Genomic and functional profiling of human Down
syndrome neural progenitors implicates S100B
and aquaporin 4 in cell injury
Giuseppe Esposito1, Jaime Imitola2, Jie Lu5, Daniele De Filippis6, Caterina Scuderi1,
Vijay S. Ganesh5, Rebecca Folkerth3, Jonathan Hecht4, Soojung Shin7, Teresa Iuvone6,
Jonathan Chesnut7, Luca Steardo1and Volney Sheen5,?
1Department of Human Physiology and Pharmacology, ‘Vittorio Erspamer’ Faculty of Pharmacy, University of Rome
‘La Sapienza’, Rome, Italy,2Center for Neurologic Diseases, Division of Neurology, Brigham and Women’s Hospital,
Harvard Medical School, Boston MA 02115 USA,3Department of Pathology, Division of Neuropathology, Brigham and
Women’s Hospital, Harvard Medical School, Boston MA 02114 USA,4Department of Neuropathology, Beth Israel
Deaconess Medical Center, Harvard Medical School, Boston, MA 02115, USA,5Department of Neurology, Beth Israel
Deaconess Medical Center, Harvard Medical School, Boston, MA 02115, USA,6Department of Experimental
Pharmacology, Faculty of Pharmacy, University Federico II of Naples Via Domenico Montesano 49, 80131 Rome, Italy
and7Invitrogen, Carlsbad, CA 92008, USA
Received August 8, 2007; Revised and Accepted November 1, 2007
Down syndrome (DS) is caused by trisomy of chromosome 21 and is characterized by mental retardation,
seizures and premature Alzheimer’s disease. To examine neuropathological mechanisms giving rise to
this disorder, we generated multiple human DS neural progenitor cell (NPC) lines from the 19–21 week frontal
cortex and characterized their genomic and functional properties. Microarray profiling of DS progenitors indi-
cated that increased levels of gene expression were not limited to chromosome 21, suggesting that increased
expression of genes on chromosome 21 altered transcriptional regulation of a subset of genes throughout
the entire genome. Moreover, many transcriptionally dysregulated genes were involved in cell death and oxi-
dative stress. Network analyses suggested that upregulated expression of chromosome 21 genes such as
S100B and amyloid precursor protein activated the stress response kinase pathways, and furthermore,
could be linked to upregulation of the water channel aquaporin 4 (AQP4). We further demonstrate in DS
NPCs that S100B is constitutively overexpressed, that overexpression leads to increased reactive oxygen
species (ROS) formation and activation of stress response kinases, and that activation of this pathway
results in compensatory AQP4 expression. In addition, AQP4 expression could be induced by direct
exposure to ROS, and siRNA inhibition of AQP4 resulted in elevated levels of ROS following S100B exposure.
Finally, elevated levels of S100B-induced ROS and loss of AQP4 expression led to increased programmed
cell death. These findings suggest that dysregulation of chromosome 21 genes in DS neural progenitors
leads to increased ROS and thereby alters transcriptional regulation of cytoprotective, non-chromosome
21 genes in response to ongoing cellular insults.
Numerous studies have demonstrated the ability to expand plur-
ipotent human embryonic stem cells (hESCs) or more restricted
multipotent human neural progenitor cells (hNPCs) and direct
their differentiation into various neuronal phenotypes (1–5).
These neural progenitors can undergo the same migratory
pathway as endogenous precursors. They are able to integrate
into the parenchyma. They also assume neuronal and glial
?To whom correspondence should be addressed. Tel: þ1 6176672699; Fax: þ1 6176670800; Email: email@example.com
# The Author 2007. Published by Oxford University Press. All rights reserved.
For Permissions, please email: firstname.lastname@example.org
Human Molecular Genetics, 2008, Vol. 17, No. 3
Advance Access published on November 5, 2007
by guest on June 2, 2013
phenotypes appropriate for the region of engraftment, if trans-
planted during the period of developmental neurogenesis
(3,6). Some studies have subsequently developed models of
human central nervous system (CNS) disorders by deriving
hNPCs from diseased tissues, while other studies have proposed
developing models of human diseases using various manipula-
tions of hESCs (7–10). These models are thought to be of par-
ticular benefit, when animal models do not clearly reflect the
human phenotype as in Parkinson’s disease, Alzheimer’s
disease (AD) or in Down syndrome (DS). Although such
studies have raised the potential for using human reagents in
the study of human diseases, few studies to date have actually
sought to identify and examine potential aberrant pathways
that could contribute to the disease phenotypes.
DS (OMIM 190685) is a common genetic variation caused
by a duplication of chromosome 21 (8,11,12). The most pro-
found neurological features of DS include mental retardation,
seizures and early onset Alzheimer disease. DS brains weigh
less than age-matched normal brains. They demonstrate simpli-
fied gyral patterning. The neuronal density and lamination are
also reduced in the DS brain (13–15). Furthermore, there
appears to be a preferential loss in GABA-ergic neurons and
the neuronal density, while normal in early gestation, decreases
later in gestation (.23 weeks) (16). Finally, the developing DS
cortical neurons have shorter dendrites, fewer dendritic spines
and abnormal dendritic morphology (17,18). Taken in sum,
these observations are consistent with alterations in cell fate
or reduced proliferation in neuronal subpopulations. Subsequent
impairments in neuronal differentiation also appear to be
reflected in DS brain. However, the subset of genes which are
constitutively overexpressed on chromosome 21 and lead to
these impairments are not known.
Analyses of DS neurons or neural progenitor cells (NPCs) in
culture have allowed for characterization of some candidate,
disease-causing genes in this disorder. Primary cortical
neurons from gestational 16–18 week fetal DS brain fail to
survive long-term in comparison to normal cultures, even in
the presence of serum with survival factors (19). The
neurons are more vulnerable to intracellular reactive oxygen
species (ROS), in part due to mitochondrial dysfunction in
DS neural cells, causing aberrant processing of amyloid pre-
cursor protein (APP). The impairment in APP processing
results in increased neuronal cell death (20). More recent
studies have also begun to examine renewable neural
progenitors derived from DS individuals (21). Differential
display PCR used on NPCs from gestational 17–19 week
growth-associated protein regulated by the neuron-restrictive
silencer factor REST, was almost undetectable in the DS
sample. While specific genes have been implicated in DS
neural development, the underlying molecular pathways that
contribute to the increased neuronal cell death and decreased
neuronogenesis are not entirely clear.
To pursue molecular mechanisms contributing to the DS
phenotype, we generated multiple human DS neural precursor
lines, derived from the frontal cortex and examined their gene
expression profiles and functional characteristics. Microarray
profiling of these DS cells demonstrated upregulation of
genes, located both on and off of chromosome 21. The
altered gene expressions strongly implicated genes involved
in cell death and oxidative stress. Network analyses suggested
that overexpression of genes such as S100B and APP on
chromosome 21 led to activation of stress response kinases,
generation of ROS and compensatory upregulation of the
water channel aquaporin 4 (AQP4) within DS NPCs. Further-
more, functional data presented here indicate that upregulation
of AQP4 may serve to mitigate cell damage and death due to
S100B induction of ROS. Overall, these studies provide some
of the first evidence of altered molecular pathways in contri-
buting to the DS neurologic phenotype.
Characterization of human DS NPCs
To generate human neural precursor cell lines, progenitors
were isolated from the cerebral wall of the frontal cortex of
19–21 week postmortem fetuses with DS. Clonal populations
were obtained from single cells, isolated by dissociation from
the subventricular zone of the frontal cortex, passaged through
a cell strainer, diluted to low density and then allowed to grow
into multiple neurospheres. On gross observation, the DS
NPCs appeared to proliferate more slowly compared with
control at early time points after dissociation (NPC diameter
for DS ¼ 77+9 mM, control ¼ 125+9 mM from three inde-
pendent lines each within one month). To exclude potential
mosaicism within individual samples, we performed fluor-
escent in situ hybridization (50/50 cells analyzed from each
of three independent samples) demonstrating the additional
chromosome 21 copy in metaphase and non-dividing inter-
phase (Supplementary Fig. S1A). Gene expression profiling
and culture studies were performed on three independent DS
fetal brain and three independent age-matched control neural
precursors obtained at gestational weeks 19–21 and cultured
for no more than five passages. Examination of mRNA
expression levels with the DS critical region (?3400 kb
region on 21q22) revealed, on average, a 1.5-fold increase in
expression of genes in this region for DS as compared to
control progenitors (Supplementary Fig. S1B).
To better define the characteristics of the control and DS
NPCs, we performed several levels of analyses. Pairwise com-
parisons of the expression levels for the ?54 000 probes in the
HU133 plus 2.0 Affymetrix microarray showed a high corre-
lation on inter-control and inter-DS comparisons, indicating
that the progenitor cells were fairly uniform in their expression
profiles. Moreover, significant differences were seen on
control–DS sample comparison, suggesting the presence of
dysregulated gene pathways in the DS precursors (Fig. 1A).
Within the control and DS cell lines, we also analyzed the
expression levels of multiple genes responsible for self-
renewal and neural progenitor identity (Fig. 1B). The
expression and conservation of this subset of genes were
related to their intrinsic capacities such as self-renewal
capacity and help characterize these cells as progenitors
(22,23). No significant differences in the levels of expression
between DS and wild-type (WT) control were seen in the tran-
scription factors SOX2, NCOR2, ITGB1 and NEK6, which are
associated with NPC self-renewal capacity (23–26). In
addition, the expression of neural progenitor genes FLNA,
FOXG1B and DCX (27–29) was not significantly changed.
Human Molecular Genetics, 2008, Vol. 17, No. 3 441
by guest on June 2, 2013
However, we found a slight increase in the expression of
GFAP in DS progenitors compared with controls. GFAP is
an intermediate filament protein gene, that is, expressed in
hNPCs and mature astrocytes (30,31). Several of the genes
involved in cancer and cell proliferation (PTPRZ1 and
EEF2) were also increased in DS NPCs, which generally
showed increased rates of proliferation with serial passaging.
These data suggested that while DS neural progenitors were
fairly uniform between human samples and largely retained
the core features of gene expression seen in neural progenitors.
