| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
OPEN ACCESS ARTICLE
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
EMBRYONIC STEM CELLS |
aBiology Department, Wesleyan University, Middletown, Connecticut, USA;
bSection of Islet Transplantation and Cell Biology, Joslin Diabetes Center, Boston, Massachusetts, USA
Key Words. Shh • Embryonic stem cells • Neurogenesis • Sox1 • Proliferation • Apoptosis
Correspondence: Laura Grabel, Ph.D., Biology Department, Wesleyan University, 52 Lawn Ave, Middletown, Connecticut 06457, USA. Telephone: 860-685-3238; Fax: 860-685-3279; e-mail: lgrabel{at}wesleyan.edu
Received August 21, 2007;
accepted for publication January 30, 2008.
First published online in STEM CELLS EXPRESS February 28, 2008.
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
primitive ectoderm neurectoderm transitions observed in vivo. Selective death of the Sox1-GFP-negative cells contributes to the enrichment of Sox1-GFP-positive NSCs. Interestingly, Shh is expressed exclusively by the NSCs themselves and elicits distinct downstream gene expression in Sox1-GFP-positive and -negative cells. Suppression of Hh signaling by antagonist treatment leads to different responses from these two populations as well: increased apoptosis in Sox1-GFP-positive NSCs and decreased proliferation in Sox1-GFP-negative primitive ectoderm cells. Hedgehog agonist treatment, in contrast, inhibits apoptosis and promotes proliferation of Sox1-GFP-positive NSCs. These results suggest that Hh acts as a mitogen and survival factor during early ESC neurogenesis, and evidence is presented to support a novel autocrine mechanism for Hh-mediated effects on NSC survival and proliferation. Disclosure of potential conflicts of interest is found at the end of this article.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
Within the vertebrate central nervous system (CNS), Shh acts as a secreted morphogen that determines ventral neuron identities in a concentration-dependent manner throughout the neural tube [2, 3]. Recently, however, multiple groups have provided evidence that Shh can also act as a mitogen that regulates neural stem cells (NSCs) proliferation and survival factor in both embryonic and adult brains [4–8]. In vivo genetic fate mapping experiments show that Shh responsive cells in the subventricular zone of the lateral ventricles and subgranular zone of the dentate gyrus are the quiescent NSCs in the adult mouse brain [9]. These findings indicate that Shh is important in regulating CNS stem cell behavior during embryonic and adult neurogenesis.
Although Shh signaling has been studied extensively in vivo, little is known about its role in embryonic stem cell (ESC) neurogenesis in vitro, an area of great interest given the potential use of ESC-derived neural derivatives to treat neurological disorders. ESCs are pluripotent cells derived from the inner cell mass of the blastocyst that can self-renew indefinitely in vitro in the presence of serum and feeder cells. Upon feeder and leukemia inhibitory factor (LIF) withdrawal, the cells form aggregates in suspension called embryoid bodies (EBs), which give rise to lineages from all three germ layers [10, 11]. Multiple protocols have been devised to derive NSCs or different subtypes of neurons from ESCs, and Shh can induce ventral neural fates in ESC-derived NSCs [12–14].
