Nuclear Receptor Nr4a1 Regulates Striatal Striosome Development and Dopamine D1 Receptor Signaling

Abstract The GABAergic medium-size spiny neuron (MSN), the striatal output neuron, may be classified into striosome, also known as patch, and matrix, based on neurochemical differences between the two compartments. At this time, little is known regarding the regulation of the development of the two compartments. Nr4a1, primarily described as a nuclear receptor/immediate early gene involved in the homeostasis of the dopaminergic system, is a striosomal marker. Using Nr4a1-overexpressing and Nr4a1-null mice, we sought to determine whether Nr4a1 is necessary and/or sufficient for striosome development. We report that in vivo and in vitro, Nr4a1 and Oprm1 mRNA levels are correlated. In the absence of Nr4a, there is a decrease in the percentage of striatal surface area occupied by striosomes. Alterations in Nr4a1 expression leads to dysregulation of multiple mRNAs of members of the dopamine receptor D1 signal transduction system. Constitutive overexpression of Nr4a1 decreases both the induction of phosphorylation of ERK after a single cocaine exposure and locomotor sensitization following chronic cocaine exposure. Nr4a1 overexpression increases MSN excitability but reduces MSN long-term potentiation. In the resting state, type 5 adenylyl cyclase (AC5) activity is normal, but the ability of AC5 to be activated by Drd1 G-protein-coupled receptor inputs is decreased. Our results support a role for Nr4a1 in determination of striatal patch/matrix structure and in regulation of dopaminoceptive neuronal function.

Nr4a1 is expressed in dopaminergic and dopaminoceptive neurons, including in the dorsal striatum, nucleus accumbens, olfactory tubercle, and prefrontal and cingulate cortex (Zetterström et al., 1996;Beaudry et al., 2000;Werme et al., 2000a); and at lower levels, in SN and ventral tegmental area (VTA). Dopamine receptor antagonists, psychostimulants, or DA denervation induce the expression of Nr4a1 in the midbrain dopaminergic SN and VTA and increase its expression in the striatum, where it acts as an immediate early gene (IEG; Beaudry et al., 2000;Werme et al., 2000a,b;St-Hilaire et al., 2003a;Ethier et al., 2004). Murine Nr4a1 genetic deletion is associated with an increase in tyrosine hydroxylase, dopamine turnover (Gilbert et al., 2006), baseline locomotor activity (Gilbert et al., 2006;Rouillard et al., 2018), and tardive dyskinesia (Ethier et al., 2004), but a reduction in levodopa induces dyskinesia [levodopa-induced dyskine-sia (LID)] in both rodent and nonhuman primate models of Parkinson's disease (St-Hilaire et al., 2003a,b;Mahmoudi et al., 2009Mahmoudi et al., , 2013. We began our studies in the Nr4a1-eGFP mouse (#GY139Gsat/Mmucd, GENSAT) to assay dopamine signal transduction in the striosomes, but found that the responses of these mice to dopamine agonists differed from those of nontransgenic littermates. We determined that the baseline Nr4a1 mRNA level in this line is twice the wild-type (WT) level. Herein, comparing the Nr4a1-eGFP mouse to the previously characterized Nr4a1-null mouse (Lee et al., 1995), we sought to determine the role of Nr4a1 in striosome development and regulation of the physiology of MSNs, and the dopamine signal transduction pathway. Our data indicate that Nr4a1 is necessary for, and promotes, the complete maturation of the striosome compartment, and its constitutive overexpression alters the D 1 R signaling pathway and response to cocaine.
Postnatal day 3 (P3) mice were rapidly killed by decapitation, and brains were removed, washed in ice-cold PBS, and postfixed for 24 h at 4°C in 4% PFA. The brains were then incubated in 30% sucrose/1ϫ PBS for 24 h at 4°C and cryopreserved in OCT embedding medium (catalog #4583, Tissue-Tek, Sakura). Serial coronal section (16 m) were cut on a Leica cryostat, collected on Superfrost Plus Microscope Slides (Thermo Fisher Scientific), and frozen at Ϫ20°C.

Striosome quantification
Striosome number and area, as a percentage of total striatal area, were measured in coronal sections from either matched adult or P3 mice immunolabeled with anti-MOR or anti DARPP-32, respectively. Using Fiji (version 2.0.0), the images were set at identical thresholds, and the regions of interest (ROIs) were outlined by manual tracing and managed with ROI manager function. The area and number of the ROIs selected were calculated using the Fiji "measure" function.

