Temporal regulation of cerebellar EGL migration through a switch in cellular responsiveness to the meninges

www.elsevier.com/locate/ydbio

Developmental Biology 267 (2004) 153–164

Temporal regulation of cerebellar EGL migration through a switch in

cellular responsiveness to the meninges

Yan Zhu,1 Tao Yu, and Yi Rao*

Department of Anatomy and Neurobiology, Washington University School of Medicine, St. Louis, MO 63110, USA

Received for publication 2 November 2002, revised 1 October 2003, accepted 28 October 2003

Abstract

We have studied the temporal and spatial control of cell migration from the external germinal layer (EGL) in the mammalian cerebellum

as a model for cortical migration. Our results have demonstrated that embryonic EGL cells do not migrate into internal layers because they

respond to a diffusible attractant in the meninges, the nonneural tissues covering the nervous system, and to a repellent in the

neuroepithelium. Two developmental changes are important for postnatal EGL migration: the disappearance of the repellent in the inner

layers and a switch in cellular responsiveness of EGL cells so that the postnatal EGL cells respond to the repellent, but not the attractant in the

meninges. Besides revealing the signaling role of meninges in cortical development, our study suggests that an active mechanism is required

to prevent cell migration, and that mechanisms of cell migration should be studied even in the absence of apparent changes in cell positions.

We propose a model for the developmental control of neuronal migration in the cerebellar cortex.

D 2003 Elsevier Inc. All rights reserved.

Keywords: Cell migration; Neuronal migration; Cortex; Cerebellum; Neural development

Introduction

Neuronal migration is essential for the formation of

cerebral and cerebellar cortices in the mammalian brain. A

classic model for cortical lamination involves the migration

of cells from the external germinal layer (EGL) into the

internal granule layer (IGL) in the cerebellum (Altman and

Bayer, 1997; Hatten, 1999; Hatten and Heintz, 1995; Pearl-

man et al., 1998; Rakic, 1971a,b, 1990; Rice and Curran,

1999).

In the embryo, EGL cells are derived from cells in the

upper rhombic lip (URL) (Alder et al., 1996; Altman and

Bayer, 1997; Hatten, 1999; Wingate and Hatten, 1999).

Cells in the URL migrate tangentially to form the EGL

(Wingate and Hatten, 1999). Once in the EGL, the URL

derived cells do not migrate towards the IGL in the embryo

(Altman and Bayer, 1997; Hatten, 1999). Postnatally, EGL

cells are induced to proliferate by the sonic hedgehog

protein secreted from the Purkinje cells underlying the

0012-1606/$ - see front matter D 2003 Elsevier Inc. All rights reserved.

doi:10.1016/j.ydbio.2003.10.037

* Corresponding author. Department of Anatomy and Neurobiology,

Washington University School of Medicine, Box 8108, 660 South Euclid

Avenue, St. Louis, MO 63110. Fax: +1-314-362-3446.

E-mail address: [email protected] (Y. Rao).1 Current address: HHMI at UCLA, 675 Charles E. Young Drive South,

5-784 MRL, Los Angeles, CA 90095.

EGL cells (Wechsler-Reya and Scott, 1999). Postnatal

EGL cells migrate radially from the more superficial layer

through the Purkinje cell layer to IGL, giving rise to the

granule cells (Rakic, 1971a, 1990). At the cellular level, the

migration of EGL cells towards the IGL in vivo is thought to

be dependent on radially aligned glial fibers (Rakic, 1971a).

Mutations in several genes affect neuronal positioning in

the cerebellum of rodents (Dhavan and Tsai, 2001; Rice and

Curran, 1999, 2001; Ross and Walsh, 2001). These include

reeler (reviewed in Rice and Curran, 2001), scrambler

(Howell et al., 1997; Sheldon et al., 1997; Sweet et al.,

1996; Ware et al., 1997), disabled-1 (Dab-1), weaver

(Hatten et al., 1984), cdk5 and its activator p35 (Chae et

al., 1997; Dhavan and Tsai, 2001; Nikolic et al., 1996; Tsai

et al., 1994). The reeler gene encodes a secreted protein

Reelin (D’Arcangelo et al., 1995). Reelin binds to the very

low density lipoprotein (VLDL) and ApoE2 receptors and

activates a pathway involving Dab-1 (D’Arcangelo et al.,

1999; Hiesberger et al., 1999; Trommsdorff et al., 1999). In

the neocortex, the precise functional role of the Reelin

pathway is still not clear (Rice and Curran, 2001), although

one possibility is that Reelin provides a stop signal for

neurons migrating from the ventricular zone to the cortical

plate (Pearlman et al., 1998). In the cerebellum, the Reelin

pathway does not seem to act on the EGL cells; Reelin is

Y. Zhu et al. / Developmental Biology 267 (2004) 153–164154

secreted by granule precursor cells and acts on Purkinje cells

to promote the formation of the Purkinje cell layer (D’Ar-

cangelo et al., 1995; Howell et al., 1997; Sheldon et al.,

1997).

Despite extensive studies, the understanding of mecha-

nisms controlling the spatial and temporal aspects of EGL

migration is still limited. For temporal control, there have

been no experimental studies designed to address the

question why the embryonic EGL cells do not migrate into

the IGL whereas the postnatal EGL cell do. For spatial

control, it is not known what directs postnatal EGL cells

migrate towards the IGL. We have now investigated the

mechanisms underlying temporal and spatial control of EGL

migration by studying the migration of embryonic and

postnatal EGL cells in rodents. The study documents the

developmental properties of both guidance sources and

targets neurons, and provides a framework for understand-

ing molecular basis of EGL migration. Our analyses have

revealed attractive and repulsive forces that control the

migration of EGL cells. Our result uncovered a signaling

Fig. 1. Migration of EGL cells in three dimensional gel matrices. (A) A diagram of

the URL and their placements in collagen cultures. Note that cells migrate from thre

the edge of URL. (B) A diagram of E17 rat cerebellar primordium, showing

neuroepithelium (NE). (C) A diagram of part of the neonatal cerebellum, showing

(D) A bright field view of an E15 URL explant cultured alone for about 36 h. As in

field view of an E17 EGL explant cultured in collagen gels for the same length of p

gels for about the same length of period as D. The scale bar is 100 Am.

role for the meninges in neural development. We discover

that an attractive activity is involved in halting intrinsically

migratory neurons. This is the first time in any system that

such mechanism is required to prevent cells from premature

migration. We have found that temporal control of cell

migration is accomplished by regulation of spatial control.

