Neural Stem Cells Express Non-Neural Markers During Embryoid Body Coculture

THE STEM CELL NICHE Neural Stem Cells Express Non-Neural Markers During Embryoid Body Coculture MARK DENHAM,a TRIEU HUYNH,a MIRELLA DOTTORI,a GREG ALLEN,b ALAN TROUNSON,c RICHARD MOLLARDd a

Centre for Reproduction and Development, Monash Institute of Medical Research, Clayton, Australia; bDepartment of Cytogenetics, Monash Medical Centre, Clayton, Australia; cMonash Immunology and Stem Cell Laboratories and d Department of Biochemistry and Molecular Biology Monash University, Clayton, Australia Key Words. Embryonic stem cells • Neural stem cells • Smooth muscle • Cell fusion • Neural crest

ABSTRACT The capacity of neural stem cells (NSC) to transdifferentiate into a wide range of non-neuronal lineages is the subject of debate. One approach to test NSC plasticity is to ectopically place NSCs in permissive or instructive microenvironments in which the signals driving differentiation of multiple cell types are being elicited. Here we produce embryoid body neurosphere aggregates by combining neurosphere derivatives from fetal mice constitutively expressing green fluorescent protein with embryonic stem (ES) cells isolated from Zin40 mice constitutively expressing nuclear ␤-galacosidase. Under these conditions, we assess neurosphere-derivative– immunoreactivity to anti-neurofilament heavy chain, anti-

pan-cytokeratin, anti-smooth muscle ␣-actinin and anti-␣fetoprotein–specific antibodies. Furthermore, we determine lineage-specific transgene expression and undertake fluorescence in situ hybridization to assess ES cell–neural stem cell–fusion indices. Our data demonstrate that following coculture in hanging drops with ES cells, neurosphere derivatives display immunoreactivity to non-neural markers, in particular smooth muscle, which is not dependent upon cell– cell fusion. These results suggest that given an appropriate environment, NSC may lose their in vivo restrictions and display non-neuronal phenotypes. STEM CELLS 2006; 24:918 –927

INTRODUCTION

In vivo, stem cells are proliferative, self-renewing, and terminally undifferentiated cells that reside in specific tissue niches throughout the body. Such stem cells respond to autonomous and heterologous stimuli to generate new cells for fetal development and the maintenance of homeostasis in the adult. Examples of this type of stem cell are the embryonic neural stem cells (NSCs) of the forebrain germinal zones, the adult NSCs of the hippocampal subgranular zone, and the subventricular zone of the lateral ventricles [5–9]. Although NSCs are capable of self-renewal and multi-lineage differentiation in vivo, they are traditionally accepted to possess a more limited developmental potential than ES cells, being capable only of producing neurons, astrocytes, and oligodendrocytes in accordance with the requirements of their neural-specific microenvironmental niche [10, 11]. More recently an intrinsic plasticity permissive to NSC reprogramming, transdifferentiation, or transdetermination has been suggested following suggestions that NSC derivatives can reconstitute the haematopoietic system, differentiate into endothelial lineages, or contribute to all three embryonic germ layers following intrablastocoelic injection [12–15]. Although theories

Embryonic stem (ES) cells are an in vitro generated pluripotential and immortal cell type derived from the inner cell mass (ICM) of the developing blastocyst [1, 2]. When injected into blastocysts, ES cell derivatives contribute to all tissues of the developing fetus, thus attesting to their pluripotency. The demonstration of ES cell pluripotentiality is not restricted to in vivo studies with differentiation into representatives of all three embryonic germ layers (ectoderm, mesoderm, and endoderm) being similarly achieved in vitro in accordance with a number of different protocols. These protocols include: embryoid body (EB) formation, tissue coculture, the addition of specific factors to media, and spontaneous differentiation following high-density plating [3]. EBs are produced when ES cells are grown either in methylcellulose containing medium, at high density in suspension culture, or in hanging drops. Their formation and subsequent in vitro growth provide a model similar to in vivo embryonic development with the sequential and regional differentiation of endodermal-like, ectodermal, and mesodermal cell layers [4].

Correspondence: Richard Mollard, Ph.D., Department of Biochemistry and Molecular Biology, Monash University, Clayton, 3800, Australia. Telephone: 61-3-9905-5753; Fax: 61-3-9905-3726; e-mail: [email protected] Received April 5, 2005; accepted for publication October 13, 2005; first published online in STEM CELLS EXPRESS October 27, 2005. ©AlphaMed Press 1066-5099/2006/$20.00/0 doi: 10.1634/stemcells.2005-0151

STEM CELLS 2006;24:918 –927 www.StemCells.com

Denham, Huynh, Dottori et al. concerning stem cell plasticity are not unequivocally accepted [16, 17] and the ability of the blastocyst environment to alter NSC fate is an abnormal situation, mounting evidence supports a more subtle plasticity whereby under certain conditions, such as BMP mediated differentiation, NSCs are capable of smooth muscle differentiation [18]. With the demonstration that EBs form permissive environments to the regionalized differentiation of endoderm, ectoderm, and mesoderm derivatives in vitro and that NSC differentiation programs can be altered according to local environmental stimuli, EB aggregates composed of ES cells and neurosphere derivatives were cultured for the analysis of in vitro mediated reprogramming events. To this end, neurospheres produced in vitro from single cell suspensions of green fluorescent protein (GFP) transgenic E12.5 neural tissue, were disaggregated and recombined in hanging drop culture with ES cells expressing the lacZ transgene. GFP expressing neurosphere derivatives isolated subsequent to ES/neurosphere aggregate EB formation were assessed for immunoreactivity to the non-neural–associated markers ␣-fetoprotein, smooth muscle ␣-actin and specific cytokeratin isotypes. Furthermore, fluorescence activated cell sorting (FACS) for GFP expression followed by immunohistochemistry for LacZ expression, in addition to fluorescence in situ hybridization (FISH) for the detection of chromosomal aneuploidies, were performed to quantify cell fusion indices. In this manner, neurosphere derivatives are demonstrated to respond to the EB microenvironment and express marker representatives of endoderm, ectoderm, and mesoderm lineages. These studies demonstrate that pleiotropic cues of the developing EB microenvironment effect the fate of neurosphere derivatives in vitro into non-neuronal phenotypes.

