|Home | About | Journals | Submit | Contact Us | Français|
Hypoxia promotes neural stem cell proliferation, the mechanism of which is poorly understood. Here, we have identified the nuclear orphan receptor TLX as a mediator for proliferation and pluripotency of neural progenitors upon hypoxia. We found an enhanced early protein expression of TLX under hypoxia potentiating sustained proliferation of neural progenitors. Moreover, TLX induction upon hypoxia in differentiating conditions leads to proliferation and a stem cell-like phenotype, along with coexpression of neural stem cell markers. Following hypoxia, TLX is recruited to the Oct-3/4 proximal promoter, augmenting the gene transcription and promoting progenitor proliferation and pluripotency. Knockdown of Oct-3/4 significantly reduced TLX-mediated proliferation, highlighting their interdependence in regulating the progenitor pool. Additionally, TLX synergizes with basic FGF to sustain cell viability upon hypoxia, since the knockdown of TLX along with the withdrawal of growth factor results in cell death. This can be attributed to the activation of Akt signaling pathway by TLX, the depletion of which results in reduced proliferation of progenitor cells. Cumulatively, the data presented here demonstrate a new role for TLX in neural stem cell proliferation and pluripotency upon hypoxia.
The adult brain retains a reservoir of stem progenitor cells in the hippocampal “neurogenic zone” capable of proliferative activity throughout life (1). These undifferentiated precursors that retain the ability to proliferate and self-renew can give rise to both neuronal and glial lineages (2). Recent studies have emphasized the role of hypoxia in maintaining pluripotency and increased proliferation of neural stem cells (3,–5). However, the molecular mechanisms underlying the increased proliferation and pluripotency of neural stem cells are yet unexplored. TLX (NR2E1), an orphan nuclear receptor expressed in vertebrate forebrains (6), is an essential regulator of adult neural stem cell self-renewal (7). TLX maintains neural stem cells in an undifferentiated and self-renewable state by complexing with histone deacetylases to repress TLX downstream target genes, such as p21 and Pten, promoting cellular proliferation (8). Furthermore, TLX is expressed in the subventricular neural stem cells in embryonic brains and plays an important role in neural development by regulating cell cycle progression of neural stem cells (9,–11). Also, TLX-positive neural stem cells play an important role in spatial learning and memory in adult brains (12).
This study aimed to investigate the role of TLX in promoting neural progenitor population under differentiating conditions, considering the fact that TLX acts as a hypoxic sensor in retinal astrocytes (13), and hypoxia promotes proliferation/dedifferentiation of progenitor cells. Our results demonstrate that TLX is responsive to hypoxia in both differentiating and proliferating conditions. The knockdown of TLX attenuates hypoxia-mediated progenitor proliferation and induces differentiation. Further investigations identified Oct-3/4 as a target for TLX binding in progenitor cells endogenously and upon hypoxia. In our studies, we demonstrate that hypoxia and overexpression of TLX in differentiating progenitors indeed promotes a like phenotype, mediated by Oct-3/4 up-regulation.
AHPs3 were provided by Dr. Gage (Salk Institute, La Jolla, CA) and maintained as described previously (1). Clonal progenitor cells were used between passages 14 and 20 postcloning. For propagation, cells were cultured in Dulbecco's modified Eagle's medium/Ham's F-12 (1:1) containing N2 supplement (Invitrogen), l-glutamine (Cambrex), and 20 ng/ml of recombinant human bFGF (PeproTech EC). When used for experiments, cells were plated at different densities on polyornithine/laminin-coated plates in either proliferating condition, medium with 20 ng/ml bFGF, or in differentiating condition, medium without bFGF. For differentiation assays, FGF was withdrawn from cultures 1 day after seeding and supplemented with 0.2% fetal calf serum. Medium was changed every 2nd day and grown for 7 days when most of the cells appeared differentiated. Unless indicated otherwise, all the experiments were performed in the proliferation media containing bFGF. Ad-TLX was a kind gift of Dr. A. Uemura (RIKEN, Japan). Infection efficiency was judged by staining for β-galactosidase and green fluorescence protein. For shRNA transfection, 48 h after seeding, cells were transfected with shRNA negative control, TLX shRNA or Oct-4 shRNA (Superarray Biosciences), using FuGENE HD reagent (Roche Applied Science) according to the manufacturer's protocol and cells collected after 72 h.
Cells were cultured on chamber slides and fixed for 20 min with 4% paraformaldehyde in phosphate-buffered saline (PBS). For Oct-3/4 staining, cells were permeabilized in 0.5% Triton X-100 for 3 min prior to staining. The cells were incubated in 10% FBS containing PBS for 1 h at room temperature. The cells were rinsed with TBS and incubated with the primary antibodies anti-TLX (1:500 MBL, 1:200 Life Span Technologies), anti-prominin1 (1:1000, Miltenyi Biotec), anti-GFAP (1:1000, Dako), anti-Map2a (1:1000), anti-active caspase 3 (1:1000) (Pharmingen), and anti-Oct-3/4 (1:200, Santa Cruz Biotechnology) diluted in the same blocking buffer overnight at 4 °C. After three washes with PBS, the cells were incubated with Alexa Fluor 488/594 secondary antibody (Molecular Probes) at a 1:2000 dilution. For nuclear counterstaining, the cells were incubated in 1 mg/ml Hoechst 33258 (Molecular Probes) for 30 min before being mounted in Dako fluorescent medium (Dakopatts AB). Cell counting was performed using Image 2000 analysis software (Zeiss) applying appropriate masks.
