| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Unité de Recherche Associée CEA-CNRS 2210, Service Hospitalier Frédéric Joliot, Département de Recherches Médicales, Direction des Sciences du Vivant, Commissariat à l’Energie Atomique, Orsay, France; and
Laboratoire de Neurobiologie cellulaire et moléculaire, CNRS, UPR 9040, Gif-sur-Yvette, France
1Correspondence: URA CEA-CNRS 2210, Service Hospitalier Frédéric Joliot, DRM, DSV, CEA, 4 place du Général Leclerc, 91401 Orsay cedex, France. Email: brouille{at}shfj.cea.fr
ABSTRACT
According to the "indirect" excitotoxicity hypothesis, mitochondrial defects increase Ca2+ entry into neurons by rendering NMDA-R hypersensitive to glutamate. We tested this hypothesis by investigating in the rat striatum and cultured striatal cells how partial mitochondrial complex II inhibition produced by 3-nitropropionic acid (3NP) modifies the toxicity of the NMDA-R agonist quinolinate (QA). We showed that nontoxic 3NP treatment, leading to partial inhibition of complex II activity, greatly exacerbated striatal degeneration produced by slightly toxic QA treatment through an "all-or-nothing" process. The potentiation of QA-induced cell death by 3NP was associated with increased calpain activity and massive calpain-mediated cleavage of several postsynaptic proteins, suggesting major neuronal Ca2+ deregulation in the striatum. However, Ca2+ anomalies probably do not result from NMDA-R hypersensitivity. Indeed, brain imaging experiments using [18F]fluorodeoxyglucose indirectly showed that 3NP did not increase QA-induced ionic perturbations at the striatal glutamatergic synapses in vivo. Consistent with this, the exacerbation of QA toxicity by 3NP was not related to an increase in the QA-induced entry of 45Ca2+ into striatal neurons. The present results demonstrate that the potentiation of NMDA-R-mediated excitotoxicity by mitochondrial defects involves primarily intracellular Ca2+ deregulation, in the absence of NMDA-R hypersensitivity.—Jacquard, C., Trioulier, Y., Cosker, F., Escartin, C., Bizat, N., Hantraye, P., Cancela*, J. M., Bonvento, G., Brouillet, E. Brain mitochondrial defects amplify intracellular [Ca2+] rise and neurodegeneration but not Ca2+ entry during NMDA receptor activation.
Key Words: 3-nitropropionic acid • calpain • striatum • Huntington’s disease
THERE IS COMPELLING EVIDENCE that abnormal glutamate receptor activity (1)
and defects in mitochondrial function (2)
are involved in acute neurological conditions, such as stroke, and in chronic neurodegenerative illnesses, such as Huntington’s disease (HD). Mitochondrial defects and the overactivation of glutamate receptors, including the ionotropic NMDA subtype in particular, may act together to trigger neurodegeneration (3
4
5)
.
Excitotoxicity, mediated by the NMDA receptor (NMDA-R), rapidly damages mitochondria by leading to Ca2+ accumulation in the organelles (6
,7)
, an increase in free radical production (8
,9)
, the collapse of mitochondrial membrane potential, and a decrease in ATP production (10
,11)
.
Acute impairment of oxidative energy metabolism can potentiate NMDA-R-mediated excitotoxicity (12
13
14
15)
. Potentiation is likely associated with major cytosolic Ca2+ deregulation. In support of this hypothesis, fluorescence imaging techniques showed that, in cultured neurons and brain slices, NMDA-R activation produces cytosolic Ca2+ rise that is exacerbated in neurons with respiratory chain defects (16
,17)
.
Three important aspects of the mechanisms underlying the potentiation between NMDA-R activation and mitochondrial defects remain to be elucidated. First, evidence for an increase in cytosolic Ca2+ concentrations in the detrimental coupling of mitochondria and NMDAR has never been provided in vivo. Second, the mechanism leading to an increase in cytosolic Ca2+ rise in neurons with mitochondrial defects is unknown. However, the so-called "weak" or "indirect" excitotoxicity hypothesis is often suggested (3
4
5
,18)
. According to this hypothesis, mitochondrial alterations may decrease ATP synthesis, impairing Na+/K+-ATPase function, thereby decreasing plasma membrane potential and relieving the voltage-dependent Mg2+ blockade of NMDA-R, rendering the receptor hypersensitive to glutamate (5)
. Hypersensitive NMDA-R could increase the entry of Ca2+ into neurons, leading to Ca2+ deregulation. However it is also possible that cytosolic Ca2+ deregulation results from an alteration of the intracellular sequestration and/or extrusion of Ca2+. Third, the relationship linking the potentiation of NMDA-R-mediated excitotoxicity and partial chronic mitochondrial dysfunction has never been characterized. One possibility is that potentiation of excitotoxicity is proportionate to the magnitude of mitochondrial defects. Alternatively, mitochondrial anomalies may potentiate NMDA-R-mediated excitotoxicity through an "all-or-nothing" process.
The aim of the present study was to investigate the mechanisms underlying the potentiation of NMDA-R-mediated neurodegeneration by chronic mitochondrial dysfunction. We determined how the toxicity of the NMDA-R agonist quinolinate (QA) was amplified by chronic administration of 3-nitropropionic acid (3NP), a selective inhibitor of mitochondrial complex II. We investigated these mechanisms in the rat striatum in vivo and in cultured striatal neurons. QA and 3NP produce striatal lesions that are reminiscent of those seen in HD (19)
. QA injection reproduces the NMDA-R hypersensitization to glutamate observed in neurons overexpressing a mutated form of huntingtin (m-Htt), the protein implicated in HD (20
21
22)
. The systemic administration of 3NP replicates the mitochondrial alterations reported in HD patients and transgenic mouse models (23
24
25
26)
. We report that QA toxicity is exacerbated by 3NP only for a particular range of mitochondrial complex II inhibition, >35% and involves Ca2+ deregulation, leading to calpain activation, in the absence of NMDA-R hypersensitivity.
MATERIALS AND METHODS
Animals
Twelve-week-old male Lewis rats (Charles River) weighing 320–350 g were used for these studies. All animals were housed under standard conditions (12 h light cycle), with free access to food and water. For all surgical procedures, rats were anesthetized with a mixture of ketamine (15 mg/kg) and xylazine (1.5 mg/kg). Protocols were performed in compliance with EEC directives (86/609/EEC) and the guidelines of the French National Committee (87/848) for Care and Use of Laboratory Animals.
