|Home | About | Journals | Submit | Contact Us | Français|
Francisella tularensis requires iron (Fe) for growth, but the biologic sources of Fe for this organism are largely unknown. We found that Francisella sp. growing in broth culture or within human macrophages can acquire Fe from the two major host Fe-binding proteins, lactoferrin (Lf) and transferrin (Tf). Fe acquisition is a potential target for novel therapies. Gallium (Ga) is a transition metal that interferes with cellular Fe metabolism by competing with Fe for uptake/utilization. Growth of either F. tularensis live vaccine strain (LVS) or Francisella novicida was inhibited by ≥2 μM Ga chelated to Tf or Lf, with GaLf being somewhat more potent. Francisella spp. express two Fe-containing antioxidant enzymes, catalase (KatG) and Fe cofactored superoxide dismutase (FeSOD). Growth of LVS with 10 μM GaTf or GaLf led to a dramatic decrease in bacterial catalase activity and in FeSOD activity that was associated with an increased susceptibility to H2O2. Ga also protected mice from intranasal challenge with F. novicida. Whereas 100% of the F. novicida-infected mice died by day 9, 75% of the mice receiving Ga continued to survive to at least day 15. Thus, a single intranasal dose of Ga followed by daily intraperitoneal Ga at a dose tolerated by the animals resulted in prolonged survival. These data support the potential utility of Ga as a therapy for F. tularensis infection of the lung.
Francisella tularensis is a highly virulent, Gram-negative pathogen that is the causative agent of tularemia (rabbit fever). F. tularensis is found worldwide, with rabbits, rodents, and deer serving as its principal natural mammalian reservoirs (17). There are four identified subspecies of F. tularensis, with two subspecies, Francisella tularensis subsp. tularensis (type A) and Francisella tularensis subsp. holarctica (type B), causing significant illness in humans (17). Type A strains are considerably more virulent for humans than type B strains (17).
In the United States, human infection by F. tularensis is rare, usually occurring as a skin infection of hunters exposed to F. tularensis-infected animals or ticks (17). Cutaneous tularemia is subacute and associated with low mortality. In contrast, when acquired by aerosols, type A strains have an extremely low 50% infectious dose (ID50) (estimated at <10 CFU) for humans and cause a rapidly progressing pulmonary infection with case fatalities of 30 to 60% (16, 17). This attribute has led to the development of type A strains as bioweapons. The World Health Organization estimated that exposure of 5 million people to an aerosol dispersal of 50 kg of a virulent strain of F. tularensis would lead to the incapacitation of 250,000 individuals and 19,000 deaths (15). In contrast, type B strains rarely cause mortality regardless of the route of inoculation. An additional species, Francisella novicida, is essentially avirulent for humans but causes a lethal disease in mice, making it an attractive model for tularemia in the laboratory.
F. tularensis is a facultative intracellular pathogen of macrophages (3, 41). Following ingestion by the macrophage, F. tularensis escapes the phagosome and gains access to the cytoplasm (12, 22, 41). It eventually causes macrophage cell death and/or escapes the cell, allowing infection of other naïve cells. Following its initial interaction with tissue macrophages, F. tularensis can disseminate throughout the reticuloendothelial system (41).
Limiting Fe availability is an evolutionary strategy for host defense against pathogenic organisms (18, 25). Consequently, there is essentially no “free” Fe in vivo (18). Instead, the Fe is chelated to transferrin (Tf) and lactoferrin (Lf), markedly decreasing its accessibility to pathogenic microbes. There is also a shift of Fe from serum to macrophages of the reticuloendothelial system mediated by the release of hepcidin, a member of the defensin family of cationic peptides, by hepatocytes in response to interleukin-6 (IL-6) and perhaps other proinflammatory factors (21). Hepcidin inhibits release of Fe from macrophages and other cells by inducing the internalization and breakdown of the plasma membrane Fe-exporting protein, ferroportin (32). Although this limits the availability of Fe for use by extracellular pathogens, its effect on intracellular pathogens that grow and replicate within macrophages is not clear and could actually enhance Fe availability for the microbe.
Bacterial pathogens acquire host Fe using a variety of strategies, e.g., production of siderophores, low-molecular-weight Fe chelators that compete with and/or remove Fe3+ from host Fe-binding proteins (31, 37). Whereas most bacterial pathogens have access to extracellular Fe, intracellular pathogens must either acquire Fe from within their intracellular compartment or establish adequate Fe stores before being phagocytosed.