As initial comparisons on the microarray platform raised the
possibility that the DS progenitors shared similarities with
glial progenitor phenotypes, we performed cell-type-specific
marker analysis of the undifferentiated NPCs by immunostain-
ing to further characterize the identity of these NPCs (Figs 1C
and 2). Although both the control and DS progenitors were
age-matched, isolated from the same regions of the brain
and cultured under the same conditions, the DS progenitors
from each of the three independent DS lines exhibited
increased expression for glial phenotypes, including vimentin
(radial glia), glial fibrillary acid protein (astrocytes/NPCs), O4
nucleotide 30-phosphodiesterase (oligodendrocytes). No sig-
nificant difference was observed in NG2 chondroitin sulfate
proteoglycan (oligodendrocyte precursors/NPCs), whereas
control NPCs showed increased expression of nestin (neuroe-
pithelial precursors) and MAP2 (neurons). Thus, while the DS
NPCs retained progenitor characteristics and could express
neuronal markers, they also appeared to adopt skewed glial
progenitor phenotypes compared with control.
to control progenitors reflected differences due to maturation
Figure 1. Human Down syndrome neural stem cells represent a fairly homogeneous cell population on inter-sample comparison, but differ significantly on com-
parison with age-matched normal controls. (A) Homogeneity of mRNA expression in DS and control human neural progenitors. Pairwise comparisons are per-
formed between levels of expression for each of the 50K transcripts on the Affymetrix H133þ Array. Significant increases (red) or decreases (green) in
expression are noted on comparison of a DS neural cell line versus a human control neural line (bottom graph). The high degrees of similarity on comparison
between control-control (top graph) and DS-DS (middle graph) neural lines are indicated by the relatively small regions of increased (red) or decreased (green)
gene expression. These observations suggest that the neural precursor cell lines derived from different individuals are fairly uniform and that the differences seen
in the DS lines are likely attributable to the disease process. (B) Human DS neural stem cells preserve core expression of neural progenitors genes. Microarray
profiling of mRNA expression comparing the expression of core NPC and progenitor genes demonstrate a non-statistical significance difference between DS
NPCs and controls. Note that the only gene with a significant increase was GFAP (averaged between n ¼ 3 control and n ¼ 3 DS lines). (C) Quantitative
summary of immunostaining of the three DS and three control NPC lines with regards to various neural, neuronal and glial markers. The DS progenitors
exhibit increased expression for glial precursor markers including vimentin, CNPase and GFAP, suggesting a predominant glial progenitor phenotype. Fluor-
escent photomicrograph images of representative immunostaining patterns are shown in Figure 2.?P , 0.01;??P , 0.005;???P , 0.002.
442Human Molecular Genetics, 2008, Vol. 17, No. 3
by guest on June 2, 2013
levels in cultures or adoption of more restricted glial progenitor
between the DS progenitors and previously published datasets
[NCBI, Dr Clive Svendsen GSM51360-51368 lines (32)]
from long-term cultured NPCs, astrocyte progenitor cells
(APCs) and oligodendrocyte progenitor cells (OPCs, Sup-
plementary Fig. S2). Longer term cultured NPCs expressed sig-
nificantly increased levels of differentiated cell markers,
including GFAP (GFAP), MAPT (neuronal) and OMG/MPB
(oligodendrocyte), when compared with the DS NPCs. Conver-
sely, the APC and OPC lines expressed significantly reduced
levels of these mature cell markers compared with the DS
NPCs. Thus, while the DS progenitors had adopted certain
glial progenitor characteristics, these features did not appear
to result from increased NPC maturation or adoption of astro-
cyte or oligodendrocyte progenitor phenotypes alone.
Gene ontology comparison analysis
To identify functional modules of gene expression and inter-
acting partners that were relevant to the DS gene expression
signature, we applied network-based analysis through the
Ingenuity Pathways Knowledge Base (IPA) (33). The initial
pairwise comparisons between multiple DS and control NPC
microarray profiles yielded 1902 candidate transcripts with
P-value thresholds of ,0.05. Of this group, only 608 candi-
dates were shown also to have at least an absolute 1.5-fold
expression change relative between the DS and control
samples. Fisher’s exact tests were then performed to determine
the likelihood that these genes of interest participated in a
given function or pathway, relative to the total number of
occurrences of these genes in all functional/pathway annota-
tions stored in the IPA. These analyses produced 334 candi-
date genes that could be placed into 46 networks. The top
four networks incorporated 90 of the candidate genes within
a network of 140 molecules with significance P-values
between 10E-27 and 10E-37 (Fig. 3A). These networks were
predicted to be involved in cell death and oxidative stress
through various signaling mechanisms including the JAK/
STAT canonical pathway (Fig. 3B).
Our observation that the trisomy 21 in the DS NPCs resulted
in an approximate 1.5-fold increase in expression of the DS
Figure 2. DS NPCs adopt a more glial-predominant progenitor phenotype. Fluorescent photomicrographs show vimentin and nestin immunostaining of both DS
and control NPCs. Increased expression of nestin (neuroepithelial precursor,) is seen in the control NPCs, while DS NPCs show slightly greater expression of the
vimentin (radial glia). DS NPCs also show diminished levels of staining for the neuronal marker MAP2. Conversely, the DS NPCs exhibit increased expression
of various glial-associated markers including CNPase (oligodendrocytes), O4 sulfatide (oligodendrocyte precursor) and GFAP (astrocyte, neural precursor
marker), while no significant difference is observed for NG2 expression (oligodendrocyte precursor) between DS and normal controls. The respective cell
specific markers are seen under rhodamine fluorescence.The nuclei of cells are labelled with Hoechst (DAPI fluorescence) to demonstrate cell density and uni-
formity of the cultures.
Human Molecular Genetics, 2008, Vol. 17, No. 3443
by guest on June 2, 2013
critical region genes on microarray profiling suggested that the
majority of significantly differentially expressed genes could
result from downstream effects of increased gene dosage on
chromosome 21. This possibility was consistent with the
array data showing that relatively small 1.5–2-fold upregula-
tion of genes on chromosome 21 resulted in a global effect on
gene expression throughout the DS NPC transcriptome (Sup-
plementary Fig. S3). Of the four networks identified by IPA
analyses, the fourth network implicated two chromosome 21
genes, S100B and APP, in functional pathways that also incor-
porated multiple genes that were found to be dysregulated
within the DS NPC gene expression profiles (Fig. 4). Both
S100B and APP have previously been implicated in DS neuro-
toxicity from elevated expression of free radicals (34–37). Of
these proteins, S100B is also found predominantly in the glial
cell lineages, promotes astrocytosis and is increased in DS
brain (38–42). Collectively, these findings suggested that
DS NPCs harboring duplications of S100B possessed elevated
levels of ROS and activated the JAK/STAT and MAPK stress
response kinase pathways.
S100B-induced oxidative stress promotes activation
of stress response kinases in human DS NPCs
As a first step in investigating this potential pathway, we
wanted to confirm that the profiling changes seen on compari-
son of the DS and control progenitors did not merely reflect
differences in cell types between the experimental and control
NPC lines. More specifically, S100B expression has been
associated with gliogenesis, and given that the DS progenitors
appeared to adopt more gliocentric progenitor phenotypes, we
wanted to exclude the possibility that the increased glial pro-
Figure 3. Ingenuity Pathway Analyses of differential gene expression in DS NPCs. (A) Interactome mapping predicts specific functional networks containing
differentially expressed genes within the DS NPCs. The score reflects the statistical probability (P-value ¼ 10E-27–10E-37) that the genes of interest participate
in the specific pathway, relative to the total number of occurrences of these genes in all functional/pathway annotations stored in the IPA Knowledge Base. The
number of focus genes indicates the number of dysregulated genes within each network of 35 genes. The top function describes the predicted function of the
network. (B) Graphical summary of the predicted biological functions, canonical pathways and resultant toxicity expected from the differentially expressed
444Human Molecular Genetics, 2008, Vol. 17, No. 3
by guest on June 2, 2013
genitor phenotype was responsible for differences seen in
expression levels for these genes within the IPA predicted
network. Expression profiling of long-term cultured NPCs
showed markedly increased expression of GFAP, MBP and
OMG compared with the DS progenitors. Despite this increase
in glial marker levels, expression of genes within the pathway
(i.e. S100B, APP, STAT2/5) was still disproportionately
greater in the DS NPCs, indicating that cell type specificity
alone was not responsible for differential gene expression (Sup-
plementary Fig. S2). Similarly, the upregulation of these
specific genes within the DS NPC network was not appreciated
within the expression profiles for the APC or OPC cell lines.
As the microarray and network analyses provided a plat-
form to identify potential pathways important for DS, we
next sought to examine whether these processes could be
appreciated functionally on a cellular level. Increased oxi-
dative stressors and free radical scavenging have previously
been reported in differentiated DS neurons and genes involved
in these same pathways were apparent in DS NPCs by micro-
array profiling (19,20,43). We therefore asked whether hydro-
gen peroxide (H2O2) generation and malondialdehyde (MDA)
formation, markers of ROS formation and cell lipid membrane
peroxidation products, respectively, were actually altered in
DS neural progenitors. Prior to and up to 48 h after induction
of differentiation, both H2O2and MDA levels increased to a
greater degree and in a parallel and time-dependent fashion
within DS as opposed to control neural precursor cells
(Fig. 5A and B). Moreover, S100B was increased in the DS
NPCs and stimulation of the DS progenitors with S100B led
to further increases in oxidative stress generation after 24 h.
This response was mitigated by pre-incubation with the
specific S100B neutralizing antibody, antibodies to the
S100B RAGE receptor, or free radical scavengers, tocopherol
and ascorbic acid.
Figure 4. Schematic of network #4 demonstrates that many of the differentially expressed genes in the DS NPCs form tight interrelationships within known
pathways involved in cell signalling, cell assembly and organization and gene expression. Multiple upregulated genes (darker shapes) form known interactions
with each other suggesting a striking robustness of altered gene expression in DS NPCs. Constitutively overexpressed genes on chromosome 21 (S100B and
APP) are predicted to alter STAT/JAK and MAPK activity.
