Uncoupling Protein 2 Prevents Neuronal Death Including that ...

0013-7227/03/$15.00/0 Printed in U.S.A.

Endocrinology 144(11):5014 –5021 Copyright © 2003 by The Endocrine Society doi: 10.1210/en.2003-0667

Uncoupling Protein 2 Prevents Neuronal Death Including that Occurring during Seizures: A Mechanism for Preconditioning SABRINA DIANO, RUSSELL T. MATTHEWS, PETER PATRYLO, LICHUAN YANG, M. FLINT BEAL, COLIN J. BARNSTABLE, AND TAMAS L. HORVATH Departments of Obstetrics and Gynecology (S.D., T.L.H.), Neurobiology (R.T.M., C.J.B., T.L.H.), Neurosurgery (P.P.), and Ophthalmology-Visual Science (C.J.B.), Yale University School of Medicine, New Haven, Connecticut 06520; and Department of Neurology and Neuroscience, Weill Medical College of Cornell University (L.Y., M.F.B.), New York, New York 10021 The mitochondrial uncoupling protein (UCP2) is expressed in selected regions of the brain. Here we demonstrate that upregulation of UCP2 is part of a neuroprotective set of responses to various cellular stresses in vitro and in vivo. PC12 cells, when transfected with UCP2, were protected against free radical-induced cell death. Seizure activity was associated with elevated UCP2 levels and mitochondrial uncoupling activity. In transgenic mice that expressed UCP2 constitutively in the hippocampus before seizure induction, a robust

M

ITOCHONDRIAL ENERGY HOMEOSTASIS is an important determinant of a cell’s ability to withstand stress. Although adequate levels of ATP are essential for cell survival, mitochondrial energy production also generates free radicals that are toxic. Hence, to minimize cell death it would be advantageous to dissociate cellular ATP production from that of free radicals. Uncoupling proteins (UCPs) can provide a controlled leak of protons across the inner membrane of the mitochondrion and thus uncouple oxidative phosphorylation from respiration, with a concomitant decrease in the inner mitochondrial membrane potential (1– 6) and free radical generation (7, 8). In vivo activation of UCPs that may require fatty acids, cofactors, and superoxides (9 –15) will also alter ATP and calcium pools within neurons, with potential effects on activity and susceptibility to toxic stimuli. UCP2 is the only one of the three closely related, suggested physiological uncouplers (UCP1 to UCP3) (14) that is expressed in many regions of the central nervous system of both rodents and primates (16 –18). UCP2 expression and activity are sensitive to environmental conditions and are increased, for example, after either acute brain injury (19) or irradiation of tumor cell lines (20). However, it is not yet clear whether this is part of a cell death response or a cell rescue response (see Refs. 19 vs. 20). Although increasing expression of UCP2 may be correlated with initiation of apoptotic responses in lymphoma cell lines (20), Nicholls and colleagues (21–24) found that a controlled decrease in neuronal mitoAbbreviations: BAC, Bacterial artificial chromosome; GFP, green fluorescence protein; h, human; NO, nitric oxide; RCR, respiratory control ratio; SSC, standard saline citrate; TBA, thiobarbituric acid; TBARS, TBA-reactive substances; UCP, uncoupling protein.

reduction in cell death was seen. Because UCP2 increased mitochondrial number and ATP levels with a parallel decrease in free radical-induced damage, it is reasonable to suggest that mitochondrial UCPs precondition neurons by dissociating cellular energy production from that of free radicals to withstand the harmful effects of cellular stress occurring in a variety of neurodegenerative disorders, including epilepsy. (Endocrinology 144: 5014 –5021, 2003)

chondrial membrane potential (an event that could occur due to functioning UCP2) can diminish damaging calcium influx during excitotoxic insults or repetitive, high frequency firing, such as seizure activity. In addition, the ability of UCP2 to diminish free radicals in macrophages (8) suggests that if that were the case in neurons as well, UCP2 activation may protect neurons in degenerative processes associated with free radical toxicity, such as occurring during epilepsy (25, 26). In agreement with our proposition that UCP2 may be neuroprotective in the brain (17), we found a negative correlation between activation of an apoptotic signal, caspase 3, and UCP2 expression levels in neurons during acute brain injury (19). We also revealed that in the primate substantia nigra, where UCP2 is expressed (27), elevation of nigral state 4 respiration precedes protection of dopamine cells by a cofactor of UCP2, CoQ10, in a model of Parkinson’s disease (27). Thus, we hypothesize that induction of UCP2 is a general defense mechanism in neurodegenerative processes (28). The goal of our present study was to demonstrate conclusively whether UCP2 expression supports neuronal survival in degenerative processes. We have used a combination of in vitro and in vivo models. An animal model of epilepsy was chosen in which, before the insult, UCP2 induction could be accomplished. The data reveal that increasing UCP2 expression before the onset of increased cellular stress robustly enhances the survival of neurons. Materials and Methods PC12 cells PC12 cells were maintained in Ultraculture medium (BioWhittaker, Walkersville, MD) with 10% fetal bovine serum, 2 mm glutamine, and 1 ␮g/ml gentamicin. Cultures in 96-well plates were transfected with

5014

Diano et al. • UCP2 Prevents Neuronal Death

500 ng green fluorescence protein (GFP; Pway, a gift from Dr. T. Hughes) or 500 ng GFP plus 500 ng human UCP2 (hUCP2) (2) expression plasmids using Lipofectamine 2000 (Life Technologies, Inc., Grand Island, NY). After 2–3 d cells were washed with serum-free medium and exposed for 3 h to either 50 ␮m H2O2 or 1–5 mm sodium nitroprusside as a source of nitric oxide (NO). Cell survival was scored as the number of fluorescing cells in three random microscopic fields per well and was expressed as a percentage of untreated wells. Expression of hUCP2 in cells was tested by RT-PCR using isoform-specific primers.

