Oasi Institute for Research on Mental Retardation and Brain Aging (IRCCS), Neuropharmacology Section, Enna, Italy
Department of Clinical and Molecular Biomedicine, Pharmacology Section, University of Catania Medical School, Viale A. Doria, 95125 Catania, Italy
Correspondence: Bianca Marchetti, Ph.D., Professor of Clinical Pharmacology, Department of Clinical and Molecular Biomedicine, Pharmacology Section, University of Catania, Viale A. Doria, 95125 Catania, Italy. Telephone/Fax: +39-095-378-4248; e-mail: email@example.com
Wnt/β-catenin signaling is required for specification and neurogenesis of midbrain dopaminergic (mDA) neurons, the pivotal neuronal population that degenerates in Parkinson's disease (PD), and in 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mouse model of PD. Wnt/β-catenin signaling plays a vital role in adult neurogenesis but whether it might engage DA neurogenesis/neurorepair in the affected PD brain is yet unresolved. Recently, the adult midbrain aqueduct periventricular regions (Aq-PVRs) were shown to harbor multipotent clonogenic neural stem/progenitor cells (mNPCs) with DA potential in vitro, but restrictive mechanisms in vivo are believed to limit their DA regenerative capacity. Using in vitro mNPC culture systems we herein demonstrate that aging is one most critical factor restricting mNPC neurogenic potential via dysregulation of Wnt/β-catenin signaling. Coculture paradigms between young/aged (Y/A) mNPCs and Y/A astrocytes identified glial age and a decline of glial-derived factors including Wnts as key determinants of impaired neurogenic potential, whereas Wnt activation regimens efficiently reversed the diminished proliferative, neuronal, and DA differentiation potential of A-mNPCs. Next, in vivo studies in wild (Wt) and transgenic β-catenin reporter mice uncovered Wnt/β-catenin signaling activation and remarkable astrocyte remodeling of Aq-PVR in response to MPTP-induced DA neuron death. Spatio-temporal analyses unveiled β-catenin signaling in predopaminergic (Nurr1+/TH−) and imperiled or rescuing DAT+ neurons during MPTP-induced DA neuron injury and self-repair. Aging inhibited Wnt signaling, whereas β-catenin activation in situ with a specific GSK-3β antagonist promoted a significant degree of DA neurorestoration associated with reversal of motor deficit, with implications for neurorestorative approaches in PD. Stem Cells2014;32:2147–2163
Parkinson's disease (PD) is a common, age-dependent neurodegenerative movement disorder characterized by the progressive loss of midbrain dopaminergic (mDA) neurons of the substantia nigra pars compacta (SNpc) and their projections into the corpus striatum leading to substantial decreases in dopamine levels . Current pharmacological treatments for PD are comprised mainly of dopamine replacement for symptom alleviation, but lack of effectiveness in preventing clinical progression, therefore neuroprotective therapies are actively investigated .
Recently, stem cell-based replacement therapies have emerged as potential strategies [3-5] and the molecular mechanisms that govern neurogenesis and differentiation of mDA neurons have attracted intense investigations for the identification of intrinsic DA neuron determinants. Particularly, Wingless-type MMTV integration site (Wnt) pathway is recognized as one central player in mDA neurogenesis [6-14]. Besides its recognized role in regulating multiple processes in developing brain, accumulating evidence points to Wnt signaling as vital regulator of adult neuron maintenance, while Wnt dysregulation has been linked to neuron demise in major neurodegenerative disorders, including PD [15-25].
Aging, the strongest risk factor for PD, and the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mouse model of PD are associated with a gradual decline of the capacity of mDA neurons to recover upon injury [26-28]. This is at least in part due to impaired neurogenic potential [29-33].
Neural progenitors in neurogenic areas such as the subventricular zone of the lateral walls of the lateral ventricles (SVZ) and the subgranular zone (SGZ) of the hippocampus are in intimate contact with astrocytes which helps to generate an instructive “niche” that promotes neurogenesis ([34-37], and references herein). Notably, astrocyte-derived Wnts and Wnt/β-catenin signaling activation contribute to the regulation of adult neurogenesis [22, 38-40], but with age, a dysfunctional Wnt/β-catenin signaling impairs both hippocampus SGZ [41, 42] and SVZ  niches.
The occurrence of neurogenesis in the adult SNpc still remains controversial [29, 32, 43, 44]. However, the midbrain-aqueduct periventricular regions (Aq-PVRs) harbor multipotent clonogenic neural stem/progenitor cells (mNPCs) with an exclusive nigral DA differentiation potential [45-47], that might provide DA neurons that ultimately migrate to reach the SN [44, 47-49]. However, strong cell-intrinsic and extrinsic restrictive factors in vivo are believed to limit their neurogenic and DA regenerative capacities.
In the light of our recent studies indicating a potential role for astrocyte-derived factors and Wnt/β-catenin signaling in promoting neurogenesis and DA neurogenesis from adult mNPCs , we herein investigated the potential of Wnt/β-catenin signaling activation for endogenous mDA cell replacement or repair in the MPTP mouse model of PD, and started addressing the following questions: (a) which factors/mechanisms regulate the behavior of these Aq-PVRs midbrain progenitors and/or the permissiveness of the Aq microenvironment in promoting neurogenesis and DA neurogenesis during the process of aging and MPTP-induced nigrostriatal injury, in vitro?. (b) Are Wnt/β-catenin-responsive progenitors present in periaqueductal regions in vivo, and if so, will they respond to MPTP-induced DA neuron injury? (c) Is it possible to activate these progenitors to rescue dopaminergic plasticity in aged-MPTP mice? We herein identify the midbrain Aq-PVR as novel Wnt responsive region and present evidences suggesting the potential for mDA neuron restoration by activating Wnt/β-catenin signaling in endogenous sources, with therapeutical implications for PD.
