The cellular requirement for ATP as a high-energy intermediate

Cellular Biology Uncoupling Protein Impacts Endothelial Phenotype via p5-mediated Control of Mitochondrial Dynamics Yukio Shimasaki, Ning Pan, Louis M. Messina, Chunying Li, Kai Chen, Lijun Liu, Marcus P. Cooper, Joseph A. Vita, John F. Keaney Jr Downloaded from http://circres.ahajournals.org/ by guest on July, 8 Rationale: Mitochondria, although required for cellular ATP production, are also known to have other important functions that may include modulating cellular responses to environmental stimuli. However, the mechanisms whereby mitochondria impact cellular phenotype are not yet clear. Objective: To determine how mitochondria impact endothelial cell function. Methods and Results: We report here that stimuli for endothelial cell proliferation evoke strong upregulation of mitochondrial uncoupling protein (UCP). Analysis in silico indicated increased UCP expression is common in highly proliferative cell types, including cancer cells. Upregulation of UCP was critical for controlling mitochondrial membrane potential (Δψ) and superoxide production. In the absence of UCP, endothelial growth stimulation provoked mitochondrial network fragmentation and premature senescence via a mechanism involving superoxide-mediated p5 activation. Mitochondrial network fragmentation was both necessary and sufficient for the impact of UCP on endothelial cell phenotype. Conclusions: These data identify a novel mechanism whereby mitochondria preserve normal network integrity and impact cell phenotype via dynamic regulation of UCP. (Circ Res. ;:89-9.) Key Words: angiogenesis endothelium endothelial function ischemia mitochondria superoxides mitochondrial uncoupling proteins The cellular requirement for ATP as a high-energy intermediate to fuel many critical cellular reactions is universal. The production of ATP depends on an extracellular source of carbohydrates or lipids that are initially metabolized in the cytosol to products that are then transported into mitochondria for oxidative phosphorylation to produce a proton gradient (ie, membrane potential, Δψ) that powers ATP generation. Known mitochondrial functions extend beyond ATP production to include other cellular processes, such as apoptosis, heme synthesis, calcium homeostasis, inflammation, 5 and development. 6 Thus, mitochondria have the capacity to impact cellular phenotype via energy-dependent and energy-independent mechanisms. Editorial, see p 86 One means by which mitochondria can impact cellular phenotype is through participation in signaling paradigms, and there is ample evidence from multiple organisms that mitochondria signal to the nucleus. Respiratory deficiency induced by either pharmacologic 7 or genetic 8 means is known to affect gene expression in the nucleus. Mitochondrial-derived heme coordinates nuclear regulation of genes encoding heme-dependent proteins, such as cytochrome c, 9 catalase, and cytochrome oxidase. The mitochondrial release of proteins, such as apoptosis-inducing factor and the second mitochondrial activator of caspase/diablo gene product, into the cytosol also regulates genes important for apoptosis. Thus, mitochondrial events are readily communicated to the nucleus to affect gene regulation. The cellular responses to environmental stimuli also involve mitochondria. Proteolytic processing of cell-surface receptors, such as epidermal growth factor receptor tyrosine kinase B, can release intracellular domains that trigger mitochondrial release of proapoptotic proteins. Energy deprivation responses, such as AMP kinase activation, induce mitochondrial gene upregulation that is critical for cellular stress adaptation. Upregulation of endothelial vascular endothelial growth factor involves perinuclear mitochondrial clustering and the local production of mitochondrial O. 5 The latter has also been linked to the control of mitogen-activated protein kinase phosphatases and the coordination of nuclear factor κb dependent inflammatory responses. 6 Finally, mitochondrial electron transport is known to impact the activation state of cell-surface growth factor receptors. 7 Thus, Original received March 5, ; revision received June 7, ; accepted July,. In May, the average time from submission to first decision for all original research papers submitted to Circulation Research was 5 days. From the Division of Cardiovascular Medicine, Department of Medicine (Y.S., N.P., C.L., K.C., L.L., M.P.C., J.F.K.); Division of Vascular Surgery, Department of Surgery, University of Massachusetts Medical School, Worcester, MA (L.M.M.); and the Department of Medicine, Boston University School of Medicine (J.A.V.). The online-only Data Supplement is available with this article at http://circres.ahajournals.org/lookup/suppl/doi:.6/circresaha..9/-/dc. Correspondence to John F. Keaney Jr, MD, Division of Cardiovascular Medicine, University of Massachusetts Medical School, 55 Lake Ave N, Room S-855, Worcester, MA, 655. E-mail john.keaney@umassmed.edu American Heart Association, Inc. Circulation Research is available at http://circres.ahajournals.org DOI:.6/CIRCRESAHA..9 89

89 Circulation Research September, Downloaded from http://circres.ahajournals.org/ by guest on July, 8 Nonstandard Abbreviations and Acronyms Δψ mitochondrial membrane potential BAEC bovine aortic endothelial cell Drp dynamin-related protein Fis fission MLEC murine lung endothelial cell Mfn Mitofusin Mfn Mitofusin Opa optic atrophy ROS reactive oxygen species UCP uncoupling protein the mitochondrion is emerging as an important organelle for the coordination of cellular environmental responses. Although mitochondria clearly impact how cells respond to the environment, the mechanisms involved in this process are not well-understood. Many mitochondrial actions are linked to the electron transport chain and the resultant membrane gradient, Δψ. Control of Δψ involves the relative availabilities of substrates for electron transport (nicotinamide adenine dinucleotide and flavin adenine dinucleotide-), respiration (O ), and ATP synthesis (ADP), as well as any proton leak. 8 The latter is largely regulated via uncoupling proteins that belong to the mitochondrial carrier superfamily. 8 Herein, we report that endothelial cell proliferation and angiogenesis involve upregulation of mitochondrial uncoupling protein (UCP) to reduce Δψ and limit mitochondrial O that otherwise promotes p5-dependent mitochondrial fragmentation resulting in premature senescence. These findings suggest a new function for UCP that has broad implications for processes that involve the vascular endothelium. Methods Endothelial Cell Culture and Transfection Bovine aortic endothelial cells (BAECs) were purchased from Genlantis (San Diego, CA) and were cultured in endothelial basal medium supplemented with EGM-MV Bullet Kit from Lonza (Walkersville, MD). Before each experiment, BAECs were made quiescent by -hour incubation in low-serum medium (endothelial basal medium supplemented with.% or.% fetal bovine serum). Human aortic endothelial cells were from Lonza and were cultured as described. Murine lung endothelial cells (MLECs) from mice of both sexes were isolated as described, 9 and cultured on gelatinor collagen-coated plates and grown in MLEC medium containing % fetal bovine serum, 8% Dulbecco's modified eagle medium, 8% Ham s F- with -μg/ml endothelial cell growth supplement (ECGS, Biomedical Technologies, Stoughton, MA), -mmol/l L-glutamine, -μg/ml heparin, and penicillin/streptomycin. Endothelial purity was confirmed by staining with,'-dioctadecyl,,','-tetramethyl-indocarbocyanine perchlorate acetylated lowdensity lipoprotein (Biomedical Technologies, Stoughton, MA) and anti-cd antibody (BD Biosciences, San Jose, CA). Quantitative Real-Time Polymerase Chain Reaction Total RNA was extracted from cells and tissues with the RNeasy Mini Kit (Qiagen) or TRIzol reagent (Invitrogen), and μg of total RNA was reverse transcribed with oligo(dt) primers for cdna synthesis. Realtime polymerase chain reaction was performed in the iq5 real-time polymerase chain reaction detection system (Bio-Rad Laboratories) and the products were detected using either SYBR Green dyes (Bio-Rad Laboratories) or TaqMan probes of the TaqMan Gene Expression Assays for specific genes (Applied Biosystems, Foster City, CA). Transfections For gene overexpression, BAECs or MLECs were incubated with adenoviruses at 5 to multiplicity of infection as described., For gene silencing, BAECs were transfected with. μg of small interfering RNA oligonucleotides (Thermo Scientific Dharmacon, Lafayette, CO) against human UCP (Cat. Nos. D-5-, -, -, and -); human superoxide dismutase (SOD; D-978-, -, -9, and -); or negative control (D--, -, -, and -5) as described. Likewise, MLECs were transfected with mouse mitofusin (Mfn) (J-6599-9, -, -, and -); mouse mitofusin (Mfn) (J-6-5, -6, -7, and -8); mouse p (J-5866-5, -6, -7, and -8); mouse p5 (J-6-9, -, -, and -); or the nontargeting pool (D-8--5). Cell Proliferation and Migration Assays DNA synthesis was directly measured via the H-thymidine incorporation assay. Cell proliferation was determined with CyQUANT GR fluorescent dye (Invitrogen Molecular Probes, Eugene, OR) to determine the relative cell number with a fluorescence microplate reader (Gemini XPS, Molecular Devices). Cell migration was assayed with the in vitro scratch assay in which a monolayer of quiescent cells was uniformly scratched and the rate of cell migration to close the void was evaluated hours after wounding using ImageJ. Mitochondrial Membrane Potential and Superoxide Measurements Mitochondrial membrane potential (Δψ) was determined using complementary fluorescent methods. Cells were incubated with -μmol/l 5, 5, 6, 6 -tetrachloro-,,, -tetraethylbenzimidazolcarbocyanine iodide (JC-) for minutes and Δψ estimated as the fluorescence ratio of JC- aggregates (red, excitation 55 nm, emission 6 nm) to monomers (green, excitation 85 nm, emission 55 nm) formed as a function of inner mitochondrial membrane potential. For measurement of mitochondrial superoxide ( O ), cells were loaded with 5-μmol/L MitoSOX Red (Invitrogen) for minutes and the fluorescence (excitation 96 nm; emission 5 nm) determined. Live cells were labeled with.5-μmol/l MitoSOX for minutes and the fluorescence images at both 5- and 5-nm excitation were captured using an Eclipse TE-S fluorescence microscope (Nikon, Melville, NY) with a CCD camera using a SPOT Insight MP Firewire Color Mosaic (Diagnostic Instruments, Sterling Heights, MI). Antibodies and Immunoblotting Antibodies against cytochrome c oxidase subunit IV, p cip/waf, p6 ink6a, phospho-p5 (mouse Ser8), phospho-c-jun (Ser6), phospho p8 mitogen-activated protein kinase (Thr8/Tyr8), and -rabbit or mouse IgG were purchased from Cell Signaling Technology. Antibodies against cytochrome c were from BD Biosciences. We obtained antibodies against UCP, UCP, and UCP from Fitzgerald Industries International (Acton, MA) and specific UCP antibodies (N-9 and A-9) and native p5 antibody from Santa Cruz Biotechnology (Santa Cruz, CA). We obtained SOD antibody from Millipore (Temecula, CA) and actin antibody from Sigma- Aldrich (St. Louis, MO). Protein extracts in DTT-containing SDS sample buffer were separated in % or % SDS-polyacrylamide gels and transferred to Hybond ECL nitrocellulose membranes (GE Healthcare, Piscataway, NJ). Immunoblotting was then performed and quantified as described. In cases where loading controls produced paired bands, both were used for quantification. Oxygen Consumption and Lactate Using a Clark-type oxygen electrode (Hansatech), respiration in whole cells was quantified as previously described., Briefly, 6 cells were resuspended in ml of respiration medium (Dulbecco`s phosphate-buffered saline, mmol/l glucose, mmol/l pyruvate, % fatty acid-free bovine serum albumin). To obtain proton leak, oligomycin was added to a final concentration of µmol/l. To measure maximal respiration, carbonyl cyanide -(trifluoromethoxy)

