Lactoferrin and Bacteria

Lactoferrin has long been recognized for its antimicrobial properties, initially attributed primarily to iron sequestration. It has since become apparent that interaction between the host and bacteria is modulated by a complex series of interactions between lactoferrin and bacteria, lactoferrin and bacterial products, and lactoferrin and host cells.

Perspectives on interactions between lactoferrin and bacteria

Lactoferrin has long been recognized for its antimicrobial properties, initially attributed primarily to iron sequestration. It has since become apparent that interaction between the host and bacteria is modulated by a complex series of interactions between lactoferrin and bacteria, lactoferrin and bacterial products, and lactoferrin and host cells. The primary focus of this review is the interaction between lactoferrin and bacteria, but interactions with the lactoferrin-derived cationic peptide lactoferricin will also be discussed. We will summarize what is currently known about the interaction between lactoferrin (or lactoferricin) and surface or secreted bacterial components, comment on the potential physiological relevance of the findings, and identify key questions that remain unanswered.

Physiological production of lactoferrin

Lactoferrin is an iron-binding glycoprotein produced in mammals that is generally considered to be a significant component of the host innate immune response. It is predominantly found in glandular secretions, on mucosal surfaces, and at sites of infection. Lactoferrin is synthesized by glandular epithelial cells and secreted into the mucosal fluids on mucosal surfaces (Ward et al. 2005). Lactoferrin is also a major component of the secretory granules of neutrophils, and is released at significant levels in response to inflammatory signals. The production of lactoferrin appears to be both constitutive and differentially regulated in a tissue specific manner (Teng 2002).

Several studies have demonstrated the lactoferrin gene response to developmental cues during the differentiation of myeloid cells and to estrogen in uterine tissue (Teng et al. 1989, 2002). In spite of our understanding of lactoferrin production by neutrophils and specific glandular tissue, many questions remain regarding the production of lactoferrin on various mucosal surfaces. A recent study demonstrated that siderocalin, a normal constituent of neutrophil secretions, is synthesized and secreted by nasal mucosal cells in mice upon colonization with the human pathogens Streptococcus pneumoniae or Haemophilus influenzae (Nelson et al. 2005). By virtue of its capacity to bind complexes of iron and various bacterial siderophores (Holmes et al. 2005), siderocalin restricts the ability of various bacterial species to replicate in the host and to cause disease (Flo et al. 2004). Clearly, lactoferrin is similar to lipocalin in its role in iron sequestration and innate immunity, and lactoferrin is colocalized with siderocalin in human neutrophils (Kjeldsen et al. 1994). The obvious question of whether mucosal epithelial cells produce lactoferrin in response to signals from various bacteria has not yet been answered. This has implications regarding the quantity and source of lactoferrin on mucosal surfaces and will be important to consider when attempting to assess the role of various lactoferrin-bacterial interactions in vivo.

Lactoferrin is present in milk at concentrations of up to 7 mg/mL, and levels on mucosal surfaces are reported to be as high as 2 mg/mL (Masson and Heremans 1971). However, it is difficult to determine the actual or effective concentrations of lactoferrin on mucosal surfaces, and it is likely that levels of lactoferrin vary substantially. Microbes that colonize mucosal surfaces in the upper respiratory tract, the gastrointestinal tract, and the genitourinary tract will likely be exposed to different concentrations of lactoferrin, to different complexes of lactoferrin and other proteins, and to different levels of lactoferrin derivatives, such as the bactericidal peptide lactoferricin. Thus, it is important to consider the ecological niche that various bacterial species normally inhabit when studying their interactions with lactoferrin. The majority of studies on the interactions between lactoferrin and bacteria have been conducted with bacterial pathogens, and inferences are often made that the interaction is unique to the pathogenic bacteria. However, the distinction between pathogenic and commensal bacteria tends to be overemphasized because many pathogenic species, such as Neisseria meningitidis, H. influenzae, and S. pneumoniae, are highly adapted to live in the mucosal environment of the host and only sporadically cause disease. In this paper, we will review various reported interactions between lactoferrin or lactoferrin byproducts and bacteria, and relate them to the niche in which the bacteria are predominantly found.

