Microarray Analyses Support a Role for Nurr1 in Resistance to Oxidative Stress and Neuronal Differentiation in Neural Stem Cells

Authors

  • Kyle M. Sousa,

    1. Laboratory of Molecular Neurobiology, Department of Medical Biochemistry and Biophysics, Karolinska Institute, Stockholm, Sweden
    Search for more papers by this author
  • Helena Mira,

    1. Laboratory of Molecular Neurobiology, Department of Medical Biochemistry and Biophysics, Karolinska Institute, Stockholm, Sweden
    2. Stem Cell Regulation Laboratory, Prince Felipe Research Center, Valencia, Spain
    Search for more papers by this author
  • Anita C. Hall,

    1. Laboratory of Molecular Neurobiology, Department of Medical Biochemistry and Biophysics, Karolinska Institute, Stockholm, Sweden
    2. Division of Cell and Molecular Biology, Imperial College London, London, United Kingdom
    Search for more papers by this author
  • Lottie Jansson-Sjöstrand,

    1. Laboratory of Molecular Neurobiology, Department of Medical Biochemistry and Biophysics, Karolinska Institute, Stockholm, Sweden
    Search for more papers by this author
  • Moriaki Kusakabe,

    1. Experimental Animal Research Center, Institute for Animal Reproduction, Ibaraki, Japan
    Search for more papers by this author
  • Ernest Arenas M.D., Ph.D.

    Corresponding author
    1. Laboratory of Molecular Neurobiology, Department of Medical Biochemistry and Biophysics, Karolinska Institute, Stockholm, Sweden
    • Karolinska Institute, Division of Molecular Neurobiology, MBB, Scheeles väg 1, A1 Plan 2, Stockholm 17177, Sweden. Telephone: 46-8-728-7663; Fax: 46-8-341960
    Search for more papers by this author

Abstract

Nurr1 is an orphan nuclear receptor required for the development of midbrain dopaminergic neurons. To better understand the molecular consequences of Nurr1 expression, we compared the transcriptomes of two independent control and Nurr1-expressing NSC lines using Affymetrix cDNA microarrays. These data reveal the regulation of genes involved in promoting cell survival (trophic/growth factors and stress response genes) and in preventing cell death (decreased caspase-3 and caspase-11 expression). We found that conditioned medium from Nurr1-expressing NSC lines enhanced the survival of midbrain dopaminergic neurons in primary cultures and that Nurr1-expressing NSC lines themselves were more resistant to oxidative stress. These findings are accompanied by a dynamic pattern of gene regulation that is consistent with a role for Nurr1 in promoting both the acquisition of brain-region-specific identity (Engrailed-1) and neuronal differentiation (tubulin β III). Interestingly, our gene expression profiles suggested that tenascin-C was regulated by Nurr1 in developing dopaminergic neurons. This was further confirmed in vitro and in Nurr1 knockout mice where low levels of tenascin-C mRNA were observed. Analysis of tenascin-C-null mice revealed an increase in the number of Nurr1+ cells that become tyrosine hydroxylase-positive (TH+) dopaminergic neurons at embryonic day 11.5, suggesting that tenascin-C normally delays the acquisition of TH by Nurr1+ precursors. Thus, our results confirm the presence of both secreted and cell-intrinsic survival signals modulated by Nurr1 and suggest that Nurr1 is a key regulator of both survival and dopaminergic differentiation.

Introduction

Parkinson disease (PD), the second most common human neurodegenerative disorder, primarily results from the selective and progressive degeneration of ventral midbrain dopamine-synthesizing neurons. Cell-replacement strategies based on the transplantation of stem cells manipulated to efficiently generate functional dopaminergic neurons have recently raised hope for the improved treatment of PD patients. In addition, genetic engineering of mouse neural stem cells and ESCs using nuclear receptor related-1 (Nurr1/NR4A2) has proven successful for generating cell populations enriched in dopaminergic (DA) neurons [1, [2]–3]. Despite the therapeutic potential of these engineered cells for patients with PD, the function of Nurr1 in genetically engineered stem cells remains unclear.

Nurr1 is a transcription factor of the nuclear receptor superfamily that is highly expressed in the developing and adult ventral midbrain [4]. In the midbrain, Nurr1 is detected prior to the expression of tyrosine hydroxylase (TH) in dopaminergic neurons in vivo [5, [6]–7]. Moreover, in Nurr1-null mice, midbrain TH+ neurons are not born, and in their usual position, cells die by apoptosis at embryonic day 18.5 (E18.5) [5, 8]. These studies suggest that Nurr1 is required for both the specification of neurotransmitter phenotype and for neuronal differentiation/survival. In agreement with these data, we have previously demonstrated that Nurr1 overexpression induces neuronal differentiation in the c17.2 NSC line [3]. However, Nurr1 NSCs adopt a TH+ dopaminergic phenotype only upon coculture with embryonic/newborn midbrain astrocytes, suggesting that additional non-cell-autonomous signals, including Wnt-5a, are required for dopaminergic differentiation [3, 9, 10]. The c17.2 NSC line consists of multipotent neural progenitors isolated from the external germinal layer of the cerebellum [11]. This cell line was initially immortalized using retroviral v-myc transfer and differentiates into both glial and neuronal cell types upon transplantation into various central nervous system regions.

