HAL Id: hal-02342679
https://hal.archives-ouvertes.fr/hal-02342679
Submitted on 2 Jun 2020HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
AIF Deficiency Induces Early Mitochondrial
Degeneration in Brain Followed by Progressive
Multifocal Neuropathology
Vincent El Ghouzzi, Zsolt Csaba, Paul Olivier, Benjamin Lelouvier, Leslie
Schwendimann, Pascal Dournaud, Catherine Verney, Pierre Rustin, Pierre
Gressens
To cite this version:
Vincent El Ghouzzi, Zsolt Csaba, Paul Olivier, Benjamin Lelouvier, Leslie Schwendimann, et al.. AIF Deficiency Induces Early Mitochondrial Degeneration in Brain Followed by Progressive Multifocal Neuropathology. Journal of Neuropathology and Experimental Neurology, Lippincott, Williams & Wilkins, 2007, 66 (9), pp.838-847. �10.1097/NEN.0b013e318148b822�. �hal-02342679�
AIF deficiency induces early mitochondrial degeneration in brain followed by
progressive multifocal neuropathology
Vincent El Ghouzzi, PhD1,2†§, Zsolt Csaba, MD, PhD3†, Paul Olivier, PhD1,2, Benjamin Lelouvier, MS1,2, Leslie Schwendimann, BS1,2, Pascal Dournaud, PhD1,2, Catherine Verney, PharmD, PhD1,2,
Pierre Rustin, PhD1,2, Pierre Gressens, MD, PhD1,2,4
1 Inserm, U676, Paris, F-75019, France
2 Université Paris 7, Faculté de Médecine Denis Diderot, IFR02 and IFR25, Paris, France
3 Neuroendocrine Research Laboratory, Hungarian Academy of Sciences and Semmelweis University, Department of Human Morphology and Developmental Biology, Budapest, Hungary
4 AP HP, Hôpital Robert Debré, Service de Neurologie Pédiatrique, Paris, France
†
VEG and ZC contributed equally to this work
§ Correspondence to V. El Ghouzzi, Inserm U676, Hôpital Robert-Debré, 48 Boulevard Sérurier, F-75019, Paris,
France. E-mail : elghouzzi@rdebre.inserm.fr, Phone: +331 4003 1973, Fax: +331 4003 1995
Running Head: Multifocal neuropathology in AIF-deficient mice
Author’s contribution:
Vincent El Ghouzzi: Conception of the study, western blots, histochemical studies, interpretation of data, paper writing Zsolt Csaba: Ultrastructural studies, mitochondrial degeneration analyses, interpretation of data
Paul Olivier: Cell counting and statistics Benjamin Lelouvier: Immunohistochemistry
Leslie Schwendimann: Phenotype study, genotyping and FluoroJade staining Pascal Dournaud: Ultrastructural studies and interpretation
Catherine Verney: Anatomical expertise and interpretation of data relevance Pierre Rustin: Interpretation of data
Pierre Gressens: Conception of the study, interpretation of data
Sources of financial support:
This study was supported by the Institut National pour la Santé et la Recherche Médicale (INSERM), the Centre National de la Recherche Scientifique (CNRS), the Université Paris7, the Association Française contre les Myopathies (Project No.11639), the Hungarian National Research Fund (OTKA F046542 to Z.C.), and the Fondation Grace de Monaco. Z.C. is a recipient of Bolyai Fellowship.
2
Abstract
Apoptosis-inducing factor (AIF) deficiency compromises oxidative phosphorylation (OXPHOS) and Harlequin mice, where AIF is downregulated, develop a severe mitochondrial complex I (CI) deficiency, suggesting that Harlequin mice may represent a natural model of the most frequent OXPHOS disorders. However, the brain phenotype reported to date specifically involves the cerebellum whereas human CI deficiencies often manifest as complex multifocal neuropathologies. To evaluate whether this model can be used as a valuable tool to study CI-deficient disorders, the whole brain of Harlequin mice was investigated during the course of the disease. Neurodegeneration was not restricted to the cerebellum but progressively affected thalamic, striatal and cortical regions as well. Strong astroglial and microglial activation with extensive vascular proliferation was observed by 4 months of age in thalamic, striatal and cerebellar nuclei associated with somatosensory-motor pathways. At 2 months of age, degenerating mitochondria were observed in most cells in these structures, even in non-degenerating neurons, a finding that indicates mitochondrial injury to be rather a cause than an effect of neuronal cell death. Thus, AIF deficiency induces early mitochondrial degeneration followed by progressive multifocal neuropathology, a phenotype much wider than previously described, resembling histopathological features of devastating human neurodegenerative mitochondriopathies associated with CI deficiency.
Key words: Apoptosis-inducing factor, Progressive neurodegeneration, Astrogliosis, Complex I
3
Introduction
Genetic defects of the mitochondrial respiratory chain components impair oxidative phosphorylation (OXPHOS), a pathomechanism recognized increasingly as an important cause of neurodegeneration (1). Complex I (NADH:ubiquinone-oxidoreductase, CI) deficiency represents the most common enzyme defect among OXPHOS disorders. Due to its high degree of complexity in structure, function and regulation, treating CI-associated disorders faces the arduous challenge of high phenotype variability (2). In particular, the neurological features of CI-deficiency diseases are associated with a wide range of pathological changes which may be either generalized or restricted to specific regions of the brain. However, some neuropathological features like neuronal loss, focal necrosis and gliosis are common to most mitochondrial encephalopathies and cerebellar, striatal and thalamic involvements are frequently observed in these disorders (3).
