Up-regulation of mitochondrial uncoupling ... - The FASEB Journal

3 downloads 0 Views 1MB Size Report
Sep 18, 2003 - Luc Dupuis,* Franck di Scala,* Frédérique Rene,* Marc de Tapia,* ... Neurodégénérescence, EA 3433, Université Louis Pasteur, Faculté de ...
The FASEB Journal express article 10.1096/fj.02-1182fje. Published online Sept. 18, 2003.

Up-regulation of mitochondrial uncoupling protein 3 reveals an early muscular metabolic defect in amyotrophic lateral sclerosis Luc Dupuis,* Franck di Scala,* Frédérique Rene,* Marc de Tapia,* Hugues Oudart,† Pierre-François Pradat,‡ Vincent Meininger,‡ and Jean-Philippe Loeffler* *Laboratoire de Signalisations Moléculaires et Neurodégénérescence, EA 3433, Université Louis Pasteur, Faculté de Médecine, 67085 Strasbourg Cedex; †CEPE, CNRS, 67087 Strasbourg; and ‡ Service de Neurologie, Hôpital de la Pitié Salpêtrière, 75651 Paris Cedex 13, France Corresponding author: Jean-Philippe Loeffler, Laboratoire de Signalisations Moléculaires et Neurodégénérescence, EA 3433, Université Louis Pasteur, Faculté de Médecine, 11 rue Humann, 67085 Strasbourg Cedex, France. E-mail: [email protected] ABSTRACT Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disorder affecting primarily motor neurons. Growing evidence suggests a mitochondrial defect in ALS. The precise molecular mechanisms underlying those defects are unknown. We studied the expression of mitochondrial uncoupling proteins (UCPs), key regulators of mitochondrial functions, in tissues from a mouse model of ALS (SOD1 G86R transgenic mice) and from muscular biopsies of human sporadic ALS. Surprisingly, in SOD1 G86R mice, UCPs, and particularly UCP3, were upregulated in skeletal muscle but not in spinal cord. Consistent with this pattern of expression, ATP levels were selectively depleted in muscle but not in neural tissues 1 month before disease onset and the respiratory control ratio of isolated mitochondria is decreased. UCP3 up-regulation was not observed in experimentally denervated muscles, suggesting that changes in muscular UCP3 expression are associated with the physiopathological processes of ALS. This is further supported by our observation of increased UCP3 levels in human ALS muscular biopsies. We propose that UCP3 up-regulation in skeletal muscle contributes to the characteristic mitochondrial damage of ALS and to the onset of the disease. Moreover, since skeletal muscle is a key metabolic tissue, our findings suggest that ALS may not solely arise from neuronal events but also from more generalized metabolic defects. Key words: mitochondria • neurodegeneration • motor neuron • metabolism

A

myotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disorder affecting specifically the motor system. The death of spinal motor neurons leads to limb weakness with muscle atrophy and to a progressive involvement of respiratory muscles leading to death within 3 to 5 years after the onset (1, 2). Several mechanisms have been put forward to explain motor neuron degeneration, including oxidative stress, glutamate excitotoxicity, and apoptosis, suggesting that the disease is multifactorial in origin (3). However, the precise molecular mechanisms responsible for the selective loss of motor neurons remain obscure. Five