Animals For the rodent studies, C57BL/6J male mice were used. Transgenic animals were produced as described previously (29). In short (described in great detail in Ref. 29), an 80-kb human bacterial artificial chromosome (BAC) clone containing the natural genes and promoters was used to produce transgenic mice that slightly overexpress UCP2 and -3. Transgenic animals and wild-type littermates were used from the sixth or higher backcross generation. All experimental procedures were approved by the institutional animal care and use committee of Yale University.

Induction of epileptic seizures Animals were injected ip with 1 mg/kg scopolamine, followed 30 min later by ip pilocarpine (280 mg/kg) or vehicle. Animals and their seizure patterns were closely monitored in the following 24 h. Observers were blinded to the experimental group. These animals were killed 24 h after the induction of seizures.

Analysis of UCP2 mRNA RT-PCR was used to assess UCP2 mRNA levels in seized and control animals and transgene (hUCP2 and hUCP3) expression in various brain regions. Total RNA was extracted, and cDNA was prepared as previously described (30). The primers used to amplify both mouse and hUCP2 were 5⬘-CTACAAGACCATTGCACGAGAGG-3⬘ (forward) and 5⬘-AGCTGCTCATAGGTGACAAACAT-3⬘ (reverse) and gave a 396-bp product. The primers used to detect hUCP2 were 5⬘-CCGTGAGACCTTACAAAGC-3⬘ (forward) and 5⬘-GGGAGCCTCTCGGGAAGTG-3⬘ (reverse) and gave a 1046-bp product. The primers used to detect hUCP3 were 5⬘-AGGCGGTCCAGACGGCCCGGCT-3⬘ (forward) and 5⬘GGGGCTGAAGTACTGGCCTGGAGGT-3⬘ (reverse). A 603-bp fragment of ␤-actin was amplified as a control (30). The PCR reaction was performed as previously described (17, 19, 29). Reactions were analyzed by gel electrophoresis, and the identity of the bands was confirmed by sequencing of selected samples.

Detection of UCP2 in the hippocampus of transgenic animals In situ hybridization for UCP2 was performed as described previously (17). The PCR product was inserted in Bluescript vector and subcloned. Linearized DNA was transcribed using T7 polymerase (antisense cRNA probe) and T3 polymerase (sense cRNA probe; Riboprobe Combination System T3/T7, Promega, Madison, WI). The radiolabeled cRNA probe was purified by passing the transcription reaction solution over a G-50 column (Amersham Pharmacia Biotech, Piscataway, NJ), and fractions were collected and counted using a scintillation counter. The purified cRNA probes were heated at 80 C for 2 min with 500 mg/ml yeast tRNA and 50 mm dithiothreitol in water before being diluted to an activity of 5.0 ⫻ 107 dpm/ml with hybridization buffer containing 50% formamide, 0.25 m sodium chloride, 1⫻ Denhardt’s solution, and 10% dextran sulfate. The brains of wild-type (n ⫽ 3) and transgenic (n ⫽ 3) animals were perfused and frozen on dry ice. The frozen brains were allowed to equilibrate in a cryostat at ⫺20 C. Coronal sections containing the hippocampus were cut at 16 ␮m and then mounted onto poly-llysine-coated slides. Sections with this hybridization solution (150 ␮l/ slide) were incubated overnight at 50 C. After hybridization, the slides were washed four times (10 min each) in 4⫻ standard saline citrate (SSC) before ribonuclease digestion (20 ␮g/ml for 30 min at 37 C) and rinsed at room temperature in decreasing concentrations of SSC that contained 1 mm dithiothreitol (2, 1, and 0.5⫻; 10 min each) to a final stringency of

Endocrinology, November 2003, 144(11):5014 –5021 5015

0.1⫻ SSC at 65 C for 30 min. After dehydration in increasing alcohols, the sections were exposed to ␤-max Hyperfilm (Amersham) for 5 d before being dipped in Kodak (Eastman Kodak, Rochester, NY) NTB-2 liquid emulsion diluted 1:1 with distilled water. The dipped autoradiograms were developed 21 d later with Kodak D-19 developer and fixed, and the sections were counterstained through emulsion with hematoxylin. Sections were examined under bright- and darkfield illumination. As a control experiment, sections were incubated as described above with hybridization solution containing the sense strand probe synthesized with T3 polymerase to transcribe the coding strand of the DNA insert. Hippocampi of transgenic and wild-type mice were immunostained with UCP2 antisera using the avidin-biotin peroxidase method in which diaminobenzidine was used to visualize tissue-bound peroxidase (17– 19). We used different antisera, including one (C-20, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) that we have previously observed to label a specific band around 33 kDa. This labeling was missing in UCP2 knockout animals, supporting the specificity of this antiserum (31).

Analysis of uncoupling activity To test the level of mitochondrial uncoupling, oxygen consumption in state 3 and state 4 respiration and the respiratory control ratio (RCR: the ratio of these levels) were assessed as described previously (29, 31) from hippocampi of mice that were either intact or underwent seizures. In short, dissected hippocampi of control and seized animals (24 h after the initiation of seizures) were homogenized in a medium containing 320 mm sucrose, 10 mm Tris, and 1 mm EGTA adjusted to pH 7.4 with HCl. The samples were diluted 10 times with isolation medium. Oxygen consumption was determined using a Clark-type oxygen electrode in an incubation medium containing 80 mm KCl, 50 mm HEPES (pH 7), 1 mm EGTA, 5 mm K2HPO4, 4 ␮m rotenone, 80 ng/ml nigericine, and 1 ␮g/ml oligomycin using a saturating amount of succinate as substrate at 37 C. State 3 respiration was initiated by the addition of ADP (300 ␮m), and subsequently, state 4 and 3 respiration and RCR were assessed.

