INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IRON OVERLOAD PRODUCES CARDIAC TOXICITY affecting both electrical conduction and muscle contractility (9, 10, 24). Decreased survival in patients with hereditary hemolytic anemias and iron overload secondary to multiple transfusions of red blood cells is due primarily to iron-induced cardiac disease (21, 28). Chelation therapy with deferoxamine can produce a negative iron balance, decrease the incidence and severity of iron-induced cardiotoxicity, and prolong survival. However, to be effective, deferoxamine must be infused subcutaneously for 8-10 h/day (29). The low compliance associated with this onerous administrative schedule is associated with decreased patient survival (13, 21, 28).

Iron overload is confirmed by measuring the iron concentration in biopsied liver as well as by increases in the ratio of serum iron to total iron binding capacity (7, 11). An elevated serum ferritin is generally reflective of increases in tissue iron. However, ferritin concentrations are poorly correlated with liver iron concentrations in sickle cell patients (18). In addition, ferritin may increase nonspecifically with inflammation (3, 11). Noninvasive methods that accurately reflect tissue iron concentrations are needed to help guide therapy. Liver relaxation times (T2) measured by using magnetic resonance (MR) spectroscopy are decreased with increases in iron load (8, 17). However, the relationships between these indicators of increased iron and the onset of iron-induced cardiotoxicity are not well defined, which has hampered the development of specific clinical guidelines for initiating, monitoring, and terminating chelation therapy with deferoxamine.

This report uses the guinea pig model of iron overload to investigate the threshold of iron-induced delay or blockade of myocardial electrical conduction measured at the Purkinje fiber-papillary muscle (PF-PM) junction (1, 30). The decrease in electrical conduction was correlated with tissue iron concentrations and iron-induced decreases in liver MR-T2. The effect of iron on ventricular contractility was evaluated by using more than four times the load of iron that blocked electrical conduction. Unexplained sudden death occurred in 27% of the animals receiving these large iron loads. However, with the large doses of iron, no change in ventricular contractility was observed, suggesting that iron-induced changes in electrical conduction are the earliest detectable manifestations of iron cardiac toxicity.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Iron loading. Female Hartley guinea pigs, weighing 200-250 g and obtained from The State of Michigan's Public Health Department (Lansing, MI), were fed a commercial diet with vitamin C (Purina Guinea Pig Chow no. 5206). When group housed, female guinea pigs were much less combative than male guinea pigs. The iron-loaded animals received 0.083 g/kg iron dextran (Fe-D) (Sigma Chemical, St. Louis, MO) by intraperitoneal injection three times per week. Control animals were treated with a comparable volume of intraperitoneal dextran (Sigma Chemical). Evaluation of cardiac electrical conduction and ventricular contractility was performed 2 wk after the guinea pigs received their last injection (1, 30). Experimental procedures and animal housing conditions were approved by the Michigan State University committee for animal experimentation.

Experimental design. This work was performed by using two different series of experiments. Initially, we wished to determine whether iron overload in the guinea pig could block or delay conduction across the PF-PM junction. We wanted to know the lowest iron dose that could affect conduction. Hence, these animals received relatively low loads of iron. Serial liver MR-T2 measurements were performed on these animals and account for the larger number of MR-T2 determinations performed on guinea pigs that had received relatively lower amounts of iron (Table 1).

Evaluation of the effect of iron on myocardial electrical conduction. Because of the known effects of anesthetics on myocardial electrical conduction, guinea pigs were killed by using blunt trauma to the head. The hearts were rapidly removed and placed in cool Tyrode's solution with the following composition (in mM): 137 NaCl, 5.4 KCl, 0.5 MgCl₂, 12 NaHCO₃, 0.42 NaH₂PO₄, 1.8 CaCl₂, and 10 glucose. Papillary muscles (0.5-1.5 mm in diameter, 2-3 mm long) along with their attached Purkinje fibers (~0.1-0.2 mm in diameter and 0.5-1.5 mm in length) were carefully dissected from both the right and left ventricles and mounted in a tissue bath superfused with 95% O₂-5% CO₂ and maintained at 36 ± 0.5°C.

MR spectroscopic measurement of T2. To perform serial liver MR-T2 measurements with increasing iron loads on the same animal, the guinea pigs were anesthetized with 2.5 mg/kg im of acepromazine and 90 mg/kg ip of ketamine. Animals were imaged 48-72 h after dosing with Fe-D or dextran. After reaching the planned total load of iron or dextran control, these animals were used for either investigation of myocardial conduction or contractility.