Cell counting
The induction of phosphorylated ERK (pERK) differs by region, so the numbers of pERK and c-fos cells were counted specifically in the dorsomedial area of the striatum at rostral ϩ1.18 mm, relative to bregma, and at caudal 0.86 mm. Cells were quantified in a fixed area using CaseViewer software by an observer blinded to treatment.
To determine the number and percentage of Drd1, Nr4a1-eGFP, and Drd2 cells in striosomes, we used coronal sections from double-hemizygous Nr4a1-eGFP/ Drd1-Tomato (Shuen et al., 2008) at bregma 0.86 mm. The striosomal area was outlined by manual tracing, and the cells were counted as GFP ϩ , Tomato ϩ , and GFP/ Tomato ϩ . The percentage of each population was calculated relative to the total number of cells indicated by 4[prime],6[prime]-diamidino-2-phenylindole dihydrochloride (DAPI) staining.

Primary neuronal cultures
Nr4a1-eGFP hemizygous and WT mice were timed mated, and the striatum was removed from E16.5 embryos by microdissection in cold Invitrogen Leibovitz's medium (L-15, Thermo Fisher Scientific). The tissue was incubated in Ca 2ϩ /Mg 2ϩ -free HBSS for 10 min at 37°C. The incubation mixture was replaced with 0.1 mg/ml papain in Hibernate E/Ca 2ϩ (BrainBits), incubated for 8 min, and rinsed in DMEM with 20% fetal bovine serum and twice in Leibovitz's medium (L-15). The tissue was then suspended in DMEM with 10% fetal calf serum, glucose (6 mg/ml), glutamine (1.4 mM), and penicillin/streptomycin (100 U/ml). Cells were triturated through a glass-bore pipette and plated onto either Lab Tek eight-well slides (125,000 cells/well) for immunocytochemistry or 24-well plates (1 ϫ 10 6 cells/well) for RT-PCR analysis, each previously coated with polymerized polyornithine (0.1 mg/ml in 15 mM borate buffer, pH 8.4) and air dried. One hour later, the media were replaced with Invitrogen Neurobasal/B27 medium (Thermo Fisher Scientific) with GLU-TAMAX and penicillin/streptomycin, and select wells were treated with brain-derived neurotrophic factor (BDNF; catalog #248-BD, R&D Systems) 25 ng/ml in 0.1% BSA/1ϫ PBS. Media change and BDNF treatment were performed every 2 d, and the cells were kept in culture until day in vitro 7 (DIV7).

Nr4a1-eGFP adenovirus transduction
Adenovirus (ADV)-CMV-Nr4a1-eGFP and ADV-CMV-eGFP were produced by SignaGen Laboratories. Nr4a1-eGFP ADV was produced using the human Nr4a1 cDNA sequence. Viral transduction was performed after cells had attached for 48 h with a multiplicity of infection (MOI) of 20. The virus was added in fresh medium, and the medium was changed 18 h later. Cells were harvested or fixed 96 h following the addition of virus. ADV transduction in primary neurons was performed in four independent sets of cultures.
For the ADV transduction experiments human iPSCderived NSCs were plated at 100,000 cells per well of a 6-well plate in 2 ml NPM supplemented with 10 ng/ml FGF2 (Peprotech) and 10 ng/ml Activin A (Peprotech). At 70% confluency, they were transduced with ADV-CMV-NR4A1-eGFP or ADV-CMV-eGFP at MOI of 20 suspended in 2 ml NPM without penicillin/streptomycin antibiotic. Nontransduced NSCs and NSCs transduced with ADV-eGFP were used as controls. A complete media change was performed 24 h post-transduction. Cells were harvested 14 d after transduction for gene expression and immunolabeling assays. ADV transduction in a human iPSC-derived NSC culture was performed twice, with three replicates each.

Quantitative real-time PCR
Snap-frozen samples for gene expression assays were homogenized in QIAzol Lysis Reagent (Qiagen). Total RNA was extracted with the miRNeasy Mini Kit (Qiagen) according to the manufacturer instructions. RNAs, 500 ng, were reversed transcribed using the High Capacity RNA-to-cDNA Kit (Applied Biosystems). Quantitative realtime PCR (qRT-PCR) was performed in a Step-One Plus System (Applied Biosystems) using All-in-One qPCR Mix (GeneCopoeia).
For qRT-PCR analysis of prepatterned Activin A-treated human NSCs, total RNA was isolated using the ISOLATE II RNA Mini Kit (Bioline). cDNA was prepared from 1 g of RNA in a total reaction volume of 20 l using the Sensi-FAST cDNA synthesis kit (Bioline). RT-PCR reactions were set up in a 384-well format using 2ϫ SensiFAST Probe No-ROX Kit (Bioline) and 1 l of cDNA per reaction in a total volume of 10 l. RT-PCR was performed on the Roche LightCycler 480 instrument.
Quantitative PCR consisted of 40 cycles, 15 s at 95°C and 30 s at 60°C each, followed by dissociation curve analysis. The ⌬Ct was calculated by subtracting the Ct for the endogenous control gene GAPDH from the Ct of the gene of interest. Mouse and human primer sequences are listed in Table 1 and in Table 2, respectively. Relative quantification was performed using the ⌬⌬Ct method (Livak and Schmittgen, 2001) and is expressed as a fold change relative to control by calculating 2 -⌬⌬Ct .