Materials and Methods
Mice and Treatments
Young (2–5 months old) and aged (9–12 months old) male C57BL and BATGAL  reporter mice were used. Transgenic BATGAL mice express nuclear beta-galactosidase under the control of the beta-catenin-activated transgene (BAT) promoter . The BATGAL reporter mice used in this study were purchased from the Jackson laboratory (Tg(Bat-lac-Z)3 Picc/Tg(Bat-lacZ)3Picc, http://www.informatics.jax.org/allele/MGI:3697064) on a C57Bl background. Mice were maintained in standard laboratory conditions and all studies carried out in strict accordance with the the Guide for the Care and Use of Laboratory Animals (NIH), approved by the Institutional Animal Care and Use Committee guidelines. All surgeries were performed under anesthesia. The MPTP-(1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine, Sigma, St. Louis, MO, http://www.sigmaaldrich.com)-based mouse model of PD was used [33, 51]. At the indicated days post-MPTP treatment (dpt), mice were given bromodeoxyuridine (BrdU, Sigma, 50 mg/kg, injected four times, 2 hours apart) and killed 2 hours after the last injection [22, 33].
Neural Stem/Precursor Cells
For isolation and processing of stem/neuroprogenitors from midbrain-Aq-PVRs (mNPCs) of young and older mice, we used the neurosphere culturing techniques similar to those used for propagation of NPCs derived from the SVZ, as previously reported [20, 22, 33, 52] and detailed in Supporting Information. Neurospheres at passage numbers 3–4 and 8–10 were used in all in vitro experiments. For proliferation and differentiation studies, the neurospheres were mechanically dissociated into single cells and plated at a final density of 1 × 105 cells per centimeter on poly(D-lysine)-coated 24-well plates (Sigma). Proliferation was studied after exposure to proliferative medium for 3 days in vitro (DIV), by addition of the nucleotide analogue BrdU (Sigma, 5 µM) at 2 DIV and the cells fixed after 24 hours at 3 DIV. The differentiation of mNPCs was initiated by removal of mitogens and plating the cells on PDL (monotypic cultures), or onto astrocyte monolayers (cocultures) in the absence or the presence of the indicated treatments, as described. mNPCs were left to differentiate for 10 DIV.
Glial Cell Cultures, NPC-Glial Cocultures, and Cell Treatments
Purified astrocyte cell cultures obtained from postnatal days 2 (P2) and older (2–24 months) mouse midbrain-Aq region were used for coculture paradigms with Y/A-NPCs, either untreated (phosphate-buffered saline, PBS) or treated [20, 22, 33]. For DA differentiation, mNPCs were grown alone, or layered on the top of astrocytes. The differentiation medium contained 2.5% fetal calf serum (FCS) instead of BSA (bovine serum albumin, all from Sigma). After 3 DIV, the growth medium was changed and replaced with fresh serum-free differentiation medium (neural basal medium, N2 from Gibco, Tulsa, OK, http://www.invitrogen.com), containing 1 mg/ml BSA and 200 µM ascorbic acid (Sigma). For pharmacological activation of Wnt/β-catenin signaling, we used the specific GSK-3β (glycogen synthase kinase-3β) inhibitor, AR-AO14418 [N-(4-methoxybenzyl)-N0-(5-nitro- 1,3-thiazol-2-yl) urea] (AR, 5 mM, Sigma), or the Wnt ligand, Wnt1 (100 ng/ml, Abcam, Cambridge, U.K., http://www.abcam.com) [20-22]. For Wnt antagonism, we used Dickkopf-1 (Dkk-1, 100 ng/ml), or Frizzled-1-cysteine-rich domain (Fzd-A, 200 ng/ml), both from R&D Systems, Minneapolis, http://www.rndsystems.com [20-22].
Cell cultures were fixed in 4% paraformaldehyde in PBS or with paraformaldehyde/PBS followed by ice-cold acidic ethanol and HCL for BrdU staining [20, 22, 33]. Analyses performed using a confocal laser microscope and computer-assisted image analysis (LEICA TCS NT, Version 1.0, Leica Lasertechnik GmBH, Heidelberg, Germany, http://www.leica-microsystems.com/), equipped with an argon/krypton laser. Quantification of the percentage of cells expressing a given marker or marker combinations was determined relative to the total number of DAPI-(4,6-diamidino-2-phenylindole, Vector Laboratories, https://www.vectorlabs.com/)-labeled nuclei or Tuj1+ (neuron-specific beta-III tubulin-positive, Covance, Princeton, NJ, http://www.covance.com/) cells using the Leica lite Software and three-dimensional overlay to avoid false-positive/negative overlay and double counting.
Caspase-3 activity was evaluated as a marker of cell death [19, 20], using the fluorogenic substrate DEVD-AFC (Ac-Asp-Glu-Val-Asp [DEVD]-pNA, Millipore, Temecula, CA 92590 USA, http://www.millipore.com, Upstate Biotechnology). Samples were analyzed in a plate reader at 405 nm and enzymatic activity is expressed as arbitrary fluorescent units [20, 21, 33].
Immunohistochemistry and Cell Counting
Serial coronal sections (14-µm thick), encompassing the striatum (Bregma 1.54 to bregma −0.46) and the SNpc-Aq regions (Bregma −2.92 to bregma −4.84 mm) according to Franklin and Paxinos  were collected, mounted on poly(L-lysine)-coated slides and preabsorbed primary antibodies reported in Supporting Information Table S1. Quantification of the percentage of cells expressing a given marker or marker combination, in any given experiment was determined as above [20, 22, 33]. For semiquantitative analyses, cells (∼100) were counted within the complete region of interest using the Leica lite Software and three-dimensional overlay as above, in sections from five different animals, and each region quantified in at least five separately stained coronal slices. All cells (DAPI-stained nuclei) within the first 300 µm of the Aq round were counted.
