www.elsevier.com/locate/ydbio

Developmental Biology 270 (2004) 455–473

Dynamics of neural crest-derived cell migration

in the embryonic mouse gut$

H.M. Young,a,* A.J. Bergner,a R.B. Anderson,a H. Enomoto,b J. Milbrandt,c

D.F. Newgreen,d and P.M. Whitingtona

aDepartment of Anatomy and Cell Biology, University of Melbourne, 3010 Parkville, Victoria, AustraliabLaboratory for Neuronal Differentiation and Regeneration, RIKEN Center for Developmental Biology, Chuo, Kobe, Hyogo 650-0047 Japan

cDepartments of Pathology and Internal Medicine, Washington University School of Medicine, St. Louis, MO 63110, USAdMurdoch Childrens Research Institute, Royal Children’s Hospital, 3052 Parkville, Victoria, Australia

Received for publication 13 January 2004, revised 3 March 2004, accepted 5 March 2004

Available online 15 April 2004

Abstract

Neural crest-derived cells that form the enteric nervous system undergo an extensive migration from the caudal hindbrain to colonize the

entire gastrointestinal tract. Mice in which the expression of GFP is under the control of the Ret promoter were used to visualize neural crest-

derived cell migration in the embryonic mouse gut in organ culture. Time-lapse imaging revealed that GFP+ crest-derived cells formed chains

that displayed complicated patterns of migration, with sudden and frequent changes in migratory speed and trajectories. Some of the leading

cells and their processes formed a scaffold along which later cells migrated. To examine the effect of population size on migratory behavior, a

small number of the most caudal GFP+ cells were isolated from the remainder of the population. The isolated cells migrated slower than cells in

large control populations, suggesting that migratory behavior is influenced by cell number and cell–cell contact. Previous studies have shown

that neurons differentiate among the migrating cell population, but it is unclear whether they migrate. The phenotype of migrating cells was

examined.Migrating cells expressed the neural crest cell marker, Sox10, but not neuronal markers, indicating that themajority of migratory cells

observed did not have a neuronal phenotype.

D 2004 Elsevier Inc. All rights reserved.

Keywords: Migration; Enteric; Neural crest; Chain migration

Introduction Newgreen and Young, 2002a,b; Parisi and Kapur, 2000;

Neural crest cell migration into and along the gastroin-

testinal tract is crucial for the establishment of the enteric

nervous system. Most enteric neurons and glial cells arise

from vagal level crest cells (Le Douarin and Teillet, 1973).

Sacral crest cells also contribute some enteric neurons and

glial cells (Burns and Le Douarin, 1998). In mice, vagal

crest-derived cells enter the foregut around E9.5 and then

migrate caudally (Kapur et al., 1992). Although many

molecules required for the colonization of the gut by

crest-derived cells have been identified (Gershon, 1999;

0012-1606/$ - see front matter D 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.ydbio.2004.03.015

$ Supplementary data associated with this article can be found, in the

online version, at doi: 10.1016/j.ydbio.2004.03.015.

* Corresponding author. Fax: +613-9347-5219.

E-mail address: h.young@unimelb.edu.au (H.M. Young).

Taraviras and Pachnis, 1999), little is known about the

behavior of enteric crest-derived cells at the cellular level.

The behavior of enteric crest-derived cells is likely to

be influenced by intrinsic factors (the motile nature of crest

cells) and molecules expressed by the gut mesenchyme

including GDNF and endothelin-3 (Barlow et al., 2003;

Iwashita et al., 2003; Jiang et al., 2003; Kruger et al.,

2003; Natarajan et al., 2002; Young et al., 2001). Further-

more, when the number of vagal crest cells is surgically

reduced, they do not migrate as far caudally (Burns, 2000;

Peters-van der Sanden et al., 1993; Yntema and Ham-

mond, 1954), suggesting that interactions between cells

affect migratory behavior. To date, the only information

about the behavior of enteric crest cells has been obtained

from fixed samples of gut. Some of these studies have

provided a timetable by which crest-derived cells colonize

Table 2

Secondary antisera used

Species in which primary

antisera were raised

Secondary antisera

Rabbit + Human Donkey antihuman Texas Red (1:200, Jackson

Immunoresearch, West Grove, PA, USA), and

goat anti-rabbit Alexa 488 (1:200, Molecular

Probes) or biotinylated donkey anti-rabbit

(1:100, Jackson Immunoresearch) followed by

streptavidin Cy5 (1:100, Amersham)

Rabbit + Sheep Donkey anti-sheep Alexa 594 and biotinylated

donkey anti-rabbit followed by streptavidin

Cy5

Rabbit + Rat Goat anti-rabbit Alexa 488 and donkey anti-rat

Alexa 594 (1:100)

H.M. Young et al. / Developmental Biology 270 (2004) 455–473456

the gut of many species (Burns and Le Douarin, 1998; Fu

et al., 2003; Kapur et al., 1992; Newgreen and Hartley,

1995). Several inferences have been made about the

dynamics of migrating crest-derived cells in the gut from

static studies (Conner et al., 2003). However, to understand

the mechanisms that influence the behavior of crest-de-

rived cells, it is imperative to know, rather than infer, how

they are behaving dynamically (Lichtman and Fraser,

2001).

Time-lapse imaging is a powerful tool to study neuronal

(Nadarajah et al., 2001, 2002) and neural crest cell

migration (Gilmour et al., 2002; Krull et al., 1995; Kulesa

and Fraser, 1998, 2000; Kulesa et al., 2000). The behavior

of DiI-labeled crest cells in chick embryos has been

examined in trunk and cranial regions (Krull et al.,

1995; Kulesa and Fraser, 1998, 2000; Kulesa et al.,

2000). Here, we used time-lapse imaging to analyze

crest-derived cells migrating through the gut. Migration

of enteric crest-derived cells differs from other crest cells

in two important aspects. First, the scale of migration is

greater. The distance they migrate to colonize the entire

gastrointestinal tract exceeds that of any other crest cell

population (Allan and Newgreen, 1980), and the duration

of migration is longer; the colonization of the mouse gut

by crest-derived cells takes more than one quarter of the

gestation period (Kapur, 1999). Second, in comparison to

other environments through which neural crest cells mi-

grate, the gut mesenchyme appears to be relatively uni-

form. Both of these qualities require and facilitate cell

biological studies.

Using transgenic mice in which every enteric crest-

derived cell expresses GFP, we made direct observations

of crest-derived cells within the embryonic gut. We show

that the cells migrate in chains that follow complex and

unpredictable trajectories, but in population terms the mi-

gration is predictable. Some of the leading cells and their

processes form a scaffold along which later cells migrate,

which is reminiscent of axon tract formation. Furthermore,

we found that migrating cells undergo cell division, that

reducing the number of crest-derived cells decreases the

migration speed, and that migrating cells do not exhibit a

Table 1

Primary antisera used

Antiserum Host Dilution Source or reference

B-FABP Rabbit 1:2000 Kurtz et al., 1994

Hu Human 1:2000 Gift of Dr. M. Epstein

Neurofilament M,

145 kDa

Rabbit 1:100 Chemicon, Temecula, USA

PECAM-1 Rat 1:1000 BD Biosciences Pharmingen,

San Diego, USA

PGP9.5 Rabbit 1:1000 Ultraclone, Isle of Wight, UK

PGP9.5 Sheep 1:50 The Binding Site, UK

Phox2b Rabbit 1:700 Pattyn et al., 1997

Sox10 Rabbit 1:1000 Chemicon

neuronal phenotype, although neurons are found among the

migrating cells.

Materials and methods

Animals

Mice in which cDNA encoding tau-EGFP-myc (TGM)

had been inserted into the first coding exon of the receptor

tyrosine kinase gene, Ret, were used (Enomoto et al., 2001).

Male mice heterozygous for the Ret-TGM mutation

(RetTGM/+ mice) were mated to wild-type (C57Bl6) females.

The genotype of adult RetTGM mice was determined by PCR

using the primers and conditions reported by Enomoto et al.

(2001). Time-mated, pregnant mice were killed by cervical

dislocation.

Time-lapse video microscopy

The gastrointestinal tract was dissected from each em-

bryo and placed in tissue culture medium (DMEM contain-

ing 10% foetal bovine serum, 2 mM glutamine, 0.075%

sodium bicarbonate and penicillin/streptomycin sulfate so-

lution). Guts in which GFP+ cells were present were then set

up as suspended explants as described by Hearn et al.

(1999). The region of gut to be imaged was suspended

across a ‘‘V’’ cut in a piece of black Millipore filter paper

and was held in place by attaching the mesentery to filter

paper. The preparation was placed in tissue culture medium

in a chamber of a multi-well cell culture plate (Molecular

Probes, Eugene, OR, USA) and coverslipped. Time-lapse

observations were made using a Zeiss Axioskop fixed stage

microscope equipped with a heated stage, a Zeiss 09 filter

set, FLUAR objective lenses (Zeiss, Germany), a UniBlitz

shutter and a COHU 4913 video camera connected to a

Scion LG-3 frame grabber. The hardware was controlled

with a modified version of Scion Image software (Murray et

al., 1998). Images were captured every 1–10 min, and

between 10 and 30 video frames were integrated for each

H.M. Young et al. / Developmental

exposure (giving exposure times of 0.4–1.2 s). The plane of

focus was altered manually when required.

