INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE INVOLVEMENT OF REACTIVE oxygen species in postischemic injury has received considerable attention. The requirement for redox-active transition metals in this process has been inferred from the prerequisite for catalytic traces necessary for formation of the more destructive species, such as hydroxyl radicals, from less reactive species, such as superoxides. By far, the transition metal that has received the most attention is iron. Numerous studies have demonstrated increased mobilization of iron during ischemia (24, 47) and that affecting iron reactivity by increasing tissue levels enhances postischemic injury (38) or that, conversely, chelating iron with deferoxamine (DEF) decreases postischemic injury (9). The ability of Cu to promote this pathological process has not been studied to the same extent. We have demonstrated that increasing tissue content of Cu renders the isolated perfused heart more sensitive to postischemic injury (36). Chevion et al. (5) have demonstrated that after prolonged ischemia Cu may become mobilized and available for catalysis. To our knowledge, this is the extent of research into this problem.

A difficulty inherent in the study of the role of Cu in postischemic injury is the measurement of the redox-active or unbound form of this metal. Redox-active iron has been demonstrated to exist as low-molecular-weight compounds bound to phosphate esters of ATP, ADP, or GTP, and to organic acids, such as citrate (11, 15), and is easily measured as such. To the contrary, intracellular redox-active Cu is generally not found as a low-molecular-weight species but tends to be attached to macromolecular structures, such as DNA, carbohydrates, enzymes, peptides, and proteins, thus possibly rendering it inaccessible to detector molecules (6, 16). However, the existence of this pool can be inferred, as Cu is known to catalyze a variety of oxidative phenomena including, but not limited to oxidation of lipoproteins (30), DNA (8), and site-specific modifications to amino acids, proteins, and enzymes (32, 46).

The ability of Cu to promote site-specific oxidations in macromolecular structures has led several investigators to propose a novel hypothesis to hinder this process (39, 42). According to this hypothesis, zinc, by virtue of its similarities in coordination chemistry to Cu, is used to displace this metal from low-affinity binding sites on macromolecular structures. Zinc has been shown to compete effectively for Cu²⁺-site-specific binding in several heme-proteins (25, 41) and because it is nonredox-active has been found to interfere with Cu-mediated site-specific oxidations to DNA (22). We have used this property of zinc to protect the isolated perfused rat heart (35) and, more recently, the in vivo swine heart (38) from postischemic oxidative injury. We propose that zinc might represent a means of displacing redox-active Cu from low-affinity binding sites on macromolecular structures into the aqueous phase where it might be more readily accessible for detection. In this study, we demonstrate the existence of an intracellular pool of redox-active Cu by using zinc and metal chelators to displace the metal from its intracellular low-affinity binding sites. In addition, we demonstrate possible biological relevance of this pool of Cu as an important mediator of tissue oxidative injury.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals. All studies were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (NIH publication No. 85-23, Revised 1985) and were approved by the Institutional Animal Care and Use Committee. Male Sprague-Dawley rats (275-450 g) were obtained from Charles River Laboratory (Wilmington, MA) or Taconic Farms (Germantown, NY) and were allowed at least 3 days of in-house acclimatization before experimental use. During this time, all animals were allowed ad libitum access to Purina lab chow (Ralston Purina, St. Louis, MO) and water.

Chemicals and reagents. Sodium bicarbonate, sodium chloride, potassium chloride, HEPES, magnesium sulfate, magnesium chloride, D-(+)-glucose, calcium chloride, zinc sulfate, histidine and DEF were obtained from Sigma Chemical (St. Louis, MO). Pepstatin, aprotinin, leupeptin and phenylmethylsulfonyl chloride were obtained from Boehringer Mannheim (Indianapolis, IN). Sodium heparin and pentobarbital sodium were obtained from the North Shore University Hospital pharmacy. Therapeutic grade 95% O₂-5% CO₂ was obtained from General Welding Supply (Westbury, NY). All other reagents were of laboratory grade and obtained from reputable sources.