Preparation of neurotoxins
3NP (Fluka, France) was dissolved in 5 ml deionized water, and the solution was brought to pH 7.4 with 10 N NaOH. The pH was stabilized with 2.5 ml 0.1 M phosphate buffer (pH 7.4). The final volume was adjusted to 50 ml. For preparation of the stock solution of QA (180 mM), 30 mg of the excitotoxin (Sigma) were dissolved in 0.9 ml 0.1 M phosphate buffer (pH 7.4) with 0.9% NaCl (PBS) and 100 µl 1N NaOH. Solutions of lower QA concentrations were prepared by appropriate dilutions in PBS.
Experimental design of rat studies
We carried out time-course experiments to characterize the effect of QA in naive rats and rats chronically treated with 3NP (Fig. 1
). Rats were subcutaneously implanted with osmotic minipumps delivering 3NP or were sham-operated (T0–5 days). Five days later (T0), we simultaneously administered QA to the right striatum and PBS to the left striatum by stereotactic injection. In [18F]fluorodeoxyglucose (FDG) uptake studies, animals were killed 1 h after intrastriatal injection. In other experiments, rats were killed 24 h after intrastriatal injection (T0+1 day) for biochemical and histological analysis of "ongoing" cell degeneration. In this case, striatal degeneration was assessed using the vital dye triphenyltetrazolium chloride (TTC) for lesion volume determination, by biochemical and in situ detection of DNA fragmentation ("intensity" of degeneration), and using biochemical markers (fluorescence assay for calpain activity and calpain-dependent protein cleavage analysis by Western blotting). We evaluated the extent and intensity of the neurodegeneration induced by the different treatments in the long term in animals killed 2 wk (T0+15 days) after intrastriatal injection of QA or vehicle, by immunohistochemistry.
|
Chronic 3NP treatment
3NP was prepared as described previously and systemically administered via subcutaneous osmotic minipumps (10 µl/h, 2ML1 model, Alzet, Palo Alto, CA) (27)
delivering 10–45 mg/kg/day for 6 days, the dose depending on the group of animals considered. Sham-treated rats (no 3NP treatment) were similarly anesthetized and subjected to the same surgery for the implantation of empty osmotic pumps.
Intrastriatal administration of QA
On the fifth day after surgery for osmotic pump implantation, all animals (3NP-treated rats and sham-operated littermates) were anesthetized and placed in a stereotactic frame. Stereotactic intrastriatal injections of 1 µl QA (4, 10, 15, 20, 40, 80, 120, or 180 nmol) or 1 µl PBS were made simultaneously in the right and left striatum, respectively, using two blunt-tipped 25 gauge Hamilton syringes placed in parallel in a needle holder. Each injection was made over 1 min, and the needle was left in place 2 min before being slowly withdrawn. Stereotactic coordinates (+0.8 mm rostral from bregma; 3.5 mm lateral from midline, and 5 mm ventral from dura, with tooth bar set at –3.3 mm) were chosen according to the atlas of Paxinos and Watson (28)
.
FDG uptake experiments
Five days after the implantation of osmotic pumps, rats fasted for 24 h were anesthetized and the femoral vein was catheterized for intravenous administration of the radiotracer. The femoral artery was also catheterized for the removal of blood samples for PaO2 and PaCO2 determinations. The animals were then placed in the stereotactic frame. We injected QA (40 nmol) into the striatum as described above, followed by FDG (3NP-treated rats, n=5, 1.2±0.4 mCi; sham-treated rats, n=5, 1.1±0.2 mCi) 15 min later. Rats were maintained at 37°C, using a heating blanket. Glycemia was determined with a Onetouch glucose (Glc) meter (Lifescan Inc., Milpitas, CA). PaO2 and PaCO2 were determined with a blood gas analyzer (ABL5 Radiometer, Copenhagen, Denmark). Physiological parameters were monitored during this experiment and were found not to change significantly over a period beginning just before and continuing until 45 min after FDG injection. Similarly, no significant intergroup differences were noted. The parameters for rats (n=5) without 3NP treatment were (before/after FDG injection): glycemia 1.36 ± 0.25/1.41 ± 0.10 g/l; PaO2 76 ± 4/79 ± 6 mm Hg; PaCO2 44 ± 5/44 ± 6 mm Hg. The parameters for 3NP-treated rats (n=5) were (before/after FDG injection): glycemia 1.36 ± 0.12/1.48 ± 0.12 g/l; PaO2 74 ± 13/89 ± 9 mm Hg; PaCO2 47 ± 2/42 ± 3 mm Hg. Rats were killed 50 min after FDG administration. Brains were rapidly removed, frozen in isopentane at –40°C and cut into 20 µm coronal sections with a cryostat. Sections were serially collected, at intervals of 100 µm, on Superfrost plus slides (Fischer, Elancourt, France). Slides were then placed against BIOMAX MR films (Kodak) overnight at room temperature. Autoradiographs were digitized using a scanner (Amersham Pharmacia Biotech, New Castle, UK), and striatal optical densities (subtracted from background) were measured using an image analysis system (MCID, Imaging Research Inc., St. Catharines, Ontario, Canada). For all animals, we calculated the mean striatal optical density (OD) of 20 sections (4 mm rostrocaudal extension) centered on the site of injection of QA and PBS. In our experimental conditions, we checked that brain OD values increased linearly as a function of increasing section thickness (from 3 to 40 µm) even in the cerebral cortex, which showed high rates of FDG uptake. This made it possible to rule out saturation of the film by high levels of radioactivity in the 100% range. In addition, striatal OD increased linearly as a function of the radioactivity injected (mCi/kg). This made it possible to determine the increase in FDG uptake induced by QA injection semiquantitatively. The relative QA-induced increase in FDG uptake in the right striatum was determined by expressing FDG uptake in the right striatum as a percentage of that in the PBS-injected contralateral (left) striatum.
Anatomical and histological evaluation
For short-term evaluation of striatal lesions, the brain was dissected out, rapidly washed in cold PBS, placed in a brain matrix (Pelco Inc.), and 1 mm fresh slices were cut. Slices were stained with the vital red dye TTC and stored in 4% paraformaldehyde (PFA) (29)
. The volume of striatal lesions was assessed after image acquisition for all sections, using a high-resolution scanner (Amersham Pharmacia Biotech) and Total Lab image analysis software (3.1 version, Amersham). For each coronal brain section, the lesioned area was measured by manually delineating the external border of the lesion, seen as a pale staining on digitized images. From these areas, the volume of the striatal lesion was determined for all animals using the Cavalieri method (30)
.