The mechanism by which F. tularensis acquires Fe is only now becoming understood, with little information available as to the nature of biologically available Fe chelates that can be utilized by the organism. Both F. tularensis live vaccine strain (LVS) and F. novicida produce a siderophore when grown under Fe-limiting conditions (14, 27, 38, 40). The siderophore appears to be structurally similar to the polycarboxylate siderophore rhizoferrin made by Rhizopus and other fungal species (40). The F. tularensis genome appears to lack genes known to encode components associated with Fe metabolism in other bacteria, including the Fe transport facilitator, TonB, or membrane receptors for Tf, Lf, ferric siderophore complexes, hemoglobin, or hemopexin (2, 28).
Isogenic Francisella mutants unable to produce this siderophore grow poorly in Fe-limited broth culture (14, 40), and addition of purified rhizoferrin enhances F. tularensis growth under Fe-limited conditions (27). However, no significant difference in growth between an F. novicida mutant unable to produce the siderophore and wild-type bacteria was observed in J774 cells, although there was a trend toward slower growth of the mutant (14). Neither of these studies directly measured Fe uptake under extracellular or intracellular growth conditions, nor did the authors examine bacterial growth in human cells (14, 40). In addition, neither study was performed using the less bioavailable forms of Fe (bound to Lf or Tf) that F. tularensis would most likely encounter during infection of the lung and other sites in humans (14, 40).
Upon introduction into humans, Francisella species would have several different Fe chelates available to them to meet their metabolic and growth requirements—most notably Fe chelated to Tf and Lf, as well as smaller amounts of Fe in the form of low-molecular-weight chelates. In the present work, we assess the ability of Francisella to utilize these various forms of Fe when it is replicating extracellularly and/or intracellularly within human macrophages.
Since Fe acquisition is critical for microbial growth and virulence, it is a potential target for therapeutic intervention. Gallium (Ga) is a transition metal that interferes with cellular Fe metabolism by competing with Fe for uptake/utilization and is an FDA-approved drug (7). We have previously shown that Ga elicits Fe-reversible growth inhibition of Mycobacterium tuberculosis and Mycobacterium avium complex growing extracellularly or in human macrophages (33). Ga is also acquired by M. tuberculosis and inhibits its ability to acquire Fe (33). We also find that Ga causes Fe-reversible inhibition of the planktonic (free-living) growth of the opportunistic pathogen Pseudomonas aeruginosa and its ability to form biofilms, another Fe-dependent process (23). Given the desperate need for new antimicrobials that could be used in the event of a bioterrorism attack using an F. tularensis strain genetically engineered for resistance to conventional antibiotics, we also examined the ability of Ga to inhibit growth and Fe acquisition of F. tularensis and tested the efficacy of Ga in a murine model of pulmonary Francisella infection.
Francisella novicida (U112) and Francisella tularensis LVS were obtained from the ATCC. Bacteria were grown on Trypticase soy agar (TSA) for 24 to 48 h prior to use. The bacteria were scraped into appropriate medium and vortexed to produce a single cell suspension. Aliquots of the suspension were diluted in the desired medium and used for initial inocula.
Macrophages were obtained from healthy adult volunteers by a University of Cincinnati College of Medicine-approved institutional review board (IRB) protocol. Mononuclear cells were aseptically separated from heparinized blood through Ficoll-Hypaque and then cultivated in RPMI 1640 (Gibco, Grand Island, NY) supplemented with 20% pooled human serum for 5 days in Teflon wells in a 5% CO2 atmosphere at 37°C. Monocyte-derived macrophage (MDM) monolayers were then formed in 24- or 6-well culture plates, supplemented with 20% autologous serum, and incubated at 37°C for 1 day prior to incubation with the bacteria.
