Contribution of Dithiol Ligands to In Vitro and In Vivo Trypanocidal Activities of Dithiaarsanes and Investigation of Ligand Exchange in an Aqueous Solution

Abstract

Twelve new dithiaarsanes were evaluated for their in vitro and in vivo trypanocidal properties in regard to their three parent molecules, 4-amino-phenylarsenoxide, melarsenoxide, and 4-dansylamino-phenylarsenoxide. The most potent dithiaarsane, compound 2b, had a minimum effective concentration of 1.5 nM after 48 h of incubation and at a dose of 0.39 μmol/kg of body weight (0.2 mg/kg) administered subcutaneously cured 100% of mice acutely infected with Trypanosoma brucei brucei CMP. With this model, the chemotherapeutic index of compound 2b was 512, compared to 256 for melarsamine dihydrochloride (Cymelarsan) under the same conditions. With a chronic infection produced by T. brucei brucei GVR, compound 2b cured 100% of mice after treatment at a dose of 25 μmol/kg (12.5 mg/kg) for 4 consecutive days, whereas melarsamine dihydrochloride and potassium melarsonyl (Trimelarsan) cured less than 50% mice at this dose. For both acute and late-stage infections, dithiaarsanes having a melaminophenyl ring exhibited the most-potent trypanocidal activity. Compound 2b is thus one of the most active organoarsenicals described in a mouse trypanosomiasis model. Considering that the main intracellular targets of organoarsenicals are thiol groups, we studied the possibility of ligand exchange between Cymelarsan and several dithiols. In aqueous solution, we observed a rapid exchange of cysteamine from melarsamine with free cysteamine and also with various dithiols always in favor of more stable cyclic derivatives. These ligand exchanges suggest the ability of trivalent organoarsenicals to react with targets such as trypanothione and dihydrolipoic acid. Among several ligands, a 1,3-dimercaptopropane moiety appeared the most suitable for trypanocidal activity.

Treatment of the cerebral stage of African trypanosomiasis has relied primarily on the melaminophenylarsine melarsoprol (5), but melarsoprol is insoluble in water and must be administered intravenously as a propylene glycol solution (1) with consequent side effects, particularly, an arsenical agent-induced encephalopathy (6, 12). Furthermore, the solvent is a powerful irritant that often causes thrombophlebitis (1). Water-soluble organoarsenicals such as potassium melarsonyl showed little improvement over melarsoprol (1), but melarsamine dihydrochloride (Cymelarsan) has been licensed for use against trypanosomiasis in animals because of its activity on various models of Trypanosoma brucei brucei, T. evansi, and T. equiperdum in camels, buffalo, goats, and pigs and in vitro (10, 11, 16–18). Immediately after dissolving in water, melarsamine dihydrochloride exists as an equilibrium mixture containing melarsamine (43%), melarsamine having lost one cysteamine moiety (24%), melarsen oxide (33%), and free cysteamine (2). Small amounts (<1%) of the oxidation products derived from the last two components were also formed (cystamine and sodium melarsen). The equation in Fig. 1 summarizes these data.

FIG. 1.

Equilibrium between melarsamine and melarsen oxide in water.

The purpose of this study was firstly to appreciate the contribution of various dithiol ligands for in vitro and in vivo trypanocidal activities of dithiaarsanes and secondly to investigate the possibility that ligand exchange in aqueous solution explained the relative affinity and competition of these ligands toward the arsenic moiety. Among these ligands, dihydrolipoic acid is thought to be a biological target of organoarsenicals in African trypanosomes (4), so that it was of interest to evaluate the effect of lipoic acid on in vitro and in vivo trypanocidal activity of a dithiaarsane.

MATERIALS AND METHODS

Chemicals.

Melarsamine dihydrochloride and potassium melarsonyl were provided by Specia (Paris, France). Cysteamine, glucose, sodium hydrogenocarbonate, thymidine, hypoxanthine, 2-mercaptoethanol, carboxymethylcellulose, and sodium pyruvate were acquired from the Sigma Chemical Co. (Saint Quentin Fallavier, France). Gentamicin was from Dakota Pharm-Sanofi (Gentilly, France). l-Glutamine and HEPES were from Gibco BRL (Cergy-Pontoise, France).

