J Mel
Cell
Gardiol24,
(1992)
1031-1038
Iron-Catalyzed
Edward
Reactions
Cause Lipid Heart
J. Lesnefsky*, Kenneth and Lawrence
Peroxidation
G. D. Allent, D. Horwitz~
in the Intact
Frank
P. Camea:,
*Case Western Reserve Universip, Cleveland, OH, USA, “yColorado State University, Ft Collins, CO, USA and f University of Colorado Health Sciences Center, Denver, CO, 1JSA (Received 25 October 1991; acceped in ~ev~sed~orn~ 21 April
1992)
ii. G. D. ALLEN, F. P. CARREA, AND L. D. HORWITZ. Iron-Catalyzed Reactions Cause Lipid Per(jxjd~tioIl in tbe Intact Heart. Journal ~~~~~~~c~~QY and Cellulur C~~d~~~o~~ (1992) 24, 1031-1038. The chemical targets and mechanisms of iron-catalyzed oxidative injury in myocardium are poorly understood. Oxygen metaholites, in the presence of iron, can initiate free-radical chain reactions in unsaturated membrane lipids, generating lipid peroxides and causing membrane injury. We examined whether exposure to iron-catalyzed oxidative injury would increase myocardial lipid peroxide levels as injury evolved in the intact heart. Isolated, huKer perfused rabbit hearts were exposed for 30 min to 100 uu Fez+/500 uu ADP and 10 UM H,O, (IRON group, rz= 51, saline vehicle (CON group, n=6) or 500 UM ADP and 10 UI H,O, without iron (ADP, II= 5). Lipid peroxides were measured in cytosol and membrane fractions by a new method, using the Iipid prroxidcinduced oxidation of exogenous GSH to GSSG, catalyzed by the enzyme glutathione peroxidase. The results indicated that iron-catalyzed lipid peroxidation occurs in the intact heart during chemically-mediated oxidativt injury.
E. J. LESNEFSKY,
KEY WORDS: Oxygen
radicals;
Iron;
Glutathione
peroxidase;
Introduction There is considerable evidence that iron-catalyzed reactions involving oxygen metabolites exacerbate myocardial injury during ischaemia and reperfusion [I-4]. Potential targets of iron-catalyzed injury in reperfiused myocardium include membrane lipids [.5J, cellular proteins [6J and nucleic acids [II. Iron-catalyzed reactions, including formation of the hydroxyl radical (‘OH), initiate free radical chain reactions in unsaturated lipids and causelipid peroxidation in isolated membrane systems [R, 93. However, direct evidence that iron-catalyzed reactions causelipid peroxidation in myocardial tissue has been lacking. Efforts to isolate the lipid peroxidation marker, malondialdehyde, from oxidativelyinjured myocardial tissuehave yielded inconsistent results,possibly becauseit is a relatively insensitive and non-specific index of lipid peroxidation [MJ. Thus, the potential contribution of iron-catalyzed lipid peroxidation to
Lipid
peroxidation;
Glutathionr
myocardial oxidative injury has remained unclear. To determine if iron-catalyzed reactions causelipid peroxidation in the intact heart, we adapted a recently described, highly sensitive lipid peroxide assay[i1] for measurementsin myocardial tissue. Lipid peroxides in tissue sampleswere assayed by adding glutathione (GSH) and giutathione peroxidase to generate oxidized glutathione (GSSG) [ll], which was then quantitated [12]. Isolated rabbit hearts were perfused with an Fez’-ADP-H,O, system that generates ‘OH [13]. Myocardial lipid peroxide levels in these iron-treated hearts (IRON group) were compared to levels in saline-treated controls (CON group) and to hearts that received H,O, and ADP, but no iron (ADP group). Using these methods, we present definitive evidence that iron-mediated reactive oxygen metabolite reactions are capable of inducing substantial lipid peroxidation in tissue (Table 1)
Abbreviations used: IRON, iron-H,O, ADP treated hearts; CON, saline treated controls; ADP group, H,O? ADP treated hearts without iron; TRIS, buffer system for cytosoi and membrane isolation; ROOH, free Fatty arid hydroperoxides; PLOOH, phospholipid hydroperoxides; PC:, phosphatidyl choline. (ML’2 -28%;92,‘0911J3 1 + 08 $08.00,‘0 :f’ 1992 Academic Press I,imiwtl
1032
E. J. Lesnefsky
TABLE 1. Iron-catalyzed lipid peroxidation in the intact heart during chemicallymediatedoxidative injury
et al.