Western blotting
Snap-frozen striatum samples dissected from Nr4a1-eGFP, Nr4a1-null, and WT mice were lysed in Pierce RIPA buffer (Thermo Fisher Scientific) containing freshly added Table 1

Patch-clamp recordings
Patch-clamp whole-cell recordings were performed from the dorsal striatum of the 300-m-thick coronal brain slices obtained from 4-month-old mice. An upright microscope with differential interference contrast, fluorescence, and IR (Nikon Eclipse E600FN, Morrel Instrument) was used to visualize the neurons. The cells were voltage clamped at Ϫ70 mV, using patch pipettes (resistance, 3-6 M⍀) filled with an internal solution containing the following (in mM): 115 K-gluconate, 20 NaCl, 1.5 MgCl2, 10 phosphocreatine-Tris, 2 Mg-ATP, 0.5 Na-GTP, and 10 HEPES, pH 7.3 and 286 mmol/kg osmolarity. The hemislices were transferred to a recording chamber constantly perfused with oxygenated aCSF at a flow rate of ϳ4 ml/min for gravity. Experiments were performed at 28.0 Ϯ 0.1°C. Series resistance was monitored through the experiments, and cells with a Ͼ10% change in series resistance were excluded from analysis. Recordings were acquired with a Multiclamp 700B Amplifier (Molecular Devices) and Digidata 1440A digitizer (Molecular Devices). Current-clamp protocols were designed and performed using pClamp 10.3 Electrophysiology Data Acquisition and Analysis Software. Output signals were acquired at 5 kHz, filtered at 2.4 kHz, and stored on-line using pCLAMP 10.3 Electrophysiology Data Acquisition and Analysis Software (Molecular Devices). Single-cell long-term depression (LTD) was induced in the presence of 10 M 1(S)9(R)(Ϫ) bicuculline methiodide using a highfrequency stimulation (HFS) protocol (Calabresi et al., 1997) consisting of four 1-s-duration, 100 Hz trains deliv-ered at a frequency of one train every 10 s. Square-wave current pulses (60 s pulse width) were delivered with a concentric bipolar electrode placed above the corpus callosum through a stimulus isolator (Isoflex, AMPI). Output signals were acquired at 5 kHz, filtered at 2.4 kHz, and stored on-line using pCLAM 10.3Electrophysiology Data Acquisition and Analysis Software (Molecular Devices). In all cases, the experimenter was blind to genotype and or/treatment.

