Generation of Dopamine Neurons with Improved Cell Survival and Phenotype Maintenance Using a Degradation-Resistant Nurr1 Mutant

Nurr1 is a transcription factor specific for the development and maintenance of the midbrain dopamine (DA) neurons. Exogenous Nurr1 in neural precursor (NP) cells induces the differentiation of DA neurons in vitro that are capable of reversing motor dysfunctions in a rodent model for Parkinson disease. The promise of this therapeutic approach, however, is unclear due to poor cell survival and phenotype loss of DA cells after transplantation. We herein demonstrate that Nurr1 proteins undergo ubiquitin-proteasome-system-mediated degradation in differentiating NP cells. The degradation process is activated by a direct Akt-mediated phosphorylation of Nurr1 proteins and can be prevented by abolishing the Akt-target sequence in Nurr1 (Nurr1Akt). Overexpression of Nurr1Akt in NP cells yielded DA neurons in which Nurr1 protein levels were maintained for prolonged periods. The sustained Nurr1 expression endowed the Nurr1Akt-induced DA neurons with resistance to toxic stimuli, enhanced survival, and sustained DA phenotypes in vitro and in vivo after transplantation.


INTRODUCTION
Midbrain dopamine (DA) neurons play essential roles in the control of voluntary movement and the regulation of emotion. Degeneration/dysfunction of this neuronal subtype underlies clinical features of many neurological and psychiatric disorders. Nurr1 (NR4A2), a transcription factor belonging to the orphan nuclear receptor family, is expressed in the developing midbrain and is critical for midbrain DA neuron development [1,2]. Nurr1 is also expressed in the DA neurons of the adult midbrain, and sustained expression of this factor has been reported to be crucial for the maintenance of dopaminergic phenotypes [1,3] and survival [2,4,5]. Reduced levels and genetic alterations of Nurr1 in adult midbrain DA neurons have been found in midbrain DA pathologies [6,7], indicating that a potential therapeutic strategy could be established through manipulation of Nurr1 protein level and function in patients with those disorders.
Neural precursor (NP) cells can be isolated from developing and adult brains, and cultured for the purpose of generating large numbers of donor cells to treat neurodegenerative disorders. Interest in Nurr1 has intensified due to its in vitro role in DA neuron generation from cultured NP cells. Exogenous Nurr1 expression in the absence [8] or presence of neurogenic factor coexpressions drives naïve nondopaminergic NP cells to differentiate into DA neurons that exhibit presynaptic functionalities capable of reversing dopaminergic deficits in a rodent model of Parkinson disease (PD). However, poor cell survival [8] and loss of DA phenotype of donor cells [9] after transplantation are the most critical concerns in these procedures. The proteasomal degradation system is a critical regulator of protein activity in a cell, with various cellular proteins targeted to the proteasome for degradation by the covalent addition of multiple molecules of ubiquitin, a 76-amino acid polypeptide. In this report, we demonstrate that Nurr1 proteins undergo ubiquitin-proteasome-system (UPS)-mediated degradation in differentiating NP cells. Intracellular signals responsible for the protein degradation were defined and this molecular understanding of the degradation process led us to generate a ubiquitylation-resistant Nurr1 mutant. Induction of mutant protein expression in NP cells yielded DA neurons with Nurr1 protein levels that were stably maintained for a prolonged period while preserving native Nurr1 functions. As a consequence, DA neurons generated by the mutant Nurr1 were resistant to toxic stimuli and exhibited enhanced cell survival in vitro and in vivo in rat brains after transplantation. These findings represent a substantial technical advance in stem/precursor cell-derived DA neuron generation for PD cell therapy and provide important cues for developing strategies to prevent PD progression.

Retroviral Construction and Infection
Retroviral vectors expressing Flag-tagged wild-type Nurr1 (Nurr1 WT ), Nurr1 mutant (Nurr1 Akt ), dominant negative form of Raf (dn-raf; kindly provided by Dr. Kang-Yell Choi, Yonsei University, Seoul, Korea), Wnt5a, and Notch intracellular domain (kindly provided by Dr. Jaesang Kim, Ewha Woman University, Seoul, Korea) were constructed by inserting each cDNA fragment into the multicloning sites of pCL [10]. Viral particles were produced by transfecting the retrovirus packaging cell line 293gpg with each vector using Lipofectamine (Invitrogen) and supernatants containing viral particles were harvested 72 hours after incubation. For viral transduction, prepared NP cells were incubated with the viral soup (5 Â 10 6 particles/ml) containing polybrene (1 lg/ml; Sigma-Genosys) for 2 hours, followed by a medium change with bFGF-supplemented N2. Coexpression studies were carried out by infecting cells with mixtures of the individual viral constructs (1:1).

