Abstract
Metabolism and pharmacokinetics of 1-(2′-trimethylacetoxyethyl)-2-ethyl-3-hydroxypyridin-4-one (CP117) were studied in the rat. Urinary recovery studies were conducted in normal (oral and intravenous) and iron-overloaded rats (500 mg Fe/kg body weight; oral only). In normal rats, the majority of the dose recovered in the urine was as the hydrophilic metabolite, CP102, accounting for 69.7 ± 9.4% (oral) and 80.7 ± 7.9% (intravenous) of the administered dose. There was, however, a dramatic decrease in the amount of CP102 recovered (47.7 ± 5.9%) (p ≤ 0.05) in the iron-loaded group of animals. The amount of CP102 glucuronide conjugate recovered in the normal and iron-overloaded rats after oral administration of CP117 did not differ significantly with values of 6.5 ± 2.5% and 7.1 ± 2.5%, respectively. There was, however, a significant increase in CP102 glucuronide conjugate (13.7 ± 3.0%) (p ≤ 0.05) after intravenous administration of CP117. Urinary iron content was determined in the iron-overloaded and normal (oral) animals. Negligible levels of the CP117 iron complex and only 0.6 ± 0.2% was present as the corresponding CP102 complex in the urine of normal animals. Less than 0.1% of the administered dose was recovered as CP117-iron complex and only 1.3 ± 0.2% as CP102-iron complex in the iron-overloaded animals. Total recovery of the administered dose was significantly different between normal (po) and iron-overloaded groups of animals, decreasing from 76.4 ± 11.7% to 57.2 ± 9.6%, respectively (p ≤ 0.05). There was no significant difference between the two routes of administration in normal animals, with total recovery of the administered dose of CP117 being 96.1 ± 9.1% by the intravenous route.
Intravenous and oral pharmacokinetics of CP117 was studied in the rat at a fixed dose of 450 μmol/kg. The AUC of the drug was 43.2 ± 9.1 μmol/liter · hr and 4.1 ± 1.8 μmol/liter · hrvia the intravenous and oral routes, respectively, thus indicating that the systemic bioavailability of the drug is <10%. Pharmacokinetic parameters of the drug determined by the intravenous route indicate that CP117 has a plasma clearance of 10.9 ± 3.0 μmol/liter · hr, a mean residence time of 0.14 ± 0.05 hr, and volume of distribution at steady-state of 1.54 ± 0.52 liters · kg−1. The Cmax andtmax of CP117 were 12.1 ± 2.5 μmol/liter and 7.0 ± 2.7 min, respectively. The AUC of the main metabolite, CP102, decreased from 425.3 ± 8.5 μmol/liter · hr to 282 ± 31 μmol/liter · hr via the intravenous and oral routes, which is presumed to reflect differences in hepatic extraction and routes of elimination of the drug. Parallel absorption studies conducted using the in situ isolated rat gut loop model indicate that the majority of the administered dose was absorbed intact as the parent drug with mesenteric vein AUC values of 3.1 ± 1.7 mmol/liter · hr and 0.3 ± 0.04 mmol/liter · hr for CP117 and CP102, respectively.
The only effective treatment of β-thalassemia major, a hemoglobinopathic disorder, is to increase the hemoglobin levels by regular blood transfusion, without which the majority of the patients die within the first year of life (1, 2). Repeated blood transfusions, however, lead to an excess of iron due to the inability of humans to excrete iron produced from the breakdown of hemoglobin (2, 3). Excess iron found in thalassemic patients is distributed throughout the body, but is found in the highest concentrations within the liver and other highly perfused organs (4). The unregulated accumulation or iron causes tissue damage and failure of organs, such as the liver and heart and eventually death (5). Complications associated with the toxicity of iron after blood transfusion can, to a large extent, be alleviated by the use of specific metal scavenging agents or chelating agents to trap and allow excretion of excess and potentially toxic forms of iron from the body (1, 4-7).
DFO1 has been available for the treatment of iron overload for >30 years. The major limiting factor of DFO is that it is inactive when administered orally and only causes sufficient iron excretion to keep pace with transfusion regimes when given either subcutaneously or intravenously over 8–12 hr for 5–7 days/week (8,9). For this reason, many patients find it difficult to comply with the treatment, and some even stop taking the drug altogether. There is, therefore, an urgent need for the development of an orally active alternative to DFO.
The HPO class of compounds (fig. 1) are currently one of the main candidates for development of orally active iron chelating alternatives to desferral (DFO) (2, 10, 11). The 1,2-dimethyl (CP20) and 1,2-diethyl (CP94) derivatives from the HPO class of chelating agents have over the past few years been evaluated in the clinic for metabolism, pharmacokinetics, and urinary iron excretion in thalassemic patients (12-14). Both the aforementioned first-generationHPO compounds unfortunately undergo extensive phase II metabolism to form the nonchelating 3-O-glucuronide conjugate, severely restricting the clinical use of these compounds at acceptable dose levels (12, 15).