Perfused heart preparation. Rats were injected with sodium heparin (500 U ip) 30 min before being anesthetized with pentobarbital sodium (60 mg/kg ip). Hearts were removed rapidly and placed in ice-cold heparinized saline. The hearts were then orthogradely perfused through the coronary arteries as previously described at a constant pressure of 95 cmH₂O.

Perfusate. The perfusate was a modified Krebs-Henseleit (KH) buffer consisting of (in mM) 118 NaCl, 6.1 KCl, 2.5 CaCl₂, 1.2 MgSO₄, 25 NaHCO₃, 1.0 HEPES, and 11.1 glucose. Complete buffer was prepared the day of the experiment by mixing the proper amounts of concentrated stock solutions to which was added the appropriate quantity of glucose and calcium chloride. All concentrated solutions, with the exception of the magnesium sulfate, were treated with iminodiacetic acid chelating resin beads, 50-100 mesh (Chelex 100, Bio-Rad, Hercules, CA) obtained from Sigma Chemical, as previously described. Zinc was added to the buffer as the zinc-bis-histidinate (Zn-His₂) complex (1 zinc:2 histidines) and was prepared daily (35).

Indexes of cardiac function. There were five indicators of cardiac function determined in this study. Left ventricular systolic pressure development and end-diastolic pressure were determined by way of a latex balloon (0.1 ml) that was expanded to exert a physiological end-diastolic pressure of 5 mmHg, as previously described (35). Developed systolic pressure or pulse pressure was calculated as the peak systolic pressure minus the end-diastolic pressure. The rate-pressure product was calculated as the heart rate multiplied by the developed systolic pressure and is expressed as mmHg/min. Coronary flow was determined by a timed collection of coronary effluent (data not shown). Heart rate was calculated from the R-to-R interval of the electrocardiogram (data not shown).

Exclusion criteria. Hearts were excluded from the study if they failed to maintain developed systolic pressure of at least 70 mmHg, or a heart rate of at least 220 beats/min during the 10-min pretreatment equilibration period. Furthermore, hearts were excluded if a persistent arrhythmia was present during the equilibration period.

Biochemical assays. Protein carbonyls were analyzed using a commercially available kit (OxyBlot, ONCOR, Gaithersburg, MD) that is based on the immunoblot technique described by Shacter et al. (43). Basically, cardiac tissue was homogenized in HEPES suspension buffer containing protease inhibitors (leupeptin, 5 µg/ml; aprotinin, 5 µg/ml; pepstatin, 7 µg/ml; and phenylmethylsulfonyl fluoride, 40 µg/ml) and then centrifuged at 10,000 g to obtain the soluble fraction. An aliquot of the soluble fraction was then treated with dinitrophenylhydrazine supplied with the kit. After neutralization, the proteins (4 µg) were separated on a 4-20% gradient gel (Ready Gel, Bio-Rad) using standard SDS-PAGE techniques. The separated proteins were then translated onto a polyvinylidene difluoride membrane, which was then incubated with the antibodies supplied in the kit. Autoradiography film was then exposed to the membrane that had been treated with chemiluminescent reagents (Renaissance, NEN Life Sciences, Boston, MA). The film was then developed and the bands quantitated using computer-assisted densitometry (Molecular Analyst Hardware and Software, Bio-Rad).

Statistical analysis. Analysis of the differences between multiple groups was made with a repeated-measures ANOVA (RMANOVA) in which the within factor was time. Differences between two individual groups were analyzed with an independent Student's t-test for independent variables. In all cases, results were considered to be significant at the P < 0.05 level. All statistics were performed with the SPSS/PC+ (SPSS, Chicago, IL) statistical analysis package.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cardiac protection by zinc, deferoxamine, and neocuproine. Studies examined the relative efficacies of zinc, neocuproine (NEO), and DEF to affect postischemic injury in the Langendorff preparation. In hearts perfused under experimental conditions, postischemic function, as measured by systolic pressure development, the rate-pressure product, and end-diastolic pressure, was significantly improved (P < 0.05, RMANOVA) in the presence of DEF and both concentrations of Zn-His₂ (Fig. 1), a result that is in general agreement with what has been previously published (35, 47). We have previously shown that histidine, at concentrations up to 100 µM, has no effect on recovery of postischemic function (35). In contrast to a previous study in the Langendorff regional ischemia model (2) that uses arrhythmias as an endpoint, NEO had no effect in this model of global ischemia (Fig. 1).