Dithiaarsane synthesis.

Twelve new dithiaarsanes were synthesized as previously described (9) via condensation of the appropriate dithiols with 4-amino-phenylarsenoxide (compound 1), melarsenoxide (compound 2) and 4-dansylamino-phenylarsenoxide (compound 3) after reduction of the corresponding arsonic acids. The dithiol ligands were 1,2-dimercaptoethane (suffix a), 1,3-dimercaptopropane (suffix b), 2,2′-dimercaptodiethylether (suffix c), 2,3-dimercaptobutane (suffix d), dihydrolipoic acid (suffix e), and 2,3-dimercaptopropanol or British antilewisite (BAL) (suffix f). Thus, for example, compounds 1a, 2a, and 3a involved the a substitution of compounds 1, 2, and 3; compounds 1b, 2b, and 3b involved the b substitution of compounds 1, 2, and 3.

The chemical structure of the compounds are shown in Fig. 2.

FIG. 2.

Structures of the dithiaarsanes and their parent arsenoxides.

Trypanosome strains.

T. brucei brucei CMP (Châtenay-Malabry Parasitologie) was obtained from the Institut Pasteur, Paris, in 1973. This strain was kept frozen in liquid nitrogen, and an aliquot was passaged in female CD1 mice before the experiments.

T. brucei brucei GVR 35/cl.2 was kept frozen in liquid nitrogen and was kindly supplied by F. W. Jennings (Glasgow, United Kingdom). This strain had been derived from trypanosomes isolated from a wild beast in the Serengeti (Tanzania) in 1966 (Serengeti/66/SVRP/10). It produces a chronic infection in mice, allowing them to survive for at least 30 days.

In vitro evaluation. (i) DIIT.

The drug incubation infectivity test (DIIT) was used for compound evaluation (8). The culture medium used throughout the study consisted of minimum essential medium (catalog No. 32360; Gibco BRL) including 25 mM HEPES and Earle's salts (per liter: 264 mg of CaCl₂ · 2H₂O, 400 mg of KCl, 200 mg of MgSO₄ · 7H₂O, 6.3 g of NaCl, 2.2 g of NaHCO₃; 158 mg of NaH₂PO₄ · 2H₂O) and supplemented with 2 mM l-glutamine, 1 g of additional glucose per liter, 10 ml of minimal essential medium nonessential amino acids (100×; Gibco BRL) per liter, 0.2 mM 2-mercaptoethanol, 2 mM sodium pyruvate, 0.1 mM hypoxanthine, 0.016 mM thymidine, 15% heat-inactivated horse serum (Gibco BRL), and 50 μg of gentamicin/ml. The culture medium was sterilized by filtration. Blood was collected aseptically from the retro-orbital sinus of infected mice having about 10⁸ trypanosomes/ml of blood, diluted with medium, and kept on ice. The trypanosomes were isolated from this suspension by centrifugation at 900 × g for 10 min at 4°C; the supernatant was discarded and the trypanosomes were at the top of the pellet. The number of trypanosomes was determined by hemocytometer counting and adjusted to 2 × 10⁵ parasites/ml. The bloodstream forms of T. brucei brucei CMP were maintained in vitro without loss of infectivity for 48 h in the dark at 37°C in a 5% CO₂ atmosphere. Screening was performed in 96-well tissue culture plates in a final volume of 200 μl containing 2 × 10⁴ parasites and the compounds to be tested (previously diluted in water-dimethyl sulfoxide). Drug concentrations were evaluated in triplicate. The minimum effective concentration (MEC) was defined as the minimum concentration at which no viable parasite was observed microscopically and with which naive mice injected intraperitoneally (i.p.) with 150 μl of treated trypanosome suspension withdrawn from the wells after a 48-h period were aparasitemic 30 days postinfection. Thus, the MEC was assessed both visually using an optical microscope after 1, 24, and 48 h of incubation and in vivo since the mice were checked for parasitemia weekly for 30 days.