The Harlequin (Hq) mouse bears a spontaneous X chromosome-linked murine proviral insertion in the gene encoding the mitochondrial flavoprotein apoptosis-inducing factor (AIF) (4), a key component in caspase-independent apoptosis (5-7). As a consequence, AIF expression is reduced to 20% of normal levels (4). The recent discovery that AIF deficiency causes CI dysfunction and thereby compromises OXPHOS, has conferred a novel function to AIF within mitochondria (8). Even though the molecular mechanisms through which AIF modulates CI assembly and/or stability are unknown, this non-apoptotic function for AIF in sustaining mitochondrial respiration has pointed out the Hq mouse as a promising model for the study of mitochondrial diseases.
Hq mice develop progressive cerebellar ataxia and blindness as a major trait, and localized neurodegeneration has been reported in both the cerebellum and the optic nerve (4), two features frequently observed in patients with mitochondrial diseases (9). However, CI deficiencies often manifest as a more complex and variable multifocal neuropathology. The present study aimed to (i) investigate the neuroanatomical features of Hq mice during the course of the disease, (ii) delineate the extent of the neuropathological changes in the whole brain of Hq mice, and (iii) evaluate whether this natural model can be used as a valuable tool to study CI-deficient disorders in vivo.
4
Material and Methods Animals
Hq mice were obtained from the Jackson Laboratory (strain B6CBACaAw-J/A-Pdc8Hq/J) and housed at 21°C, 60% humidity with a 12-hour-light/dark cycle and free access to food and water. Experimental protocols were approved by the institutional review committee, meet the INSERM guidelines, and were carried out in accordance with the Guide for the Care and use of Laboratory Animals as adopted by the U.S. National Institutes of Health.
Genotyping and sex identification
Hemizygous males were obtained by mating heterozygous females and control males. Genomic DNA was purified from tail samples, using MasterPure Kit (Epicentre, Madison, WI, USA). For sex determination, PCR amplification was performed using primers specific to the Sry gene (5’-TGGGACTGGTGACAATTGTC-3’ and 5’-GAGTACAGGTGTGCAGCTCT-3’). The affection status (wild-type or mutant) of Aif alleles was determined by multiplex PCR using primers located in the first intron of Aif: 5’-AGTGTCCAGTCAAAGTACCGG-3’ (forward), 5’-CTATGCCCTTCTCCATGTAGTT-3’ (backward) and in the proviral sequence (5’-CCAGAAACTGTCTCAAGGTTCC-3’ as previously described (4).
Immunohistochemistry
Wild-type and hemizygous males were sacrificed at 2, 4 and 6 months of age by decerebration (n=5 per group). Brains were fixed in formalin and embedded in paraffin. Sagittal plane was selected in order to observe the caudo-rostral extension of the brain including the cerebellum. Sections of 10µm were mounted on gelatine-coated slides and reacted with anti-GFAP (1:500, Z0334-Dako, Glostrup, Denmark), anti-NeuN (1:500, MAB377-Chemicon, Billerica, MA, USA), anti-S100β (1:500, Swant, Bellinzona, Switzerland), anti-Iba1 (1:500, 019-19741-Wako, Neuss, Germany), anti-Glut1 (1:1000, 4670-1601-Biogenesis, Poole, UK) or FluoroJade B (AG310-Chemicon, Billerica, MA). Detection of primary antibodies was performed with avidin–biotin horseradishperoxidase kits (Vector, Burlingame,
5 CA, USA) according to manufacturer'sinstructions. Control and experimental groups were processed in parallel in all immunohistochemical procedures.
Western blot analysis
Tissue samples were dissected out from hemispheres of five wild-type and five Hq mice, pooled, sonicated in ice-cold lysis buffer (50mM Tris-HCl,pH7.5, 150mM NaCl, 0.5% NP40, 0.25% sodium deoxycholate, protease inhibitors) and centrifuged. Protein concentrations were determined according to Bradford (10). Lysates were denatured, separated by SDS-PAGE and transferred to PVDF membranes (Millipore, Billerica, MA, USA). Membranes were probed with anti-AIF (sc-9416-Santa-Cruz, CA, USA), anti-p30 and 39kDa (Molecular probes, Carlsbad, CA, USA), anti-GFAP (CBL411-abcys, Paris, France), anti-S100β (Swant, Bellinzona, Switzerland), or anti-Actin (MAB1501-Chemicon, Billerica, MA, USA) and signal was detected by enhanced chemiluminescence.
Electron Microscopy analysis
Two Hq mutant and wild-type mice at 2, 4 and 6 months of age were perfused transaortically with 20ml saline followed by 100ml of 2% paraformaldehyde with 2% glutaraldehyde at 4°C. Brains were post-fixed in 2% paraformaldehyde at 4°C. Sections were cut on a vibratome at 50µm thickness and post-fixed in 1% glutaraldehyde for 10min, treated with 1% osmium for 10min, dehydrated in an ascending series of ethanol, including 1% uranyl-acetate in 70% ethanol. Sections were then treated with propylene oxyde, equilibrated overnight in Durcupan ACM (Fluka,Buchs, Switzerland), and flat-embedded on glass slides for 48h at 60°C. Areas of interest were cut out, re-flat-embedded in Durcupan blocks and ultrathin sections were cut on a Reichert ultramicrotome. Ultrathin sections were collected on formvar-coated single-slot grids, stained with lead citrate, and examined with a Hitachi H-7600 electron microscope. Mitochondria were counted in the perikarya of granule and Purkinje cells of the cerebellum (lobules VI-IX), large projection neurons of the thalamus ventral posterolateral nucleus (VPL), and large neurons of the caudate-putamen which were embedded in striatal fibre tracts. A total
6 number of 4513 mitochondria were counted in Hq mice, and pathological changes in mitochondrial ultrastructure (prominent swelling, vacuolization, disruption of crystal architecture) were recorded.
