Evidence of oligodendrogliosis in 1‐methyl‐4‐phenyl‐1,2,3,6‐tetrahydropyridine (MPTP)‐induced Parkinsonism

V. Annese, C. Barcia, F. Ros‐Bernal, A. Gómez, C. M. Ros, V. De Pablos, E. Fernández‐Villalba, M.‐E. De Stefano and M. T. Herrero (2013) Neuropathology and Applied Neurobiology39, 132–143


Introduction
Parkinson's disease (PD) is a neurodegenerative disease primarily characterized by the progressive degeneration of neurones of the Substantia Nigra pars compacta (SNpc) and of their fibres projecting to the striatum [1]. Recently, the idea that axonal damage of the SNpc neurones may be a critical step in PD pathology has been put forward [2][3][4]. Degeneration of the nigrostriatal fibres in human PD is irreversible and the possibility of their partial or complete regeneration is one of the main challenges for researchers in neurosciences. The most adopted toxin-based model of PD uses 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), a selective dopaminergic neurotoxin for humans, primates and mice, which induces clinical symptoms in both humans and monkeys mimicking idiopathic PD [5,6]. The neurotoxic action of MPTP begins at the level of the dopaminergic terminals in the striatum from where, once its metabolite N-methyl-4-phenylpyridine-positive (MPP + ) is captured by these terminals [7,8], it induces retrograde degeneration of the dopaminergic axons [9,10] that, if not impeded, leads to a dying-back of the dopaminergic neurones [3,9,10]. In the nervous system, maintenance of healthy axons depends on neurones, as well as on both myelinating and non-myelinating glial cells [11] [oligodendrocytes (OL) and Schwann cells]. Therefore, severely injuring a neurone, or its axon, may induce axonal degeneration, damage to the periaxonal glial cells and, when present, progressive demyelination [12]. In several neurodegenerative diseases, axon damage and demyelination is followed by microglial and astroglial activation, which, in turn, activate OL precursor cells localized in the vicinity by releasing pro-inflammatory cytokines. In other neurodegenerative scenarios, after migration to the injured area, OL precursor cells differentiate into mature OLs by increasing the number and complexity of their processes, up-regulating the expression of myelin proteins and forming new membrane wraps around axons [13][14][15]. Less is known on the behaviour of non-myelinating OLs.
The involvement of microglial cells and astrocytes in the nigrostriatal pathway of PD patients and animals with experimental Parkinsonism is a well-established phenomenon [16][17][18]. However, apart from biochemicalbased reports on the damage to OLs in the striatum after MPTP toxicity in mice (reviewed in [19]) and the description of a-synuclein-positive inclusions in non-myelinating OLs of the degenerating nigrostriatal pathway of PD patients [20], the correlation between OL and Parkinsonism has been scarcely explored [17]. In the present work, we investigated in vivo whether experimental Parkinsonism induced by acute and chronic administration of MPTP in mice and monkeys (Macaca fascicularis), respectively, is associated with OL morphological changes consequent to the primary damage of the distal axon of the dopaminergic neurones and to the secondary alteration of the striatal circuits in which the dopaminergic input is integrated.

Parkinsonian macaques
We used brain sections derived from a colony of chronic parkinsonian macaques (M. fascicularis), previously established and studied in our Primate Unit. Analysed samples were obtained from seven young adult animals: four of the seven monkeys had been treated weekly with low intravenous doses of MPTP (0.3 mg/kg) and progressive motor alterations were assessed using a rating scale ranging from 0 to 25, as previously described [16]. MPTP-intoxicated monkeys showing clear impairment of their motor score were classified as parkinsonian; untreated animals were considered as controls.
Tissue preparation from mice Mice for each of the following experimental categories were used: control, 72 h and 2 weeks after the last MPTP injection. Mice were deeply anaesthetized with an intraperitoneal injection of ketamine (50 mg/kg body weight, b.w.) and xylazine (50 mg/kg b.w.) and perfused transcardially with a Ringer's oxygenated solutions (pH 7.3), followed by a fixative composed by 4% freshly depolymerized paraformaldehyde in phosphate buffer (PB). Brains were dissected and cryoprotected, for 72 h at 4°C, in 30% sucrose in saline until they sank. Free-floating coronal sections (25 mm-thick), spanning the entire midbrain and striatum, were cut at a cryostat and stored at -20°C in a cryoprotectant composed by 0.5 M sucrose diluted in ethylene glycol and 0.2 M PB (pH 7.4) in a 1:1 proportion.