Determination of migratory speed

The net speed of forward movement of the migratory

wavefront was determined by measuring the distance be-

tween the location of the most caudal GFP+ cell visible at one

time point and the location of the most caudal GFP+ cell at a

later time using Scion Image, and dividing by the time (a

minimum of 8 h). To estimate the mean rate of migration of an

individual cell, the total path length that the cell moved

(including forward, backward and circumferential compo-

nents) was determined using Scion Image, and divided by the

time.

Cell trajectory analysis

The trajectory of the leading (most caudal) cell was

determined at 10-min intervals by constructing a straight

line from its position with its position 10 min earlier and

measuring the angle that each vector made with the ros-

trocaudal axis of the gut, which is the axis of population

advance. Where necessary, the autofluorescence of the gut

epithelium was used to locate the rostrocaudal axis. Angles

were measured using Scion Image, and polar histograms

plotted using Origin software.

Fig. 1. (A–F) Confocal micrographs of wholemount preparations of hindgut from an

are also GFP+, although some Phox2b+ (arrows) are only weakly GFP+. In themid hi

chains and all of the Phox2b+ cells are also GFP+. (G– I) E13.5 midgut. B-FABP+

Removal of mesenchyme from the circumference of a small

region of the embryonic gut

Segments of E12.5 or E13.5 hindgut were examined

under a fluorescence microscope using a �20 lens. The

location of the most caudal GFP+ cells was determined

relative to landmarks such as the shape of the cut mesentery.

Using fine forceps and a dissecting microscope, as much as

possible of the mesenchyme was removed from the circum-

ference of a small section of gut, just rostral to the most

caudal GFP+ cells, leaving the epithelial tube intact (Fig. 9).

The behavior of the GFP+ cells that were isolated from the

rest of the population by removal of the mesenchyme was

then examined using time-lapse microscopy for 8–12 h.

Some lesioned preparations were photographed immediately

after removal of the mesenchyme to determine the number

and location of GFP+ cells present on the top surface of the

gut caudal to the lesion, placed in an incubator, fixed 8 h

later and then examined to determine the distance that the

population of cells caudal to the lesion had migrated.

Immunohistochemistry

Wholemount preparations of gut were processed for

immunohistochemistry using the methods described in

Young et al. (2002) and the antisera shown in Tables 1

and 2.

Biology 270 (2004) 455–473 457

E12.5RetTGM/+mouse. In the rostral hindgut (A–C), all of the Phox2b+ cells

ndgut, close to themigratorywavefront (D–F), most of the cells are present in

glial precursors (arrows) are weakly GFP+. Scale bar: 20 Am.

ental Biology 270 (2004) 455–473

Phenotype of migrating cells

Short, time-lapse sequences (2–10 images taken every

2.5–5 min) of GFP+ cells were taken using a �20 lens.

The gut was then fixed immediately and processed for

double-label immunohistochemistry. The region containing

H.M. Young et al. / Developm458

Fig. 2. Selected frames of the hindgut of an E12.5 mouse showing the caudal prog

cells are present in chains that follow a variety of trajectories. A polar plot showing

is shown in Fig. 5A. Scale bar: 100 Am. A movie (movie 1) of this preparation i

the migrating GFP+ cells was then re-located and exam-

ined using a MRC-1024 BioRad confocal microscope

attached to a Zeiss Axioplan microscope. We established

whether GFP+ cells that had migrated in the 5 min

immediately before fixation showed Hu-, PGP9.5-,

Sox10- or neurofilament-immunoreactivity.

ression of GFP+ neural crest cells. Time is noted in minutes (V). Most of the

the trajectories of the most caudal cell in this preparation at 10-min intervals

s available at the Developmental Biology website.

H.M. Young et al. / Developmental Biology 270 (2004) 455–473 459

Results

Use of Ret-TGM mice to visualize enteric neural

crest-derived cells

Neural crest-derived cells that colonize the embryonic

gut express the tyrosine kinase, Ret (Pachnis et al., 1993).

Recently, mice were generated in which cDNA encoding

tau-EGFP-myc (TGM) was inserted into the first coding

exon of the Ret gene (Enomoto et al., 2001). To visualize

neural crest-derived cells in the gut, embryonic mice that

were heterozygous for the Ret-TGM mutation (RetTGM/+

mice) were generated by breeding male RetTGM/+ mice with

wild-type females. Approximately 50% (145/301) of the

E10.5–E13.5 embryos screened for this study showed TGM

expression in the gut. GFP was ubiquitously expressed in

the neural crest-derived cells, including the nuclei, cyto-

plasm and processes (Figs. 1B, E, H). The intensity of GFP

was very weak at E10.5, and although it increased with age,

even at E12.5 and E13.5, the intensity of GFP was low.

To determine whether all enteric neural crest-derived cells

in the embryonic gut expressed TGM, the presence of GFP

was compared to the immunolocalization of the transcription

factor, Phox2b. Phox2b appears to be expressed by all crest-

derived cells in the E11.5–E14.5 mouse gut because all cells

labeled with other neuronal or neural crest cell markers are

Phox2b+ (Young et al., 1999, 2003). In the current study, all

Phox2b+ cells in the E11.5–E12.5 gut were GFP+ and vice

versa (Fig. 1). Behind the migratory wavefront, some

Phox2b+ cells were found to be only weakly GFP+ (Figs.

1A–C). The differentiation of enteric glial cells appears to be

associated with a down regulation of Ret (Natarajan et al.,

1999). In the embryonic mouse gut, glial precursors can be

identified using an antiserum to brain-derived fatty acid

binding protein (B-FABP). B-FABP+ cells are not present

at the migratory wavefront and cannot be detected until at

least 1 day after the passage of the migratory wavefront

through each gut region (Young et al., 2003). In the present

specimens, all B-FABP+ cells were weakly GFP+, but there

were also cells that were weakly GFP+ that were not B-

FABP+ (Figs. 1G–I).

RetTGM/TGM homozygotes lack enteric neurons and

kidneys and have defects in the sympathetic nervous sys-

tem, but RetTGM/+ mice have been reported to show no

deficits (Enomoto et al., 2001). However, to examine

whether the timetable by which neural crest-derived cells

Fig. 3. Selected frames of the midgut of an E10.5–E11 mouse showing the

caudal progression of GFP+ neural crest cells. The intensity of GFP is very

low at this age. The most caudal cell at each interval is arrowed. Time is

noted in minutes (V). During the sequence, the cells migrated around the

curve in the midgut adjacent to the umbilical artery. The caecum is just out

of the field of view. The epithelial lining of the gut (asterisks) is

autofluorescent. A polar plot showing the trajectories of the most caudal

cell in this preparation at 10-min intervals is shown in Fig. 5D. Scale bar:

100 Am. A movie (movie 4) of this preparation is available at the

Developmental Biology website.

Fig. 4. Net speed at which the migratory wavefront of GFP+ neural crest

cells migrated caudally along the gut under control conditions (left), and

when there was a reduced population (right). The net speed was determined

by measuring the distance between the location of the most caudal GFP+

cell visible at the beginning of the sequence and the location of the most

caudal GFP+ cell at the end of the sequence (a minimum of 8 h later).

Normally, the population of GFP+ cells moved caudally at around 35 Am/h

(left). A rim of mesenchyme was removed to isolate a small population of

GFP+ neural crest cells from the remainder of the population. When the

population was reduced, the net speed was significantly reduced (right).

The number in brackets is the estimated number of GFP+ neural crest cells

on the top surface of the gut caudal to the zone where the mesenchyme was

removed in each preparation.

H.M. Young et al. / Developmental Biology 270 (2004) 455–473460

colonize the gut in RetTGM/+ mice is the same as that in

wild-type mice, the location of Phox2b+ cells in the gut of

three E12.5 RetTGM/+ mice was compared to that of six wild-

type littermates, and to the well-established timetable of

neural crest-derived cell migration through the embryonic

mouse gut (see Young and Newgreen, 2001). There was no

obvious difference in the distribution of Phox2b+ cells along

the gastrointestinal tract between RetTGM/+ and wild-type

embryos, and the most caudal Phox2b+ cells were approx-

imately midway along the post-caecal hindgut in both

genotypes (data not shown). In addition, the location of

GFP+ cells along the gut of E10.5–E13.5 RetTGM/+ embryos

was similar to that reported previously using other neural

crest cell markers (Kapur et al., 1992; Young et al., 1998). In

summary, at E10.5, the most caudal GFP+ cells in RetTGM/+

mice were colonizing the midgut, at E11.5 they were

colonizing the caecum and post-caecal hindgut, at E12.5

they were midway along the post-caecal hindgut, and at

E13.5, they were in the caudal hindgut, which is the same as

in wild-type mice.

These observations show that although expression varies

between cells, TGM is expressed by all enteric neural crest-

derived cells in E10.5–E13.5 RetTGM/+ mice. Furthermore,

the timetable by which neural crest-derived cells colonize

the gut of RetTGM/+ mice is indistinguishable from that in

wild-type mice.