MATERIALS

AND

METHODS

Cell Culture Zin40 ES cells and EBs were cultured as previously described [19, 20]. Upon reaching ⬃70% confluency, Zin40 ES cells were dispersed as single cells in 0.25% trypsin/1 mM EDTA and resuspended in LIF-free Dulbecco’s modified Eagle’s medium (DMEM; Gibco BRL, Gaithersburg, MD, http://www.gibcobrl. com) and 10% fetal calf serum (FCS; Gibco BRL). For the production of neurospheres, C57LB/6-TgN(ACTbEGFP)10sb males (Jackson Laboratory, Bar Harbor, ME, http://www.jax. org) were mated with CBA/CaH WEHI females and fetuses recovered at embryonic day (E)12.5. GFP heterozygotes were identified by fluorescence microscopy (488 nm) of tail clippings. Brains were extracted in artificial cerebral spinal fluid (aCSF) [21], and meninges and blood vessels were removed. Neural tissue was disaggregated into single cell suspension and cultured in neural basal media (NBM) consisting of neurobasal-A (Gibco BRL) supplemented with 1 ⫻ ITS-G, 1 ⫻ N-2, 0.5 ⫻ B-27, chemically defined lipid concentrated diluted 1:20, 1 mM L-glutamine (GIBCO-BRL), 20 ng/ml epidermal growth factor (EGF; Research Diagnostics, Concord, MA, http://www. researchd.com/) and 20 ng/ml basic fibroblast growth factor (FGF2) at a concentration of less than 10 cells/␮l to ensure that resulting neurospheres were clonally derived [10, 17]. For chimeric EB formation, neurospheres were disaggregated into single-cell suspension after 3 days by gentle aspiration in aCSF and the resulting NSC and neural progenitors combined with single www.StemCells.com

919

cell suspensions of ES cells in the indicated ratios to a total cell number of 300 and cultured as hanging drops in 30 ␮l of DMEM and 10% FCS. Differentiation of neurospheres towards neurons was performed by plating dissociated NSC onto polyD-lysine coated dishes and cultured for 2–3 weeks in NBM without supplementation of growth factors. For glial differentiation, neurospheres were dissociated and cultured onto fibronectin coated dishes and cultured in NBM (see above) and 20 ng/ml PDGF-AA. For oligodendrocyte differentiation, neurospheres were dissociated and cultured onto poly-D-lysine coated dishes and cultured in DMEM/F12 supplemented with B27, 1% FCS and 40 ng/ml T3 (Research Diagnostic).

Reverse Transcription and PCR After 3 days of culture, neurospheres generated from E12.5 neural tissue were harvested and RNA isolated according to standard procedures [22]. RNA (5 ␮g) was reverse transcribed using Superscript II (Gibco BRL) and PCR performed for 35 cycles at an annealing temperature of 60°C with: nestin, forward primer 5⬘-CAGCTGGCGCACCTCAAGATG-3⬘, reverse primer 5⬘-AGGGAAGTTGGGCTCAGGACTGG-3⬘, expected product of 208 base pairs and musashi, forward primer 5⬘-CAGCCAAAGGAGGTGATGTC-3⬘, reverse primer 5⬘-CGCTGATGTAACTGCTGACC3⬘, expected product of 451 base pairs. Amplified products were resolved by agarose gel electrophoresis.

Immunofluorescence and Confocal Microscopy After the designated culture period, EB/neurosphere aggregates were subjected to whole-mount immunofluorescence. Aggregates were fixed in 4% paraformaldehyde (PFA) for 2 hours at 4°C, washed three times in wash buffer (150 mM NaCl, 1 mg/ml bovine serum albumin, 0.5% Nonidet P-40, 50 mM Tris pH 6.8) and cell membranes were permeabilized with two 10-minunte incubations in radioimmunoprecipitation buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycolate, 0.1% SDS, 1 mM ethylenediaminetetraacetic acid ([EDTA], 50 mM Tris pH 8.0). Cells were subsequently refixed in 4% PFA for 30 minutes at 4°C, blocked in wash buffer for 1 hour at 4°C, and incubated overnight at 4°C with either mouse anti-human ␣-fetoprotein (AFP), anti-human 200 kd neurofilament (NFH), anti-human pankeratin (PK) or antihuman smooth muscle ␣-actin (SMA) monoclonal antibodies (Biogenesis Ltd., Poole, Dorset, U.K., www.biogenesis.co.uk). Cells were once again washed, and then incubated with Alexa Fluor 546 goat antimouse IgG (Molecular Probes, Eugene, OR, http://probes.invitrogen.com) for 4 hours at 4°C, washed and mounted in Vectashield mounting medium (Vectashield, Burlingame, CA, http://www.vectorlabs.com). Cryosections of neurospheres were treated in the same fashion for detection of vimentin using a mouse anti-vimentin monoclonal antibody (Biogenesis Ltd.). Plated cells and cryosections were assessed by conventional microscopy on a Leica DMR immunofluorescent microscope and images were captured using the Leica MPS60 photo system; (Leica Microscope and Scientific Instruments Group, http://www.leicamicrosystems.com). Aggregates were analyzed by confocal scanning laser microscopy (CSLM) on a Bio-Rad (Hercules, CA, http://www.biorad.com) MRC1000 confocal scanning laser microscope and images captured using Bio-Rad Lasersharp 2000 imaging software (Bio-Rad). All cells in each third 2.5 ␮m interval (at