The critical dependence of chelator efficacy on metabolic behavior has led the authors to design HPO compounds that do not undergo extensive conjugation reactions with glucuronic acid (16). This will permit the use of lower doses, because the bulk of the absorbed dose would retain its chelating ability. 1-Hydroxyalkyl derivatives of HPOs, such as CP102 that do not undergo such unfavorable biotransformation reactions in a variety of animal species (including humans), have since been identified (16, 17). Although the use of 1-hydroxyalkyl derivatives of HPOs may offer a significant improvement over previously evaluated HPOs, a possible disadvantage of these compounds, especially with some of the more hydrophilic analogs, is poor oral absorption and extraction by the liver—the major iron storage organ. A strategy to improve chelation efficiency further while minimizing drug-induced toxicity is to deliver the drug selectively to target iron pools such as the liver. The development of 1-hydroxyalkyl derivative esters with increased hydrophobicity is one route that has been considered to improve both drug absorption and hepatic extraction (18). To fulfill this design objective, it is essential that 1-hydroxyalkyl esters are absorbed intact from the GIT. Ideally, selective hydrolysis by hepatic carboxyesterases will allow generation of hydrophilic metabolites within hepatocytes. These metabolites will thus have the potential to chelate liver iron and, once in the systemic circulation, reticuloendothelial-derived iron without significant permeation of other tissues (fig. 2).
A variety of ester types are presently being considered for use to alter the physicochemical properties of hydrophilic hydroxyalkyl hydroxypyridinones chelating agents to improve both drug absorption and allow hepatic targeting. The present study describes the metabolism and pharmacokinetics of one such compound, CP117, in the rat.
Materials and Methods
Chemicals.
The HPOs used in the present study (CP102, CP94, and CP41) were synthesized as described previously (19). The synthesis of CP117 and CP162 is described herein. Characterization of the compounds was conducted using IR spectroscopy (Perkin-Elmer 298), proton NMR [Perkin-Elmer R32 (90 MHz)], MS [Vacuum Generators 16F (35 eV)], and elemental analysis (University of Manchester, Analytical Laboratory). Sodium dihydrogen orthophosphate, disodium hydrogen orthophosphate, sodium azide, β-glucuronidase (type 1X-A fromEscherichia coli), atomic absorption standard iron (1020 μg/ml in 0.1 M HCl), and iron dextran (100 mg Fe/ml) were purchased from Sigma Chemical Company (Poole, Dorset, UK). EDTA (trisodium salt Convol) was obtained from BDH (Poole, Dorset, UK). PBS was prepared and sterilized in-house. HPLC-grade acetonitrile (MeCN) was purchased from BDH. All other reagents used were of analytical grade. Sodium phosphate buffer (10 mM) was prepared using deionized water (Waters Milli-Q system). All compounds administered to animals used in the present study were prepared in PBS.
Synthesis of CP117.
CP102 prepared from the aforementioned method (19) was treated with trimethylacetyl chloride (pivaloyl chloride–Aldrich) and triethylamine in dimethylformamide at 75°C for 18 hr under an inert atmosphere to afford the diester 1-(2′-trimethylacetoxyethyl)-2-ethyl-3-(trimethylacetoxy)-pyridin-4-one; m.p. 133°–134°C; C19H29O5N: % calculated—C, 64.9; H, 8.3; N, 3.9; % found—C, 65.2; H, 8.1; N, 3.9. The diester was subsequently hydrolyzed in water at 65°C for 4 hr to form the monoester (CP117) as a free base; m.p. 141°–144°C; IR (nujol): 3130, 1710, 1620, and 1580 cm−1; ′H NMR (DMSO-d6): δ1.18 [s, 9H, C(CH3)3], 1.27 (t, 3H, CH2CH3), 2.85 (q, 2H, CH2CH3), 4.02–4.45 (m, 4H, NCH2, and CH2O), 5.25 (b, 1H, OH), 6.37 (d, 1H, 5-H), 7.25 (d, 1H, 6-H); MS: m/z 268 [(M+H)+ (100)], 267 (34); C14H21O4N (CP117): % calculated—C, 62.8; H, 7.9; N, 5.2; % found—C, 62.7; H, 7.6; N, 5.3. The HCl salt for the ester was prepared by passing HCl gas through a solution of the free base in dry ethyl acetate. Subsequent addition of dry diethyl ether afforded CP117 as the hydrochloride salt; m.p. 184.5°–186°C; ′H NMR (DMSO-d6): δ1.07 [s, 9H, C(CH3)3], 1.18 (t, 3H, CH2CH3), 3.01 (q, 2H, CH2CH3), 4.41 (t, 2H, CH2O), 4.72 (t, 2H, N+CH2), 7.46 (d, 1H, 5-H), 8.26 (d, 2H, 6-H), 8.5 (b, 2H, 2OH). Purity of compounds was also confirmed by HPLC.
Synthesis of CP162.
Use of 1-(2′-hydroxyethyl)-2-methyl-3-benzyloxy-pyridin-4-one hydrochloride, trimethylacetyl chloride, and pyridine in dimethylformamide at 60°C afforded the diester of CP40; m.p. 163°–164.5°C: C18H27NO5: % calculated—C, 64.0; H, 8.0; N, 4.1; % found—C, 64.2; H, 8.3; N, 4.0.
