Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Hypoxia-inducible Factor Prolyl 4-Hydroxylase Inhibition: A TARGET FOR NEUROPROTECTION IN THE CENTRAL NERVOUS SYSTEM *

Hypoxia-inducible Factor Prolyl 4-Hydroxylase Inhibition: A TARGET FOR NEUROPROTECTION IN THE... THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 280, NO. 50, pp. 41732–41743, December 16, 2005 © 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A. Hypoxia-inducible Factor Prolyl 4-Hydroxylase Inhibition A TARGET FOR NEUROPROTECTION IN THE CENTRAL NERVOUS SYSTEM Received for publication, May 5, 2005, and in revised form, October 13, 2005 Published, JBC Papers in Press, October 13, 2005, DOI 10.1074/jbc.M504963200 ‡§¶ ‡ §¶ ‡ ‡ Ambreena Siddiq , Issam A. Ayoub , Juan C. Chavez , Leila Aminova , Sapan Shah , Joseph C. LaManna , ‡‡ ‡‡ ‡‡§§ Stephanie M. Patton**, James R. Connor**, Robert A. Cherny , Irene Volitakis , Ashley I. Bush , ¶¶ ¶¶ ¶¶ ‡§¶1 Ingrid Langsetmo , Todd Seeley , Volkmar Gunzler , and Rajiv R. Ratan From the Department of Neurology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts § ¶ 02115, Burke/Cornell Medical Research Institute, White Plains, New York 10605, the Department of Neurology and Neuroscience, Weill Medical College of Cornell University, New York, New York 10021, the Department of Anatomy and Neurology, Case Western Reserve University, Cleveland, Ohio 44106, the **Department of Neurosurgery, George M. Leader Family Laboratory for Alzheimer ‡‡ Disease Research, Penn State College of Medicine, Hershey, Pennsylvania 17033, the Department of Pathology, the University of §§ Melbourne, Mental Health Research Institute of Victoria, Parkville 3052, Australia, Laboratory of Oxidation Biology, Genetics and Aging Research Unit, Department of Psychiatry, Harvard Medical School, Massachusetts General Hospital, Charlestown, ¶¶ Massachusetts 02129, and Fibrogen, Inc., South San Francisco, California 94080 Hypoxia-inducible factor (HIF) prolyl 4-hydroxylases are a family with Parkinson disease (6–9). Alzheimer disease has also been found to of iron- and 2-oxoglutarate-dependent dioxygenases that nega- be associated with an increase in the iron content of senile plaques tively regulate the stability of several proteins that have established (10–15). Accumulation of mitochondrial iron has been shown to play a roles in adaptation to hypoxic or oxidative stress. These proteins role in Friedrich ataxia (16, 17). Similarly, changes in intracellular free include the transcriptional activators HIF-1 and HIF-2. The abil- iron levels have been observed in cerebral ischemia (18–20). Direct ity of the inhibitors of HIF prolyl 4-hydroxylases to stabilize pro- evidence that disrupted iron homeostasis contributes to injury rather teins involved in adaptation in neurons and to prevent neuronal than simply being caused by it has been obtained by treatment with low injury remains unclear. We reported that structurally diverse low molecular weight iron chelators or by overexpression of iron storage molecular weight or peptide inhibitors of the HIF prolyl 4-hydroxy- proteins. Small molecule iron chelators such as deferoxamine mesylate lases stabilize HIF-1 and up-regulate HIF-dependent target genes (DFO) inhibit neuronal injury in rodent models of stroke (21), Parkin- waf1/cip1 (e.g. enolase, p21 , vascular endothelial growth factor, or son disease (22), and multiple sclerosis (23). Moreover, DFO and some erythropoietin) in embryonic cortical neurons in vitro or in adult rat other metal chelators such as clioquinol have been shown to slow the brains in vivo. We also showed that structurally diverse HIF prolyl progression of Alzheimer disease in humans (24, 25). Similarly, the 4-hydroxylase inhibitors prevent oxidative death in vitro and ische- forced expression of the iron binding and storage protein ferritin in the mic injury in vivo. Taken together these findings identified low substantia nigra diminishes iron accumulation and prevents neuronal molecular weight and peptide HIF prolyl 4-hydroxylase inhibitors loss in a rodent model of Parkinson disease (26). Together, these find- as novel neurological therapeutics for stroke as well as other dis- ings suggest that iron-dependent toxicity is part of the affector pathway eases associated with oxidative stress. of injury in a host of neurological conditions. How do iron chelators prevent neuronal injury? Iron is generally believed to participate in neuronal dysfunction and death through its ability to catalyze (via electron donation) the generation of highly reac- Iron maintains a unique role in physiology via its ability to change tive hydroxyl radicals via Fenton chemistry (1). In this model hydroxyl readily its oxidation state in response to changes in its local environ- ment. A general simplification of its primary function is that it mediates radicals modify lipid, protein, and DNA targets to induce cell dysfunc- one-electron redox reactions. This chemical property of iron enables it tion in these cellular constituents (27). It has been proposed that iron chelators can inhibit Fenton chemistry and prevent neuronal cell injury to act as an essential component in several biological activities, includ- by sequestering accumulated free iron; however, this pathway of pro- ing as a cofactor for enzymes such as tyrosine hydroxylase. Oxygen binding to biomolecules such as hemoglobin and myoglobin is also tection remains controversial. Indeed, prior studies from our laboratory coordinated by iron. Indeed iron deficiency can lead to a host of disor- have suggested alternative pathways of protection by iron chelators. In an in vitro model of oxidative stress, we correlated the protective ders, including anemia and restless legs syndrome (1). effects of iron chelators with their ability to activate the transcriptional Paradoxically, the biochemical properties that make iron beneficial in many biological processes appear to be a drawback when the balance activator hypoxia-inducible factor-1 (28). In this scheme, iron chelators between its accumulation/sequestration within cellular compartments inhibit the activity of a class of iron-dependent enzymes called the HIF prolyl 4-hydroxylases. Inhibition of HIF prolyl 4-hydroxylases prevents and its release is disturbed in favor of iron accumulation (2). Indeed, iron overload is associated with several neurological conditions (3–5). For example, the iron content of nigral Lewy bodies is elevated in patients The abbreviations used are: DFO, deferoxamine mesylate; 3,4-DHB 3,4-dihydroxyben- zoate; hypoxia-inducible factor; VEGF, vascular endothelial growth factor; HCA, * This work was supported by National Institutes of Health Grants NS39170, NS40591, homocysteate; DIV, day in vitro; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetra- and NS46239 (to R. R. R.). The costs of publication of this article were defrayed in part zolium bromide; AM, acetoxymethyl ester; PBS, phosphate-buffered saline; RT, by the payment of page charges. This article must therefore be hereby marked reverse transcription; MCAO, middle cerebral artery occlusion; TTC, 2,3,5-tri- “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. phenyltetrazolium chloride; DAPI, 4,6-diamidino-2-phenylindole; ANOVA, analysis of To whom correspondence should be addressed: Burke/Cornell Medical Research Insti- variance; FITC, fluorescein-isothiocyanate; HRE, hypoxia-response element; ICP-MS, tute, 785 Mamaroneck Ave., White Plains, NY 10605. Tel.: 914-597-2851; Fax: 914-597- inductively coupled plasma-mass spectrometry; HIV, human immunodeficiency 2225; E-mail: [email protected] or [email protected]. virus. 41732 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 This is an Open Access article under the CC BY license. HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury the attachment of an OH group (hydroxylation) to phylogenetically and 100 M cystine, containing the glutamate analog homocysteate conserved proline residues at amino acid 402 and 564 in the protein (HCA; 5 mM). HCA was diluted from 100-fold concentrated solutions HIF-1 (29). Unhydroxylated HIF-1 does not bind the ubiquitin-pro- that were adjusted to pH 7.5. Viability was assessed by calcein-ace- tein isopeptide ligase, von Hippel-Lindau protein (30–32). HIF-1 is toxymethyl ester (AM)/ethidium homodimer-1 staining (live/dead thus not ubiquitinated and degraded by the proteasome. Once stabi- assay) (Molecular Probes, Eugene, OR) under fluorescence microscopy lized, nuclear HIF-1 can heterodimerize with its partner HIF-1, bind and the MTT assay (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazo- to the pentanucleotide hypoxia-response element (5-TGATC-3)in lium bromide) method (36). To evaluate the effects of the drugs on gene regulatory regions, and transactivate the expression of established cytotoxicity, we added DFO, 3,4-DHB, compound A, and the HIF pep- protective genes, including vascular endothelial growth factor (VEGF) tide simultaneously with HCA. and erythropoietin (33, 34). According to this model, the protective Immunofluoroscence Staining—Indirect labeling methods were used effects of iron chelators are not exclusively the result of suppression of to determine the levels of HIF-1 protein in cortical neuron cultures. Fenton chemistry, but may also result from the inhibition of iron-de- Dissociated cells from cerebral cortex were seeded onto poly-D-lysine- pendent prolyl hydroxylases. However, direct evidence that HIF-prolyl coated 8-well culture slides (BD Biosciences) and treated with Tat-HIF/ hydroxylase inhibition is a relevant target for protection from oxidative ODD/mut, Tat-HIF/ODD/wt, and DFO overnight. Cells were washed death by iron chelators has yet to be presented. In this paper we show with warm PBS and fixed at room temperature for 15 min with 4% that small molecule or peptide inhibitors of the HIF prolyl 4-hydroxyl- paraformaldehyde. After washing, cells were incubated with blocking ases are sufficient to inhibit neuronal death because of oxidative stress in solution containing 0.3% Triton X-100 and 5% goat serum in PBS for 1 h, vitro and because of permanent focal ischemia in vivo. These results followed by incubation with rabbit HIF-1 antibody (Upstate Cell Signal- implicate HIF prolyl 4-hydroxylases as targets for neuroprotection in ing, Lake Placid, NY) (1:100 dilution) overnight. After three washes with the central nervous system. PBS, cells were incubated with rhodamine-conjugated goat anti-rabbit IgG antibody (Molecular Probes, Eugene, OR) (1:200 dilution) and EXPERIMENTAL PROCEDURES DAPI. The slides were washed three times with PBS and mounted with fluorochrome mounting solution (Vector Laboratories). Images were Compound A, a proprietary compound, was obtained from Fibrogen analyzed under a confocal fluorescence microscope (Zeiss LSM50, Ger- Inc. DFO and 3,4-DHB were purchased from Sigma. Tat-HIF/ODD/wt many). Control experiments were performed in the absence of primary (YGKKRRQRRRDLDLEMLAPYIPMDDDFQL) and Tat-HIF/ODD/ antibody. mut (YGKKRRQRRRDLDLEMLAAYIAMDDDFQL) peptides were HIF Prolyl 4-Hydroxylase Activity Assay—The effects of Tat-HIF/ synthesized by Biopolymers Laboratory, Harvard Medical School. 40, ODD/mut, Tat-HIF/ODD/wt on purified HIF-prolyl 4-hydroxylase 100, and 10 M of compound A, DFO, and 3,4-DHB, respectively, were activity was analyzed by the method based on the hydroxylation-cou- used throughout the studies unless otherwise stated. Peptides were used pled decarboxylation of 2-oxo[1- C]glutarate (37). at a concentration of 100 M unless otherwise stated. ICP-MS Analysis—Cortical neurons were plated and treated with Primary Neurons—Cell cultures were obtained from the cerebral 3,4-DHB, DFO, or compound A. Cells were collected, lyophilized, and cortex of fetal Sprague-Dawley rats (embryonic day 17) as described stored at 80 °C until further testing. 50 l of concentrated nitric acid previously (35). All experiments were initiated 24 h after plating. Under (HNO ) (Aristar, BDH) was added to each lyophilized sample. Samples these conditions, the cells are not susceptible to glutamate-mediated 3 were allowed to digest overnight at room temperature followed by heat- excitotoxicity. ing for 20 min at 90 °C using a heating block. Samples were diluted 1:25 Immunoblot Analysis—Cell lysates were obtained by rinsing cortical with 1% HNO prior to analysis. Analysis was carried out on a Varian neurons with cold PBS and adding 0.1 M potassium phosphate contain- 3 UltraMass ICP-MS instrument using operating conditions suitable for ing 0.5% Triton X-100. Samples were then boiled in Laemmli buffer and routine multielement analysis. The instrument was calibrated using 0, electrophoresed under reducing conditions on 12% polyacrylamide gels. 10, 50, and 100 ppb of a certified multielement ICP-MS standard solu- Proteins were then transferred to a polyvinylidene difluoride membrane tion (ICP-MS-CAl2-1, AccuStandard). A certified internal standard (Bio-Rad). Nonspecific binding was inhibited by incubation in Tris- solution containing 100 ppb of yttrium ( Y) as an internal control (ICP- buffered saline with Tween 20 (TBST: 50 mM Tris-HCl, pH 8.0, 0.9% MS-IS-MIX1-1, AccuStandard) was also used. NaCl, and 0.1% Tween 20) containing 5% nonfat dried milk for 1.5 h. Calcein-AM Assay—Primary neurons were treated with the drugs Primary antibodies against aldolase (rabbit muscle; Rockland), -tubu- 3,4-DHB, DFO, and compound A overnight as described previously. lin (Sigma), HIF-1 (Novus Biologicals, Littleton, CO), VEGF (Santa Cruz Biotechnology), and p21 (BD Biosciences) were diluted 1:2000, Fluorescence was measured from intact cells using an excitation wave- length of 485 nm and an emission wavelength of 535 nm (38) after 1:2000, 1:500, 1:400 and 1:500, respectively, in TBST containing 1% milk addition of 0.25 M calcein-AM (Molecular Probes, Eugene, OR) was overnight at 4 °C followed by incubation with respective horseradish peroxidase-conjugated secondary antibodies for2hat room tempera- added to cultured neuronal cultures for 30 min at room temperature. IRP/IRE Binding Assay—Cells were washed with PBS, trypsinized, ture. Immunoreactive proteins were detected according to the and collected. To separate membrane-bound organelles from cytosol, enhanced chemiluminescent protocol (Amersham Biosciences). Luciferase Assay—Primary neurons infected with adenovirus (multi- the cells were washed and homogenized in lysis buffer (10 mM/liter HEPES, pH 7.4, 40 mmol/liter KCl, 5% glycerol, 3 mmol/liter MgCl , 0.1 plicity of infection  100) containing luciferase gene driven by an HRE mmol/liter EDTA, 1 mmol/liter dithiothreitol, 1 mmol/liter phenyl- promoter (a kind gift from Dr. Robert Freeman) were treated with DFO, 3,4-DHB, compound A, and the HIF-peptides overnight. The luciferase methylsulfonyl fluoride, and 0.1 g/liter leupeptin). Cells were sonicated assay was performed using the bioluminescent method (Promega to disrupt membranes. The homogenates were centrifuged for 5 min at Corp.). 4 °C and 4000  g to sediment nuclei. The resulting supernatant was Viability Assays—For cytotoxicity studies, cells were rinsed with centrifuged at 4 °C and 145,000  g for 1 h, and the cytosolic fraction warm PBS and then placed in minimum essential medium (Invitrogen) was collected. For preparation of RNA transcripts, the pGEM7Zf()- containing 5.5 g/liter glucose, 10% fetal calf serum, 2 mML-glutamine, 5L containing the complete 5-untranslated region of rat liver L-ferritin DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41733 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury cDNA (39) was provided by R. S. Eisenstein (University of Wisconsin, tion. This method placed the tip of the suture at the origin of the ante- Madison). The ferritin IRE RNA transcripts were prepared from maI- rior cerebral artery, thereby occluding the MCA. The placement of the suture tip was monitored by laser Doppler flowmetry measurements of digested pGEM7Zf()-5L containing 100 bases of 5-untranslated region and 40 bases from the vector. RNA transcripts were synthesized regional cerebral blood flow. MCAO caused a sharp drop in regional from 1 g of linearized plasmid DNA in the presence of 100 Ci of cerebral blood flow to less than 30% of preischemic base line. The suture was left in place, and the animals were allowed to awaken from the [- P]CTP (800 Ci/mmol), 12 M unlabeled CTP, 0.5 mM of the three unlabeled nucleotides, 100 mM dithiothreitol, 20 units of T7 RNA anesthesia following closure of the operation sites. polymerase in a 20-l reaction volume. The reaction was carried out for Drug Administration—3,4-DHB was dissolved in ethanol (vehicle) for a total volume of 100 l and was administered intraperitoneally to 1 h at 37 °C. The yield of the transcript synthesized was determined by trichloroacetic acid precipitation. Under these conditions, the specific animals (n  6) at a dose of 180 mg/kg of body weight. 