Tissue preparation from macaques Two years after the last MPTP administration, monkeys were sacrificed with a lethal injection of pentobarbital after a pre-anaesthesia with an intramuscular injection of ketamine (8 mg/kg b.w.). Brains were quickly removed, fixed for 3 days in 4% freshly depolymerized paraformaldehyde in 0.1 M PB and cut into 40 mm-thick coronal sections at a sliding microtome (Microm, HM400).
Immunohistochemistry and immunofluorescence Brain sections were processed for TH immunohistochemistry by using the avidin-biotin complex (ABC) procedure. The primary antibody was diluted in 1% normal horse serum, 0.5% Triton X-100 and 0.1% NaN 3 in phosphate-buffered saline (PBS) (pH 7.4), for 48 h at 4°C, under constant shaking. After a rinse in buffer, sections were first incubated with a biotinylated donkey anti-sheep immunoglobulin G (IgG) secondary antibody (diluted 1:500, Jackson Immuno Research Laboratories, Inc., West Grove, PA, USA), and successively with the ABC, diluted 1:100 (ABC kit; Vectastain Elite, Vector Labs, Burlingame, CA, USA). The peroxidase activity associated with the immune complexes was revealed by incubating the sections in 0.25 mg/ml 3,3′-diaminobenzidine (DAB) (Sigma-Aldrich) and 0.03% H 2O2 (Sigma-Aldrich) in PBS, pH 7.4. Some of the mouse brain sections immunolabelled for MBP were also processed with the ABC immunohistochemistry, using as the secondary antibody a biotinylated donkey anti-mouse IgG. The following procedures were similar to those described above. Negative controls were obtained by omitting the primary antibodies. Floating sections were mounted on gelatin-coated slides and dehydrated in an ascending series of ethanol alcohol and xylene before being coverslipped.

Quantitative analysis on immunohistochemical sections
Quantification of immunopositive cells in the SNpc was performed on two coronal sections from each monkey (+1 mm and +2 mm from the anterior commisure according to macaque brain atlas) and three coronal sections (0.86 mm, -046 mm and -3.08 mm from Bregma according to mouse brain atlas) from each mouse. Images of the SNpc and striatum from both hemispheres were acquired, by using a 20¥ objective, at a light microscope (Zeiss Axioplan2) connected to a digital camera (AxioCam HRc, Zeiss). TH-immunopositive (TH + ) cells were counted in each photographic field and expressed as the mean number of cells/mm 2 Ϯ standard error of the mean (SEM). Specifically, each photographic field corresponds to 260 ¥ 370 mm sided rectangles and all cells, including those falling at the border of the photographic rectangle, were counted. TH immunoreactivity in the striatum was evaluated by densitometric analysis of the immunoposi-tive areas and expressed as optical density (arbitrary grey units), calculated in relation to the background staining of each section by using the ImageJ software.
Quantitative analysis on confocal microscope images Brain sections were examined by using an immersion oil 63¥ objective. Each section was scanned in 0.5 mm-thick optical section, the series range being determined by setting the upper and the lower thresholds with the Z/Y Position for Spatial Image Series setting. All quantifications were done blindly. Images were converted into black and white by using the ImageJ software and number and area occupied by the immunopositive cells in each photographic field (three image stacks per animal taken through the whole thickness of the SNpc and striatum) was measured.
Statistical analysis Statistical analyses were performed by using either the Student's t-test (when comparing two groups at a time) or the one-way anova test following the post hoc Dunnet's analysis (when comparing multiple groups against a single reference group). Differences were considered statistically significant for P Յ 0.05. Data were expressed as the mean Ϯ SEM. A Pearson coefficient was used to establish the correlation analysis. The significance of the correlations was determined by using the critical values of the Pearson coefficient table. Differences were considered statistically significant for P < 0.05.

