Mild Cerebellar Neurodegeneration of Aged Heterozygous PCD Mice Increases Cell Fusion of Purkinje and Bone Marrow-Derived Cells

Abstract

Bone marrow-derived cells have different plastic properties, especially regarding cell fusion, which increases with time and is prompted by tissue injury. Several recessive mutations, including Purkinje Cell Degeneration, affect the number of Purkinje cells in homozygosis; heterozygous young animals have an apparently normal phenotype but they undergo Purkinje cell loss as they age. Our findings demonstrate that heterozygous pcd mice undergo Purkinje cell loss at postnatal day 300, this slow but steadily progressing cell death starting sooner than has been reported previously and without massive reactive gliosis or inflammation. Here, transplantation of bone marrow stem cells was performed to assess the arrival of bone marrow-derived cells in the cerebellum in these heterozygous mice. Our results reveal that a higher number of cell fusion events occurs in heterozygous animals than in the controls, on days 150 and 300 postnatally. In sum, this study indicates that mild cell death promotes the fusion of bone marrow-derived cells with surviving Purkinje neurons. This phenomenon suggests new therapies for long-lasting neurodegenerative disorders.

Introduction

Several murine mutations lead to cerebellar degeneration, affording mutants such as Lurcher, hot foot, staggerer, nervous, reeler, weaver, and Purkinje Cell Degeneration (23). All these mutants exhibit subsequent impairments of motor capacity and other cerebellar functions. These genetic alterations have been known for over 50 years and, with a few exceptions, most of these mutations elicit the death of Purkinje cells (PCs) (23). The Purkinje Cell Degeneration mutation (pcd) is characterized by an inhibited expression of the Agtpbp1 gene (ATP/GTP binding protein 1; a.k.a. Nna1) (13). Homozygous PCD mice (pcd/pcd) undergo a dramatic PC loss that begins around postnatal day 20 (P20) and rapidly progresses over the next 2 weeks (27). Some cerebellar areas, such as the lobule X, seem to be more resistant to degeneration, but after 2 months virtually all the PCs of the cerebellum have disappeared (34). In addition, pcd/pcd mice undergo the loss of other specific neuronal populations, such as the mitral cells of the olfactory bulb, some thalamic neurons, and photoreceptors (17,34). Moreover, previous reports have demonstrated that aged (17 months) heterozygous pcd mice (+/pcd) are also subject to a decrease in the number of PC in comparison with age-matched controls (+/+) (11).

Cerebellar PCs are selective targets for fusion with bone marrow-derived cells (BMCs), which can reach and integrate with different tissues, as previously demonstrated (2). In addition, a degenerative environment can foster the migration of these BMCs to injured regions, presumably in an attempt to restore the damaged cytoarchitecture and their functions (19,24,29). There is now hope for the development of a new array of therapies for different disorders, including neurodegenerative diseases (8,25,36).

Previous reports employing homozygous pcd/pcd mice were not able to demonstrate whether the neurodegenerative microenvironment of such animals could increase the cell fusion process because of the fast and dramatic PC loss (29). Accordingly, in the present work, we have employed +/pcd mice to analyze the aging-initiated mild cerebellar neurodegeneration and to see if this slow and long-lasting process can increase cell fusion of BMCs with the remaining PCs.

Materials and Methods

Four green fluorescent protein (GFP) transgenic BALBc mice (Mus musculus, L. 1758; Jackson Bar Harbor, ME) were used as donors of BMCs. +/+ and +/pcd mice (eight animals each) of the C57BL/DBA strain (Jackson) were employed as recipients of the bone marrow transplants. Mice were housed at the Animal Facilities of the University of Salamanca at constant temperature and humidity, with a 12/12-h photoperiod, and were fed water and special rodent chow ad libitum (Rodent toxicology diet, B&K Universal G.J., S.L. Molins de Rei, Barcelona).

+/+ and +/pcd mice are phenotypically indistinguishable and in addition, pcd/pcd animals are not suitable for breeding. Therefore, a segregation of both pcd and normal alleles and a subsequent genotyping method were necessary to identify the different animals needed to perform these experiments and to expand the colony. C57BL/6J male mice heterozygous for the pcd^1J mutation were obtained from Jackson Laboratories and were mated with DBA/2J wild-type females. The pcd^1J allele was associated with the genetic background of the C57BL/6J strain, whereas the normal allele was associated with the genetic background of the DBA/2J strain. PCR was performed for the microsatellite regions D13Mit250 and D13Mit283 associated with the locus of the pcd (or +/+) allele. The primers employed were proposed in the web resource of Jackson Laboratories: Mouse Genome Informatics (see: http://www.informatics.jax.org/searches/probe.cgi?38700 and http://www.informatics.jax.org/searches/probe.cgi?41581). PCR products of the D13Mit250 and D13Mit283 regions have different sizes in both the C57BL/6J and DBA/2J strains, allowing the differentiation of the mice genotype by electrophoresis in a 3% agarose gel.

