|
Items in cart [1] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 20
Colloquium
Screening for mammalian neural genes via
fluorescence-activated cell sorter purification of
neural precursors from Sox7-gfp knock-in mice
Jerome Aubert*t, Marios P. Stavridis*t, Susan Tweedie*, Michelle O'Reilly*, Klemens Vierlingert, Meng Li*,
Peter Ghazalt, Tom Pratt§, John 0. Mason§, Douglas Royt, and Austin Smithson
*Institute for Stem Cell Research, University of Edinburgh, King's Buildings, West Mains Road, EH9 3JQ Edinburgh, Scotland; tThe Scottish Centre for
Genomic Technology and Informatics, University of Edinburgh, Chancellor's Building, 49 Little France Crescent, EH16 4SB Edinburoh, Scotland; and
§Division of Biomedical Sciences, University of Edinburgh, Hugh Robson Building, George Square, EH8 9XD Edinburgh, Scotland
The transcription factor Sox1 is the earliest and most specific
known marker for mammalian neural progenitors. During fetal
development, Sox1 is expressed by proliferating progenitor cells
throughout the central nervous system and in no tissue but the
lens. We generated a reporter mouse line in which egfp is inserted
into the Sox1 locus. Sox1GFP animals faithfully recapitulate the
expression of the endogenous gene. We have used the GFP
reporter to purify neuroepithelial cells by fluorescence-activated
cell sorting from embryonic day 10.5 embryos. RNAs prepared from
Sox16FP+ and Sox1GFP- embryo cells were then used to perform a
pilot screen of subtracted cDNAs prepared from differentiating
embryonic stem cells and arrayed on a glass chip. Fifteen unique
differentially expressed genes were identified, all previously as-
sociated with fetal or adult neural tissue. Whole mount in situ
hybridization against two genes of previously unknown embry-
onic expression, Lrrn1 and Musashi2, confirmed the selectivity of
this screen for early neuroectodermal markers.
Neural stem cells are promising candidates for the development
of cellular and genetic therapies for neurodegenerative disor-
ders such as Parkinson's disease and Huntington's disease (1), and
for creation of in vitro drug discovery and toxicological screens (24.
However, the biomedical application of neural stem cells will
require the generation of large homogenous populations of these
cells in vitro. One source of neural stem cells is embryonic stem (ES)
cells (34. ES cells are derived from the inner cell mass of the
preimplantation blastocyst-stage embryo and can be propagated
indefinitely in an undifferentiated, pluripotent state (44. The for-
mation of multicellular aggregates called embryoid bodies permits
the commitment of pluripotent ES cells into multiple cellular
lineages in vitro (5), mimicking aspects of cellular differentiation in
early embryos (6~. This provides a powerful system for the discovery
of genes induced early during development and for functional
validation of candidate genes (74. A favored protocol for the
commitment of ES into neural lineage is the treatment of embryoid
bodies with all-trans retinoic acid (~-104. After induction and
outgrowth onto an adhesive substratum, up to 50% of cells express
the neural precursor markers Soxl and Sox2 and can generate
neurons and glia (114.
Key advances in defining the optimal conditions for generating
and propagating neural stem cells are likely to come from a
proper understanding of the molecular mechanisms controlling
the fate decisions of pluripotent cells and of fetal and ES
cell-derived neural precursors. Here we describe a refined
approach to identify genes induced during neural specification
and/or maintained in neural progenitor cells in vivo and in vitro.
Transgenic mice (SoXlGFP) were generated in which the en-
hanced GFP (e~p) reporter is inserted into the Soxl gene via
gene targeting (12~. Soxl is the earliest specific marker of neural
11836-11841 1 PNAS 1 September 30, 2003 1 vol. 100 1 suppl. 1
precursors in the mouse embryo (134. It is present in prolifer-
ating neural precursors from the neural plate stage onwards. The
only other site of expression in the mid-gestation fetus is the lens
(Fig. L43. Exit from mitosis and neuronal or glial differentiation
is accompanied by down-regulation of Soxl (144. The GFP
knock-in allows the visualization of Soxl expression in these
animals by fluorescence microscopy, and the purification of
Soxl -positive cells by f luorescence-activated cell sorting
(FACS). This purification allows preparation of RNAs for highly
selective differential screening of microarrays. We have tested
this approach by application to a custom microarray of a
subtractive cDNA population prepared from retinoic acid-
induced embryoid bodies.
Materials and Methods
Gene Targeting. A phage genomic DNA library from 129/Ola
strain mouse was screened with a 2-kb probe containing the Soxl
ORF (generously provided by Larysa Pevny). From a resulting
phage containing 12 kb of Soxl genomic sequence, 5.5- and
2.5-kb fragments flanking the Soxl ORF were taken as 5' and 3'
homology arms to prepare a targeting vector. The gene for e~p
was fused in-frame into the second of three consecutive ATGs
at the Soxl translation initiation site (15, 16) via PCR. The fusion
product, linked via an internal ribosome entry site (IRKS) (17)
to the gene encoding puromycin acetyltransferase (pac), was
cloned between the homology arms. A cytomegalovirus promot-
er-driven hygromycinR-thymidine kinase dual selection cassette
(18) flanked by loxP sites was inserted downstream of the
GFPirespac cassette (Fig. 1A). After electroporation in E14Tg2a
ES cells and selection in hygromycin, three targeted clones were
identified by Southern analysis with flanking 5' and 3' probes,
and unique integration was confirmed for two of these by using
an egfp probe. Transient transfection with a Cre recombinase
expression vector was used to remove the selection cassette.