Shimasaki et al UCP and Endothelial Function 89 Downloaded from http://circres.ahajournals.org/ by guest on July, 8 phenylhydrazone was added to a final concentration of. µmol/l. Nonmitochondrial respiration, obtained by adding myxothiazol to a final concentration of µmol/l, was subtracted from total respiration measurements. For lactate, accumulation was determined by a lactate fluorometric assay kit (MBL International, Woburn, MA) according to the manufacturer s instructions and the concentration of secreted lactate was normalized to the cell number in each sample. Capillary Sprouting Assay in Aortas Thoracic aorta was placed in endothelial basal medium- as described, 5 periaortic tissue was carefully removed, and then the aorta was cleaned and sliced into mm-long rings. Rings were then embedded in liquid collagen gel in 8-well plates (BD Biosciences), incubated at 7 C for hour to polymerize the collagen, and the solid gel covered with MLEC medium diluted : with Dulbecco's modified eagle medium. Each aortic ring was examined daily and capillary sprouts counted along the sample perimeter under and magnification. Vascular sprouts were distinguished from fibroblasts via morphology as described 6 and CD staining. To infect aortic rings with adenoviral vectors, each ring was embedded in liquid collagen gel containing respective vectors (adenovirus UCP, adenovirus SOD, etc) at. 8 pfu before polymerized. Transfection was validated in separate experiments with green fluorescent protein adenovirus ( GFP) transfection and after days by fixation of collagen gel-embedded tissue with % formaldehyde in phosphate-buffered saline for minutes at C. Fixed sections underwent immunofluorescence staining with anti-gfp and anti-cd antibody and counterstained with ',6-diamidino--phenylindole before fluorescence imaging. Animals and Hindlimb Ischemia Model Heterozygous UCP-null animals 7 on the C57 background were obtained from Dr Bradford Lowell (Harvard Medical School) and bred to homogeneity with age-matched controls. SOD heterozygous mice were obtained from Jackson Laboratories and bred to obtain heterozygous and control animals. All experimental protocols were approved by the Institutional Animal Care and Use Committee of University of Massachusetts Medical School. Male mice at 8 to weeks of age were anesthetized with intraperitoneal injection of combination of mg/kg ketamine hydrochloride and 5 mg/kg xylazine (Webster Veterinary, Devens, MA) before surgery. Unilateral hindlimb ischemia in the left leg was introduced in the mice as described. 8 In selected experiments, µl (. 8 pfu) of adenoviral vectors encoding UCP, or β-glactosidase (LacZ) was injected into 5 different sites of the ischemic thigh muscles, such as adductor longus, adductor magnus, and adductor brevis muscles. Hindlimb tissue perfusion was assessed with Moor LDI-IR laser-doppler imaging system (Moor Instruments, Devon, UK). Blood flow images were obtained under conditions of constant body temperature (6±. C) and average hindlimb blood flow was expressed as the ratio of ischemic to nonischemic foot flow to account for minor variations in imaging conditions. Cell Cycle Analysis Endothelial cells at 5% confluence were synchronized in.% serum overnight. Cells were then cultured in complete medium and at and 8 hours harvested, stained with propidium iodide and subjected to fluorescence-activated cell sorting analysis as described. 9 Cellular Senescence Assay MLECs were seeded onto.% gelatin-coated -well plates and maintained for days under normal conditions. Senescence was assayed as senescence-associated β-galactosidase activity using the senescence β-galactosidase staining kit (Cell Signaling Technology). Both bright-field and phase-contrast pictures were viewed through or objective on a microscope on the Eclipse TE-S microscope as above and images processed with SPOT Advanced Version.6 imaging software (Diagnostic Instruments Inc). Senescence-associated β- galactosidase positive cells were quantified with the ImageJ software. Mitochondrial Length Measurements Mitochondrial length was measured as an index of mitochondrial fragmentation. Live endothelial mitochondria were labeled with -nmol/l MitoTracker Green FM for minutes and the fluorescence images were captured using a oil immersion objective to acquire high-resolution images of mitochondria as described. To determine mitochondrial length, ImageJ was applied in a blinded fashion to the measurement for each frame of a selected region of interest as described for determining DNA contour lengths. Statistical Analysis All data are expressed as mean±se and the numbers of independent experiments are indicated. Statistical comparisons were conducted between groups by use of Student t test or Mann Whitney U test as appropriate. Multiple groups were compared with either -way Kruskal Wallis or ANOVA with a post hoc Tukey Kramer multiple comparisons test as indicated in legends. A P value <.5 was considered significant. All statistics were done using StatView version 5. (SAS Institute, Cary, NC) or GraphPad Prism version 5 (GraphPad Software, La Jolla, CA). Results Dynamic Modulation of UCP and Δψ With Endothelial Proliferation Quiescent BAEC monolayers stimulated to proliferate with fetal bovine serum exhibited a reduction in Δψ (Figure A), whereas increasing cell confluence was associated with an increase in Δψ (Figure B and C). Because mitochondrial UCPs are implicated in Δψ regulation, 8 we probed endothelial UCP expression and observed mrna and protein only for UCP and UCP (Figure D). We then observed that UCP is upregulated with endothelial proliferation (Figure A) and downregulated with increasing confluence, with no dynamic regulation of UCP (Figure B and E). We could recapitulate the proliferation-induced changes in Δψ by molecular manipulation of UCP (Figure F) and UCP-null cells had increased Δψ (Figure G). Thus, our data indicate that the proliferation-induced changes in Δψ can be explained, at least in part, by dynamic regulation of UCP. UCP and Endothelial Metabolism Because we found that UCP is dynamically regulated in the endothelium, we probed its implications for basic mitochondrial functions. We found that UCP-null cells had a trend for lower basal respiration rate (P<.7) than wild-type cells and similar rates of basal proton leak. UCP-null cells also had similar maximal respiration (.8±.) than wildtype cells (.6±.58; P=. by -tailed t test; Figure A). Proliferating UCP-null cells also produced lesser lactate than wild-type cells (Figure B), consistent with a lower rate of glycolysis. Suppression of UCP was associated with reduced ATP levels (Figure C) and endothelial ATP levels seemed greatest in endothelium with the highest proliferation rate (Figure D). Collectively, these data indicate that UCP- null cells have more prominent perturbations in glycolysis than respiration, and that endothelial ATP levels vary as a function of proliferation. UCP Regulates Endothelial Phenotype via Changes in Δψ To gain insight into the potential function of UCP isoforms, we examined UCP mrna in silico as a function of cell type. We found that UCP exhibited relatively homogeneous mrna expression across multiple cell types as did UCP, with the exception of a -fold higher expression in skeletal muscle