Time-lapse videomicroscopy of neural crest-derived cell

migration within the gut

To characterize the migratory behavior of neural crest-

derived cells within the gut, time-lapse video microscopy

was performed. GFP+ cells in explants of gut from E10.5–

E13.5 RetTGM/+ mice were imaged for up to 17 h (Figs. 2, 3;

movies 1–4). Long-term (minimum of 4 h) GFP+ cell

migration behavior was examined in preparations of mid-

and hindgut from E10.5 (n = 3), E11 (n = 2), E11.5 (n = 5)

and E12.5 (n = 20) mice. As only one preparation per litter

could be imaged at a time, the preparations that were not

imaged were left in tissue culture medium in an incubator.

At the end of imaging sessions that lasted > 8 h, the location

of the GFP+ cells in the imaged preparation was compared

to that of preparations left in the incubator. The locations of

GFP+ cells in the gut preparations that underwent time-lapse

analysis were always similar to that in gut preparations left

in the incubator, indicating that the conditions to which the

cells were exposed specifically during time-lapse microsco-

py did not noticeably affect the speed of neural crest-derived

cell migration. Importantly, we did not observe any GFP+

cells undergoing cell death in the imaged preparations, nor

did we observe signs that any of the imaged GFP+ cells were

unhealthy (e.g., blebbing). Furthermore, the behavior of

GFP+ cells imaged for 12 h showed no detectable differ-

ences from the behavior of cells in the first hour of imaging.

It is likely that the lack of any detectable phototoxic effects

is due to the low levels of GFP expressed by the crest-

derived cells in the Ret-TGM mice. In contrast, we have

observed blebbing and cell death when imaging migrating

crest-derived cells exhibiting a high-intensity GFP signal

(from a different transgenic mouse—data not shown).

Based on the rate at which the embryonic gut increases in

length, and the appearance of neuron precursors rostrocau-

dally in fixed samples of gut, it has been estimated that the

neural crest-derived cell wavefront moves down the embry-

onic chick gut in vivo at around 40 Am/h (Allan and

Newgreen, 1980). We determined the net speed at which

the wavefront of GFP+ cells migrated along explants of

E10.5 midgut (n = 3) and E12.5 hindgut (n = 4) by

measuring the distance between the location of the most

caudal GFP+ cell visible at one time point and the location

of the most caudal GFP+ cell a minimum of 8 h later. The

wavefront of GFP+ cells within the E12.5 hindgut and E10.5

midgut moved caudally at a net speed of around 35 Am/

h (Fig. 4). At E11.5, GFP+ cells were colonizing the

caecum, but we were unable to follow accurately the most

caudal cell because of the wide range of focal planes in

which cells were found within the caecum.

At E12.5, the migratory wavefront of GFP+ cells was in

the middle part of the hindgut. When explants of E12.5 gut

were removed and grown in organ culture, the wavefront of

GFP+ crest-derived cells reached the anal end of the hindgut

H.M. Young et al. / Developmental Biology 270 (2004) 455–473 461

after approximately 24–30 h (i.e., equivalent to E13.5).

However, in vivo, the migratory wavefront of crest-derived

cells does not reach the anal end until E14.5 (Young et al.,

1998). As the speed of crest cell migration in vitro is similar

to that reported in vivo (see above), the earlier colonization

Fig. 5. Polar histograms representing the trajectories of the most caudal cell at 10-

E10.5 midgut (D). The trajectories were determined by drawing a straight line from

0j being the rostrocaudal axis of the gut. Explants A and D are illustrated in Figs.

moved rostrally for a short period of time. The numbers shown on the inner arcs ar

at the Developmental Biology website.

of cultured segments of hindgut is almost certainly due to

the reduced rate of growth of the gut that occurs in vitro

compared to in vivo (Hearn et al., 1999); that is, in vitro, the

crest-derived cells do not have to migrate as far to reach the

anal end of the gut segments.

min intervals in three explants of E12.5 hindgut (A–C) and one explant of

the position of the most caudal cell with its position 10 min previously, with

2 and 3, respectively. Note that in explant B (movie 2), the most caudal cell

e frequencies (percentages) of cells. Movies of these explants can be viewed

ental Biology 270 (2004) 455–473

Dynamic behavior of migrating neural crest-derived cells

within the gut

The rate of migration of GFP+ cells, and the trajectories

they followed, were highly variable (Figs. 2, 3; movies 1–

4). The trajectory of the leading (i.e., most caudal) cell was

determined at 10-min intervals by constructing a straight

line from its position with its position 10 min previously,

with horizontal being the rostrocaudal axis of the gut

(E10.5, n = 1; E12.5, n = 3). Thus, cells moving straight

down the gut would have a trajectory of 0j. Polar histo-

grams of the angle of the vectors show that the trajectories

of the leading (most caudal) GFP+ cells were variable in

detail. The proportion of the vectors that pointed directly

down the gut varied between preparations, but there was

also a strong, but not absolute, bias against backward

(rostrally directed) movement (Fig. 5).

Static images of fixed preparations suggest that enteric

neural crest-derived cells usually occur in chains (Conner et

al., 2003; Pomeranz et al., 1991; Young et al., 1998).

Imaging of living cells showed that both at, and behind,

H.M. Young et al. / Developm462

Fig. 6. (A–D) Single time point images at 100V intervals of GFP+ cells colonizing(outlined in AV, BV, CV, DV and superimposed on the images taken at later times) o

caudally. Some parts of the original networks become slightly distorted or go out

cells. Although most of the original networks remain intact, there do appear to b

the migratory wavefront, the migrating GFP+ cells were in

close contact with neighboring GFP+ cells most of the time,

and formed chains or networks. Fixed preparations of gut

processed for immunohistochemistry using antisera to

Phox2b to label crest cells and PECAM-1 (CD31) to label

vascular endothelial cells showed that the chains of Phox2b+

crest cells were not associated with the network of blood

vessels in the gut, which were deeper (closer to the lumen)

than the crest cells (data not shown).

At the migratory wavefront, individual cells could gen-

erally not be resolved because the nuclei and cytoplasm of

individual cells showed approximately equal GFP intensity

and there was little variation between cells in the intensity of

GFP. We therefore could not determine whether the leading

cell within a chain changed identity, or whether individual

cells retained their relative position within individual chains.

However, different chains of cells moved at variable speeds,

and the leading (most caudal) chain of cells often changed.

A single chain of cells could be considerably in advance of

(caudal to) the rest of the population. In particular, in late

the E12.5 hindgut. The networks laid down by GFP+ cells at the wavefront

ften remained intact for many hours after the wavefront had moved further

of focus. Gaps in the original networks can be later filled in by other GFP+

e some regions that are not retained. Scale bar: 100 Am.

H.M. Young et al. / Developmental Biology 270 (2004) 455–473 463

E11.5 mice, there was often a long chain of cells that

extended from the caecum into the post-caecal hindgut

along the mesenteric border, which preceded the entry of

other GFP+ cells into the post-caecal hindgut. The chains of

GFP+ cells that colonized the hindgut at E12.5 and E13.5

tended to be longer and broader than the chains of cells that

colonized the midgut at E10.5.

Chains of GFP+ cells at, or close to, the migratory wave-

front showed complicated and unpredictable patterns of

migration, with changes in migratory speed and trajectories.

A chain of GFP+ cells that was migrating in one direction

could suddenly slow or stop, change direction, branch or

retract. However, cells that retracted usually did so along

their previous trajectory of advance. Moreover, cells imme-

diately behind the leading chains often appeared to follow

the same pathway through the mesenchyme as the cells that

preceded them. The network laid down by chains of cells at

the wavefront often remained intact for many hours after the

wavefront had moved further caudally (Fig. 6). The con-

stancy in the form of the network may be because (i) some of

the neural crest-derived cells that are transiently at the

wavefront (pioneers) become nonmigratory and they and

their processes form a scaffold along which more rostral cells

(followers) migrate, or (ii) there are preexisting cues in the

gut environment which guide the detailed migration of all

Fig. 7. Selected images of GFP+ cells in the E12.5 hindgut, just behind the migra

Over a period of 3.5 h, this cell-free region was ‘‘filled in’’ both by cells from behi

bar: 25 Am.

cells, both pioneer and follower cells, or (iii) pioneer cells

leave a ‘‘molecular trail’’ or otherwise alter the environment

to favor the migration of following cells.

Behind the migratory wavefront, there were some differ-

ences between cells in the intensity of GFP. Differences

between cells in the levels of expression of Ret are also seen

in fixed preparations of gut using antisera to Ret (e.g.,

Schiltz et al., 1999; Young et al., 2003) and in gut from

RetTGM/+ embryos that are fixed immediately (Figs. 1B, H).

