Determination of the Mechanism of Free Radical in ... - Semantic Scholar

3 downloads 0 Views 807KB Size Report
human cells and tissues. The enzyme xanthine oxidase can be an important source of radical generation; how- ever, it has been reported that this enzyme may ...
Vol. 269, No 39, Issue of September 30,pp 24156-24162, 1994 Printed in U S A .

Determination of the Mechanismof Free Radical Generation in Human Aortic Endothelial Cells Exposed to Anoxia and Reoxygenation* (Received for publication, May 9, 1994)

Jay L. Zweierl, Raymond BroderickO, Periannan Kuppusamy, Susan Thompson-Gorman, and Gerard A. LuttyY From the EPR Laboratories, Department of Medicine, and nThe Wilmer Ophthalmological Institute, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21224 and the $Department of Pharmacology, University of Maryland School of Medicine, Baltimore, Maryland 21201

Endothelial cell-derived oxygen free radicals are im- zymes or free radical-scavengingdrugs can prevent reperfusion portant mediators of postischemic injury; however, the injury and enhance the recovery of postischemic function (1,2). mechanisms that trigger this radical generation are not These studies have provided indirect evidence of free radical known, and it is not known if this process can occur in generation in postischemic tissues. More recently, free radical human cells and tissues. The enzyme xanthine oxidase generation has been measured in postischemic tissues using can be an important source of radical generation; how- electron paramagnetic resonance (EPR)’ spectroscopy. Both diever, it has been reported that this enzyme may not be rect and spin trapping EPR techniques have demonstrated that present in human endothelium. To determine the pres- there is a burst of oxygen free radical generation afterpostisence and mechanisms of radical generation in human chemic reperfusion of the heart (2-11). It was further demonvascular endothelial cells subjected to anoxia and reoxy- strated that the administration of copper-zinc superoxide disgenation, electron paramagnetic resonance measurecontractile ments were performed oncultured human aortic endo- mutase, solely upon reperfusion, canprevent (5, 6, 9, 11). dysfunction and quench free radical generation thelial cells usingthe spin trap 5,5-dimethyl-l-pyrroline N-oxide (DMPO). These measurements were correlated Since superoxide dismutase would not have been expected to with cellular injury, xanthine oxidase activity, and alter- rapidly traverse the endothelial and myocyte membranes due ations in cellular nucleotides. Upon reoxygenation after to its 32-kDa molecular mass, this data suggested that signif60 min of anoxia, large DMPO-OH (aN= a, = 14.9 G) and icant free radical generation occurs at the cell surface of the smaller DMPO-R (aN= 15.8 G, a, = 22.8 G ) signals were endothelium or within the vascular lumen adjacent to the enseen. Superoxidedismutase totally quenched this radi- dothelium. Subsequently in culturedbovine aortic endothelial cal generation. The ferric iron chelator deferoxamine cells, it was demonstrated that endothelial cells subjected to prevented cell death and totally quenched the DMPO-R anoxia and reoxygenation generate a burst of superoxidesignal with a 40% decrease in theDMPO-OH