For long-term immunohistochemical evaluation of striatal lesions, rats were anesthetized with pentobarbital and perfused transcardially with 4% PFA. The brain was removed, postfixed overnight, and cryoprotected in sucrose solutions (12, 16, and 18%) in PBS. Brain sections (40 µm) were stained with gallocyanin or processed for immunohistochemistry using mouse monoclonal antineuronal nucleus antibody (Ab) (NeuN, MAB377, diluted 1:10000, Chemicon International Inc.) or a rabbit anti-DARPP32 (AB1656, 1:3000, Chemicon International Inc.). The primary Ab was detected with biotinylated-conjugated antimouse or anti-rabbit Ab (1:3000, Vector), followed by double amplification with avidin-biotin and tyramine-biotin. The final staining pattern was obtained by incubation for 2 min with VIP substrate (Vector). Before DARPP32 immunostaining, we performed NADPH-diaphorase histochemistry by incubating sections for 120 min at 37°C in a buffer containing 0.1 M Tris-HCl (pH 7.4), 0.4 g/l NADPH, 1 g/l nitroblue tetrazolium (Sigma), and 0.5% Triton-X100.
For in situ detection of DNA fragmentation, brains were processed as for immunohistochemistry and 10 µm thick coronal sections were processed for the terminal deoxynucleotidyl transferase-mediated biotinylated uridine triphosphate nick end labeling (TUNEL) method, using the in situ cell death detection-fluorescein kit (Roche). For all animals, the apparent density of TUNEL-positive cells was estimated over the entire surface of the striatum in two to three sections close to rostrocaudal concentration, 0.8 mm anterior to the bregma. Sections were scanned with an x20 objective, using a Zeiss Axioplan2 imaging microscope motorized for X, Y, and Z displacements and an image acquisition and analysis system (Morphostar 5.12, IMSTAR, Paris, France) (27)
. Results are expressed as the number of TUNEL-positive cells per section.
Measurement of SDH activity
SDH activity was evaluated in situ, using a histochemical procedure (31)
and a previously described quantification method (32)
in frozen coronal brain sections. This procedure allows for quantitative determination of changes in the regional Vmax of SDH (32)
.
Brain processing for biochemical analysis
Brains were collected 24 h after QA injection, rapidly rinsed in cold PBS, and cut into 2 mm fresh slices using a steel brain matrix. The striatum was dissected out and homogenized, using a 1 ml glass Potter homogenizer (800 rpm, 20 strokes), in 300 µl of buffer containing 25 mM HEPES-NaOH, pH 7.4, 150 mM NaCl, 1.3 mM EDTA, 1 mM EGTA, 10 mM benzamidine, and 1 mM PMSF. An aliquot of the total homogenate was used for Western blotting (60 µl). The remaining homogenate was centrifuged at 15,000 g for 30 min to obtain a supernatant enriched in cytosolic proteins for proteolytic activity assays and oligonucleosome detection. Aliquots of total homogenate and of the supernatant fraction were stored at –80°C until analysis. For all samples, protein concentrations were quantified with the microBCA kit (Pierce), according to the manufacturer’s instructions.
Calpain activity measurement in brain extracts
Calpain activity was determined using the fluorogenic substrate N-succinyl-Leu-Tyr-AMC as described previously, but with minor modifications (27)
. The selectivity of the assay was increased (elimination of the Ca2+-dependent release of AMC unrelated to calpain) by incubating half of the sample with Z-Leu-Leu-CHO (10 µM), a specific inhibitor of calpain. AMC release was quantified using a standard curve after 30 min of incubation at 37°C in a 96-well plate reader (Fusion plate reader, Perkin-Elmer), with excitation/emission wavelengths set at 380 nm/460 nm. Specific calpain activity was calculated as the difference between AMC released in presence of Ca2+ and that released in presence of Ca2+ and z-Leu-Leu-CHO. For each group of rats, specific activity was normalized according to protein content. Activity was expressed in picomoles of AMC released per minute per milligrams of protein.
Western blotting
Equal amounts of protein from striatal extract samples were pooled for each animal group (7 or 8 striata per group), and 30 µg of the pooled sample were loaded in triplicate on acrylamide gels and separated by SDS-PAGE (6, 8, or 10% acrylamide). The protein bands were blotted onto nitrocellulose or PVDF membranes, which were then incubated overnight at 4°C with an Ab against one of the following proteins: fodrin (1:1000, MAB 1622, non erythroid 
-spectrin, Chemicon), huntingtin (1/1000, hu4C8, Euromedex), PSD-95 (1/250, P46520, Beckton Dickinson), NR-2B (1/500, NR-2B C-terminal, 1469, Santa Cruz), and actin (1/10000, A2066, Sigma). Membranes were incubated with horseradish-peroxidase-conjugated antimouse, anti-rabbit (1:2000, Amersham Biosciences), or anti-goat (1:2000, Vector Laboratories) Ab, and peroxidase activity was detected using enhanced chemiluminescence (ECL) reagent (Amersham Biosciences). The membranes were placed against X-ray films (Kodak), and the resulting autoradiographs were scanned, and OD values were measured with Total Lab software. For all experimental groups, the mean OD of all studied bands (full-length proteins and breakdown products) was measured from observations made in triplicate.
Oligonucleosome detection
The oligonucleosome content of the cytosolic supernatant fraction was determined spectrophotometrically, using the cell death detection ELISA plus kit (Roche Applied Science), according to the manufacturer’s instructions. The presence of oligonucleosomes was assessed by measuring OD, which was normalized according to the protein content of the sample.
Primary striatal culture
Primary striatal neurons were obtained from E14-E15 rat embryos. Timed-pregnant Sprague-Dawley rats (Janvier, Le Genest-St-Isle, France) were killed with a lethal dose of pentobarbital, and embryos were quickly removed and dissected on cooled Hank’s balanced sodium salts (without Ca2+ and Mg2+, Sigma). Ganglionic eminences were isolated and incubated for 15 min at 37°C in 0.3 mg/ml DNase I (Sigma). Tissues were mechanically dissociated with a fire-polished Pasteur pipette, and debris was removed after decantation of the suspension. Cells were finally concentrated by centrifugation (20°C, 5 min, 1500 g) and resuspended in serum-free Neurobasal medium supplemented with 2% B27 supplement (Life Technologies), 1% antibiotic-antifungal mixture (Life Technologies), and 0.5 mM L-glutamine (Sigma). Cells were plated at a density of 400,000 cells/well in 24-well-Costar plates coated with 50 µg/ml 30–70 kDa poly-D-lysine (Sigma). The cultures were kept in a humid incubator (5% CO2, 37°C), and half the medium was changed once a week.
