Substantia nigra dopaminergic neurons and striatal interneurons are engaged in three parallel but interdependent postnatal neurotrophic circuits

Abstract The striatum integrates motor behavior using a well‐defined microcircuit whose individual components are independently affected in several neurological diseases. The glial cell line‐derived neurotrophic factor (GDNF), synthesized by striatal interneurons, and Sonic hedgehog (Shh), produced by the dopaminergic neurons of the substantia nigra (DA SNpc), are both involved in the nigrostriatal maintenance but the reciprocal neurotrophic relationships among these neurons are only partially understood. To define the postnatal neurotrophic connections among fast‐spiking GABAergic interneurons (FS), cholinergic interneurons (ACh), and DA SNpc, we used a genetically induced mouse model of postnatal DA SNpc neurodegeneration and separately eliminated Smoothened (Smo), the obligatory transducer of Shh signaling, in striatal interneurons. We show that FS postnatal survival relies on DA SNpc and is independent of Shh signaling. On the contrary, Shh signaling but not dopaminergic striatal innervation is required to maintain ACh in the postnatal striatum. ACh are required for DA SNpc survival in a GDNF‐independent manner. These data demonstrate the existence of three parallel but interdependent neurotrophic relationships between SN and striatal interneurons, partially defined by Shh and GDNF. The definition of these new neurotrophic interactions opens the search for new molecules involved in the striatal modulatory circuit maintenance with potential therapeutic value.

that together constitute the striatal modulatory circuit. The relevance of these neuronal populations on striatal function is highlighted by the alteration of distinct neuronal types in several neurological disorders, including chorea (MSN), parkinsonism (DA SNpc), Tourette syndrome (ACh and FS), and dystonia (ACh). Although it is well known that acute alteration of a specific striatal component induces changes in the other elements of the circuit (Girasole & Nelson, 2015;Salin et al., 2009) (Ibáñez & Andressoo, 2016;Kumar et al., 2015;Pascual et al., 2008).
However, the role of GDNF in adult SNpc survival is still controversial Pascual & López-Barneo, 2015) and needs further clarification. Concomitantly, Sonic hedgehog (Shh), an extracellular ligand involved in cellular specification during development (Ingham & McMahon, 2001), is produced by DA SNpc and is Finally, we searched for alternatives sources of Shh that could be involved in ACh postnatal maintenance.

