Metal Chelation and Inhibition of Bacterial Growth in Tissue Abscesses

See allHide authors and affiliations

Science  15 Feb 2008:
Vol. 319, Issue 5865, pp. 962-965
DOI: 10.1126/science.1152449


Bacterial infection often results in the formation of tissue abscesses, which represent the primary site of interaction between invading bacteria and the innate immune system. We identify the host protein calprotectin as a neutrophil-dependent factor expressed inside Staphylococcus aureus abscesses. Neutrophil-derived calprotectin inhibited S. aureus growth through chelation of nutrient Mn2+ and Zn2+: an activity that results in reprogramming of the bacterial transcriptome. The abscesses of mice lacking calprotectin were enriched in metal, and staphylococcal proliferation was enhanced in these metal-rich abscesses. These results demonstrate that calprotectin is a critical factor in the innate immune response to infection and define metal chelation as a strategy for inhibiting microbial growth inside abscessed tissue.

Abscesses represent an immune response to infection that helps confine the spread of disease through the restriction of microbial growth and dissemination to neighboring tissues (1). S. aureus infection results in the formation of abscesses characterized by the extensive accumulation of host neutrophils. Although a role for neutrophils in abscess development is established, the specific host factors that limit microbial growth within the abscess are incompletely defined.

To identify host factors that limit microbial growth within the abscess, we applied imaging mass spectrometry (IMS) to a mouse model of S. aureus infection (2). Kidneys from uninfected and infected animals were sectioned and analyzed by means of IMS. Among the proteins that were present exclusively in abscessed tissue was a protein exhibiting a mass-to-charge ratio (m/z) of 10,165 that displayed the strongest mass-to-charge intensity observed in these experiments; this protein was subsequently identified as S100A8, a component of the calprotectin heterodimer (S100A8/S100A9) (Fig. 1, A and B) (2). Immunohistochemistry with antisera to S100A8 as a probe confirmed that S100A8 was only detectable in tissue from infected mice and localized coordinately with abscesses (Fig. 1C and fig. S1). Furthermore, IMS revealed that S100A8 and S100A9 colocalize to the site of infection, which supports heterodimeric calprotectin as the functional form of these proteins inside tissue abscesses (Fig. 1D). Calprotectin is an S100 EF-hand Ca2+-binding protein that accounts for ∼ 40% of the cytosolic protein pool of neutrophils (3). This protein was originally identified through its ability to inhibit the growth of a variety of fungal and bacterial pathogens in vitro (4). Calprotectin's antimicrobial activity has been proposed to be due to the calprotectin-mediated chelation of nutrient Zn2+ (57). However, the considerable body of data on the antimicrobial mechanism of calprotectin is largely indirect, and the contribution of calprotectin to the host-pathogen interaction has not been evaluated.

Fig. 1.

S100A8 is recruited to staphylococcal abscesses in a neutrophil-dependent manner. (A) IMS of S. aureus–infected and uninfected murine kidneys. Optical images of kidney sections mounted on a gold-coated matrix-assisted laser desorption/ionization plate are shown in the far left column. Two-dimensional ion density maps of representative proteins expressed in murine tissue are shown in the four right panels. CFU, colony-forming units. (B) SDS–polyacrylamide gel electrophoresis analysis of material extracted from kidneys of uninfected (lane 1) or infected (lane 2) mice followed by tandem mass spectrometry–based identification of proteins. The band at ∼10 kD was excised and subjected to further analysis. Residues in blue denote peptides that matched the sequence of S100A8. (C) Immunohistochemistry with S100A8 antisera localizes S100A8 to S. aureus–infected murine abscesses. Arrowhead denotes abscess. (D) IMS of S. aureus–infected and uninfected murine kidneys prepared as described in (A) showing the distribution of S100A8 (m/z 10,165) and S100A9 (m/z 12,976). Arrows denote abscesses. H&E, hematoxylin and eosin. (E) IMS analysis of S100A8 expression. In each organ set, the top row shows matrix-treated organs and the bottom row shows IMS. Mice were either depleted of neutrophils (RB6) or treated with control antibody (SRF3). Asterisks denote sites of abscess formation.