Clones that had undergone excision were selected for in the
presence of ganciclovir and screened by Southern analysis (Fig.
1B). One such clone (46C) was injected into blastocysts and
This paper results from the Arthur M. Sackier Colioquium of the Nationai Academy of
Sciences, "Regenerative Medicine," held October 18-22, 2002, at the Arnoid and Mabei
Beckman Center of the National Academies of Science and Engineering in irvine, CA.
Abbreviations: ES, embryonic stem; FACS, fluorescence-activated ceil sorting; En, embry-
onic day n.
~J.A. and M.P.S. contributed equally to this work.
~A.S. is a scientific adviser to Stem Cell Sciences, Ltd., and holds nonvoting equity in the
company. Stem Cell Sciences funds research in the laboratory and has patents granted and
pending on technology used in this article.
IlTo whom correspondence should be addressed. E-mail: austin.smith@ed.ac.uk.
2003 by The National Academy of Sciences of the USA
www.pnas.org/cgi/doi/10.1 073/pnas.1 734197100
OCR for page 21
B
C 14 46 ~ cat
~ Sox1 ORf it/ ~ l`= ~ 8~t
_ | . Genomic locus ~
i;;= ~ Aft\ ~ EcoRI: 12Kb
X ~71ntemay ~ Lox\
~=~ Targeting vector
egfiP pllC C~V try-=
En
Xbal Xbal Score I
1 ~ 1 Targeted locus On
deal: 9Kb ~~ ~ ~ ~
EcoRI: SKb CO 14 46 53 46C
+ Cry —~ ~ ~~> _ .~ ~ ~ 1 2Kb
L_' ,_ 46C locus ~-~ ~
6.6Kb
_ _ ~ 5Kb
TV
14 46 S3 cot
—1~~ _ ~ 9Kb
Fig. 1. Targeting Sox1 and Sox1GFP expression. (A) In situ hybridization of an E10.5 embryo with a Sox1 riboprobe, showing expression restricted to and
throughout the neuraxis. (B) Schematic showing the design of the targeting vector and the screening strategy for identification of correctly targeted clones. (C)
Southern blots showing the correct targeting events and the excision of the cytomegalovirus (CMV) HyTK cassette after transient transfection with a Cre
expression plasmid. (Top) 5' probe. (Middle) Internal EGFP probe. (Bottom) 3' probe. Clone 14 is correctly targeted but has multiple integrations; clones 46 and
53 are correctly targeted, single integration clones. Clone 46C is a derivative of clone 46 after Cre-mediated deletion of the CMV Hy-TK cassette. Control is DNA
from the parental E14Tg2a ES cells. (D) Sox1GFP expression in embryos and adult animals. (] Sox1GFP embryo at E9.5 showing expression throughout the length
of the neural tube. (i/) Dorsal anterior view of a Sox1GFP embryo at E11.5 showing exclusion of GFP fluorescence from the midline, indicating no expression in
roofplate and floorplate. (ii/) Coronal section through the head of an E12.5 embryo showing Sox1GFP expression in the ventricular and subventricular zone and
the lens. (iv) Section through an adult SoxlGFP mouse brain, showing Sox1GFP-expressing cells in the subgranular layer of the hippocampus. (Inset) Higher
magnification of boxed area, showing a GFP-expressing cell in the granular layer.
Tissue Preparation. For analysis of SoxlGFP expression, heterozy-
gous Sox1GFP males were crossed with wild-type females. Midday
after vaginal plug was considered as embryonic day 0.5 (E0.54.
Females were killed by cervical dislocation, and the embryos
were dissected free of the uterus, washed in PBS, and observed
under a fluorescence microscope. For cryosectioning, embryos
were fixed in 4% paraformaldehyde (PFA), cryoprotected in
30% sucrose in PBS, and embedded in OCT compound before
cryosectioning at 10-12 ,um. Adult (4-7 weeks old) heterozy-
gous brains were dissected out and fixed in 4% PEA before
embedding in 2% agarose in PBS and sectioning at 50 ,um by
using a vibratome. Sections were counterstained with propidium
iodide and analyzed by confocal microscopy.
passed through the germ line of chimeras to generate the RT-PCR.To eliminate contaminating genomic DNA, 1 ,ugof total
Soxl GFP mouse line. Mice were maintained on a mixed 129xMF1 RNA was treated with 1 unit of DNase I (GIBCO/BRL) for 15
background by breeding of heterozygotes to outbred MF1 mice. min at 25°C. DNase I was inactivated with 25 mM EDTA (pH
8.0, GIBCO/BRL) at 65°C for 10 min and chilled on ice. First
strand random-primed cDNA was synthesized by using Super-
script II Preamplification System (GIBCO/BRL) as described
by the supplier. The cDNA was analyzed by PCR amplification
using individual primer pairs for specific marker genes. The PCR
cycling sequence used was 94.0°C for 3 min. followed by 20-35
cycles of 94.0°C for 30 s, 58.0-60.0°C for 30 s and 72.0°C for 1
min. This was followed by a final extension time of 7 min. All
PCR samples were analyzed by electrophoresis on No agarose
gels stained with ethidium bromide.
FACS Purification and RNA Preparation. E10.5 Soxl-GFP-positive
embryos were dissected free of extraembryonic membranes,
digested in 0.1% trypsin, and resuspended in cold 10% FCS in
PBS. The cells were sorted by flow cytometry to give two cell
populations, Soxl-GFP-positive (SoxlGFP+) cells and Soxl-
GFP-negative (SoxlGFP-) cells. The sample was kept cold at all
times to minimize RNA degradation and cell death during
sorting. Viable cells were gated by their forward and side scatter
characteristics, and gates were set to sort positive and negative
cell populations.