Recipient animals were irradiated at P19 with a 137Cs source for mice, model Gammacell 1000 Elite (MDS Nordion, Ottawa, Canada). This device provides a radiation rate of 243 cGy/min, with an energy of 0.662 MeV. The dose employed was 7.5 Gy, sufficient to ablate the bone marrow of irradiated animals and allow subsequent transplantation (9,29).

Donors were sacrificed by cervical dislocation, their hind paws were removed, and the femurs and tibias were dissected. Bone marrow extraction was performed with an IMDM injection medium (Iscove's modified Dulbecco's medium; Invitrogen, Carlsbad, CA, USA) with an insulin syringe at both epiphyses. The bone marrow wash was filtered through a 70-μm pore size filter (Becton Dickinson, Franklin Lakes, NJ, USA) and centrifuged at 1,500 rpm for 5 min. The supernatant was removed and the pellet was resuspended in lysis buffer (140 mM NH₄Cl, 17 mM Tris-base, pH 7.4) for 5 min to break up the erythrocytes. The reaction was then stopped by adding 45 ml of 0.1 M phosphate-buffered saline (PBS), pH 7.4, to each collection tube. The number of cells in each vial was estimated from an aliquot using a Thoma chamber. The cell suspension was centrifuged again at 1,500 rpm for 5 min and the pellet was resuspended in PBS for transplantation into the recipient animals. Transplants were performed 24 h after the irradiation (at P20), each mouse receiving a single injection in the tail vein containing a suspension of 7.5 million bone marrow cells from the GFP animals. The constitutive expression of GFP in donor mice makes it possible to trace the cells from the reconstituted bone marrow because of their distinctive green fluorescent signal (29).

After irradiation and transplantation, the mice were housed separately for 2 weeks in an insulated rack used specifically for immunodepressed animals. Then, animals were returned to their cages in the original environment until they were sacrificed at P150 or at P300. The mice were deeply anaesthetized with 1 μl/g body weight (b.wt.) of a mixture containing ketamine hydrochloride (Ketolar; Parke-Davis, Barcelona, Spain) and xylazine hydrochloride (Rompun; Bayer, Leverkussen, Germany) at a proportion of 3:4. Then, the animals were perfused intracardially with PBS, followed by 5 ml/g b.wt. of Zamboni's fixative (4% w/v depolymerized paraformaldehyde and 15% v/v saturated picric acid in PBS). After perfusion, the brains were dissected out and rinsed for 2 h with PBS.

Tissue blocks were cryoprotected by immersing them in 30% (w/v) sucrose in PBS, frozen, and sectioned in 40-μm-thick parasagittal slices using a freezing–sliding microtome (Jung SM 2000, Leica Instruments, Nussloch, Germany) attached to a freezing unit (Frigomobil, Leica Instruments).

Immunofluorescence was performed by washing the tissue slices in PBS and incubating them overnight at room temperature under continuous rotary shaking in a medium containing 0.2% (w/v) Triton X-100, 5% (v/v) normal donkey serum, the primary anti-GFP polyclonal goat antibody (1:2,000; Abcam, Cambridge, UK), and the primary anti-calbindin D28k (CB) monoclonal mouse antibody (1:2,000; Swant, Bellinzona, Switzerland) in PBS. An anti-glial fibrillary acidic protein (GFAP) polyclonal mouse IgG (1:1,000; Sigma-Aldrich, St. Louis, MO, USA) to label astrocytes and an anti-ionized calcium binding adaptor molecule 1 (Iba-1) polyclonal rabbit IgG (1:1,000; Wako, Osaka, Japan) to label microglia were also employed. Following this, slices were washed in PBS and incubated in a second medium with cyanine 2 (Cy2)-conjugated donkey anti-goat fluorescent antibody (1:500; Jackson Laboratories, West Grove, PA, USA) and Cy3-conjugated donkey anti-mouse fluorescent antibody (1:500; Jackson) in PBS for 2 h at room temperature under continuous rotary shaking. Finally, sections were rinsed in PBS and mounted with coverslips and a freshly prepared antifade medium.

The terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) technique was employed to detect apoptotic cells, as previously described (33). Tissue slices were washed with PBS and treated for 15 min with 0.2% (v/v) Triton X-100 and 0.1% (p/v) sodium citrate in distilled water. They were washed again with PBS and incubated with TUNEL buffer containing 30 mM Tris-HCl, 140 mM sodium cacodylate, 1 mM CoCl₂, and 0.3% Triton X-100 (v/v) for 30 min. Then, the tissue slices were incubated in a medium with terminal transferase (800 U/ml; Roche Diagnostics, Mannheim, Germany) and biotinylated dUTP (1 μM; Roche Diagnostics) in TUNEL buffer for 2 h. The reaction was terminated by the addition of saline sodium citrate buffer (0.15 M sodium chloride and 0.015 M sodium citrate). Finally, the slices were washed with PBS and developed with a medium containing Cy2-conjugated streptavidin (1:200; Jackson) in PBS. Slide mounting was performed as described above.

PCs were counted with the Neurolucida (V8.23, MicroBrightField, Colchester, VT, USA) and Neuroexplorer programs (V4.70.3, MicroBrightField) following the same procedure as described in previous reports (15,32) to determine the PC density (expressed as number of PCs/mm). Three cerebellar regions were analyzed: the vermis and two levels of the lateral hemispheres at 2.52 and 3.55 mm from the midline. The vermis is the first cerebellar region that undergoes neurodegeneration in pcd/pcd mice (27,33) and contains all the cerebellar lobules. The cerebellar hemispheres were analyzed to determine whether the changes in cell density were also distinguishable in the entire cerebellum. Once homoscedasticity had been checked (Kolmogorov-Smirnov test), Student's t-test was employed to analyze possible differences between the experimental groups at the different survival ages (SPSS 17.0 for Windows, SPSS, Chicago, IL, USA).

All animals were housed, manipulated, and sacrificed in accordance with current European (86/609/ECC and directive 2003/65/CE) and Spanish legislation (RD 1201/2005 and RD 32/2007). The experiments were approved by the Bioethical Committee of the University of Salamanca.

Results

The immunofluorescent technique against CB was performed to detect PCs in the vermis and in two regions of the cerebellar hemispheres. Four mice of each group were employed to perform this analysis. Student's t test showed no differences in the density of PCs between control and +/pcd mice at P150 in any of the three analyzed regions (p > 0.05) (Fig. 1a–e). However, at P300 a significant decrease of the PC density was detected in both the vermis (p < 0.05) (Fig. 1a, f, g) and in the most lateral (3.55 mm) of the studied regions of the hemispheres (p < 0.05) (Fig. 1c). In the region 2.52 mm away from the midline, +/pcd mice showed a decrease in the PC density, although in this case the p-value was exactly 0.05 (Fig. 1b). Summarizing, a ~20% reduction in the general PC density was detected between the control and experimental groups at P300 (Fig. 1), indicating increased cell death in +/pcd mice, not only in the vermis, but throughout the entire cerebellum. Additionally, no differences in the cerebellar volume were detected between experimental groups either at P150 or at P300, which confirms previous findings (11). Therefore, PC loss in heterozygous mice is not severe enough to cause shrinkage of the cerebellum, as it does in the homozygous animals (27).

Figure 1.

Analysis of the Purkinje cell (PC) density. (a–c) Charts showing the density of PCs/mm between control and +/pcd mice at P150 and at P300; analysis of vermis (a) and two regions of the lateral hemispheres (b, 2.52 mm; c, 3.55 mm from midline respectively); +/pcd mice at P300 undergo a general decrease in this cell density (n = 4 for each group. Bars represent mean ± SEM. *p < 0.05; #p < 0.05 for statistically significant differences (Student's t-test). (d–g) Neurolucida drawings of parasagittal sections of mouse vermis at P150 and at P300; each round dot represents one PC. Note the depletion of PCs in the P300 +/pcd mouse.

