Introduction

Top Abstract Introduction Experimental Procedures Results and Discussion References

Iron metabolism is characterized by a highly efficient recycling process (Finch et al., 1970; Hallberg, 1981; Finch and Huebers, 1982, 1986) with no specific mechanism for elimination. The introduction of "excess iron" (O'Connell et al., 1985; Thomas et al., 1985; Seligman et al., 1987) into this closed metabolic loop leads to chronic overload and ultimately to peroxidative tissue damage. Iron excretion can be promoted by therapy with ferric ion-specific chelators such as the hydroxamate desferrioxamine B (DFO)¹, which is a bacterial siderophore and the drug of choice for the treatment of transfusional iron overload. Because DFO is poorly absorbed from the gastrointestinal tract and rapidly eliminated from the circulation, prolonged parenteral infusion is needed (Pippard, 1982, 1989; Lee et al., 1993). Patient compliance with such a demanding, expensive, and unpleasant regimen is a problem. Thus, considerable interest has developed in the search for an iron chelator that does not require parenteral administration.

Our laboratory has considerable experience in the study of iron chelators in the iron-loaded Cebus apella primate model. This model is highly predictive for the performance of ligands in patients with transfusional iron overload (Peter et al., 1994). We recently reported an extensive structure-activity relationship for orally active iron che-lators based on the desferrithiocin (DFT) pharmacophore [(S)-4,5-dihydro-2-(3-hydroxy-2-pyridinyl)-4-methyl-4-thiazolecarboxylic acid, 1, Fig. 1] (Bergeron et al., 1999a,b). This and the other studies (Bergeron et al., 1991, 1993, 1994, 1996b) have made it possible to construct DFT analogs that are still orally effective, yet have substantially reduced toxicity compared with the parent ligand. Nonetheless, an even broader therapeutic window would be desirable, and it also may be advantageous to design chelators that selectively target one of the body's two main pools of chelatable iron, the hepatocellular (parenchymal) and reticuloendothelial (RE) stores. Thus, an understanding of the pharmacokinetic behavior of iron chelators may allow the development of improved strategies of chelation therapy, including combination regimens targeting specific physiological compartments.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Structures of DFT (1), DMDFT (2), 4-(S)-hydroxydesazaDMDFT (3), and the oxazoline analog of desazaDMDFT (4).

This paper describes the distinctive pharmacokinetic behavior of three prospective oral iron chelators based on the DFT pharmacophore (Fig. 1): desmethyldesferrithiocin [(S)-4,5-dihydro-2-(2-hydroxyphenyl)-4-thiazolecarboxylic acid, DMDFT, 2]; 4-(S)-hydroxydesazaDMDFT [(S)-4,5-dihydro-2-(2,4-dihydroxyphenyl)-4-thiazolecarboxylic acid, 3]; and the oxazoline analog of desazaDMDFT [(R)2-(2-hydroxyphenyl)-4-oxazolinecarboxylic acid, 4].

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results and Discussion References

Materials. Ligands 2, 3, and 4 were synthesized as described previously (Bergeron et al., 1994, 1999a,b). C. apella monkeys were purchased from World Wide Primates (Miami, FL). All reagents and standard iron solutions were obtained from Aldrich Chemical Co. (Milwaukee, WI). All hematological and serum chemical tests were carried out by Allied Clinical Laboratories (Gainesville, FL). Cremophor RH-40 was obtained from BASF (Parsippany, NJ).

Collection of Pharmacokinetic Samples in Monkeys. After an overnight fast, sedation with either ketamine and scopolamine or Telazol and atropine (3 and 4), and insertion of a lubricated 5 French, 16-inch urethral catheter, the compounds were solubilized in 40% (v/v) Cremophor RH-40/water and administered p.o. by gavage at a dose of 300 µmol/kg via an 8 French feeding tube. Sedation was maintained by either additional i.m. ketamine (2) or isoflurane gas given by intubation (3 and 4).

Plasma and Urine Analytical Methods. Plasma and urine pharmacokinetic samples were prepared for HPLC analysis by mixing with an equal volume of methanol, chilling at 0°C for 30 min, centrifugation at 3000g, and filtration of the methanolic supernatant through a 0.2-µm filter before injection.

Pharmacokinetic Analyses. The model-independent pharmacokinetic parameters, including the area under the time-concentration curve (AUC) from time zero to the time of the last measured plasma concentration (8 or 24 h), the total area under the first-moment curve (AUMC), mean residence time (MRT), and renal clearance (CL_R) were estimated from plasma and urine concentration-time data as reported in Bergeron et al. (1995, 1996a, 1998).