Bacteria were suspended in Trypticase soy broth (TSB) supplemented with 0.1% cysteine (Sigma Chemical, St. Louis, MO) to an initial A600 of 0.01 and incubated in Teflon wells for 24 h at 37°C and 5% CO2. 59Fe (10 μM) chelated to Tf (Sigma Chemical, St. Louis, MO, and Athens Research & Technology, Athens, GA), Lf (kindly provided by Ventria Bioscience, Sacramento, CA), or citrate (Fischer Chemical, Fairlawn, NJ), prepared as previously described (35), was added to the medium, and the bacterial culture was incubated for specified time periods. Triplicate aliquots were withdrawn and centrifuged (10,000 × g, 10 min, 4°C) to pellet the bacteria. The pellets were washed three times in TSB containing 5 mM ascorbate, pH 4.5. The pellet was finally suspended in 0.5 ml of TSB, added to an 0.2-μm Spin-X filter tube (Costar-Corning, Inc., Corning, NY), and centrifuged. The bottom of the filter containing trapped bacteria was cut off and placed into a gamma counter tube, and the bacterium-associated 59Fe was then determined using a gamma counter.
For Fe limitation experiments, bacteria were scraped from plates as above into Chamberlain's defined medium minus added Fe (F. novicida) or 10% TSB (LVS) at an initial inoculum of an optical density at 600 nm (OD600) of 0.01. Incubation continued in the presence or absence of exogenous Fe at 37°C and 5% CO2. At specified times, the absorbance of the culture at 600 nm (A600) was determined.
Our previously developed assay for the acquisition of Fe by M. tuberculosis residing within human MDMs (33, 34) was adapted for use with F. tularensis. MDM monolayers in RPMI containing 1% serum were incubated with the bacteria at a multiplicity of infection (MOI) of 50 for 2 h. Control cultures received medium only. The monolayers were washed twice, repleted with RPMI containing 1% pooled human serum and 50 μg/ml gentamicin, incubated for 1 h, and washed an additional three times. The monolayers were finally repleted with RPMI containing 1% serum and 2 μg/ml gentamicin. After 24 h of incubation, 10 μM 59Fe chelated to Tf, Lf, or citrate was added to the monolayers and incubated for defined time periods. These chelates of 59Fe were prepared as previously described (35). After desired periods, the monolayer was lysed using 0.1% sodium dodecyl sulfate (SDS). Sample aliquots were withdrawn for the estimation of total MDM-associated Fe. After each tube was pulse-vortexed, the contents were transferred into a 1.5-ml O-ring conical test tube and centrifuged (10,000 × g, 10 min, 4°C) to pellet the bacteria. The pellets were washed three times in RPMI containing 0.01% SDS. The final bacterial suspension was then filtered through an 0.2-μm Spin-X centrifuge tube filter (Costar-Corning). Bacterium-associated radioactivity on the filter was measured with a gamma counter. 59Fe content of the control was subtracted from wells containing bacteria to calculate the amount of bacterium-associated 59Fe.
The MDM monolayer was infected with bacteria as above and incubated overnight. After 24 h of incubation, 10 μM [59Fe] chelated with Tf, Lf, or citrate and 10 or 100 μM Ga(NO3)3 were added to the monolayers for 48 h. The monolayer was processed as above, and the 59Fe contents of both the MDM and isolated bacteria were determined with a gamma counter.
Bacterial lysates were generated by first incubating the bacteria in appropriate medium in the presence of Ga2-Tf or Ga2-Lf until the cell density reached an OD600 of 0.5 to 0.8. The suspension was centrifuged, and the pellet was washed three times in phosphate buffer. The pellet was incubated with lysis buffer (50 mM Tris plus 0.1 mM EDTA containing 0.2 μg/ml lysozyme) at 37°C for 4 h with constant tumbling. Alternatively, bacteria were disrupted after 25 strokes of a French press at 1,200 lb/in2. The suspension was centrifuged, and the lysate was added to Spin-X centrifuge tubes and centrifuged. The protein concentration of each lysate was determined using the bicinchoninic acid (BCA) assay (Pierce, Rockford, IL).
Catalase activity of bacterial lysates was measured by following the decomposition of H2O2 as monitored spectrophotometrically as previously described (6). The isoenzyme profiles of Francisella catalase and superoxide dismutase (SOD) activity were also assessed using native gel electrophoresis followed by activity staining for both enzymes (9, 11, 42). For analysis of catalase in nondenaturing gels, the gels were extensively rinsed with deionized H2O and soaked in 10 mM H2O2 for 10 min, followed by incubation in a staining solution consisting of 1% potassium ferricyanide and ferric chloride. Areas of catalase activity appear as achromatic bands on a blue-green background.
