First human studies with a high-molecular-weight iron chelator.

The release of free, reactive iron from cellular iron stores has been implicated as an important contributor to tissue damage in a variety of clinical situations, including ischemia and reperfusion injury, hemorrhagic shock, and burn injury. Deferoxamine mesylate (DFO), the only iron chelator currently approved for clinical use, is used for the treatment of iron overload, including acute iron poisoning and treatment of chronic iron overload in transfusion-dependent anemias such as beta-thalassemia. However, it is not suitable for acute care situations because of its toxicity, primarily hypotension when given at high intravenous doses, and its short plasma half-life. We have produced a high-molecular-weight iron chelator by chemically coupling DFO to hydroxyethyl starch. This novel chelator (HES-DFO) was administered to healthy male subjects by intravenous infusion over a 4-hour period. The drug was well tolerated, and signs of DFO acute toxicity were not observed. Maximum plasma chelator levels of approximately 3 mmol/L were achieved with HES-DFO, which is more than an order of magnitude higher than has been reported with injections of DFO. Drug residence time in plasma was markedly prolonged, with an initial half-life of 22 to 33 hours. Urinary iron excretion was 7.1 +/- 2.2 mg in 48 hours in the highest dose group, as compared with 0.06 +/- 0.15 mg in control subjects who received normal saline infusions. Intravenous infusion of HES-DFO is well tolerated, produces substantial and prolonged plasma chelator levels, and markedly stimulates urinary iron excretion.

The long-acting parenteral iron chelator, hydroxyethyl starch-deferoxamine, fails to protect against alcohol-induced liver injury in rats.

We studied the effect of the long-acting parenteral iron chelator, hydroxyethyl starch deferoxamine (HES-DFO) on liver nonheme iron, lipid peroxidation and pathologic changes in the liver in the intragastric feeding rat model for alcoholic liver disease. Male Wistar rats (225-250 g) were fed liquid diet and ethanol for 2 months. In control pair-fed animals, ethanol was isocalorically replaced by dextrose. Two additional groups of animals (dextrose and ethanol fed) received HES-DFO (25 mg deferoxamine equivalents/kg, three times a week). The blood ethanol level in the ethanol-fed animals was maintained between 150 and 350 mg/dl. For each animal, the levels of hepatic nonheme iron, lipid peroxidation and pathologic changes were evaluated. Ethanol administration caused fatty liver, necrosis and inflammation. Addition of HES-DFO to the ethanol diet increased the severity of pathologic changes, particularly necrosis and inflammation. The nonheme iron in alcohol-fed animals was significantly higher (18.3 +/- 4.3 microg liver) than in pair-fed dextrose controls (12.5 +/- 1.5 microg, P < .05). Addition of HES-DFO significantly increased nonheme iron levels in the dextrose-fed rats (17.1 +/- 2.0 microg/g, P < .02) but not in ethanol-fed rats (20.0 +/- 2.0). Ethanol increased levels of conjugated dienes; these levels were not altered by HES-DFO. The most significant observations in this study were: 1) the higher hepatic nonheme iron content in ethanol-fed rats compared with pair-fed dextrose controls; 2) the absence of changes in hepatic nonheme iron levels or lipid peroxidation in ethanol-fed groups treated with HES-DFO; and 3) the worsening of liver injury in ethanol-fed rats by HES-DFO.

Polymer conjugation reduces deferoxamine induced retinopathy in an albino rat model.

PURPOSE: The iron chelating agent deferoxamine mesylate USP (Desferal, Ciba, Summit, NJ) is commonly used in the treatment of acute iron intoxication and chronic iron overload (associated with the transfusion-dependent anemias). When used for prolonged periods of time or in high doses deferoxamine is attended by a range of ocular toxicities. The visual symptoms associated with deferoxamine administration often limit effective iron chelation therapy and can result in permanent vision loss. Deferoxamine has recently been conjugated to certain high molecular weight biocompatible polymers without altering its iron-binding properties. Here the effect of conjugation of deferoxamine to hydroxyethyl starch on retinal toxicity is examined. METHODS: An albino rat model of electroretinographically determined, deferoxamine-induced retinal toxicity has been previously described. We use this model to evaluate and compare both native deferoxamine and hydroxyethyl starch conjugated deferoxamine. RESULTS: Our data show that retinal function, as assessed by the electroretinogram b-wave, is significantly depressed 1 day after a single dose of native deferoxamine, while the b-waves of rats receiving a single dose of hydroxyethyl starch-deferoxamine, are not significantly depressed at any time during the study. In addition, the administered dose of hydroxyethyl starch-deferoxamine resulted in plasma deferoxamine concentrations up to five times greater than those achieved with native deferoxamine. CONCLUSION: These results suggest that hydroxyethyl starch conjugated deferoxamine is associated with less retinal toxicity than native deferoxamine and that it may be a safer alternative for iron chelation therapy.