View larger version (38K): [in this window] [in a new window]   Fig. 1.   Effect of zinc, deferoxamine (DEF), and neocuproine (NEO) on postischemic hemodynamic recovery. Isolated rat hearts were perfused with Krebs-Henseleit buffer (KH) alone (), or KH plus 20 () or 30 () µM Zn-His2, or KH plus 0.5 mM DEF (), or KH plus 42 µM NEO () for 20 min, then subjected to 20 min of normothermic global ischemia, followed by 30 min of reperfusion with same buffer. Developed systolic pressure (A), the rate-pressure product (B), and end-diastolic pressure (C) were monitored throughout. Values are means ± SE of 6-12 observations.

Excretion of Cu and iron from the heart. Initial studies tested the hypothesis that zinc, as the Zn-His₂, can displace Cu from low-affinity binding sites on macromolecular structures into the aqueous phase. In the presence of 30 µM Zn-His₂, the rate of Cu excretion from hearts into coronary venous effluent was increased by 225% during the initial preischemic loading period (Fig. 2A). In the immediate postischemic interval, there was a further increase in the rate of excretion that represents Cu that is displaced during the period of stop-flow ischemia (ischemia is represented by the break in the axis), which then returns to preischemic levels (Fig. 2A). At all determination points, the rate of Cu release was significantly higher than initial prezinc values in the treated hearts (P < 0.05, RMANOVA). The redox nature of the excreted metal was examined by determination of its reactivity toward ascorbate. Despite the presence of zinc, which can inhibit this reaction (29), it is clear that this Cu is redox-active as there is a substantial rate of ascorbate oxidation, both pre- and postischemic (Fig. 2C). What is most striking is that both the rates of Cu excretion and ascorbate oxidation increase as soon as zinc is added, providing the first evidence for the existence of an intracellular pool of Cu that can be considered to be catalytic in nature. Histidine (100 µM) had no effect on Cu excretion (Fig. 2A).

View larger version (42K): [in this window] [in a new window]   Fig. 2.   Excretion of Cu and iron from the isolated perfused rat heart. Isolated rat hearts were perfused with KH buffer alone (), or KH plus 30 µM Zn-His2 (), or KH plus 0.5 mM DEF (), or KH plus 42 µM NEO () for 20 min, then subjected to 20 min of normothermic global ischemia, followed by 30 min of reperfusion with same buffer. At the indicated times, aliquots of pulmonary artery effluent were analyzed for Cu (A) or iron (B) using atomic absorption spectroscopy. Additionally, redox activity of the excreted metal was examined by determining reactivity toward ascorbate (C). Values are means ± SE of 6-10 observations.

Effects on protein carbonyls. Studies were conducted to analyze protein carbonyls, a product of protein oxidation, using an immunoblotting technique (OxyBlot, ONCOR). This technique is more sensitive and specific than the spectrophotometric method that is routinely used by many investigators. The protein ladder for these determinations is supplied in the kit and consists of phosphorylase B (97.4 kDa), bovine serum albumin (68 kDa), ovalbumin (43 kDa), carbonic anhydrase (29 kDa), and trypsin inhibitor (21 kDa), all conjugated with dinitrophenylhydrazine to react with the antibody. Computer-assisted densitometry was used to analyze four protein bands that were consistently present at the same location in all samples. The molecular weights of these proteins correspond to ~29 kDa, 36 kDa, 43 kDa, and 85 kDa (see arrows in Fig. 3). The identity of these proteins is not known at this time, and they do not correspond with those in the ladder. Exposure to 20 min of ischemia results in a clear increase in protein oxidation, a result consistent with previous studies (34) (Table 1). As shown in Table 1 and illustrated by Fig. 3A, the optical density of the protein bands in the zinc-treated hearts appears to be intermediate between the nonischemic and ischemic hearts. The effect of DEF on protein oxidation is tabulated in Table 2. Unlike Zn-His₂, DEF appears to increase protein carbonyl formation, in particular, the 29 kDa, 43 kDa, and 85 kDa proteins. This is particularly evident in the representative immunoblot depicted in Fig. 3B. This was an unexpected observation, the possible significance of which is discussed in the following section.