Immunofluorescent Staining
Cultured cells and brain tissues were fixed with 4% paraformaldehyde, blocked in 0.1% bovine serum albumin (BSA)/10% goat serum/0.3% Triton X-100 and incubated with primary antibodies overnight at 4 C. For detecting Nurr1-expressing cells grafted in brain sections, an antigen retrieval procedure was applied by treating cells with sodium dodecyl sulfate (1% in phosphate-buffered saline [PBS]) at room temperature for 5 minutes before the blocking procedure. The following primary antibodies were used: Nurr1

Nurr1 Protein Stability
HEK293 cells were transfected with Nurr1 WT (or Nurr1 Akt ) and harvested during 6 hours of cycloheximide (100 lg/ml; Calbiochem) treatment. Nurr1 protein levels were determined by Western blot analyses.

In Vivo Transplantation
The 6-OHDA-lesioned rats were generated as described [12]. Neural precursor cells were harvested 2 days after Nurr1 WT or mutant transduction and dissociated into single cells in HBSS. Using a 22-gauge needle, 3 ll of cell suspension (1.5 Â 10 5 cells/ll in N2þbFGF) was deposited at the striatum (coordinates in AP, ML, and V relative to bregma and dura: -0.09, -0.42, -0.66 incisor bar set at 3.5 mm). The needle was left in place for 10 minutes following each injection. For histological analysis, animals were anesthetized with ketamine (4.5 mg/kg) mixed with rompun (93.28 lg/kg) and perfused transcardially with 4% paraformaldehyde in PBS. Brains were equilibrated with 30% sucrose in PBS and sliced on a freezing microtome (CM 1850, Leica). Free-floating brain sections (35-lm thick) were subjected to immunohistochemistry as described above. The total numbers of cells positive for Nurr1, TH, and DAPI in the graft were estimated by the Abercrombie correction factor [13].

Cell Counting and Statistic Analysis
Cell counting was performed in microscopic fields randomly chosen (fractionator) across the culture area. Data are expressed as mean AE SEM of three independent experiments. Statistical comparisons were made by Student's t test or one-way analysis of variance (ANOVA) with post-hoc test using SPSS software (version 13.0; SPSS Inc., Chicago, IL, http://www.spss.com).

Degradation of Exogenous Nurr1 Proteins During In Vitro Precursor Differentiation
Nurr1 expression was induced in cultured NP cells derived from rat fetal cortices. As demonstrated previously [10,14,15], exogenous Nurr1 expression yielded cells positive for TH, a key marker for DA neurons, from naïve nondopaminergic NP cells. Nurr1 immunoreactivity was localized in the nucleus of virtually all THþ cells for up to 4 days of differentiation (Fig. 1F). The number of Nurr1-immunoreactive cells and levels of Nurr1 protein gradually decreased during the longer differentiation period (Fig. 1A-H, 1M, 1N) but without a significant change in Nurr1 mRNA levels (Fig. 1O), consequently yielding increased populations of THþ cells which were negative for Nurr1 immunoreactivity (insets of Fig. 1G, 1H). In addition, reduced Nurr1 expression was followed by a substantial reduction in THþ cell numbers after prolonged differentiation (Fig. 1F-H , 1M). As a control, protein levels of exogenous LacZ, expressed in a vector construct identical to that of Nurr1, were observed to be uniform and without variation throughout the cell differentiation period (Fig. 1I-L, 1M, 1N). Treatment of the cells with proteasome inhibitors MG132 or lactacystin significantly blocked Nurr1 protein degradation ( Fig. 1P and 1Q). At differentiation day 6, Nurr1þ cells accounted for 15.2 AE 10.4% of total cells in untreated control versus 54.1 AE 10.6% in MG132 (10 lM)treated cells (total 13,833 and 12,573 cells counted from three sets of independent cultures, p < 0.01, Student's t test). In an IP assay, Nurr1 proteins bound directly to ubiquitin (Ub) and slower migrating forms that corresponded to polyubiquitinylated species were visible (Fig. 1R). Leptomycin B, an irreversible inhibitor of CRM-1-dependent nuclear export [16], had no effect on Nurr1 decay (data not shown), suggesting degradation of Nurr1 in the nucleus.

bFGF Prevents Nurr1 Protein Degradation
In contrast to exogenous Nurr1 protein decay during cell differentiation, the protein levels of Nurr1 in proliferating NP cells were maintained in the continued presence of bFGF in culture (Fig. 2D, 2E, 2G). These findings prompted us to determine if Nurr1 protein decay can be prevented by other mitogens acting on NP cells. Proliferation of NP cells was similarly induced by EGF treatment or activation of Notch signaling via Notch intracellular domain transduction (data not shown). However, neither of these factors was able to maintain the Nurr1 protein stability elicited by bFGF (Fig. 2D-H). Nurr1 protein levels were also not significantly altered by treatments with factors regulating NP cell differentiation (BDNF, NT3, FGF20, Wnt-5a) and survival (Fas ligand, pan-caspase inhibitor; Fig. 2H). The bFGF-sustained Nurr1 protein levels were abolished by treatment of cells with SU5402, an FGF receptor blocker (Fig. 3c). Together, these results suggest that the maintenance of Nurr1 proteins is specifically mediated by bFGF.

Counteracting Regulatory Actions of Raf-and Akt-Mediated Intracellular Signals in Nurr1 Protein Stability
We next sought intracellular signals that act downstream of bFGF to maintain Nurr1 protein stability. To this end, we explored time-course changes of protein levels of activated (phosphorylated) forms of potential signaling molecules for the period after bFGF withdrawal. An immediate decrease of pERK levels was observed within 15 hours after bFGF withdrawal and pERK was present at reduced levels for the remainder of the differentiation period tested (Fig. 3A, 3B). In contrast, levels of pAkt, another potential signaling molecule downstream of bFGF [17,18], were slightly decreased during the initial period of bFGF withdrawal, but gradually and substantially increased for the rest of the differentiation period. The activation of Akt signaling was likely caused by the decrease in Raf/ERK activation, as the pAkt and Raf-ERK signals have been shown to mutually regulate each other in an inhibitory manner [19,20].
Inhibition of Raf-Erk signaling by the specific inhibitors PD98059 and U0126 or transduction of a dn-raf resulted in a marked reduction of Nurr1 protein levels ( Fig. 3C and data not shown). On the contrary, the PI3K-Akt signal blockers LY294002 and wortmannin resulted in a striking increase of Nurr1 protein levels (Fig. 3D). To determine if the Nurr1 protein level changes were caused by Erk-or Akt-mediated regulation of protein degradation, we compared the stabilities of Nurr1 proteins in the absence and presence of inhibitors for these signaling molecules. Nurr1 proteins were readily degraded within 6 hours of cycloheximide treatment (Fig.  3E). Further drastic reduction of Nurr1 was observed to occur rapidly after PD98059 treatment. In contrast, Nurr1 protein levels in the cultures treated with LY294002 were stable after 6 hours of cycloheximide treatment, confirming counterregulatory roles of Raf-Erk and Akt signals in Nurr1 protein degradation.

Direct Akt Phosphorylation Is Responsible for Nurr1 Ubiquitylation
We next investigated the possibility that Raf-Erk and Akt molecules function through direct interaction with Nurr1 proteins. Direct protein interactions of Akt (Fig. 4A) and Erk1/2/5 with Nurr1 were observed in IP assays [21,22] (data not shown). Differentiation-dependent decreases of a Nurr1 mutant protein, in which all three Erk phosphorylation consensus sites were abolished, was comparable and insignificantly different from that of wild-type Nurr1 (data not shown), ruling out the possibility of direct Erk phosphorylation of Nurr1 mediating the maintenance effect. The Nurr1 protein contains a consensus site for Akt phosphorylation at Ser 347 (Fig. 4B). As shown in Figure 4C, pAkt substrate antibody, which recognizes phosphorylated peptides and proteins at the Akt target motif (RXRXXS/T), readily binds to Nurr1 WT , but not to the Nur-r1 Akt in which serine 347 is substituted by alanine, indicating Akt-mediated phosphorylation of Nurr1 and abolishment of this Akt-mediated phosphorylation in the mutant.
The effect of the mutation was dramatic, with a clear difference in the protein stability of Nurr1 Akt compared to that of Nurr1 WT after cycloheximide treatment (Fig. 4D). Nurr1 ubiquitylation was significantly reduced in Nurr1 Akt -transfected cells, indicating that the effect of the Akt mutation is elicited by preventing initiation of UPS-mediated Nurr1 protein degradation (Fig. 4E).

Phenotype Maintenance and Cell Survival of TH1 DA Cells Generated by Nurr1 Akt Transduction
Consistent with the sustained protein stability of Nurr1 Akt , no significant decreases in the percentage of Nurr1þ cells were seen in the cortical precursor cells transduced with Nurr1 Akt during 12 days of differentiation in vitro (percent Nurr1þ cells out of total cells: 67.5 AE 1.8% (Diff0), 69.2 AE 4.0% (Diff4), 74.5 AE 12.8% (Diff8), and 63.9 AE 1.4% (Diff12); p ¼ 1.0 compared to Diff0 and Diff12, n ¼ 3 sets of independent cultures, one-way ANOVA with post-hoc test, total 6,012-13,260 cells counted), whereas a greater than 60% loss in the percentage of Nurr1þ cells was observed during the same period of time in Nurr1 WT -transduced cultures (Fig. 4F-K). We further confirmed that the effect of the Akt mutation is not likely due to enhanced transcription of the Nurr1 gene, as mRNA levels of Nurr1 Akt and Nurr1 WT were indistinguishable (data not shown). Along with the sustained Nurr1 protein levels, THþ cell numbers in the Nurr1 Akt -transduced cultures were stably maintained during the in vitro differentiation period. For instance, the percentages of THþ cells were 66.5 AE 2.8% at Diff4, 60.3 AE 7.4% at Diff8, and 71.5 AE 2.0% at Diff 12 (total 7,005, 11,985, and 13,260 cells counted, respectively, from three independent experiments, p ¼ .873 compared to Diff0 and Diff12 by ANOVA with post-hoc Bonferroni test) in the Nurr1 Akt -transduced cultures (Fig. 4L-Q, 4S).
Messenger RNA expression of DA phenotype genes such as TH and DA transporter were increased in Nurr1 Akt -transduced cultures compared to control cultures (data not shown). Maintenance of the DA phenotype in Nurr1 Akt -transduced cultures is likely to be achieved by the continuous transcriptional activation resulting from sustained levels of Nurr1 Akt proteins. In addition, Nurr1 may have a cell survival effect in THþ cells because Nurr1 has been shown to act as a cell survival factor [2,4,5]. Cell apoptosis was markedly decreased in Nurr1 Akttransduced cultures than in those with Nurr1 WT (percent cells with apoptotic nuclei: 2.7 AE 0.7% in Nurr1 Akt vs. 13.2 AE 1.3% in Nurr1 WT -transduced cultures at Diff6, total 11,397 and 11,462 cells counted, n ¼ 3, p < .01). Furthermore, Nurr1 Akt -  transduced cells were more resistant to the cellular toxicity induced by H 2 O 2 and 6-OHDA treatments based on our estimation of cell viability using the MTT assay (Fig. 5I, 5K), the percentage of THþ cells (Fig. 5J, 5L), and the PI staining (Fig. 5E,  5H, data not shown). Thus, sustained Nurr1 protein stability in Nurr1 Akt -transduced precursor cells preserves DA phenotypes and improves cell survival. These events contribute in turn to the maintenance of THþ cells during differentiation.