The first step in our approach was to ascertain the cell population responsible for recruitment and/or expression of S100A8 in the staphylococcal abscess. IMS revealed that S100A8 localizes coordinately with staphylococcal kidney and liver abscesses in neutrophil-replete mice. In contrast, infected neutropenic mice do not express S100A8 in the kidney or liver in spite of S. aureus–induced lesion formation in these organs (Fig. 1E and figs. S2 and S3). These findings indicate that the presence of S100A8 in infected kidney and liver abscesses is dependent on an intact neutrophil population.

Consistent with a previous report (4), we found that purified calprotectin inhibits S. aureus growth in a dose-dependent manner, with complete growth arrest observed at 75 μg/ml (Fig. 2A). The anti-staphylococcal activity of calprotectin is augmented when the protein is purified in the presence of Ca2+ (Fig. 2B). This result is consistent with calprotectin's assignment as an S100 EF-hand Ca2+-binding protein (8). We found that neither the S100A9 homodimer, nor the S100A9C3S homodimer, nor the S100A8/S100A9C3S heterodimer (9) was able to inhibit staphylococcal growth (Fig. 2C), indicating that heterodimerization and the cysteine residue at position 3 in S100A9 are also critical to calprotectin's antimicrobial activity.

Fig. 2.

Calprotectin inhibits S. aureus growth by chelating an essential component from the media. Error bars in (A) to (C), (E), and (F) (some of which are smaller than the symbols) represent the SD of at least three replicates, and all experiments were performed in triplicate. +, P ≤ 0.05; ++, P ≤ 0.005 (Student's t test). (A) Effect of recombinant calprotectin on S. aureus growth. O.D.600, optical density at 600 nm. (B) Analysis of S. aureus growth in the presence of calprotectin purified with (Ca2+) or without (no Ca2+) calcium as compared with that in buffer alone (buffer). (C) Analysis of S. aureus growth in the presence of S100A8/A9 heterodimers (A8/A9), S100A9 homodimers (A9), S100A8/S100A9C3S heterodimers (A8/A9*), or S100A9C3S homodimers (A9*) as compared with that in buffer alone (buffer). (D) Assessment of calprotectin's ability to coprecipitate with staphylococcal cells. (E) Ability of calprotectin (A8/A9) to inhibit S. aureus growth across a dialysis membrane as compared with that in buffer alone (buffer). (F) Growth kinetics of S. aureus in media that were pretreated with calprotectin (A8/A9) or buffer control (buffer) and filtered through a centricon column with a 5-kD cutoff.

The antimicrobial activity of calprotectin is proposed to occur through metal-ion chelation, suggesting that physical contact between calprotectin and S. aureus is not required for growth inhibition. We were unable to detect calprotectin-staphylococci interactions using a coprecipitation assay (Fig. 2D), which is consistent with this notion. Moreover, calprotectin (500 μg/ml) inhibited staphylococcal growth in the absence of physical contact (Fig. 2E), and S. aureus growth was decreased in media transiently treated with calprotectin (Fig. 2F). To determine whether calprotectin chelates nutrient cations, we quantitated calprotectin-dependent cation removal from media using inductively coupled plasma mass spectrometry (ICPMS). These analyses revealed no statistical difference in the number of Fe2+ or Mg2+ atoms remaining in the medium upon transient treatment with calprotectin (Fig. 3A). In contrast, Mn2+ and Zn2+ atoms were not detected after calprotectin treatment (Fig. 3A). A significant increase in detectable Ca2+ in the medium after calprotectin treatment is likely due to the release of excess Ca2+ ions bound to calprotectin in secondary binding sites as a by-product of purification in a Ca2+-rich buffer.

Fig. 3.