Quantification of immunoreactive cells
Quantitative analysis of immunoreactive cells was performed in the cerebellum (posterior lobules), the thalamus (VPL and ventro-lateral nuclei (VL)), the caudate-putamen (dorso-medial area of the striatum) of 2-, 4- and 6-month-old Hq mutant mice and wild-type littermates. A blinded procedure was used to count labeled cells at x400 magnification in 0.065 mm² (0.01625 mm² for NeuN in the cerebellum) of 4 sections from at least 4 animals in each group.
Statistical Analysis
All data are reported as mean +/- S.E.M. The data were evaluated by variance analysis with age (from 2 to 6 months) and group (Hq vs wild-type) as factors. Bonferroni post-test was used for multiple pairwise comparisons (Prism4.0, SAS Institute Inc).
Results
We studied mice at 2, 4 and 6 months of age, three stages that corresponded to mice without ataxia, mice with first signs of ataxia and ataxic mice, respectively. The main structures of the telencephalon, the diencephalon and the mesencephalon were examined. For all following data, we report on anatomical regions where significant abnormalities were detected, i.e. the cerebellum, the thalamus, the cerebral cortex, and the striatum.
Reduced CI expression in AIF-deficient brain areas from ataxic Hq mice
Following fine dissection of cerebellum, thalamus, cortex and striatum from 6-month-old animals, the total amount of AIF protein analyzed by western blot in each tissue was found reduced by a factor of 2 to 6 in Hq mice as compared to age-matched wild-type littermates (Fig.1). In parallel, a strong
7 reduction in the abundance of the respiratory chain p30 and p39 subunits of CI (ranging from 50% to 75%) was detected in all four areas analyzed. Thus, AIF-deficient cerebellum, thalamus, cortex, and striatum from ataxic animals all display a reduced CI expression.
Progressive astrogliosis in Hq mutant mice
Astrogliosis often occurs in response to various brain injuries and is frequently encountered in human mitochondriopathies. Therefore, we asked whether it could also occur in Hq mice. Brains at 2, 4 and 6 months of age were used for immunohistochemistry against the glial fibrillary acidic protein (GFAP), a major hallmark of the astroglial response. The thalamus of wild-type animals poorly expressed GFAP at all ages observed whereas that of Hq mutant animals showed an increasing number of GFAP-immunoreactive (IR) astrocytes from 2 to 6 months of age (Fig.2A). Strongly GFAP-IR cells were characterized by a large cell body and processes typical of reactive astroglia. A unique pattern of dense GFAP-IR cells distribution was detected in Hq mice. The ventral thalamic areas including the motor-related nuclei (VA, VL) and the somatosensory nucleus (VPL) showed strong GFAP staining, spreading progressively throughout the whole thalamus but sparing a small dorsal region even at 6 months (Fig.2A). Examination of the thalamus at medial, intermediate and lateral sagittal levels indicated that the motor, somatosensory and visual areas were highly GFAP-positive (Fig.2B and 3A) whereas medial (rather limbic) and dorsal (rather associative) areas appeared more spared.Consistent with this, cell counting in the motor-related VL nucleus of the thalamus showed a progressive increase of the number of GFAP-IR cells by a factor of 5 to 10 between 2 and 6 months of age (Fig.3B). Similar results were obtained when counting GFAP-IR cells in the VPL nucleus (not shown). Therefore, progressive astrogliosis in the thalamus of Hq mutant mice selectively affects motor and somatosensory nuclei.
The cerebellum and the striatum also displayed a dense GFAP positive astrogliosis at 2 months increasing strongly at 4 and 6 months of age. In the cerebellum, the astroglial reaction was massive as a high number of enlarged GFAP-IR cell bodies filled the granule cell layer and numerous processes of Bergmann glial cells were GFAP-positive as well (Fig.4A). At 4 months of age, the astroglial
8 reaction was visible in the deep nuclei of the cerebellum to an extent similar to that measured in cerebellar lobules (not shown). In the striatum, astrogliosis occurred within and around the fibre tracts and was restricted to its sensorimotor territory, i.e. the caudate-putamen (dorso-medial area, Fig.5A) whereas the limbic territories, which are rather ventral, were not stained. Counting of GFAP-IR cells in the granular layer of cerebellar lobules and the caudate-putamen confirmed that astrogliosis was progressive in both regions (Fig.4B and 5B). Moreover, the globus pallidus - which gives rise to the basal ganglia afferences toward the motor thalamus - was also strongly positive for GFAP in 4-month-old Hq mice (not shown).
Astrogliosis was less obvious is the cerebral cortex and varied between animals. Globally, Hq mice displayed more GFAP-IR cells than their wild-type counterparts in the deep cortical layers of the somatosensory cortical areas (not shown).