Histological analyses Twenty-four hours after the initiation of seizures, groups of animals were transcardially perfused under ether anesthesia with saline, followed by fixative containing 4% paraformaldehyde, 15% picric acid, and 0.1% glutaraldehyde. The unbiased stereological methods for cell number assessment used a systematic random sampling method and met the statistical requirements necessary to ensure an unbiased estimate of the feature of interest. We chose to sample every 10th section (30 ␮m thick) to assess cell number; the starting point of the series for each animal was randomly selected from the first 10 sections. Each electron micrograph of 1 section covered 1,800 ␮m2, and each animal provided 8 samples. Thus, cell number was calculated for each experimental group as number per 14,400 ␮m2. Calculations of mitochondrial number were performed in the stratum radiatum and the principal layer of the CA1 region of the hippocampus using unbiased ultrastructural analyses. In this, to obtain a comparable measure of cell and mitochondria numbers, unbiased for possible changes in size, the dissector technique was used. The number of mitochondria was counted in parallel consecutive serial sections with the help of a reference grid superimposed on the electron microscope photographs. The dissector volume (volume of reference) was the area of the reference grid multiplied by the distance between the upper faces of the reference and look-up sections. Section thickness was determined using Small’s fold method. Because an F test analysis of cell and mitochondria counts in different experimental and control groups revealed a significant nonhomogeneity of variances between groups, the Kruskal-Wallis one-way nonparametric ANOVA test was selected for statistical comparisons. A level of confidence of P ⬍ 0.05 was adopted.

Biochemical measurements Assay for thiobarbituric acid (TBA)-reactive substances (TBARS). To assess tissue levels of superoxides in the cortex and hippocampus, the TBARS assay was employed as described previously (32). Homogenates of

5016

Endocrinology, November 2003, 144(11):5014 –5021

dissected areas (1:10, wt/vol) were prepared immediately before use in ice-cold aqueous 20 mm Tris-HCl buffer, pH 7.4, using a Teflon pestle motor homogenizer. Lipid peroxidation was stimulated in assays containing 100 ␮l homogenate by the addition of 10 ␮l 100 ␮m Fe3⫹ (ammonium ferric sulfate) solution, and the mixture was incubated for 30 min at 37 C. The reaction was stopped by the addition of sodium dodecyl sulfate [100 ␮l of an 8.1% (wt/vol) solution] and 750 ␮l 20% acetic acid (adjusted to pH 3.5 using NaOH). Precipitated proteins were then removed by centrifugation at 10,000 ⫻ g for 15 min. Aliquots (500 ␮l) of the clear supernatant were heated in glass tubes capped with aluminum foil with an equal volume of TBA solution (0.8%, wt/vol; heated at 60 C to dissolve) at 95 C for 30 min. Samples were cooled on ice, 100 ␮l of each were pipetted into 96-well plates, and absorbance was read at 532 nm using a Ceres UV 900C microplate reader. Blank values representing samples containing 20 mm Tris-HCl were subtracted from all values. Each assay was performed in triplicate. Means of triplicates from each assay were pooled to obtain the final mean and se. Typical TBARS formation in control brain and brain homogenates plus Fe3⫹ (100 ␮m) were calculated from the absorbance at 532 nm using 1,1,3,3-tetramethoxypropane as an external standard. Assay sensitivity was estimated from the standard curve, and a t test with two-tailed probabilities was employed for statistical comparisons. Measurement of high energy phosphates in the brain. ATP, ADP, and AMP levels were assessed by HPLC analysis of regions of rapidly frozen tissues dissected at ⫺30 C as described previously (33). In short, immediately after decapitation, whole brains and skull were immersed in liquid nitrogen and then dissected at ⫺30 C to minimize the metabolism of high energy phosphates due to stress or declining oxygen levels. The apparatus consisted of a PerkinElmer (Norwalk, CT) gradient HPLC pump, a Waters Associates (Milford, MA) 409 multiple UV wavelength detector, and a Shimadzu 501 integrator (Tokyo, Japan). Frozen tissue was dissected on a freezing cold plate (⫺20 C), placed in 0.4 m perchloric acid (10 ␮l/mg wet weight), homogenized, and centrifuged. The supernatant was neutralized with 25 ␮l 2 m K2CO3 added to 200 ␮l supernatant and centrifuged. The supernatant was stored at ⫺80 C until injected. Standards were prepared in 0.4 m perchloric acid at concentrations of 5 ␮m. Samples were separated on a 15-cm 3-␮m Nikko Bioscience C18 HPLC column (ESA, Inc., Chelmsford, MA) at a flow rate of 1 ml/min using a gradient. Buffer A was 25 mm Na2HPO4 with 100 mg/liter TBA, pH 5.5, and buffer B was 200 mm Na2HPO4 with 100 mg/liter TBA, pH 4.0, and 10% acetonitrile. The gradient was 100% buffer A for 0 –5 min, 100% buffer A to 100% buffer B for 5–25 min, and 100% buffer B for 25–34 min. Samples were monitored at 214 nm for 0 –9 min, 260 nm for 9 –25 min, and 214 nm for 25–34 min.


UCP2 transcripts in the hippocampus of control mice, but these were clearly present 24 h after seizure induction (Fig. 2). To test whether UCP2 induction in the mouse hippocampus has metabolic consequences, we assessed the RCRs of control and experimental mice. The RCR is the oxygen consumption at state 3 respiration (phosphorylating mitochondria) divided by the oxygen consumption at state 4 respiration (nonphosphorylating mitochondria in which oxygen consumption is due exclusively to proton leak, i.e. uncoupling). Lower RCR values correspond to higher mitochondrial uncoupling. In support of UCP2 being functionally active, the mean RCR of animals that seized was significantly lower than that of controls (1.73 ⫾ 0.1 vs. 2.34 ⫾ 0.06; P ⬍ 0.05). UCP2 expression in the hippocampus of transgenic animals. hUCP2 RNA was detected in the cerebellum, midbrain, brainstem, and cortex of transgenic, but not wild-type, mice (Fig. 3A). In the hippocampus, including the CA1 region, no UCP2 mRNA could be detected by in situ hybridization in wild-type mice (Fig. 3, B and C). In contrast, in transgenic mice, we found that UCP2 mRNA was associated with the pyramidal cell layer (Fig. 3, D and E). UCP2-like immuno-

Lactate measurements. Tissue lactate was measured from the same samples as those for high energy phosphate measurements using a commercially available ELISA kit (Sigma-Aldrich Corp., St. Louis, MO).