3,4-DHB was administered 6 h before MCAO. Compound A was administered to activity was5 10 cpm/g RNA. For the RNA band shift assay, 5 g animals (n 9) at a dose of 100 mg/kg of body weight using a gavage 6 h of protein for the cytoplasmic fractions were incubated with the syn- thetic radiolabeled IRE probe. The RNA-protein complexes were sepa- before the MCAO. The control animals received an equivalent volume of the vehicle (ethanol for 3,4-DHB and carboxymethylcellulose for rated on a 6% native polyacrylamide gel and visualized by autoradiogra- compound A) on a similar administration schedule. phy. The autoradiograms were scanned, and the resulting digital image Infarct Measurement—24 h after MCAO, the animal were anesthe- was subjected to densitometric band analysis using the Labworks Anal- tized with ketamine (100 mg/kg, intraperitoneally) and xylazine (50 ysis software (UVP Inc., Upland, CA). The results of the RNA band shift mg/kg, intraperitoneally) and decapitated. The brain was rapidly assays were expressed as active IRP binding from each treatment as a removed, sliced into seven 2-mm coronal sections using a rat matrix ratio of control. (RBM 4000C, ASI Instrument Inc., Warren, MI), and stained according RT-PCR—The levels of p21, enolase, VEGF, and -actin were ana- to the standard 2,3,5-triphenyltetrazolium chloride (TTC) method (40, lyzed by semiquantitative RT-PCR by using the one-step RT-PCR kit 41). Each slice was drawn using a computerized image analyzer (Scion ReddyMix version (Abgene). The following primers were used in this Corp.). The calculated infarction areas were then compiled to obtain the study: enolase 5-GGTTCTCATGCTGGCAACAAGT-3 (sense) and TM infarct volumes per brain (in mm ). Infarct volumes were expressed as a 5-TAAACCTCTGCTCCAATGCGC-3 (antisense) (GenBank percentage of the contralateral hemisphere volume to compensate for accession number M012554); VEGF 5-TGCAATGATGAAGCCCT- edema formation in the ipsilateral hemisphere (42, 43). GGA-3 (sense) and 5-TGCTATGCTGCAGGAAGCTCA-3 (anti- TM sense) (GenBank accession number M32167); p21 5-ATGTCCGA- RESULTS TCCTGGTGATGT-3 (sense) and 5-ACTTCAGGGCTTTCTCTT- TM GC-3 (antisense) (GenBank accession number U24174); -actin 5- Low Molecular Weight HIF Prolyl 4-Hydroxylase Inhibitors Stabilize CCTCATGAAGATCCTGACCG-3 (sense) and 5-TGCCAATAGT- HIF and Activate HIF-dependent Gene Expression in Embryonic Rat TM GATGACCTGG-3 (antisense) (GenBank accession number Cortical Neurons—Oxidative glutamate toxicity in immature cortical NM007393). neurons results from the ability of glutamate to inhibit competitively the In Vivo Experiments—All experimental procedures were approved by uptake of cystine via its plasma membrane transporter (system X ) the Harvard Medical Area Standing Committee on Animals and meet (44). Inhibition of cystine uptake into the cell via system X  results in the standards of the Federal and State reviewing organizations. depletion of intracellular cysteine, the rate-limiting precursor of gluta- Animal Preparation and Monitoring—Eleven adult male Sprague- thione synthesis. Cyst(e)ine deprivation thus leads to depletion of the Dawley rats (n  11) weighing 250–280 g (Charles River Breeding Lab- major cellular antioxidant glutathione, an imbalance in cellular oxidants oratories, Wilmington, MA) were operated on. Nine rats survived and and antioxidants and oxidative stress-induced death. Indeed, a host of were analyzed for this study. Mortality was 18% (2:11) and occurred agents with known antioxidant capacity can completely abrogate gluta- equally in the groups of animals (one rat in each group). Death occurred mate toxicity in this paradigm. To determine whether prolyl 4-hydrox- during the first 20 h postoperatively. Animals were allowed free access ylase inhibition is sufficient to prevent oxidative glutamate toxicity, we to food and water before and after surgery. Briefly, rats were anesthe- treated immature cortical neurons with established, structurally tized by an intraperitoneal injection of 400 mg/kg chloral hydrate fol- diverse, low molecular weight inhibitors of the prolyl 4-hydroxylases, lowed 45 min later by a maintenance intraperitoneal infusion at a rate of DFO, 3,4-DHB, and compound A. DFO, 3,4-DHB, and compound A 120 mg/kg/h using a butterfly needle set. The animals were free breath- were selected for testing because their mechanisms of prolyl hydroxy- ing. Their body temperatures were kept stable at 36.5  0.5 °C using a lase inhibition are believed to be distinct. 3,4-DHB has been shown feedback-regulating heating pad and a rectal probe (Harvard Apparatus, previously to be an inhibitor of collagen prolyl 4-hydroxylases with MA). The right femoral artery was cannulated for measurement of arte- respect to the 2-oxoglutarate and ascorbate cofactors and to be neutral rial blood gases, glucose, and mean arterial blood pressure. These phys- with respect to Fe (45). By contrast, the ability of DFO to inhibit iological parameters were monitored before and after middle cerebral recombinant prolyl 4-hydroxylase activity can be overcome by addition artery occlusion (MCAO). In addition, laser Doppler flowmetry (Moor of exogenous iron to the test tube (45). Compound A has been reported Instruments, Devon, UK) was used to monitor the regional cerebral to be an inhibitor of HIF prolyl 4-hydroxylases (30). Immunoblotting blood flow through a burr hole 2 mm in diameter created in the right using a specific antibody to HIF-1 established that, like hypoxia, all of parietal bone (2 mm posterior and 6 mm lateral to bregma). these agents (100 M DFO, 10 M dihydroxybenzoic acid, 40 M com- Surgery—All rats were subjected to right MCAO. Under the operat- pound A) stabilize HIF-1 but do so in normoxic neurons (Fig. 1A). ing microscope, the right common carotid artery was exposed through Stabilization of HIF-1 by each of these low molecular weight com- a midline incision in the neck. A 4-0 nylon suture with its tip rounded by pounds was associated with significant transactivation of a reporter heating over a flame and subsequently coated with poly-L-lysine (Sigma) gene driven by hypoxia-response elements (boldface in the sequence) was introduced into the external carotid artery and then advanced into from the promoter region of the enolase gene (5-ACGCTGAGT- the internal carotid artery for a length of 18–19 mm from the bifurca- GCGTGCGGGACTCGGAGTACGTGACGGA-3) (Fig. 1B; p  0.05). 41734 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury A control reporter gene without these response elements or one with the hypoxia-response elements mutated were not up-regulated by these compounds (not shown). To verify that increased reporter activity was reflected in increased protein expression, immunoblotting was performed to monitor levels of waf1/cip1 aldolase, VEGF, and p21 , known HIF-1-regulated genes (33). -Tubulin was used as a control because its transcriptional regulation appears to be HIF-independent. As expected, DFO, 3,4-DHB, and com- waf1/cip1 pound A all increased protein levels of aldolase, VEGF, and p21 but not -tubulin (Fig. 1C). Taken together these results established that like non-neuronal cells, DFO, 3,4-DHB, and compound A can all inhibit HIF prolyl 4-hydroxlyases, stabilize HIF, and lead to the increased transcription of established HIF-dependent genes. Structurally Diverse Inhibitors of the HIF Prolyl 4-Hydroxylases Can Abrogate Oxidative Stress-induced Death in Cortical Neurons—To determine whether structurally diverse inhibitors of the HIF prolyl 4-hydroxylases can abrogate death because of oxidative glutamate tox- icity at concentrations at which they activate HIF, we added these com- pounds at the time of exposure to glutamate (not shown) or the gluta- mate analog homocysteate (HCA, 5 mM). As expected, all three agents significantly prevented oxidative stress-induced death, suggesting that HIF prolyl 4-hydroxylases are a target for protection from oxidative death in neurons (Fig. 2, A and B). The number of live and dead cells in the dish were monitored using several methods, including MTT reduc- tion (Fig. 2A), visualization by phase contrast microscopy (not shown), and calcein labeling or ethidium homodimer labeling to tag live cells or dead cells, respectively (Fig. 2, B–I). In all cases, the results obtained were similar. The absolute level of protection from oxidative death by the three compounds did not seem to correlate with their transcriptional efficacy. For example, compound A is the most effective in stimulating HRE- driven luciferase activity and VEGF protein levels; and despite signifi- cantly preventing oxidative death, it also induces a small but reproduc- ible basal toxicity in the cultures (Fig. 2H). These findings suggest that although HIF-dependent transcription is a good marker of HIF prolyl 4-hydroxylase inhibition, other factors besides HIF transcriptional activity may contribute to neuroprotection by agents that inhibit HIF prolyl 4-hydroxylases. Of course, one cannot rule out other prolyl 4-hy- droxylase-independent effects of the drugs on viability as an explanation for the quantitative differences in neuroprotection from the oxidative death we observed between DFO, DHB, and compound A Some but Not All Prolyl 4-Hydroxylase Inhibitors Tested Reduce Total Cellular Iron in Embryonic Rat Cortical Neurons—HIF prolyl 4-hy- droxylases are iron-dependent enzymes. To verify that at least one of the PH inhibitors tested does not alter iron levels in cortical neurons, we FIGURE 1. Structurally diverse, low molecular weight prolyl 4-hydroxylase inhibi- tors stabilize HIF-1 protein, enhance activity of a hypoxia-response element- monitored total iron levels by inductively coupled plasma mass spec- driven luciferase reporter, and increase the expression of established HIF-regu- trometry. This technique allowed us to perform sensitive analysis of lated genes in embryonic cortical neurons. A, cortical neuronal cultures (1 day in vitro chelatable and nonchelatable metals. Total cellular iron was decreased (DIV)) were treated with a vehicle control (C,Me SO (0.01%)), DFO (100 M), 3,4-DHB (10 M), or compound A (Comp-A;40 M) or exposed to 1% hypoxia (H) overnight. Proteins by overnight incubation with DFO (100 M) and compound A (40 M) from cell lysates were separated using gel electrophoresis and immunodetected using but not 3,4-DHB (10 M) treatment in cultured neurons (TABLE ONE). an antibody against HIF1-. -Actin was monitored as a loading control. B, cortical neu- ronal cultures (1 DIV) were infected with an adenovirus harboring the hypoxia-response This does indicate some ability to extract iron or prevent its uptake, in elements from the enolase promoter linked to a luciferase reporter (multiplicity of infec- this case, from cells and perhaps suggests the possibility of direct iron tion  100). Parallel infection with adenovirus containing the green fluorescent protein chelation. These findings are in agreement with prior in vitro studies (GFP) indicated reproducible infection efficiencies of30 – 40%. 24 h following infection with adeno-HRE-luciferase, cortical neuronal cultures were treated with a vehicle control demonstrating that DFO inhibits prolyl 4-hydroxylase activity by bind- (C), DFO (100 M), 3,4-DHB (10 M), or compound A (40 M). Lysates from each treatment ing its critical iron cofactor (46), whereas 3,4-DHB inhibits prolyl 4-hy- group were generated, and luciferase activity was measured. Graph depicts mean per- centage luciferase activity (compared with control) S.E. calculated from three separate droxylases by displacing 2-oxoglutarate or ascorbate and not iron (45). experiments for each group (* denotes p 0.05 by ANOVA and Student-Newman-Keuls The effect of DFO and compound A showed selectivity for the trace tests for DFO, 3,4-DHB, or compound A). C, DFO (100 M), 3,4-DHB (10 M), and com- pound A (40 M) increase expression of established HIF target genes, aldolase, VEGF, or metal iron as neither these agents nor 3,4-DHB had any effect on total waf1/cip1 p21 in cultured cortical neurons as compared with vehicle-treated control. zinc in rat neuronal cultures (TABLE ONE). Taken together, these Immunoblots are representative examples of three experiments. results are consistent with the notion that DFO and compound A act to DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41735 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury FIGURE 2. Structurally diverse, low molecular weight prolyl 4-hydrdoxylase inhibitors abro- gate oxidative glutamate toxicity in embryonic cortical neuronal cultures. A, the glutamate ana- log HCA (5 mM) was added to cortical neuronal cul- tures (1 DIV). In parallel, DFO (100 M), 3,4-DHB (10 M), and compound A (Comp-A) (40 M) were added with and without HCA. Twenty four hours later, cell viability was determined using the MTT assay. Graph depicts mean  S.E. for three experi- ments performed in triplicate (* denotes p  0.05 from HCA-treated cultures by ANOVA and Stu- dent-Newman Keuls tests for control, DFO, 3,4-DHB, or compound A). B–I, live/dead assay of cortical neuronal cultures (2 DIV). Live cells are detected by uptake and trapping of calcein-AM (green fluorescence). Dead cells fail to trap calcein but are freely permeable to the highly charged DNA intercalating dye, ethidium homodimer (red fluorescence). B, control. C, HCA (5 mM). D, DFO (100 M). E, DFO  HCA. F, 3,4-DHB (10 M). G, 3,4- DHB HCA. H, compound A (40 M). I, compound A  HCA. 41736 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury inhibit the prolyl hydroxylases by removing the Fe cofactor or binding inhibit collagen prolyl hydroxylases, HIF prolyl 4-hydroxylases, and to it within the enzyme, whereas 3,4-DHB inhibits prolyl 4-hydroxylase other 2-oxoglutarate-dependent dioxygenases in the intact cell. In order activity by iron-independent mechanisms. to determine whether HIF prolyl 4-hydroxylases are relevant for pro- A Peptide Inhibitor of the HIF Prolyl 4-Hydroxylases Delivered into tection from glutathione depletion-induced death by DFO, 3,4-DHB, Neurons Using an HIV Tat Protein Transduction Domain Activates HIF- and compound A, we used the following strategy. Myllyharju and co- dependent Gene Expression and Prevents Oxidative Death in Vitro— workers (37) recently described the identification of a 19-amino acid Low molecular weight inhibitors of the HIF prolyl 4-hydroxylases were peptide that is able to act as a substrate for all the human HIF prolyl initially identified via a search for inhibitors of the collagen prolyl 4-hy- 4-hydroxylases(K 5–10M).Thispeptidecontainspartoftheoxygen- droxylases. The collagen prolyl hydroxylases catalyze the formation of dependent domain including the C-terminal hydroxylation site of 4-hydroxyproline in collagen and certain other proteins that possess human HIF at proline 564 (DDLDEMLAPYIPMDDDFQL; boldface P collagen-like sequences by hydroxylating prolines in X-Pro-Gly motifs indicates proline 564). We hypothesized that such a peptide delivered (47). Hydroxyprolines facilitate the formation of stable triple helical into neurons could act as a competitive inhibitor of HIF prolyl 4-hy- collagen. Unlike HIF prolyl 4-hydroxylases, which are believed to be droxylases while leaving other 2-oxoglutarate-dependent dioxygenases cytoplasmic or nuclear (48), the collagen prolyl hydroxylase is a soluble unperturbed. Indeed, previous studies have shown that this peptide is endoplasmic reticulum luminal protein (47). However, like HIF prolyl not a substrate for recombinant collagen prolyl hydroxylases (30). Phy- 4-hydroxylases, collagen hydroxylases employ ferrous ions, 2-oxoglut- logenetic conservation (from worms to humans) of the critical residues arate, molecular oxygen, and ascorbate as cofactors. Thus by disrupting from HIF prolyl 4-hydroxylases that serve as ligands for Fe and 2-oxo- one or more of these cofactors, DFO, 3,4-DHB, and compound A could glutarate suggested that a peptide deemed to be a potent and effective substrate of all three known human HIF prolyl 4-hydroxylases would TABLE ONE also act to competitively inhibit HIF prolyl 4-hydroxylases expressed in Total zinc and iron levels in cultured cortical neurons nontreated rat cortical neurons (46). Our HIF/ODD/wt peptide and a peptide con- (control) or treated with DFO (100 M) 3,4-DHB (10 M), or compound trol with mutation of both prolines to alanine (HIF/ODD/mut; A (40 M) measured by ICP-MS DDLEMLAAYIAMDDDFQL) were rendered cell-permeant by fusing Values are shown as mean  S.E. percent relative to control levels that were arbitrarily designated as 100%. Levels of individual metals in a particular cul- each of these peptides to the cell membrane transduction domain of the ture were normalized to protein. human immunodeficiency virus, type 1 (HIV-1), Tat protein Control DFO 3,4-DHB Compound A (YGRKKRRQRR) to obtain two 30-amino acid peptides Tat-HIF/ ODD/wt and Tat-HIF/ODD/mut (49). To verify that the addition of Tat Zinc (%) 100 93.6  2.8 120.3  2.9 110.7  2.2 a a Iron (%) 100 68.8  9.5 93.3  3.1 70.04  2.4 p  0.05. 