Calcium uptake measurement with 45Ca2+
Experiments were performed on cultures after 21 days in vitro (DIV). PBS or 3NP (75 µM) was added to the medium 5 h before 45Ca2+ uptake measurements. Ca2+ uptake assays were performed as described by Hartley et al. (33)
but with minor modifications. Cells were rinsed with buffer A, which contained 25 mM HEPES (pH 7.4), 120 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl2, 1.8 mM CaCl2, and 15 mM Glc. Cells were incubated with 5 mM QA in buffer A supplemented with 45Ca2+ (2x106 cpm/well, Amersham CES3) for 10 min at 24°C. For pharmacological blockade of QA-induced Ca2+ entry, 10 µM MK801 was added for 15 min before addition of QA and 45Ca2+. We stopped 45Ca2+ uptake by replacing buffer A with fresh, non radioactive buffer A containing 2 µM MK-801, for the rapid blocking of NMDA-R-mediated 45Ca2+ influx (34)
. Cells were rapidly rinsed four times in 900 µl of this buffer. The fourth wash lasted 6 min. Extracellular buffer was collected for evaluation of Ca2+ efflux. Then cells were extracted from culture wells using 300 µl SDS 0.2%. Protein concentrations were determined on 20 µl aliquots. The radioactivity present in 200 µl of cell extract (i.e., intracellular 45Ca2+) and 200 µl of washing buffer (i.e., extracellular 45Ca2+) was determined by three rounds of scintillation counting (Beckman counter LS5801) for 2 min and corrected for background. Intracellular accumulation of Ca2+ was expressed as counts per minute per milligram of protein. The experiment was repeated twice and similar results were obtained in each case.
SDH activity in neuronal cell culture
Cells in a six-well plate were incubated with 75 µM 3NP or PBS for 5 h, rinsed, treated with trypsin, and centrifuged. We homogenized 
1.5 x 106 cells in 200 µl of a buffer containing 20 mM Tris (pH 7.2), 250 mM sucrose, 2 mM EGTA, 40 mM KCl, and 1 mg/ml BSA and protease inhibitor cocktail (Roche). SDH activity was measured as described by Munujos et al. (35)
. Samples were incubated in triplicate with 50 mM succinate and in duplicate without succinate (for determination of nonspecific activity) in the presence of 2 mM iodo-nitrotetrazolium chloride (INT) in 50 mM Tris buffer (pH 8.3) supplemented with 0.5 mM EDTA at 37°C for 90 min. The formation of red formazan was monitored spectrophotometrically at 485 nm. Specific activity was calculated by subtracting nonspecific formazan formation in absence of succinate from total activity measured in presence of succinate. Activity was normalized according to protein concentration.
Cell culture cytotoxicity
Cytotoxicity experiments were performed on sister cultures after 21 DIV. LDH release, an index of cell disruption, was measured in duplicate in each well. Four wells were used for each experimental condition. 3NP (75 µM) or PBS was applied 5 h (T0–5 h) before QA treatment. QA (5 mM) or PBS in culture medium was then applied (T0). For the determination of LDH release, aliquots of culture medium were collected immediately after QA application (T0, 50 µl), and every day thereafter (T0+24 h, T0+48 h, T0+72 h, 100 µl). After centrifugation to remove floating cells, aliquots were incubated at 37°C with the LDH kit (Roche) for 60 min, according to the manufacturer’s instructions. LDH release was calculated as the percentage of LDH activity in the medium normalized according to the total LDH activity (100%) in wells after the complete lysis of cells with 2% Triton X-100.
In situ detection of calpain activity on living cells
For in situ determination of calpain activity in cell cultures (19 DIV), the volume of extracellular medium was set to 500 µl for each well the day preceding the experiment. The treatment with 3NP (75 µM) started 5 h before QA treatment. The fluorogenic calpain substrate SLY-AMC (5 mM in DMSO) was added at a final concentration of 50 µM final (0.1% DMSO final) in the culture medium. QA (5, 1, or 0.5 mM final) or vehicle (PBS) is added to the medium and fluorescence intensity in cells was determined at different time points (0, 15, and 20 h after QA addition) using a plate reader (fluorimeter Fusion, Packard) set in "bottom" mode to selectively detect fluorescence of AMC (Ex 380/Em 460 nm) at the levels of plated cells. Fluorescence at time "0" was considered to represent background fluorescence unrelated to substrate cleavage. We checked that the fluorescence signal detected at 20 h after QA (4 wells, mean±SE: 4661±940) was markedly inhibited by pretreating cells with 10 µM calpain inhibitor I (Sigma) (4 wells, mean±SEM: 1440±490), demonstrating that at least 70% of the accumulation of AMC resulted from calpain activity.
ATP measurement
ATP concentrations was determined using the method of Lust et al. (36)
in four sister culture wells for each condition. Cultured cells (19 DIV) were incubated in 75 µM or 1 mM 3NP for 5 h. As positive control for massive ATP depletion, sodium azide (0.2%) plus 1 mM 3NP were added to massively block oxidative phosphorylation. Cultures were rinsed with PBS, and 200 µl per well of the ATP releasing buffer (Sigma) were rapidly added at 37°C for 5 min. Detection of ATP concentrations in samples was performed right after lysis. Triplicates (50 µl) of each sample are put in black plastic 96-well plates (Nunc). To start the reaction, 50 µl of the reaction medium containing 100 mM glycine, 15 mM de dithiothreitol, 4 mM EGTA, 1 mM EDTA, 20 mM MgSO4, 0.08% BSA, and 30 µM luciferin (Promega) with luciferase (38,400 LU/mg, Sigma) were sequentially added to each sample using the automatic injector of a luminometer (MITHRAS LB940, Berthold technologies, Bad Wildbad, Germany). Reading was started 1 s after the injection and lasted 1 s. Light that was detected in each sample was transformed in ATP concentrations using a standard curve (from 10 picoM to 1 mM). Results were expressed as (mean±SE) pmol of ATP/well.
Calcium imaging
For imaging studies cells were plated on glass coverslips coated with poly-D-lysine. Cells were incubated in culture medium with Fura-2-AM (5 µM final) in a plexiglass chamber set at 37°C for 30 min. Fura-2-loaded cells were perfused (1 ml/min) with buffer A during 10 min to establish the baseline of fluorescence at rest. Then a 5 mM QA solution in buffer A was perfused. The measurements were performed at room temperature using an inverted epifluorescence microscope (Leica) with an x40 oil immersion lens and equipped with a cooled CCD camera (Hamamatsu ORCA ER). Fura-2 was sequentially excited at 340 nm and 380 nm, and emitted fluorescence was collected at 510 nm. All settings of the Lambda DG4, filters, and microscope and the complete data acquisition were controlled by SimplePCI6.0 software (Compix Inc., Imaging Systems). Ratio images 340/380 were recorded at 12 images/min right before and for 2 min after starting QA perfusion and at 0.5 image/min thereafter. To determine the effects of 3NP, we used the 340/380 ratio values as an index of the variation of cytosolic Ca2+ levels at rest and during QA stimulation. For each condition (control and 3NP), three coverslips were studied. A total of 50 and 35 neurons were analyzed in 3NP-treated and untreated cells respectively.