View larger version (78K): [in this window] [in a new window]   Fig. 3.   Representative immunoblot illustrating the effect of zinc (A) and DEF (B) on protein oxidation during ischemia. Isolated rat hearts were equilibrated for 10 min with Krebs-Henseleit buffer, then perfused with buffer ± 30 µM Zn-His2 (A) or 0.5 mM deferoxamine (B) for 20 min, and finally subjected to 20 min of normothermic global ischemia (total time = 50 min). A subset of hearts were perfused with Krebs-Henseleit buffer for an equivalent period of time (50 min) (nonischemic control). Hearts were then analyzed for protein carbonyls using SDS-PAGE followed by immunoblotting.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present studies demonstrate the importance of Cu as a catalyst of postischemic cardiac oxidative injury. Zinc preferentially displaced Cu, but not iron, from intracellular binding sites. This process might be related to some unique properties of the Zn-His₂ complex as the formation constant of the bis-histidinate complex of Cu is six orders of magnitude higher than that of zinc (18.33 vs. 12.88, respectively) (3). This raises the possibility that a ligand exchange reaction occurs in which zinc exchanges with Cu, at the metal-specific site, which then complexes with histidine. This appears not to be true for iron, as the formation constant of the histidinate complex (9.3) is much lower (3). The other possibility is that iron, being mostly complexed with low-molecular-weight ligands, may not take part in this type of reaction. From the point of view of postischemic functional preservation, interference with the catalytic properties of this pool of intracellular Cu may be beneficial, since in chemical systems Cu is 60-fold more effective than iron in promoting · OH formation from H₂O₂ (k₂ iron for H₂O₂ = 76 mol¹ · s¹; k₂ Cu for H₂O₂ = 4.7 × 10³ mol¹ · s¹)(21). This disparity in reactivity between Cu and iron was most clearly illustrated by a study (45) that extended our previous report examining the Cu-ascorbate free radical generating system (37), with the exception that Fe-ascorbate was examined. Approximately 50 times more iron was necessary to produce cardiac dysfunction that was approximately equivalent to what we observed with 1 µM Cu. Treatment with zinc resulted in a net efflux of redox-active Cu; thus postischemic cardiac function was preserved, and an index of oxidative damage, i.e., protein oxidation, was decreased. On the other hand, NEO, a known Cu chelator, had no effect on metal efflux, and its effect on postischemic function was indistinguishable from control conditions. Initial studies with the spectrophotometric protein carbonyl assay were unable to detect an effect of NEO on postischemic protein carbonyl formation (data not shown). This result, although disappointing, may be related to lack of cell penetration by this compound as well as to the effects of concentration. Observations that limiting Cu redox-activity might prevent protein oxidation are in agreement with studies demonstrating that ceruloplasmin can inhibit carbonyl formation in cellular proteins (26).

The observation that DEF resulted in a net efflux of Cu was not unexpected. DEF has a formation constant for iron of 10³², however, its formation constant for Cu is, at 10¹⁴, still very significant (1). When iron stores are depleted or available iron is totally bound, DEF would be free to chelate Cu. What was totally unexpected was the effect of DEF on iron efflux, as no change was detected. However, this phenomenon has been observed in other organ preparations, such as the isolated lung. A similar treatment with DEF results in an increase in tissue DEF-bound iron, yet no increase in perfusate content of iron (personal communication from Dr. Aron B. Fisher, University of Pennsylvania). These results raise some doubts as to whether DEF should be used as a "probe" for iron-mediated oxidative injury. DEF-bound iron is known to increase during ischemia, but as can be inferred from our studies, this complex does not exit the cell to any great extent. Perhaps the reason that intracellular DEF-bound iron increases during ischemia is that the cell membrane is impenetrable, or only slightly penetrable, by this complex. If so, the use of DEF to determine increases in iron during ischemia (24) may actually overestimate the degree of the change. Moreover, the significance of even this supposedly "inactive" form of iron accumulating within the cell is unclear. This is particularly troublesome in light of reports suggesting that under certain conditions DEF may exhibit prooxidant properties through autoreduction of bound ferric iron resulting in formation of redox-active ferrous iron or possibly redistribution of nonredox-active, sequestered iron (4, 17, 23). A possible explanation for the observed increase in protein oxidation is that DEF-treatment resulted in redistribution of iron to protein binding sites from nonprotein binding sites. This explanation becomes more plausible in light of reports that infusion of iron into a DEF-treated isolated heart increases · OH production (31) and that increasing the soluble iron content of cardiac tissue, although enhancing postischemic injury, does not in itself increase protein oxidation (28). These observations and published reports suggest that oxidative protein damage is not the only explanation for postischemic myocardial dysfunction. In addition, a combination of other factors, including removal of redox-active Cu, as well as the reported nonspecific free radical scavenging effect of DEF, must account for its observed sparing of postischemic cardiac function.

