High-performance liquid chromatographic method for the separation and quantitative estimation of anti-parasitic melaminophenyl arsenical compounds

TRANSACTIONS

OFTHEROYAL

SCCIE~

OFTROPICAL MEDICINE AND HYGIENE (1994)88,357-359

High-performance liquid chromatographic quantitative estimation of anti-parasitic Bradley J. Berger and Alan H. Fairlamb*

357

method for the separation and melaminophenyl arsenical compounds

Department of Medical Parasitology, London School of Hygi&e and Tropical

Medicine, Keppel Street, London, WCIE 7HT, UK Abstract

A sensitive high-performance liquid chromatographic method has been developed for the separation and quantitative estimation of melaminophenyl arsenical drugs used in the treatment of human and veterinary trypanosomiases. Five clinically relevant compounds (melarsoprol, melarsonyl potassium, cymelarsan, melarsen oxide, and sodium melarsen) can be separatedwith an octadecylsilane reverse-phasecolumn and detected down to 10 pmol per injection. The chromatographic system utilizes a propanol gradient in water, with lithium camphor sulphonate as ion modifier, and detection by ultraviolet absorbance at 280 nm. Preliminary results indicate that the arsenical compounds can be extracted and concentrated from serum using octadecylsilane solid-phase extraction cartridges. This method represents the first specific and sensitive assaysystem for separating and quantitatively estimating this important classof antiparasitic agents. Introduction

African trypanosomiasis remains both a major public health and veterinarv hazard. with 50 million ueoule at risk and 1.0~ 107k&2 of Africa made unsuitable fdr rearing livestock (WHO, 1990). At present, treatment of secondary (cerebral) trypanosomiasls in humans is normallv limited to 2 classesof comnound: the ornithine decarbbxylase inhibitor difluoromethylornithine (DFMO, eflornithine) and the melaminophenyl arsenicals (WBRY, 1991). While many arsenicals, including sodium melarsen, melarsen oxide, and melarsonyl potassium have been tested in humans, only melarsoprol is currently used clinically. This latter compound, introduced in 1947 (FRIEDHEIM. 1949). is associatedwith a number of ierio& ii& effect;, incl&ing arsenical-induced encephalopathy in approximately 5% of cases(PBPIN& MILORD, 1991), but remains the drug of choice for treating secondary human trypanosomiasis in rural Africa. The existing therapeutic regimens have been developed largely on a trial and error basis and may therefore be contributing to the high incidence of drug toxicity or relapse after treatment. Even the most basic pharmacological properties of the arsenical drugs are almost completely unknown, mainly due to the lack of an easy and reliable method for their separation and quantitative estimation. To date, there have been descriptions of an enzyme-linked immunsorbent assayfor melarsoprol (MAES et al., 1988a, 1988b)! bioassaysinvolving incubation of biological sampleswith cultured Tppanozoon trypanosomes (HAWKING, 1962; BURRI & BRUN, 1992), and a polarographic method for measuring organic arsenicals ~CRISTAUet al., 1972). Several of these methods have be& used to deteimine’distribution and excretion of arsenicals (HAWKING, 1962; CRISTAU et al., 1975; BURRIet al., 1993), but none of the existing assaysis particularly convenient to use and none can separatearsenical compounds from each other or from potential metabolites. In order to address some of these problems, we have developed a simple high-performance liquid chromatographic (HPLC) assaywhich is capable of resolving and quantitatively estimating several arsenical trypanotides. Materials

and Methods

Sodium melarsen, melarsen oxide, melarsonyl potassium (Mel W), cymelarsan (Mel Cy), and melarsoprol (Mel B) were all obtained from Specia RhBne-Poulenc (Paris, France). HPLC grade 1-propanol was obtained from BDH (Poole, UK) and camphor sulphonate and lithium hydroxide from the Aldrich Chemical Co. (Gillingham, UK). All water was filtered and deionized by a Milh-QSO@system (Millipore, Watford, UK). A Beckman@HPLC system (Beckman Instruments, High Wycombe, UK), consisting of 2 model 1lOB *Author for correspondence.