Since astrogliosis could result from either an astrocyte proliferation or an increased GFAP expression, the S100b protein was used as an alternative marker to assess astrocytic density. Interestingly, S100b staining and counting of S100b-IR cells gave similar results in Hq and wild-type mice in all tissues tested, suggesting that the strong GFAP reactivity resulted from an increased expression of GFAP without astrocytic proliferation (Fig.3-5). Consistent with this, semi-quantitative measurements of GFAP and S100b expression levels using western blot on crude protein extracts showed a strong increase of GFAP expression in the cerebellum, the thalamus and the striatum of Hq mutant mice as compared to wild-type, whereas that of S100b showed no difference between the groups (Fig.1). Therefore, the astroglial reaction observed in Hq brains is not associated with astrocytes proliferation.
Since astrocytes are known to contribute actively to metabolic support of neurons through nutrients supply such as glucose, and vasomodulation of blood flow, we asked whether the vascular network was changed in the cerebellum, the thalamus, the cortex, and the striatum of Hq mutant mice. Using the Glucose Transporter-1 (Glut-1) as a metabolic marker specific for glial cells (11), we found that the microvessels were much thicker, longer, more numerous, and highly splitting in the somatosensory and motor-related thalamic nuclei (but not in limbic areas) of Hq animals compared to
9 wild-types (Fig.3). This was also obvious in the cerebellum where both the granular and the molecular layers displayed an enlarged vascular network (Fig.4) and to a lesser extend in the caudate-putamen (Fig.5) and in the cerebral cortex (not shown). Strong Glut-1-positive staining was also observed in the globus pallidus and in the deep cerebellar nuclei, in the same areas where dense astrogliosis had been detected (not shown).
Microglial reaction in Hq mutant mice
Microglial cells, the tissue macrophage of the central nervous system (CNS), rapidly activate in response to even minor pathological changes in the brain and their activation is among the earliest signs of CNS dysfunction (12, 13). Distribution of microglia-macrophage was analyzed in Hq mice, using Iba1, a macrophage-specific calcium-binding protein which labels resident and activated microglia (14). While wild-type animals displayed a dispersed labelling of small Iba1-IR resident microglia in the different areas examined at any age, Hq mutant mice presented with clusters of large Iba1-positive activated microglia. As for astroglia, activated microglial cells were numerous in the cerebellar granule cell layer (Fig.4A), as well as in the fibre tracts of the caudate-putamen (Fig.5A) and in the somatosensory and motor-related thalamic nuclei (Fig.3A).
Neuronal cell death in Hq mutant mice
As previously reported, Hq mutant mice have progressive degeneration of terminally differentiated cerebellar neurons, the granule cells being affected primarily (4). To visualize neuron loss in cerebellar structures and other brain areas in Hq mice, sections were stained with the selective neuron marker NeuN and with Fluoro-Jade B, a high affinity fluorescent marker specific for neuronal degeneration during acute neuronal distress. Fluoro-Jade B allows detection of neurons which are undergoing the degenerative process and stains both neurites and cell bodies (15). At 2 months of age, although the cerebella from wild-type and Hq mice were undistinguishable in size, structure and neuron number, a restricted number of granule cells were Fluoro-Jade-positive in most mutant animals, suggesting that
10 the degenerative process was already ongoing at 2 months (not shown). No neuron loss was detected in other structures at this stage. At 4 months of age, as cerebellar granule cells degenerated massively (Fig.4A), numerous Fluoro-Jade B-positive cell bodies and neurites were detected in the deep cerebellar nuclei (Fig.6), from where motor outputs leave the cerebellum towards the thalamus. At the same age, neurite degeneration was detected in the fibre tracts of the caudate-putamen (Fig.5A), in the globus pallidus (Fig.6) and in the motor-related thalamic nuclei (Fig.3A). However, we did not detect any degenerating cell bodies in these structures after 9 months of age (not shown), and counting of NeuN-IR cells in these areas failed to reveal significant neuronal loss, suggesting that resident neurons were more resistant to AIF/CI deficiency in the basal ganglia and in the thalamus than in the cerebellum. Neurodegenerative cells and processes were also detected in the frontal and the parietal cortex, as restricted Fluoro-Jade B-positive areas in the deep cortical layers, which matched with the somatosensory and motor cortical regions, respectively (Fig.6).
By 6 months of age, more than half of the neurons had died in the granular layer of the cerebellum and Fluoro-Jade B staining was hardly detectable in most mutant mice. However, sustained neurite degeneration was still visible in the deep nuclei, suggesting an ongoing retrograde degenerative process. In the caudate-putamen, neuritic degeneration was more pronounced than at 4 months while in motor-related thalamic nuclei it appeared similar (not shown).
Mitochondria degeneration in Hq mutant mice
Consistent with the progressive nature of the Hq phenotype, the decrease of CI levels and activity are progressive (8, 16), suggesting that not all mitochondria are affected at the same time. Mitochondrial morphology was therefore analyzed in wild-type and Hq mice at 2, 4 and 6 months of age, using electron microscopy. While no structural abnormalities of mitochondria were observed in any of the studied regions in wild-type mice, large and swollen mitochondria were observed in most neurons from the cerebellum, the somatosensory thalamic nucleus VPL and the caudate-putamen in Hq animals (Fig.7). Enlarged mitochondria were degenerated almost completely with partial or total disappearance of the crests. Surprisingly, in both the cerebellum and the thalamus, the rate of neurons
11 having at least one affected mitochondrion was extremely high, ranging between 70 and 100%, even at 2 months, when no neurodegeneration was detectable except in a few cerebellar granule cells (Table I). In particular, Purkinje cells, which do not die before 7-8 months of age, all displayed degenerated mitochondria. Interestingly, the overall rate of degenerating mitochondria in cerebellar granule cells reached 41% at 2 months of age and then shifted to 62% and 82% at 4 and 6 months, respectively, whereas that of Purkinje cells was of 3-4% at 2 and 4 months and reached only 13% at 6 months. Similar rates of degenerating mitochondria were found in the thalamic somatosensory nucleus VPL (Table I). In the caudate-putamen, where no abnormal mitochondrion was detected before 6 months of age, most neurons finally harboured abnormal mitochondria and 23% of mitochondria showed signs of degeneration by 6 months of age. These results demonstrate that the percentage of degenerating mitochondria is well correlated to the neuronal degeneration process in the Hq model and that the degeneration of mitochondria precedes neuronal cell death.