Based on our study, we propose a systematic model on

temporal and spatial control of EGL migration.

Materials and methods

Materials

Timed pregnant Sprague–Dawley rats were obtained

from Charles River. Collagen gels were made from rat tail

tendon. Matrigel basement membrane matrix was from

Collaborative Biomedical Products. Mouse anti-TuJ1 mono-

clonal antibody was from Covance Research Products; anti-

nestin monoclonal antibody (RAT401) was from PharMin-

E15 rat cerebellar primordium (CEP) showing the isolation of explants from

e sides of the explants (b, c and d), but not from side a because this side was

the relative positions of the EGL, the differentiation zone (DZ) and the

the EGL, the Purkinje cell layer (PL) and the internal granule layer (IGL).

dicated in panel A, few cells migrated out from the lower edge. (E) A bright

eriod as D. (F) A bright field view of a P0 EGL explant cultured in collagen

Y. Zhu et al. / Developmental Biology 267 (2004) 153–164 155

gen International; monoclonal anti-GFAP antibody was

from Sigma (G 3893). Cy3-conjugated donkey anti-goat

antibodies and Cy3-conjugated goat anti-rabbit antibodies

were from Jackson ImmunoResearch Laboratories. Pertussis

toxin (PTX, P7208), Hoechst 33258, and paraformaldehyde

(PFA) were obtained from Sigma.

Explant assay for EGL migration

Cocultures of EGL explants were performed similarly as

previously reported with some modifications (Zhu et al.,

1999). Explants were isolated from the EGL of E17 rats or

the URL of E15 rats. Explants from postnatal day 0 (P0) and

P5 behaved similarly. For explants of postnatal EGL, para-

sagittal cerebellar slices of 300 Am were cut on a Vibratome,

and the EGL explants were dissected by a fine tungsten

needle from the anterior cerebellum. Meninges were taken

from the area directly covering the cerebellum. Isolated

regions of proper layers had been confirmed by immunos-

taining with anti-MATH1, anti-TuJ1, anti-Calbindin, and

anti-nestin antibodies on frozen sections of explants that

were fixed immediately after dissection. Aggregates of HEK

were prepared as described previously (Li et al., 1999; Zhu

et al., 1999). Explants and cell aggregates were positioned at

a distance of 300 to 400 Am in a 1:2 mixture of Matrigel and

collagen gel. For the embryonic EGL explants, few cells

migrated out of the edge of rhombic lip (side a in Figs. 1A

and D), and we therefore avoided to have this edge face

toward or away from the source: the explants were posi-

tioned in a way that either of the cutting edges of the

medial– lateral orientation in the developing cerebellum

faced the source (side b or d in Fig. 1A). For the postnatal

EGL explants, the explants were positioned in a way that

either of the vibratome-sectioned edges faced the source.

The cocultures were performed in 10% FCS (fetal calf

serum), 1�P/S (penicillin and streptomycin), DMEM (Dul-

becco’s minimal essential medium, Gibco) for 24 to 48 h in

5% CO2 at 37jC. To calculate P/D and D/P ratios, cultures

were fixed and stained with the nuclear dye. Pictures were

taken with a CCD camera and the images were analyzed

with Scion Image software (Release Beta 4.0.2, Scion

Corporation) to obtain the ratios of numbers of migratory

cells on the proximal and distal sides. Statistic analyses were

carried out with Student’s t test.

Immunostaining of the explants cultured in collagen

matrix was performed similarly as previously reported

(Zhu et al., 1999). Confocal images were taken on an

Olympus fluoview laser scanning microscope. SYTOX

Green nucleic acid stain (S-7020, Molecular Probes) instead

of Hoechst dye was used as the nuclear dye in the confocal

images.

PTX blocking experiment

Explants were dissected and then incubated in the culture

medium with 0.1 Ag/ml of PTX for 4 h at 37jC. The

controls were explants cultured at the same time in the

medium without PTX. After pretreatment, cocultures of the

EGL explants with meninges were performed as described

above except that PTX was added to both the mixture gel

and the culture medium. PTX in control cocultures did not

affect the production of secreted proteins from HEK cell

aggregates (data not shown).

Results

Migration of neurons from the EGL in three dimensional gel

matrices

It is known that cells in the URL migrate tangentially

into the EGL of the mammalian cerebellum (Adler et al.,

1996; Altman and Bayer, 1997; Hatten, 1999; Wingate and

Hatten, 1999). However, the embryonic EGL cells do not

migrate internally whereas the postnatal EGL cells do

migrate into the IGL (Altman and Bayer, 1997; Hatten,

1999). The mechanism for the lack of migration of embry-

onic EGL cells into the internal layers has not been

previously investigated. During the course of our studies

of EGL cell migration to address the question why embry-

onic EGL cells do not move into internal layers, we

established an in vitro culture assay with three dimensional

collagen matrices.

Three dimensional matrices are a reliable system for

studying neuronal migration (Letinic and Rakic, 2001;

Marin et al., 2001; Pleasure et al., 2000; Powell et al.,

2001; Wong et al., 2001; Wu et al., 1999; Zhu et al., 1999)

and were used here to culture explants of the URL and

different stages of EGL (Figs. 1 and 2). Cells could migrate

out of the URL explants from embryonic day 15 (E15) rats,

EGL explants of E17 and postnatal day 0 (P0) rats (Fig. 1).

Immunocytochemistry with the neuronal specific anti-TuJ1

antibody showed that these migrating cells were neurons

(Figs. 2B and E). Antibodies to the glial fibrillary acidic

protein (GFAP) can be used to examine postnatal glial cells

in the cerebellum, but they do not stain for embryonic glial

precursor cells. Antibodies to nestin could reveal glial

precursor cells as well as neuronal precursor cells in

embryonic and postnatal cerebellum (Hopfield and McKay,

1985; Sotelo et al., 1994). Results from immunocytochem-

istry with antibodies to nestin and GFAP showed that glial

cells and their precursor cells stayed within the explants and

did not migrate out of either embryonic or postnatal EGL

explants (Figs. 2H, K and N). All cells that migrated out of

the explants are TuJ1 positive neurons (Fig. 2E). These

results indicate that neurons in the EGL do not rely on glial

fibers for their migration.

