Int. J. Devl Neuroscience 28 (2010) 611–620

Contents lists available at ScienceDirect

International Journal of Developmental Neuroscience

journa l homepage: www.e lsev ier .com/ locate / i jdevneu

Glial cells modulate heparan sulfate proteoglycan (HSPG) expression byneuronal precursors during early postnatal cerebellar development

Ana Paula B. Araujo a,b, Maria Emília O.B. Ribeiro a, Ritchelli Ricci a, Ricardo J. Torquato a,Leny Toma a, Marimélia A. Porcionatto a,∗

a Departamento de Bioquímica, Universidade Federal de São Paulo, Rua Três de Maio, 100, 04044­020 São Paulo, SP, Brazilb Investiga Institutos de Pesquisa, Avenida Dr. Romeu Tórtima, 452, Barão Geraldo, 13084­791 Campinas, SP, Brazil

a r t i c l e i n f o

Article history:

Received 21 May 2010

Received in revised form 24 June 2010

Accepted 9 July 2010

Keywords:

Cerebellum

Glypican

Syndecan

Runx

Granule cell precursor

a b s t r a c t

Cerebellum controls motor coordination, balance, eye movement, and has been implicated in memory

and addiction. As in other parts of the CNS, correct embryonic and postnatal development of the cere­

bellum is crucial for adequate performance in the adult. Cellular and molecular defects during cerebellar

development can lead to severe phenotypes, such as ataxias and tumors. Knowing how the correct devel­

opment occurs can shed light into the mechanisms of disease. Heparan sulfate proteoglycans are complex

molecules present in every higher eukaryotic cells and changes in their level of expression as well as in

their structure lead to drastic functional alterations. This work aimed to investigate changes in heparan

sulfate proteoglycans expression during cerebellar development that could unveil control mechanisms.

Using real time RT­PCR we evaluated the expression of syndecans, glypicans and modifying enzymes

by isolated cerebellar granule cell precursors, and studied the influence of soluble glial factors on the

expression of those genes. We evaluated the possible involvement of Runx transcription factors in the

response of granule cell precursors to glial factors. Our data show for the first time that cerebellar granule

cell precursors express members of the Runx family and that the expression of those genes can also be

controlled by glial factors. Our results also show that the expression of all genes studied vary during

postnatal development and treatment of precursors with glial factors indicate that the expression of

heparan sulfate proteoglycan genes as well as genes encoding heparan sulfate modifying enzymes can

be modulated by the microenvironment, reflecting the intricate relations between neuron and glia.

© 2010 ISDN. Published by Elsevier Ltd. All rights reserved.

1. Introduction

During normal cerebellar development neuronal precursors

undergo proliferation, apoptosis, migration and differentiation, cel­

lular events orchestrated by the sum of intrinsic gene expression

programs and extrinsic informative cues provided by surrounding

cells (Hatten and Heintz, 2005). Defects in one or more of steps of

these developmental events can originate severe phenotypes such

as ataxias, intellectual disability and pediatric tumors (Gulino et

al., 2008; Steinlin, 2008; Behesti and Marino, 2009; Vaillant and

Monard, 2009).

Proteoglycans are complex molecules composed of sulfated

glycosaminoglycan chains covalently linked to a core protein. Pro­

teoglycans are divided in distinct classes based on the sugar chain

∗ Corresponding author. Tel.: +55 11 5579 3175; fax: +55 11 5579 3175.

E­mail addresses: apbergamo@gmail.com (A.P.B. Araujo), ebrenha@gmail.com

(M.E.O.B. Ribeiro), ritchelli@gmail.com (R. Ricci), torquato.bioq@epm.br

(R.J. Torquato), leny.bioq@epm.br (L. Toma), marimelia.bioq@epm.br,

marimelia.porcionatto@gmail.com (M.A. Porcionatto).

composition and amino acid sequence of the core protein. Heparan,

chondroitin, and keratan sulfate proteoglycans are components of

the CNS extracellular matrix, and are also present at neuronal and

glial plasma membranes. In the CNS, proteoglycans are involved

in many different events such as proliferation, migration, differen­

tiation, axonal outgrowth, and synapse formation (Pearlman and

Sheppard, 1996; Kroger and Schroder, 2002; Porcionatto, 2006;

Busch and Silver, 2007; Galtrey and Fawcett, 2007). Among all hep­

aran sulfate proteoglycans, members of two families are highly

expressed in the brain, syndecans and glypicans. Both families are

localized at the cell surface by their core protein, either as inte­

gral proteins, in the case of syndecans (Gallagher et al., 1990) or

GPI­anchored proteins, in the case of glypicans (Fransson et al.,

2004). During postnatal cerebellar development, heparan sulfate

proteoglycans are critical for granule cell precursors proliferation

induced by sonic hedgehog (SHH) (Rubin et al., 2002; Chan et al.,

2009). A few studies have shown that glypican 3 controls cell prolif­

eration promoted by hedgehog (HH) indicated by the overgrowth

of glypican 3 null mice (Capurro et al., 2008), and there are only

a few studies about the involvement of syndecans in cerebellar

development.

0736­5748/$36.00 © 2010 ISDN. Published by Elsevier Ltd. All rights reserved.

doi:10.1016/j.ijdevneu.2010.07.228

612 A.P.B. Araujo et al. / Int. J. Devl Neuroscience 28 (2010) 611–620

Heparan sulfate proteoglycans are present in every cell type

of organisms with tissue organization (Dietrich et al., 1998), and

syndecans and glypicans have been implicated in a great number

of biological processes, from proliferation to differentiation and

migration of several cell types. The fine structure of the heparan

sulfate chains, mainly the sulfation pattern, is known to be modi­

fied in response to external stimuli [for a review, see Sugahara and

Kitagawa, 2002]. The modifications lead to the complex structure

of the polysaccharide chain which exhibits a considerable number

of sequences with specific sulfation profiles. Those sequences are

recognized by specific proteins that belong to a group known as

“heparin­binding proteins”, and in turn, the sugar sequences sup­

port unique functions of the respective proteins through specific

interactions.