Discussion
The Hq mouse which has a reduced (but not abolished) AIF expression (4), spontaneously develops a progressive and severe neuronal degeneration associated with mitochondrial CI deficiency (8). Hence, the present study aimed to delineate the neuropathological features in Hq mutant mice as to evaluate whether this natural model is close to human CI-deficient disorders and can serve as a valuable tool to study CI-deficient disorders in vivo.
By investigating the whole brain of Hq mice, we have shown that neurodegeneration is not restricted to the cerebellum but progressively affects thalamic, striatal and cortical regions as well. We found a reduced expression of CI subunits in all these brain areas. As previously shown in whole brain extracts, the decrease of CI levels and activity are progressive and tightly correlated (8, 16). Our findings that most neurons contained at least one swollen degenerating mitochondrion, even at 2 months of age in areas where no neuron loss is observed, further suggest that mitochondrial injury
12 precedes the onset of neuronal degeneration. Hence, mitochondrial degeneration is rather a cause than a consequence of the degenerating process, even if we can not exclude that this latter might enhance in turn mitochondrial injury. Consistent with our findings, mitochondrial dysfunction and oxidative stress occur early in all major neurodegenerative diseases. In Alzheimer’s disease, oxidative damage is the earliest event observed, before the onset of significant plaque pathology (17, 18). In Cu/Zn-superoxide dismutase mutant mice, a model of amyotrophic lateral sclerosis (ALS) where progressive toxicity of the SOD1-G93A mutation poisons motor neurons by impairing mitochondrial energy metabolism (19), mitochondrial calcium capacity impairment occurs long before the onset of motor symptoms (20). In Huntington’s disease as well, mitochondrial calcium abnormalities have been characterized as early events triggered by mutant huntingtin (21). Moreover, there is now strong evidence that these dysfunctions of mitochondria have a causal role in the pathogenesis of most neurodegenerative diseases (1). Thus, in the Hq mutant mice, AIF deficiency in brain induces a typical mitochondrial dysfunction characterized by early mitochondrial degeneration.
Investigating the whole brain of Hq mice using neuronal, glial and cell death markers, we observed a massive degeneration of cerebellar granule cells, as previously reported (4), but found it started earlier, as the degenerative process was already ongoing at 2 months. Interestingly, deep cerebellar neurons which send motor outputs outside the cerebellum were also dying from 4 months of age. Degenerating neurites were also detected at the same age within the thalamo-cortical fibres of the caudate-putamen and in the somatosensory and motor areas of both the thalamus and the cortex. Beside neuronal cell death, the most evident pathological change was the massive glial reaction that occurred specifically in different areas of the brain, namely the ventro-lateral and ventro-postero-lateral thalamic nuclei, the cerebellum as a whole, and the caudate-putamen. Of interest, these regions are functionally involved in the two main sensorimotor loops that relay efferent information in the CNS, i.e. the striato-thalamo-cortical pathway and the cerebello-thalamo-cortical pathway. This suggests a specific impairment of the somatosensory-motor capacities in the Hq mouse. Moreover, strong astrogliosis also occurred both in the globus pallidus and in the deep cerebellar nuclei which serve as intermediates to relay the motor information between the thalamus and the caudate-putamen
13 or the cerebellum, respectively. Pathological abnormalities in Hq mice thus affect more brain areas than previously thought and involve both the somatosensory and the motor systems in the CNS. Taken together, these data show that AIF downregulation in the Hq mouse results in a progressive CI deficiency-induced multifocal neuropathology.