The leading processes of neurons migrating from E15

URL and postnatal EGL explants are quite long and give the

appearance that migrating neurons are associated with, or

perhaps even move along, the processes of other neurons

(Figs. 1D and F). However, although neurons from E17

Fig. 2. Distribution of neuronal and glial cells in cultures of EGL explants. Embryonic URL explants (A–C, G–I) and postnatal EGL explants (D–F, J–O)

were stained with anti-TuJ1 (B, E), anti-nestin (H, K), or anti-GFAP antibodies (N). All antibody stainings were shown in red. The color of counter-staining

with the nuclear dye Hoechst 33258 was converted from blue to green for better visualization (A, D, G, J, M). Combining red and green signals of the same

explants results merged pictures (C, F, I, L, O). E17 EGL explants were similar to E15 URL and are therefore not shown here.


EGL explants had leading processes, they were shorter than

those of E15 and postnatal EGL explants. It is clear that E17

neurons migrated individually, without relying on the fibers

Fig. 3. Guidance of embryonic EGL cells by the meninges and the neuroepithelium

More cells were present in the quadrant proximal to the eM than those in the qua

Coculture of an E15 EGL explant with postnatal meninges (pM); (D) Coculture o

same stage. (E–H) Confocal images of the cocultures of embryonic EGL explants

Dots mark the edge of the meningeal explants. An optical section of the explant sta

another explant stained with the nuclear dye (G) and anti-nestin antibody (H).

explants so that dotted lines were straight. (I) Statistical analysis of results of cultu

the control, 49 for eM, 94 for pM, 11 for DZ, 6 for NE, and 12 for DZ + NE. The

and the P–D ratios were calculated (mean F S.E.M.) (Zhu et al., 1999), a P–D rat

distal/proximal (D–P) ratio was shown to indicate repulsion. Columns eM and pM

whereas column DZ is not ( P = 0.916).

of other neurons (Figs. 1E and 3B and D). These results

indicate that, although it is difficult to rule out the possibility

of neuronal migration along other neuronal fibers for E15

. (A) Coculture of an E15 EGL explant with the embryonic meninges (eM).

drant distal to the eM; (B) Coculture of an E17 EGL explant with eM; (C)

f an E17 EGL explant with an explant of the neuroepithelium (NE) of the

with embryonic meninges. White dashes indicate the outline of the explants.

ined with a nuclear dye (E) and anti-TuJ1 antibody (F). An optical section of

For confocal studies, a permeable filter was placed around the meningeal

ring EGL explants alone (C) or cocultures; numbers of explants used: 31 for

numbers of cells in the proximal (P) and distal (D) quadrants were counted

io larger than 1 indicates attraction. When P–D ratio is much smaller than 1,

are very significantly different from the control ( P < 0.001 in both cases)


and postnatal EGL explants, in principle, migrating neurons

do not have to move along tracks formed by fibers of other

neurons, at least in the cases of neurons from E17 explants.

With the exception of the length of neuronal processes, the

URL cells and embryonic EGL cells responded similarly in

all assays described below.


Attractive and repulsive activities in the Meninges and in the

neuroepithelium

With the establish assay, we investigated whether the

migration of embryonic URL and EGL cells could be

affected by their neighboring tissues. The outer neighbor

of the EGL is the nonneuronal meninges, which cover the

nervous system. Layers internal to the EGL in the embry-

onic cerebellum include the differentiation zone (DZ) and

the neuroepithelium (NE) (Fig. 1B).

The intuitively obvious way for the neighboring

tissues to control EGL migration is to inhibit the

intrinsic motility of the embryonic EGL cells. However,

coculture experiments showed that cells migrated from

the embryonic EGL explants in the presence of the

neighboring tissues (Fig. 3), indicating that none of the

neighboring tissues could inhibit EGL migration. Inter-

estingly, cells from both the E15 URL (Fig. 3A) and

the E17 EGL (Fig. 3B) migrated towards the meninges,

as indicated by the presence of a larger number of cells

in the quadrant proximal to the meninges than that in

the distal quadrant (Figs. 3A, B and I). These results

indicate that the meninges attract EGL cells, and do not

inhibit EGL migration. Because the attraction was not

dependent on cell contact (Figs. 3A and B), the

attractive cue(s) in the meninges must be diffusible.

The neuronal nature of EGL cells attracted by the

meninges was shown by positive staining for the

TuJ1 antibody (Fig. 3F). The attractive effect of the

meninges is likely to be direct on the neurons, rather

than being indirectly through glial fibers because stain-

ing for glial precursor cells with an anti-nestin antibody

showed that glial precursor cells stayed within the

explants (Fig. 3H).

When EGL explants were cocultured with the NE, it was

found that the NE repelled the EGL cells (Figs. 3D and I).

The repellent(s) in the NE was also diffusible. The DZ

neither attracted nor repelled the EGL cells (Fig. 3I).

Explants containing both the NE and the DZ (labeled as

NE + DZ in Fig. 3I) were still repulsive to the EGL cells,

indicating that DZ did not mask the repulsive activity

diffusing from the NE.

In summary, the embryonic EGL cells respond to a

diffusible attractive activity in the meninges and a diffus-

ible repellent in the NE, suggesting that the strategic

positioning of the guidance cues may make the EGL cells

unable to move into the inner layers in the embryonic

cerebellum.

Fig. 4. Responses of postnatal EGL cells to guidance activities. (A) Cocultur

magnification view of the distal side of the EGL explant in A. (C) A high magnif

100 Am. (D) Coculture of a postnatal EGL explant with the embryonic meninges.

and internal granular layer. (F) Staining with the nuclear dye. (G) Staining of the sa

Staining of the same explant as that in H with the anti-nestin antibody. The scale ba

numbers of explants used: 28 in pEGL alone (C), 134 in eM, 215 in pM, and 22

control ( P < 0.001 in both cases), whereas column PL + IGL is not ( P = 0.773

Developmental changes in the responsiveness of EGL cells

to guidance cues in the meninges

The presence of attractive and repulsive activities in

tissues surrounding the EGL raises the question of how

postnatal EGL cells can migrate into the IGL. We first

investigated the activity in the inner layers. Postnatally, the

neuroepithelium is absent. But do the Purkinje cell layer

(PL) and the IGL have any attractive or repulsive activities?