Based on those characteristics of heparan sulfate proteoglycans,

we wanted to know if the expression of the two major heparan

sulfate proteoglycan families present in the brain, syndecan and

glypican, is regulated during early postnatal cerebellar develop­

ment. We also investigated if factors secreted by glial cells could

modulate the expression of those proteoglycans. In order to achieve

our goal, we have used real time RT­PCR to analyze the expression

of the members of the syndecan and glypican families as well as

of the main enzymes that modify the polysaccharide chains, the

sulfotransferases and the extracellular sulfatases. Primary cultures

of neuronal precursors were used to investigate both the develop­

mental regulation of gene expression and the possible influence

of glial secreted factors on the expression of syndecans, glypicans,

sulfotransferases and sulfatases.

2. Experimental procedures

2.1. Animals

C57Bl/10 mice used in this study were maintained in ventilated microisolators,

under a cycle of light/dark (12 h/12 h), with free access to water and food. Pups were

kept with the mother until euthanasia. All experimental protocols and handling of

the animals were approved by the Committee for Ethics in Research of Universidade

Federal de São Paulo (CEP 0445/05).

2.2. Isolation of cerebellar granule cell precursors

Granule cell precursors were isolated from neonatal mouse cerebellum as previ­

ously described (Klein et al., 2001; Zhou et al., 2007). Briefly, cerebella were dissected

and meninges were removed. After incubation with 0.1% (v/v) Trypsin in Hank’s solu­

tion (EBSS, 36 mM glucose, 15 mM HEPES) with 125 U/ml DNase for 20 min at 37 ◦C,

enzymatic activity was stopped by the addition of DMEM/10% fetal bovine serum

(FBS; Cultilab, Brazil). Tissue fragments were collected using a clinical centrifuge

by spinning the cell suspension for 4 min at 2370 × g. Cell pellets were washed

three times with Hank’s solution and centrifuged for 4 min at 2370 × g. The final

cell suspension was plated in 35 mm tissue culture dishes for 30 min to remove

adherent glial cells. The cell suspension containing granule cell precursors was plat­

ted at a density of 106 cells/35 mm dish on poly­l­lysine coated dishes. Cells were

maintained in DMEM/F12 1:1, containing 20 mM KCl, 36 mM glucose and peni­

cillin/streptomycin, at 37 ◦C in a humidified incubator with 5% CO2 overnight. For

real time RT­PCR analysis of gene expression during early development, cells were

harvested and RNA was extracted after the overnight incubation period. For treat­

ment of cells with glial conditioned medium, the culture medium was removed after

the overnight incubation and replaced by the conditioned medium. Cells were then

incubated for an additional 24 h period, after which cells were harvested and RNA

was extracted for real time RT­PCR analysis.

2.3. Conditioned medium from cerebellar glial cell cultures

Adherent glial cells obtained from cerebella were maintained in DMEM/F12 1:1,

20 mM KCl, 36 mM glucose and penicillin/streptomycin at 37 ◦C in a humidified incu­

bator with 5% CO2 for 24 h. After this period of time, culture medium was aspirated

to remove non­adherent granular precursors and dead cells in suspension. Fresh

medium was added and maintained for 3 days. Glial conditioned medium was recov­

ered after this period and filtered to remove any debris and added to the granule

cell precursors.

2.4. Evaluation of axonal outgrowth

DIC images of granule cell precursors cultured in the presence of control medium

or glial conditioned medium were acquired using a Axyovert100M Microscope (Carl

Zeiss, Oberkochen, Germany). The size of the processes was measured using tools

from the Software LSM 510 Meta.

2.5. Quantitative RT­PCR

Total RNA was Trizol® (Invitrogen, USA) extracted from granule cell precur­

sors isolated from postnatal day 3 (P3), P6 and P9 mice cerebella. Seven mice at

each developmental stage were used for RNA extraction and RNA was quantified

using NanoDrop ND­1000 Spectrophotometer (Thermo Fisher Scientific, USA). Two

micrograms of total RNA were reverse­transcribed with Oligo(dT)15 Primer and

ImProm­II Reverse Transcription System with Recombinant RNasin® Ribonuclease

inhibitor (Promega, USA) and MgCl2 . The primer sequences were verified to be spe­

cific using GenBank’s BLAST (Altschul et al., 1997). The specifications of primers

are given in Table 1 (Ford­Perriss et al., 2003; Yabe et al., 2005); primers that were

synthesized according to previously published papers are indicated in the table,

all the other primers were designed by us. Quantitative real time RT­PCR was per­

formed using SYBR®­Green PCR Master Mix, including AmpliTaq­GOLD polymerase

(Applied Biosystems, USA). Reactions were performed on ABI PRISM 7500 real time

PCR System (Applied Biosystems). The relative expression levels of genes were calcu­

lated using the 2−11CT method (Livak and Schmittgen, 2001). The amount of target

genes expressed in a sample was normalized to the average of the endogenous

control. This is given by 1CT , where 1CT is determined by subtracting the average

endogenous gene CT value from the average target gene CT value [CT target gene − CT

average (endogenous gene)]; where 2−1CT is the relative expression of the target

gene compared to the endogenous gene. The calculation of 11CT was done by sub­

tracting 1CT value for the controls from the 1CT value for a given developmental

age [1CT target gene(treated) − 1CT target gene(control)]; where 2−11CT is the relative

expression of the target gene at different developmental ages compared to controls.

In Figs. 1–3 and 8, “relative expression” is the value of 2−1CT and in Figs. 5–7 and 9,

“relative expression” is the value of 2−11CT . Three different genes were tested to be

used as endogenous control: b­actin, GAPDH, and HPRT. Expression levels of b­actin

and GAPDH by granule precursors did not change over the postnatal ages, whereas

HPRT expression was lower than the two other genes, but no temporal difference

was observed (Supplementary Fig. S1). Based on this result, we chose to use b­actin

as the endogenous gene.

