OUP user menu

Bacteriostatic activity of human lactoferrin against Staphylococcus aureus is a function of its iron-binding properties and is not influenced by antibiotic resistance

Antonio Aguila, Ana G. Herrera, Donald Morrison, Bonnie Cosgrove, Alicia Perojo, Isis Montesinos, Julián Pérez, Gustavo Sierra, Curtis G. Gemmell, Jeremy H. Brock
DOI: http://dx.doi.org/10.1111/j.1574-695X.2001.tb00511.x 145-152 First published online: 1 August 2001

Abstract

The in vitro antistaphylococcal activity of lactoferrin and the antibiotic resistance of clinical Staphylococcus aureus isolates obtained from three different sites of infection were examined. Antibiotic, but not lactoferrin resistance correlated with selective antibiotic pressure, and nosocomial and most community isolates were antibiotic resistant, whereas only a third of each group was resistant to lactoferrin. The antimicrobial activity of lactoferrin, both in defined medium and in normal human plasma serum, was dependent upon its ferrochelating properties. Therapeutic approaches based on the use of ferrochelating agents such as lactoferrin combined with antimicrobial drugs may help to counteract the reduced efficacy of current antibiotics.

Keywords
  • Antibiotic
  • Antimicrobial activity
  • Iron
  • Lactoferrin
  • MRSA
  • Staphylococcus aureus

1 Introduction

Staphylococcus aureus is the leading cause of surgical wound infections and the second most frequent cause of intrahospital infections world-wide [1]. The appearance of multidrug-resistant and highly transmissible strains [2,3], and increasing morbidity and mortality due to staphylococcal infections have focused attention on host immune mechanisms active against S. aureus as an alternative to antibiotics.

Among the human innate immune defence mechanisms is lactoferrin, an iron-binding protein (IBP), which occurs in neutrophil secondary granules and external secretions [4]. This protein displays anti-inflammatory [5] and antimicrobial [4,6] effects partly by means of binding iron, which limits the amount of ‘free’ iron available. Consequently, increased lactoferrin concentrations due to neutrophilia induced during bacterial infections and recruitment of neutrophils to sites of inflammation [7] may, in combination with other immune defence mechanisms, limit microbial growth and inflammation [5,8].

The capacity to acquire iron from the various host IBPs, including lactoferrin, correlates with microbial virulence. Several bacterial mechanisms for iron uptake are induced under iron deprivation [4,9]. Some bacteria synthesise siderophores, which can sequester host and environmental iron. Three structurally distinct siderophores, staphyloferrins A and B, and aureochelin, have been reported for S. aureus[10], but little information is available about their ability to acquire iron bound to IBPs [4,9,11]. As an alternative strategy, several pathogenic bacteria express receptors for host-specific IBPs. S. aureus had been reported to express receptors for both lactoferrin [12] and transferrin [13], but the former has not been well characterised and the latter, although it can mediate iron removal from transferrin in vitro, it unlike most microbial transferrin receptors is of relatively low affinity (Kd~0.27 µM), does not distinguish between the apo- and diferric forms of human transferrin, is expressed constitutively, and also has glycolytic activity [13]. Its role in vivo has not been established. Thus, it is far from clear how S. aureus acquires host iron in vivo.

Lactoferrin, and lactoferrin-derived peptides comprising parts of the N-terminal residues 1–47 of human lactoferrin [14] or 1–48 of bovine lactoferrin [15], also possess a non-ferrochelating antimicrobial mechanism based on their direct bactericidal activity. They are reported to act against S. aureus both in vitro [16,17] and in vivo [16].

The potential use of lactoferrin as an alternative antibiotic against S. aureus is suggested by reports of its antimicrobial effects in vitro [4,6] and in murine models of systemic diseases [5,18,19]. Furthermore, resistance to antibiotics and to lactoferrin may be linked as sub-inhibitory concentrations of specific antibiotics modulate staphylococcal siderophore production [20].