We have previously shown that endogenous Hh signaling is required during early ESC neurogenesis. EBs derived from Smo–/– ES cells, which cannot respond to Hh signaling, arrest at the primitive ectoderm stage and fail to form neurectoderm [15]. However, it is not clear whether the lack of neurectoderm is due to a failure of neural fate specification or to poor expansion and/or survival of the NSCs. Using the Sox1-GFP ESC line [16], we separated Sox1-GFP-positive and -negative populations during early ESC neurogenesis using fluorescence activated cell sorting (FACS). Reverse transcription-polymerase chain reaction (RT-PCR) analysis of mRNA acquired from these two populations shows that Shh is expressed exclusively in the Sox1-GFP-positive NSCs. Treatment of the differentiating ESCs with the Hh pathway antagonist Cur199691 did not alter the percent of Sox1-GFP-positive cells, but increased apoptosis of Sox1-GFP-positive cells and decreased proliferation of Sox1-GFP-negative cells. Hh appears to act directly on the NSCs, without input from additional cell types, based on our observations that treatment of FACS isolated Sox1-GFP-positive NSCs with Hh antagonist increased apoptosis levels, whereas Hh agonist promoted both NSC survival and proliferation. These data suggest Hh signaling is not required for the fate determination of NSCs but rather supports the survival and proliferation of NSCs and their precursors.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
The Smoothened Mutant and Heterozygous ESC Lines. For the Smo+/– and Smo–/– ESC lines, undifferentiated stem cells were maintained on mitomycin C inactivated STO fibroblast feeder layer in DMEM (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 15% fetal bovine serum (Atlanta Biologicals, Norcross, GA, http://www.atlantabio.com), 2 mM L-glutamine, 1% non-essential amino acids, 1 mM sodium pyruvate, 100 unit/ml Pen/Strep (Invitrogen), 0.1 mM 2-mercaptoethanol and recombinant LIF. For neural induction, methods are modified from defined medium protocol [17]. Briefly, ESCs were dissociated by TrypLE (Invitrogen) into single cell suspension and allowed to form EBs in Petri dishes. EBs were grown and differentiated in EB medium (DMEM +15% fetal bovine serum + 2 mM L-glutamine + 100 unit/ml pen/strep +1% non-essential amino acid) for 4–5 days and then transferred into tissue culture (TC) dishes (Corning Life Sciences, Acton, MA, http://www.corning.com/lifesciences) in EB medium and allowed to attach overnight. To select for NSCs, the medium was then changed to ITSFn (DMEM/F12 (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) + ITS-X (Invitrogen) + 5 ug/ml fibronectin (Sigma) + Pen/strep + L-glutamine) the next day and renewed daily thereafter. After 6–7 days, cells in ITSFn were dissociated by TrypLE into single cells, resuspended in mN3FL medium (DMEM/F12 + ITS-x + 1 ug/ml Laminin + 20 ng/ml fibroblast growth factors (FGF2) (Sigma) + 16 ug/ml putrescine + 6 ng/ml progesterone + L-glutamine) and plated onto laminin coated tissue culture dishes. The cells were maintained in mN3FL for 5–10 days to expand the neural progenitor population, followed by immunofluorescent staining.
The Sox1-GFP ESC (46C) Line. The Sox1-GFP mouse ESC (46C) line was provided by the Austin Smith laboratory [16]. Undifferentiated ESCs were maintained on 0.1% gelatin coated tissue culture dishes without feeder cells in Glasgow minimal essential medium (Invitrogen) medium with 10% fetal bovine serum (Atlanta Biologicals), 2 mM L-glutamine, 1% non-essential amino acids, 1 mM sodium pyruvate, 100u/ml penicillin/streptomycin (Invitrogen), 0.1 mM β-mercaptoethanol (Sigma) and LIF (derived from CHO-LIF cells). ESCs were passaged 1/8 every two days.
For neural induction, ESCs were trypsin dissociated and plated onto 0.1% gelatin coated tissue culture dishes at 0.5–1.5 x 104 cells/cm2 in N2B27 medium supplemented with 0.1 mM β-mercaptoethanol. Cells were allowed to differentiate under this condition for up to 12 days with medium renewal every 2 days. N2B27 is a 50/50 mixture of DMEM/F12+ N2 supplement (1% Insulin-Transferrin-Selenium-X (Invitrogen) + 50 µg/ml bovine serum albumin (Sigma), 16 µg/ml putrecine (Sigma), 6 ng/ml progestrone (Sigma) and neural basal medium + B27 + L-glutamine (all from Invitrogen). For the Hh antagonist experiments, 750 nM or 1.5 uM Cur199691 (Genentech, Inc., San Francisco, http://www.gene.com/gene/index.jsp) were prepared as 1,000X stocks in dimethyl sulphoxide (Sigma-Aldrich) and added to differentiation medium from the first day that ESCs were plated. The same amount of DMSO was added to control dishes. For Hh agonist experiment, 2.5 nM Ag1.4 (Cur199567, Curis Inc., Cambridge, MA, http://www.curis.com) was prepared as 1,000X stocks in DMSO and added to differentiation medium on specified dates.