RNA Extraction, Reverse Transcription, and Real-Time PCR
Tissue/cell samples were processed for total RNA isolation as described in full details [20, 22, 33]; and Supporting Information] using Taqman Assay Reagents (Applied Biosystems, (http://www.lifetechnologies.com), q-RT-PCR performed according to manufacturers protocol. The assay IDs are reported in Supporting Information Table S2. Results are expressed as arbitrary units (A.U.).
Western Blot Analysis
Protein extracts prepared from tissue/cells isolated from saline or MPTP mice (n = 6–8 mice/age-group/tp) and from cell cultures (run in triplicates) within the different experimental groups [20, 22, 33]. Primary antibodies sources and dilutions used are detailed in Supporting Information Table S1. Bands from the Western blots were densitometrically visualized, the signals quantified on X-ray films using the data subjected to statistical analysis of variance (ANOVA).
Effect of Manipulation of Wnt/β-catenin Signaling, In Vivo
Pharmacological activation of Wnt/β-catenin signaling was carried out with AR [18-20]. Mice anesthetized with chloral hydrate (600 mg/kg) were positioned in a stereotaxic apparatus. The following stereotaxic coordinates were used: AP= 3.5 mm posterior from the bregma, L = 0.2 mm lateral to the midline . The needle was placed at a distance of 2.8 mm from the dura . Groups of mice received unilateral infusion of either the Wnt/β-catenin signaling activator AR or the vehicle (PBS), which served as control, into the Aq-PVR. Mice were let to recover for 24 hours, then treated with MPTP. At different time-intervals, part of the mice was subjected to the Rotarod analysis for motor coordination. On the day of sacrifice, mice were given BrdU and killed 2 hours after the last injection.
Dopaminergic Endpoints in Striatum and Ventral Midbrain
The principal DA endpoints at SNpc and striatal levels were investigated before and at the indicated time-intervals upon MPTP administration. Briefly, specific high-affinity neuronal dopamine uptake in striatum expressed as fentomoles of dopamine uptake per microgram of protein minus the fentomoles of mazindol uptake  was determined before and 21 and 45 dpt. Values are represented as percentage changes in dopamine uptake versus control. Striatal dopamine (DA) was determined by HPLC as detailed elsewhere  in striata of both sides. The loss of TH+ (tyrosine hydroxylase-positive)/DAT+ (dopamine transporter-positive) SNpc (subtantia nigra pars compacta) neurons was determined by serial section analysis of the total number of TH+ cells counted through the entire rostro-caudal axis of the murine SNpc , and counting performed as detailed [20, 21] either before and 1–3, 7–14, 21–30, and 45–65 days after MPTP. DAT- or TH-fluorescence intensity (FI) in striatum was assessed at the same time points in n = 3 coronal sections, in n = 6 mice/group/time point, and changes in average FI (mean ± SD) expressed as percentage (%) of saline-injected controls .
For the analysis of motor coordination, a five-lane accelerating rotarod; Ugo Basile, Comerio, Italy (http://www.ugobasile.com/catalogue/product/47600_mouse_rota_rod.html) was used . Mice have to keep their balance on a horizontal rotating rod (diameter, 3 cm) and rotation speed was increased every 30 seconds by 4 rpm. Five mice were tested at the same time, separated by large disks. A trial starts when the mouse is placed on rotating rod, and it stops when the mouse falls down or when 5 minutes are completed. Falling down activates a switch that automatically stops a timer. The testing day, each mouse is submitted to five trials with an intertrial interval of 30 minutes. Mice housed five per cage were acclimated to a 12 hours shift in light/dark cycle so that the exercise occurred during the animals normal wake period. Saline- and MPTP-treated mice undergoing PBS or AR infusions as described were assessed for their Rotarod performance at −7 days and + 3, +7, +14, +28, +35, and +45 post-MPTP (10/experimental group).
Statistical significance between means ± SEM was analyzed by a two-way ANOVA, and Student's t test for paired or unpaired data. Experimental series performed on different days were compared by the Student-Newman-Keuls t test. A value of p < .05 was considered to be statistically significant.
Age and MPTP Impair Stem/Neuroprogenitors from Midbrain Aq-PVRs
The In Vitro Model
Stem/neuroprogenitors from adult mouse midbrain caudal ventricular regions (mNPCs) (Fig. 1A) express proliferation (bromodeoxyuridine, BrdU), precursor (nestin, Musashi1), proneural (oligodendrocyte transcription factor 2, Olig2; neurogenin2, Ngn2), and astrocyte cell markers (glial fibrillary acid protein, GFAP) during in vitro clonal expansion (Fig. 1B, 1C), both at a protein and/or mRNA levels, confirming previous studies including our own [20, 45, 46]. The high expression of the midbrain marker engrailed 2 (En2) as opposed to very low forebrain (Dlx2) and hindbrain (Hoxd3) markers, supported their midbrain phenotype. The proliferative, differentiation, and survival properties of mNPCs from 2 to 5 (young, Y) and 9 to 12 month-(M)-old (aged, A) mice were then assessed in basal conditions and during MPTP-induced injury and recovery (Fig. 1, 1D). Y-mNPCs of three to four passages grown in proliferative medium for 3 DIV express nestin, of which a certain percentage incorporate BrdU (Fig. 1E, 1G). When shifted in differentiation medium and stained after 10 DIV, a certain proportion of cells express the neuronal marker Tuj1 (Fig. 1E, 1H), thereby proving their NPC behavior. By contrast, A-mNPCs show significantly decreased percentages of BrdU+ and Tuj1+ cells (Fig. 1F–1H), indicating that with age, mNPC proliferative and neuron differentiation potential are impaired.
