The dual personality of iron chelators: growth inhibitors or promoters?

Iron is an essential nutrient for almost all living cells, and pathogenic bacteria are faced with severe iron limitation in the mammalian host (1). To make iron available during infection, most microorganisms produce and transport into the cell specific iron chelators (siderophores) or acquire iron bound to exogenous iron carriers, like xenosiderophores, heme, or the host’s transferrins (2). Due to their capacity to promote microbial growth, siderophores act as virulence factors in several models of infection (3). In a recent article published in Antimicrobial Agents and Chemotherapy, Thompson et al. investigated the susceptibilities of common nosocomial pathogens (Acinetobacter baumannii, Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Staphylococcus aureus) to iron chelators that have recently been developed for treatment of iron overload in humans, as well as to two U.S. Food and Drug Administration-approved drugs, namely, deferoxamine (Desferal; Novartis) and deferiprone (3-hydroxy-1,2-dimethylpyridin-4(1H)-one; Sigma-Aldrich) (4). In the same issue of the journal, de Léséleuc et al. assessed the activity of these two chelators against A. baumannii (5). Thompson et al. emphasize the potential of deferiprone as an antibacterial agent, but they overlook the problem that some chelators behave as Janus-faced molecules; on one face, chelators are essential to treat iron overload in humans, while on the other, they can act as growth promoters for invading pathogens (reviewed in reference 6). A paradigmatic example of this duality is offered by deferoxamine, whose xenosiderophore activity has been proven for several pathogenic bacteria, including P. aeruginosa (7, 8), consistent with clinical evidence indicating that prolonged administration to humans, as in cases of -thalassemia, increases susceptibility to infection (6). Thompson et al. showed that high deferiprone concentrations inhibit P. aeruginosa growth in RPMI medium and ascribed the antibacterial activity to iron withholding. Deferiprone MICs were 128 to 512 g/ml, equivalent to 0.92 to 3.67 mM, depending on the strain (4). However, no evidence that iron could reverse growth inhibition was provided, raising concern about the mechanism of deferiprone activity and its intrinsic toxicity. Conversely, no inhibition was reported for deferoxamine up to 512 g/ml, equivalent to 0.78 mM (4). Since deferiprone plasma levels in human therapy are 0.1 mM (http://www.accessdata.fda.gov /drugsatfda_docs/label/2011/021825lbl.pdf), the above findings do not rule out the possibility that deferiprone acts as an iron carrier to P. aeruginosa at lower, yet physiologically meaningful, concentrations. P. aeruginosa is endowed with an impressively broad iron uptake capability, producing two endogenous siderophores, pyoverdine and pyochelin, and is capable of acquiring iron via a multiplicity of exogenous chelators (9). Consistently, P. aeruginosa has the coding potential for up to 34 TonB-dependent siderophore receptors, whose ligand specificities have only in part been determined (10). Siderophore receptor redundancy denotes the importance of iron acquisition for P. aeruginosa metabolism and suggests the existence of still-unexplored uptake capabilities in this species. Here, we tested the antimicrobial activities of unsaturated and iron-saturated deferiprone against three prototypic P. aeruginosa strains: PAO1 (ATCC 15692 type strain isolated from a wound), PA14 (burn isolate [11]), and TR1 (cystic fibrosis isolate [12]). The deferiprone MICs in the iron-poor media M9-succinate (M9-S) and RPMI 1640 were 0.92 mM for all three strains, in line with results by Thompson et al. (4). However, saturation of deferiprone with iron (1:3 Fe[III]/deferiprone ratio) reversed but did not abrogate the antibacterial activity, resulting in MIC values of 1.84 to 3.67 mM, depending on the strain. At these or higher ferric-deferiprone concentrations, growth was still inhibited. Thus, while the anti-P. aeruginosa activity of deferiprone can partly be ascribed to its iron-chelating properties, some intrinsic, chelation-independent toxicity should be taken into account at high concentrations ( 3.67 mM). Then, we analyzed the growth response of P. aeruginosa to deferiprone. Supplementation of M9-S with increasing deferiprone concentrations (0 to 0.16 mM) was paralleled by an increase of P. aeruginosa PAO1 growth rates and a decrease of pyoverdine yields at the stationary phase, suggesting that deferiprone can act as an iron carrier, thereby promoting bacterial growth and reducing the need for endogenous siderophore production (Fig. 1A). To gain further insight into this activity, we generated a P. aeruginosa PAO1 pvdA pchD double-null mutant impaired in the synthesis of both endogenous siderophores, pyoverdine and pyochelin, and analyzed its growth response to the addition of endogenous or exogenous siderophores (20 M). Compared with wild-type PAO1, the pvdA pchD mutant grew poorly in M9-S and even less well in M9-S containing 400 M 2,2=-dipyridyl (a chelator of the intracellular Fe[II] pool), while growth was restored by the addition of 50 M FeCl3, reflecting a severe defect in iron acquisition (Fig. 1B). As expected, growth of the mutant was strongly promoted by the addition of pyoverdine or deferoxamine mesylate salt to the medium. Pyochelin had a similar though less pronounced effect, likely due to its poor affinity for iron (9). Most remarkably, supplementation with deferiprone resulted in significant growth promotion of the pvdA pchD mutant, higher than that caused by the homologous siderophore pyochelin (Fig. 1B). Siderophores are actively transported in P. aeruginosa via specific receptors, namely, FpvA and