and 500~~~NaADP. The rate of infusion was not changed during the experiment. These solutions were passedthough a 0.22~~ filter (Millipore, Bedford, MA, USA) prior to Lipid peroxides Cytosol Membrane being infused. Control hearts (CON, n= 6) (pmol/mgtissue) received an infusion of normal saline vehicle at the same rate. The H,O,-ADP group CON 6.4f3.0 0.1 fO.l (n= 5) received an identical infusion to the ADP 14.6f7.1 0.2f0.1 experimental group, except that FeCl, was IRON 64.7i7.F 10.9*4.7* omitted. *P 430 mmHg) as previously described genate was centrifuged at 100000 g for [14]. The balloon volume was adjusted to 60 min at 4°C. The supernatant was collected yield an initial left ventricular end-diastolic and represents the soluble (cytosol) fraction pressure of 5 mmHg and was held constant [I5f. The pellet, which included the membrane fraction, was resuspendedin 2 ml TRIS throughout the experiment. The coronary sinus was cannulated to provide effluent for buffer that contained 0.1% Triton X-100, and measurement of coronary flow. Flow was was recentrifuged at 100OOOg at 4°C for measuredevery 5 min throughout the experi- 60 min. This supernatant yields an unwashed membrane fraction [ISJ. Cytosol and memment by timed collection of coronary effluent. The hearts were paced at 240 beats/min via brane fractions were assayedfor lipid peroxides using the method described below. epicardial electrodes. After 15 min of equilibration, the experimental hearts (IRON, n = 5) were perfused Lipid peroxide assay for 30 min with H,O, and a FeCl,-NaADP chelate, each separately infused at 0.5 ml/min An aliquot of the cytosol fraction (100 ~1) was added to 330 ~1 of buffer (73 mM sodium via independent infusion systems using a phosphate pH 7.4,5 mM NaEDTA and 5 mgjl double Harvard infusion pump (Harvard, Billerica, MA, USA). Thus, the H,O, and Tween-20 as emulsifier), To increase the Fe’+-ADP did not mix until addition to the sensitivity of the assay for membrane peroxKrebs-Henseleit buffer. The concentrations of ides, 250~1 of the membrane fraction was H,O,, FeCl, and ADP were adjusted basedon added to 180,uI of buffer. Fifty microlitres of 1 mM GSH and 20~1 of glutathione peroxithe coronary flow during the equilibration period to yield final concentrations in Krebs- dase (10 U/ml) were sequentially added and the mixture incubated at 37°C for 20 min. Henseleit buffer of 10PM H,O,, 100,UM FeCl,
Myocardial
Lipid
Fifty microlitres of 12.5 mM N-ethyl maleimide (NEM) was then added to eliminate unreacted GSH. After 15 min incubation at 28°C excessNEM was inactivated by adding 8 1wKOH, with vortexing, to raise the pH to 11.O,and the sample left at room temperature for 5 min [11, 16’j. Some 8~1 4 M HCl was added with vortexing to return the pH to 6.5 7.0. Tert-butyl hydroperoxide was used asthe lipid peroxide standard, with a range of O4500 pm01 per reaction mixture. The GSSG generated by this assay was measured immediately on a multiwell platereader (Bio-Tek, Highland Park, VT, USA) using the method of Sies adapted for use in a multiwell platereader [12, 171. Ten microlitres of the reaction mixture from the lipid peroxide assay was added to 140~1 of 0.1 M potassium phosphate buffer pH 7.4, 5Og1 of 3rnM dinitrothiobenzene (DTNB)-5 mM NaEDTA, and 10~1 glutathione reductase !0.25 mg/ml, Sigma Type III). 40 ~1 of 1.25 mM NADPH was added to start the reaction. ‘fotal well volume was 250~1. Absorbance at 412 nm (OD,,,) was followed at 2 min intervals at 28°C. GSSG standards were used with concentrations of O-500 pmol per well. Standard curves of concentration vs net OD,,, :ODmnplr-ODlllarrk) were constructed at each timr point and used to calculate the GSSG content (pmol per well) of samples.AI1 chemicals were reagent grade and were purchased from Sigma Chemical Co., St. Louis, MO, L-SA. Myocardial lipid peroxide levels were calculated asfollows. The GSSG content in pmol per well was calculated for lipid peroxide standards and samples.The smail amount of GSSG present in the lipid peroxide assay blanks was subtracted from all samplesand standards to yield n GSSG. Linear regressions of n GSSG vs tert-butyl hydroperoxide standards were constructed for each assay. The mean correlation coefficient for these tertbutyl standard curves was 0.98 (a= 16). Since tissue homogenates have a small amount of GSSG already present prior to incubation in the iipid peroxide assay, each tissue fraction was also directly assayedfor GSSG f17j. The pre-existing GSSG was subtracted from LiGSSG, to yieId a net “AGSSG effective”. This “AGSSG effective” is the GSSG generated by the glutathione peroxidase-catalyzed