Human Serum Albumin Binding Assay. The binding was carried out by incubation (30 min at 37°C) in 100 mM TRIS-chloride buffer, pH 7.4, such that the final concentrations of human serum albumin (Sigma Chemical Co., St. Louis, MO) and chelator were 1 and 300 µM, respectively. Ultrafiltration was performed with Centricon 3 filters (Amicon, Beverly, MA) to separate bound chelator from free ligand.

	Results and Discussion

Top Abstract Introduction Experimental Procedures Results and Discussion References

Compounds 2 (Bergeron et al., 1993) and 3 (Bergeron et al., 1999a) administered p.o. were at least as active in promoting iron excretion as an equivalent dose of parenteral DFO, whereas the oxazoline analog 4 was unexpectedly ineffective in inducing iron excretion (Table 1). Pharmacokinetic analysis provides some insight into the origin of the differences in iron clearance among the analogs.

Although 2 and 3 were comparably effective in inducing iron excretion, they exhibited markedly different plasma pharmacokinetics (Fig. 2). The C_max and AUC_{0-8 h} of 3 were ~4 and <10%, respectively, of the corresponding values for 2 (Table 1). The amount of 3 eliminated in the urine in 24 h was only 10% of the amount of 2 eliminated in the first 4 h; this mirrored the low plasma AUC for 3. The CL_R of the two compounds was similar, 492 ± 55 versus 312 ± 161 ml/h/kg for 2 and 3, respectively. The low plasma levels observed for 3 are most plausibly the result of first-pass clearance by the liver. This explanation is consistent with the efficacy of 3 in inducing iron excretion and the predominantly fecal mode of that excretion (Table 1). In contrast, compound 2 achieved high plasma levels and was found to be extensively eliminated in the urine (50.7 ± 4.2% in 4 h), comparing favorably with the 58% urinary mode of excretion observed for 2 in iron clearance studies (Table 1). The different modes of excretion are not readily attributable to different distribution patterns of loaded iron because liver parenchymal and hepatocyte compartments appear to be loaded to a similar, massive extent in the primate model (R.J.B., G.M. Brittenham, H. Fujioka, W.R.W., and J.W., unpublished data).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Plasma pharmacokinetics of DMDFT (2, ), 4-(S)-hydroxydesazaDMDFT (3, ), and the oxazoline analog of desazaDMDFT (4, ). Compounds were administered by oral gavage at 300 µmol/kg to male C. apella monkeys. For clarity, data are plotted as means of five (), eight (), or four () experiments.

Given the inability of oxazoline 4 to promote iron excretion (Table 1), the plasma pharmacokinetics of 4 were remarkable (Fig. 2, dashed line). This compound was well absorbed orally with a C_max value of 103 µM and a very prolonged presence in the plasma; a terminal elimination phase was difficult to discern from these 0 to 8 h experiments. Also remarkable was the very poor CL_R of 4 compared with 2 and 3 (Table 1). We suspected that plasma protein binding might be the basis for the prolonged presence of 4 in the plasma. Thus, the ability of human serum albumin to bind the ligands was evaluated (Table 2). Compound 4 was >99.5% bound to the albumin; this was at least a 10-fold greater extent than 2 or 3. Therefore, plasma protein binding may provide an explanation of the prolonged residence of 4, its poor CL_R, and its lack of clinical efficacy, i.e., such binding may simply render the drug unavailable to chelatable iron pools in the body. Alternatively, it may be that 4 is promoting iron excretion, but the iron chelate is eliminated at such a slow rate as to not be detected in our iron balance experiments. Certainly, the ultimate fate of 4 in the body remains to be determined.

Although the three ligands analyzed in these experiments are all DFT analogs, their iron-clearing properties and pharmacokinetic parameters are markedly different. It is clear that the oral bioavailability of a ligand does not guarantee iron-clearing efficacy. Because the pharmacokinetic behavior of the DFT pharmacophore can be manipulated, it follows that the site of iron chelation also can be altered. To find whether a chelator targets either of the two main pools of chelatable iron, the selective radiolabeling of hepatocellular and RE iron stores can be accomplished with a hypertransfused rat model (Zevin et al., 1992). In studies using this model, iron mobilized from hepatocellular pools was excreted directly into the bile. About one-third to one-half the iron derived from RE stores was eliminated in the urine; the remainder of this iron was recycled into the liver and excreted in the bile (Zevin et al., 1992). If this is the case, then our findings suggest that 3 particularly targets hepatocellular pools, whereas 2 promotes excretion of RE iron. Thus, it may be possible to design nontoxic, orally effective iron chelators that select either of the two major pools of chelatable iron.