The Use of Zn-Desferrioxamine for Radioprotection in Mice, Tissue Culture, and Isolated DNA¹

Redox-active metals, such as copper and iron, potentiate the production of reactive oxygen-derived species after various kinds of oxidative stress (1, 2). The effect of iron and copper, either chelatable or structured in metalloproteins, on radiation damage has also been widely studied. Iron, even in ferritin, can act as a radiosensitizer and increase the radiation-induced killing of Chinese hamster ovary cells (3). The role played by copper in radiation-induced damage is more complex and was found to involve protective, sensitizing (4, 5, 6, 7, 8), and recovery effects (9, 10, 11, 12, 13, 14). In addition, oxidative stress might also mobilize metal ions from their cellular stores to the pool of catalytically active metal. Such a mobilization serves as an index of the extent of injury and might further increase the damage (15, 16).

To minimize the metal effect, chelating agents are used widely. DFO,³ which avidly binds Fe(III), is used clinically for prevention and treatment of iron overload in subjects who have ingested toxic oral doses of iron salts or in thalassemic patients (17, 18, 19). DFO inhibits generation of hydroxyl radicals and iron-dependent peroxidation of lipids and can serve at mm concentrations as a radical scavenger and a reducing agent (20, 21, 22). In addition to binding metals, a key property of iron chelators is their ability to cross the biological membrane, enabling them to be absorbed from the intestinal tract and enter cells of a range of tissues including liver and heart. Most drugs enter cells by simple diffusion through the hydrophobic region of the cell membrane. Consequently, uncharged drugs permeate more rapidly than charged molecules (23). It is possible to assess the ease with which molecules penetrate membranes by measuring their partition coefficients between the aqueous and organic phases (usually 2-octanol). A partition coefficient in the range of 0.2–1, for a chelator, was found optimal for oral activity, ability to penetrate hepatocytes, and minimal acute toxicity. DFO, which has low lipid solubility and a partition coefficient of 0.01 (24), diffuses slowly (20) and predominantly (25) enters the cell via nonspecific fluid phase pinocytosis (26), a mechanism that underlies the transfer of DFO into lysosomes.

Unlike its protective effect against nonradiative injury, the radioprotective efficacy of DFO is limited and controversial. DFO reportedly attenuates radiation-induced lipid peroxidation of platelets and erythrocytes and was found by several authors (27, 28) to protect mice subjected to whole-body irradiation (29). On the other hand, it was also reported to radiosensitize cells in culture (30). Recently, a Zn complex of DFO has been suggested as a protective agent against postischemic reperfusion injury in the cat retina (31), against corneal alkali injury in rabbits (32), and against X-ray-induced damage to salivary glands (33). Zn-DFO has a globular form that might enable it to enter cells faster. It could prolong the persistence of the drug in the circulation, and furthermore, zinc can displace and substitute catalytically active iron (34). The present study tests the potential effect of Zn-DFO on radiation-induced injury in isolated DNA, cultured mammalian cells, and mice.

DFO was purchased from Sigma Chemical Co. Zn-DFO was prepared as described previously (35).

C3H female mice were supplied through the Frederick Cancer Research and Development Center Animal Production Area (Frederick, MD). The animals were housed five per cage in climate-controlled, circadian rhythm-adjusted rooms and allowed food and water ad libitum. The animals were 12 weeks of age and weighed 22–27 g at the time of irradiation. Experiments were conducted according to the principles outlined in the Guides for the Care and Use of Laboratory Animals prepared by the Institute of Laboratory Animal Resources, National Research Council. The animal protocol was approved by a National Cancer Institute Animal Care and Use Subcommittee.

To determine its maximal tolerated dose, Zn-DFO dissolved in saline [e.g., 40 mg in 10 ml for a 5 mm stock solution, 5 mmol/kg b.w. (pH 7.4)], was administered i.p. to C3H mice in a volume of 1% of their weight. Four groups of five mice each received Zn-DFO doses of 50, 100, 150, and 200 μmol/kg b.w., respectively, and survival was assessed up to 14 days after injection.