Neural Stem Cell Plasticity

920

Figure 1. NSC isolation and aggregate EB culture. (A): Bright field of 3 day neurosphere. (B): Reverse transcriptase-polymerase chain reaction of Musashi and Nestin transcripts from the neurospheres. Negative controls were prepared in the same fashion but in the absence of reverse transcriptase. (C): Immunoreactivity to the NSC marker vimentin within the neurospheres prior to dissociation for ES cell aggregation culture (V and arrows point to the vimentin immunoreactivity that is stained red; Hoechst 33342 nuclear counterstain is blue). (D–F): Differentiation of NSC to neural, glial, and oligodendrocyte lineages as shown by immunoreactivity to MAP2AB (D), GFAP (E), and O4 (F). (G, H): The morphological appearance of an EB and a 50% ES/50% NSC representative aggregate (respectively) cultured in hanging drops for 7 days. (I): Histogram depicting the effect of different NSC/ES percentage contributions upon EB formation in hanging drops after 5 days of culture. In each group, 480 hanging drops were assayed. Error bars represent standard errors of the mean. No significant difference in the ability of the various NSC/ES aggregates to form EBs was observed. (J–L): EB/neurosphere aggregates following 7 days of culture as hanging drops and 7 days after plating. Arrows indicate: a cardiac muscle–like cell mass (J) that rhythmically contracted, stellate neuronal–like cells (K) with axonal-like projections, and (L) mixed populations of cells. Scale bars ⫽ 50 ␮m. Abbreviations: EB, embryoid body; ES, embryonic stem cell; M, Musashi; N, Nestin; ⫺ve, negative controls; NSC, neural stem cell.

least 20 intervals per aggregate) of at least 4 chimeric EBs were assayed per lineage specific marker group. Differentiated NSC were fixed in 4% PFA for 10 minutes at 4°C and analyzed by immunofluorescence staining as described above using the primary antibodies mouse anti-MAP2AB (AP20; NeoMarkers; Lab Vision Corporation, Freemont, CA, http:// www.labvision.com), rabbit anti-GFAP (DAKO, Glostrup, Denmark, http://www.dako.com). Staining to detect oligodendrocyte differentiation was performed on live cells using mouse anti-O4 (a gift from D. Anderson, California Institute of Technology, Pasadena, CA).

http://www.roche-applied-science.com) and NBT (4-Nitroblue tetrazolium chloride; Roche Diagnostics) following manufacturers’ instructions. After the whole-mount in situ, those aggregates that were to be subjected to immunofluorescence for GFP were frozen in optimal cutting temperature, sectioned at 8 ␮m, and following the same protocol for immunofluorescence, the primary antibody rabbit anti-GFP antibody was applied (Molecular Probes, 1:500) followed by the Alexa Fluor 568 goat antirabbit IgG secondary antibody (2 ␮g/ml; Molecular Probes).

FACS and FISH Combined Whole-mount In Situ Hybridization and Immunofluorescence Mouse EB/neurosphere aggregates were grown according to the above procedure and whole-mount–in situ hybridization and riboprobe synthesis were performed as previously described [22–24]. A 2.1 kb HindIII full-length Foxd3 cDNA [25], a 240 bp BamHI Slug cDNA fragment (a gift from Martyn Goulding, The Salk Institute for Biological Studies, La Jolla, CA), a 1.5 kb XbaI Snail fragment [26], and a 1 kb XhoI Mammalian achaete-scute homolog 1 (Mash1) fragment (a gift from Francois Guillemot, National Institute for Medical Research, Mill Hill, London) were used as templates for whole-mount in situ hybridization. Alkaline phosphatase coloration was achieved using 5-bromo-4-chloro-3-indolylphosphate (BCIP; Roche Diagnostics, Basel Switzerland,

EB/neurosphere aggregates were dissociated in 0.25% trypsin/1 mM EDTA for 5 minutes at 37°C. Enzymatic activity was neutralized in FCS and single-cell suspensions were sorted for GFP activity on a Mo-Flo FACS (Cytomation, Inc., Fort Collins, CO, http://www.cytomation.com) equilibrated to include any tetraploid cells in the GFP-positive stream. Dissociated GFP neurospheres served as positive controls and Zin40 ES cells as negative controls. A total of 106 GFP- negative and 104 GFPpositive cells were sorted and aliquots of respective cell fractions were fixed to 0.1% gelatinized slides in 4% paraformaldehyde for 4 hours. For FISH, fixed cells were hybridized with a biotinylated IDbright mouse chromosome 18 (37 cM locus) specific point probe (Applied Genetics Laboratories, Inc., Melbourne, FL, http://www.appliedgenetics.com) according to manufacturer’s instructions. For immunofluorescence, the same

Denham, Huynh, Dottori et al. protocol was followed as described above, however, a rabbit anti-␤-galactosidase polyclonal primary antibody (ICN Pharmaceuticals, Costa Mesa, CA, http://www.valeant.com), biotinylated goat anti-rabbit IgG secondary antibody and streptavidin, Alexa Fluor 350 conjugate were used to assay lacZ expression.