The differences between cells in the intensity of GFP

enabled us to visualize, within the networks of GFP+ cells,

a subpopulation of cells that was streaming caudally along a

scaffold of nonmigratory cells and cell processes (see movie

5). Most migrating cells within a chain of cells appeared to

follow the same pathway along the scaffold as the cell in

front. This suggests that at least some of the consistency in

the shape of the network is due to a scaffold formed by

nonmigratory GFP+ cells and cell processes. The relation-

ship between this early network and the interconnected

ganglia seen in mature animals in unclear, as distinct

myenteric ganglia do not form in the mouse hindgut until

late embryonic stages (E16.5).

Caudally migrating GFP+ cells were commonly observed

in regions of the gut that had been populated by crest-

derived cells for up to 8–12 h, but were less frequently

tory wavefront. (A) A small region (dotted) had remained cell-free. (B–D)

nd moving caudally and cells from in front moving rostrally (arrows). Scale

Table 3

Behavior of selected individual GFP+ cells (cells not part of a chain)

Cell # Approximate

length of time not

in contact with

other cells (min)

Outcome

1 30–40 Divided, one daughter cell rejoined

original chain, the other went in the

opposite direction around the gut

and out of focus

2 30–40 Rejoined original chain

3 120 Stopped migrating after approximately

80 min and was joined by another

chain of cells

4 60 Left one chain and joined another

5 30–40 Rejoined original chain

H.M. Young et al. / Developmental Biology 270 (2004) 455–473464

observed in regions that had been colonized for over 12 h.

In regions well behind (rostral to) the wavefront that had

been colonized by neural crest-derived cells for over 24

h (for example, the midgut of E12.5 mice when the

migratory wavefront is well into the hindgut), GFP+ cells

formed web-like networks with cell-free spaces. Within the

networks, individual GFP+ cells were sometimes observed

Fig. 8. The most caudal cell visible in this preparation (arrow) of E12.5 hindgut ro

opposite directions circumferentially around the gut. Scale bar: 25 Am.

to migrate, across cell-free spaces, from one cluster of GFP+

cells to another. However, we did not observe streams of

cells migrating caudally through the network. Hence, we did

not find any evidence for significant numbers of late,

caudally migrating cells. In addition, the cell-free spaces

were often in constant locations for many hours.

Although the population of GFP+ cells moved caudally,

some chains of cells, or what appeared to be individual cells,

moved rostrally transiently. Rostrally directed movement

was observed by cells both at (see Fig. 5B and movie 2), and

behind, the wavefront (Fig. 7).

In fixed preparations, some crest-derived cells within the

gut do not appear to be in contact with other crest-derived

cells (Young et al., 1998). In the current time-lapse study,

individual GFP+ cells were sometimes observed to break off

from a chain and lose all visible contact with other GFP+

cells. These cells migrated at very variable speeds of up to

90 Am/h when calculated over a 30-min timebase. (Note: the

rates of migration of individual cells were determined by

measuring the total path length of individual cells, unlike the

calculations of net speed of the migratory wavefront, which

only measured the total forward movement.) The behavior

of selected individual cells is shown in Table 3. All

unded up (B) and divided (C), and then the daughter cells (arrows) went in

H.M. Young et al. / Developmental Biology 270 (2004) 455–473 465

individual cells we observed (a) later rejoined the same

chain of GFP+ cells, (b) joined another chain, (c) migrated

for a short time and then ceased migrating and were later

contacted by other GFP+ cells, or (d) rounded up and

divided (see below) and then re-contacted other GFP+ cells.

Thus, GFP+ neural crest-derived cells only appeared to lose

contact with other GFP+ cells transiently.

Cells undergoing mitosis were easy to distinguish (see

cells arrowed in movie 5). Just before undergoing cell

division, most cells at the migratory wavefront would first

appear to break contact with neighboring cells, then round

up and divide (Fig. 8). In some cases of cell division we

observed at the wavefront (5/9), immediately after dividing,

the daughter cells would elongate and then migrate in

Fig. 9. (A) Diagram showing that removal of a rim of mesenchyme just rostral to t

rest of the population. (B–H) Selected images of a preparation in which a rim of

caudal to the region in which the mesenchyme was removed (dotted). One cell (a

position. Another cell (open arrow) migrates for a short period of time, rounds u

another cell. Scale bar: 50 Am.

opposite directions. However, in three other instances, the

two daughter cells appeared to remain in contact, and in

another case, one of the daughter cells migrated out of the

plane of focus, so its direction of migration could not be

ascertained. Behind the wavefront, most cells undergoing

mitosis remained within chains, but the behavior of their

daughter cells could not be followed (see movie 5).

At E11.5–E13.5, GFP+ cells were observed in the

mesentery and tissue adjacent to the distal hindgut (data

not shown). These GFP+ cells presumably arise from the

sacral neural crest (Burns and Le Douarin, 1998; Kapur,

2000). We did not observe any migration of sacral GFP+

cells into the hindgut before the colonization of the distal

hindgut by vagal GFP+ cells.

he most caudal cells will isolate a small group of neural crest cells from the

mesenchyme had been removed. There were only a small number of cells

rrow) extends processes in many different directions, but does not alter its

p and is stationary for over 2 h, and then migrates to the same location as

ental Biology 270 (2004) 455–473

Effects of reducing neural crest-derived cell number on the

dynamic behavior of neural crest-derived cells

The above observations indicate a high degree of inter-

action between neural crest-derived cells during migration.

Previous studies have shown that following a reduction in

the starting population of vagal neural crest cells, they do

not migrate as far distally along the gut (Burns, 2000;

Peters-van der Sanden et al., 1993; Yntema and Hammond,

1954). To test the hypothesis that cell number influences the

migratory behavior of enteric crest-derived cells, we wished

to isolate a small group of the most caudal GFP+ cells. In

preliminary experiments in which the gut was completely

severed just rostral to a small group of the most caudal

H.M. Young et al. / Developm466

Fig. 10. Confocal images of wholemount preparations of E12.5 (A–D) and E11.

strong Hu immunostaining in both the nucleus and cytoplasm and also neurofila

cytoplasmic Hu staining and no neurofilament staining. (E–G) The subpopulati

(arrows), but the cells that show strong Hu staining in both the nucleus and cytopl

bar: 10 Am.

GFP+ cells, it was not possible to image the cells because it

was impossible to prevent the unattached (severed) end

from moving. Therefore, a rim of mesenchyme just rostral

to the most caudal neural crest-derived cells was removed,

leaving the epithelium intact, in the E12.5 or E13.5 hindgut

(Fig. 9). In some of the preparations (n = 4), the behavior of

the isolated cells was examined by time-lapse microscopy

for a minimum of 8 h. In other preparations (n = 3), the

region caudal to the lesion was photographed immediately

after the lesion was performed, the preparations were then

left in an incubator for 8 h and then re-photographed. The

distance between the location of the most caudal GFP+ cell

visible immediately after the lesion and the location of the

5 (E–G) midgut. (A–D) A subpopulation of GFP+ cells (asterisks) shows

ment staining. Another subpopulation of GFP+ cells (arrows) shows weak

on of cells that shows weak Hu staining only in the cytoplasm is Sox10+

asm are Sox10� (asterisk). Some Sox10+ cells are Hu� (open arrow). Scale

Table 4

The phenotype of migrating neural crest cells

Combination of

antisera used

Number of GFP+

migrating cells

examined

Phenotype

Hu and neurofilament

M (NF)

27 93% weak cytoplasmic

Hu+/NF�

3.5% weak cytoplasmic

Hu+/?NF

3.5% nuclear Hu+/NF�

Hu and PGP9.5 8 88% weak cytoplasmic

Hu+/weak PGP9.5+

12% weak cytoplasmic

Hu+/PGP9.5�

Neurofilament M

(NF) and PGP9.5

2 100% NF�/weak PGP9.5+

PGP9.5 and Sox10 18 94% Sox10+/weak PGP9.5+

6% Sox10+/?PGP9.5

Short time-lapse sequences (2–10 images taken every 2.5–5 min) of

migrating GFP+ cells were taken, the gut was then fixed immediately and

processed for immunohistochemistry using four different combinations of

antisera (left hand column). The regions containing the GFP+ cells that had

migrated in the 5-min period immediately before fixation were relocated

and examined using confocal microscopy. ‘‘?’’ indicates that we were

unable to decide if the level of immunostaining was above background.

H.M. Young et al. / Developmental Biology 270 (2004) 455–473 467

most caudal GFP+ cell a minimum of 8 h later was measured

(using the method described above for determining the

speed of the migratory wavefront). The net migratory speeds

of the populations of isolated cells were significantly lower

than that of control populations (Fig. 4; t test: two-sample

assuming unequal variances, P < 0.05). The number of

GFP+ cells on the top surface of the explants, caudal to the

lesion, was estimated. The total size of the isolated popu-

lation of GFP+ cells in each explant could not be determined

because the GFP+ cells on the sides and bottom of the each

preparation could not be visualized, and thus we could not

establish the number of cells required for normal migratory

speed. However, in general, the greater the number of GFP+

cells on the top surface, the faster the migration rate (Fig. 4);

in one preparation, the migration rate was not very different

from control preparations.