2.6. Statistical analysis

One­way analysis of variance (ANOVA) followed by Dunnett’s test was used to

evaluate differences among ages. Differences between groups of genes were eval­

uated by two­way (ANOVA) followed by Bonferroni post­tests using the GraphPad

Prism software version 5.01 (GraphPad Software, USA). Results are expressed as

mean ± error and were considered to be significant if p < 0.05.

3. Results

In order to investigate if heparan sulfate proteoglycan expres­

sion by neuronal precursors is regulated during early postnatal

cerebellar development we first analyzed the expression of genes

encoding the core protein of members of the syndecan and glyp­

ican families, the two major heparan sulfate proteoglycans found

in brain, as well as the heparan sulfate chain modifying enzymes,

sulfotransferases and sulfatases.

3.1. Expression of syndecans and glypicans by neuronal

precursors varies during early postnatal cerebellar development

Expression of syndecan and glypican family members by iso­

lated cerebellar granule cell precursors have shown that the

expression of all members of the two heparan sulfate proteoglycans

is regulated during early postnatal development. While syndecan 1

expression decreased from postnatal day 3 (P3) to P6–9, the expres­

sion of the other three members of the syndecan family, syndecans

2, 3 and 4, increased from P3 to P6–9 (Fig. 1).

The expression of the members of the glypican family also var­

ied over time. Glypican 1 was the only member of the family to

show increased expression from P3–6 to P9. Expression of glypi­

cans 2, 3, 5 and 6 decreased from P3 to P6–9, whereas the expression

of glypican 4 did not change during early postnatal development

(Fig. 1).

A.P.B. Araujo et al. / Int. J. Devl Neuroscience 28 (2010) 611–620 613

Table 1

Specifications of the primers used in real time RT­PCR.

Gene (Acc No)a Sequence Tmb Product (bp) Reference

Syndecan­1 (NM 011519) F: GAC TCT GAC AAC TTC TCT GGC

R: GCT GTG GTG ACT CTG ACT GTT G

58 ◦C 319 Yabe et al. (2005)

Syndecan­2 (NM 008304) F: CAG GAG CTG ATG AAG ACA TAG AGA

R: ATG AGG AAA ATG GCA AAG AGA A

55 ◦C 324 Yabe et al. (2005)

Syndecan­3 (NM 011520) F: TCG TTT CCT GAT GAT GAA CTA GAC

R: GTG CTG GAC ATG GAT ACT TTG TT

55 ◦C 301 Yabe et al. (2005)

Syndecan­4 (NM 011521) F: AGA GCC CAA GGA ACT GGA AGA GAA

R: ATC AGA GCT GCC AAG ACC TCA GTT

60 ◦C 147

Glypican­1 (NM 016696) F: ACT CCA TGG TGC TCA TCA CTG ACA

R: TTT CCA CAG GCC TGG ATG ACC TTA

60 ◦C 151

Glypican­2 (NM 172412) F: TCT TTG GCT CAG CTC TTC TCG CAT

R: ACT GTA TTG TGG GTG CAG CAA AGG

60 ◦C 189

Glypican­3 (NM 016697) F: AAT GAT ACC CTG TGC TGG AAC GGA

R: AGA ACT TAC CCT TGG GCA CAG ACA

60 ◦C 199

Glypican­4 (NM 008150) F: ACA ACC CAG AAG TCC AGG TTG ACA

R: ACT CCG AAG GGC ACT GCT GAT ATT

60 ◦C 195

Glypican­5 (NM 175500) F: CAC GTG CTC CTG AAC TTC CAC TTG

R: AAC ACA AAG CTG GTC AGC CAG GCC

60 ◦C 282 Yabe et al. (2005)

Glypican­6 (AF105268) F: GTC CGG ACC TAC GGG ATG CTG TAC

R: TCT TCC ATT TCT GTG GTG CAG CAGG

60 ◦C 199

2­O­Sulfotrasferase­1 (NM 011828) F: AAT CTT CTC CTG GTG CAG GGA CAT

R: AGA AGG GTA AGT GGC AAA CGGACT

55 ◦C 158

3­O­Sulfotransferase­1 (NM 010474) F: AGA TTC CTG AAG CTT TCT CCA C

R: ATT TGG CTC ATG AAA GTA TTC GT

60 ◦C 160

3­O­Sulfotransferase­2 (NM 001081327) F: CAA GAG ATT CAT CAC AGA CAA GC

R: GTC TAT AAA ATT CCC GGA GCT G

60 ◦C 170 Ford­Perriss et al. (2003)

3­O­Sulfotransferase­3 (NM 018805) F: CTT CTA CTT CAA CCA GAC CAA GG

R: ATC TGG TAA AAC TTG CGG TTG

60 ◦C 183

3­O­Sulfotransferase­4 (NM 006040) F: CTA CTC AGA GAT GGC ATT CAT TG

R: GAT TCA ACT CTC CCC ACT CAT AC

60 ◦C 192

3­O­Sulfotransferase­5 (NM 001081208) F: TGT TGA TCA TTG TCA GGG AGC CGA

R: ATA TGC TGG TCC TAA CCG CCT TGT

60 ◦C 168

6­O­Sulfotransferase­1 (NM 015818) F: CGC CCA GAA AGT TCT ACT ACA TC

R: GGT TGT TAG CCA GGT TAT AGG G

60 ◦C 203

6­O­Sulfotransferase­2 (NM 015819) F: ACC TCT TTC TGC AAA GGT ATC AG

R: GTT CTG ATT TGG GTT AGG ATT TG

60 ◦C 174

6­O­Sulfotransferase­3 (NM 015820) F: GGA CAT GCA GCT TTA TGA GTA TG

R: GCT GTT GTA GTC CTC AGT GAC AG

60 ◦C 190 Ford­Perriss et al. (2003)

N­Deacetylase­N­Sulfotransferase­1 (NM 008306) F: TCC TAT CAG CAT CTT CAT GAC AC

R: TCT TCT CCT TAG ACC AGA TGT CC

55 ◦C 236 Ford­Perriss et al. (2003)