Nurr1 is expressed in adult neurons and is regulated by different types of brain insults, including ischemia [12] and seizures [13, 14]. These results, together with an increased loss of dopaminergic neurons in Nurr1+/− mice [15], have suggested a role for Nurr1 in neuronal survival after the birth of dopaminergic neurons. To gain more insights into the molecular events elicited by exogenous Nurr1 expression in stem cells, we have analyzed the transcription profile of c17.2 neural stem cells stably transfected with Nurr1. Here, we show that Nurr1 exerts multiple functions, including facilitating the differentiation of multipotent stem cells, maintaining postmitotic neuronal precursor populations, promoting survival, and conferring resistance to oxidative stress.

Materials and Methods

Cell Culture and Treatment

c17.2 neural stem cells were maintained and passaged as previously described [11]. To generate accumulated cell growth curves, the total number of viable cells was assessed at each passage by trypan blue exclusion, and 500,000 viable cells were plated. At the time of the next passage (after 2 days in vitro), the number of cells generated was determined, and the ratio of cell production was multiplied by the total number of cells in the previous point of the curve. Growth data are represented by the equation y = a·2sx, where y is the total number of cells, x is the time in days in vitro (DIV), s is the rate of doubling, and a is a constant [16]. For differentiation, c17.2 cells were plated on poly(d-lysine)-coated wells in a defined, serum-free medium (N2, consisting of a 1:1 mixture of Ham's F-12 and minimal essential medium supplemented with 15 mM HEPES buffer, 5 μg/ml insulin, 100 μg/ml apo-transferrin, 100 μM putrescine, 20 nM progesterone, 30 nM selenium, 6 mg/ml glucose, and 1 mg/ml bovine serum albumin (BSA), and grown for 3 DIV. Cells were fixed with 4% paraformaldehyde (PFA) for 20 minutes prior to immunocytochemical analysis. For primary cultures, ventral mesencephala from E12.5 mouse embryos were dissected, treated with dispase/DNase, mechanically dissociated, and plated at a density of 125,000 cells per cm2 in N2 medium supplemented with 20 ng/ml basic fibroblast growth factor for 2 hours. Medium was then changed to 0.2-μm-filtered N2 that had been conditioned by c17.2 cells for 24 hours.

RNA Preparation, Real-Time Reverse Transcription-Polymerase Chain Reaction, and Quantification of Gene Expression

RNA from c17.2 cells was extracted using RNeasy Mini Kit (Qiagen, Hilden, Germany, http://www1.qiagen.com). Five μg of total RNA from two control (puroB [pB] and puroD [pD]) and two Nurr1-overexpressing cell lines (clone 42 [c42] and clone 48 [c48]) was treated with RQ1 RNase-free DNase (Promega, Madison, WI, http://www.promega.com), and RNA integrity was assessed by electrophoresis. RNA was also prepared from dissected ventral mesencephala of E11 wild-type and Nurr1-null embryos. Briefly, 1 μg of RNA was reverse-transcribed using SuperScript II Reverse Transcriptase (Invitrogen) and random primers (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) (reverse transcription-positive [RT+] reaction), and parallel reactions without reverse transcriptase enzyme were done as a control (RT reaction). Quantitative polymerase chain reaction (PCR) was performed as previously described [10].

High-Density Oligonucleotide Array Hybridization

RNA probes were labeled according to the supplier's instructions (Affymetrix, Santa Clara, CA, http://www.affymetrix.com). First-strand synthesis was carried out by a T7-(dT)24 primer and SuperScript II reverse transcriptase (Gibco-BRL, Gaithersburg, MD, http://www.gibcobrl.com) using 10 μg of total RNA sample. Second-strand synthesis was performed according to the SuperScript Choice System (Gibco-BRL) by Escherichia coli DNA-polymerase I, E. coli ligase, and RNase H. Fragment end polishing was performed using T4-polymerase. An in vitro transcription reaction was used to incorporate biotin-11-CTP and biotin-16-UTP into the cRNA probe (BioArray HighYield RNA Transcript Labeling Kit; Enzo Life Sciences, Farmingdale, NY, http://www.enzolifesciences.com). The fragmented cRNA was hybridized overnight (45°C) to the DNA chips. Washes were done in the GeneChip Fluidics Station (Affymetrix) according to the manufacturer's protocol. Staining was performed with R-phycoerythrin streptavidin (Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com), followed by an antibody amplification procedure using biotinylated anti-streptavidin antibody (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) and goat IgG (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com). The scanning was carried out at 3 μm resolution, 488 nm excitation, and 570 nm emission wavelengths using the GeneArray Scanner (Hewlett Packard, Palo Alto, CA, http://www.hewlettpackard.com). Two control (pB and pD) and two Nurr1-overexpressing (c42 and c48) cell lines were processed independently. Data were scaled based on total intensity, and data analysis was performed with Affymetrix software (MAS 5.0). The p values were calculated by applying the Wilcoxon's signed rank test and then used to generate the complex calls. Complete microarray data were registered at the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) and hybridization array data repository as GSE3571 (http://www.ncbi.nlm.nih.gov/geo).