Bacterial lactoferrin-binding proteins

Lactoferrin is an 80 kDa iron-binding glycoprotein, consisting of 2 lobes that possess essentially identical structural folds and iron-binding sites (Anderson et al. 1987; Baker et al. 2002). Lactoferrins from different host species vary in amino-acid sequence and in the location of the N-linked oligosaccharide side chains, but commonly possess a relatively high isoelectric point with positively charged amino acids clustered in the N-terminal region of the N-terminal lobe. Because of its net positive charge at physiological pH, lactoferrin has a tendency to interact with negatively charged bacterial surface components, which can complicate the interpretation of findings from binding studies. Thus, a simple demonstration of lactoferrin binding does not necessarily indicate functional significance. Because the cationic properties are shared by lactoferrins from various host species, if binding activity is limited to lactoferrin from a specific host, it suggests a greater likelihood of functional significance. However, the most direct evidence of functional significance is the demonstrated loss of function in a strain with a knockout of the gene encoding the protein that lactoferrin interacts with.

Clearly, this can only be performed for known functions, such as iron acquisition, and the significance of many reported findings might only become apparent when we have a better overall appreciation of the variety of functions that can be mediated by this interaction. Previous studies have identified lactoferrin-binding proteins in Bordetella pertussis (Redhead et al. 1987), Mycobacterium pneumoniae (Tryon and Baseman 1987), and Treponema spp. (Staggs et al. 1994), but the functional role of these proteins has not been determined. More recently, a 70 kDa lactoferrin-binding protein was identified in Helicobacter pylori (Dhaenens et al. 1997). Its specificity for human lactoferrin coupled with its loss of ability to use human lactoferrin as a source of iron for growth after protease treatments provided reasonably compelling preliminary evidence of a functional receptor. However, there have been no further reports on this protein and there are no obvious homologues of established lactoferrin-binding proteins in the H. pylori genome. There have been a number of reports of lactoferrin-binding proteins in Gram-positive bacteria, including bovine and human lactoferrin-binding proteins in Staphylococcus aureus (Naidu et al. 1990, 1992), but these preliminary reports have not been followed by more definitive studies.

More recently, a lactoferrin-binding protein was identified in Streptococcus uberis that is specific for bovine lactoferrin and, based on sequence analysis, belongs to the M protein family (Moshynskyy et al. 2003). However, a deletion mutant in the gene encoding the binding protein did not affect bacterial growth or adhesion to bovine mammarygland epithelial cells, so the function of the binding protein is still uncertain. The identification and characterization of lactoferrin receptors in members of the Neisseriaceae family have been more definitive and extensive. Lactoferrin-binding proteins were originally identified in N. meningitidis, Neisseria gonorrhoeae, Moraxella catarrhalis, and Moraxella bovis, based on specific binding of host lactoferrin and affinity isolation of the proteins under high pH and high salt conditions to minimize nonspecific interactions (Lee and Schryvers 1988; Schryvers and Morris 1988; Schryvers and Lee 1989; Yu and Schryvers 2002). The role of these proteins in iron acquisition from lactoferrin has been conclusively demonstrated by the inability of isogenic mutant strains, defective in the expression of the binding proteins, to grow when lactoferrin was provided as the sole source of iron (Bonnah and Schryvers 1998; Bonnah et al. 1999; Quinn et al. 1994).

Analogous to the bacterial transferrin receptors found in the Neisseriaceae and Pasteurellaceae families, the lactoferrin receptor is comprised of 2 distinct proteins (Fig. 1A), which are expressed from an iron-repressible operon. Lactoferrinbinding protein A (LbpA) is a homologue of TonB-dependant outer-membrane receptors, for which detailed structural information is available (Ferguson et al. 1998; van der Helm et al. 2002), and, thus, shares the overall structural features of a C-terminal beta-barrel filled with an N-terminal plug domain. However LbpA is approximately 20 kDa larger in size, primarily owing to the much more extensive predicted surface loops (Prinz et al. 1999), which cannot be effectively modeled using the available structural information. Studies with recombinant bovine-transferrin-human-lactoferrin hybrid proteins demonstrate that LbpA primarily recognizes regions in both domains of the C-terminal lobe of lactoferrin (Fig. 1A) (Wong and Schryvers 2003). It is hoped that a more detailed appreciation of the structure of LbpA and its interaction with lactoferrin will be attained through protein crystallography. The presence of the second receptor protein, lactoferrinbinding protein B (LbpB), was initially overlooked because of difficulties isolating it with the conventional affinity isolation protocol (Schryvers et al. 1998). Indeed, the properties and function of LbpB are largely inferred from comparison with its homologue, transferrin-binding protein B (TbpB). Studies with isogenic mutants defective in the expression of this protein do not provide any evidence of its role in iron acquisition (Bonnah and Schryvers 1998; Bonnah et al. 1999).