Coculture experiments indicated that the PL and the IGL

were neither attractive nor repulsive to the postnatal EGL

cells (Figs. 4E and J).

We then examined the guidance activities in the meninges.

If the meninges are attractive to the EGL, how could EGL

cells move away from it? One possibility is that there is a

developmental change in the signal sent by the meninges, so

that only the embryonic, but not the postnatal, meninges

send the attractant. Alternatively, there could be a change in

the responsiveness of the EGL cells to the meningeal

signal(s), so that postnatal EGL cells do not respond to

the attractive signal. It is also possible that both of these

changes take place.

To distinguish among possibilities, we cocultured em-

bryonic and postnatal meningeal explants with EGL

explants of the same or different developmental stages.

Similar to the embryonic meninges (Figs. 3A and B), the

postnatal meninges also attracted the embryonic EGL cells

(Fig. 3C). On the other hand, the postnatal EGL cells were

repelled by both the postnatal meninges (Fig. 4A) and the

embryonic meninges (Fig. 4D). The repelled postnatal EGL

cells were neurons (Fig. 4G) whereas glial cells stayed

within the explants (Fig. 4I) and glial fibers grew a very

short distance outside the explants (Fig. 4I).

These results indicate that both attractive and repulsive

activities are present in the meninges and that these activ-

ities are present in both the embryonic and postnatal

meninges. By contrast, a critical developmental change

has taken place in the EGL cells: the embryonic EGL cells

respond predominantly to the attractant in the meninges,

whereas the postnatal EGL cells respond predominantly to

the repellent in the meninges.

Differential sensitivities of the attractive and repulsive

responses to PTX

Because it is known that the same cells can change their

responses to the same molecular cue (Ming et al., 1997;

Poznansky et al., 2000; Song et al., 1997, 1998) and because

e of a postnatal EGL explant with the postnatal meninges. (B) A high

ication view of the proximal side of the EGL explant in A. The scale bar is

(E) Coculture of a postnatal EGL explant with the postnatal Purkinje layer

me explant as that in F with anti-TuJ1. (H) Staining with the nuclear dye. (I)

r is 100 Am. (J) Statistical analysis with the response of postnatal EGL cells;

in PL + IGL. Columns eM and pM are very significantly different from the

).


SDF-1 is the attractant for the embryonic EGL cells in the

meninges (Lu et al., 2001; Zhu et al., 2002), it is possible

that postnatal EGL cells are repelled by SDF-1 in the

meninges. In other words, the same molecule that attracts

the embryonic EGL cells may also repel the postnatal EGL

cells.


We used the pertussis toxin (PTX), an inhibitor of the

heterotrimeric G proteins, to test the sensitivities of the

EGL cells. We found that, although the attractive response

of embryonic EGL cells to the meninges was blocked by

PTX (Figs. 5A and C), the repulsive response of the

postnatal EGL cells to the postnatal meninges could not

be blocked by PTX (Figs. 5B and C). These results

indicate that the attractant and the repellent are distinct

and that the attractant signals through G-protein coupled

receptors sensitive to PTX, whereas the repellent in the

meninges does not.

Fig. 5. PTX sensitivity of the attractive and repulsive responses in the EGL. (A

embryonic meninges in the presence of PTX. The attractive response was blocked

the postnatal meninges in the presence of PTX. The repulsive response was not

response of the EGL cells toward the meninges in the presence of PTX; numbers o

pM, and 12 in pEGL + pM + PTX. eEGL + eM is very significantly different from

from pEGL + pM + PTX ( P = 0.630).

Discussion

Our studies have provided insights into two important

aspects of cell migration. First, our results have provided a

basis to understand cellular mechanisms controlling both the

timing and the direction of EGL cell migration in the

cerebellum. The model illustrated in Fig. 6 describes major

cellular and molecular forces that provide directional guid-

ance for neuronal migration in the cerebellar cortex. The

general principle of cellular interactions in controlling

neuronal migration may be applicable in other parts of the

) Nuclear staining of a coculture of an embryonic EGL explant with the

by PTX. (B) Nuclear staining of a coculture of a postnatal EGL explant with

blocked by PTX. The scale bar is 100 Am. (C) Statistical analysis of the

f explants used: 34 in eEGL + eM, 39 in eEGL + eM + PTX, 42 in pEGL +

eEGL + eM + PTX ( P < 0.001). pEGL + pM is not significantly different

Fig. 6. A model for mechanisms underlying temporal and spatial control of EGL cell migration. The left panel is a diagram of the embryonic cerebellum.

Diffusible attractant (+) and repellent (� ) are present in the embryonic meninges. The embryonic EGL cells respond to the attractive cue in the meninges.

There is a diffusible repulsive activity in the neuroepithelium (NE). The combination of the meningeal attractant and the neuroepithelial repellent prevents the

EGL cells from migrating into the internal layers of the embryonic cerebellum. PPL: prospective Purkinje layer. The right panel is a diagram of the postnatal

cerebellum. Both the attractive and repulsive cues are still present in the postnatal meninges. However, the postnatal EGL cells respond predominantly to the

repulsive cue in the meninges. The neuroepithelium is not present in the postnatal cerebellum and there is no repulsive activity in the Purkinje layer and the

IGL. EGL cells will thus migrate towards the IGL.


nervous system such as the neocortex. Secondly, our results

demonstrate that an active mechanism is required to keep

inherently mobile cells in place, which can be generally

important for controlling cell positioning not only in the

nervous system but also in other systems.

A model for temporal and spatial control of cell migration

in the cerebellum

We propose here the first mechanistic model for the

temporal and spatial control of cortical migration (Fig. 6).

Cortical migration is controlled by strategic positioning of

multiple attractive and repulsive cues and developmental

changes in the cellular responsiveness to these cues. In the

embryo, although the meninges contain both attractants and

repellents, the embryonic EGL cells respond predominantly

to the meningeal attractant. There is also a diffusible repel-

lent from the neuroepithelium. The combination of the

meningeal attractant and the neuroepithelial repellent pre-

vents the EGL cells from migrating into the internal layers.