signal. Xan- derived free radicals upon reoxygenation (12). thine oxidase was shown to be present in these cells and It hasbeen proposed that theenzyme xanthine oxidase may to be the primary source of free radicals. While the con- be a central mechanism of free radical generation in a variety centration of this enzyme did not change after anoxia, of postischemic cells and tissues (13). In particular, it was hythe concentration of its substrate, hypoxanthine, mark- pothesized that in ischemic tissues xanthine dehydrogenase, edly increased, resulting in increased free radical gen- which reduces NAD to NADH, is converted via proteolytic eration upon reoxygenation. Thus, reoxygenated human cleavage to xanthine oxidase, which reduces 0, to 0;.This vascular endothelial cells generate superoxidefree radi- increase in xanthineoxidase activity has been proposed t o trigcals, which further reactwith iron to form the reactive ger the generation of 0, upon reperfusion. During ischemia, hydroxyl radical, which in turn causes cell death. Xan- elevated concentrations of the substrates xanthine andhypoxthine oxidase wasthe primary source of radical generaanthine may occur due to the breakdown of ATP. It has been tion with this process triggered by the breakdown of demonstrated in the isolated rat heart that the enzyme and ATP to the substratehypoxanthine during anoxia. substrate are present and account for a burst of free radicals upon postischemic reperfusion (10).Recently, however, it was reported that xanthine oxidase may not be present at imporOxygen free radicals have been proposed to be central me- tant vascular sites in human tissues. Based on a failure to diators of the cellular injury that occurs upon postischemic detectenzymeactivity in heart tissue homogenates, it was reperfusion. Studies ina variety of tissues including the heart, postulated that xanthine oxidase is not present in the human lung, kidney, gastrointestinal tract, and brain have demon- heart and that thismechanism may occur only in animal spestrated that intravascular administration of antioxidant en- cies such as therat or cow, which have high concentrations of the ehzyme (14). This has been questioned by other investiga* This work was supported by National Institutes of Health Grants tors who, based on immunohistochemistry, asserted that the HL-17655, HL-38324,and HL-07227 and the American Heart Associa- enzyme is present and primarily located within the vascular tion. Thecosts of publication of this article were defrayed in part by the endothelium, and thus, activity might not be easily measured be hereby marked payment of page charges. This article must therefore in whole-heart tissue homogenates, which are mostly com“advertisement”in accordance with 18 U.S.C.Section1734solelyto indicate this fact. The abbreviationsused are: EPR, electron paramagnetic resonance; $ Recipient of an American Heart Association Established InvestigaN-oxide; HPLC, high performanceliqtor Award. To whom correspondence should be addressed: Johns Hop- DMPO, 5,5-dimethyl-l-pyrroline kins Asthma and Allergy Center, EPR Laboratories, Rm. LA-14, 5501 uid chromatography; HAEC, human aortic endothelial cell; PBS, phosphate-buffered saline. Hopkins Bayview Circle, Baltimore, MD 21224. Tel.: 410-550-0339.