Human Molecular Genetics, 2008, Vol. 17, No. 3445
by guest on June 2, 2013
Figure 5. Increased free radicals and lipid peroxidation by-products in human DS NPCs. (A) Levels of hydrogen peroxide (H2O2) and malondialdehyde (MDA,
marker of lipid peroxidation) are increased in unstimulated DS NPCs compared with control. Stimulation of both control and DS NPCs by S100B, a calcium/
zinc-binding protein implicated in the activation of ROS generation, leads to further increases in free radical and lipid peroxidation by-products. S100B-induced
generation of ROS can be inhibited by free radical scavengers, alpha-tocopherol and ascorbic acid in a dose-dependent fashion (???P , 0.001 compared with
unstimulated; ###P , 0.001 and oooP , 0.001 compared with 5 mM S100B stimulus). (B) Free radicals and lipid peroxidation by-products also increase follow-
ing induction of neural differentiation in DS progenitors. Both control and DS progenitors gradually accumulate increasing levels of intracellular H2O2and lipid
peroxidation by-products (MDA) following withdrawal of growth factors and maintenance in serum-containing media. DS progenitors, however, exhibit much
greater increases in H2O2and MDA compared with the control cells. A significant proportion of the increase in free radicals and lipid peroxidate by-products can
be mitigated by incubation of cultures with antibody to S100B (???P , 0.001). Generation of reactive oxygen species (ROS) is also dependent on S100B acti-
vation through its RAGE receptor (receptor for advanced glycation end products). Treatment of control progenitors with S100B leads to a rise in both free radical
production and lipid peroxidation, which can be inhibited by RAGE blockade in a dose-dependent fashion (1:10000 or 1:1000 dilution). (?P , 0.05,??P , 0.01
and???P , 0.001 compared with control;ooP , 0.01 andoooP , 0.001 compared to 5 mM S100B stimulus in absence of RAGE Ab.).
446 Human Molecular Genetics, 2008, Vol. 17, No. 3
by guest on June 2, 2013
Both STAT/JAK and MAP kinase pathways take part in
response to stress stimuli, such as cytokines, free radical
injury and osmotic shock (44,45). In this respect, the findings
of S100B-induced oxidative stress within DS NPCs could lead
to activation of these kinase pathways, as predicted by the
network analyses. To address whether S100B exposure could
induce activation of these stress activated protein kinases,
control NPCs were pre-incubated in S100B and phosphory-
lation states of the proteins were evaluated by western blot
analyses. We observed a dose-dependent increase in the
expression levels of the activated forms of both the MAP
(p38MAP and JNK) and STAT kinases after incubation in
S100B for 1 h. Furthermore, pre-incubation with the neutraliz-
ing antibody for RAGE before S100B exposure significantly
inhibited S100B-induced p38MAP kinase, JNK and STAT
phosphorylation in a dose-dependent manner (Fig. 6).
Finally, both SB203580 and SP600125, specific and selective
S100B-induced expression of these proteins (data not
shown). Given the constitutive increase in S100B from the
duplication of chromosome 21 in DS, these findings suggested
that overexpression of S100B likely induces ROS generation
and consequent activation of MAPk and STAT/JAK pathways.
Downstream induction of AQP4 in response to S100B
overexpression in DS NPCs
Oxidative stresses and stress response kinase activation are
known to alter cell proliferation, differentiation and apoptosis
(44,45). Each of these developmental functions should be
reflected in downstream activation of non-chromosome 21
genes in the DS NPCs. In evaluating the other highly
scored, predicted IPA networks (#1–3, Fig. 3A), the top func-
tions of these pathways were involved cell cycle (prolife-
ration), cellular function and maintenance (differentiation
and survival) and cellular compromise (apoptosis). Because
the S100B and APP have both been implicated in cell toxicity
from free radical generation, we focused on the second and
third networks. Network #3 already shared significant overlap-
ping interactors with network #4 (including the JAK/STAT
and p38 MAPK molecules). Network #2 could similarly be
linked to network #4 (Supplementary Fig. S4), but several
observations also suggested that proteins in this functional
pathway might serve a compensatory role in DS. Among the
25 focus genes that were differentially expressed in the array
profiling for this network, several genes have previously
demonstrated functions in the inhibition of inflammatory cyto-
kines (GDF15), promoting cytokine degradation (EIF4G1), as
well as regulation of hypoxia inducible factors (CITED2) and
maintenance of water homeostasis (AQP4). As stress kinase
activation has been associated with cytokine, ROS and
osmotic changes, activation of these genes would be consistent
with a secondary response to S100B induction of STAT/JAK
Several additional observations reinforced a potential link
between S100B and AQP4 in DS during brain development.
Both S100B and AQP4 proteins are found predominantly in
glial cell lineages, associated with elevated levels of ROS
and increased in various neuropathological conditions invol-
ving cell injury and death (38–40). Recent studies have also
shown a concomitant increase in S100B and AQP4 after trau-
matic spinal cord by microarray profiling (46,47). To further
explore this predicted relationship, we first confirmed the
increase in AQP4 expression seen on profiling of the DS
NPCs. The mRNA transcripts for AQP4 demonstrated
upwards of a 3-fold increase compared with controls on mul-
tiple probe sets (P , 0.0002 on pairwise comparison). An
increase in both the AQ4 mRNA and protein levels was also
seen in the DS NPCs as well as the human and mouse DS
brain tissues, further confirming upregulation of this water
channel in response to duplication of some subset of gene(s)
on human chromosome 21 (Supplementary Fig. S5). Second,
21-located S100B following induction of differentiation
Figure 6. Activation of the stress activated protein kinases through S100B–RAGE interaction. S100B incubation of control progenitors leads to the phosphoryl-
ation of p38 (A), JNK (B) and STAT (C) pathways. Blocking of S100B with the RAGE receptor antibody inhibits activation of these stress response pathways.
As S100B leads to ROS generation and increases in ROS have previously been shown to activate p38 MAP kinase and JAK/STAT phosphorylation, constitutive
overexpression of S100B in DS NPCs likely promotes free radical production and stress activated protein kinase activation. (?P , 0.05;??P , 0.01;???P ,
0.001 compared with control;ooP , 0.01 andoooP , 0.001 compared with 5 mM S100B stimulus in absence of RAGE Ab).
Human Molecular Genetics, 2008, Vol. 17, No. 3 447
by guest on June 2, 2013
within DS progenitors led to a consequent increase in
expression for the chromosome 18-located AQP4. Within
DS neural progenitors, a time-dependent and parallel increase
in AQP4 and S100B expression was seen up to 48 h after
induction of differentiation (Fig. 7A). Moreover, pre-
incubation with S100B neutralizing antibody almost comple-
tely abolished the time-dependent increase in AQP4 protein,
suggesting a causal relationship independent of neural differ-
entiation. Third, we asked whether S100B and/or ROS
exposure could similarly promote AQP4 expression within
control neural progenitors. By 24h after S100B (0.05–5 mM)
stimulation, both AQP4 transcription and protein expression
was increased in a dose-dependent fashion (Fig. 7B). Concur-
rently, neutralization of the S100B receptor by pre-incubation
with RAGE antibody significantly inhibited S100B-induced
AQP4 mRNA and protein upregulation. A similar induction
of AQP4 expression was achieved by direct stimulation of pro-
genitors with H2O2. Finally, the introduction of free radical
scavengers could mitigate the S100B-induced expression of
AQP4 in both control and DS NPCs (Supplementary
Fig. S6). Collectively, these results suggested that constitutive
overexpression of S100B in DS progenitors led to increased
intracellular ROS and stress activated kinase activation, fol-
lowed by the induction of AQP4 water channels.
AQP4 mitigates S100B-induced programmed cell
death through clearance of ROS
While the IPA networks and cellular studies demonstrated
interactions between S100B, the stress activated kinases and
AQP4, we wanted to understand the functional consequences
of these gene changes. Prior work had indicated that aquaporin
water channels might maintain water homeostasis by facilita-
ting clearance of water from the extravascular space and also
served to transfer intracellular ROS to the extracellular space
(38,39,48). As our observations suggested a role for these
genes in oxidative stress, we asked whether AQP4 could
serve as a compensatory response to the S100B-induced
elevation in ROS seen in the DS NPCs. We measured
changes in free radicals levels caused by S100B after inhibit-
ing AQP4 water channel expression. AQP4 mRNA transcripts
and protein expression were inhibited by siRNA within human
SHSY5Y neuroblastoma cells (Fig. 8) (49). Human SHSY5Y
neuroblastoma cells were used given their high transfection
efficiency. Decreased AQP4 mRNA levels were appreciated
by 24 h with a significant inhibition of AQP4 protein levels
apparent by 48 h. Loss of AQP4 function resulted in a dra-
matic increase in free radicals (ROS and lipid peroxidation
by-products) following S100B exposure, suggesting that
these water channels reduce buildup of intracellular ROS
levels within neural cells.
As increases in free radicals within progenitors cause apop-
tosis (Supplementary Fig. S6), an upregulation in AQP4 could
serve to mitigate the effects of these free radicals and provide
a protective mechanism against cell death. To address this
possibility, we initially examined whether S100B application
to NPCs promoted caspase3 activity (a mediator of apoptosis),
as predicted in the network analyses (Supplementary Fig. S4).
Treatment of NPCs with 5 mM S100B-induced cleavage of
pro-caspase into active caspase3 and promoted cell death, as
evidence by trypan blue incorporation (Fig. 9). As expected,
inhibition of AQP4 through RNA interference led to an
increase in both caspase3 activation and cell death within
SHSY5Y neuroblastoma cells.
Several observations arise from the current studies examining
the cellular and molecular characteristics of human DS neural
progenitors. First, genetic profiling of the human DS NPCs
suggest that these cell lines retain progenitor properties,
differ significantly from age-matched control NPCs and
adopt a more gliocentric progenitor phenotype. Second, gene
ontology comparison analyses further show that networks of
interacting, differentially expressed genes in the DS progenitor
were associated with cell cycle (proliferation), cell function
and maintenance (differentiation and survival) and cellular
compromise (cell death). Moreover, some of these genes
were implicated in oxidative stress toxicity and JAK/STAT
canonical signaling. Third, significant but specific changes in
gene transcriptional levels were seen throughout the genome,
indicating that primary changes in gene expression on chromo-
some 21 lead to secondary, potentially compensatory changes
on non-chromosome 21 genes. For example, the chromosome
21 located S100B could induce ROS generation, stress
response kinase activation and upregulation of chromosome
18 located AQP4 water channel expression. While S100B pro-
moted cell death through ROS generation, AQP4 overexpres-
sion led to a reduction in intracellular ROS and overall cell
death. More broadly, these studies provide some of the first
evidence that NPCs derived from a human disorder can be
used to explore disrupted cellular and molecular mechanisms
that give rise to a disease phenotype.