Role of iron and oxygen radicals in hemorrhage and shock.

This contribution focuses on the role of iron as a critical component in the genesis of oxygen radical mediated tissue injury occurring after global ischemia associated with severe hypovolemic shock. Conventional colloid or crystalloid fluid resuscitation does not adequately protect organs susceptible to reperfusion injury. One approach aimed at attenuating such post-trauma reperfusion injury is systemic, high dose, iron chelation used in combination with colloid fluid replacement.

Effects of hydroxyethyl starch conjugated deferoxamine on myocardial functional recovery following coronary occlusion and reperfusion in dogs.

The effect of hydroxyethyl starch-conjugated deferoxamine (HES-DFO) on the recovery of regional myocardial function after 15 min of coronary artery occlusion followed by 3 h of reperfusion of the left anterior descending coronary artery (stunned myocardium) was investigated in anesthetized dogs. Regional myocardial blood flow was measured by radioactive microspheres and regional myocardial segment shortening (%SS) by sonomicrometry. HES-DFO (equivalent of 50 mg/kg DFO), iron saturated HES-DFO (HES-FO), deferoxamine (DFO, 50 mg/kg), or saline were administered by intravenous infusion starting 30 min before occlusion and throughout occlusion. Ischemic bed size and collateral blood flow were similar in all four groups. HES-DFO significantly improved %SS in the ischemic-reperfused region during reperfusion; however, HES-FO and DFO had no effect on %SS as compared to the saline-treated group. HES-DFO and HES-FO had no effect on hemodynamics; however, DFO produced a marked reduction in systemic blood pressure. Since HES-FO had no effect on the recovery of %SS, the beneficial effect of HES-DFO is thought to be due to its iron chelating characteristics. Plasma concentrations of HES-DFO not only reached a higher peak level but also had a longer half life (3 h) than that of regular DFO (20 min). Thus, high-molecular-weight HES-DFO is effective in enhancing the recovery of regional wall motion in stunned myocardium. The reason for the lack of efficacy of DFO compared to HES-DFO at this high dose may be related to the formation of a toxic deferoxamine free radical species.

Hemodynamic effects of intraatrial administration of deferoxamine or deferoxamine-pentafraction conjugate to conscious dogs.

Deferoxamine (DFX) is a specific Fe3+ chelator that is used to manage iron overload, and is being evaluated as an agent to reduce ischemic organ damage that involves iron-mediated OH formation. However, high intravascular doses cause significant hemodynamic changes that may limit or counteract beneficial effects. We used conscious, closed-chest dogs to test the hypothesis that conjugating DFX to pentafraction, a high molecular weight fraction of pentastarch, could reduce such hemodynamic changes. We infused 50 mg/kg of body weight of native DFX, or an equivalent dose as DFX-pentafraction, intraatrially over 15 min. Within 10 min of starting the infusion. DFX increased heart rate from pre-drug values of 105 +/- 11 (mean +/- SEM; N = 9) to 158 +/- 13 beats/min, and reduced left ventricular (LV) systolic pressure from 131 +/- 3 to 99 +/- 16 mm Hg, LV end-diastolic pressure from 12 +/- 3 to 3 +/- 3 mm Hg, and mean arterial pressure (MABP) from 101 +/- 5 to 74 +/- 13 mm Hg. In two dogs, MABP decreased to less than or equal to 35 mm Hg. These parameters returned to predrug values by 60 min after infusion. All of these changes were statistically significant (p less than 0.05). In contrast, infusing DFX-pentafraction (N = 9) caused no significant cardiac or hemodynamic changes other than a transient and slight (approximately 7%) increase in systolic arterial pressures. This conjugate, which prolongs the plasma half-life and does not alter the iron-chelating activity of native DFX, eliminates many undesirable hemodynamic actions. It may be a useful therapeutic alternative to native DFX in some settings.

High-dose iron-chelator therapy during reperfusion with deferoxamine-hydroxyethyl starch conjugate fails to reduce canine infarct size.