ACh on DA SNpc
Previous studies suggested that Shh produced by DA SNpc is required for the maintenance of both striatal ACh and FS (Gonzalez-Reyes et al., 2012). However, no alterations in the number of striatal interneurons have been associated with the progressive neurodegeneration of DA SNpc observed in Parkinson's disease (Fahn, 2009;Lang & Lozano, 1998). To evaluate the postnatal neurotrophic potential of DA SNpc over striatal interneurons, we decreased the number of dopaminergic neurons and, consequently, the production of Shh and other neurotrophic molecules synthesized by these cells (Figure 1a Scale bars: 25 µm. Graphs represent the total striatal number of FS (upper panels) and ACh (lower panels) interneurons measured by stereological methods from control (black bars) and Th-Sdhd (green bars) mice. In the entire figure, values represent mean ± SEM n = 6 per group. *p < 0.05 (Student's t test).
where the mitochondrial Sdhd gene is specifically inactivated in the DA SNpc (Diaz-Castro et al., 2012). Th-Sdhd mice showed a normal number of DA SNpc at birth but suffered a specific dopaminergic postnatal neurodegeneration as they acquired functionality. Sixmonth-old mice lacked more than 95% of DA SNpc, a reduction that remained stable in older mice and correlated with motor-related behavioral defects and with a strong decrease in striatal DA levels (Diaz-Castro et al., 2012). The authors also showed that the Th-Sdhd model did not present any nonspecific recombination in striatal neurons and that MSN and FS were not altered at the initial stages of the neurodegeneration (2 months old; Diaz-Castro et al., 2012). We extended this observation to older mice in order to establish the postnatal role of DA SNpc-derived signals over the integrity of striatal interneurons. Stereological unbiased quantifications of ACh and FS in 8-month-old mice revealed a clear decrease (42%) in the number of striatal FS without altering the ACh population ( Figure 1c).
Although Th-Sdhd mice presented reduced mobility and a difficulty for independent feeding, we were able to age a small group of mice  (Ingham & McMahon, 2001). Striatal ACh and FS acquire their phenotype during the first 2-3 postnatal weeks in rodents along with the expression of specific markers such as choline acetyltransferase (ChAT, Ach; Phelps, Brady, & Vaughn, 1989) and parvalbumin (PV, FS;Schlösser, Klausa, Prime, & Bruggencate, 1999). To conditionally delete Smo in ACh or FS, we selected the ChAT-Cre (Rossi et al., 2011) and PV-Cre (Hippenmeyer et al., 2005) lines (Figure 2a,e). To validate the specificity of these mouse lines, we used the Rosa26R-YFP reporter mouse. Almost all YFP-positive cells coexpressed the ChAT marker by immunohistochemistry in ChAT-Cre; R26R-YFP mice (Supporting Information Figure S1a) and more than 50% of cells positive for PV were also positive for YFP in the PV-Cre; R26R-YFP line (Supporting Information Figure S1b Figure S1c). In the PV- Smo model, we did not observe any changes in the mRNA levels of Smo at 1 month of age (Supporting Information Figure S1d). However, as time advanced, we found a significant decrease of around 50% in the striatal Smo levels (4-month-old mice), and this reduction was further confirmed at 10 months of age (Supporting Information Figure S1d). Notably, Smo is not only expressed within the striatum by ACh and FS but also by other neuronal and non-neuronal cells  Figure S2a). These mice developed a strong phenotype with a very reduced lifespan and body weight (Supporting Information Figure S2b,c) and fast neurodegeneration was observed in striatal FS (Supporting Information Figure S2d; 45 days old). As a control, we verified that the recombination in this new mouse model using the Rosa26R-Tomato reporter was similar to control mice and did not produce any nonspecific recombination (Supporting Information Figure S1e). Collectively, these data strongly suggest that Shh signaling is postnatally required in the striatum to preserve ACh but is not involved in the postnatal maintenance of FS.

| Shh is expressed by striatal interneurons
To Shh observed in our models (Supporting Information Figure S4a), we tried to replicate the effect of SAG and Cyc over GDNF expression.
No differences were found after stereological injection of both molecules using the protocol described in Ref. (Gonzalez-Reyes et al., 2012), although a trend to decrease with SAG treatment was detected (Supporting Information Figure S4b), altogether suggesting that the cross-regulation between GDNF and Shh may be of smaller magnitude than previously proposed.

| DISCUSSION
We define three parallel and interdependent neurotrophic relation- either that DA SNpc loss is not enough in the patients to provoke a decrease in the number of FS interneurons or that the dopamine replacement therapy used to treat those patients could prevent the loss of these cells. The first option is unlikely as DA SNpc loss is estimated to be 60%-70% at the onset of symptoms (Fearnley & Lees, 1991;Lang & Lozano, 1998). Favoring the second option is the report that, after binding to D 2 receptors, dopamine can be internalized to form a signaling complex (including ß-arrestin and protein phosphatase 2) that regulates the Akt pathway (Beaulieu et al., 2008), a cascade involved in neuroprotection (Dudek et al., 1997;Soler et al., 1999). Dopamine inhibits GABA A -mediated synaptic inputs to intrinsic striatal neurons (Bracci, Centonze, Bernardi, & Calabresi, 2002;Momiyama & Koga, 2001;Pisani, Bonsi, Centonze, Calabresi, & Bernardi, 2000) through presynaptic D 2 receptors (Centonze et al., 2003; for a review, see Berke, 2011). All together, we speculate that dopamine could be involved in the acute and chronic protection of FS striatal terminals and, therefore, could con-  (Chinta et al., 2007;Lin et al., 2012;Tillack, Aboutalebi, & Kramer, 2015).
Shh signaling has been proposed as a key factor for the physiol- | 9 of 14 normal physiology of these brain nuclei and for the fight against neurodegenerative disorders.