89 Circulation Research September, A (kda) JC- (Δψ).5 5 FBS (%) B (kda) UCP UCP UCP Actin 6 COX IV Actin JC- (Δψ) 5 6 7 8 9 Cell Confluence (%) C (TMRE / MitoGreen) (%) 5 5 5 6 7 8 9 Cell Confluence (%) D (UCPs mrna / GAPDH mrna) (kda) 5.. UCP UCP Downloaded from http://circres.ahajournals.org/ by guest on July, 8 E F G UCP / Actin (% of 5) 8 6 5 6 7 8 9 Cell Confluence (%) Figure. Regulation of endothelial uncoupling protein (UCP) and Δψ with proliferation. Bovine aortic endothelial cells (BAECs) were (A) stimulated with fetal bovine serum (FBS) or (B) cultured to specific confluence before assessment of Δψ by JC- fluorescence and expression of UCP, UCP, cytochrome c oxidase IV (COX IV), or actin by immunoblotting. Carbonyl cyanide m-chlorophenyl hydrazone (CCCP) of 5-μmol/L was used as a control for the lower limit of Δψ. Data represent n=; P<.5 for trend from -way ANOVA. C, Mitochondrial Δψ by tetramethylrhodamine ethyl ester (TMRE) corrected for mitochondrial mass. n=; P<.5 for trend from -way ANOVA. D, Expression of UCP and UCP mrna and protein in BAECs. E, Densitometric analysis of UCP protein by immunoblot as a function of confluence. n=5; P<.5 vs 5% by ANOVA with Tukey Kramer post hoc test. F, BAECs (control [CTL]) were transfected with UCP (UCP at the indicated multiplicity of infection [MOI]), β-galactosidase (LacZ; MOI), or the indicated small interfering RNA (sirna) followed by assessment for Δψ or protein levels of transfected or endogenous UCP, COX IV, or actin by immunoblotting. n=; P<.5 vs LacZ or sirna control (sictl). G, Mitochondrial Δψ by TMRE as in C. n=; P<.5 by unpaired t test. Ad indicates adenovirus; GADPH indicates glyceraldehyde -phosphate dehydrogenase; siucp, small interfering UCP; and, wild-type. (Online Figure I). In contrast, UCP mrna expression varied more than 5-fold as a function of cell type with the highest transcript levels in rapidly dividing cells, including those harboring erythroid (CD7), endothelial progenitor (CD), and endothelial angiogenic (CD5) markers (Online Figure A Relative O uptake C (Cellular ATP Level) (%) 8 6 Basal Oligo FCCP SiCTL siucp B Lactate (pmol/cell) D (Cellular ATP Level) 5 5 5 (%) 8 6 I). Based on this latter association, we tested known stimuli for angiogenesis such as vascular endothelial growth factor and AMP kinase and found UCP upregulation (Figure A). Serum- and vascular endothelial growth factor mediated UCP upregulation was associated with c-jun N-terminal 5 6 7 8 9 Oligo Cell Confluence (%) Figure. Metabolic signature of uncoupling protein (UCP)-null endothelium. A, Oxygen consumption was determined in wild-type () and UCP-null murine lung endothelial cells (MLECs) in the presence or absence of oligomycin (oligo) or carbonyl cyanide -(trifluoromethoxy) phenylhydrazone (FCCP) using a Clark electrode as described in Methods section. B, Media from wild-type or UCP-null MLECs in.% serum was analyzed for lactate content and expressed as a function of cell count. n=; P<. vs by unpaired t test. C, Bovine aortic endothelial cell (BAEC) ATP levels as a function of UCP status as described in Methods section. n=; P<. vs sirna control (sictl) by unpaired t test. D, BAEC ATP levels as a function of confluence or oligomycin (oligo) treatment. n=; P<. for trend by -way ANOVA. siucp indicates small interfering UCP.

Shimasaki et al UCP and Endothelial Function 895 Downloaded from http://circres.ahajournals.org/ by guest on July, 8 A (kda) B C D E Cell Migration (cell #) F (Sprouts / aorta) G 6 5 (Sprouts / aorta) 6 8 6 8 LacZ UCP UCP SOD COX IV Cyt c Actin Control sirna 5 6 7 8 (days) UCP sirna 9 + UCP + LacZ 5 7 (days) 9 Cell Proliferation (%CTL) Cells / well (x ) 5 8 6 5 8 CTL VEGF Control UCP LacZ UCP sirna sirna 5 7 9 (days) H Perfusion (ischemic/non-ischemic) UCP -/-...8.6.. Cell Migration (cell #) 5 Pre 7 8 (Post-Operative Day) Pre-op POD POD 7 POD POD 8 N I N I N I N I Figure. Uncoupling protein (UCP) modulates endothelial cell proliferation and migration. A, Wild-type () murine lung endothelial cells (MLECs) were treated with aminoimidazole carboxamide ribonucleotide (AICAR;.5 mmol/l) or vascular endothelial growth factor (VEGF; 5 ng/ml) as indicated for hours followed by immunoblots for the indicated proteins. Actin, cytochrome c (Cyt c), and cytochrome c oxidase subunit IV were loading controls for cytosol and mitochondria. B, Bovine aortic endothelial cell (BAEC) proliferation by [ H]-thymidine incorporation in response to VEGF or UCP manipulation as indicated. Data are normalized to the VEGF vehicle control. n= to 5; P<.5 vs respective controls by Mann Whitney U test. C, BAECs underwent UCP manipulation as indicated and migration assessed hours after scratch. n= to 5; P<. vs respective control by Mann Whitney U test. D, Proliferation of and UCP / MLECs by cell count. n=; P<.5 vs by -way repeated measures ANOVA. E, Migration of and UCP / MLECs after scratch wounding. n=5; P<. vs by Mann Whitney U test. F, Aortic segments from the indicated genotypes implanted in collagen gel after 7 days (bar=5 μm) with capillary sprout counts as a function of time. n=6; P<. vs by -way repeated-measures ANOVA. G, Capillary sprouting in UCP / aortic segments in collagen gel containing no additions (contol [CTL]) or the indicated adenoviral vector. n=6 per group; P<. vs vehicle or LacZ by -way repeated measures ANOVA. (H) Blood flow recovery in and UCP / mice with unilateral femoral artery excision expressed as a fraction ratio of the ischemic (I) vs nonischemic (N) limbs with representative images of laser-doppler tissue perfusion in hindlimbs. n=; P=. vs by -way repeated measures ANOVA. Ad indicates adenovirus; POD, postoperative day; sirna, small interfering RNA; and SOD, superoxide dismutase. kinase activation and inhibition attenuated both UCP upregulation and vascular endothelial growth factor mediated endothelial cell proliferation (Online Figure II). We also manipulated endothelial UCP levels and found that endothelial cell proliferation (Figure B) and migration (Figure C) were directly related to UCP. Consistent with these observations, UCP-null endothelium exhibited impaired proliferation (Figure D) and migration (Figure E), without any compensatory upregulation of UCP (Online Figure IIIA). Because endothelial proliferation and migration are features of angiogenesis, we examined the impact of UCP in a capillary sprout formation assay, an in vitro model of early angiogenesis. Wild-type aortic segments exhibited a higher rate and total extent of capillary sprout formation than did UCP-null aortic segments (Figure F), and reconstitution of UCP into UCP-null endothelium rescued the defect in capillary sprout formation (Figure G). Similarly, capillary sprout formation from UCP-null adipose tissue was impaired relative to wildtype, and UCP reintroduction also rescued this defect (Online Figure IIIB and IIIC). To determine the physiological relevance of our in vitro observations, we examined hindlimb ischemiainduced angiogenesis and observed a lower rate of blood flow recovery in UCP-null mice compared with wild type controls (Figure H) that was largely rescued by adenoviral-mediated restoration of UCP expression (Online Figure IIID). UCP Modulates Endothelial Proliferation Via Δψ- Dependent Changes in Mitochondrial O Because Δψ modulates mitochondrial O and UCP can impact Δψ, 8 we investigated mitochondrial O as a function of cell proliferation. Both BAECs (Figure A) and murine endothelial cells (Figure B) exhibited an inverse relation between mitochondrial O and cell proliferation rate and this response was accentuated in UCP-null endothelium (Figure B). Reconstitution of UCP into UCP-null endothelium (Figure C) normalized mitochondrial O and forced overexpression of UCP in wild-type cells attenuated mitochondrial O (Figure C). Treatment of either wild-type