The diester was hydrolyzed in water to afford the monoester CP162; m.p. 139°–140°C; IR (nujol): 3150, 1720, 1620, 1570 cm−1′H NMR (DMSO-d6): δ 1.1 [s, 9H, C(CH3)3], 2.33 (s, 3H, CH3), 4.29 (s, 4H, NCH2CH2O), 6.15 (d 1H, 5-H), 6.35 (b, s, 1-OH), 7.56 (d, H, 6-H); MS: m/z 253 M+ (92) 129 (100); C13H19O4N: % calculated—C, 61.6; H, 7.5; N, 5.5; % found—C, 61.7; H, 7.5; N, 5.4. The HCl salt of CP162 was prepared as previously described for synthesis of CP117 HCl salt. m.p. 186.5°–188°C; ′H NMR (DMSO-d6): δ 1.07 [s, C(CH3)3], 2.58 (s, 3H, CH3), 4.4 (t, 2H, CH2O), 4.69 (t, 2H, N+CH2), 7.45 (d, 1H, 5-H), 8.26 (d, 1H, 6-H), 9.8 (b, 2OH). Purity of the compound was also confirmed by HPLC.
HPLC System.
A Hewlett-Packard model 1090 M-II HPLC system with an autoinjector, autosampler, and diode array detector linked to a HP 900-300 data station was used to measure drug and metabolite concentrations. Treated urine samples were analyzed by reversed-phase HPLC using a Hypercarb column packed porous graphatized carbon column (10 × 0.46 cm) (Shandon Scientific Ltd., Runcorn, Cheshire, UK). A polymer PLRP-S column (15 cm × 4.6 mm, i.d. 5 μm) (Polymer Laboratories Ltd., Church Stretton, Shropshire, UK) was used to analyze treated plasma samples to allow separation of the sacrificial ester and hydrolyzed product from that of drug and its metabolite (refer to pharmacokinetic studies). The mobile phase was a mixture of aqueous NaH2PO4:MeCN. NaH2PO4solution (10 mM) was prepared by dissolving the 1.56 g of the monobasic sodium salt of phosphoric acid in 800 ml of deionized water; 2 ml of 1 M EDTA Tris-sodium was then added and the pH adjusted to 3.0 with orthophosphoric acid before being made up to volume (1 liter). The mobile phase was pumped through the column at a flow rate of 1 ml/min. The following gradient systems were used (min/% MeCN): 0/16, 2/16, 4/55, 6/55, 8/16, 12/16, and 1/2, 5/16, 6/24, 13/27, 14/2, and 18/2 for urine and plasma samples, respectively. The eluant was monitored at 285 nm.
Urinary Recovery Studies.
Urinary recovery studies were conducted in normal (intravenous and oral) and iron-overloaded rats (oral only) to compare the different routes of administration. For the oral dosing studies wherein the influence of iron overload was determined, two groups of five male Wistar rats (200–240 g) were used. One of the aforementioned groups were iron-overloaded by the administration of 4 intraperitoneal injections of iron dextran over a period of 14 days (days 1, 5, 8, and 12) such that a final total body iron loading of 500 mg Fe/kg was achieved. The other group represented an age-/weight-matched control of the iron-overloaded animals. Animals were housed in a 12-hr light/dark cycle temperature-controlled (22°C) metabolic room maintained at a humidity level of 50%. Food and water were provided ad libitum.
After an equilibration period of 14 days, rats were fasted overnight. The test chelator, CP117, was administered by gavage at a dose of 450 μmol/kg, and rats were placed in perspex metabolic cages. The chelating agent was prepared from its corresponding hydrochloride salt in PBS at a concentration of 450 μmol/ml. A standard laboratory diet and water was provided 1-hr postadministration of the chelating agents. Urine was collected for a total period of 72 hr in 24-hr fractions in containers surrounded by dry ice. On completion of the collection period, residual urine adhering to the metabolic cages was carefully washed with a small volume of saline. This was then pooled with the collected urine. Samples were stored frozen at −20°C until required for analysis. Stability studies (data not shown) indicate that negligible (<1%) loss of CP102 and CP117 occurs during the collection period.
Administration of CP117 via the intravenous and oral routes were compared to assess how the different routes affect the resultant recoveries of drug/metabolites in the urine. Five male Wistar rats (300–350 g) had only their jugular vein cannulated as described for the pharmacokinetic study. Rats were then administered the HCl salt of CP117 as single bolus dose (450 μmol/kg, iv), and urine was collected and processed as described previously.
Analysis of Urine Samples.
Iron complexes of bidentate chelating agents, such as the HPOs unlike hexadentate ligands, have a tendency to dissociate on chromatographic columns, thus preventing simple analysis of both the free ligand and that of the intact bidentate metal-chelate complex (20). To allow quantification of the drug and its metabolites, both chelating and nonchelating, three separate assays were necessary due to differences in extractability into DCM.
Assay A: Quantification of Bound and Unbound Drug {[(CP117)3-Fe] + [CP117]} and Chelating Metabolite {[(CP102)3-Fe] + [CP102]}.