to 10% of ALS cases are dominantly inherited, and genetic linkage studies revealed that 20% of these are associated with point mutations in the gene encoding cytosolic Cu/Zn superoxide dismutase (SOD1; ref 4), a major free radical scavenging enzyme that protects cells against oxidative stress (5). The identification of such mutations has allowed the production of transgenic mice that develop neurological disorders characteristic of ALS (6–8). It has been demonstrated that the deleterious mechanism of these mutations is related to a gain of function (9) possibly mediated by the mitochondrial fraction of the mutated SOD1 (10, 11). A growing body of evidence involves mitochondrial dysfunction in ALS: morphological abnormalities of mitochondria (swelling) are very early events in ALS, (12) and several reports demonstrated a decrease in mitochondrial DNA as well as in respiratory chain enzyme activities both in ALS patients and in ALS transgenic mice (13–16). Furthermore, the involvement of mitochondria in ALS pathology is reinforced by the observation that supplementation of creatine, a mitochondrial energy buffering agent, is the most efficient treatment known to delay motor dysfunction and extend survival of ALS mice (17). The mitochondrial respiratory chain is the major site of superoxide production, and there is increasing evidence for the hypothesis of an oxidative stress-related mitochondrial involvement as a key determinant of motor neuron degeneration. Increased markers of lipid and protein damage due to oxidative stress have been detected in mitochondria of ALS mice (13). These findings support the concept that the mitochondrial fraction of SOD1 might gain its toxic effect via an aberrant catalytic function leading to excessive production of free radical species (13). In addition, markers of oxidative damage to protein, lipid, and DNA have been detected in the nervous system of both sporadic and familial ALS (18–21). Uncoupling proteins (UCPs) are members of the growing family of mitochondrial carrier proteins (22). The founding member of UCP family is UCP1, formerly known as thermogenin. UCP1 is almost exclusively expressed in the brown adipose tissue (BAT) and is responsible, through the diversion of energy from ATP synthesis to thermogenesis, for the thermogenic function of this tissue, especially in rodents. UCP2, -3, -4, and -5 (also known as BMCP1) were further cloned by homology with UCP1, and their thermogenic activity in vivo is still controversial. Mitochondria from UCP3 knockout mice display a higher conductance rate although no changes are noted in the overall respiratory rate of the animal (23, 24). Such data suggest that UCPs might have a function in the fine regulation of mitochondrial respiration, and it was recently suggested that this function deals with the resistance to oxidative stress (24, 25). Furthermore, it was very recently demonstrated that UCP2 controlled a particular form of cell death called oncosis (26). This shows that UCPs are linked to oxidative stress resistance, mitochondrial function, and ultimately cell survival, these three functions being defective in ALS. UCPs are thus candidates for possible effectors in ALS, and we decided to monitor their expression pattern in ALS. MATERIALS AND METHODS Animals and tissue preparation We used transgenic male mice with the missense mutation Gly86Arg (G86R, human G85R equivalent) in the SOD1 gene (7) that were 75, 90, and 105 days old. Nontransgenic littermates served as wild-type controls and for denervation experiments. All mice were from the FVB/N