Statistical Analysis A one-way analysis of variance followed by Tukey post tests was used to determine significant differences between the ability of chimeric cell aggregates to form chimeric EBs and the expression of each lineage marker in aggregates composed of different neurosphere-derivative/ES cell ratios.

921

the interior of the neurosphere, further suggestive of the presence of NSCs and/or neural progenitors (Fig. 1C). The potential of the clonally expanded neurospheres to differentiate into neural, glial, and oligodendrocyte lineages was demonstrated by in vitro differentiation and immunostaining (Fig. 1D–1F). Fetal blood and endothelial tissue sampled from the fetal brain and cultured in the same fashion failed to generate spheres (data not shown). Taken together, the clonal generation of neurospheres isolated from fetal mouse brain tissue, the criteria for NSC characteristics shown, and the use of conditions used to favor expansion of pure neurosphere populations from neural tissue [17]; these data demonstrate that all spheres formed were neurospheres and not the product of contaminating non-neuronal cell types.

RESULTS The Production of Chimeric EBs Isolation of NSC from Fetal Mouse Brain Tissue The telencephalic germinal zone of E12.5 embryonic–neural tissue was dissociated into single cells and neural-stem cell populations were clonally expanded in vitro as neurospheres in the presence of EGF and FGF2 (Fig. 1A) [27, 28]. After 3 days of culture, reverse transcriptase-polymerase chain reaction with primers specific for Musashi and Nestin sequences amplified the correct sized single cDNA fragments from neurosphere RNA, demonstrating the presence of NSCs and/or neural progenitors (Fig. 1B). Immunoreactivity to an anti-vimentin specific antibody was observed mainly in the periphery of the neurospheres, but also scattered throughout

To determine the effect of combining neurosphere derivatives with ES cells upon EB formation, neurospheres were dissociated after 3 days of culture by gentle aspiration in aCSF and the cellular derivatives were mixed at concentrations of 0%, 10%, 50%, 90%, and 100% with single suspensions of ES cells in hanging drops in DMEM plus 10% FCS. Cell mixtures were then cultured for a further 7 days in hanging drops. Cell aggregates formed and expanded in each group, with the exception of the 100% neurosphere– derivative group where the cells did not aggregate and failed to grow (Figs. 1G, 1H, and data not shown). From a total of 480 hanging drop cultures analyzed in each 0%, 10%, 50%, and 90% neuro-

Figure 2. Marker immunoreactivity of aggregate EBs in hanging drops. Representative images showing GFP-expressing cells (A, E, and I,) neurofilament heavy chain (B), pan-keratin (F), and smooth-muscle actin (J) and their colocalization (C, G, K, respectively), in chimeric EBs composed of 10%, 50%, and 90% NSCs at the onset of culture. Green depicts GFP-positive–NSC derivatives (white arrowheads), red depicts marker antibody immunoreactivity in the absence of GFP colocalization (blue arrow), and yellow depicts colocalization of the marker antibody with a NSC derivative (white arrows). Colocalization of GFP was observed with each marker antibody and in each EB/neurosphere aggregate group. Control immunoreactivity for primary antibodies neurofilament heavy chain expression in adult-mouse brain (D), pan-keratin expression in humanendometrium tissue (H) and smooth-muscle actin expression in adult mouse gut (L). Abbreviations: GFP, green fluorescent protein; NSC, neural stem cell. Scale bars ⫽ 10 ␮m.

www.StemCells.com

922

Neural Stem Cell Plasticity

Figure 3. Percentage of neural stem cell derivatives showing colocalization of each marker antibody and GFP. (A–D): No significant difference in GFP/marker colocalization was observed within each marker specific group (n ⱖ 4 aggregates per percentage point). (E): A significant difference between marker groups was observed (p ⬍ .0001; n ⬎ 15 aggregates per group), antiSMA demonstrated significantly greater GFP colocalization than anti-AFP, antiNFH, and anti-PK antibodies (*p ⬍ .001, **p ⬍ .01, ***p ⬍ .05, respectively) and the anti-AFP antibody displayed significantly reduced GFP colocalization than the anti-pan-keratin antibody (**p ⬍ .01). Abbreviations: EB, embryoid body; GFP, green fluorescent protein; NSC, neural stem cell.

sphere-derivative group, no significant difference in the ability to form aggregates was observed (Fig. 1I). To analyze the differentiation potential of the chimeric aggregates, after 7 days of culture as hanging drops, 56 aggregates from each group forming EB-like aggregates were plated as monolayers and cultured for an additional 7 days. Morphological signs of pluripotential differentiation were observed within aggregates of each group with clusters of synchronously beating cells (Fig. 1J and data not shown), neuronal like cells (Fig. 1K) and mixtures of various cell

populations (Fig. 1L) being readily identifiable. The addition of neurosphere derivatives does not therefore inhibit EB formation or multilineage differentiation of specific cell types within the developing aggregate.

NSC Derivatives Express Non-Neuronal Markers in Hanging Drop EB Cultures To investigate reprogramming of neurosphere derivatives by the pleiotropic EB microenvironment during neurosphere/EB aggregate formation, whole-mount CSLM for GFP

Denham, Huynh, Dottori et al.