(ii) Lysis assay.

The lysis assay method used was described by Yarlett et al. (15). Briefly, trypomastigotes were suspended at 5 × 10⁷ per ml in heat-inactivated fetal calf serum, pH 7.4 (Gibco BRL), containing 50 μg of gentamicin per ml. Suspensions were held at 4°C, and aliquots were warmed to 37°C prior to assay. Trypomastigote suspension was added to six cuvettes containing aliquots of the arsenical in a final volume of 1.5 ml, and the absorbance was monitored at 500 nm for 30 min (at 1-min intervals) using a thermostatically controlled recording spectrophotometer (Shimadzu UV-120-02; Roucaire, Paris, France). Calf serum was used as the zero absorbance reference. In cuvettes, the absorbance of trypanosome suspension should be in a range from 0.700 to 0.800, corresponding to a suspension of about 15 × 10⁶ parasites/ml. The time of 50% lysis was determined for each compound. A lysis constant (L₅₀) was determined by plotting the reciprocal of the time of lysis versus the drug concentration and extrapolating the slope to intercept the abscissa. Lysis curves always included appropriate controls containing no drug.

Animals.

Female CD1 mice (18 to 20 g) were purchased from Charles River Ltd. (Cléon, France).

In vivo evaluation. (i) Acute infections with T. brucei brucei CMP.

Mice were infected i.p. with 10⁴ bloodstream trypomastigotes taken from an infected mouse and suspended in 0.1 ml of phosphate-buffered saline, pH 7.2. The infection was allowed to develop for 24 h before treatment was begun. Ten infected mice were used as controls and received only excipient, 1% carboxymethylcellulose by the i.p. route in a 0.1 ml volume. The other mice received a single dose of the diluted or suspended compounds in the same manner. Six mice were used per dose. The trypanocidal activity was evaluated by the mean survival time of treated mice for each dose. Treatment was considered to be successful when the mean survival time exceeded 30 days and the mice remained aparasitemic. Control mice (infected and untreated) do not survive more than 4 days postinfection. Cure rates were calculated and are expressed as percentages.

(ii) Late-stage infections with T. brucei brucei GVR 35.

Mice were similarly inoculated as above but with 1.0 × 10³ trypanosomes per mouse, and treatment was begun 21 days postinfection by the subcutaneous route. Animals were checked daily for deaths, and tail vein blood smears were examined weekly for parasites. Animals were considered cured if they survived and were aparasitemic for at least 16 weeks after the end of treatment. Cure rates were calculated and are expressed as percentages.

Ligand exchange study.

Initially, ligand exchange was measured in different aqueous mixtures of melarsamine dihydrochloride and cysteamine hydrochloride using ¹³C nuclear magnetic resonance (NMR). Spectra were obtained using a spectrometer (Spectrospin; Bruker, Wissembourg, France) operating at 75.5 MHz for ¹³C. The other dithiols were then studied for their ability to exchange with cysteamine groups of melarsamine.

RESULTS

In vitro activity.

The DIIT evaluates the ability of trypanosomes to be infective after a 48-h in vitro treatment, whereas the lysis assay provides a measure of rapid destruction of parasites provoked by the compounds (Table 1). Except for compounds 1, 3, 3a, 3b, and 3f, the compounds exhibited better in vitro activity in the DIIT than potassium melarsonyl (MEC, 0.012 μM). Although eight compounds exhibited the same MEC as melarsamine dihydrochloride (MEC, 0.003 μM), none was more active than this reference compound.

TABLE 1.

