Alkoxyl Radicals Produced from Anoxia/Reoxygenated Endothelial Cells. low'hal off ... (hydroxyl, superoxide anion) during post-anoxic reoxygenation, but little ...
l Mot (;ell Cardiol 27, 371-~;81 (1995)
Phospholipid Hydroperoxides are Precursors of Lipid Alkoxyl Radicals Produced from Anoxia/Reoxygenated Endothelial Cells laY H. Kramer, Benjamin F. Dickens, Vladimir Migtk:* and William B. Weglicki Departments q[ Medicine and Physiolo~l~], Divisiotl of E:cperimer,tal MediciHe, The (;eorqe WashiHqtol~ Utfiversit!t Medical CeHten 2 3 0 0 E je Street, N.W., Washingtotl, DC 20037, USA (Received 4 March 1994, acceptedin revisedform 15 June 1994) J. H. KRAMER,13. F. DICKENS, V. MIS1K AND W. 13. WV.C;HCK1.Phospholipid Hydroperoxides are Precursors of I~ipid Alkoxyl Radicals Produced from Anoxia/Reoxygenated Endothelial Cells. low'hal offMolecularand Celhdar Cardiologist (1995) 27, 371-381. Endolhelial cells have been shown to generate primary oxygen-centered free radicals (hydroxyl, superoxide anion) during post-anoxic reoxygenation, but little evidence is available concerning subsequent initiation of lipid peroxidalive iniury in this model. Electron spin resonance (ESP,) spectroscopy with ~-phenyl-N-tert-butylnitrone (PBN) spin trapping was used to monitor lipid peroxidation (Ll~O)-derived free radicals tbrmed by cultured bovine aortic endothelial cell suspensions exposed (~7°C) to anoxia (A, 45 min, N2 gas) and reoxygenation (R, 15 rain, 95% 0j5% (702). In some studies, superoxide dismutase (SOl), 1()yg/ml) was introduced just prior to R to assess the effects of this primary free radical scavenger on LPO-derived free radical production. At. various times, aliquots were removed and PBN was introduced to either the cell suspension aliquot (8 mM PBN final, 1 min), or to the corresponding cell-free filtrate (60 mM PI]N final), prior to extraction with toluene and ESR spectroscopy. A LPO-derived alkoxyl radical adduct of PBN (PBN/RO., hyperfine splitting c~N= 13.63 G and ~ o - - 1 . 9 4 - / . 9 8 G) was observed during R using both trapping procedures, with maximal production at 4-5 rain and a second minor peak at 10 rain. SO1) effectively reduced PBN/RO. production and improved viability of A/R cells. In parallel studies, lipid hydroperoxide production was assessed in lipid extracts of A/R cells by high-performance liquid chromatography. Their separation profiles revealed a peak of oxidized lipid occurring between phosphatidylethanolamine (PE) and phosphatidylcholine (PC) in samples taken at 4-5 rain and 10 rain of R. Resolubilizing cell lipid extracts in oxygen-free benzene conlaining cobalt(ll) acetylacetonate and PBN led to alkoxyl radical production, but only in the oxidized lipid samples, contirming the presence of hydroperoxides. These results suggest that A/R leads to primary free radical induced-lipid peroxidative injury to endothelial cells, as indicated by alkoxyl radical production originating from oxidized membrane phospholipids. KeY WOI/DS:Endothelial cells; Anoxia/Reoxygenation; Free radicals; ESR spin trapping; Lipid hydroperoxides; Superoxide dismutase.
Introduction
m y o c a r d i u m , brain, liver, g a s t r o i n t e s t i n a l tract, kid.* hey, a n d l u n g (Greenwald, 1990). The p a r t i c i p a t i o n of p r i m a r y o x y g e n free radicals (hydroxyl, "OH; superoxide anion, 02 ) in reperfusion i n j u r y is suggested by the prolection provided following treat-
Toxic o x y g e n free radicals p r o d u c e d d u r i n g postis(heroic reperfusion, m a y initiate similar peroxidative events in v a r i o u s tissues, including the
Please address all correspondence to: Jay [L Kramer, Ph.I)., Deparlments of Medicine and Physiology, Division of Experimental Medicine. 414A Ross Hall, The George Washington University Medical ('.enter, 2300 Eye Street, N.W., Washington, l)C 20037, USA. Current address: Radiation On(elegy Branch, Building 10, NCl/National Institutes of Health, Bethesda, MD 20892, USA. 0022-2828/95/010371 + 1.1 $08.00/0
371
© 1995 Academic Press [Amited
372
]. H. Kramer et al.
