Dopamine Neuron Agenesis in Nurr1-Deficient Mice

See allHide authors and affiliations

Science  11 Apr 1997:
Vol. 276, Issue 5310, pp. 248-250
DOI: 10.1126/science.276.5310.248


Dopamine neurons of the substantia nigra and ventral tegmental area regulate movement and affective behavior and degenerate in Parkinson’s disease. The orphan nuclear receptor Nurr1 was shown to be expressed in developing dopamine neurons before the appearance of known phenotypic markers for these cells. Mice lacking Nurr1 failed to generate midbrain dopaminergic neurons, were hypoactive, and died soon after birth. Nurr1 expression continued into adulthood, and brains of heterozygous animals, otherwise apparently healthy, contained reduced dopamine levels. These results suggest that putative Nurr1 ligands may be useful for treatment of Parkinson’s disease and other disorders of midbrain dopamine circuitry.

Dopamine (DA) midbrain neurons function in the control of movement and their loss is the cause of Parkinson’s disease. Mechanisms underlying dopaminergic development are largely unknown. Whereas general induction of ventral cell fate requires the secreted factor sonic hedgehog (1), additional factors must contribute to differentiation of specific cell types such as DA neurons. Glial cell line–derived neurotrophic factor (GDNF) and brain-derived neurotrophic factor have been demonstrated to promote survival of DA cells (2), but the absence of these factors does not affect prenatal DA neuron development (3). In contrast, Nurr1, an orphan member of the steroid/thyroid hormone receptor family (4), which is expressed in DA cells (5), was found to be absolutely required for development of midbrain DA neurons.

Nurr1 genomic DNA was used to construct a targeting vector (Fig.1A) in which a neomycin phosphotransferase gene replaced the NH2-terminal transactivation and DNA-binding domains of Nurr1. From cross-breeding of heterozygous Nurr1 mutant mice, we obtained homozygous animals (Nurr1 / ; Fig. 1, B and C) born at the expected frequency of 25% without apparent abnormalities. Homozygous mice could be distinguished after birth by hypoactivity and the lack of milk in their stomachs and died within the first 2 days after birth. Histology of mutants did not reveal any gross morphological abnormalities of brain, adrenals, thymus, lungs, heart, liver, kidneys, or intestine. Peripheral nerves, spinal cord, and dorsal root ganglia appeared normal in Nurr1–/– mice as judged by laminin, protein gene product 9.5, neurofilament, and calcitonin gene-related peptide immunohistochemistry (6).

Figure 1

Targeting the mouse Nurr1 locus (21). (A) Schematic representation of the Nurr1 wild-type locus, targeting vector, and targeted locus. Untranslated and translated exons are indicated by open and filled boxes, respectively. In the targeted locus, a thymidine kinase promoter–driven neomycin resistance gene (neo) replaces exons 2, 3, and part of exon 4. The expected fragment lengths in diagnostic Southern blot experiments are indicated by arrows for the wild-type (above) or targeted (below) locus, respectively. Probes located 5′ or 3′ (black bars) of the neo cassette were used. Restriction endonucleases are Sph I (Sp), Eco RI (E), Bam HI (B), Sal I (S), and Not I (N). For convenience, only those Sph I sites relevant for results in Southern blots are indicated. (B) Southern blot analysis was performed with the 5′ probe of Eco RI–cut DNA derived from mice originating from one of the injected ES cell lines. Lanes are homozygote (−/−), wild-type (+/+), and heterozygote (+/−) DNA. (C) Shown is a Southern blot of Sph I–cut DNA using the 3′ probe.

Nurr1 mRNA is expressed in substantia nigra pars compacta (SNC) and the ventral tegmental area (VTA) (5). Expression appears at embryonic day 10.5 (E10.5) in the ventral aspect of the mesencephalic flexure where DA neurons later develop but before known phenotypic markers for DA neurons such as tyrosine hydroxylase (TH) can be detected (Fig. 2). TH immunoreactivity (IR) (6) was absent in Nurr1–/– mice in the ventral midbrain where TH-positive cell groups (defined as A8 through A10) (7) are normally present (Fig. 2, E and H). TH-IR was present in newborn Nurr1–/– mice in other areas known to express TH (Fig. 2, F, G, I, and J). TH-IR, normally appearing in the developing midbrain at approximately E11.5 in mouse (8), was absent throughout all prenatal stages (from E11.5 to newborn; Fig. 2, K through N). Therefore, the data suggest that Nurr1 is critical for midbrain dopaminergic cell differentiation.

Figure 2

In situ hybridization (22) of Nurr1 and TH mRNA (A throughD), and TH immunohistochemistry (E throughN) (6). Nurr1 mRNA was found in the ventral aspect of E10.5 mouse embryo midbrain (A) (indicated by arrow), at which time no TH mRNA could be detected within this region (B). At E11.5, when Nurr1 mRNA was strongly expressed in the mesencephalic flexure (C) (14), TH mRNA was also detected (D). (E through N) TH-IR in wild-type (+/+) and Nurr1 (−/−) mutants are compared. Coronal (E through J) and sagittal (K through N) sections of different stages are shown. In SNC (A9), TH-IR is strong in newborn wild-type mice (E) but completely absent in Nurr1 mutants (H). Distribution of TH-IR in the dopaminergic cell groups A11/A13 (7) of diencephalon (F and I) and the pontine locus coeruleus [A6 (G and J)], showing nuclei where no differences in TH-IR can be seen between wild-type and Nurr1 / newborn brains. TH-IR was completely absent in ventral mesencephalon of Nurr1 mutants both at E11.5 (K and L) and E15.5 (M and N). At E15.5, TH-positive fibers, detected by peroxidase anti-peroxidase IR (6), extended rostrally from the midbrain in wild-type mice (M). A diencephalic TH-positive cell group was detected in the Nurr1 / brain (N). Note the absence of TH in ventral mesencephalon (N). Arrows indicate the mesencephalic flexures (M and N). Scale bars, 300 μm.

