Lactoferricin influences early events of Listeria monocytogenes infection in THP-1 human macrophages

Listeria monocytogenes is an invasive Gram-positive bacterium associated with severe food-borne diseases in humans that infects a wide variety of host cells, including non-professional and professional phagocytic cells (especially macrophages) (Vazquez-Boland et al., 2001). C3bi and C1q complement receptors are involved in uptake by phagocytic cells (Alvarez-Dominguez et al., 1993; Drevets et al., 1993), although non-opsonic receptorligand interactions also occur (Pierce et al., 1996). The bacterial surface protein ActA, essential for intracellular actin-based motility, may be a L. monocytogenes ligand that mediates recognition of heparan sulphate proteoglycan receptors present on the surface of both professional and non-professional phagocytes (Alvarez-Dominguez et al., 1997; Henry-Stanley et al., 2003).

Naturally occurring antimicrobial substances that are resistant to gastric fluid and capable of preventing listerial infection at the level of enterocytes and/or macrophages may be useful in hindering the spread of this pathogen to host tissues. Among biologically active milk proteins, lactoferrin (Lf), a multifunctional iron-binding protein found in milk and colostrum, as well as in various other secretions (Vorland, 1999), and lactoferricin (Lfcin), a 25-residue, pepsin-generated Lf fragment (Tomita et al., 1994), have been shown to possess antibacterial activity towards Gram-negative and Gram-positive bacteria, including L. monocytogenes (Arnold et al., 1980; Payne et al., 1990; Tomita et al., 1994; Branen & Davidson, 2000). Decreased L. monocytogenes invasion of human intestinal cultured cells in the presence of Lf from bovine milk (BLf) (Antonini et al., 1997) has been correlated to an interaction with listerial surface proteins (Conte et al., 1999).

Although different mechanisms of Lf antimicrobial defence have been identified, the antibacterial action of Lfcin has not yet been fully characterized. Research data provide evidence that BLf-derived lactoferricin (LfcinB) shows similarities to several major classes of cationic antimicrobial peptides and possesses an amphipathic structure capable of depolarizing the membrane of susceptible bacteria (Hwang et al., 1998). Hence, direct interaction of the LfcinB net positive charge with charged elements in the bacterial membranes, such as lipopolysaccharide in Gram-negative bacteria and lipotechoic acid in Gram-positive bacteria, is required for its lethal effect (Appelmelk et al., 1994; Elass-Rochard et al., 1995).

To add further information on the properties of LfcinB, research has been carried out to analyse whether this peptide could affect the invasion of macrophages by L. monocytogenes. The effect of non-bacteriostatic and non-bactericidal concentrations of BLf, LfcinB and LfcinB residues 49 (LfcinB₄₉), termed the antimicrobial centre of LfcinB (Tomita et al., 1994), towards the invasion of IFN-γ-activated human macrophages (THP-1) by L. monocytogenes was investigated.

L. monocytogenes LM2, isolated previously in our laboratory from the spinal fluid of a newborn affected by listeriosis, was used in this study (Conte et al., 1994). Bacteria were routinely grown in brain heart infusion broth (BHI; Oxoid), pH 7.2, and maintained by serial passage on tryptone soy agar (TSA; Oxoid).

BLf was purchased from Sigma, LfcinB was prepared according to Bellamy et al. (1992) and amidated hexapeptide LfcinB₄₉ (RRWQWRNH₂) was synthesized commercially (Inbios). Protein purity and concentration were determined as described previously (Antonini et al., 1997). Heparin (170 USP units mg^-1) was purchased from Sigma. All reagents were dissolved as stock solutions in pyrogen-free PBS, pH 7.2.

Minimum inhibitory concentrations (MICs) were determined in 96-well plates (Sterilin) after 1 h incubation of L. monocytogenes suspensions with proteins. MICs were defined as the lowest concentrations of protein that completely inhibited growth, determined by OD₆₂₀. To assess minimum bactericidal concentrations (MBCs), aliquots of 100 µl were taken from all wells, spread on TSA and evaluated by c.f.u. counts. MIC and MBC values were obtained by repeating the procedures on each protein at least four times.

THP-1 cells, a myelomonocytic cell line, were maintained in RPMI 1640 medium as described previously (Conte et al., 2002). THP-1 cells were differentiated by incubation with phorbol myristate acetate (0.16 µM; Sigma) for 48 h at 37 °C in 24-well Nunc plates and activated by exposure to recombinant human IFN-γ (Roche Diagnostics) [100 U ml^-1; specific activity 2x10⁷ U (mg protein)^-1] for 24 h at 37 °C.

The maximal non-cytotoxic dose of proteins was determined by incubating serial dilutions of each substance in RPMI at 37 °C with confluent THP-1 cells grown in 96-well microplates (Flow Laboratories) and by evaluating cell morphology, viability and proliferation after 24, 48 and 72 h, as described previously (Antonini et al., 1997). The highest protein concentration that did not affect any cytotoxicity parameter was 30 µM for BLf, 8 µM for LfcinB and 628 µM for LfcinB₄₉.