This assay measures both the free ligand and iron-bound forms of both CP117 and its chelating metabolite, CP102. Addition of a large excess of a high-affinity ligand, such as EDTA, was required to allow quantitative dissociation of the iron complexes of CP117 and CP102. Because the bidentate HPOs bind in a strict stoichiometric manner (three ligands are always associated with an iron atom), each ligand provides 2 of a total of 6 coordination positions around the metal ion (17). Data obtained from this assay can be expressed as in eq. 1, where ligand (L) is applicable to both CP117 and its metabolite, CP102.
Assay B: Quantification of Glucuronide Conjugates.
Analysis of nonchelating metabolites, such as the glucuronide conjugates of CP117 and CP102 that are not extracted into DCM, was conducted by incubating urine samples with a sufficient excess of β-glucuronidase to ensure complete hydrolysis to reform the parent drug. The procedure for assay B is similar to assay A, with the exception that the combined internal standard and EDTA solution also contained 5000 units/ml β-glucuronidase (type IX-A from E. coli). Samples were incubated for 16 hr at 37°C to ensure complete hydrolysis of the glucuronide conjugate. This assay quantitates any species that on dissociation or hydrolysis leads to the reformation of the parent drug and measures the total drug recovered from the urine sample, including unchanged drug, its corresponding iron complex, and glucuronide conjugate (eq. 2). Differences between assays A and B, in the presence ([L]T) and absence of β-glucuronidase ([L]t), accounts for the glucuronide conjugate (L-GLUC) present in the urine sample (eq. 3).
Data obtained from assay B can be expressed as in eq. 3.
Assay C: Quantification of the Iron Complex.
To determine the relative contributions of the free ligand and the corresponding iron complexes in assay A, a method that directly measures one of the aforementioned species is required. Iron content in the urine samples from both normal and iron-overloaded animals was measured using an atomic absorption spectrometer. Values obtained are corrected for background excretion for both normal and iron-overloaded animals. Data obtained from the direct measurement of iron in assay C can, therefore, be used to solve eq. 1 and provide information on the free ligand concentrations in urine (eq. 4) (L3-Fe = iron complex):
Aliquots of urine samples (1 ml) from normal and iron-overloaded animals were centrifuged at 1000g for 2 min using 0.4-μm centrifuge filters. Urinary iron content was then determined using a Pye Unicam SP9 atomic absorption spectrometer. A calibration curve was constructed over the range of 0–2.0 mM by adding known quantities of the HPO iron complexes prepared by adding three equivalents of the ligands to one equivalent of atomic absorption standard solutions of ferric chloride to blank urine samples. Iron content was measured using atomic absorption spectrometry described previously. Calibration curves obtained were linear with correlation coefficient values ≥0.995. The CV% and M%D of assay C was <5% over the entire concentration range. The MQL for iron using this method was 0.5 μM.
Calibration curves for assay A were constructed over the range of 0 to 1 mM by spiking known quantities of drug and chelating metabolite into blank urine collected from rats. These samples were then processed and analyzed using the sample preparation procedure described in the preceding section. Peak area ratios (HPO/internal standard) were plotted against concentration of the HPO. Calibration curves obtained were linear over the entire range, with correlation coefficient values ≥0.995. Accuracy and precision of the assays, as indicated by CV% and M%D was <3%. The MQL for CP117 and CP102 were 0.5 and 1.0 μM, respectively.
Calibration curves for assay B were constructed as for assay A, by spiking known quantities of the glucuronide conjugates into blank urine collected from rats. The presence of known quantities of iron complex and free ligand spiked into blank urine samples only marginally affected quantification of the glucuronide conjugates. The CV% and M%D of the assay in blank urine was <3% for the compound. In the presence of 50 μM iron complex and free ligand spiked into blank urine, the CV% and M%D of the glucuronide conjugate for the compound was <5% between the concentration range 0.25–1 μmol/ml and <10% from 0–10 nmol/ml. The MQL for the glucuronide conjugates of both CP117 and CP102 was 2.0 nmol/ml.
Pharmacokinetic Studies.
Male Wistar rats weighing ∼350–400 g were housed under conditions described for urinary recovery studies. Animals were surgically prepared before administration of CP117. The jugular vein and carotid artery cannulae (0.58 mm i.d., polythene; Portex, Hyte, Kent, UK) were implanted under light anesthesia using a mixture of fentanyl and fluanisone (Hypnorm, Janssen Pharmaceuticals Ltd., Grove, Oxford, UK) and midazolam (Hypnovel, Roche Products Ltd., Welwyn Garden City, Herts, UK) at a dose of 2.5, 0.08, and 5 mg/kg, respectively. The inserted cannulae were secured with sutures before being exteriorized in the dorsal intracapsular area. The surgical incision was closed using biodegradable sutures and the wound swabbed with heparinized saline-dipped cotton wool. The patency of the cannulae was maintained by flushing with heparinized saline (200 μl, 100 units/ml). Rats (N = 6) were allowed a recovery period of 24 hr before the administration of 450 μmol/kg of CP117 prepared in normal saline (450 μmol/ml) via the jugular cannulae as a single bolus injection. For oral pharmacokinetic studies, only the carotid artery was cannulated and the drug administered by gavage.