The mouse Oct-3/4 core promoter and the different mutants of the Oct-3/4 promoter-luciferase plasmids were a gift from Dr. Scholer (Max Planck Institute for Molecular Biomedicine, Germany). TLX expression vector was made with mouse TLX cDNA under the control of the CAG promoter. Cells were seeded at a density of 3.5 × 104 cells per well in 12-well plates and were transfected with the reporter plasmid, GFP control, and expression vectors using FuGENE HD reagent (Roche Applied Science). The total amount of DNA was 0.5–1.0 μg/well. Equal amount of protein was used for the assay, and GFP count was used for normalization.
Total RNA extraction and cDNA synthesis were done according to methods described previously (14). PCR was carried out using standard protocol with DreamTaq polymerase (Fermentas). The following nucleotide primers were used: TLX (62 °C) sense, 5′-GGCCCA TTG TGC TAT TCC TA, and antisense, 5′-TGA ATG GGA CCC CAA TGT AT; Oct-3/4 (68 °C) sense, 5′-ATGGCTGGACACCTGGCTT-3′, and antisense 5′-GGAGTTGGTTCCACCTTCT-3′; actin (62 °C) sense, 5′-AAG ATG ACC CAG ATC ATG TTT GAG, and antisense, 5′-AGG AGG AGC AAT GAT CTT GAT CTT. The samples were run on 1.5% agarose gel containing ethidium bromide and analyzed with FLA 2000 plate reader (Fujifilm). For real time PCR using SYBR Green mix from Applied Biosystems, diluted cDNA/ChIP DNA was used as template, and quantitation was performed by ΔΔCT method.
AHPs were cultured on polyornithine-laminin-coated 24-well plates as described under proliferating or differentiation conditions. The cells were harvested; protein was separated on SDS gel, electroblotted onto a PVDF membrane, and incubated with 5% bovine serum albumin (BSA) in TBS with 0.1% Tween 20, and the membranes were probed with the monoclonal antibodies anti-GFAP (1:1000, Dako), anti-TLX (1:1000, Life Span Technologies), anti-Map2a (1:2000, Millipore), anti-GAPDH (1:2000, Sigma), anti-Ηif2α (1:1000), anti-pAkt (1:1000), anti-total Akt (1:1000), anti-Ki67 (1:2000, Millipore), and anti-Oct-3/4 (1:1000, Santa Cruz Biotechnology). Horseradish peroxidase-conjugated anti-mouse or anti-rabbit immunoglobulin G was used as secondary antibody at a concentration of 1:15,000 (Amersham Biosciences). Signals were detected by enhanced chemiluminescence (Amersham Biosciences).
For ChIP assay, cells were cultured as described above. For the ChIP assay, 1–5 × 106 AHP cells were treated with DMEM containing 1% formaldehyde for 10 min at room temperature for cross-linking, which was stopped by a 10-min incubation with 1.5 m glycine. After washing twice, the cells were resuspended in 300 μl of SDS lysis buffer (50 mm Tris-HCl (pH 8.0), 10 mm EDTA, 1% SDS, and protease inhibitors) by pipetting and kept on ice for 20 min. The chromatin was then sonicated into fragments with an average length of 0.5–1 kb. After centrifugation at 13,000 rpm for 10 min, the supernatants were diluted with dilution buffer (50 mm Tris-HCl (pH 8.0), 1.1% Nonidet P-40, 167 mm NaCl, and protease inhibitors). The extracts were precleaned by incubation with 30 μl of protein G-Sepharose beads (Amersham Biosciences) for 6 h. The supernatants were mixed with antibodies for 16 h and incubated with protein G-Sepharose beads for 3 h. The incubated beads were then washed once with low salt buffer (50 mm Tris-HCl (pH 8.0), 2 mm EDTA, 2% Nonidet P-40, and 0.2% SDS) containing 150 mm NaCl, once with high salt buffer containing 500 mm NaCl, and once with LiCl wash solution (10 mm Tris-HCl (pH 8.0), 250 mm LiCl, 1 mm EDTA, and 0.5% Nonidet P-40). The washed beads were incubated in elution buffer (10 mm Tris-HCl (pH 8.0), 300 mm NaCl, 5 mm EDTA, and 0.5% SDS) at 65 °C for 12 h, followed by phenol/chloroform treatment and ethanol precipitation. ChIP DNA was amplified by standard PCR using Taq polymerase with the primers from Cyber-gene, as listed.