Adhesion of L. monocytogenes was determined by adding exponentially grown bacteria to IFN-γ-activated THP-1 cells (10 bacteria per cell), followed by incubation at 4 °C for 1 h in the presence or absence of different protein dilutions in RPMI. Cells were then washed extensively in RPMI and lysed with ice-cold 0.1 % Triton X-100. Invasion assays were performed by incubating exponentially grown bacteria (1 bacterium per cell) for 1 h at 37 °C (infection step) in the presence or absence of protein dilutions. In some assays, proteins were pre-incubated with macrophages or bacteria before the infection step. After the infection step, macrophages were washed in RPMI and lysed or, to evaluate intracellular growth, incubated at 37 °C with the same medium supplemented with 5 µg gentamicin ml^-1 and lysed at different times after infection. Viable adherent or intracellular bacteria were determined by plate counts. Adhesion and invasion were expressed as percentages of inoculated bacteria that adhered to or invaded TPH-1 cells after 1 h incubation at 4 or 37 °C. Intracellular growth was evaluated as the replication index (RI), corresponding to the number of c.f.u. at 4 h post-infection divided by the number of c.f.u. at 1 h post-infection.

Macrophages infected with L. monocytogenes LM2 (100 bacteria per cell) in the presence or absence of LfcinB were fixed at 30 and 60 min after addition of bacteria and processed for electron microscopy as described previously (Conte et al., 2002).

Statistical analysis was performed using Student's t-test for unpaired data. Results were expressed as means ± SD and P values < 0.05 were considered significant.

Effect of BLf and BLf-derived peptides on L. monocytogenes invasion
To determine the MICs of BLf, LfcinB and LfcinB₄₉ hexapeptide towards strain LM2, different concentrations of protein in BHI were incubated at 37 °C with L. monocytogenes suspensions for 1 h. The lowest MIC was obtained with LfcinB (31 µM), whereas BLf and LfcinB₄₉ gave MICs of >120 and >500 µM. None of the proteins exerted bactericidal activity.

To evaluate the effect of BLf and BLf-derived peptides on entry and survival of L. monocytogenes, mixtures of bacteria and different non-cytotoxic, non-bacteriostatic, non-bactericidal concentrations of proteins were incubated with IFN-γ-activated THP-1 macrophages for 1 h at 37 °C (infection step). Table 1 shows the results of invasion assays carried out with BLf, LfcinB and LfcinB₄₉ at the same concentration (1.6 µM). Only when THP-1 macrophages were infected in the presence of LfcinB was the invasion rate after 1 h incubation affected significantly (0.7 versus 4.4 %; P < 0.05). Slight decreases only in invasion efficiency were observed with 6 µM BLf and 50 µM LfcinB₄₉ (invasion of 3.5 ± 0.2 and 3.0 ± 0.2 %, compared with 4.2 ± 0.4 % for untreated controls). When monolayers were incubated further for 4 h in the presence of gentamicin, RIs showed that IFN-γ-treated TPH-1 cells were non-permissive for bacterial growth in all experimental conditions described above (data not shown).

Table 1. Effect of BLf and BLf-derived peptides on invasion of L. monocytogenes LM2 in IFN-γ-activated THP-1 macrophages Proteins were each added at 1.6 µM. Invasion is expressed as the percentage of the initial bacterial inoculum that was internalized in THP-1 cells 1 h post-infection. Data are means ± SD of at least four experiments.

To investigate the mechanisms involved in inhibition of invasion by LfcinB further, bacterial suspensions or cell monolayers were pre-incubated with LfcinB (1.6 µM) before the infection step. The results showed that LfcinB was effective under both these experimental conditions (0.3 ± 0.3 % invasion for LfcinB-treated bacteria and 1.6 ± 0.2 % for LfcinB-treated THP-1 cells compared with 4.5 ± 0.2 % for untreated controls; P < 0.05).

In order to verify whether inhibition of invasion by LfcinB could be due to competition of this peptide with the listerial surface protein ActA for common heparan sulphate receptors (Alvarez-Dominguez et al., 1997; Shimazaki et al., 1998), infection of THP-1 cells was performed in the presence of increasing concentrations of heparin (1, 10 and 100 µg ml^-1), alone or in combination with a fixed concentration of LfcinB (1.6 µM). After 1 h incubation at 37 °C, heparin alone decreased L. monocytogenes invasion of THP-1 cells in a dose-dependent fashion, whereas heparin in the presence of LfcinB was unable to achieve complete prevention of peptide inhibition (Fig. 1).

(27K): Fig. 1. Effect of heparin and LfcinB in combination on L. monocytogenes infection. Increasing concentrations of heparin, alone or in combination with LfcinB (1.6 µM), were incubated with IFN-γ-activated macrophages during the infection step (1 h at 37 °C). Invasion was monitored by c.f.u. counts and expressed as percentage of internalized bacteria. Bars indicate SD.