Results In vitro cell culture

UCP2 protects PC12 cells. To test the role of UCP2 directly on cell survival, we transfected excitable pheochromocytoma PC12 cells with a full-length hUCP2 cDNA and measured their responses to H2O2 and NO, agents known to induce rapid cell death via free radical formation (34 –36). As PC12 cells express little or no UCP2 under normal growth conditions, only the transfected cells contained UCP2 mRNA (Fig. 1A). We found significantly less death in UCP2-transfected vs. control PC-12 cells after exposure to either H2O2 or NO (Fig. 1, B–E), providing direct in vitro evidence that the expression of UCP2 is protective. In vivo epilepsy model using transgenic mice expressing hUCP2

UCP2 mRNA and mitochondrial uncoupling are induced in hippocampus of mice after seizures. RT-PCR analysis did not detect

FIG. 1. UCP2 is protective against H2O2 and NO in PC12 cells. A, RT-PCR amplification of UCP2 from a plasmid positive control (hUCP2), PC12 cells transfected with GFP alone (Con), and PC12 cells transfected with GFP plus hUCP2 (GFP⫹hUCP2), confirming expression only in the transfected cells. B–D, Images of cells transfected with GFP alone before (B) and after (C) treatment with H2O2 or transfected with both GFP and hUCP2 (D). Many more cells survive after hUCP2 transfection. E, Summary of three independent experiments, in each of which six separate wells were counted. Cell death is presented as the mean ⫾ SE. After treatment with either H2O2 or NO, there is significantly less (P ⬍ 0.05) cell death in cultures transfected with hUCP2.

FIG. 2. UCP2 is induced in the hippocampus of mice after seizures. RT-PCR analyses showed no presence of UCP2 gene product in the hippocampus of intact (Int) mice, but it was induced after pilocarpine (Pil)-induced seizures.


labeling was also present in pyramidal cells of the hippocampus of transgenic (Fig. 3, H and I), but not wild-type (Fig. 3, F and G), animals. The observations on wild-type mice are consistent with earlier reports on the lack UCP2 mRNA expression in the cortex and hippocampus of mice (16). The hUCP2 transgene expression in the hippocampus of intact animals provides an excellent model to test whether elevated UCP2 levels before initiation of seizures may be protective or detrimental for hippocampal neurons. UCP2 is neuroprotective in a mouse model of epileptic seizures. The UCP2 transgenic animals are on a C57/BL6 background that is resistant to kainic acid-induced seizures (37) and repeatedly delivered nonhomogenous responses to steady doses of kainic acid among wild-type animals. We therefore studied an alternative model in which seizure activity is induced by systemic scopolamine (1 mg/kg) and subsequent pilocarpine (280 mg/kg) injections, allowing noninvasive induction of motor seizures (38). In both wild-type (n ⫽ 12) and transgenic (n ⫽ 12) littermates, ip administration of pilocarpine resulted in the development of motor seizures within 10 –20 min after the initiation of treatment. The severity of motor seizures did not significantly differ between experimental groups, and both groups continued to manifest motor seizures in the hours following treatment, with seizure activity in this period showing no obvious difference between animal types (data not shown). The observed seizure patterns in surviving transgenic and wild-type animals predict massive neuronal death in the hippocampal formation within 24 h, including the CA1 region (38). Indeed, tissue damage in the CA1 region was obvious in wild-type animals even by low power light microscopic survey (Fig. 4B). Strikingly, however, the structure of the hippocampus of equally seized, transgenic mice was robustly less damaged (Fig. 4C), more closely resembling that of the intact (Fig. 4A) than the wild-type seized animals (Fig. 4B). This differential tissue damage was also confirmed by electron microscopic analysis (Fig. 4, D–F). Quantitative ultrastructural analyses of the CA1 region revealed that the number of degenerating or apoptotic pyramidal cells was higher (Fig. 5) in wild-type animals (n ⫽ 5) than transgenic littermates (n ⫽ 5; P ⬍ 0.05). In randomly selected electron micrographs of the CA1 pyramidal cell layer (14,400 ␮m2/ animal), wild-type animals had 724 ⫾ 34 (⫾sem) neurons, of which 365 ⫾ 18 (50.4%) were necrotic or apoptotic, and the transgenic animals had 808 ⫾ 41 total neurons, of which 75 ⫾ 5 (9.2%) were dying. In the transgenic animals, no significant cell death was apparent in other parts of the hippocampus, including the dentate gyrus and CA3 and CA2 areas. Although there were slightly more neurons in the principal layer of CA1 region of transgenic animals compared with wild-type mice (808 ⫾ 41 vs. 724 ⫾ 34), this difference did not reach significance (P ⬎ 0.05). No differences were detected in the volume of the entire brain or the hippocampal formation between transgenic and wild-type animals. UCP2 decreases lipid peroxidation. A potentially significant effect of UCP2 in neuroprotection may be its ability to control free radical production (7, 8). To test whether free radical levels are affected in hUCP2-expressing mice, we analyzed


lipid peroxidation in hippocampal and cerebral cortical homogenates from wild-type (n ⫽ 6) and hUCP2-expressing (n ⫽ 6) transgenic animals. TBARS assay revealed significantly lower levels of free radical content in the hippocampus of transgenic mice compared with wild-type littermates (15.14 ⫾ 1.106 vs. 20.18 ⫾ 1.81 nm 1,1,3,3-tetramethoxypropane/mg tissue; P ⬍ 0.05; Fig. 6A). The vulnerability of neurons to oxidative and free radical stress may be lower in transgenic animals, due in part to decreased production and consequent increased buffering capacity of UCP2-expressing cells to reactive oxygen species. UCP2 elevates ATP and ADP levels. When oxidation is uncoupled from phosphorylation, it might be anticipated that