The following antibodies were used: anti-Sp1, anti-p300, anti-HDAC1, anti-EPAS1, anti-H3K9/14Ac (Santa Cruz Biotechnology), anti-TLX (Life Span Technologies), anti-pol II, anti-H3K9Me3 (Abcam), or mouse/rabbit IgG; rOct4 499, GGTTGGAAAGCTACGCTCTG; rOct4 805, GCAACCAAGTGGACCTTCAT; rOct4 137, CCCATTCAAGGGTTGAGTA; rOct4 389, TAGGCCTCGCTTTCTACGAG; rOct4 59, TGAGGTGTCCAGGGACCTA; rOct4 318, TGCTGGAAGGACGACTCAC; actin promf1, GTGTGGTCCTGCGACTTCTAAGTG; and actin promr1, TCCTAGGTGTGGACATCTCTTGGG.
DNA probes were labeled with [α-32P]ATP (ICN) using Klenow polymerase. The amplicons from Oct-3/4 promoter primers were gel-purified, and 500 ng was used for random labeling. The consensus TLX-binding site was synthesized as custom-made oligos and annealed in the presence of salt. After further purification, they were used for isotope labeling. The sequence was as follows: Rat1 forward, TGAACCTGAAGTTCAGATTTTT, and Rat1 reverse, AAAAATCTGACTTCAGGTTC. Nuclear extract was incubated in DNA binding buffer (10 mm Tris-HCl (pH 8.0), 50 mm KCl, 5 mm MgCl2, 1 mm dithiothreitol, 0.05% Nonidet P-40, and 2.5% glycerol) for 30 min at room temperature with 32P-labeled DNA probe and 50 μg/ml poly(dI-dC)(Amersham Biosciences), with or without a 100-fold molar excess of cold competitors. Samples were separated on a 6–8% native polyacrylamide gel, and radiolabeled bands were visualized with BAS-2000.
For supershift assays, the binding activity of TLX on the consensus elements was detected using the EZ-TFA (universal transcription factor assay-colorimetric) kit (Millipore), according to the manufacturer's instructions. Briefly, the consensus elements was biotinylated at the 5′ end on the forward strand alone and annealed. Supershift assays were performed using samples from normoxic or hypoxic nuclear lysates from AHP cells, treated with or without FGF. Preimmune sera and a nonself-capture oligo were employed to rule out nonspecificity. Absorbance was measured at 450/560, and mean value from three independent experiments was plotted as the binding activity of TLX.
After incubating in hypoxic or normoxic conditions for 48 h, the cells were labeled with 10 μm BrdU for 2 h. The cells were fixed in 4% paraformaldehyde in PBS and processed with immunocytochemical staining. In brief, the cells were incubated in 2 n hydrochloric acid (HCl) for 30 min at 37 °C and 0.1 m sodium borate (pH 8.5) for 10 min. Cells were washed with PBS and incubated overnight with mouse monoclonal anti-BrdU antibody (1:200, Chemicon) in PBS containing 0.1% Triton X-100 and 2% BSA at 4 °C. After washing in PBS, the cells were reacted with Alexa Fluor 594-conjugated anti-mouse antibody (1:200) for 1 h at room temperature. Labeled cells were further counter-stained using DAPI and mounted. BrdU-labeled cells were counted using fluorescence microscopy and normalized to the total cell number (DAPI-stained cells). Proliferation rate was obtained as the ratio of cells between hypoxia and normoxia.
Statistical analysis was performed using Student's t test and Pearson rank correlation. All data are expressed as mean ± S.D. The p < 0.05 was considered statistically significant. All calculations were performed using SigmaPlot.
Several studies demonstrate the role of nuclear orphan receptor TLX in maintaining the undifferentiated state and self-renewal capacity of neural stem cells. We wanted to check if TLX could also play a role in committed AHP cells (cultured in differentiating media) back to undifferentiated cells. For this, we proposed to use the well established model of mild hypoxia, after subjecting cells to differentiation conditions by withdrawal of bFGF from media and addition of 0.2% FCS (differentiating medium). After 7 days of withdrawal of FGF, AHP cells appeared differentiated (defined processes compared with the rounded morphology of progenitor cells) (Fig. 1A). A further confirmation was obtained by the monitoring the number of cells expressing progenitor markers nestin and prominin. Both these markers declined after 7 days of culturing in differentiating medium, although Map2a (neuron) and GFAP (glial) increased (Fig. 1B). The protein expression of TLX in undifferentiated and differentiated progenitor cells showed that that the levels of TLX decreased ~4-fold upon differentiation (Fig. 1C). This result was supported by immunofluorescence analysis, where the percentage of cells coexpressing nestin and TLX was 87 ± 3.3% in the presence of FGF and decreased to 28 ± 1.2% upon FGF withdrawal (Fig. 1D and supplemental Fig. 1A).