In Y-mNPCs isolated 7 days post-MPTP treatment, both proliferation and neuron differentiation were only transiently decreased (Fig. 1E, 1G, 1H), to return back to pre-MPTP levels by 45 dpt (Fig. 1E, 1G, 1H). In A-mNPCs, BrdU+ and Tuj1+ cells at both 7 and 45 dpt were significantly (°, p < .05) reduced as compared to Y-mNPC counterparts (Fig. 1F–1H). Of note, MPTP further reduced neuron differentiation potential of A-mNPC at 7 dpt with respect to their saline-treated counterparts, with a return back to aged control levels by 45 dpt (Fig. 1H). Conversely, MPTP did not changed the already low levels of A-mNPCs proliferation at both time points (Fig. 1G).
Using the fluorogenic substrate DEVD-AFC to measure Caspase3-like activity as death marker, we found increased DEVD-signal in mNPC isolated from A-MPTP mice as compared to Y-counterparts (Fig. 1I), indicating reduced mNPC survival capacity.
When Y/A-mNPCs were expanded for 8–10 passages in the presence of FGF-II and GF and neuron differentiation potential studied after withdrawal of growth factors, addition of serum and N2 supplements for 10 DIV, we found that Y- and A-mNPCs expressed nestin, Tuj1, and GFAP, however in A-mNPC cultures, fewer Tuj1+ cells were counted (Fig. 1F, 1J) as compared to Y-mNPCs (Fig. 1E, 1J) with a large proportion of GFAP+ astrocytes (Fig. 1F). Altogether, aging and MPTP impair mNPCs plasticity with no recovery after in vitro expansion which correlate with the failure of TH+ neuron to recover from MPTP insult.
Aging Inhibits Canonical Wnt/β-catenin Signaling in mNPCs: Effect of Glial Age
Using q-RT-PCR and specific primers for β-catenin, the hallmark of “canonical” Wnt/β-catenin pathway , we found that Y-mNPC expresses moderately high β-catenin both at a mRNA and protein levels, while these levels are sharply reduced in A-mNPCs (Fig. 2A, 2C). In mNPCs isolated 45 days post-MPTP, β-catenin mRNA and protein levels were still decreased in A-mNPC as opposed to younger counterparts showing a significant upregulation of β-catenin (Fig. 2B, 2C). Additionally, CyclinD1, a β-catenin target gene regulating the progression of the G1 to the S phase of the cell cycle, was reduced in A- as compared to Y-mNPCs counterparts, both after saline or MPTP exposure (Fig. 2A, 2B), thus supporting reduced canonical Wnt/β-catenin signaling. In keeping with these findings, Axin2, a Wnt target gene regarded as one most reliable endogenous index for measuring Wnt/β-catenin pathway activation , was significantly decreased in A- as compared to Y-mNPCs, with or without MPTP treatment (Fig. 2A, 2B). Next, the expression of glycogen synthase kinase-3β (GSK-3β) which is part of a destruction complex leading to β-catenin ubiquitination and degradation by the proteasome  was significantly increased both at mRNA and protein levels in A- compared to Y-mNPCs (Fig. 2A–2C), thus indicating that GSK-3β upregulation may play a role in the depletion of β-catenin observed in A-mNPCs. Likewise, using immunocytochemistry, decreased β-catenin-IF in A- as compared to Y-mNPC cultures was observed (Supporting Information Fig. S2A, S2B) whereas exposure to Wnt1 (100 ng/ml) sharply increased BrdU-, β-catenin-, and Tuj1-IF cell formation in both Y- and A-mNPC, albeit this effect was stronger in younger counterparts (Supporting Information Fig. S2C–S2H).
Astrocytes within the adult neurogenic niches produce Wnts that act in a paracrine manner on neuronal progenitors to induce expansion and/or neuron differentiation [8, 20, 22, 38-40]. Given that our previous study documented robust proneurogenic abilities of midbrain astrocytes , we used different coculture paradigms between young/old Aq-PVR astrocytes and Y/A- mNPC (Fig. 2D), in the absence or the presence of Wnt/β-catenin activators/inhibitors to address the effect of glial age and glial-derived Wnt. In accord with previous studies [8, 20], coculture of Y-mNPCs with purified astrocytes markedly stimulated neurogenesis. Here, we further show that this process depends on age, with 2-day and 2-month (Fig. 2E, 2M)-old astrocytes sharply increasing the proportion of BrdU+ and Map2a+ out of the Dapi+ nuclei, compared to mNPC exposed to a nonspecific cell type, whereas exposure to 10- or 24-month-old astrocytes reversed astrocytes proneurogenic abilities (Fig. 2F, 2M). Conversely, A-mNPCs layered on the top of astrocyte monolayers of 2-day and 2-month-old, showed a remarkable increase Map2a+ cell number as compared to A-mNPC grown alone (Fig. 2I, 2N), indicating that while glial is a limiting factor A-mNPCs still retain their neurogenic potential if adequately stimulated. Reciprocally, when A-mNPCs were exposed to 10-month- (Fig. 2J) or 24-month-old (Fig. 2N) astrocytes a further inhibition of proliferative potential and the formation of Map2a+ cells were observed, suggesting a synergy between astrocyte age and mNPC age in promoting neurogenic impairment.