3 A. Siddiq and R. Ratan, unpublished observations. FIGURE 3. A cell-permeant peptide inhibitor of the HIF prolyl 4-hydroxylases, but not a mutant control, is a substrate for purified, recombinant HIF prolyl 4-hydroxylases 1, 2, and 3 isoforms. A, Tat-HIF/wt peptide (100 M, composed of the oxygen-dependent domain, including the C-ter- minal hydroxylation site of human HIF-1 at pro- line 564 (DDLDEMLAPYIPMDDDFQL; boldface P indicates proline 564) linked to a Tat protein trans- duction domain YGKKRRQRRR) but not a corre- sponding peptide with the C-terminal proline hydroxylation site of HIF-1-mutated (Tat-HIF/ mut, DLDLEMLAAYIAMDDDFQL; boldface A indi- cates the mutation site) acts as a substrate for the HIF prolyl 4-hydroxylases and releases [ C]CO from -ketoglutarate. DLD 19 HIF peptide, com- posed of the C-terminal hydroxylation site of human HIF-1 at proline 564 without the TAT sequence was used as a positive control. B, repre- sentative low magnification image (40) of corti- cal neurons treated with a FITC-labeled Tat-HIF/wt peptide (100 M)(green fluorescence) and the DNA intercalating dye, DAPI, to label nuclei (blue fluo- rescence). Note all blue nuclei have green fluores- cence in the cytoplasm indicating a high efficiency of transduction of the peptide into cells. C, a rep- resentative high magnification image (60) of cortical neurons treated with FITC-Tat-HIF/wt pep- tide (100 M) and labeled with the DNA intercalat- ing dye DAPI. DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41737 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury FIGURE 4. The peptide inhibitor of the HIF prolyl 4-hydroxylases, but not a mutant control, sta- bilizes HIF protein in primary cortical neurons. A, Tat-HIF/mut peptide (f and n), Tat-HIF/wt pep- tide (100 M)(g and o), DFO (100 M)(h and p), or vehicle control (e and m) were added to the bath- ing medium of cortical neurons. Panel 1 represents lower magnification (20), and panel 2 represents higher magnification (60). Cells were fixed with 4% paraformaldehyde for 15 min and immuno- stained for HIF-1 (red) as described under “Exper- imental Procedures.” DAPI staining was per- formed to label nuclei (blue)(a– d, 20; and i–l, 60). B, immunoblot analysis with an antibody to HIF-1 in cell lysates from nontreated (lane 1)or after treatment with Tat-HIF/mut peptide (100 M, lane 2), Tat-HIF/wt peptide (20 M, lane 3; 100 M, lane 4), or DFO (100 M, lane 5, positive control). to the 19-amino acid HIF peptide does not alter its specificity for the its cell bodies and processes reflecting the uptake of the Tat-HIF/ recombinant HIF prolyl 4-hydroxylases, we compared activities of the ODD/wt FITC peptide (Fig. 3, B and C). Similar high efficiency trans- Tat-HIF/ODD/wt, Tat-HIF/ODD/mut, and the native HIF/ODD/wt. duction was observed when cortical neuronal cultures were incubated We employed a commonly used assay for the HIF prolyl 4-hydroxylase with the Tat-HIF/ODD/mut FITC peptide (not shown). In parallel, via- that is based on the hydroxylation of a synthetic substrate and the asso- bility measurements verified that the Tat-HIF/ODD peptides do not ciated measurement of the radioactivity of CO formed during the negatively alter basal viability in our cortical neuronal cultures (not hydroxylation coupled decarboxylation of 2-oxo[1- C]glutarate. This shown). assay confirmed that the Tat-HIF/ODD/wt peptide but not the Tat- To determine whether the Tat-HIF/ODD/wt peptide stabilizes HIF/ODD/mut is a substrate with similar activity to the isoforms of HIF HIF-1 in cortical neurons, we performed immunocytochemical stain- prolyl 4-hydroxylases as a standard synthetic 19-amino acid peptide ing with an antibody to HIF-1 (Fig. 4, e–h,20; m–p,60) using in all with a sequence around the C-terminal hydroxylation site of HIF-1 cases a rhodamine-labeled secondary antibody that fluoresces red. In (DLD 19 HIF peptide (HIF/ODD/wt), Fig. 3A). parallel, the same cultures were treated with the nuclear stain DAPI, To evaluate the ability of our Tat-HIF/ODD peptides to be delivered which fluoresces blue (Molecular Probes, Eugene, OR) (Fig. 4, a–d, to cortical neurons efficiently and without toxicity, we conjugated the 20; i–l, 60). The HIF-1 antibody fluorescence confocal images chromophore fluorescein isothiocyanate to the Tat-HIF/ODD/wt pep- (20) confirmed that the Tat-HIF/ODD/wt peptide increases HIF-1 tide. Tat-HIF/ODD/wt FITC peptides (10 M) were bath applied to rat staining in cortical neurons (Fig. 4g), but the Tat-HIF/ODD/mut pep- cortical neuronal cultures from embryonic day 17 rat embryos. These tide does not (Fig. 4f). Addition of vehicle was used as a negative control cultures are 90% neurons after 1 day in vitro (35). The balance of the (Fig. 4e), and DFO was used as a positive control (Fig. 4h). High magni- cells in the culture is glial in origin. Twenty four hours after addition of fication confocal microscopic images of HIF- antibody fluorescence the peptide to the bathing medium, the level of intracellular accumula- from the nuclei of cortical neurons (60) (Fig. 4o) and immunoblot tion of the peptide was monitored by fluorescence microscopy. In every analysis of nuclear extracts using the same HIF-1 antibody (Fig. 4B, field examined ( 10), each cell nucleus, identified by DNA intercalating lanes 3 and 4) verified that the increases in HIF-1 levels induced by chromophore DAPI, was found to be associated with fluorescein label in Tat-HIF/ODD/wt are cytoplasmic and nuclear. 41738 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury To confirm that stabilization of HIF-1 is associated with activation of HIF-dependent gene expression, we monitored the levels of a lucif- erase reporter gene that is regulated by a hypoxia-response element found in the enolase gene promoter. As expected, we found that addi- tion of Tat-HIF/ODD/wt peptide significantly enhanced levels of a hypoxia-response element-regulated reporter gene in a concentration- dependent manner (not shown). The level of induction of the HIF-de- pendent reporter gene by 100 M Tat-HIF/ODD/wt peptide was similar to that stimulated by 40 M compound A, a low molecular weight inhib- itor of the HIF prolyl 4-hydroxylase (Fig. 5A). By contrast, 100 M Tat- HIF/ODD/mut peptide with proline 564 and its adjacent proline mutated to alanine did not induce the HIF-dependent reporter gene. Because the unhydroxylatable peptide did not increase target gene expression, it does not likely serve as a competitive inhibitor of the HIF prolyl 4-hydroxylases. To verify that the changes in the HIF-dependent reporter gene reflect changes in endogenous HIF-dependent genes, we waf1/cip1 monitored enolase, p21 , and VEGF by RT-PCR. -Actin was measured as a control (Fig. 5B). As expected, the Tat-HIF/ODD/wt waf1/cip1 peptide increased levels of enolase, p21 , and VEGF message but did not alter levels of -actin. Accordingly, we found that the Tat-HIF/ waf1/cip1 ODD/mut did not affect levels of p21 or -actin and actually diminished the basal levels of enolase and VEGF, two known HIF-de- pendent genes. Real time PCR was utilized to examine quantitatively the difference in message levels between the HIF/ODD/wt- and HIF/ODD/ mut-treated neurons. These studies confirmed the ability of the wild type but not the mutant peptide to increase HIF-dependent gene expression (VEGF, 1.46  0.218-fold induction (wt/mut); erythropoie- waf1/cip1 tin, 1.84  0.068-fold induction (wt/mut); p21 , 1.52  0.16-fold induction). Taken together, these results establish that the Tat-HIF/ FIGURE 5. A peptide inhibitor of the HIF prolyl 4-hydroxylases, but not a mutant ODD/wt peptide can be delivered to neurons and can enhance the tran- control, induces expression of established HIF-dependent genes. A, Tat-HIF/wt pep- tide (100 M) but not a corresponding peptide with the C-terminal proline hydroxylation scriptional activity of well characterized HIF target genes. By contrast, a site of HIF-1 mutated (Tat-HIF/mut, 100 M) significantly enhances the activity of a Tat/HIF/ODD/mut, in which the conserved proline hydroxylation site hypoxia-response element-driven reporter in cortical neurons (* corresponds to p 0.05 compared with control by paired t test). Note that the Tat-HIF/wt peptide induces HIF (proline 564) and an adjacent proline have been mutated to alanine, reporter activity to a similar extent as the low molecular weight inhibitor, compound A does not up-regulate an HIF-dependent reporter gene or endogenous (Comp-A) (40M). B, RT-PCR using specific primers to determine the mRNA level of estab- waf1/cip1 waf1/cip1 lished HIF target genes, enolase, p21 , or VEGF in vehicle-treated (control) or in HIF-regulated genes (e.g. enolase, p21 , or VEGF). neuronal cultures treated with Tat-HIF/wt peptide (100 M) and Tat-HIF/mut (100 M) To test whether Tat-HIF/ODD/wt, a peptide inhibitor of HIF prolyl overnight. -Actin was monitored as a housekeeping gene. 4-hydroxylases, can prevent oxidative neuronal death in cortical neu- FIGURE 6. A cell-permeant peptide inhibitor of the HIF prolyl 4-hydroxylases, but not a mutant control, prevents oxidative glutamate toxicity. The glutamate analog, HCA (5 mM) was added to cortical neurons (1 DIV) with or without Tat-HIF/wt peptide (30, 40, 50, 100, and 200 M), Tat-HIF/mut peptide (200 M), DFO (100 M), and 3,4-DHB (10 M). Twenty four hours later cell viability was determined using the MTT assay. Graph depicts mean  S.E. for three experiments performed in triplicate (* denotes p  0.05 from HCA-treated cultures by ANOVA and Student-Newman Keuls tests for control, Tat-HIF/wt, Tat-HIF/mut, DFO and 3,4-DHB). DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41739 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury FIGURE 7. Structurally diverse, low molecular weight prolyl 4-hydroxylase inhibitors stabilize HIF protein, increase the expression of established HIF-regulated genes, and reduce infarct volume induced by permanent focal ischemia in vivo. Three rats were treated in parallel with compound A (lanes 1, 3, and 5) with a dose of 100 mg/kg of body weight (by gavage) 6 h prior to MCAO. As controls, three rats (lanes 2, 4, and 6) received equal volume of vehicle (0.5% carboxymethylcellulose) in a blinded study. Whole brain lysates 41740 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury TABLE TWO Physiologic data obtained from both control and treated animal groups are represented as means  S.E. The following abbreviations are used: Hct, hematocrit; Glu, blood glucose; MABP, mean arterial blood pressure; HR, heart rate. Animals were maintained at 36.5  0.5 °C (rectal temperature). There was no statistically significant difference between control and treated groups in both pre- and post-MCAO categories. Pre-ischemia Post-ischemia Physiological parameters Control group Compound A group Control group Compound A group Compound A pH 7.35  0.05 7.36  0.02 7.32  0.03 7.34  0.03 PCO (mm Hg) 33 535 430 232  3 PO (mm Hg) 99  12 100  10 95  10 100  10 Hct (%) 40 440 338 440  4 MABP (mm Hg) 91  12 89  18 95  12 88  23 3,4-DHB pH 7.35  0.05 7.33  0.06 7.35  0.06 7.33  0.06 PCO (mm Hg) 34 234  2 34.2 235  2 PO (mm Hg) 93 395 493 396  3 Hct (%) 42 143 241 243  2 MABP (mm Hg) 93 392 394 493  4 rons induced by the glutamate analog, HCA, we applied the peptides concentrations of 3,4-DHB and compound A, and we determined that 24 h prior to the addition of HCA. Addition of Tat-HIF/ODD/wt pep- administration of 100 mg/kg of compound A (via gavage) or 160 mg/kg tide but not its mutant control (Tat-HIF/ODD/mut) protected cortical of 3,4-DHB (via intraperitoneal administration) was effective at stabi- neurons from oxidative glutamate (glutathione depletion-induced) tox- lizing the HIF protein and inducing an increase in some HIF-dependent icity in a concentration-dependent manner as assessed by MTT assay genes, including erythropoietin in lysates from brains of individual vehi- (Fig. 6). The morphology of the neurons that were protected from oxi- cle-treated or HIF prolyl 4-hydroxylase inhibitor-treated brains (Fig. 7, dative glutamate toxicity by Tat-HIF/ODD/wt was indistinguishable A, B, and E). The stabilization of HIF in brain by compound A was from the morphology of control neurons as assessed by phase contrast observed within 3 h (Fig. 7A, whereas the stabilization of HIF by 3,4- microscopy or calcein/ethidium homodimer staining (not shown). DHB was observed within6hofadministration (Fig. 7E). Inhibition of These results suggest that inhibition of HIF prolyl 4-hydroxylases is HIF prolyl 4-hydroxylases was associated with marked reduction in sufficient to prevent oxidative neuronal death in cortical neurons in infarct volume as measured histologically by TTC staining by com- vitro and argue that 3,4-DHB, DFO, and compound A are influencing pound A (67%, Fig. 7, C and D) or by 3,4-DHB (46%, Fig. 7F). Protection cell viability by inhibiting HIF prolyl 4-hydroxylases rather than a dis- by these agents could not be attributed to the effects of these agents on tinct 2-oxoglutarate-dependent dioxygenase. body temperature, plasma glucose levels, plasma pH, or blood flow Low Molecular Weight Inhibitors of HIF Prolyl 4-Hydroxylases (TABLE TWO, compound A and 3,4-DHB). Together, these findings Reduce Neuronal Damage Induced by Permanent Focal Ischemia suggest that HIF prolyl 4-hydroxlyases are a target for protection in in Vivo—Oxidative stress is an established mediator of neuronal loss in stroke and suggest that the salutary effects of these treatments in cere- cerebral ischemia (50). Prior studies have also demonstrated that the bral ischemia are due in part to HIF prolyl 4-hydroxylase inhibition. putative antioxidant DFO can ameliorate metabolic failure in dogs sub- DISCUSSION jected to cerebral ischemia or induce ischemic tolerance in neonatal or adult rodents (51–53). Some of these studies have suggested that the Prior studies from our laboratory correlated the protective effects of effects of DFO may be related to HIF prolyl 4-hydroxylase inhibition. iron chelators in an in vitro model of oxidative stress with their ability to However, these studies did not provide any experimental support for stabilize HIF-1 and activate HIF-dependent genes (28). These findings the notion that HIF prolyl 4-hydroxylases are a target for ischemic neu- supported the hypothesis that iron chelators not only prevent neuronal roprotection by DFO. The protective effects of HIF prolyl 4-hydroxylase injury by inhibiting hydroxyl radical formation via Fenton chemistry but inhibitors in preventing oxidative neuronal death in vitro herein (Figs. also via inhibition of the iron-dependent HIF prolyl 4-hydroxylases that 1–6) provided additional rationale for testing these agents in an in vivo regulate HIF stability. In this study we utilize low molecular weight or model of cerebral ischemia. We therefore examined the ability of 3,4- peptide inhibitors of the HIF prolyl 4-hydroxylases in vitro and in vivo to DHB or compound A to prevent neuronal damage in response to per- define this family of enzymes as targets for neuroprotection in the cen- manent focal ischemia. In this model, a filament is placed permanently tral nervous system, and also to suggest that the inhibition of these into the distal aspect of the internal carotid artery to occlude the origin enzymes is one way in which these treatments exert their salutary effects of the middle cerebral artery (43). After 24 h of occlusion, there is sig- against oxidative stress. The HIF prolyl 4-hydroxylases require iron, nificant infarction in the hemisphere ipsilateral to the occlusion; few 2-oxoglutarate, oxygen, and ascorbate