Effect of LfcinB on early interaction between L. monocytogenes and macrophages
The ability of bacteria to adhere to IFN-γ-activated macrophages was assayed in the presence of equal concentrations of BLf, LfcinB and LfcinB₄₉. The percentage of bacteria that adhered to macrophages was similar whether or not they were exposed to BLf or hexapeptide, whereas a slight increase in adherent bacteria was observed when assays were carried out with LfcinB (2.1 ± 0.1 versus 1.6 ± 0.2 %).

Further experiments were carried out by submitting infected monolayers to a temperature shift from 4 to 37 °C (Fig. 2): after the infection step in the presence of 1.6 µM LfcinB (1 h at 4 °C), monolayers were incubated at 37 °C for 15, 30 and 60 min and then lysed immediately for bacterial counts. The results obtained showed that inhibition of bacterial invasion occurred within 30 min after the temperature shift to 37 °C.

(16K): Fig. 2. L. monocytogenes invasion of IFN-γ-activated macrophages during the infection step with () or without () LfcinB (1.6 µM). The infection step was performed at 4 °C (1 h) and cells were then incubated at 37 °C for 15, 30 or 60 min and immediately lysed for c.f.u. counts. Bars indicate SD.

Transmission electron microscopy of macrophages infected with L. monocytogenes in the presence of LfcinB
After 30 min infection, bacteria were detected in the cytoplasm of untreated cells (Fig. 3a), whereas, in the presence of LfcinB (1.6 µM), a delay in bacterial internalization was observed and L. monocytogenes was detected close to the plasma membrane (Fig. 3b). When ultrastructural investigations were carried out after 60 min infection, different behaviour was also observed in LfcinB-treated cells compared with untreated controls: infection of IFN-γ-activated THP-1 cells with strain LM2 in the absence of LfcinB resulted in numerous bacteria surrounded by a network of actin filaments, and bacteria with actin tails were frequently observed (Fig. 3c). However, in LfcinB-treated infected cells, bacteria were observed just internalized in the cytoplasm (Fig. 3d).

(171K): Fig. 3. Electron micrographs of IFN-γ-activated THP-1 macrophages infected with L. monocytogenes. Micrographs were taken 30 (a, b) or 60 (c, d) min after addition of bacteria. (a) Untreated bacteria internalized into the cytoplasm; (b) LfcinB-treated L. monocytogenes in proximity of cells; (c) untreated L. monocytogenes in multiplication with actin tail; (d) LfcinB-treated bacteria internalized into cytoplasm. Bars, 1.3 µm.

Significant inhibition of L. monocytogenes invasion of THP-1 cells took place when LfcinB, a 25-residue BLf-derived peptide, was included during the bacterial infection step. This effect was significantly greater than that caused by intact BLf or the LfcinB₄₉ fragment. Similar to the effect observed for BLf (Conte et al., 1999), the reduced entry following pre-incubation of LfcinB with bacteria suggested an interaction of this peptide with Listeria cell surfaces. Moreover, the decreased invasion obtained with LfcinB-treated macrophages could be due to competition for common heparan sulphate receptors (Alvarez-Dominguez et al., 1997; Shimazaki et al., 1998).

However, the failure of heparin to cause a complete elimination of LfcinB inhibition, together with the inability of LfcinB to decrease bacterial adherence, suggests the involvement of other receptor ligands in the interaction of bacteria with macrophages and/or the possibility that LfcinB may also influence early phases of the post-absorption step. Ultrastructural studies showed that a delay in bacterial entry may occur when invasion assays were performed in the presence of LfcinB. Invasion experiments involving a temperature shift from 4 to 37 °C supported this hypothesis. At this time, the decrease in intracellular bacteria could depend on the intracellular generation of reactive oxygen species induced by LfcinB treatment in THP-1 cells (Yoo et al., 1997). However, studies performed with IFN-γ-non-activated THP-1 macrophages showed that listerial cells were also able to multiply in the presence of LfcinB (unpublished results). Furthermore, since, similar to BLf (Moriishi et al., 1999), we have observed previously that LM2 haemolysis was inhibited by LfcinB (unpublished data), interference of this peptide in listeriolysin O-mediated release of bacteria in the cytoplasm could not be excluded.

The short hexapeptide fragment of LfcinB, LfcinB₄₉, was virtually unable to affect bacterial invasion in macrophages. Since the hexapeptide interacts with hydrophobic bacterial membranes, similar to LfcinB (Schibli et al., 1999), it could be argued that the mere binding of a small amphipathic peptide to bacterial membranes is not sufficient to hinder bacterial invasion in macrophages.

These findings, as well as the well-known antimicrobial effects of LfcinB, highlight a new antibacterial property for this peptide, i.e. the ability to affect early steps of infection by an intracellular facultative pathogen through an interaction with both bacteria and target cells, thus suggesting the possibility of hindering in vivo the spread of bacteria to preferred localization sites within the host. The present results, obtained in the THP-1 model system, may have future therapeutic potential, but they need to be extended to a cell system closer to that encountered in vivo by L. monocytogenes, such as human peripheral monocytes.

This work was supported by MIUR grants to L. S., M. P. C. and G. A. and by a grant from the National Institute of Health to F. S.