Postnatally, we have shown that the inner layers are not

repulsive. However, both the attractant and the repellent are

still present in the postnatal meninges. Due to a change in

cellular responsiveness, postnatal EGL cells respond to the

repellent in the meninges, but not to the attractive activity.

In other parts of the nervous system, the model outlined

in Fig. 6 may also be applicable. For example, in the

neocortex, it is possible that cells from the ventricular zone

are repelled by cues in the ventricular zone and move

towards the outer layers by attractive cues in the outer

layers or in the meninges. In the spinal cord, the same

principle may also apply.

Molecular identities of the chemoattractive and

chemorepulsive factors

Our model for mechanisms underlying temporal and

spatial control of EGL cell migration provides a framework

to put molecular players involving in EGL migration into

proper context. We have uncovered both attractive and

repulsive activities in the meninges and a repulsive activity

in the neuroepithelium. The molecular identity of the

attractant in the meninges has been reported recently (Lu

et al., 2001; Zhu et al., 2002).

The role of SDF-1 as the meningeal attractant has been

proven by several findings. (1) Previous findings from mice

lacking SDF-1 and its receptor CXCR4 have led to the

observation of abnormal neuronal positioning in the IGL

(Ma et al., 1998; Tachibana et al., 1998; Zou et al., 1998).

(2) The expression patterns of SDF-1 and CXCR4 are

consistent of a role for SDF-1 in anchoring the EGL cells

(Klein et al., 2001; Lu et al., 2001; Zhu et al., 2002; Zou et

al., 1998). (3) Transwell assays indicate that SDF-1 can

attract postnatal granule cells (Klein et al., 2001; Lu et al.,

2001). (4) In vitro explant coculture assays demonstrate that

embryonic EGL cells are attracted by SDF-1 whereas

postnatal EGL cells are not (Zhu et al., 2002). (5) Meninges

from SDF-1 knock-out mice no longer attract embryonic

EGL cells (Zhu et al., 2002). These results indicate that

SDF-1 is either the predominant or the sole attractant in the

meninges.

The repellent in the meninges is not known, but it is not

SDF-1 because its activity is still present in SDF-1 knock-

out mice (Zhu et al., 2002) and cannot be blocked by PTX

(Fig. 5).

Anchoring as a new mechanism for controlling cell

migration, and perhaps also for regulating axonal and

dendritic connections

Our finding of the use of an attractant to hold responsive

neurons in place is related to a question of general interest

for cell migration in all systems. Cell migration is usually

inferred from changes in cell positioning during develop-

ment or in adult life. In fact, lack of migration is most often

thought to imply the inability of cells to move, and there is


little interest in studying cell migration when the apparent

positions of cells do not change. We propose a new

mechanism named anchoring for actively preventing inher-

ently mobile cells from migrating in a spatially or tempo-

rally erroneous manner.

Multiple potential mechanisms can be used to anchor

cells or neuronal processes. In theory, factors that make cells

lose motility transiently or permanently should anchor cells

for a corresponding period of time. Regulation of the

permissiveness of a migratory pathway could also be used

to anchor cells. Adhesive anchoring is another possible

means to anchor cells. This would involve increased cell

adhesion to substrates or neighboring cells.

Our results with the meninges indicate that an attractive

guidance cue plays an unusual role: not by directing cells to

their target region, but by anchoring the cells beneath the

meninges and preventing them from inward migration. This

mode of anchoring is termed attractive anchoring. There can

also be a mode of repulsive anchoring, whereby a repulsive

cue is placed to prevent cells from entering an otherwise

permissive pathway for migration. Thus, when attractive

and repulsive cues are operating in anchoring, the strategic

localization of the guidance cue is crucial.

If anchoring is a general mechanism inside and outside

the nervous system, a conceptual change is required. Not

only should we continue to investigate why cells move

towards their targets, but we will also have to study why

cells do not move away from certain regions. When cells are

apparently stationary for a short or a long time, they may

still be regulated by active mechanisms. The apparently

nonmigratory behavior of these cells or processes makes it

difficult to pinpoint where these mechanisms function in

normal development. Experimental manipulations will be

required to provide important starting points for exploring

functional roles of anchoring.

Similar mechanism may also exist in other neuronal

processes including axons and dendrites that are known to

undergo long- or short-range navigation. Because there are

common mechanisms regulating neuronal migration, and

axonal or dendritic projections (Polleux et al., 2000; Wu et

al., 1999; Yee et al., 1999; Zhu et al., 1999), it will be

interesting to examine whether anchoring can control the

beginning of outgrowth or temporary stops during axonal or

dendritic navigation. The most obvious step for anchoring to

function is at the termination of axonal or dendritic projec-

tion. Anchoring between the pre- and postsynaptic compo-

nents could be an important step in synaptogenesis. It is also

possible for anchoring to play roles in neural plasticity;

competition for anchoring cues may underlie some forms of

plasticity.

A signaling role for the meninges

While the intact meninges provide physical support to

the nervous system, traditionally they have not been thought

to play any important roles in neural development. Recent

work by Hatten et al. suggested that collagen IV expressed

in the pia could be a ligand for the discoidin domain

receptor 1 (DDR1) and that the interaction between collagen

IV and DDR1 might regulate axon extension of cerebellar

granule neurons (Bhatt et al., 2000).

Our results demonstrate that the meninges provide an

anchoring signal for cerebellar EGL cells. Since the meninx

covers other parts of the nervous system, it may also play

essential roles in the development of other regions. In cell

migration, those roles do not have to be anchoring cells near

the surface, they could also be attracting cells from the

deeper layers into more superficial layers, or repelling cells,

axons or dendrites from more superficial layers to deeper

layers. Although we only have evidence for its role in cell

migration, it may not be too speculative to think of other

possible roles for the meninges.

There are three layers in the meninges, the dura mater,

the arachnoid mater, and the pia, with the pia closest to the

EGL (Angelov and Vasilev, 1989; Kamiryo et al., 1990;

O’Rahilly and Mueller, 1986). Because it is technically not

possible to separate the layers and their precursors in

development (Angelov and Vasilev, 1989; Kamiryo et al.,

1990), we could not test the roles of individual layers of the

meninges for their signaling roles. However, the presence of

SDF-1 in all three layers suggests that there may not be

much difference (Zhu et al., 2002).