as cofactors. Two of the low animals survive beyond this time point. We first examined a range of molecular weight agents we tested, DFO and compound A, appear to from individual animals in the treated (n 9) (lanes 1, 3, and 5) and control group (n 5) (lanes 2, 4, and 6) were separated using gel electrophoresis and immunodetected using an antibody against HIF1- (A) and erythropoietin (Epo)(B). Whole brain lysates of 1% hypoxia-exposed animals were used as a positive control for HIF-1 and erythropoietin (lane 7). -Actin was monitored as a housekeeping gene. C, histologically measured infarct volume of brains of compound A (Comp-A)(n  9) and vehicle-treated animals (n  5) after TTC staining. The values are shown as % infarct volume relative to total brain volume (*  p  0.05 by paired t test). D, representative pictures of vehicle-treated (left) and compound A-treated (right) brains showing infarctions. E, animals were treated with 3,4-DHB (n 6) (lanes 2 and 3) with a dose of 180 mg/kg of body weight (intraperitoneally) 6 h prior to MCAO. Control animals (n 5) (lanes 1 and 4) received equal volume of vehicle (ethanol, intraperitoneally) in a blinded study. Whole brain lysates were separated using gel electrophoresis and immunodetected using an antibody against HIF1-. -Actin was monitored as a housekeeping gene. F, histologically measured infarct volume of brains of 3,4-DHB (n  6) and vehicle-treated animals (n  5) after TTC staining. The values are shown as % infarct volume relative to total brain volume (  p  0.05 by paired t test). DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41741 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury inhibit prolyl 4-hydroxylases by removing the iron cofactor. Indeed, rat HIF prolyl 4-hydroxylase. Future studies will determine which one both agents reduced total iron in embryonic rat cortical neurons by 30% or more of the described rat HIF prolyl 4-hydroxylases needs to be inhibited in order to protect rat embryonic cortical neurons from oxi- after 24 h (TABLE ONE). Despite these measurable changes in total iron, we could not detect statistically significant changes in IRP-IRE dative stress. binding or calcein fluorescence (independent measures of the labile iron What is the mechanism of neuroprotection by inhibitors of the HIF prolyl 4-hydroxylases? These enzymes stabilize HIF leading to tran- pool) in cortical neuronal cultures treated with DFO or compound A scriptional up-regulation of a cassette of genes involved in hypoxic (not shown). As some of the proteins established to play a role in iron homeostasis are regulated by prolyl hydroxylases (e.g.IRP2 (50)) or HIF adaptation, including VEGF and erythropoietin (56, 58). The estab- lished ability of these peptide growth factors to prevent neuronal injury (e.g. transferrin receptor (51)), the lack of change in IRP-IRE binding and death in vitro and in vivo makes them attractive candidates to medi- raises the intriguing possibility that prolyl hydroxylase inhibition is a ate the neuroprotective effects of HIF prolyl 4-hydroxylase inhibition. compensatory response to cellular iron deficiency that permits the labile Additional studies are needed to determine whether HIF stabilization is iron levels to remain normal despite the presence of a chelator that necessary for the protective effects of HIF prolyl 4-hydroxylase inhibi- reduces total iron. Given the oral availability of compound A (DFO tion. Indeed, growing awareness of other substrates such as IRP2 (59) cannot be taken orally) and its efficacy against oxidative stress in vitro and RNA polymerase (60), whose stability can be regulated by the HIF and permanent focal ischemia in vivo, the findings also suggest that this prolyl 4-hydroxylases, suggests other plausible schemes in addition to or molecule may have advantages over DFO as a therapeutic agent for exclusive of HIF that may account for protection by HIF prolyl 4-hy- acute and chronic neurological conditions where DFO has been shown droxylase inhibition. Even if HIF-1 stabilization is not necessary for the to be effective (24, 26). protection against oxidative stress or ischemia provided by low molec- Prior studies in non-neuronal systems have suggested that 3,4-DHB ular weight or peptide inhibitors of the HIF prolyl 4-hydroxylases, also stabilizes HIF via its ability to bind iron (54, 55). However, several HIF-1 stabilization appears to be a good marker of HIF prolyl 4-hy- observations herein argue against this possibility in embryonic rat cor- droxylase inhibition. tical neurons. First, the concentrations of 3,4-DHB (10 M) required to The ability of inhibitors of HIF prolyl 4-hydroxylases to prevent neu- stabilize HIF and protect cortical neurons were not found to bind iron in ronal injury when given prior to permanent focal ischemia suggests that one study where 3,4-DHB was reported to act as an iron chelator (54). they be considered as neuroprotective agents in clinical situations Second, in vitro studies using recombinant collagen prolyl 4-hydroxyl- where the imminent risk of ischemic or oxidative neuronal injury is ase showed that 3,4-DHB inhibitory actions occur via the displacement high. Coronary bypass surgery, abdominal aortic aneurysm repair, and of 2-oxoglutarate and ascorbate but not iron (45). Third, total iron levels acute anterior wall myocardial infarction with a thrombus are examples and IRP-IRE binding in 3,4-DHB-treated neurons were unchanged as of clinical situations where increasing the threshold for oxidative or compared with controls (TABLE ONE and data not shown). Despite the hypoxic-ischemic brain injury is desirable. Studies are currently under- inability of 3,4-DHB to bind iron, it was still capable of stabilizing HIF, way to define whether HIF prolyl 4-hydroxylase inhibitors are effective activating HIF-dependent gene expression, and protecting cortical neu- in preventing neuronal damage when given after the onset of ischemia rons from oxidative stress in vitro and cerebral ischemia in vivo (Figs. 1 or during the period of stroke recovery. The ability of HIF prolyl 4-hy- and 4). Thus, inhibition of the HIF prolyl 4-hydroxylases is the likely droxylase inhibitors to induce expression of cellular, local, and systemic relevant consequence of treatment with 3,4-DHB, compound A, and homeostatic responses to ischemia suggests that these agents may be DFO. For chronic neurodegenerative conditions where daily use of an useful in post-event treatment in addition to the preventative treatment HIF prolyl 4-hydroxylase inhibitor that chelates iron may result in rest- described herein. less legs syndrome (56) or anemia (57), the use of an HIF prolyl 4-hy- droxylase inhibitor that targets 2-oxoglutarate or ascorbate rather than Acknowledgments—We thank Robert Freeman and Gregg Semenza for provid- iron has potential advantages. ing the HRE reporters. We thank Brett Langley, Philipp Lange, Kyungsun Suh, The 2-oxoglutarate-dependent prolyl 4-hydroxylases are a family of and JoAnn Gensert for their helpful discussions. We are especially thankful to iron- and ascorbate-dependent enzymes that include not only HIF Johanna Myllyharju for input and advice. We thank Yixin Ben for excellent prolyl 4-hydroxylases but collagen prolyl 4-hydroxylases as well. DFO, technical assistance and Wayne Kleinman for excellent editorial assistance. 3,4-DHB, and compound A would be expected to inhibit both subfam- We also acknowledge the support of the Spinal Cord Injury Research Board of the New York State Department of Health in this study. ilies of enzymes, although recent studies suggest the K of these agents for collagen prolyl 4-hydroxylases and HIF prolyl 4-hydroxylases may be distinct (37). In an attempt to develop a more specific inhibitor of the REFERENCES HIF prolyl 4-hydroxylases, we conjugated a 19-amino acid C-terminal 1. Zecca, L., Stroppolo, A., Gatti, A., Tampellini, D., Toscani, M., Gallorini, M., Giaveri, portion of the oxygen-dependent degradation domain of HIF to an G., Arosio, P., Santambrogio, P., Fariello, R. G., Karatekin, E., Kleinman, M. H., Turro, N., Hornykiewicz, O., and Zucca, F. A. (2004) Proc. Natl. Acad. Sci. U. S. A. 101, 11-amino acid protein transduction domain from the HIV-tat protein. 9843–9848 Recent studies have shown that this 19-amino acid peptide (ODD/wt) 2. Kress, G. J., Dineley, K. E., and Reynolds, I. J. (2002) J. Neurosci. 22, 5848–5855 but not shorter peptides bind all three human isoforms of the HIF-prolyl 3. Connor, J. R., Menzies, S. L., Burdo, J. R., and Boyer, P. J. (2001) Pediatr. Neurol. 25, 4-hydroxylases with high affinity and are efficiently hydroxylated by 118–129 4. Gerlach, M., Ben-Shachar, D., Riederer, P., and Youdim, M. B. (1994) J. Neurochem. these enzymes (37). We verified that our Tat-HIF/ODD/wt peptide is a 63, 793–807 substrate with similar affinity for the recombinant HIF prolyl 4-hy- 5. Thompson, K. J., Shoham, S., and Connor, J. R. (2001) Brain Res. Bull. 55, 155–164 droxylases as HIF/ODD/wt (Fig. 3A). We also show that intracellular 6. Castellani, R. J., Siedlak, S. L., Perry, G., and Smith, M. A. (2000) Acta Neuropathol. delivery of Tat-HIF/ODD/wt (but not a control peptide with the proline 100, 111–114 7. Dexter, D. T., Wells, F. R., Lees, A. J., Agid, F., Agid, Y., Jenner, P., and Marsden, C. D. hydroxylation sites mutated to alanine) stabilizes HIF-1 (Fig. 4), effec- (1989) J. Neurochem. 52, 1830–1836 tively activates HIF-dependent gene expression (Fig. 5), and protects 8. Kienzl, E., Jellinger, K., Stachelberger, H., and Linert, W. (1999) Life Sci. 65, embryonic rat cortical neurons from oxidative stress (Fig. 6). These 1973–1976 findings suggest that Tat-HIF/ODD/wt is also an effective inhibitor of 9. Sofic, E., Riederer, P., Heinsen, H., Beckmann, H., Reynolds, G. P., Hebenstreit, G., 41742 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury and Youdim, M. B. (1988) J. Neural Transm. 74, 199–205 37. Hirsila, M., Koivunen, P., Gunzler, V., Kivirikko, K. I., and Myllyharju, J. (2003) J. Biol. 10. Thompson, C. A. (2001) Am. J. Health Syst. Pharm. 58, 649–650 Chem. 278, 30772–30780 11. Smith, M. A., Rottkamp, C. A., Nunomura, A., Raina, A. K., and Perry, G. (2000) 38. Breuer, W., Epsztejn, S., and Cabantchik, Z. I. (1996) FEBS Lett. 382, 304–308 Biochim. Biophys. Acta 1502, 139–144 39. Bettany, A. J., Eisenstein, R. S., and Munro, H. N. (1992) J. Biol. Chem. 267, 12. Mazur-Kolecka, B., Frackowiak, J., Le Vine, H., III, Haske, T., and Wisniewski, H. M. 16531–16537 (1997) Brain Res. 760, 255–260 40. Bederson, J. B., Pitts, L. H., Germano, S. M., Nishimura, M. C., Davis, R. L., and 13. Good, P. F., Perl, D. P., Bierer, L. M., and Schmeidler, J. (1992) Ann. Neurol. 31, Bartkowski, H. M. (1986) Stroke 17, 1304–1308 286–292 41. Belayev, L., Alonso, O. F., Busto, R., Zhao, W., and Ginsberg, M. D. (1996) Stroke 27, 14. Deibel, M. A., Ehmann, W. D., and Markesbery, W. R. (1996) J. Neurol. Sci. 143, 1616–1623 137–142 42. Swanson, R. A., Morton, M. T., Tsao-Wu, G., Savalos, R. A., Davidson, C., and Sharp, 15. Connor, J. R., Menzies, S. L., St Martin, S. M., and Mufson, E. J. (1992) J. Neurosci. Res. F. R. (1990) J. Cereb. Blood Flow Metab. 10, 290–293 31, 75–83 43. Ayoub, I. A., Lee, E. J., Ogilvy, C. S., Beal, M. F., and Maynard, K. I. (1999) Neurosci. 16. Schapira, A. H. (1999) Biochim. Biophys. Acta 1410, 159–170 Lett. 259, 21–24 17. Waldvogel, D., van Gelderen, P., and Hallett, M. (1999) Ann. Neurol. 46, 123–125 44. Ratan, R. R., Ryu, H., Lee, J., Mwidau, A., and Neve, R. L. (2002) Methods Enzymol. 18. Aisen, P. (1994) Adv. Exp. Med. Biol. 356, 31–40 352, 183–190 19. Bralet, J., Schreiber, L., and Bouvier, C. (1992) Biochem. Pharmacol. 43, 979–983 45. Majamaa, K., Gunzler, V., Hanauske-Abel, H. M., Myllyla, R., and Kivirikko, K. I. 20. Lipscomb, D. C., Gorman, L. G., Traystman, R. J., and Hurn, P. D. (1998) Stroke 29, (1986) J. Biol. Chem. 261, 7819–7823 487–493 46. Epstein, A. C., Gleadle, J. M., McNeill, L. A., Hewitson, K. S., O’Rourke, J., Mole, D. R., 21. Hurn, P. D., Koehler, R. C., Blizzard, K. K., and Traystman, R. J. (1995) Stroke 26, Mukherji, M., Metzen, E., Wilson, M. I., Dhanda, A., Tian, Y. M., Masson, N., Ham- 688–694 ilton, D. L., Jaakkola, P., Barstead, R., Hodgkin, J., Maxwell, P. H., Pugh, C. W., 22. Kaur, D., and Andersen, J. K. (2002) Aging Cell 1, 17–21 Schofield, C. J., and Ratcliffe, P. J. (2001) Cell 107, 43–54 23. Bowern, N., Ramshaw, I. A., Clark, I. A., and Doherty, P. C. (1984) J. Exp. Med. 160, 47. Myllyharju, J. (2003) Matrix Biol. 22, 15–24 1532–1543 48. Metzen, E., Berchner-Pfannschmidt, U., Stengel, P., Marxsen, J. H., Stolze, I., Klinger, 24. McLachlan, D. R., Smith, W. L., and Kruck, T. P. (1993) Ther. Drug Monit. 15, M., Huang, W. Q., Wotzlaw, C., Hellwig-Burgel, T., Jelkmann, W., Acker, H., and 602–607 Fandrey, J. (2003) J. Cell Sci. 116, 1319–1326 25. Ritchie, C. W., Bush, A. I., Mackinnon, A., Macfarlane, S., Mastwyk, M., MacGregor, 49. Aarts, M., Liu, Y., Liu, L., Besshoh, S., Arundine, M., Gurd, J. W., Wang, Y. T., Salter, L., Kiers, L., Cherny, R., Li, Q. X., Tammer, A., Carrington, D., Mavros, C., Volitakis, M. W., and Tymianski, M. (2002) Science 298, 846–850 I., Xilinas, M., Ames, D., Davis, S., Beyreuther, K., Tanzi, R. E., and Masters, C. L. 50. Liang, D., Dawson, T. M., and Dawson, V. L. (2004) Curr. Mol. Med. 4, 207–225 (2003) Arch. Neurol. 60, 1685–1691 51. Bergeron, M., Gidday, J. M., Yu, A. Y., Semenza, G. L., Ferriero, D. M., and Sharp, F. R. 26. Kaur, D., Yantiri, F., Rajagopalan, S., Kumar, J., Mo, J. Q., Boonplueang, R., Viswanath, (2000) Ann. Neurol. 48, 285–296 V., Jacobs, R., Yang, L., Beal, M. F., DiMonte, D., Volitaskis, I., Ellerby, L., Cherny, 52. Davis, S., Helfaer, M. A., Traystman, R. J., and Hurn, P. D. (1997) Stroke 28, 198–204 R. A., Bush, A. I., and Andersen, J. K. (2003) Neuron 37, 899–909 53. Prass, K., Ruscher, K., Karsch, M., Isaev, N., Megow, D., Priller, J., Scharff, A., Dirnagl, 27. Winterbourn, C. C. (1995) Toxicol. Lett. 82, 969–974 U., and Meisel, A. (2002) J. Cereb. Blood Flow Metab. 22, 520–525 28. Zaman, K., Ryu, H., Hall, D., O’Donovan, K., Lin, K. I., Miller, M. P., Marquis, J. C., 54. Wang, F., Sekine, H., Kikuchi, Y., Takasaki, C., Miura, C., Heiwa, O., Shuin, T., Fujii- Baraban, J. M., Semenza, G. L., and Ratan, R. R. (1999) J. Neurosci. 19, 9821–9830 Kuriyama, Y., and Sogawa, K. (2002) Biochem. Biophys. Res. Commun. 295, 657–662 29. Willam, C., Masson, N., Tian, Y. M., Mahmood, S. A., Wilson, M. I., Bicknell, R., 55. Warnecke, C., Griethe, W., Weidemann, A., Jurgensen, J. S., Willam, C., Bachmann, Eckardt, K. U., Maxwell, P. H., Ratcliffe, P. J., and Pugh, C. W. (2002) Proc. Natl. Acad. S., Ivashchenko, Y., Wagner, I., Frei, U., Wiesener, M., and Eckardt, K. U. (2003) Sci. U. S. A. 99, 10423–10428 FASEB J. 17, 1186–1188 30. Ivan, M., Kondo, K., Yang, H., Kim, W., Valiando, J., Ohh, M., Salic, A., Asara, J. M., 56. Wang, X., Zhu, C., Gerwien, J. G., Schrattenholz, A., Sandberg, M., Leist, M., and Lane, W. S., and Kaelin, W. G., Jr. (2001) Science 292, 464–468 Blomgren, K. (2004) J. Neurochem. 91, 900–910 31. Jaakkola, P., Mole, D. R., Tian, Y. M., Wilson, M. I., Gielbert, J., Gaskell, S. J., Krieg- 57. White, K. C. (2005) Pediatrics 115, 315–320 sheim, A., Hebestreit, H. F., Mukherji, M., Schofield, C. J., Maxwell, P. H., Pugh, C. W., 58. Shi, Y. H., Wang, Y. X., Bingle, L., Gong, L. H., Heng, W. J., Li, Y., and Fang, W. G. and Ratcliffe, P. J. (2001) Science 292, 468–472 (2005) J. Pathol. 205, 530–536 32. Yu, F., White, S. B., Zhao, Q., and Lee, F. S. (2001) Cancer Res. 61, 4136–4142 59. Hanson, E. S., Rawlins, M. L., and Leibold, E. A. (2003) J. Biol. Chem. 278, 33. Semenza, G. L., Agani, F., Feldser, D., Iyer, N., Kotch, L., Laughner, E., and Yu, A. 40337–40342 (2000) Adv. Exp. Med. Biol. 475, 123–130 60. Kuznetsova, A. V., Meller, J., Schnell, P. O., Nash, J. A., Ignacak, M. L., Sanchez, Y., 34. Bruick, R. K. (2003) Genes Dev. 17, 2614–2623 Conaway, J. W., Conaway, R. C., and Czyzyk-Krzeska, M. F. (2003) Proc. Natl. Acad. 35. Murphy, T. H., Schnaar, R. L., and Coyle, J. T. (1990) FASEB J. 4, 1624–1633 36. Mosmann, T. (1983) J. Immunol. Methods 65, 55–63 Sci. U. S. A. 100, 2706–2711 DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41743 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Biological Chemistry American Society for Biochemistry and Molecular Biology

Hypoxia-inducible Factor Prolyl 4-Hydroxylase Inhibition: A TARGET FOR NEUROPROTECTION IN THE CENTRAL NERVOUS SYSTEM *

Download PDF
12 pages

Loading next page...