Several studies demonstrate the role of hypoxia in promoting dedifferentiated morphology of cancer cells and proliferation/pluripotency of progenitor cells (3, 5). In this context, we examined if TLX induced by hypoxia could in turn influence the reprogramming of differentiated AHP cells. For this, AHP cells were grown in normoxia (21% O2 concentration) or hypoxia (1.7% O2 concentration) in proliferation medium or differentiating medium (for 7 days). Immunoblot analysis of TLX in cells exposed to different time intervals of hypoxia (1.7% O2) and normoxia (21% O2) showed that under both proliferating and differentiating conditions, hypoxia increased the protein levels of TLX, although there was a difference in the kinetics of TLX induction between the two conditions (Fig. 1E). There was a 2.5–3-fold increase in TLX from 2 h that increased to ~7-fold after 24 h under proliferating conditions. Under differentiating conditions, TLX induction was much lower, about ~3–4-fold at 12 and 24 h (Fig. 1F). Transcript levels remained the same upon hypoxia, suggesting the role of degradation pathways/post-translational modifications in TLX stabilization after hypoxia (supplemental Fig. 1B).
We next examined the morphology of cells upon exposure to hypoxia when cultured in proliferating and differentiating media. As expected, cells in differentiating medium showed dendritic and glia-like outgrowths unlike the more rounded morphology of proliferating progenitors (Fig. 2A, panels a and b). Upon hypoxia, cells cultured in proliferating medium formed neurosphere-like colonies after 24 h, whereas FGF-depleted cells had just started neurosphere formation, indicating dedifferentiation process (Fig. 2A, panels c and d). After 48 h of hypoxia, cells cultured in both the conditions formed neurosphere-like colonies, although the number of colonies in proliferating medium was 5-fold more (Fig. 2A, panels e and f). These neurospheres stained positively for TLX readily differentiated into GFAP-positive and Map2a-positive cells after subjecting to differentiating conditions (supplemental Fig. 1C). To further assess the role of TLX in maintaining the progenitor-like phenotype, we used both overexpression and knockdown of TLX (shRNA-mediated approach) (supplemental Fig. 1D). TLX overexpressing and knockdown cells were cultured in proliferating media and checked for the colony formation in response to hypoxia (Fig. 2B). Strikingly, AHP cells depleted of TLX and exposed to hypoxia exhibited differentiated morphology, while the overexpression of TLX resulted in an increased colony formation (Fig. 2C). Quantitation of the size of these spheres showed that TLX resulted in an increased average sphere size, similar to the production of more and larger neurospheres, with a higher fraction of Tlx-positive cells per neurosphere observed by Liu et al. (15) (Fig. 2D). Subsequently, the proliferation rate of these cells was quantified by the fold expansion of cell number from day 1 to day 7 of hypoxia. These results show that hypoxia doubles cell population in 24 h in the presence of FGF, whereas the withdrawal leads to the reduced ability of progenitor cells to proliferate (Fig. 2E and supplemental Fig. 1E). The next step was to examine the proliferation rates of progenitor cells upon hypoxia under proliferating and differentiating conditions in the absence of TLX. Upon normoxic conditions, the lack of FGF resulted in a 2–3-fold decrease in the proliferation rate as witnessed by lowered BrdU incorporation. Upon hypoxia, however, the difference was reduced to 1.2-fold. The knockdown of TLX, however, led to an unaltered proliferation between hypoxia and normoxia in the absence of FGF, while it was 1.4-fold in the presence of FGF (Fig. 2F).
Interestingly, we observed only a 30% cell survival upon knockdown of TLX in differentiating conditions, witnessed by an increased apoptosis with condensed or fragmented nuclei (Fig. 2G). Also, transfection of adenoviral Tlx in differentiating conditions reduced the number of apoptotic nuclei marked by active caspase 3 (supplemental Fig. 2, A and B).
To confirm that hypoxia induced dedifferentiation of cells cultured in differentiating medium, we examined the expression profiles of several well known neural stem cell markers and differentiation markers along with TLX upon hypoxia. For this, we performed immunofluorescence analysis for TLX along with prominin, nestin, GFAP, and MAP2a (Fig. 3A). Upon normoxia, TLX was predominantly coexpressed with stem cell markers in the presence of FGF (74% prominin and 76% nestin), although upon differentiation, the percentages decreased to 32 and 28%, respectively (Fig. 3B, 1st and 3rd bars in each group). Under proliferating conditions, 16% of cells stained for GFAP and 12% for MAP2a (Fig. 3C, 1st bar in each group). Although there was a 17% overlap of cells expressing both MAP2a and TLX, only scant cells coexpressed TLX with GFAP, upon FGF withdrawal. In the presence of FGF, 6% cells coexpressed MAP2a and TLX, although we could not observe coexpression of TLX and GFAP (supplemental Fig. 3). This can be attributed to the promotion of neuronal lineage by TLX. Exposure to hypoxia in the presence of FGF leads to an increased expression of TLX with the neural stem cell markers (82% prominin and 89% nestin) (Fig. 3B, 1st and 2nd bars in each group) and a decrease of differentiation markers (8% GFAP and 5% Map2a) (Fig. 3C, 1st and 2nd bars in each group). Upon FGF withdrawal in normoxia, AHP cells exhibited a differentiated morphology as expected, expressing 45% GFAP and 34% MAP2a (Fig. 3C, 3rd bar in each group). The withdrawal of FGF, however, seemed to be compensated for by hypoxia as TLX costained with 53 and 58% of prominin and nestin (Fig. 3B, 4th bar in each group), and the differentiation markers decreased to 21% GFAP and 12% MAP2a (Fig. 3C, 4th bar in each group).