Peroxidation
1033
reaction of Iipid peroxides and GSH. The “A GSSG effective” was used to calculate the amount of lipid peroxide present, using the tert-butyl vs AGSSG standard curve. Results were expressed as picomoles (pmol) per mg tissue. Three tissue samples from each heart were homogenized, and the resulting membrane and cytosol fractions were analyzed in duplicate in the lipid peroxide assay. Triplicate aliquots of each lipid peroxide assay reaction mixture were assayedfor GSSG. Critique qf the lipid peroxideasscp! In the O-4500 pmol per reaction mixture range the lipid peroxide assayis linear. Recovery of added tert-butyl peroxide from tissue using this assay is 80%. Reaction blanks do not vary by more than 10%. Blanks containing GSH and glutathione peroxidase but no lipid peroxide have the samelow GSSG values as standards with lipid peroxide and GSH, but no glutathione peroxidase. This confirms that GSSG formation is due to a reaction catalyzed by glutathione peroxidase between GSH and lipid peroxide, rather than nonspecific oxidation of GSH. In separate experiments on cytosol fractions, where glutathione peroxidase was omitted, all measured GSSG was accounted for by GSSG present in the fraction prior to lipid peroxide assay (“ a GSSG effective” = 0). This again confirms that non-specific oxidation of GSH did not occur in the lipid peroxide assay. Since lipid peroxide levels were not significantly increased in the H,O,-ADP group, it is unlikely that H,O, trapped in myocardimm explains the elevated lipid peroxide levels found in the IRON group. Iron levels were measured in cytosol fractions from hearts in the IRON and CON groups. Measurement of iron was performed by quantitation of absorbance at 429 nm of the iron-saturated complex ferrioxamine [28J. An excess of deferoxamine (50 mM) was added to cytosol fractions. Cytosolic iron levels were elevated in the IRON group (22.3 f 6.4pM, n= 3) compared to the CON group (6.1&6. I PM, n=3). To assure that increased cytosolic iron did not contribute to lipid peroxide formation during assayprocrdures or tissue processing, cytosol fractions from ADP-H,O, treated hearts were assayed
1034
E. J. Lesnefsky
in the presence of added iron, ( 10 ,ul of 160 PM FeCl, added to the cytosol fraction) in order to increase iron levels to those present in the IRON hearts. Tissue lipid peroxide levels were not increased by the additional iron (PcO.45, 72~6, paired t test). Reactivity offree fat0 acid and phospholipid peroxides The reactivity of free fatty acid hydroperoxides (ROOH) and phospholipid hydroperoxides (PLOOH) in the lipid peroxide assay using bovine erythrocyte glutathione peroxidase (Sigma) [11] was determined. Phosphatidyl choline (PC) (Sigma, 500 pg, purity confirmed by thin layer silica gel chromatography with hexane:diethyl ether:acetic acid 83:15:2 as solvent) was dispersed in 1 .O ml 5 mM HEPES buffer, pH 7.4, 0.15 M NaCl, containing BHT as an antioxidant. When unperoxidized PC was assayed for lipid peroxides [II], none was detected as expected. PC was then hydrolyzed with 10 U porcine pancreas phospholipase A, in the presence of 10 mti Ca” and 1.25 mM Na-deoxycholate at 37°C for 30 min. BHT and argon flushing of the headspace were utilized to prevent oxidation. The reaction was stopped by 20 mM EGTA. No lipid peroxides were detected in the hydrolysate. Thin layer chromatography confirmed release of fatty acids from PC. PC was then photo-oxidized using a fluorescent lamp with methylene blue as a photosensitizer. When the photo-oxidized PC was assayed for lipid peroxides, again none were detected. Thin layer chromatography confirmed that no free fatty acids were present. Photo-oxidized PC was then hydrolyzed by phospholipase A,. Lipid peroxides were now easily detectable. Approximately 35% of the maximal theoretical yield of ROOH from the SN-2 position (arachidonic acid) was obtained. Thin layer chromatography showed that the photo-oxidized phospholipase-treated sample now had easily demonstrable levels of free fatty acids present. These experiments show that the glutathione peroxidase-based lipid peroxide assay detects ROOH, but not PLOOH. Other biochemical assays Reduced
(GSH)
and oxidized
(GSSG)
gluta-
et al.