(TE/T2)

Analysis of tissue iron concentrations. To determine tissue concentrations of iron, iron was measured in heart and liver by using inductively coupled plasma-atomic emission spectroscopy (ICP-AES) (32). Tissue samples (1 g) were obtained from animals used for electrical conduction studies and were digested overnight at 95°C with 2 ml of concentrated HNO₃ (instra-analyzed, JT Baker, Phillipsburg, NJ) in 15-ml Teflon screw-capped vials (Tuf-tainers, Savillex, Minnetonka, MN). Samples were quantitatively transferred to 10-ml volumetric flasks, 100 µg of yttrium internal standard (in 1 ml of 2% HNO₃) was added, and the sample was diluted to volume with H₂O. Accuracy was assured by concurrent digestion and analysis of National Institute of Standards and Technology Standard Reference Material 1577a.

Left ventricular pressure measurements. To evaluate left ventricular contractility, a separate set of experiments was performed by using a large, 1.5 g/kg, iron load. A solution of 50 mg/kg ketamine (Fort Dodge Labs, Fort Dodge, IA) and 5 mg/kg xylazine (Mobay, Shawnee, KS) was administered intramuscularly for anesthesia. This regimen does not inhibit respiration in the guinea pig. Less than 0.5 ml of 1% lidocaine was administered to the incision site for topical anesthesia and to the carotid artery-vagus nerve track to avoid hypoventilation from nerve stimulation. A PE50 catheter was inserted into the carotid artery and threaded through to the left ventricle. The catheter was connected to a Grass 7B polygraph via a micropressure transducer. Before volume challenges, resting left ventricular pressure and change in pressure over time (dP/dt) tracings were recorded three times during 0.5 h and averaged.

Statistical analysis. Comparisons among iron-loaded groups and controls were statistically analyzed with analysis of variance. Fisher's exact test was used to compare death rates between iron-loaded and control guinea pigs. Statistical analysis and correlations were performed by using the Number Cruncher Statistical System (Kaysville, UT) and Graphpad Prism (San Diego, CA). A P value of <0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Evaluation of myocardial electrical conduction. Increased iron load produced progressive delay and blockade of electrical conduction at the PF-PM junction (Table 2). As the iron load increased, the delay of electrical conduction was demonstrated with relatively lower concentrations of extracellular potassium. Prolongation of conductivity across the PF-PM junction was first observed in animals loaded with 0.33 g/kg of iron when the concentration of potassium in the external fluid was increased to 13.5 mM. Increasing potassium in the external fluid to 14.85 mM in animals loaded with 0.33 g/kg Fe-D completely blocked electrical conduction across the PF-PM junction.

Decreased liver MR-T2 with increased iron loads. The sensitivity of MR spectroscopy to iron was investigated by using serial T2 performed on anesthetized animals. The threshold of loaded iron that first produced a decrease in liver MR-T2 was 0.25 g/kg (Fig. 1). Mean MR-T2 was significantly decreased compared with either the preinjection (0 g/kg iron) measurement or with the concurrent dextran control group. Increasing iron load to a maximum of 0.42 g/kg further decreased the MR-T2.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Liver magnetic resonance (MR)-relaxation time (T2) from guinea pigs serially imaged after increasing doses of iron (Fe)-dextran or dextran. Values are means ± SE. * MR-T2 in animals loaded with 0.25 g/kg or more of Fe-dextran are significantly shortened compared with 0 g/kg and matched dextran controls (P < 0.05).

Iron concentrations in livers and hearts. Total tissue iron concentrations were performed by using ICP-AES (Table 3). Increasing the total iron load produced a progressive increase in tissue iron concentrations in both liver and heart. Iron-loaded animals receiving 0.42 g/kg had a ×6.4 increase in the concentration of iron in their livers and nearly a ×2 increase in heart iron concentration.

Relationships among electrical conduction at PF-PM junction, tissue iron concentrations, and liver MR-T2. The relationship between liver MR-T2 to both heart and liver iron concentrations is shown in (Fig. 2). Increasing amounts of iron produced a prolongation and blockade of electrical conduction at the PF-PM junction that was first observed with a total iron load of 0.33 g/kg. This dose of iron almost doubled the cardiac tissue iron concentration to 395 ± 65 µg/g dry wt compared with control of 213 ± 4 µg/g dry wt. At this same iron dose (0.33 g/kg), the increased liver iron concentration of 4,183 ± 13 µg/g dry wt resulted in a markedly shortened MR-T2 of 8.0 ms (control = 14.4 ms). Hence, the threshold of iron load that produced a delay or blockade of electrical conduction at the PF-PM junction was heralded by a decrease in the liver MR-T2.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Relationships among liver MR-T2, iron loads, and tissue iron concentrations. Increases in iron load result in a decrease in liver MR-T2 that is plotted with the tissue iron concentrations in the heart (A) and in the liver (B). Amount of iron loaded is indicated next to the data point.