Although Nurr1 may directly activate transcription of the TH gene, this seems unlikely as most TH-positive cell groups do not express Nurr1 mRNA (5), and these other cell groups were TH immunoreactive in Nurr1−/− mice (Fig. 2). Instead, Nurr1 appears to determine the specific dopaminergic cell fate. Analysis of cresyl violet–stained sections (6) revealed a disorganized distribution of cells in the ventral midbrain of Nurr1–/–mice, suggesting that TH-expressing neurons fail to differentiate. In situ hybridization showed that Nurr1 mRNA was expressed in SNC and VTA in wild-type animals, whereas no hybridization signal was detected in Nurr1–/– mice (Fig. 3, A and B). Neither TH-IR nor TH mRNA (Fig. 3, C and D) was found in mutant midbrains. Additional DA neuron markers, including aldehyde dehydrogenase 2 (ADH2) (9), mRNA for the GDNF signal transducing receptor c-Ret (Fig. 3, E through H) (10), and the DA D2receptor were also absent in SNC and VTA in Nurr1–/–mice.

Figure 3

Markers for DA neurons in the newborn midbrain. In situ hybridization (22) (A throughF) and immunohistochemistry (6) (Gand H) are shown for coronal sections of wild-type (A, C, E, and G) and Nurr1 / (B, D, F, and H) mice. Nurr1 was expressed in wild-type (A) (5) but not in Nurr1 / ventral midbrain (B). TH mRNA was detected in wild-type (C) but not in Nurr1 / (D) ventral-midbrain. c-Ret mRNA was detected in wild-type DA neurons (E) but was virtually absent in Nurr1 / midbrain (F). Immunohistochemistry of DA neuron marker ADH2 showed labeling in wild-type SNC (G) but not in Nurr1 / midbrain (H). Scale bars, 300 μm.

The cause of death in Nurr1–/– mice is not clear although it appears to involve inability to suckle. Mutant mice display movement disturbances including severe difficulties in turning when placed on their backs and abnormal flexion-extension movements of limbs. The cause of death does not seem to involve competition with normal littermates, because litter restriction did not influence survival of the mutant pups. In mice where the TH gene was disrupted in dopaminergic cells, homozygotes survived 2 weeks after birth (11). Such delayed death is consistent with the postnatal course of development of striatal DA innervation (12) and suggests that perinatal lack of midbrain DA is not immediately lethal. However, absence of DA neurons, as is the case in the Nurr1–/– mice, rather than lack of transmitter, is accompanied by additional deficits in neuropeptides and neuronal circuitry (13). Alternatively, Nurr1 could have vital functions at other sites of expression (5, 14).

Striatum develops many organotypic features in genetically engineered DA-deficient mice (11) after perinatal ablation of dopaminergic neurons (15) or when grafted into 6-hydroxy-DA–lesioned adult rat brains (16). Our data support these findings. We found that mRNAs encoding enkephalin, substance P (Fig. 4, A and B), and choline acetyltransferase were normally distributed in striatum of Nurr1 mutant brains, although substance P mRNA levels appeared somewhat reduced, in agreement with previous studies of mice deficient in the DA D1 receptor, as well as DA (11, 17).

Figure 4

(A and B) In situ hybridization of coronal sections at the striatal level of newborn brains shows normal distribution of enkephalin (A) and substance P (B) mRNA in Nurr1 / mice. Scale bar, 300 μm. (C) DA was absent in Nurr1 / mice and reduced in Nurr1+/ mice. Shown is HPLC analysis (18) of dissected striata (orange bars) and ventral midbrains (blue bars) of wild-type (+/+), heterozygote (+/−), and Nurr1 (−/−) mice. The decreases of DA are statistically significant (P = 0.01). (DA detected in nigra areas in two of five Nurr1 / brains in all probability corresponds to contamination from adjacent nonaffected DA neurons.)

High-pressure liquid chromatography (HPLC) was used to measure levels of DA and the DA metabolite DOPAC in pieces of striatum and ventral midbrain (18). DA was absent in Nurr1–/–animals (Fig. 4C). Newborn heterozygotes showed reduced levels of DA (Fig. 4C) and DOPAC, indicating that nigrostriatal DA levels are affected by Nurr1 gene dosage. Adult heterozygotes also manifested reduced striatal DA levels (P < 0.01) but no apparent histological or behavioral abnormalities. This indicates that Nurr1 helps maintain the differentiated DA neuron phenotype in mature animals and suggests that DA levels could be regulated by Nurr1 ligands.

Understanding DA neuron development will involve elucidating the relationship of Nurr1 to sonic hedgehog signaling and to other regulatory components involved in midbrain patterning and cell fate specification (1, 19). Nurr1 can promote signaling through heterodimerization with the 9-cis retinoic acid receptor, RXR (20). The enzyme ADH2, used here as a DA neuron marker, may have importance for retinoid synthesis (9). Thus, retinoids may activate RXR-Nurr1 heterodimers in developing DA neurons. Further insight into how retinoids, as well as a putative ligand for Nurr1, may influence the development and survival of DA neurons may provide opportunities for pharmacological intervention to manage the function of dopaminergic neurons.

  • * To whom correspondence should be addressed. E-mail: thomas.perlmann{at}


View Abstract

Navigate This Article