Time-lapse analysis revealed that in preparations with

only two or three GFP+ cells on the top surface caudal to the

lesion, most GFP+ cells caudal to the lesion failed to migrate

at all, although some stationary cells did send out processes

in a variety of directions (Fig. 9). Other cells on the caudal

side of the lesion migrated intermittently (Fig. 9). In

preparations that had larger numbers of cells present caudal

to the lesion, the population of isolated cells migrated

caudally at a slower rate than control populations but no

difference in the behavior of cells was observed—the cells

showed a large degree of cell–cell contact and formed

chains that displayed complicated patterns of migration.

Do migrating neural crest-derived cells show a neuronal

phenotype?

Previous studies have hypothesized that the migratory

ability of enteric neural crest-derived cells is reduced once

the cells differentiate into neurons (Hearn et al., 1998; Wu et

al., 1999). Cells with a neuronal phenotype are present

among crest-derived cells at the migratory wavefront in

the embryonic gut (Young et al., 1999). We therefore

examined the phenotype of migrating GFP+ neural crest-

derived cells. Neural crest cells undergo many changes in

phenotype between their emigration from the neural tube

until their final differentiated state. To define cells at

different states of differentiation, antisera to the transcription

factor, Sox10, and to three different pan-neuronal markers,

PGP9.5, neurofilament (NF) and Hu, were used. Sox10 is

expressed by neural crest-derived cells before emigration

from the neural tube and is rapidly down-regulated in

differentiating neurons (Lo et al., 2002; Young et al.,

2002, 2003). PGP9.5 and NF antisera have been used

previously to label neurons in the embryonic mouse and

rat gut (Chalazonitis et al., 1998, 2001; Sidebotham et al.,

2001; Young et al., 2003). Hu antiserum, which labels all

differentiated neurons in the adult mouse gut (Stewart et al.,

2003), has been used to label neurons in the embryonic quail

and chick gut (Burns and Le Douarin, 1998; Conner et al.,

2003). We first compared the presence of Hu immunostain-

ing with other neural crest-derived cell and neuronal

markers in the E11.5 and E12.5 mid- and hindgut. Two

populations of Hu+ cells were observed: one population

showed strong Hu immunostaining in both the nucleus and

cytoplasm (termed ‘‘nuclear Hu+’’), and a second population

showed weak immunostaining that was present only in the

cytoplasm (‘‘weak cytoplasmic Hu+’’). Nuclear Hu+ cells

were Sox10-negative, but all weak cytoplasmic Hu+ cells

were Sox10+ (Figs. 10E–G). There were also Sox10+ cells

that did not show any Hu staining (Figs. 10E–G). At the

migratory wavefront, most of the cells showed weak cyto-

plasmic Hu+. Only nuclear Hu+ cells showed NF immunos-

taining (Figs. 10A–D). All nuclear Hu+ cells were strongly

PGP9.5+, and many of the weak cytoplasmic Hu+ cells

including those at the migratory wavefront were also weakly

PGP9.5+ (not shown; Sidebotham et al., 2001).

In summary, these markers seem to distinguish between

two populations of cells; one is Sox10�/nuclear Hu+/NF+/

strong PGP9.5+ and can be defined as neurons, and the other

is Sox10+/weak cytoplasmic Hu+/NF�/weak PGP9.5+ and

represent cells that have not commenced differentiation into

neurons or glial cells. Glial cell and glial precursor markers

were not used for this part of the study because cells

showing mature glial markers such as S100b cannot be

detected before E14.5, and although glial precursors (B-

FABP+ cells) are present in the foregut of E12.5 mice, they

cannot be detected in the caudal midgut or hindgut (Young

et al., 2003).

To determine the phenotype of migrating neural crest-

derived cells, short time-lapse sequences (2–10 images

taken every 2.5–5 min) of GFP+ cells were taken. The

gut was then fixed immediately, processed for immunohis-

tochemistry, and the region containing the migrating GFP+

cells re-located and examined using confocal microscopy.

Fig. 11. (A) Image of a GFP+ cell (arrow and outlined in pink dots) that had migrated from a previous location (outlined in white dots) over a period of 10 min.

The tissue was fixed immediately after the image was taken and processed for immunohistochemistry. The GFP+ cell was then re-located and examined using

confocal microscopy (B–D). The migrating GFP+ cell (arrow) was weakly PGP9.5+ (C) and showed Hu staining in the cytoplasm, but not the nucleus (D). The

migrating cell was approximately 150 Am from the migratory wavefront. Scale bar: 10 Am.

H.M. Young et al. / Developmental Biology 270 (2004) 455–473468

As it was difficult to identify individual cells within chains,

only cells that were at the leading edge of a chain, or that

formed a branch from a chain, were chosen for analysis so

that they could be unequivocally identified. Most of the

fields of view analyzed were within 250 Am of the wave-

front. The phenotypes of 55 GFP+ cells from E12.5 hindgut

Fig. 12. (A, B) Selected images of a time-lapse series showing a GFP+ cell (asteri

moved circumferentially from its original position (dotted outline). The tissue was

immunohistochemistry. The GFP+ cell (arrow) was then re-located and examined

(D) and Sox10+, and was about 100 Am from the migratory wavefront. Scale bar

preparations that had migrated in the 5 min immediately

before fixation are shown in Table 4. Most of these

migrating cells were Sox10+, weakly PGP9.5+ and also

showed weak cytoplasmic Hu+ (Figs. 11, 12; movie 6).

Sox10 is expressed by undifferentiated neural crest-derived

cells and then subsequently by glial precursors. The migrat-

sk) that migrated over a period of 12.5 min. (B) The GFP+ cell (arrow) had

fixed immediately after the image shown in B was taken and processed for

using confocal microscopy (C–F). The migrating cell was weakly PGP9.5+

: 10 Am. See movie 6 the Developmental Biology website.

H.M. Young et al. / Developmental Biology 270 (2004) 455–473 469

ing Sox10+ cells observed are unlikely to be glial precursors

because glial precursors are not present in the caudal midgut

or hindgut of E12.5 mice (the preparations used for this

series of experiments) (Young et al., 2003). A total of 130

GFP+ cells in the wavefront fields of view were scored for

Hu labeling and 16% had a Hu+ nucleus (i.e., a neuronal

phenotype). In contrast, only 3% (1/35) of migrating GFP+

cells had a Hu+ nucleus, and none showed neurofilament

immunostaining (n = 29). Taken together, less than 2% of all

migrating cells examined (n = 55) expressed any of the

neuronal markers used. Some cells extended processes

during the short, time-lapse sequences, and some of these

processes showed neurofilament immunostaining. In the

fields of view examined, we did not observe any cell

showing Sox10, PGP9.5, Hu or NF staining that were not

GFP+, confirming that GFP appears to be expressed by all

crest-derived cells at the embryonic stages examined.

Discussion

Most neural crest-derived cells that enter the developing

gut undergo an extensive migration before forming the

enteric nervous system. To examine the dynamic behavior

of crest-derived cells as they migrate through the gut mesen-

chyme, we used time-lapse microscopy and Ret-TGM mice,

in which enteric crest-derived cells express GFP (Enomoto et

al., 2001). We showed that the developing gut is a technically

tractable system to examine long-range cell migration in a

real tissue milieu. Most previous time-lapse studies of neural

crest cells used DiI applied to premigratory crest cells, and

thus only a subpopulation of cells was labeled. In contrast,

GFP appears to be expressed by every enteric crest-derived

cell of E10.5–E13.5 Ret-TGMmice, and hence we were able

to observe the entire population.

Our data have revealed some novel and hitherto unex-

pected behavior—for example, the network laid down by

leading cells is maintained for many hours, in part because a

subpopulation of the cells forms a scaffold. Paradoxically,

although the migration is predictable in population terms,

we showed that the behavior is unpredictable at the single-

cell level, at least at current levels of analysis. Previous

studies have shown that vagal crest cell number has an

effect on migratory behavior, but our study has revealed that

the likely underlying cellular mechanism is that cell–cell

contact is required for migration. Although many inferences

have been made about the behavior of migrating crest-

derived cells in the gut from static studies, it is extremely

important that these inferences are tested by examining the

dynamic behavior of the cells (Lichtman and Fraser, 2001).

Our data have demonstrated directly some behaviors that

had previously only been inferred, such as a high degree of

cell–cell interactions and migration of cells in chains.

Finally, it has been generally assumed that there is an

inverse correlation between neuronal differentiation and

migratory ability, yet to our knowledge, the phenotype of

migrating cells has never been directly examined anywhere

in the nervous system. We show that only an extremely

small proportion of the migrating crest-derived cells

expressed neuronal markers.

Migratory behavior of enteric crest-derived cells

Like crest cells in other locations (Krull et al., 1995;

Kulesa and Fraser, 1998, 2000; Kulesa et al., 2000), migrat-

ing enteric crest-derived cells formed chains that displayed

episodic and complicated patterns of movement, with un-

predictable changes in speed and trajectory. However, the

network laid down by the leading migratory cells remained

constant for many hours. This appears to be because some of

the leading cells and their processes form a scaffold along

which the following cells migrated. This phenomenon shows

some similarities to peripheral nerve formation where mi-

grating crest-derived cells follow axons (Gilmour et al.,

2002) and axon tract formation, where follower axons follow

a scaffold laid down by pioneer neurons (Bate, 1976; Easter

et al., 1994). We also observed that when chains of cells

retracted, they usually did so along the same pathway

followed during their advance. The molecular mechanisms

underlying the detailed pathway choice made by the migrat-

ing cells remain to be determined. Important influences on

pathway choice could include (i) interactions between crest-

derived cells (both migratory and postmigratory cells), (ii)

preexisting local variations in the stationary environment of

mesenchymal cells or the extracellular matrix (Gershon et

al., 1993; Howard and Gershon, 1998; Newgreen and

Hartley, 1995), and/or (iii) local environmental variations

induced by pioneering crest-derived cells.