Calprotectin inhibits staphylococcal growth by chelating Mn2+ and Zn2+. Error bars in (A) and (C) to (F) represent the SD of at least three replicates. *, P ≤ 0.05 (Student's t test). (A) ICPMS analyses of metals in media after exposure to calprotectin (A8/A9) or buffer alone (buffer). Percent atoms in solution represents the percent of each metal present in medium after treatment as compared with the corresponding percentages in untreated medium. (B) Fluorescence emission spectra of S100A8/S100A9 in the absence (diamonds) and presence (squares) of Mn2+ (250 μM). (C) The sensitivity of S. aureus to growth in metal-rich media (“metal rep.”), media lacking all cations (“metal dep.”), media depleted for Zn2+ (“Zn2+ dep.”), or media depleted for Mn2+ (“Mn2+ dep.”). (D) Effect of calprotectin (A8/A9) on S. aureus growth in vitro in the presence of excess Zn2+ or Mn2+. (E) The growth of S. aureus upon exposure to neutrophil lysates in the presence or absence of excess Mn2+ as compared with that in buffer alone. (F) The growth of S. aureus upon exposure to polymorphonuclear leukocyte (PMN) lysates that have been immunodepleted of calprotectin (α-cal), treated with a nonspecific control antibody (mAb), or treated with buffer control (buffer).

Our data revealed a previously unrecognized role for calprotectin in binding Mn2+. To validate this hypothesis, we examined the Mn2+-binding properties of calprotectin using a fluorescence spectroscopy assay. The shift in the intensity of wavelength of the peak maximum as a result of addition of Mn2+ (Fig. 3B) confirmed that calprotectin does bind this metal ion. To determine the sensitivity of S. aureus to Zn2+ and Mn2+ starvation, we grew S. aureus in media deplete in either cation. These experiments revealed that S. aureus is acutely sensitive to Mn2+ deprivation, whereas the bacteria proliferate in media that contain trace levels of Zn2+ (Fig. 3C). Addition of excess Mn2+ or Zn2+ to growth medium rescues calprotectin-mediated inhibition of staphylococcal proliferation, confirming that Mn2+ and Zn2+ chelation is responsible for the anti-staphylococcal activity of calprotectin (Fig. 3D). Presumably, the ability of individual metals to rescue staphylococcal growth is due to saturation of the Zn2+/Mn2+-binding sites of calprotectin by either excess Zn2+ or Mn2+. As a further demonstration that S. aureus exposed to calprotectin is Mn2+-starved, S. aureus strains were created that are inactivated for the Mn2+-dependent transcriptional repressor (ΔmntR) or the Mn2+ uptake system (ΔmntA and Δ mntB). MntR is a repressor of the MntABC transport system responsible for Mn2+ acquisition, and its disruption renders S. aureus sensitive to Mn2+-mediated toxicity (10). We found that coincubation with calprotectin alleviates the sensitivity of ΔmntR to excess Mn2+, whereas strains defective in Mn2+ uptake (ΔmntA or Δ mntB) exhibit increased sensitivity to calprotectin toxicity (fig. S4). Furthermore, comparison of the global gene expression profiles of S. aureus exposed to calprotectin with those observed upon Mn2+ starvation, Zn2+ starvation, heat shock, cold shock, alkaline shock, and acid shock, as well as stringent response–inducing and SOS response–inducing conditions (11), demonstrated that S. aureus treated with subinhibitory calprotectin elicits an expression profile most closely resembling that of staphylococci starved for Zn2+ and Mn2+ (fig. S5 and tables S4 to S9). Taken together, these data support a model whereby calprotectin prevents S. aureus growth through the chelation of nutrient Mn2+ and Zn2+.