Immunofluorescent labeling against GFP was also performed in sections of the same mice (four animals each group) to detect neural cells derived from the transplanted bone marrow (Fig. 2). Both microglial (data not shown) and PCs (Fig. 2b–d) derived from the transplanted bone marrow were detected. These GFP-positive PCs also expressed calbindin D28k, suggesting a standard neurochemical pattern, and they were spread throughout the different parts of the cerebellum without any specific location. Accordingly, the total number of GFP-positive PCs in cerebella of both experimental groups at P150 and P300 was analyzed. The number of GFP-positive PCs increased with time from P150 to P300 in both the control and +/pcd mice, indicating a progressive increase in the formation of these neurons derived from fusion with BMCs (Fig. 2a, d). These GFP-positive PCs showed normal morphological features but many had clearly two nuclei (Fig. 2b, c, arrowheads), confirming that a fusion process had occurred. The percentage of binucleated cells of the total GFP-positive PCs was also estimated. At P150 in control animals, 48% of GFP-positive PCs were binucleated, whereas in +/pcd animals it was 50%. At P300, both percentages in control and +/pcd mice appeared slightly reduced: 38.5% and 37.5% respectively. Moreover, on comparing both experimental groups Student's t-test revealed a higher number of GFP-positive PCs in +/pcd mice compared to the controls at both survival times (p < 0.05) (Fig. 2a, c), pointing to a relationship between the decrease in PC numbers in +/pcd mice and an increase in cell fusion events.

Figure 2.

Analysis of the transplant. (a) Chart showing the number of green fluorescent protein (GFP)-positive PCs at P150 and at P300; +/pcd mice show high numbers of such cells at both ages (n = 4 for each group). Bars represent mean ± SEM. *p < 0.05 for statistically significant differences (Student's t-test). (b) GFP-positive PC (green) derived from a fusion event. Note the two nuclei (arrowheads) labeled with DAPI (blue). (c) Z-axis-stack representation of focal planes of a fusion-derived Purkinje cell (GFP, green); the two nuclei (arrowheads) labeled with propidium iodide (PI, red) are in the same focal plane; note their x-axis and y-axis projections (arrows). (d) Image of the cerebellar cortex of a +/pcd mouse. Note two GFP-positive PCs (green). All Purkinje cells are labeled with calbindin (CB, red). (e) Apoptotic PC of a +/pcd animal at P300 labeled with TUNEL (green). All PCs are labeled with calbindin (CB, red). (f–i) Immunostaining against glial fibrillary acidic protein (GFAP) (f, h) and ionized calcium binding adaptor molecule 1 (Iba1) (g, i) to label astrocytes and microglia, respectively. Control animals (f, g) show no differences in comparison with +/pcd mice (h, i) for either marker.

The TUNEL technique was performed to analyze the neurodegenerative process, in particular cell death by apoptosis. A few TUNEL-positive PCs were detected only in the cerebella of P300 +/pcd mice (Fig. 2e). No TUNEL-positive cells were detected in the control animals at either age or in the +/pcd mice at P150 postnatal age.

In order to determine whether reactive gliosis (a glial response to neuronal death) occurred, a double immunofluorescent labeling was performed against Iba1 and GFAP (Fig. 2f–i). However, no differences either in the distribution pattern of microglia and astrocytes or in the morphology of the labeled cells were detected between the experimental groups, either at P150 or at P300.

These results suggest a lack of the massive glial response that characterizes the neurodegeneration of homozygous pcd animals (unpublished data). These results point to an age-delayed degenerative process undergone by +/pcd mice, with few apoptotic events.

Discussion

The PC population in +/pcd mice is known to diminish with age, with a reduction of ~18% in the total PC number at P510 (17 months) (11). Our experiments demonstrate that this cell loss starts sooner, being clearly detectable at P300 (i.e., more than 200 days earlier than previously reported). However, in this previous work it was not specified if the analysis was performed before P510. In addition, the different estimation of the PC quantity (total number vs. density) could vary the age at which the PC loss was detected. On the other hand, our quantifications were performed in both +/+ and +/pcd irradiated and transplanted mice. Previous findings have demonstrated that irradiation does not reduce the number of adult PCs (12). However, we cannot discard possible DNA damage in these cells that can lead to a subsequent cell death in +/pcd animals since pcd/pcd mice show defects in the DNA reparation machinery (5,32). Moreover, although our findings demonstrated that this PC loss was spread through the entire cerebellum, no differences in the cerebellar volume between the experimental groups were detected either at P150 or at P300, corroborating previous results obtained at P510 (11). The relatively mild neurodegeneration of the +/pcd mice is insufficient to affect the overall cerebellar volume.

Apoptotic PCs were scarce in the cerebella of the P300 +/pcd mice. In addition, no reactive gliosis was detected in either experimental group, contrary to the case of homozygous mutants (unpublished data). Our data suggest that PC death in +/pcd mice is rare and occurs as isolated events, since PC loss spreads slowly from the vermis to the cerebellar hemispheres. Gradual PC apoptosis has been observed in other neurodegenerative models involving heterozygous animals with recessive mutations; these include both in vitro (37) and in vivo (10) examples. In these models, cell death was usually detected in old animals, suggesting that it is an age-related effect (6,10,11,30). Thus, gene silencing of the wild-type allele or any other age-related alteration of the genome may provide a possible explanation for the observed apoptotic PC death subpopulation in heterozygous animals (16). As discussed above, radiation could affect the DNA of PCs. Due to the lack of one wild-type allele in +/pcd mice and their possible deficiencies in the DNA reparation (5,32), they could be more susceptible than control animals for radiation damage.