To further address the antimicrobial potential of lactoferrin against S. aureus and its relation to antibiotic resistance, we have examined the nature of the in vitro antistaphylococcal activity of lactoferrin and the antibiotic resistance of nosocomial and community clinical S. aureus isolates obtained from three different sites of infection: bloodstream (BS), skin/soft tissues (SST) and conjunctiva (CJ). It was found that antibiotic, but not lactoferrin resistance correlated with selective antibiotic pressure and that nosocomial and most community isolates were antibiotic resistant, whereas only a third of each group was resistant to lactoferrin. The antimicrobial activity of lactoferrin was mainly microbiostatic and dependent on its ferrochelating properties. These results show that lactoferrin has a significant microbiostatic activity against clinical S. aureus isolates, irrespective of their antibiotic resistance. Therapeutic approaches based on the use of ferrochelating agents such as lactoferrin combined with antimicrobial drugs may help to counteract the alarmingly reduced efficacy of current antibiotics.

2 Materials and methods

2.1 Laboratory strains

S. aureus Cowan I, a standard laboratory strain of human origin, was kindly provided by Dr. D. Corcoran, Department of Immunology and Bacteriology, Western Infirmary, Glasgow, Scotland. Strain LS1, originally isolated from a mouse with arthritis, and extensively used in a murine model of septic arthritis [21], was kindly provided by Dr. A. Tarkowski, University of Göteborg, Sweden. S. aureus AIRD strain was obtained from a patient with septic arthritis in Glasgow Royal Infirmary, Glasgow, Scotland.

2.2 Clinical isolates

A total of 42 clinical S. aureus isolates were collected between 1998 and 2000 at three Havana hospitals. Nosocomial isolates came from ICU patients at the Centre for Medical-Surgical Research (10 BS isolates) and the Paediatric Hospital ‘Juan Manuel Márquez’ (eight SST isolates), and community isolates came from out-patients at the University Hospital ‘Gral. Calixto García’ (24 CJ isolates).

2.3 Reagents

Human colostral lactoferrin was purified from pooled fresh human colostrum by cation exchange chromatography using S-Sepharose FF (Pharmacia, Sweden) [22]. The N-terminus of the purified protein was intact, as determined by N-terminal analytical Mono-S HR HPLC [22]. The iron chelator ethylenediamine-di(o-hydroxyphenylacetic acid) (EDDHA) and Chelex-100 resin were obtained from Sigma (St. Louis, MO, USA). Iron-nitrilotriacetate (FeNTA) was prepared by mixing a freshly prepared solution of FeCl3 with Na-nitrilotriacetate in a 1:4 molar ratio, as described [22].

2.4 Antibiotic resistance

Clinical isolates and laboratory strains were tested against a panel of 19 antibiotics (penicillin, methicillin, cefuroxime, erythromycin, clindamycin, streptomycin, trimetropin, sulfamethoxazole, rifampicin, fusidic acid, ciproflaxin, chloramphenicol, kanamycin, tobramycin, gentamicin, tetracycline, amikacin, mupirocin, and vancomycin) by the Kirby-Bauer disk diffusion method according to the National Committee for Clinical Laboratory Standards (NCCLS) guidelines [23].

2.5 Culture media

Chemically defined medium (CDM) treated with Chelex-100 to remove traces of iron was prepared as originally reported by Kadurugamuwa et al. [24] and subsequently modified by us [22]. Normal human plasma serum (NHPS) was obtained by a special procedure involving immediate separation of blood cells by serial centrifugation at low temperatures of freshly collected blood [25] to minimise release of staphylocidal platelet factors during clotting [26], which can interfere with antimicrobial assays. Sterile disposable plasticware (Corning) was used to prevent iron contamination.

2.6 Culture conditions and measurement of growth

All isolates and strains were kept freeze-dried in 12% skim milk (Oxoid) at 4°C. To reduce stored iron levels, they were streaked on BHI agar (Oxoid) plates supplemented with 800 µM EDDHA. After three successive passages (12 h at 37°C), a colony was inoculated into 1 ml CDM and cultured overnight in a shaker incubator (G76D New Brunswick Scientific) at 150 rpm, 37°C. This culture was harvested (6000×g, 10 min, 4°C), washed with CDM, suspended in 1 ml fresh CDM and used to inoculate CDM and NHPS cultures as described below.