We next checked the contribution of TLX to the hypoxia-driven pool of progenitor cells. Partial knockdown of Tlx using shRNA was performed to assess the percentage of cells expressing stem cell markers (supplemental Fig. 4A). The percentage of cells expressing prominin and nestin decreased by ~2-fold (to 34 and 28%) under normoxia upon Tlx knockdown, and the absence of FGF further decreased this population to 15 and 13%, respectively (Fig. 3, D and E). Upon hypoxia, the knockdown of Tlx decreased progenitor population by 2-fold (92 and 89% to 51 and 42%), although the knockdown of Tlx in the absence of FGF led to the depletion of this population to 4 and 3% (Fig. 3, F and G), because of cell death (supplemental Fig. 4B). These results demonstrate the role of TLX in maintaining the progenitor population and aiding in hypoxia-driven dedifferentiation of progenitor cells.
Several recent studies point out the role of Oct-3/4 in maintaining the stem cell population, and it has been deemed important to induce neural stem cells to embryonic stem cells. In this context, our studies were driven by the fact that Oct-3/4 is a hypoxia-driven factor (16), and it has also been shown to be expressed in neural stem cells albeit to a lower degree (17, 18). Hence, our next step was to investigate if TLX could regulate the expression of Oct-3/4.
AHP cells were transiently transfected with Tlx cDNA construct, and we checked for the transcript of Oct-3/4 in these cells. Semi-quantitative and real time PCR analysis showed that the basal levels of Oct-3/4 in AHP cells were very low, and Tlx overexpression indeed increased the transcript levels of Oct-3/4 by 2.1–4-fold (Fig. 4, A and B). These results were the n verified by immunoblot analysis where we could observe a concomitant increase in the Oct-3/4 protein levels upon ectopic expression of TLX (Fig. 4C). Immunofluorescence analysis also confirmed these results (Fig. 4D). To check for TLX-driven Oct-3/4 gene regulation, we first performed reporter assays where different segments of the Oct-3/4 promoter, including the proximal enhancer (PE) and the distal enhancer (DE), were cloned upstream of a luciferase gene (depicted in Fig. 4E). Cotransfection of Tlx with various promoter truncations showed that TLX increased the luciferase activity of the core promoter by 3.8-fold. Mutation in the retinoic receptor region and enhanced Sp1 activity both led to an increase in the promoter activity by 4- and 6.7-fold by TLX. On the contrary, the mutation of the Sp1 site led to a decrease in the promoter activity by 2-fold (Fig. 4F). These findings are suggestive of the role of TLX in synergistically acting with Sp1 and opposing the effects of the retinoic acid receptor element on Oct-3/4 promoter.
To identify the binding of TLX to a specific region on the promoter, we scanned the 1.0-kb region upstream of transcription start site of Oct-3/4 and segmented it into an immediate region upstream of transcription start site (C), middle region (A), and the distal region (B). EMSA with these regions as probes showed that TLX formed a distinct complex with the proximal core promoter (C), and there was no discernable complex with the middle probe (A) in the presence of dI-dC; however, there was a weak binding to the distal region B (Fig. 5B, lanes 1–3). Overexpression of TLX and knockdown using siRNA were used to check the specific binding of the complex (Fig. 5B, lanes 5 and 6). Specific (C) and nonspecific (A, B, Nsp) cold competitors were used to document the specificity of TLX binding to the core promoter (Fig. 5B, lanes 7–10).
A detailed analysis of the proximal core promoter of rat Oct-3/4 showed the presence of a conserved TLX-binding site. To check TLX binding to this site, we performed EMSA studies with a synthetic oligo spanning 21 bp encompassing this site. Purified FLAG-TLX showed complex formation with the consensus oligo witnessed by gel retardation studies (Fig. 5C, lane 2). Moreover, competition with 50× molar excess cold self and core promoter sequence could ablate the complex formation (Fig. 5C, lanes 3 and 4), although nonspecific oligo did not affect the complex formation (Fig. 5C, lanes 5 and 6). The specific binding activity of TLX on the consensus oligo was quantified by employing EZ-TFA, the universal transcription factor assay kit, as described under “Materials and Methods.” Our results show that upon hypoxia, TLX bound 2.5 times more to the consensus oligo, both in the presence or absence of FGF, demonstrated by an increased absorbance upon addition of TLX antibody. Preimmune control and a nonspecific capture oligo was employed to demonstrate the specificity (Fig. 5D). We next proceeded to check the in vivo recruitment of TLX to the proximal promoter of Oct-3/4 in proliferating conditions (with bFGF). For this, we performed chromatin immunoprecipitation experiments both endogenously and upon overexpression of TLX in AHP cells. Interestingly, TLX recruitment occurred both endogenously (1.2-fold IP) and upon TLX overexpression (3.2-fold IP) in the presence of FGF. The active nature of chromatin was marked by H3K9/14 acetylation. Apart from the 10% input as positive control and immunoglobulin as negative control (Fig. 5E), unrelated actin promoter and the distal promoter element from Oct-3/4 were employed to rule out nonspecific binding (Fig. 5F). To further determine the recruitment of TLX and associated complex on the Oct-3/4 promoter in the presence and absence of FGF, ChIP experiments were performed (Fig. 5G). Upon withdrawal of FGF, the recruitment of TLX and associated active histone marks drastically decreased when compared with FGF treatment (a decrease of ~20-fold for TLX, ~6-fold for H3K9/14Ac, ~10-fold for pol II, and ~1.3-fold for Sp1). There was an increase in the trimethylation at H3K9 residue and recruitment of HDAC1 on the Oct-3/4 promoter upon FGF withdrawal, suggestive of a closed chromatin conformation (Fig. 5H).