Cell–cell contact

Like other populations of neural crest cells, and some

populations of migrating neurons in the CNS (Lois et al.,

1996), enteric crest-derived cells migrated in chains. The

chains were sometimes broken, but only transiently. The

propensity to form chains strongly suggests the presence of

cell–cell adhesive interactions mediated by cell adhesion

molecules. A variety of cell adhesion molecules are

expressed by enteric crest-derived cells (Chalazonitis et

al., 1997; Iwashita et al., 2003). Furthermore, Ret, which

appears to be expressed by all crest-derived cells in the gut,

contains cadherin-related sequences in the extracellular

domain (see Airaksinen and Saarma, 2002). However, the

identity of the molecules, and the way that they mediate the

formation of chains of crest-derived cells rather than other

forms of aggregates, has yet to been elucidated.

Cell number

Our time-lapse observations of enteric crest-derived cells

revealed high cell density and almost continuous cell–cell

H.M. Young et al. / Developmental Biology 270 (2004) 455–473470

interactions. Previous studies have shown that partial abla-

tion of the vagal crest cell population in vivo reduces the

distance that the cells migrate along the gut (Burns, 2000;

Peters-van der Sanden et al., 1993; Yntema and Hammond,

1954), and that high cell density promotes migration of

trunk neural crest cells in vitro (Newgreen et al., 1979;

Rovasio et al., 1983; Thomas and Yamada, 1992). This has

lead to the idea that cell density or ‘‘population pressure’’

may, by increasing cell–cell contact, stimulate cell migra-

tion in a direction other than the contact direction (New-

green, 1990). Alternatively, or additionally, crest-derived

cells could alter their local microenvironment, for example,

by growth/chemotactic factor depletion, thereby promoting

cell movement into uncolonized regions (see below). We

found that when a small population of the most caudal crest-

derived cells was isolated from the population behind

(rostral to) them, the cells migrated slower. Furthermore,

some of the isolated cells that were not in contact with any

other crest-derived cells were immobile, although they did

extend projections. Our observations suggest that cell–cell

contact is required for ongoing motility of enteric neural

crest-derived cells, and provide at least a partial explanation

for unpredictable trajectories that individual cells follow.

To produce small, isolated populations of crest-derived

cells, we removed a rim of mesenchyme. We cannot rule out

the possibility that the removal of mesenchyme affected the

migration of neural crest cells, for example, by release of

substances from lesion site. However, any perturbation that

reduces the number of neural crest-derived cells could have

indirect influences on the crest-derived cells. It is extremely

unlikely that the retarded crest cell migration we observed

following removal of a rim of mesenchyme was due to

isolation of the cells from a rostral source of chemorepellent

because in both avian (Burns et al., 2002) and mouse (HMY,

AJB, RBA unpublished data) gut, vagal neural crest cells

will migrate rostrally if introduced to the distal hindgut.

It would be very interesting to determine how large a

population is required for normal migration rate. In the

current study, the number of GFP+ cells on the top surface of

the isolated segments of gut was estimated. In general, the

greater the number of GFP+ cells on the top surface, the

faster the migration rate. However, because crest-derived

cells are likely to be present around the entire circumference

of the explants, and because it was not possible to count

GFP+ cells on the sides and underneath surface of the

explant, the exact relationship between cell number and

migration rate could not be determined. The idea that enteric

crest cell migratory behavior is influenced by cell number is

supported by (a) studies showing that partial ablation of the

vagal premigratory crest cell population in avian embryos

results in aganglionosis of the distal gut (Burns, 2000;

Peters-van der Sanden et al., 1993; Yntema and Hammond,

1954), and (b) the phenotype of Et-3�/� mice, which have

reduced numbers of crest cells in the gut from E10.5 and

aganglionosis (Barlow et al., 2003). However, the relation-

ship between crest cell number and migratory ability seems

to be complex as gdnf+/� mice have reduced numbers of

crest cells within the gut, but do not exhibit aganglionosis

(Shen et al., 2002).

It is possible that the combined outcome of a cell density

effect (cell–cell contact stimulating motility away from

contact) and of cell–cell adhesion is that crest-derived cells

migrate in chains. However, trunk neural crest cells in vitro

advance as a broad front with no observable chains, but with

numerous cell–cell contacts (Newgreen et al., 1979), so

chain formation is not an inevitable consequence of a

combination of cell–cell adhesion and migration.

The molecular signals underlying these cellular interac-

tions (cell–cell adhesion and motility induced by cell–cell

contact) are unknown. It is possible that the same cell

surface molecule mediates both adhesion and motility. For

example, while the extracellular domain of cadherins medi-

ates cell–cell adhesion, the cytoplasmic tail can interact

with the actin cytoskeleton via intracellular attachment

proteins, such as catenins.

Chemotactic molecules

GDNF is expressed by the gut mesenchyme and is

chemoattractive to enteric crest-derived cells (Iwashita et

al., 2003; Natarajan et al., 2002; Young et al., 2001). Thus,

the concentration of GDNF to which each crest-derived cell

is exposed could influence their migratory behavior. We

have previously proposed that, even if the differences in

levels of GDNF rostrocaudally along the gut are not high,

crest-derived cells may still migrate towards regions lacking

other crest-derived cells because of the presence of available

(unbound) GDNF in the mesenchyme (Young et al., 2001).

The current study suggests that the local concentration

differences of unbound GDNF are not the only influence

on migratory behavior, as cells would be expected to

separate from each other and ‘‘tile’’ the mesenchyme evenly.

However, the response to GDNF will be occurring concur-

rently with other mechanisms influencing migratory behav-

ior such as endothelin-3 concentration (Barlow et al., 2003;

Kruger et al., 2003), cell–cell contact and cell density. It is

unknown to what extent the chemoattractive effect of

GDNF is dependent on cell density, and whether a steep

gradient of GDNF can counteract the effect of low cell

numbers on migration.

Other factors affecting migratory behavior

Mice lacking endothelin-3 (Et-3�/� mice) have aganglio-

nosis of the distal hindgut. Based on the location of crest-

derived cells in fixed preparations of wild-type and Et-3�/�

mice (Coventry et al., 1994), and the high levels of Et-3

expression in the caecum of wild-type mice (Leibl et al.,

1999), it has been proposed that migration of crest cells

through the caecum is important for the colonization of the

post-caecal hindgut (Barlow et al., 2003; Coventry et al.,

1994; Kruger et al., 2003). Since the caecal wall is thick and

H.M. Young et al. / Developmental Biology 270 (2004) 455–473 471

the crest-derived cells are in many different focal planes

within the caecal wall, we were unable to examine the

migratory behavior of the crest cells as they migrated

through the caecum in the current study. Future studies

using confocal microscopy should reveal whether any

aspect of the migratory behavior changes as crest cells enter

and migrate through the caecum.

Phenotype of migrating cells

Vagal crest-derived cells that enter the foregut migrate

caudally. However, at variable distances along the gut,

individual cells must then cease migrating so that enteric

neuron precursors are distributed relatively evenly along the

entire gastrointestinal tract. The ‘‘stop’’ signals influencing

crest-derived cell migration have not been identified. Since

in vitro studies have shown that endothelin-3 inhibits the

differentiation of enteric crest-derived cells into neurons

(Hearn et al., 1998; Wu et al., 1999), and since mice lacking

endothelin-3 or its receptor, endothelin receptor B, lack

enteric neurons in the distal part of the intestine (Baynash

et al., 1994; Hosoda et al., 1994), it has been hypothesized

that in the absence of endothelin-3 signaling, crest-derived

cell differentiate into neurons before colonizing the distal

part of the intestine (Hearn et al., 1998; Wu et al., 1999).

This hypothesis assumes that the differentiation of cells into

neurons inhibits their migratory ability. This assumption has

not previously been tested.

In the current study, we examined the phenotype of

migrating crest-derived cells and found that less than 2% of

migrating cells examined expressed any of the neuronal

markers used, although cells with neuronal phenotypes were

present nearby. It is possible that neurons may migrate at

significantly slower rates than other cells, and had not

migrated detectable distances in the 5- to 10-min time period

used for this study. Also, to make re-localization of individual

cells possible, we only examined migrating cells that were at

the leading edge of a chain. We therefore cannot rule out the

possibility that neurons do migrate, but only within chains.