However, as demonstrated in studies with TbpB from the porcine pathogen Actinobacillus pleuropneumoniae (Baltes et al. 2002), failure to demonstrate a role in laboratory growth experiments does not preclude it from performing a critical role in vivo. The inherent bilobed nature of LbpB, revealed by weak internal amino-acid homology, has been experimentally confirmed for its homologue TbpB (Retzer et al. 1999). LbpBs have clusters of negatively charged amino acids near the C-terminal end of the C-terminal lobe that are unique and of unknown function, but that have an obvious potential for interaction with the cationic N-terminal region of lactoferrin (Fig. 1A). Laboratory experiments establishing the role of lactoferrin receptor proteins in the acquisition of iron from lactoferrin have been augmented by studies in a human gonococcal-infection model, demonstrating that they can also serve this functional role in the host (Anderson et al. 2003). In strains lacking a transferrin receptor, the presence of a lactoferrin receptor was sufficient to support the growth of gonococci and the initiation of an infection in human male volunteers. This confirms that there is a sufficient amount of iron contained in lactoferrin on the mucosal surface of the male urethra to support growth, and that potential iron sources, other than transferrin, were not sufficient in quantity to overcome the iron limitation for growth. Although the extent of growth and the ability to cause infection were even greater with a functional transferrin receptor, suggesting that transferrin is a more readily available iron source, the presence of the lactoferrin receptor provided a selective advantage.

The relatively frequent occurrence of clinical gonococcal isolates that are defective in the expression of the lactoferrin receptor proteins is not readily reconciled with the demonstrated selective advantage they confer, suggesting that there might be opposing selective forces in effect in alternate niches. Although there is no direct experimental evidence demonstrating that lactoferrin can serve as an iron source for growth on the nasopharyngeal mucosa, the invariant presence of functional lactoferrin receptors in clinical isolates of N. meningitidis and M. catarrhalis support the extrapolation of the findings from the gonococcal-infection model. However, lactoferrin receptors are not present in members of the Pasteurellaceae family, which occupy the same physiological niche in humans and other mammals (H. influenzae, Mannheimia haemolytica, and A. pleuropneumoniae). Clearly, lactoferrin is not a critical source of iron for these species. Although there is compelling evidence that lactoferrin receptors are required for iron acquisition from lactoferrin in the host for some bacterial species, it does not preclude them from serving additional roles important for survival in vivo.

Antimicrobial activities of lactoferrin

Early studies on the antibacterial effects of lactoferrin implicated iron sequestration as the primary bacteriostatic activity because iron-free lactoferrin was more effective at inhibiting bacterial growth than iron-loaded lactoferrin (Arnold et al. 1977; Kalmar and Arnold 1988). The observation that a supply of additional iron can overcome the bacteriostatic effect of iron-free lactoferrin, as shown for clinical isolates of S. aureus (Aguila et al. 2001), supports this conclusion. However, lactoferrin also clearly possesses bacterial killing properties. Studies comparing the bactericidal and bacteriostatic effects of lactoferrin on Streptococcus mutans have demonstrated that the irreversible inhibition of S. mutans in vitro is not attributed to iron deprivation (Arnold et al. 1982). The bactericidal activity was shown to be associated with an N-terminal peptide fragment, lactoferrin, which was released from lactoferrin by proteolysis and has since been intensively studied (Bellamy et al. 1992; Gifford et al. 2005). The cationic antimicrobial peptide released from lactoferrin has been compared with other antimicrobial peptides, such as human a- and b-defensins, suggesting that the mode of action is through their effect on the integrity of membranes (Brogden 2005). Although there is an increasing body of evidence demonstrating the bactericidal activity of lactoferricins, the degree to which they influence bacteria colonizing various mucosal surfaces has not been clearly established. Because bactericidal peptides can be generated by pepsin digestion of lactoferrin, significant oral intake of lactoferrin in the diet or possibly lactoferrin from oral and nasopharyngeal secretions are potential sources of lactoferricin on the gastrointestinal mucosa.