Female C3H mice 5 months of age were divided into three groups numbering at least four animals and anesthetized with an i.m. injection of ketamine (90 mg/kg) and xylazine (9 mg/kg) and placed on a heating pad to keep their body temperature at ∼37.5°C. The right carotid artery was exposed, occluded with a small clamp, and cannulated with Abbocath T 24G × 3/4-inch radio-opaque FEP Teflon i.v catheter (Abbott Hospitals, Inc., North Chicago, IL). The catheter was attached to a 20-inch extension set (Abbott Laboratories) that was connected to a P23XL pressure transducer (Viggo-Spectramed). Sterile saline containing 2 unit heparin/ml was dripped through the transducer, tubing, and catheter at a constant flow rate of 3 ml/h, controlled by a Sorenson Intraflo II (Abbott Critical Care Systems). Either 200 μmol DFO/kg b.w, 200 μmol Zn-DFO/kg b.w., or saline alone was administered i.p. to the mouse. Hemodynamic parameters, including blood pressure and heart rate, were recorded simultaneously throughout the experiment and saved for later analysis and evaluation.

Solutions of 5–20 mm Zn-DFO and DFO were freshly prepared in 0.9% NaCl. Control mice were given 0.9% sodium chloride. All injections were administered i.p. The drugs were injected in volumes equivalent to 1% of each animal’s weight (0.22–0.27 ml), 30 min before irradiation. Groups of five mice were placed in round (30.5-cm diameter by 10.5-cm height) Plexiglass containers with holes for ventilation. Two separate containers were placed in the sample tray of the irradiator. A ¹³⁷Cs Gamma Cell 40 (Norion International, Inc., Kanata, Ontario, Canada) was used as the ionizing radiation source. The irradiator was calibrated with thermoluminescent dosimetry chips planted in phantom mice, and the radiation dose was determined according to methodology described previously (36). The dose rate was 0.98 Gy/min. Immediately after irradiation, the mice were returned to their cages and were assessed daily for survival.

CCC plasmid DNA (pUC19 2686 bp) was prepared and isolated as described previously (37). To induce strand breaks, 0.2 μg of DNA was exposed to various doses of γ-radiation using a ¹³⁷Cs Gamma Cell (Radiation Machinery Corporation, Parisppany, New Jersey) source having a dose rate of 10 Gy/min. The samples were applied to horizontal, 0.9% agarose containing 0.3 μg/ml ethidium bromide, slab gels in pH 8 Tris-borate EDTA buffer or Tris-acetate EDTA, and electrophoresed at a potential gradient of 5.7 V/cm. After electrophoresis, the gels were illuminated with UV light and photographed using a Bio Imaging System 202D and a Fujifilm, Thermal Imaging System FTI-500. The extent of DNA breaks was assessed by comparing the relative intensities of the migrating bands of supercoiled and relaxed CCC DNA. Under these experimental conditions, the linear and circular relaxed double-strand plasmid DNA of pUC19 hardly differed by their electrophoretic mobilities. Analysis and quantitation of the relative intensities of the migrating bands was performed using a Fluor-S-MultiImager (Bio-Rad).

Chinese hamster V79 cells were grown in F12 medium supplemented with 10% FCS, penicillin, and streptomycin. Survival was determined by the clonogenic assay (38). The control plating efficiency ranged between 62 and 72%. Stock cultures of exponentially growing cells were trypsinized, rinsed, and plated (5 × 10⁵ cells/dish) in Petri dishes (60 or 100 mm) and incubated overnight at 37°C before experimental protocols. Zn-DFO or DFO (100 μm) was added to exponentially growing cells in complete F12 medium at room temperature 60 min before irradiation.

Cells were irradiated at room temperature with a ⁶⁰Co source at a dose rate of 2 Gy/min. Immediately after irradiation, the cells were trypsinized, counted, plated in triplicates, and incubated for 7 days for macroscopic determination of colony formation. Colonies were fixed with methanol:acetic acid (3:1) and stained with crystal violet. Colonies containing >50 cells were scored.

To assess survival of mice after irradiation, logistic regression analysis was used to fit smooth curves to the survival proportions as a function of radiation dose for each treatment. The doses were logarithmically transformed for analysis to obtain an improved overall fit. The transformation had little effect upon the significance of the results. Differences between pairs of curves were assessed by using the likelihood ratio test of the logistic model, which is a two-tailed test. The maximum likelihood estimates of LD_50/30, radiation dose which caused 50% lethality at 30 days, were derived from the parameters of the logistic fit, and confidence intervals based on the profile likelihoods were calculated.