N­Deacetylase­N­Sulfotransferase­2 (NM 010811) F: AAA AGT GCC ACC TAC TTT GAC TC

R: AGC TGA AAT CAC TTG GTA GAA GG

60 ◦C 160

N­Deacetylase­N­Sulfotransferase­3 (NM 031186) F: TTA TGC ATC CTT CCA TCC TTA GT

R: AGT ATG GTG ATG ATC TTG GCT TT

60 ◦C 212

N­Deacetylase­N­Sulfotransferase­4 (NM 022565) F: GGA GAA AAC CTG TGA CCA TTT AC

R: CCT TGT GAT AGT TGT TGC CAT TA

60 ◦C 147

mSulfatase­1 (NM 172294) F: TGC TGA ACA GTC ACC CTG ATC CAA

R: TCA GAT GCA GGG TTT GGA GGT TGA

60 ◦C 195

mSulfatase­2 (NM 028072) F: TCA AAG TGA CCC ATC GGT GCT ACA

R: AGT CAC ATT CTT CCG GTC GCT TCT

60 ◦C 195

RunX 1 (NM 001111021) F: AGA TGG ACG GCA GAG TAG GGA ACT

R: TGA TGA CAC TGC CAC CTC TGA CTT

60 ◦C 100

RunX 2 (NM 001145920) F: TGA TGA CAC TGC CAC CTC TGA CTT

R: AGG GAT GAA ATG CTT GGG AAC TGC

60 ◦C 123

RunX 3 (NM 019732) F: TGG CCG GCA ATG ATG AGA ACT ACT

R: TTG GTG AAC ACG GTG ATT GTG AGC

60 ◦C 148

HPRT F: CTC ATG GAC TGA TTA TGG ACA GGA C

R: GCA GGT CAG CAA AGA ACT TAT AGC

58 ◦C 180

GAPDH F: AAG AAG GTG GTG AAG CAG GCA TCT

R: ACC CTG TTG CTG TAG CCG TAT TCA

60 ◦C 200

b­Actin F: ACT CTT CCA GCC TTC CTT C

R: ATC TCC TTC TGC ATC GTG TC

55 ◦C 200

a GenBank accession numbers are listed in parenthesis.b Tm, predicted melting temperature.

3.2. Expression of heparan sulfate sulfotransferases and

extracellular sulfatases is also regulated during development

The analysis of heparan sulfate O­sulfotransferases expres­

sion revealed that most of them are downregulated dur­

ing early development (3­O­sulfotransferases 1, 3, 4, 5; 6­

O­sulfotransferases 1 and 2; N­deacetylase­N­sulfotransferase

2), whereas the expression of 2­O­sulfotransferase 1, 6­O­

sulfotransferase 3 and N­deacetylase­N­sulfotransferases 1, 3

and 4 increased from P3 to P6–9, and of 3­O­sulfotransferase

2 did not change over time (Fig. 2). The expression of the

two extracellular sulfatases analyzed showed similar profile,

with decreased expression from P3 to P9 for both of them

(Fig. 3).

614 A.P.B. Araujo et al. / Int. J. Devl Neuroscience 28 (2010) 611–620

Fig. 1. The expression of glypicans and syndecans by neuronal precursors changes

during early postnatal cerebellar development. The expression level of each gene

was evaluated by real time RT­PCR in samples obtained from granule cell precur­

sors isolated from cerebellum of mice at different postnatal ages (3, 6 and 9 days).

Values were normalized to that of b­actin transcript. Data were obtained in triplicate

and are given as the mean ± error. One­way analysis of variance (ANOVA) followed

by Dunnett’s test was used to evaluate differences among ages. Differences were

considered statistically significant if p < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001).

The gene expression data presented so far has shown that the

core proteins of the two families of heparan sulfate proteoglycan,

syndecan and glypican, as well as the modifying enzymes involved

in post­translational and post­secretion modifications of the hep­

aran sulfate chains is regulated in neuronal precursors during early

postnatal cerebellar development.

3.3. Soluble factors secreted by glial cells modulate axonal

outgrowth and gene expression by cerebellar neuronal precursors

We then asked if glial cells would interfere with the expression

pattern of these genes by the means of secreted factors. To inves­

tigate that we first analyzed the effects of glial secreted factors on

axonal outgrowth by neuronal precursors in vitro. Using glial condi­

tioned medium we were able to observe that neuronal precursors

cultured in the presence of culture medium conditioned by glial

cells extended longer axons when compared to precursors cultured

in control medium (Fig. 4). We then investigated if the glial solu­

ble factors would also modulate gene expression of genes encoding

syndecan, glypican and the heparan sulfate modifying enzymes.

Glial conditioned medium induced increased expression of all

four syndecans by P3 neuronal precursors without affecting their

expression by P9 precursors, whereas the expression of all six

glypicans was decreased in both ages by the action of the sol­

uble factors (Fig. 5). On the other hand, the expression of the

sulfotransferases was affected in different ways, depending on the

enzyme and age of cells. We observed stimulation of expression

of 3­O­sulfotransferases 2, 3 and 4, and decreased expression of 6­

O­sulfotransferase 3, N­acetylase­N­sulfotransferases 3 and 4 when

cells were cultured in glial conditioned medium (Fig. 6). The expres­

sion of the extracellular sulfatases mSulfatase 1 and 2 was increased

upon treatment with glial conditioned medium (Fig. 7).

3.4. The expression of the transcription factors RUNXs is

modulated by glial factors and correlates to syndecan expression

In order to evaluate if members of a family of recently described

transcription factors, the runx family, known to be involved in the

transition from proliferation to differentiation in different neuronal

lineages, could be regulated by glial secreted factors, we analyzed

the expression of the three Runx genes by granule cell precursors.

All three Runx genes known in mice, Runx1, 2 and 3, are expressed

by granule cell precursors, with a marked increase from P3 to P9

(Fig. 8). The expression of Runx1 and Runx2 is low both in P3 and P9

granule precursors and highly enhanced in P3 cells exposed to glial

secreted factors (Fig. 9). On the other hand, Runx3 is not expressed

at P3, has very low levels of expression at P9 and is completely

inhibited after treatment with glial conditioned medium at P9.