Field electrophysiology
Coronal brain slices containing the striatum were prepared from 9-week-old WT or Nr4a1-eGFP mice. Animals were anesthetized with isoflurane, and brains were rapidly removed from the skull and placed in ice-cold modified solution (aCSF) containing the following (in mM): 215 sucrose, 2.5 KCl, 1.6 NaH 2 PO 4 , 4 MgSO4, 1 CaCl2, 4 MgCl2, 20 glucose, and 26 NaHCO3, pH 7.4, and equilibrated with 95% O 2 and 5% CO 2 . Coronal brain slices (250 m thick) were prepared with a VT1000S Vibratome (Leica Microsystems), incubated at 31°C for 30 min, and then stored at room temperature for Ն1 h in normal aCSF containing the following (in mM): 120 NaCl, 3.3 KCl, 1.2 Na 2 HPO 4 , 26 NaHCO3, 1.3 MgSO4, 1.8 CaCl2, and 11 glucose, pH 7.4 equilibrated with 95% O 2 and 5% CO 2 , 280 -300 mmol/kg osmolarity. The hemi-slices were transferred to a recording chamber constantly perfused with oxygenated aCSF at a flow rate of ϳ4 ml/min using a peristaltic pump (Masterflex C/L); experiments were performed at 28.0 Ϯ 0.1°C. Recordings were acquired with a GeneClamp 500B Amplifier (Molecular Devices) and a Digidata 1440A digitizer (Molecular Devices). All signals were low-pass filtered at 2 kHz and digitized at 10 kHz. For extracellular field recordings (fEPSP recordings), a patch-type pipette was fabricated on a micropipette puller (Sutter Instrument), filled with normal aCSF (resistance, 3-6 M⍀), and placed in the dorsomedial striatum to measure long-term potentiation (LTP). A Concentric Bipolar Electrode stimulator (FHC) was placed immediately above the corpus callosum. Before and after HFS, the stimulus intensity was set to the level at which an evoked population spike was around half of the amplitude of the maximal obtainable response. Stimulus intensity was adjusted to a level evoking a maximal response during HFS. Stimulus intensity ranged from 0.3 to 1.2 mA (Partridge et al., 2000). Paired-pulse facilitation was measured by delivering two stimuli at 20, 50, and 100 ms interstimulus intervals before HFS. Each interstimulus interval was repeated three times, and the resulting potentials were averaged. LTP was induced using an HFS protocol, as follows: four 1 s duration, 100 Hz trains delivered at a frequency of one Table 2
cAMP measurement and adenylyl cyclase activity assay Flash-frozen striatal tissue punches from adult Nr4a1-eGFP and WT mice were homogenized in 250 l 0.1N HCl, centrifuged at 1000 ϫ g for 15 min, and supernatants diluted 20-fold for cAMP quantification using a cAMP enzyme immunoassay kit (cAMP Direct EIA) following the acetylated protocol (Enzo). The activity of adenylyl cyclase (AC) in striatal membrane preparations was determined as described previously (Xie et al., 2012). Briefly, striatal tissue punches were flash frozen in liquid nitrogen before homogenization in a buffer containing the following (in mM): HEPES, pH 8.0 (20); EDTA (1); NaCl (150); MgCl 2 (2), dithiothreitol (1); and 1ϫ complete protease inhibitor cocktail (Roche). After centrifugation at 1000 ϫ g for 15 min, the supernatant was subject to ultracentrifugation at 25,000 rpm for 35 min in a Beckman SW41 rotor over a 23%/43% sucrose gradient. The plasma membrane fraction was isolated from the sucrose interface, and the concentration was determined using the Pierce 660 nm Protein Assay Reagent (Thermo Fisher Scientific). Striatal membranes (2 g/reaction) were treated with vehicle (basal) or indicated stimulator for 10 min at 30°C in AC assay buffer (50 mM HEPES, pH 8.0; 0.6 mM EDTA; 100 g/ml BSA; 100 M 3-isobutyl-1-methylxanthine; 3 mM phosphoenolpyruvate potassium; 10 g/ml pyruvate kinase; 5 mM MgCl2; 10 M GTP; and 100 M ATP). Reactions were stopped by adding an equal volume of 0.2N HCl. The resulting cAMP in the sample was determined by cAMP Direct EIA kit.

Statistics
Statistical analysis was performed using GraphPad software version 6. One-way ANOVA followed by Sidak's multiple-comparisons test was performed for one-factor comparisons versus control. For 2 ϫ 2 comparisons, two-way ANOVA was used with repeated measures for the appropriate factor followed by Bonferroni's or Holm-Sidak's post hoc comparisons. For the RT-PCR and Western blot densitometry, we performed one-way ANOVA with genotype factor analysis, followed by post hoc tests with multiple comparisons versus control or WT mice. To analyze locomotor data and AC5 activity results, two-way ANOVA with treatment and genotype factors were used. For the electrophysiology experiments, statistical comparisons of pooled data were performed by unpaired t test or ANOVA (one-way or two-way). Results were considered significant at p Ͻ 0.05. Values are presented as the mean Ϯ SEM based on the number of samples that were used in each experiment.

Nr4a1 mRNA expression in Nr4a1-eGFP and Nr4a1null mice
Some GENSAT mice created using BAC technology [e.g. Drd2-eGFP (Kramer et al., 2011) and ChAT-Cre (Chen et al., 2018)] show expression from the transgene, increasing the total expression level of the BAC-encoded gene. Similarly, Nr4a1-eGFP adult hemizygous mice express twice as much Nr4a1 mRNA in the striatum relative to wild-type littermates ( Fig.1; F (2,14) ϭ 23.42, p Ͻ 0.0001; Sidak's multiple-comparisons test, t (14) ϭ 2.787, p ϭ 0.0145). The exact cause of Nr4a1 overexpression remains unknown. However, despite being engineered to prevent the increased levels of the reporter gene, several BAC mice show increased levels of the gene under study, or of other genes encoded on the BAC or impacted by insertion of the BAC (Kolisnyk et al., 2013;Ting and Feng, 2014). Confirming previous results (Lee et al., 1995), Nr4a1 mRNA expression is abolished in Nr4a1 homozygote-null [knock-out (KO)] mice ( Fig.1; Sidak's multiple-comparisons test, t (14) ϭ 4.166, p ϭ 0.0010). The level of mRNA of a second Nurr family member, Nr4a2, mRNA levels is unchanged in both genotypes ( Fig. 1; F (2,20) ϭ 0.1871, p ϭ 0.8308). We were unable to confirm the specificity of commercially available anti-Nr4a1 antibodies for immunocytochemistry or Western blotting.