CDM (2 ml) was inoculated with sufficient bacteria of each strain to give an initial optical density of 0.1 at 470 nm. Growth was monitored by optical density measurements of samples taken every hour during a 6 h incubation at 150 rpm, 37°C, as above.

For NHPS cultures, about 6×106 cfu ml−1 of each strain were inoculated and the serum incubated for 12 h at 37°C in a water bath. Growth was monitored by counting colonies of samples taken every 2 h, streaked on blood agar plates (Oxoid) and grown aerobically for 18 h at 37°C.

2.7 Mechanism of action of lactoferrin against laboratory S. aureus strains

The three laboratory strains were grown in CDM or NHPS as described above, supplemented with 1.25–20 µM lactoferrin. Lactoferrin-free cultures were used as controls. The bacteriostatic and ferrochelating-dependent activity of lactoferrin was tested by the capacity of iron supplementation (addition of 50 µM FeNTA) and iron deprivation (addition of 800 µM EDDHA) at the time of inoculation to restore or inhibit respectively the growth of S. aureus.

Lactoferrin-derived peptides were obtained by acid proteolysis at 120°C, a procedure previously shown to give rise to fragments with antibacterial activity [27]. After cooling to room temperature, the acid-treated lactoferrin (1.5–20 µM) was added to PBS, CDM or NHPS inoculated with 6×106 cfu ml−1 of each laboratory strain, and incubated for 12 h at 37°C in a water bath. Growth was monitored by colony counting as above. A higher inoculum was used for CDM than in the experiments described above in order to demonstrate bactericidal as opposed to bacteriostatic activity.

2.8 Lactoferrin activity against clinical S. aureus isolates

This was assessed by monitoring growth in CDM, as described for the laboratory strains. Susceptibility to lactoferrin was defined as the ability of 20 µM lactoferrin to inhibit bacterial growth by at least 50% after 10 h of incubation as determined by optical density measurements at 470 nm, compared with the same isolate grown in lactoferrin-free CDM.

3 Results

3.1 Antibiotic resistance

Although all isolates were fully sensitive to four of the antibiotics (clindamycin, fusidic acid, rifampicin, and vancomycin) some nosocomial isolates showed decreased methicillin susceptibility, and one BS and one SST isolate were resistant to methicillin and mupirocin, respectively. In each case the mecA and mupA genes, which are associated with resistance to these antibiotics, were detected by PCR (data not shown). Overall, 93% (39/42) of isolates were resistant to penicillin (Table 1), suggesting that this antibiotic, commonly used in local health facilities, has poor therapeutic value. Nosocomial isolates showed both higher antibiotic resistance and multidrug resistance (defined as resistance to more than one antibiotic) compared to community CJ isolates (Table 1). Thus, 87% (7/8) of SST and 80% (8/10) of BS isolates were multiresistant with some isolates being resistant to up to 12 antibiotics, whereas only 42% (10/24) of CJ isolates were resistant to two antibiotics, and 50% (12/24) of them to only one antibiotic, usually penicillin.

View this table:
Table 1

Resistance of clinical S. aureus isolates to a panel of 19 antibiotics

AntibioticsPercentage of antibiotic-resistant clinical isolates
Nosocomial infectionsCommunity-acquired infections
Bloodstream (n=10)Skin/soft tissues (n=8)Conjuntivitis (n=24)
Amikacin0130
Cefuroxime100130
Ciproflaxin601313
Chloramphenicol0130
Erythromycin40508
Gentamicin0380
Kanamycin0500
Methicillin1000
Mupirocin0130
Penicillin10010088
Rifampicin000
Streptomycin0130
Sulfamethoxazole20504
Tetracycline806313
Trimetropin0384
Tobramycin0500

As expected the laboratory strains showed a relatively lower antibiotic resistance, being susceptible to 13 antibiotics. S. aureus Cowan I and LS1 strains were resistant to penicillin, tetracycline and mupirocin (mupA+), with Cowan I additionally resistant to kanamycin/tobramycin, and LS1 to ciprofloxacin. The AIRD strain was resistant only to penicillin, so only the first two laboratory strains can be classified as multidrug resistant.