Left ventricular pressure measurements. Guinea pigs loaded with maximal amounts of iron (1.5 g/kg) had left ventricular pressures measured in isovolemic, hypovolemic, and hypervolemic states (Fig. 3). No significant differences were noted between iron-loaded and control animals. No change in myocardial contractility was observed with iron loads that were more than four times greater than the dose of iron that produced blockade of electrical conduction at the PF-PM junction. However, arrhythmias with premature ventricular contractions were frequently observed in the iron-loaded guinea pigs and rarely in the dextran controls. In addition, compared with animals receiving lesser loads of iron, tissue iron concentration measurements in 21 animals loaded with 1.5 g/kg showed increased iron in liver 10,700 ± 1,180 µg iron/g dry wt and in heart 800 ± 88 µg iron/g dry wt.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Left ventricular pressures in 1.5 g/kg Fe-dextran and dextran guinea pigs in isovolemic, hypovolemic, and hypervolemic states: A: end-systolic pressures. B: end-diastolic pressures. C: positive and negative changes in pressure over time.  and filled bars, Fe-dextran;  and open bars, dextran.

Sudden death. Unexplained, sudden death was observed in 9 of 33 (27%) guinea pigs loaded with 1.5 g/kg of iron; none of the 19 controls died (P < 0.05). Because death in the iron-loaded animal occurred without weight loss, decrease in activity, change in fur texture, other signs of illness, or necropsy evidence of disease or hemorrhage, we suspect that the animals died of a cardiac arrhythmia.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The initial effects of iron-induced cardiac toxicity in this animal model preferentially inhibited electrical conductance without producing a demonstrable decrease in myocardial contractility. Iron blockade of myocardial conduction at the PF-PM junction was observed with a relatively small iron load (0.33 g/kg). The larger iron load (1.5 g/kg) resulted in 27% of the animals dying from an unexplained sudden death, presumably from iron-related cardiac arrhythmias, but did not affect left ventricular contractility.

The dose of iron that resulted in an increase in liver iron concentration and decreased the liver MR-T2 was also associated with prolongation and blockade of myocardial electrical conduction at the PF-PM junction. In this animal model, the decrease in the liver MR-T2 could be used as a noninvasive marker for the earliest demonstrable defect in myocardial electrical conduction.

Translation of these observations to iron overload patients depends on the similarity of the guinea pig model of iron overload to human disease. Hepatic iron concentrations measured in iron-loaded guinea pigs were in the same range as those determined in humans with iron overload (6, 8). Prior reports validating the guinea pig as a model of iron overload show that, like humans with iron-induced organ toxicity, increased iron in liver, heart, and bone marrow was observed both histologically and with analytic measurements (1). The ratio of serum iron to total iron binding capacity was also increased in the iron-loaded guinea pigs (1). Furthermore, studies with the guinea pig model confirm iron-induced lysosomal membrane fragility and increased peroxidation of membranes (30). In humans, iron-induced cardiac disease develops after years of a gradual increase in tissue iron concentration. The guinea pig model of iron overload demonstrated changes in electrical conduction, electrocardiogram-documented arrhythmias, and a high incidence of sudden death after a few weeks of iron load. In addition, guinea pigs take 8 wk to complete the iron loading process. The guinea pig lifespan is ~3 yr or 156 wk. Hence, the ratio of iron loading to the total life span of the guinea pig computes to ~0.05. Similar computations in humans by using a transfusion iron overload model show that an adult not producing red cells is transfused with ~2 U of red blood cells every 2 wk. After ~2 yr, patients have received ~100 U of red blood cells, which computes to an iron load of ~25 g. It is at this time period that these patients are felt to be at risk from iron overload toxicity secondary to the red cell transfusions. Hence, in humans, iron overload occurs via transfusion in ~2 yr. If the human lifespan is judged to be ~70 yr, the ratio of iron loading to total lifespan is approximately 0.02, which is comparable to the ratio of time of iron loading to life span computed for the guinea pig. One of the virtues of the guinea pig model is its similarities to the human disease of transfusion iron overload and ease of obtaining an animal model in a relatively short period of time.