| Tissue preparation
For monitoring Cre recombinase expression, brains were embedded in gelatin and 50-µm-thick sections were obtained using a vibratome (Leica).
For in situ hybridization, 2-month-old wild-type mice were processed following RNAscope protocols (ACD). In brief, immediately following dissection the brain was fixed in 10% neutral-buffered formalin (Sigma) for 24 hr at room temperature. Fixed samples were paraffin-embedded using an automatic tissue processor (ASP300S; Leica).

| Light microscopy immunohistochemistry
Serial sections from control and transgenic mice were processed in parallel for light microscopy immunostaining using the same batches of solutions to minimize variability in the immunohistochemical labeling conditions. Coronal brain sections were deparaffinized using Imaging was performed with a BX-61 microscope (Olympus) or with a A1R Confocal (Nikon).

| In situ hybridization
Coronal mouse brain sections of 10 µm were obtained using a microtome (RM2255; Leica) and were then incubated overnight at

| Stereological analysis
The total number of FS (PV-immunoreactive) and ACh (ChAT-immunoreactive) was obtained by stereology-based quantification of the striatum from different transgenic mice (n is described in the figure legends) using an Olympus BX61 microscope and the NEWCAST software package (Olympus

| Quantitative RT-PCR
To determine mRNA levels, brain RNA was extracted from the striatum or ventral mesencephalon using TRIzol reagent (Invitrogen) in a

| GDNF ELISA
Striatal GDNF protein content was estimated using a commercial ELISA kit (GDNF Emax Immunoassay System; Promega). Brain was removed and immediately frozen in liquid nitrogen. Hemicortex and striatum were processed as described (Pascual et al., 2008). Absorbance from hemicortical sample extracts was subtracted to each individual striatal measurement.

| Preparation of acute brain slices
For the preparation of brain slices, we used the N-methyl-D-glucamine (NMDG) protective recovery method described by the laboratory of

| Electrophysiology
For whole-cell patch-clamp recordings, slices were transferred into a recording chamber that was perfused with 33 ± 1°C bubbled ACSF at and acquired using PatchMaster software (HEKA). All miniature postsynaptic currents were analyzed with the program Stimfit (Guzman, Schlögl, & Schmidt-Hieber, 2014). Recordings were first digitally filtered at 1 kHz. For each cell, all events were inspected to avoid false positives, and then, an average of all events detected was taken. with PBS, slices were incubated with secondary antibody (Alexa 488 donkey anti-rabbit, 1:500, Jackson Immunoresearch) in PBS and 0.3% Triton X-100 at 4°C overnight. Finally, slices were embedded with fluorescent mounting medium (Dako) and visualized in a confocal microscope (Zeiss LSM 7 Duo).

| Western blot
The striatum and ventral mesencephalon from one hemisphere were dissected on PBS, pH 7.4 on ice, and fast-frozen on liquid nitrogen.

| Stereotactic injection
Mice were anaesthetized using a solution of ketamine/xylazine at 100, 8 mg/kg, respectively. SAG and CYP were unilaterally injected into the right striatum (−0.5 mm anteroposterior; ±2.4 mm mediolateral; −2.5 mm dorsoventral) using a 1-µl neurosyringe (7001 KH; Hamilton). The needle was slowly lowered to coordinates, and it was left in place for 5 min after the injections before it was slowly withdrawn. Cyclopamine (C-8700; LC Laboratories) was diluted at 2 µg/ µl in 45% HBC (2-hydroxypropyl-beta-cyclodextrin; Sigma) in PBS, and 0.5 µl of the prior solution was injected. Smoothened Agonist SAG (Calbiochem) was diluted in PBS at 50 nM and 32 nl of the prior solution was injected. Left striatum was injected with Sham solution (vehicle). Animals were sacrificed 30 hr after injections, striata were dissected, and RNA was extracted.

| Statistical analysis
Data are presented as mean ± standard error of the mean. In the case of one variable, statistical significance was assessed by Student's t test with a Levene test for homogeneity of variance (normal distribution) or by the nonparametric Mann-Whitney U test (nonnormal distribution). For two variables, data were analyzed by twoway ANOVA followed by Sidak's multiple comparisons test. Survival curves were analyzed by Mantel-Cox test. GRAPHPAD PRISM 5.0 and SPSS 22.0 Software were used for statistical analysis.