pumps and a model 167 variable wavelength ultraviolet spectrophotometric detector, was used for all analyses. The column was a Beckman octadecylsilane 250x4.6 mm reverse-phasecolumn with 5 pm particle size kept at room temperature. Samples were injected through an Altex@ 210A manual injection valve (Beckman Instruments) equipped with either a 10 PL or 100 PL loop. All compounds were detected using a wavelength of 280 nm and data were collected, stored, and analysed using Beckman SystemGold@operating software. The mobile phase consisted of 0.25% camphor sulphonate (diluted from a 1% stock solution adjusted to pH 2.0 with LiOH) as buffer A and 0.25% camphor sulphonate/25% 1-propanol as buffer B. Using a flow rate of 1.0 mlimin, runs were initiated at 10% buffer B and held at this concentration for 5 min. A linear gradient from 10% to 85% buffer B was then maintained over 75 min, followed by a reverse gradient of 85-10% B over 2 min. This final concentration was then held for 10 min. Arsenical stock solutions were made at a concentration of 10 mM using distilled water (Mel Cy, Mel W), dimethylformamide (Mel B, melarsen oxide) or 0.1 M HEPESbuffer, pH 7.8 (sodium melarsen). A mixture containing 0.1 nmol/vL of each arsenical compound was then made from these stocks using 20% v/v propylene glvcol as diluent. A uortion of this mixture was oxidized 6; mixing with hydrogen peroxide at a ratio of 20 PL HzOz per mL of drug solution followed by 60 min incubation at room temperature. A total of 10 PL was then injected to give 1.0 nmol of each arsenical per injection. For standard addition curves, a mixture of melarsen, melarsen oxide, Mel W, and Mel B was made at concentrations of 500, 100, 50, 10, 5.0, and 1.0 nmol/mL in 20% propylene glycol. A total of 10 PL of each dilution was injected to give 5000, 1000, 500, 100, 50, and 10 pmol per injection. Foetal calf serum (Gibco, Uxbridge, UK) was ‘spiked’ with a mixture of melarsen oxide, sodium melarsen, Mel W, and Mel B to give a final concentration of 100 pmol each per mL of serum. Two mL of this sample were then passeh through a Prep-Sep@octadecylsilane solid-phase extraction cartridee (Waters. Watford. UK) which had been previously wisged witi 5 mL aceionithle and then 5 mL water. The cartridge was then washed with 5 mL water before elution with 2 mL of acetonitrile. This eluate was brought to near dryness under vacuum and resuspended to 1.0 mL with HPLC buffer A before injection of 100 yL aliquots. Results

To determine the relative retention times of the various arsenical drugs, standard solutions of the compounds were analysed by HPLC individually (data not shown) and as a mixture containing 1.O nmol of each compound (Fig. 1, B). All the drugs produced sharp peaks which were free from any interfering background

358

0.1 10

100

1000

10000

pmol Injected

II

ill

60

ill

sa

Time (mid

Fig. 1. Separation of arsenical drugs by high performance liquid chromatography (HPLC). A mixture of melaminophenyl arsenicalswas made in 20% propylene glycol so that a 10 I.LLinjection contained 1.0 nmol of each drug. (A) Blank chromatogram, 10 FL of HPLC buffer A injected. (B) Injection of 1-Onmol sodium melarsen, 1.0 nmol melarsen oxide, I.0 nmol Mel W, and 1.0 nmol Mel B. (C) Injection of the samesample as in (B) which had been incubated with 20 pL/mL Hz02 for 60 min at room temperature before analysis. The chromatographic conditions are ontlined in the Materials and Methods section. The labelled peaks and struttnres are: (1) sodium melarsen, (2) melarsen oxide, (3) Mel W, and (4) Mel B.