Analysis and statistics
Results are mean ± SE. We used Mann-Whitney tests to compare QA-induced Glc uptake in rats in the presence and absence of 3NP. Other data were analyzed by unpaired Student’s t tests in experiments, with groups being compared in pairs. If several groups were to be compared in the same experiment, statistical analysis included one-way ANOVA followed by Fisher’s post hoc PLSD test. The number of 3NP-treated rats and sham-rats with lesions after QA injection (15 and 20 nmol) as compared by means of a 
2 test.
RESULTS
Exacerbation of QA-induced striatal degeneration by chronic 3NP treatment
We investigated the potential synergistic effects of mitochondrial defects on QA toxicity in vivo by determining the dose of 3NP leading to respiratory chain inhibition without triggering overt striatal degeneration. We studied the effects of various doses of 3NP, using staining with the vital dye TTC to detect striatal lesions. Treatment with <40 mg/kg 3NP for 6 days did not, in any case, lead to striatal degeneration (Fig. 2
A). Striatal lesions were seen in animals receiving higher doses (56 mg/kg/day; Fig. 2A
). Complex II is homogeneously inhibited by 3NP throughout the brain (32
,37)
, but in cases in which 3NP also causes degeneration, complex II activity in the striatum is reduced by both the inhibitory effects of the neurotoxin and by cell death. In this dose-dependent experiment, we therefore measured complex II activity in the parietal cortex. The degree of SDH/complex II inhibition increased steeply as a function of 3NP dose (correlation coefficient, r=0.982, P<0.01; Fig. 2B
), consistent with the irreversible mechanism of action of 3NP on the complex II catalytic site (32
,38
,39)
. Consistent with published results (27
,32
,40
,41)
, striatal lesions occurred only when inhibition of complex II was >50% (i.e., 57±4% of inhibition for 54 mg/kg/day).
|
To study the effects of partial complex II inhibition on QA toxicity, we chose a dose of 35 mg/kg/day 3NP, which does not lead to overt striatal degeneration. Complex II inhibition on the sixth day of 3NP administration at this dose was similar in the striatum and cerebral cortex (44±3 and 49±2% respectively, n=7 per group, n.s. by ANOVA and post hoc PLSD test, not shown).
We then compared the effect of QA injection in sham-operated and 3NP-treated rats (35 mg/kg/day). Stereotactic intrastriatal injections of QA and PBS were performed on the fifth day of 3NP treatment. PBS injection did not produce striatal degeneration in 3NP-treated animals (Fig. 2C-D
). One day after PBS injection, TTC staining showed no striatal lesion. The biochemical quantification of soluble oligonucleosomes indicated no detectable nuclear DNA fragmentation. In line with this, in situ detection of nuclei with fragmented DNA using the TUNEL method showed that, in the vicinity of the needle track, only a few cells were damaged by the intrastriatal injection of PBS (not shown). Immunohistochemical evaluation of 3NP-treated rats 2 wk after the intrastriatal injection of PBS showed no detectable cell loss in the striatum (Fig. 2D
).
We examined the presence of striatal lesions using TTC staining 24 h after QA injection (Fig. 3
). The injection of doses of QA doses <15 nmol left the striatum undamaged in all animals (Fig. 3A
). The injection of 15 and 20 nmol QA produced lesions in 10% (2/20) of the naive rats (Fig. 3A
) and 60% (12/20) of the 3NP-treated rats (df=1, 2=10.98, P<0.001). The injection of 40 nmol QA or more resulted in lesions in all animals (Fig. 3A
). The volume of QA-induced striatal lesions was greater in 3NP-treated animals than in rats not treated with 3NP (Fig. 3B
). For example, in rats injected with 40 nmol QA, the volume of striatal lesions in 3NP-treated rats was double that in rats not treated with 3NP (Fig. 3B
), with the lesion extending throughout the striatum (Fig. 3H
). At higher doses of QA (80–180 mmol), this difference between 3NP-treated rats and sham-rats was apparently reduced when using the volume of lesion as an index of degeneration, the entire striatum being almost entirely lesioned at high doses of QA (Fig. 3B
). Interestingly, regardless of the size of the lesion, the "intensity" of QA-induced cell death was higher in 3NP-treated rats than in rats not treated with 3NP. QA-induced lesions appeared as a gray pallor in the striatum in rats without 3NP, whereas those in 3NP-treated rats were white and seemed to be more necrotic (Fig. 3E and H
).
|
We determined the "intensity" of ongoing neurodegeneration rather than exclusively the spatial extent of the lesion 24 h after QA injection by studying the intensity of DNA fragmentation. QA (40 nmol) treatment increased free oligonucleosome levels 10 times more strongly in 3NP-treated rats than in rats without 3NP (Fig. 3C
). In situ DNA fragmentation analysis by TUNEL confirmed that the number of dying cells 24 h after QA injection was about 10 times larger in 3NP-treated animals than in rats not treated with 3NP (Fig. 3D
). Histological evaluation of rats 2 wk after stereotactic injections (i.e., when the cell death process was complete) showed that 40 nmol of QA led to only partial cell loss in rats not treated with 3NP, with residual neurons detected in the lesioned area (Fig. 3F and G
). In contrast, QA-induced lesions in 3NP-treated rats were associated with the complete loss of neuronal markers (Fig. 3I and J
).
Thus, 3NP markedly increases the severity of QA-induced striatal lesions and renders low doses of QA (not in themselves overtly toxic) able to trigger excitotoxic striatal lesions.