Blood samples (∼200 μl) were obtained from the conscious ratsvia the intraarterial cannulae at 5, 10, 15, 20, 25, 30, 45, and 60 min and at 2, 3, 4, 5, 6, 7, and 8 hr into heparinized microcentrifuge tubes containing 10 μl of “sacrificial ester” (CP162; 100 mg/ml). The pivaloyl ester of a structurally related HPO (CP162) at high concentrations was added to minimize plasma hydrolysis of CP117. After each sample collection, blood volume was replenished with an equal volume of heparinized saline (200 μl, 100 units/ml). The total amount of blood removed from the rats during the study period was ∼10% of the total blood volume of the animals. Blood samples obtained were immediately centrifuged at 4°C after collection at 1000g for 2 min using a microfuge to obtain plasma. Plasma (100 μl) was then quickly transferred to extraction tubes containing 10 ml DCM and 100 μl of internal standard (100 μg/ml CP41). Assessment of the possible extent of plasma hydrolysis of CP117 during the aforementioned procedure was conducted by mimicking, as closely as possible, the conditions of blood collection and the subsequent delay in preparation of plasma and finally precipitation of plasma proteins in DCM. Negligible levels of the hydrolytic product, CP102, were detected at high (100 μg/ml) and low (5 μg/ml) concentrations of CP117 in the presence of the sacrificial ester, accounting for <1% and 3% of CP117 as CP102, respectively.
CP117 and CP102 were quantified by constructing calibration curves over the range of 0–500 μM by spiking 10 μl of various amounts of drug and metabolite into 90 μl of blank plasma containing 5 μl of sacrificial ester (CP162, 100 mg/ml) to which was subsequently added 100 μl of the internal standard, CP41 (100 μg/ml). Samples were then extracted with 1 × 10 ml of DCM, evaporated to dryness, reconstituted with mobile phase (100 μl), and finally analyzed by reversed-phase HPLC. Calibration curves obtained were linear over the entire range with correlation coefficient values ≥0.995. The CV% and M%D was <5%. The MQL for CP117 and CP102 were 0.5 and 1.0 μM, respectively.
Pharmacokinetic Analysis.
Data obtained was analyzed using noncompartmental analysis. AUC0−∞ after intravenous bolus and oral doses were determined by using the trapezoidal rule from time 0 (determined by extrapolation) to the last sampling time point t*, and the AUC from the last sampling time point t* to infinity was determined by dividing the last plasma concentration C* by the terminal elimination rate constant (λz). Terminal elimination rate (λz) was determined from the slope of the regression line fitted to the log plasma concentration-time data of the terminal phase by the method of least squares. t1/2 was calculated by dividing ln 2 by the elimination rate constant. The total plasma CL was calculated using eq. 5. Contribution of CLR toCL was determined by dividing the amount of drug recovered unchanged in the urine (oral) in time t* divided by the AUC in the same time, obtained via the same route of administration. AUMC is the area under the curve of the product of concentration and time vs. time. The AUMC from time 0 tot* was calculated by means of the trapezoidal rule and fromt* to infinity being estimated using eq. 6, whereC* is the concentration at last sampling time t*.Vdss was calculated from eq. 7, the MRT was determined by eq. 8 and MAT was calculated by eq. 9.tmax and Cmax after oral administration were assessed by visual inspection of the blood concentration-time curve. The oral bioavailability (F) was calculated by the AUC ratio method [F = (AUCpo)/(AUCiv)] after the administration of 450 μmol/kg of appropriate compound.
Absorption Studies.
Absorption studies were conducted using the in situ isolated rat gut loop model. Male Wistar rats were anesthetized as described previously in the pharmacokinetic studies. Although still under anesthesia, a 10-cm section of the jejunum was isolated and the two ends ligated with nylon sutures. Mesenteric vein branching from the isolated vein was cannulated (0.58 mm i.d., polythene; Portex). CP117 at a dose of 450 μmol/kg (prepared in 1 ml PBS) was administeredvia a ligated cannulae into the isolated gut loop. Blood samples were collected both before and after drug administrationvia the mesenteric vein for 60 min at 5-min intervals into heparinized microcentrifuge tubes containing 10 μl of sacrificial ester (CP162, 100 mg/ml) and processed as described for pharmacokinetic studies.
After 60 min, the animals were killed, and the ligated gut section removed, homogenized (in the presence of residual contents), and analyzed for CP117 and CP102 to determine the proportion of dose yet to be absorbed. Approximately 1 g of gut tissue was homogenized in 20 ml PBS containing 0.1% Triton X-100. The supernatant obtained on centrifugation of homogenized tissue was quantified for CP117 and CP102 using the method described for plasma samples. Calibration curves were constructed by spiking known quantities of CP117/CP102 to an equivalent amount of homogenized tissue obtained from control animals. The CV% and M%D were <5%. The MQL for CP117 and CP102 was 2.5 and 5 nmol/g tissue.
Statistical Analysis.
Statistical analysis was determined by Student’s paired and unpairedt test (Statworks). p ≤ 0.05 was considered statistically significant.