Although degenerating neurites were detected by 4-months of age within the fibre tracts of the sensorimotor territory of the basal ganglia and in the somatosensory and motor-related thalamic areas, we did not find any dying cell bodies in these regions up to an age of 9 months. This suggests that, unlike cerebellar granule cells, other neurons are less vulnerable to AIF-induced oxidative stress. Our findings that the number of swollen degenerating mitochondria is much higher in cerebellar granule cells and well correlated in time with the degeneration process argues in favour of a threshold rate of active mitochondria below which the degeneration process is triggered. Interestingly, the hypothesis of a threshold level of the neuronal damage has been proposed previously in the pathogenesis of ALS, where progressive toxicity of SOD1 mutation affects motor neurons (22). Since both the thalamus and the striatum of Hq mice exhibited a reduction in the abundance of the CI subunits and activity similar to that of the cerebellum, we hypothesize that cerebellar granule cells display a greater susceptibility than other cells to CI deficiency. Of interest, a number of studies have shown that brain cells differ in their susceptibility to the same oxidative stress insult (23-25). Especially, a differential regulation of CI activity between glia (up-regulation) and neurons (dramatic decrease) has been observed when subjected to oxidative stress induced by glutathione depletion (26). Differences in susceptibility to respiratory chain deficiency might also occur between subtypes of neurons. Alternatively, astrogliosis may influence differently the neurodegeneration process in the cerebellar granule cells as compared to that in other brain areas. Similar to what we report here for the Hq mice, astrogliosis occurs as one of the earliest pathological event in the mouse model of ALS (27, 28). A close temporal correlation between the onset of neuronal degeneration and the beginning of astrogliosis has been evidenced in this model, indicating that astrogliosis occurs in response to the initiation of neuronal degeneration and actively participates in the neuronal degeneration process (22). Although many functions of astrocytes are known, the significance of astrogliosis is still debated (29). Depending upon the
14 pathophysiological context it might be either beneficial or detrimental for neuronal survival (30). In the thalamus and the basal ganglia of Hq mutant mice, the glial reaction starts long before (2 months) the first degenerating neurites can be detected (4 months). The question is therefore raised of whether gliosis could limit neuronal cell death in these areas, thus avoiding the early onset of degeneration as observed in the cerebellum. Of interest, histopathology features of Leigh and MELAS syndromes, two encephalomyelopathies often associated with CI deficiency (31, 32), involve vascular proliferation, gliosis with relative sparing of neurons in both the basal ganglia (caudate-putamen) and the thalamus, i.e., in the same areas as those involved in Hq mice (3, 33).
Close to what we described for the Hq model, the neurological features of patients with CI deficiency often include neurodegenerative processes with neuronal loss and gliosis (3). Multifocal patterns of neuropathology are frequently encountered in these patients and progressive pathological changes in the cerebellum, the caudate-putamen and the thalamus have been reported recurrently in overall clinical pictures of CI-deficient syndromes (34-41). Analogies between neuropathological findings in human CI-related disorders and those we described in the Hq mouse make this animal model valuable to investigate neurodegenerative mitochondriopathies associated with CI deficiency in vivo.
Acknowledgement
This study was supported by the Institut National pour la Santé et la Recherche Médicale (INSERM), the Centre National de la Recherche Scientifique (CNRS), the Université Paris7, the Association Française contre les Myopathies (Project No.11639), the Hungarian National Research Fund (OTKA F046542 to Z.C.), and the Fondation Grace de Monaco. Z.C. is a recipient of Bolyai Fellowship. The authors thank Cécile Martel for her support as well as Kristy Ilinsky for advice and fruitful discussions during preparation of the manuscript.
15
References
1. Lin MT, Beal MF. Mitochondrial dysfunction and oxidative stress in neurodegenerative diseases. Nature. 2006;443:787-795
2. Janssen RJ, Nijtmans LG, van den Heuvel LP, Smeitink JA. Mitochondrial complex I: structure, function and pathology. J Inherit Metab Dis. 2006;29:499-515
3. Brown GK, Squier MV. Neuropathology and pathogenesis of mitochondrial diseases. J Inherit Metab Dis. 1996;19:553-572
4. Klein JA, Longo-Guess CM, Rossmann MP et al. The harlequin mouse mutation downregulates apoptosis-inducing factor. Nature. 2002;419:367-374
5. Cande C, Vahsen N, Garrido C, Kroemer G. Apoptosis-inducing factor (AIF): caspase-independent after all. Cell Death Differ. 2004;11:591-595
6. Cregan SP, Dawson VL, Slack RS. Role of AIF in caspase-dependent and caspase-independent cell death. Oncogene. 2004;23:2785-2796
7. Susin SA, Lorenzo HK, Zamzami N et al. Molecular characterization of mitochondrial apoptosis-inducing factor. Nature. 1999;397:441-446
8. Vahsen N, Cande C, Briere JJ et al. AIF deficiency compromises oxidative phosphorylation. Embo J. 2004;23:4679-4689
9. Zeviani M. Mitochondrial disorders. Suppl Clin Neurophysiol. 2004;57:304-312
10. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem. 1976;72:248-254
11. Yu S, Ding WG. The 45 kDa form of glucose transporter 1 (GLUT1) is localized in oligodendrocyte and astrocyte but not in microglia in the rat brain. Brain Res. 1998;797:65-72 12. Carson MJ, Thrash JC, Lo D. Analysis of microglial gene expression: identifying targets for
CNS neurodegenerative and autoimmune disease. Am J Pharmacogenomics. 2004;4:321-330 13. Kreutzberg GW. Microglia: a sensor for pathological events in the CNS. Trends Neurosci.