Because Oct-3/4 expression is induced by hypoxia and the gene regulated by Hif2α, we investigated the recruitment of TLX on the Oct-3/4 promoter in a time course experiment upon hypoxia in the presence and absence of FGF. First, we performed EMSA studies on Oct-3/4 core promoter, with lysates from proliferating and differentiating progenitor cells upon hypoxia. The binding of TLX to the core promoter increased in time-dependent manner upon hypoxia in undifferentiated cells, although there was only a basal binding in the absence of FGF (Fig. 6A). In vivo chromatin recruitment studies are indicative of an early recruitment of TLX on Oct-3/4 promoter (from 1 h and remains steady until 48 h) in the presence of FGF, although upon FGF withdrawal, the recruitment kinetics is much slower and starts at 12 h (Fig. 6B). An elaborate analysis and comparison of the factor recruitment on the Oct-3/4 promoter showed that the recruitment of TLX and its associated activator complex increased in the presence of hypoxia compared with normoxia, more so in the case of FGF treatment (Fig. 6B). The withdrawal of FGF upon hypoxia had a similar effect, although to a lesser extent (Fig. 6B). The basal recruitment of TLX in differentiated cells upon normoxia was negligible, although we could observe a 1.2-fold increase in TLX recruitment upon hypoxia (Fig. 6, C and D). We could also see a 1.2- and 1.8-fold increase in the recruitment of pol II and H3K9/14 acetylation upon hypoxia. There was a 1.5-fold decrease in the recruitment of HDAC1 and a 1.6-fold decrease in the trimethylation at the H3K9 residue suggesting the active conformation of chromatin (Fig. 6, C and D). The recruitment of Hif2α to the proximal Oct-3/4 promoter served as a positive control for hypoxia (supplemental Fig. 5A).
We then correlated the recruitment and increased expression of Oct-3/4 by TLX upon hypoxia in differentiating and proliferating conditions. Time kinetics of Oct-3/4 induction following hypoxia correlated positively to TLX induction under both proliferating and differentiating conditions. As expected, induction of Oct-3/4 was significantly lowered in differentiating conditions (Fig. 7A). Furthermore, immunofluorescence analysis upon hypoxia showed that the number of cells expressing Oct-3/4 increased from 10 to 57% under proliferating conditions, and upon differentiating conditions, the number of cells expressing Oct-3/4 increased from 3 to 21%, suggesting an alternative FGF-independent pathway for TLX-mediated induction of Oct-3/4 (Fig. 7B and supplemental Fig. 5B).
We then checked if the overexpression/knockdown of Tlx in differentiating or proliferating conditions altered Oct-3/4 levels in AHP cells. The knockdown of Tlx in proliferating medium mimicked the differentiation marker profile in differentiating medium, i.e. knockdown cells had increased expression of GFAP and Map2a, although a decrease in Oct-3/4 was observed (Fig. 7C, left panel). Similarly, overexpression of Tlx in differentiating medium could rescue the effect by inducing Oct-3/4 and down-regulating GFAP and Map2a (Fig. 7C, right panel).
Next, to identify if the induction of Oct-3/4 by TLX leads to an altered proliferation rate, cells cultured in differentiating medium were transfected with TLX, TLX after knockdown of Oct-3/4, and TLX shRNA alone or in combination with Oct-3/4 shRNA. Although the expression of Ki67, a known proliferation marker for cells increased ~3-fold upon TLX transfection, in the absence of Oct-3/4, there was a significant reduction on Ki67 expression (1.2-fold), while knockdown of both Oct-3/4 and TLX, exhibited an additive effect in the reduction of Ki67 expression. Moreover, the expression of GFAP was increased ~2-fold upon knockdown of both TLX and Oct-3/4, although down-regulation mediated by TLX was rescued by TLX shRNA. Also, knockdown of Oct-3/4 had a partial rescue effect.