Results and Discussion
Metabolism and pharmacokinetic studies are crucially important for the design of safe and effective chelating agents due to their narrow therapeutic safety margin. One such requirement is to ensure that chelating agents are not metabolically degraded to metabolites that lack the ability to bind iron. This would allow the use of lower doses and correspondingly decreasing the risk of inducing toxic side effects. The use of 1-hydroxyalkyl derivatives of HPOs identified by the authors that are either nonmetabolized or selectively metabolized will overcome the aforementioned limitations. Further therapeutic benefit can be gained by ensuring that the chelating moiety, either as the drug or chelating metabolite, is delivered to target site(s) at an appropriate concentration, rate, and duration to ensure interception of iron from targeted body stores of the metal.
Hepatic iron is one of the two major pools of this metal in the body and is therefore an obvious target for the design of new therapeutic agents. Data generated by us (18) and others (19, 21) suggest that iron bound within hepatocytes and from the extracellular space is strictly compartmentalized and behave independently of each other due to their inability to permeate biological membranes. Chelation of hepatic iron leads to exclusive excretion via the bile and that of extracellular iron via the urine (22). Recovery data obtained from iron-overloaded animals in comparison to normal animals are potentially useful, because they provide information on the ability of chelating agents to access the hepatic iron pool. The difference in urinary recovery of the drug/metabolite between the two groups arises, because chelation of hepatic iron will lead to a decrease in the recovery of the drug/metabolite in the urine. Such studies also provide an indication on the likely in vivo efficacy of a chelating agent.
Urinary Recovery Study.
Urinary recovery studies of CP117 were conducted in normal (intravenous and oral, N = 5) and iron-overloaded animals (oral only, N = 5) to assess the influence of iron. Iron loading was achieved by using iron dextran, a polymeric form of iron that is selectively taken up by reticuloendothelial cells and subsequently redistributed via plasma transferrin and when transferrin saturation occurs, as nontransferrin bound iron to parenchymal iron stores such as hepatocytes. CP117 was also injected as an intravenous bolus to normal animals to compare routes of drug administration.
The proportion of dose recovered in the urine of both normal (intravenous and oral) and iron-overloaded (oral only) rats postadministration of 450 μmol/kg CP117 is depicted in fig.3. Unchanged drug was not detected in the urine of normal animals, and only very low levels were detected in iron-loaded animals (0.8 ± 0.7%). In normal rats, the majority of the dose recovered in the urine was as the hydrophilic metabolite, CP102, accounting for 69.7 ± 9.4% (oral) and 80.7 ± 7.9% (intravenous) of the administered dose. There was, however, a dramatic decrease in the amount of CP102 recovered (47.7 ± 5.9%) in the iron-loaded group of animals (p ≤ 0.05). Lower but similar amounts of glucuronide conjugate of CP102 (6.5 ± 3.0%) and (7.1 ± 2.5%) were also present in normal and iron-overloaded rats, respectively. However, there was a significant increase in CP102 glucuronide conjugate (13.7 ± 3.0%) (p ≤ 0.05) after intravenous administration of CP117.
Urinary recovery of CP102, the metabolite of CP117, after administration of CP102 itself (450 μmol/kg, po) in normal animals was 68.4 ± 12.2%. This value is virtually identical to that obtained when CP117 was administered at a similar dose and route. In the overloaded animals, there was also a small but significant rise in the amount of metabolite recovered as the corresponding iron complex, with 1.3 ± 0.2% being excreted compared with 0.64 ± 0.2% in normal animals (oral) (p ≤ 0.05). Total recovery of the administered dose was significantly different between normal (oral) and iron-overloaded groups of animals, decreasing from 76.4 ± 11.65% to 57.2 ± 9.6%, respectively (p ≤ 0.05). There was no significant difference between the two routes of administration in normal animals with total recovery of the administered dose being 96.1 ± 9.1% when CP117 was given intravenously.
Results obtained indicate that extensive hydrolysis of CP117 occurs because only negligible levels of unchanged drug was recovered in the urine of both normal (oral and intravenous) and iron-overloaded animals. From urinary recovery data alone, it is difficult to predict with any degree of certainty the likely site for CP117 hydrolysis. Ideally, hydrolysis within hepatocytes is favored, but esterases present elsewhere in the body could be equally responsible for the extensive excretion of CP102 in the urine. A likely explanation for the significant reduction of CP102 recovered in the urine of iron-overloaded animals compared with normal animals is that hepatic extraction of CP117 is occurring subsequent to which ester hydrolysis/chelation of intracellular iron leads to a decrease in the amount of drug recovered in the urine as a result of biliary excretion of the corresponding iron complexes of CP117 and CP102.
Conversely, formation of the iron complex in the extracellular space can also dramatically reduce the amount of drug available for metabolism. This will, in turn, lead to a greater proportion of the drug being available for renal excretion. In a similar study conducted previously using relatively hydrophilic compounds (18), increased urinary recovery of drug/metabolites were observed. However, in this study, the extraction of CP117 into the liver is presumed to be both extensive and rapid. Iron in the extracellular space is therefore unlikely to significantly affect the recovery of CP117.