923

Figure 4. FACS and FISH analyses for cell fusion. Bright (A, C) and dark (B, D) field photomicrographs of GFP-positive (A, B) and negative (C, D) Fluorescence activated cell sorting fractions immunostained for embryonic stem (ES) cell specific ␤-galactosidase localization. GFP NSC derivatives are green and ES cell derivatives are blue. No green/blue colocalization was seen in any cell from either fraction from a total of 1065 cells analyzed. (E): Fluorescence in situ hybridization analysis with a mouse chromosome 18 specific point probe. Light blue dots represent hybridization events. Two dots are seen per cell, indicating diploidy. Of 707 GFP-positive and -negative cells analyzed, no tetraploid cells were scored. Abbreviations: GFP, green fluorescent protein; ⫹ve, GFP positive; –ve, GFP negative; NSC, neural stem cell.

and colocalization of either anti-␣-fetoprotein (endoderm), anti-neurofilament heavy chain (ectoderm), anti-pan-keratin (endoderm and ectoderm), or anti-smooth muscle ␣-actin (mesoderm) antibodies was assessed (Fig. 2). After 7 days of hanging drop culture, GFP-positive cells demonstrated immunoreactivity to the endodermal specific marker ␣-fetoprotein at an incidence of 0.5%, 0.7%, and 0.7% within the aggregates constructed from 10%, 50%, and 90% neurosphere derivatives, respectively (Fig. 3A and data not shown; n ⬎ 4 aggregates per group). No significant difference in ␣-fetoprotein immunoreactivity between these groups was observed. Immunoreactivity to the neuronal marker neurofilament heavy chain similarly colocalised with GFP with an incidence of 2.5%, 7.1% and 3.9% (Figs. 2A–2C, 3C), pankeratin immunoreactivity colocalised with an incidence of 3.8%, 4.8% and 11.2% (Figs. 2E–2G, 3B), and immunoreactivity to the mesoderm-specific smooth muscle ␣-actin antibody colocalised at an incidence of 12.9%, 15.0% and 16.5%, respectively (Figs. 2I–2K, 3D; n ⬎ 5 aggregates per group). Similarly, the percentage contribution of neurosphere derivatives to the aggregates at the outset of culture had no significant effect upon colocalization of the neurofilament heavy chain, pan-keratin, or smooth muscle ␣-actin antibodwww.StemCells.com

ies with the GFP signal. A variation in colocalization between the different lineage-specific marker antibodies and GFP, however, was observed (p ⬍ .0001; Fig. 3D). The anti-␣-fetoprotein antibody displayed significantly reduced GFP colocalization when compared with the anti-pan-keratin antibody (p ⬍ .01) and the anti-smooth muscle ␣-actin antibody displayed significantly greater GFP colocalization when compared with the anti-␣-fetoprotein, anti-neurofilament, and anti-pan-keratin antibodies (p ⬍ .001, p ⬍ .01, p ⬍ .05, respectively). Neurospheres included as negative controls demonstrated no detectable immunoreactivity to the ␣-fetoprotein, neurofilament heavy chain, pan-keratin, or smooth muscle ␣-actin antibodies (data not shown).

NSC Derivatives Expressing Markers of Reprogramming are Diploid The above immunoreactivity of GFP expressing NSC derivatives to non-neuronal markers could be interpreted either as a reprogramming event of diploid NSC derivatives or a fusion event between ES and NSC derivatives, thus creating a tetraploid cell co-expressing individual traits of each founder cell. To discriminate between these two possibilities, chimeric EBs constructed from 50% NSC/50% ES cells were trypsinized into

924

single cell suspension and sorted by FACS for GFP-positive and -negative fractions. Of 265 individual cells analyzed from the GFP-positive cell fraction, all displayed green fluorescence following excitation at 488 nm, whereas no ␤-galactosidase immunoreactivity, characterizing Zin40 ES cell derivatives, was detected (Fig. 4A, 4B). Of 800 individual cells analyzed from the GFP negative fraction, all demonstrated immunoreactivity to nuclear ␤-galactosidase, whereas no GFP excitation at 488 nm was detected (Fig. 4C, 4D). In order to rule out the possibility that potential cell fusion events had silenced expression of the lacZ transgene, FISH was undertaken with a mouse chromosome 18 specific point probe on both GFP-positive and -negative cell fractions. Of 231 and 476 individual cells analyzed from the GFP-positive and -negative cell fractions, respectively, no tetraploid cells were scored (Fig. 4E). With 15.0% of GFP cells showing immunoreactivity to smooth muscle ␣-actin in chimeric EBs constructed from 50% NSC/50% ES cell derivatives, if fusion were to account for colocalization of GFP activity and smooth muscle ␣-actin immunoreactivity, then ⬃35 tetrapoid cells should be observed in the GFP-positive FACS fraction. Thus, the colocalization of GFP expression and non-neural cell marker expression does not appear to result from cell fusion.