 
/lp/american-society-for-biochemistry-and-molecular-biology/hypoxia-inducible-factor-prolyl-4-hydroxylase-inhibition-a-target-for-gpCYLY35Lq

References (62)

Publisher
American Society for Biochemistry and Molecular Biology
Copyright
Copyright © 2005 Elsevier Inc.
ISSN
0021-9258
eISSN
1083-351X
DOI
10.1074/jbc.m504963200
Publisher site
See Article on Publisher Site

Abstract

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 280, NO. 50, pp. 41732–41743, December 16, 2005 © 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A. Hypoxia-inducible Factor Prolyl 4-Hydroxylase Inhibition A TARGET FOR NEUROPROTECTION IN THE CENTRAL NERVOUS SYSTEM Received for publication, May 5, 2005, and in revised form, October 13, 2005 Published, JBC Papers in Press, October 13, 2005, DOI 10.1074/jbc.M504963200 ‡§¶ ‡ §¶ ‡ ‡ Ambreena Siddiq , Issam A. Ayoub , Juan C. Chavez , Leila Aminova , Sapan Shah , Joseph C. LaManna , ‡‡ ‡‡ ‡‡§§ Stephanie M. Patton**, James R. Connor**, Robert A. Cherny , Irene Volitakis , Ashley I. Bush , ¶¶ ¶¶ ¶¶ ‡§¶1 Ingrid Langsetmo , Todd Seeley , Volkmar Gunzler , and Rajiv R. Ratan From the Department of Neurology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts § ¶ 02115, Burke/Cornell Medical Research Institute, White Plains, New York 10605, the Department of Neurology and Neuroscience, Weill Medical College of Cornell University, New York, New York 10021, the Department of Anatomy and Neurology, Case Western Reserve University, Cleveland, Ohio 44106, the **Department of Neurosurgery, George M. Leader Family Laboratory for Alzheimer ‡‡ Disease Research, Penn State College of Medicine, Hershey, Pennsylvania 17033, the Department of Pathology, the University of §§ Melbourne, Mental Health Research Institute of Victoria, Parkville 3052, Australia, Laboratory of Oxidation Biology, Genetics and Aging Research Unit, Department of Psychiatry, Harvard Medical School, Massachusetts General Hospital, Charlestown, ¶¶ Massachusetts 02129, and Fibrogen, Inc., South San Francisco, California 94080 Hypoxia-inducible factor (HIF) prolyl 4-hydroxylases are a family with Parkinson disease (6–9). Alzheimer disease has also been found to of iron- and 2-oxoglutarate-dependent dioxygenases that nega- be associated with an increase in the iron content of senile plaques tively regulate the stability of several proteins that have established (10–15). Accumulation of mitochondrial iron has been shown to play a roles in adaptation to hypoxic or oxidative stress. These proteins role in Friedrich ataxia (16, 17). Similarly, changes in intracellular free include the transcriptional activators HIF-1 and HIF-2. The abil- iron levels have been observed in cerebral ischemia (18–20). Direct ity of the inhibitors of HIF prolyl 4-hydroxylases to stabilize pro- evidence that disrupted iron homeostasis contributes to injury rather teins involved in adaptation in neurons and to prevent neuronal than simply being caused by it has been obtained by treatment with low injury remains unclear. We reported that structurally diverse low molecular weight iron chelators or by overexpression of iron storage molecular weight or peptide inhibitors of the HIF prolyl 4-hydroxy- proteins. Small molecule iron chelators such as deferoxamine mesylate lases stabilize HIF-1 and up-regulate HIF-dependent target genes (DFO) inhibit neuronal injury in rodent models of stroke (21), Parkin- waf1/cip1 (e.g. enolase, p21 , vascular endothelial growth factor, or son disease (22), and multiple sclerosis (23). Moreover, DFO and some erythropoietin) in embryonic cortical neurons in vitro or in adult rat other metal chelators such as clioquinol have been shown to slow the brains in vivo. We also showed that structurally diverse HIF prolyl progression of Alzheimer disease in humans (24, 25). Similarly, the 4-hydroxylase inhibitors prevent oxidative death in vitro and ische- forced expression of the iron binding and storage protein ferritin in the mic injury in vivo. Taken together these findings identified low substantia nigra diminishes iron accumulation and prevents neuronal molecular weight and peptide HIF prolyl 4-hydroxylase inhibitors loss in a rodent model of Parkinson disease (26). Together, these find- as novel neurological therapeutics for stroke as well as other dis- ings suggest that iron-dependent toxicity is part of the affector pathway eases associated with oxidative stress. of injury in a host of neurological conditions. How do iron chelators prevent neuronal injury? Iron is generally believed to participate in neuronal dysfunction and death through its ability to catalyze (via electron donation) the generation of highly reac- Iron maintains a unique role in physiology via its ability to change tive hydroxyl radicals via Fenton chemistry (1). In this model hydroxyl readily its oxidation state in response to changes in its local environ- ment. A general simplification of its primary function is that it mediates radicals modify lipid, protein, and DNA targets to induce cell dysfunc- one-electron redox reactions. This chemical property of iron enables it tion in these cellular constituents (27). It has been proposed that iron chelators can inhibit Fenton chemistry and prevent neuronal cell injury to act as an essential component in several biological activities, includ- by sequestering accumulated free iron; however, this pathway of pro- ing as a cofactor for enzymes such as tyrosine hydroxylase. Oxygen binding to biomolecules such as hemoglobin and myoglobin is also tection remains controversial. Indeed, prior studies from our laboratory coordinated by iron. Indeed iron deficiency can lead to a host of disor- have suggested alternative pathways of protection by iron chelators. In an in vitro model of oxidative stress, we correlated the protective ders, including anemia and restless legs syndrome (1). effects of iron chelators with their ability to activate the transcriptional Paradoxically, the biochemical properties that make iron beneficial in many biological processes appear to be a drawback when the balance activator hypoxia-inducible factor-1 (28). In this scheme, iron chelators between its accumulation/sequestration within cellular compartments inhibit the activity of a class of iron-dependent enzymes called the HIF prolyl 4-hydroxylases. Inhibition of HIF prolyl 4-hydroxylases prevents and its release is disturbed in favor of iron accumulation (2). Indeed, iron overload is associated with several neurological conditions (3–5). For example, the iron content of nigral Lewy bodies is elevated in patients The abbreviations used are: DFO, deferoxamine mesylate; 3,4-DHB 3,4-dihydroxyben- zoate; hypoxia-inducible factor; VEGF, vascular endothelial growth factor; HCA, * This work was supported by National Institutes of Health Grants NS39170, NS40591, homocysteate; DIV, day in vitro; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetra- and NS46239 (to R. R. R.). The costs of publication of this article were defrayed in part zolium bromide; AM, acetoxymethyl ester; PBS, phosphate-buffered saline; RT, by the payment of page charges. This article must therefore be hereby marked reverse transcription; MCAO, middle cerebral artery occlusion; TTC, 2,3,5-tri- “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. phenyltetrazolium chloride; DAPI, 4,6-diamidino-2-phenylindole; ANOVA, analysis of To whom correspondence should be addressed: Burke/Cornell Medical Research Insti- variance; FITC, fluorescein-isothiocyanate; HRE, hypoxia-response element; ICP-MS, tute, 785 Mamaroneck Ave., White Plains, NY 10605. Tel.: 914-597-2851; Fax: 914-597- inductively coupled plasma-mass spectrometry; HIV, human immunodeficiency 2225; E-mail: [email protected] or [email protected]. virus. 41732 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 This is an Open Access article under the CC BY license. HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury the attachment of an OH group (hydroxylation) to phylogenetically and 100 M cystine, containing the glutamate analog homocysteate conserved proline residues at amino acid 402 and 564 in the protein (HCA; 5 mM). HCA was diluted from 100-fold concentrated solutions HIF-1 (29). Unhydroxylated HIF-1 does not bind the ubiquitin-pro- that were adjusted to pH 7.5. Viability was assessed by calcein-ace- tein isopeptide ligase, von Hippel-Lindau protein (30–32). HIF-1 is toxymethyl ester (AM)/ethidium homodimer-1 staining (live/dead thus not ubiquitinated and degraded by the proteasome. Once stabi- assay) (Molecular Probes, Eugene, OR) under fluorescence microscopy lized, nuclear HIF-1 can heterodimerize with its partner HIF-1, bind and the MTT assay (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazo- to the pentanucleotide hypoxia-response element (5-TGATC-3)in lium bromide) method (36). To evaluate the effects of the drugs on gene regulatory regions, and transactivate the expression of established cytotoxicity, we added DFO, 3,4-DHB, compound A, and the HIF pep- protective genes, including vascular endothelial growth factor (VEGF) tide simultaneously with HCA. and erythropoietin (33, 34). According to this model, the protective Immunofluoroscence Staining—Indirect labeling methods were used effects of iron chelators are not exclusively the result of suppression of to determine the levels of HIF-1 protein in cortical neuron cultures. Fenton chemistry, but may also result from the inhibition of iron-de- Dissociated cells from cerebral cortex were seeded onto poly-D-lysine- pendent prolyl hydroxylases. However, direct evidence that HIF-prolyl coated 8-well culture slides (BD Biosciences) and treated with Tat-HIF/ hydroxylase inhibition is a relevant target for protection from oxidative ODD/mut, Tat-HIF/ODD/wt, and DFO overnight. Cells were washed death by iron chelators has yet to be presented. In this paper we show with warm PBS and fixed at room temperature for 15 min with 4% that small molecule or peptide inhibitors of the HIF prolyl 4-hydroxyl- paraformaldehyde. After washing, cells were incubated with blocking ases are sufficient to inhibit neuronal death because of oxidative stress in solution containing 0.3% Triton X-100 and 5% goat serum in PBS for 1 h, vitro and because of permanent focal ischemia in vivo. These results followed by incubation with rabbit HIF-1 antibody (Upstate Cell Signal- implicate HIF prolyl 4-hydroxylases as targets for neuroprotection in ing, Lake Placid, NY) (1:100 dilution) overnight. After three washes with the central nervous system. PBS, cells were incubated with rhodamine-conjugated goat anti-rabbit IgG antibody (Molecular Probes, Eugene, OR) (1:200 dilution) and EXPERIMENTAL PROCEDURES DAPI. The slides were washed three times with PBS and mounted with fluorochrome mounting solution (Vector Laboratories). Images were Compound A, a proprietary compound, was obtained from Fibrogen analyzed under a confocal fluorescence microscope (Zeiss LSM50, Ger- Inc. DFO and 3,4-DHB were purchased from Sigma. Tat-HIF/ODD/wt many). Control experiments were performed in the absence of primary (YGKKRRQRRRDLDLEMLAPYIPMDDDFQL) and Tat-HIF/ODD/ antibody. mut (YGKKRRQRRRDLDLEMLAAYIAMDDDFQL) peptides were HIF Prolyl 4-Hydroxylase Activity Assay—The effects of Tat-HIF/ synthesized by Biopolymers Laboratory, Harvard Medical School. 40, ODD/mut, Tat-HIF/ODD/wt on purified HIF-prolyl 4-hydroxylase 100, and 10 M of compound A, DFO, and 3,4-DHB, respectively, were activity was analyzed by the method based on the hydroxylation-cou- used throughout the studies unless otherwise stated. Peptides were used pled decarboxylation of 2-oxo[1- C]glutarate (37). at a concentration of 100 M unless otherwise stated. ICP-MS Analysis—Cortical neurons were plated and treated with Primary Neurons—Cell cultures were obtained from the cerebral 3,4-DHB, DFO, or compound A. Cells were collected, lyophilized, and cortex of fetal Sprague-Dawley rats (embryonic day 17) as described stored at 80 °C until further testing. 50 l of concentrated nitric acid previously (35). All experiments were initiated 24 h after plating. Under (HNO ) (Aristar, BDH) was added to each lyophilized sample. Samples these conditions, the cells are not susceptible to glutamate-mediated 3 were allowed to digest overnight at room temperature followed by heat- excitotoxicity. ing for 20 min at 90 °C using a heating block. Samples were diluted 1:25 Immunoblot Analysis—Cell lysates were obtained by rinsing cortical with 1% HNO prior to analysis. Analysis was carried out on a Varian neurons with cold PBS and adding 0.1 M potassium phosphate contain- 3 UltraMass ICP-MS instrument using operating conditions suitable for ing 0.5% Triton X-100. Samples were then boiled in Laemmli buffer and routine multielement analysis. The instrument was calibrated using 0, electrophoresed under reducing conditions on 12% polyacrylamide gels. 10, 50, and 100 ppb of a certified multielement ICP-MS standard solu- Proteins were then transferred to a polyvinylidene difluoride membrane tion (ICP-MS-CAl2-1, AccuStandard). A certified internal standard (Bio-Rad). Nonspecific binding was inhibited by incubation in Tris- solution containing 100 ppb of yttrium ( Y) as an internal control (ICP- buffered saline with Tween 20 (TBST: 50 mM Tris-HCl, pH 8.0, 0.9% MS-IS-MIX1-1, AccuStandard) was also used. NaCl, and 0.1% Tween 20) containing 5% nonfat dried milk for 1.5 h. Calcein-AM Assay—Primary neurons were treated with the drugs Primary antibodies against aldolase (rabbit muscle; Rockland), -tubu- 3,4-DHB, DFO, and compound A overnight as described previously. lin (Sigma), HIF-1 (Novus Biologicals, Littleton, CO), VEGF (Santa Cruz Biotechnology), and p21 (BD Biosciences) were diluted 1:2000, Fluorescence was measured from intact cells using an excitation wave- length of 485 nm and an emission wavelength of 535 nm (38) after 1:2000, 1:500, 1:400 and 1:500, respectively, in TBST containing 1% milk addition of 0.25 M calcein-AM (Molecular Probes, Eugene, OR) was overnight at 4 °C followed by incubation with respective horseradish peroxidase-conjugated secondary antibodies for2hat room tempera- added to cultured neuronal cultures for 30 min at room temperature. IRP/IRE Binding Assay—Cells were washed with PBS, trypsinized, ture. Immunoreactive proteins were detected according to the and collected. To separate membrane-bound organelles from cytosol, enhanced chemiluminescent protocol (Amersham Biosciences). Luciferase Assay—Primary neurons infected with adenovirus (multi- the cells were washed and homogenized in lysis buffer (10 mM/liter HEPES, pH 7.4, 40 mmol/liter KCl, 5% glycerol, 3 mmol/liter MgCl , 0.1 plicity of infection  100) containing luciferase gene driven by an HRE mmol/liter EDTA, 1 mmol/liter dithiothreitol, 1 mmol/liter phenyl- promoter (a kind gift from Dr. Robert Freeman) were treated with DFO, 3,4-DHB, compound A, and the HIF-peptides overnight. The luciferase methylsulfonyl fluoride, and 0.1 g/liter leupeptin). Cells were sonicated assay was performed using the bioluminescent method (Promega to disrupt membranes. The homogenates were centrifuged for 5 min at Corp.). 4 °C and 4000  g to sediment nuclei. The resulting supernatant was Viability Assays—For cytotoxicity studies, cells were rinsed with centrifuged at 4 °C and 145,000  g for 1 h, and the cytosolic fraction warm PBS and then placed in minimum essential medium (Invitrogen) was collected. For preparation of RNA transcripts, the pGEM7Zf()- containing 5.5 g/liter glucose, 10% fetal calf serum, 2 mML-glutamine, 5L containing the complete 5-untranslated region of rat liver L-ferritin DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41733 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury cDNA (39) was provided by R. S. Eisenstein (University of Wisconsin, tion. This method placed the tip of the suture at the origin of the ante- Madison). The ferritin IRE RNA transcripts were prepared from maI- rior cerebral artery, thereby occluding the MCA. The placement of the suture tip was monitored by laser Doppler flowmetry measurements of digested pGEM7Zf()-5L containing 100 bases of 5-untranslated region and 40 bases from the vector. RNA transcripts were synthesized regional cerebral blood flow. MCAO caused a sharp drop in regional from 1 g of linearized plasmid DNA in the presence of 100 Ci of cerebral blood flow to less than 30% of preischemic base line. The suture was left in place, and the animals were allowed to awaken from the [- P]CTP (800 Ci/mmol), 12 M unlabeled CTP, 0.5 mM of the three unlabeled nucleotides, 100 mM dithiothreitol, 20 units of T7 RNA anesthesia following closure of the operation sites. polymerase in a 20-l reaction volume. The reaction was carried out for Drug Administration—3,4-DHB was dissolved in ethanol (vehicle) for a total volume of 100 l and was administered intraperitoneally to 1 h at 37 °C. The yield of the transcript synthesized was determined by trichloroacetic acid precipitation. Under these conditions, the specific animals (n  6) at a dose of 180 mg/kg of body weight. 