Astrocyte-derived Wnts decline with age within the midbrain , the hippocampus , and the SVZ . Here again, Wnt1 mRNA measured in Y-astrocytes (A.U. = 0.6 ± 0.05) decreased to almost undetectable levels in 10–24-month astrocytes of the midbrain-Aq region. Then, to activate β-catenin signaling in aged cocultures, we used a specific antagonist of GSK-3β, AR-AO14418 (AR) that stabilizes β-catenin in adult NPCs [20, 22, 57] and observed that aged-astrocyte-induced decreased neuron production was efficiently counteracted in both Y- and A-mNPC cocultures (Fig. 2O). Conversely, using the Wnt/β-catenin antagonists, Dikkopf-1 (Dkk1, 100 ng/ml), or Frizzled-1-cysteine-rich domain (Fzd-A, 200 ng/ml), that binds to the Wnt ligand preventing its interaction with the Fzd receptor, we found a significant reversal of young astrocyte proneurogenic capacities, thereby supporting Wnt/β-catenin signaling intermediacy (Fig. 2O).
In the study of Hermann et al. , using a multistep differentiation protocol in coculture conditions with PA6 stromal cells, a small number of mNPCs acquired morphological and functional properties of mDA neurons in culture . In our previous studies , we showed that coculture with astrocytes or activated astrocytes (that received a pretreatment with the chemokines CCL3, CXCL11, or CXCL10 for 24 hours) did stimulate tyrosine hydroxylase-positive (TH+) neuron formation. We then exposed both Y/A-mNPCs to astrocytes or activated astrocytes as described  and observed significantly higher numbers of Map2a+ neuron formation (Fig. 2G, 2K), of which a certain percentage of TH+ out of the Tuj1+ cells was counted (Fig. 2H, 2L, 2P). We next verified the role of astrocyte-derived Wnt with Dkk1, and found a significant inhibition in the number of TH+ cells (Fig. 2P), whereas the GSK-3β antagonist, AR, sharply increased TH+ neuron production (Fig. 2P), thus supporting a permissive role of astroglial-derived factors together with canonical Wnt ligands in TH+ neuron formation.
Together, Wnt/β-catenin signaling is dysfunctional in aged mNPC. Astrocytes with age loose proneurogenic and DA differentiation abilities via decline of astrocyte-derived factors including Wnts. Additionally, astrocyte-coculture paradigms coupled to Wnt-activation regimens can rescue these progenitors and promote TH+ neuron formation.
MPTP Injury Promotes GFAP and β-Catenin Remodeling in Aq-PVRs: Effect of Age
The In Vivo Model
Next, we addressed the role of Wnt/β-catenin signaling in vivo. First, using fluorescence immunohistochemistry coupled to confocal laser microscopy applied to coronal midbrain sections we identified β-catenin-IF cells in midbrain Aq-PVRs of both Y- (Fig. 3A1–A4, 3E) and A- (C1-C4, E) saline-treated mice, albeit they were reduced with age. β-Catenin-IF signal was localized in the ventricular lining and the Aq round. Bright GFAP+ cells were seen bordering the ventricular wall all along the Aq-PVRs. Upon MPTP-induced DA neuron injury time course analyses uncovered a remarkable increase β-catenin+ cells by 1 day postlesion (Fig. 3B1, 3B2, 3E), whereas A-mice showed only a slight increase of β-catenin+ cells (Fig. 3D1-3D4, 3E), suggesting an impairment of the aged Aq-PVR β-catenin response to injury. In Y-MPTP mice, β-catenin+ cells were delineated by a strongly immunolabeled and thick GFAP+ astrocyte cell layer (Fig. 33B2-B4), while A-mice showed smaller increases in both β-catenin+- and GFAP+ cells (Fig. 3D2-D4). By 3 dpt, GFAP+ astrocytes were seen almost segregating the β-catenin+ cell domain (Fig. 2B3, 2B4) in young mice, showing elongated cell bodies and long processes, coursing together/intermingled with β-catenin+ cells (Fig. 3B4). Especially, a remarkable increase of polysialylated neural cell adhesion molecule, PSA-Ncam+ cells, within the β-catenin+ domain was observed (Supporting Information Fig. S3), suggesting stem/neuroprogenitor activation early upon DA neuron injury. By 7 postinjury, β-catenin+ cells were downregulated, to recover over pre-MPTP levels by 21 dpt, indicating that changes in Aq microenvironment might influence cell survival or migration. In stark contrast, in A-mice, after a slight increase of β-catenin+ cells 1–3 dpt, a significant reduction was observed at all times studied (Fig. 3D1-3D4, 3E). By applying q-RT-PCR to freshly isolated midbrain periaqueductal tissues (Fig. 3F) we found that Wnt1 and β-catenin mRNAs were significantly increased in young mice 1 day after MPTP, while by 7 days a downregulation of Wnt1 and β-catenin transcripts occurred to recover back by 21 dpt (Fig. 3F), whereas in older counterparts Wnt1 and β-catenin mRNAs were reduced at all time-studied, in the face of upregulated GSK-3β transcript levels (Fig. 3F), thereby denoting failure of Wnt1, to be induced in response to DA neuron injury in A-mice.