Relationship between levels of SDH inhibition and potentiation of QA toxicity
We then investigated whether the cell death potentiation phenomenon observed was proportional to the mitochondrial defects induced by 3NP. We plotted a toxicity curve for 40 nmol QA, using oligonucleosome levels as a function of the complex II inhibition measured in rats following treatment with various doses of 3NP (Fig. 4
). No potentiation was seen at levels of SDH/complex II inhibition <35%. At inhibition levels of 35–50%, the potentiation of QA-induced striatal degeneration increased sharply with inhibition. These results indicated that QA toxicity was increased only for a particular range of complex II defects, a situation not toxic in itself but probably associated with a decrease in the capability of neurons to cope with a deleterious event triggered by QA.
|
Massive activation of calpain in the potentiation of QA toxicity by 3NP
It was shown in vitro that 3NP increases the cytosolic accumulation of Ca2+ provoked by NMDA-R stimulation (16
,17)
. We wanted to test whether 3NP could produce similar effects in striatal cells in vivo. To this purpose, we studied the Ca2+-activated protease calpain, which we used as an in vivo index of Ca2+ deregulation. We used fluorimetric assays to measure the proteolytic activity of the protease in striatal extracts. We also used Western blotting to study fodrin, as the rate of calpain-dependent cleavage of this protein increases during excitotoxic cell death (42)
. In addition, we analyzed synaptic proteins of the NMDA-R multiprotein complex: the NR2B subunit, PSD-95, and huntingtin, both of which interact with the receptor (21)
.
The injection of 40 nmol of QA into rats without 3NP led to a nonsignificant increase in the proteolytic activity of calpain, as seen in fluorimetric assays (Fig. 5
A). However, significant accumulation of the products of fodrin digestion by calpain was observed, as a 145–150 kDa doublet on Western blots (Fig. 5B
), indicating that significant calpain activation occurred following QA injection.
|
In rats treated with 3NP only, no significant change was observed in the levels of calpain-like proteolytic activity in fluorimetric assays (Fig. 5A
). Consistently, fodrin and the other proteins studied were not cleaved (Fig. 5B
).
In 3NP-treated rats receiving QA, fluorimetric assays showed a significant increase in the proteolytic activity of calpain (Fig. 5A
). Western blot analysis showed that levels of the full-length forms of fodrin, huntingtin, PSD-95, and NMDA-R NR2B subunit were significantly lower in 3NP-treated than in untreated rats. The potential breakdown products of these substrates accumulated in significantly larger amounts in 3NP-treated than in untreated rats, demonstrating significant potentiation (Fig. 5B
, Table 1
). Many of these breakdown products were probably generated by calpain digestion as they migrated at the same apparent molecular weight as bands appearing after the in vitro incubation of striatal homogenates from control rats with purified µ-and m-calpain (Fig. 5B
).
|
Thus, 3NP treatment potentiated the calpain activation induced by QA, suggesting that the exacerbation of QA-induced neuronal cell death by mitochondrial impairment was associated with major Ca2+ deregulation.
Mitochondrial defects do not increase QA-induced neuronal activation in vivo
We then investigated whether the exacerbation of Ca2+ deregulation by a combination of QA and 3NP treatment was dependent on NMDA-R hypersensitivity. We reasoned that if mitochondrial defects increase the sensitivity of NMDA-R to agonists, then the striatal neuron activation/firing (and the corresponding energy demand) triggered after QA injection should be enhanced in 3NP-treated rats. We tested this hypothesis by determining whether 3NP treatment exacerbated the QA-induced increase in Glc uptake in the striatum, a direct index displaying 1:1 stoichiometry with glutamatergic activity in vivo (43
44
45)
. We injected the positron emitter FDG 10 min after intrastriatal injection of QA to study the early pharmacological effects of QA. The basal uptake of FDG in the PBS-injected striatum was similar in 3NP-treated rats and in rats not treated with 3NP (mean OD/mCi/kg ±SE; sham, 0.373±0.024; 3NP rats, 0.308±0.019; nonsignificant; Fig. 6
A and B). The administration of 40 nmol of QA significantly increased FDG uptake (mean OD/mCi/kg±SE; sham-treated rats, 0.617±0.049; 3NP rats, 0.564±0.058; P<0.001 over that observed in PBS-injected striatum, as shown by ANOVA and post hoc PLSD Fisher test). This increase was of similar extent in both groups (sham-treated, +65±5%; 3NP, +81±8%; nonsignificant by Mann-Whitney test; Fig. 6C
).
|
These results indirectly suggest that 3NP did not markedly change the early effect of QA on NMDA-R in vivo.
Mitochondrial defects do not increase the amount of Ca2+ entering the cell in response to NMDA-R stimulation
We wanted to determine more directly how intracellular Ca2+ homeostasis alterations were related to 3NP-induced potentiation of NMDA-R-mediated excitotoxicity. Fluorescence imaging techniques can be used to assess variation of cytosolic Ca2+ concentration. To quantitatively determine the modifications of the net influx/accumulation of Ca2+ from the extracellular space to the intracellular space of neurons, measurement of radioactive 45Ca2+ uptake is also particularly suitable (33)
. Since these assays can not be performed in vivo, we assessed the effect of 3NP on NMDA-R sensitivity to QA in cultured striatal cells.
We first set up in vitro conditions reproducing our in vivo observations. We previously showed that striatal cells cultured in similar conditions and continuously treated with 100 µM 3NP degenerate within 48–72 h (46)
. In the present study, we identified a 3NP concentration that substantially inhibited complex II without triggering overt degeneration. Measurements of LDH release, a sensitive index of degeneration, showed that 75 µM 3NP did not increase cell death levels over those in control cultures (Fig. 7
A). This concentration of 3NP produced substantial inhibition of (74%) complex II inhibition in cultured striatal cells (Fig. 7B
). At this nontoxic dose of 3NP, ATP concentrations were not modified compared with control condition. Higher concentrations of 3NP (1 mM) or a combination of respiratory chain inhibitors (sodium azide+3NP) were necessary to produce significant loss of ATP (Fig. 7C
). In striatal cells not treated with 3NP, 5 mM QA increased cell death levels slightly over those observed in the control (by a factor of 1.7; Fig. 7A
). QA-induced neurodegeneration levels in cultures treated with 75 µM 3NP were 2.3 times higher than those in cells treated with QA alone and 3.2 times higher than those in control cells (Fig. 7A
). At a concentration of 50 µM, 3NP inhibited complex II activity by 54% and failed to potentiate QA toxicity (not shown), suggesting that potentiation occurred only over a particular range of inhibition. Thus, subtoxic 3NP treatment in cell culture can exacerbate QA-induced cell death, as seen in rat striatum in vivo. However, the amplitude of potentiation was lower in vitro compared with the in vivo situation.