Loss of dopaminergic neurones and axons induced by MPTP treatment in mice is associated with a transient oligodendrogliosis in both SNpc and striatum
We performed detailed confocal and quantitative analysis on sections of striatum and SNpc immunolabelled for MBP, to evaluate changes in cell number, morphology and area of occupancy in MPTP-treated mice compared with control ( Figure 1). Seventy-two hours after MPTP injection, a significant increase in the MBP + cell number/mm 2 (SNpc: control = 205.  Figure 1A,B,D-F, SNpc; Figure 1AЈ,BЈ,DЈ-FЈ, striatum). To date, this is the first quantitative evaluation, which takes into account this set of parameters in MPTP-treated mice and macaques for evaluating alteration of the normal steady-state condition of OLs. OLs were characterized by hypertrophic cell bodies, thickened proximal processes and elevated arborization of the distal ones ( Figure 1A,AЈ, control; Figure 1B,BЈ, 72 h MPTP), which accounted for the noticeable enlargement of their cell body size ( Figure 1E,F, SNpc; Figure 1EЈ,FЈ, striatum). Two weeks after MPTP treatment, confocal ( Figure 1C,CЈ) and quantitative ( Figure 1D Except for the number of the MBP + cells in the SNpc and their average area they occupied in the striatum, all these parameters were also significantly lower than, or equal to, those measured in control animals, suggesting a reversion of this process. These results were confirmed by camera lucida reconstructions of anti-OL-immunopositive cells, revealed with DAB ( Figure 2A1,2,3), performed on sections of control and MPTP-treated mouse striatum. As expected, 72 h after MPTP treatment, a significant decrease in both cell number/mm 2 (control = 228.83 Ϯ 10.39, 72 h MPTP = 98.66 Ϯ 5.84) and intensity of immunolabelling of TH + neurones and neurites in the SNpc, as well as in the intensity of the TH immunopositivity in the striatum, expressed as optical density (control = 22.06 Ϯ 3.13, 72 h MPTP = 9.58 Ϯ 2.89), were observed ( Figure S1). This decrease partially, but significantly, recovered 2 weeks after MPTP injection. OL phenotypical changes inversely correlated with the levels of TH immunoreactivity along the dopaminergic pathway, that is, a higher degree of dopaminergic loss was associated with higher levels of oligodendrogliosis ( Figure 2B). analysis was carried out on sections of macaque striatum immunolabelled for MBP (Figure 3). Chronic Parkinsonian macaques showed a striking and significant increase in the number of MBP + cells in both caudate ( Figure 3B,BЈ, confocal images; Figure 3C, cell count: ctrl = 43.02 Ϯ 3.6 and MPTP = 94 Ϯ 10.3 MBP + cells/ mm 2 ) and putamen ( Figure 3G,GЈ, confocal images; Figure 3H, cell count, ctrl = 224.07 Ϯ 15.9 and MPTP 540.12 Ϯ 41.01 MBP + cells/mm 2 ) with respect to control. This increase was also appreciated by using DAB as the revelation system in place of fluorescence ( Figure 3A1,2 caudate; Figure 3F1,2 putamen). OLs in macaque striatum also showed an apparent increase in their cell body size, which, however, we were technically unable to measure as  cell processes were so intermingled that the software could not identify individual cells. After MPTP treatment, the number of MBP + cells in both caudate and putamen inversely correlated with the number of surviving dopaminergic neurones in the SNpc (Figure 3D,I) and with the densitometric values corresponding to the amount of TH + fibres projecting into both areas of the striatum (Figure 3E,J). Chronic macaque MPTP treatment elicited a marked decrease in TH + cells and neurites in the SNpc, as well as in the dopaminergic projections to the striatum ( Figure S2). However, differently from mice, these parameters were persistent 2 years after the last MPTP injection.

Discussion
Our results demonstrate that loss of dopaminergic neurones and of their axons projecting to the striatum, which is consequent upon MPTP-induced Parkinsonism in mice and macaques, is accompanied by a prominent oligodendrogliosis along the nigrostriatal pathway.