thione were measured in myocardium by the method of Sies adapted for use in a multiwell platereader [12, 171. Adenosine triphosphate (ATP) and phosphocreatine (PCR) were measured by the flourometric method of Lowry and Passonneau [I$]. Tissue protein was measured by the dye binding method of Bradford [20]. Statistical methods Data were expressed as mean (S.E.). Comparisons among chemical and hemodynamic one way analysis of variance the Student-Newman-Keuls vidual comparisons. Within sons of hemodynamic results by two way ANOVA [Zl]. considered significant.
f standard error groups of bioresults were by (ANOVA), with test used for indigroup compariover time were A P~0.05 was
Results Lipid peroxide levels were increased in both cytosol (PC 0.01) and membrane (PC 0.02) fractions in the iron-H,O, (IRON) group compared to saline control (CON) and H,O,ADP (ADP) control groups (Table 2). Lipid peroxide levels in the ADP group (no iron) were similar to those in the CON group (P= NS) . Similar low lipid peroxide levels in the CON and ADP hearts suggest that the catalytic effect of iron was important. The similar levels in the CON and ADP groups do not support a possible reaction between H,O, and GSH in the lipid peroxide assay as a cause of the elevated lipid peroxide levels found in the IRON group. Furthermore, excess iron added directly to the tissue fractions did not increase lipid peroxide formation, confirming that lipid peroxidation during sample processing did not occur. The increase in both cytosol and membrane lipid peroxides in the IRON hearts was the most prominent biochemical evidence of oxidative injury in these hearts. However, there was other tissue evidence suggestive of oxidative injury in IRON hearts. ATP levels (P~0.05) were decreased. The IRON group also had modest trends toward elevated tissue GSSG and decreased tissue GSH, GSH/GSSG ratio and phosphocreatine (PCR) (Table 2), though these were not statistically significant.
Myocardiai
Lipid
1035
Peroxidation
TABLE 2. Biochemicalresults Iron-E&O, Cytosol LP (prno~~rn~ tissue) Membrane LP (pmol/mg tissue) GSH (nmol/g tissue)
GSSG (nmol/g tissue) GSH/GSSG :YTP jnmol/mg prot) PCR (nmol/mg prot) Protein (mg/g tissue)
(n = 5)
64.7 f 7.5’ 10.9*4.7 798 rk 93.0 27.4f 11.0 67 f 24.0 16.2f2.52 39.3f2.6 66.9zt 7.5
Saline (n = 6)
H,O,-ADP In= 5)
6.4 i 3.0 0.1 fO.l 92 1* 64.0 9.7fO.7 96f 7.0 27.9f 3.6 56.8 jc 8.2 67.7i4.6
10.1 f 1.7 0.2JtO.l 662 + 50.0 6.5iO.8 103f 14.0 21.91 1.7 51.3rf: 11.8 71.1 f5.1
’ P< 0.05 “S otlw groups. “PC 0.05 vs saline.