Distinct functional pathways in DS NPCs
DS NPCs differ from WT control NPCs. Microarray profiling
suggests that the DS progenitors across several different indi-
viduals appear fairly uniform in their gene expression patterns,
but significantly differ on comparison to normal, age-matched
controls. Network analysis further implicates several expected
pathways involving gene expression, cellular growth and pro-
liferation and cell death/compromise. Given duplication of
chromosome 21, DS progenitors might be expected to show
increased expression of genes involved in transcription and
translation (50). While the DS NPCs maintain core expression
of genes associated with self-renewing progenitors, they more
surprisingly share features typically seen in glial progenitors
and upregulate genes involved in cell proliferation. The glio-
centric features may in part arise from a self-selection
process, whereby neuronal progenitors are more susceptible
to cell death or impaired development (35). The similar gene
expression patterns seen across multiple DS as compared to
control NPC lines and the identical culturing conditions,
however, would also argue that intrinsic activation of
genetic pathways in the DS cells are likely responsible for
these changes. In retrospect, induction of glial and cell
growth networks might even be expected as cell injury and
death through ROS can promote cytokine release (such as
448Human Molecular Genetics, 2008, Vol. 17, No. 3
by guest on June 2, 2013
EGF, FGF) and consequent proliferation and astrocytosis
(51,52). Further investigation of the cellular and molecular
mechanisms underlying these differences in progenitor charac-
teristics, however, clearly will be necessary to understand
these processes. Finally, while inflammation and neurodege-
neration are often invoked in the later onset AD seen in DS,
the current findings would suggest that ongoing cell injury
and damage already can be appreciated early in development.
The upregulated expression of genes involved in oxidative
stress and cell death (network #4) provides a partial mecha-
nistic basis for the DS phenotype.
Genome-wide dysregulation of gene expression
in DS NPCs
While primary gene dosage effects have been shown in
various DS tissue (53), the nature and extent of trisomy 21
in effecting secondary disomic genes is unclear. Some
studies have suggested that while changes in transcript levels
occur in the trisomic genes, no pervasive effect is seen on
gene expression on chromosomes other than 21 (54–56).
Alternatively, other work has argued that overexpression of
trisomy 21 can lead to profound disruption of the entire tran-
scriptome (57–59). The present study suggests that trisomic
genes do lead to specific changes in transcript expression on
each of the non-chromosome 21 genes (so termed, global dys-
regulation), but only in fairly specific pathways. Of the
?50 000 transcripts, only 1900 transcripts (3.8%) were
found to be significantly different on pairwise comparison of
multiple control and DS samples. Moreover, of these 1900
transcripts, only 330 transcripts (0.07%) were shown to signifi-
cantly interact on network analysis and the levels of abnormal
expression were generally only 1.5–2-fold in excess/deficit
compared with normal. Thus, our findings are compatible
with both views in that we observe a global dysregulation in
gene expression with increases in non-chromosome 21 genes
Figure 7. S100B induction of AQP4 expression in neural progenitors. (A) Increased S100B protein expression in DS versus control neural progenitors is seen
following induction of neural differentiation and protein extraction for western blot analyses. Neural differentiation is induced by withdrawal of EGF and FGF
and culturing of neural progenitors in serum-containing media. Increases in AQP4 expression can also be inhibited by the addition of soluble antibody to S100B
to the culture media, suggesting a causal relationship (???P , 0.001 versus control NPC at same time point). (B) S100B through its RAGE receptor can promote
transcriptional and translational induction of AQP4 through ROS generation. S100B treatment of control human progenitors induces AQP4 expression at both the
mRNA (left, by RT–PCR) and protein (middle by western blot) levels. Blocking of the RAGE receptors (1:10000 and 1:1000 dilution vol/vol of RAGE anti-
body) leads to a reduction in AQP4 expression. Direct application of H2O2onto neural progenitor cultures leads to a dose-dependent increase in AQP4 expression
(right) and free radical scavengers can block S100B-induced expression of AQP4 (Supplementary Fig. S10), suggesting that increased free radicals due to con-
stitutively overexpressed S100B in the DS neural progenitors contributes to the increase in AQP4 levels. (?P , 0.05;??P , 0.01 and???P , 0.001 compared
with control; 88P , 0.01 and 888P , 0.001 compared with 5 mM S100B stimulus in absence of RAGE Ab.).
Human Molecular Genetics, 2008, Vol. 17, No. 3449
by guest on June 2, 2013
at both the mRNA and protein levels, but find that these
changes occur only in a very specific and selective subset of
genes. These observations would be consistent with the
interpretation that increased expression of genes on chromo-
some 21 within neural cells leads to specific compensatory
changes in expression of genes found off of chromosome 21.
The relatively specific changes seen on the secondary non-
chromosome 21 transcriptional profiling from these studies, as
compared to prior reports, may extend from several factors:
(1) tissue specificity, (2) modest changes seen on genes
expression from DS transcriptional profiling and (3) use of
cultured NPCs as opposed to tissue. First, prior studies demon-
strating no clear changes in secondary transcriptional effects
from DS typically used human or mouse samples, composed
of mixed cell populations from whole brain, cerebellum or
heart tissues (55,56). The current studies rely on temporal,
regional and cellular specificity for mRNA profiling. NPCs
were generated from a fairly narrow window in human cortical
development (gestational ages 19–21 weeks), isolated from a
specific area of the ventricular zone along the frontal cortex,
and expanded by EGF and FGF to generate primary neural
progenitors. In this manner, the secondary changes seen in
non-chromosome 21 genes for the current studies may
reflect specific differences for a particular cell population in
the brain during development. Second, we also find that
1.5–2-fold difference changes in chromosome 21 gene
expression result in only modest fold changes in expression
of non-chromosome 21 genes. These findings are similar to
those by FitzPatrick et al. (57) who used DS amniocyte cul-
tures and found only modest changes in the average level of
transcription. Presumably, the whole tissue containing mixed
cellular populations used in some prior studies might mask
these more subtle findings, given the complexity of gene
expression for different cell types. Finally, transcriptional pro-
filing for these experiments was performed on cultured NPCs.
These experiments used pairwise comparisons between DS
and normal NPCs, which are generated and maintained
under identical conditions in vitro, thereby suggesting that
the observed differences likely have important functional
implications. Moreover, some of the observed changes do cor-
relate with gene and protein expression in tissue such as AQP4
and S100B upregulation. However, the extent to which cell
culture modifies gene transcriptional levels in DS is not
S100B and AQP4 regulation of free radical mediated
cell death in DS NPCs
While overexpression of such proteins as S100B and APP are
thought to contribute to the neurodegenerative features of DS
and AD (19,20,42), we now find that these same processes are
apparent very early in neural development and within the
Figure 8. Inhibition of AQP4 expression leads to increased S100B-induced ROS levels in human SHSY5Y neuroblastoma cells. (A) Decreased expression levels
of AQP4 mRNA can be seen by RT–PCR at 24h following transfection of the siRNA AQP4 construct, as compared to control (Cyclophilin B siRNA) and
untransfected cells. (B) Decreased expression levels of AQP4 protein can be seen by western blot analysis at 48h following transfection of the siRNA
AQP4 construct as compared to control (Cyclophilin B siRNA) and untransfected cells. (C and D) S100B-induced ROS (H2O2and MDA generation) are poten-
tiated with inhibition of AQP4 channel expression, suggesting that the water channels assist in the clearance of free radicals (???P , 0.001).
450 Human Molecular Genetics, 2008, Vol. 17, No. 3
by guest on June 2, 2013
Figure 9. Inhibition of AQP4 expression leads to increased cell death in human SHSY5Y neuroblastoma cells. (A) S100B exposure results in activation of
caspase3 and (B) increased cell death by trypan blue incorporation. Control NPCs are harvested for western blot analysis of caspase3 activity and examined
in culture for trypan blue exclusion following 6 h incubation in 5 mM S100B. (C) S100B can induce a similar activation of caspase activity in human neuroblas-
toma SHSY5Y cells in a dose-dependent fashion. (D) S100B-induced caspase3 activity is increased with inhibition of AQP4 by RNA interference. Differences in
levels of activation of this apoptotic protease following AQP4 inhibition are best appreciated following 0.5 mM S100B incubation for 6 h. (E) S100B-induced
ROS levels and caspase activity correspond to a decline in SHSY5Y cell viability. Inhibition of AQP4 prior to S100B exposure leads to a further decline in cell
viability (???P , 0.001;oooP , 0.001;ooP , 0.01 versus untreated). IC50 ¼ minimum concentration of S100B required to achieve 50% cell viability.
Human Molecular Genetics, 2008, Vol. 17, No. 3451
by guest on June 2, 2013
neural precursor population. More specifically, the induction
of ROS generation through S100B expression seen in the
neural progenitors during development parallels the mecha-
nisms of cell injury postulated for the calcium-binding
protein in neurodegeneration. S100B is constitutively overex-
pressed in glia throughout life given its location on chromo-
some 21(37,42). Activated
calcium-binding protein which activates the receptor for
advanced glycation (RAGE), induces ROS generation and
p38 MAPkinase upregulation (40,60). This presumed inflam-
matory process from S100B is implicated in plaque formation
and neuronal degeneration in AD (36,38,61). A nearly identi-
cal response can be seen in the neural progenitors. That said,
these observations may not be so surprising, given that
insults from many degenerative processes are thought to be
progressive and cumulative.
The increase in S100B and free radicals leads to the conse-
quential upregulation of AQP4 water channels. Consistent
with some role in water clearance and neuroprotection,
increased AQP4 expression has been associated with brain
swelling from strokes, head injuries, brain tumors and brain
abscesses (62). AQP4 expression in glial foot processes near
blood vessels, the neuroependyma (and NPCs) and pial
surface has led to the paradigm that these channels serve to
maintain water balance in the brain by shunting water out of
the extracellular space and into the vasculature or cerebrospinal
fluid (63–66). However, prior reports show that null AQP4
mice do not develop severe brain edema following injury,
raising the question of whether induction of these water chan-
nels after injury is intended for the maintenance of water
homeostasis (67,68). The current network analyses suggest
that AQP4 is linked to mediators of ROS and stress response
kinases. Moreover, cell oxidative stressors appear to be suffi-
cient to induce water channel expression, perhaps in promotion
of cell survival. A similar transfer of intracellular ROS to the
extracellular space has been proposed in plant cells (48).