Thus, many of the original studies primarily demonstrated bactericidal activities against enteric pathogens that could be exposed to lactoferricins generated in the upper gastrointestinal tract. Because the bactericidal effects of lactoferricins are only evident in vitro in the absence of significant divalent cation concentrations, the composition of the diet can have a significant impact. In view of the potential for relatively frequent exposure of enteric bacteria, such as Escherichia coli and Salmonella typhimurium, to lactoferricin in vivo, one might expect the development of mechanisms of resistance, but none have been described to date in these enteric species. On other mucosal surfaces, such as the upper respiratory tract, it is not obvious how lactoferricin is generated in the absence of an inflammatory response. Perhaps this explains why few studies to date have examined the effects of lactoferricin on bacteria that colonize the nasopharynx, such as H. influenzae, M. catarrhalis, and N. meningitidis.

Recently, it has been shown that a surface protein on S. pneumoniae, PspA, is responsible for mediating resistance to lactoferricin (Shaper et al. 2004), and that a surface bacterial protease, PrtA, is responsible for the production of lactoferricin (Briles and Mizra 2006). An obvious question that arises is whether LbpB in N. meninigitidis or M. catarrhalis serves a role similar to that of PspA in S. pneumoniae, in view of the clusters of negatively charged amino acids in the C-terminal region of these proteins. The study with PrtA suggests one mechanism for the generation of lactoferricin on mucosal surfaces, and studies evaluating whether surface proteases from other microbes colonizing the mucosal surface result in lactoferricin production are warranted.

Alteration of bacterial surfaces by lactoferrin

In addition to having a direct effect on the viability and growth of bacteria, lactoferrin is capable of interfering with processes that involve bacterial surface components through a variety of mechanisms that are gradually being unearthed. Lactoferrin binds avidly to lipopolysaccharides (LPS) from Gram-negative bacteria (Elass-Rochard et al. 1995), and is capable of modulating the host response to LPS (Elass-Rochard et al. 1998). Lactoferrin binds to the lipid A portion of LPS, and this binding activity is significantly reduced in the presence of an outer membrane destabilizing cationic peptide such as polymyxin B (Appelmelk et al. 1994). Competitive-binding assays and site-directed mutagenesis experiments with human lactoferrin identified the N-terminal loop region as being essential for lactoferrin binding to LPS (Elass-Rochard et al. 1995). This region corresponds to the lactoferricin domain (Fig. 1D). Direct binding of lactoferrin to outer membranes was originally considered to be one possible mechanism for the observed loss in integrity of the outer membrane of enteric bacteria by lactoferrin (Ellison et al. 1988; Ellison 1994).

However, the potential contribution of sublethal concentrations of lactoferricin has to also be considered. Lactoferrin was shown to release the Hap adhesin and IgA protease from the surface of H. influenzae (Qiu et al. 1998), factors considered important for colonization of the mucosal surface. These proteins are members of the autotransporter protein family, and their release from the bacterial surface was shown to be due to cleavage of the anchoring peptide region by an intrinsic protease activity present in lactoferrin (Fig. 1E) (Hendrixson et al. 2003). The use of site-directed mutagenesis of the proposed residues responsible for the proteolytic activity provided compelling evidence that the protease activity was due to lactoferrin and not a contaminating protease in the recombinant lactoferrin preparation. Because members of the autotransporter family can also serve as adhesins for other nasopharyngeal bacteria, such as M. catarrhalis and N. meningitidis, lactoferrin could have a broader effect on the colonization of nasopharyngeal mucosa. However, this effect was only observed with relatively high concentrations of lactoferrin and required considerable exposure time, owing to the relatively weak proteolytic activity of lactoferrin.