896 Circulation Research September, A B C D 8 6 5% 9% Superoxide (RFU) 5 6 7 8 9 Cell Confluence (%) Superoxide (RFU) LacZ UCP SOD Proliferation (%CTL) 5 5 Downloaded from http://circres.ahajournals.org/ by guest on July, 8 I Cell Migration (cell #) E Superoxide (RFU) 5 5 5 SOD+/- LacZ SOD Control sirna F SOD sirna Cells / well (x ) 6 8 Superoxide (RFU) 6 8 6 SOD+/ 7 (days) CTL SOD sirna sirna J (Sprouts / aorta) 5 5 5 G Superoxide (RFU) 6 Wild-type SOD+/- 5 6 7 (days) H Proliferation (%CTL) K 5 5 Perfusion (ischemic/non-ischemic)...8.6.. CTL LacZ UCP SOD SOD+/- SOD+/- SOD+/- +UCP SOD+/- + LacZ Pre 7 8 (Post-Operative Day) Figure. Uncoupling protein (UCP) dictates endothelial phenotype via mitochondrial O. A, Bovine aortic endothelial cells (BAECs) stained with mitochondrial-targeted hydroethidine at the indicated level of confluence to assess mitochondrial O (bar=5 μm). B, Mitochondrial O in murine lung endothelial cells (MLECs) from the indicated genotype as a function of confluence expressed as relative fluoresence units (RFU). n=; P<. for wild-type () vs UCP / by -way factorial ANOVA. C, MLEC mitochondrial O in the indicated genotypes transfected with UCP (UCP), SOD (SOD), or LacZ. n=; P<.5 vs UCP / LacZ; P<.5 vs LacZ. D, Proliferation of UCP-null MLECs transfected as in C with data normalized to control (CTL). n=; P<.5 vs LacZ by Kruskal Wallis ANOVA. E, Mitochondrial O in vs SOD +/ endothelium. n= to 8; P<.5 vs by Mann Whitney U test. F, Proliferation of and SOD +/ MLECs as a function of time. n=; P<.5 for trend by -way repeated measures ANOVA. Mitochondrial O (G) and proliferation (H) in SOD +/ MLECs as a function of transfection with control (LacZ), UCP, or SOD adenovirus. n=; P<.5 vs LacZ by Kruskal Wallis ANOVA. I, left, Migration of BAECs with manipulated SOD levels via the indicated adenovirus or small interfering RNA (sirna). n=6; P=. vs respective controls by Mann Whitney U test. I, right, BAEC Mitochondrial O as a function of treatment with CTL or SOD sirna. n=; P<. vs control sirna by Student t test. J, Capillary sprouting in aortic segments from the indicated genotypes. n=6; P<. vs by -way repeated measures ANOVA. K, Blood flow recovery in and SOD +/ mice after unilateral hindlimb ischemia with or without hindlimb transfection with UCP or LacZ adenovirus. n=5 per group; P<. vs and P<. vs LacZ by -way repeated measures ANOVA. Ad indicates adenovirus; and SOD, superoxide dismutase. or UCP-null endothelium with mitochondrial SOD (SOD) also reduced the mitochondrial O flux (Figure C). Finally, attenuation of mitochondrial O in UCP-null cells with either UCP or SOD (Figure C) enhanced endothelial cell proliferation (Figure D). These data indicate UCP modulates endothelial cell proliferation via mitochondrial O. If mitochondrial O explains the UCP-null phenotype, then independent mitochondrial O manipulation should produce qualitatively similar effects. To this end, we found that SOD +/ endothelium exhibited increased mitochondrial O (Figure E) and impaired proliferation (Figure F) that were both corrected by either UCP or SOD transfection (Figure G and H). Forced overexpression of SOD in BAECs enhanced migration in the scratch assay (Figure I), whereas SOD suppression inhibited migration and increased mitochondrial O (Figure I). Capillary sprouting in aortic segments from SOD +/ animals was impaired compared with wild-type animals (Figure J). In the hindlimb ischemia model, we found less blood flow recovery in SOD +/ mice than in wild-type mice that was rescued via adenoviral overexpression of UCP (Figure K) in a manner that suppressed mitochondrial O (Figure G) and improved proliferation (Figure H) in the endothelium. Thus, excess mitochondrial O produces impaired endothelial cell proliferation, migration, and angiogenesis.

Shimasaki et al UCP and Endothelial Function 897 Downloaded from http://circres.ahajournals.org/ by guest on July, 8 A JC- (Δψ) C Cell Migration (Cell #)..5..5 5 5 5 LacZ UCP LacZ UCP 5 UCP Does Not Limit Endothelial Antioxidants or Impact Basal NO Bioactivity Because endothelial antioixdants and O can impact nitric oxide (NO ) bioactivity, we investigated both as a function of UCP. Acute suppression of UCP had no material impact on BAEC enzymatic antioxidant levels (Online Figure IVA). In addition, stimulated NO bioactivity with acetylcholine was not different between wild-type and UCP-null vessels (Online Figure IVB) and basal NO bioactivity was actually greater in UCP-null vessels than wild-type vessels (Online Figure IVC and IVD). Thus, the impact of mitochondrial O on endothelial phenotype is not dependent on reduced antioxidant levels or basal NO bioactivity. Limiting Δψ Is Sufficient to Explain the Impact of UCP To confirm that Δψ changes can account for the impact of UCP effects on endothelial phenotype, we force expressed UCP, a protein not expressed in endothelium, to decrease endothelial Δψ. We found that forced expression of UCP decreased Δψ (Figure 5A), increased proliferation (Figure 5B), increased migration (Figure 5C), and limited mitochondrial O (Figure 5D). Thus, independently decreasing Δψ with UCP produced qualitatively similar changes in endothelial cell phenotype as Δψ manipulation with UCP. These data suggest that the impact of UCP on endothelial phenotype is because, in part, of its effect on mitochondrial Δψ. UCP Regulates Cell Cycle Progression via Mitochondrial O Because NO bioactivity was not impaired, we examined other mechanisms of impaired endothelial function. Proliferating UCP-null endothelium exhibited more cells in the G-phase cells than wild-type endothelium (Figure 6A), suggesting B Cell Proliferation (RFU) D 5 5 LacZ LacZ UCP UCP Figure 5. Independent manipulation of Δψ impacts endothelial phenotype. Bovine aortic endothelial cells (BAECs) were transfected with irrelevant (LacZ) or uncoupling protein adenovirus (UCP) and examined for (A) Δψ, (B) proliferation, (C) migration, and (D) mitochondrial O as described in Methods section. n= to 6; P<.5 vs LacZ by unpaired t test. (MitoSOX /MitoGreen) impaired G-S cell cycle transition. This part of the cell cycle is controlled, in part, by cyclin-dependent kinases, and we found that UCP-null cells exhibited higher expression of the cyclindependent kinase inhibitors, such as p6 inka and p cip/waf, than did wild-type cells (Figure 6B and 6C). Because cell cycle inhibition can lead to senescence, we examined the senescence indicator, β-galactosidase, and found a.5-fold increase in senescence in UCP-null endothelium compared with wild-type cells (Figure 6D and 6E). Transfection of UCP-null cells with either UCP or SOD (Figure 6E) attenuated senescence. We observed a similar senescence increase in SOD +/ cells (Figure 6E) and transfection of either UCP or SOD also inhibited the development of senescence (Figure 6E). Because p5 is both sensitive to reactive oxygen species (ROS) 9 and can control p6 inka and p cip/waf expression, we investigated the role of p5 in the UCP-null phenotype. We found that p5 suppression by small interfering RNA rescued the proliferation defect of UCP-null endothelium (Figure 6F). Moreover, we observed that proliferating and hypoxic UCP-null endothelium exhibited p5 phosphorylation at serine 8 (Figure 6G), a residue involved in redox-sensitive p5 activation. 9 Thus, UCP-null endothelium exhibits a dysfunctional phenotype manifest as premature senescence via excess mitochondrial O in a p5-dependent manner. UCP Regulates Endothelial Phenotype via Superoxide-Dependent Control of Mitochondrial Fragmentation Because mitochondria undergo dynamic changes in fusion and fission during the G-S cell cycle transition, 5 we examined endothelial mitochondrial morphology as a function of UCP. Compared with wild-type endothelium, proliferating UCP- null cells exhibited superoxide-dependent mitochondrial fragmentation that was rescued by transfection with either UCP or SOD (Figure 7A and 7B). Similarly, SOD heterozygous cells exhibited greater mitochondrial fragmentation than wildtype cells that was largely rescued by transfection with SOD or UCP (Figure 7A and 7B). This fragmentation was linked to p5, because p5 RNA interference attenuated fragmentation (Figure 7C), whereas p cip/waf forced overexpression recapitulated mitochondrial fragmentation (Figure 7D). To gain insight into the mechanism of mitochondrial fragmentation, we examined gene expression important for normal mitochondrial morphology (Figure 7E) and found that genes required for fusion of mitochondria, such as mitofusin and (Mfn, Mfn) and optic atrophy (Opa), were significantly lower than genes required for mitochondrial fission, such as fission (Fis) and dynamin-related protein (Drp) a pattern associated with mitochondrial fragmentation. 6 To determine whether mitochondrial fragmentation was sufficient to cause the UCP-null phenotype, we suppressed mitofusin and expression and found that induction of mitochondrial fragmentation significantly reduced endothelial cell proliferation (Figure 7F). Taken together, these results indicate that endothelial UCP is necessary for maintaining mitochondrial morphology, and that disrupting the normal balance between mitochondrial fusion and fission has implications for endothelial function.