Tissue GSH tended to be lower (P=O.O6) in the ADP group than in CON (Table 2). Administration of 10~~ H,O, oxidizes GSH to GSSG. The GSSG formed is then actively transported from the myocyte [ZZ, I#]. This maintains normal tissueGSSG at the cost of a reduction in tissue GSH [lZ], since GSSG is released rather than reduced to GSH. Myocardial GSH depletion was lessprominent in the IRON
group,
possibly
since most of the
infused H,O, reacted with the excess iron to form ‘OH. Hydroxyl radical reacts rapidly with most adjacent moleculesincluding unsaturated lipids [.?2]. In contrast, H,02 is poorly reactive with unsaturated lipids and other biomolecules [22], but diffuses into the cytosoi and oxidizes GSH [12, .?Z]. Thus, more H,O, may have reached the cytosol and oxidized GSH in the ADP group than in the IRON group, accounting for the lower tissue GSH values. ~et~~xi~cation of intracellular H,O, by GSH may also account for the absenceof an increase in lipid peroxides in the ADP group. It is likely that H,O, was eliminated by reaction with GSH before it could react to form ‘OH in this group, especially since levels of “free” redox-active iron are very low in the non-ischaemic heart [ZZ, 231. Thus, reduction ofH,O, to water by GSH rather than reaction of H,O, with ‘
Cell
Gardiol24,
(1992)
1031-1038
Iron-Catalyzed
Edward
Reactions
Cause Lipid Heart
J. Lesnefsky*, Kenneth and Lawrence
Peroxidation
G. D. Allent, D. Horwitz~
in the Intact
Frank
P. Camea:,
*Case Western Reserve Universip, Cleveland, OH, USA, “yColorado State University, Ft Collins, CO, USA and f University of Colorado Health Sciences Center, Denver, CO, 1JSA (Received 25 October 1991; acceped in ~ev~sed~orn~ 21 April
1992)
ii. G. D. ALLEN, F. P. CARREA, AND L. D. HORWITZ. Iron-Catalyzed Reactions Cause Lipid Per(jxjd~tioIl in tbe Intact Heart. Journal ~~~~~~~c~~QY and Cellulur C~~d~~~o~~ (1992) 24, 1031-1038. The chemical targets and mechanisms of iron-catalyzed oxidative injury in myocardium are poorly understood. Oxygen metaholites, in the presence of iron, can initiate free-radical chain reactions in unsaturated membrane lipids, generating lipid peroxides and causing membrane injury. We examined whether exposure to iron-catalyzed oxidative injury would increase myocardial lipid peroxide levels as injury evolved in the intact heart. Isolated, huKer perfused rabbit hearts were exposed for 30 min to 100 uu Fez+/500 uu ADP and 10 UM H,O, (IRON group, rz= 51, saline vehicle (CON group, n=6) or 500 UM ADP and 10 UI H,O, without iron (ADP, II= 5). Lipid peroxides were measured in cytosol and membrane fractions by a new method, using the Iipid prroxidcinduced oxidation of exogenous GSH to GSSG, catalyzed by the enzyme glutathione peroxidase. The results indicated that iron-catalyzed lipid peroxidation occurs in the intact heart during chemically-mediated oxidativt injury.
E. J. LESNEFSKY,
KEY WORDS: Oxygen
radicals;
Iron;
Glutathione
peroxidase;
Introduction There is considerable evidence that iron-catalyzed reactions involving oxygen metabolites exacerbate myocardial injury during ischaemia and reperfusion [I-4]. Potential targets of iron-catalyzed injury in reperfiused myocardium include membrane lipids [.5J, cellular proteins [6J and nucleic acids [II. Iron-catalyzed reactions, including formation of the hydroxyl radical (‘OH), initiate free radical chain reactions in unsaturated lipids and causelipid peroxidation in isolated membrane systems [R, 93. However, direct evidence that iron-catalyzed reactions causelipid peroxidation in myocardial tissue has been lacking. Efforts to isolate the lipid peroxidation marker, malondialdehyde, from oxidativelyinjured myocardial tissuehave yielded inconsistent results,possibly becauseit is a relatively insensitive and non-specific index of lipid peroxidation [MJ. Thus, the potential contribution of iron-catalyzed lipid peroxidation to
Lipid
peroxidation;
Glutathionr
myocardial oxidative injury has remained unclear. To determine if iron-catalyzed reactions causelipid peroxidation in the intact heart, we adapted a recently described, highly sensitive lipid peroxide assay[i1] for measurementsin myocardial tissue. Lipid peroxides in tissue sampleswere assayed by adding glutathione (GSH) and giutathione peroxidase to generate oxidized glutathione (GSSG) [ll], which was then quantitated [12]. Isolated rabbit hearts were perfused with an Fez’-ADP-H,O, system that generates ‘OH [13]. Myocardial lipid peroxide levels in these iron-treated hearts (IRON group) were compared to levels in saline-treated controls (CON group) and to hearts that received H,O, and ADP, but no iron (ADP group). Using these methods, we present definitive evidence that iron-mediated reactive oxygen metabolite reactions are capable of inducing substantial lipid peroxidation in tissue (Table 1)
Abbreviations used: IRON, iron-H,O, ADP treated hearts; CON, saline treated controls; ADP group, H,O? ADP treated hearts without iron; TRIS, buffer system for cytosoi and membrane isolation; ROOH, free Fatty arid hydroperoxides; PLOOH, phospholipid hydroperoxides; PC:, phosphatidyl choline. (ML’2 -28%;92,‘0911J3 1 + 08 $08.00,‘0 :f’ 1992 Academic Press I,imiwtl
1032
E. J. Lesnefsky
TABLE 1. Iron-catalyzed lipid peroxidation in the intact heart during chemicallymediatedoxidative injury
et al.