Total RNA was extracted from both cell populations by using
the RNeasy Minikit (Qiagen, Valencia, CA), according to the
manufacturer's instructions. RNA yield was determined by
measuring absorbance at 260 nm. RNA quality was assessed by
electrophoresis of 1 ,~g of RNA on a standard 1.2% formalde-
hyde agarose gel.
Aubert et al.
Microarray and Sequence Analysis. A subtracted library enriched
for genes expressed during retinoic acid-induced neural com-
mitment of ES cells (7) was spread out on LB plates containing
ampicillin and 5-bromo-4-chloro-3-indolyl ,l3-D-galactoside (X-
gal). A total of 384 white bacterial colonies were randomly
picked and cultured in 96-well plates. One microliter of each
bacterial culture was amplified by using the Advantage cDNA
PCR kit (CLONTECH) and PCR primers that are homologous
to the flanking regions of the cDNA insert (30-35 cycles: 30 s at
95°C, followed by 3 min at 68°C). PCR fragments were analyzed
by electrophoresis on 2% agarose gel. The average insert size was
between 200 and 800 bp. PCR products were printed onto
poly~lysine)-coated glass slides by using an Affymetrix (Santa
Clara, CA) 417 arrayer. Each cDNA insert was spotted in
triplicate. A collection of marker genes (see Fig. 2B) whose
expression was predetermined by RT-PCR analysis was included
to act as control for the specificity of neural expression in the
sorted Soxl+/- RNA populations.
Twenty micrograms of Soxl+ or Soxl- total RNA obtained
from disaggregated embryos was reverse transcribed with Su-
PNAS | September 30, 2003 I vol. 100 | suppl. ~ | 11837
OCR for page 22
A
489692
357194
sly
o 244796
122398
it. i ~ i.: -
P~ ..... .
art L2''::::::f:,:: '.:'
~ ~ ''.2''~''-:.:'."'''''''.",''
~~'~'~'~'~'~'~'~'~''''~''
N:: .. .....
6 Hi I. I I
~ - ~ ..~:..:.,.:..i..,~.,.,.,.,.~..
~~ it.:
;. .
Neslin _
Patch
sFRP2
R1
R2
~ _'
0° 10' 1o2 103
Sox1-GFP
FAX \,
~0* ~0*
P311
Riken cDNA
2810027019
ShcC
Soxl O
Bmp4
p-aclin
Fig. 2. FACS purification of neural precursors. (A) FACS profile showing cell
sorting of the Sox1 + neural precursor population from the Sox1- cell popu-
lation from E10.5 mouse embryos. (B) Positive (R2, 30.35%) and negative (R1,
44.02%) populations were sorted and analyzed by RT-PCR for marker gene
expression.
perscript II (GIBCO/BRL) in the presence of Cy3 dCTP after
priming with poly~dT) according to methods described by Brown
et al. (http://brownlab.stanford.edu). Cy3-labeled Soxl+ and
Soxl- cDNA was hybridized (19, 20) to separate microarrays.
Table 1. Genes identified by microarray analysis
Clone
Fold
Increase
P5C3 2.40
P2H5 2.40
P2C5 1.98
P3B7
P5C1 1 1.96
P5H8
P3C1
P3B1
P2E5
P5G9
P5D1 1
P5D5
P5H2
P5A9
P2E7
Scanning was performed with an Affymetr~x 428 scanner, and
the program QUANTARRAY (Perkin-Elmer) was used for image
analysis. Multiple scans were taken to define the optimal dy-
namic range of signal for subsequent analysis. The background
was first subtracted for each respective probe element on the
array. The median value for the triplicate probe elements
representing each gene or insert was then defined. A scaling
factor was applied to the arrays representing normalization to
the 75th percentile of the global signal distribution. Ratios of
expression were then calculated by using these normalized
median values. A ratio value (Soxl+/Soxl-) of 1.5 or greater
was used for the selection of neural specific expression.
Clones encoding 15 up-regulated transcripts (Table 1) were
sequenced, and the corresponding genes were identified by
BLASTN searches of nonredundant (nr), dbEST, and mouse
genomic databases. Protein domains were identified by using
SMART.
In Situ Hybridization. Subtracted library clones were used to
generate antisense RNA probes labeled with digoxygenin-UTP.
Automated in situ detection was carried out on E8.5, E10.5, and
E11.5 outbred mouse embryos by using an InsituPro machine
(Abimed, Langenfeld, Germany).
Results and Discussion
Expression of Sox1GFP. In undifferentiated ES cells in vitro there is
no detectable activity of the SoXlGFP allele but expression is
specifically activated on induction of neural differentiation as
described elsewhere (12, 14~. In viva after germ-line transmis-
sion, the GFP reporter is faithfully expressed in the nervous
system and lens, with no apparent ectopic expression. SoxlGFP
fluorescence is first detected around E8.5 throughout the neural
plate and headfolds (ref. 12 and data not shown). This is slightly
GenBank
accession
no.