For SOD activity staining in nondenaturing gels, the gel was incubated in potassium phosphate buffer (50 mM, pH 7.8) containing 2.6 μM nitroblue tetrazolium, 28 mM riboflavin, and 28 mM tetramethylethylenediamine in the dark for 30 min. The gel was then rinsed, placed on an overhead transparency, and illuminated on a white-light transilluminator (Fisher BioTech). SOD enzyme activities were visualized as achromatic (white) bands against a dark purple background. SOD activity was also measured by a biochemical assay using the SOD assay kit WST (Dojindo Molecular Technologies, Inc., Rockville, MD). In this case, protein determination was made using the Bradford method (8).
F. novicida strain U112 was grown to exponential phase in Chamberlain's broth. The suspensions were then divided into three equal subcultures that were then supplemented with 0, 2, or 10 μM GaTf and incubated for an additional 1 h. All subcultures were then treated with 1 mM H2O2 for 30 min. In some cases, additional experiments were performed under the same conditions, except that Ga(NO3)3 or apoTf was used in lieu of GaTf. Cultures were diluted serially onto Mueller-Hinton agar, and CFU were determined after 24 to 48 h at 37°C. The CFU of respective cultures without H2O2 treatment served as a control.
Groups of four 5- to 6-week-old, female, BALB/c mice were infected by intranasal inoculation of a suspension of 100 CFU of F. novicida in 20 μl of phosphate-buffered saline (PBS) or PBS lacking F. novicida. The animals were then randomized to receive a single 50-μl dose (250 mg/ml) of intranasal Ga(NO3)3 20 min after Francisella administration that was then followed by daily intraperitoneal (i.p.) Ga(NO3)3 at a dose of 10 mg/kg of body weight, or they were left untreated. Animal survival was monitored and recorded daily.
All results are expressed as means ± 1 standard error of the mean (SEM). For analysis limited to two groups, Student's t test was employed (P < 0.05). Statistical comparisons among treatment groups were accomplished using analysis of variance (ANOVA) and the least significant difference test (P < 0.05) to determine differences between individual means.
Upon introduction into humans, Francisella species would encounter several different Fe chelates—most notably Fe chelated to Tf and Lf, as well as smaller amounts of low-molecular-weight Fe chelates. It seemed likely that the ability of Francisella to acquire Fe from the various chelates would differ in magnitude or kinetics. Therefore, we measured the ability of F. novicida (strain U112) and the LVS strain of F. tularensis to acquire Fe from Tf, Lf, and citrate when grown in broth culture. Plate-grown F. novicida or LVS bacteria were scraped into TSB medium supplemented with 0.1% cysteine, and acquisition of 59Fe chelated to Tf, Lf, or citrate was determined over time. Both F. novicida and LVS demonstrated the ability to readily acquire Fe from citrate and Tf but at 24 h were much less capable of acquiring it from Lf (Fig. 1A and B).
We found that Fe availability is growth rate limiting when F. novicida is grown in Chamberlain's medium (a defined minimal medium ) or when LVS is grown in 10% TSB. Addition of 5 μM Fe bound to Tf or Lf increased the growth rate of F. novicida in Chamberlain's medium and of LVS in 10% TSB 1.5- to 2-fold (not shown). Interestingly, when grown in Chamberlain's medium both F. novicida and LVS exhibited similar magnitudes of Fe acquisition at 24 h and the patterns of relative uptake from Tf, Lf, or citrate were also similar (Fig. 1C and D). Fe acquisition at 24 h from citrate remained greater than that from Tf or Lf for LVS and for Lf with F. novicida, but there was no longer a significant difference between Fe acquisition from Tf and that from Lf. The same was true for LVS grown in 10% TSB (Fig. 1C and D). Although the amount of 59Fe taken up at 24 h under Fe-limited conditions (Fig. 1C and D) was greater in magnitude than that in Fe-replete medium (Fig. 1A and B), this should not be interpreted to mean that more total Fe was taken up under Fe-limited conditions. The amount of 59Fe taken up is negatively impacted by the presence of “cold” Fe, which competes with it for acquisition by the organism. Without knowing and adjusting for the amounts of “cold” Fe present under the two growth conditions, one cannot accurately calculate and compare the total amounts of Fe taken up. These experiments indicate that both F. novicida and LVS appear to be capable of utilizing Fe bound to either Tf or Lf for growth under Fe-limited conditions in vitro.