Immunostaining, BrdU Uptake, and RT-PCR
Immunohistochemistry was performed using the following antibodies: Nestin (Rat 401, Chemicon, Temecula, CA, http://www.chemicon.com; 1:1,000), βIII-tubulin (Covance, Princeton, NJ, http://www.covance.com; mouse monoclonal, 1:500), anti-green fluorescent protein (Molecular Probes, Eugene, OR, http://probes.invitrogen.com; 1:1,000), Oct3/4 (Santa Cruz, Biotechnology, Inc., Santa Cruz, CA, http://www.scbt.com; H-134 rabbit polyclonal, 1:100), cleaved-caspase3 (Cell Signaling Technology, Inc., Danvers, MA, http://www.cellsignal.com, rabbit monoclonal, 1:300), Patched1 (Santa Cruz, G19 goat polyclonal 1:100), anti-phospho-Histone 3 (pH3) (Upstate, Lake Placid, NY, http://www.upstate.com, rabbit polyclonal 1:500), FGF5 (Santa Cruz, fl-268, rabbit polyclonal, 1:100). For nuclear counter stain, we used Hoechst 33,342 1 ug/ml (Molecular Probes).
For 5-bromo-2-deoxyuridine (BrdU) uptake experiments, 10 uM BrdU (Sigma) was added to the culture 30 minutes before the cells were fixed. The dishes were then double-stained with anti-BrdU (Sigma, mouse monoclonal, 1:1,000) and anti-GFP (Molecular Probes, 1:1,000).
Total RNA was extracted from ESCs as well as cells at various stages of differentiation following the user's manual of Ultraspec-RNA (Biotecx Laboratories, Inc., Houston, http://www.biotecx.com). RNA was treated with RNAse free DNaseI (Invitrogen) and reverse-transcribed with MMLV-RT (Invitrogen). cDNAs were amplified with various primers using REDTaq-ReadyMix (Sigma). β-actin was used as a loading control. Primer sequences for Sox1, Oct4, β-actin were as published [16]. Primer sequences for Shh, Ihh, Ptc1, Gli1, Gli2, Gli3 and Smo were as published [18]. Dhh primers and conditions were: forward 5'-caagcagtttgtgcccagta-3'; reverse 5'-ggccttcgtagtggagtgaa-3', 299 bp, 55°C, 35 cycles. Cycle number for each primer pair was determined individually so that amplification was in the exponential range and had not reached a plateau.
Flow Cytometry Analysis and FACS
For Sox1-GFP and Annexin V analysis, cells were dissociated with TrypLE into single cell suspensions and filtered through nylon meshes with 64-um diameter pore (Small Parts Inc., Miramar, FL, http://www.smallparts.com). Staining procedure for Annexin V-APC (BD Biosciences Pharmingen, San Jose, CA, http://www.bdbiosciences.com) was as described in the user's manual. Dead cells were labeled by Propidium Iodine (PI) (Sigma) and excluded from the assay by gating. Cells were analyzed by a FACSCalibre flow cytometer and CellQuest Software (BD Biosciences).
For FACS sorting, cells at day 6 of differentiation were dissociated with TrypLE and filtered through nylon meshes with 40 um diameter pores. The cells were then resuspended in PBS+3% FCS at 1 x 107 cell/ml. The cell suspensions were analyzed and sorted on a FACSVantage flow cytometer and analyzed with FACS Diva software (BD Biosciences). Undifferentiated 46C ESCs were used as GFP-negative control. PI positive cells and doublets were excluded from sorting. Sorted GFP positive and negative cells were collected with 15 ml conical tubes containing N2B27 with 10% FCS. Sorted cells were than reanalyzed for percent of GFP-positive or negative cells, spun down, and re-suspended either in RNA extraction buffer for total RNA isolation or in N2B27 + FGF2+ epidermal growth factor (EGF) for replating and further experiments.