Under these conditions, we checked for the Akt signaling cascade that has been shown to be sufficient for stem cell maintenance (21), and we found that the knockdown of TLX significantly reduced pAkt levels, although there was only a moderate influence on Oct-3/4 knockdown, suggesting the involvement of Akt signaling cascade in TLX-mediated progenitor proliferation and survival (Fig. 7D). In a converse approach, we overexpressed Oct-3/4 in combination with TLX knockdown to check the degree of synergism between TLX and Oct-3/4. Although both TLX and Oct-3/4 showed elevated Ki67 levels, the knockdown of TLX resulted in moderately low levels of Ki67. However, in the case of pAkt, the knockdown of TLX has a pronounced effect, but Oct-3/4 overexpression was not sufficient to induce phosphorylation of Akt (Fig. 7E). The involvement of Akt pathway in TLX-Oct-3/4-mediated proliferation and survival of neuroprogenitors was confirmed by use of Akt inhibitor (B2311, Sigma). The overexpression of Tlx and/or Oct-3/4 failed to show an increased Ki67 level in the absence of active Akt signaling (Fig. 7E), showing their dependence on Akt signaling. The effects of modulation of these molecules were then examined by following in vivo proliferation rates. In the absence of FGF (differentiating condition), BrdU incorporation was ~2-fold lesser compared with the proliferating condition, although TLX overexpression rescued this to normal levels (similar to proliferating condition upon normoxia). The knockdown of Oct-3/4 prior to Tlx overexpression, however, attenuated this response under both proliferating and differentiating conditions. Upon hypoxia, the number of proliferating cells increased by 2-fold both in proliferating and differentiating conditions. The knockdown of TLX, prior to hypoxia exposure however, led to a drastic decrease of 4- and 5-fold in proliferating and differentiating media. Knockdown of both Oct-3/4 and TLX prior to hypoxia attenuated the proliferative potential of the progenitors under both the conditions (Fig. 7F). These results point at the synergism between TLX and Oct-3/4 in maintaining the proliferative potential of progenitor cells upon hypoxia.
In this study, we identify the role of TLX in promoting hypoxia-dependent proliferation of neural stem cells. We also present evidence for the direct binding of TLX to the Oct-3/4 promoter, thereby regulating the expression levels in neural stem cells upon hypoxia. Our studies also reflect that bFGF and TLX synergize in promoting the cell proliferation and survival, and upon hypoxia, TLX and Oct-3/4 are sufficient to promote the stemness of differentiating progenitors.
This study stems from the fact that Tlx is an essential gene for neural stem cell renewal and proliferation. Several recent studies identify the increased proliferation of neural stem cells under hypoxia (3, 5), and the hypoxic environment leads to down-regulation of genes associated with a neuronal phenotype and up-regulation of genes associated with a neural crest-like phenotype (22,–24). In this context, our studies aimed to define the role of TLX in hypoxia-driven dedifferentiation. This was strengthened by the observation that in retinal astrocytes, TLX acts as a proangiogenic switch in hypoxia. This prompted us to test the expression and level of TLX upon mild hypoxia and the bearing on neural progenitor maintenance. Once we had established the expression level differences between progenitors in proliferating and differentiating conditions (by combination of the stem cell mitogen bFGF withdrawal and addition of 0.2% FCS for 7 days), we proceeded to check the levels of TLX upon hypoxia in these conditions. Interestingly, both in the presence and absence of bFGF, TLX was elevated in hypoxia, although the kinetics and the extent of up-regulation was much lower in differentiating conditions. This was also reflected in the slower proliferation of cells upon hypoxia in the absence of TLX and bFGF. Interestingly, cells depleted of TLX upon FGF withdrawal sustained for a very short time in culture and hypoxia augmented cell death. Increased pyknotic nuclei upon hypoxia in the absence of TLX and bFGF are indicative of synergistic roles in the cell survival pathway apart from the well established role in proliferation (25,–27) for which detailed studies are needed. Moreover, ectopic expression of TLX led to increased neurosphere-like colonies (in size and number) contrary to TLX knockdown, which showed a differentiated phenotype.
Upon hypoxia, there was an increased co-expression of TLX and neural stem cell markers prominin and nestin, although the number of cells expressing differentiation markers GFAP and Map2a was much smaller. Interestingly, even in the absence of FGF, we could observe this effect, although the number of cells coexpressing the stem cell markers was significantly smaller compared with hypoxic cells with FGF. We found a striking overlap of cells expressing TLX and Map2a, which can be attributed to the promotion of neuronal lineage by TLX, upon differentiation. This was in contrast to cells expressing GFAP, where TLX expression could not be tracked and can be explained because TLX represses GFAP expression thereby promoting neural stem cell maintenance (7).