Iron catalyzes reactions during ischemia and reperfusion that contribute to myocardial injury. The iron-chelator deferoxamine blocks these reactions, but undesirable side effects limit the clinical potential of deferoxamine to decrease injury. We tested whether intravenous (i.v.) administration of high doses of a well-tolerated deferoxamine-hydroxyethyl starch (DEFHES) iron-chelator during the last 10 min of ischemia and the first 10 min of reperfusion would decrease canine infarct size. Fourteen chloralose-anesthetized mongrel dogs were randomized to therapy in a blinded fashion with deferoxamine conjugate (75 mg/kg deferoxamine) or hydroxyethyl starch (HES) vehicle alone. Nine other untreated dogs served as controls. Infarct size as a percentage of area at risk (MI/RISK) was not reduced by therapy with deferoxamine conjugate. The deferoxamine dose was five times greater than the maximally tolerated dose of free deferoxamine. Arterial deferoxamine concentrations in the deferoxamine-conjugate group were 0.69 +/- 0.09 mM at onset of reperfusion and 1.37 +/- 0.05 mM at 10 min of reperfusion. Area at risk, ischemic collateral blood flow, and heart rate-blood pressure (HR/BP) product were similar in the groups. Chelation of intravascular iron at the time of reperfusion does not reduce myocardial necrosis in an in vivo model of myocardial ischemia-reperfusion injury.

Modulation of deferoxamine toxicity and clearance by covalent attachment to biocompatible polymers.

A class of high molecular weight iron chelators has been prepared by covalently attaching deferoxamine (DFO), by its amino group, to a variety of biocompatible polymers such as dextran and hydroxyethyl-starch. The iron-binding properties of DFO are virtually unchanged after the attachment procedure, but the toxicity and circulatory half-life are profoundly altered. Competitive iron-binding experiments indicate that the conjugates retain a high affinity for ferric iron. In addition, the derivatives inhibit iron-driven lipid peroxidation as effectively as the parent drug. However, the LD50 in mice (based on DFO equivalents) is approximately 4000 mg/kg for dextran-DFO as compared to 250 mg/kg for free DFO. Consistent with the greatly decreased LD50, intravenous administration of the conjugates in dogs at a dose of 100 mg/kg (body weight) does not cause the severe hypotension associated with intravenous administration of DFO. The plasma half-lives of these adducts are increased greater than 10-fold for dextran-DFO and hydroxyethyl-starch-DFO compared to the free drug. Finally, and most importantly, the conjugates are effective in mediating in vivo iron mobilization and excretion. Because recent evidence implicates iron as an important component of tissue injury in many disease states, these high molecular weight iron chelators may have potential for improved therapy, allowing higher sustained plasma concentrations of the active drug.

Acute iron poisoning. Rescue with macromolecular chelators.

Acute iron intoxication is a frequent, sometimes life-threatening, form of poisoning. Present therapy, in severe cases, includes oral and intravenous administration of the potent iron chelator, deferoxamine. Unfortunately, high dose intravenous deferoxamine causes acute hypotension additive with that engendered by the iron poisoning itself. To obviate this problem, we have covalently attached deferoxamine to high molecular weight carbohydrates such as dextran and hydroxyethyl starch. These macromolecular forms of deferoxamine do not cause detectable decreases in blood pressure of experimental animals, even when administered intravenously in very large doses, and persist in circulation much longer than the free drug. These novel iron-chelating substances, but not deferoxamine itself, will prevent mortality from otherwise lethal doses of iron administered to mice either orally or intraperitoneally. Further reflecting this enhanced therapeutic efficacy, the high molecular weight iron chelators also abrogate iron-mediated hepatotoxicity, suppressing the release of alanine aminotransferase. We conclude that high molecular weight derivatives of deferoxamine hold promise for the effective therapy of acute iron intoxication and may also be useful in other clinical circumstances in which control of free, reactive iron is therapeutically desirable.

Antimalarial properties of orally active iron chelators.

The appearance of widespread multiple drug resistance in human malaria has intensified the search for new antimalarial compounds. Metal chelators, especially those with high affinity for iron, represent one presently unexploited class of antimalarials. Unfortunately the use of previously identified chelators as antimalarials has been precluded by their toxicity and, in the case of desferrioxamine, the necessity for parenteral administration. The investigators now report that a new class of orally active iron chelators, namely the derivatives of alpha-ketohydroxypyridines (KHPs), are potent antimalarials against cultured Plasmodium falciparum. The KHPs evidently exert this effect by sequestering iron because a preformed chelator:iron complex has no antimalarial action. The pool(s) of iron being sequestered by the chelators have not been identified but may not include serum transferrin. Preincubation of human serum with KHPs followed by removal of the drug results in the removal of greater than 97% of total serum iron. Nonetheless, this serum effectively supports the growth of P falciparum cultures. Therefore the KHPs may exert antimalarial effect through chelation of erythrocytic rather than serum iron pool(s). The investigators conclude that these powerful, orally active iron chelators may form the basis of a new class of antimalarial drugs.