Immunocytochemical Analysis

Cultures were blocked for 1 hour at room temperature in PBST (1× phosphate-buffered saline [PBS], 1% BSA, and 0.3% Triton X-100) and overnight at 4°C with the corresponding primary antibody diluted in blocking buffer. The following antibodies were used: mouse anti-β-tubulin type-III (TuJ1), 1:2,000 (Sigma-Aldrich); mouse anti-TH, 1:10,000 (Diasorin, Stillwater, MN, http://www.diasorin.com); mouse anti-Engrailed-1 (1:50), rabbit anti-glial fibrillary acidic protein, 1:500 (DAKO, Glostrup, Denmark, http://www.dako.com); rabbit anti-tenascin-C (TN-C), 1:250 (Chemicon, Temecula, CA, http://www.chemicon.com); mouse anti-cleaved caspase-3 (Asp175), 1:100 (Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com); and mouse Nestin (Rat-401), 1:500 (Developmental Studies Hybridoma Bank, University of Iowa). After washing, cultures were incubated with biotinylated secondary antibodies diluted 1:500 for 2 hours, and immunostaining was visualized using Vector Laboratories ABC immunoperoxidase kit and Supergrey substrate. For immunofluorescence, 1:100 dilutions of Cy2- or rhodamine-coupled secondary antibodies (Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com) were used. Cultures were then rinsed twice in PBS and analyzed.

Cell Death and Oxidative Stress Assays

For oxidative stress experiments, c17.2 cells were grown to 60% confluence, plated at a density of 5,000 cells per cm2, and allowed to grow overnight. The following day, fresh medium was added, and cells were treated with freshly diluted 25 μM H2O2 for 0 or 6 hours. Following peroxide treatment, cells were fixed with 4% PFA for 20 minutes, rinsed, and blocked with PBS/5% goat serum for 30 minutes at room temperature. Immunocytochemistry for cleaved caspase-3 was performed as described above. Next a 1:500 dilution of Hoechst 33328 (0.5 mg/ml) was added for 5 minutes at room temperature. Cells were washed twice in PBS, visualized under a Zeiss HBO100 microscope (Carl Zeiss, Jena, Germany, http://www.zeiss.com), and counted. For cell viability assays, cell lines were grown, plated, and treated as described above. Cells were stained with 100 μg/ml Hoechst 33342 and 2 μg/ml propidium iodide in ice-cold PBS. After 10 minutes, the unfixed, stained cells were immediately counted.

Tenascin-C Knockout Mice and Immunohistochemistry

TN-C-null mice were previously described [17, 18]. Embryos were fixed in ice-cold 4% PFA for 6 hours, cryoprotected in 20% sucrose, and frozen in OCT compound at −70°C. Next, serial coronal 14 μM sections of the entire ventral midbrain were cut on a cryostat. Immunohistochemistry was performed on 4% paraformaldehyde-postfixed slides. Incubations were carried out at 4°C overnight with the following antibodies diluted in PBS, pH 7.3, 1% BSA, and 0.3% Triton X-100 solution: rabbit anti-Nurr1 (1:500 dilution; Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com), rabbit anti-TH (1:1,000; Pel-Freeze Biologicals, Rogers, AR, http://www.pelfreez-bio.com), rat anti-bromodeoxyuridine (BrdU) (1:200; Abcam, Cambridge, U.K., http://www.abcam.com), mouse anti-phosphorylated histone H3 (Ser10) (1:100; Upstate, Charlottesville, VA, http://www.upstate.com), mouse anti-RC2 (1:200; DSHB), guinea pig anti-glutamate transporter, GLAST (1:2,000; Chemicon). After washes, the sections were blocked for 30 minutes in dilution buffer and then incubated for 2–4 hours with a secondary antibody (Cy2/3 [cyanine]-coupled donkey anti-mouse IgG, Cy2-coupled donkey anti-rabbit IgG [1:200 dilution; Jackson Immunoresearch Laboratories]). Immunostaining was visualized by using a Zeiss HBO100 microscope. For BrdU analysis, pregnant mice were injected with 0.5 ml of a 1 mg/ml BrdU/saline solution 6 hours before sacrifice. Pretreatment with 2 N HCl for 15 minutes prior to preincubation with primary antibody was needed for detection of BrdU. Stainings were visualized using a Zeiss HBO100 microscope, and images were collected with a Hamamatsu C4742-95 camera (Bridgewater, NJ, http://www.hamamatsu.com).