DIIT using T. brucei brucei CMP and L₅₀ determination

Compoundd	MECa (μM) at:			L50b (μM) ± SD
	1 h	24 h	48 h
1	25	0.4	0.2	NDc
1a	50	0.006	0.003	ND
1b	25	0.003	0.003	ND
1c	50	0.003	0.003	ND
1d	50	0.003	0.003	ND
1e	25	0.003	0.003	ND
2	0.1	0.006	0.003	15.5 ± 2.1
2a	0.05	0.003	0.0015	11.2 ± 1.4
2b	0.025	0.003	0.0015	10.7 ± 1.8
2c	0.05	0.003	0.003	12.1 ± 2.2
2e	0.05	0.003	0.003	13.4 ± 2.5
3	50	0.050	0.012	ND
3a	50	0.050	0.012	ND
3b	50	0.050	0.025	ND
3f	>100	0.050	0.050	ND
Melarsamine dihydrochloride	0.05	0.006	0.003	16.2 ± 1.9
Potassium melarsonyl	10	0.1	0.012	45.6 ± 3.7

^aThe MEC. was defined as the minimum concentration at which no viable parasite was observed microscopically and with which naive mice injected i.p. with 150 μl of treated trypanosome suspension withdrawn from the wells after a 1-, 24-, or 48-h period were aparasitemic 30 days, postinfection. 

^bL₅₀ determination: the time of 50% lysis was determined for each compound. A lysis constant (L₅₀) was determined by plotting the reciprocal of the time of lysis versus the drug concentration and extrapolating the slope to intercept the abscissa. 

^cND, not determined. 

^dEach compound was initially evaluated at 100 μM and further tests were at progressive two-fold dilutions. 

Structure-activity relationships are clear for arsenoxides: melarsenoxide (compound 2) was more active than 4-aminophenylarsenoxide (compound 1), which was more potent than 4-dansylaminophenylarsenoxide (compound 3).

The presence of dithiol ligands greatly enhanced trypanocidal activity, since dithiaarsanes from 1a to 1e were 100 times more active than 4-aminophenylarsenoxide (compound 1).

The melarsen and phenylarsen derivatives (MEC in a range from 0.0015 to 0.003 μM) were significantly more active than the 4-dansylaminophenylarsen derivatives.

Considering the rapid activity of the melarsen derivatives, only these compounds were studied in the lysis assay. This series gave homogeneous results in regard to the L₅₀ in a range from 10.7 to 15.5 μM, and the molecule having the best ability to lyse trypanosomes within 30 min was compound 2b.

Concerning melaminyl derivatives, the presence of a dithiol on melarsenoxide enhanced the trypanocidal activity by only two-fold. The lysis assay is a complementary method to the determination of drug sensitivity and structure-activity relationships since the L₅₀ determination is more precise than the MEC determination. Thus, the lysis assay allowed us to conclude that compound 2b was the most potent trypanocidal compound of this series.

In vivo activity.

For acute infections (Tables 2 to 4), melarsen derivatives exhibited the most-potent trypanocidal activity (Table 3). The most-efficient compound of this series was compound 2b, which cured 100% of mice with a single dose of 0.39 μmol/kg administered subcutaneously.

TABLE 2.

Treatment of T. brucei brucei CMP acute infections with dithiaarsanes derived from 4-amino-phenylarsenoxide

Compound	Dosea (μmol/kg)	Mean survival time (day) ± SD	Parasitemiab on day:		No. of mice cured/ total no. of mice	Cure rate (%)	Toxicityc
			8	30
1	400	0.5 ± 0.2	NDd	ND	0/6	0	++
	200	0.8 ± 0.3	ND	ND	0/6	0	++
	100–12.5	>30	0	0	6/6	100	0
	6.25	12.4 ± 3.3	+	ND	2/6	33	0
	3.12	3.6 ± 0.7	ND	ND	0/6	0	0
1a	400	0.8 ± 0.2	ND	ND	0/6	0	++
	200	2.6 ± 0.3	ND	ND	0/6	0	++
	100–12.5	>30	0	0	6/6	100	0
	6.25	11.3 ± 4.5	+	ND	1/6	16	0
	3.12	4.3 ± 2.1	ND	ND	0/6	0	0
1b	400	>30	0	0	6/6	100	±
	200–100	>30	0	0	6/6	100	0
	50	11.5 ± 4.7	+	ND	2/6	33	0
	25	7.4 ± 2.3	+	ND	0/6	0	0
1c	400	>30	0	0	6/6	100	0
	200	17.5 ± 3.9	0	ND	0/6	0	0
	100	5.3 ± 2.1	+	ND	0/6	0	0
1d	400–100	>30	0	0	6/6	100	0
	50	7.5 ± 2.2	+	ND	0/6	0	0
	25	3.5 ± 0.5	ND	ND	0/6	0	0
1e	400	0.8 ± 0.2	ND	ND	0/6	0	++
	200	2.8 ± 0.5	ND	ND	0/6	0	++
	100	3.5 ± 0.4	ND	ND	0/6	0	0