In regions of the embryonic gut that have been colonized

by crest-derived cells, not all of the cells show a neuronal

phenotype. Furthermore, undifferentiated crest-derived cells

are present in the mouse gut even after birth (Kruger et al.,

2002; Young et al., 2003). Thus, although our data support

the idea that neuronal differentiation may be a ‘‘stop’’ signal

for some cells, other mechanisms must also operate to halt

the migration of a subpopulation of undifferentiated crest-

derived cells throughout the gut.

Conclusions

This study presents the first direct examination of crest-

derived cell migration in the gastrointestinal tract. The

behavior of crest-derived cells described in the current study

provide essential information, which was previously un-

known or unproven, about the dynamic behavior of normal

enteric crest-derived cells, and will provide an important

baseline for proposed future studies in which the effects of

perturbing endothelin-3 or GDNF levels (or any other mol-

ecule of interest) on migratory behavior of cells can be

examined. The results emphasize the importance of cell–cell

interactions in the migratory process, which affect rate,

directionality and persistence of migration. As cell interac-

tions imply a population, they remind us that in some circum-

stances, cell behavior cannot be simply divided into ‘‘cell

autonomous’’ and ‘‘non-cell autonomous’’, but a concept of

‘‘cell population autonomous’’ should be considered.

Acknowledgments

This work was supported by the National Health and

Medical Research Council of Australia (project grant

145628 to HMY and PMW, project grant 237144 to DFN,

Senior Research Fellowship 170224 to HMY and C.J.

Martin Fellowship 007144 to RBA) and the ANZ Trust. We

thank Dr. Heather Wheat for assistance with plotting the

polar histograms, Dr. Colin Anderson and Professor John

Furness for sharing equipment, and Drs. Jean-Franc�oisBrunet, Miles Epstein, Mark Murphy and Thomas Muller

for supplying antisera.

References

Airaksinen, M.S., Saarma, M., 2002. The gdnf family: signalling, biolog-

ical functions and therapeutic value. Nat. Rev., Neurosci. 3, 383–394.

Allan, I.J., Newgreen, D.F., 1980. The origin and differentiation of enteric

neurons of the intestine of the fowl embryo. Am. J. Anat. 157, 137–154.

Barlow, A., de Graaff, E., Pachnis, V., 2003. Enteric nervous system pro-

genitors are coordinately controlled by the G Protein-coupled receptor

EDNRB and the receptor tyrosine kinase RET. Neuron 40, 905–916.

Bate, C.M., 1976. Pioneer neurones in an insect embryo. Nature 260, 54–56.

Baynash, A.G., Hosoda, K., Giaid, A., Richardson, J.A., Emoto, N., Ham-

mer, R.E., Yanagisawa, M., 1994. Interaction of endothelin-3 with

endothelin-B receptor is essential for development of epidermal mela-

nocytes and enteric neurons. Cell 79, 1277–1285.

Burns, A.J., Le Douarin, N.M., 1998. The sacral neural crest contributes

neurons and glia to the post-umbilical gut: spatiotemporal analysis of

the development of the enteric nervous system. Development 125,

4335–4347.

Burns, A.J., Champeval, D., Le Douarin, N.M., 2000. Sacral neural crest

cells colonise aganglionic hindgut in vivo but fail to compensate for

lack of enteric ganglia. Dev. Biol. 219, 30–43.

Burns, A.J., Delalande, J.M., Le Douarin, N.M., 2002. In ovo transplanta-

tion of enteric nervous system precursors from vagal to sacral neural

crest results in extensive hindgut colonisation. Development 129,

2785–2796.

Chalazonitis, A., Tennyson, V.M., Kibbey, M.C., Rothman, T.P., Gershon,

M.D., 1997. The alpha1 subunit of laminin-1 promotes the develop-

ment of neurons by interacting with LBP110 expressed by neural crest-

derived cells immunoselected from the fetal mouse gut. J. Neurobiol.

33, 118–138.

Chalazonitis, A., Rothman, T.P., Chen, J., Gershon, M.D., 1998. Age-de-

pendent differences in the effects of GDNF and NT-3 on the develop-

H.M. Young et al. / Developmental Biology 270 (2004) 455–473472

ment of neurons and glia from neural crest-derived precursors immu-

noselected from the fetal rat gut: expression of GFRalpha-1 in vitro and

in vivo. Dev. Biol. 204, 385–406.

Chalazonitis, A., Pham, T.D., Rothman, T.P., DiStefano, P.S., Bothwell, M.,

Blair-Flynn, J., Tessarollo, L., Gershon, M.D., 2001. Neurotrophin-3 is

required for the survival-differentiation of subsets of developing enteric

neurons. J. Neurosci. 21, 5620–5636.

Conner, P.J., Focke, P.J., Noden, D.M., Epstein, M.L., 2003. Appearance of

neurons and glia with respect to the wavefront during colonization of

the avian gut by neural crest cells. Dev. Dyn. 226, 91–98.

Coventry, S., Yost, C., Palmiter, R.D., Kapur, R.P., 1994. Migration of

ganglion cell precursors in the ileoceca of normal and lethal spotted

embryos, a murine model for Hirschsprung disease. Lab. Invest. 71,

82–93.

Easter Jr., S.S., Burrill, J., Marcus, R.C., Ross, L.S., Taylor, J.S., Wilson,

S.W., 1994. Initial tract formation in the vertebrate brain. Prog. Brain

Res. 102, 79–93.

Enomoto, H., Crawford, P.A., Gorodinsky, A., Heuckeroth, R.O. , Johnson

Jr., E.M., Milbrandt, J. 2001. RET signaling is essential for migration,

axonal growth and axon guidance of developing sympathetic neurons.

Development 128, 3963–3974.

Fu, M., Chi Hang Lui, V., Har Sham, M., Nga Yin Cheung, A., Kwong

Hang Tam, P., 2003. HOXB5 expression is spatially and temporarily

regulated in human embryonic gut during neural crest cell colonization

and differentiation of enteric neuroblasts. Dev. Dyn. 228, 1–10.

Gershon, M.D., 1999. Lessons from genetically engineered animal models

II. Disorders of enteric neuronal development: insights from transgenic

mice. Am. J. Physiol. 277, G262–G267.

Gershon, M.D., Chalazonitis, A., Rothman, T.P., 1993. From neural crest to

bowel: development of the enteric nervous system. J. Neurobiol. 24,

199–214.

Gilmour, D.T., Maischein, H.M., Nusslein-Volhard, C., 2002. Migration

and function of a glial subtype in the vertebrate peripheral nervous

system. Neuron 34, 577–588.

Hearn, C.J., Murphy, M., Newgreen, D., 1998. GDNF and ET-3 differen-

tially modulate the numbers of avian enteric neural crest cells and

enteric neurons in vitro. Dev. Biol. 197, 93–105.

Hearn, C.J., Young, H.M., Ciampoli, D., Lomax, A.E., Newgreen, D.,

1999. Catenary cultures of embryonic gastrointestinal tract support or-

gan morphogenesis, motility, neural crest cell migration, and cell dif-

ferentiation. Dev. Dyn. 214, 239–247.

Hosoda, K., Hammer, R.E., Richardson, J.A., Baynash, A.G., Cheung, J.C.,

Giaid, A., Yanagisawa, M., 1994. Targeted and natural (piebald-lethal)

mutations of endothelin-B receptor gene produce megacolon associated

with spotted coat color in mice. Cell 79, 1267–1276.

Howard, M.J., Gershon, M.D., 1998. Development of LBP110 expression

by neural crest-derived enteric precursors: migration and differentiation

potential in ls/ls mutant mice. J. Neurobiol. 35, 341–354.

Iwashita, T., Kruger, G.M., Pardal, R., Kiel, M.J., Morrison, S.J., 2003.

Hirschsprung disease is linked to defects in neural crest stem cell func-

tion. Science 301, 972–976.

Jiang, Y., Liu, M.T., Gershon, M.D., 2003. Netrins and DCC in the guid-

ance of migrating neural crest-derived cells in the developing bowel and

pancreas. Dev. Biol. 258, 364–384.

Kapur, R.P., 1999. Early death of neural crest cells is responsible for total

enteric aganglionosis in Sox10(Dom)/Sox10(Dom) mouse embryos.

Pediatr. Dev. Pathol. 2, 559–569.

Kapur, R.P., 2000. Colonization of the murine hindgut by sacral crest-

derived neural precursors: experimental support for an evolutionarily

conserved model. Dev. Biol. 227, 146–155.

Kapur, R.P., Yost, C., Palmiter, R.D., 1992. A transgenic model for study-

ing development of the enteric nervous system in normal and agan-

glionic mice. Development 116, 167–175.

Kruger, G., Mosher, J., Bixby, S., Joseph, N., Iwashita, T., Morrison, S.,

2002. Neural crest stem cells persist in the adult gut but undergo

changes in self-renewal, neuronal subtype potential, and factor respon-

siveness. Neuron 35, 657.

Kruger, G.M., Mosher, J.T., Tsai, Y.H., Yeager, K.J., Iwashita, T., Gariepy,

C.E., Morrison, S.J., 2003. Temporally distinct requirements for endo-

thelin receptor B in the generation and migration of gut neural crest

stem cells. Neuron 40, 917–929.

