Effects of Synthetic Siderophores on Proliferation of Plasmodium falciparum in Infected Human Erythrocytes

Abstract

Because iron is essential for Plasmodium falciparum, we investigated the in vitro potential of various synthetic siderophores to kill P. falciparum in infected human erythrocytes. The substances with the most promising profile, i.e., low 50% lethal doses for plasmodia and minimum toxicity towards mammalian cells, were siderophores with an acylated monocatecholate or a triscatecholate as substituent.

The spread of drug resistance among Plasmodium falciparum strains is a global health concern. Many antimalarial drugs affect the release of iron from hemoglobin by the parasite. The importance of iron for the clinical course of malaria has been shown in several studies (5, 9). Since the malaria parasite is dependent on a sufficient supply of iron, compounds that induce iron deprivation are potentially useful antimalarial agents (10, 11). Chelators suppress parasite growth by affecting iron availability within the parasite (8, 14), presumably by interfering with hemoglobin breakdown (3).

This resulted in placebo-controlled studies investigating the effect of the iron chelator desferrioxamine B (DFO B) on the clinical course of P. falciparum infection (4). Although these studies showed some clinical benefit, the overall mortality was not reduced (16). However, these studies suggested that iron deprivation is a useful therapeutic approach for malaria. Moreover, the hydrophilic DFO B displays poor permeation across membranes (8) and thus is relatively slow in eliciting biological effects on intracellular parasites (17).

Since the degree of antimalarial activity seems to correlate with the degree of lipophilicity, i.e., the ability of the compounds to cross over cell membranes (1, 7, 15, 19), the identification of new iron-chelating substances with improved pharmacokinetics may increase the antimalarial efficacy of the drugs. This strategy is exemplified by a class of synthetic hydrophobic siderophores which feature an iron(III)-binding cavity mimicking that of ferrichrome or by N-terminal derivatives of DFO, which fully retain the iron-binding capacity of the parent compound.

We examined artificial synthetic siderophores as monocatecholates (dihydroxybenzoates) containing aminobenzoic acid (Fig. 1, compound 7), amino acids (compounds 8 to 10), and hydrazones of glyoxylic acid (compounds 1 and 2), phenylglyoxylic acid (compounds 3 and 6), or formylbenzoic acids (compounds 4 and 5) as scaffold or as triscatecholates based on triamines (compounds 11 to 16) and tris-aminopropyl-methyl-α-d-glucopyranoside (compounds 17 and 18), possessing one or three catecholate moieties, respectively, as chelating groups. The catecholate moieties were used as free catecholates in the 2,3-OH derivatives (compounds 1, 3, 4, 5, 11, 12, and 13) and the 3,4-OH derivative (compound 2), as acylated catecholates in the 2,3-OCOCH₃ derivatives (compounds 7 and 17) and the 3,4-OCOOCH₃ derivative (compound 6), or in heterocyclic form as 2,4-dioxo-1,3-benzoxazine derivatives (compounds 8, 9, 14, 15, 16, and 18). The acylated catecholates possess higher lipophilic properties. For iron chelation these compounds have to be cleaved by acylases to the free catecholates. Iron complexes need three catecholate groups, corresponding to one molecule of triscatecholate or three molecules of monocatecholates. The triscatecholate is a trimer of N-2,3-dihydroxybenzoyl l-serine with an unstable trilactone backbone. The structures of these siderophores are given in Fig. 1.

FIG. 1.

Structures of the synthetic siderophores 1 to 18 used in the study. Ac = COCH₃; Moc = COOCH₃.

The following compounds were synthesized according to the literature: compounds 4, 7, 6, 8, and 9 (18); compounds 1 and 2 (12); compounds 11, 12, 13, 15, and 16 (13); and compounds 17 and 18 (6).

As standards for comparison, we used the iron chelators DFO (Sigma, Munich, Germany) and dexrazoxane (ICRF-187).

The culture of P. falciparum was carried out exactly as described elsewhere (2).

The modulation of parasite growth by the different siderophores was studied by adding increasing concentrations of the respective drugs to human erythrocytes infected with P. falciparum. The relative parasitemia was evaluated by preparing blood smears after an incubation period of 24 h with the respective drugs, and a 50% lethal dose (LD₅₀) was calculated for each substance investigated by comparing these results with the counts obtained for untreated control samples. None of the solvents used for the siderophores, i.e., dimethyl sulfoxide or dimethylformamide, altered parasitemia compared to that for untreated controls (data not shown).

To assess the toxicity towards mammalian cells, we studied the effects of the siderophores on viability of K562 cells. K562 cells were incubated with increasing concentrations of the respective siderophores for 24 h. Then, MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide; 5 mg/ml; Sigma) was added, and after another 4 h, 200 μl of dimethyl sulfoxide was used to solubilize the colored formazan product, which was then quantified with an enzyme-linked immunosorbent assay reader. An MTT value of 100% indicates that the respective siderophore did not reduce cell viability compared to that for untreated control cells, whereas an MTT value of 10% indicates that 90% of cells died compared to the control.