3,4-DHB was administered 6 h before MCAO. Compound A was administered to activity was5 10 cpm/g RNA. For the RNA band shift assay, 5 g animals (n 9) at a dose of 100 mg/kg of body weight using a gavage 6 h of protein for the cytoplasmic fractions were incubated with the syn- thetic radiolabeled IRE probe. The RNA-protein complexes were sepa- before the MCAO. The control animals received an equivalent volume of the vehicle (ethanol for 3,4-DHB and carboxymethylcellulose for rated on a 6% native polyacrylamide gel and visualized by autoradiogra- compound A) on a similar administration schedule. phy. The autoradiograms were scanned, and the resulting digital image Infarct Measurement—24 h after MCAO, the animal were anesthe- was subjected to densitometric band analysis using the Labworks Anal- tized with ketamine (100 mg/kg, intraperitoneally) and xylazine (50 ysis software (UVP Inc., Upland, CA). The results of the RNA band shift mg/kg, intraperitoneally) and decapitated. The brain was rapidly assays were expressed as active IRP binding from each treatment as a removed, sliced into seven 2-mm coronal sections using a rat matrix ratio of control. (RBM 4000C, ASI Instrument Inc., Warren, MI), and stained according RT-PCR—The levels of p21, enolase, VEGF, and -actin were ana- to the standard 2,3,5-triphenyltetrazolium chloride (TTC) method (40, lyzed by semiquantitative RT-PCR by using the one-step RT-PCR kit 41). Each slice was drawn using a computerized image analyzer (Scion ReddyMix version (Abgene). The following primers were used in this Corp.). The calculated infarction areas were then compiled to obtain the study: enolase 5-GGTTCTCATGCTGGCAACAAGT-3 (sense) and TM infarct volumes per brain (in mm ). Infarct volumes were expressed as a 5-TAAACCTCTGCTCCAATGCGC-3 (antisense) (GenBank percentage of the contralateral hemisphere volume to compensate for accession number M012554); VEGF 5-TGCAATGATGAAGCCCT- edema formation in the ipsilateral hemisphere (42, 43). GGA-3 (sense) and 5-TGCTATGCTGCAGGAAGCTCA-3 (anti- TM sense) (GenBank accession number M32167); p21 5-ATGTCCGA- RESULTS TCCTGGTGATGT-3 (sense) and 5-ACTTCAGGGCTTTCTCTT- TM GC-3 (antisense) (GenBank accession number U24174); -actin 5- Low Molecular Weight HIF Prolyl 4-Hydroxylase Inhibitors Stabilize CCTCATGAAGATCCTGACCG-3 (sense) and 5-TGCCAATAGT- HIF and Activate HIF-dependent Gene Expression in Embryonic Rat TM GATGACCTGG-3 (antisense) (GenBank accession number Cortical Neurons—Oxidative glutamate toxicity in immature cortical NM007393). neurons results from the ability of glutamate to inhibit competitively the In Vivo Experiments—All experimental procedures were approved by uptake of cystine via its plasma membrane transporter (system X ) the Harvard Medical Area Standing Committee on Animals and meet (44). Inhibition of cystine uptake into the cell via system X  results in the standards of the Federal and State reviewing organizations. depletion of intracellular cysteine, the rate-limiting precursor of gluta- Animal Preparation and Monitoring—Eleven adult male Sprague- thione synthesis. Cyst(e)ine deprivation thus leads to depletion of the Dawley rats (n  11) weighing 250–280 g (Charles River Breeding Lab- major cellular antioxidant glutathione, an imbalance in cellular oxidants oratories, Wilmington, MA) were operated on. Nine rats survived and and antioxidants and oxidative stress-induced death. Indeed, a host of were analyzed for this study. Mortality was 18% (2:11) and occurred agents with known antioxidant capacity can completely abrogate gluta- equally in the groups of animals (one rat in each group). Death occurred mate toxicity in this paradigm. To determine whether prolyl 4-hydrox- during the first 20 h postoperatively. Animals were allowed free access ylase inhibition is sufficient to prevent oxidative glutamate toxicity, we to food and water before and after surgery. Briefly, rats were anesthe- treated immature cortical neurons with established, structurally tized by an intraperitoneal injection of 400 mg/kg chloral hydrate fol- diverse, low molecular weight inhibitors of the prolyl 4-hydroxylases, lowed 45 min later by a maintenance intraperitoneal infusion at a rate of DFO, 3,4-DHB, and compound A. DFO, 3,4-DHB, and compound A 120 mg/kg/h using a butterfly needle set. The animals were free breath- were selected for testing because their mechanisms of prolyl hydroxy- ing. Their body temperatures were kept stable at 36.5  0.5 °C using a lase inhibition are believed to be distinct. 3,4-DHB has been shown feedback-regulating heating pad and a rectal probe (Harvard Apparatus, previously to be an inhibitor of collagen prolyl 4-hydroxylases with MA). The right femoral artery was cannulated for measurement of arte- respect to the 2-oxoglutarate and ascorbate cofactors and to be neutral rial blood gases, glucose, and mean arterial blood pressure. These phys- with respect to Fe (45). By contrast, the ability of DFO to inhibit iological parameters were monitored before and after middle cerebral recombinant prolyl 4-hydroxylase activity can be overcome by addition artery occlusion (MCAO). In addition, laser Doppler flowmetry (Moor of exogenous iron to the test tube (45). Compound A has been reported Instruments, Devon, UK) was used to monitor the regional cerebral to be an inhibitor of HIF prolyl 4-hydroxylases (30). Immunoblotting blood flow through a burr hole 2 mm in diameter created in the right using a specific antibody to HIF-1 established that, like hypoxia, all of parietal bone (2 mm posterior and 6 mm lateral to bregma). these agents (100 M DFO, 10 M dihydroxybenzoic acid, 40 M com- Surgery—All rats were subjected to right MCAO. Under the operat- pound A) stabilize HIF-1 but do so in normoxic neurons (Fig. 1A). ing microscope, the right common carotid artery was exposed through Stabilization of HIF-1 by each of these low molecular weight com- a midline incision in the neck. A 4-0 nylon suture with its tip rounded by pounds was associated with significant transactivation of a reporter heating over a flame and subsequently coated with poly-L-lysine (Sigma) gene driven by hypoxia-response elements (boldface in the sequence) was introduced into the external carotid artery and then advanced into from the promoter region of the enolase gene (5-ACGCTGAGT- the internal carotid artery for a length of 18–19 mm from the bifurca- GCGTGCGGGACTCGGAGTACGTGACGGA-3) (Fig. 1B; p  0.05). 41734 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury A control reporter gene without these response elements or one with the hypoxia-response elements mutated were not up-regulated by these compounds (not shown). To verify that increased reporter activity was reflected in increased protein expression, immunoblotting was performed to monitor levels of waf1/cip1 aldolase, VEGF, and p21 , known HIF-1-regulated genes (33). -Tubulin was used as a control because its transcriptional regulation appears to be HIF-independent. As expected, DFO, 3,4-DHB, and com- waf1/cip1 pound A all increased protein levels of aldolase, VEGF, and p21 but not -tubulin (Fig. 1C). Taken together these results established that like non-neuronal cells, DFO, 3,4-DHB, and compound A can all inhibit HIF prolyl 4-hydroxlyases, stabilize HIF, and lead to the increased transcription of established HIF-dependent genes. Structurally Diverse Inhibitors of the HIF Prolyl 4-Hydroxylases Can Abrogate Oxidative Stress-induced Death in Cortical Neurons—To determine whether structurally diverse inhibitors of the HIF prolyl 4-hydroxylases can abrogate death because of oxidative glutamate tox- icity at concentrations at which they activate HIF, we added these com- pounds at the time of exposure to glutamate (not shown) or the gluta- mate analog homocysteate (HCA, 5 mM). As expected, all three agents significantly prevented oxidative stress-induced death, suggesting that HIF prolyl 4-hydroxylases are a target for protection from oxidative death in neurons (Fig. 2, A and B). The number of live and dead cells in the dish were monitored using several methods, including MTT reduc- tion (Fig. 2A), visualization by phase contrast microscopy (not shown), and calcein labeling or ethidium homodimer labeling to tag live cells or dead cells, respectively (Fig. 2, B–I). In all cases, the results obtained were similar. The absolute level of protection from oxidative death by the three compounds did not seem to correlate with their transcriptional efficacy. For example, compound A is the most effective in stimulating HRE- driven luciferase activity and VEGF protein levels; and despite signifi- cantly preventing oxidative death, it also induces a small but reproduc- ible basal toxicity in the cultures (Fig. 2H). These findings suggest that although HIF-dependent transcription is a good marker of HIF prolyl 4-hydroxylase inhibition, other factors besides HIF transcriptional activity may contribute to neuroprotection by agents that inhibit HIF prolyl 4-hydroxylases. Of course, one cannot rule out other prolyl 4-hy- droxylase-independent effects of the drugs on viability as an explanation for the quantitative differences in neuroprotection from the oxidative death we observed between DFO, DHB, and compound A Some but Not All Prolyl 4-Hydroxylase Inhibitors Tested Reduce Total Cellular Iron in Embryonic Rat Cortical Neurons—HIF prolyl 4-hy- droxylases are iron-dependent enzymes. To verify that at least one of the PH inhibitors tested does not alter iron levels in cortical neurons, we FIGURE 1. Structurally diverse, low molecular weight prolyl 4-hydroxylase inhibi- tors stabilize HIF-1 protein, enhance activity of a hypoxia-response element- monitored total iron levels by inductively coupled plasma mass spec- driven luciferase reporter, and increase the expression of established HIF-regu- trometry. This technique allowed us to perform sensitive analysis of lated genes in embryonic cortical neurons. A, cortical neuronal cultures (1 day in vitro chelatable and nonchelatable metals. Total cellular iron was decreased (DIV)) were treated with a vehicle control (C,Me SO (0.01%)), DFO (100 M), 3,4-DHB (10 M), or compound A (Comp-A;40 M) or exposed to 1% hypoxia (H) overnight. Proteins by overnight incubation with DFO (100 M) and compound A (40 M) from cell lysates were separated using gel electrophoresis and immunodetected using but not 3,4-DHB (10 M) treatment in cultured neurons (TABLE ONE). an antibody against HIF1-. -Actin was monitored as a loading control. B, cortical neu- ronal cultures (1 DIV) were infected with an adenovirus harboring the hypoxia-response This does indicate some ability to extract iron or prevent its uptake, in elements from the enolase promoter linked to a luciferase reporter (multiplicity of infec- this case, from cells and perhaps suggests the possibility of direct iron tion  100). Parallel infection with adenovirus containing the green fluorescent protein chelation. These findings are in agreement with prior in vitro studies (GFP) indicated reproducible infection efficiencies of30 – 40%. 24 h following infection with adeno-HRE-luciferase, cortical neuronal cultures were treated with a vehicle control demonstrating that DFO inhibits prolyl 4-hydroxylase activity by bind- (C), DFO (100 M), 3,4-DHB (10 M), or compound A (40 M). Lysates from each treatment ing its critical iron cofactor (46), whereas 3,4-DHB inhibits prolyl 4-hy- group were generated, and luciferase activity was measured. Graph depicts mean per- centage luciferase activity (compared with control) S.E. calculated from three separate droxylases by displacing 2-oxoglutarate or ascorbate and not iron (45). experiments for each group (* denotes p 0.05 by ANOVA and Student-Newman-Keuls The effect of DFO and compound A showed selectivity for the trace tests for DFO, 3,4-DHB, or compound A). C, DFO (100 M), 3,4-DHB (10 M), and com- pound A (40 M) increase expression of established HIF target genes, aldolase, VEGF, or metal iron as neither these agents nor 3,4-DHB had any effect on total waf1/cip1 p21 in cultured cortical neurons as compared with vehicle-treated control. zinc in rat neuronal cultures (TABLE ONE). Taken together, these Immunoblots are representative examples of three experiments. results are consistent with the notion that DFO and compound A act to DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41735 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury FIGURE 2. Structurally diverse, low molecular weight prolyl 4-hydrdoxylase inhibitors abro- gate oxidative glutamate toxicity in embryonic cortical neuronal cultures. A, the glutamate ana- log HCA (5 mM) was added to cortical neuronal cul- tures (1 DIV). In parallel, DFO (100 M), 3,4-DHB (10 M), and compound A (Comp-A) (40 M) were added with and without HCA. Twenty four hours later, cell viability was determined using the MTT assay. Graph depicts mean  S.E. for three experi- ments performed in triplicate (* denotes p  0.05 from HCA-treated cultures by ANOVA and Stu- dent-Newman Keuls tests for control, DFO, 3,4-DHB, or compound A). B–I, live/dead assay of cortical neuronal cultures (2 DIV). Live cells are detected by uptake and trapping of calcein-AM (green fluorescence). Dead cells fail to trap calcein but are freely permeable to the highly charged DNA intercalating dye, ethidium homodimer (red fluorescence). B, control. C, HCA (5 mM). D, DFO (100 M). E, DFO  HCA. F, 3,4-DHB (10 M). G, 3,4- DHB HCA. H, compound A (40 M). I, compound A  HCA. 41736 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury inhibit the prolyl hydroxylases by removing the Fe cofactor or binding inhibit collagen prolyl hydroxylases, HIF prolyl 4-hydroxylases, and to it within the enzyme, whereas 3,4-DHB inhibits prolyl 4-hydroxylase other 2-oxoglutarate-dependent dioxygenases in the intact cell. In order activity by iron-independent mechanisms. to determine whether HIF prolyl 4-hydroxylases are relevant for pro- A Peptide Inhibitor of the HIF Prolyl 4-Hydroxylases Delivered into tection from glutathione depletion-induced death by DFO, 3,4-DHB, Neurons Using an HIV Tat Protein Transduction Domain Activates HIF- and compound A, we used the following strategy. Myllyharju and co- dependent Gene Expression and Prevents Oxidative Death in Vitro— workers (37) recently described the identification of a 19-amino acid Low molecular weight inhibitors of the HIF prolyl 4-hydroxylases were peptide that is able to act as a substrate for all the human HIF prolyl initially identified via a search for inhibitors of the collagen prolyl 4-hy- 4-hydroxylases(K 5–10M).Thispeptidecontainspartoftheoxygen- droxylases. The collagen prolyl hydroxylases catalyze the formation of dependent domain including the C-terminal hydroxylation site of 4-hydroxyproline in collagen and certain other proteins that possess human HIF at proline 564 (DDLDEMLAPYIPMDDDFQL; boldface P collagen-like sequences by hydroxylating prolines in X-Pro-Gly motifs indicates proline 564). We hypothesized that such a peptide delivered (47). Hydroxyprolines facilitate the formation of stable triple helical into neurons could act as a competitive inhibitor of HIF prolyl 4-hy- collagen. Unlike HIF prolyl 4-hydroxylases, which are believed to be droxylases while leaving other 2-oxoglutarate-dependent dioxygenases cytoplasmic or nuclear (48), the collagen prolyl hydroxylase is a soluble unperturbed. Indeed, previous studies have shown that this peptide is endoplasmic reticulum luminal protein (47). However, like HIF prolyl not a substrate for recombinant collagen prolyl hydroxylases (30). Phy- 4-hydroxylases, collagen hydroxylases employ ferrous ions, 2-oxoglut- logenetic conservation (from worms to humans) of the critical residues arate, molecular oxygen, and ascorbate as cofactors. Thus by disrupting from HIF prolyl 4-hydroxylases that serve as ligands for Fe and 2-oxo- one or more of these cofactors, DFO, 3,4-DHB, and compound A could glutarate suggested that a peptide deemed to be a potent and effective substrate of all three known human HIF prolyl 4-hydroxylases would TABLE ONE also act to competitively inhibit HIF prolyl 4-hydroxylases expressed in Total zinc and iron levels in cultured cortical neurons nontreated rat cortical neurons (46). Our HIF/ODD/wt peptide and a peptide con- (control) or treated with DFO (100 M) 3,4-DHB (10 M), or compound trol with mutation of both prolines to alanine (HIF/ODD/mut; A (40 M) measured by ICP-MS DDLEMLAAYIAMDDDFQL) were rendered cell-permeant by fusing Values are shown as mean  S.E. percent relative to control levels that were arbitrarily designated as 100%. Levels of individual metals in a particular cul- each of these peptides to the cell membrane transduction domain of the ture were normalized to protein. human immunodeficiency virus, type 1 (HIV-1), Tat protein Control DFO 3,4-DHB Compound A (YGRKKRRQRR) to obtain two 30-amino acid peptides Tat-HIF/ ODD/wt and Tat-HIF/ODD/mut (49). To verify that the addition of Tat Zinc (%) 100 93.6  2.8 120.3  2.9 110.7  2.2 a a Iron (%) 100 68.8  9.5 93.3  3.1 70.04  2.4 p  0.05. 