Results

Microarray Analysis of Nurr1 Neural Stem Cells

Although overexpression of Nurr1 facilitates the differentiation of neural stem cells [3] and ESCs [1, 2] into midbrain DA neurons, the effects of Nurr1 on downstream gene expression in a proliferating cell remain unclear. To investigate this, we determined the transcription profile of proliferating c17.2 neural stem cell lines stably transfected with Nurr1 [3]. Two Nurr1-overexpressing clones, clone 42 and clone 48, and two control clones transfected with a control vector, pB and pD, were chosen for this study. Clone 42 and clone 48 were selected among all the Nurr1-transfected cell lines generated because of their high potential to become TH+ dopaminergic neurons upon coculture with immature ventral mesencephalic glia [3]. Accumulated in vitro cell growth curves were generated (supplemental online Fig. 1A), and population doubling rates (PoDs) (the reciprocal of the doubling time in DIV) for the clones were estimated from these curves (PoD = 0.57 ± 0.05 for clone 42; PoD = 0.84 ± 0.02 for clone 48; PoD = 0.53 ± 0.09 for clone pB; PoD = 0.75 ± 0.01 for clone pD). To minimize the effects of proliferation on the study, the Nurr1 and corresponding control clones were matched (c42 with pB and c48 with pB) according to their doubling rates and further processed for gene expression comparison (supplemental online Fig. 1A).

The overall gene expression pattern of c42 versus clone pB and of c48 versus clone pD was compared using the Affymetrix murine U74v2 type A microarray. Total RNA from the c17.2 clones was isolated, reverse-transcribed, and hybridized to the chip. Of the >12,000 genes queried, a total of 1,949 genes were upregulated (1,081 in the c42/pB subtraction and 868 in the c48/pD subtraction), whereas 2,502 genes were downregulated (1,472 in c42/pB and 1,030 in c48/pD). However, only those genes regulated in a coordinate manner between the two independent comparisons were considered for further validation and quantitative analysis. These genes included 190 activated genes and 203 repressed genes (supplemental online Fig. 1B).

Classification and Confirmation of Genes Regulated by Nurr1 in Neural Stem Cells

The distribution of identified genes with known or putative functions is shown in supplemental online Figure 1C, and a complete list is available online (NCBI/GEO GSE3571). The highest number of genes up- and downregulated represented expressed sequence tags (23%) whose functions are unknown. The majority of characterized genes regulated encoded transcription factors (11%) and proteins involved in intracellular signaling (15%). The expression of several secreted ligand (3%)-inducing growth factors and extracellular matrix/cell adhesion molecules (13%) was also regulated. In addition, Nurr1 overexpression downregulated a number of cell cycle genes (4%), including cyclin B2, D1, and p21, whose regulation has been implicated in the differentiation of ventral progenitors [19]. Nurr1 overexpression also resulted in the upregulation of trophic/survival signals that coincided with the downregulation of a number of proapoptotic genes (10%) (downregulation of caspase-3 and -11). For a more detailed analysis, we focused our studies on those genes either up- or downregulated at least twofold or an average of 2.5-fold in both of the independent comparisons and classified them by broad biological function (Tables 1, 2, respectively). Many genes were represented several times in the chip and showed concordant expression, thereby validating the hybridization. To confirm the results of the microarray, 11 genes were randomly selected for quantitative PCR analysis either in the respective clones or in c17.2 neural stem cells transiently transfected with Nurr1 (data not shown). We found that in virtually all cases analyzed (10 of 11), the selected genes were corroborated (Table 3). To extend the validation of our microarray analysis further, we analyzed the expression of eight selected genes in three additional Nurr1-expressing clones: clones 12 (c12), 19 (c19), and 26 (c26). We observed that in all cases except one, the genes regulated in c42 and c48 were also regulated in c12, c19, and c26 (supplemental online Fig. 2). These data, obtained in five independent Nurr1-expressing clones, confirm that several of the genes identified in the array are regulated by Nurr1.