1.96 CB968098
1.96 CB9681 07
1.95 CB968099
1.94 CB968097
1.81 CB968094
1.76 CB9681 05
1.69 CB9681 03
1.65 CB9681 04
1.63 CB9681 06
1.60 CB9681 00
1.51 CB968095
Gene symbol (synonyms), name Expression
CB968102 Nhlh2 (HenZ, Nsc1-2, NSCL2), nescient helix-loop-helix 2
CB968096 Mm.1 56164
CB968093 Khdrbs3 (SLM-2, Et/e, T-STAR), KH domain containing,
RNA binding, signal transduction associated 3
Slc2a1 (Glut-1), solute carrier family 2 (facilitated
glucose transporter), member 1
CB968101 sFRP2 (Sdf5), secreted frizzled-related sequence
protein 2
Lrrn1 (NLRR-1), leucine-rich repeat protein 1, neuronal
Sox4, SRY-box-containing gene 4
Zic1, zinc finger protein of the cerebellum 1
Vim, vimentin
Rtn1, reticulon 1
Msi2h (msi2), musashi homolog 2*
Sox11, SRY-box-containing gene 11*
Qk, Quaking
Hrmt113, HMT1 hnRNP methyltransferase-like 3*
Tuba 1 (Talpha 1), tubulin or1
Dev NS and PN cerebellum
Dev and mature NS, epididymis
Ad B and S muscle/testes
E10 NT, OV, heart and gut
E neuroepithelium
Dev CNS, DRG and cartilage
Dev CNS, thymus, BM
E cerebellum and dorsal 1/2
of NT
Radial glia in Dev hindbrain
Dev and mature NS
Dev and mature NS, ubiquitous
by NB
Dev NS, sites of Ep-Mes
interactions
Myelin-forming cells
Dev and mature NS
NPCs
Domains
HLH
KH
TM
CRD
LRR/TM
HMG
ZnF
CC
TM
RRM
HMG
KH
Data are fold increase in expression, accession number of each clone sequence, gene symbols/names, characterized expression pattern, and protein domain
information (given where known). Ad, adult; B. brain; BM, bone marrow; CC, coiled coil; CRD, cysteine-rich domain; Dev, developing; DRG, dorsal root ganglia;
E, embryo; Ep, epithelial; HLH, helix-loop-helix; HMG, high-mobility group; KH, K homology RNA-binding domain; LRR, leucine-rich repeat; M, muscle; Mes,
mesenchymal; NB, Northern blot; N PCs, neural precursor cells; NS, nervous system; NT, neural tube; OV, optic vesicle; PN, postnatal; RRM, RNA recognition motif;
S. skeletal; TM, transmembrane; ZnF, zinc finger.
*The predicted transcripts from these genes are available in the Third Party Annotation Section of the DDBJ/EMBL/GenBank databases under the accession
numbers TPA: BK001349, BK001483, and BK001484.
11838 1 www.pnas.org/cgi/doi/10. 1 073/pnas. 1734197100
~ ~ ~ G
Aubert et a/.
,.
OCR for page 23
later than the reported onset of expression of Soxl mRNA and
protein (13, 14), presumably because of the time required for
correct folding and accumulation of GFP to detectable levels. At
E9.5 SoxlGFP is expressed along the entire neuraxis but in no
other tissue (Fig. lDi). At mid-gestation, SoxlGFP is maintained
throughout the brain and the neural tube but is excluded from
the roofplate and floorplate (Fig. lDii). At this stage, SoxlGFP
expression also becomes evident in the lens where Soxl has been
shown to regulate the ~y-crystallin genes and to be necessary for
lens fiber cell elongation (214. The distribution of SoxlGFP is in
agreement with the published expression of Soxl mRNA and
protein (14~. In later stages of embryonic development, SoxlGFP
is excluded from most differentiated neurons and glia but is
maintained in the proliferative ventricular zone and in the lens
(Fig. lDiii).
In the brains of adult animals SoxlGFP is prominently ex-
pressed in the subgranular layer of the dentate gyrus (Fig. lDiv).
Numerous GFP-expressing cells are seen in the inner subgranu-
lar layer, the area where adult neural progenitors have been
shown to reside (22~. Neurons born in this area subsequently
migrate through the granular layer of the dentate gyrus. At
higher magnification, smaller numbers of GFP-positive cells can
be observed in the granular layer in SoxlGFP mice.
Heterozygous SoxlGFP animals are viable and apparently healthy
with no obvious phenotype. Homozygous null SoxlGFP mice have
small eyes with opaque lenses and suffer from spontaneous sei-
zures, as described for Soxl mutants (21~.
FACS Purification and RNA Probe Generation. E10.5 SoXlGFP em-
bryos were trypsinised, and the resultant pool of cells was sorted
by flow cytometry based on GFP expression. Cell sorting yielded
a SoxlGFP+ neural precursor population (gate R1) and a
SoxlGFP- control population (gate R2) (Fig. 2A). RNA was
extracted from the Soxl+ and Soxl- cells. To confirm the
identity of the two RNA populations, cDNA was produced by
reverse transcription and analyzed by RT-PCR. We examined
the presence in both populations of a number of known genes
whose expression pattern is both spatially and temporally de-
fined during development (Fig. 2B).
As expected, Soxl mRNA was restricted to SoxlGFP+ cells.
Genes known to be restricted to neural progenitors populations
such as Ngn2 and Pax6 (23) displayed a similar restriction to the
GFP+ population. Pax7, Nestin, and sFRP2 all show strong
expression in the GFP+ population but are also represented in
the GFP- fraction. This is in agreement with a predominant
expression of these genes in the developing CNS with additional
expression in the somitic mesoderm (Nestin and Pax7) and in the
mesonephros (sFRP2) (23-25~. RT-PCR analysis revealed weak
expression of the ShcC gene in the Sox1 + cell population. ShcC
is an adapter protein that is predominantly expressed in mature
neurons (26, 273. As neurogenesis in the neural tube begins at
E9.5, the expression seen here could be indicative of a small
number of early neurons present in the Soxl+ cell population.