1996;19:312-318
14. Imai Y, Kohsaka S. Intracellular signaling in M-CSF-induced microglia activation: role of Iba1. Glia. 2002;40:164-174
15. Schmued LC, Hopkins KJ. Fluoro-Jade B: a high affinity fluorescent marker for the localization of neuronal degeneration. Brain Res. 2000;874:123-130
16 16. Zhu C, Wang X, Huang Z et al. Apoptosis-inducing factor is a major contributor to neuronal
loss induced by neonatal cerebral hypoxia-ischemia. Cell Death Differ. 2006
17. Nunomura A, Perry G, Aliev G et al. Oxidative damage is the earliest event in Alzheimer disease. J Neuropathol Exp Neurol. 2001;60:759-767
18. Pratico D, Uryu K, Leight S et al. Increased lipid peroxidation precedes amyloid plaque formation in an animal model of Alzheimer amyloidosis. J Neurosci. 2001;21:4183-4187 19. Mattiazzi M, D'Aurelio M, Gajewski CD et al. Mutated human SOD1 causes dysfunction of
oxidative phosphorylation in mitochondria of transgenic mice. J Biol Chem. 2002;277:29626-29633
20. Damiano M, Starkov AA, Petri S et al. Neural mitochondrial Ca2+ capacity impairment precedes the onset of motor symptoms in G93A Cu/Zn-superoxide dismutase mutant mice. J Neurochem. 2006;96:1349-1361
21. Panov AV, Gutekunst CA, Leavitt BR et al. Early mitochondrial calcium defects in Huntington's disease are a direct effect of polyglutamines. Nat Neurosci. 2002;5:731-736 22. Levine JB, Kong J, Nadler M, Xu Z. Astrocytes interact intimately with degenerating motor
neurons in mouse amyotrophic lateral sclerosis (ALS). Glia. 1999;28:215-224
23. Beretta S, Wood JP, Derham B et al. Partial mitochondrial complex I inhibition induces oxidative damage and perturbs glutamate transport in primary retinal cultures. Relevance to Leber Hereditary Optic Neuropathy (LHON). Neurobiol Dis. 2006;24:308-317
24. Bolanos JP, Heales SJ, Land JM, Clark JB. Effect of peroxynitrite on the mitochondrial respiratory chain: differential susceptibility of neurones and astrocytes in primary culture. J Neurochem. 1995;64:1965-1972
25. Galvani P, Fumagalli P, Santagostino A. Vulnerability of mitochondrial complex I in PC12 cells exposed to manganese. Eur J Pharmacol. 1995;293:377-383
26. Vasquez OL, Almeida A, Bolanos JP. Depletion of glutathione up-regulates mitochondrial complex I expression in glial cells. J Neurochem. 2001;76:1593-1596
27. Bruijn LI, Becher MW, Lee MK et al. ALS-linked SOD1 mutant G85R mediates damage to astrocytes and promotes rapidly progressive disease with SOD1-containing inclusions. Neuron. 1997;18:327-338
28. Wong PC, Pardo CA, Borchelt DR et al. An adverse property of a familial ALS-linked SOD1 mutation causes motor neuron disease characterized by vacuolar degeneration of mitochondria. Neuron. 1995;14:1105-1116
29. Pekny M, Nilsson M. Astrocyte activation and reactive gliosis. Glia. 2005;50:427-434
30. Markiewicz I, Lukomska B. The role of astrocytes in the physiology and pathology of the central nervous system. Acta Neurobiol Exp (Wars). 2006;66:343-358
17 31. Benit P, Slama A, Cartault F et al. Mutant NDUFS3 subunit of mitochondrial complex I causes
Leigh syndrome. J Med Genet. 2004;41:14-17
32. Triepels RH, Van Den Heuvel LP, Trijbels JM, Smeitink JA. Respiratory chain complex I deficiency. Am J Med Genet. 2001;106:37-45
33. Tanji K, Kunimatsu T, Vu TH, Bonilla E. Neuropathological features of mitochondrial disorders. Semin Cell Dev Biol. 2001;12:429-439
34. Benit P, Chretien D, Kadhom N et al. Large-scale deletion and point mutations of the nuclear NDUFV1 and NDUFS1 genes in mitochondrial complex I deficiency. Am J Hum Genet. 2001;68:1344-1352
35. Crimi M, Papadimitriou A, Galbiati S et al. A new mitochondrial DNA mutation in ND3 gene causing severe Leigh syndrome with early lethality. Pediatr Res. 2004;55:842-846
36. Esteitie N, Hinttala R, Wibom R et al. Secondary metabolic effects in complex I deficiency. Ann Neurol. 2005;58:544-552
37. Huang MY, Jong YJ, Tsai JL et al. Mitochondrial NADH-coenzyme Q reductase deficiency in Leigh's disease. J Formos Med Assoc. 1996;95:325-328
38. Larsson NG, Andersen O, Holme E et al. Leber's hereditary optic neuropathy and complex I deficiency in muscle. Ann Neurol. 1991;30:701-708
39. Laugel V, This-Bernd V, Cormier-Daire V et al. Early-onset ophthalmoplegia in Leigh-like syndrome due to NDUFV1 mutations. Pediatr Neurol. 2007;36:54-57
40. Mizuno Y, Ohta S, Tanaka M et al. Deficiencies in complex I subunits of the respiratory chain in Parkinson's disease. Biochem Biophys Res Commun. 1989;163:1450-1455
41. Solano A, Roig M, Vives-Bauza C et al. Bilateral striatal necrosis associated with a novel mutation in the mitochondrial ND6 gene. Ann Neurol. 2003;54:527-530
18
Legends to figures (7 figures)
Figure 1: Expression of AIF, CI and astrocytic proteins in ataxic mice (Hq) and wild-type littermates (WT). (A) Western blots of cerebellum, cortex, thalamus, and striatum homogenates
pooled from four 6-month-old WT and Hq mutant males. Actin was used as control of equal loading. (B) Semi-quantitative analysis of the expression AIF, OXPHOS Complex I subunits p39 and p30, GFAP and S100β shown in A, expressed in percentage relative difference with WT signal intensity. AIF and OXPHOS complex I subunits (p39 and p30) are significantly reduced in Hq samples as compared to WT, whereas the glial fibrillary acidic protein (GFAP) is strongly increased.