Imaging, Quantification, and Statistics
For BrdU, pH3, and activated caspase 3 quantification, data were from two independent triplicate experiments, pictures of eight fields of cells from each 35 mm tissue culture dishes were taken and counted using Nikon NIS-Elements AR software (Nikon Instruments Inc., Melville, NY, http://www.nis-elements.com). For flow cytometry experiments, data were from one triplicate experiment. p values for statistical significance are described in the corresponding figure legends. Values shown on graphs represent mean ± SEM.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
When maintained in ESC medium, the 46C ESCs expressed high levels of the POU transcription factor, Oct4, based on immunocytochemistry (Fig. 2A, 2B). Upon transfer to N2B27, a neural basal medium used to promote neural induction (see Materials and Methods), there was an initial expansion of ESC colonies, followed by a rapid decrease in Oct4-positive cells after 3 days and a corresponding increase in GFP-positive cells, which first emerged as single cells or clusters (Fig. 2C, D). These cells expressed low levels of GFP and were tightly packed without apparent polarity (Fig. 2D arrow, D'). They expanded in a rosette pattern, and the outer most ring of cells elongated and assumed a bipolar morphology (Fig. 2D arrow-head, 2D''). These cells turned bright green and completely lost Oct4 expression (data not shown). Consistent with previous descriptions [16], by day 5, nearly 80% of the cells were Sox1-GFP-positive and associated with rosettes (Fig. 2E–2F, 2M), and nestin staining showed processes on both ends of these bipolar cells (Fig. 2F' arrows). These rosettes subsequently expanded, and the bipolar cells developed longer processes (Fig. 2G, 2H, 2G', 2H'). By day 7, the non-polar, low GFP-expressing cells in the core diminished, either due to apoptosis or differentiation, leading to the formation of a lumen (Fig. 2G–2H). When we dissociated the differentiating cells at days 7–8 and replated them onto laminin-coated TC dishes, we generated a nearly pure population of Sox1-GFP-positive cells with the typical bipolar NSC morphology (Fig. 2K, 2L). These cells were maintained in FGF2- and EGF-containing medium, as observed previously [24].
|
Interestingly, there was a window of massive cell death at days 4–5 in monolayer culture, coincident with the enrichment of rosettes. At this time, the colonies were predominantly composed of either GFP-positive cells or Oct4-positive cells (see supplemental online Fig. 1A–1D). There was minimal overlap between GFP-positive and Oct4-positive cell clusters, and the boundary between these two populations was usually distinct (see supplemental online Fig. 1A–1D). To determine which cells were dying, we examined the expression of caspase 3, a key protease activated in cells that irreversibly undergo apoptosis [25]. Double labeling of caspase 3 and GFP revealed that the GFP-negative, Oct4-positive cells between rosettes had a high apoptosis rate and often died off as groups, leaving empty spaces between rosettes (see supplemental online Fig. 1E–1H). This carving-like selection process resulted in a highly enriched Sox1-GFP and nestin-positive rosette population after day 5 (Fig. 2E, 2F).
We next used the combination of PI and Annexin V staining with flow cytometry analysis to quantify the percent of apoptosis. PI is a DNA binding dye that labels dead cells and Annexin V is a phospholipid binding protein that binds to phospho-serine molecules on the surface of dead and dying cells [26]. Since dead Sox1-GFP-positive cells lose GFP expression and can not be distinguished from dead GFP-negative cells, we excluded dead cells (PI-positive) from the analysis and quantified only early apoptotic cells (PI-negative/Annexin V-positive) within the GFP-positive and negative populations. GFP-positive cells had a significantly lower percent of apoptotic cells than GFP-negative cells at day 4 through day 9 (see supplemental online Fig. 1I–1J). The caspase 3 and Annexin V data together suggested that NSCs had survival advantages over other cell lineages under serum-free conditions and that selective apoptosis contributes to NSC enrichment.