^aTreament administered subcutaneously in one single dose. 

^b+, presence of trypanosomes; 0, absence of trypanosomes. 

^c++, death of all the mice within 3 days after treament; +, death of some mice within 3 days after treament; ±, some apparent signs of acute toxicity without any death; 0, no apparent acute toxicity. 

^dND, not determined 

TABLE 4.

Treatment of T. brucei brucei CMP acute infections with dithiaarsanes derived from 4-dansylamino-phenylarsenoxide and reference drugs

Compound	Dosea (μmol/kg)	Mean survival time (day) ± SD	Parasitemiab on day:		No. of mice cured/ total no. of mice	Cure rate (%)	Toxicityc
			8	30
3	400	0.8 ± 0.2	NDd	ND	0/6	0	++
	200	2.6 ± 0.5	ND	ND	0/6	0	++
	100	3.7 ± 0.4	ND	ND	0/6	0	0
	50	3.5 ± 0.9	ND	ND	0/6	0	0
3a	400–200	3.5 ± 0.5	ND	ND	0/6	0	0
3b	400–200	3.5 ± 0.5	ND	ND	0/6	0	0
3f	400–200	3.5 ± 0.6	ND	ND	0/6	0	0
Melarsamine dihydrochloride	400	1.3 ± 0.2	ND	ND	0/6	0	++
	200–0.78	>30	0	0	6/6	100	0
	0.39	12.6 ± 4.6	0/+	ND	3/6	50	0
	0.19	5.2 ± 2.1	ND	ND	0/6	0	0
Potassium melarsonyl	400	0.5 ± 0.2	ND	ND	0/6	0	0
	200–12.5	>30	0	0	6/6	100	0
	6.25	8.4 ± 2.5	+	0	1/6	16	0
	3.12	3.6 ± 0.7	ND	ND	0/6	0	0
Excipient only		3.5 ± 0.8	ND	ND	0/10	0	0

^aTreatment administered subcutaneously in one single dose. 

^b+, presence of trypanosomes; 0, absence of trypanosomes; 0/+, presence of trypanosomes in three of six mice. 

^c++, death of all the mice within 3 days after treatment; +, death of some mice within 3 days after treatment; ±, some apparent signs of acute toxicity without any death; 0, no apparent acute toxicity. 

^dND, not determined. 

TABLE 3.

Treatment of T. brucei brucei CMP acute infections with dithiaarsanes derived from melarsenoxide