strain. Peripheral nerve injury in normal male mice was accomplished by either crush or axotomy of sciatic nerve as described previously (27, 28). Sham-operated mice served as control. We assessed motor dysfunction by monitoring fingerprint traces after dipping hindleg extremities in ink. Lumbar spinal cords and gastrocnemius muscles were dissected either from transgenic mice or sciatic nerve-injured mice. Tissues were immediately frozen in liquid nitrogen and stored at -80°C until use. Animal manipulation followed current European Union regulations. Human tissue preparation We obtained muscle samples from patients submitted to either a standard surgical muscle biopsy under local anesthesia or an orthopaedic operation under general anesthesia. All patients gave written informed consent. Tissues were immediately frozen in liquid nitrogen and stored at 80°C until use. The ALS group consisted of 21 patients (7 for RT-PCR studies, 14 for Western blot studies) with probable or definite sporadic ALS according to the World Federation of Neurology criteria (29) and who were attending the ALS Center of the Pitié Salpétrière Hospital (Paris, France). Control groups included the following: 10 patients without any significant neurological history who underwent an orthopaedic surgery (Western blot studies), 6 patients with symptomatic sensorimotor polyneuropathies, and 3 patients with primary muscle diseases (RT-PCR studies). Phylogenetic analysis of UCP family Proteins homologous to UCP1 were defined using a PSI-BLAST algorithm (NCBI, Bethesda, MD). The 15 most highly homologous proteins were further selected, and their sequences were aligned using CLUSTALW algorithm (Infobiogen). The results were further analyzed using PHYLIP software and Fitsch-Margolias algorithm. Tree was rooted using the hemoglobin αchain as root. RT-PCR Two micrograms of total RNA was reverse transcribed using 200 U MoMuLV reverse transcriptase and 0.5 µg random primers (Promega, Charbonnieres, France) as described previously (30). After PCR amplification, products were separated with a 2% agarose gel, and visualized by ethidium bromide staining. The degree of denervation in transgenic and sciatic nerve-injured mice was assessed by monitoring the levels of nicotinic acetylcholine receptor-α subunit (AchRα) mRNA (31). We used either glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA levels or 18S rRNA levels as internal control. PCR primers were as follows: murine UCP1 5′ primer: 5′-GCCTTCAGATCCAAGGTGAA-3′; murine UCP1 3′ primer: 5′-TAAGCCGGCTGAGATCTTGT-3′; murine UCP2 5′ primer: 5′-GGGGAGAGTCAAGGGCTAGT-3′; murine UCP2 3′ primer: 5′-GCTCTGAGCCCTTGGTGTAG-3′; murine UCP3 5′ primer: 5′-GGCCCAACATCACAAGAAAT-3′; murine UCP3 3′ primer: 5′-GCGTTCATGTATCGGGTCTT-3′; murine UCP4 5′ primer 5′-CTTGCTAGGTTGGGAGATGG-3′;

murine UCP4 3′ primer 5′-TCACTTTTGCCAAACACGAC-3′; murine BMCP1 5′ primer: 5′-TTGATGGTGCCATATGATGC-3′; murine BMCP1 3′ primer: 5′-CGGCAGTGATTCATCAGAAA-3′; human and murine GAPDH 5′ primer: 5′-ACCACAGTCCATGCCATCAC-3′; human and murine GAPDH 3′ primer: 5′-TCCACCACCCTGTTGCTGTA-3′; and human UCP3: 5′ primer: 5′-ATTTCAGGCCAGCATACACC-3′ and 3′ primer: AGTCCAGCAGCTTCTCCTTG-3′

5′-

Anti-UCP3 Western blot Tissues were homogenized in PBS containing 1% Nonidet P40, 0.5% sodium deoxycholate, 0.1% SDS, 0.5% aprotinine, 5 mM DTT, and 1 mM PMSF. Homogenates were then boiled for 5 min and sonicated for 30 s. After centrifugation, equal amounts of protein, according to Bradford protein assay (Bio-Rad), were electrophoresed through a 13% SDS-polyacrylamide gel. Separated proteins were then electrotransferred to nitrocellulose membranes and immunostained with the 3046 anti-UCP3 antibody (Chemicon) diluted 1/1000. Membranes were then incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (Pierce, Bezons, France) diluted 1/2000 and developed by enhanced chemiluminescence detection. To ensure equal loading, membranes were stained with Ponceau S stain (Sigma). Mitochondrial oxygen consumption For each experiment, total hindlimb muscles of two SOD1 G86R and two control littermate mice were pooled by genotype, minced with scissors, and gently homogenized using a glass-Teflon potter in a medium (pH=7.4) containing (mM): 250 sucrose, 20 TES, 1 EDTA, and 1% (w/v) of bovine serum albumin. Homogenates were centrifuged at 800 g, pellets were resuspended in the same medium and centrifuged again at 800 g, and then supernatants were pooled and centrifuged at 10,000 g for 20 min. Mitochondrial pellets were washed in the previous buffer devoid of BSA and resuspended in the same buffer to yield a final protein concentration of 10-15 mg/ml. All steps were performed at 4°C. Oxygen consumption was measured polarographically using a Clark-type electrode (Hansatech, Norfolk, UK) with the chambers maintained at 37°C and continuously stirred. One-hundred and fifty micrograms of mitochondrial proteins were diluted in 0.3 ml of medium (pH=7.2) containing (mM): 250 sucrose, 10 TES, 5 potassium phosphate, 1 EDTA, 2 magnesium chloride, and 0.2% (w/v) of bovine serum albumin in the presence of 5 µM of rotenone (inhibitor of complex I). All the experiments were performed using 10 mM succinate as the substrate (state 2 respiration). Three-hundred micrograms of ADP were further added to yield state 3 respiration. Each measure was performed at least in duplicate. Cytochrome c oxydase activity was then measured polarographically using the same Clark-type electrode. Briefly, 0.6 mg/ml of cytochrome c solubilized in Tris-HCl 4.10−2 M, pH = 7.4, 37°C was reduced by addition of 1.2 mg/ml of sodium dithionite. Then the excess of dithionite was removed by bubbling with air. Twenty-five micrograms of mitochondrial proteins were added to 500 µl of reduced cytochrome c solution.