and 500~~~NaADP. The rate of infusion was not changed during the experiment. These solutions were passedthough a 0.22~~ filter (Millipore, Bedford, MA, USA) prior to Lipid peroxides Cytosol Membrane being infused. Control hearts (CON, n= 6) (pmol/mgtissue) received an infusion of normal saline vehicle at the same rate. The H,O,-ADP group CON 6.4f3.0 0.1 fO.l (n= 5) received an identical infusion to the ADP 14.6f7.1 0.2f0.1 experimental group, except that FeCl, was IRON 64.7i7.F 10.9*4.7* omitted. *P 430 mmHg) as previously described genate was centrifuged at 100000 g for [14]. The balloon volume was adjusted to 60 min at 4°C. The supernatant was collected yield an initial left ventricular end-diastolic and represents the soluble (cytosol) fraction pressure of 5 mmHg and was held constant [I5f. The pellet, which included the membrane fraction, was resuspendedin 2 ml TRIS throughout the experiment. The coronary sinus was cannulated to provide effluent for buffer that contained 0.1% Triton X-100, and measurement of coronary flow. Flow was was recentrifuged at 100OOOg at 4°C for measuredevery 5 min throughout the experi- 60 min. This supernatant yields an unwashed membrane fraction [ISJ. Cytosol and memment by timed collection of coronary effluent. The hearts were paced at 240 beats/min via brane fractions were assayedfor lipid peroxides using the method described below. epicardial electrodes. After 15 min of equilibration, the experimental hearts (IRON, n = 5) were perfused Lipid peroxide assay for 30 min with H,O, and a FeCl,-NaADP chelate, each separately infused at 0.5 ml/min An aliquot of the cytosol fraction (100 ~1) was added to 330 ~1 of buffer (73 mM sodium via independent infusion systems using a phosphate pH 7.4,5 mM NaEDTA and 5 mgjl double Harvard infusion pump (Harvard, Billerica, MA, USA). Thus, the H,O, and Tween-20 as emulsifier), To increase the Fe’+-ADP did not mix until addition to the sensitivity of the assay for membrane peroxKrebs-Henseleit buffer. The concentrations of ides, 250~1 of the membrane fraction was H,O,, FeCl, and ADP were adjusted basedon added to 180,uI of buffer. Fifty microlitres of 1 mM GSH and 20~1 of glutathione peroxithe coronary flow during the equilibration period to yield final concentrations in Krebs- dase (10 U/ml) were sequentially added and the mixture incubated at 37°C for 20 min. Henseleit buffer of 10PM H,O,, 100,UM FeCl,
Myocardial
Lipid
Fifty microlitres of 12.5 mM N-ethyl maleimide (NEM) was then added to eliminate unreacted GSH. After 15 min incubation at 28°C excessNEM was inactivated by adding 8 1wKOH, with vortexing, to raise the pH to 11.O,and the sample left at room temperature for 5 min [11, 16’j. Some 8~1 4 M HCl was added with vortexing to return the pH to 6.5 7.0. Tert-butyl hydroperoxide was used asthe lipid peroxide standard, with a range of O4500 pm01 per reaction mixture. The GSSG generated by this assay was measured immediately on a multiwell platereader (Bio-Tek, Highland Park, VT, USA) using the method of Sies adapted for use in a multiwell platereader [12, 171. Ten microlitres of the reaction mixture from the lipid peroxide assay was added to 140~1 of 0.1 M potassium phosphate buffer pH 7.4, 5Og1 of 3rnM dinitrothiobenzene (DTNB)-5 mM NaEDTA, and 10~1 glutathione reductase !0.25 mg/ml, Sigma Type III). 40 ~1 of 1.25 mM NADPH was added to start the reaction. ‘fotal well volume was 250~1. Absorbance at 412 nm (OD,,,) was followed at 2 min intervals at 28°C. GSSG standards were used with concentrations of O-500 pmol per well. Standard curves of concentration vs net OD,,, :ODmnplr-ODlllarrk) were constructed at each timr point and used to calculate the GSSG content (pmol per well) of samples.AI1 chemicals were reagent grade and were purchased from Sigma Chemical Co., St. Louis, MO, L-SA. Myocardial lipid peroxide levels were calculated asfollows. The GSSG content in pmol per well was calculated for lipid peroxide standards and samples.The smail amount of GSSG present in the lipid peroxide assay blanks was subtracted from all samplesand standards to yield n GSSG. Linear regressions of n GSSG vs tert-butyl hydroperoxide standards were constructed for each assay. The mean correlation coefficient for these tertbutyl standard curves was 0.98 (a= 16). Since tissue homogenates have a small amount of GSSG already present prior to incubation in the iipid peroxide assay, each tissue fraction was also directly assayedfor GSSG f17j. The pre-existing GSSG was subtracted from LiGSSG, to yieId a net “AGSSG effective”. This “AGSSG effective” is the GSSG generated by the glutathione peroxidase-catalyzed