Nr4a1 striatal distribution and effect of its overexpression or deletion on striosome compartment maturation in vivo
We analyzed several aspects of striatal phenotype in Nr4a1-eGFP and Nr4a1-null mice, including the compartmentalization of EGFP ϩ neurons, the surface area occupied by striosomes; and the mRNA levels of striosome, matrix, and markers of common MSNs. Ppp1r1b/ DARPP-32 served as an early marker of striosomes, and as a general MSN marker in adult striatum. Calb1 (i.e. calbindin 28 kDa) was used as a marker of the matrix compartment, and Oprm1/MOR, Rasgrp1/Caldag-GEFII, and, to a lesser extent, Foxp2, as markers of striosomes in the adult (Crittenden et al., 2009;Crittenden and Graybiel, 2011). We quantitated the distribution of Nr4a1-eGFP, dMSNs, and iMSNs in the two compartments, using doublehemizygous Nr4a1-eGFP/Drd1-tdTomato adult mice (Fig.  2a). In the striosomes, we counted a total of 506 cells, of which 50% were tdTomato ϩ , 43% EGFP ϩ , and 23% were double labeled. In the matrix, from a total of 1362 cells, only 17.1% were EGFP ϩ and only 5.8% were double labeled (Fig. 2b). Therefore, in the striosomes, assuming that the majority of Drd1-tdTomato Ϫ /DAPI ϩ cells are Drd2 ϩ , iMSNs and dMSNs are present at equal levels (i.e. 50% each). Focusing on Nr4a1 coexpression, 24% of the cells are Drd1 ϩ /Nr4a1 ϩ and 20% are Drd2 ϩ /Nr4a1 ϩ (Fig. 2c).

Nr4a1 promotes maturation of the medium spiny neuron in vitro
Given that Nr4a1 overexpression increases the striatal level of Oprm1 mRNA relative to wild type in vivo, we investigated whether Nr4a1 overexpression impacts the maturation of MSNs in vitro. Lateral ganglionic eminence primary neuronal cultures from E16.5 Nr4a1-eGFP embryos, compared with cultures from wild-type mice, have higher levels of Ppp1r1b/DARPP-32 ( Fig.3a; Sidak's multiple comparisons test, t (27) ϭ 2.910, p ϭ 0.0072) and Oprm1 mRNAs ( Fig. 3a; t (13) ϭ 3.338, p ϭ 0.0053). BDNF promotes the maturation of MSNs and requires Egr-1 (Keilani et al., 2012). We found that the BDNF and Nr4a1 effects on the induction of DARPP-32 are additive (Fig 3a; t (27) ϭ 5.479, p ϭ 0.0001; Fig. 3b) and that BDNF does not induce Nr4a1 mRNA (t (15) ϭ 1.448, p ϭ 1.682), implying the use of alternate signal transduction pathways. Importantly, EGFP fluorescence was visible in these cultures at the time of plating and Nr4a1 mRNA was already increased (Fig. 3b), indicating expression from the BAC transgene at this early age.

Impact of Nr4a1 overexpression on electrophysiological properties of dorsomedial striosomal MSNs and on striatal synaptic plasticity
Altered dopaminergic transmission at corticostriatal synapses is associated with impaired bidirectional synaptic plasticity, including LTP, LTD, and depotentiation (Shen et al., 2008;Cerovic et al., 2015;Trusel et al., 2015). Specifically, ERK has a crucial role in LTP induction as ERK inhibitors attenuate or even eliminate LTP in dorsomedial striatum (Xie et al., 2009). In Nr4a1-EGFP mice, we performed single-cell patch clamp recordings of EGFP ϩ and EGFP Ϫ neurons located in the center of striosomes. Based on the percentages of EGFP ϩ neurons that are also Drd1 ϩ ( Fig. 2a-c), we estimated that half of the EGFP ϩ cells from which we recorded were dMSNs and half were iMSNs. In the WT mice, in which MSN subtypes are indistinguishable morphologically, we assumed that the majority of the neurons from which we recorded are located in the matrix, which represents ϳ90% of the striatum, with a 1:1 distribution between dMSNs and iMSNs. Overall, EGFP ϩ neurons were more excitable than those recorded in WT mice, as determined by left-shifted current-frequency plots and lower rheobase currents (Fig.  6a,b; two-way ANOVA with genotype factor, F (2,57) ϭ 7.421, p ϭ 0.0014). This might be due to an intrinsic increased excitability of striosomal neurons compared with matrix (Crittenden et al., 2017). Membrane resistance and spike threshold were equal in WT, and EGFP ϩ and EGFP Ϫ MSNs (F (2,50) ϭ 0.7388, p ϭ 0.4828 and F (2,49) ϭ 0.5219, p ϭ 0.5967), but resting membrane potential (rheobase) was more depolarized in EGFP ϩ neurons (F (2,55) ϭ 4.303, p ϭ 0.0183; Fig. 6c-f). Notably, we observed the difference in excitability in mixed populations of MSNs despite the fact that D 1 dMSNs are less excitable than D 2 iMSNs (Gertler et al., 2008;Planert et al., 2010) and the EGFP ϩ neurons are equally likely to be D 1 or D 2 .
Using a standard high-frequency stimulation protocol (Calabresi et al., 1997), we observed that LTD induction was equivalent in WT neurons (cells registered from wildtype mice) and EGFP ϩ neurons (Fig. 6g,h; unpaired twotailed t test, t (15) ϭ 1.868, p ϭ 0.0815). LTP, however, could not be induced in the majority of EGFP ϩ neurons (only two of seven) using the same protocol, whereas it was readily induced in five of seven WT MSNs (data not shown). As we were unable to reliably obtain LTP in EGFP ϩ neurons with whole single-cell recordings, we also used field recordings to confirm this genotype-dependent effect, and assayed LTP, LTD, and paired-pulse ratio. In field recordings, LTD and paired-pulse ratio in Nr4a1-eGFP striatum were equivalent to those in WT (two-way ANOVA with genotype factor, F (1,19) ϭ 0.2400, p ϭ 0.6298; data not shown), but LTP, albeit present, was significantly decreased in amplitude (Fig. 6i,j; unpaired two-tailed t test, t (12) ϭ 3.011, p ϭ 0.0108). The normal paired-pulse ratio suggests that the corticostriatal glutamate release and AMPA receptor function are unaltered by Nr4a1 overexpression. The differences in LTP induction observed both by single-cell and field recordings strongly suggest an alteration in the activation of MSNs by constitutive overexpression of Nr4a1, but these data do not allow us to distinguish between the effects of genotype and compartment.