ATP levels measurements ATP levels were measured using ATP bioluminescence assay kit CLS II (Roche) and a FB12 Berthold luminometer following the recommendations of the provider. Data were standardized vs. the total protein content using a BCA kit (Roche). Statistical analysis Data are means ± SE. Statistical significance was determined by unpaired Student's t test. Paired Student’s t test was used to analyze mitochondrial oxygen consumption data as indicated in the figure legends. RESULTS UCPs family pattern of expression in tissues involved in ALS To date, murine UCP1 to 4 have been cloned. We identified murine BMCP1 sequence as RIKEN cDNA NM028711 by homology with human BMCP1 mRNA. To further clarify the evolutionnary relationships between UCPs and mitochondrial carrier proteins, we built a phylogenic tree of murine mitochondrial carrier proteins (Fig. 1A). UCP1, -2, and -3 are very close relatives. BMCP1 and UCP4 are more distant from UCP1 but share a significant homology. On the contrary, other mitochondrial carrier proteins such as adenine nucleotide translocator proteins (ANT) displayed lower homology and are evolutionnary more distant. We decided to study the five mitochondrial carrier proteins more closely related to UCP1 namely UCP1, -2, -3, -4, and BMCP1 (Fig. 1A). Using semiquantitative RT-PCR, we studied the expression pattern of UCPs in tissues primarily affected in ALS (lumbar spinal cord and skeletal muscle) of wild-type animals and extended those studies to liver, BAT, and white adipose tissues (WAT) as control tissues. As expected, UCP1 is primarily expressed in BAT and hardly detectable in the other tissues tested. UCP2 is mainly expressed in WAT but is also highly expressed in the spinal cord. UCP3 expression is high in the skeletal muscle and very low in BAT, WAT, and spinal cord. UCP4 and BMCP1 were only hardly detectable in the spinal cord (Fig. 1B). UCPs pattern of expression in ALS mice tissues We next studied the expression levels of UCPs in tissues of ALS-affected mice carrying a mutated form of SOD1 (G86R mice) using semiquantitative RT-PCR. We focused on three tissues: liver regarding its key importance in energy homeostasis and spinal cord and skeletal muscle because they are primary targets of ALS disease. We measured UCP expression levels at three different ages (75, 90, and 105 days of age) in both G86R mice and control littermates. At 75 and 90 days of age, G86R mice were healthy, with no obvious locomotor symptom. End stage disease set up at ~105 days of age. We assessed the expression levels of the five UCPs along with GAPDH as a control in the liver and found that neither UCP1, -3, -4, nor BMCP1 was significantly expressed in the liver of G86R or control littermates animals. As expected, UCP2 expression was detectable in the liver, but its expression levels were unchanged in G86R animals at the three ages tested (data not shown). UCP2, -3, -4, and BMCP1 were expressed at various

