Iron Chelators and Antioxidants Regenerate Neuritic Tree and Nigrostriatal Fibers of MPP+/MPTP-Lesioned Dopaminergic Neurons

Neuronal death in Parkinson’s disease (PD) is often preceded by axodendritic tree retraction and loss of neuronal functionality. The presence of non-functional but live neurons opens therapeutic possibilities to recover functionality before clinical symptoms develop. Considering that iron accumulation and oxidative damage are conditions commonly found in PD, we tested the possible neuritogenic effects of iron chelators and antioxidant agents. We used three commercial chelators: DFO, deferiprone and 2.2’-dypyridyl, and three 8-hydroxyquinoline-based iron chelators: M30, 7MH and 7DH, and we evaluated their effects in vitro using a mesencephalic cell culture treated with the Parkinsonian toxin MPP+ and in vivo using the MPTP mouse model. All chelators tested promoted the emergence of new tyrosine hydroxylase (TH)-positive processes, increased axodendritic tree length and protected cells against lipoperoxidation. Chelator treatment resulted in the generation of processes containing the presynaptic marker synaptophysin. The antioxidants N-acetylcysteine and dymetylthiourea also enhanced axodendritic tree recovery in vitro, an indication that reducing oxidative tone fosters neuritogenesis in MPP+-damaged neurons. Oral administration to mice of the M30 chelator for 14 days after MPTP treatment resulted in increased TH- and GIRK2-positive nigra cells and nigrostriatal fibers. Our results support a role for oral iron chelators as good candidates for the early treatment of PD, at stages of the disease where there is axodendritic tree retraction without neuronal death.


Introduction
A large body of evidence shows that disturbed iron homeostasis, often coupled to mitochondrial dysfunction, plays an important role in the development of common neurodegenerative diseases such as Alzheimer's disease, Parkinson's disease and Huntington's disease [1][2][3].
Committee of the Faculty of Sciences, Universidad de Chile. All surgery was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize animal suffering. Animals were monitored once a day during the overall length of the experiment. Examination included general aspect, possible loss of body weight, spontaneous and behavior upon prodding. No animal died as a result of the treatment with MPTP and later with M30. Because of male-male aggressive behavior, occasionally there were individual aggressions that resulted in injuries. In these cases, the injured animal was terminated according to approved protocols. Mildly affected animals were given the analgesic ketoprofen (2-5 mg/Kg body weight). Typically 3 victims of aggression died in a cohort of 24 animals. The protocol used to determine euthanization was based in a punctuation of the following observations: 1) General aspect: normal 0; uneven fur: 1; ocular or nasal secretions: 2; abnormal posture: 3. 2) Body weight loss: none: 0; less than 10%: 1; between 10 and 20%: 2; over 20%: 3. 3) Spontaneous behavior: normal: 0; small changes: 1; inactivity: 2; very unquiet on no movement: 3. 4) behavior upon prodding: normal: 0; small changes: 1; moderate changes: 2; aggressive or comatose animals: 3. Mice were euthanized when having an accumulate score of 10-12 points. Animals were euthanized by an i.p. injection of sodium pentobarbital. The structures and synthetic strategies for 7-morpholinylmethyl-8-hydroxyquinoline (7MH) and 7-dimethylaminomethyl-8-hydroxyquinoline (7DH) are described in S1 Fig.

Mesencephalic cell culture
Mesencephalic cells were prepared as described [27]. On day 14 of gestation, pregnant Sprague-Dawley rats were exposed to CO 2 followed by laparotomy. The fetuses were collected in cold L-15 medium and the brains were isolated. The mesencephalic dopaminergic region (A8, A9 and A10 dopaminergic nuclei) was dissected and dispersed by repeated pipetting in DMEM/F12 medium containing 0.1% bovine albumin, 5 mg/ml insulin, 30 nM L-thyroxin, 20 nM progesterone, 30 nM sodium selenite, 100 unit/ml penicillin, 100 mg/ml streptomycin and 5% fetal bovine serum. Cells were plated on glass cover slips pre-coated with 1 mg/ml poly-L-lysine, at a density of 55,000 viable cells/cm 2 . On the first day in vitro (DIV1), the medium was changed, and then half of the medium was changed at DIV3, DIV5 and DIV7. Cells were used at DIV7.

Determination of the total length of the axodendritic tree
The total length of the axodendritic tree was determined using the HCA-Vision software (CSIRO, Australia). The total length of the axodendritic tree of DIV7 dopaminergic neurons ranged from 2400 to 2800 μm.

