structure, regulation and biological
functions. Eur. J. Cell Biol. 74, 111-122.
Griffith, C. M. and Hay, E. D. (1992). Epithelial-mesenchymal
transformation during palatal fusion: carboxyfluorescein traces cells at light
and electron microscopic levels. Development 116, 1087-1099.
24 Download full-text
Hadjantonakis, A. K., Gertsenstein, M., Ikawa, M., Okabe, M. and Nagy,
A. (1998). Generating green fluorescent mice by germline transmission of
green fluorescent ES cells. Mech. Dev. 76, 79-90.
Holmbeck, K., Bianco, P., Caterina, J., Yamada, S., Kromer, M.,
Kuznetsov, S. A., Mankani, M., Robey, P. G., Poole, A. R., Pidoux, I. et
al. (1999). MT1-MMP-deficient mice develop dwarfism, osteopenia,
arthritis, and conective tissue disease due to inadequate collagena turnover.
Cell 99, 81-92.
Kaartinen, V., Cui, X. M., Heisterkamp, N., Groffen, J. and Shuler, C. F.
(1997). Transforming growth factor-beta3 regulates transdifferentiation of
medial edge epithelium during palatal fusion and associated degradation of
the basement membrane. Dev. Dyn. 209, 255-260.
Knäuper, V., Bailey, L., Worley, J. R., Soloway, P., Patterson, M. L. and
Murphy, G. (2002). Cellular activation of proMMP-13 by MT1-MMP
depends on the C-terminal domain of MMP-13. FEBS Lett. 532, 127-130.
Knäuper, V., Cowell, S., Smith, B., Lopez-Otin, C., O’Shea, M., Morris,
H., Zardy, L. and Murphy, G. (1997). The role of C-terminal domain of
human collagenase-3 (MMP-13) in the activation of procollagenase-3,
substrate specificity, and tissue inhibitor of metalloproteinase interaction. J.
Biol. Chem. 272, 7608-7616.
Knight, B., Laukaitis, C., Akhtar, N., Hotchin, N. A., Edlund, M. and
Horwitz, A. R. (1999). Visualizing muscle cell migration in situ. Curr. Biol.
Levkau, B., Kenagy, R. D., Karsan, A., Weitkamp, B., Clowes, A. W., Ross,
R. and Raines, E. W. (2002). Activation of metalloproteinases and their
association with integrins: an auxiliary apoptotic pathway in human
endothelial cells. Cell Death Differ. 9, 1360-1367.
Lochter, A., Galosy, S., Muschler, J., Freedman, N., Werb, Z. and Bissell,
M. J. (1997). Matrix metalloproteinase stromelysine-1 triggers a cascade of
molecular alterations that leads to stable epithelial-to-mesenchymal
conversion and a premalignant phenotype in mammary epithelial cells. J.
Cell. Biol. 139, 1861-1872.
Markwald, R. R., Fitzharris, T. P. and Smith, W. N. (1975). Structural
analysis of endocardial cytodifferentiation. Dev. Biol. 42, 160-180.
Martinez-Alvarez, C., Tudela, C., Perez-Miguelsanz, J., O’Kane, S.,
Puerta, J. and Ferguson, M. W. (2000). Medial edge epithelial cell fate
during palatal fusion. Dev. Biol. 220, 343-357.
Mori, C., Nakamura, N., Okamoto, Y., Osawa, M. and Shiota, K. (1994).
Cytochemical identification of programmed cell death in the fusing fetal
mouse palate by specific labelling of DNA fragmentation. Anat. Embryol.
Morris-Wiman, J., Burch, H. and Basco, E. (2000). Temporospatial
distribution of matrix metalloproteinases and tissue inhibitors of matrix
metalloproteinases during murine secondary palate morphogenesis. Anat.
Embryol. 202, 129-141.
Nagase, H. (1997). Activation mechanisms of matrix metalloproteinases. Biol.
Chem. 378, 151-160.
Nguyen, A. V. and Pollard, J. W. (2000). Transforming growth factor β3
induces cell death during the first stage of mammary gland involution.
Development 127, 3107-3118.
Opperman, L. A., Adab, K. and Gakunga, P. T. (2000). Transforming factor-
beta 2 and beta 3 regulate fetal rat cranial suture morphogenesis by
regulating rates of cell proliferation and apoptosis. Dev. Dyn. 219, 237-247.