In addition to hydrolysis of the ester moiety of the parent drug, phase II biotransformations can in principle take place within hepatocytes to form the 3-O-glucuronide conjugate of both CP117 and CP102. However, after oral administration in both the urine of normal and iron-loaded animals, the glucuronide conjugate of CP117 was absent, and only small amounts were present after intravenous administration (normal) (0.8 ± 0.4%). Of particular interest is the dramatic increase in CP102 glucuronide levels after intravenous administration in normal animals and the fact that iron overload did not significantly affect this metabolic route.
Despite iron loading, only very small amounts of the iron complexes of drug and metabolite were recovered in the urine. A likely explanation for this is due to the failure of all animal models of iron overload to saturate transferrin fully, because the erythropoietic activity and, consequently, the plasma iron turnover is normal, despite high total body iron loading. The effect of this will be that iron can be donated by the metal complex to apotransferrin and this in turn will prolong the mean residence time of the drug/metabolite in the extracellular pool, thereby increasing the likelihood of urinary excretion of the free drug/metabolite and smaller quantities as the iron complex. This process will not occur in thalassemic patients, because the transferrin is almost continually saturated and, therefore, iron complexes formed within the extracellular space will be excreted unchangedvia the urine.
Pharmacokinetic Study.
Urinary recovery studies, although useful, do not provide information to account for the quantitative behavior of the drug with respect to time. Information relating to these aspects can be obtained from pharmacokinetic investigations. The pharmacokinetics of CP117 was studied at a fixed dose of (450 μmol/kg, N = 6)via both the intravenous and oral routes. Plasma concentration-time plots of CP117 and its hydrolytic product, CP102, postintravenously and orally are depicted in fig. 4, a andb, respectively. The pharmacokinetic parameters of CP117 postintravenous and oral administrations are tabulated in tables 1 and 2.
The intravenous plot (fig. 4a) clearly shows that CP117 has only a limited plasma life-time and is rapidly metabolized by ester hydrolysis to the corresponding 1-(2′-hydroxyethyl) metabolite. The AUC of CP117 when administered by the intravenous route was 43.2 ± 9.1 μmol/liter · hr. In comparison, the AUC of the metabolite (CP102), postintravenous administration of CP117 was significantly higher, with a value of 425 ± 9.0 μmol/liter · hr. The AUC of the metabolite is virtually identical to that obtained when CP102 was itself administered by an intravenous bolus at a similar dose (458 ± 38 μmol/liter · hr). The Vdss of CP117 was 1.54 ± 0.52 liters/kg. The plasma protein binding of CP117 is relatively low (≤40%); therefore, no significant changes in distribution are expected to occur as a result of saturation of protein binding sites. The MRT of CP117 was 0.14 ± 0.05 hr (t1/2 = 0.1 ± 0.04 hr). The rapid elimination of CP117 from plasma is also reflected by its CLof 10.9 ± 3.0 liters/kg/hr.
CP117, when administered by the oral route (fig. 4b) is rapidly absorbed from the gut with a correspondingtmax of 12.1 ± 2.5 min and aCmax of 7.00 ± 2.5 μmol/liter. The MRT of CP117 was 0.22 ± 0.03 hr (t1/2 = 0.15 ± 0.02 hr), giving a MAT of 0.08 hr (∼5 min). The AUC of both CP117 and CP102 were also significantly reduced, compared with the intravenous route, with values of 4.1 ± 1.8 and 282 ± 32 μmol/liter · hr, respectively. In comparison, the AUC of CP102 when administered as an oral bolus was 318 ± 46 μmol/liter · hr. AUC values of CP117 obtained by the intravenous and oral routes suggest that <10% of the administered oral dose reaches the systemic circulation, presumably due to extensive first-pass metabolism.
An alternative explanation for the decreased AUC values for both the drug and the metabolite is ester hydrolysis of the parent drug before absorption from the GIT. To discount this possibility, a limited number of animals (N = 3) were studied using an in situ isolated rat gut loop model. To complement measurements of drug/metabolite levels within the mesenteric vein, direct measurement of CP117/CP102 present within the gut loop and residual contents to determine the proportion of dose yet to be absorbed at the end of the study period (60 min) was conducted. In all cases, <5% of the dose was accounted for, further discounting the possibility of CP117 hydrolysis to CP102 before absorption.
Figure 5 shows the mesenteric vein concentration-time profile of both CP117 and CP102. The AUC0–60 min of CP117 was 3.1 ± 1.7 mmol/liter · hr. The AUC of CP102 in contrast was ∼10% of that of the parent drug, with a value of 0.3 ± 0.04 mmol/liter · hr, thus indicating that CP117 is predominantly absorbed intact from the GIT. Comparison of the AUC values of CP117 and CP102 alone, however, is unlikely to be truly reflective of the relative ratio of drug/metabolite absorption from the GIT, because further enhancement of the mesenteric vein AUC of the metabolite could arise from postabsorption hydrolysis of CP117.