Natural variation in passive sodium permeability in human erythrocytes.

The rate of influx of 22Na+ into human erythrocytes (RBC) varies greatly depending upon the donor. A high rate of influx may be related to a congenital predisposition to essential hypertension. In Northern Europeans, we find a threefold difference in the rate of 22Na+ influx between those with the least (LP) and most highly permeable (HP) RBC (from less than 0.15 to greater than 0.60 mmol Na+/liter RBC/hr). In order to further define determinants of these apparently hereditary differences in passive membrane Na+ transport, we identified two groups of normal laboratory and hospital personnel differing markedly (greater than twofold) in RBC 22Na+ influx rate. We find that the loop diuretics furosemide and bumetanide decrease by about 50% the influx of 22Na+ into HP RBC, but have a lesser influence on LP RBC. Impermeant polyanions such as citrate and pyrophosphate also specifically diminish 22Na+ influx into HP, but not LP, RBC. Therefore, the exaggerated 22Na+ influx into HP RBC probably occurs through a discrete pathway (perhaps via "Na/K/Cl cotransport"), which appears to be almost absent in LP RBC. The differences between HP and LP RBC most likely do not involve polymorphisms of RBC anion transport per se. The rate of RBC anion (35SO4(2-)) transport is the same in HP and LP RBC and is equally inhibited by furosemide and (to a lesser extent) bumetanide. Furthermore, the potent inhibitor of RBC anion transport, DIDS (diisothiocyanostilbene disulfonate) does not affect RBC Na+ permeability in either group. Nonetheless, the preferential reduction of Na+ permeation of HP RBC by loop diuretics may be of help in experimentally distinguishing HP from LP phenotypes. This information may be crucial in unraveling the structural basis of intrinsic differences in cell membrane Na+ permeability and their possible relationship to essential hypertension.

Hemoglobin potentiates central nervous system damage.

Iron and iron compounds--including mammalian hemoglobins--catalyze hydroxyl radical production and lipid peroxidation. To determine whether hemoglobin-mediated lipid peroxidation might be important in hemorrhagic injuries to the central nervous system (CNS), we studied the effects of purified hemoglobin on CNS homogenates and injected hemoglobin into the spinal cords of anesthetized cats. Hemoglobin markedly inhibits Na/K ATPase activity in CNS homogenates and spinal cords of living cats. Hemoglobin also catalyzes substantial peroxidation of CNS lipids. Importantly, the potent iron chelator, desferrioxamine, blocks these adverse effects of hemoglobin, both in vitro and in vivo. Because desferrioxamine is not known to interact with heme iron, these results indicate that free iron, derived from hemoglobin, is the proximate toxic species. Overall, our data suggest that hemoglobin, released from red cells after trauma, can promote tissue injury through iron-dependent mechanisms. Suppression of this damage by desferrioxamine suggests a rational therapeutic approach to management of trauma-induced CNS injury.

Erythrocyte catalase. A somatic oxidant defense?

Mammalian erythrocytes have large amounts of catalase, an enzyme which catabolizes hydrogen peroxide (H2O2). Because catalase has a low affinity for H2O2, others have suggested that glutathione peroxidase clears most H2O2 within the erythrocyte and that catalase is of little import. We hypothesized that erythrocyte catalase might function to protect heterologous somatic cells against challenge by high levels of exogenous H2O2 (e.g., in areas of inflammation). We find that, whereas nucleated cells (L1210 murine leukemia) are readily killed by an enzymatically generated flux of superoxide (and, therefore, H2O2), the addition of human and murine erythrocytes blocks lethal damage to the target cells. Inhibition of erythrocyte superoxide dismutase, depletion of glutathione, and lysis of the erythrocytes do not diminish this protection. However, inhibition of erythrocyte catalase abrogates the protective effect and the addition of purified catalase (but not superoxide dismutase) restores it. Furthermore, erythrocytes derived from congenitally hypocatalasemic mice (in which other antioxidant systems are intact) do not protect L1210 cells. Our results raise the possibility that the erythrocyte may serve as protection against by-products of its own cargo, oxygen.

Hypohaptoglobinemia associated with familial epilepsy.