Table Table 1.. Functional classification of genes upregulated in c17.2-Nurr1 stem cells
original image
Table Table 2.. Functional classification of genes downregulated in c17.2-Nurr1 stem cells
original image
Table Table 3.. Validation of microarray genes by quantitative PCR
original image

Nurr1 Induces the Expression of Neuronal Markers in NSCs

Several lines of evidence, including our microarray, have suggested distinct functions for Nurr1 in neuronal differentiation [20, 21]. To investigate the effects of Nurr1 in this process, we stained both control and Nurr1 NSC clones with several markers that define stages of progressive neuronal differentiation. Both control and Nurr1 NSCs stained positive for the neural stem cell marker Nestin (Fig. 1A). However, Nurr1-overexpressing clones exhibited higher levels of immunoreactivity for the neuronal cytoskeletal protein Tuj1 and Engrailed-1 (Fig. 1A), both upregulated 4- and 2.4-fold, respectively, in the microarray. Nurr1-overexpressing NSCs were also immunoreactive for TN-C, an extracellular matrix protein critical for the development and maintenance of several neuronal types [17, 18] (Fig. 1B). Consistent with these observations was the finding that virtually all Nurr1-overexpressing cells, but not control cells, acquired a bipolar neuronal morphology and were double-labeled with antibodies against Tuj1 and TN-C upon differentiation in defined medium without mitogens (Fig. 1B). Similar results were observed with c48 and pD (data not shown). In addition, both Nurr1 clones secreted significantly higher levels of TN-C protein compared with their respective control cell lines (data not shown). These results confirm that elevated Nurr1 levels in undifferentiated neural stem cells activate the expression of several genes (tubulin β III, neurofilament, and TN-C) that define a neuronal phenotype as identified by the microarray analysis. Next, we sought to determine whether the strong upregulation observed in Nurr1-expressing clones of TN-C was associated with Nurr1 expression. We addressed this by assessing TN-C transcript levels in ventral mesencephala isolated from E11 Nurr1-null embryos. Our quantitative PCR analysis indicates that ablation of Nurr1 expression resulted in diminished TN-C levels in vivo, confirming that TN-C is a downstream target of Nurr1 (Fig. 1C).

Figure Figure 1..

Nurr1 regulates the expression of stem/progenitor cell markers and promotes differentiation. (A, B): Immunocytochemical detection of Nestin, Engrailed-1, and Tuj1 (A) and tenascin-C (TN-C) (B) in both control (pB) and Nurr1 (c42) NSCs proliferating in the presence of serum. (B): Immunocytochemical detection of Tuj1 and TN-C in c42 differentiated for 3 days in the absence of serum. (C): TN-C transcripts levels were significantly reduced in the ventral mesecephala of Nurr1-null mice. For each experiment (n = 3), statistical analysis was performed using the paired t test (∗∗∗, p < .001). Abbreviations: c42, clone 42; pB, puroB; TN-C, tenascin-C; Tuj1, tubulin β III.

Nurr1 Overexpression Promotes Survival and Resistance to Oxidative Stress

The results of the GeneChip analysis revealed several trophic factors and survival signals upregulated by Nurr1, including transforming growth factor-β (TGF-β) and nerve growth factor (upregulated at least twofold in both clones). To assess whether the trophic factors expressed by Nurr1-overexpressing cells were biologically active, conditioned medium from control and Nurr1-overexpressing cell lines was collected. Addition of conditioned medium from c42 or c48 cells to ventral mesencephalic E12.5 primary cultures increased the number of TH+ dopaminergic neurons compared with the addition of conditioned medium from both control cell lines (Fig. 2A, 2B), suggesting that the trophic factors expressed by Nurr1 clones do indeed promote the survival and/or differentiation of DA neurons.

Figure Figure 2..

Nurr1 NSCs secrete trophic and survival signals that increase the number of dopaminergic neurons in mouse primary cultures. (A): Dissociated cells from embryonic day 12.5 (E12.5) mouse ventral mesencephala (VM) were treated with filtered N2 medium that had been conditioned by Nurr1 or control clones for 24h. Cells were fixed after 3 days in vitro and immunostained for TH. (B): Total numbers of TH+ neurons in VM E12.5 primary cultures treated with conditioned medium. Values represent mean ± SEM of the total number of TH+ cells per square centimeter of three independent experiments done in triplicate. Statistical analysis was performed by one-way analysis of variance with Fisher's post hoc test (∗, p < .05; ∗∗, p < .001). Abbreviations: c42, clone 42; pB, puroB; pD, puroD; TH, tyrosine hydroxylase.