Given that Francisella is an intracellular pathogen of macrophages and that the ability of the organism to access Fe bound to extracellular chelates from within macrophages may be limited, we examined the ability of F. novicida and LVS residing within human MDMs to acquire extracellular Fe bound to Tf, Lf, and citrate. MDMs were infected with F. novicida or LVS. 59Fe chelated to Tf, Lf, or citrate was then added. After 24 h, intracellular F. novicida bacteria were harvested (33). F. novicida-associated 59Fe was detectable with each of the Fe chelates employed. In contrast to results with organisms grown in broth culture, 59Fe acquisition appeared lowest when 59Fe was initially bound to Tf followed by Lf, with a marked increase when bound to citrate (Fig. (Fig.2A).2A). Similar results were obtained when Fe acquisition by LVS growing in human MDMs was examined (Fig. (Fig.2B).2B). Fe acquisition by MDMs from each of the chelates was not significantly different (not shown), consistent with our previous observations (35, 36), and thus did not explain the chelate-dependent differences in Fe acquisition by the bacteria.
Given our previous experience with Ga inhibition of M. tuberculosis and P. aeruginosa Fe metabolism and growth (26, 33), we initiated studies to assess the ability of this metal to disrupt Fe acquisition by Francisella sp. In order to accomplish this goal, it was desirable to perform such studies of a defined growth medium in which Fe availability is growth limiting. Using Chamberlain's medium for F. novicida and 10% TSB for LVS, we examined the effect of Ga bound to Tf or Lf on bacterial growth. We focused on Ga chelated to Tf and Lf as these are the forms in which Ga would most likely be presented to the organism within the human lung if administered systemically or by aerosol—i.e., Ga would not be present as a free metal. Both LVS and F. novicida showed a concentration-dependent growth inhibition when exposed to Ga chelated to Tf or Lf (Fig. (Fig.3).3). Growth inhibition was observed at Ga concentrations of ≥2 μM. GaLf appeared to be somewhat more potent than GaTf (Fig. (Fig.3).3). Ga-mediated growth inhibition was decreased by the addition of exogenous Fe to the medium, suggesting that Ga was acting by competitively interfering with the organism's ability to acquire and/or use Fe (Fig. (Fig.33).
We also found that Ga inhibited the ability of both F. novicida and LVS to acquire Fe from Lf or Tf (Fig. (Fig.4).4). The dose-response curves for inhibition of bacterial growth and Fe acquisition were quite similar (Fig. (Fig.4),4), suggesting a relationship between the two events. For both F. novicida and LVS, Ga bound to Tf inhibited bacterial acquisition of Fe initially bound to Tf (Fig. (Fig.4)4) or Lf (Fig. (Fig.5).5). Similarly, Ga bound to Lf inhibited bacterial acquisition of Fe regardless of whether the Fe was initially bound to Tf (Fig. (Fig.4)4) or Lf (Fig. (Fig.5).5). Finally, the addition of Ga to MDMs infected with F. novicida or LVS resulted in a decrease in bacterial Fe acquisition (Fig. (Fig.6A)6A) without significantly altering that of the MDMs (data not shown). Bacterial growth also decreased as a function of the concentration of Ga present (Fig. (Fig.6B6B).
Catalase and FeSOD are key antioxidant enzymes expressed by Francisella that contain Fe in their active sites. We hypothesized that if Ga was depleting the bacteria of Fe, Ga exposure would decrease bacterial catalase and FeSOD activity. Consistent with this hypothesis, growth of LVS in the presence of GaTf or GaLf decreased bacterial catalase activity 40 to 70%, as measured by biochemical assay (rate of disappearance of H2O2, Fig. Fig.7A).7A). That this was indeed secondary to an effect on KatG activity and not another mechanism of H2O2 catabolism was confirmed using the semiquantitative activity gel staining technique for KatG. The intensity of visualized KatG activity was considerably decreased with exposure to Ga (Fig. (Fig.7A).7A). Gallium also decreased bacterial FeSOD activity (Fig. (Fig.7B).7B). A GaTf-induced decrease in antioxidant enzyme activity would be expected to increase bacterial susceptibility to H2O2, and this was found to be the case (Fig. (Fig.88).