MPTP Injury Promotes β-Catenin Signaling in Aq-PVR Niche of Young BATGAL Reporters
With LacZ driven by a siamois promoter and seven T-cell factor/lymphoid enhancer factor binding sites, transgenic (Tg) BATGAL mice allow the visualization of cells with transcriptionally active β-catenin signaling through the expression of Xgal and β-galactosidase (β-gal) the protein product of the LacZ transgene . Wt and Tg mice were treated with either saline or MPTP and midbrain sections were stained with Xgal and anti-β-gal antibodies (Abs). Confirming previous studies [38, 39], BATGAL+ cells were found in SVZ and hippocampal SGZ of Tg (not shown) but not Wt controls. Within the Aq-PVRs of young saline-treated young reporters (at Bregma −3.08 to −3.64), we identified Xgal+ nuclei as opposed to Wt controls, where no Xgal staining was detected (Fig. 4A1, 4A2). As soon as after 1 day post-MPTP-injury, Xgal staining sharply increased in the Aq round and the periaqueductal area (Fig. 4A3, 4A4). β-Gal staining colocalized with dapi counterstained nuclei (Fig. 4B1, 4B2), occupying the domain of the β-catenin+ cells, with no colocalization of β-gal with GFAP+ cells (Fig. 4B1, 4B2). Here again, MPTP-induced a profound remodeling of Aq architecture, with GFAP+ astrocytes emanating long processes, as they were “embrassing” β-gal+ nuclei (Fig. 4B1, 4B2). Using the Axin2-LacZ mouse strain [56, 58] (Tg Axin, Fig. 4B3) where the Wnt target gene Axin2 is mutated by insertion of LacZ, MPTP induced a comparable increase in β-Gal+ nuclei and GFAP activation, mirroring BATGAL cell activation, thus corroborating endogenous Wnt/β-catenin signaling activation in Aq progenitors in response to MPTP-induced DA neuron death in the two reporter lines. Increased β-catenin signaling of BATGAL mice correlated with β-catenin and Wnt1 mRNAs expression, confirming Wnt/β-catenin activation observed both in vitro and in vivo (Figs. 2, 3). Conversely, when BrdU was injected on the day of sacrifice and mice killed after 2 hours, we did not observe BrdU-expressing nuclei, possibly indicating that these β-gal+ cells may represent postmitotic precursors [46, 59, 60].
During development, astrocyte-derived Wnts and β-catenin control DA neurogenesis by maintaining the integrity of the neurogenic niche and promoting the progression from Nurr1+/TH− postmitotic progenitors to Nurr1+/TH+ neurons with a pivotal role exerted by Wnt1 [6-10, 61-63]. In BATGAL mice, Nurr1/NR4A2 transcript levels were sharply increased (Fig. 4D). Interestingly, waves of Nurr1+/TH− cells were seen as they were migrating off the aqueductal region (Fig. 4C1, 4C2). Traking X-Gal/β-gal spatiotemporally, we found maximal X-Gal/β-gal in Aq-PVR region at the peak of DA neurodegeneration (i.e., 1–3 dpt, Fig. 4F). Later on, Xgal stained nuclei and β-gal+/Nurr1+ neurons emanated from the ventral midline (Fig. 4C3, 4C4, 4E, 4F). Together, Wnt/β-catenin signaling is dynamically regulated in Aq-PVR in young reporters in response to mDA neuron death. In the studied conditions, β-catenin signaling is not associated to proliferation but is spatio-temporally activated in postmitotic Nurr1+/TH− neurons.
MPTP-Induced Neuron Injury/Repair Activates Wnt/β-Catenin Signaling Outside the Aq-PVR Niche
The question, then, arised as to whether/how activation of β-catenin signaling in young mice might have functional consequences for DA neuron injury and repair after MPTP exposure. In Wt mice (Fig. 5A1), bright DAT+ neurons of the SNpc (A9) and the ventral tegmental area (VTA, A10) can be observed . DAT+ positive neurons are also seen emerging from the Aq (inset in A1) and GFAP+ astrocytes are clearly visualized both within the Aq-PVRs and SNpc (Fig. 5A1). In Tg saline-injected mice, discrete β-gal staining was observed in the SNpc (Fig. 5A2). Immunofluorescence coupled to confocal laser microscopy of dual staining with β-gal and DAT, followed by quantitative analysis of the percentage of DAT+/β-gal+/over the total number of DAT+ in SNpc, indicated β-catenin signaling in approximately 6% of mature DAT+ neurons (Fig. 5A2, 5A3, 5D1), likely involved in the maintenance of adult SNpc DA neuron survival.
In Wt mice, MPTP-induced an approximate 60% loss of SNpc TH+ neurons by 7 dpt, associated by an almost complete disappearance of TH+ neurons along the midline of the ventral tegmentum, and a sharp increase of GFAP+ astrocytes in both Aq-PVRs and SNpc levels (Fig. 5B1 and Supporting Information Table S3). In Tg mice, tracking Xgal/β-gal in SNpc as a function of time and DA degeneration showed numerous Xgal+ nuclei by 7 dpt (Fig. 5B2, 5B3). Dual staining with DAT and β-Gal (Fig. 5B4) showed active β-catenin signaling in a certain proportion of DAT+ neurons, but no colocalization between GFAP and β-gal (Fig. 5B5). Spatio-temporal analyses carried out during the active degeneration phase (1–3 dpt), the stabilization of the lesion (7–14 dpt), and the recovery phase (>21 dpt), indicated a significant increase of β-catenin signaling 7–45 dpt (Fig. 5D1), thus suggesting increased β-catenin signaling in DAT+ neurons surviving and recovering from MPTP insult.
In Wt mice, with time, a slow nigrostriatal DA neuron recovery is observed between 45 and 65 dpt , as reflected by the significant increase of DA cell body numbers in SNpc, of DAT-IF fibers in striatum, and of DAT+/Nurr1+ in SNpc (Supporting Information Table S3). By 65 dpt, an intense trafficking of bright TH+ neurons and GFAP+ astrocytes was still seen coursing from the Aq and midline as they were migrating to reach the VTA and SNpc DA subfields (Figs. 5C1, 6B1-6B9). In Tg mice, the recovery of nigral cell bodies started already from 21 to 30 dpt on, as reflected by the significant increase in percentages of DAT+ neurons from almost 38% (at 7–14 dpt) to approximately 58% (at 21–30 dpt) (Fig. 5D1). Likewise, Nurr1+/DAT+ neurons increased from almost 28% (7–14 dpt) to 60% (21–30 dpt) (Fig. 5D2).