Questions about the precise role of glial fibers in neuronal

migration

The migration of EGL cells towards the IGL has long

been thought to depend on radial glial fibers (Hatten, 1999;

Hatten and Heintz, 1995; Rakic, 1971a,b, 1990). Because of

the linear shapes of glial processes, the mere action of

climbing along glial processes may greatly reduce the

possible migratory directions into two. We intended to

address whether radial glia also provide directional guidance

cues for neurons to move one direction vs. another one. Our

results with cells migrating out of EGL explants in vitro

raise questions about the precise role of the glial fibers in

neuronal migration.

In vitro, EGL cells can bind to and migrate along the

glial fibers (Edmondson and Hatten, 1987; Gregory et al.,

1988; Hatten and Liem, 1981; Hatten and Mason, 1990;

Mason et al., 1988). The binding of EGL cells to the glial

fiber requires the cell adhesion molecule astrotactin

(Edmondson et al., 1988; Stitt and Hatten, 1990; Fishell

and Hatten, 1991; Zheng et al., 1996). Because the same

neuron can migrate in both directions along the same glial

fiber (Edmondson et al., 1987; Gregory et al., 1988; Hatten

and Liem, 1981; Mason et al., 1988), the glial fibers

unlikely provide information to guide the direction of

neuronal migration under these conditions. In our studies

of EGL cell migration, the three dimensional matrices are

known to contain extracellular matrix proteins (Kleinman et

al., 1982). Neurons migrating from URL or EGL explants


were not associated with glial fibers. When neurons were

attracted or repelled by the meninges, they were not asso-

ciated with glial fibers. The glial fibers therefore are not

required in mediating EGL responses to the guidance cues,

which must act directly on neurons.

The conclusions reached from in vitro studies also agree

with in vivo results. Altman and Bayer (1997, pp. 341–347)

reported that migrating EGL cells appeared to move sepa-

rately in vivo and were not associated with glial or neuronal

fibers in postnatal rats of 5, 7 or 10 days of age. After P12,

neurons seem to migrate along the fibers of other neurons in

vivo and along glial fibers later (Altman and Bayer, 1997).

Recently, Nadarajah et al. (2001, 2002) report that, while

neuronal precursor cells in the rat neocortex after E15

prevalently move along glial fibers, early migrating neurons

are not associated with glial fibers.

Our results indicate that the presence of glial fibers is not

absolutely required for EGL neurons to migrate. In vitro

studies of live granule cells indicate that, when glass fibers

were coated with extracellular matrix proteins such as

laminin or fibronectin, granule cells can migrate along the

coated glass fibers (Fishman and Hatten, 1993). This sug-

gests that deposition of extracellular matrix proteins is

sufficient to replace glial fibers. Under our culture condi-

tions, migrating neurons from E15 URL and postnatal EGL

explants appear to be associated with tracks of neuronal

fibers. Although neurons from E17 EGL explants have

leading processes, cell bodies of individual neurons of

E17 explants were not associated with tracks of fibers of

other neurons. At least in the case of E17 neurons, the effect

of guidance cues is unlikely to be mediated by their action

on tracks of neuronal fibers.

The precise roles of the glial fibers are thus not very

clear. Are they facilitating neurons to move faster? Is the

association of neurons and glial fibers used for coordinating

their differentiation than for migration? Are they limiting

neurons to migrate along one dimension among the three

dimensional space? It seems that ablation of specific glial

fibers along which neuronal migration could be traced in

live brain slices is one way to investigate the precise role of

glial fibers.

A related question is the subcellular structure in a single

migrating neuron that senses the guidance cues: is it the cell

body or the leading process? To answer this question, one

may have to develop a method to specifically deliver the

guidance cue to the cell body without affecting the leading

process or to the tip of the leading process without affecting

the cell body.

Acknowledgments

We are grateful to Huashun Li for participating in the

early phase of this project; to Jennifer Jones for assistance;

to Jane E. Johnson for the anti-Math1 antibody; to Guofa

Liu, Harumi Saito and Yong Yin for help; to Alan Pearlman,

Joshua Sanes, Joseph Price, Huashun Li and Bagi Nadaraja

for helpful discussions; to the NIH, the Klingenstein

foundation, and the John Merck fund for support (to Y. R.).

References

Alder, J., Cho, N.K., Hatten, M.E., 1996. Embryonic precursor cells from

the rhombic lip are specified to a cerebellar granule neuron identity.

Neuron 17, 389–399.

Altman, J., Bayer, S.A., 1997. Development of the Cerebellar System. CRC

Press, New York.

Angelov, D.N., Vasilev, V.A., 1989. Morphogenesis of rat cranial meninges.

A light- and electron-microscopic study. Cell Tissue Res. 257, 207–216.

Bhatt, R.S., Tomoda, T., Fang, Y., Hatten, M.E., 2000. Discoidin domain

receptor 1 functions in axon extension of cerebellar granule neurons.

Genes Dev. 14, 2216–2228.

Chae, T., Kwon, Y.T., Bronson, R., Dikkes, P., Li, E., Tsai, L.H., 1997.

Mice lacking p35, a neuronal specific activator of Cdk5, display cortical

lamination defects, seizures, and adult lethality. Neuron 18, 29–42.

D’Arcangelo, G., Miao, G.G., Chen, S.-C., Soares, H.D., Morgan, J.I.,

Curran, T., 1995. A protein related to extracellular matrix proteins

deleted in the mouse mutant reeler. Nature 374, 719–723.

D’Arcangelo, G., Homayouni, R., Keshvara, L., Rice, D.S., Sheldon, M.,

Curran, T., 1999. Reelin is a ligand for lipoprotein receptors. Neuron

24, 471–479.

Dhavan, R., Tsai, L.H., 2001. A decade of cdk5. Nat. Rev., Mol. Cell Biol.

2, 749–759.

Edmondson, J.C., Hatten, M.E., 1987. Glial-guided granule neuron migra-

tion in vitro: a high-resolution time-lapse video microscopic study.