Neural Stem Cell Derivatives Do Not Express Neural Crest Stem Cell Markers in Chimeric EBs A possible explanation for the high percentage of smooth muscle ␣-actin immunopositive cells observed to differentiate from the NSC population, is a de-differentiation of NSCs to a neural

Neural Stem Cell Plasticity crest stem cell (NCSC) phenotype followed by linear differentiation down a neural crest fate. Neurospheres and chimeric EBs were therefore examined for expression of the NSSC markers: Foxd3, Snail, and Slug [25, 26, 29] (Fig. 4). Neurospheres were first analyzed and all were negative for Foxd3, Snail, and Slug expression (n ⬎ 50 aggregates per group; Figs. 5D-5F), demonstrating no NCSC derivatives within the initial neurosphere culture. The aggregates were then examined for NCSC marker transcripts at days 4 and 7 (n ⬎ 50 aggregates per group). On day 4 chimeric EBs were all positive for both Foxd3 and Snail transcripts. Similarly, day 7 chimeric EBs were also positive for Foxd3 and Snail transcripts, although the percentage of positive cells had reduced by this time. All aggregates were negative for the migratory neural crest marker Slug (Figs. 5H-5J). To determine whether these Foxd3⫹ and Snail⫹ cells were originating from NSC, immunofluorescene for GFP was conducted on cryostat sections of the aggregates (n ⬎ 300 cells per marker). None of the cells positive for either Foxd3 or Snail were found to be derivatives of the NSCs (Fig. 6).

Mash1 Transcripts Were Not Identified Within Neurospheres or Chimeric EBs It may be likely that if the NSC were differentiating into NCSCs that other NCSCs progeny may result. Therefore we assayed the neurospheres, days 4 and 7 chimeric EBs for Mash1 a precursor marker of neurogenic differentiating NCSCs and found none as having positive transcripts for Mash1 [30] (n ⬎ 50 aggregates per group; Fig. 7). These results together with the absence of

Figure 5. Analysis of neural crest transcripts within neurospheres and chimeric EBs. E10.5 mouse embryos hybridized with Foxd3, Snail, and Slug riboprobes respectively (A–C). Neurospheres show no transcripts for Foxd3, Snail, and Slug respectively (D–F). Day 4 chimeric EBs show cells positive for Foxd3 and Snail transcripts (G, H, arrows) but not Slug (I). Day 7 chimeric EBs show some EBs with positive cells for Foxd3 and Snail, but no Slug staining, respectively (J, K, L, arrows). Abbreviation: EBs, embryoid bodies.

Denham, Huynh, Dottori et al.

925

Figure 6. NSC progeny tested for Foxd3 and Snail expression. Immunofluorescence for green fluorescent protein (GFP) combined with whole-mount in situ for Foxd3 (C, D) or Snail (A, B) on cryostat sections of day 4 chimeric EBs. Bright field of Snail and Foxd3, respectively, showing positive transcripts stained purple (A, C). Inverse image of bright field for Snail and Foxd3 (white staining) respectively, with immunofluorescence for GFP (red staining) to detect neural stem cell derivatives (B, D). (E–H): Neurosphere immunostained for GFP after wholemount in situ procedure. Positive detection of GFP staining throughout whole sphere (F). (E, G): Dapi stains of (F) and (H), respectively. (H): Negative IgG control staining. Scale bars ⫽ 50 ␮m. Abbreviations: EGFP, enhanced green fluorescent protein; NSC, neural stem cell.

NCSC precursor markers in neurospheres, days 4 or 7 chimeric EBs, suggests NSCs may not utilize a NCSC pathway to smooth-muscle differentiation.

DISCUSSION Whether NSCs follow strictly linear differentiation pathways, producing only neurons, oligodendrocytes, and astrocytes, or whether NSCs are also capable of nonlinear differentiation into functionally significant cellular derivatives of endodermal and mesodermal lineages has been widely debated [31, 32]. For example, adult NSCs have been reported to contribute to all www.StemCells.com

embryonic lineage derivatives following blastocyst injection; furthermore, clonally derived adult NSCs have been reported to reconstitute the haematopoietic system of mice following sublethal irradiation [12, 13]. Similar contributions to haematopoiesis have been reported following the injection of human NSCs into severe combined immunodeficient-hu mice [14]. These examples of NSC transdifferentiation or plasticity, however, have been questioned with suggestions that proof of NSC transdifferentiation functionality following blastocyst injection awaits the birth of healthy neonates and that NSC in vivo neurohaematopoietic potential represents an inconsistent com-

Neural Stem Cell Plasticity

926

Figure 7. Mash1 expression in neurospheres and chimeric embryoid bodies (EBs). E10.5 mouse embryo hybridized with Mash1 riboprobe, positive cells are stained purple and are observed in regions of the telencephalon (arrow) and trunk neural tube (A). Neurospheres, days 4 and 7 chimeric EBs all show no expression for Mash1 (B, C, D).

petence gained by genetic and epigenetic change following long-term in vitro culture [17]. Evidence for NSC plasticity has been similarly investigated in vitro, with transdifferentiation to muscle syncytia and the expression of the filamentous muscle marker protein, myosin heavy chain, being reported following coculture with EBs [12]. The possibility of cell fusion, however, has also been raised following reports detailing coculture and antibiotic-specific selection studies of hygromycin resistant ES cells with puromycin resistant neurosphere derivatives [33]. These latter studies suggested that the apparent expression of extra-embryonic endodermal and myocardial phenotypes by NSC derivatives in ES coculture could be attributed to cell– cell fusion producing polyploid cells expressing characteristic differentiation markers and antibiotic resistance of both ES- and NSCfounder cell populations. However, Wurmser and colleagues demonstrated in cocultures experiments of adult mouse NSC with human endothelial cells differentiation of NSC into endothelial-like cells via a mechanism independent of cell fusion, suggesting NSC plasticity can be observed without cell fusion [15]. Here we further investigate NSC plasticity and describe a coculture system whereby neurosphere derivatives are subjected to a differentiation environment created by embedment within ES aggregates during EB formation as hanging drop cultures. In this manner we demonstrate that E12.5 NSC derivatives can be induced to express markers of all three germ layers (ectoderm, mesoderm, and endoderm). No examples of tetraploidy and no examples of cells co-expressing the ES cell derivative lacZ marker and the neural GFP marker were observed. Differentiation as described here is, therefore, unlikely to result from neurosphere-derivative–ES cell fusion, does not require the maintenance of cell– cell fusion and appears to be the direct result of instructive cues affected by the developing EB-microenvironmental niche. In this study, an average of 15% of all neurosphere derivatives displayed immunoreactivity to the smooth muscle marker ␣-actin, whereas only 3.2% displayed immunoreactivity to the neuronal marker NFH. Although smooth muscle of the viscera and blood vessels derives from the lateral plate mesoderm,