This could arise from perdurance of GFP after differentiation
and Soxl down-regulation.
RNAs for bone morphogenic protein 4 (BMP4), a marker for
early mesodermal differentiation (28), and SoxlO, a key regu-
lator in the differentiation of peripheral glial cells, with high
expression in neural crest cells and cells of the melanocyte
lineage, were preferentially expressed in SoxlGFP- cells. To-
gether, these results confirm the efficient separation of neural
and nonneural cell populations by flow cytometry.
P311 and Riken cDNA2810027019 represent two genes that are
induced in retinoic acid treated embryoid bodies (74. These genes
show similar expression levels in Soxl+ and Soxl- cell popula-
tions. This result is consistent with the expression profile revealed
by in situ hybridization analyses. Both mRNAs are abundant in the
neural tube, but P311 is also found in the somites and apical
Aubert et a/.
ectodermal ridge (7) and Riken cDNA 2810027019 mRNA is
present in migrating neural crest cells, the apical ectodermal ridge,
and in condensing mesenchymal cells (data not shown). These two
examples illustrate a limitation in the selectivity of screens based on
total cell populations from retinoic acid-induced differentiating
embryoid bodies (see below).
Microarray Analysis. We have previously generated a subtracted
library enriched for genes induced during retinoic acid-induced
differentiation of ES cells (74. Analysis by differential filter
hybridization indicated substantial enrichment for genes of
interest. Of 480 clones, 138 (29%) were preferentially expressed
during retinoic acid-mediated differentiation. These corre-
sponded to 96 unique genes, 40% of which showed enriched
expression in the developing or adult CNS (7~. However, most of
these clones also showed appreciable expression in nonneural
tissues of the developing fetus as exemplified by P311 and Riken
cDNA 2810027019. We reasoned that a more stringent screen
of the SSH library would be to perform the differential hybrid-
ization with SoxlGFP-purified cell populations from embryos
rather than whole cell populations from embryoid bodies.
To test this idea, we generated a cDNA microarray from 384
randomly picked clones from the SSH library and carried out a
differential hybridization screen with RNAs generated from the
SoxlGFP embryos. Also included on the array were a number of
the marker genes used for characterization of the Soxl GFP
RNA isolated from embryos (Fig. 2~.
The Soxl + and Soxl - RNA samples were labeled and
hybridized to the array. The majority of the marker genes shown
to be preferentially expressed in Soxl+ RNA by RT-PCR
analysis (e.g., Soxl, Ngn2, Nestin, and Pax6; Fig. 2B) showed
expression fold changes of >1.5 when analyzed by microarray
analysis. The expression of SoxlO, P311, and the motor neuron
precursor marker Isletl was not elevated in the Soxl+ popula-
tion, consistent with additional sites of expression in ventral
mesoderm and endoderm (29~. The expression of these marker
genes determined by the microarray screen is in a broad agree-
ment with the RT-PCR profiles of the sorted RNA populations
and known expression within and outwith the embryonic CNS.
Clone Identification. Fifteen clones that demonstrated the highest
differential Soxl+/Soxl- expression ratios (>1.5) were taken
for sequencing. Fourteen sequences correspond to known genes,
all of which had previously been associated with either devel-
oping or adult CNS (Table 1~. Clones PSD11 (Msi2h) and P5D5
(Soxl l ~ do not match known genes directly but are derived from
extended 3' UTR regions that lie downstream of the current gene
annotation. In each case, the sequence can be linked to the
identified gene via a contiguous assembly of expressed se-
quences. The discovery of additional 3' UTR sequences is
consistent with the fact that a Musashi2 (Msi2h) 3' UTR probe
detects a 7.1-kb transcript and suggests that the original 2.3-kb
Msi2h cDNA sequence (AB056103) is incomplete (30~. We also
noted that the first 513 base pairs of the original Msi2h cDNA
sequence do not align with other Msi2h ESTs or the human MSI2
sequence. In fact, this presumed 5' UTR sequence has a 100%
match with a genomic sequence on mouse chromosome 1
suggesting that the original cDNA is a hybrid of two clones.
Msi2h is thought to play a role in the maintenance and prolif-
eration of CNS precursor cells (30, 31~.
The relationship between clone P5D5 and Soxl l was detected
by combining sequences from two unigene clusters Mm.41704
and Mm.254253 into a single contiguous sequence of ~7 kb. This
relatively long predicted transcript is consistent with human
SOX11, which has a 3' UTR of >8 kb (32~. Interestingly, the
screen identified a second SRY-box-containing gene, Sox4,
which is expressed in the differentiating subventricular zone
progenitors during neurogenesis (33) and has a role during B cell
PNAS | September 30, 2003 | vol. 100 | suppl. ~ | 11839
OCR for page 24
differentiation and heart development (34). Sox4 and Soxll :~ ~ ~ ~ IRK
encode class C Sox proteins, are closely related at the sequence ~~ ~~ ~~ ~ ~ ~ ~
level, and have similar expression pattern. It has been proposed ~ ~~-~ ~ ~~ ~ i`)
that they may be functionally redundant during CNS develop- ~ DID 1~1~t '6~
Hrmtll3 (clone P5A9) has not been described in mouse, but is ~ ~ ~~ ~~ ~~ ~~ ~~ ~ ~ Be ~~]~ ~ ~;
presumed to be the ortho ague of human hnRNP me by rans-
ferase- ke 3 HRMTIL3~-based high identity (>99%) and
conserved synteny. The gene encoding P2H5 remains to be
identified; this sequence matches a unigene cluster containing
ESTs that are derived exclusively from neural tissues and have
moderate similarity to Neural Wiskott-Aldrich syndrome pro-
tein (N-WASP).