3.2 Activity of lactoferrin against laboratory S. aureus strains

Addition of lactoferrin (1.25–20 µM) to CDM caused a similar dose-dependent retardation of bacterial growth of strains Cowan I, LS1 and AIRD, maximal inhibition occurring at concentrations above 5 µM (Fig. 1). Inhibition was reversed by addition of 50 µM FeNTA, and could be mimicked by addition of the iron chelator EDDHA (800 µM) to lactoferrin-free cultures. The inhibitory effect of lactoferrin was dependent on the time of addition, and was lost if lactoferrin was added 3 h or more after inoculation of bacteria (Fig. 2). In contrast, addition of iron to cultures whose growth was inhibited by lactoferrin caused an immediate increase of bacterial growth irrespective of the length of exposure to lactoferrin.

Figure 1

Growth of S. aureus Cowan I strain in CDM, supplemented with: control, medium alone (♦); 5 µM lactoferrin (●); 800 µM EDDHA (○), 50 µM FeNTA (□), and 50 µM FeNTA plus 5 µM lactoferrin (|). Points are individual readings from a representative experiment, which was performed at least three times.

Figure 2

Growth of S. aureus Cowan I strain in CDM, supplemented with: control, medium alone (♦); 5 µM lactoferrin added at 0 h (●) and at 3 h (○), and 50 µM FeNTA added at 0 h (1) and at 3 h (□) to CDM supplemented with 5 µM lactoferrin at 0 h. Points are individual readings from a representative experiment, which was performed at least three times.

The Cowan I strain grew slowly in NHPS, but lactoferrin concentrations above 5 µM caused a 3–4 log reduction in viable count after 10 h of incubation (Fig. 3). Addition of 50 µM FeNTA to NHPS reversed the reduction of viable counts observed in the presence of lactoferrin, but addition of FeNTA to lactoferrin-free NHPS did not greatly increase growth of S. aureus, suggesting that the endogenous transferrin in the serum did not inhibit growth of S. aureus under these conditions. The AIRD strain showed a similar response to Cowan I, but the LS1 strain grew even more slowly and inhibition was not observed until after 36 h of growth (data not shown).

Figure 3

Growth of S. aureus Cowan I strain in NHPS, supplemented with: control, medium alone (♦); 5 µM lactoferrin (●); 800 µM EDDHA (○), 50 µM FeNTA (□), and 50 µM FeNTA plus 5 µM lactoferrin (1). Points are individual readings from a representative experiment, which was performed at least three times.

Lactoferrin-derived peptides showed a strong staphylocidal activity after 2 h of incubation in PBS, but this effect was not observed when they were added to growth-supportive media such as CDM or NHPS (Fig. 4).

Figure 4

Effect of lactoferrin-derived peptides on growth of S. aureus Cowan I. Bacteria were inoculated into NHPS (circles), CDM (triangles) and PBS (squares). Solid symbols represent cultures supplemented with 20 µM lactoferrin-derived peptides obtained by acid proteolysis, and empty symbols represent non-supplemented control cultures. Points are individual readings from a representative experiment, which was performed at least three times.

3.3 Activity of lactoferrin against clinical S. aureus isolates

The inhibitory effect of lactoferrin varied between clinical isolates from different sites of infection. However, compared with the laboratory strains they showed higher overall lactoferrin resistance, requiring 20 µM lactoferrin to reliably demonstrate inhibition of lactoferrin-sensitive isolates, a concentration 4 times higher than that needed to inhibit the laboratory strains under similar conditions. BS isolates showed the highest resistance (50% of isolates) to inhibition by 20 µM lactoferrin, compared with 33% for CJ isolates and only 13% for SST isolates (Table 2). Lactoferrin-sensitive isolates were further divided into two groups: isolates whose growth was inhibited by 5 µM lactoferrin as for the laboratory strains were classed as low resistance, whereas those requiring 20 µM lactoferrin to show a similar effect were classed as high resistance (Fig. 5). BS isolates contained the lowest proportion of low resistance isolates (20%), whereas SST isolates showed the highest proportion (63%). The lactoferrin-sensitive CJ isolates were equally distributed into low (38%) and high (30%) resistance groups. When grouped into nosocomial and community isolates, both groups showed the same level of resistance: 33%. Thus under our conditions lactoferrin resistance was not related to local selective antibiotic pressure.