4. Discussion

Heparan sulfate proteoglycans play key roles in a great variety of

biological events, and although many studies have been conducted

in this field, there is still limited information about the involvement

of these molecules in CNS development. We have used the devel­

oping cerebellum as our working model due to the large amount

of cellular and molecular data available regarding the postnatal

development of this region of the CNS. Adding new elements to

the control mechanisms of the cellular events during cerebellar

development is important to better understand the mechanisms

underlying the appearance of human diseases such as ataxias,

cognition impairment and brain tumors. The data presented here

describe the expression pattern of syndecans, glypicans and hep­

aran sulfate modifying enzymes by cerebellar neuronal precursors

at different postnatal ages. Expression of some of those genes in

the whole cerebellum has been analyzed before by other authors

A.P.B. Araujo et al. / Int. J. Devl Neuroscience 28 (2010) 611–620 615

Fig. 2. Developmental changes in heparan sulfate sulfotransferases expression by neuronal precursors. The expression level of each gene was evaluated by real time RT­PCR

in samples obtained from granule cell precursors isolated from cerebellum of mice at different postnatal ages (3, 6 and 9 days). Values were normalized to that of b­actin

transcript. Data were obtained in triplicate and are given as the mean ± error. One­way analysis of variance (ANOVA) followed by Dunnett’s test was used to evaluate

differences among ages. Differences were considered statistically significant if p < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001; 2­OST­1, 2­O­sulfotransferase­1; 3­OST­1, 2, 3, 4,

and 5, 3­O­sulfotransferase­1, 2, 3, 4 and 5; NDST­1, 2, 3 and 4, N­deacetylase­N­sulfotransferase­1, 2, 3 and 4; 6­OST­1, 2 and 3; 6­O­sulfotransferase­1, 2 and 3).

Fig. 3. Extracellular heparan sulfate 6­O­endosulfatases expression by cerebellar neuronal precursors changes during postnatal development. The expression level of each

gene was evaluated by real­time RT­PCR in samples obtained from granule cell precursors isolated from cerebellum of mice at different postnatal ages (3, 6 and 9 days).

Values were normalized to that of b­actin transcript. Data were obtained in triplicate and are given as the mean ± error. One­way analysis of variance (ANOVA) followed by

Dunnett’s test was used to evaluate differences among ages. Differences were considered statistically significant if p < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001; mSulf­1 and ­2,

extracellular heparan sulfate 6­O­endosulfatase­1 and ­2).

616 A.P.B. Araujo et al. / Int. J. Devl Neuroscience 28 (2010) 611–620

Fig. 4. Glial cell conditioned medium induces axonal outgrowth by cerebellar neuronal precursors. Granule cell precursors obtained from mice cerebella at post­natal days

3, 6 and 9 were cultured in the absence or presence of medium conditioned by glial cells obtained from cerebella of mice at the same ages. After 1 or 7 days in culture, cells

were photographed and axonal length was measured. Graph shows axonal length measured at day 7. Data are from 3 independent experiments. Differences were considered

statistically significant if p < 0.05 (***p < 0.01; **p < 0.05).

using different approaches (Smith et al., 2005; Sato et al., 2008).

We present, for the first time to our knowledge, data regarding

the expression of all syndecans, glypicans and enzymes evaluated

in isolated granule cell precursors, the most abundant neuron in

the cerebellum. The developmental window selected for this work

was from P3 to P9, for two reasons. First, we were interested in

collecting information about the expression of heparan sulfate pro­

teoglycans in the period of higher proliferation and migration of

neuronal precursors as well as studying the possible modulation

of their expression by glial secreted factors. Second, in order to

study the expression of genes by cerebellar granule cell precur­

sors only, we needed to be able to isolate and culture them, and

this is possible only up to around P9, after that cells are too dam­

aged during the isolation procedure and usually do not survive in

culture.

Our results establish that the heparan sulfate proteoglycan core

proteins and the enzymes are regulated in the early postnatal

period by intrinsic and extrinsic mechanisms. Although we did not

address the question of how gene expression is regulated, it is clear

that during early postnatal cerebellar development, the expression

of the heparan sulfate proteoglycan core protein genes as well as

the genes encoding the modifying enzymes is controlled by a set of

intracellular factors and/or by factors secreted by cells surround­

ing the granule cell precursors. In the first 2 weeks after birth, the

cerebellum undergoes massive changes in size and in neural cir­

cuit formation. The growth of the cerebellum in the early postnatal

period is achieved by intense proliferation of granule cell precursors

in the external granular layer, followed by migration of all granule

cells inwards to form the internal granular layer.

A large number of papers in the literature accounts for the

involvement of heparan sulfate proteoglycans in several cellular

events, including proliferation, migration and differentiation. In

the CNS, conditional inactivation of a heparan sulfate­polymerizing

enzyme, Ext1, leads to hypoplasia of cerebral hemispheres, accom­

panied by absence of the olfactory bulbs and cerebellum, in addition

to severe guidance errors in major commissural tracts (Inatani

et al., 2003). The heparan sulfate modifying enzymes, the sulfo­

transferases, are also extremely important for proper development

of the brain as well as other organs, as showed by the hypopla­

sia and craniofacial defects observed in mice lacking the enzyme

N­deacetylase­N­sulfotransferase 1 (NDST1), responsible for N­

sulfation of N­acetylglucosamine residues (Grobe et al., 2005),

and kidney malformation in mice lacking 2­O­sulfotransferase

(2OST) (Wilson et al., 2002). The results we present here show

that expression levels of the modifying enzymes change over

time, more specifically, most sulfotransferases are downregulated

from P3 to P9 (Fig. 2). Also, the extracellular 6­O­endosulfatases

mSulf1 and mSulf2 expression is higher in younger cells, sug­

gesting that heparan sulfates would be more 6­O­sulfated as the

cerebellum develops. This result is in accordance with published

information about gene expression during cerebellar development

(www.cdtdb.brain.riken.jp) (Sato et al., 2008). We attempted to

sequence the heparan sulfates purified from granule cell pre­

cursors, and although we did not achieve to get the complete

structures, our results indicate that the sulfation profile of hep­

aran sulfate chains changes with postnatal development (data not

shown). Changes in the sulfation pattern of heparan sulfate chains

have been implicated in modifications of a great number of activ­

ities of those molecules, including changes in their growth factors

binding properties.