peak and were well resolved from each other. The large, broad shoulder from 7 to 22 min present in blank runs (Fig. 1, A) was not associatedwith any arsenical compound, and probably resulted from a contaminant in the water supply. In addition, there were several peaks detected before 3 min in all samples, including blanks (Fig. l), which were solvent front injection artifacts. Mel Cy was unstable, as 4 peaks were detected by ultraviolet absorption at 280 mm (data not shown). For this reason, Mel Cy was not used in any subsequent experiment. A small second peak was also detected in the Mel B sample (Fig. 1, B), which represented the separation of optical isomers, as trivalent arsenic has an approximately tetrahedral configuration (FAIRLAMB et aZ., 1989). Oxidation of an aliquot of this mixed sample with hydrogen peroxide converted all the arsenicals to sodium melarsen (Fig. 1, C). The final melarsen peak, which theoretically contained 4.0 nmol arsenical, was calculated from the peak area to contain 4.15 nmol (an error of 3.75%). Therefore, sample oxidation can quantitatively convert mixed samples to a single known compound for measurement. The arsenicals were detected down to less than 10 pmol per injection, and gave a linear response over a range of at least 10-5000 pmol per injection (Fig. 2). The inter-day variation with samplesof 1.0 nmoliinjection of melarsen oxide was 9.93% (n=5; relative standard deviation) and the intra-day variation with samplescontaining 0.1 nmol/injection of melarsen oxide was 4.75% (n=4; relative standard deviation). Studies were also performed on methods for extracting arsenicals from biological matrices. A preliminary experiment demonstrated that an octadecylsilane solid-phase extraction cartridge could extract Mel B and Mel W from

Fig. 2. Standard addition curves for the arsenical compounds. Mixtures of sodium melarsen, melarsen oxide, Mel W, and Mel B were made so that 10 FL injections contained 5000, 1000, 500, 100, 50, and 10 pmol of each drug. The chromatographic conditions are outlined in the Materials and Methods section. The peak area for each drug was plotted against the amount injected in a log-log plot. Sodium melarsen (0), melarsen oxide (0) Mel W (A), and Mel B (A). The coefficient of variation for the linear regression of each cnrve exceeds 0.993 for each drug.

a 20% propylene glycol solution with a recovery of >80% (data not shown). Therefore, these cartridges were used to recover melarsen, melarsen oxide, Mel W, and Mel B from serum which had been ‘spiked’ at a concentration of 0.1 nmol/mL to simulate concentrations which might be found in viva many days after dosing. The reported range for plasma concentration of Mel B is 2.5-9.0 nmol/mL 24 h after doses of 1.2-3.6 mgikgld (HAWKING, 1962; BURRI & BRUN, 1992). While the arsenicals could be extracted from serum (Fig. 3), the recoveries were variable-14.1% for sodium melarsen, 61.5% for melarsen oxide, 35.6% for Mel W, and 65.2% for Mel B. In addition, an unknown peak was detected, near the melarsen oxide peak, which was also found in octadecylsilane extracts of ‘blank’ serum. Therefore, this peak represented an unknown compound in calf serum and was not associatedwith the arsenical drugs. O.lZO,

o,,y , ‘IO , 60 , SO , 0 20 Time (min)

Fig. 3. Extraction and detection of arsenical drugs from a ‘spiked’ serum sample. Foetal calf serum was fortified with stock sodium melarsen, melarsen oxide, Mel W, and Mel B to a final concentration of 100pmol/mL and then extracted over an octadecylsilane extraction cartridge before HPLC analysis. Details of the extraction and chromatographic systems are contained in the Materials and Methods section. (1) Sodium melarsen, (2) melarsen oxide, (3) Mel W, (4) Mel B, and (x) an unknown compound also present in extracts of foetal calf serum only.

359 Medicine and Parasitology, 43,223-225.

Discussion

The HPLC assayoutlined here is very sensitive, estimating as little as 10 pmol of arsenical per injection, and is accurate and reproducible (variations of 9.93% and 4.75% for inter- and intra-day variation). In addition, the assay can be easily used to estimate the total arsenical content of complex samples by oxidation of compounds to sodium melarsen by incubation with hydrogen peroxide. Therefore, the total amount of known arsenical in a mixture or extract can be correlated with the total arsenical content after oxidation. This processmay be important, as the trivalent arsenical drugs are known to form conjugates with sulphydryl-containing compounds such as trypanothione, lipoic acid and proteins (JOHNSTONE, 1963; FAIRLAMB et al., 1989, 1992), which may render them unextractable, undetectable, or give them unknown chromatographic behaviour. Octadecylsilane solid-phase extraction cartridges have been shown in this study to be useful in the extraction of arsenical drugs from ‘spiked’ serum. However, the extraction efficiencv is low (14.1% to 65.2%) and at least one serum component is ah extracted. Small alterations in the HPLC eradient utilized should allow comolete resolution of th: arsenical drugs from serum peaks. Although we recognize that this aspect of the assayprocedure requires further investigation, we are reporting our existing extraction system so that others may pursue this line of research. Despite the potential problems associated with extracting the compounds from biological samples and the fact that the 80 min run time would probably preclude its use for routine clinical monitoring, this assayremains the most promising for elucidating the pharmacological and chemical properties of this class of compounds. We are currently using this method to examine the stability of arsenical drugs in solution, the uptake of arsenicals by trypanosomes, and the interaction of these drugs with possible target molecules.