In addition to secreted factors, several survival, stress response/DNA repair, and antiapoptotic factors were also induced (e.g., LRG-21, gadd45β, and protein kinase C) in the Nurr1 clones. Notably, a number of proapoptotic genes were downregulated (e.g., ASM-like phosphodiesterase, Mekk1, caspase-3, and caspase-11). An upregulation of several genes (e.g., oxidative stress-induced protein and cytochrome b5) involved in regulating the cellular response to oxidative stress was also observed (upregulated approximately 1.5-fold in both clones). Oxidative stress has been implicated in the pathophysiology of several neurodegenerative diseases, including PD [22, [23]–24]. We therefore examined whether Nurr1 expression protected cells from death induced by oxidative stress. Both sets of Nurr1-overexpressing neural stem cells and controls were treated with several concentrations of H2O2, and viable cells were counted after 6 hours. Peroxide treatments above 25 μM H2O2 were effective at inducing cell death in both sets of clones (data not shown). Cultures from both sets of control and Nurr1-overexpressing cells were treated with 25 μM peroxide and then stained with propidium iodide to identify dead cells. Significantly (40%) less cell death was observed in cells overexpressing Nurr1 compared with controls (Fig. 3A, 3B). To confirm the extent of apoptosis upon peroxide treatment, we assessed the presence of cleaved caspase-3 by immunocytochemistry. Fewer Nurr1-overexpressing cells were immunoreactive for cleaved caspase-3 compared with controls (Fig. 3C, 3D). Next, we examined whether Nurr1-overexpressing cells were resistant to peroxide-induced oxidative stress through secreted extrinsic factors or through cell-autonomous mechanisms. Control clones pB and pD were cultured with conditioned medium from Nurr1-expressing clones c42 and c48 and subjected to peroxide treatment. Conditioned medium from Nurr1 failed to rescue the peroxide-induced death associated with oxidative stress (data not shown). Together, these results demonstrate that Nurr1 promotes cell survival through secreted factors but helps confer resistance to oxidative stress through cell-intrinsic mechanisms by downregulating several proteins involved in the apoptotic response.

Figure Figure 3..

Nurr1 expression protects cells from oxidative stress-induced death. (A, B): Nurr1 clones were resistant to oxidative-stress induced death. Control and Nurr1 clones were treated with 25 μM H2O2 for 6 hours, stained with PI and Hoechst 33342, and counted (n = 4). Data represent the mean percent increase in cell death after 6 hours of peroxide treatment (Hoechst/PI). (C, D): Fewer Nurr1 NSCs were immunoreactive for cleaved caspase-3 following exposure to oxidative stress (n = 4). Statistical analysis was performed by one-way analysis of variance with Fisher's post hoc test (∗, p < .05; ∗∗, p < .001; ∗∗∗, p < .0001). Abbreviations: c42, clone 42; pB, puroB; pD, puroD; PI, propidium iodide.

Premature Differentiation of DA Neurons in Tenascin-C-Null Mice

TN-C expression has been previously detected throughout the ventral mesencephalon from E9 to E13. However, the role of TN-C in the development of dopaminergic neurons remains largely unknown [25]. Our gene expression profiles suggested an important role for TN-C in developing DA neurons given its upregulation in Nurr1 clones (Table 1). To investigate the role of TN-C in vivo, we analyzed tenascin-C knockout mice. At the onset of dopaminergic neurogenesis (E11), the number of tyrosine hydroxylase-immunoreactive cells was increased in TN-C-null mice (Fig. 4A, 4B). This increase was particularly confined to the more lateral dopaminergic neuron population. However, the excess of dopaminergic neurons was corrected for by E15 and postnatal day 1, a time period where neurogenesis has ended in vivo (Fig. 4B). We hypothesized that the early phenotype observed in TN-C knockouts could reflect an overproduction of proliferating DA progenitors or a change in the proliferative capacity of the ventral mesencephala. To assess this, we compared the numbers of proliferating cells between genotypes using BrdU, a marker that labels cells in the S phase of the cell cycle. BrdU analysis revealed no change in the number of proliferating cells in the ventral mesencephalon between the wild-type and TN-C-null mice at E11 (767.66 ± 59.88 and 708.00 ± 37.32, respectively; Fig. 4C) or E12.5 (data not shown). These results were further confirmed using an additional marker of proliferating cells, phosphorylated histone H3 (data not shown). Because the number of active caspase-3-immunoreactive cells did not differ between wild-type and TN-C-null mice (data not shown), these data suggest that deletion of TN-C does not increase the number of DA neurons by altering proliferation or cell survival. We next examined radial glia, cells that can act as stem cells and direct region-specific neural differentiation in the central nervous system [24, 26], including the developing midbrain (A.C. Hall et al., unpublished data). However, no change in RC2 or GLAST immunoreactivity was observed in TN-C-null mice at E11 (Fig. 4D; data not shown). In addition, TN-C knockout mice were analyzed for changes in Nurr1+ DA precursors. At E11, no alterations in Nurr1+ cells were observed between wild-type and TN-C knockout mice (Fig. 4D, 4E). These data indicate that TN-C normally acts to delay the acquisition of TH in Nurr1+ precursors in vivo and that upon its deletion, the DA differentiation process is accelerated at E11, but the total number of TH+ cells at the end of neurogenesis is not altered. Thus, our results suggest that Nurr1+ cells, by regulating TN-C expression, delay their own differentiation into TH+ neurons.

Figure Figure 4..