The expression of syndecans 2, 3 and 4 core proteins is upreg­

ulated during the first 9 days after birth (Fig. 1), and glial secreted

factors can stimulate the expression of those genes earlier, at P3

(Fig. 5). The increase in syndecan expression correlates with the

increase in axonal extension promoted by glial secreted factors

(Fig. 4). The upregulation of syndecans 2 and 3 could be related

to neuronal precursor differentiation and extension of axons for

synapse formation. It has been described that syndecan 2 is present

at the synapse and is an important element for synapse formation

and maturation (Ethell and Yamaguchi, 1999; Lucido et al., 2009).

Parallel to the increased expression of syndecans we observed

a variety of profiles for glypican expression. No changes in glypi­

cans 4 and 5 expression were observed, whereas glypicans 2, 3, and

6 had their expression downregulated (Fig. 1). Glypican 1 was the

only member of the glypican family that was highly upregulated

at P9 (Fig. 1). Glypican 1 seems to be key for neuronal precur­

sor proliferation, as Gpc1−/− mice present reduced brain size and

cerebellar foliation defects (Jen et al., 2009). When granule cell

precursors were treated with glial conditioned medium all glyp­

icans were downregulated in both ages tested (Fig. 5). These data,

together with the increased expression of syndecans promoted by

glial conditioned medium, suggest that glial secreted factors could

be repressing proliferation and inducing differentiation of neuronal

precursors.

A.P.B. Araujo et al. / Int. J. Devl Neuroscience 28 (2010) 611–620 617

Fig. 5. Glial cell conditioned medium upregulates the expression of syndecans and

downregulates the expression of glypicans by neuronal precursors. Granule cell pre­

cursors obtained from mice cerebella at post­natal days 3 and 9 were cultured in

the absence or presence of medium conditioned by glial cells obtained from cere­

bella of mice at the same ages. After 24 h in culture, neuronal precursors RNA was

extracted and syndecan and glypican expression was measured by real time RT­PCR.

Data were obtained in triplicate and are given as the mean ± error.

Besides changes observed in the expression of genes encoding

syndecans and glypicans core proteins, the expression of heparan

sulfate modifying enzymes was also affected by glial soluble fac­

tors. The main trend observed was an upregulation of the majority

of sulfotransferases when P9 cells were treated with glial condi­

tioned medium (Fig. 6) and upregulation of mSulf1 at P3 and P9, and

upregulation of mSulf2 at P3 (Fig. 7). Regulation of heparan sulfate

sulfatases has been previously linked to mechanisms of differen­

tiation of neural progenitors into oligodendrocytes in embryonic

chick ventral spinal cord (Danesin et al., 2006).

We then asked if there were any transcription factors that

would be able to promote changes in the expression of a variety

of genes. The runt­related transcription factors, Runx, appealed as

interesting candidates because they compose a family of evolution­

ary conserved transcription factors that play essential roles during

development, regulating the transition from proliferation to differ­

entiation in several cell types (Zagami et al., 2009). Mammalians

have three members in the family, Runx1, Runx2 and Runx3. The

RUNX proteins bind other DNA­binding proteins, transcriptional

coactivators and corepressors leading to up or downregulation of

different target genes (Durst and Hiebert, 2004; Miyazono et al.,

2004; Katoh, 2007). Runx2 has been implicated in bone develop­

ment (Lian et al., 2004; Komori, 2008), whereas both Runx1 and

Runx3 are involved in neuronal proliferation and differentiation

(Inoue et al., 2008; Zagami et al., 2009). Although, to our knowl­

edge, there is no information regarding the expression of Runx

in the developing cerebellum, our results show that granule cell

precursors express all three members of the family, Runx1, Runx2

and Runx3 (Fig. 8). At P3 we detected very low levels of all three

Runx genes, whereas at P9 all of them were upregulated. Exposing

the granule cell precursors to glial cell secreted factors altered the

expression of all three genes. While Runx1 and Runx2 expression

was upregulated in P3 cells, their expression did not change at P9.

On the other hand, the expression of Runx3 was not altered at P3

and was completely abolished at P9 (Fig. 9).

The expression profile of Runx1 and Runx2 upon treatment of

granule cell precursors with glia conditioned medium nicely cor­

relates with the expression of all four syndecans (Figs. 1 and 9),

suggesting that these genes could either be regulated by a common

activation pathway upstream of Runx1 and Runx2, or alternatively

these transcription factors could regulate the expression of synde­

cans themselves. There is no information regarding any regulatory

mechanism implicating Runxs and syndecans, but there is data

showing that 12 out of the 13 known small leucine rich proteo­

glycans (SLRPs) have at least one HOX–RUNX binding site (Tasheva

et al., 2004). Also, it has been shown that the expression of Runx2 by

mesenchymal stem cells is upregulated upon treatment with BMP­

2, which also leads to upregulation the SRLP biglycan (Knippenberg

et al., 2006). A recent publication has identified RUNX proteins as

the crucial transcription regulators for luteinizing hormone (LH)­

induced hyaluronan and proteoglycan link protein 1 (HAPLN1)

expression by rat ovaries (Liu et al., 2010).

Syndecans are involved in the mechanisms of cell attachment

and axonal outgrowth (Kinnunen et al., 1998). In the same way,

Runx3 has also been implicated in the regulation of axonal out­

growth and axonal guidance of proprioceptive DRG (dorsal root

ganglion) neurons (Inoue et al., 2002). Our results show that Runx3

expression is low in postnatal ages studied, but the expression of

Runx1 and Runx2, low in control conditions, can be highly stimu­

lated by soluble factors secreted by glial cells, similar to what we

observed for syndecans, and correlated to what we observed for

axonal outgrowth.