Both the guinea pig and the gerbil models of iron overload use Fe-D as an iron loading reagent (1, 5). Iron loading is via injection into the peritoneum in the guinea pig and under the skin in the gerbil. Validation of the gerbil model relied on demonstration of increased iron in hearts and livers, myocardial necrosis, fibrosis, and hepatic fibrosis (5). Similarly, the guinea pig model also demonstrated increased hepatic and cardiac iron after iron loading (1). In addition, the guinea pig model confirmed that iron caused cellular toxicity with increased fragility of the lysosomal membrane as well as peroxidation of membranes (30). Hence, the guinea pig model of iron overload is similar to the human condition with respect to morphological and membrane pathological changes.

Conduction defects at the PF-PM junction can lead to reentrancy and circus movement (14-16). Like the guinea pig model of iron overload, cardiac arrhythmias were observed in gerbils after iron loading; three of four gerbils tested had cardiac arrhythmias (5). Furthermore, gerbil cardiomyocytes isolated after 4-12 wk of iron loading had an increase in iron content, a decrease in overshoot and duration of cardiac action potential, and changes in sodium and potassium currents (22). In iron-loaded guinea pigs, the observed delay and blockade of electrical conduction at the PF-PM junction as well as the observed cardiac arrhythmias suggest that the high rate of unexplained sudden death recorded for the guinea pigs receiving higher iron loads may be secondary to cardiac arrhythmias.

Abnormalities of myocardial electroconductivity and myocardial function in humans were correlated morphologically with increased deposition of iron in both myocardium and conduction tissues (4, 9, 31). Long-term studies in thalassemic patients related marked cardiac hypertrophy and dilation with increased iron deposition. Tissue iron concentrations evaluated by using endocardial biopsy may be misleading because the distribution of iron in the heart was not homogeneous; endocardial tissue contained half of the iron concentration that was present in the epicardial layer (12). Bundle conduction studies demonstrated decreased electrical transmission in iron-loaded patients (10, 27). This report confirms that iron-induced defects in myocardial conduction occur in guinea pigs and shows that the earliest detectable conduction abnormalities at the PF-PM junction were observed with a load of iron that was signaled by a decrease in liver MR-T2.

Iron cardiac toxicity was related to oxidation of biological membranes (23). Peroxidation of membrane lipids in isolated guinea pig hearts with tert-butyl hydroperoxide caused conduction disturbances and arrhythmias with an associated increase in myocardial malondialdehyde, confirming increased tissue peroxidation (25). Pretreatment with an antioxidant inhibited these effects. Increased concentrations of malondialdehyde were detected in the guinea pig model of iron overload that was used in these investigations (1). This suggests the hypothesis that conduction abnormalities recorded at low levels of loaded iron may be related to the sensitivity of the PF-PM junction to iron-facilitated peroxidation of membrane lipids.

Abnormal T2-weighted images and shortened T2 were reported for MR hepatic scans performed on iron-overloaded patients (20). Decreased T2 observed in iron-overloaded patients correlated with increases in serum ferritin and iron concentration in the liver (20). Another MR measurement used to gauge the amount of iron overload was the MR signal intensity ratios between liver and subcutaneous fat. This ratio correlated with serum ferritins and changed toward normal as the iron burden was reduced with phlebotomy (2, 19). Decreased hepatic iron concentrations after chelation therapy were monitored noninvasively by using superconducting quantum-interference device biomagnetometry (26). However, before this report, the relationship of changes in hepatic iron concentration to the onset of iron-induced defects in myocardial conduction had not been investigated.

Investigations in the past relied on long-term clinical outcomes to evaluate the efficacy of chelation therapy in iron-overloaded patients. The results of this study, which used the guinea pig model of iron overload, demonstrate that the onset of prolongation and blockade of cardiac electrical conduction across the PF-PM junction occurs with a modest increase in myocardial iron concentration. The inhibition of electrical conductance observed at lower iron loads may be related to the high incidence of unexplained sudden death, presumably from arrhythmias that were observed with the larger load of iron. The threshold of iron load that initially delayed or blocked cardiac conduction produced sufficient increases in liver iron to signal a decrease in the liver MR-T2. Larger iron loads failed to decrease left ventricular contractility, demonstrating that the earliest detectable iron-induced cardiac toxicity preferentially inhibits electrical conduction. These data suggest that this animal model will be useful in evaluating new chelating agents for their ability to decrease tissue iron concentrations and to reverse iron-induced electrical conduction defects. Because of the similarities between the guinea pig model of iron overload and human iron overload states, the utility of serial liver MR-T2 to monitor iron-induced cardiac conduction defects in patients needs to be investigated.