mesoepithelial cells of the neural crest also give rise to smooth-muscle derivatives [34, 35]. In the embryo, neuroepithelial cells give rise to neurons, astrocytes and oligodendrocytes as well as neural crest derivatives [36]. However, their differentiation towards these lineages is restricted and developmentally regulated by temporal and spatial cues. In vitro studies have shown NSC can give rise to a broader range of cell types, including non-neuronal lineages [12, 15, 16]. It is likely that inhibitory factors present in vivo may limit the NSCs to neurons, astrocytes and oligodendrocytes. Removed from these constraints, NSC show a broader developmental potential in vitro, which includes smooth muscle (SM) cells normally produced by the neural crest [18]. Therefore there remain two obvious possibilities for this outcome: 1) SM cells are an additional cell type that NSC can give rise to through linear differentiation, and 2) NSCs undergo a de-differentiation into a NCSC precursor cell. The potential of NSC may be broadened when cultured in vitro resulting from the absence of in vivo restrictions. Rat oligodendrocyte precursor cells when cultured in vitro under certain conditions can acquire properties of CNS stem cells differentiating into neurons and type-1 astrocytes [37] and early-stage rat NSCs are capable of neural crest differentiation [38, 39]. If the NSCs were undergoing a de-differentiation into a NCSC phenotype there exists the potential for other neural crest lineages to arise. Previous research has demonstrated that NCSCs, when subject to BMPs, differentiate into Mash1 expressing cell types and subsequent autonomic neuronal differentiation. Neither neurospheres, days 4 or 7 chimeric EBs were identified to have positive transcripts for Mash1. Early specific NCSC markers were also not detected within the original neurospheres cultures, nor in the NSCs within days 4 and 7 chimeric EBs. Although the mechanism of SM differentiation via a NCSC pathway still remains plausible within this coculture system, taken together the results favor the notion that NSCs can produce an SM cell type through linear differentiation. From other studies, SM differentiation may act via a BMP mediated linear differentiation, with the down stream targets previously shown to be SMADs [18].

CONCLUSION This in vitro data corroborate previously described evidence demonstrating NSCs possess a differentiation capability broader than that seen in vivo [12]. The high percentage of smoothmuscle differentiation seen from these NSCs may be attributed to the removal of in vivo restrictions placed upon the fate of NSCs, and demonstrate a linear differentiation pathway for NSCs not seen in vivo. The ability of the aggregate hanging drop system described here to induce linear marker immunoreactivity of NSC derivatives unseen in normal development in the absence of cell– cell fusion, therefore, provides a relevant in vitro system with which to study further cell fate, reprogramming, dedifferentiation, and subsequent function.

ACKNOWLEDGMENTS We thank Peter Mountford for the ZIN40 ES cell line; David Finkelstein for the C57LB/6-TgN(ACTbEGFP)10sb mouse line; Peter Foley for the mouse chromosome 18 specific point probe; and Martin Pera, Rodney Rietze, Gary Peh, and Perry

Denham, Huynh, Dottori et al.

927

Bartlett for discussions. This work was supported by grants from the Strategic Monash University Research Fund.

DISCLOSURES

REFERENCES

21 Gritti A, Frolichsthal-Schoeller P, Galli R et al. Epidermal and fibroblast growth factors behave as mitogenic regulators for a single multipotent stem cell-like population from the subventricular region of the adult mouse forebrain. J Neurosci 1999;19:3287–3297.

1

Nagy A, Gocza E, Diaz EM et al. Embryonic stem cells alone are able to support fetal development in the mouse. Development 1990;110:815– 821.

2

Nagy A, Rossant J, Nagy R et al. Derivation of completely cell culturederived mice from early-passage embryonic stem cells. Proc Natl Acad Sci U S A 1993;90:8424 – 8428.

3

Rodda SJ, Kavanagh SJ, Rathjen J et al. Embryonic stem cell differentiation and the analysis of mammalian development. Int J Dev Biol 2002;46:449 – 458.

4

Wobus AM. Potential of embryonic stem cells. Mol Aspects Med 2001; 22:149 –164.

5

Bayer SA, Yackel JW, Puri PS. Neurons in the rat dentate gyrus granular layer substantially increase during juvenile and adult life. Science 1982; 216:890 – 892.

6

Lois C, Alvarez-Buylla A. Proliferating subventricular zone cells in the adult mammalian forebrain can differentiate into neurons and glia. Proc Natl Acad Sci U S A 1993;90:2074 –2077.