Tenascin-C-null mice display premature dopaminergic differentiation in vivo. (A): Coronal sections through embryonic day 11 (E11) ventral mesencephala (VM) revealed that the number of DA neurons (assessed by TH immunoreactivity) is increased in TN-C-null mice compared with controls. (B): Quantifications of TH+ cells through serial coronal sections of E11, E15, and P1 control and TN-C-null mice. A correction in the number of TH+ cells was observed by E15 and at P1 in TN-C-null mice. (C): Proliferating progenitors were identified by bromodeoxyuridine (BrdU) uptake in vivo at E11 in the VM. The proliferative capacity of E11 DA precursors was not affected in TN-C-null mice compared with control animals, as assessed by immunohistochemistry against BrdU. (D, E): The total numbers of Nurr1+ cells and radial glia (GLAST) were not altered at E11 in TN-C-null mice compared with WT controls. For each experiment (n = 5), statistical analysis was performed using the paired t test (∗, p < .05). Abbreviations: P1, postnatal day 1; TH, tyrosine hydroxylase; TN-C, tenascin-C; WT, wild-type.

Discussion

To gain more insights into the molecular pathways regulated by Nurr1 during development, we performed an analysis of genes regulated by Nurr1 in a proliferating neural stem cell line using Affymetrix microarrays. Of more than 12,000 genes queried, we identified approximately 400 genes that showed regulation by Nurr1. The findings we report here support a role for Nurr1 in reducing the potential of a neural stem/precursor cell by initiating its differentiation into a neuronal phenotype. Interestingly, this process was accompanied by an increase in the expression of genes that provide regional identity in the brain and confer cell survival and oxidative stress resistance. Thus, Nurr1 appears to coordinate the function of diverse genes during development.

The role of Nurr1 in promoting stem cell differentiation is supported by data from this microarray indicating that expression of neural stem cell markers, mitogens and their cognate receptors, and proteins that control cell cycle progression are downregulated (NCBI/GEO GSE3571). The downregulation of these genes could contribute to the coupling of cell cycle exit and differentiation programs in Nurr1+ cells. Interestingly, Nurr1 induces cell cycle arrest and postmitotic maturation in cooperation with the cyclin dependent kinase inhibitor p57kip2 in the dopamine-synthesizing neuronal cell line MN9D [27]. Accordingly, we have found several cell cycle genes downregulated in our study, in particular cyclin B2, D1, and p21 (all downregulated at least 1.5-fold in both comparisons and confirmed by quantitative PCR; data not shown).

Our results suggest that proliferating c17.2-Nurr1 cells represent a transition state from the multipotent neural stem cell to a precursor/immature neuron. Accordingly, the cytoskeletal protein encoding the neurofilament medium polypeptide was strongly induced in the Nurr1 clones. This result, together with the colocalization of tubulin β III (an early neuronal marker) and tenascin-C (involved in neurogenesis) in Nurr1 clones, suggests that proliferating c17.2-Nurr1 cells have initiated a precocious neuronal differentiation program. A strong repression of B-FABP expression was also detected: a fatty acid transporter that is highly expressed in neuroepithelial and radial glia cells in the developing brain and enriched in NSC cultures [28, 29]. Accordingly, the expression of several genes involved in non-neuronal development (such as that encoding the muscle cytoskeletal protein Myosin X, which is downregulated 2.5-fold in both clones) was repressed. Together, these results indicate that Nurr1 expression has positive effects on the development of a neuronal phenotype in NSCs.

Another important set of genes controlled by Nurr1 in a coordinated manner consists of those involved in promoting cell survival. Several studies have identified proteins critical for the induction and survival of mesencephalic dopaminergic neurons [30, 31]. Furthermore, retinoid X receptor (RXR) ligands that activate Nurr1-RXR heterodimers have been demonstrated to be critical for the survival of dopaminergic neurons [32]. Here, we observed an upregulation of several neurotrophic factors and downregulation of proapoptotic genes in Nurr1-overexpressing NSCs. Among the genes upregulated in Nurr1-overexpressing NSCs were two trophic signals required for the induction (TGF-β) and postnatal development (glial cell line-derived neurotrophic factor [GDNF]) of dopaminergic neurons [33, 34]. In fact, both TGF-β and GDNF have been shown to alleviate functional deficits in animal models of PD [35, 36]. Consistent with an upregulation of trophic factors was the observation that conditioned medium from c42 or c48 cells increased the number of TH+ dopaminergic neurons in ventral mesencephalic E12.5 primary cultures compared with control conditioned medium. Our results suggest that Nurr1 overexpression induces the expression of several signals later required for the survival of dopaminergic neurons.