Dual staining with β-gal and DAT showed a remarkable increase of DAT+ neurons, of which a significant percentage coexpressed β-gal (Fig. 5C2, 5C3, 5D1, arrowhead), while other β-gal+ cells were DAT− (Fig. 5D1). Analyses of the percentages of DAT+/β-gal+ as above indicated that at both 21–30 and 45–65 dpt, β-catenin signaling occurred in >50% of repairing DA neurons (Fig. 5D1). Coupled to the quantitative analysis of Nurr1+/β-gal+ and Nurr1+/DAT+ cells (Fig. 5D2), these results suggest ongoing β-catenin signaling activation both in Nurr1+/DAT− predopaminergic and recovering DAT+ neurons during the neurorescue process (Fig. 5D1, 5D2, 5E).
Aging-Induced Failure to Activate Wnt/β-Catenin Signaling: Effect of Direct β-Catenin Activation In Situ in Nigrostriatal DA Recovery
In aged mice, after MPTP exposure, Xgal/β-gal staining as well as mRNAs for Wnt1, β-catenin, and Nurr1 were all significantly reduced (Fig. 7A1-7A4), in the face of exacerbated GSK-3β transcript levels (Fig. 7C3), supporting dysregulated Wnt1 and β-catenin expression with age. The effect of intracerebral infusion of the GSK-3β antagonist, AR or PBS, in midbrain Aq-PVR 24 hours before MPTP was next addressed. AR infusion in Aq-PVRs resulted in β-catenin signaling activation as revealed by a remarkable increase of β-gal+ (Fig. 7A2, 7A3) nuclei and robust BrdU incorporation (Fig. 7B2, 7B3) in the Aq round two days post-MPTP. Strongly immunolabeled GFAP+ astrocytes and β-catenin+ cells were observed (Fig. 7B1-7B4) all around the Aq, of which a certain number coexpressed BrdU (Fig. 7B1-7B4, 7A3, inset), indicating the ability of direct GSK-3β antagonism to promote proliferation and to induce Aq remodeling as observed in young mice.
At midbrain level, in Wt aged mice, upon MPTP treatment, DA cell body numbers, striatal DAT innervation, and DAT+/Nurr1+ cell numbers were all significantly reduced with no recovery until 45 dpt as compared to younger counterparts (Supporting Information Table S3). In Tg mice, spatio-temporal analyses in midbrain sections of PBS/AR-infused mice were carried out as above to monitor nigrostriatal recovery upon MPTP (Fig. 7C1-7C6 and Supporting Information Fig. S7). Hence, dual staining of midbrain sections with DAT and Nurr1, and time course analysis of the percentages of DAT+/Nurr1+ showed that almost 65% of DAT+ neurons coexpress Nurr1, whereas AR infusion increased this percentage to almost 85% (Fig. 7C1, 7C4). MPTP induced a sharp decrease of DAT+/Nurr1+ neurons for all the experimental period (Fig. 7C2, 7C4), whereas AR infusion induced a less severe decrease of DAT+/Nurr1+ neurons 2 dpt, and with time, AR-infused mice did show progressive increases of SNpc Nurr1+/DAT+ neurons (Fig. 7C4). Likewise, β-gal signal increased with time in DAT+ neurons (Fig. 7C5-7C7), and by 45 dpt, the percentages of DAT+/Nurr1+and DAT+/β-gal+ neurons were comparable to younger counterparts, suggesting that activation of β-catenin signaling in midbrain Aq niche can promote neurorecovery. At striatal level, the markers of DA functionality were significantly increased between 21 and 45 dpt in AR compared to PBS mice (Supporting Information Fig. S7A–S7C). Additionally, motor behavior analyses with the Rotarod indicated that motor coordination deficit of aged MPTP mice was reversed by AR infusion, whereas PBS-infused MPTP mice failed to recover from the motor deficit (Supporting Information Fig. S7D), supporting the requirement of β-catenin signaling activation to promote nigrostriatal recovery of aged mice.
Using in vitro cell culture systems and in vivo models applied to β-catenin reporters, this study uncovered the adult rodent midbrain Aq-ventricular region as a novel Wnt-responsive niche. MPTP-induced DA neuron death promoted a remarkable astrocyte-dependent remodeling and Wnt/β-catenin signaling activation in Nurr1+ postmitotic DA precursors, in surviving and repairing SNpc-DA neurons, a process correlated with a robust time-dependent DA neurorescue. By contrast, the changing properties of midbrain-Aq microenvironment with age impact in DA neurogenic potential of Aq-mNPCs via loss of astrocytic Wnt1 and failure of Wnt/β-catenin signaling activation both inside and outside the niche, in turn associated to failure to recover from MPTP insult. Importantly, aged mNPCs still retain their neurogenic and DA differentiation potential when Wnt/β-catenin signaling is restored via “astrocyte rejuvenation”-induced Wnt1 expression or under Wnt/β-catenin activation regimens, such as GSK-3β antagonism, leading to DA neuron formation. Together, these findings suggest that disruption of a key neurodevelopmental signaling pathway with age may predispose to loss of mDA plasticity via inhibition of Wnt/β-catenin signaling as a prelude for PD development and vulnerability. These results may indicate the potential to restore mDA neuron functionality by activating Wnt/β-catenin signaling in endogenous Wnt-responsive sources, through either pharmacological/cellular approaches aimed at activating/recruiting endogenous progenitors and rescuing the imperiled/diseased DA neurons [44, 47].