The remaining up-regulated genes were identified as being
expressed predominantly in developing or adult nervous system.
The identification of Reticulon 1, an endoplasmic reticulum
protein of unknown function, in this screen is consistent with the
fact that this protein is localized in neuroepithelial progenitors
at the pallio-subpallial boundary of the developing telencepha-
lon (354.
The gene for the intermediate filament protein, vimentin,
which is expressed in radial glia (36), was up-regulated in the
Soxl+ fraction. Radial glia are thought to serve as precursor
cells in the developing forebrain (374. The neuronal precursor
cell marker tubulin on (38) was also enriched in Soxl + cells.
Nescient helix-loop-helix 2 (~Nhlh2) is a basic helix-loop-helix
transcription factor that is reported to be transiently expressed
in subependymal cells throughout the CNS at mid-gestation and
also transiently in the postnatal cerebellum (39, 404. Expression
in the Soxl + fraction could reflect the perdurance of GFP or an
earlier onset of Nhlh2 expression than previously described.
sFRP2, encoding the Wnt antagonist secreted frizzled-related
protein-2, is expressed in the embryonic neuroepithelium (25,
414. This gene was also isolated in the previous screen of this
library and demonstrated to promote neural differentiation of
ES cells (74. Detection of sFRP2 in the microarray screen
demonstrates the potential of this approach for identification of
functionally significant players in neural differentiation.
The KH domain containing, RNA binding, signal transduc-
tion-associated 3 (Khdrbs3) gene is a predominantly nuclear
RNA-binding protein which heterodimerizes with Sam68 (68-
kDa Src substrate associated during mitosis). Khdrbs3 expression
has been observed in adult brain and also skeletal muscle (424.
Its embryonic expression and functional role have yet to be
defined.
Expression of Lrrnl (leucine-rich repeat protein 1, neuronal)
has been reported in the CNS at E11.5 by Northern-blot, and
whole mount in situ hybridization on E13.5 revealed a predom-
inant expression in the developing nervous system (434. Al-
though its role remains unknown, the ERR domain is proposed
to function in cell adhesion and has been implicated in a variety
of events in neural development.
In Situ Hybridization. We used whole mount in situ hybridization
to examine the embryonic expression of two genes emerging
from the microarray screen, Lrrnl and Msi2h. Both have been
suggested to play significant roles in neural development (Fig. 3~.
At E8.5, Msi2h hybridization is readily detected in the hindbrain
and the otic vesicle, and is not evident in any other tissue. Lrrnl
mRNA is detectable along the entire antero-posterior axis of the
neuroectoderm, with additional faint expression in somites. At
E10.5, Msi2h expression is maintained in the hindbrain and otic
vesicle, but also extends along the neural tube. Hybridization is
also apparent in dorsal root ganglia and limb bud (Fig. 3D).
Lrrnl mRNA on E10.5 is present in the ventral-most neural tube
as well as the hindbrain and the telencephalic vesicle, and is also
prominent in the somites.
11840 1 www.pnas.org/cgi/doi/10.1 073/pnas.17341 97100
;~ ~^
Fig. 3. Expression of Lrrn1 and Musashi2 mRNAs in mouse embryos. (A)
Lateral view of E8.5 embryo showing Msi2h hybridization in the hindbrain and
the otic vesicle. (B and C) Lateral and dorsal view, respectively, of an E10.5
embryo showing Msi2h hybridization in the neural tube, hindbrain, and otic
vesicles. (D) Transverse section at E11.5 showing specific Msi2h hybridization
in the ventral half of the neural tube and the dorsal root ganglia. (E) Lateral
view of an E8.5 embryo showing Lrrn1 hybridization in the neural tube and
weakly in the somites. (F and G) Lateral and dorsal view, respectively, of an
E10.5 embryo showing Lrrn1 hybridization in the telencephalic vesicle, the
hindbrain, the otic vesicle, and the somites. (H) Transverse section at E10.5
shows Lrrn1 hybridization in the ventral part of the neural tube and the
som ites.
Conclusion
In this study, we have shown that Soxl-~p knock-in mice allow
reliable visualization and purification of pan-neural progenitor
cells from mid-gestation mouse embryos. Importantly, Soxl is
expressed in neuroepithelial cells throughout the entire neuraxis,
labeling all categories of regionally specified neural precursor.
The particular advantage of Soxl over the other well established
pan-neural marker nestin is that there is no detectable expres-
sion outwith the CNS during early to mid-fetal development
apart from in the well defined structure of the lens. Examination
in whole mount embryos shows that the SoxlGFP reporter
reproduces faithfully the expression of Soxl. Interestingly, pre-
liminary analyses of adult brains have highlighted expression in
the subgranular layer of the dentate gyrus (Fig. 1), a region for
which there is now overwhelming evidence of the persistence of
neural precursor cells (22, 444. Thus, SoxlGFP may be a useful
marker of adult neural precursors. Further studies are required
to test this directly.