levels in both wild-type and G86R lumbar spinal cords, but their expression did not change throughout the disease (data not shown). On the contrary, in skeletal muscle we found that UCP2 expression was heavily increased at 105 days of age while UCP3 mRNA was upregulated at 90 and 105 days of age (Fig. 2). In this tissue, UCP1 and BMCP1 expression were not detectable, while UCP4 mRNA levels were low and did not show any variation throughout the disease (data not shown). UCP2 but not UCP3 up-regulation is linked to denervation processes We and others have previously observed that characteristic features of denervation processes occur in G86R animals before the onset of locomotor symptoms. We asked whether UCP2 and UCP3 changes in expression in G86R muscles were part of the typical pattern of expression occurring in response to denervation. As expected, sciatic nerve crush or axotomy provoked hindlimb disabilities (Fig. 3A), axonal degeneration (Fig. 3B), and high levels of AchRα mRNA (Fig. 3C) indicative of denervation. RT-PCR analysis revealed a significant up-regulation of UCP2 mRNA both in sciatic nerve crushed and in axotomized (data not shown) animals, thus reproducing the situation observed in G86R animals (Fig. 3C and D). This was not the case for UCP3 expression, which was down-regulated in nerve-crushed and axotomized animals, a situation opposite to the one observed in G86R transgenic mice. UCP3 up-regulation in G86R mice is therefore not linked to denervation processes. UCP3 pattern of expression in ALS mice is specific to the mutation of SOD1 rather than to overexpression of SOD1 To determine whether the pattern of UCP2 and UCP3 expression in muscle of G86R animals is due to the overexpression of a superoxide dismutase rather than to expression of a mutated form of this enzyme, semiquantitative RT-PCR analyses were performed in C57Bl/6xDBA/2 mice overexpressing the human wild-type SOD1. No variations in muscular UCP2 expression were noted, whereas a sharp down-regulation of UCP3 mRNA was observed (Fig. 3E and F). These data strongly suggest that the early up-regulation of UCP3 mRNA in the skeletal muscle of G86R animals is specifically linked to the expression of the ALS-associated mutant enzyme. Muscular UCP3 up-regulation is associated with human ALS We next assessed the expression levels of UCP3 in human ALS biopsies. As shown in Fig. 4A, total levels of UCP3 mRNA were increased in the seven ALS biopsies tested as compared with control patients affected either by neurogenic denervation diseases (polyneuropathies, n=6) or primary muscle diseases (n=3). Furthermore, UCP3 protein levels were massively increased in 14 tested ALS biopsies as compared with normal subjects as demonstrated by Western blotting in Fig. 4C. Collectively, these data show a massive dysregulation of UCP3 in the skeletal muscle of sporadic ALS. Mitochondrial dysfunction in skeletal muscle tissue of ALS mice To determine whether the detected UCP2 and UCP3 up-regulations in skeletal muscle tissue of ALS mice were able to lead to a functional uncoupling, we isolated skeletal muscle mitochondria of ALS and control mice and measured their oxygen consumption. No significant changes in

oxygen consumption were observed either in state 2 (substrate alone) or in state 3 respiration (substrate and ADP; data not shown). However, a significant decrease in the respiratory control ratio (RCR, defined as the ratio between state 3 and state 2 respiration) was observed (Fig. 5A). It should be noted that this decrease in RCR was not associated with a decrease in mitochondrial respiratory chain enzyme activities because cytochrome c oxidase (cox) activity measured in the same mitochondrial preparations was unchanged between ALS and control mice (Fig. 5B). These data suggest that mitochondria of ALS mice skeletal muscles are more uncoupled than mitochondria of their wild-type counterparts. To determine whether this mitochondrial dysfunction leads to a decrease of ATP levels, we measured ATP levels in gastrocnemius muscle, liver, and spinal cord of ALS mice. Measured ATP levels in 105-day-old mice gastrocnemic muscles were 28.8 ± 2.2 µmol ATP/µg of proteins in wild-type mice and 17.9 ± 2.0 µmol ATP/µg of proteins in G86R muscle tissues (n=4; P