Compound	Dosea (μmol/kg)	Mean survival time (day) ± SD	Parasitemiab on day:		No. of mice cured/total no. of mice	Cure rate (%)	Toxicityc
			8	30
2	400	1.2 ± 0.8	NDd	ND	0 /6	0	++
	200	>30	0	0	4 /6	100	+
	100	>30	0	0	6 /6	100	±
	50–1.56	>30	0	0	6 /6	100	0
	0.78	9.8 ± 3.5	+	ND	3 /6	33	0
	0.39	3.5 ± 0.8	ND	ND	0 /6	0	0
2a	400	2.4 ± 0.7	ND	ND	0 /6	0	++
	200	>30	0	0	2 /6	100	±
	100–0.78	>30	0	0	6 /6	100	0
	0.39	12.5 ± 5.8	+	ND	0 /6	0	0
	0.19	3.4 ± 0.8	ND	ND	0 /6	0	0
2b	400	3.2 ± 1.9	ND	ND	0 /6	0	++
	200–0.39	>30	0	0	6 /6	100	0
	0.19	5.6 ± 1.0	ND	ND	0 /6	0	0
2c	400–12.5	>30	0	0	6 /6	100	0
	6.25	11.4 ± 3.1	+	0	0 /6	0	0
	3.12	6.2 ± 1.4	+	0	0 /6	0	0
2e	400–1.56	>30	0	0	6 /6	100	0
	0.78	8.2 ± 3.1	+	0	0 /6	0	0
	0.39	5.4 ± 2.1	ND	0	0 /6	0	0

^aTreament administered subcutaneously in one single dose. 

^b+, presence of trypanosomes; 0, absence of trypanosomes. 

^c++, death of all the mice within 3 days after treament; +, death of some mice within 3 days after treament; ±, some apparent signs of acute toxicity without any death; 0, no apparent acute toxicity. 

^dND, not determined. 

Concerning the parent arsenoxides, melarsenoxide (compound 2) was the most-active compound, with a minimum curative dose at 1.56 μmol/kg, whereas that of 4-aminophenyl-arsenoxide (compound 1) was 12.5 μmol/kg and 4-dansylaminoyl-arsenoxide (compound 3) was completely inactive (Tables 2 to 5).

TABLE 5.

Chemotherapeutic index of dithiaarsanes in the acute-infection model

Compound	MTDa (μmol/kg)	MCDb (μmol/kg)	CIc
1	100	12.5	8
1a	100	12.5	8
1b	200	100	2
1c	>400	400	>1
1d	>400	100	>4
1e	100	>100	<1
2	100	1.56	64
2a	100	0.78	128
2b	200	0.39	512
2c	>400	12.5	>32
2e	>400	1.56	>256
3	100	>100	<1
3a	>400	>400	NDd
3b	>400	>400	ND
3f	>400	>400	ND
Melarsamine dihydrochloride	200	0.78	256
Potassium melarsonyl	>400	12.5	>32

^aMTD, maximal tolerated dose. 

^bMCD, minimal curative dose. 

^cCI, chemotherapeutic index (calculated as MTD/MCD). 

^dND, not determined. 

The four most-potent compounds in the acute infection model (i.e., compounds 2, 2a, 2b, and 2e) were further evaluated in a chronic-infection model (Table 6). Compound 2b was the most active since 100% of mice were cured at a dose of 25 μmol/kg for 4 consecutive days whereas melarsamine dihydrochloride (Cymelarsan) and potassium melarsonyl (Trimelarsan) cured less than 50% mice at this dose. Thus, melarsamine dihydrochloride cured all the mice at 50 μmol/kg and potassium melarsonyl at 60 μmol/kg.

TABLE 6.

Treatment of T. brucei brucei GVR chronic infections with dithiaarsanes and reference drugs^a

Compound	Daily doseb (μmol/kg)	No. of mice cured/ total no. of mice	Cure rate (%)	Toxicityc
2	50	6 /6	100	0
	25–12.5	0 /6	0	0
2a	50	6 /6	100	0
	25	3 /6	50	0
	12.5	0 /6	0	0
2b	50–25	6 /6	100	0
	12.5	0 /6	0	0
2e	50	6 /6	100	0
	25	1 /6	17	0
	12.5	0 /6	0	0
Melarsamine dihydrochloride	50	6 /6	100	0
	25	2 /6	33	0
	12.5	0 /6	0	0
Potassium melarsonyl	60	6 /6	100	0
	40	3 /6	50	0
	20	0 /6	0	0
Excipient only		0 /10	0	0

^aTreatment was performed on 4 consecutive days beginning at 21 days postinfection. 

^bEach dose was administered on four consecutive days. 