J. Neurosci. 7, 1928–1934.

Edmondson, J.C., Liem, R.K., Kuster, J.E., Hatten, M.E., 1988. Astro-

tactin: a novel neuronal cell surface antigen that mediates neuron–

astroglial interactions in cerebellar microcultures. J. Cell Biol. 106,

505–517.

Fishell, G., Hatten, M.E., 1991. Astrotactin provides a receptor system for

CNS neuronal migration. Development 113, 755–765.

Fishman, R.B., Hatten, M.E., 1993. Multiple receptor systems promote

CNS neural migration. J. Neurosci. 13, 3485–3495.

Gregory, W.A., Edmondson, J.C., Hatten, M.E., Mason, C.A., 1988. Cy-

tology and neuron–glial apposition of migrating cerebellar granule cells

in vitro. J. Neurosci. 8, 1728–1934.

Hatten, M.E., 1999. Central nervous system neuronal migration. Annu.

Rev. Neurosci. 22, 511–539.

Hatten, M.E., Liem, R.H.K., 1981. Astroglia provide a template for the

positioning of developing cerebellar neurons in vitro. J. Cell Biol. 90,

622–630.

Hatten, M.E., Mason, C.A., 1990. Mechanisms of glial-guided neuronal

migration in vitro and in vivo. Experientia 46, 907–916.

Hatten, M.E., Liem, R.K., Mason, C.A., 1984. Defects in specific associ-

ations between astroglia and neurons occur in microcultures of weaver

mouse cerebellar cells. J. Neurosci. 4, 1163–1172.

Hiesberger, T., Trommsdorff, M., Howell, B.W., Goffinet, A., Mumby,

M.C., Cooper, J.A., Herz, J., 1999. Direct binding of Reelin to VLDL

receptor and ApoE receptor 2 induces tyrosine phosphorylation of the

adaptor protein Disabled-1 and modulates tau phosphorylation. Neuron

24, 481–489.

Hopfield, S., McKay, R.D., 1985. Identification of major cell classes in the

developing mammalian nervous system. J. Neurosci. 5, 3310–3328.

Howell, B.W., Hawkes, R., Soriano, P., Cooper, J.A., 1997. Neuronal

position in the developing brain is regulated by mouse disabled-1.

Nature 389, 733–737.

Kamiryo, T., Orita, T., Nishizaki, T., Aoki, H., 1990. Development of the

rat meninx: experimental study using Bromodeoxyuridine. Anat. Rec.

227, 207–210.


Klein, R.S., Rubin, J.B., Gibson, H.D., DeHaan, E.N., Alvarez-Hernandez,

X., Segal, R.A., Luster, A.D., 2001. SDF-1a induces chemotaxis and

enhances Sonic hedgehog-induced proliferation of cerebellar granule

cells. Development 128, 1971–1981.

Kleinman, H.K., McGravey, M.L., Liotta, L.A., Robey, P.G., Tryggvason,

K., Martin, G.R., 1982. Isolation and characterization of type IV pro-

collagen, laminin, and heparan sulfate proteoglycan from the EHS sar-

coma. Biochemistry 21, 6188–6193.

Letinic, K., Rakic, P., 2001. Telencephalic origin of human thalamic GA-

BAergic neurons. Nat. Neurosci. 4, 931–936.

Li, H.S., Chen, J.H., Wu, W., Fagaly, T., Zhou, L., Yuan, W., Dupuis,

S., Jiang, Z.H., Nash, W., Gick, C., Ornitz, D.M., Wu, J.Y., Rao,

Y., 1999. Vertebrate slit, a secreted ligand for the transmembrane

protein roundabout, is a repellent for olfactory bulb axons. Cell 96,

807–818.

Lu, Q., Sun, E., Klein, R.S., Flanagan, J.G., 2001. Ephrin-B reverse signal-

ing is mediated by a novel PDZ-RGS protein and selectively inhibits G

protein-coupled chemoattraction. Cell 105, 69–79.

Ma, Q., Jones, D., Borghesani, P.R., Segal, R.A., Nagasawa, T., Kishi-

moto, T., Bronson, R.T., Springer, T.A., 1998. Impaired B-lymphopoi-

esis, myelopoiesis, and derailed cerebellar neuron migration in

CXCR4- and SDF-1-deficient mice. Proc. Natl. Acad. Sci. U. S. A.

95, 9448–9453.

Marin, O., Yaron, A., Bagri, A., Tessier-Lavigne, M., Rubenstein, J.L.,

2001. Sorting of striatal and cortical interneurons regulated by sema-

phorin–neuropilin interactions. Science 293, 872–875.

Mason, C.A., Edmondson, J.C., Hatten, M.E., 1988. The extending astro-

glial process: development of glial cell shape, the growing tip, and

interactions with neurons. J. Neurosci. 8, 3124–3134.

Ming, G.-L., Song, H.-J., Berninger, B., Holt, C., Tessier-Lavigne, M., Poo,

M.-M., 1997. cAMP-dependent growth cone guidance by netrin-1.

Neuron 19, 1225–1235.

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

A.L., 2001. Two modes 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.

Nikolic, M., Dudek, H., Kwon, Y.T., Ramos, Y.F., Tsai, L.H., 1996. The

cdk5/p35 kinase is essential for neurite outgrowth during neuronal dif-

ferentiation. Genes Dev. 10, 816–825.

O’Rahilly, R., Mueller, F., 1986. The meninges in human development. J.

Neuropathol. Exp. Neurol. 45, 588–608.

Pearlman, A.L., Faust, P.L., Hatten, M.E., Brunstrom, J.E., 1998. New

directions for neuronal migration. Curr. Opin. Neurobiol. 8, 45–54.

Pleasure, S.J., Anderson, S., Hevner, R., Bagri, A., Marin, O., Lowenstein,

D.H., Rubenstein, J.L., 2000. Cell migration from the ganglionic emi-

nences is required for the development of hippocampal GABAergic

interneurons. Neuron 28, 727–740.

Polleux, F., Morrow, T., Ghosh, A., 2000. Semaphorin 3A is a chemo-

attractant for cortical apical dendrites. Nature 404, 567–573.

Powell, E.M., Mars, W.M., Levitt, P., 2001. Hepatocyte growth factor/

scatter factor is a motogen for interneurons migrating from the ventral

to dorsal telencephalon. Neuron 30, 79–89.