24156

Mechanism of Radical Generation by Human Endothelial Cells prised of myocytes (15,16). Thus, it is not known if free radical generation occurs after ischemia in human vascular endothelium. In addition, many questions remain regarding the presence and control of mechanisms that cause this free radical generation. In particular, the presence and role of the enzyme xanthine oxidase in this process have been hotly contested. There has also been controversy regarding whether xanthine oxidase-mediated radical generation is enzyme- or substratelimited (17). In thismanuscript, we report studies that demonstrate that human vascular endothelial cells intrinsically generate oxygen free radicals when subjected to the stimuli of anoxia and reoxygenation in theabsence of metabolic substrates, conditions observed in ischemic and reperfused tissues. EPR spectroscopyin the presence of the spin trap DMPO was applied to measure the magnitude and mechanisms of radical generation in cultured human aortic endothelial cells. Experiments werealso performed to determine the presence of xanthine oxidase in these human vascular endothelial cells and whether this enzyme was a source of free radical generation. In addition, high performance liquid chromatography (HPLC) was applied to measure the alterations that occur in thecellular nucleotide pool during anoxia to determine if xanthine oxidase-mediated free radical generation is triggered after anoxia by the degradation of ATP to the substrates hypoxanthine and xanthine. MATERIALSANDMETHODS CellCulture-Human aortic endothelial cells (HAECs) were purchased from Clonetics Corp. (San Diego, CA).The lines were established by a modification of the method of Hoshi and McKeehan (18).Cells are provided as cryopreserved vials of a tertiary culture with 2% fetal bovine serum in the medium. These cells are positive for factor VIIIrelated antigen (von Willebrand's factor) and acetylated low density lipoprotein uptake and remain viable and responsive to growth factors up to passage 6.Cultures were maintained on endothelial cell growth medium (EGM-W, Clonetics)with 2% fetal bovine serum and 0.4ml of bovine brain extract in a 37 "C, 5% CO,, 95% air incubator. For the experimental studies, the cellmonolayerswere gently washed with 10 ml of phosphate-buffered saline (PBS), and then the cells wereharvested from the culture flasks by incubation with 2 ml of a 0.1% trypsin solution for 10 min followed byaddition of 8 ml of quench solution (minimum essential medium (I media with 10% fetal bovine serum, Life Technologies, Inc.).The cells were then centrifuged at 100 x g for 5 min, washed twice with 10 ml of PBS, and then suspended in the desired final volume of PBS. Cell counts were performed either with a Coulter counter (Coulter Electronics) or manually with a hemocytometer. Cell viability was assessed by trypan blueexclusion.Cellswere stained with 0.02% trypan blue in PBS, and cell counts of 100 cells were performed after 2 min using a Zeiss standard laboratory light microscope. ExogenousFree Radicul-generating Systems-Phorbol ester-activated polymorphonuclearleukocyteswere utilized as an exogenous known cellular source of superoxide free radical generation. These cells were purified from human blood by dextran gradient sedimentation as previously described (19, 20). The cells were activated with 200 ng/ml 12-0-tetradecanoylphorbol-13-acetate. Spin trapping measurements of free radical generation were performed in PBS with lo6 celldml in the presence of 50 m~ DMPO. An iron redox hydroxyl radical-generating system consisting of 1.0 mM H,O, and 20 p~ Fe3+-nitrilotriacetate(1:2) was utilized as previously described (21)in the presence of 50 m~ DMPO. EPR Spectroscopy and Spin napping-The spin trapping studies were performedusing the spin trap DMPO a t a final concentration of 50 m ~The . DMPO (>97%pure) was purchased from Aldrich and further purified by double distillation (12).Anaerobic cellular preparations were achieved by purging under a gentle stream of pure nitrogen gas, and reoxygenation was achieved by reexposure to air. EPR spectra were recorded in flat cells a t room temperature with a Bruker-IBM ER 300 spectrometer operating at X-band with 100 kHz modulation frequency and a TM 110 cavity. The microwave frequency and magnetic field were precisely measured using an EIP 575 microwave frequency counter and Bruker ER035M NMR gaussmeter. Spec-