The presence of GFP-positive PCs is mainly accepted to be originated through a process of cell fusion with BMCs (1,2,12,19,24,29,35). This phenomenon increases with time, since progressively more binucleated bone marrow-derived PCs are observed (24,29). Accordingly, the number of detected bone marrow-derived PCs increased throughout time in both +/+ and +/pcd mice. However, not all of these GFP-positive PCs were binucleated. One possible explanation is that it is an artifact due to tissue sectioning, where some cells were split and only one of their two nuclei remained in the analyzed sections. On the other hand, the cytoplasmic fusion could be followed by a nuclear fusion as it has been demonstrated in epithelial cells (31). Finally, it has been suggested that cell fusion between PCs and BMCs could be a transitory process (28).

Our findings demonstrate that +/pcd mice undergo more fusion events than control animals, both at P150 and at P300. Previous studies have reported that damaged tissue can foster both the proliferation and attraction of stem cells, including those derived from bone marrow (3,12,18,19,22,24). Recent findings from our laboratory have revealed that although the severe neurodegeneration in homozygous PCD mice attracts microglial cells derived from the bone marrow (29), PC degeneration is too rapid to allow rescue by fusion with BMCs (29). However, we have demonstrated that the milder activity of the pcd mutation in heterozygosis fosters BMC fusion.

Our experiments show that +/pcd mice undergo more fusion events than control animals both at P150 and at P300, demonstrating the influence of the neurodegenerative microenvironment. It also supports the possibility of employing cell transplantation as a therapeutic tool to alleviate long-term neurodegeneration, since PCs derived from fusion events have two functional nuclei expressing their genetic information (one from the original neuron and the other from the BMC) (35). In fact, the transplantation of BMCs with a healthy or a modified genotype and the ensuing cell fusion has been employed previously to alleviate different neurodegenerative diseases. Previous works employed this methodology, rescuing cells from cell death by promoting the expression of wild-type or other beneficial genes, both in the spinal cord (8,25) and in the encephalon (36), in particular the cerebellum (3,4,7,20,21). In addition, BMCs can release antiapoptotic, neuroprotective or proliferating factors, which provide additional therapeutic benefits to the surrounding tissue (1,21). Previous findings have also demonstrated regular electrophysiological and synaptic properties of such fused cells even in neurodegeneration models, highlighting the possible benefits of cell transplantation (4). Tissue inflammation encourages the fusion of BMCs and PCs (19,24), but it can be also harmful for the surrounding tissue (14,26). However, our results suggest that PC death in +/pcd mice is a slow and long-term process, with only sporadic apoptosis, and undetectable inflammation and reactive gliosis at the time of sacrifice. Moreover, we identified a decrease in PC numbers only at P300 but not at P150, although the fusion in +/pcd mice was increased at both ages. Therefore, a different signal must be responsible for the migration of BMCs to the encephalon in +/pcd mice; a signal that is not related with a glial reaction or massive apoptosis. Accordingly, previous findings have demonstrated that a cerebellar disease associated with a noninflammatory condition, increases the capacity of bone marrow-derived mesenchymal stem cells to fuse with PCs (3). The increase of cell fusion without a process of inflammation should be taken into account to aim different therapeutic cell types (see above) at the damaged regions without the undesirable, reactive gliosis that can further damage the affected structures (14,26).

Previous experiments in our laboratory have revealed that preneurodegeneration phenomena are present before cell death takes place, both in the olfactory bulb and in the cerebellum of homozygous PCD mice (5,32). These phenomena involve transcriptional changes and gene silencing. It is therefore possible that PC loss in +/pcd mice may start before it becomes discernible (e.g., by P150) and that cells in the process of neurodegeneration (which appear healthy) release putative signals that promote cell fusion.

Our findings indicate that the relatively slow degeneration of the +/pcd mutation can promote cell fusion and attract healthy BMCs to degenerating PCs without a perceptible inflammation, this being useful to develop efficient cell therapies. More experiments will be necessary to clarify and improve this attraction so that functional regenerative therapies can be developed for severe neurodegenerative disorders.