Poznansky, M.C., Olszak, I.T., Foxall, R., Evans, R.H., Luster, A.D., Scad-

den, D.T., 2000. Active movement of T cells away from a chemokine.

Nat. Med. 6, 543–548.

Rakic, P., 1971a. Neuron–glia relationship during granule cell migration in

developing cerebellar cortex. J. Comp. Neurol. 141, 283–312.

Rakic, P., 1971b. Guidance of neurons migrating to the fetal monkey neo-

cortex. Brain Res. 33, 471–476.

Rakic, P., 1990. Principles of neural cell migration. Experientia 46,

882–891.

Rice, D.S., Curran, T., 1999. Mutant mice with scrambled brains: under-

standing the signaling pathways that control cell positioning in the

CNS. Genes Dev. 13, 2758–2773.

Rice, D.S., Curran, T., 2001. Role of the Reelin signaling pathway in

central nervous system development. Annu. Rev. Neurosci. 24,

1005–1039.

Ross, M.E., Walsh, C.A., 2001. Human brain malformations and their

lessons for neuronal migration. Annu. Rev. Neurosci. 24, 1041–1070.

Sheldon, M., Rice, D.S., D’Arcangelo, G., Yoneshima, H., Nakajima, K.,

Mikoshiba, K., Howell, B.W., Cooper, J.A., Goldowitz, D., Curran, T.,

1997. Scrambler and yotari disrupt the disabled gene and produce a

reeler-like phenotype in mice. Nature 389, 730–733.

Song, H.-J., Ming, G.-L., Poo, M.-M., 1997. A cAMP-induced switching

of turning direction of nerve growth cones. Nature 388, 275–279.

Song, H.-J., Ming, G.-L., He, Z., Lehmann, M., McKerracher, L., Tessier-

Lavigne, M., Poo, M.-M., 1998. Conversion of neuronal growth cone

responses from repulsion to attraction by cyclic nucleotides. Science

281, 1515–1518.

Sotelo, C., Alvarado-Mallart, R.M., Frain, M., Vernet, M., 1994. Molecular

plasticity of adult Bergmann fibers is associated with radial migration of

grafted Purkinje cells. J. Neurosci. 14, 124–133.

Stitt, T.N., Hatten, M.E., 1990. Antibodies that recognize astrotactin block

granule neuron binding to astroglia. Neuron 5, 639–649.

Sweet, H.O., Bronson, R.T., Johnson, K.R., Cook, S.A., Davisson,

M.T., 1996. Scrambler, a new neurological mutation of the mouse

with abnormalities of neuronal migration. Mamm. Genome 7,

798–802.

Tachibana, K., Hirota, S., Iizasa, H., Yoshida, H., Kawabata, K., Kataoka,

Y., Kitamura, Y., Matsushima, K., Yoshida, N., Nishikawa, S., Kishi-

moto, T., Nagasawa, T., 1998. The chemokine receptor CXCR4 is

essential for vascularization of the gastrointestinal tract. Nature 393,

591–594.

Trommsdorff, M., Gotthardt, M., Hiesberger, T., Shelton, J., Stockinger,

W., Nimpf, J., Hammer, R.E., Richardson, J.A., Herz, J., 1999. Reeler/

disabled-like disruption of neuronal migration in knock-out mice lacking

the VLDL receptor and ApoE receptor 2. Cell 97, 689–701.

Tsai, L.H., Delalle, I., Caviness Jr., V.S., Chae, T., Harlow, E., 1994. p35 is

a neural-specific regulatory subunit of cyclin-dependent kinase 5. Na-

ture 371, 419–423.

Ware, M.L., Fox, J.W., Gonzale, J.L., Davis, N.M., Lambert de Rouvroit,

C., Russo, C.J., Chua Jr., S.C., Goffinet, A.M., Walsh, C.A., 1997.

Aberrant splicing of a mouse disabled homolog, mdab1, in the scram-

bler mouse. Neuron 19, 239–249.

Wechsler-Reya, R.J., Scott, M.P., 1999. Control of neuronal precursor pro-

liferation in the cerebellum by sonic Hedgehog. Neuron 22, 103–114.

Wingate, R.J.T., Hatten, M.E., 1999. The role of the rhombic lip in avian

cerebellum development. Development 126, 4395–4404.

Wong, K., Ren, X.R., Huang, Y.Z., Xie, Y., Liu, G., Saito, H., Tang, H.,

Wen, L., Brady-Kalnay, S.M., Mei, L., Wu, J.Y., Xiong, W.C., Rao, Y.,

2001. Signal transduction in neuronal migration: roles of GTPase acti-

vating proteins and the small GTPase Cdc42 in the Slit–Robo pathway.

Cell 107, 209–221.

Wu, W., Wong, K., Chen, J.H., Jiang, Z.H., Dupuis, S., Wu, J.Y., Rao, Y.,

1999. Directional guidance of neuronal Migration in the olfactory sys-

tem by the secreted protein Slit. Nature 400, 331–336.

Yee, K.T., Simon, H.H., Tessier-Lavigne, M., O’Leary, D.D.M., 1999.

Extension of long leading processes and neuronal migration in the

mammalian brain directed by the chemoattractant Netrin-1. Neuron

24, 607–622.

Zheng, C., Heintz, N., Hatten, M.E., 1996. CNS gene encoding astrotactin,

which supports neuronal migration along glial fibers. Science 272,

417–419.

Zhu, Y., Li, H.S., Zhou, L., Wu, J.Y., Rao, Y., 1999. Cellular and molecular

guidance of GABAergic neuronal migration from the striatum to the

neocortex. Neuron 23, 473–485.

Zhu, Y., Yu, T., Zhang, X.-C., Nagasawa, T., Wu, J.Y., Rao, Y., 2002. Role

of the chemokine SDF-1 as the meningeal attractant for embryonic

cerebellar neurons. Nat. Neurosci. 5, 719–720.

Zou, Y.R., Kottmann, A.H., Kuroda, M., Taniuchi, I., Littman, D.R., 1998.

Function of the chemokine receptor CXCR4 in haematopoiesis and in

cerebellar development. Nature 393, 595–599.

Documents

Temporal regulation of cerebellar EGL migration through a switch in cellular responsiveness to the meninges