Discussion
Nr4a1 is a member of the Nur family of nuclear receptors, which are expressed in specific patterns in the CNS and periphery. In the brain, a high level of Nr4a1/Nur77 is found in dopaminoceptive striatal MSNs, where it is enriched in the striosomes (Davis and Puhl, 2011). The Nr4a1-null mouse has been characterized to some extent (Gilbert et al., 2006), but its striosomal architecture has not been described, and the effects of Nr4a1 overexpression have not been reported. In this study, we show that the GENSAT Nr4a1-eGFP reporter mouse expresses twice the normal level of Nr4a1 mRNA in the striatum, allowing us to examine the effects of Nr4a1 overexpression and deletion on specific aspects of striatal development and function. However, the mechanism leading to Nr4a1 overexpression in this animal model remains unknown.
We found that in the presence of increased Nr4a1, several markers of striosomal MSNs are increased both in the early postnatal period and in the adult, whereas markers of matrix MSNs are largely unchanged. Although striosomes are clearly demarcated in the absence of Nr4a1, they are smaller and occupy a lower percentage of the total area. Moreover, Nr4a1 overexpression in humanderived NSCs differentiated with Activin A further promotes their maturation toward a general MSN phenotype. Notably, Nr4a1 promotes maturation of the medium spiny neuron in vitro, including several striosome markers. The exact mechanism via which Nr4a1 regulates striosome formation is unknown, but a microarray study of hippocampal neurons in which Nr4a1 is overexpressed revealed the upregulation of several transcription factors also involved in striatal development, including Sp8, Meis1, and Gsx1 (Chen et al., 2014). We conclude that a   9) ϭ 1.340, p ϭ 0.3094). n ϭ 8 mice/genotype. One-way ANOVA corrected for multiple comparisons (Sidak's test). Data are presented as the mean Ϯ SEM. d, Calbindin and pERK immunolabeling shows that the induction of pERK occurs predominantly in the matrix compartment after a single intraperitoneal injection of cocaine (20 mg/kg). Scale bars, 100 m. The graph shows the quantification of pERK ϩ cells in matrix and striosomes in sections from bregma 0.86 mm. n ϭ3 mice, unpaired t test: t (2.003) ϭ 4.702, ‫ء‬p ϭ 0.0423. Data are wild-type level of Nr4a1 is required for normal striosome development and maintenance, suggesting unique functions of Nr4a1 and the absence of compensation by other members of the Nur family.
Nr4a1 deletion alters striatal response to dopamine agonists and antagonists, and the data herein show that overexpression leads to dysregulation of striatal plasticity and response to external stimuli. It remains to be determined how much of this is due to constitutive overexpression and/or the increased induction of Nr4a1 as an IEG due to the BAC transgene. Notably, the acute effects may be mitigated to some extent by the decreased induction of pERK, which is required for the induction of Nr4a1 transcription (Bourhis et al., 2008;Yue et al., 2017). Genetic disruption of Nr4a1 in rats has recently been shown to reduce the development of LIDs (Rouillard et al., 2018), but pERK level was not assayed in these animals. We are attempting to distinguish between the effects of these opposing activities on ERK phosphorylation.