^c0, no apparent acute toxicity. 

No compound was active after treatment with a single dose of 50 μmol/kg.

In summary, the range of activity of arsenoxides is conserved in the acute-infection model. As was the case for compounds 2, 2a, 2b, 2c, and 2e, dithiaarsanes having a melaminophenyl ring exhibited the most-potent trypanocidal activities for both acute and late-stage infections.

Toxicity.

No compound was responsible for the death of mice at single doses of less than 200 μmol/kg. Chemotherapeutic index data are gathered in Table 5. On the acute-infection model, the chemotherapeutic index of compound 2b was 512, whereas that of melarsamine dihydrochloride (Cymelarsan) was 256 under the same conditions.

Ligand exchange. (i) Cysteamine exchange in an aqueous solution of melarsamine.

The synthesis of melarsamine from melarsenoxide and cysteamine may lead to products containing small excess amounts of cysteamine. All attempts to measure directly the amount of this impurity failed since usual methods (high-performance liquid chromatography, UV, etc.) necessitate a highly diluted solution, and at these concentrations only the starting materials were observed. In solution melarsamine is in equilibrium with its starting products as shown by the reaction in Fig. 3.

FIG. 3.

Equilibrium between melarsamine and its starting products in aqueous solution.

Study by ¹³C NMR of an aqueous solution containing both melarsamine dihydrochloride and cysteamine chloride did not exhibit separated peaks of linked and free cysteamine but a single C-N peak and a single C-S peak having intermediate chemical shifts resulting from the exchange depicted in Fig. 4.

FIG. 4.

Ligand exchange between melarsamine and free cysteamine. The star respresents the cysteine moiety which has been exchanged.

Thus, the carbon atoms of the SCCN moiety had ¹³C NMR characteristics which were highly sensitive to the excess of cysteamine. The most reliable value was found by determining the shift difference between the CS and the CN carbons, Δδ = δ_CN − δ_CS, which corresponds to an internal probe of this equilibrium. Experimental determination of this value in the range 0 to 25% cysteamine excess and curve fitting led to the quantification of the cysteamine impurity in the synthesis using a reliable assay (Fig. 5). Thus, the concentration of cysteamine in excess was obtained from the curve by measuring the Δδ value by ¹³C NMR.

FIG. 5.

¹³C NMR data of an aqueous solution of melarsamine chloride and cysteamine chloride mixture (Δδ = δ_CN − δ_CS) as a function of percentage of cysteamine in excess in melarsamine chloride solution.

(ii) Ligand exchange with other thiols in an aqueous solution of melarsamine.

In addition we were interested in the fact that the exchange of ligand in aqueous solution could occur with other thio moieties. Thus, we tried to obtain classical organoarsenicals such as melarsoprol and melarsonyl by starting with melarsamine and dithiols such as BAL and 2,3-dimercapto-succinic acid (Fig. 6).

FIG. 6.

Ligand exchange between melarsamine and other ligands.

In all cases, with heating in solution, we observed the quantitative formation of the cyclic dithio derivative. Furthermore, whatever other dithiols were used—1,2-dimercaptoethane (a), 1,3-dimercaptopropane (b), 2,2′-dimercaptodiethylether (c), 2,3-dimercaptobutane (d), dihydrolipoic acid (e), and 2,3-dimercaptopropanol or BAL (f)—we always observed by ¹³C NMR a ligand exchange in favor of cyclic compounds, and this exchange was complete.

DISCUSSION

Despite their toxicity, organoarsenicals remain of high interest in the treatment of late-stage African trypanosomiasis (12). Considering that arsenoxide is the active form of trivalent arsenicals, we suggest that the dithiol moiety should be important for drug access to the target, for example, bypassing the membranes. Our contribution relies on the pharmacomodulations of dithiols able to facilitate either trivalent organoarsenical uptake or reactivity on biochemical targets in African trypanosomes. We synthesized and examined a series of dithiaarsanes as potential chemotherapeutic agents for the treatment of infections caused by T. brucei brucei and, by extension, other African trypanosomes.