The above results with Ga suggested that it might have potential as an antimicrobial agent against Francisella during airway infection. As a proof of principle, we tested the ability of Ga(NO3)3 to protect mice from pulmonary infection by F. novicida. Ga(NO3)3 rather than GaTf or GaLf was chosen for several reasons. First, Ga(NO3)3 is the FDA-approved form of the drug and was used in our prior work in a P. aeruginosa lung infection model. Second, after administration of Ga(NO3)3 to animals, including humans, the drug is rapidly bound to Tf and Lf. Third, administration of Tf or Lf to the animals would have dramatically increased the expense and complexity of the experiments and required additional controls to be certain that the Fe-binding capacity of these proteins, rather than the Ga, was not responsible for any effect seen.
Consistent with the data of others (43), daily i.p. doses of 10 mg/kg Ga(NO3)3 were well tolerated by BALB/c mice, even in the absence of daily fluid challenge (not shown). We next tested the ability of Ga to prevent the death of animals that were inoculated with F. novicida into their lungs via intranasal challenge. Sets of 4 female BALB/c mice received intranasally (i) 100 CFU (10 times the lethal dose) of F. novicida, (ii) 100 CFU of F. novicida plus a single dose of intranasal Ga(NO3)3 that was then followed by daily i.p. Ga(NO3)3, or (iii) no bacteria but Ga(NO3)3 per the treatment regimen. As shown in Fig. Fig.9,9, F. novicida-infected mice that received Ga i.p. exhibited a far better outcome than did untreated mice. Whereas 100% of the F. novicida-infected mice who did not receive Ga died by day 9, 75% of the mice receiving Ga remained alive at day 15. Limiting the Ga dose to a single intranasal administration followed by a single i.p. dose 48 h postinfection showed no protective efficacy; all Ga-treated mice died at the same rate as did infected mice who were left untreated (data not shown). These results demonstrate that a single intranasal dose of Ga followed by sustained i.p. Ga treatment, at a dose tolerated by the animals, resulted in prolonged survival and support the potential utility of Ga as a therapy for F. tularensis infection of the lung.
Acquisition of Fe is critical for most bacterial pathogens to infect humans and cause disease. F. tularensis is one of the most virulent intracellular bacterial pathogens. When introduced into the lung of humans, F. tularensis produces disseminated disease with significant mortality, even with rapid antibiotic treatment (16, 17). Therefore, F. tularensis must have inherent mechanisms to acquire Fe from host sources. Previous work provided indirect evidence that Fe is important for the growth and replication of F. tularensis within macrophages (14, 20, 40). Our current studies further demonstrate that Fe is growth limiting for Francisella species replicating in vitro.
Some information has emerged regarding the processes utilized by F. tularensis to acquire Fe (13, 14, 27, 30, 38, 40). Two groups have shown that Francisella produces a siderophore (14, 40) that is structurally similar to the polycarboxylate siderophore rhizoferrin made by Rhizopus and other fungal species (40). Francisella siderophore production is mediated through the products of a series of genes (fslA to fslE) (14, 40). This siderophore appears to be very important for the organism to grow in Fe-limited medium (14, 40). However, to our knowledge, the ability of the organism to use host Fe chelates available in vivo, how availability of different forms of Fe might influence Francisella pathogenesis, and the role of rhizoferrin production in Fe acquisition from biologic sources of Fe have not been assessed.
In this study, we found that the capability of F. novicida and the LVS strain of F. tularensis to acquire Fe varies with the form in which the Fe is present, as well as the conditions under which the organisms are studied. Both F. novicida and LVS exhibited a greater ability to acquire Fe from citrate and Tf than from Lf when grown in TSB, whereas Fe acquisition was similar from all three chelates when the studies were performed in Fe-limited medium.
In contrast to Fe acquisition in broth medium, Fe acquisition by bacteria growing within macrophages is more complicated, as an organism is less likely to have direct contact with extracellular Fe. Thus, the source of Fe used by the bacteria might differ during extracellular versus intracellular growth. In support of this possibility, in spite of its decreased ability to grow in Fe-limited medium, an fslA mutant unable to produce rhizoferrin exhibited similar growth in the murine macrophage-like J774 cell line relative to wild-type organisms (14). Measurements of Fe uptake under extracellular or intracellular growth conditions by the wild type or the fslA mutant have not been reported, nor has Fe acquisition of Francisella growing in human cells been examined in detail (14, 40).