In select kindreds afflicted with familial idiopathic epilepsy, most individuals suffering seizures also have low levels of the plasma hemoglobin-binding protein, haptoglobin. This hypohaptoglobinemia may be causally associated with a tendency to develop epilepsy. Our experimental results indicate that artificially-induced hypohaptoglobinemia in mice causes retarded clearance of free hemoglobin from the central nervous system, and that such free hemoglobin may engender the peroxidation of brain lipids. We hypothesize that hypohaptoglobinemia, either inherited, or acquired via traumatic processes, may prevent efficient clearance of interstitial hemoglobin from the central nervous system, thereby predisposing these people to encephalic inflammation and the appearance of seizure disorders.

Hemoglobin. A biologic fenton reagent.

Iron and iron compounds may facilitate hydroxyl-radical generation from activated oxygen species. Earlier work on the generation of this radical has been focused on simple, low-molecular-weight iron compounds. We hypothesized that free hemoglobin, like other iron-rich substances, might also mediate hydroxyl-radical generation. We find: 1) In the presence of a superoxide anion-generating system (hypoxanthine and xanthine oxidase), hemoglobin promotes hydroxyl-radical formation in a dose-dependent fashion. 2) This generation of hydroxyl radical is greatly decreased by prior oxidation of the hemoglobin, equilibration of hemoglobin with carbon monoxide, or addition of catalase, while added superoxide dismutase has little effect. Therefore, hydroxyl radical probably arises primarily via reaction between the ferrous heme iron and H2O2. 3) In further support of this, hydroxyl radical forms as readily upon the addition of H2O2 to hemoglobin. 4) Hemoglobin also increases hypoxanthine/xanthine oxidase-driven peroxidation of poly-unsaturated fatty acids such as arachidonic acid and human red cell membrane lipids. 5) The addition of sufficient haptoglobin (the plasma hemoglobin-binding protein) suppresses both hemoglobin-driven hydroxyl radical and malondialdehyde generation. Thus, free hemoglobin may be biologically hazardous, in part because it acts as a "Fenton" reagent, having the potential to catalyze hydroxyl-radical generation in areas of inflammation. Haptoglobin, which binds hemoglobin very tightly, blocks this through a presently unknown mechanism. An important physiologic function of haptoglobin may be prevention of such hemoglobin-mediated oxidation.

Comparison of the state of deoxyhemoglobin S molecules in solution and in fibers by hydrogen exchange kinetics.

Hydrogen exchange kinetics of deoxyhemoglobin S gel and deoxyhemoglobin A solution were compared at 4.8 mM tetramer concentration, 25 degrees C, and in sodium phosphate buffer at pH 7.0 with gamma/2 = 0.2 by means of microdialysis using tritium as a trace label. Cyanomethemoglobin A in solution and as crosslinked crystals were compared under the same conditions. The exchange values from 15 to 10(4) min were fitted to a power law, and the distribution function of exchange rates was calculated. There was no significant difference in the distribution for deoxyhemoglobin S gel and deoxyhemoglobin A. Exchange from crosslinked cyanomethemoglobin crystals was less in the early time region than for the solution state, but after 600 min the exchange curves were the same. This resulted in a larger area for the distribution function, although the predominate rates were nearly the same. The effect of polymerization on conformational fluctuations was very small, smaller than the effect of crosslinking hemoglobin crystals.

Study of hydrogen exchange in hemoglobin as a function of fractional ligand saturation.

We have studied the hydrogen-exchange kinetics of hemoglobin A0 as a function of ligand, CO, saturation. In the noncooperative system, azide binding to methemoglobin, the alterations in exchange kinetics are proportional to the average degree of ligation. However, in the case of CO binding to deoxyhemoglobin the changes in hydrogen-exchange pattern run ahead of the degree of ligation. The data can be best fitted assuming that all the liganded species, regardless of the number of ligands, show the same exchange properties. This two-state behavior must be the consequence of the fact that all the conformational changes leading to increased solvent accessibility take place when the first ligand is bound. Studies of the effect of pH changes and carbamoylation on the relationship between ligand binding and hydrogen exchange show that the observed differences of hydrogen exchange between deoxy and the liganded state are linked to the alkaline Bohr effect and to the state of the alpha-N-termini. As a consequence, at pH 9 despite a highly cooperative ligand binding isotherm the differences in hydrogen exchange between the deoxy and fully liganded species have vanished. We have to conclude that the hydrogen exchange is mirroring only the first part of the overall R to T transition. In all the experiments with pH changes and carbamoylation it is the liganded form that shows changes becoming more like the deoxy state. This is not consistent with a model where ligand binding removes a structural restriction in the less accessible deoxy state.