We next sought to determine the contribution of calprotectin to neutrophil-mediated bacterial killing. Neutrophils extracted from wild-type (WT) and calprotectin-deficient mice kill S. aureus in vitro with similar efficiency, suggesting that calprotectin does not contribute to phagocytic killing (fig. S6). However, the antimicrobial activity of the cytoplasmic compartment from purified neutrophils is reversible upon the addition of excess Mn2+ or immunodepletion of calprotectin (Fig. 3, E and F) (2). These results indicate that neutrophil cytoplasmic extracts containing calprotectin inhibit bacterial growth through metal chelation. In keeping with this, calprotectin is likely directed at bacterial pathogens through targeted secretion systems or upon neutrophil lysis as a posthumous killing mechanism. To determine whether calprotectin-mediated cation chelation inhibits bacterial growth in tissue abscesses, we compared the growth of S. aureus in abscesses from WT and calprotectin-deficient (S100A9–/–) mice (2). Livers from mice lacking calprotectin exhibit increased abscess formation and bacterial burden as compared with livers from infected WT animals (Fig. 4, A and B). Although we did not detect a significant difference in neutrophil recruitment to the livers of infected animals (fig. S7), additional immunomodulatory functions attributed to calprotectin could contribute to the sensitivity of calprotectin-deficient animals to S. aureus infection (12).

Fig. 4.

S100A9 is a vital component of the neutrophil arsenal to inhibit S. aureus growth in vivo. (A) Photographs of livers dissected from C57BL/6 WT and S100A9Δ mice infected with S. aureus 96 hours after infection. (B) S. aureus multiplication in infected livers. Each circle represents a positive culture from a single liver. The mean of each group is represented by a horizontal line (n = 16 mice for S100A9Δ, n = 19 mice for C57BL/6 wild type). *, P ≤ 0.02 (Student's t test). (C) LA-ICPMS of infected organs from WT and calprotectin-deficient (S100A9Δ) mice. Top panel shows H&E stains. Bottom panels show LA-ICPMS maps for Ca2+ (“calcium-44”), Mn2+ (“manganese-55”), and Zn2+ (“zinc-67”). Arrows denote the site of abscesses. Scales are presented in arbitrary units. (D to F) Quantitative determinations of metal concentrations in abscessed tissue from WT or calprotectin-deficient (S100A9Δ) mice extracted via LCM. The error bars in (D) to (F) represent the SD of three replicates, and the asterisk in (E) denotes statistical significance (P < 0.05) as determined by the Student's t test.

To assess the effect of calprotectin on nutrient metal availability inside the abscess, we applied laser ablation ICPMS (LA-ICPMS) mapping to image metal distribution in infected animal tissue (2). LA-ICPMS revealed that staphylococcal liver abscesses from WT mice are enriched in Ca2+, which is consistent with robust immune cell infiltrate to the infection site (Fig. 4C). In contrast, these abscesses are devoid of detectable Zn2+ and Mn2+, unequivocally establishing the abscess as a cation-starved environment (Fig. 4C and fig. S8). Abscesses from mice lacking calprotectin contain appreciable levels of Ca2+; however, these levels appear diminished as compared with the levels in abscesses from WT mice, potentially reflecting the Ca2+ contribution of calprotectin to the abscess. Furthermore, calprotectin-deficient abscesses contain levels of Mn2+ that are equivalent to those in the surrounding healthy tissue, demonstrating the in vivo requirement for calprotectin in Mn2+ chelation and removal from the abscess. The absence of calprotectin did not have a noticeable impact on Zn2+ levels in these experiments. To corroborate these results, we extracted abscessed material from WT and calprotectin-deficient animals using laser capture microdissection (LCM) and quantitated Zn2+, Mn2+, and Ca2+ concentrations in isolated abscesses using ICPMS (Fig. 4, D to F). These experiments confirmed that abscesses from mice lacking calprotectin are significantly enriched in Mn2+ as compared with those from WT mice (Fig. 4E). Taken together, these results establish calprotectin-mediated metal chelation as an immune defense strategy to prevent bacterial outgrowth in tissue abscesses.

The inhibition of bacterial nutrient uptake represents a promising alternative area of research for the design of new antimicrobials, and the observed calprotectin-mediated metal chelation provides a specific direction to assess the therapeutic potential of this concept. The results reported here suggest that beyond direct treatment of abscesses with calprotectin protein, noncytotoxic bioavailable metal-ion chelators represent a promising area of investigation for inhibitors of microbial growth.

Supporting Online Material

Materials and Methods

Figs. S1 to S8

Tables S1 to S9


References and Notes

View Abstract

Navigate This Article