Burri, C., Balm, T., Giroud, C., Doua, F., Welker, H. A. & Brun, R. (1993). Pharmacokinetic properties of the trypanocida1drug melarsoprol. Chemotherapy, 39,225-234. Cristau, B., Placidi, M. & Audibert, I’. (1972). Voies et cinetiques d’elimination de l’arsenic chez le rat apres administration de medicaments organoardnies. Medecine Tropical, 32,

Acknowledgements We acknowledge support from the IJNDPiWorld Bank/WHO Special Programme for Research and Training in Tropical Disease and the Wellcome Trust (A.H.F.) and the NATO ScienceFellowship Program (B. J.B.).

W&y, M. (1991). Therapy for African trypanosomiasis. Current Opinion in Infectious Disease,4,838-843. WHO (1990). African trypanosomiasis. In: Tropical Diseases. Progress in Research 1989-1990. Geneva: World Health Organization, pp. 59-68.

References Burri, C. & Brun, R. (1992). An in vitro bioassayfor quantification of melarsoprol in serum and cerebrospinal fluid. Tropical

367-771 --,

II_.

Cristau, B., Placidi, M. & Legait, J. P. (1975). Etude de l’excretion de l’arsenic chez le trypanosome train5 au melarsoprol (arsobal). Medecine Tropicale, 35,389-401. Fairlamb. A. H., Henderson, G. B. & Cerami. A. (1989). Trvpanothione is the primary’target for arsenical drugs against African trypanosomiasis. Proceedings of the National Academy of Sciences of the USA, 86,2607-2611.

Fairlamb, A. H., Smith, K. & Hunter, K. J. (1992). The interaction of arsenical drugs with dihvdrolinoamide and dihvdroliuoamide dehvdro&nase from arsenical resistant and sensiiive strains 0; Tryianosoma bracei brucei. Molecular and Biochemical Parasitoloev. 53,223-232.

Friedheim, E. A. H. (lq49). ‘Mel B in the treatment of human trypanosomiasis. American Journal of Tropical Medicine, 29, 173-180. Hawking, F. (1962). Estimation of the concentration of melarsopro1 (Mel B) and Mel W in biological fluids by bioassay with trypanosomes in vitro. Transactions of the Royal Society of Tropical Medicine and Hygiene, 56,354-363. Johnstone, R. M. (1963). Sulfhydryl agents: arsenicals. In: MetabolicZnhibitors, vol. 2, Hochster, R. M. & Quastel, J. H. (editors). New York: Academic Press, pp. 99-118. Maes, L., Doua, F. & Hamers, R. (1988a). ELISA assayfor melarsoprol. Bulletin de la Socibte de Pathologie Exotique, 81,

557-560.

Maes, L., Vanderveken? M:, Hamers, R., Doua, F. & Cattand, I’. (1988b). The momtormg of trypanocidal treatment with a sensitive ELBA method for measuring melarsoprol levels in serum and in cerebrospinal fluids. Annales de la Socibte Belge de Medecine Tropicale, 68,219-231.

I’epin, J. & Milord, F. (1991). African trypanosomiasis and drug-induced encephalopathy: risk factors and pathogenesis. Transactionsof the Royal Society of Tropical Medicine and Hygiene, 85222-224.

Received 19 August 1993; revised 27 September accepted for publication 29 September 1993

Announcement International

Symposium on Tropical

Alexandria,

Haematology

Egypt: 25-26 August 1994

This is a joint meeting of the Egyptian Branch of the Royal Society of Tropical Medicine and Hygiene, the Egyptian Society of Haematology and the World Health Organization Regional Office for the Eastern Mediterranean. Further information can be obtained from the conveners: Professor S. M. Kabil and Professor M. ElSawy, Department of Clinical Pathology, Faculty of Medicine, University of Alexandria, Alexandria, Egypt.

1993;