898 Circulation Research September, Downloaded from http://circres.ahajournals.org/ by guest on July, 8 A hr G G 65.% G.% S 9.7% S 7.7% G 7.% G.5% G 8 hr D F # Cells # Cells Cell Proliferation (RFU) 5 5 G 7.5% S 6.% G.% PI Content Wild type No Treatment Control sirna G 9.% S.% G.5% PI Content p5 sirna Discussion The principal finding of this work is that endogenous UCP modulates endothelial mitochondrial network morphology that dictates, in part, endothelial cell function. Through UCP, endothelial Δψ decreases with cell proliferation and, as endothelial growth slows, reduced UCP levels produce an increase in Δψ. We found these changes in Δψ moved in parallel with mitochondrial O production that, if left uncontrolled, resulted in p5-dependent mitochondrial network fragmentation that limited endothelial functions important for angiogenesis, including proliferation, migration, and blood flow recovery from tissue ischemia. The central role of Δψ was supported by observations that Δψ manipulation with UCP could phenocopy the effect of UCP. Moreover, the key role of mitochondrial O was consistent with observations that SOD +/ endothelium, with excess mitochondrial O, demonstrated mitochondrial network fragmentation and impaired endothelial cell function. Collectively, these data indicate that endogenous UCP is an important modulator of endothelial cell function, in part, via its impact on mitochondrial network integrity. The precise biological functions of uncoupling proteins remain a matter of debate. UCP, the predominant uncoupling protein in brown adipose tissue, mediates Δψ proton leak that is critical for adaptive thermogenesis. 7 In contrast, UCP 8 and UCP 9 can mediate proton leak, but are not necessary for adaptive thermogenesis or normal energy metabolism. 8, These latter uncoupling proteins modulate Δψ to a lesser extent than UCP, 8 B p6 mrna G E SA-β-gal stained cells (%) p-p5 p5 Actin GFP UCP SOD GFP UCP SOD GFP Normoxia p mrna C p6 p Actin Hypoxia SOD+/- (kda) (Ser8) 5 (kda) 6 Figure 6. Uncoupling protein (UCP) and endothelial senescence. A, Cell cycle analysis in wild-type () and UCP-null murine lung endothelial cells (MLECs) at the indicated time based upon propidium iodide (PI) content. B, Expression of and UCP-null mrna for p6 inka and p cip/waf. n=; P<.5 vs by Mann Whitney U test. C, Expression of and UCP- null endothelial p6 inka and p cip/waf by immunoblotting. D, and UCP-null senescence-associated β-galactosidase (SA-β-gal) via x-gal staining. Bar=75 μm. E, Endothelial SA-β-gal staining as a function of genotype and transfection with either UCP or SOD. n= to 6; P<.5 vs +GFP, P<.5 vs UCP / + GFP, and P<.5 vs SOD +/ +GFP by Kruskal Wallis ANOVA and post hoc comparison. F, Cell proliferation in and UCP / MLECs treated with none, control, or p5 small interfering RNA (sirna). n=; P<.5 vs with control sirna; P<. vs with no treatment; P<. vs with control sirna; and P<. vs UCP / with control sirna by -way ANOVA with Tukey Kramer test. G, MLECs of the indicated genotype under normoxic or hypoxic (% O ) conditions for 6 hours were lysed and probed for the indicated proteins by immunoblot. Ad indicates adenovirus; and SOD, superoxide dismutase. even in the presence of requisite coactivators, such as fatty acids and O -derived alkenals., Multiple models of UCP or UCP overexpression suggest a protective role against oxidative stress 8,, and endothelial dysfunction from diet-induced obesity. 5 Considering that endothelial cell proliferation can be associated with excess O and oxidative stress, 6 one could argue that our data identify UCP as a stress-responsive protein. However, UCP overexpression is prone to artifact from improper membrane insertion, casting doubt on the ultimate role of UCPs in limiting O -mediated toxicity. Our data strongly support a role of UCP to limit mitochondrial O and deserve particular attention, because we focused on endogenous UCP regulation. Moreover, we used loss-of-function models with rescue by UCP complementation that are less prone to artifact. Consistent with the idea that UCP limits mitochondrial O, we found that independent manipulation of mitochondrial O (via SOD) copied the UCP-null phenotype, increasing confidence that endogenous UCP importantly regulates mitochondrial ROS. One might expect that mitochondrial O would limit NO bioactivity in our system to explain the UCP-null phenotype. Rather, we found intact NO bioactivity in unstressed UCP-null endothelium, but the cells exhibited premature senescence and upregulation of p cip/waf and p6 inka. These gene products have been linked, in part, to activation of the tumor suppressor, p5. 7 Our observations that correction of excess mitochondrial O prevented senescence is consistent with data that ROS can upregulate p5 to induce both cell cycle arrest 8 and genes that tend to lower cellular ROS levels. 9,5 In this context, it is germane to