Oct-3/4 is a POU family transcription factor essential for maintaining pluripotency of stem cells. More importantly, the precise expression level of Oct-3/4 is crucial for determining the cell differentiation fate (28). For example, sustained expression of Oct-3/4 in serum-free LIF-deficient medium leads to neuroectoderm formation and subsequently neuronal differentiation. Therefore, a finely tuned control of Oct-3/4 expression is crucial to maintain undifferentiated status of stem cells. There have been several contrasting reports regarding the expression of Oct-3/4 in neural stem cells (29,–31), although several studies unequivocally demonstrate the role of Oct-3/4 in maintaining their pluripotency (17, 19, 20, 32, 33). It is widely agreed that Oct-3/4 is essential to maintain the stemness, and its ectopic expression in adult neural stem cells could generate induced pluripotent stem cells (19, 20, 34). Oct-3/4 core promoter contains multiple cis-acting elements with developmental stage-specific activities, thereby resulting in a number of epigenetic modifications on the promoter, controlling the gene expression. For example, the distal promoter controls Oct-3/4 in intracellular mass and ES cells, whereas the proximal enhancer and retinoic acid response elements control the epiblast/embryonic carcinoma state (35,–38). Because several nuclear orphan receptors bind to the Oct-3/4 promoter and TLX is a retinoid X acid receptor-related nuclear orphan receptor (6, 40), we checked the effect of TLX on Oct-3/4 gene levels. Reporter assays using TLX and core Oct-3/4 promoter showed an increased promoter activity in the presence of TLX. Strikingly, the mutation of the proximal retinoic acid response element resulted in a pronounced increase in the TLX-mediated up-regulation of the Oct-3/4 promoter activity. This is in contrast to the autologous regulation of retinoic acid response β promoter where TLX synergizes with retinoic acid and alleviates the silencing of a Su(var)3-9, Enhancer of Zeste (SET) element (40). Thus, similar to Sirtuin 1 (SIRT1) promoter, TLX could act as both transcriptional repressor and activator (41).
A detailed analysis of the proximal sequence revealed the presence of a core TLX-binding site, and in vitro binding studies showed a positive binding of TLX on this region of the core promoter. In vivo binding assays employing the chromatin from cells overexpressing TLX showed that TLX recruitment on the core promoter correlates with the active H3K9 acetylation and increased pol II recruitment. Even in unstimulated cells (progenitors), we could observe TLX recruitment, suggestive of its role in basal level maintenance of Oct-3/4 levels. When the growth factor was withdrawn (bFGF removal), TLX recruitment is diminished, along with the hypoacetylation at H3K9. We also observed H3K9 trimethylation, indicative of silenced chromatin, promoting the differentiation event.
Upon hypoxia, however, we observed a reinforced TLX recruitment and active histone modifications, overriding the differentiation caused by growth factor deprivation. The trimethylation at H3K9 was also reversed, pointing out the involvement of a demethylase component. Interestingly, TLX has been associated with LSD1, a methyltransferase repressing gene transcription (42). However, here we did not find LSD1 recruitment but rather observed a p300 recruitment. A simultaneous amplification under these conditions showed that TLX recruitment is reversed on GFAP and Oct-3/4 promoter, pointing out to the presence of TLX-activating element favored by another transcription factor on Oct-3/4 promoter (supplemental Fig. 5A). Interestingly, upon hypoxia, a few cells showed coexpression of Oct-3/4 with a neuronal marker, as documented previously (supplemental Fig. 5B) (16, 43). This could be interesting, because TLX has also been shown to influence neuronal differentiation of progenitors when grown in the absence of growth factors (44).
Immunoblot analysis of cells grown in differentiating conditions upon hypoxia showed that TLX up-regulation correlated with Oct-3/4 protein expression. Moreover, immunofluorescence analysis of AHP cells under these conditions showed that Oct-3/4 was localized in the nucleus, and we could not observe any discernable subnuclear distribution. Additionally, the induced Oct-3/4 protein was functional as demonstrated by knockdown studies, where absence of Oct-3/4 significantly reduced the proliferative potential mediated by TLX. Although Oct-3/4 alone was sufficient to induce proliferation as demonstrated by increased expression of Ki67 marker, the survival pathways governed by Akt signaling seemed to be drastically affected by the absence of TLX. Considering that Akt signaling is sufficient to maintain stem cell population even in the absence of LIF, our finding that TLX activates Akt in differentiating conditions is of significance. This can be attributed to the down-regulation of phosphatase and tensin homolog by TLX (8, 42) that in turn activates pAkt. On the other hand, activation of Oct-3/4 by TLX could up-regulate Tcl1 that substantially increases Akt kinase activity and mediates its nuclear translocation (39, 45). This is in support of our finding that the knockdown of Tlx in differentiating conditions leads to increased pyknosis and cell death. In summary, our findings here demonstrate a novel role for TLX in regulating neuronal progenitor proliferation upon hypoxia by regulating the expression of the Oct-3/4 gene.
We thank Prof. F. Gage (Salk Institute) for AHP cells; Prof. H. Scholer (Max Planck Institute, Germany) for the Oct-3/4 promoter luciferase constructs and CMV-Oct-3/4 construct; and Prof. A. Uemura, (RIKEN, Japan) for Ad-TLX construct.
*This work was supported in part by grants from the Swedish Research Council, Swedish Cancer Society, Swedish Childhood Cancer Research Foundation, Hjalmar Svensson Foundation, ALF Funds at the Västra Götaland Region, and in part by Assar Gabrielson Foundation, Sweden.
3The abbreviations used are: