Nurr1, an orphan nuclear receptor with essential functions in developing dopamine cells
- First Online:
- Cite this article as:
- Perlmann, T. & Wallén-Mackenzie, Å. Cell Tissue Res (2004) 318: 45. doi:10.1007/s00441-004-0974-7
- 396 Views
Nurr1 is a transcription factor that is expressed in the embryonic ventral midbrain and is critical for the development of dopamine (DA) neurons. It belongs to the conserved family of nuclear receptors but lacks an identified ligand and is therefore referred to as an orphan receptor. Recent structural studies have indicated that Nurr1 belongs to a class of ligand-independent nuclear receptors that are unable to bind cognate ligands. However, Nurr1 can promote signaling via its heterodimerization partner, the retinoid X receptor (RXR). RXR ligands can promote the survival of DA neurons via a process that depends on Nurr1–RXR heterodimers. In developing DA cells, Nurr1 is required for the expression of several genes important for DA synthesis and function. However, Nurr1 is probably also important for the maintenance of adult DA neurons and plays additional less-well-elucidated roles in other regions of the central nervous system and in peripheral tissues.
KeywordsNurr1NR4A2Nuclear receptorOrphan receptorTranscription factorDopamine cell developmentParkinson’s disease
Midbrain dopamine (DA) neurons are formed in the ventral tegmentum during development. DA cells have been extensively studied, mostly because of their critical involvement in disorders such as Parkinson’s disease and schizophrenia. The characterization of the DA neurotransmitter system was greatly facilitated by discoveries of the Falck-Hillarp histofluorescence methodology allowing the visualization of catecholaminergic neurons and their inneravation targets (Falck et al. 1962). These studies established that the majority of DA cells are localized in the ventral midbrain (VMB), which contains approximately 75% of the total number of brain DA cells. Within the VMB, the DA neurons are located in the lateral groups of the retrorubral field and the substantia nigra compacta and in the medially located ventral tegmental area. The neurons of the substantia nigra compacta project to the dorsolateral striatum and the caudate putamen, forming the nigrostriatal pathway involved in voluntary movements. Neurons of the ventral tegmental area project to the ventromedial striatum and subcortical and cortical areas, forming the mesolimbocortical system, which is involved in emotional behavior, motivation, and reward. Finally, among other functions, retrorubral field neurons are involved in connections between the substantia nigra and the ventral tegmental area (Ungerstedt 1971).
Midbrain DA cells are generated in the vicinity of the mid/hindbrain boundary (isthmus) and the floor plate and are a specialized cell type in the ventral midline of almost the entire neural tube. Several critical events in the rodent have been defined that influence the formation of the midbrain and its distinct cell types. Early midbrain patterning events, including the secretion of fibroblast growth factor 8 and Sonic hedgehog (described elsewhere in this issue), give rise to proliferating DA progenitor cells (Hynes et al. 1995; Hynes and Rosenthal 1999). The first postmitotic differentiating DA cells appear at approximately embryonic day (E) 10–10.5 in the mouse. The cessation of proliferation is followed by the upregulation of general neuronal and specific dopaminergic markers such as tyrosine hydroxylase (TH). Newly formed DA neurons migrate into medial and lateral positions to form the various DA-neuron-containing areas, and axon outgrowth toward rostral innervation targets is initiated.
Several transcription factors are expressed in the early mid/hindbrain region, including the homeobox-containing engrailed (En1 and En2), Lmx1b, Otx2, and Gbx2 (Wurst and Bally-Cuif 2001). However, no studies of transcription factors specifically expressed in proliferating DA cell progenitors have been reported. In contrast, several transcription factors, including Nurr1, Lmx1b, En1/2, and Pitx3, are expressed in postmitotic differentiating DA neurons and have been shown to influence their differentiation (Simon et al. 1998; Smidt et al. 2000, 1997; Wallén et al. 1999). In this review, we focus on our current understanding of the functional properties and the in vivo role of Nurr1, a transcription factor that is involved in the generation of midbrain DA neurons and that is, so far, the most well characterized.
Nurr1, a member of the nuclear receptor family of transcription factors
Nurr1 (also referred to as NR4A2, NOT, RNR-1, HZF-3) is structurally related to members of the nuclear receptor family of transcription factors. Nurr1 forms a highly conserved subfamily of nuclear orphan receptors together with Nur77 (NR4A1) and Nor1 (NR4A3). Nuclear receptors are ligand-regulated transcription factors that bind steroid hormones, thyroid hormone, retinoids, and other small lipophilic signaling molecules (Mangelsdorf et al. 1995). These proteins consist of several functionally distinct domains including evolutionary conserved DNA and ligand-binding domains, situated in the central and C-terminal regions of the proteins, respectively. Most nuclear receptors are constitutively localized in the cell nucleus, even in the absence of ligand. Upon ligand binding, a structural transition within the ligand-binding domain allows the dissociation of so-called co-repressors and the recruitment of proteins referred to as co-activators. These events trigger transcriptional activation of regulated target genes. Nuclear receptors bind sequence-specific promoter elements either as monomers or homodimers; however, many nuclear receptors require heterodimerization with the common dimerization partner, retinoid X receptor (RXR).
Nuclear receptors are interesting proteins for several reasons. First, they are excellent tools for understanding the way in which genes are regulated, since small molecule ligands can be used to switch these transcription factors between active and inactive states. Second, classical nuclear receptor signaling pathways, e.g., steroid hormone and retinoid receptors, influence many biological pathways important in development and adult physiology. Accordingly, their significance in disease is critical. Third, nuclear receptors include a large number of related but less well characterized orphan receptors lacking identified ligands. The existence of these proteins is intriguing and suggests that additional unexplored nuclear receptor-mediated signaling pathways remain to be characterized (Chawla et al. 2001; Kliewer et al. 1999).
Nurr1 in developing DA cells
Nurr1 is mainly expressed within the central nervous system (CNS), both during development and in the adult (Law et al. 1992; Saucedo-Cardenas and Conneely 1996; Zetterström et al. 1996a, 1996b). Within the CNS, Nurr1 is expressed in a number of areas including the cortex, hippocampus, brain stem, and spinal cord. In the mouse VMB, strong Nurr1 expression is detected as early as E10.5 in newly born postmitotic DA neurons (Wallén et al. 1999; Zetterström et al. 1997); this expression continues throughout adulthood. In contrast, other DA neurons located in, for example, the paraventricular/periventricular hypothalamic nucleus and noradrenergic TH-positive neurons do not express Nurr1 (Backman et al. 1999). Although not expressed during the development of olfactory DA neurons, Nurr1 appears to be expressed in these cells in the adult brain (Backman et al. 1999).
Reports describing data from three independent strains of Nurr1 knockout mice have shown that it is critical for the generation of DA neurons (Castillo et al. 1998; Le et al. 1999a, 1999b; Saucedo-Cardenas et al. 1998; Zetterström et al. 1997). In these animals, essentially no analyzed DA neuron markers, e.g., TH, GFRα1, and the aromatic amino acid decarboxylase (AADC), can be detected at birth. In contrast, and as expected from the expression patterns of Nurr1, all other catecholaminergic cell groups are intact in Nurr1 knockout mice. Although Nurr1 can induce cell cycle arrest when expressed in certain cell lines (Castro et al. 2001), analyses of cell proliferation and of Nurr1 knockout embryos indicate that Nurr1 does not regulate cell cycle exit in vivo (Wallén et al. 1999).
Only a few early DA neuron markers, including TH and the receptor tyrosine kinase subunit Ret, fail to be induced in Nurr1 knockout embryos. In contrast, although essentially all DA neuron markers are absent in knockout mice at birth, several genes, e.g., those encoding the transcription factors Pitx3, En1, En2, GFRα1, and Lmx1b, are normally induced even in the absence Nurr1 at earlier stages of development (see Saucedo-Cardenas et al. 1998; Smidt et al. 1997; Thuret et al. 2004; Wallén et al. 1999). These observations indicate that Nurr1 is primarily important for the continued development and maintenance of early developing DA cells and that it is not a prerequisite for the induction of all dopaminergic genes. Analyses have also indicated that Nurr1 is important for cell migration, as remaining markers are localized at a medial position in developing Nurr1 mutant midbrains (Wallén et al. 1999). Moreover, Nurr1-deficient neurons seem unable to innervate their forebrain target areas normally, as has been demonstrated by retrograde fluorogold tracing in newborn mutant pups (Wallén et al. 1999). Nevertheless, one research group has reported preserved innervation and cellularity in newborn Nurr1 knockout mice (Castillo et al. 1998; Witta et al. 2000). One explanation for this discrepancy might be their use of the DiI tracing method, which does not distinguish between ascending and descending pathways. Thus, adjacent nondopaminergic pathways might have been detected. Alternatively, differences in the gene targeting strategy and/or strain variability may also have contributed to the discrepancy. Finally, increased cell death is detected at late gestation in Nurr1 knockout mice (Saucedo-Cardenas et al. 1998; Wallén et al. 1999). At this developmental time-point, essentially no dopaminergic markers are expressed, cells fail to migrate to their final destinations, and target innervation is undetectable. Thus, increased cell death is most likely a secondary consequence of severe cellular deficiencies, but as will be described below, other data indicate that Nurr1 might be important for the long-term survival of DA neurons.
Limited information is currently available concerning the Nurr1 target genes to explain the severe cellular deficiency resulting from Nurr1 deficiency. As described elsewhere in this issue, knockout experiments have shown that En1/2 and Lmx1b are required for the normal development of DA neurons (Simon et al. 1998; Smidt et al. 2000). However, as mentioned above, these genes are induced normally in Nurr1 knockout VMB. Conversely, and in contrast to observations in Nurr1 knockouts, TH remains expressed at early time-points in both En1/2 and Lmx1b knockouts. Thus, Nurr1 seems to function in a developmental pathway that, at least initially, is independent of En1/2 and Lmx1b. Importantly, these observations also provide evidence indicating that Nurr1 does not affect DA cell development via the regulation of these homeobox transcription factors.
One of the few genes that has been shown to be absent at early developmental time-points in Nurr1 knockout VMB is the tyrosine kinase Ret (Wallén et al. 2001). Ret is critical as a signaling subunit of receptors responding to trophic factors such as the glial cell line-derived neurotrophic factor (GDNF). However, gene ablation of Ret, GDNF, or related neurotrophic factors has shown that they are not crucial for the generation of DA neurons during embryonic development (see references in Riddle and Pollock 2003). Thus, whereas Nurr1 regulation of Ret gene expression is likely to be of significance in postnatal brains (Granholm et al. 2000; Oo et al. 2003), Ret is unlikely to mediate essential Nurr1 functions during embryogenesis.
Nurr1 has recently been shown to be important for the expression of p57Kip2 in developing DA cells (Joseph et al. 2003). p57Kip2 belongs to the Cip/Kip family of cyclin-dependent kinase inhibitors. This protein is expressed in both proliferating and postmitotic DA cells, and its induction in postmitotic cells is strictly dependent on Nurr1. Moreover, some abnormalities observed in Nurr1 knockout VMBs are also seen in p57Kip2 knockout embryos, indicating that p57Kip2 contributes to DA cell differentiation in a pathway downstream of Nurr1. Interestingly, p57Kip2 can also interact physically with Nurr1 suggesting that it might serve as a Nurr1 co-factor in DA cell development. Thus, these findings suggest a rather unorthodox function for p57Kip2 in cellular differentiation.
Nurr1 and DA neuron engineering from stem cells
As Parkinson’s disease results from the loss of DA neurons, the prospect of utilizing cell replacement therapies has attracted substantial interest. Clinical trials involving embryonic tissue transplantation have provided proof-of-concept for this strategy, although one recent human trial has been disappointing (Dunnett et al. 2001). Thus, the elucidation of regulatory cascades influencing the development of DA neurons is of great interest in attempts to engineer DA cells from stem cells. The identification of transcription factors, such as Nurr1, that influence DA neuron development opens up the possibility of genetically manipulating stem cells to induce the dopaminergic fate. A preliminary study demonstrating the utility of this approach involved an immortalized neural stem cell line derived from mouse embryonic cerebellum. Nurr1 expression in these cells was shown to result in the robust differentiation and formation of TH-expressing neurons when these cells were co-cultured with VMB type I astrocytes (Wagner et al. 1999). The nature of the astrocyte factors remains elusive. Overexpression of Nurr1 in another immortalized neural stem cell line has provided additional evidence that this approach can result in the expression of dopaminergic markers in cultured stem cells (Kim et al. 2003b).
Several studies have indicated that Nurr1 overexpression can positively influence the ability of embryonic stem (ES) cells to differentiate into DA neurons (Chung et al. 2002; Kim et al. 2002; Sonntag et al. 2004). In one study, Nurr1-overexpressing differentiated ES cells were shown to restore dopaminergic functions in rodents exposed to 6-hydroxy-DA treatment (Kim et al. 2002). Although an ideal scenario would be to avoid the need of gene transfer and instead to depend exclusively on extrinsic signals to manipulate stem cells in culture, these studies illustrate the value of identifying key developmental mechanisms underlying DA neuron differentiation.
Does Nurr1 have a function in mature DA neurons?
As mentioned above, Nurr1 continues to be expressed in DA neurons in the adult brain. Since Nurr1 knockout mice die at birth, the consequences of Nurr1 deficiency have not yet been elucidated in postnatal brains. Studies of Nurr1 knockout animals have indicated, however, that Nurr1 may be important for the regulation of DA neurotransmission. Accordingly, the DA synthesizing enzymes TH and AADC are both regulated by Nurr1 (Castillo et al. 1998; Hermanson et al. 2003; Saucedo-Cardenas et al. 1998; Smits et al. 2003; Zetterström et al. 1997). Functional binding sites for Nurr1 have been identified in the TH promoter (Iwawaki et al. 2000; Kim et al. 2003a; Sakurada et al. 1999). Interestingly, an age-related decrease of Nurr1 immunoreactivity correlates with decreasing TH expression in the normal process of aging in the human substantia nigra, suggesting a possible regulatory relationship between Nurr1 and TH in the human brain (Chu et al. 2002). Nurr1 gene expression is also significantly upregulated in mice lacking functional copies of the DA D2 receptor gene indicating that Nurr1 might mediate autoreceptor functions in mesencephalic DA neurons (Tseng et al. 2000). Other genes that are regulated by Nurr1 include the dopamine transporter (DAT) and VMAT2 (Hermanson et al. 2003; Sacchetti et al. 1999, 2001; Smits et al. 2003). The regulation of these genes is probably of significance in the adult brain in which Nurr1 might influence the function of the dopaminergic system. Evidence for this hypothesis has been derived from studies of Nurr1 heterozygous mice. In these animals, DA content is reduced at birth, although the mice have the normal number of midbrain DA neurons (Eells et al. 2002; Le et al. 1999b; Zetterström et al. 1997). Moreover, Nurr1 heterozygous animals display abnormalities that might be associated with the dopaminergic system. For example, these animals show an altered response in certain models of addiction and have altered locomotor behaviors in response to novel environments, stress, and metamphetamine treatment (Backman et al. 2003; Eells et al. 2002; Werme et al. 2003). Mutations in the human Nurr1 gene have furthermore been associated with schizophrenia and manic depression (Buervenich et al. 2000). Although these phenotypes are complex and influenced by several neurotransmitter systems, some contribution from the abnormal function of DA neurons seems likely.