FIG. 3. hUCP2 transgene is present in intact transgenic animals. A, hUCP2 mRNA expression in the cerebellum (Ce), midbrain (Mid), brainstem (BS), and cortex (Ctx) of intact transgenic and wild-type animals. hUCP3 mRNA was undetectable in any of the regions. B–E, In situ hybridization histochemistry demonstrated no UCP2 mRNA expression in the CA1 region of wild-type animals (B and C), but it was expressed in the pyramidal layer of the CA1 region of transgenic animals (D and E). C and E, High-power magnifications of the pyramidal layer shown on D and E, respectively. Silver grains exceeding three times the background level around counterstained nuclei were considered as a cell with UCP2 mRNA labeling. The bar scale in D represents 50 ␮m for B and D. The bar scale in E represents 10 ␮m for C and E. F–I, Immunocytochemistry for UCP2 using antisera that cross-react with hUCP2 resulted in no specific labeling in the CA1 region of wild-type animals (F and G). In transgenic animals, immunolabeling was detected in principal cells (H and I) that resembled pyramidal cells (pyr; arrowheads). A high-power magnification (I) showed that immunoperoxidase was also associated with dendrites extending to the stratum radiatum (str). The bar scale in H represents 50 ␮m for F and H. The bar scale in I represents 10 ␮m for G and I.

5018



FIG. 4. CA1 integrity is maintained after seizures of transgenic animals. A–F, Light (A–C) and electron (D–F) microscopic images show extensive damage in the CA1 area of wild-type animals (B and E) 24 h after seizures that were less apparent in hUCP2-expressing animals (C and F). Nonapoptotic or nonnecrotic pyramidal cells (pyr) were more abundant in the CA1 area of transgenic animals, similar to intact controls (A and D), and mostly glial cells, including microglia (mic) showed degeneration. A–C, Light microscopic images of unstained sections processed for electron microscopic embedding. The apparent labeling of necrotic cells in B is due to the effect of osmium. The bar scales in D (for A–C) and F (for D–F) represent 100 and 10 ␮m, respectively.

FIG. 5. Epileptic seizure-induced cell death is diminished in hUCP2expressing mice. Cell death in the CA1 area of the hippocampus was significantly reduced in surviving transgenic animals compared with surviving wild-type littermates 24 h after seizures.

ATP production would be compromised in mitochondria, which, in turn, could lead to cell death rather than survival (26). To determine ATP and ADP levels in hUCP2-expressing mice (n ⫽ 5) and wild-type littermates (n ⫽ 5), we analyzed freeze-clamped brain tissue for AMP, ADP, and ATP by HPLC. Strikingly, we found a nearly 60% increase in ATP (P ⬍ 0.05) and a 30% increase in ADP (P ⬍ 0.05) levels in the brains of hUCP2-expressing mice compared with wild-type littermates (Fig. 6C). One explanation for this may be that decreased yield of ATP production by respiration is balanced by an increased rate of ATP production by glycolysis, but no significant difference was found between brain lactate levels in transgenic (n ⫽ 5) and wild-type (n ⫽ 5) animals (29.36 ⫾ 3.87 vs. 33.16 ⫾ 9.80 nmol/mg protein; P ⫽ 0.55). Although these measurements cannot detect more complex changes in the rate of lactate metabolism, they do not support the argument that increased ATP levels in UCP2-overexpressing animals are the result of increased glycolysis. UCP2 increases mitochondrial number. In a recent study, PGC-1 (peroxisome-proliferator activated receptor ␥ coactivator-

1␣), a cold-induced transcription factor, was shown to upregulate the expression of UCPs, including UCP2, and to induce mitochondrial proliferation (39). To test whether by partially uncoupling respiration from oxidative phosphorylation, UCP2 might trigger mitochondrial proliferation, we assessed mitochondria number and size in the CA1 region of the hippocampus by stereological quantification. We found that the number of mitochondria was significantly higher in the hippocampus of UCP2 transgenic animals (n ⫽ 5) compared with wild-type littermates (n ⫽ 5; Fig. 6, B, D, and E). In the principal layer of CA1 of wild-type animals, 651 ⫾ 38 (wild-type) and 1425 ⫾ 72 (transgenic) mitochondria were counted within a 3600-␮m2 region. This represented a 218% increase (P ⬍ 0.05) in mitochondria number. In the stratum radiatum of the CA1 region, the mean numbers of mitochondria were 1584 ⫾ 82 (wild-type) and 2364 ⫾ 102 (transgenic), which also represented a significant (P ⬍ 0.05) increase (150%) in transgenic animals. Because hUCP2 mRNA and protein were expressed in pyramidal cell and their dendrites, these observations suggest that UCP2 can lead to the production of more mitochondria, and this may be the basis for the observed increases in ATP and ADP levels. Discussion

Our in vitro and in vivo results both show that increased expression of UCP2 protects neurons against a variety of noxious insults. Although the functionality of UCP2 remains controversial in different tissue types, for example, in the spleen (40, 41), our present observations are supportive of the idea that UCP2 has significant impact on physiological and pathological processes. Transfection of PC12 cells by UCP2 decreased the death of cells exposed to either H2O2 or NO, substances shown to be toxic via their free radical metabolites (34 –36). Similar mech-



FIG. 6. Decreased lipid peroxidation and increased mitochondria number in hUCP2-expressing mice. A, TBARS analyses of the brain revealed a significantly lower level of lipid peroxidation in the hippocampus of transgenic mice (f) compared with wild-type littermates (䡺). B, D, and E, Mitochondria number was significantly higher in the hippocampus of transgenic mice compared with wild-type littermates. Arrowheads in the electron micrographs of C (wild-type) and D (transgenic) point to mitochondria. The bar scale in D represents 1 ␮m (for C and D). C, Mitochondrial number was elevated in transgenic mice compared with their wild-type littermates (P ⬍ 0.05).