Shimasaki et al UCP and Endothelial Function 899 A + AdUCP SOD+/- SOD+/- + AdSOD Downloaded from http://circres.ahajournals.org/ by guest on July, 8 B Mitochondrial Length (μm) D LacZ UCP SOD LacZ UCP SOD + AdLacZ Mitochondrial Length (μm) sictl si-p + p SOD+/- Adp LacZ E F Relative Expression to GAPDH Cell Number / Field C Mitochondrial Length (μm)..5..5 5 Mfn Mfn CTL sirna Mfn sirna sictlsi-p5 sictlsi-p5 sictlsi-p5 5 Days in Culture Opa 5 Fis Drp SOD+/- Ucp CTL sirna Mfn+ sirna 5 Days in Culture Figure 7. Mitochondrial morphology and endothelial phenotype. A, Representative images of murine lung endothelial cell (MLEC) mitochondrial morphology by Mitotracker green staining as a function of the indicated genotype and transfection with the indicated adenovirus (Ad). B, Composite data indicating average mitochondrial length as an index of fragmentation in MLECs from the indicated genotype and transfected with the indicated adenovirus. n= to 6; P<. vs respective wildtype () or LacZ control by Kruskal Wallis ANOVA. C, Mitochondrial length as an index of fragmentation in the indicated genotypes with and without suppression of p5. n= to 6; P<.5 vs respective control (CTL). D, MLEC mitochondrial morphology and average length with forced expression of p cip/waf. Bar=5 μm; n=9 to ; P<.5 vs LacZ by Kruskal Wallis ANOVA. E, MLEC expression of mitochondrial genes as a function of the indicated genotype. n= to ; P<.5, P<. vs by Mann Whitney U test. F, MLEC proliferation with suppression of mitofusin (Mfn) or both mitofusin (Mfn) and Mfn. n= each; P<.5 vs CTL sirna by -way repeated measures ANOVA. Drp indicates dynamin-related protein ; Fis, fission ; opa, optic atrophy ; sictl, sirna control; and SOD indicates superoxide dismutase. note that RNA interference mediated suppression of prohibitin, an inner mitochondrial protein, also upregulates mitochondrial O, producing premature senescence and impaired endothelial proliferation. 5 Studies in other cell types indicate prohibitins are required for proliferation, 5 prompting speculation that suppressing mitochondrial O (perhaps to prevent p5 activation) may be a general requirement of cellular proliferation, perhaps via prevention of premature senescence. In this regard, we need to interpret our data with caution as we have not determined whether senescence alone could explain all of our experimental findings. The notion that UCP suppression of mitochondrial O may be important for cell proliferation is supported by our observation that UCP mrna is most prominent in very proliferative cells and tissues, such as bone marrow, lymphomas, leukemias, erythroid precursors, T-cells, and CD5 + endothelial cells (Online Figure I). Moreover, UCP gene silencing prevented epidermal cell tumor induction in a p5-dependent manner. 5 Thus, data from diverse cell types link UCP upregulation (and mitochondrial O suppression) to rapid cell proliferation, suggesting that UCP impacts some fundamental component of the proliferative response. Because UCP is a target gene for peroxisome proliferator γ coactivator-α, the principal determinant of mitochondrial biogenesis, 5 UCP may be needed to expand mitochondrial mass during cell proliferation, perhaps by limiting mitochondrial ROS-mediated damage. Alternatively, UCP is known to promote oxygeninsensitive glycolysis (the Warburg effect) 55 and glycolysis is required for the high proliferation rate of many cancer cells. Thus, UCP may be required for glycolysis-dependent growth, a finding consistent with our observations of reduced lactate production in UCP-null endothelium (Figure ). We observed excess mitochondrial fragmentation with UCP-null mice that was linked to mitochondrial O, because it was recapitulated in the SOD +/ mice and attenuated with maneuvers that reduce mitochondrial O. These findings suggest a novel relation between mitochondrial redox state and the balance between mitochondrial fusion and fission. This relation is consistent with our observations that mitochondrial

9 Circulation Research September, Downloaded from http://circres.ahajournals.org/ by guest on July, 8 fragmentation in UCP-null cells involved a p5- and p cip/waf - dependent mechanism, and literature that p5 can be activated by ROS via p5 phosphorylation on serine (mouse Ser8). 9 This latter event is known to involve the ataxia-telangiectasia mutated kinase. Thus, one potential explanation for our data would be ataxia-telangiectasia mutated kinase activation by mitochondrial O. This contention is supported by observations that ataxia-telangiectasia mutated and its p5 target (Ser8/) have important implications for metabolism. 56,57 The findings presented here suggest a role for UCP in cell proliferation. With the increased metabolic flux that occurs as a result of growth stimulation, it is not surprising that mitochondrial Δψ and O would increase. The fact that forced expression of UCP limited Δψ and recapitulated the effect of UCP suggests that UCP upregulation is an important mechanism for limiting Δψ and the resultant mitochondrial O flux that, if left unchecked, would activate pathways to limit cell proliferation. With regards to the endothelium, this paradigm has important therapeutic implications. For example, our findings suggest that targeting UCP may prove to be an important means of limiting tumor angiogenesis. Conversely, promotion of mitochondrial uncoupling could enhance the angiogenic response to ischemia and this strategy could prove helpful in the setting of occlusive vascular diseases. Thus, our data highlight the importance of mitochondrial uncoupling and its dynamic regulation in the endothelial cell proliferative response. Acknowledgments We kindly thank Dr Bradford Lowell for the UCP-null mice. Sources of Funding This work was supported by National Institutes of Health grants DK8985 (M.P. Cooper), HL9 and HL987 (J.F. Keaney). Informatics and imaging core services were supported by University of Massachusetts D.E.R.C. grant DK5. None. Disclosures References. Mitchell P. Coupling of phosphorylation to electron and hydrogen transfer by a chemi-osmotic type of mechanism. Nature. 96;9: 8.. Jiang X, Wang X. Cytochrome C-mediated apoptosis. Annu Rev Biochem. ;7:87 6.. Ajioka RS, Phillips JD, Kushner JP. Biosynthesis of heme in mammals. Biochim Biophys Acta. 6;76:7 76.. Berridge MJ, Bootman MD, Roderick HL. Calcium signalling: dynamics, homeostasis and remodelling. Nat Rev Mol Cell Biol. ;:57 59. 5. Seth RB, Sun L, Ea CK, Chen ZJ. Identification and characterization of MAVS, a mitochondrial antiviral signaling protein that activates NFkappaB and IRF. Cell. 5;:669 68. 6. Chen H, Detmer SA, Ewald AJ, Griffin EE, Fraser SE, Chan DC. Mitofusins Mfn and Mfn coordinately regulate mitochondrial fusion and are essential for embryonic development. J Cell Biol. ;6: 89. 7. Siemens TV, Nichols DL, Zitomer RS. Effect of mitochondrial functions on synthesis of yeast cytochrome c. J Bacteriol. 98;:99 57. 8. Parikh VS, Morgan MM, Scott R, Clements LS, Butow RA. The mitochondrial genotype can influence nuclear gene expression in yeast. Science. 987;5:576 58. 9. Forsburg SL, Guarente L. Communication between mitochondria and the nucleus in regulation of cytochrome genes in the yeast Saccharomyces cerevisiae. Annu Rev Cell Biol. 989;5:5 8.. Hörtner H, Ammerer G, Hartter E, Hamilton B, Rytka J, Bilinski T, Ruis H. Regulation of synthesis of catalases and iso--cytochrome c in Saccharomyces cerevisiae by glucose, oxygen and heme. Eur J Biochem. 98;8:79 8.. Saltzgaber-Müller J, Schatz G. Heme is necessary for the accumulation and assembly of cytochrome c oxidase subunits in Saccharomyces cerevisiae. J Biol Chem. 978;5:5.. Jiang X, Wang X. Cytochrome C-mediated apoptosis. Annu Rev Biochem. ;7:87 6.. Naresh A, Long W, Vidal GA, Wimley WC, Marrero L, Sartor CI, Tovey S, Cooke TG, Bartlett JM, Jones FE. The ERBB/HER intracellular domain ICD is a BH-only protein promoting apoptosis of breast cancer cells. Cancer Res. 6;66:6 6.. Schulz E, Dopheide J, Schuhmacher S, Thomas SR, Chen K, Daiber A, Wenzel P, Münzel T, Keaney JF Jr. Suppression of the JNK pathway by induction of a metabolic stress response prevents vascular injury and dysfunction. Circulation. 8;8:7 57. 5. Al-Mehdi AB, Pastukh VM, Swiger BM, Reed DJ, Patel MR, Bardwell GC, Pastukh VV, Alexeyev MF, Gillespie MN. Perinuclear mitochondrial clustering creates an oxidant-rich nuclear domain required for hypoxiainduced transcription. Sci Signal. ;5:ra7. 6. Kamata H, Honda S, Maeda S, Chang L, Hirata H, Karin M. Reactive oxygen species promote TNFalpha-induced death and sustained JNK activation by inhibiting MAP kinase phosphatases. Cell. 5;:69 66. 7. Chen K, Thomas SR, Albano A, Murphy MP, Keaney JF Jr. Mitochondrial function is required for hydrogen peroxide-induced growth factor receptor transactivation and downstream signaling. J Biol Chem. ;79: 579 586. 8. Krauss S, Zhang CY, Lowell BB. The mitochondrial uncoupling-protein homologues. Nat Rev Mol Cell Biol. 5;6:8 6. 9. Kuhlencordt PJ, Rosel E, Gerszten RE, Morales-Ruiz M, Dombkowski D, Atkinson WJ, Han F, Preffer F, Rosenzweig A, Sessa WC, Gimbrone MA Jr, Ertl G, Huang PL. Role of endothelial nitric oxide synthase in endothelial activation: insights from enos knockout endothelial cells. Am J Physiol Cell Physiol. ;86:C95 C.. Yamagishi SI, Edelstein D, Du XL, Kaneda Y, Guzmán M, Brownlee M. Leptin induces mitochondrial superoxide production and monocyte chemoattractant protein- expression in aortic endothelial cells by increasing fatty acid oxidation via protein kinase A. J Biol Chem. ;76:596 5.. Smiley ST, Reers M, Mottola-Hartshorn C, Lin M, Chen A, Smith TW, Steele GD Jr, Chen LB. Intracellular heterogeneity in mitochondrial membrane potentials revealed by a J-aggregate-forming lipophilic cation JC-. Proc Natl Acad Sci USA. 99;88:67 675.. Chen K, Kirber MT, Xiao H, Yang Y, Keaney JF Jr. Regulation of ROS signal transduction by NADPH oxidase localization. J Cell Biol. 8;8: 9 9.. Liu L, Sanosaka M, Lei S, Bestwick ML, Frey JH Jr, Surovtseva YV, Shadel GS, Cooper MP. LRP protein remodels mitochondria and stimulates fatty acid oxidation. J Biol Chem. ;86:5 6.. Cooper MP, Qu L, Rohas LM, Lin J, Yang W, Erdjument-Bromage H, Tempst P, Spiegelman BM. Defects in energy homeostasis in Leigh syndrome French Canadian variant through PGC-alpha/LRP complex. Genes Dev. 6; :996 9. 5. Gealekman O, Burkart A, Chouinard M, Nicoloro SM, Straubhaar J, Corvera S. Enhanced angiogenesis in obesity and in response to PPARgamma activators through adipocyte VEGF and ANGPTL production. Am J Physiol Endocrinol Metab. 8;95:E56 E6. 6. Nicosia RF, Ottinetti A. Growth of microvessels in serum-free matrix culture of rat aorta. A quantitative assay of angiogenesis in vitro. Lab Invest. 99;6:5. 7. Zhang CY, Baffy G, Perret P, Krauss S, Peroni O, Grujic D, Hagen T, Vidal-Puig AJ, Boss O, Kim YB, Zheng XX, Wheeler MB, Shulman GI, Chan CB, Lowell BB. Uncoupling protein- negatively regulates insulin secretion and is a major link between obesity, beta cell dysfunction, and type diabetes. Cell. ;5:75 755. 8. Craige SM, Chen K, Pei Y, Li C, Huang X, Chen C, Shibata R, Sato K, Walsh K, Keaney JF Jr. NADPH oxidase promotes endothelial angiogenesis through endothelial nitric oxide synthase activation. Circulation. ;:7 7. 9. Chen K, Albano A, Ho A, Keaney JF Jr. Activation of p5 by oxidative stress involves platelet-derived growth factor-beta receptor-mediated ataxia telangiectasia mutated (ATM) kinase activation. J Biol Chem. ;78:957 95.. Song W, Bossy B, Martin OJ, Hicks A, Lubitz S, Knott AB, Bossy- Wetzel E. Assessing mitochondrial morphology and dynamics using