General role of NPCs derived from CNS developmental
disorders in studying human disease
Several approaches have been used to study the molecular
genetics and pathology of DS. Mouse aneuploidies have
been used to model DS, either trisomy 16 (embryonic
lethal), partial trisomy 16 (trisomy strains Ts65Dn and
Tc1Cje) and most recently, a mouse containing an almost
complete copy of human chromosome 21 (strain Tc1)
(69,70). Chromosome 16 in the mouse contains approximately
two-thirds of the homologous genes found on human chromo-
some 21. An alternative model is now provided by the use of
human DS neural progenitors and neurons. The ability to
expand and generate large numbers of human progenitors
affords the opportunity to study the molecular and cellular
aspects of DS. Human progenitors offer the added advantage
that three complete copies of genes on human chromosome
21 including the upstream and downstream regulatory
elements are replicated in each cell. The complete genetic
makeup of the human cells will likely serve to enhance the
validity of altered molecular pathways observed from
changes in expression of genes found both on and off chromo-
Previous analyses of human DS neural precursors and
neurons in culture have allowed for characterization of some
potential aberrant pathways in this disorder. Primary cortical
neurons from gestational 16–18 week fetal DS brain fail to
survive long-term in comparison to normal cultures, even in
the presence of serum with survival factors, due to increased
ROS, mitochondrial dysfunction and aberrant APP processing
(19,20). More recent studies using DS neural progenitors
showed that SCG10, a neuron-specific growth-associated
protein regulated by the neuron-restrictive silencer factor
REST, was almost undetectable in the DS sample, consistent
with a reduction of cortical neurons in DS brain (21).
However, we noted minimal expression of REST in either
the DS or the control NPC population, in support of findings
reported by Sohn et al. (71). The discrepancy may be due to
different culturing techniques although the current work and
the Bahn et al. (21) studies both rely on EGF and FGF for
NPC propagation. More likely, the current work utilized pro-
genitors derived from a later stage in fetal development,
19–22 weeks as opposed to 8–18 weeks reported in the
prior work. In this respect, our findings could support the
Bahn results, as an early decline in neural precursors, which
have a greater capacity to generate neurons, could give rise
to the neural precursors derived from the current studies,
which exhibit more glial-related properties.
DS is a complex genetic condition arising from an altered
dosage of WT genes on human chromosome 21. As seen in
the current study, variations in gene dosage on chromosome
21 can lead to altered gene expression on non-chromosome
21 genes. These changes outside of chromosome 21 may
reflect both compensatory cyto-protective mechanisms, as
well as effectors of neuropathology. In this respect, and par-
ticularly in the case of the human nervous system, the dis-
rupted signaling cascades in human neural precursors and
neurons may not necessarily be shared across other mamma-
lian species. Isolation of hNPCs as models of human disease
provides a means to functionally test and correct potential
altered cellular and molecular pathways distinct for DS as
well as other various disorders of the CNS.
MATERIALS AND METHODS
Human tissue, ethical and licensing considerations
The study has been approved by the Institutional Review
Board (IRB) at the Beth Israel Deaconess Medical Center
and Brigham and Women’s Hospital. De-identified human dis-
carded tissue was obtained from pathological samples during
autopsy. Neural progenitors from three DS and three age-
matched control brains (gestational ages 19–21 weeks) were
used in this study for each of the experiments described.
Antibodies and reagents
Antibodies used for immunostaining and western blot analyses
are as follows. Anti-GFAP (1:100; DAKO, Carpinteria, CA,
USA); anti-O4, anti-CNPase, anti-NG2 and anti-nestin
452Human Molecular Genetics, 2008, Vol. 17, No. 3
by guest on June 2, 2013
(1:100; Chemicon, Temecula, CA, USA); anti-vimentin and
anti-MAP2 (1:100; Sigma, St. Louis, MO, USA), anti-mouse
AQP4 (1:100,Novus Biologicals),
(1:1000; Acris antibodies, Hiddenhous, Germany) anti-S100B
(1:250; AbCam, Cambridge, UK), p-p38MAPK(1:500; Phar-
mingen, Milan, Italy), anti-pJNK (1:500; Santa Cruz Biotech-
nology, La Jolla, CA, USA), anti-pSTAT (1:500; Santa Cruz
Biotechnology) and anti-caspase3 (1:500; Sigma, Milan, Italy).
Tissue dissociation and isolation. Methodsforventricular zone
dissection and dissociation follow general guidelines used in
murine samples (72). In brief, samples were obtained along
the periventricular zone within the frontal cortex (Supplemen-
tary Fig. S1A), minced and washed in cold Hank’s buffered
saline solution and placed in trypsin solution at 378C for
30 min.Thesamplewasthenstrainedthrougha40 mcellstrainer
(Falcon, San Jose, CA, USA), and washed in Dulbecco’s modi-
fied eagle medium (DMEM) with 10% fetal calf serum (FCS)
(Hyclone, Logan, UT, USA) to inactivate the trypsin. The disso-
ciated cells were spun down, the media aspirated and cells were
placed in at low dilution (1 ? 105–1 ? 106per 5 ml) in neuro-
tures were maintained in a 378C/5% CO2incubator and media
was renewed weekly. To differentiate cells and with initiation
of the culture experiments, NM was replaced by DMEM with
10% FCS or neurobasal/B27 media supplemented with
L-glutamine and streptomycin/penicillin (5 mM).
SHSY5Y cell cultures were maintained as follows. Cells
were cultured in DMEM with 10% fetal bovine serum (FBS)
for 3 days. The FBS content of the culture medium was then
reduced to 5%, and the cells were exposed to 10 mM retinoic
acid. The cells were kept under these conditions for 5 days,
changing the medium every 2 days. Subsequently, the cells
were cultured in Neurobasal/B27 medium supplemented with
B-27, retinoic acid (10 mM), L-glutamine and streptomycin/
penicillin (5 mM).
Genetic analysis. Cytogenetic analysis of the expanded DS
neural precursor population was performed using standard flu-
orescent in situ hybridization techniques against chromosome
21 (73). Labeling of the BAC probes followed standard pro-
cedures with dUTP containing rhodamine and fluoroscein
tags (methods outlined in the Vysis nick translation kit,
Grove, IL, USA). Hybridization was performed by denaturing
the slides in 70% formamide/2?SSC, dehydrating the slides
with serial ethanol washes and applying the probe to the
slide samples. Post-hybridization, the slides were washed, cov-
erslipped and examined under fluorescence microscopy (Zeiss
Axioskop). Analysis of each of the progenitor lines was per-
formed to exclude potential mosaicism in the samples.
Microarray analyses. Methods for Affymetrix gene chip pro-
filing followed previously published work on gene expression
in human NPCs (74). Microarray analysis was performed by
Expression Analysis (Affymetrixw Genomic Processing Faci-
lity in Durham, NC, USA) using the Affymetrix HU 133 plus
2.0 chip. Total RNA (10 mg) was extracted from the neural
precursors, and labeled. The target of labeled cRNA was
hybridized to the GeneChip, expression values were calculated
based on the difference of the perfect match oligos and
mismatch in the probe set, the signal values from each array
was normalized and a data file generated. Data sets were ana-
lyzed by a pairwise comparison and the Wilcoxon’s
signed-rank test to derive biologically meaningful results.
To ensure reproducibility and biological significance, RNA
samples were collected from three DS (gestational age 19–21
weeks) and age-matched control neural progenitor lines
(biological replicates). Samples were obtained within 12 h
postmortem (in collaboration with the Boston hospitals and
in collaboration with Drs Folkerth and Hecht). Data from
methods with Rosetta Resolver software. Statistical signifi-
cance of gene expression differences between neuronal sub-
types was determined by pairwise comparisons at each age
using Significance Analysis of Microarrays (75). Microarray
data from the biological replicates were combined in Rosetta
Resolver for trend plots.
For comparison studies against early and late control
NPCs, APCs and OPCs, data sets from publicly available
domains were used [NCBI, GSM51360-51368 from Dr Clive
Svendesen and Shin and Chesnut (32)]. Relative ratios of
the expression levels for genes in network #4 (APP, CPE,
EPOR, FGF12, GAP43,LIG3, LRP1, JAK1, NAGA,
PICALM, PKD1, PRKCA, S100B, SRC, STAT2, STAT5B,
TM2D1, TUB), cell specific markers (GFAP, MBP, MAPT,
DCX, OMG) and AQP4 were calculated for the respective
NPC lines. Each of the datasets contained early NPC array
profiling, allowing for calculation of relative rates or
increase/decrease on comparison of DS versus early control,
53 week NPC versus early 15 week control and APC/OPC
versus early control. The genes evaluated in network #4 rep-
resented the subset of genes that were found in common
across each of the different array platforms.
Identification of significant pathways in biological processes
and altered interactome of DS NPCs genes. The IPA is a
curated database of previously published findings on mamma-
lian biology from the public literature (Ingenuity Systems).
Reports on individual studies of genes in human, mouse or rat
were first identified from peer-reviewed publications, and find-
ings were then encoded into an ontology by content and mode-
ling experts. Manual extraction and curation probably results in
more specific and comprehensive interactions, with far fewer
false-positives than automated alternatives. The following
steps were used (1): Genes identified as significant from the
experimental data sets were overlaid onto the interactome.
Focus genes were identified as the subset having direct inter-
action(s) with other genes in the database (2). The specificity
of connections for each focus gene was calculated by the per-
centage of its connections to other significant genes. The
initiation and growth of pathways proceeded from genes with
the highest specificity of connections. Each pathway had a
maximum of 40 genes (3). Pathways of highly interconnected
genes were identified by statistical likelihood using equations
as previously described. The IPA database was used according
to the company instructions (www.ingenuity.com).
Real time-PCR protocol. The RT–PCR protocol followed pre-
viously described methods (76). The cDNA templates were
reverse-transcribed in vitro (Invitrogen) using total RNA
Human Molecular Genetics, 2008, Vol. 17, No. 3453
by guest on June 2, 2013
extracted from human DS and control neural progenitors.