3 A. Siddiq and R. Ratan, unpublished observations. FIGURE 3. A cell-permeant peptide inhibitor of the HIF prolyl 4-hydroxylases, but not a mutant control, is a substrate for purified, recombinant HIF prolyl 4-hydroxylases 1, 2, and 3 isoforms. A, Tat-HIF/wt peptide (100 M, composed of the oxygen-dependent domain, including the C-ter- minal hydroxylation site of human HIF-1 at pro- line 564 (DDLDEMLAPYIPMDDDFQL; boldface P indicates proline 564) linked to a Tat protein trans- duction domain YGKKRRQRRR) but not a corre- sponding peptide with the C-terminal proline hydroxylation site of HIF-1-mutated (Tat-HIF/ mut, DLDLEMLAAYIAMDDDFQL; boldface A indi- cates the mutation site) acts as a substrate for the HIF prolyl 4-hydroxylases and releases [ C]CO from -ketoglutarate. DLD 19 HIF peptide, com- posed of the C-terminal hydroxylation site of human HIF-1 at proline 564 without the TAT sequence was used as a positive control. B, repre- sentative low magnification image (40) of corti- cal neurons treated with a FITC-labeled Tat-HIF/wt peptide (100 M)(green fluorescence) and the DNA intercalating dye, DAPI, to label nuclei (blue fluo- rescence). Note all blue nuclei have green fluores- cence in the cytoplasm indicating a high efficiency of transduction of the peptide into cells. C, a rep- resentative high magnification image (60) of cortical neurons treated with FITC-Tat-HIF/wt pep- tide (100 M) and labeled with the DNA intercalat- ing dye DAPI. DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41737 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury FIGURE 4. The peptide inhibitor of the HIF prolyl 4-hydroxylases, but not a mutant control, sta- bilizes HIF protein in primary cortical neurons. A, Tat-HIF/mut peptide (f and n), Tat-HIF/wt pep- tide (100 M)(g and o), DFO (100 M)(h and p), or vehicle control (e and m) were added to the bath- ing medium of cortical neurons. Panel 1 represents lower magnification (20), and panel 2 represents higher magnification (60). Cells were fixed with 4% paraformaldehyde for 15 min and immuno- stained for HIF-1 (red) as described under “Exper- imental Procedures.” DAPI staining was per- formed to label nuclei (blue)(a– d, 20; and i–l, 60). B, immunoblot analysis with an antibody to HIF-1 in cell lysates from nontreated (lane 1)or after treatment with Tat-HIF/mut peptide (100 M, lane 2), Tat-HIF/wt peptide (20 M, lane 3; 100 M, lane 4), or DFO (100 M, lane 5, positive control). to the 19-amino acid HIF peptide does not alter its specificity for the its cell bodies and processes reflecting the uptake of the Tat-HIF/ recombinant HIF prolyl 4-hydroxylases, we compared activities of the ODD/wt FITC peptide (Fig. 3, B and C). Similar high efficiency trans- Tat-HIF/ODD/wt, Tat-HIF/ODD/mut, and the native HIF/ODD/wt. duction was observed when cortical neuronal cultures were incubated We employed a commonly used assay for the HIF prolyl 4-hydroxylase with the Tat-HIF/ODD/mut FITC peptide (not shown). In parallel, via- that is based on the hydroxylation of a synthetic substrate and the asso- bility measurements verified that the Tat-HIF/ODD peptides do not ciated measurement of the radioactivity of CO formed during the negatively alter basal viability in our cortical neuronal cultures (not hydroxylation coupled decarboxylation of 2-oxo[1- C]glutarate. This shown). assay confirmed that the Tat-HIF/ODD/wt peptide but not the Tat- To determine whether the Tat-HIF/ODD/wt peptide stabilizes HIF/ODD/mut is a substrate with similar activity to the isoforms of HIF HIF-1 in cortical neurons, we performed immunocytochemical stain- prolyl 4-hydroxylases as a standard synthetic 19-amino acid peptide ing with an antibody to HIF-1 (Fig. 4, e–h,20; m–p,60) using in all with a sequence around the C-terminal hydroxylation site of HIF-1 cases a rhodamine-labeled secondary antibody that fluoresces red. In (DLD 19 HIF peptide (HIF/ODD/wt), Fig. 3A). parallel, the same cultures were treated with the nuclear stain DAPI, To evaluate the ability of our Tat-HIF/ODD peptides to be delivered which fluoresces blue (Molecular Probes, Eugene, OR) (Fig. 4, a–d, to cortical neurons efficiently and without toxicity, we conjugated the 20; i–l, 60). The HIF-1 antibody fluorescence confocal images chromophore fluorescein isothiocyanate to the Tat-HIF/ODD/wt pep- (20) confirmed that the Tat-HIF/ODD/wt peptide increases HIF-1 tide. Tat-HIF/ODD/wt FITC peptides (10 M) were bath applied to rat staining in cortical neurons (Fig. 4g), but the Tat-HIF/ODD/mut pep- cortical neuronal cultures from embryonic day 17 rat embryos. These tide does not (Fig. 4f). Addition of vehicle was used as a negative control cultures are 90% neurons after 1 day in vitro (35). The balance of the (Fig. 4e), and DFO was used as a positive control (Fig. 4h). High magni- cells in the culture is glial in origin. Twenty four hours after addition of fication confocal microscopic images of HIF- antibody fluorescence the peptide to the bathing medium, the level of intracellular accumula- from the nuclei of cortical neurons (60) (Fig. 4o) and immunoblot tion of the peptide was monitored by fluorescence microscopy. In every analysis of nuclear extracts using the same HIF-1 antibody (Fig. 4B, field examined ( 10), each cell nucleus, identified by DNA intercalating lanes 3 and 4) verified that the increases in HIF-1 levels induced by chromophore DAPI, was found to be associated with fluorescein label in Tat-HIF/ODD/wt are cytoplasmic and nuclear. 41738 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury To confirm that stabilization of HIF-1 is associated with activation of HIF-dependent gene expression, we monitored the levels of a lucif- erase reporter gene that is regulated by a hypoxia-response element found in the enolase gene promoter. As expected, we found that addi- tion of Tat-HIF/ODD/wt peptide significantly enhanced levels of a hypoxia-response element-regulated reporter gene in a concentration- dependent manner (not shown). The level of induction of the HIF-de- pendent reporter gene by 100 M Tat-HIF/ODD/wt peptide was similar to that stimulated by 40 M compound A, a low molecular weight inhib- itor of the HIF prolyl 4-hydroxylase (Fig. 5A). By contrast, 100 M Tat- HIF/ODD/mut peptide with proline 564 and its adjacent proline mutated to alanine did not induce the HIF-dependent reporter gene. Because the unhydroxylatable peptide did not increase target gene expression, it does not likely serve as a competitive inhibitor of the HIF prolyl 4-hydroxylases. To verify that the changes in the HIF-dependent reporter gene reflect changes in endogenous HIF-dependent genes, we waf1/cip1 monitored enolase, p21 , and VEGF by RT-PCR. -Actin was measured as a control (Fig. 5B). As expected, the Tat-HIF/ODD/wt waf1/cip1 peptide increased levels of enolase, p21 , and VEGF message but did not alter levels of -actin. Accordingly, we found that the Tat-HIF/ waf1/cip1 ODD/mut did not affect levels of p21 or -actin and actually diminished the basal levels of enolase and VEGF, two known HIF-de- pendent genes. Real time PCR was utilized to examine quantitatively the difference in message levels between the HIF/ODD/wt- and HIF/ODD/ mut-treated neurons. These studies confirmed the ability of the wild type but not the mutant peptide to increase HIF-dependent gene expression (VEGF, 1.46  0.218-fold induction (wt/mut); erythropoie- waf1/cip1 tin, 1.84  0.068-fold induction (wt/mut); p21 , 1.52  0.16-fold induction). Taken together, these results establish that the Tat-HIF/ FIGURE 5. A peptide inhibitor of the HIF prolyl 4-hydroxylases, but not a mutant ODD/wt peptide can be delivered to neurons and can enhance the tran- control, induces expression of established HIF-dependent genes. A, Tat-HIF/wt pep- tide (100 M) but not a corresponding peptide with the C-terminal proline hydroxylation scriptional activity of well characterized HIF target genes. By contrast, a site of HIF-1 mutated (Tat-HIF/mut, 100 M) significantly enhances the activity of a Tat/HIF/ODD/mut, in which the conserved proline hydroxylation site hypoxia-response element-driven reporter in cortical neurons (* corresponds to p 0.05 compared with control by paired t test). Note that the Tat-HIF/wt peptide induces HIF (proline 564) and an adjacent proline have been mutated to alanine, reporter activity to a similar extent as the low molecular weight inhibitor, compound A does not up-regulate an HIF-dependent reporter gene or endogenous (Comp-A) (40M). B, RT-PCR using specific primers to determine the mRNA level of estab- waf1/cip1 waf1/cip1 lished HIF target genes, enolase, p21 , or VEGF in vehicle-treated (control) or in HIF-regulated genes (e.g. enolase, p21 , or VEGF). neuronal cultures treated with Tat-HIF/wt peptide (100 M) and Tat-HIF/mut (100 M) To test whether Tat-HIF/ODD/wt, a peptide inhibitor of HIF prolyl overnight. -Actin was monitored as a housekeeping gene. 4-hydroxylases, can prevent oxidative neuronal death in cortical neu- FIGURE 6. A cell-permeant peptide inhibitor of the HIF prolyl 4-hydroxylases, but not a mutant control, prevents oxidative glutamate toxicity. The glutamate analog, HCA (5 mM) was added to cortical neurons (1 DIV) with or without Tat-HIF/wt peptide (30, 40, 50, 100, and 200 M), Tat-HIF/mut peptide (200 M), DFO (100 M), and 3,4-DHB (10 M). Twenty four hours later cell viability was determined using the MTT assay. Graph depicts mean  S.E. for three experiments performed in triplicate (* denotes p  0.05 from HCA-treated cultures by ANOVA and Student-Newman Keuls tests for control, Tat-HIF/wt, Tat-HIF/mut, DFO and 3,4-DHB). DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41739 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury FIGURE 7. Structurally diverse, low molecular weight prolyl 4-hydroxylase inhibitors stabilize HIF protein, increase the expression of established HIF-regulated genes, and reduce infarct volume induced by permanent focal ischemia in vivo. Three rats were treated in parallel with compound A (lanes 1, 3, and 5) with a dose of 100 mg/kg of body weight (by gavage) 6 h prior to MCAO. As controls, three rats (lanes 2, 4, and 6) received equal volume of vehicle (0.5% carboxymethylcellulose) in a blinded study. Whole brain lysates 41740 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury TABLE TWO Physiologic data obtained from both control and treated animal groups are represented as means  S.E. The following abbreviations are used: Hct, hematocrit; Glu, blood glucose; MABP, mean arterial blood pressure; HR, heart rate. Animals were maintained at 36.5  0.5 °C (rectal temperature). There was no statistically significant difference between control and treated groups in both pre- and post-MCAO categories. Pre-ischemia Post-ischemia Physiological parameters Control group Compound A group Control group Compound A group Compound A pH 7.35  0.05 7.36  0.02 7.32  0.03 7.34  0.03 PCO (mm Hg) 33 535 430 232  3 PO (mm Hg) 99  12 100  10 95  10 100  10 Hct (%) 40 440 338 440  4 MABP (mm Hg) 91  12 89  18 95  12 88  23 3,4-DHB pH 7.35  0.05 7.33  0.06 7.35  0.06 7.33  0.06 PCO (mm Hg) 34 234  2 34.2 235  2 PO (mm Hg) 93 395 493 396  3 Hct (%) 42 143 241 243  2 MABP (mm Hg) 93 392 394 493  4 rons induced by the glutamate analog, HCA, we applied the peptides concentrations of 3,4-DHB and compound A, and we determined that 24 h prior to the addition of HCA. Addition of Tat-HIF/ODD/wt pep- administration of 100 mg/kg of compound A (via gavage) or 160 mg/kg tide but not its mutant control (Tat-HIF/ODD/mut) protected cortical of 3,4-DHB (via intraperitoneal administration) was effective at stabi- neurons from oxidative glutamate (glutathione depletion-induced) tox- lizing the HIF protein and inducing an increase in some HIF-dependent icity in a concentration-dependent manner as assessed by MTT assay genes, including erythropoietin in lysates from brains of individual vehi- (Fig. 6). The morphology of the neurons that were protected from oxi- cle-treated or HIF prolyl 4-hydroxylase inhibitor-treated brains (Fig. 7, dative glutamate toxicity by Tat-HIF/ODD/wt was indistinguishable A, B, and E). The stabilization of HIF in brain by compound A was from the morphology of control neurons as assessed by phase contrast observed within 3 h (Fig. 7A, whereas the stabilization of HIF by 3,4- microscopy or calcein/ethidium homodimer staining (not shown). DHB was observed within6hofadministration (Fig. 7E). Inhibition of These results suggest that inhibition of HIF prolyl 4-hydroxylases is HIF prolyl 4-hydroxylases was associated with marked reduction in sufficient to prevent oxidative neuronal death in cortical neurons in infarct volume as measured histologically by TTC staining by com- vitro and argue that 3,4-DHB, DFO, and compound A are influencing pound A (67%, Fig. 7, C and D) or by 3,4-DHB (46%, Fig. 7F). Protection cell viability by inhibiting HIF prolyl 4-hydroxylases rather than a dis- by these agents could not be attributed to the effects of these agents on tinct 2-oxoglutarate-dependent dioxygenase. body temperature, plasma glucose levels, plasma pH, or blood flow Low Molecular Weight Inhibitors of HIF Prolyl 4-Hydroxylases (TABLE TWO, compound A and 3,4-DHB). Together, these findings Reduce Neuronal Damage Induced by Permanent Focal Ischemia suggest that HIF prolyl 4-hydroxlyases are a target for protection in in Vivo—Oxidative stress is an established mediator of neuronal loss in stroke and suggest that the salutary effects of these treatments in cere- cerebral ischemia (50). Prior studies have also demonstrated that the bral ischemia are due in part to HIF prolyl 4-hydroxylase inhibition. putative antioxidant DFO can ameliorate metabolic failure in dogs sub- DISCUSSION jected to cerebral ischemia or induce ischemic tolerance in neonatal or adult rodents (51–53). Some of these studies have suggested that the Prior studies from our laboratory correlated the protective effects of effects of DFO may be related to HIF prolyl 4-hydroxylase inhibition. iron chelators in an in vitro model of oxidative stress with their ability to However, these studies did not provide any experimental support for stabilize HIF-1 and activate HIF-dependent genes (28). These findings the notion that HIF prolyl 4-hydroxylases are a target for ischemic neu- supported the hypothesis that iron chelators not only prevent neuronal roprotection by DFO. The protective effects of HIF prolyl 4-hydroxylase injury by inhibiting hydroxyl radical formation via Fenton chemistry but inhibitors in preventing oxidative neuronal death in vitro herein (Figs. also via inhibition of the iron-dependent HIF prolyl 4-hydroxylases that 1–6) provided additional rationale for testing these agents in an in vivo regulate HIF stability. In this study we utilize low molecular weight or model of cerebral ischemia. We therefore examined the ability of 3,4- peptide inhibitors of the HIF prolyl 4-hydroxylases in vitro and in vivo to DHB or compound A to prevent neuronal damage in response to per- define this family of enzymes as targets for neuroprotection in the cen- manent focal ischemia. In this model, a filament is placed permanently tral nervous system, and also to suggest that the inhibition of these into the distal aspect of the internal carotid artery to occlude the origin enzymes is one way in which these treatments exert their salutary effects of the middle cerebral artery (43). After 24 h of occlusion, there is sig- against oxidative stress. The HIF prolyl 4-hydroxylases require iron, nificant infarction in the hemisphere ipsilateral to the occlusion; few 