anisms of apoptosis induction have been implicated in epileptic insults (25, 42– 44), and thus, they could have contributed to the in vivo results as well. In addition, the increased in vivo survival of neurons in the hippocampus of epileptic rodent brain expressing hUCP2 constitutively may be mediated by an increase in the number of mitochondria. UCP levels are also positively correlated with mitochondrial number in brown fat (UCP1) (45) and with ATP amounts in muscle (46), but are negatively correlated with ATP levels in pancreatic ␤-cells in vitro (47). Thus, it is not unlikely that UCP2’s effect on mitochondrial proliferation and ATP production may be tissue specific. What the downstream molecular events are after UCP2 activation to trigger mitochondrial proliferation remains to be resolved. Nevertheless, it is tempting to speculate that the subtle lowering of inner mitochondrial membrane potential by UCP2 activation compromises the efficacy of a mitochondrion to produce ATP, which may, in turn, activate signaling cascade between the mitochondrion to the host nucleus for the need of more ATP sources, i.e. mitochondria. The idea that mitochondrial proliferation is important in neuronal survival during epilepsy is supported by the observation that in epileptic hippocampus of humans, the number of mitochondria in surviving neurons is elevated (48). The partially uncoupled, more numerous mitochondria thus can produce the same or larger amounts of ATP than the normal number of tightly coupled mitochondria of wild-type animals without a corresponding increase in free radical production (Fig. 7). Indeed, the suppressed lipid peroxida-

tion that we observed in the brains of transgenic animals with higher mitochondrial numbers and ATP levels are likely to be the consequences of diminished free radical levels, because UCP2 uses these reactive oxygen species to increase their proton conductance (13, 14), a process that also eliminates these very free radicals (present results, but also see Refs. 7 and 8). Because of this and the fact that UCP2 can be induced by increased cellular stress during brain injury (19), it is reasonable to suggest that the mechanism depicted in Fig. 7 can underlie preconditioning of neuronal tissue by repeated low exposure of neurons to cellular stress in which UCP2 is induced. Whether the same mechanism of preconditioning by UCP2 may exist in other tissues, for example, in the heart, needs to be tested in future studies. The last decade has provided overwhelming evidence that mitochondrial proteins, in addition to participating in energy production, act as initiating elements of cellular death by either apoptosis or necrosis (49 –51). Failure of the integrity of the inner and outer membranes (and the collapse of membrane potentials) of the mitochondria are among the first critical steps in initiating neuronal cell death. Our present data together with the pioneering work of Nicholls and coworkers (21–24) suggest that UCP2, by stabilizing mitochondrial membrane potential, may avoid disintegration of the inner mitochondrial membrane, which, in turn, prevents the likelihood of cell death. The cytotoxic and/or excitotoxic damage to neurons induced by epileptic seizures has similarities to that seen in a variety of other neurological disorders ranging from stroke to Alzheimer’s disease. We suggest

5020



FIG. 7. Schematic illustration of UCP2 action based on our results. In control animals (left), a proton gradient in tightly coupled mitochondria drives ATP synthesis through ATP synthase, and as a consequence, free radical oxygen species are produced. In transgenic animals (right), increased expression of UCP2 leads to leakage of protons across the inner mitochondrial membrane, which more efficiently neutralizes the free radical oxygen. Although this may be associated with less efficient ATP production per mitochondrion, the increased number of mitochondria results in an overall increased level of ATP with reduced free radical oxygen production.

that increased UCP2 activity could provide protection in these disorders. In fact, recently we have established that increased mitochondrial uncoupling predicts decreased cell death in acute brain injury (19) as well as in a nonhuman primate model of Parkinson’s disease (27). Because there is good evidence that UCP2 activity can be regulated at both transcriptional and posttranslational levels (13–15), it is an excellent candidate to target for the development of novel therapies for these disorders. On the other hand, when UCP2 down-regulation is designed to affect peripheral mechanisms, including insulin secretion (43) or the innate immune system (8), profound side-effects of those treatments might occur that could trigger or worsen neurodegenerative processes. Acknowledgments We thank Drs. Gerald Shulman and Gary Cline for the initial ATP measurements, Luis Miguel-Garcia and Sergio Vega for help with seizure induction in mice for oxygen consumption measurements, and Tamas Bartfai for helpful discussions. We are indebted to Erzsebet Borok, Marya Shanabrough, and Adrienne LaRue for excellent technical support. Received May 29, 2003. Accepted August 8, 2003. Address all correspondence and requests for reprints to: Dr. Tamas L. Horvath, Department of Obstetrics and Gynecology, Yale Medical School, 333 Cedar Street, FMB 339, New Haven Connecticut 06520. E-mail: [email protected]. This work was supported by NIH Grants NS-41725 (to T.L.H.) and EY-013865 (to C.J.B.). Current address for P.P.: Department of Physiology, Southern Illinois University School of Medicine, Carbondale, Illinois 62901.

References 1. Bouillaud F, Ricquier D, Thibault J, Weissenbach, J 1985 Molecular approach to thermogenesis in brown adipose tissue: cDNA cloning of the mitochondrial uncoupling protein. Proc Natl Acad Sci USA 82:445– 448