In the present study, we used FACS to separate Soxl-GFP-
positive and -negative populations from whole E10.5 mouse
embryos. Analysis of a panel of markers by RT-PCR yielded
expression data consistent with the substantial elimination of
nonneural cells from the GFP-positive population and con-
versely the absence of neural precursors from the GFP-negative
population.
We then carried out a pilot microarray screen with the aim of
identifying genes specifically expressed both during neural com-
mitment of ES cells and in neural progenitor cells in vivo. From
384 arrayed SSH clones, we identified 15 unique clones showing
preferential expression in the GFP-positive cell population. Of
these, 11 represent known genes previously reported as ex-
pressed in embryonic and/or adult neural tissues, particularly in
the brain. The remaining genes were ESTs, each of which has
originated from libraries derived from neural tissues. We have
been able to identify three of the ESTs as corresponding to
Musashi2, Soxll, and Hrmtll3. Musashi2 has previously been
described as ubiquitously expressed based on Northern analyses
Aubert et a/.
OCR for page 25
~ al r~ ~ i ~
of adult tissues (30~. However, our in situ hybridization data show
that this gene is preferentially expressed in neural tissue in the
fetus. It is noteworthy that marker genes with expression in the
developing nervous system but substantial additional nonneural
expression (i.e., Soxl O. Isletl, P311, and Riken cDNA
2810027019) were not significantly enriched in the Soxl +
population. Furthermore, several clones were found to be
present at higher levels in the Soxl - RNA population (data not
shown). These clones most likely correspond to transcripts
expressed in nonneural tissues induced by retinoic acid treatment
of embryoid bodies.
The fold enrichment value is not an absolute measure of
differential expression, and in many cases may be a considerable
underestimate of the selectivity of expression caused by the
heterogeneity of the Soxl+ population. Soxl marks the entire
pool of proliferating precursors in the neural tube, whereas all
of the genes identified have a regionally restricted expression. A
previous study has indicated that complex tissues such as the
brain are prone to a "dilution effect" when analyzed by microar-
ray, yielding lower levels of fold change and smaller numbers of
differentially expressed genes compared with studies using cell
lines (45~. Nonetheless, larger-scale screening may identify genes
with broader neural expression and consequent higher fold
enrichment values.
1. Rossi, F. & Cattaneo, E. (2002) Nat. Rev. Ne~rosci. 3, 401-409.
2. Gorba, T. & Allsopp, T. (2003) Pharn~acol. Res. 47, 269-278.
3. Svendsen, C. N. & Smith, A. G. (1999) Trends Neurosci. 22, 357-364.
4. Evans, M. J. & Kaufman, M. H. (1981) Nature 292, 154-156.
5. Doetschman, T. C., Eistetter, H., Katz, M., Schmidt, W. & Kemler, R. (1985)
J. Embroil. Exp. Morphol. 87, 27-45.
6. Keller, G. M. (1995) Cars. Opin. Cell Biol. 7, 862-869.
7. Aubert, J., Dunstan, H., Chambers, I. & Smith, A. (2002) Nat. Biotechnol. 20,
1240-1245.
8. Strubing, C., Ahnert-Hilger, G., Shan, J., Wiedenmann, B., Hescheler, J. &
Wobus, A. M. (1995) Mech. Dev. 53, 275-287.
9. Fraichard, A., Chassande, O., Bilbaut, G., Dehay, C., Savatier, P. & Samarut,
J. (1995) ]. Cell Sci. 108, 3181-3188.
10. Bain, G., Kitchens, D., Yao, M., Huettner, J. E. & Gottlieb, D. I. (1995) Dev.
Biol. 168, 342-357.
1l. Li, M., Pevny, L., Lovell-Badge, R. & Smith, A. (1998) Cuter. Biol. 8, 971-974.
12. Ying, Q., Stavridis, M., Griffiths, D., Li, M. & Smith, A. (2003) Nat. Biotechnol.
21, 183-187.
13. Wood, H. B. & Episkopou, V. (1999) Mech. Dev. 86, 197-201.
14. Pevny, L. H., Sockanathan, S., Placzek, M. & Lovell-Badge,
Development (Cambridge, U.K) 125, 1967-1978.
15. Collignon, J., Sockanathan, S., Hacker, A., Cohen-Tannoudji, M., Norris, D.,
Rastan, S., Stevanovic, M., Goodfellow, P. N. & Lovell-Badge, R. (1996)
Development (Cambridge, U.K) 122, 509-520.
16. Kamachi, Y., Uchikawa, M., Collignon, J., Lovell-Badge, R. & Kondoh, H.
(1998) Development (Cambridge, U.K) 125, 2521-2532.
17. Mountford, P., Zevnik, B., Dowel, A., Nichols, J., Li, M., Dani, C., Robertson,
M., Chambers, I. & Smith, A. (1994) Proc. Natl. Acad. Sci. USA 91, 4303-4307.
18. Lupton, S. D., Brunton, L. L., Kalberg, V. A. & Overell, R. W. (1991) Mol. Cell.
Biol. 11, 3374-3378.
19. Schena, M., Shalon, D., Davis, R. W. & Brown, P. O. (1995) Science 270,
467-470.
20. Schena, M., Shalon, D., Heller, R., Chai, A., Brown, P. O. & Davis, R. W.
(1996) Proc. Natl. Acad. Sci. USA 93, 10614-10619.
21. Nishiguchi, S., Wood, H., Kondoh, H., Lovell-Badge, R. & Episkopou, V.
(1998) Genes Dev. 12, 776-781.
22. Kuhn, H. G., Dickinson-Anson, H. & Gage, F. H. (1996) J. Neurosci. 16,
2027-2033.