These studies confirm the existence of an intracellular pool of redox-active Cu that appears to play an important role as a mediator in postischemic cardiac oxidative injury. The existence of this intracellular pool of Cu has proven difficult to confirm. Various investigators have attempted to quantify this pool under both physiological and pathological conditions. One of the principal techniques that has been used to quantify this pool has been the Cu-phenanthroline-dependent degradation of DNA method developed by Gutteridge (13). With the use of this technique, significant redox-active or phenanthroline-detectable Cu has been measured in human sweat (13, 19), cerebrospinal fluid (13, 14), and in freeze-thawed (13), but not fresh, plasma (10, 13) despite reports that up to 5% of human plasma Cu is bound to albumin (7). Under pathological conditions, phenanthroline-detectable Cu has been measured in cerebrospinal fluid from patients suffering from Parkinson's disease (33), in synovial fluid from rheumatoid patients (13), and in freshly prepared plasma from some patients suffering from fulminant hepatic failure, but curiously not in plasma from patients with uncomplicated Wilson's disease (10). The common element in most of these studies is not the use of the phenanthroline assay, but rather the type of sample that was analyzed. Those samples in which Cu was readily detected represent, in the case of cerebrospinal fluid, synovial fluid, and sweat, cell-free protein-deficient ultrafiltrates to which the detector was added. In fresh plasma, which contains high amounts of protein, phenanthroline-detectable Cu is difficult to measure (10, 13). Only when plasma proteins would be depleted, such as in fulminant hepatic failure (10), or when the proteins might be degraded, as in freeze-thawed plasma (13) was Cu detectable. To our knowledge, the presence of phenanthroline-detectable Cu has never been demonstrated in freshly prepared organ or tissue homogenates or extracts. Using the phenanthroline assay, we have been unable to consistently determine this pool of Cu in freshly prepared heart homogenates and ultrafiltrates (unpublished results).

The assumption made when trying to detect any intracellular substance is that it is available to the detector, an assumption which may be erroneous in the case of intracellular redox-active Cu. Intracellular redox-active iron has been extensively studied, and a significant intracellular pool has been demonstrated to exist as a low-molecular-weight species bound to phosphate esters of ATP, ADP, or GTP and to organic acids, such as citrate (11, 15). Intracellular iron in this state is clearly redox active (15) and is accessible to, and detectable by, a variety of detectors, such as bleomycin (18) and ferrozine (12). To the contrary, intracellular redox-active Cu is generally not found in a low-molecular-weight state but tends to be attached to macromolecular structures, such as DNA, carbohydrates, enzymes, peptides, and proteins (6, 16). Cu bound to macromolecular structures might not be accessible to detector molecules, thus accounting for the inability to consistently measure this pool.

The potential mediatory role of Cu in ischemic oxidative injury has been a subject of much debate. The results of these studies suggest that removal of redox-active Cu before ischemia may be beneficial. Both Zn-His₂ and DEF promoted a net efflux of Cu and improved postischemic cardiac function. On the other hand, NEO had no effect on metal efflux or postischemic function. Although these studies may not be conclusive, when considered in light of our previous studies (35-37) they strongly implicate Cu as a major catalyst of postischemic cardiac oxidative injury.