Focusing on direct pathway function, we found that Nr4a1 overexpression impairs Drd1 signaling, reducing cocaine-induced phosphorylation of ERK, locomotor sensitization to cocaine, and LTP. We found dysregulation of multiple components of this signal transduction pathway that may contribute to decreased Drd1 signaling Girault, 2012a,b), but not all in the same direction, making it difficult to attribute any changes to one specific molecule. Overall, we conclude that Nr4a1 overexpression compromises AC5 availability reducing the efficacy of its responses to stimulatory GPCR inputs. As AC5 is the predominant cyclase in MSNs (Lee et al., 2002), it is unlikely that decreased AC activation reflects the dysfunction of an alternate cyclase. Interestingly, an increase in G i␣ inputs could reduce AC5 activation, and Oprm1, which is increased in Nr4a1-eGFP MSNs, is a G i␣ coupled receptor (Chakrabarti et al., 1995;Saidak et al., 2006;Lamberts et al., 2011;Traynor, 2012). Activation of AC5 via Oprm1 is required for morphine-induced locomotor activity (Kim et al., 2006). Therefore, chronically increased MOR tone could impact cyclase activation. Likewise, with our current knowledge, we are unable to pinpoint the etiology of ERK dysregulation. Increased activity of PP1 could decrease ERK phosphorylation (for review, see Pascoli et al., 2014), as could a genetic, albeit compensatory, increase in STEP61, as constitutive deletion of STEP leads to an increase in pERK1/2 levels (García-Forn et al., 2018). Conversely, CalDEG-GEFII (called Rasgrp1) is increased in striosomes in a rat model of LIDs, and its dysregulation in the presence of an increase in Nr4a1 may also contribute to the regulation of ERK phosphorylation (Crittenden et al., 2009), but the possible increase in activity of this pathway due to increased Nr4a1 clearly does not overcome whatever is inhibiting the phosphorylation of ERK.
Not surprisingly, the motor and signal transduction abnormalities in the Nr4a1-eGFP mice are associated with striatal electrophysiological abnormalities, which require further investigation. The decreased induction of pERK, the apparent reduction in LTP, and the lack of locomotor sensitization to cocaine are internally consistent. Changes in excitability may contribute to alteration in networks and plasticity, thereby impacting the response to psychostimulant (for review, see Crittenden and Graybiel, 2011;Cao et al., 2018). How striosomes may impact movement and response to psychostimulants remains an open question that should be further studied using methods in which striosomes and Drd1 are delineated in the absence of any molecular changes, so that effects of both compartmentalization and MSN subtypes can be distinguished. The D 1 R proportion of striosomes is highly dependent on their location and the relative expression of Oprm1 and the neuropeptides SP and ENK (Tajima and Fukuda, 2013;Miyamoto et al., 2018), the latter of which corresponds with the distribution of Nr4a1. Here we concluded that Nr4a1-EGFP is expressed equivalently in dMSNs and iMSNs, whereas Davis and Puhl (2011) reported an enrichment of Nr4a1-EGFP in the dMSNs. This discrepancy may arise from their use of Drd1 immunolabeling to identify dMSNs and in the location of the striosomes.
The composite effects of Nr4a1 overexpression are extremely complex and may also alter Drd2-mediated function and cholinergic interneuron activity (e.g., via opioidergic stimulation; Ponterio et al., 2013). We did not examine the morphology of MSNs, but, in the hippocampus, Nr4a1 overexpression eliminates neuronal spines (Chen et al., 2014) via transcriptional regulation of cytoskeletal proteins. Nr4a1 also regulates spine density in the striatum (Tian et al., 2010), which impacts Parkinson's disease and addiction phenotypes (Villalba and Smith, 2013). Finally, Nr4a1 is expressed in glia and is a key regulator of the inflammatory response in microglia and astrocytes (Ipseiz et al., 2014;Rothe et al., 2017;Popichak et al., 2018), another mechanism via continued presented as the mean Ϯ SEM. e, c-fos and GFP immunolabeling in the dorsal striatum of Drd1-eGFP and Nr4a1-eGFP adult mice 1 h after a single intraperitoneal injection of cocaine (20 mg/kg), indicating relatively reduced c-fos induction in Nr4a1-eGFP mice. Arrows indicate c-fos labeling in the GFP ϩ striosomes in the Nr4a1-eGFP