Some reports indicate that Nurr1 contributes to the maintenance of mature DA neurons. Accordingly, DA neurons of Nurr1 heterozygous mice are more vulnerable to the neurotoxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (Le et al. 1999a). Moreover, mutations in the human Nurr1 gene have been identified in patients with familial Parkinson’s disease (Le et al. 2003). These mutations are situated in a region encoding the 5′ untranslated region of human Nurr1 mRNA and have been shown to decrease the expression of Nurr1. Although these findings have not been confirmed in independent studies, they indicate that Nurr1 might contribute to the survival of mature DA cells in vivo.
A recent study has provided some clues regarding the manner in which Nurr1 might promote neuronal survival (Wallén-Mackenzie et al. 2003). As mentioned above, Nurr1 can form heterodimers with RXR, another nuclear receptor that serves a central role in the nuclear receptor family. RXR is a promiscuous heterodimer partner of many nuclear receptors including retinoid and thyroid hormone receptors. In this capacity, RXR functions as an obligatory and essential partner required for efficient DNA binding of these receptor heterodimers. RXR can also bind its own ligands and activate transcription in a ligand-dependent manner. Natural RXR ligands include the retinoid metabolite 9-cis retinoic acid and the omega-3 polyunsaturated fatty acid docosahexaenoic acid (Heyman et al. 1992; Mata de Urquiza et al. 2000). Nurr1-RXR heterodimers are activated by endogenous RXR ligands in the embryonic CNS (Wallén-Mackenzie et al. 2003). In addition, these heterodimers promote the survival of neurons, e.g., DA neurons, via a mechanism that strictly depends on Nurr1, RXR, and ligands that can activate RXR. These results may help to explain the mechanism whereby Nurr1 promotes the survival of DA neurons in the adult mouse and human brain.
Nurr1 may have a more general role in neuronal protection. Such a function might explain the dramatic Nurr1 upregulation seen after hypoxic stress and other stressful insults to the brain (Crispino et al. 1998; Honkaniemi et al. 1995; Honkaniemi and Sharp 1996, 1999; Honkaniemi et al. 1997; Ojeda et al. 2003; Pena de Ortiz and Jamieson 1996; Xing et al. 1997). Conditional knockout analyses, allowing characterization of adult mice with Nurr1 deficiency, should ultimately help to resolve this important question.
Non-dopaminergic functions of Nurr1
As Nurr1 expression is not solely confined to VMB DA neurons, it may have functions unrelated to these cells. Indeed, the severity of the knockout phenotype supports this conclusion. Nurr1 knockout mice die within 24 h after birth, a time-point at which the dopaminergic system probably does not as yet play essential roles in wildtype animals. Animals lacking DA, as a consequence of the selective ablation of the TH gene in DA neurons, survive several weeks after birth and can be rescued to adulthood by L-Dopa replacement (Zhou and Palmiter 1995). Thus, early on in the analyses of Nurr1 mutant mice, it became evident that these animals die for reasons that are unrelated to the lack of DA. Recent results showing that newborn Nurr1 knockout mice have severe respiratory abnormalities and lack a normal response to hypoxia might explain their early demise (Nsegbe et al. 2004). There is as yet no clear understanding of the cellular deficiency related to this phenotype. However, Nurr1 is expressed in several areas associated with respiratory control, including the nucleus of the solitary tract, the nucleus ambiguus, and the dorsal motor nucleus X (DMN-X; Nsegbe et al. 2004). Although the morphology and cellularity of the brain stem is grossly normal in Nurr1 knockout mice at birth, the expression of Ret is diminished in the DMN-X, and associated nerve bundles exiting the brain stem appear slightly disorganized (Nsegbe et al. 2004; Wallén et al. 2001). A more thorough phenotypic characterization will be necessary to define the cellular abnormalities contributing to respiratory failure, since the expression of Nurr1 in the DMN-X alone cannot explain the severity of the phenotype. Moreover, whether the phenotype reflects a developmental abnormality or an acute role of Nurr1 in respiratory function remains unclear.