Thus, the extent to which this inhibition of colonization occurs on mucosal surfaces is uncertain. Lactoferrin has been shown to inhibit the function of the type III secretory systems in enteropathogenic E. coli and Shigella flexneri (Gomez et al. 2003; Ochoa et al. 2003). These systems play a significant role in the adherence or invasion of mammalian cells. The effect of lactoferrin was attributed to the loss of the proteins involved in the formation of the secretory apparatus and of the effector proteins (Fig. 2A). Lactoferrin was shown to decrease the ability of S. flexneri to invade HeLa cells, probably by inducing the degradation of invasion plasmid antigens IpaB and IpaC, which are key proteins responsible for the bacteria-directed apoptosis and phagocytosis of mammalian cells (Gomez et al. 2003). In E. coli, the loss of effector proteins was inhibited by protease inhibitors and was shown to occur when incubating isolated preparations of the effector protein EspB and recombinant lactoferrin produced by Aspergillus (Ochoa et al. 2003). The obvious inference is that this effect is due to the intrinsic protease activity of lactoferrin, but this should be confirmed by studies with site-directed mutants of lactoferrin to exclude the possibility that it is due to contaminating proteases isolated from Aspergillus.

If these effects can be attributed to the intrinsic protease activity of lactoferrin, it would suggest that this property of lactoferrin affects a wide range of bacteria. Human lactoferrin was shown to inhibit the growth of Porphyromonas gingivalis when hemoglobin was the sole iron source by binding to and removing a hemoglobin receptor protein (HbR) from the bacterial surface (Shi et al. 2000). The HbR receptor is a 19 kDa domain of a larger multidomain protein encoded by the hagA gene. Processing by either of the 2 cysteine proteases, RgpA and Kgp, disrupts the covalent connection between the domains, which nevertheless remain associated. The removal of HbR by lactoferrin involves a negatively charged region of HbR, and is thought to occur when the interaction with the other domains is disrupted. The selective removal of HbR can also be accomplished with the addition of lactoferricin (Fig. 2B). It has not been established whether lactoferricin is generated by exposure of lactoferrin to P. gingivalis, which raises the question of whether the effect of intact lactoferrin is attributed to released lactoferricin. Nevertheless, these effects were observed at concentrations of lactoferrin that are found in the gingival crevicular fluids of patients with periodontal disease and, thus, the inhibition by lactoferrin might be physiologically relevant. Lactoferrin has also been shown to inhibit the formation of biofilms by Pseudomonas aeruginosa at concentrations well below those required to kill planktonic bacteria (Singh 2004).

The bacteria exposed to lactoferrin appeared to be able to attach to the surface and multiply, but they had increased twitching motility and failed to remain anchored at the original parental anchoring site. This effect was not observed when the bacteria were exposed to the iron-loaded form of lactoferrin, suggesting that iron sequestration is responsible for the impairment of biofilm formation. Unfortunately, the authors did not perform the experiments with transferrin or with the addition of excess exogenous iron to confirm this interpretation. In view of the increased susceptibility of the apo form of lactoferrin to proteolytic cleavage, the possibility that the impairment of biofilm formation was due to the generation and release of lactoferricin cannot be excluded.

Synopsis and future directions

There appears to be a complex interplay between lactoferrin and bacteria that modulates the ability of bacteria to successfully colonize mucosal surfaces or cause disease in the host. Although ongoing studies have begun to reveal some of the molecular details of these interactions, there is a need for further detailed characterization of the specific interactions, for a better appreciation of how widespread the interactions are among bacterial species, and for studies confirming their relevance in the host. The use of recombinant mutant and hybrid lactoferrins in conjunction with specific bacterial mutants has the potential to provide definitive detailed information on the specific interactions between lactoferrin and bacteria (Hendrixson et al. 2003; Wong and Schryvers 2003). The recombinant lactoferrins would be effective tools for probing the interaction of lactoferrin with other bacterial species, and efforts should be made to use these as reagents in various studies for comparative purposes. A number of the bacterial species under study are human-specific pathogens, making it difficult to provide direct compelling evidence for the relevance of the findings from in vitro studies. However, studies with comparable nonhuman pathogens might provide experimental evidence that can be extrapolated to human pathogens. For instance, studies with the porcine pathogen A. pleuropneumoniae provided evidence that TbpB is essential for survival (i.e., iron acquisition) in vivo (Baltes et al. 2002), a finding that almost certainly applies to human pathogens such as H. influenzae and N. meningitidis. A similar strategy could be considered to demonstrate the importance of specific lactoferrin-bacterial interactions by selecting appropriate nonhuman pathogens (e.g., M. bovis in lieu of M. catarrhalis).