DISCUSSION
In this report, we showed that Nurr1 proteins undergo UPSmediated degradation during precursor cell differentiation and that the process is regulated by the opposing actions of ERK and Akt intracellular signals. We further demonstrated that direct Akt phosphorylation of Nurr1 protein controls the ubiquitinylation of this protein and that stability of Nurr1 proteins can be sustained by abolishing the Akt phosphorylation site of Nurr1. These findings are novel and are supported by previous studies exemplified as follows. It is thought that protein phosphorylation plays a critical role in the initiation of protein ubiquitylation [23][24][25]. ERK and Akt have recently been specified as the critical protein phosphorylation pathways controlling several protein degradations [26][27][28]. In support of our findings, opposing cellular responses mediated by Raf and Akt signals have also been shown to control UPSdegradation of p53 proteins [29]. Furthermore, similar to our results, direct Akt-mediated phosphorylation of androgen receptor, a member of the same nuclear-hormone receptor family of proteins that Nurr1 belongs to, activates ubiquitinylation of this protein [28]. Current knowledge supports the idea that the nucleus is the primary target for degradation of nuclear receptors [30]. Consistent with this theory, treatment with the nuclear export inhibitor, leptomycin B, did not affect Nurr1 degradation, indicating localization of Nurr1-specific UPS-mediated degradation to the nucleus. Further molecular understanding of Nurr1 degradation requires studies to define Nurr1-specific E3 ligases and deubiquitinylation enzymes.
The most critical problem with current methods of cell transplantation for PD treatment is the low viability of donor cells. Only a minor portion of the DA neurons survive transplantation [8,31]. Furthermore, recent studies have demonstrated that the diseased PD environment transmits a toxic sig-nal to the grafted neurons [32], indicating that improvement of host environment will also be required to improve the survival of grafted neurons. Fundamental corrections to the environment created by the diseased state, however, do not seem to be easily achievable. An alternative would be to provide transplanted cells with resistance to the toxic environment. In addition to the key issue of survival, unstable phenotypes of transplanted DA neurons [9] also contribute to a low yield of DA neurons after transplantation. The present study shows an example of genetic manipulation yielding donor DA cells with improved cell survival and better-maintained phenotypes in vitro and in vivo after transplantation. The single mutation of the Akt-phosphorylation site of Nurr1 in this study resulted in a dramatic effect on Nurr1 protein stability, and in turn maintenance of THþ cells. It is manifest that the continued Nurr1 expression in the Nurr1 Akt -transduced cells accounts for the observed effect of DA phenotype maintenance, as Nurr1 is a transcription factor that activates expression of genes involved in the DA phenotype [1,2,6]. The majority of the THþ cells derived from Nurr1 Akt -transduced precursors expressed Nurr1 for a prolonged period during differentiation. Consistent with a previous study [33], the Nurr1-expressing DA cells maintained their DA phenotype better and survived longer. Cell transplantation in clinical level, however, requires long period of donor cell survival. While enhanced Nurr1þ/ THþ cell yield was clear and striking by Nurr1 Akt -transduced cell transplantation at 2 weeks after transplantation (Fig. 5M-V), the beneficial effect of the Nurr1 mutation was not continued for a longer period after transplantation: only few THþ cells (less than 200 THþ cells/graft) were detected in the brains grafted with Nurr1 Akt -cells 8 weeks after transplantation. We found that lack of the long-term effect is mainly due to unstable exogene (Nurr1) mRNA expression, but not due to loss of the Akt mutation effect in maintaining Nurr1 protein stability in the transplanted brains in vivo (data not shown). This is consistent to the previous studies demonstrating loss of exogene (GFP) expression in donor cell after neural transplantation [34,35]. We further found that promoters of expression vectors (LTR, CMV, EF1a) commonly contain cAMP-response element (CRE) and the promoter-driven expressions are highly dependent on activated CRE-binding protein (CREB) intracellular signal (data not shown). Thus inactivation of CREB signal in donor cells long after transplantation, with unidentified mechanism, is likely to be responsible for loss of Nurr1 mRNA expression and in turn loss of Nurr1-induced TH phenotype expression in those cells. We are now investigating to develop methods to achieve stable exogene expression in transplanted donor cells.
DA neurons yielded by the mutant Nurr1 were more resistant to toxic stimuli. These findings indicate an advance in cell therapeutic approaches for PD through generation of donor DA cells that are able to sustain their phenotype after transplantation and that survive by overcoming the pathologic host environment of PD. Nurr1 is a susceptible factor in PD and decreased levels of Nurr1 have been found in PD patients [7], suggesting that Nurr1 is a target molecule for the treatment and prevention of PD. We anticipate that the information we presented in this study regarding Nurr1 protein degradation and stability can be used to develop medical treatments to prevent the progression of PD.