The snapshots of the expression of syndecans, glypicans and

heparan sulfate modifying enzymes show that the final structure

of these proteoglycans vary during early postnatal development of

the cerebellum, certainly serving different roles in the dynamics of

618 A.P.B. Araujo et al. / Int. J. Devl Neuroscience 28 (2010) 611–620

Fig. 6. Glial cell conditioned medium regulates the expression of heparan sulfate sulfotransferases by neuronal precursors. Granule cell precursors obtained from mice

cerebella at post­natal days 3 and 9 were cultured in the absence or presence of medium conditioned by glial cells obtained from cerebella of mice at the same ages. After 24 h

in culture, neuronal precursors RNA was extracted and heparan sulfate sulfotransferases expression was measured by real time RT­PCR. Data were obtained in triplicate and are

given as the mean ± error (2­OST­1, 2­O­sulfotransferase­1; 3­OST­1, 2, 3, 4, and 5, 3­O­sulfotransferase­1, 2, 3, 4 and 5; NDST­1, 2, 3 and 4, N­deacetylase­N­sulfotransferase­1,

2, 3 and 4; 6­OST­1, 2 and 3; 6­O­sulfotransferase­1, 2 and 3).

Fig. 7. Glial cell conditioned medium regulates the expression of extracellular heparan sulfate 6­O­sulfatases by neuronal precursors. The expression level of each gene was

normalized to that of b­actin transcript. Data were obtained in triplicate and are given as the mean ± error (mSulf­1 and 2; extracellular heparan sulfate 6­O­endosulfatase­1

and 2).

Fig. 8. Developmental changes in Runx gene expression by neuronal precursors. The expression level of each gene was evaluated by real­time RT­PCR in samples obtained

from granule cell precursors isolated from cerebellum of mice at different postnatal ages (3, 6 and 9 days). Values were normalized to that of b­actin transcript. Data were

obtained in triplicate and are given as the mean ± error. One­way analysis of variance (ANOVA) followed by Dunnett’s test was used to evaluate differences among ages.

Differences were considered statistically significant if p < 0.05 (**p < 0.01; ***p < 0.001).

A.P.B. Araujo et al. / Int. J. Devl Neuroscience 28 (2010) 611–620 619

Fig. 9. Glial cell conditioned medium regulates the expression of Runx family genes by neuronal precursors. The expression level of each gene was normalized to that of

b­actin transcript. Data were obtained in triplicate and are given as the mean ± error.

neuronal proliferation and differentiation. Our data also show for

the first time that cerebellar neuronal progenitors express mem­

bers of the Runx family and that the expression of those genes

can also be controlled by extrinsic factors secreted by glial cells.

The treatment of isolated granule cell precursors with soluble fac­

tors secreted by glia indicate that the expression of those genes

is also regulated by the surroundings of the neuronal precursors

and reflect the intimate relations between neuron and glia during

development.

Contributions

APBA and MAP conceived the studies, designed the experiments,

interpreted the data and wrote the manuscript. APBA, MEOBR, RR,

RJT carried out the experiments. LT contributed with real time RT­

PCR experimental design and interpretation.

Conflict of interest statement

The authors declare that they have no conflict of interest.

Acknowledgments

This work was granted by FAPESP (Research Grant to MAP; Grad

Student’s Fellowship to MEOBR, RR, and RJT); CNPq (Research Fel­

lowship to MAP; PhD Fellowship to APBA); and Investiga Institutos

de Pesquisa (Research Grant to MAP; Fellowship to APBA).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in

the online version, at doi:10.1016/j.ijdevneu.2010.07.228.

References

Altschul, S.F., Madden, T.L., et al., 1997. Gapped BLAST and PSI­BLAST: a new genera­tion of protein database search programs. Nucleic Acids Res. 25 (17), 3389–3402.

Behesti, H., Marino, S., 2009. Cerebellar granule cells: insights into proliferation,differentiation, and role in medulloblastoma pathogenesis. Int. J. Biochem. Cell.Biol. 41 (3), 435–445.

Busch, S.A., Silver, J., 2007. The role of extracellular matrix in CNS regeneration. Curr.Opin. Neurobiol. 17 (1), 120–127.

Capurro, M.I., Xu, P., et al., 2008. Glypican­3 inhibits Hedgehog signaling duringdevelopment by competing with patched for Hedgehog binding. Dev. Cell 14(5), 700–711.

Chan, J.A., Balasubramanian, S., et al., 2009. Proteoglycan interactions with SonicHedgehog specify mitogenic responses. Nat. Neurosci. 12 (4), 409–417.

Danesin, C., Agius, E., et al., 2006. Ventral neural progenitors switch toward anoligodendroglial fate in response to increased Sonic hedgehog (Shh) activity:involvement of Sulfatase 1 in modulating Shh signaling in the ventral spinalcord. J. Neurosci. 26 (19), 5037–5048.

Dietrich, C.P., Tersariol, I.L., et al., 1998. Structure of heparan sulfate: identificationof variable and constant oligosaccharide domains in eight heparan sulfates ofdifferent origins. Cell Mol. Biol. (Noisy­le­grand) 44 (3), 417–429.

Durst, K.L., Hiebert, S.W., 2004. Role of RUNX family members in transcriptionalrepression and gene silencing. Oncogene 23 (24), 4220–4224.

Ethell, I.M., Yamaguchi, Y., 1999. Cell surface heparan sulfate proteoglycansyndecan­2 induces the maturation of dendritic spines in rat hippocampal neu­rons. J. Cell Biol. 144 (3), 575–586.

Ford­Perriss, M., Turner, K., et al., 2003. Localisation of specific heparan sulfate pro­teoglycans during the proliferative phase of brain development. Dev. Dyn. 227(2), 170–184.

Fransson, L.A., Belting, M., et al., 2004. Novel aspects of glypican glycobiology. CellMol. Life Sci. 61 (9), 1016–1024.

Gallagher, J.T., Turnbull, J.E., et al., 1990. Heparan sulphate proteoglycans. Biochem.Soc. Trans. 18 (2), 207–209.