Immunocytochemistry and immunohistochemistry
For immunocytochemistry, cells grown on cover slips were fixed in PBS containing 4% paraformaldehyde, 4% sucrose; after permeation with 0.2% Triton-X-100 in PBS, cultures were incubated overnight at 4°C with the primary antibody followed by 2 h incubation with the secondary antibody. The primary antibodies used in this study were rabbit anti-TH and goat anti-G-protein-regulated inward-rectifier potassium channel 2 (GIRK2) from Sigma-Aldrich, mouse anti-synaptophysin from EMD Millipore and mouse monoclonal anti-4-hydroxynonenal antibody HNE-J2 from Abcam. The secondary antibodies (Invitrogen) were anti-mouse  IgG labeled with Alexa Fluor 488 or 546, anti-rabbit IgG labeled with Alexa Fluor 488 or 546 and anti-goat IgG labeled with Alexa Fluor 633. TOPRO-3 staining (Invitrogen) was used for nuclei identification. Cells were viewed with a Zeiss LSM Meta confocal laser scanning microscope (Carl Zeiss, Göttingen, Germany).
Immunohistochemistry was performed as described [28]. Briefly, mice brains were fixed by transcardiac perfusion of 4% paraformaldehyde in PBS, dissected and post-fixed for 24 h in the same fixation solution. Parasagittal slices of 75 μm width, with a 10-degree tilt, were obtained in a cryostat as described [29]. This technique, named the "Basal Ganglia Slice" technique, allows for sequential cuts along the nigrostriatal pathway. Free-floating slices were permeabilized, blocked for nonspecific binding sites and incubated with primary antibody. Immunolabeling was visualized using Alexa 488-or Alexa-546-tagged secondary antibodies. For detection of the nigrostriatal pathway, the slice with the highest TH-labeling intensity was selected as representative.

Axodendritic tree recovery protocol
Mesencephalic cells at DIV7 were treated with 0.5 μM MPP+ for 24 h in complete culture medium. No indication of TH-positive neurons death was evident as determined by the ratio of TH-positive neurons to overall cells, specified by TOPRO3 staining [26]. MPP+ was removed and cells were subsequently incubated for 48 h in complete medium with or without (Control) an iron chelator at the following concentrations: DFO: 5 μM; DFP: 50 μM; DPD: 10 μM; 7MH: 50 nM; 7DH: 50 nM and M30: 0.25 μM. The concentrations used for the different chelators were determined as optimal on the basis of experiments of dose of chelator vs. axodendritic tree length in the absence of MPP+ treatment (chelator toxicity test) and dose of chelator vs. axodendritic tree recovery after MPP+ treatment (recovery activity test). NAC was used at either 0.5 or 1 mM and DMTU at 1 or 2 mM. A scheme of the axodendritic tree recovery protocol is shown in Fig 1A.

Recovery of nigrostriatal fibers protocol
We used a modification of the protocol described by Gal et al. [30], schematically shown in Fig  1B. In brief, C57BL/6 mice were subjected to sub-acute MPTP intoxication, in which mice were injected i.p. once a day for four consecutive days with either 25 mg/kg body weight of MPTP or with the same volume of saline. Mice were left to recover for four days. Subsequently, two saline-treated groups and two MPTP-treated groups were dosed by gavage once a day for 14 consecutive days either with 2.5 mg/kg of body weight of M30 or with the same volume of saline. All the experiments were performed with three mice per experimental group. For quantification of TH labeling, the sum of all the pixels in fixed areas encompassing either SNc or nigrostriatal fibers was determined using the Quantity One software program (Bio-Rad Inc). This protocol was applied twice with similar results.

Statistical analysis
The Shapiro-Wilk test was used for the determination of normal distribution of replicates. One-way ANOVA was used to test for differences in mean values, and Tukey's post-hoc test was used for comparisons between mean values. A value of P<0.05 was taken as statistically significant.