An interesting property of Nurr1, and its closely related orphan nuclear receptors Nur77 and Nor1, is that they are all encoded by immediate early genes and are rapidly induced by various stimuli, e.g., growth factors. Thus, in addition to its developmental functions, Nurr1 seems to participate in adaptive processes, e.g., in response to ischemia and other types of neuronal stress (Crispino et al. 1998; Honkaniemi et al. 1995, 1997; Honkaniemi and Sharp 1996, 1999; Ojeda et al. 2003; Pena de Ortiz and Jamieson 1996; Xing et al. 1997).
Several studies have indicated that Nurr1 is functionally important outside of the nervous system. For example, Nurr1 is upregulated in regenerating liver and in inflammatory tissues (Arkenbout et al. 2002, 2003; Murphy et al. 2001; Scearce et al. 1993). Nurr1 is robustly upregulated in endothelial and smooth muscle cells in inflammatory rodent models of atherogenesis and in atherosclerotic human tissue (Arkenbout et al. 2002, 2003). Moreover, Nurr1, Nur77, and Nor1 are strongly induced by vascular endothelial growth factor in human endothelial cells in vitro (Liu et al. 2003). Interestingly, Nurr1 has been shown to regulate peripheral corticotropin-releasing hormone (CRH) in human inflammatory arthritis and is thus likely to modulate inflammatory responses (McEvoy et al. 2002). However, whether its function is pro-inflammatory remains unclear as Nur77 mediates the growth inhibition of smooth muscle cells in models of atherosclerosis (Arkenbout et al. 2002). One might speculate that Nurr1 participates in similar growth inhibitory control, not least as cell cycle arrest in G1 is induced in several cell lines overexpressing Nurr1 (Castro et al. 2001). In addition to playing roles in peripheral inflammatory responses, Nurr1, Nur77, and Nor1 may also contribute to the central control of inflammation via the pituitary. Accordingly, Nurr1, together with Nur77 and Nor1, can regulate the expression of the pro-opiomelanocortin gene in response to CRH in the anterior pituitary (Philips et al. 1997). Finally, Nurr1 has also been implicated in other functions, such as bone homeostasis via the regulation of the osteopontin gene in osteoblasts and in adrenal synthesis of aldosterones via the regulation of CYP11B2 gene expression in zona glomerulosa cells (Bassett et al. 2004; Lammi et al. 2004).
Although Nurr1 has been extensively studied, several important questions remain unanswered. First, relatively few target genes of Nurr1 have been identified, and the severity of the developmental DA neuron phenotype indicates that Nurr1 regulates more than DA synthesis, reuptake, and storage. The identification of additional targets may reveal novel critical players in DA neurogenesis. Second, an understanding of the function of Nurr1 in adult DA neurons and other cell types has been hampered by the early lethality of Nurr1 knockout mice. Thus, it will be important to establish conditional knockout models or to design alternative experimental approaches to address the role of Nurr1 in the adult. Third, a few studies have identified mutations in the human Nurr1 gene linked to Parkinson’s disease, schizophrenia, and manic depression (Buervenich et al. 2000; Le et al. 2003). On the other hand, other studies have failed to find such an association (see Hering et al. 2004; Tan et al. 2003). Thus, the significance of mutations in the human Nurr1 gene requires further studies. Finally, as mentioned above, Nurr1 is not regulated by cognate ligands as this transcription factor lacks a cavity for ligand binding. Therefore, crucial studies are required to define the way in which Nurr1 is regulated, to identify associated co-factors, and to understand the manner in which other signaling pathways utilize Nurr1 as a downstream nuclear target.
As indicated in this review, the identification of Nurr1 as a critical component in developing DA cells has triggered many studies with clinical implications. For example, improved methods for generating DA neurons from stem cells have been developed, potential novel links to disease-causing mutations have been revealed, and novel ideas concerning therapeutic intervention in Parkinson’s disease have been presented (Wallén-Mackenzie et al. 2003). Accordingly, the identification of additional critical components in the process of DA cell development will surely have implications that extend far beyond the understanding of the generation of a single cell type during embryogenesis.
The authors wish to thank Gerard Benoit for the preparation of Fig. 1. This work as supported by Swedish Foundation for Strategic Research and the Wallenberg Foundation.