Shimasaki et al UCP and Endothelial Function 9 Downloaded from http://circres.ahajournals.org/ by guest on July, 8 fluorescence wide-field microscopy and D image processing. Methods. 8;6:95.. Lysetska M, Knoll A, Boehringer D, Hey T, Krauss G, Krausch G. UV light-damaged DNA and its interaction with human replication protein A: an atomic force microscopy study. Nucleic Acids Res. ;:686 69.. Vander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science. 9;:9.. Nagata D, Mogi M, Walsh K. AMP-activated protein kinase (AMPK) signaling in endothelial cells is essential for angiogenesis in response to hypoxic stress. J Biol Chem. ;78: 6.. Beckman JS, Beckman TW, Chen J, Marshall PA, Freeman BA. Apparent hydroxyl radical production by peroxynitrite: implications for endothelial injury from nitric oxide and superoxide. Proc Natl Acad Sci USA. 99;87:6 6. 5. Mitra K, Wunder C, Roysam B, Lin G, Lippincott-Schwartz J. A hyperfused mitochondrial state achieved at G-S regulates cyclin E buildup and entry into S phase. Proc Natl Acad Sci USA. 9;6:96 965. 6. Detmer SA, Chan DC. Functions and dysfunctions of mitochondrial dynamics. Nat Rev Mol Cell Biol. 7;8:87 879. 7. Puigserver P, Wu Z, Park CW, Graves R, Wright M, Spiegelman BM. A cold-inducible coactivator of nuclear receptors linked to adaptive thermogenesis. Cell. 998;9:89 89. 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. Disruption of the uncoupling protein- gene in mice reveals a role in immunity and reactive oxygen species production. Nat Genet. ;6:5 9. 9. Gong DW, Monemdjou S, Gavrilova O, Leon LR, Marcus-Samuels B, Chou CJ, Everett C, Kozak LP, Li C, Deng C, Harper ME, Reitman ML. Lack of obesity and normal response to fasting and thyroid hormone in mice lacking uncoupling protein-. J Biol Chem. ;75:65 657.. Brand MD, Esteves TC. Physiological functions of the mitochondrial uncoupling proteins UCP and UCP. Cell Metab. 5;:85 9.. Echtay KS, Murphy MP, Smith RA, Talbot DA, Brand MD. Superoxide activates mitochondrial uncoupling protein from the matrix side. Studies using targeted antioxidants. J Biol Chem. ;77:79 75.. Murphy MP, Echtay KS, Blaikie FH, Asin-Cayuela J, Cocheme HM, Green K, Buckingham JA, Taylor ER, Hurrell F, Hughes G, Miwa S, Cooper CE, Svistunenko DA, Smith RA, Brand MD. Superoxide activates uncoupling proteins by generating carbon-centered radicals and initiating lipid peroxidation: studies using a mitochondria-targeted spin trap derived from alpha-phenyl-n-tert-butylnitrone. J Biol Chem. ;78:85 855.. Barreiro E, Garcia-Martínez C, Mas S, Ametller E, Gea J, Argilés JM, Busquets S, López-Soriano FJ. UCP overexpression neutralizes oxidative stress rather than nitrosative stress in mouse myotubes. FEBS Lett. 9;58:5 56. What Is Known? Mitochondria are known to participate in a number of cellular functions beyond that of energy production. The main driving force for many mitochondrial functions is the electrochemical gradient (Δψ) produced by the mitochondrial electron transport chain. The regulation of Δψ involves both the electron transport chain and uncoupling proteins that facilitate protein leak across the inner mitochondrial membrane. What New Information Does This Article Contribute? Mitochondrial Δψ is dynamically regulated in endothelial cells by uncoupling protein (UCP) in response to external stimuli such signals to promote proliferation. The decrease in endothelial cell Δψ by UCP upregulation during proliferation limits mitochondrial superoxide generation. In the absence of UCP, excess mitochondrial superoxide leads to p5-dependent mitochondrial fragmentation that is necessary and sufficient to impair endothelial cell proliferation and angiogenesis. Novelty and Significance. Pecqueur C, Alves-Guerra MC, Gelly C, Levi-Meyrueis C, Couplan E, Collins S, Ricquier D, Bouillaud F, Miroux B. Uncoupling protein, in vivo distribution, induction upon oxidative stress, and evidence for translational regulation. J Biol Chem. ;76:875 87. 5. Tian XY, Wong, Xu A, Lu Y, Zhang Y, Wang L, Cheang WS, Wang Y, Yao X, Huang Y. Uncoupling protein- protects endothelial function in diet-induced obese mice. Circ Res. ;: 6. 6. Abid MR, Shih SC, Otu HH, Spokes KC, Okada Y, Curiel DT, Minami T, Aird WC. A novel class of vascular endothelial growth factor-responsive genes that require forkhead activity for expression. J Biol Chem. 6;8:55 555. 7. Gottlieb E, Vousden KH. p5 regulation of metabolic pathways. Cold Spring Harb Perspect Biol. ;:a. 8. Levine AJ. p5, the cellular gatekeeper for growth and division. Cell. 997;88:. 9. Bensaad K, Tsuruta A, Selak MA, Vidal MN, Nakano K, Bartrons R, Gottlieb E, Vousden KH. TIGAR, a p5-inducible regulator of glycolysis and apoptosis. Cell. 6;6:7. 5. Budanov AV, Sablina AA, Feinstein E, Koonin EV, Chumakov PM. Regeneration of peroxiredoxins by p5-regulated sestrins, homologs of bacterial AhpD. Science. ;:596 6. 5. Schleicher M, Shepherd BR, Suarez Y, Fernandez-Hernando C, Yu J, Pan Y, Acevedo LM, Shadel GS, Sessa WC. Prohibitin- maintains the angiogenic capacity of endothelial cells by regulating mitochondrial function and senescence. J Cell Biol. 8;8:. 5. Sievers C, Billig G, Gottschalk K, Rudel T. Prohibitins are required for cancer cell proliferation and adhesion. PLoS One. ;5:e75. 5. Wang F, Fu X, Chen X, Chen X, Zhao Y. Mitochondrial uncoupling inhibits p5 mitochondrial translocation in TPA-challenged skin epidermal JB6 cells. PLoS One. ;5:e59. 5. Wu Z, Puigserver P, Andersson U, Zhang C, Adelmant G, Mootha V, Troy A, Cinti S, Lowell B, Scarpulla RC, Spiegelman BM. Mechanisms controlling mitochondrial biogenesis and respiration through the thermogenic coactivator PGC-. Cell. 999;98:5. 55. Samudio I, Fiegl M, McQueen T, Clise-Dwyer K, Andreeff M. The Warburg effect in leukemia-stroma cocultures is mediated by mitochondrial uncoupling associated with uncoupling protein activation. Cancer Res. 8;68:598 55. 56. Schneider JG, Finck BN, Ren J, Standley KN, Takagi M, Maclean KH, Bernal-Mizrachi C, Muslin AJ, Kastan MB, Semenkovich CF. ATMdependent suppression of stress signaling reduces vascular disease in metabolic syndrome. Cell Metab. 6;:77 89. 57. Armata HL, Golebiowski D, Jung DY, Ko HJ, Kim JK, Sluss HK. Requirement of the ATM/p5 tumor suppressor pathway for glucose homeostasis. Mol Cell Biol. ;:5787 579. Mitochondria have long been known to be critical for cellular ATP production. More recently, mitochondrial have been implicated in other cellular processes, but their role in the endothelium is largely unknown. Thus, we sought to probe the role of mitochondria in endothelial proliferative responses. We found that endothelial cell proliferation was associated with a reduced Δψ because of upregulation of UCP, a protein known to regulate Δψ. In the absence of UCP, proliferating endothelial cells had higher Δψ and excess mitochondrial superoxide that prevented their proliferation, migration, and participation in angiogenesis. Similarly, endothelium lacking mitochondrial superoxide dismutase had excess mitochondrial superoxide and mimicked the UCP-null phenotype. This excess mitochondrial superoxide lead to p5-dependent mitochondrial fragmentation and endothelial cell senescence. Correction of excess mitochondrial superoxide corrected both the mitochondrial fragmentation and endothelial dysfunction. These data indicate that endothelial UCP is important to maintain both normal mitochondrial dynamics and endothelial function.