Results
In order to study putative axodendritic tree regeneration, we first analyzed whether sub-lethal doses of MPP+ promote axodendritic tree shortening in dopaminergic neurons without causing cell death. As illustrated in Fig 2, we found that treatment for 24 h with 0.5 μM MPP+ markedly decreased neuritic tree length. Fig 2A shows a control TH-positive neuron, with a long axon and numerous neurites, and Fig 2B shows two neurons subjected to incubation for 24 h with 0.5 μM MPP+, which produced a 65% decrease in neuritic tree length relative to the control (Fig 2C). Under the assay conditions used (24 h, 0.5 μM MPP+), no cell death was observed. Evaluation of TH-positive cell versus total (TOPRO-positive) cells indicated a ratio of 0.0072 ± 0.0012 for control cultures and 0.0067 ± 0.0010 for MPP+-treated cultures.
Considering that the effect of MPP+ on neuritic tree length most probably depends on the intracellular iron concentration [26], it was of interest to investigate whether commonly used iron chelators could regenerate the axodendritic tree of cultured dopaminergic neurons previously damaged by MPP+. The chelators DFO, DFP and DPD were all effective in inducing neuritogenesis, restoring the axodendritic tree of MPP+-damaged neurons after 48 h of treatment (Fig 3).
The putative functionality of regenerated neurites was evaluated by analyzing the presence of the presynaptic marker synaptophysin, a synaptic vesicle protein that participates in neurotransmitter secretion [31]. Control TH+ neurons presented synaptophysin-positive spots in all new thin neurite processes; synaptophysin was also evident in neurites of cells treated with MPP+, and with MPP+ plus DFO, DFP or DPD (S2 Fig). Because TH-positive neurons represent less than 1% of the total cells in the culture, it was not possible to evaluate the changes in synaptophysin levels by quantitative methods like Western blot analysis.
We tested next the regenerative capacity of three 8-hydroxyquinoline-based iron chelators. Treatment with M30 resulted in significant regeneration of the axodendritic tree of MPP+damaged neurons (MPP+ vs MPP+/M30, P< 0.01) (Fig 4). Significant regeneration activity was observed also upon treatment with 7MH (MPP+ vs MPP+/7MH, P< 0.01) and 7DH (MPP+ vs MPP+/7DH P< 0.01), whereas treatment with 50 nM 7DH showed some toxic activity (Control vs /DH, P<0.05) (Fig 4A). Chelators also reduced lipoperoxidation, as determined by decreased levels of protein-HNE adducts (Fig 4B. Given the critical role played by ROS in PD neurodegeneration [32], we tested the possible mediation of reducing the cellular oxidative tone in the process of neuritogenesis. To this end, we examined the effects of the antioxidants NAC and DMTU on the recovery of the axodendritic tree after MPP+ treatment (Fig 5). Treatment for 48 h with 1 mM NAC or 2 mM DMTU  resulted in almost complete recovery of the axodendritic tree of MPP+-damaged neurons, while treatment with 0.5 mM NAC or 1 mM DMTU afforded significant recoveries to 59% and 62% of control, respectively ( Fig 5A). As expected, antioxidants also reduced MPP+-induced lipoperoxidation, as determined by decreased levels of protein-HNE adducts (Fig 5B).
Next we evaluated in an animal model of PD whether oral treatment with the 8-hydroxyquinoline-based chelator M30 could regenerate nigrostriatal fibers of neurons previously damaged by MPTP, using the protocol outlined in Fig 1B. The relative abundance of fibers and somas was determined by the intensity of TH immunolabeling. The nigrostriatal pathway was detected in 10-degree-tilted parasagittal sections obtained as described in Fig 6A. TH staining Fig 6. Recovery of SNc-striatum fibers by M30 treatment. A) Scheme of the "basal ganglia slice" procedure. Brains were placed on an acrylic base with a 10-degree slope in one of its faces, and were cut into 75 μm parasagittal slices. Image modified from [29]. B) Low-amplification photomontage of the nigrostriatal pathway. Shown is a five-frame reconstitution of the nigrostriatal pathway from a control animal. STR, striatum; SNc: substantia nigra pars compacta. Dashed frames highlight the regions used to quantify TH intensity in fibers and SNc (see below). C) Parasagittal slices. Shown are frames from individual mice subjected to the conditions described in Fig 1B,  showed abundant fibers in the nigrostriatal pathway ( Fig 6B). As compared to the Control (saline/saline) condition, treatment with MPTP reduced the intensity of TH labeling in the fibers and in the cell somas of the SNc (Fig 6C). Post-treatment with M30 (MPTP/M30) largely recovered this reduction. Quantification of the fluorescence intensity of the frames in Fig 6C revealed significant reductions in TH staining after MPTP treatment for both SNc neuronal cell bodies (52.1 ± 6.8%, P<0.05 compared to the Control condition) and fibers (17.3 ± 5.1, P< 0.01 compared to the Control condition) (Fig 6D). As compared with the MPTP condition, post-treatment with M30 (MPTP/M30) induced a significant increase of TH labeling in fibers (P< 0.01) and in SNc somas (P< 0.05).
The above results suggest that M30 given after MPTP treatment results in the regeneration of nigrostriatal fibers and SNc cell somas. Nevertheless, it is possible that after MPTP treatment, fibers and cell somas were not lost but merely expressed decreased levels of TH [33][34][35][36]. Therefore, we analyzed a second axonal marker to test actual fiber and soma regeneration. The K + channels GIRK1 and GIRK2 are important mediators of neuronal excitability in vivo [37]. In the central nervous system, GIRK2 is found in TH-positive neurons of the SNc and in a fraction of neurons of the ventral tegmental area [38]. We found that MPTP treatment diminished GIRK2 immunostaining in the nigrostriatal pathway and SNc neuronal somas, which recovered after M30 treatment (Fig 7A). GIRK immunofluorescence recovered to control levels after M30 treatment, in a way similar to TH immunofluorescence (Fig 7B). Neurons treated with M30 exhibited a tendency to increase in both TH and GIR2 immunofluorescence, but differences with controls did not reach significance. Because of low signal to noise ratio, no quantitative analysis of GIRK-positive nigrostriatal fibers was attempted.