|
We next investigated whether potentiation was associated with intracellular alterations of Ca2+ homeostasis. We studied fodrin cleavage by Western blot analysis. Results showed a decrease in levels of full length fodrin and an apparition of a calpain-mediated 145/150 kDa doublet at 6 h after QA treatment (Fig. 8
A). At later time points, a similar pattern was seen, suggesting that calpain activation was rapidly maximal during QA treatment. In 3NP-treated cells, QA treatment produced a total loss of full-length fodrin and the intensity of the 145/150 kDa doublet was slightly increased compared with QA treatment in absence of 3NP (Fig. 8A
). We also measured as an index of calpain activity the in situ accumulation of AMC in living cells incubated with the calpain substrate SLY-AMC. In all conditions, a net accumulation of fluorescence was detected 15 h after addition of the substrate (earlier reading gave no reliable detection; Fig. 8B
). In cells treated with 5 mM QA, fluorescence increased to 450% of control values. Interestingly, 3NP treatment produced no substantial increase as if a ceiling effect occurred at this QA concentration (Fig. 8B
). A similar experiment was performed with lower concentrations of QA (0.5 and 1 mM). Results showed that with the lowest QA concentration (0.5 mM), AMC accumulation was still substantial (300% of control) and could be significantly potentiated by 3NP treatment (400% of control).
|
We reasoned that potentiation of QA-induced alterations of intracellular Ca2+ by 3NP might be subtle in vitro and probably occur early after the beginning of QA treatment. We used microscopy Fura-2 ratio imaging to dynamically study variation of cytosolic Ca2+ concentrations before and during QA treatment. Using such ratio measurement, we showed that basal levels of cytosolic Ca2+ concentrations were similar in control and 3NP-treated cells (Fig. 8C and D
). Perfusion of control striatal cultures with 5 mM QA produced a rapid rise in cytosolic Ca2+ that remained relatively stable thereafter. In cultures that were pretreated with 75 µM 3NP, the QA-induced rise in cytosolic Ca2+ was significantly higher than that seen in cells not treated with 3NP (Fig. 8C and E
), suggesting that partial blockade of complex II reduced the ability of neurons to cope with cytosolic Ca2+ signals. Ca2+"deregulation" (i.e., sudden and massive elevation of cytosolic Ca2+ concentrations) as seen by others in studies where glutamate rapidly induces excitotoxicity (see, for example, refs 47
,48
) did not occur within the 2 h of the experiments, consistent with the observation that QA produced a relatively slow excitotoxic death in our cultures.
We finally investigated whether potentiation was associated with an increase in 45Ca2+ entry. Control cells showed low basal levels of 45Ca2+ accumulation (Fig. 9
A). Preincubation with the NMDA-channel inhibitor MK801 did not markedly affected basal Ca2+ entry. Incubation with QA produced marked elevation in 45Ca2+ accumulation, which was totally inhibited by MK801 (Fig 9A
). Nontoxic concentrations of 3NP (75 µM) did not significantly affect basal Ca2+ accumulation in cells. Incubation with 5 mM QA increased 45Ca2+ accumulation by a factor of seven over that in basal conditions in control cells (Fig. 9B
). In 3NP-treated cells, QA increased 45Ca2+ accumulation by a factor of only three with respect to control conditions (Fig. 9B
). Net 45Ca2+ efflux evaluated at the end of the period of accumulation after stimulation by QA was lower in 3NP-treated cells than in untreated cells (mean±SE, 3NP+QA, 0.75±0.18 103 cpm/mg of protein; QA, 1.17±0.12 103 cpm/mg of protein; P<0.025). Relative efflux (extruded/accumulated ratio) was not statistically different in QA-treated cells with or without 3NP (mean±SE; 3NP+QA, 20.53±1.29%; QA, 16.51±0.58%). This indicated that the reduction in 45Ca2+ accumulation produced by 3NP plus QA as compared with QA alone was therefore not due to increased efflux, but reduced QA-induced entry of 45Ca2+ into neurons.
|
These results show that the exacerbation of QA toxicity induced by complex II inhibition was therefore not associated with NMDA-R hypersensitivity.
DISCUSSION
The present results identify three key aspects of the potentiation of NMDA-R-mediated excitotoxicity by chronic defects in the respiratory chain. First, potentiation occurs for a restricted range of complex II defects in vivo and in vitro. Second, potentiation leads to exacerbation of the activation of the protease calpain, indicating Ca2+ deregulation in vivo. Third, Ca2+ deregulation associated with potentiation is likely not linked to NMDA-R hypersensitivity.
NMDA-R excitotoxicity is exacerbated for a particular "window" of respiratory chain inhibition
This study demonstrates that NMDA-receptor-mediated excitotoxicity can be strongly increased in vivo by 3NP-induced partial blockade of the respiratory chain, which is not overtly toxic in itself. It also shows that 3NP significantly decreases, by a factor of up to two, the threshold dose of QA required to trigger striatal lesions. Thus, nontoxic NMDA-R activation combined with nontoxic mitochondrial complex II defects can synergistically trigger striatal degeneration. This result is consistent with the results of previous studies in cultured neurons, in which the toxicity of low doses of NMDA or glutamate was increased by prior treatment with 3NP (13
,16
,49)
. Only two pioneering studies have demonstrated that this synergy operates in vivo by providing evidence that the size of striatal lesions produced by NMDA is significantly increased by injecting the reversible complex II inhibitor malonate into the striatum (15)
or acute intraperitoneal injection of 3NP (14)
. These studies did not investigate the relationship between the magnitude of potentiation and the concentration of complex II inhibition. The results presented here are novel in that they show that QA toxicity is exacerbated for a particular "window" of complex II inhibition. Indeed, no potentiation was detected for levels of inhibition <35%. Once this threshold had been exceeded, QA toxicity rapidly increased with the concentration of complex II inhibition through an almost "all-or-nothing" process. In this case, mitochondria may be in a particular state rendering striatal cells highly vulnerable to excitotoxic stress.
Possible mechanisms leading to calpain activation and Ca2+ deregulation during potentiation
We show here by fluorimetric proteolysis assays and Western blotting that QA activates calpain more strongly in 3NP-treated rats than in untreated rats. Doses of 3NP potentiating QA toxicity when applied alone did not activate calpain. We obtained similar results in cultured striatal cells, although the amplitude of the potentiation of QA-induced cell death and calpain activation by 3NP were found of limited amplitude compared with the in vivo situation. As the deleterious activation of calpains requires high concentrations of Ca2+ (42)
, our results suggest that the potentiation of QA toxicity by 3NP is associated with a major increase in cytosolic Ca2+ concentrations in vivo. This hypothesis is consistent with the present intracellular Ca2+ imaging study with Fura-2 and studies performed earlier by others (16
,17)
showing that the activation of NMDA-R leads to higher elevation in cytosolic concentrations of Ca2+ in 3NP-treated cultured neurons and brain slices as compared with untreated preparations.