2-oxoglutarate, oxygen, and ascorbate as cofactors. Two of the low animals survive beyond this time point. We first examined a range of molecular weight agents we tested, DFO and compound A, appear to from individual animals in the treated (n 9) (lanes 1, 3, and 5) and control group (n 5) (lanes 2, 4, and 6) were separated using gel electrophoresis and immunodetected using an antibody against HIF1- (A) and erythropoietin (Epo)(B). Whole brain lysates of 1% hypoxia-exposed animals were used as a positive control for HIF-1 and erythropoietin (lane 7). -Actin was monitored as a housekeeping gene. C, histologically measured infarct volume of brains of compound A (Comp-A)(n  9) and vehicle-treated animals (n  5) after TTC staining. The values are shown as % infarct volume relative to total brain volume (*  p  0.05 by paired t test). D, representative pictures of vehicle-treated (left) and compound A-treated (right) brains showing infarctions. E, animals were treated with 3,4-DHB (n 6) (lanes 2 and 3) with a dose of 180 mg/kg of body weight (intraperitoneally) 6 h prior to MCAO. Control animals (n 5) (lanes 1 and 4) received equal volume of vehicle (ethanol, intraperitoneally) in a blinded study. Whole brain lysates were separated using gel electrophoresis and immunodetected using an antibody against HIF1-. -Actin was monitored as a housekeeping gene. F, histologically measured infarct volume of brains of 3,4-DHB (n  6) and vehicle-treated animals (n  5) after TTC staining. The values are shown as % infarct volume relative to total brain volume (  p  0.05 by paired t test). DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41741 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury inhibit prolyl 4-hydroxylases by removing the iron cofactor. Indeed, rat HIF prolyl 4-hydroxylase. Future studies will determine which one both agents reduced total iron in embryonic rat cortical neurons by 30% or more of the described rat HIF prolyl 4-hydroxylases needs to be inhibited in order to protect rat embryonic cortical neurons from oxi- after 24 h (TABLE ONE). Despite these measurable changes in total iron, we could not detect statistically significant changes in IRP-IRE dative stress. binding or calcein fluorescence (independent measures of the labile iron What is the mechanism of neuroprotection by inhibitors of the HIF prolyl 4-hydroxylases? These enzymes stabilize HIF leading to tran- pool) in cortical neuronal cultures treated with DFO or compound A scriptional up-regulation of a cassette of genes involved in hypoxic (not shown). As some of the proteins established to play a role in iron homeostasis are regulated by prolyl hydroxylases (e.g.IRP2 (50)) or HIF adaptation, including VEGF and erythropoietin (56, 58). The estab- lished ability of these peptide growth factors to prevent neuronal injury (e.g. transferrin receptor (51)), the lack of change in IRP-IRE binding and death in vitro and in vivo makes them attractive candidates to medi- raises the intriguing possibility that prolyl hydroxylase inhibition is a ate the neuroprotective effects of HIF prolyl 4-hydroxylase inhibition. compensatory response to cellular iron deficiency that permits the labile Additional studies are needed to determine whether HIF stabilization is iron levels to remain normal despite the presence of a chelator that necessary for the protective effects of HIF prolyl 4-hydroxylase inhibi- reduces total iron. Given the oral availability of compound A (DFO tion. Indeed, growing awareness of other substrates such as IRP2 (59) cannot be taken orally) and its efficacy against oxidative stress in vitro and RNA polymerase (60), whose stability can be regulated by the HIF and permanent focal ischemia in vivo, the findings also suggest that this prolyl 4-hydroxylases, suggests other plausible schemes in addition to or molecule may have advantages over DFO as a therapeutic agent for exclusive of HIF that may account for protection by HIF prolyl 4-hy- acute and chronic neurological conditions where DFO has been shown droxylase inhibition. Even if HIF-1 stabilization is not necessary for the to be effective (24, 26). protection against oxidative stress or ischemia provided by low molec- Prior studies in non-neuronal systems have suggested that 3,4-DHB ular weight or peptide inhibitors of the HIF prolyl 4-hydroxylases, also stabilizes HIF via its ability to bind iron (54, 55). However, several HIF-1 stabilization appears to be a good marker of HIF prolyl 4-hy- observations herein argue against this possibility in embryonic rat cor- droxylase inhibition. tical neurons. First, the concentrations of 3,4-DHB (10 M) required to The ability of inhibitors of HIF prolyl 4-hydroxylases to prevent neu- stabilize HIF and protect cortical neurons were not found to bind iron in ronal injury when given prior to permanent focal ischemia suggests that one study where 3,4-DHB was reported to act as an iron chelator (54). they be considered as neuroprotective agents in clinical situations Second, in vitro studies using recombinant collagen prolyl 4-hydroxyl- where the imminent risk of ischemic or oxidative neuronal injury is ase showed that 3,4-DHB inhibitory actions occur via the displacement high. Coronary bypass surgery, abdominal aortic aneurysm repair, and of 2-oxoglutarate and ascorbate but not iron (45). Third, total iron levels acute anterior wall myocardial infarction with a thrombus are examples and IRP-IRE binding in 3,4-DHB-treated neurons were unchanged as of clinical situations where increasing the threshold for oxidative or compared with controls (TABLE ONE and data not shown). Despite the hypoxic-ischemic brain injury is desirable. Studies are currently under- inability of 3,4-DHB to bind iron, it was still capable of stabilizing HIF, way to define whether HIF prolyl 4-hydroxylase inhibitors are effective activating HIF-dependent gene expression, and protecting cortical neu- in preventing neuronal damage when given after the onset of ischemia rons from oxidative stress in vitro and cerebral ischemia in vivo (Figs. 1 or during the period of stroke recovery. The ability of HIF prolyl 4-hy- and 4). Thus, inhibition of the HIF prolyl 4-hydroxylases is the likely droxylase inhibitors to induce expression of cellular, local, and systemic relevant consequence of treatment with 3,4-DHB, compound A, and homeostatic responses to ischemia suggests that these agents may be DFO. For chronic neurodegenerative conditions where daily use of an useful in post-event treatment in addition to the preventative treatment HIF prolyl 4-hydroxylase inhibitor that chelates iron may result in rest- described herein. less legs syndrome (56) or anemia (57), the use of an HIF prolyl 4-hy- droxylase inhibitor that targets 2-oxoglutarate or ascorbate rather than Acknowledgments—We thank Robert Freeman and Gregg Semenza for provid- iron has potential advantages. ing the HRE reporters. We thank Brett Langley, Philipp Lange, Kyungsun Suh, The 2-oxoglutarate-dependent prolyl 4-hydroxylases are a family of and JoAnn Gensert for their helpful discussions. We are especially thankful to iron- and ascorbate-dependent enzymes that include not only HIF Johanna Myllyharju for input and advice. We thank Yixin Ben for excellent prolyl 4-hydroxylases but collagen prolyl 4-hydroxylases as well. DFO, technical assistance and Wayne Kleinman for excellent editorial assistance. 3,4-DHB, and compound A would be expected to inhibit both subfam- We also acknowledge the support of the Spinal Cord Injury Research Board of the New York State Department of Health in this study. ilies of enzymes, although recent studies suggest the K of these agents for collagen prolyl 4-hydroxylases and HIF prolyl 4-hydroxylases may be distinct (37). In an attempt to develop a more specific inhibitor of the REFERENCES HIF prolyl 4-hydroxylases, we conjugated a 19-amino acid C-terminal 1. Zecca, L., Stroppolo, A., Gatti, A., Tampellini, D., Toscani, M., Gallorini, M., Giaveri, portion of the oxygen-dependent degradation domain of HIF to an G., Arosio, P., Santambrogio, P., Fariello, R. G., Karatekin, E., Kleinman, M. H., Turro, N., Hornykiewicz, O., and Zucca, F. A. (2004) Proc. Natl. Acad. Sci. U. S. A. 101, 11-amino acid protein transduction domain from the HIV-tat protein. 9843–9848 Recent studies have shown that this 19-amino acid peptide (ODD/wt) 2. Kress, G. J., Dineley, K. E., and Reynolds, I. J. (2002) J. Neurosci. 22, 5848–5855 but not shorter peptides bind all three human isoforms of the HIF-prolyl 3. Connor, J. R., Menzies, S. L., Burdo, J. R., and Boyer, P. J. (2001) Pediatr. Neurol. 25, 4-hydroxylases with high affinity and are efficiently hydroxylated by 118–129 4. Gerlach, M., Ben-Shachar, D., Riederer, P., and Youdim, M. B. (1994) J. Neurochem. these enzymes (37). We verified that our Tat-HIF/ODD/wt peptide is a 63, 793–807 substrate with similar affinity for the recombinant HIF prolyl 4-hy- 5. Thompson, K. J., Shoham, S., and Connor, J. R. (2001) Brain Res. Bull. 55, 155–164 droxylases as HIF/ODD/wt (Fig. 3A). We also show that intracellular 6. Castellani, R. J., Siedlak, S. L., Perry, G., and Smith, M. A. (2000) Acta Neuropathol. delivery of Tat-HIF/ODD/wt (but not a control peptide with the proline 100, 111–114 7. Dexter, D. T., Wells, F. R., Lees, A. J., Agid, F., Agid, Y., Jenner, P., and Marsden, C. D. hydroxylation sites mutated to alanine) stabilizes HIF-1 (Fig. 4), effec- (1989) J. Neurochem. 52, 1830–1836 tively activates HIF-dependent gene expression (Fig. 5), and protects 8. Kienzl, E., Jellinger, K., Stachelberger, H., and Linert, W. (1999) Life Sci. 65, embryonic rat cortical neurons from oxidative stress (Fig. 6). These 1973–1976 findings suggest that Tat-HIF/ODD/wt is also an effective inhibitor of 9. Sofic, E., Riederer, P., Heinsen, H., Beckmann, H., Reynolds, G. P., Hebenstreit, G., 41742 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 280 • NUMBER 50 •DECEMBER 16, 2005 HIF Prolyl 4-Hydroxylase Inhibition Prevents Neuronal Injury and Youdim, M. B. (1988) J. Neural Transm. 74, 199–205 37. Hirsila, M., Koivunen, P., Gunzler, V., Kivirikko, K. I., and Myllyharju, J. (2003) J. Biol. 10. Thompson, C. A. (2001) Am. J. Health Syst. Pharm. 58, 649–650 Chem. 278, 30772–30780 11. Smith, M. A., Rottkamp, C. A., Nunomura, A., Raina, A. K., and Perry, G. (2000) 38. Breuer, W., Epsztejn, S., and Cabantchik, Z. I. (1996) FEBS Lett. 382, 304–308 Biochim. Biophys. Acta 1502, 139–144 39. Bettany, A. J., Eisenstein, R. S., and Munro, H. N. (1992) J. Biol. Chem. 267, 12. Mazur-Kolecka, B., Frackowiak, J., Le Vine, H., III, Haske, T., and Wisniewski, H. M. 16531–16537 (1997) Brain Res. 760, 255–260 40. Bederson, J. B., Pitts, L. H., Germano, S. M., Nishimura, M. C., Davis, R. L., and 13. Good, P. F., Perl, D. P., Bierer, L. M., and Schmeidler, J. (1992) Ann. Neurol. 31, Bartkowski, H. M. (1986) Stroke 17, 1304–1308 286–292 41. Belayev, L., Alonso, O. F., Busto, R., Zhao, W., and Ginsberg, M. D. (1996) Stroke 27, 14. Deibel, M. A., Ehmann, W. D., and Markesbery, W. R. (1996) J. Neurol. Sci. 143, 1616–1623 137–142 42. Swanson, R. A., Morton, M. T., Tsao-Wu, G., Savalos, R. A., Davidson, C., and Sharp, 15. Connor, J. R., Menzies, S. L., St Martin, S. M., and Mufson, E. J. (1992) J. Neurosci. Res. F. R. (1990) J. Cereb. Blood Flow Metab. 10, 290–293 31, 75–83 43. Ayoub, I. A., Lee, E. J., Ogilvy, C. S., Beal, M. F., and Maynard, K. I. (1999) Neurosci. 16. Schapira, A. H. (1999) Biochim. Biophys. Acta 1410, 159–170 Lett. 259, 21–24 17. Waldvogel, D., van Gelderen, P., and Hallett, M. (1999) Ann. Neurol. 46, 123–125 44. Ratan, R. R., Ryu, H., Lee, J., Mwidau, A., and Neve, R. L. (2002) Methods Enzymol. 18. Aisen, P. (1994) Adv. Exp. Med. Biol. 356, 31–40 352, 183–190 19. Bralet, J., Schreiber, L., and Bouvier, C. (1992) Biochem. Pharmacol. 43, 979–983 45. Majamaa, K., Gunzler, V., Hanauske-Abel, H. M., Myllyla, R., and Kivirikko, K. I. 20. Lipscomb, D. C., Gorman, L. G., Traystman, R. J., and Hurn, P. D. (1998) Stroke 29, (1986) J. Biol. Chem. 261, 7819–7823 487–493 46. Epstein, A. C., Gleadle, J. M., McNeill, L. A., Hewitson, K. S., O’Rourke, J., Mole, D. R., 21. Hurn, P. D., Koehler, R. C., Blizzard, K. K., and Traystman, R. J. (1995) Stroke 26, Mukherji, M., Metzen, E., Wilson, M. I., Dhanda, A., Tian, Y. M., Masson, N., Ham- 688–694 ilton, D. L., Jaakkola, P., Barstead, R., Hodgkin, J., Maxwell, P. H., Pugh, C. W., 22. Kaur, D., and Andersen, J. K. (2002) Aging Cell 1, 17–21 Schofield, C. J., and Ratcliffe, P. J. (2001) Cell 107, 43–54 23. Bowern, N., Ramshaw, I. A., Clark, I. A., and Doherty, P. C. (1984) J. Exp. Med. 160, 47. Myllyharju, J. (2003) Matrix Biol. 22, 15–24 1532–1543 48. Metzen, E., Berchner-Pfannschmidt, U., Stengel, P., Marxsen, J. H., Stolze, I., Klinger, 24. McLachlan, D. R., Smith, W. L., and Kruck, T. P. (1993) Ther. Drug Monit. 15, M., Huang, W. Q., Wotzlaw, C., Hellwig-Burgel, T., Jelkmann, W., Acker, H., and 602–607 Fandrey, J. (2003) J. Cell Sci. 116, 1319–1326 25. Ritchie, C. W., Bush, A. I., Mackinnon, A., Macfarlane, S., Mastwyk, M., MacGregor, 49. Aarts, M., Liu, Y., Liu, L., Besshoh, S., Arundine, M., Gurd, J. W., Wang, Y. T., Salter, L., Kiers, L., Cherny, R., Li, Q. X., Tammer, A., Carrington, D., Mavros, C., Volitakis, M. W., and Tymianski, M. (2002) Science 298, 846–850 I., Xilinas, M., Ames, D., Davis, S., Beyreuther, K., Tanzi, R. E., and Masters, C. L. 50. Liang, D., Dawson, T. M., and Dawson, V. L. (2004) Curr. Mol. Med. 4, 207–225 (2003) Arch. Neurol. 60, 1685–1691 51. Bergeron, M., Gidday, J. M., Yu, A. Y., Semenza, G. L., Ferriero, D. M., and Sharp, F. R. 26. Kaur, D., Yantiri, F., Rajagopalan, S., Kumar, J., Mo, J. Q., Boonplueang, R., Viswanath, (2000) Ann. Neurol. 48, 285–296 V., Jacobs, R., Yang, L., Beal, M. F., DiMonte, D., Volitaskis, I., Ellerby, L., Cherny, 52. Davis, S., Helfaer, M. A., Traystman, R. J., and Hurn, P. D. (1997) Stroke 28, 198–204 R. A., Bush, A. I., and Andersen, J. K. (2003) Neuron 37, 899–909 53. Prass, K., Ruscher, K., Karsch, M., Isaev, N., Megow, D., Priller, J., Scharff, A., Dirnagl, 27. Winterbourn, C. C. (1995) Toxicol. Lett. 82, 969–974 U., and Meisel, A. (2002) J. Cereb. Blood Flow Metab. 22, 520–525 28. Zaman, K., Ryu, H., Hall, D., O’Donovan, K., Lin, K. I., Miller, M. P., Marquis, J. C., 54. Wang, F., Sekine, H., Kikuchi, Y., Takasaki, C., Miura, C., Heiwa, O., Shuin, T., Fujii- Baraban, J. M., Semenza, G. L., and Ratan, R. R. (1999) J. Neurosci. 19, 9821–9830 Kuriyama, Y., and Sogawa, K. (2002) Biochem. Biophys. Res. Commun. 295, 657–662 29. Willam, C., Masson, N., Tian, Y. M., Mahmood, S. A., Wilson, M. I., Bicknell, R., 55. Warnecke, C., Griethe, W., Weidemann, A., Jurgensen, J. S., Willam, C., Bachmann, Eckardt, K. U., Maxwell, P. H., Ratcliffe, P. J., and Pugh, C. W. (2002) Proc. Natl. Acad. S., Ivashchenko, Y., Wagner, I., Frei, U., Wiesener, M., and Eckardt, K. U. (2003) Sci. U. S. A. 99, 10423–10428 FASEB J. 17, 1186–1188 30. Ivan, M., Kondo, K., Yang, H., Kim, W., Valiando, J., Ohh, M., Salic, A., Asara, J. M., 56. Wang, X., Zhu, C., Gerwien, J. G., Schrattenholz, A., Sandberg, M., Leist, M., and Lane, W. S., and Kaelin, W. G., Jr. (2001) Science 292, 464–468 Blomgren, K. (2004) J. Neurochem. 91, 900–910 31. Jaakkola, P., Mole, D. R., Tian, Y. M., Wilson, M. I., Gielbert, J., Gaskell, S. J., Krieg- 57. White, K. C. (2005) Pediatrics 115, 315–320 sheim, A., Hebestreit, H. F., Mukherji, M., Schofield, C. J., Maxwell, P. H., Pugh, C. W., 58. Shi, Y. H., Wang, Y. X., Bingle, L., Gong, L. H., Heng, W. J., Li, Y., and Fang, W. G. and Ratcliffe, P. J. (2001) Science 292, 468–472 (2005) J. Pathol. 205, 530–536 32. Yu, F., White, S. B., Zhao, Q., and Lee, F. S. (2001) Cancer Res. 61, 4136–4142 59. Hanson, E. S., Rawlins, M. L., and Leibold, E. A. (2003) J. Biol. Chem. 278, 33. Semenza, G. L., Agani, F., Feldser, D., Iyer, N., Kotch, L., Laughner, E., and Yu, A. 40337–40342 (2000) Adv. Exp. Med. Biol. 475, 123–130 60. Kuznetsova, A. V., Meller, J., Schnell, P. O., Nash, J. A., Ignacak, M. L., Sanchez, Y., 34. Bruick, R. K. (2003) Genes Dev. 17, 2614–2623 Conaway, J. W., Conaway, R. C., and Czyzyk-Krzeska, M. F. (2003) Proc. Natl. Acad. 35. Murphy, T. H., Schnaar, R. L., and Coyle, J. T. (1990) FASEB J. 4, 1624–1633 36. Mosmann, T. (1983) J. Immunol. Methods 65, 55–63 Sci. U. S. A. 100, 2706–2711 DECEMBER 16, 2005• VOLUME 280 • NUMBER 50 JOURNAL OF BIOLOGICAL CHEMISTRY 41743

Journal

Journal of Biological Chemistry – American Society for Biochemistry and Molecular Biology

Published: Dec 5, 2005

There are no references for this article.