View this table:
Table 2

Lactoferrin and multidrug resistance of S. aureus isolates collected from different types of infections

Percentage of resistant isolates
Nosocomial infectionsCommunity-acquired infections
Bloodstream (n=10)Skin/soft tissues (n=8)Conjunctivitis (n=24)
Multidrug808842
Lactoferrin501333
Figure 5

Lactoferrin resistance of S. aureus clinical isolates varied according to the site of infection. The growth of low resistance isolates (empty bars) was inhibited by at least 50% by addition of 5 µM lactoferrin, whereas high resistance isolates (striped bars) required 20 µM lactoferrin, and insensitive isolates (solid bars) were not inhibited at any of the concentrations tested (up to 40 µM lactoferrin).

4 Discussion

Lactoferrin exerted a strong growth inhibitory effect against the laboratory S. aureus strains used in this study. The antibacterial activity in low iron CDM was essentially bacteriostatic and dependent on its ferrochelating properties, since iron supplementation or deprivation respectively inhibited or mimicked the effects of lactoferrin. The immediate resumption of bacterial growth after iron addition, irrespective of the length of exposure to lactoferrin, also points to an iron-dependent bacteriostatic effect. The inability of lactoferrin to inhibit S. aureus once growth had commenced suggests that initial iron deprivation, due to the reduction of intracellular iron storage by serial passages in EDDHA-deferrated BHI plates, and the low iron content of the medium induce bacterial iron uptake mechanisms, such as siderophore secretion or expression of lactoferrin receptors, thus restoring microbial growth. Since S. aureus is able to grow at extremely low iron concentrations (0.04 µM) [28], lactoferrin must be able to reduce levels of available iron to below this concentration in order to exert antistaphylococcal activity.

Although the effect of lactoferrin on S. aureus growth in CDM was clearly bacteriostatic rather than bactericidal, in NHPS addition of lactoferrin caused an initial iron-reversible reduction in colony count. While this might suggest that lactoferrin exerted a bactericidal rather than a bacteriostatic effect, it seems more likely that the effect is due to the combined activity of iron deprivation by lactoferrin and the bactericidal activity of traces of staphylocidal platelet-derived proteins in the serum. Inocula lower than about 106 cfu ml−1 were killed by NHPS (data not shown), indicating the presence of residual staphylocidal activity.

Although transferrin is present at high concentrations (50 µM) in NHPS, it was, unlike 5 µM lactoferrin, apparently unable to inhibit the growth of laboratory S. aureus strains, as iron supplementation of NHPS caused only a modest increase in growth of S. aureus. Thus the antimicrobial effect of human transferrin against laboratory S. aureus strains grown in NHPS was not very effective, and certainly weaker than that of lactoferrin. The ability of lactoferrin to exert an antimicrobial effect against S. aureus in NHPS suggests that it may also exert such an effect in vivo.

The capacity of S. aureus to efficiently counteract the bactericidal mechanisms exerted by lactoferrin-derived peptides, provided culture conditions allow bacterial growth, suggests that S. aureus can repair cell wall damage that may be caused by these peptides in vivo, if conditions present at the site of infection allow bacterial growth. This conclusion agrees with recent work on the effect of lactoferricin B, the N-terminal peptide derived from bovine lactoferrin, on S. aureus, Escherichia coli and Proteus mirabilis, which shows that the capacity of bacterial metabolism to repair damage to the cell envelope is the limiting step for the bactericidal activity [17]. Sensitivity to lactoferricin B increased if bacterial metabolism was slowed by incubation of bacteria at low temperature (20°C or 4°). These results suggest that lactoferrin-derived peptides are unlikely to act under complex in vivo conditions, although in vivo antimicrobial activity in a mouse model has been reported [16]. However, it is also possible that the peptides act indirectly in vivo, for example by upregulating immune responses such as the phagocytic activity of neutrophils [29].