In mice, acute MPTP administration triggers a significant OL response within 72 h. This event is exacerbated in the striatum, possibly connected with the fact that MPP + , the toxic product of MPTP metabolism in glial cells [23], is captured by the dopaminergic axon terminals and retrogradely transported to the cell body, affecting neurone activity and viability [8]. This mechanism of retrograde degeneration is not peculiar to MPTP, as it is also observed in other pathological conditions induced by administration of different neurotoxins [24,25]. The idea that, at least in some of the common neurodegenerative pathologies, deregulation of axonal activity could be primary with respect to cell body damage and neurone death is interesting, if considering that in both human PD and animal model of Parkinsonism, axons are severely affected [26,27]. In fact, it has been hypothesized that axonal degeneration, probably involving OLs, may indeed precede protein aggregation in nerve fibres (Lewy neurites) and neuronal apoptosis, becoming one of the initial causes of neuronal loss [28]. In our experimental model, oligodendrogliosis may reflect different degrees of axonal damage consequent upon MPTP administration, as previously shown by ultrastructural studies [29]. The anti-MBP antibody we used to recognize OLs labels both early and mature OLs. This means that myelinating and nonmyelinating cells can be equally immunolabelled. Based on our correlation studies, showing higher levels of oligo-dendrogliosis when the MPTP-induced loss of the dopaminergic innervation to the striatum is more prominent, we suggest that the first OL response is in part a consequence of the direct lesion of the dopaminergic neurone axons, which are none or scarcely myelinated in different species, including humans [30][31][32][33]. On the other hand, we envision that striatal spiny neurones, deprived of the dopaminergic input, may undergo a series of changes characteristic of the neuronal response to denervation. These imply remodelling of their own axons, as well as alteration of other inputs impinging on their cell body and coming, for instance, from the cortical layers (represented by highly myelinated axons). This hypothesis is supported by numerous studies demonstrating how the striatum responds to fluctuation in dopaminergic signalling and how diseases that alter this signalling change striatal functions (reviewed in [34]). An intense and bidirectional cross-talk between glia and axons has been demonstrated both during development [34,35] and in adulthood [34]. Specifically, while glia maintains axolemmal organization, axonal diameter and neurone health, axons maintain glial differentiation and, when present, myelin integrity [34], by means of axonal contact, diameter, electrical activity and different types of molecular signalling (reviewed in [13]). In the peripheral nervous system, altering this signalling by axonal damage initiates an active Schwann cell reprogramming [36]. A similar mechanism can, therefore, be advocated in the central nervous system, where damage to axonal transport and neurone physiology, characteristic of dying-back pathologies (here mimicked by MPTP administration), influences OL activity and myelinating properties. On the other hand, one should consider that cytokines released by both microglia and astrocytes, involved in a parallel inflammatory process ( [16,17]; V. Annese, C. Barcia A question arising from our results is whether the features we described are due to effective OL morphological changes, or are just a representation of a relocalization of the MBP protein within the cells. It has been reported that mature non-myelinating OLs, identified as MBP + cells, are characterized by a highly ramified morphology, which is reduced when the formation of new wraps around nearby axons, to form myelinated internodes, occurs [37,38]. These data strongly support the hypothesis that, in our experimental paradigm, axonal degeneration induced by MPTP triggers the shift to an MBP + highly ramified nonmyelinating OL phenotype, which will be reverted with the restoration of peri-axonal wraps. To further investigate this hypothesis, we performed a double immunostaining for MBP and CNPase, a protein expressed in myelin-forming cells throughout their lineage, in the striatum of control mice. Interestingly, in contrast to what is reported in the mouse cortex, where CNPase is localized in both cell bodies and processes and MBP is mostly present in cell processes [39] in the mouse striatum MBP immunostaining overlapped that of CNPase ( Figure S3). This result does not allow addressing the idea of a putative change in the distribution of MBP after MPTP treatment. However, we cannot exclude the possibility that, beside OL morphological changes, a general increase in MBP immunoreactivity after the MPTP treatment could also occur.
Two weeks after acute MPTP administration in mice, the partial recovery in TH immunoreactivity and the reduction in the inflammatory response (V. Annese, C. Barcia, M. Di Pentima, F. Ros Bernal, V. De Pablos, E. Fernandez-Villalba, M.T. Herrero and M.E. De Stefano, in preparation) in both SNpc and striatum was concomitant to the reduction of oligodendrogliosis, suggesting that re-establishment of some dopaminergic connectivity reflects a more general restoration of striatal circuitry. This result opens up an important line of investigation concerning OL behaviour, that is, their possible involvement in the remodelling and guidance of distal axons of those dopaminergic neurones that survived MPTP treatment, as well as of other neurone axons projecting to the striatum and affected by the disruption of the nigrostriatal circuit. Recurrent OL activation is a peculiar feature of demyelinating diseases, a potentiality that is, however, decreased during ageing [40]. As idiopathic PD is a pathology characteristic of the elderly, a decrease in axon plasticity and remyelination capabilities with age may indeed further exacerbate progressive neuronal death. To corroborate this hypothesis, future studies on aged mice could give precious information on the development of Parkinsonism, by combining both pathological aspect and physiological ageing.