24157

tral simulations were performed using simulation programs as previouslydescribed (22).Quantitation of the free radical signals was performed by comparing the double integral of the observed signal with that of a known concentration of the 2,2,6,6,-tetramethyl-l-piperidinyloxy free radical in aqueous solution as previously described (5). Assays ofxunthine Oxidase Activity-For measurements of cellular xanthine oxidase, cells were removed from the flasks by scraping using a rubber policeman instead o f trypsin. The pellets from twoT-150flasks were resuspended in 5 ml of 0.05 M potassium phosphate buffer, pH 7.8, containing 1 m~ phenylmethylsulfonyl fluorideand 10 m~ dithiothreitol, which prevents the in uitro conversionof xanthine dehydrogenase to xanthine oxidase (10).The cell suspension was then sonicated four times for 15 s using a Microson ultrasonic cell disrupter (HeatSystems Ultrasonics) at output level 12. The sonicated suspension was then centrifuged at 105,000 x g for 60 min at 4 "C, and the supernatantwas passed through a Sephadex G-25column equilibrated with the phosphate buffer containing 1 mM phenylmethylsulfonyl fluorideand 10 m~ dithiothreitol. The eluent was spectrophotometricallyassayed at 295 nm for uric acid production in the presence of 60 p~ xanthine using a Hewlett-Packard 8452A diode array spectrophotometer as previously described (10).The reaction mixture contained 0.8ml of eluent and 0.2 ml of 0.3 mM xanthine for measurements of xanthine oxidase activity. For measurement of total xanthine oxidase and xanthine dehydrogenase activity, 12 111of 0.6m~ NAD+was also present. Enzyme activity is expressed in milliunits/g protein, where 1unit of activity equals 1 pmol of substrate converted to uric acidlmin. HPLC Measurement ofthe Cellular Nucleotide Pool-Cells were harvested and pooled from two T-150 flasks for each measurement, The cells were manually disrupted using a glass homogenizer with Teflon pestle (25 strokes) in 0.5 ml of perchloric acid at 4 "C. Acid extraction continued on ice for 15 min, at which time cellular debris was pelleted by centrifugation at 14,000x g for 5 min. The acid extract was neutralized by mixing with 2 ml of Freodtrioctylamine (4:l)for 30 s. The mixture was centrifuged at 14,000x g for 5 min, and the upper aqueous layer was recovered, rapidly frozen in liquid nitrogen, and stored at -80 "C for later analysis. Reverse-phase HPLC was performed as described by Hull-Ryde etal. (23) using a Waters Bondapak C18 column and a Waters HPLC system (Waters Associates, Milford, M A ) with a 510 reciprocating pumps, and model 484 W detector,twomodel Maxima software. Reagents-Recombinant human copper-zinc superoxide dismutase (>99%purity, 3,000 unitdmg) was obtained from BiotechnologyGeneral Corp. (New York,NY). Purified bovine liver catalase was obtained from Sigma with an assayed activity of 11,000 unitdmg protein or from Boehringer Mannheim with an assayed activity of 65,000unitdmg. In repeat experiments using either of these catalase preparations, similar results were observed. Deferoxaminemesylate was obtained from Ciba Pharmaceuticals, Inc.Allopurinol was purchased fromAldrich, and oxypurinol was obtained from Sigma. Superoxide dismutase was denatured by a modification of the procedure of Hodgson and Fridovich (24)in which the protein was titrated to a pH of 10 and incubated at room temperature for 12 h in the presence of a 10-fold excess of hydrogen peroxide followed by two-step dialysis. Catalase was denatured by boiling for 30 min. RESULTS

Measurements of Free Radical Generation-WCs were harvested from the culture flasks, washed, and resuspended in PBS. Concentrated suspensions of 4 x lo6 cells in 1 ml were incubated under anaerobic conditions at 37 "C for 60 min and then reoxygenated by addition of aerobic solution of DMPO and exposure to air. The cells were immediately transferred to an EPR flat cell, and spectra were acquired. EPR signals consisting of a prominent 1:2:2:1 quartet signal (a, = a, = 14.9 G), indicative of DMPO-OH, and a small 1:l:l:l:l:l sextet signal (a, = 15.8 a, = 22.8 G), indicative of DMPO-R, were observed (Fig. lA).In matched control preparations of aerobic cells, no signals were seen (Fig. 1B). Computer simulation confirmed that thespectrum of the reoxygenated cellsconsisted of a linear combination of the DMPO-OH and DMPO-R signals (Fig. 2). The time course of endothelial free radical generation following 60 min of anoxia was measured (Fig. 3). EPR spectra were recorded every 10 min for a period of 120 min. After 10 min of reoxygenation, a signal was observed that continued to gradu-

Mechanism of Radical Generation by Human Endothelial

24158

Cells

0.1 5

-h

I

3

0.10

I O I

0

2

0.05

n

-b 0.00 0

25

50

75

100

Time (min) 0.15

- I 2 I

I 3435

1

I

I

3460

I

I 3485

0.10

3510

3535 0

Magnetic Field (Gauss) FIG.1. EPR spectra of preparations of HAECs. A, spectrum observed from HAECs subjected to 60 rnin anoxia followed by reoxygenation; E , identical preparation of cells not subjected to anoxia.Spectra were recorded on 4 x lo6 celldm1 in the presence of 50 mM DMPO at a microwave frequency of 9.77 GHz, a microwave power of 20 milliwatts, and a modulation amplitude of 0.5 G .