23. Walther, C. & Gruss, P. (1991) Development (Cambridge, U.K) 113, 1435-1449.
24. Sejersen, T. & Lendahl, U. (1993) J. Cell Sci. 106, 1291-1300.
Aubert et a/.
R. (1998)
Overall, this study demonstrates the potential of combining in
vitro ES cell differentiation and in vivo lineage purification with
microarray technology to achieve rapid, efficient identification
of genes expressed selectively in tissues and stages of interest.
Previously described expression profiles and our in situ hybrid-
ization data of the differentially regulated clones examined
confirms the underlying principle of using RNA prepared from
Soxl-selected cells to screen custom-built microarrays enriched
for neural genes. This pilot scale screen has been sufficient to
highlight several genes, notably Nhlh2, Lrrnl, Hrmtll3, Rtnl, and
the unknown gene corresponding to the P2H5 EST, for further
investigation as potential regulators of neural development. A
rapid means of assessing the significance of these genes would be
via episomal gain-of-function analyses in ES cells (7~.
We thank Hannah Dunstan for in situ hybridization, Steve le Moenic for
MoFlo operation, Jenny Nichols for chimaera production, and Carol
Manson and staff for mouse husbandry. This research was supported by
the Biotechnology and Biological Sciences Research Council, the Med-
ical Research Council of the United Kingdom, and the Human Frontiers
Science Program Organisation. J.A. was the recipient of an Institut
National de la Sante et de la Recherche Medicale fellowship and a Marie
Curie individual fellowship. M.O'R. is supported by the Wellcome Trust
Ph.D. Program. The Scottish Centre for Genomic Technology and
Informatics was established with support from the Scottish Higher
Education Funding Council.
25. Leimeister, C., Bach, A. & Gessler, M. (1998) Mech. Dev. 75, 29-42.
26. O'Bryan, J. P., Songyang, Z., Cantley, L., Der, C. J. & Pawson, T. (1996) Proc.
Natl. Acad. Sci. USA 93, 2729-2734.
27. Conti, L., Sipione, S., Magrassi, L., Bonfanti, L., Rigamonti, D., Pettirossi, V.,
Peschanski, M., Haddad, B., Pelicci, P., Milanesi, G., et al. (2001) Nat. Neurosci.
4, 579-586.
28. Streit, A. & Stern, C. D. (1999) Trends Genet. 15, 20-24.
29. Pfaff, S. L., Mendelsohn, M., Stewart, C. L., Edlund, T. & Jessell, T. M. (1996)
Cell 84, 309-320.
30. Sakakibara, S., Nakamura, Y., Satoh, H. & Okano, H. (2001) J. Neurosci. 21,
8091-8107.
31. Sakakibara, S., Nakamura, Y., Yoshida, T., Shibata, S., Koike, M., Takano, H.,
Ueda, S., Uchiyama, Y., Noda, T. & Okano, H. (2002) Proc. Natl. Acad. Sci.
USA 99, 15194-15199.
32. Azuma, T., Ao, S., Saito, Y., Yano, K., Seki, N., Wakao, H., Masuho, Y. &
Muramatsu, M. (1999) DNA Res. 6, 357-360.
33. Cheung, M., Abu-Elmagd, M., Clevers, H. & Scotting, P. J. (2000) Brain Res.
Mol. Brain Res. 79, 180-191.
34. Schilham, M. W., Oosterwegel, M. A., Moerer, P., Ya, J., de Boer, P. A., van
de Wetering, M., Verbeek, S., Lamers, W. H., Kruisbeek, A. M., Cumano, A.
& Clevers, H. (1996) Nature 380, 711-714.
35. Hirata, T., Nomura, T., Takagi, Y., Sato, Y., Tomioka, N., Fujisawa, H. &
Osumi, N. (2002) Brain Res. Dev. Brain Res. 136, 17-26.
36. Yoshida, M. (2001) J. Neurosci. Res. 63, 284-289.
37. Malatesta, P., Hartfuss, E. & Gotz, M. (2000) Development (Cambridge, U.K)
127, 5253-5263.
38. Wang, S., Wu, H., Jiang, J., Delohery, T. M., Isdell, F. & Goldman, S. A. (1998)
Nat. Biotechnol. 16,196-201.
39. Haire, M. F. & Chiaramello, A. (1996) Brain Res. Mol. Brain Res. 36, 174-178.
40. Gobel, V., Lipkowitz, S., Kozak, C. A. & Kirsch, I. R. (1992) Cell Crowth DifJer.
3, 143-148.
41. Kim, A. S., Anderson, S. A., Rubenstein, J. L., Lowenstein, D. H. & Pleasure,
S. J. (2001) J. Neurosci. 21, RC132 1-5.
42. Di Fruscio, M., Chen, T. & Richard, S. (1999) Proc. Natl. Acad. Sci. USA 96,
2710-2715.
43. Hayata, T., Uochi, T. & Asashima, M. (1998) Gene 221, 159-166.
44. van Praag, H., Schinder, A. F., Christie, B. R., Toni, N., Palmer, T. D. & Gage,
F. H. (2002) Nature 415, 1030-1034.
45. Wurmbach, E., Gonzalez-Maeso, J., Yuen, T., Ebersole, B. J., Mastaitis, J. W.,
Mobbs, C. V. & Sealfon, S. C. (2002) Neurochem. Res. 27, 1027-1033.
PNAS 1 September 30, 2003 1 vol. 100 1 suppi. 1 1 11841
Representative terms from entire chapter:
situ hybridization