The best activity of melaminophenyl arsenicals relative to the others (4-aminophenyl- and 4-dansylaminophenyl-arsenicals) could be ascribed to the fact that a specific adenosine transporter recognizes the melaminophenyl moiety of these molecules (13). However, the dithiol ligand is very important, since it was described that melarsoprol can enter trypanosomes by a route other than through an adenosine transporter (13). Moreover, 2,3-dimercaptopropinol was pointed out as playing an important role in the absorption of melarsoprol through the skin and/or blood-brain barrier into the central nervous system and/or into the trypanosome (7). The fact that compound 2b is slightly more active than melarsenoxide, the active principle, could indicate that the passage of various pharmacokinetics barriers is enhanced by using the 1,3-dimercaptopropane ligand. Contrary to melarseno-arsenicals, 1,3-dimercaptopropane is not the most-relevant ligand to enhance trypanocidal activity of 4-aminophenyl-arsenicals. Such a discrepancy certainly involves the adenosine transporter, which is able to concentrate melarseno-arsenicals within the parasite, whereas 4-aminophenyl-arsenicals are not recognized by this transporter and have an unknown mechanism of uptake.

The trypanocidal activities of dithiaarsanes are probably achieved by a specific interaction with the thiol metabolism of the parasites. It was shown previously that several trypanosomatids possess a defense pathway against oxidative damage, with unique features compared to the mammalian host (3). The key molecule in this system is the dithiol N¹,N⁸-bis(glutathionyl)spermidine called trypanothione. Trypanothione is transformed to trypanothione disulfide by a peroxidase system, and trypanothione disulfide is reduced back to trypanothione by the corresponding trypanothione reductase. Trypanothione is reactive in its reduced dithiol form towards some arsenical compounds. Thus, trivalent arsenicals such as melarsenoxide are able to form adducts (3); the addition product itself acts as an inhibitor of trypanothione reductase. Moreover, arsenicals have been described to also interfere with lipoic acid metabolism in the trypanosome, by the formation of stable addition products (4). Thus, d,l-dihydrolipoamide and d,l-dihydrolipoic acid react to form stable complexes with melarsenoxide (4). The ligand linked to the arsenic atom is of high importance in the reactivity toward the previous targets. In this study, we have demonstrated the potential of cysteamine to exchange with itself and also with dithiols, when bound to an arsenical moiety, the reaction being always in favor of cyclic derivatives. These data explain the high reactivity of the most relevant arsenical drug presently known, melarsamine, towards targets described above. A previous study had demonstrated that arsenoxides easily bound to the Escherichia coli puruvate dehydrogenase complex with inhibition of the enzyme (14). The addition of 2,3-dithiopropanol reactivated the complex, showing a ligand exchange between the arsenoxide bound to the lipoic acid of the enzyme and the added dithiopropanol and the striking preference in stability between five- and six-membered rings in favor of the former. The present study shows that noncyclic dithiaarsanes rapidly exchange cysteamine ligand with free thiols to give more-stable cyclic dithiaarsanes. Moreover, we observed by ¹³C NMR the presence of the cis-isomer for six-membered rings, which is less stable than the trans-isomer observed for five-membered rings. Thus, the stability in the exchange of the thioarsenical derivatives is in the range of noncyclic derivatives to six-membered rings to five-membered rings (least to greatest).

Nevertheless, concerning their biological activity, other factors should be taken into account, mainly the lipophilicity of the product and the size of the enzymatic site. Thus, compounds 1e and 2e, whose the ligand was dihydrolipoic acid, were active in vitro but inactive in vivo. The in vitro activities of these compounds suggest the ability of ligand exchange within the parasite between the dihydrolipoic acid moiety of these two dithiaarsanes and the biological targets (another dihydrolipoic acid moiety or trypanothione).

Further study should be focused on compound 2b, which appears to be a promising compound and must be evaluated on other trypanosomiasis animal models.