ANALYT,CALB,OCHEM,STRY 203, 116-120

(1992)

Quantification of Desferrioxamine Metabolites by High-Performance Chromatography and Simultaneous Radioactive Detection S. Singh,i

N. Mohammed,

Chelsea Department of Pharmacy, and *Department of Haematology. 98 Chenies Mews, London WClE

R. Ackerman,

and Its Iron Chelating Liquid Ultraviolet-Visible/

J. B. Porter,*

KingB College. 1imuersit.v LJniuersQ College Hosoital 6HX. linked Kingdom

and

R. C. Hider

of London. Monresa Road, London and Middlesex School of Medicine,

SW3

6LX,

United

Kingdom;

Received October 24, 1991

An HPLC-based method for quantification of desferrioxamine (DFO) and its iron chelating metabolites in plasma has been developed. This assay overcomes stability problems associated with DFO by the addition of radioactive iron to convert unbound drug and metabolites to radio-iron-bound species. A dual detection system utilizing uv-vis absorption and radioactive (8particle) detector was used to quantify total and radioiron-bound species. The use of octadeeyl silanol solid phase extraction cartridges permits concentration of samples and allows accurate quantification of drug and metabolites down to 0.1 nmollml. #cm 19% ~~~~~~~~~~~~~~ rnc.

Desferrioxamine (I; Fig. 1) is currently the only widely adopted drug to treat iron overload disorders which result from frequent blood transfusions. DFO* is also increasingly being used to treat aluminium intoxication (1,2). Over the last 10 years DFO has been shown to produce negative iron balance and prolong life expectancy in iron-overloaded patients with remarkably few side effects. However, for a variety of reasons, including poor patient compliance with subcutaneous DFO, many thalassemic patients still die before their third decade of life. With the more recent use of intensive chelation protocols (2100 mg/kg/day), a number of DFO-induced

’ To whom all correspondence should be directed. *Abbreviations used: DFO. desferrioxamine: FO, fernoxamine: ODS, oetadecyl silanol: M&N, acetonitnle: MeOH. methanol; NTA, mtrilotriaeetic acid, NTBI. nontransferrin-bound iron: Mops, I-morn pbolinepropanesulfonic acid; CV W, coefficient of variation; MID, mean percentage difference; IS, internal standard: *FO. radioactive ferrioxamine

toxicities have emerged such as retinal and auditory toxicity, although iron overload appears to protect patients from at least some of these effects (3). At the time of initial clinical introduction of DFO, it was not possible to study the metabolism andpharmacokinetics of the drug. This was largely due to the limitations of the analytical techniques then available. With development of reliable HPLC-based assays, detailed urinary metabolic profiles of DFO have been carried out (4,5). These studies were performed using a combination of HPLC and fast atom bombardment mass spectrometry, which led to the identification of four iron chelating metabolites of the drug. The metabolism of DFO was found to parallel that of the amino acid lysine in that oxidative deamination, +.xidation, and decarboxylation reactions take place. In addition, N-hydroxylation of the terminal group also occurs (5). While the long term goal of iron chelation therapy is to produce negative iron balance, it is also important to remove nontransferrin-bound iron (NTBI), a potentially toxic species found in plasma of iron-overloaded patients. DFO is able to chelate this iron fraction, leading to the excretion of ferrioxamine in urine. Preliminary clinical investigations indicate that NTBI levels have a marked bearing on DFO metabolism. Larger concentrations of the major iron chelating metabolites relative to DFO are detected at low/zero NTBI levels (5). The proportion of drug available for cellular uptake and subsequent metabolism is apparently directly dependent on the proportion of unbound DFO available in plasma. In order to more reliably assess patient progress to chelation therapy and to establish relationships between drug and iron metabolism, a reproducible and ac-

CHROMATOGRAPHIC

QUANTIFICATION

117

OF DESFERRIOXAMINE

I FIG. 1.