There are three main mechanisms that could explain abnormal Ca2+ concentration increases within a neuron (50)
: 1) increases in Ca2+ entry, 2) decreases in Ca2+ extrusion, and 3) changes in the sequestration of Ca2+ to intracellular stores.
The potentiation of QA toxicity by 3NP may depend on an increase in Ca2+ entry into striatal neurons. Indeed, according to the hypothesis of "indirect excitotoxicity" in energy-deficient situations, ATP availability is reduced and the plasma membrane Na+/K+-ATPase cannot maintain the resting membrane potential. Partial depolarization renders NMDA-R hypersensitive to agonists by relieving the voltage-dependent Mg2+ blockade of the cation channel. In the presence of even low (physiological) concentrations of glutamate, this would lead to an increase in Ca2+ influx and cell death. Supporting this hypothesis, the neurotoxic effects of many mitochondrial toxins are reduced by drugs blocking the NMDA-R, such as MK-801 (5
,19)
. Electrophysiological studies on brain slices have also shown that strong complex II inhibition by 3NP (>75%) leads to a decrease in plasma membrane resting potential, resulting in the hypersensitivity of NMDA-R to agonists (17)
.
Our results showed that NMDA-R did not seem to become hypersensitive to QA in 3NP-treated rats. We directly showed that the QA-induced influx of 45Ca2+ into cultured striatal neurons was decreased rather than increased by 3NP, suggesting the absence of a hypersensitivity of NMDA-R. As a similar analysis of Ca2+ entry could not be performed in vivo, we used an indirect index of NMDA-R activation. One of the primary effects of NMDA-R activation in the striatum is ionic disturbance and an increase in firing/discharges (51)
. NMDA-R activation leads to an influx of Na+ and Ca2+, increasing the activity of the plasma membrane Na+/K+-ATPase, which accounts for a large proportion of the ATP used in neurons (44)
. Consistent with this, we found that Glc uptake increased markedly in the striatum after QA injection. We found that 3NP did not enhance the QA-induced increase in Glc consumption, indicating that the primary effect of NMDA-R activation by QA (ionic imbalance across the plasma membrane) remains essentially unchanged by 3NP-induced complex II inhibition in vivo. This is only indirect evidence that, in vivo, NMDA-R sensitivity is not modified by 3NP treatment. However, the in vitro experiments we carried out more directly demonstrate that the exacerbation of the QA-induced cytosolic Ca2+ perturbation by 3NP is not related to increased entry of Ca2+ into the cells. This is consistent with the observation in brain slices, showing that when 3NP-induced complex II inhibition is in the range of 50%, cell death can be triggered in absence of plasma membrane depolarization (52)
.
An alternative mechanism contributing to Ca2+ deregulation is based on a reduced extrusion of Ca2+ from the cell. The efficiency of the Ca2+-ATPase of the plasma membrane may be reduced if ATP stores are decreased by 3NP. However, this is unlikely in the early phase of synergy. Indeed, we show in the present study that in conditions for which 3NP inhibits SDH and potentiates QA toxicity, ATP levels are essentially normal. In addition, we show using 45Ca2+ that Ca2+ efflux is not modified by 3NP. This is consistent with previous studies that have strongly suggested that ATP levels are not markedly modified by 35 to 50% inhibition of complex II in vivo (see Discussion in ref 27
). Another possible mechanism involving a decrease in Ca2+ extrusion in 3NP-treated rats is calpain-mediated cleavage of the type 3 Na+/Ca2+ exchanger (NCX3) present in the striatum (47)
. However, as our results show that subtoxic doses of 3NP in themselves do not enhance cytosolic Ca2+ concentrations, activate calpain, or increase Ca2+ accumulation into neurons, the role of NCX3 cleavage in Ca2+ deregulation during potentiation is likely secondary to altered capacity of neurons to sequester Ca2+ within intracellular stores.
Finally, complex II inhibition may modify the sequestration of Ca2+ entering through NMDA-R in intracellular pools. As the capacity of mitochondria to buffer Ca2+ in vitro is dependent on the energy substrates used to energize mitochondria (53)
and is reduced by 3NP (54)
, it is likely that subtoxic 3NP treatment modifies the ability of mitochondria to properly sequester Ca2+ entering through NMDA-R, leading to an elevation of the cytosolic Ca2+ concentrations. Thus, even if less Ca2+ enters into 3NP-treated neurons, the cytosolic concentrations of the cation will sufficiently increase to initiate calpain activation and the cascade of deleterious events leading to cell death. The probable role of mitochondria in the deregulation of Ca2+ homeostasis during potentiation is further supported by the present findings. Indeed, we showed that synaptic proteins (including PSD-95, NR2B, fodrin, and huntingtin) are massively cleaved by calpain in 3NP-treated rats receiving QA. In striatal cells in vivo, these proteins are preferentially localized in dendrites. As dendrites are remote from the reticulum and nucleus (2 important pools of Ca2+) but contain many mitochondria, calpain-dependent cleavage of synaptic proteins indicates preferential Ca2+ deregulation in vicinity of mitochondria.
Thus, the main mechanism of "weak/indirect" excitotoxicity is associated with massive calpain activation, likely resulting from a decrease in the ability of neurons to cope with cytosolic increase in Ca2+ concentration rather than an exacerbated increase in Ca2+ entry.
Possible implications for neurodegenerative diseases
These in vivo results provide support for the view that mitochondrial defects and NMDA-R activation may have synergistic effects in acute and chronic neurodegenerative disorders, such as HD in particular. Indeed, mutated huntingtin increases the sensitivity of NMDA-R to glutamate (20
21
22
,55
,56)
and partial mitochondrial defects have been reported in HD (23
24
25
,57
,58)
. Our data suggest that intracellular Ca2+ anomalies in HD (59)
, possibly leading to calpain activation in the striatum of patients (60)
, may result from a combination of moderate NMDA-R and mitochondrial defects.
ACKNOWLEDGMENTS
We thank Dr. Philippe Gervais for providing FDG. We would also like to thank Philippe Champeil (DBJC, CEA Saclay) for help with 45Ca2+ uptake experiments and Dr. Raphael Boisgard (ERIT-M 0103, CEA Orsay) for help in setting up the ATP assay. This work was supported by the CEA and the CNRS. C. Jacquard holds a Ph.D. fellowship from the CEA. Y. Trioulier holds a postdoctoral fellowship from Hereditary Disease Foundation.
Received for publication September 16, 2005. Accepted for publication December 22, 2005.
REFERENCES
P < 0.0001 vs. QA+MK801. #P < 0.001 vs. QA, ANOVA, and Fisher’s post hoc PLSD test.