mice. Scale bars, 50 m. f, Quantification of c-fos ϩ cells in Drd1-eGFP and Nr4a1-eGFP adult mice shown in e. n ϭ 3 mice/genotype and treatment, unpaired t test: t (4) ϭ 6.721, ‫‪p‬ءء‬ ϭ 0.0026. Data are presented as the mean Ϯ SEM. g, pERK immunostaining in WT and Nr4a1-eGFP mice after the injection NMDA antagonist MK-801(0.1 mg/kg) followed 30 min later by a single injection of cocaine (20 mg/kg) shows the abolition of pERK induction. Scale bars, 50 m. h, Schematic representation of treatments used for the induction of locomotor sensitization to cocaine. i, Nr4a1-eGFP mice show decreased locomotor sensitization to chronic cocaine use relative to WT mice. One-way ANOVA corrected for multiple comparisons (Bonferroni's correction; F (7,108) ϭ 8.639, p Ͻ 0.0001; WT cocaine day 1 vs WT cocaine day 5: t (108) ϭ 3.705, ‫‪p‬ءء‬ ϭ 0.003; Nr4a1-eGFP cocaine day 1 vs Nr4a1-eGFP cocaine day 5: t (108) ϭ 0.6920, p ϭ 0.4904). n ϭ 15 mice/genotype, One-way ANOVA corrected for multiple comparisons (Bonferroni's correction). Data are presented as the mean Ϯ SEM. ns ϭ non-significant. Figure 6. Characterization of electrophysiological properties of dorsomedial striosomal MSNs and the impact of Nr4a1 overexpression on striatal synaptic excitability and plasticity. a, Action potential sample traces from single cells derived from the dorsal striatum in WT mice and in the center of the striosomes for GFP ϩ and GFPneurons from Nr4a1-eGFP mice. b, Number of action potentials as a function of injected current intensity in WT, GFP ϩ , and GFPneurons indicate that neuronal action potentials are increased in GFP ϩ neurons. Two-way ANOVA with genotype factor (F (2,57) ϭ 7.421, p ϭ 0.0014). n Ն 17/genotype and cell type; ‫‪p‬ءء‬ Ͻ 0.01. Data are presented as the mean Ϯ SEM. c, Resting membrane potential is more depolarized in GFP ϩ neurons compared with WT (WT MSNs: average, Ϫ75.8 mV, n ϭ 17 cells; GFP ϩ MSNs: average, Ϫ71.7mV, n ϭ 22; GFP Ϫ MSNs: average, Ϫ74.2 mV, n ϭ 21). One-way ANOVA followed by Bonferroni's multiple-comparisons test: F (2,55) ϭ 4.303, p ϭ 0.0183, ‫ء‬p Ͻ 0.05. Data are presented as the mean Ϯ SEM. d, Rheobase current is lower in GFP ϩ neurons relative to WT and GFP Ϫ MSNs (WT MSNs: average, 171.1 pA, n ϭ 17 cells; GFP ϩ MSNs: average, 134.1 pA, n ϭ 22; GFP Ϫ MSNs: average, 164.7 pA, n ϭ 21). One-way ANOVA-Bonferroni's multiple-comparison test, F (2,51) ϭ 4.342, p ϭ 0.0181; ‫ء‬p Ͻ 0.05. Data presented as Ϯ SEM. e, Membrane resistance recorded in voltage-clamp experiments is equal in the three cell types. WT, n ϭ 10; GFP ϩ and GFP Ϫ , n ϭ 16. One-way ANOVA-Bonferroni's multiple-comparison test: F (2,50) ϭ 0.7388, p ϭ 0.4828. Data are presented as the mean Ϯ SEM. f, Spike threshold is equal in the three cell types. WT, n ϭ 15; GFP ϩ and GFP Ϫ , n ϭ 19. One-way ANOVA-Bonferroni's multiple-comparison test: F (2,49) ϭ 0.5219, p ϭ 0.5967. Data are presented as the mean Ϯ SEM. g, Whole-cell patch-clamp recordings of long-term synaptic depression (LTD) induced by high-frequency stimulation in WT and GFP ϩ MSNs showing overlapping traces for both genotypes. Data are presented as the mean Ϯ SEM. h, Bar graph representing the average of the last 5 min after LTD induction in WT and GFP ϩ MSNs indicating no significant differences between the two groups: WT, 10 recordings/7 mice; GFP ϩ , 7 recordings/3 mice. Unpaired two-tailed t test, t (15) ϭ 1.868, p ϭ 0.0815. Data are presented as the mean Ϯ SEM. i, LTP assayed in field recordings in WT and Nr4a1-eGFP shows reduced LTP in Nr4a1-eGFP mice after high-frequency stimulation. Data are presented as the mean Ϯ SEM. j, Bar graph representing the average of the last 5 min after LTP induction in field recordings (7 recordings and 3 mice for both genotypes). Unpaired two-tailed t test: t (12) ϭ 3.011, p ϭ 0.0108; ‫ء‬p Ͻ 0.05; Data are presented as the mean Ϯ SEM.