Characterization of Cell Types Present During the ESC to NSC Transition in Monolayer
During embryogenesis, neurectoderm derives from primitive ectoderm, also known as epiblast. To determine whether ESCs go through a primitive ectoderm stage in monolayer culture, we examined expression of FGF5, a secreted protein that is transiently expressed in primitive ectoderm prior to gastrulation in vivo [27]. Both ESCs and primitive ectoderm cells express high levels of Oct4 in vitro, but only primitive ectoderm cells express FGF5 [28]. RT-PCR data show that, in monolayer culture, fgf5 is expressed transiently, but strongly, during day 3–6 (Fig. 3A), when Sox1-GFP-positive cells emerge. At day 4, FGF5 protein is expressed at high levels in both GFP-positive (arrow in Fig. 3D) and GFP-negative cell clusters (arrowhead, Fig. 3D). Based on this expression pattern, we propose that undifferentiated ESCs (Oct4-positive, FGF5-negative) first turn into primitive ectoderm (Oct4-positive, FGF-positive), then go on to differentiate into neurectoderm (Oct4-negative, FGF5-positive, Sox1-GFP-positive), and finally NSCs (Oct4-negative, FGF5-negative, Sox1-GFP-positive) (Fig. 3A–3E, 3J).
|
During the early neurogenesis phase of monolayer culture (day 3–6), the composition of the cell population included primitive ectoderm (GFP-negative, Oct4-positive, FGF5-positive), neurectoderm (Oct4-negative, FGF5-positive, GFP-positive), and primitive endoderm (GFP-negative, Oct4-negative, Troma1-positive) (Fig. 3J). At day 4, there were typically about 50% Oct4-positive cells, 30% GFP-positive cells, and 20% Troma1-positive cells (Fig. 3K). The GFP-positive population expanded quickly and reached 70–80% by day 6, whereas other lineages diminished (Fig. 3K).
Hh Signaling Is Active During Monolayer Differentiation and Shh Is Expressed by Sox1-GFP-Positive NSCs
In order to investigate whether Hh signaling is active during neurogenesis in monolayer culture, we performed RT-PCR analysis on mRNA extracted from cells at various stages of differentiation (Fig. 4A). Trends described were observed in at least three separate experiments. Oct4 levels decreased as Sox1 levels increased, as observed by immunocytochemistry. Low levels of Sox1 message present in ESCs may be due to the presence of some differentiating cells, whereas decreasing levels at day 12 are likely attributed to neuronal differentiation [16]. Shh expression was first detected on day 4 and increased thereafter, following the emergence of Sox1-GFP-positive cells in culture. Low levels of Ihh were detected on day 6 and also increased slightly thereafter. Dhh was expressed in ESCs and decreased significantly with differentiation. Hh pathway target genes Ptc1 and Gli1 were expressed in ESCs and at early stages in monolayer culture before Shh was expressed, perhaps due to the presence of Dhh. Ptc1 and Gli1 expression levels increased coincident with the increase in Shh expression. In other studies examining ESC neural differentiation, increased levels of Hh pathway components also accompany the production of neural derivatives [12, 13, 31, 32]. These results suggest that Hh signaling is active during ESC differentiation, regardless of protocol or cell line used.
|
Although RT-PCR data suggest that undifferentiated ESCs and differentiating cells in monolayer culture express similar levels of Ptc1 mRNA (Fig. 4A), immunostaining for Ptc1 in ESCs showed only low background levels of protein (Fig. 4C–4D). At day 4, clusters of cells expressing elevated levels of Ptc1 appeared (Fig. 4E, 4F). Some of these clusters also expressed Sox1-GFP, indicating differentiation into NSCs. Ptc1-expressing clusters also expressed FGF5, suggesting that the receptor is initially upregulated in primitive ectoderm, prior to NSC differentiation (see supplemental online Fig. 2). By day 6, however, mature Sox1-GFP positive mature rosettes appeared to downregulate expression of Ptc1 protein, whereas adjacent Sox1-GFP-negative domains robustly expressed the protein (Fig. 4G, 4H). Surprisingly, at least some cells with a neuronal morphology, which expressed βIII tubulin (data not shown), also expressed Ptc1 (Fig. 4H arrow). These data suggest that dynamic regulation of expression of Hh pathway components occurs as domains of Hh expression and response are established during monolayer differentiation.