Figure 2: GFAP immunoreactivity on sagittal sections from Hq mutant mice and wild-type littermates at thalamic level. (A) GFAP labelling on intermediate sagittal sections at 2, 4 and 6
months of age reveals a growing astroglial reaction in the thalamus from Hq mutant mice except in a small dorsal region which includes the latero-dorsal nucleus at 4 months (black box). (B) Higher magnifications of the dorsal thalamic area at 4 months corresponding to the region boxed in A and shown at medial, intermediate and lateral sagittal sections. The dorsal area spared by astrogliosis is wider in medial sections and disappears in lateral section (*). Th = Thalamus. Hi = Hippocampus. mo = months of age. WT = wild-type. Scale bars = 1000µm (A); 200µm (B).
Figure 3: Immunohistochemical analysis of the ventro-lateral area of the thalamus in Hq mutant mice and wild-type littermates at 4 months of age. (A) GFAP, Glucose Transporter-1 (GLUT-1),
IBA-1, NeuN, S100β and Fluoro-Jade B labelling. In the ventro-lateral area of the thalamus, GFAP immunostaining is stronger in Hq mutant mice than in wild-type animals. GLUT-1 immunostaining reveals a highly developed vascular network and IBA-1 immunostaining shows active microglia in the same region. Density of S100β positive astrocytes and NeuN positive neurones is similar in Hq and wild-type animals. Some degenerating neurites are detected with Fluoro-Jade B (white arrows). Scale bar = 50µm. (B) Quantitative analysis of the number of GFAP-, S100β- and NeuN- immunoreactive
19 cells in the ventro-lateral area of the thalamus from Hq and wild-type mice at 2, 4 and 6 months of age. Bars represent the mean ± standard deviation. Asterisks indicate significant differences between the Hq and the wild-type groups (***, p<0.001 by 2-way analysis of variance with the Bonferroni correction).
Figure 4: Immunohistochemical analysis of the cerebellum in Hq mutant mice and wild-type littermates at 4 months of age. (A) GFAP, Glucose Transporter-1 (GLUT-1), IBA-1, NeuN, S100β
and Fluoro-Jade B labelling. In the lobules, GFAP immunostaining is stronger in Hq mutant mice than in wild-type animals. The staining involved both the granular layer (GL) where numerous hypertrophic GFAP-positive astrocytes were found and the molecular layer (ML) where the Bergmann glia was strongly positive (black arrows). GLUT-1 immunostaining reveals a highly developed vascular network and IBA-1 immunostaining shows active microglia in the same region. Density of S100β positive astrocytes appears similar in Hq and wild-type animals. Fluoro-Jade B detects a massive degeneration of the granule cells, consistent with the loss of NeuN staining observed in the granular layer. Scale bar = 100µm. (B) Quantitative analysis of the number of GFAP-, S100β - and NeuN- immunoreactive cells in the cerebellar granular layer from Hq and wild-type mice at 2, 4 and 6 months of age. Note the progressive increase of GFAP immunoreactivity and the progressive loss of NeuN-positive cells. Bars represent the mean ± standard deviation. Asterisks indicate significant differences between the Hq group and the wild-type group (*, p<0.05, **, p<0.01 and ***, p<0.001 by 2-way analysis of variance with the Bonferroni correction).
Figure 5: Immunohistochemical analysis of the caudate-putamen in Hq mutant mice and wild-type littermates at 4 months of age. (A) GFAP, Glucose Transporter-1 (GLUT-1), IBA-1, NeuN,
S100β and Fluoro-Jade B labelling. In the thalamo-cortical fibre tracts, GFAP immunostaining is stronger in Hq mutant mice than in wild-type animals. GLUT-1 immunostaining reveals a highly developed vascular network and IBA-1 immunostaining shows active microglia in the same region (black arrows). Density of S100β positive astrocytes and NeuN positive neurones appears similar in
20 Hq and wild-type animals. Fluoro-Jade B detects degenerating neurites specifically located within the fibre tracts (white arrows). Scale bar = 100µm. (B) Quantitative analysis of the number of GFAP-, S100β- and NeuN-immunoreactive cells in the caudate-putamen from Hq and wild-type mice at 2, 4 and 6 months of age. Bars represent the mean ± standard deviation. Asterisks indicate significant differences between the Hq group and the wild-type group (**, p<0.01, and ***, p<0.001 by 2-way analysis of variance with the Bonferroni correction).
Figure 6: Fluoro-Jade B labelling in the somatosensory and motor cortical regions, in the globus pallidus and in the deep cerebellar nuclei in Hq mutant mice at 4 months of age. Degenerating
neurons and processes are detected in the deep layers of the motor cortex and the somatosensory cortex from Hq mutant mice at 4 months (arrows). Degenerating neurites are detected in the globus pallidus and in the deep cerebellar nuclei at the same age (arrows). Scale bar = 20µm.
Figure 7: Electron microscopy micrographs of the thalamic, cerebellar and striatal regions in Hq mutant mice at 2, 4 and 6 months of age. Neurons from the ventro-postero-lateral area of the
thalamus (Th), the cerebellar granule cells (GCL), the Purkinje cells (Pkj) and the Caudate-Putamen (CPu) show swollen degenerating mitochondria in all cell types. Insets depict higher magnifications of boxed areas. Most neurons examined harboured degenerating mitochondria even at 2 months of age (white arrows). Scale bar = 0.5µm.