merit with free radical scavenging enzymes and drugs (Greenwald, 1990). However, while there is general agreement that free radicals are involved in reperfusion injury, their cellular origin(s) in rive remains uncertain. Ratych et aL (1988) suggested that the ubiquitous endothelial cell may be the primary source of reperfusion-generated free radicals, alter noting the similarities between the oxidative component of reperfusion injury in diftL'rent organs. Using cultured endothelial cells, several laboratories (Zweier et al., 1988; Schinetti et al., 1989; Arroyo et aL, 1990) demonstrated that anoxia followed by reoxygenation resulted in production of oxygen free radicals, thus, supporting the hypothesis that vascular endothelial cells may be a prominent source of free radical generation during ischemia/reperfusion; moreover, endothelial cells may be a principal site of oxidative injury. We have previously shown that cultured endothelial cells are particularly vulnerable to exogenously-generated free radical stress, leading to decreased cell viability (Dickens et aI., 1992). Loss of viability was also associated with reduced levels of cellular antioxidants (glutathione) (Maket al., 1 9 9 2 ) and increased tbrmation of lipid peroxidation (LPO) products (malondialdehyde and lipid peroxides) (Mak et al., 1992; Weglicki et al., 1992). In the current study, electron spin resonance (ESR) spectroscopy and spin trapping methods were used to characterize production of endothelial cell LPOderived free radicals during anoxia/reoxygenation (A/R); detection of these t)ee radical species provides an index of lipid peroxidative injury (Kramer et al., ] 994) [bllowin g oxygen l)ee radical initiation of LPO. Since xanthine oxidase-rich endothelial cells (species specific) (gloner et aI., 1989; Ashraf and Samra, ] 991) may have a high propensity lowards 02 production during A/R, the involvement ofthis free radical species in the initiating events o:l' LPO was assessed with superoxide dismutase (SOD). High-performance liquid chromatography/llV analyses of endothelial cell lipid extracts provided a method for monitoring the time of peroxidized lipid formation during A/R. The time course ofl,PO-derived free radical detection was compared with that of peroxidized lipid tbrmation and their association was described.
Materials and Methods Chemicals Bovine erythrocyte SOD (3500U/mg), phosphodiesterase (bovine heart 3',5'-cyclic nucleotide 5"-nucleotidohydrolase, 0.28 lJ/mg), cumene hydro-
peroxide (CuOOt]), the spin trap ~-phenyl-N-tertbutylnitrone (PBN) and the free radic~d standard 2,2,6, 6-tetramethyl piperidine- ] -oxyl (TEMPO) were obtained from Sigma Chemicals (St. Louis, Me, USA). 2,7-Dichlorofluorescin diacetate was purchased from Eastman Kodak Co. (Rochester, NY, USA), and Cobalt(II) acetylacetonate (Co[Ill AC/\C) from Fisher Scientific (Pittsburgh, PA, IJSA). All other chemicals were purchased li~ora Sigma Chemicals and solvents from Fisher Scientific. The incubation bufti~r and spin trap solutions were prepared in double deionized, distilled water and filtered through 0.8#m pore Gehnan filters. No trace metal contamination (iron, copper) was detected by atomic absorption spectroscopy in the incubation but])~r and spin trap solutions; however, the cell culture media did contain 250nm iron (as Fe(NO~)~'9H20) as a normal constituent.
Cultured endothelial cell preparations Bovine aortic endothelial cells (AG 07684, The Coriell Institute R)r Medical Research, Camden, NJ, []SA) were cultured in minimal essential medium and grown in 150-cm 2 culture tlasks (Coming, Wexlbrd, PA, USA) as described (Dickens et al., ] 992). Following trypsinization and neutralization by addition of growth medium (Dickens et aL, 1992), suspensions were centrifuged twice (400 × g, 4 rain, 24°C) with an iso-osmotic 10 mm potassium phosphate incubation buffer (pit 7.4) coutainh]g (nlM): NaCI, 120; MgSO4'7tt20, 1.2: CaCl:z'2HeO, 1.25; and glucose, 5. The washed final pellet was resuspended in incubation buffer and cellular viabilities of >90%, based on morphology and trypan blue exclusion, were typically obtained prior to initiating experimenta] incubations.
Incubation procedures Endothelial cell suspensions were diluted (3.27 x 1()'4:().31 ce/ls/ml final) into either nitrogen-gassed (anoxic) or 95% O j 5 % CO,-gassed control incubation buffer: suspensions were then gassed again (10 s) in the presence of a Clark oxygen electrode (with a Yellow Springs Instrument 5300 monitor) to verily oxygen content betbre initiating cell incubation (Kramer et al., 1991). For the anoxia/reoxygenation (A/R) studies, cell samples were incubated (37°C shaking bath} for 45 rain in anaerobically-sealed test-tubes, and subsequently reoxygenated with 95% OJ5% CO, for ] 5 rain. At the end of anoxia and at various times ofreoxygenation,
Alkoxyl Radicals From Endothelial Cells aliquots were retrieved and were assessed for either lipid hydroperoxides (9 x 10 ~ cells/sample) or free radical production by ESR spin trapping (5 x 1() ~' cells/sample). Endothelial cell viability was determined at the end of reoxygenatkm by the trypan blue dye (0.1% final) exclusion method (Kramer et al., 1991; 1)ickens et al., 1992). In some studies, endothelial cells were exposed to SOD (10#g/ml final) 1 rain prior to reoxygenation. The choice of dose was based on previous experience (Kramer et at., 11991; Freedman et al., 1991) and the observation thai: SOl) may lose its cardioprotective effects at higher levels (Omar et al., 1990). Since SOD will generally be confined to the extracellular aqueous space, addition prior to A/R was not considered necessary. SO1) lreatment had no sign ificanI effects on normoxic control cell viability or lipid peroxide formation. In all instances, appropriatelyincubated (up to 60 rain, 37°C, aerated) normoxic control samples were run for comparison with the A/R studies.