Structure

II of the iron-complexed

forms of desferrioxamine

curate method for measuring DFO and its metabolites, in both the iron-bound and unbound forms in plasma is required. Ideally this method should also permit concentration of the drug and metabolites and the incorporation of an internal standard to allow for sample variation during preparative and analytical procedures. A human pharmacokinetic investigation was recently carried out on DFO and two of its metabolites (6). This method unfortunately suffers from a number of problems, which include poor resolution of the uncharacterized metabolites. In the above study excess iron was added to plasma and total drug and metabolite levels were measured in the metal complexed form. The behavior of DFO and its iron complex (FO), however, differs considerably with regard to their membrane permeability and renal clearance rates (7). It is therefore vital to distinguish the two forms of drug and metabolites in order to relate iron and drug metabolism. DFO, in contrast to its iron-bound species, is readily hydrolyzed in rodent plasma, although such breakdown occurs less rapidly in humans (8). DFO and its metabolites are also unstable when stored frozen (-20°C) (M. J. Pippard, personal communication). A novel HPLC assay involving the use of a dual detection system has been developed to overcome these problems. This method involves the addition of radioactive iron to plasma samples which converts unbound drug and metabolites to radio-iron-bound species (Scheme I). The above has the additional advantage of overcoming the risk of introducing iron contamination during the subsequent sample preparation and HPLC procedures. Fur. thermore, formation of the iron complexes will enable HPLC &ants to be monitored at 430 nm instead of the uv region (215 nm) required for DFO quantification. Such measurements are less susceptible to interference

(I) and desferrioxamine-E

from solvent absorbance coeluting species. MATERIALS

AND

during

(II)

gradient

elution

and

METHODS

Chemicals

DFO was obtained from Ciba-Geigy Pharmaceuticals (Horsham, UK); DFO-E (Fig. 1) and N,N-dimethylDFO were kindly provided by Dr. H. Peter (Ciba-Geigy, Bale, Switzerland); NTA (disodium salt), potassium dihydrogen phosphate, dipotassium hydrogen phosphate, Mops buffer, and hexylamine were obtained from Aldrich (Milwaukee, WI). HPLC grade methanol (MeOH) and acetonitrile (M&N) were purchased from BDH (Poole, UK). All reagents were of analytical grade. Milli-Q water was used throughout the study. Sample

Collection

and Storage

Due to the limited stability of DFO and its metabolites in the iron free form, strict guidelines for collection and storage of plasma samples must be followed in order to minimize losses of the above species. To 0.9 ml of freshly collected plasma (within 10 min of blood sampling), 75 pl of preinternal standard (133 FM DFO-E in 0.2 M phosphate buffer, pH 7) and 25 /.d of 4 rn~ radioiron saturating solution (0.1 M HC1; 2.0 pCi) was added. The plasma samples containing the internal standard can be either processed immediately or stored frozen until required. HPLC

System

A Hewlett-Packard Model 1090M HPLC system complete with an autoinjector, autosampler, and diodearray detector attached to a HP 900-300 data station

118

SINGH

ET AL

A Spherisorb ODS-2 analytical column (150 X 4.6 mm i.d.; 5 pm) complete with a guard column (20 X 4.6 mm i.d.; 5 pm) of the same packing material was used for analysis of samples. Spherisorb ODS-2 due to its low residual silanol groups as a result of its high carbon loading greatly improves the peak shape of ferrioxamine. The mobile phase composition used in the iron exchange studies consisted of 20 rn~ phosphate, pH 7.0 (+2 mu NTA), with variable amounts of organic modifier (M&N), which was typically 9%. The flow rate was 1 ml/min and the eluant was monitored at 430 nm and scanned over the range 210-600 nm (960 ms/scan). A Rheodyne 7125 injector fitted with a 100~pl sample loop was used to introduce samples onto the HPLC column. Chromatographic conditions used for the separation of DFO and its metabolites is as follows (Fig. 2): Flow rate = 1 ml/min; Pump A (20 rn~ phosphate, pH 7.0, containing 2 rn~ NTA); Pump B (20 rn~ phosphate, pH 7.0), containing 2 rn~ NTA in 50% M&N). The following gradient program was used (min/% B): O/10, l/10, 8120. 12125, 13/10,18/10. Sample

Time

(min)

FIG. 2. HPLC cbromatogram of a plasma sample containing DFO an its iron cbelatine metabolites ohtsined from a thalassemic mtient undergoing intravenous infusion of DFO 60 r&kg124 h) as d&ted by (A) uv-vis (430 nm) and(B) radioactivity (P-detectwnl The identity of the peaks with respect to an mcrease in the elution order comespands to metaholite X (6 min), metaholite A. metabolite B. FO, metaholite D, metahalite D2 (oxidized metaholite D), and FO-E (internal standard). The structures of these metabolites are as previously reported (5).

was used. A Cannbera-Packard Radiomatic B-detector (Model A250 Flo-One/Data II) connected in series was used to analyze the &ant from the LC. A 500.~1 liquid scintillation cell was used for detection of radio-ironbound species (E = 460 KeV). Optiphase MP (Pharmacia, UK) scintillant was pumped through the flow cell at a flow rate of 4 ml/min. Alternatively, a solid scintillant cell (calcium fluoride, 300 ~1) can be used.