Moreover, both survival and stress response/DNA repair/antiapoptotic factors were induced (e.g., LRG-21 and gadd45β, upregulated 2.5- and 1.6-fold, respectively), whereas proapoptotic genes were downregulated (e.g., caspase-3, and caspase-11) in Nurr1 clones. Another gene regulated by Nurr1 in c17.2 cells was LRG-21, a member of the activating transcription factor/cAMP response element-binding protein family that is involved in the cellular stress response [37] and is regulated by seizures [38] and axotomy [39]. We also detected increased levels of gadd45β, an inducible factor that belongs to a protein family associated with cell-cycle control and DNA repair [40] and downregulates proapoptotic c-Jun N-terminal kinase signaling [41]. Thus, combined, our results indicate that one of the main functions of Nurr1 is to promote survival by orchestrating the expression of stress response and antiapoptotic genes.

To further investigate the prosurvival effects of Nurr1, we examined whether Nurr1 expression protected cells from death induced by oxidative stress. The potential for oxidative damage is intrinsically associated with Nurr1+/TH+ dopaminergic neurons, since tyrosine hydroxylase activity forms oxygen radicals in a redox mechanism with its cofactor [42]. Cell survival upon oxidative treatments above 25 μM H2O2 was greater for Nurr1 clones compared with controls (Fig. 3A–3D). Microarray data indicating an upregulation in the expression of several oxidative-stress proteins important for maintaining free radical homeostasis in Nurr1-overexpressing cells (NCBI/GEO GSE3571) also supported these findings. Nurr1 NSCs also exhibited lower levels of cleaved caspase-3 protein after peroxide treatment compared with controls. This result is in agreement with previous findings showing that deletion of one Nurr1 allele decreases the survival of DA neurons in vitro [43] and that decreased Nurr1 expression increases the vulnerability of DA neurons to 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), a neurotoxin that induces oxidative stress and PD-like symptoms in rodents [14, 15, 24]. Nurr1 clones also expressed lower levels of caspase-11. Recent reports have indicated that caspase-11 knockout mice are more resistant to MPTP-mediated cell death than wild-type mice [44]. Moreover, the recent identification of Nurr1 mutations in PD patients [45] further suggests the intriguing possibility that the expression of survival and/or cell death genes regulated by Nurr1 could be altered in PD.

One of the highest induction rates in proliferating c17.2 Nurr1 NSCs corresponded to tenascin-C, an extracellular matrix glycoprotein. Tenascin-C is differentially expressed in specific patterns throughout nervous system, and although several genomic studies have previously implicated increased TN-C expression in maintaining the “stemness” of mouse neurospheres and rat neural progenitors, the precise role of TN-C remains unclear [46, 47]. Tenascin-C expression also correlates well with the migration, differentiation, and decreased proliferation of oligodendrocyte precursor cells, as well as axonal growth and guidance in neurons [48, 49]. Interestingly, analysis of TN-C levels in E11 Nurr1-null embryos revealed that TN-C is regulated by Nurr1 levels in vivo during dopaminergic neurogenesis (Fig. 1C). Moreover, adult mice lacking TN-C exhibit hyperlocomotive activities that have been attributed to altered levels of tyrosine hydroxylase and dopamine [17]. However, the observed increase in locomotor activity cannot be attributed to increased DA cell number since TN-C-null mice display higher numbers of DA neurons at the onset, but not at the end, of DA neurogenesis. Our results suggest the occurrence of a premature differentiation of Nurr1+ precursors in TN-C-null mice at E11 that normalizes by E15. These data are in agreement with other findings showing normal phenotypes elsewhere later in development [49, 50]. Our data also indicate that at the onset of neurogenesis, TN-C does not regulate the size of the proliferative progenitor pool but instead contributes to preserving Nurr1+ postmitotic precursors in a TH state. Recent reports have described TN-C as a regulator of Wnt signaling [51]. Studies from our laboratory have also shown that Wnt signals and β-catenin overexpression regulate the maintenance and conversion of Nurr1+ precursors into DA neurons [10, 52]. Thus, our results highlight a novel function for TN-C to maintain early Nurr1+ postmitotic precursors in an undifferentiated state. These results further our understanding of how Nurr1 promotes neuronal differentiation.

Disclosures

E.A. owns stock in and has served as a scientific officer of and on the board of Neurotherapeutics AB within the last 2 years.

Acknowledgements

We thank Drs. Helder Andre, Clare Parish, Tibor Harkany, Carmen Salto, and Paola Sacchetti for critical reading and help with the manuscript. We also thank Jun Inoue for technical support, Dr. Dirk Koczan for help with the microarray data analysis, and Thomas Perlmann for providing Nurr1-null mice. Financial support was obtained from the Swedish Foundation for Strategic Research, Swedish Royal Academy of Sciences, Knut and Alice Wallenberg Foundation, Swedish Medical Research Council, Michael J. Fox Foundation, European Commission, Juvenile Diabetes Foundation, EuroStemCell, Karolinska Institute, and Petrus and Augusta Hedlunds Foundation. H.M. was supported by the Karolinska Institute. A.C.H. was supported by the Human Frontier of Science Program. K.M.S. and H.M. contributed equally to this work.

Ancillary