ESR spin trapping and sample extraction ESR spin trapping was performed by our previously described technique (Kramer et al., 1991) on control and A/R-exposed endothelial cell samples in the presence or absence of SOD. Aliquots (1.5 ml) of cell suspension were removed and rapidly filtered through a vacuum filtratkm manifold (Whatman GF/C filters) into tubes coniaining 1.5ml PBN (120 mM in 0.11M NaC1). Filtration of the cell suspension should eliminate cellular metabolism of PBN or free radical adducts during sample processing (Samuni and Swartz, 1989). Since ceils are not directly exposed to the trapping agent during incubation, the potential for dose-dependent toxic (Bolli et al., 1988) or non-toxic (ttearse and Tosaki, 1987; Bolli et al., 1988) effects can also be avoided. The filtrate/PBN mixtures were immediately t~ozen in liquid nitrogen and subsequently extracted (1-vol sample: 1-vol toluene: 20°C) with nitrogengassed, high-performance liquid chromatography (HPLC)-grade toluene; extracts were recovered following centrifugation (170 x g, 4 min) and volumes reduced under nitrogen gas to 0.3 ml prior to ESR spectroscopy. For comparison (n=3), aliquots of post-anoxic cell Suspensions were directly exposed to lower PBN concentration (8 mM) during I min intervals throughout the 15 min reoxygenation period; these unfiltered smnples were then extracted prior to ESR speciroscopy. Extraction efficiency fbr PBN was 90% (Mergner et al., 1991), and the extraction procedure provided stable recovery of
?173
the spin adduct with no significant ESR signal artifact observed in extracts of PBN-supplemented, cell-free incubatkm buffer controls. For comparison with experime~tal spectra, PBN/alkoxyl (PBN/RO') adducts were chemically-generated from a mixture of CuOOH and FeSO4'7HeO prior to toluene extraction as previously described (Mergner et a/., 19911). No PBN/RO. adducts were detected in the CuOOH-supplemented incubation buftL'r mlless an exogenous source of Fe :z~ was provided.
Conditions for ESR spectroscopy ESR analyses (Kramer et ~1., 1991; Mergner et al., 1991) were perlbrmed in subdued light at 11.7°C with a Bruker ER-I O0 series, X-band spectrometer using a magnetic field modulation frequency of 100 Kifz. The ESR spectra were obtained using a quartz fiat cell ( 6 O x 1 O x O . 2 5 m m ) . The microwave power was maintained at 10 mW to avoid saturatkm, and scans were perlbrmed with a modulation amplitude equal to or lower than 1.25 G. ESR spectra of filtrate/PBN mixture extracts were obtained with the gain sel at 2.5 x 1() ~, time constant at 0.2 s, spectra width at 100 G, and sweep time. at 200 s. ESR spectra of extracts from unfiltered PBN-exposed cell suspension aIiquots were obtained using the above settings, except for a gain of 6.3 xl(Y' and speclra width of 80(;. }typerfine coupling constants were measured directly l~om the field scan using a 10 G marker signal for calibration, and spectral simulatkm was used to verify spectral parameters; l)ee radical adduct concentrations were determined by double integration of experimental spectra using TEMPO as an integration standard (Mergner et al., 1991; Tortolani et al., 1993).
Lipid analyses The possibility that the observed PBN-alkoxyl adducts produced by A/R endothelial cells arise from phospholipid hydroperoxide precursors was investigated. Lipids were extracted from 9 x 1 tY' cells with choloroform/methanol and aliquots of the lipid extract were subjected to HPLC chromatography (Altech Ultrasphere column) as previously described (Weglicki et al., 1984). The lipid extracts were screened for the presence of hydroperoxide moieties by two diftbrent assays. In one (Kramer et al., 1994), 2,7-dichlorotluorescin diacetate (3 itM) was digested for 10 min at 30°C with phosphodiesterase (0.1 rag/ ml) to produce 2,7-dichlorofluorescin (DCF). DCF
374
J.H. Kramer et al.
was added to aliquots of phospholipid extract (reconstituted in 100% ethanol) from normoxic, anoxic, and reoxygenated endothelial cells, and the resulting fluorescence formed by the oxidation of DCF to 2,7-dichloroIluorescin was measured using a Sequoia-Turner fluorometer (wavelength 488 n m excitation, with a 500 n m cutoff filter). In the second assay, the lipids were resolubilized (450/4) in de-oxygenated benzene and PBN (1.5 mM final) and Co[II] ACAC (0.6 mM final)were added prior to ESR spectroscopy. Metal-catalysed production of alkoxyl radicals will occur in the absence of oxygen if hydroperoxides are present in the peroxidized lipid extracts.
Statistical analysis Analysis of variance was used for comparison of several means and the Tukey test for all paired comparison of means. Significance was considered at P