Krull, C.E., Collazo, A., Fraser, S.E., Bronner-Fraser, M., 1995. Segmental

migration of trunk neural crest: time-lapse analysis reveals a role for

PNA-binding molecules. Development 121, 3733–3743.

Kulesa, P.M., Fraser, S.E., 1998. Neural crest cell dynamics revealed by

time-lapse video microscopy of whole embryo chick explant cultures.

Dev. Biol. 204, 327–344.

Kulesa, P.M., Fraser, S.E., 2000. In ovo time-lapse analysis of chick hind-

brain neural crest cell migration shows cell interactions during migra-

tion to the branchial arches. Development 127, 1161–1172.

Kulesa, P., Bronner-Fraser, M., Fraser, S., 2000. In ovo time-lapse analysis

after dorsal neural tube ablation shows rerouting of chick hindbrain

neural crest. Development 127, 2843–2852.

Kurtz, A., Zimmer, A., Schnutgen, F., Bruning, G., Spener, F., Muller, T.,

1994. The expression pattern of a novel gene encoding brain-fatty acid

binding protein correlates with neuronal and glial cell development.

Development 120, 2637–2649.

Le Douarin, N.M., Teillet, M.A., 1973. The migration of neural crest cells

to the wall of the digestive tract in avian embryo. J. Embryol. Exp.

Morphol. 30, 31–48.

Leibl, M.A., Ota, T., Woodward, M.N., Kenny, S.E., Lloyd, D.A., Vaillant,

D.H., Edgar, D.H., 1999. Expression of endothelin 3 by mesenchymal

cells of embryonic mouse caecum. Gut 44, 246–252.

Lichtman, J.W., Fraser, S.E., 2001. The neuronal naturalist: watching neu-

rons in their native habitat. Nat. Neurosci. , (4 Suppl), 1215–1220.

Lo, L., Dormand, E., Greenwood, A., Anderson, D.J., 2002. Comparison of

the generic neuronal differentiation and neuron subtype specification

functions of mammalian achaete-scute and atonal homologs in cultured

neural progenitor cells. Development 129, 1553–1567.

Lois, C., Garcia-Verdugo, J.M., Alvarez-Buylla, A., 1996. Chain migration

of neuronal precursors. Science 271, 978–981.

Murray, M.J., Merritt, D.J., Brand, A.H., Whitington, P.M., 1998. In vivo

dynamics of axon pathfinding in the Drosophilia CNS: a time-lapse

study of an identified motorneuron. J. Neurobiol. 37, 607–621.

Nadarajah, B., Brunstrom, J.E., Grutzendler, J., Wong, R.O., Pearlman, A.L.,

2001. Twomodes of radial migration in early development of the cerebral

cortex. Nat. Neurosci. 4, 143–150.

Nadarajah, B., Alifragis, P., Wong, R.O., Parnavelas, J.G., 2002. Ventricle-

directed migration in the developing cerebral cortex. Nat. Neurosci. 5,

218–224.

Natarajan, D., Grigoriou, M., Marcos-Gutierrez, C.V., Atkins, C., Pachis,

V., 1999. Multipotential progenitors of the mammalian enteric nervous

system capable of colonising aganglionic bowel in organ culture.

Development 126, 157–168.

Natarajan, D., Marcos-Gutierrez, C., Pachnis, V., de Graaff, E., 2002. Re-

quirement of signalling by receptor tyrosine kinase RET for the directed

migration of enteric nervous system progenitor cells during mammalian

embryogenesis. Development 129, 5151–5160.

Newgreen, D.F., 1990. Control of the directional migration of mesenchyme

cells and neurites. Semin. Dev. Biol. 1, 301–311.

Newgreen, D.F., Hartley, L., 1995. Extracellular matrix and adhesive

molecules in the early development of the gut and its innervation

in normal and spotting lethal rat embryos. Acta Anat. (Basel) 154,

243–260.

Newgreen, D., Young, H.M., 2002a. Enteric nervous system: develop-

ment and developmental disturbances—Part 1. Pediatr. Dev. Pathol.

5, 224–247.

Newgreen, D., Young, H.M., 2002b. Enteric nervous system: develop-

ment and developmental disturbances—Part 2. Pediatr. Dev. Pathol. 5,

329–349.

Newgreen, D.F., Ritterman, M., Peters, E.A., 1979. Morphology and be-

haviour of neural crest cells of chick embryo in vitro. Cell Tissue Res.

203, 115–140.

Pachnis, V., Mankoo, B., Costantini, F., 1993. Expression of the c-ret

H.M. Young et al. / Developmental Biology 270 (2004) 455–473 473

proto-oncogene during mouse embryogenesis. Development 119,

1005–1017.

Parisi, M.A., Kapur, R.P., 2000. Genetics of Hirschsprung disease. Curr.

Opin. Pediatr. 12, 610–617.

Pattyn, A., Morin, X., Cremer, H., Goridis, C., Brunet, J.F., 1997. Expres-

sion and interactions of the two closely related homeobox genes Phox2a

and Phox2b during neurogenesis. Development 124, 4065–4075.

Peters-van der Sanden, M.J., Kirby, M.L., Gittenberger-de Groot, A.,

Tibboel, D., Mulder, M.P., Meijers, C., 1993. Ablation of various

regions within the avian vagal neural crest has differential effects on

ganglion formation in the fore-, mid- and hindgut. Dev. Dyn. 196,

183–194.

Pomeranz, H.D., Rothman, T.P., Gershon, M.D., 1991. Colonization of the

post-umbilical bowel by cells derived from the sacral neural crest: direct

tracing of cell migration using an intercalating probe and a replication-

deficient retrovirus. Development 111, 647–655.

Rovasio, R.A., Delouvee, A., Yamada, K.M., Timpl, R., Thiery, J.P., 1983.

Neural crest cell migration: requirements for exogenous fibronectin and

high cell density. J. Cell Biol. 96, 462–473.

Schiltz, C.A., Benjamin, J., Epstein, M.L., 1999. Expression of the GDNF

receptors Ret and GFRalpha1 in the developing avian enteric nervous

system. J. Comp. Neurol. 414, 193–211.

Shen, L., Pichel, J.G., Mayeli, T., Sariola, H., Lu, B., Westphal, H., 2002.

Gdnf haploinsufficiency causes Hirschsprung-like intestinal obstruction

and early-onset lethality in mice. Am. J. Hum. Genet. 70, 435–447.

Sidebotham, E.L., Woodward, M.N., Kenny, S.E., Lloyd, D.A., Vaillant,

C.R., Edgar, D.H., 2001. Assessment of protein gene product 9.5 as a

marker of neural crest-derived precursor cells in the developing enteric

nervous system. Pediatr. Surg. Int. 17, 304–307.

Stewart, A.L., Anderson, R.B., Young, H.M., 2003. Characterization of

lacZ-expressing cells in the gut of embryonic and adult DbH-nlacZ

mice. J. Comp. Neurol. 464, 208–219.

Taraviras, S., Pachnis, V., 1999. Development of the mammalian enteric

nervous system. Curr. Opin. Genet. Dev. 9, 321–327.

Thomas, L.A., Yamada, K.M., 1992. Contact stimulation of cell migration.

J. Cell. Sci. 103, 1211–1214.

Wu, J.J., Chen, J.X., Rothman, T.P., Gershon, M.D., 1999. Inhibition of in

vitro enteric neuronal development by endothelin-3: mediation by endo-

thelin B receptors. Development 126, 1161–1173.

Yntema, C.L., Hammond, W.S., 1954. The origin of intrinsic ganglia of

trunk viscera from vagal neural crest in the chick embryo. J. Comp.

Neurol. 101, 515–541.

Young, H.M., Newgreen, D., 2001. Enteric neural crest-derived cells:

origin, identification, migration, and differentiation. Anat. Rec. 262,

1–15.

Young, H.M., Hearn, C.J., Ciampoli, D., Southwell, B.R., Brunet, J.F.,

Newgreen, D.F., 1998. A single rostrocaudal colonization of the rodent

intestine by enteric neuron precursors is revealed by the expression of

Phox2b, Ret, and p75 and by explants grown under the kidney capsule

or in organ culture. Dev. Biol. 202, 67–84.

Young, H.M., Ciampoli, D., Hsuan, J., Canty, A.J., 1999. Expression of

ret-, p75(NTR)-, Phox2a-, Phox2b-, and tyrosine hydroxylase-immu-

noreactivity by undifferentiated neural crest-derived cells and different

classes of enteric neurons in the embryonic mouse gut. Dev. Dyn.

216, 137–152.

Young, H.M., Hearn, C.J., Farlie, P.G., Canty, A.J., Thomas, P.Q., New-

green, D.F., 2001. GDNF is a chemoattractant for enteric neural cells.

Dev. Biol. 229, 503–516.

Young, H.M., Jones, B.R., McKeown, S.J., 2002. The projections of early

enteric neurons are influenced by the direction of neural crest cell

migration. J. Neurosci. 22, 6005–6018.

Young, H.M., Bergner, A.J., Muller, T., 2003. Acquisition of neuronal

and glial markers by neural crest-derived cells in the mouse intestine.

J. Comp. Neurol. 456, 1–11.