Hh Antagonist Treatment Decreases Total Cell Number in Differentiating Dishes, but not the Percent of Sox1-GFP-Positive Cells
To further investigate the role of Shh in NSC derivation from ESCs, we treated the cells in monolayer culture with a Hh antagonist, Cur199691, a small molecule working at the level of Smo that can suppress the activity induced by all Hh ligands. Its structure is similar to another small molecule Hh pathway inhibitor, Cur61414 [33], but Cur 199691 is more potent. For all experiments, unless otherwise noted, 750 nM antagonist was added from the first day of differentiation and replenished every day throughout the experiment. This concentration was determined by testing the effect of this antagonist on EBs derived from a Ptc1-LacZ ESC line [34]. RT-PCR analysis of the 46C cells with antagonist treatment showed a significant reduction of Ptc1 and Gli1 expression at day 4 compared to controls, but not a total suppression (Fig. 5A).
|
The colonies in the antagonist treated dishes appeared smaller and unhealthy, based on the presence of pyknotic nuclei (Fig. 5D, 5H). There also was an initial delay in neural differentiation, possibly due to the small colony size. At day 4, whereas numerous GFP-positive cells or clusters started to emerge in control dishes, very few were seen in antagonist treated dishes (Fig. 5E, 5I). However, on day 8, although total cell number and colony size differences remained between control and antagonist treated dishes, the antagonist treated dishes produced a similar percent of GFP-positive and βIII-tubulin positive cells (Fig. 5F–5G, 5J–5K, 5M). Quantification of the percent Sox1-GFP cells present each day indicated that, with a slight lag noted above, the extent of differentiation was comparable in the presence or absence of the antagonist (Fig. 5M). These data suggest that Shh was not required for NSC differentiation from ESCs. Alternatively, Hh activity remaining after antagonist treatment may be sufficient to support neural specification. The observation that Hh antagonist treatment caused a decrease in total cell number in monolayer culture led us to hypothesize that Hh signaling either supports the survival or stimulates the proliferation of NSCs.
Hh Supports the Survival of Sox1-GFP-Positive NSCs in an Autocrine Fashion
To investigate the causes of the cell number reduction in antagonist treated dishes, we analyzed the apoptosis of the differentiating cells with or without antagonist treatment using activated caspase 3 staining (Fig. 6A–6D). At day 5, antagonist treatment caused a nearly two-fold increase in caspase 3-positive cells in the GFP-positive population, 10.9 ± 1.72% in antagonist treated dishes versus 5.5 ± 0.44% (mean ± SEM, n = 6, p < .05) in controls (Fig. 6E). However, no significant differences were observed in the GFP-negative population (Fig. 6F).
|
Due to the heterogeneity of the cell population, it is not clear whether this Hh-dependent survival effect works directly on GFP-positive NSCs or is mediated by the GFP-negative cells. To test these alternatives, we isolated day 6 GFP-positive NSCs by FACS. Post-sort reanalysis showed that 99.9% sorted cells were GFP-positive. We plated these cells in FGF2- and EGF-containing N2B27 medium overnight to allow them to attach. On the next day, growth factors were withdrawn and the medium changed to N2B27 alone, with Hh antagonist, or with Hh agonist, a well-described small molecule that acts at the level of Smo [35]. Over the next 4 days, there was an increase in cell number under control conditions that was inhibited by the presence of antagonist (see supplemental online Fig. 3A, 3B). In contrast, addition of Hh agonist promoted a robust increase in total cell number, pronounced by day 4 (see supplemental online Fig. 3A, 3B). Flow cytometry analysis of control, antagonist, or agonist-treated, FACS isolated Sox1-GFP-positive cells at 2–4 days following replating demonstrated an intermediate level of dead (PI-positive) and apoptotic cells (PI-negative/Annexin V-positive) in control cultures, significantly increased by antagonist treatment and decreased by agonist treatment (see supplemental online Fig. 3C–3E). The levels of apoptotic cells in control and antagonist treated, FACS sorted GFP-positive cells were comparable to their levels in the mixed population (Fig. 6H). These data indicate that Sox1-GFP-negative cells were not required for Hh to promote survival of GFP-positive cells, suggesting that Shh supports the survival of NSCs in an autocrine fashion.