Discussion
Knowledge about factors that influence dopaminergic dendrite tree regeneration is incipient. Treatment with 9-methyl-beta-carboline afforded dendrite regeneration after rotenone or lipopolysaccharide intoxication [39]. Similarly, neurite growth was induced by immunophilin ligand GPI-1046 after 6-hydroxidopamine treatment [40]. In light of our previous report indicating that iron-generated ROS mediate the process of axodendritic tree collapse induced by MPP+ [26], here we studied whether iron chelators promote neurite regeneration both in vitro, using cell cultures of dopaminergic neurons treated with MPP+, and in vivo in an animal PD model of PD.
Incubation for a brief 48-h period with all six iron chelators tested (DFO, DFP, DPD, 7MH, 7DH and M30) effectively restored the axodendritic tree of MPP+-treated neurons, as determined by the emergence of new TH-positive processes. The recovered TH-positive neurites contained synaptophysin, a protein present in neurotransmitter-containing synaptic vesicles. These findings raise the possibility that treatment with iron chelators rescues functional axodendritic trees in MPP+-damaged dopaminergic neurons.
To test recovery of nigrostriatal fibers in vivo, we selected the iron chelator M30 in virtue of its permeability across the blood-brain barrier and its proven properties of neuroprotection and possible neurogenesis [30]. Treatment with M30 reduced iron content in the SN in two animal models of PD [8,30]. In agreement with previous studies showing about 40% death of dopaminergic neurons under similar experimental conditions [41], we found that treatment with MPTP decreased the TH staining intensity of SNc neuronal cell bodies to 52% of control. Treatment with MPTP also resulted in a more pronounced loss of fibers in the nigrostriatal pathway to 17% of control. These results indicate that fibers are more susceptible than the cell body to MPTP-induced damage, and that in the process of neurodegeneration fiber loss precedes actual neuronal death, an observation consistent with a process of axon dieback. Oral supplementation with M30 for 14 days resulted in a striking recovery of TH staining, both in SNc cell bodies and nigrostriatal fibers. The recovery of SNc neuronal cell bodies and fibers as determined by TH staining was associated with recovery of the functional marker GIRK2, suggesting that the newly formed nigrostriatal fibers are functional.
The mechanisms underlying neurite regeneration by iron chelator treatment probably engage multiple factors. Neurite regrowth may result from the inhibition of HIF-prolyl hydroxylase [42][43][44][45], a Fe-and O 2 -dependent enzyme that mediates the degradation of transcription factor HIF-1α. Inhibition of prolyl hydroxylase by iron chelators may raise the level of HIF-1α and the expression of HIF-1α-dependent pro-survival genes, including BDNF, GDNF and VEGF [44]. Indeed, treatment with the chelators M30 or HLA20 induce neuritogenesis in mouse NSC-34 motor neuron cells, which is associated with the production of BDNF and GAP43 [46].
Changes in the oxidative tone may also contribute to neurite rescue. Iron chelators, by reducing the concentration of redox-active iron, may decrease also excessive ROS production in MPP+ treated neurons [47]. In fact, we found that antioxidants NAC and DMTU promoted the recovery of the axodendritic tree of MPP+-damaged neurons to a similar extent than chelators. Hence, it is possible that the reduction of the oxidative tone by iron sequestration underlies one of the mechanisms by which chelators promote axodendritic rescue.
In addition to increased ROS levels, high intracellular calcium concentrations are intimately associated with the progress of neuronal degeneration observed in PD [48,49]. Increasing iron levels in primary hippocampal neurons promotes calcium release from the endoplasmic reticulum [50], causing significant calcium dyshomeostasis and mitochondrial fragmentation [51], two factors that presumably contribute to the impairment of neuronal function produced by iron accumulation. Accordingly, iron chelators may also induce neurogenesis by decreasing toxic intracellular calcium levels to neuritogenesis-permissive concentrations.
In conclusion, this study provides evidence that treatment with iron chelators induces the regeneration of the axodendritic tree collapsed by MPP+ treatment, and restores nigrostriatal fibers and SNc cells in MPTP-treated mice. The design of chelator supplementation strategies leading to the regeneration of the axodendritic tree in damaged neurons may be relevant for the early treatment of PD, at stages of the disease where there is axodendritic tree retraction without neuronal death.