Peroxidation
1033
reaction of Iipid peroxides and GSH. The “A GSSG effective” was used to calculate the amount of lipid peroxide present, using the tert-butyl vs AGSSG standard curve. Results were expressed as picomoles (pmol) per mg tissue. Three tissue samples from each heart were homogenized, and the resulting membrane and cytosol fractions were analyzed in duplicate in the lipid peroxide assay. Triplicate aliquots of each lipid peroxide assay reaction mixture were assayedfor GSSG. Critique qf the lipid peroxideasscp! In the O-4500 pmol per reaction mixture range the lipid peroxide assayis linear. Recovery of added tert-butyl peroxide from tissue using this assay is 80%. Reaction blanks do not vary by more than 10%. Blanks containing GSH and glutathione peroxidase but no lipid peroxide have the samelow GSSG values as standards with lipid peroxide and GSH, but no glutathione peroxidase. This confirms that GSSG formation is due to a reaction catalyzed by glutathione peroxidase between GSH and lipid peroxide, rather than nonspecific oxidation of GSH. In separate experiments on cytosol fractions, where glutathione peroxidase was omitted, all measured GSSG was accounted for by GSSG present in the fraction prior to lipid peroxide assay (“ a GSSG effective” = 0). This again confirms that non-specific oxidation of GSH did not occur in the lipid peroxide assay. Since lipid peroxide levels were not significantly increased in the H,O,-ADP group, it is unlikely that H,O, trapped in myocardimm explains the elevated lipid peroxide levels found in the IRON group. Iron levels were measured in cytosol fractions from hearts in the IRON and CON groups. Measurement of iron was performed by quantitation of absorbance at 429 nm of the iron-saturated complex ferrioxamine [28J. An excess of deferoxamine (50 mM) was added to cytosol fractions. Cytosolic iron levels were elevated in the IRON group (22.3 f 6.4pM, n= 3) compared to the CON group (6.1&6. I PM, n=3). To assure that increased cytosolic iron did not contribute to lipid peroxide formation during assayprocrdures or tissue processing, cytosol fractions from ADP-H,O, treated hearts were assayed
1034
E. J. Lesnefsky
in the presence of added iron, ( 10 ,ul of 160 PM FeCl, added to the cytosol fraction) in order to increase iron levels to those present in the IRON hearts. Tissue lipid peroxide levels were not increased by the additional iron (PcO.45, 72~6, paired t test). Reactivity offree fat0 acid and phospholipid peroxides The reactivity of free fatty acid hydroperoxides (ROOH) and phospholipid hydroperoxides (PLOOH) in the lipid peroxide assay using bovine erythrocyte glutathione peroxidase (Sigma) [11] was determined. Phosphatidyl choline (PC) (Sigma, 500 pg, purity confirmed by thin layer silica gel chromatography with hexane:diethyl ether:acetic acid 83:15:2 as solvent) was dispersed in 1 .O ml 5 mM HEPES buffer, pH 7.4, 0.15 M NaCl, containing BHT as an antioxidant. When unperoxidized PC was assayed for lipid peroxides [II], none was detected as expected. PC was then hydrolyzed with 10 U porcine pancreas phospholipase A, in the presence of 10 mti Ca” and 1.25 mM Na-deoxycholate at 37°C for 30 min. BHT and argon flushing of the headspace were utilized to prevent oxidation. The reaction was stopped by 20 mM EGTA. No lipid peroxides were detected in the hydrolysate. Thin layer chromatography confirmed release of fatty acids from PC. PC was then photo-oxidized using a fluorescent lamp with methylene blue as a photosensitizer. When the photo-oxidized PC was assayed for lipid peroxides, again none were detected. Thin layer chromatography confirmed that no free fatty acids were present. Photo-oxidized PC was then hydrolyzed by phospholipase A,. Lipid peroxides were now easily detectable. Approximately 35% of the maximal theoretical yield of ROOH from the SN-2 position (arachidonic acid) was obtained. Thin layer chromatography showed that the photo-oxidized phospholipase-treated sample now had easily demonstrable levels of free fatty acids present. These experiments show that the glutathione peroxidase-based lipid peroxide assay detects ROOH, but not PLOOH. Other biochemical assays Reduced
(GSH)
and oxidized
(GSSG)
gluta-
et al.