Breaks in pUC19 plasmid DNA have been induced radiolytically. The extent of strand breaks was assessed electrophoretically (see “Materials and Methods”). Typical electrophoretograms are presented in Fig. 1,A, wherein the migration profile of the plasmid DNA reflects the degree of breakage. The DNA migration profile seen represents the conversion of supercoiled into linear DNA after irradiation, in the presence of various doses of ZnCl₂, DFO, or Zn-DFO. The plasmid preparation contained predominantly CCC DNA, accompanied by a minor band of linear DNA. For treated samples, the intensity of the trailing band increased at the expense of that of the leading band, reflecting an extensive conversion of supercoiled into linear DNA caused by irradiation. ZnCl₂ provided only minimal protection. DFO at the respective concentrations afforded greater protection than ZnCl₂. The drug Zn-DFO had a greater protective effect than that of each of its components alone. Fig. 1 B demonstrates the quantitation of the relative intensities of the migrating bands in terms of percentage of residual supercoiled DNA (CCC DNA). γ-Irradiation-induced damage was inhibited to various extents by Zn < DFO < Zn-DFO. With no radiation and under similar conditions, ZnCl₂, DFO, and Zn-DFO had no effect on DNA breakage.

To compare the radiation response of cultured Chinese hamster V79 cells with DFO or Zn-DFO, cells were incubated for 1 h with the drugs before γ-irradiation, and the results are shown in Fig. 2. Dose-modifying factors were calculated by determining the ratio of doses at 10% survival level of treated cells and control cells. Although previous studies have shown a sensitizing factor of 1.17 for DFO (500 μm) in irradiated V79 cells, the present results indicate that 100 μm DFO and Zn-DFO had practically no effect on the cells. Zn-DFO (500 μm) demonstrated a sensitizing factor of 1.41, and under similar experimental conditions, Zn-DFO exerted no cytotoxicity to nonirradiated cells (data not shown).

Four groups of five mice each were injected i.p. with 0.2 ml of saline containing Zn-DFO in final concentrations of 50, 100, 150, and 200 μmol/kg b.w. Group 1, which received 50 μmol of Zn-DFO/kg b.w., demonstrated no external sign of convulsion, depression, or excitation. However, the mice in the groups that received higher doses demonstrated transient dose-dependent symptoms of depression of locomotor activity, which were the most prominent for group 4, having mice treated with 200 μmol of Zn-DFO/kg b.w. Those symptoms, however, were not detectable after 24 h. None of the mice in any of the groups died because of the administration of Zn-DFO.

To determine the appropriate timing for administering Zn-DFO, 48 C3H female mice at 2 months of age were subjected to 9 Gy of total body irradiation. The mice were injected i.p. with Zn-DFO 10 or 30 min before irradiation and were divided into five experimental groups differing in Zn-DFO concentrations. Survival was monitored for 30 days after irradiation. In mice exposed to 5–10 Gy whole-body γ-radiation, death was usually due to bone marrow failure and occurred predominantly between 8 and 20 days. Zn-DFO at 100 μmol/kg b.w. protected the animals. Enhanced survival was observed for mice treated with Zn-DFO 30 min before irradiation, whereas shorter pretreatment conferred lower protection. On the other hand, Zn-DFO at 200 μmol/kg b.w. had a radiosensitizing effect, which increased with the time elapsed between injection and irradiation.

C3H female mice 2–3 months of age were injected i.p. with saline, DFO, or Zn-DFO at various concentrations; 30 min later, they were subjected to various doses of total body irradiation. Survival was monitored for 30 days, and the results are displayed in Fig. 3. Although 100 μmol of Zn-DFO/kg b.w. provided radioprotection, 5 and 30 μmol had no effect, and 200 μmol of Zn-DFO/kg b.w. radiosensitized the mice to total body irradiation. Thus, a therapeutic window in terms of doses was observed. Control experiments showed that DFO at 5, 30, 100, and 200 μmol/kg b.w. did not modify mice survival (data not shown). The effects of DFO and Zn-DFO on radiation-induced lethality in mice is shown in Fig. 4. Zn-DFO administered 30 min before radiation provided significant protection compared with control animals receiving DFO or vehicle alone. The LD_50/30s were determined by using doses ranging from 7 to 11 Gy. Data were pooled from several experiments, with each data point repeated at least twice and representing 15–40 mice. The LD_50/30s were as follows: 10.3 Gy for Zn-DFO (95% confidence intervals, 9.83–10.99), 8.03 for DFO (95% confidence intervals, 7.84–8.26), and 7.91 for saline (95% confidence intervals, 7.79–8.04). This represents a dose modification factor (radiation dose that caused 50% lethality at 30 days in the Zn-DFO-treated group divided by the radiation dose that caused 50% lethality at 30 days in the control group) of 1.3 (P < 0.01). The differences between dose-modifying factors and their null values of 1 were tested using the likelihood ratio test of a logistic regression model reparameterized for that null hypothesis.