Downloaded from http://circres.ahajournals.org/ by guest on July, 8 Uncoupling Protein Impacts Endothelial Phenotype via p5-mediated Control of Mitochondrial Dynamics Yukio Shimasaki, Ning Pan, Louis M. Messina, Chunying Li, Kai Chen, Lijun Liu, Marcus P. Cooper, Joseph A. Vita and John F. Keaney, Jr Circ Res. ;:89-9; originally published online July, ; doi:.6/circresaha..9 Circulation Research is published by the American Heart Association, 77 Greenville Avenue, Dallas, TX 75 Copyright American Heart Association, Inc. All rights reserved. Print ISSN: 9-7. Online ISSN: 5-57 The online version of this article, along with updated information and services, is located on the World Wide Web at: http://circres.ahajournals.org/content//7/89 Data Supplement (unedited) at: http://circres.ahajournals.org/content/suppl//7//circresaha..9.dc Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: http://www.lww.com/reprints Subscriptions: Information about subscribing to Circulation Research is online at: http://circres.ahajournals.org//subscriptions/

CIRCRES//9/R Supplemental Material Supplemental Experimental Procedures Bioinformatics The gene expression profiles of human UCP,, and in various tissues were examined using the BioGPS database from microarrays (Genomics Institute of the Novartis Research Foundation; San Diego, CA). Capillary Sprouting Assay in Adipose Tissues Freshly harvested epididymal fat pads were placed in Krebs-Ringer solution buffered with HEPES (KRH) as described, digested for min at 7 C in KRH, ph 7., containing mg/ml collagenase type I (Worthington Biochemical; Lakewood, NJ) and.5% BSA Fraction V (Sigma-Aldrich). Digested tissue was then filtered through a µm cell strainer (Fisher Scientific BD Falcon; Pittsburgh, PA) and the captured stromal vascular fraction was cut into small ( mm ) pieces that were embedded in liquid collagen gel in 8-well plates (BD Biosciences), incubated at 7 C for h to polymerize the collagen, and the solid gel covered with MLEC medium diluted : with DMEM. Each adipose tissue was examined daily and capillary sprouts counted along the sample perimeter under and x magnification. Vascular sprouts were distinguished from fibroblasts via morphology as described, and CD staining. Vascular Ring Relaxation To study NO bioactivity, mice aged 8- weeks were euthanized and thoracic aortae were isolated, cleaned of connective tissue, and cut into mm segments. The vessel segments were mounted on stainless-steel holders in organ baths (Danish Myo Technology, Denmark) containing physiological saline solution (PSS; 9 mm NaCl,.69 mm KCl,.7 mm MgSO,.8 mm KH PO,.5 mm CaCl, 5 mm NaHCO,. mm EDTA, 5.5

CIRCRES//9/R mm Glucose) and aerated with 95% O / 5% CO for isometric force recording. Preparations were allowed to equilibrate for min under constant passive force of (~ mn) and synchronized with KCl and phenylephrine. Basal NO bioactivity was estimated as the difference in contraction to phenylephrine with or without µm L-NAME to inhibit endothelial nitric oxide synthase. Evoked NO bioactivity was estimated as the relaxation response to acetylcholine (ACh) in phenylephrine-contracted segments. Data were analyzed with PowerLab software (AD Instruments).

CIRCRES//9/R Human UCP Human UCP Human UCP Median xm 5 Median xm xm xm 5 xm Median 5 xm Online Figure I. Uncoupling protein mrna expression as a function of cell type. Gene expression profiles of human UCP,, and as a function of cell type examined with the BioGPS online database. Expression levels are indicated as a multiple of the median expression level.

CIRCRES//9/R A 5 5 6 8 5 5 6 8 min C P-c-Jun B P-p8 Actin CTL SP65 SB58 5 5 h 5 5 h 5 5 h P-c-Jun Proliferation (% DMSO) 5 5 P-p8 UCP DMSO DMSO SP.5 SP.5 (µm) VEGF Online Figure II. c-jun, N-terminal kinase and UCP upregulation. (A) Wild-type () and UCP-null endothelium was exposed to % serum for the indicated number of minutes, lysed, and immunoblots performed for actin or phosphorylated (activated) forms of c-jun (Ser6) or p8 MAP kinase (Thr8/Tyr8). (B) BAECs in the presence or absence of vehicle alone (CTL), the JNK inhibitor SP65 (.5 µm), or the p8 MAP kinase inhibitor SB58 ( µm) were exposed to % serum for the indicated time. Cells were then lysed and examined for UCP or activation of JNK or p8 MAP kinase as in (A). (C) BAECs were exposed to vehicle (DMSO) or VEGF (5 ng/ml) with or without SP65 as indicated for h followed by assessment of proliferation by [ H]-thymidine incorporation. N=; P<.5 vs DMSO; P<.5 vs. VEGF/DMSO by one-way ANOVA with Tukey post hoc test.

CIRCRES//9/R A D Perfusion (ischemic/non-ischemic) (UCPs mrna / GAPDH mrna)...8.6....5..5 UCP Wild-type UCP + UCP + LacZ B Pre 7 8 (Post-Operative Day) UCP LacZ adipose tissue (Sprouts / fat) C + LacZ + UCP + LacZ 6 (days) Online Figure III. Uncoupling protein manipulation and angiogenesis. (A) UCP and mrna levels in mouse lung endothelial cells. Dagger = not detectable. (B) Representative phase contrast images of capillary sprouts from adipose tissue explants taken with a x objective on the day culture. Scale bar, 5 µm. (C) Ex vivo capillary sprouting in fat pad explants from the indicated genotype embedded in collagen gel 5d after in vivo transfection with UCP or LacZ adenovirus. N=5/group; P <.5 vs. with LacZ by two-way ANOVA.. (D) Blood flow recovery in mice with unilateral hindlimb ischemia and treatment with control (LacZ) or UCP adenovirus. N=5; P <.5 vs. LacZ by two-way ANOVA. 5