Primers for AQP4 were designed to amplify a 200–300 bp
PCR fragment. The human GAPDH gene was used as a
control to normalize the amount of the paired of cDNA tem-
plates (for example, cDNA templates from control and DS
neural progenitors). The PCR reaction containing SYBR
Green I dye was performed in a 96-well plate using the
SYBR Green Master Mix kit (Applied Biosystem, Foster
City, CA, USA). PCR reaction for each cDNA sample was
duplicated to reduce the chance of contamination. GAPDH
was used as a control for every sample. The fluorescent inten-
sity of PCR products was measured and quantified as numbers
of amplification cycles by the ABI Prism 7700 Sequence
Detection System (Applied Biosystem). The difference of
PCR cycles between two cDNA samples for a tested gene
was then calculated into a ratio according to the method
suggested by the ABI Prism 7700 Sequence Detection System.
ELISA protocol. In brief, 96-well plates (Nunc 80040LE 0903)
were coated with AQP4 mouse-Ab (1:300; Acris, SM1811P,
Germany) (50 ml each well) at 48C overnight, washed with
PBS and blocked with 3% bovine serum albumin for 2 h at
room temperature (RT). The wells were washed with PBS,
samples were added (protein extracts from control and DS
NPCs) and the plates were incubated for 2 h at RT. The plate
was washed with PBS five times and incubated with AQP4
Rabbit-Ab (1:300; Chemicon, AB3068) (50 ml each well) for
2 h at RT. Following another PBS wash, the wells were incu-
bated with biotinylated anti-Rabbit Ab (1:300) for 2 h at RT;
washed with PBS and incubated with streptavidin-HRP from
ABC kits (1:50; Vector, Burlingame, CA, USA) for 2 h at
RT. The wells were then washed with PBS and incubated
with ABTS (Roche Diagnostics, Indianapolis, IN, USA,
1204521) substrate for 20 min at RT. The colorimetric change
in each well was measured as a change in the optical density
(OD) value with an ELISA plate reader at 405/490 nm.
Immunocytochemistry. The neural precursors were plated
onto glassslide chambers
(Beckton Dickenson, San Jose, CA, USA) for several hours
and fixed with 4% paraformaldehyde in phosphate-buffered
saline (PBS) or 2808C methanol. The cultures were placed
in blocking solution with PBS containing 10% FCS, 5%
horse serum and 5% goat serum or 10% donkey serum
(Hyclone), incubated overnight in the appropriate antibody
(Acris, aquaporin 4) and processed through standard fluor-
escent secondary antibodies (CY3, Jackson Immunoresearch
Laboratories and FITC, Sigma).
Measurement of ROS formation
The measurement of intracellular H2O2was used as a marker
of ROS formation. H2O2 production was evaluated by the
20,70-dichlorofluorescein (DCF) method according to pre-
viously described methods (40). DS and euploid NPCs, cul-
tured in DMEM with FCS to induce differentiation, were
harvested at various time points (2–48 h) after introduction
of the differentiating media. Cells were detached from Petri
dishes and seeded at a density of 5 ? 103cells/well into
Biochemicals, Costa Mesa, CA, USA) was then added directly
to the medium at a final concentration of 5 mM and the cells
were then incubated for 1 h at 378C. H2DCF-DA is a non-
fluorescent permeable molecule which diffuses passively into
cells; the acetates are then cleaved by intracellular esterases
to form H2DCF which is thereby trapped within the cell. In
the presence of intracellular H2O2, H2DCF is rapidly oxidized
to the highly fluorescent DCF.
In further experiments, euploid neural precursor cells were
cultured, seeded as described above and treated with exogen-
ous S100B (0.05–5 mM) for 0.5 h in the presence or absence of
specific blocking RAGE neutralizing antibody (1:1000–
1:10 000 dil. v/v; R&D Systems, Minneapolis, MN, USA).
After treatment, cells were washed twice with PBS and the
plates placed in a fluorescent microplate reader (LS 55 Lumi-
nescence Spectrometer, Perkin Elmer, Beaconsfield, Bucks,
UK). Fluorescence was monitored using an excitation wave-
length of 490 nm and an emission wavelength of 520 nm.
Results were expressed as relative fluorescence units.
Lipid peroxidation assay
Measurements of cell membrane lipid peroxidation were per-
formed using the thiobarbituric acid (TBA, Sigma) colori-
metric assay, which detects the most abundant lipid
peroxidation product MDA. Methods have been described pre-
viously (77). DS and euploid neural precursor cells, cultured
for different time points (2–48 h) after induction of differen-
tiation, were washed three times with 1? PBS and then det-
rached from the Petri dishes by scraping in 48C PBS. In
further experiments, euploid neural precursor cells were cul-
tured and treated with exogenous S100B (0.05–5 mM) for
1 h in the presence or absence of specific RAGE neutralizing
antibody (1:1000–1:10000 dil. v/v). Cells were then washed
and detached as described above. In both experiments, cells
were lysed by six cycles of freezing and thawing. One ml
10% (w/v) trichloroacetic acid (Sigma) was added to each
450 ml of cellular lysate. After centrifugation at 1000g for
10 min, 1.3 ml 0.5% (w/v) TBA was added and the mixture
was heated at 1008C for 20 min. After cooling, MDA formation
was recorded (Absorbance 530 nm and Absorbance 550 nm) in
a Perkin Elmer (Massachusetts, MA, USA) spectrofluorimeter
and the results are presented as ng MDA/1 ? 106cells.
Protein isolation and western blot analysis
Following the different experimental conditions described
earlier, precursor cells (1 ? 106) were washed twice with
ice-cold PBS and centrifuged at 180g for 10 min at 48C. The
cell pellet was resuspended in 100 ml of ice-cold hypotonic
lysis buffer (10 mM HEPES, 1.5 mM MgCl2, 10 mM KCl,
0.5 mM phenylmethylsulphonylfluoride, 1.5 mg/ml soybean
trypsin inhibitor, 7 mg/ml pepstatin A, 5 mg/ml leupeptin,
0.1 mM benzamidine, 0.5 mM DTT) and incubated on ice for
15 min. The cells were lysed by rapid passage (5–6 times)
through a syringe needle and the samples was centrifuged
for 1 min at 13 000g for 1 min. Protein concentration was
determined and equivalent amounts (100 mg) of each sample
was run by electrophoresis on a 12% discontinuous polyacry-
lamide gel. The proteins were transferred onto nitro-cellulose
454Human Molecular Genetics, 2008, Vol. 17, No. 3
by guest on June 2, 2013
membranes, according to the manufacturer’s instructions
(Bio-Rad, Hercules, CA, USA). The membranes were saturated
by incubation at 48C overnight with 10% non-fat dry milk in
PBS and then incubated with the antibody specified. The mem-
branes were washed three times with 1% Triton 100-X in PBS
and then incubated with anti-rabbit immunoglobulins coupled
to peroxidase (1:1000; DAKO, Golstrup, Denmark). The immu-
nocomplexes were visualized by the ECL chemiluminescence
method (Amersham, Piscataway, NJ, USA). Subsequently, the
relative expression of tau protein in cytosolic fraction was quan-
tified by densitometric scanning of the X-ray films with a GS
700 Imaging Densitometer
program (Molecular Analyst, IBM).
Cell death studies
Cell viability was measured in NPC and DS cells using the
bromide) assay according to the techniques described by
Mosman et al. (78). In brief, cells were plated at the density
of 1 ? 105cells/well in 96 well plates, treated with S100B
(0.005–5 mM), H202 (0.001–1 mM), or staurosporine (1–
625 nM; ICN) for 24 h. After 24 h, 25 ml of MTT (5 mg/ml
in DMEM/Neurobasal 1:1) was added to each well and cells
were incubated for an additional 3 h at 378C. After this inter-
val of time, cells were lysed and dark blue crystals solubilized
with 125 ml of a solution containing 50% (v/v) N,N0-dimethyl
formamide, 20% (w/v) sodium dodecylsulfate with an adjusted
pH of 4.5. The OD of each well was measured with a micro-
equipped with a 620 nm filter. Cell viability was thus calcu-
lated as % of cell viability ¼ (OD treated/OD control) ? 100.
RNA isolation and reverse transcriptase-PCR analysis
The mRNA level of AQP4 protein in euploid neural precursor
cells was determined using the semiquantitative RT–PCR
method (Invitrogen, Milan, Italy). Total mRNA was extracted
from cells by use of an ultrapure TRIzol reagent (Gibco BRL,
Carlsbad, CA, USA) as directed by the manufacturer. The con-
centration and purity of total RNA were determined from the
A260/A280 ratio using a UV spectrophotometer (DU 40,
Beckman). The primers sequences used for PCR amplification
were sense 50-TTGTGGCAACTGAAGATGGA-30, antisense
30-CTGCTCTTATGGGGCAATCT-50, and sense 50-ATGAA-
GATCCTGACCGCGCGT-30, antisense: 50-AACGCAGCT-
CAGTAACAGTCCG-30, for b-actin. 1 mg of total RNA
from each specimen was subjected to RT–PCR. RT–PCR
was carried out using a SuperScript TM One-step RT–PCR
with Platinum Taq Kit (Invitrogen, Carlsbad, CA, USA) in a
total reaction volume of 25 ml, containing 2? reaction mix
(12.5 ml), 25 mmol/l sense primer (0.5 ml), 25 mmol/l anti-
sense primer (0.5 ml), RT–PCR platinum Taq mix (0.5 ml),
and autoclaved distilled water. The b-actin and AQP4 PCR
products were electrophoresised on 1% agarose gel and visu-
alized by staining with ethidium bromide. The integrated
density values of the bands representing amplified products
were acquired and analyzed by GS 700 Imaging Densitometer
(Bio-Rad) and a computer program (Molecular Analyst IBM).
Results were expressed as the mean+SEM of n experiments.
Statistical analysis was determined with ANOVA and multiple
comparison were performed with Bonferroni’s test, with
P , 0.05 considered significant. The cell biological exper-
iments were done in each of three control and DS neural
lines, isolated from different individuals and results were con-
siRNA synthesis and transfection of siRNA
RNA duplexes of 21 nucleotides specific for human AQP4
sequence were synthesized by Dharmacon Research, Inc.
(Lafayette, CO, USA). The methods followed previously
published procedures (49). AQP4 sense sequence was
siRNA was used as control siRNA. SHSY5Y cells were
seeded the day before transfection using an appropriate
medium with 10% FBS without antibiotics. Transient transfec-
tion of siRNAs was carried out using Oligofectamine (Invitro-
gen), using the protocol suggested by the manufacturer.