Hh Promotes the Proliferation of Sox1-GFP-Positive and Negative Cells
To investigate whether Hh antagonist treatment affects proliferation, we measured uptake of BrdU, a synthetic thymidine analog that is incorporated into DNA during S-phase, and expression of pH3, which labels cells in mitosis. At day 4, with antagonist treatment the percent of BrdU labeled cells within the GFP-positive population (14.3 ± 2.0%, n = 6) was comparable to controls (17.2 ± 2.2%, n = 6) (Fig. 7A–7D, 7Q). There was, however, a decrease in the percent of BrdU labeled cells in the GFP-negative population (41.2 ± 3.3%, n = 6) compared to controls (53.5 ± 3.6%, n = 6, p = .066) (Fig. 7Q). By day 6, there was extensive BrdU incorporation into the Sox1-GFP-positive population, but no detectable difference in the level of incorporation between treated and untreated control cultures (Fig. 7I–7L).
|
To further investigate the role of Hh on NSC proliferation, we examined BrdU uptake and pH3 labeling in the FACS sorted GFP-positive and -negative cells after replating under control, antagonist-added, or agonist-added conditions. Sorted GFP-negative cells underwent extensive cell death under control conditions and did not survive beyond 3 days in culture. The addition of Hh agonist promoted proliferation of these cells but did not rescue cell death (data not shown). For GFP-positive cells, there was a statistically significant increase in BrdU incorporation upon agonist treatment but no significant decrease upon antagonist treatment (see supplemental online Fig. 3G). pH3 data did not show statistically significant differences between the three conditions.
Added together, these data suggest that the endogenous Hh signaling in monolayer culture supports survival of GFP-positive NSCs and promotes proliferation of GFP-negative primitive ectoderm cells. Addition of ectopic Hh signaling not only further decreases apoptosis of NSCs, but also promotes their proliferation.
There was a clear distinction between the proliferative behavior of the GFP-positive NSCs at day 4 versus day 6, under all conditions assayed. At day 4, the immature, non-polar GFP-positive cells were slowly dividing cells, with a very low-level of BrdU uptake (Fig. 7B, 7D). BrdU labeling of this population increased dramatically as the cells assumed their mature, bipolar morphology and radial configuration at day 6 (Fig. 7J, 7L). In addition, the pH3-positive, dividing cells were randomly distributed within the non-polar, GFP-positive population at day 4 (Fig. 7F, 7H) but, at day 6, were restricted to the core of the mature rosette, adjacent to the lumen (Fig. 7N, 7P). This observation suggests that, as seen with radial glia in the developing embryo, dividing cells are located adjacent to a lumen, the neural tube in vivo [36] and the rosette core in vitro.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
The FGF5-positive, Sox1-GFP-negative primitive ectoderm cell is the likely immediate precursor to the NSC. Our data demonstrate that the GFP-negative population has a proliferation response to Shh produced by the emerging NSCs. Based upon the elevated levels of Ptc1 and Gli1 observed in the isolated GFP-negative population, this response is likely mediated by the well-described Ptc1-Smo-Gli Hh cascade.
As the NSC proportion quickly rises to 70–80% by day 5 (Fig. 5M), we observe selective apoptosis of non-neural lineage cells, the carving phenomenon where GFP-negative cells die off, leaving fields of GFP-positive rosettes. A similar die-off of non-neural lineage cells occurs during the widely used defined medium protocol developed by Ron McKay's laboratory. EBs were plated in a serum free medium, resulting in the death of many cells and selective survival of a highly enriched nestin-positive NSC population [17]. These data suggest that NSCs have survival advantages over other cell lineages under nutrient-poor, serum-free conditions, regardless of protocol or ESC used, although the molecular mechanism has not been described.
Evidence suggests that Shh serves as a survival factor specific to the NSC population during ESC differentiation. Hh antagonist treatment significantly increases apoptosis levels in GFP-positive, but not negative cells (Fig. 6). This antagonist-induced apoptosis is still observed in replated, FACS isolated NSCs, suggesting that no additional cell types are required to relay the effect (see supplemental online Fig. 3). In addition, treatment of these FACS-isolated NSCs with the Hh agonist SAG promotes their survival, as well as a modest increase in proliferation (see supplemental online Fig. 3). Given our observation that Shh is expressed exclusively in GFP-positive NSCs at day 6, these data are consistent with an autocrine mode of action for Hh in promoting NSC survival.