25

50 75 Time (min)

100

Frc. 3. Time courses of the appearance of the EPR signals in preparation of W C s . 4 x lo6 HAECs/ml were exposed to 60 min of anoxia followed by reoxygenation in thepresence of 50 mM DMPO. The upper panel shows the time course for the DMPO-OH signal, and the lower panel shows the data for the DMPO-R signal.

shorter periods of anoxia gave rise to less radical generation upon reoxygenation. Cell preparations subjected to anoxia for 30 min followed by reoxygenation did not produce any detectable signal. Experiments performed with cells subjected to 90 min of anoxia demonstratedthat radical generation upon reoxygenation was further increased by about 30%,with total radical concentrations of 240 t 20 nM observed. With anoxia duration of 120 min, no further increase inradical production was observed (Table I). Experiments were performed in which the number of cells were varied to determineif the observed radical generation was specifically due to generation by the cells. On decreasing the number of cells in the anaerobic incubation solution, the observed radicalsignals were similarly decreased. Witha factor of of the radical signals 2 decreaseto 2 x IO6 celldml, the intensity was decreased by about 2-fold to a concentration of 86 t 12 m. With a 10-fold decrease in the cell number, no radical signals D were detectable. Experiments were performed to determine if the observed DMPO-OH or DMPO-R signals were derived from superoxide or hydrogen peroxide. Suspensions of 4 x lo6 cells were sub3421 3452 3411 3521 3502 jected to 60 min of anoxia at 37 "C in the presence of 200 Magnetic Field (Gauss) unitdm1 superoxide dismutase or 400 unitdm1 catalase followed by reoxygenation. In the absence of superoxidedisFIG.2. Computer simulation of the experimentalEPR spectra of reoxygenated HAECs. A , experimental spectrum; E , simulation of mutase or catalase, a prominent DMPO-OH signal was seen A as a linear combination of DMPO-OH ( C )and DMPO-R ( D )adducts. (Fig. 4A). In the presence of superoxide dismutase, no radical signals were observed (Fig. 4B).In thepresence of 400 unitdm1 ally increase for 40 min and then plateaued. Quantitation of catalase, 80-90% scavenging occurred (Fig. 4C). To determine the maximum total radical signal was performed in triplicate if the radical scavenging observed with superoxide dismutase specific enzyme activities, experimeasurements, and a concentration of 180 * 20 nM was ob- and catalase was due to their served for preparations of 4 x lo6celldm1 with 50 m~ DMPO. ments were performed in which W C s were exposed to simiTo determine whetherprolongation of the durationof anoxia lar periods of anoxia and reoxygenation, in the presence of in the HAECs would alter the magnitudeof free radical gen- denatured superoxide dismutase or catalase, at concentrations eration, similar preparationsof cells were subjected to varying matching those in the studies performed with the active enduration of anoxia, followed by reoxygenation in thepresence of zymes. With denatured superoxide dismutase or catalase, no DMPO. It was observed that HAEC preparations subjected to decrease in radical generation was seen.

n

n

ells

Mechanism of Radical Generation by Human Endothelial Cells

24159

TABLEI Correlation of radical concentration and cell injury Cells were counted 10 min after reoxygenation. Unless noted otherwise, cells weresubjected to 60 minanoxia prior to reoxygenation. Data are shown as mean t S.E.; n = 3. Radical Cellconcentration

DMPO-R

death (cells taking up trypan blue)

0 130 + 12

0 51 t 5

121 28 f 4"

174 t 15

67 + 7

38 f 6"

170 t 13

65 f 6

44 * I"

0

0

1*1

12 f 2 60 t 5

421 0

2t1 4t2

DMPO-OH

%

n M

Control Reoxygenated (60 min anoxia) Reoxygenated (90min anoxia) Reoxygenated (120 min anoxia) Reoxygenated + superoxide dismutase Reoxygenated + catalase Reoxygenated + deferoxamine Reoxygenated + oxypurinol

'p