All substances tested exerted toxic effects towards P. falciparum (Table 1). LD₅₀s greatly varied and were not obviously different between monocatechol- and triscatechol-substituted siderophores. However, the antiplasmodial activity of the respective parent catecholates was greatly affected by the different derivatives. The monocatecholates based on aminobenzoic acid (compound 7) or hydrazones of glyoxylic acid, phenylglyoxylic acid, or formylbenzoic acid (compounds 1, 3, 5, and 6) with 2,3-OH or 2,3-O-acylated and benzoxazine substituents showed good antiplasmodial activity (LD₅₀ ≤ 10 μM), except for compound 4. The monocatecholate with a 3,4-OH substituent (compound 2) was less toxic. The monocatecholate compounds 8 and 9—based on amino acids—were also highly active. The triscatecholates based on triamines exhibited good antiparasitic activity, especially compound 11, with 2,3-OH substituents and a shorter backbone triaminopropylamine, while the corresponding compound 13, with a longer triamine backbone, was less active. The most active compounds were the triscatecholates based on glucopyranoside, which were the most lipophilic drugs (compound 17). Overall, the antiplasmodial activities of these compounds were consistently at least equal to that of DFO and, in one case (compound 17), >1,000-fold greater.

TABLE 1.

Siderophores and iron chelators investigated^a

Siderophore no.	Catecholate type	Side group(s)	LD50 for parasite growth (μM)	Estimated toxicity towards K562 cells at LD50 (% of control) by MTT assay	Reversibility of antiplasmodial effects of siderophores by iron salts
1	Mono	2,3-OH	∼6	94	Yes
2	Mono	3,4-OH	∼20	96	Partly
3	Mono	2,3-OH	∼5	94	Partly
4	Mono	2,3-OH	∼30	90	No
5	Mono	2,3-OH	∼6	94	No
6	Mono	3,4-O-acyl	0.9	105	No
7	Mono	2,3-O-acyl	10	102	No
8	Mono	Benzoxacine	10	92	No
9	Mono	Benzoxazine	∼9	77	Partly
10	Mono	3,4-O-acyl	∼100	70	No
11	Tris	2,3-OH	1	85	Yes
12	Tris	2,3-OH	∼7	85	Yes
13	Tris	2,3-OH	∼40	85	Yes
14	Tris	Benzoxazine	∼5	85	Yes
15	Tris	Benzoxazine	∼6	64	Partly
16	Tris	Benzoxazine	∼7	64	Yes
17	Tris	2,3-O-acyl	0.01	89	No
18	Tris	Benzoxazine	∼2	86	No
Dexrazoxane			∼400	100	Yes
DFO			∼30	93	Yes

^aThe table shows types of structures, basic pharmacological descriptions, antiplasmodial activities, toxicities toward mammalian cells, and reversibility of antiplasmodial effects by exogenously added iron salts.

In parallel, we tested the siderophores for potential toxicity towards mammalian cells by means of the MTT assay. The substances with the most promising profile, i.e., low LD₅₀s for plasmodia and hardly any toxicity towards mammalian cells at these concentrations, were mono- and triscatecholates with O-acyl and 2,3-OH substitutions (compounds 1, 3, 6, 7, 11, and 17 [Table 1]).

Interestingly, iron salts at low concentrations (20 μM) were able to reverse the antiparasitic effects of most triscatechol-containing compounds, suggesting that their antiparasitic effect may be primarily due to limitation of cellular iron availability. In contrast, the toxicity towards P. falciparum exerted by most monocatechol substances was not affected by exogenous iron salts (Table 1). This indicates that the antiparasitic effect of these substances is independent of the interaction with the intracellular iron pool and may be due to alternative mechanisms, e.g., direct attack of essential enzymes within the plasmodia or interference with iron release from hemoglobin.

In general it also appeared that siderophore compounds with the potential to chelate iron exerted increased toxicity towards mammalian cells (Table 1).

We have identified a number of siderophores with good antiplasmodial activity and low toxicity towards mammalian cells. However, the potential of the various substances to eliminate the parasite differed as a function of catecholate substitutions, of the basic structures, and of lipophilicity. Moreover, the mode of action may differ between monocatechols and triscatechol derivatives, with the latter primarily acting as iron chelators. Based on the present results, more-detailed structure-activity correlations need to be studied with new directed synthesized derivatives.

However, the antiplasmodial activity of all tested siderophores was higher than that of DFO B, a compound which has been used for treatment of human malaria. Thus, investigations of the most promising compounds identified in our study in murine models for malaria may be an attractive approach in the search for new antimalarials.