Galtrey, C.M., Fawcett, J.W., 2007. The role of chondroitin sulfate proteoglycans inregeneration and plasticity in the central nervous system. Brain Res. Rev. 54 (1),1–18.

Grobe, K., Inatani, M., et al., 2005. Cerebral hypoplasia and craniofacial defectsin mice lacking heparan sulfate Ndst1 gene function. Development 132 (16),3777–3786.

Gulino, A., Arcella, A., et al., 2008. Pathological and molecular heterogeneity ofmedulloblastoma. Curr. Opin. Oncol. 20 (6), 668–675.

Hatten, M.E., Heintz, N., 2005. Large­scale genomic approaches to brain developmentand circuitry. Annu. Rev. Neurosci. 28, 89–108.

Inatani, M., Irie, F., et al., 2003. Mammalian brain morphogenesis and midline axonguidance require heparan sulfate. Science 302 (5647), 1044–1046.

Inoue, K., Ozaki, S., et al., 2002. Runx3 controls the axonal projection of proprioceptivedorsal root ganglion neurons. Nat. Neurosci. 5 (10), 946–954.

Inoue, K., Shiga, T., et al., 2008. Runx transcription factors in neuronal development.Neural Dev. 3, 20.

Jen, Y.H., Musacchio, M., et al., 2009. Glypican­1 controls brain size through regula­tion of fibroblast growth factor signaling in early neurogenesis. Neural Dev. 4,33.

Katoh, M., 2007. Networking of WNT, FGF, Notch, BMP, and Hedgehog signalingpathways during carcinogenesis. Stem Cell Rev. 3 (1), 30–38.

Kinnunen, A., Kinnunen, T., et al., 1998. N­syndecan and HB­GAM (heparin­bindinggrowth­associated molecule) associate with early axonal tracts in the rat brain.Eur. J. Neurosci. 10 (2), 635–648.

Klein, R.S., Rubin, J.B., et al., 2001. SDF­1 alpha induces chemotaxis and enhancesSonic hedgehog­induced proliferation of cerebellar granule cells. Development128 (11), 1971–1981.

Knippenberg, M., Helder, M.N., et al., 2006. Osteogenesis versus chondrogenesis byBMP­2 and BMP­7 in adipose stem cells. Biochem. Biophys. Res. Commun. 342(3), 902–908.

Komori, T., 2008. Regulation of bone development and maintenance by Runx2. Front.Biosci. 13, 898–903.

Kroger, S., Schroder, J.E., 2002. Agrin in the developing CNS: new roles for a synapseorganizer. News Physiol. Sci. 17, 207–212.

Lian, J.B., Javed, A., et al., 2004. Regulatory controls for osteoblast growth and dif­ferentiation: role of Runx/Cbfa/AML factors. Crit. Rev. Eukaryot. Gene Expr. 14(1–2), 1–41.

Liu, J., Park, E.S., 2010. Periovulatory expression of hyaluronan and proteoglycan linkprotein 1 (Hapln1) in the rat ovary: hormonal regulation and potential function.Mol. Endocrinol. 24 (6).

Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data usingreal­time quantitative PCR and the 2(−Delta Delta C(T)) Method. Methods 25(4), 402–408.

Lucido, A.L., Suarez Sanchez, F., et al., 2009. Rapid assembly of functional presynapticboutons triggered by adhesive contacts. J. Neurosci. 29 (40), 12449–12466.

Miyazono, K., Maeda, S., et al., 2004. Coordinate regulation of cell growth and dif­ferentiation by TGF­beta superfamily and Runx proteins. Oncogene 23 (24),4232–4237.

Pearlman, A.L., Sheppard, A.M., 1996. Extracellular matrix in early cortical develop­ment. Prog. Brain Res. 108, 117–134.

Porcionatto, M.A., 2006. The extracellular matrix provides directional cues for neu­ronal migration during cerebellar development. Braz. J. Med. Biol. Res. 39 (3),313–320.

Rubin, J.B., Choi, Y., et al., 2002. Cerebellar proteoglycans regulate sonic hedgehogresponses during development. Development 129 (9), 2223–2232.

Sato, A., Sekine, Y., et al., 2008. Cerebellar development transcriptome database(CDT­DB): profiling of spatio­temporal gene expression during the postnataldevelopment of mouse cerebellum. Neural Netw. 21 (8), 1056–1069.

620 A.P.B. Araujo et al. / Int. J. Devl Neuroscience 28 (2010) 611–620

Smith, F.I., Qu, Q., et al., 2005. Gene expression profiling of mouse postnatalcerebellar development using oligonucleotide microarrays designed to detectdifferences in glycoconjugate expression. Gene Expr. Patterns 5 (6), 740–749.

Steinlin, M., 2008. Cerebellar disorders in childhood: cognitive problems. Cerebel­lum 7 (4), 607–610.

Sugahara, K., Kitagawa, H., 2002. Heparin and heparan sulfate biosynthesis. IUBMBLife 54 (4), 163–175.

Tasheva, E.S., Klocke, B., et al., 2004. Analysis of transcriptional regulation of thesmall leucine rich proteoglycans. Mol. Vis. 10, 758–772.

Vaillant, C., Monard, D., 2009. SHH pathway and cerebellar development. Cerebellum8 (3), 291–301.

Wilson, V.A., Gallagher, J.T., et al., 2002. Heparan sulfate 2­O­sulfotransferase (Hs2st)and mouse development. Glycoconj. J. 19 (4–5), 347–354.

Yabe, T., Hata, T., et al., 2005. Developmental and regional expression of heparansulfate sulfotransferase genes in the mouse brain. Glycobiology 15 (10), 982–993.

Zagami, C.J., Zusso, M., et al., 2009. Runx transcription factors: lineage­specific regu­lators of neuronal precursor cell proliferation and post­mitotic neuron subtypedevelopment. J. Cell Biochem. 107 (6), 1063–1072.

Zhou, P., Porcionatto, M., et al., 2007. Polarized signaling endosomes coordi­nate BDNF­induced chemotaxis of cerebellar precursors. Neuron 55 (1), 53–68.