Another significant observation from the pharmacokinetic data was the dramatic reduction in the AUC of the metabolite (CP102), decreasing by half when CP117 was administered by the oral route. A possible explanation is that when the drug is given by the oral route, the majority of the dose is presumed to be extracted and metabolized by hepatocytes followed by excretion of the metabolite by both the urinary and biliary routes. However, when CP117 is administered by the intravenous route, extensive tissue distribution of the drug to both hepatic and nonhepatic eliminating tissues is presumed to have occurred before excretion, thus prolonging the MRT. A consequence of a rapid MAT at a dose of 450 μmol/kg could possibly result in saturation of the hepatic first-pass elimination process. This can be verified by conducting pharmacokinetic studies at lower doses or by controlling the rate of drug dissolution. These approaches may lead to the complete extraction of CP117 and minimization of potential toxicity problems relating to the extensive distribution of hydrophobic esters of iron chelating agents.
It may be possible to explain the apparent decrease in glucuronidation of CP102 after an oral dose of CP117 (6.6 ± 2.3%), compared with intravenous dosing (13.7 ± 3.0%) by estimation of the absolute bioavailability of CP102 after CP117 administration. This is performed by comparing the AUC of CP102 from oral CP117 (282 μmol/liter · hr) with the AUC of CP102 (458 μmol/liter · hr) of an equimolar intravenous dose of CP102 that can be accounted for by incomplete bioavailability. Estimated bioavailability of 61% virtually accounts for the decreased glucuronidation. This, indeed, will also explain the doubling of CP102 glucuronide after intravenous administration of CP117, which exactly mirrors the corresponding increase in the AUC of CP102 (intravenous) (425 ± 9.0 μmol/liter · hr), compared with AUC (oral) (282 ± 32 μmol/liter · hr).
Results obtained in this study indicate that CP117 is rapidly absorbed from the GIT, predominantly in the intact form. Less than 10% of the administered dose reaches the systemic circulation, and it is presumed the remainder of the absorbed dose undergoes extensive presystemic metabolism, mainly due to extraction and subsequent biotransformation to the metabolite CP102 within the liver. It is likely that only a small proportion of the absorbed dose is hydrolyzed between passage from the mesenteric vein to the liver due to both limited contact with blood and the relatively low whole blood esterase activity (4.0 ± 0.22 nmol/ml/min), compared with the liver (620 ± 250 nmol/g/min).2
One of the interesting observations made in the present study are the similarities in the AUC values of CP102 when administered either as CP117 or CP102 itself by both the intravenous and oral routes. Thus, the bioavailability of CP102 is not significantly affected by the administration of CP117 by either route. This is partly reflected by efficacy studies conducted using the 59Fe-ferritin rat model, wherein hepatic iron stores are specifically radiolabeled.3 The results indicate that CP117 is only marginally more effective than CP102 when administered (oral) at a dose of 450 μmol/kg. In this animal model, the percentage of 59Fe excreted in the feces, after correcting for background, were 5.4 ± 1.1% and 8.3 ± 1.0% for CP102 and CP117, respectively, and significantly less than CP94 (15.4 ± 2.0%) at a similar dose. However, a limitation of the ferritin model is that it uses only normal animals, and differences between chelators are not so apparent because there is only a limited amount of iron available for excretion. Further studies in iron-overloaded animals may shed more light on the likely clinical efficacy of CP117 in thalassemic patients.
Acknowledgments
We are grateful to Dr. Bob Ings for the constructive comments made during the preparation of this manuscript. R.C. thanks Engineering and Physical Sciences Research Council for providing a Ph.D. studentship.
Footnotes
-
Send reprint requests to: Dr. Surinder Singh, Department of Pharmacy, King’s College London, Manresa Road, Chelsea, London SW3 6LX, UK.
-
↵2 Choudhury et al., unpublished observations.
-
↵3 Singh et al., unpublished observations.
- Abbreviations used are::
- DFO
- desferrioxamine
- HPO
- 3-hydroxypyridin-4-one
- CP20
- 1,2-dimethyl-3-hydroxypyridin-4-one
- CP94
- 1,2-diethyl-3-hydroxypyridin-4-one
- CP102
- 1-(2′-hydroxyethyl)-2-ethyl-3-hydroxypyridin-4-one
- GIT
- gastrointestinal tract
- CP117
- 1-(2′-trimethylacetoxyethyl)-2-ethyl-3-hydroxypyridin-4-one
- CP41
- 1-(3′-hydroxypropyl)-2-methyl-3-hydroxypyridin-4-one
- CP162
- 1-(2′-trimethylacetoxyethyl)-2-methyl-3-hydroxypyridin-4-one
- PBS
- phosphate-buffered saline
- MeCN
- acetonitrile
- DMSO
- dimethylsulfoxide
- CP40
- 1-(2′-trimethylacetoxyethyl)-2-methyl-3-(trimethylacetoxy)-pyridin-4-one
- DCM
- dichloromethane
- CV%
- coefficient of variation
- M%D
- mean percentage difference
- MQL
- minimum quantifiable level
- AUC
- area under the plasma concentration-time curve
- t1/2
- terminal half-life
- CL
- clearance
- CLR
- renal clearance
- AUMC
- area under the first-moment curve
- Vdss
- volume of distribution at steady-state
- MRT
- mean residence time
- MAT
- mean absorption time
- tmax
- time to peak
- Cmax
- peak plasma
- Received July 31, 1996.
- Accepted December 6, 1996.
- The American Society for Pharmacology and Experimental Therapeutics