DED-mztabolites

+

Preparation

Samples

of Calibration

Curues

htablires 59(*h

l

-1s

+

--->

*Is

PO

FDmetabDlt~ 1. Conversmn

of Plasma

Calibration curves were constructed by spiking known amounts of DFO into plasma to which was added the preinternal standard and radio-iron saturating solution. Spiked plasma samples were put through the sample preparation procedure described above. Peak area ratios (DFO/IS) of both absorbance (430 nm) and

Etl

SCHEiVIE species.

Preparation

A single preparative step was performed which utilized solid phase extraction of plasma using ODS-Bond Elut cartridges (Anachem, UK). The cartridges were conditioned using l-ml aliquots of methanol, water, and 50 rn~ hexylamine in 20 rn~ phosphate (pH 7). The cartridges were washed with a further 1 ml of water prior to and after loading of radio-iron-saturated plasma samples containing the internal standard (*FOE). FO and its metabolites were finally eluted with 1 ml MeOH. The &ants were then evaporated to dryness using a centrifugal concentrator (JOUAN RC lo-221 at room temperature (25°C) for 30 min and finally reconstituted in 100 pl Mops buffer (20 rn~, pH 7) before analysis.

of unbound

lG!E~lit.es DFO, its metabolites,

and the preinternal

standard

(DFO-E)

to the corresponding

radio-iron-bound

CHROMATOGRAPHIC

QUANTIFICATION

OF DESFERRIOXAMINE

119

iron complexes, which are all stable even after prolonged periods of storage at -20°C. It is therefore apparent that a method for stabilization of the unbound form of the drug (and metabolites) is required. This can be achieved by the addition of excess radioactive iron to plasma samples of patients receiving DFO treatment. The radio-iron introduced is chelated only by the unbound proportion of DFO and its metabolites present (Scheme I). This has the effect of

a) w-vu

__ ____----

[mlroral

A

lSCHEME II. Schematic reprewntatian of an HPLC chromatogram of ferrioaamine (FO) and its internal standard (IS) as detected by (a) uv-vis and (b) radioactivity. Radioactive ferrioxamine (*FO) corresponds to DFO originally present in plasma.

radioactivity were plotted against known concentrations of drug. Linear behavior with correlation coefficients values >0.995 were obtained. Calibration curves of DFO are equally applicable to the iron chelating metabolites, which contain identical cbromophores and are each able to bind one iron atom. Drug and metabolite concentrations can also be expressed as concentrations per unit of whole blood by multiplying plasma values by (l-hematocrit). The assay precision as indicated by coefficient of variation (CV%,) and the accuracy by mean percentage difference (M%D) was less than 3.5 and 5% for concentrations above and 7 and 12% for concentrations below 1 nmol/ml. Day-to-day reproducibility as determined by CV% and M%D was less than 5%. Correlation coefficient values for calibration cures were greater than 0.99 for both spectrometric and radioactive data. The detection limit of the assay is 0.1 nmol/ml in plasma. RESULTS Desigrz

AND

DISCUSSIONS

of uu-ui.slRadioactiue

HPLC

Assay

Plasma samples obtained from patients undergoing DFO treatment contain both iron-bound and unbound forms of drug and metabolites. Unfortunately the unbound form of the drug and metabolite have limited stability in plasma and undergo breakdown even when stored at -20°C. This is in complete contrast to their

FIG. 3. HPLC chromatogram of an aliquot containing FO-Met B [retention time (RT) = 6 min] and radiolabeled FO (RT = 11 min) 88 detected by (A) uv-vis (430 nm) and radioactivity (B and C!). The above iron exchange experiment WBB carried out in 50 rn~ Mops, pH 7.4 125’CL and analyzed after (B) 4 h and(C) 24 h, respectively. The kinetically inert nature of the hexadentete hydroxamate moieties can be seen by only a smell increase in the amount of radioactive FO-Met B (