We found that both F. novicida and LVS can acquire Fe from both Tf and Lf while located within human MDMs. Smaller amounts of Fe were acquired from Tf than Lf, the opposite of what was observed in Fe-limited culture medium. Fe acquisition from citrate was far greater than from either Tf or Lf in MDMs, a pattern that was also seen to a smaller degree with growth in Fe-limited medium. The differences in Fe acquisition from the three chelates by intracellular Francisella could not be explained by differences in Fe acquisition by the MDMs. At this time, the mechanism(s) by which intracellular Francisella is able to acquire Fe is unclear, as is whether the strategies employed by extracellular and intracellular bacteria are the same.
Since Fe acquisition is critical for microbial growth and virulence, it is a potential target for novel therapeutic strategies. Gallium is a transition metal that interferes with cellular Fe metabolism by competing with Fe for uptake/utilization and is an FDA-approved drug for treatment of certain cancers and hypercalcemia associated with cancer (7). We have previously shown that Ga produces an Fe-reversible growth inhibition of M. tuberculosis and M. avium-intracellulare complex growing extracellularly or within human macrophages (33). Ga is acquired by M. tuberculosis and inhibits its acquisition of Fe (33). We also discovered that Ga leads to an Fe-reversible inhibition of the planktonic growth of P. aeruginosa, as well as the organism's ability to form antibiotic-resistant biofilms, another Fe-dependent process (23).
Our current studies reveal that Ga also negatively affects both extracellular and intracellular growth of Francisella in vitro by a mechanism that, based upon its ability to be abrogated by the addition of exogenous Fe, is likely related to disruption of bacterial Fe acquisition. The ability to inhibit both intracellular and extracellular growth of Francisella could be important in vivo as there is evidence that Francisella exists in both locations during infection in mice (19). Our initial studies demonstrate that daily intraperitoneal administration of Ga(NO3)3, the FDA-approved form of Ga, to mice was very effective in decreasing mortality resulting from intratracheal administration of approximately 10 times the lethal dose of F. novicida. More extensive studies are required to determine the optimal dose, timing, and formulation of Ga, as well as the extent to which Ga is active against the fully virulent SchuS4 strain of F. tularensis. Once administered, Ga is believed to be rapidly bound by Tf. In our in vitro studies, we found that Ga bound to Lf exhibited somewhat greater antibacterial activity than Ga bound to Tf. Whether this would hold true in vivo also warrants examination.
The key cellular target(s) by which Ga mediates its effects in vivo remains unclear. The amount of Ga-mediated growth inhibition of Francisella growing within macrophages, particularly with GaLf, was not as closely linked to inhibition of Fe uptake as was the case with growth in broth medium. The explanation for this observation is not clear. It is possible that in the macrophage system, Ga may be having additional effects beyond simply inhibiting bacterial Fe uptake. This could occur if Ga was altering bacterial susceptibility to macrophage host defense systems, e.g., reactive oxygen species, at concentrations below those needed for maximal inhibition of Fe uptake. Ga has a negative effect in vitro on Fe-dependent processes in Francisella even at concentrations that do not inhibit bacterial growth. A potentially key effect is a decrease in both catalase and FeSOD activity, which we find is associated with an increase in susceptibility of the organism to H2O2. The reduction in the activity of these critical antioxidants could enhance susceptibility to killing by macrophages, but the extent to which this occurs has yet to be examined. Previous studies have indicated that antioxidant enzymes play a role in virulence of Francisella sp. (4, 5, 24, 29, 39).
Our studies provide new insight into F. tularensis Fe metabolism and serve as a foundation for further exploration of Ga compounds as novel therapeutic agents against this pathogenic microorganism. This line of research could prove highly beneficial in the scenario of a bioterrorism attack with F. tularensis genetically engineered for widespread resistance to conventional antibiotics, such as those constructed by the former Soviet Union nearly 50 years ago (1). Furthermore, given the inability of most microbial systems to discriminate between Ga and Fe, this approach could have broad applicability for treating other bacterial pathogens.
This work was sponsored by the NIH/NIAID Regional Center of Excellence for Bio-defense and Emerging Infectious Diseases Research (RCE) Program. We acknowledge membership within and support from the Region V Great Lakes RCE (NIH award 1-U54-AI-057153). Additional support was provided by NIH PO1 AI44642 and a Merit Review Award from the Department of Veterans Affairs to B.E.B.
Published ahead of print on 16 November 2009.