7

Kuhn HG, Dickinson-Anson H, Gage FH. Neurogenesis in the dentate gyrus of the adult rat: Age-related decrease of neuronal progenitor proliferation. J Neurosci 1996;16:2027–2033.

8

Doetsch F, Caille I, Lim DA et al. Subventricular zone astrocytes are neural stem cells in the adult mammalian brain. Cell 1999;97:703–716.

9

Johansson CB, Momma S, Clarke DL et al. Identification of a neural stem cell in the adult mammalian central nervous system. Cell 1999;96: 25–34.

10 Reynolds BA, Weiss S. Clonal and population analyses demonstrate that an EGF-responsive mammalian embryonic CNS precursor is a stem cell. Dev Biol 1996;175:1–13. 11 Carlen M, Cassidy RM, Brismar H et al. Functional integration of adult-born neurons. Curr Biol 2002;12:606 – 608. 12 Clarke DL, Johansson CB, Wilbertz J et al. Generalized potential of adult neural stem cells. Science 2000;288:1660 –1663. 13 Bjornson CR, Rietze RL, Reynolds BA et al. Turning brain into blood: A hematopoietic fate adopted by adult neural stem cells in vivo. Science 1999;283:534 –537.

The authors indicate no potential conflicts of interest.

22 Chirgwin JM, Przybyla AE, MacDonald RJ et al. Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 1979;18:5294 –5299. 23 Mollard R, Dziadek M. Homeobox genes from clusters A and B demonstrate characteristics of temporal colinearity and differential restrictions in spatial expression domains in the branching mouse lung. Int J Dev Biol 1997;41:655– 666. 24 Mollard R, Ghyselinck NB, Wendling O et al. Stage-dependent responses of the developing lung to retinoic acid signaling. Int J Dev Biol 2000;44:457– 462. 25 Dottori M, Gross MK, Labosky P et al. The winged-helix transcription factor Foxd3 suppresses interneuron differentiation and promotes neural crest cell fate. Development 2001;128:4127– 4138. 26 Smith DE, Franco del Amo F, Gridley T. Isolation of Sna, a mouse gene homologous to the Drosophila genes snail and escargot: Its expression pattern suggests multiple roles during postimplantation development. Development 1992;116:1033–1039. 27 Tropepe V, Sibilia M, Ciruna BG et al. Distinct neural stem cells proliferate in response to EGF and FGF in the developing mouse telencephalon. Dev Biol 1999;208:166 –188. 28 Martens DJ, Tropepe V, van Der Kooy D. Separate proliferation kinetics of fibroblast growth factor-responsive and epidermal growth factorresponsive neural stem cells within the embryonic forebrain germinal zone. J Neurosci 2000;20:1085–1095. 29 Nieto MA, Bennett MF, Sargent MG et al. Cloning and developmental expression of Sna, a murine homologue of the Drosophila snail gene. Development 1992;116:227–237. 30 Guillemot F, Lo LC, Johnson J et al. Mammalian achaete-scute homolog 1 is required for the early development of olfactory and autonomic neurons. Cell 1993;75:463– 476. 31 Gritti A, Vescovi AL, Galli R. Adult neural stem cells: Plasticity and developmental potential. J Physiol Paris 2002;96:81–90.

14 Shih CC, Weng Y, Mamelak A et al. Identification of a candidate human neurohematopoietic stem-cell population. Blood 2001;98:2412–2422.

32 D’Amour KA, Gage FH. Are somatic stem cells pluripotent or lineagerestricted? Nat Med 2002;8:213–214.

15 Wurmser AE, Nakashima K, Summers RG et al. Cell fusion-independent differentiation of neural stem cells to the endothelial lineage. Nature 2004;430:350 –356.

33 Ying QL, Nichols J, Evans EP et al. Changing potency by spontaneous fusion. Nature 2002;416:545–548.

16 Anderson DJ. Stem cells and pattern formation in the nervous system: The possible versus the actual. Neuron 2001;30:19 –35. 17 Morshead CM, Benveniste P, Iscove NN et al. Hematopoietic competence is a rare property of neural stem cells that may depend on genetic and epigenetic alterations. Nat Med 2002;8:268 –273. 18 Rajan P, Panchision DM, Newell LF et al. BMPs signal alternately through a SMAD or FRAP-STAT pathway to regulate fate choice in CNS stem cells. J Cell Biol 2003;161:911–921. 19 Mountford PS, Smith AG. Internal ribosome entry sites and dicistronic RNAs in mammalian transgenesis. Trends Genet 1995;11:179 –184. 20 Munsie MJ, Michalska AE, O’Brien CM et al. Isolation of pluripotent embryonic stem cells from reprogrammed adult mouse somatic cell nuclei. Curr Biol 2000;10:989 –992.

www.StemCells.com

34 LeDouarin NM. The Neural Crest. Cambridge: Cambridge University Press. 1982. 35 Jiang X, Rowitch DH, Soriano P et al. Fate of the mammalian cardiac neural crest. Development 2000;127:1607–1616. 36 Temple S. The development of neural stem cells. Nature 2001;414:112– 117. 37 Kondo T, Raff M. Oligodendrocyte precursor cells reprogrammed to become multipotential CNS stem cells. Science 2000;289:1754 –1757. 38 Molne M, Studer L, Tabar V et al. Early cortical precursors do not undergo LIF-mediated astrocytic differentiation. J Neurosci Res 2000; 59:301–311. 39 Mujtaba T, Mayer-Proschel M, Rao MS. A common neural progenitor for the CNS and PNS. Dev Biol 1998;200:1–15.