thione were measured in myocardium by the method of Sies adapted for use in a multiwell platereader [12, 171. Adenosine triphosphate (ATP) and phosphocreatine (PCR) were measured by the flourometric method of Lowry and Passonneau [I$]. Tissue protein was measured by the dye binding method of Bradford [20]. Statistical methods Data were expressed as mean (S.E.). Comparisons among chemical and hemodynamic one way analysis of variance the Student-Newman-Keuls vidual comparisons. Within sons of hemodynamic results by two way ANOVA [Zl]. considered significant.
f standard error groups of bioresults were by (ANOVA), with test used for indigroup compariover time were A P~0.05 was
Results Lipid peroxide levels were increased in both cytosol (PC 0.01) and membrane (PC 0.02) fractions in the iron-H,O, (IRON) group compared to saline control (CON) and H,O,ADP (ADP) control groups (Table 2). Lipid peroxide levels in the ADP group (no iron) were similar to those in the CON group (P= NS) . Similar low lipid peroxide levels in the CON and ADP hearts suggest that the catalytic effect of iron was important. The similar levels in the CON and ADP groups do not support a possible reaction between H,O, and GSH in the lipid peroxide assay as a cause of the elevated lipid peroxide levels found in the IRON group. Furthermore, excess iron added directly to the tissue fractions did not increase lipid peroxide formation, confirming that lipid peroxidation during sample processing did not occur. The increase in both cytosol and membrane lipid peroxides in the IRON hearts was the most prominent biochemical evidence of oxidative injury in these hearts. However, there was other tissue evidence suggestive of oxidative injury in IRON hearts. ATP levels (P~0.05) were decreased. The IRON group also had modest trends toward elevated tissue GSSG and decreased tissue GSH, GSH/GSSG ratio and phosphocreatine (PCR) (Table 2), though these were not statistically significant.
Myocardiai
Lipid
1035
Peroxidation
TABLE 2. Biochemicalresults Iron-E&O, Cytosol LP (prno~~rn~ tissue) Membrane LP (pmol/mg tissue) GSH (nmol/g tissue)
GSSG (nmol/g tissue) GSH/GSSG :YTP jnmol/mg prot) PCR (nmol/mg prot) Protein (mg/g tissue)
(n = 5)
64.7 f 7.5’ 10.9*4.7 798 rk 93.0 27.4f 11.0 67 f 24.0 16.2f2.52 39.3f2.6 66.9zt 7.5
Saline (n = 6)
H,O,-ADP In= 5)
6.4 i 3.0 0.1 fO.l 92 1* 64.0 9.7fO.7 96f 7.0 27.9f 3.6 56.8 jc 8.2 67.7i4.6
10.1 f 1.7 0.2JtO.l 662 + 50.0 6.5iO.8 103f 14.0 21.91 1.7 51.3rf: 11.8 71.1 f5.1
’ P< 0.05 “S otlw groups. “PC 0.05 vs saline.
Tissue GSH tended to be lower (P=O.O6) in the ADP group than in CON (Table 2). Administration of 10~~ H,O, oxidizes GSH to GSSG. The GSSG formed is then actively transported from the myocyte [ZZ, I#]. This maintains normal tissueGSSG at the cost of a reduction in tissue GSH [lZ], since GSSG is released rather than reduced to GSH. Myocardial GSH depletion was lessprominent in the IRON
group,
possibly
since most of the
infused H,O, reacted with the excess iron to form ‘OH. Hydroxyl radical reacts rapidly with most adjacent moleculesincluding unsaturated lipids [.?2]. In contrast, H,02 is poorly reactive with unsaturated lipids and other biomolecules [22], but diffuses into the cytosoi and oxidizes GSH [12, .?Z]. Thus, more H,O, may have reached the cytosol and oxidized GSH in the ADP group than in the IRON group, accounting for the lower tissue GSH values. ~et~~xi~cation of intracellular H,O, by GSH may also account for the absenceof an increase in lipid peroxides in the ADP group. It is likely that H,O, was eliminated by reaction with GSH before it could react to form ‘OH in this group, especially since levels of “free” redox-active iron are very low in the non-ischaemic heart [ZZ, 231. Thus, reduction ofH,O, to water by GSH rather than reaction of H,O, with ‘