An autocrine mechanism, however, is unusual for Hh signaling. Typically, a field of cells produces a Hh ligand and adjacent cells respond. The response requires the presence of Smo and Ptc1 and is usually characterized by an increase in Ptc1 and Gli1 levels. Whereas the proposed autocrine response for Hh action on ESC-derived NSCs may well involve the Ptc1-Smo-Gli cascade, two observations suggest this conclusion is problematic. First, the isolated GFP-positive NSCs express very low levels of Gli1 mRNA (Fig. 4B). There are, however, ample levels of Gli2 in these cells (Fig. 4B), and this family member has been demonstrated to replace Gli1 function [37]. Second, whereas the GFP-positive NSC population has significant levels of Ptc1 mRNA, immuno-cytochemistry suggests that levels of the Ptc1 protein are very low (Fig. 4H). Both the Hh antagonist and agonist used in these studies act at the level of Smo, bypassing Ptc1 to inhibit or promote downstream components of the cascade, but under control conditions, it remains unclear how Shh could act on NSCs without binding to a Ptc receptor. One possibility is that Ptc2 is expressed on these cells. Alternatively, the low levels of Ptc1 present may be sufficient to mediate the response. Further investigation is needed to examine the proposed autocrine mechanism and determine whether Ptc1's identified role as a dependence receptor that induces apoptosis when not bound by ligand could be involved [38].
We set out to test whether Hh signaling is required for neural fate determination or for survival or proliferation of NSCs or intermediate cell types. Two lines of evidence from this study argue against a fate determining role. First, in monolayer protocol, there were no apparent neural inducer cells that secret Hh. Ihh was expressed after the emergence of NSCs, Shh was expressed by the NSCs themselves, and Dhh was expressed by ESCs and quickly downregulated after differentiation, suggesting no link to neurogenesis. Second, the percent of Sox1-GFP-positive NSCs in both control and Hh antagonist treated cells reached the same level, suggesting that neural fate specification was not inhibited by antagonist treatment.
Our data support Hh's roles as a mitogen and survival factor critical for the derivation of NSCs from ESCs under serum free conditions. Consistent with these findings, when mouse ESCs were plated at clonal density under serum free conditions, minimizing effects of trophic factors and cell-cell interactions, 70–80% of the cells acquired neural fate within several hours after plating [39, 40], suggesting a default mechanism. However, the viability of these NSCs was very low, only 0.02% of the ESCs plated under these conditions forming colonies [40]. Treatment of cells plated at clonal density with Shh may increase the yield of clonal forming colonies. Studies are currently underway to investigate this hypothesis.
ESC derived NSCs are an attractive source of transplantation therapies for neuro-degenerative diseases in that they can be generated in sufficient numbers in a relatively short time, and protocols are currently being developed to produce specific subtypes of neural derivatives to replace lost or damaged cells [12, 41, 42]. Many of these protocols have proved effective for generating neural derivatives from human ESCs as well [43]. Besides directly replacing the damaged neurons, it has been proposed that neuroprotection is also an important mechanism underlying functional recovery in multiple CNS disease models after transplantation therapy using adult brain-derived NSCs [44]. This neuroprotection effect is at least partially due to Shh secreted by the transplanted NSCs [45, 46]. Co-grafting brain-derived NSCs with ventral midbrain neurons to 6-OHDA lesioned rats, a model for Parkinson's disease, dramatically increased the survival of grafted dopaminergic neurons and expedited functional recovery, and Shh was found associated with the surviving dopaminergic neurons [44]. Transplanted adult-derived NSCs were also shown to have neural protective roles in mouse models for Parkinson's disease [47], stroke [48, 49], and multiple sclerosis [50]. Intrastriatal injection of Shh alone can improve motor behavior in Parkinson's disease animal models [51, 52]. We show here that, as described for adult brain-derived NSCs, ESC-derived NSCs also secret Shh. This finding broadens the potential application of ESC-derived NSCs to transplantation therapies.
|
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosure of Potential...AcknowledgmentsReferences |
|---|
| |||||||||||||||||||||||||||||