A lack of appropriate niche environmental signals is recognized to restrict the neurogenic potential of multipotent progenitors isolated from PD-related brain regions [29, 32, 43-46, 64]. Studying the PVRs throughout the whole ventricular axis of the young intact mouse brain, Hermann et al.  reported that proliferative cell populations are restricted to the PVR-LV. However, in Aq-PVRs, the PSA-Ncam+-like cells observed in vivo strongly correlated with the number of neurosphere-forming cells isolated, in vitro, suggesting that a quiescent subtype of PSA-Ncam+ cells might be the source of mNPCs endowed with neurogenic and DA differentiation potential [45, 46]. Here, the profound morphological remodeling of the Aq-PVRs and β-catenin signaling activation upon DA neuron injury may suggest a high degree of plasticity of this caudal midbrain ventricular region upon lesion of DA nigrostriatal system, while loss of plasticity was observed with age. Interestingly, such upregulated Wnt/β-catenin response in Aq-PVRs, close to the SNpc DA cell bodies, as opposed to the transient β-catenin downregulation observed in SVZ [22, 39] adjacent to striatal DA terminals [29-31], denotes a region-specific response to DA neurodegeneration. The positive correlation between NPC proliferation and β-catenin signaling in the SVZ [22, 39, 58] as opposed to the Aq-PVRs [45, 59] may be linked to factors/mechanisms restraining the proliferative potential within the Aq-PVRs, in vivo, as indicated by the robust proliferation observed after growth factor  or herein with GSK-3β antagonist infusion. Within the SVZ, the intimate contact of NPCs with surrounding glia ( and references herein), coupled to the vast array of growth/neurotrophic factors, neurotransmitters (in particular dopamine), morphogens, cytokines, and Wnt/β-catenin signaling components, contributes to the NPC homeostatic regulation via complex cell-cell interactions and signaling cascades [22, 29-33, 37, 66-69], suggesting that similar mechanisms may be at play here. Different lines of evidences point to age-dependent dysregulation of Wnt signaling as causal factor in aging-induced impairment of both SVZ and SGZ niches. In the SVZ, aging and MPTP antagonize Wnt/β-catenin signaling leading to neurogenic impairment via crosstalk with inflammatory pathways at least in part mediated by upregulation of microglial proinflammatory mediator-induced downregulation of Wnt/β-catenin signaling [22, 33, 68, 69], with potential implications for mNPC neurogenic impairment herein observed [68, 69]. In SGZ, decreased Wnt3 release from aged hippocampal astrocytes regulates the age-associated decline of adult neurogenesis . Importantly, the endogenous Wnt antagonist Dkk1 increases with age, resulting in the suppression of adult neurogenesis and proliferation, whereas Dkk1 knockout mice show increased Wnt signaling, leading to enhanced neurogenesis and improved spatial memory . Our findings add to this scenario the midbrain ventricular region and both pre-DA and DA neurons as candidate targets that respond to SNpc lesion with an intrinsic repair program against injury [57, 60, 68, 69]. In this connection, the activation of astrocyte compartment and increased Wnt/β-catenin signaling herein observed in Nurr1+/TH− precursors of Aq-PVRs appear of specific interest in the light of the pivotal roles of Wnt1 and β-catenin in the development of these neurons [6-10, 13]. β-Catenin maintains adherent junctions and cell polarity of progenitor cells, as well as the integrity of radial glia which provide scaffolds for newly generated DA neurons to migrate on toward their final destinations [61, 62, 70]. Radial glia-like progenitors express Wnt1 , while deletion of Wnt1 induces a severe loss of radial glia-like cells and DA neurons [6, 7, 10, 13, 61, 62]. Wnt signaling via β-catenin promotes the differentiation of Nurr1+/TH− DA precursors. Then, when the degradation of β-catenin is inhibited with specific GSK-3β antagonists, the size of DA neurons increases, through conversion of precursors expressing Nurr1 into TH+ . By contrast, removal of β-catenin in DA progenitors reduces the progression from committed DA progenitors to DA neuron [6, 7, 10, 61, 62]. Of specific mention, the Wnt/β-catenin-dependent  Nurr1 expression is required for the maintenance, the survival, and protection of adult mDA neurons . Conversely, active GSK-3β overexpression is critically involved in neuronal death and depletion of the neurogenic niches, whereas GSK-3β inhibitors can in part mitigate these effects [20, 21, 72-75]. Here, with age, astrocyte dysfunction with loss of endogenous Wnts leads to upregulated GSK-3β levels, engendering (a) disruption of Aq-PVR architecture and function with failure of Wnt/β-catenin/Nurr1 signaling activation and neurogenic impairment; (b) increased vulnerability of mDA neurons likely responsible for reduced of both DA survival and repair.
The observation that β-catenin signaling is active in discrete DA neuronal populations of the intact brain is of interest in the light of its recognized role in neuronal synapse regulation and remodeling as well as in DA neuron maintenance and neuroprotection, as part of a neuron-astrocyte crosstalk [17, 20-22, 25, 68, 69]. Here, stabilization of β-catenin reinstated nigrostriatal plasticity and motor coordination in aged MPTP, whereas ablation of β-catenin in DA neurons alters motor and reward-associated memories and affects striatal mRNA levels for several markers known to regulate synaptic plasticity and DA neurotransmission .
In conclusion, coupled to the increasing evidences on the key role of Wnt/β-catenin signaling cascades in neurodevelopment, neurodegeneration, and regeneration of DA neurons , developing targeted in situ pharmacological interventions/cell manipulations that boost the inherent DA regenerative potential may have implications for restorative and regenerative approaches in PD [3, 4, 9, 12, 13, 68, 69, 77-83].
This study was supported by grants from “Ricerca Corrente” of the Italian Ministry of Healtha, and the Italian Ministry of Healthc, the Ministry of Education and Universityd, the European Research Council (ERC) 2010-StGc, The Evelyn Trustc and The Baschule Charitable Trustc.
B.M.: conception and design, data analysis and interpretation, manuscript writing, and final approval manuscript; S.P.: conception and design and data analysis and interpretation; F.L'E., C.T., N.T., S.C., M.C.M., and M.F.S.: collection and assembly of data.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.