2. Fleury C, Neverova M, Collins S, Raimbault S, Champigny O, Levi-Meyrueis C, Bouillard F, Seldin MF, Surwit RS, Ricquier D, Warden CH 1997 Uncoupling protein-2: a novel gene linked to obesity and hyperinsulinemia. Nat Genet 15:269 –272 3. Boss O, Samec S, Paoloni-Giacobino A, Rossier C, Dulloo A, Seydoux J, Muzzin P, Giacobino JP 1997 Uncoupling protein-3: a new member of the mitochondrial carrier family with tissue-specific expression. FEBS Lett 408: 39 – 42 4. Vidal-Puig A, Solanes G, Grujic D, Flier JS, Lowell BB 1997 UCP3: an uncoupling protein homologue expressed preferentially and abundantly in skeletal muscle and brown adipose tissue. Biochem Biophys Res Commun 235:79 – 82 5. Mao W, Yu XX, Zhong A, Li W, Brush J, Sherwood SW, Adams SH, Pan G 1999 UCP4, a novel brain-specific mitochondrial protein that reduces membrane potential in mammalian cells. FEBS Lett 443:326 –330 6. Sanchis D, Fleury C, Chomiki N, Goubern M, Huang Q, Neverova M, Gregoire F, Easlick J, Raimbault S, Levi-Meyrueis C, Miroux B, Collins S, Seldin M, Richard D, Warden C, Bouilland F, Ricquier D 1998 BMCP1, a novel mitochondrial carrier with high expression in the central nervous system of humans and rodents, and respiration uncoupling activity in recombinant yeast. J Biol Chem 273:34611–34615 7. Negre-Salvayre A, Hirtz C, Carrera G, Cazenave R, Troly M, Salvayre R, Penicaud L, Casteilla L 1997 A role for uncoupling protein-2 as a regulator of mitochondrial hydrogen peroxide generation. FASEB J 11:809 – 815 8. Arsenijevic D, Onuma H, Pecqueur C, Raimbault S, Manning BS, Miroux B, Couplan E, Alves-Guerra MC, Goubern M, Surwit R, Bouillaud F, Richard D, Collins S, Ricquier D 2000 Disruption of the uncoupling protein-2 gene in mice reveals a role in immunity and reactive oxygen species production. Nat Genet 26:435– 439 9. Nicholls DG, Locke RM 1984 Thermogenic mechanism in brown fat. Physiol Rev 64:1– 64 10. Lowell BB, Spiegelman BM 2000 Towards a molecular understanding of adaptive thermogenesis. Nature 404:652– 660 11. Echtay KS, Winkler E, Klingenberg M 2000 Coenzyme Q is an obligatory cofactor for uncoupling protein function. Nature 408:609 – 613 12. Echtay KS, Winkler E, Frischmuth K, Klingenberg M 2001 Uncoupling proteins 2 and 3 are highly active H⫹ transporters and highly nucleotide sensitive when activated by coenzyme Q (ubiquinone). Proc Natl Acad Sci USA 98: 1416 –1421 13. Echtay KS, Murphy MP, Smith RA, Talbot DA, Brand MD 2002 Superoxide activates mitochondrial uncoupling protein 2 from the matrix side: Studies using targeted antioxidants. J Biol Chem 277:47129 – 47135 14. Echtay KS, Roussel D, St-Pierre J, Jekabsons MB, Cadenas S, Stuart JA, Harper JA, Roebuck SJ, Morrison A, Pickering S, Clapham JC, Brand MD 2002 Superoxide activates mitochondrial uncoupling proteins. Nature 415: 96 –99 15. Pecqueur C, Alves-Guerra MC, Gelly C, Levi-Meyrueis C, Couplan E, Col-


16. 17. 18.

19. 20.

21. 22. 23. 24. 25.

26. 27.

28. 29.

30. 31.

32.

33.

lins S, Ricquier D, Bouillaud F, Miroux B 2001 Uncoupling protein 2, in vivo distribution, induction upon oxidative stress, and evidence for translational regulation. J Biol Chem 276:8705– 8712 Richard D, Rivest R, Huang Q, Bouillaud F, Sanchis D, Champigny O, Ricquier D 1998 Distribution of the uncoupling protein 2 mRNA in the mouse brain. J Comp Neurol 397:549 –560 Horvath TL, Warden CH, Hajos M, Lombardi A, Goglia F, Diano S 1999 Brain uncoupling protein 2: uncoupled neuronal mitochondria predict thermal synapses in homeostatic centers. J Neurosci 19:10417–10427 Diano S, Urbanski HF, Horvath B, Bechmann I, Kagiya A, Nemeth G, Naftolin F, Warden CH, Horvath TL 2000 Mitochondrial uncoupling protein 2 (UCP2) in the non-human primate brain and pituitary. Endocrinology 141: 4226 – 4238 Bechmann I, Diano S, Warden CH, Bartfai T, Nitsch R, Horvath TL 2002 Brain mitochondrial uncoupling protein 2 (UCP2): protective stress signal in neuronal injury. Biochem Pharmacol 64:363–367 Voehringer DW, Hirschberg DL, Xiao J, Lu Q, Roederer M, Lock CB, Herzenberg LA, Steinman L, Herzenberg LA 2000 Gene microarray identification of redox and mitochondrial elements that control resistance or sensitivity to apoptosis. Proc Natl Acad Sci USA 97:2680 –2685 Budd SL, Nicholls DG 1996 Mitochondrial calcium regulation and acute glutamate excitotoxicity in cultured cerebellar granule cells. J Neurochem 67:667– 679 Castilho RF, Hansson O, Ward MW, Budd SL, Nicholls DG 1998 Mitochondrial control of acute glutamate excitotoxicity in cultured cerebellar granule cells. J Neurosci 18:10277–10286 Castilho RF, Ward MW, Nicholls DG 1999 Oxidative stress, mitochondrial function, and acute glutamate excitotoxicity in cultured cerebellar granule cells. J Neurochem 72:1394 –1401 Nicholls DG, Ward MW 2000 Mitochondrial membrane potential and neuronal glutamate excitotoxicity: mortality and millivolts. Trends Neurosci 23: 166 –174 Bellissimo MI, Amado D, Abdalla DS, Ferreira EC, Cavalheiro EA, NaffahMazzacoratti MG 2001 Superoxide dismutase, glutathione peroxidase activities and the hydroperoxide concentration are modified in the hippocampus of epileptic rats. Epilepsy Res 46:121–128 Tarnopolsky MA, Beal MF 2001 Potential for creatine and other therapies targeting cellular energy dysfunction in neurological disorders. Ann Neurol 49:561–574 Horvath TL, Diano S, Leranth C, Garcia-Segura LM, Cowley MA, Shanabrough M, Elsworth JD, Sotonyi P, Roth RH, Dietrich EH, Matthews RT, Barnstable CJ, Redmond Jr DE 2003 Coenzyme Q induces nigral mitochondrial uncoupling and prevents dopamine cell loss in a primate model of Parkinson’s disease. Endocrinology 144:2757–2760 Horvath TL, Diano S, Barnstable C 2003 Mitochondrial uncoupling protein 2 in the central nervous system: neuromodulator and neuroprotector. Biochem Pharmacol 65:1917–1921 Horvath TL, Diano S, Miyamoto S, Barry S, Gatti S, Alberati D, Livak F, Lombardi A, Moreno M, Goglia F, Mor G, Hamilton J, Kachinskas D, Horwitz B, Warden CH 2003 Uncoupling proteins-2 and 3 influence obesity and inflammation in transgenic mice. Int J Obesity 27:433– 442 Ahmad I, Leinders-Zufall T, Kocsis JD, Shepherd GM, Zufall F, Barnstable CJ 1994 Retinal ganglion cells express a cGMP-gated cation conductance activatable by nitric oxide donors. Neuron 12:155–165 Horvath B, Spies C, Horvath Gy, Kox WJ, Miyamoto S, Barry S, Warden CH, Bechmann I, Diano S, Heemskerk J, Horvath TL 2002 Uncoupling protein 2 (UCP2) lowers alcohol sensitivity and pain threshold. Biochem Pharmacol 64:369 –374 Fraga CG, Leibovitz BE, Tappel AL 1988 Lipid peroxidation measured as thiobarbituric acid-reactive substances in tissue lices: characterization and comparison with homogenates and microsomes. Free Radical Biol Med 4:155– 161 Matthews RT, Yang L, Jenkins BG, Ferrante RJ, Rosen BR, Kaddurah-Daouk