Pullan, S., Wilson, J., Metcalfe, A., Edwards, G. M., Goberdhan, N., Tilly,
J., Hickman, J. A., Dive, C. and Streuli, C. H. (1996). Requirement of
basement membrane for the suppression of programmed cell death in
mammary epithelium. J. Cell. Sci. 109, 631-642.
Ramsdell, A. F. and Markwald, R. R. (1997). Induction of endocardial
cushion tissue in the avian heart is regulated, in part, by TGF-beta 3
mediated autocrine signaling. Dev. Biol. 188, 64-74.
Roberts, L. M., Visser, J. A. and Ingraham, H. A. (2002). Involvement of
a matrix metalloproteinase in MIS-induced cell death during urogenital
development. Development 126, 1487-1496.
Ruoslahti, E. and Reed, J. C. (1994). Anchorage dependence, integrins, and
apoptosis. Cell 77, 477-478.
Shuler, C. F., Guo, Y., Majumder, A. and Luo, R. Y. (1991). Molecular and
morphologic changes during the epithelial-mesenchymal transformation of
palatal shelf medial edge epithelium in vitro. Int. J. Dev. Biol. 35, 463-472.
Shuler, C. F., Halpern, D. E., Guo, Y. and Sank, A. C. (1992). Medial edge
epithelium fate traced by cell lineage analysis during epithelial-
mesenchymal transformation in vivo. Dev. Biol. 154, 318-330.
Smiley, G. R. and Dixon, A. D. (1968). Fine structure of midline epithelium
in the developing palate of the mouse. Anat. Rec. 161, 293-310.
Song, W., Jackson, K. and McGuire, P. G. (2000). Degradation of type IV
collagen by matrix metalloproteinases is an important step in the epithelia-
mesenchyme transformation of the endocardial cushions. Dev. Biol. 227,
Sternlicht, M. D. and Werb, Z. (2001). How matrix metalloproteinases
regulate cell behavior. Annu. Rev. Cell Dev. Biol. 17, 463-516.
Sun, D., Vanderburg, C. R., Odierna, G. S. and Hay, E. D. (1998).
TGFbeta3 promotes transformation of chicken palate medial edge
epithelium to mesenchyme in vitro. Development 125, 95-105.
Taniguchi, K., Sato, N. and Uchiyama, Y. (1995). Apoptosis and
heterophagy of medial edge epithelial cells of the secondary palatine shelves
during fusion. Arch. Histol. Cytol. 58, 191-203.
Taya, Y., O’Kane, S. and Ferguson, M. W. (1999). Pathogenesis of cleft
palate in TGF-beta3 knockout mice. Development 126, 3869-3879.
Tudela, C., Formoso, M. A., Martinez, T., Perez, R., Aparicio, M., Maestro,
C., Del-Rio, A., Martinez, E., Ferguson, M. and Martinez-Alvarez, C.
(2002). TGF-beta3 is required for the adhesion and intercalation of medial
edge epithelial cells during palate fusion. Int. J. Dev. Biol. 46, 333-336.
Wallingford, J. B., Fraser, S. E. and Harland, R. M. (2002). Convergent
extension: the molecular control of polarized cell movement during
embryonic Development 2, 695-706.
Wang, Z., Juttermann, R. and Soloway, P. D. (2000). TIMP-2 is required
for efficient activation of proMMP-2 in vivo. J. Biol. Chem. 275, 26411-
Yano, H., Ohtsuru, A., Ito, M., Fuji, T. and Yamashita, S. (1996).
Involvement of c-Fos proto-oncogene during palatal fusion and interdigital
space formation in the rat. Dev. Growth. Differ 38, 351-357.
Young, D. L., Schneider, R. A., Hu, D. and Helms, J. A. (2000). Genetic
and teratologenic approaches to craniofacial development. Crit. Rev. Oral.
Biol. Med 11, 304-317.
Yu, Q. and Stamenkovic, I. (2000). Cell surface-localized matrix
metalloproteinase-9 proteolytically activates TGF-beta and promotes tumor
invasion and angiogenesis. Genes Dev. 14, 163-176.
Development 131 (1) Research article