The lactoferrin concentration required to inhibit growth of the laboratory strains (5 µM) was 4-fold lower than that required to reliably exert a similar effect against lactoferrin-sensitive clinical isolates (20 µM). This corroborates a previous report of lower siderophore production by laboratory strains [20], in which plasmid genes responsible for inducible siderophore secretion may be lost, leaving only a lower constitutive siderophore expression controlled by chromosomal genes [9,10].

The combination of reducing conditions, low pH and negative oxidation-reduction potential (Eh) values in bacterial pyogenic infections such as conjunctivitis and skin and soft tissue infections can reduce the ferrochelating properties of lactoferrin [8], which may explain the lower lactoferrin resistance shown by CJ (33%) and SST (13%) isolates compared with BS isolates. Furthermore, during inflammation the increased plasma concentration of lactoferrin and the presence of other IBPs, all with direct access to blood-borne pathogens, together with systemic hypoferraemia induced by inflammatory cytokines such as IL-1 and TNF-α may further induce lactoferrin resistance in the BS isolates.

The higher resistance of CJ isolates compared to SST isolates may be explained by two factors. Firstly, poorly vascularised inflammatory foci formed during SST infections could prevent lactoferrin from reaching the site of infection, whereas the eye conjunctiva is continuously irrigated by lachrymal fluids containing lactoferrin [30]. Secondly, iron availability increases during local tissue damage as result of release of intracellular iron, which may be greater during SST infections than during typically mild CJ infections. The higher resistance to lactoferrin shown by BS isolates (50%) compared to CJ and SST isolates may be explained by the ability of systemic homeostatic mechanisms to limit the development of highly reducing conditions which may limit the bacteriostatic effect of lactoferrin ferrochelating activity during pyogenic infections as opposed to bloodstream infections.

Antibiotic resistance was higher among nosocomial (BS and SST) isolates than community (CJ) isolates, corroborating a previous observation showing a direct and causal correlation between antibiotic resistance and selective antibiotic pressure against bacterial pathogens. The ability of sub-inhibitory concentrations of amikacin to increase siderophore production by oxacillin- and amikacin-resistant S. aureus isolates [10,20] suggests potential links between antibiotic resistance and lactoferrin resistance. However, our results did not provide any evidence of such a correlation, although oxacillin resistance was not directly tested in our work, and most of our isolates were not methicillin resistant, a trait commonly associated with susceptibility to oxacillin. Only one SST isolate was resistant to amikacin. Other combinations of antibiotics have not been found to have such an effect. We did not test directly the effect of antibiotics on lactoferrin activity, and to our knowledge the effect of antibiotics on siderophore production has not been linked to the capacity of S. aureus to obtain iron from either lactoferrin or other IBPs. Overall we conclude that among the set of isolates tested under our conditions no correlation was found between antibiotic and lactoferrin resistance, which agrees with the essential differences in their antimicrobial mechanisms. It may be possible to use lactoferrin as an alternative antimicrobial agent against S. aureus, acting on molecular targets not associated with antibiotic resistance.

Overall, these results suggest that lactoferrin has potential as a therapeutic agent against S. aureus isolates irrespective of their antibiotic resistance, and that lactoferrin is more effective than endogenous transferrin. Therapeutic approaches based on the use of ferrochelating agents, such as lactoferrin, combined with other antimicrobial drugs, and research into the molecular mechanisms underlying the antibacterial activity of lactoferrin may help to improve the alarmingly low efficacy of current antibiotics.

Acknowledgements

This research was supported by the MacFeat Bequest, Scottish MRSA Reference Laboratory and Finlay Institute Vice-presidency for Research. A.A. was supported by grants from the Royal Society, London. Bloodstream isolates were kindly provided by Dr. Margot Martínez, Centre for Medical-Surgical Research, Siboney, Havana. Technical assistance of Inaldis Chappotín and Hubert J. Ramirez is gratefully acknowledged.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
View Abstract