34. 35. 36. 37. 38. 39.

40. 41.

42.

43.

44. 45.

46.

47.

48. 49. 50.

51.

R, Beal MF 1998 Neuroprotective effects of creatine and cyclocreatine in animal models of Huntington’s disease. J Neurosci 18:156 –163 Behrman HR, Riley JCM, Aten RF 1993 Reactive oxygen species and ovarian function. In: Adashi EY, Leung PCK, eds. The ovary. New York: Raven Press; 455– 471 Gassen M, Youdim MB 1999 Free radical scavengers: chemical concepts and clinical relevance. J Neural Transm 56(Suppl):193–210 Heck DE 2001 *NO, RSNO, ONOO⫺, NO⫹, *NOO, NOx: dynamic regulation of oxidant scavenging, nitric oxide stores, and cyclic GMP-independent cell signaling. Antioxid Redox Signal 3:249 –260 Engstrom FL, Woodbury DM 1988 Seizure susceptibility in DBA and C57 mice: the effects of various convulsants. Epilepsia 29:389 –395 Turski WA, Cavalheiro EA, Bortolotto ZA, Mello LM, Schwarz M, Turski L 1984 Seizures produced by pilocarpine in mice: a behavioral, electroencephalographic and morphological analysis. Brain Res 321:237–253 Wu Z, Puigserver P, Andersson U, Zhang C, Adelmant G, Mootha V, Troy A, Cinti S, Lowell B, Scarpulla RC, Spiegelman BM 1999 Mechanisms controlling mitochondrial biogenesis and respiration through the thermogenic coactivator PGC-1. Cell 98:115–124 Krauss S, Zhang CY, Lowell BB 2002 A significant portion of mitochondrial proton leak in intact thymocytes depends on expression of UCP2. Proc Natl Acad Sci USA 99:118 –122 Couplan E, del Mar Gonzalez-Barroso M, Alves-Guerra MC, Ricquier D, Goubern M, Bouillaud F 2002 No evidence for a basal, retinoic, or superoxideinduced uncoupling activity of the uncoupling protein 2 present in spleen or lung mitochondria. J Biol Chem 277:26268 –26275 Bagetta G, Paoletti AM, Leta A, Del Duca C, Nistico R, Rotiroti D, Corasaniti MT 2002 Abnormal expression of neuronal nitric oxide synthase triggers limbic seizures and hippocampal damage in rat. Biochem Biophys Res Commun 291:255–260 Gupta RC, Milatovic D, Dettbarn WD 2001 Nitric oxide modulates highenergy phosphates in brain regions of rats intoxicated with diisopropylphosphorofluoridate or carbofuran: prevention by N-tert-butyl-␣-phenylnitrone or vitamin E. Arch Toxicol 75:346 –356 Muller M, Fontana A, Zbinden G, Gahwiler BH 1993 Effects of interferons and hydrogen peroxide on CA3 pyramidal cells in rat hippocampal slice cultures. Brain Res 619:157–162 Rossmeisl M, Barbatelli G, Flachs P, Brauner P, Zingaretti MC, Marelli M, Janovska P, Horakova M, Syrovy I, Cinti S, Kopecky J 2002 Expression of the uncoupling protein 1 from the aP2 gene promoter stimulates mitochondrial biogenesis in unilocular adipocytes in vivo. Eur J Biochem 269:19 –28 Garcia-Martinez C, Sibille B, Solanes G, Darimont C, Mace K, Villarroya F, Gomez-Foix AM 2001 Overexpression of UCP3 in cultured human muscle lowers mitochondrial membrane potential, raises ATP/ADP ratio, and favors fatty acid vs. glucose oxidation. FASEB J 15:2033–2035 Zhang CY, Baffy G, Perret P, Krauss S, Peroni O, Grujic D, Hagen T, Vidal-Puig AJ, Boss O, KimYB, Zheng XX, Wheeler MB, Shulman GI, Chan CB, Lowell BB 2001 Uncoupling protein regulates insulin secretion and is major link between obesity, ␤ cell dysfunction, and type 2 diabetes. Cell 105:745–755 Blumcke I, Zuschratter W, Schewe JC, Suter B, Lie AA, Riederer BM, Meyer B, Schramm J, Elger CE, Wiestler OD 1999 Cellular pathology of hilar neurons in Ammon’s horn sclerosis. J Comp Neurol 414:437– 453 Shimizu S, Narita M, Tsujimoto Y 1999 Bcl-2 family proteins regulate the release of apoptogenic cytochrome c by the mitochondrial channel VDAC. Nature 399:483– 487 Susin SA, Lorenzo HK, Zamzami N, Marzo I, Snow BE, Brothers GM, Mangion J, Jacotot E, Costantini P, Loeffler M, Larochette N, Goodlett DR, Aebersold R, Siderovski DP, Penninger JM, Kroemer G 1999 Molecular characterization of mitochondrial apoptosis-inducing factor. Nature 397:441– 446 Kroemer G, Reed JC 2000 Mitochondrial control of cell death. Nat Med 6:513–519