The results obtained in the macaque model of chronic MPTP administration differ, but do not contradict, those described in mice. In chronic parkinsonian macaques, both number and size of OLs in the striatum were still significantly higher than in control animals 2 years after neurotoxin administration, the oligodendrogliosis being more prominent in macaques suffering more severe dopaminergic depletion. Oligodendrogliosis was maintained throughout the years and no apparent recovery of the dopaminergic system was observed. This discrepancy in the time of occurrence of oligodendrogliosis between mice and macaca, could be partly due to species differences. Moreover, differently from MPTP acutely treated mice, in chronically induced Parkinsonism, the presence of a cascade of glia-mediated inflammatory signals, which perpetuate themselves, has been well established [16]. These signals could sustain a persistent oligodendrogliosis involving both myelinating and non-myelinating OLs, both expressing MBP [41]. A prolonged inflammatory response may also promote a glial scar, creating a hostile environment in which OL activity could be ineffective compared with acute lesions. As discussed for mice, oligodendrogliosis may not be solely associated with degeneration of dopaminergic neurone axons, as spiny neurones and neuronal populations projecting to the striatum may undergo progressive degeneration, therefore triggering and perpetuating oligodendrogliosis.
In conclusion, this study underlines that OLs may play an important role in Parkinsonism, although the function and mechanisms in experimental models and the human disease need to be further investigated. Detailed post mortem studies of the nigrostriatal pathway of PD patients will be crucial for understanding type and extent of axonal modifications occurring during the development of the disease, as well as for developing specific therapeutic strategies aimed at preventing, or minimizing, axonal degeneration in Parkinsonism.

Supporting information
Additional Supporting Information may be found in the online version of this article: Figure S1. TH depletion and partial recovery in both SNpc and striatum following acute MPTP treatment in mice. TH-DAB immunostaining in the SNpc (A-C) and striatum (D-F) of control (A,D) and MPTP-injected mice (B,C,E,F). (A,D) In control animals, TH immunolabelling is intense in both cell bodies and axons of the dopaminergic neurones. (B,E) Seventy-two hours after MPTP treatment, a drastic decrease in the intensity and number of immunopositive neurones and neurites in the SNpc (B), as well as in the intensity of TH immunolabelling in the whole striatum, is observed, especially in the dorsolateral area. (C,F) Two weeks after MPTP injection, a partial recovery of the immunolabelling is observed in both areas (C, SNpc; F, striatum). (G,H) Quantitative analysis confirms that changes observed 72 h and 2 weeks after MPTP injection, with respect to the control (ctrl), in both the number of TH + cells/mm 2 in the SNpc (G) and the optical density of the TH + fibres in the striatum (H) are statistically significant. A partial, but significant, recovery is observed 2 weeks after the MPTP treatment in both SNpc and striatum. n = 4-5 animals/time point; bars represent the mean Ϯ SEM; *P < 0.05, calculated by one-way anova and Duncan test. Figure S2. Persistent TH depletion induced by MPTP chronic treatment in monkeys. (A,B) TH-DAB immunolabelling in the SNpc (A1,2) and striatum (B1,2), both caudate (Cd) and putamen (Put), of control (A1,B1) and parkinsonian (A2,B2) monkeys. A marked decrease in TH+ cells and neurites in the SNpc (circled area in the left-hand drawing in A) and in the dopaminergic projections to the striatum (circled area in the left-hand drawing in B) is observed in parkinsonian macaques 2 years after the last MPTP injection, compared with control. (C) Quantification of the number of TH+ cells in the SNpc and measurement of the optical densities (O.D.) of the TH+-immunopositive fibres in the striatum show that these decreases are statistically significant. Control group: n = 3 monkeys; Parkinsonian group: n = 5 monkeys. Histograms represent the mean Ϯ SEM; *P < 0.05, **P < 0.01 calculated by the Student's t-test. Figure S3. CNPase and MBP co-labelling in the mouse nigrostriatal pathway. Top row of confocal images shows a mouse striatum immunostained for CNPase (red) and