Hemodynamic parameters, such as blood pressure and heart rate, were recorded throughout the experiment and analyzed as follows. Blood pressure and heart rate were measured each minute over a period of 20–30 min. Mean heart rate was calculated by adding heart rate values across the same time points. Blood pressure was expressed as mean arterial blood pressure, which was calculated using the values of diastolic pressure and systolic pressure according to the following equation: mean arterial blood pressure = diastolic pressure + 2/3 (systolic pressure-diastolic pressure). Mean arterial blood pressure was calculated by summing the arterial blood pressure values across the same time points. Fig. 5 presents the effects of Zn-DFO, DFO, and saline on heart rate and mean arterial blood pressure. Each group numbered at least four animals. Administration of saline alone, 200 μmol of DFO, or Zn-DFO/kg b.w., resulted in a drop of 25–30% in mean heart rate. Administration of 200 μmol of DFO or Zn-DFO/kg b.w. or saline alone resulted in changes in mean arterial blood pressure, which did not differ from each other.

The radioprotective activity of Zn-DFO was tested at three levels: molecular, cellular, and whole animal. The results from the present study suggest that Zn-DFO provides radioprotection to isolated DNA and to whole animals (Figs. 1 and 4). However, Zn-DFO has no effect on the radiation-induced cytotoxicity in V79 cells (Fig. 2). At the molecular level, zinc, DFO, and Zn-DFO, in an increasing order of efficiency, inhibited DNA breakage induced by radiation (Fig. 1).

Redox-active iron (or copper) ions may affect the radiation damage by: (a) catalyzing the generation of deleterious OH radicals through the Haber-Weiss reaction; (b) coordinating with the essential macromolecules that might actually determine the exact location of the damage in a site-specific manner (2, 39); (c) mobilizing metals from cellular stores because of stress; and (d) increasing the pool of catalytic iron (or copper) that might further amplify the injury. Because the coordination chemistry of the zinc ion resembles that of iron, it can displace and substitute redox-active iron from its binding sites (40).

Indeed, both organic and inorganic complexes of zinc have been reported to confer radioprotection by preventing destruction of blood cells and enhancing hemopoietic recovery (41). For instance, zinc aspartate protected against whole-body irradiation, inhibited lethality and hematological damage, and exerted a synergistic radioprotective effect with WR-2721, probably through stabilization of endogenous and exogenous thiols by the formation of zinc thiol complexes (41, 42, 43). In the present study, ZnCl₂ offered limited protection to plasmid DNA γ-irradiation. The radioprotective activity of Zn-DFO might be explained by the “push and pull mechanism” (34).

Another plausible mechanism for the protective activity of Zn-DFO is associated with its charge. Positively charged stable nitroxides have demonstrated greater radioprotection to DNA as compared with neutral nitroxides both in bacterial and mammalian cells (44, 45). Because the net charge of Zn-DFO is more polarized than that of DFO alone, (46), it might better associate with negatively charged biopolymers such as nucleic acids. Hence, it might be more efficient in chelating the iron and substituting it with zinc.

It is possible that Zn-DFO permeates cells more readily than DFO alone, presumably because of its globular structure as opposed to the steric conformation of DFO. Although this explanation is not relevant to the protective effect of Zn-DFO in a noncompartmentalized test system, such as in the case of isolated DNA, it might account for its effect in cultured cells and whole animals. The disagreement between the effects of Zn-DFO in whole animal and tissue culture systems is intriguing. The modes of action considered above do not explain its failure to protect tissue cultured cells from radiation. The difference between the effects of Zn-DFO in animals and in cultured cells might be analogous to the selective radioprotection by zinc to normal hemopoietic cells in vivo but not to tumor cells in vitro (47). The possibility of an effect on the vascular system or the level of tissue oxygenation was not supported by the effect on the hemodynamic parameters. Zn-DFO and saline did not differ in their effect on heart rate or mean arterial blood pressure of the mice (Fig. 5); therefore, it is possible that an additional elaborate systemic effect plays a role in Zn-DFO radioprotective activity. DFO complex with iron (ferrioxamine) has a stability constant of 10³¹ and a prolonged half-life as compared with DFO alone (48). Ferrioxamine is also probably extremely resistant to enzymatic degradation (18). On this basis, it is reasonable to assume that Zn-DFO, with a stability constant of 10¹¹, would be partially resistant to degradation by enzymes, thus extending the half-life of the drug but still maintaining its chelating power.

In conclusion, the present results indicate the potential in using Zn-DFO in modulating radiation injury in experimental animals. It is particularly important in this field where, unlike with other kinds of oxidative injury, chelators of metals have previously shown only limited efficacy. Nonetheless, the specific roles of DFO and Zn should be elucidated, in particular, in the context of hemopoietic protection and recovery.