Briefly, the appropriate amount of Oligofectamine was
diluted 1:5 in OPTIMEMw medium (Invitrogen) and incu-
bated at RT for 10 min. In parallel, siRNAs were diluted 1:9
in OPTIMEM medium. The two mixtures were combined
and incubated for 20 min at RT for complex formation.
After addition of the appropriate growth medium with 10%
FCS and without antibiotics, the entire mixture was added to
the cells. Specific silencing was confirmed by at least three
independent western blot and RT–PCR experiments.
Supplementary Material is available at HMG Online.
We also wish to thank Dr Roger Reeves for kindly providing
tissue sections from trisomy 16 mice.
Conflict of Interest statement: None declared.
This work was supported by grants to V.L.S from the Julian
and Carol Cohen and the Ellison Foundation. V.L.S. is a
Beckman Young Investigator and Doris Duke Clinical Scien-
tist Development Award recipient.
1. Renfranz, P.J., Cunningham, M.G. and McKay, R.D. (1991)
Region-specific differentiation of the hippocampal stem cell line HiB5
upon implantation into the developing mammalian brain. Cell, 66, 713–
2. Reynolds, B.A. and Weiss, S. (1992) Generation of neurons and astrocytes
from isolated cells of the adult mammalian central nervous system.
Science, 255, 1707–1710.
Human Molecular Genetics, 2008, Vol. 17, No. 3455
by guest on June 2, 2013
3. Gage, F.H. (2000) Mammalian neural stem cells. Science, 287, 1433–
4. Snyder, E.Y., Deitcher, D.L., Walsh, C., Arnold-Aldea, S., Hartwieg, E.A.
and Cepko, C.L. (1992) Multipotent neural cell lines can engraft and
participate in development of mouse cerebellum. Cell, 68, 33–51.
5. Snyder, E.Y. and Macklis, J.D. (1995) Multipotent neural progenitor or
stem-like cells may be uniquely suited for therapy for some
neurodegenerative conditions. Clin. Neurosci, 3, 310–316.
6. Ourednik, V., Ourednik, J., Flax, J.D., Zawada, W.M., Hutt, C., Yang, C.,
Park, K.I., Kim, S.U., Sidman, R.L., Freed, C.R. et al. (2001) Segregation
of human neural stem cells in the developing primate forebrain. Science,
7. Sheen, V.L., Ferland, R.J., Harney, M., Banham, A., Brown, P., Chenn,
A., Corbo, J., Hecht, J., Folkerth, R. and Walsh, C.A. (2006) Impaired
proliferation and migration in human Miller Dieker neural precursors.
Ann. Neurol, 60, 137–144.
8. Bhattacharyya, A. and Svendsen, C.N. (2003) Human neural stem cells: a
new tool for studying cortical development in Down’s syndrome. Genes
Brain Behav., 2, 179–186.
9. Scheffler, B., Edenhofer, F. and Brustle, O. (2006) Merging fields: stem
cells in neurogenesis, transplantation, and disease modeling. Brain.
Pathol, 16, 155–168.
10. Di Giorgio, F.P., Carrasco, M.A., Siao, M.C., Maniatis, T. and Eggan, K.
(2007) Non-cell autonomous effect of glia on motor neurons in an
embryonic stem cell-based ALS model. Nat. Neurosci, 10, 608–614.
11. Coyle, J.T., Oster-Granite, M.L. and Gearhart, J.D. (1986) The
neurobiologic consequences of Down syndrome. Brain Res. Bull, 16,
12. Scott, B.S., Becker, L.E. and Petit, T.L. (1983) Neurobiology of Down’s
syndrome. Prog. Neurobiol, 21, 199–237.
13. Becker, L., Mito, T., Takashima, S. and Onodera, K. (1991) Growth and
development of the brain in Down syndrome. Prog. Clin. Biol. Res, 373,
14. Schmidt-Sidor, B., Wisniewski, K.E., Shepard, T.H. and Sersen, E.A.
(1990) Brain growth in Down syndrome subjects 15 to 22 weeks of
gestational age and birth to 60 months. Clin. Neuropathol, 9, 181–190.
15. Wisniewski, K.E. (1990) Down syndrome children often have brain with
maturation delay, retardation of growth, and cortical dysgenesis.
Am. J. Med. Genet. Suppl, 7, 274–281.
16. Ross, M.H., Galaburda, A.M. and Kemper, T.L. (1984) Down’s
syndrome: is there a decreased population of neurons? Neurology, 34,
17. Weitzdoerfer, R., Dierssen, M., Fountoulakis, M. and Lubec, G. (2001)
Fetal life in Down syndrome starts with normal neuronal density but
impaired dendritic spines and synaptosomal structure. J. Neural. Transm.
18. Becker, L.E., Armstrong, D.L. and Chan, F. (1986) Dendritic atrophy in
children with Down’s syndrome. Ann. Neurol, 20, 520–526.
19. Busciglio, J. and Yankner, B.A. (1995) Apoptosis and increased
generation of reactive oxygen species in Down’s syndrome neurons in
vitro. Nature, 378, 776–779.
20. Busciglio, J., Pelsman, A., Wong, C., Pigino, G., Yuan, M., Mori, H. and
Yankner, B.A. (2002) Altered metabolism of the amyloid beta precursor
protein is associated with mitochondrial dysfunction in Down’s syndrome.
Neuron, 33, 677–688.
21. Bahn, S., Mimmack, M., Ryan, M., Caldwell, M.A., Jauniaux, E., Starkey,
M., Svendsen, C.N. and Emson, P. (2002) Neuronal target genes of the
neuron-restrictive silencer factor in neurospheres derived from fetuses
with Down’s syndrome: a gene expression study. Lancet, 359, 310–315.
22. Ramalho-Santos, M., Yoon, S., Matsuzaki, Y., Mulligan, R.C. and
Melton, D.A. (2002) ‘Stemness’: transcriptional profiling of embryonic
and adult stem cells. Science, 298, 597–600.
23. D’Amour, K.A. and Gage, F.H. (2003) Genetic and functional differences
between multipotent neural and pluripotent embryonic stem cells. Proc.
Natl. Acad. Sci. USA, 100 (Suppl. 1), 11866–11872.
24. Ellis, P., Fagan, B.M., Magness, S.T., Hutton, S., Taranova, O., Hayashi,
S., McMahon, A., Rao, M. and Pevny, L. (2004) SOX2, a persistent
marker for multipotential neural stem cells derived from embryonic stem
cells, the embryo or the adult. Dev. Neurosci, 26, 148–165.
25. Graham, V., Khudyakov, J., Ellis, P. and Pevny, L. (2003) SOX2
functions to maintain neural progenitor identity. Neuron, 39, 749–765.
26. Hermanson, O., Jepsen, K. and Rosenfeld, M.G. (2002) N-CoR controls
differentiation of neural stem cells into astrocytes. Nature, 419, 934–939.
27. Regad, T., Roth, M., Bredenkamp, N., Illing, N. and Papalopulu, N.
(2007) The neural progenitor-specifying activity of FoxG1 is
antagonistically regulated by CKI and FGF. Nat. Cell. Biol, 9, 531–540.
28. Corbo, J.C., Deuel, T.A., Long, J.M., LaPorte, P., Tsai, E.,
Wynshaw-Boris, A. and Walsh, C.A. (2002) Doublecortin is required in
mice for lamination of the hippocampus but not the neocortex.
J. Neurosci, 22, 7548–7557.
29. Gleeson, J.G., Lin, P.T., Flanagan, L.A. and Walsh, C.A. (1999)
Doublecortin is a microtubule-associated protein and is expressed widely
by migrating neurons. Neuron, 23, 257–271.
30. Doetsch, F., Caille, I., Lim, D.A., Garcia-Verdugo, J.M. and
Alvarez-Buylla, A. (1999) Subventricular zone astrocytes are neural stem
cells in the adult mammalian brain. Cell, 97, 703–716.
31. Sanai, N., Tramontin, A.D., Quinones-Hinojosa, A., Barbaro, N.M.,
Gupta, N., Kunwar, S., Lawton, M.T., McDermott, M.W., Parsa, A.T.,
Manuel-Garcia Verdugo, J. et al. (2004) Unique astrocyte ribbon in adult
human brain contains neural stem cells but lacks chain migration. Nature,
32. Shin, S., Sun, Y., Liu, Y., Khaner, H., Svant, S., Cai, J., Xu, Q.X.,
Davidson, B.P., Stice, S.L., Smith, A.K. et al. (2007) Whole genome
analysis of human neural stem cells derived from embryonic stem cells
and stem and progenitor cells isolated from fetal tissue. Stem Cells, 25,
33. Calvano, S.E., Xiao, W., Richards, D.R., Felciano, R.M., Baker, H.V.,
Cho, R.J., Chen, R.O., Brownstein, B.H., Cobb, J.P., Tschoeke, S.K. et al.
(2005) A network-based analysis of systemic inflammation in humans.
Nature, 437, 1032–1037.
34. Behar, T.N. and Colton, C.A. (2003) Redox regulation of neuronal
migration in a Down Syndrome model. Free Radic. Biol. Med, 35, 566–
35. Mazur-Kolecka, B., Golabek, A., Nowicki, K., Flory, M. and Frackowiak,
J. (2005) Amyloid-beta impairs development of neuronal progenitor cells
by oxidative mechanisms. Neurobiol. Aging, 27, 1181–1192.
36. Schmidt, S. (1998) S100B: pathogenetic and pathophysiologic
significance in neurology. Nervenarzt, 69, 639–646.
37. Allore, R., O’Hanlon, D., Price, R., Neilson, K., Willard, H.F., Cox, D.R.,
Marks, A. and Dunn, R.J. (1988) Gene encoding the beta subunit of S100
protein is on chromosome 21: implications for Down syndrome. Science,
38. Mrak, R.E. and Griffin, W.S. (2004) Trisomy 21 and the brain.
J. Neuropathol. Exp. Neurol, 63, 679–685.
39. Cavazzin, C., Ferrari, D., Facchetti, F., Russignan, A., Vescovi, A.L., La
Porta, C.A. and Gritti, A. (2006) Unique expression and localization of
aquaporin-4 and aquaporin-9 in murine and human neural stem cells and
in their glial progeny. Glia, 53, 167–181.
40. Esposito, G., De Filippis, D., Cirillo, C., Sarnelli, G., Cuomo, R. and
Iuvone, T. (2005) The astroglial-derived S100beta protein stimulates the
expression of nitric oxide synthase in rodent macrophages through p38
MAP kinase activation. Life Sci, 78, 2707–2715.
41. Reeves, R.H., Yao, J., Crowley, M.R., Buck, S., Zhang, X., Yarowsky, P.,
Gearhart, J.D. and Hilt, D.C. (1994) Astrocytosis and axonal proliferation
in the hippocampus of S100b transgenic mice. Proc. Natl. Acad. Sci. USA,
42. Griffin, W.S., Sheng, J.G., McKenzie, J.E., Royston, M.C., Gentleman,
S.M., Brumback, R.A., Cork, L.C., Del Bigio, M.R., Roberts, G.W. and
Mrak, R.E. (1998) Life-long overexpression of S100beta in Down’s
syndrome: implications for Alzheimer pathogenesis. Neurobiol. Aging,
43. Stabel-Burow, J., Kleu, A., Schuchmann, S. and Heinemann, U. (1997)
Glutathione levels and nerve cell loss in hippocampal cultures from
trisomy 16 mouse–a model of Down syndrome. Brain Res, 765, 313–318.
44. Hebenstreit, D., Horejs-Hoeck, J. and Duschl, A. (2005) JAK/
STAT-dependent gene regulation by cytokines. Drug News Perspect, 18,
45. Pearson, G., Robinson, F., Beers Gibson, T., Xu, B.E., Karandikar, M.,
Berman, K. and Cobb, M.H. (2001) Mitogen-activated protein (MAP)
kinase pathways: regulation and physiological functions. Endocr. Rev, 22,
46. Nesic, O., Lee, J., Ye, Z., Unabia, G.C., Rafati, D., Hulsebosch, C.E. and
Perez-Polo, J.R. (2006) Acute and chronic changes in aquaporin 4
expression after spinal cord injury. Neuroscience, 143, 779–792.
47. Nesic, O., Lee, J., Johnson, K.M., Ye, Z., Xu, G.Y., Unabia, G.C., Wood,
T.G., McAdoo, D.J., Westlund, K.N., Hulsebosch, C.E. et al. (2005)
456Human Molecular Genetics, 2008, Vol. 17, No. 3
by guest on June 2, 2013
Transcriptional profiling of spinal cord injury-induced central neuropathic Download full-text
pain. J. Neurochem, 95, 998–1014.
48. Henzler, T. and Steudle, E. (2000) Transport and metabolic degradation of
hydrogen peroxide in Chara corallina: model calculations and
measurements with the pressure probe suggest transport of H(2)O(2)
across water channels. J. Exp. Bot, 51, 2053–2066.
49. Nicchia, G.P., Frigeri, A., Liuzzi, G.M. and Svelto, M. (2003) Inhibition
of aquaporin-4 expression in astrocytes by RNAi determines alteration in
cell morphology, growth, and water transport and induces changes in
ischemia-related genes. Faseb J, 17, 1508–1510.
50. Lyle, R., Gehrig, C., Neergaard-Henrichsen, C., Deutsch, S. and
Antonarakis, S.E. (2004) Gene expression from the aneuploid
chromosome in a trisomy mouse model of Down syndrome. Genome Res,
51. Coates, P.J., Lorimore, S.A. and Wright, E.G. (2005) Cell and tissue
responses to genotoxic stress. J. Pathol, 205, 221–235.
52. Hagberg, H. and Mallard, C. (2005) Effect of inflammation on central
nervous system development and vulnerability. Curr. Opin. Neurol, 18,
53. Mao, R., Zielke, C.L., Zielke, H.R. and Pevsner, J. (2003) Global
up-regulation of chromosome 21 gene expression in the developing Down
syndrome brain. Genomics, 81, 457–467.
54. Mao, R., Wang, X., Spitznagel, E.L., Jr, Frelin, L.P., Ting, J.C., Ding, H.,
Kim, J.W., Ruczinski, I., Downey, T.J. and Pevsner, J. (2005) Primary and
secondary transcriptional effects in the developing human Down
syndrome brain and heart. Genome Biol, 6, R107.
55. Amano, K., Sago, H., Uchikawa, C., Suzuki, T., Kotliarova, S.E., Nukina,
N., Epstein, C.J. and Yamakawa, K. (2004) Dosage-dependent expression
of genes in the trisomic region of Ts1Cje mouse model for Down
syndrome. Hum. Mol. Genet, 13, 1333–1340.
56. Dauphinot, L., Lyle, R., Rivals, I., Dang, M.T., Moldrich, R.X., Golfier,
G., Ettwiller, L., Toyama, K., Rossier, J., Personnaz, L. et al. (2005) The
cerebellar transcriptome during postnatal development of the Ts1Cje
mouse, a segmental trisomy model for Down syndrome. Hum. Mol. Genet,
57. FitzPatrick, D.R., Ramsay, J., McGill, N.I., Shade, M., Carothers, A.D.
and Hastie, N.D. (2002) Transcriptome analysis of human autosomal
trisomy. Hum. Mol. Genet, 11, 3249–3256.
58. Tang, Y., Schapiro, M.B., Franz, D.N., Patterson, B.J., Hickey, F.J.,
Schorry, E.K., Hopkin, R.J., Wylie, M., Narayan, T., Glauser, T.A. et al.
(2004) Blood expression profiles for tuberous sclerosis complex 2,
neurofibromatosis type 1, and Down’s syndrome. Ann. Neurol, 56, 808–
59. Saran, N.G., Pletcher, M.T., Natale, J.E., Cheng, Y. and Reeves, R.H.
(2003) Global disruption of the cerebellar transcriptome in a Down
syndrome mouse model. Hum. Mol. Genet, 12, 2013–2019.
60. Cecil, D.L., Johnson, K., Rediske, J., Lotz, M., Schmidt, A.M. and
Terkeltaub, R. (2005) Inflammation-induced chondrocyte hypertrophy is
driven by receptor for advanced glycation end products. J. Immunol, 175,
61. Lam, A.G., Koppal, T., Akama, K.T., Guo, L., Craft, J.M., Samy, B.,
Schavocky, J.P., Watterson, D.M. and Van Eldik, L.J. (2001) Mechanism
of glial activation by S100B: involvement of the transcription factor
NFkappaB. Neurobiol. Aging, 22, 765–772.
62. Papadopoulos, M.C., Krishna, S. and Verkman, A.S. (2002) Aquaporin
water channels and brain edema. Mt. Sinai J. Med, 69, 242–248.
63. La Porta, C.A., Gena, P., Gritti, A., Fascio, U., Svelto, M. and Calamita,
G. (2005) Adult murine CNS stem cells express aquaporin channels. Biol.
64. Badaut, J., Lasbennes, F., Magistretti, P.J. and Regli, L. (2002)
Aquaporins in brain: distribution, physiology, and pathophysiology.
J. Cereb. Blood Flow Metab, 22, 367–378.
65. Venero, J.L., Vizuete, M.L., Machado, A. and Cano, J. (2001) Aquaporins
in the central nervous system. Prog. Neurobiol, 63, 321–336.
66. Griesdale, D.E. and Honey, C.R. (2004) Aquaporins and brain edema.
Surg. Neurol, 61, 418–421.
67. Ma, T., Yang, B., Gillespie, A., Carlson, E.J., Epstein, C.J. and Verkman,
A.S. (1997) Generation and phenotype of a transgenic knockout mouse
lacking the mercurial-insensitive water channel aquaporin-4. J. Clin.
Invest, 100, 957–962.
68. Manley, G.T., Fujimura, M., Ma, T., Noshita, N., Filiz, F., Bollen, A.W.,
Chan, P. and Verkman, A.S. (2000) Aquaporin-4 deletion in mice reduces
brain edema after acute water intoxication and ischemic stroke. Nat. Med,
69. Kola, I. and Hertzog, P.J. (1998) Down syndrome and mouse models.
Curr. Opin. Genet. Dev, 8, 316–321.
70. O’Doherty, A., Ruf, S., Mulligan, C., Hildreth, V., Errington, M.L.,
Cooke, S., Sesay, A., Modino, S., Vanes, L., Hernandez, D. et al. (2005)
An aneuploid mouse strain carrying human chromosome 21 with Down
syndrome phenotypes. Science, 309, 2033–2037.
71. Sohn, S.Y., Weitzdoerfer, R., Mori, N. and Lubec, G. (2003)
Transcription factor REST dependent proteins are comparable between
Down syndrome and control brains: challenging a hypothesis. J. Neural.
Transm. Suppl, 59–66.
72. Sheen, V.L. and Macklis, J.D. (1995) Targeted neocortical cell death in
adult mice guides migration and differentiation of transplanted embryonic
neurons. J. Neurosci, 15, 8378–8392.
73. Sheen, V.L., Wheless, J.W., Bodell, A., Braverman, E., Cotter, P.D.,
Rauen, K.A., Glenn, O., Weisiger, K., Packman, S., Walsh, C.A. et al.
(2003) Periventricular heterotopia associated with chromosome 5p
anomalies. Neurology, 60, 1033–1036.
74. Wright, L.S., Li, J., Caldwell, M.A., Wallace, K., Johnson, J.A. and
Svendsen, C.N. (2003) Gene expression in human neural stem cells:
effects of leukemia inhibitory factor. J. Neurochem, 86, 179–195.
75. Tusher, V.G., Tibshirani, R. and Chu, G. (2001) Significance analysis of
microarrays applied to the ionizing radiation response. Proc. Natl. Acad.
Sci. USA, 98, 5116–5121.
76. Sun, T., Patoine, C., Abu-Khalil, A., Visvader, J., Sum, E., Cherry, T.J.,
Orkin, S.H., Geschwind, D.H. and Walsh, C.A. (2005) Early asymmetry
of gene transcription in embryonic human left and right cerebral cortex.
Science, 308, 1794–1798.
77. Mihara, M. and Uchiyama, M. (1978) Determination of malonaldehyde
precursor in tissues by thiobarbituric acid test. Anal. Biochem, 86, 271–278.
78. Mosmann, T. (1983) Rapid colorimetric assay for cellular growth and
survival: application to proliferation and cytotoxicity assays. J. Immunol.
Methods, 65, 55–63.
Human Molecular Genetics, 2008, Vol. 17, No. 3457
by guest on June 2, 2013