|
|
|
|
|
|
Scientific
Publications - Work Done by Microbiology Reader Department of Clinical Veterinary Sciences, Faculty of Veterinary Medicine, University of Helsinki, Finland Role of Lactoferrin in Treatment of Bovine MastitisTaina Kutila ACADEMIC DISSERTATION To be presented with the permission of the Faculty of Veterinary Medicine, University of Helsinki, for public criticism in Auditorium Maximum, Hämeentie 57, Helsinki, on March 12th, 2004 at 12 noon. Helsinki 2004
Supervisors Professor Hannu Saloniemi, Department of Clinical Veterinary Sciences, Faculty of Veterinary Medicine, Finland Professor Satu Pyörälä, Saari Unit, Faculty of Veterinary Medicine, Finland Docent Liisa Kaartinen, National Agency for Medicines, Finland
Reviewers Honorary Senior Reserch Fellow Jeremy Brock, Department of Immunology, University of Glasgow, Scotland DVM, PhD, Professor Weihuan Fang, College of Animal Sciences/Institute of Preventive Veterinary, Medicine, Zhejiang University, Hangzhou, P.R. China
Opponent Professor Hannu Korhonen, MTT AgriFood Research, Finland
Chairman Professor Hannu Saloniemi, Department of Clinical Veterinary Sciences, Faculty of Veterinary Medicine, Finland
Vaikka sinulla olisi älyä joutuisit aina pulaan ellei sinulla ole sydäntä joka älyäsi ohjaa. Maria Jotuni Dedicated to the memory of my father
TABLE OF CONTENTS Page ABSTRACT 11 ABBREVIATIONS 13 LIST OF ORIGINAL PAPERS 15 1. INTRODUCTION 16 2. REVIEW OF THE LITERATURE 18 2.1 Structure and mechanisms of action of lactoferrin 18 2.2 Antimicrobial activity of lactoferrin 19 2.3 Lactoferrin concentrations in bovine milk and udder secretions 21 2.3.1 Lactating and non-lactating healthy cows 21 2.3.2 Mastitic cows 23 2.4 Lactoferrin alone or in combination with antimicrobial drugs in treatment of bovine mastitis 23 2.5 Treatment of coliform mastitis 25 3. AIMS OF THE STUDY 28 4. MATERIALS AND METHODS 29 4.1 Animals (I - IV) 29 4.2 Lactoferrin and citrate determinations (I, II) 29 4.3 Lactoferrin purification (II - IV) 30 4.4 Bacteria (II) 30 4.5 Analysis of bacterial growth by turbidometry (II) 31 4.6 Experimental designs (I - IV) 32 4.7 Clinical observations and collection of blood and milk samples (III, IV) 33 4.8 Determination of indicators of inflammation in milk (I - IV) 34 4.9 Other methods (IV) 34 4.10 Statistical methods (I - IV) 35 5. RESULTS 36 5.1 Lactoferrin and citrate concentrations in milk and dry cow 36 secretion (I) 5.2 Inhibitory activity of lactoferrin against udder pathogens (II) 36 5.3 Disposition kinetics of lactoferrin in lactating dairy cows (III) 37 5.4 Efficacy of lactoferrin in treatment of experimentally induced E. coli mastitis (IV) 39 6. DISCUSSION 42 7. CONCLUSIONS 49 ACKNOWLEDGEMENTS 50 REFERENCES 52
ABSTRACT The non-specific, multifunctional glycoprotein lactoferrin (Lf) is present in milk and external body secretions. It is released by the secondary granules of neutrophils and epithelial cells in high concentrations in response to inflammatory stimuli. Lf has a broad-spectrum antimicrobial activity, especially against coliform bacteria, such as Escherichia coli, which cause severe mastitis in dairy cows. Since treatment of severe coliform mastitis using antimicrobial agents is problematic, Lf as a natural protein may offer an alternative for treating this disease. The aim of this work was to evaluate the usefulness of exogenic Lf in mastitis treatment. First, normal concentrations of Lf and citrate in the early dry cow secretion of healthy cows were studied. Second, the antimicrobial efficacy of Lf against udder pathogens, particularly E. coli, was tested in vitro. Third, the disposition kinetics of Lf infused into the udder quarters of dairy cows was investigated. Finally, the antibacterial effect of Lf was studied in an experimental E. coli mastitis model. To determine the normal concentrations of Lf and citrate, milk and dry udder secretion samples were collected on the last day of lactation before the drying-off, and then 2 and 6 days later. The mean Lf concentration in the milk increased from 5.29 mg/ml on the last day of drying-off to 8.09 and 11.26 mg/ml 2 and 6 days later, respectively. Citrate concentration decreased from 1.85 mg/ml to 1.54 and 1.09 mg/ml, respectively. Concentrations varied greatly between cows and even between udder quarters at each time-point. Median molar ratio (citrate to native Lf) decreased gradually from 153 on the last day of lactation to 44 on day 6 after drying-off. The antibacterial effect of Lf was tested in vitro against E. coli, Staphylococcus aureus, and coagulase-negative staphylococci (CNS) as well as on Pseudomonas aeruginosa and Klebsiella pneumoniae, originally isolated from bovine mastitis. Concentrations of Lf used were 0.67, 1.67 and 2.67 mg/ml. The best inhibitory activity of Lf was seen against E. coli and P. aeruginosa. The inhibitory effect of Lf for E. coli was concentration-dependent, and variation between the five isolates of E. coli was small. None of the isolates was totally resistant to Lf. The growth of two isolates of P. aeruginosa was clearly inhibited in Iso-Sensitest Broth, in contrast to two K. pneumoniae isolates, which were virtually unaffected. The isolates of CNS and S. aureus showed more variation in susceptibility to Lf than E. coli. Three isolates of S. aureus were more susceptible to Lf than the two other isolates at 0.67 mg/ml of Lf. The growth of four CNS isolates was somewhat inhibited by Lf, but one isolate was totally resistant. The disposition kinetics of Lf after intramammary administration was studied in lactating dairy cows. A 1-gram dose of Lf produced elevated Lf concentrations in milk for several hours. The mean elimination half-life of total Lf was 2.2 h, and the mean maximum concentration of 6.3 mg/ml was reached between 1 and 4 h post-infusion. The average Lf concentration in milk of six cows before administration of Lf was 0.4 mg/ml. After 8 h of administration, the average Lf concentration decreased to 0.8 mg/ml. Lf caused some local tissue irritation in the udder quarters, but general signs, such as fever and anorexia, were not observed. The udder quarters of primiparous cows seemed to react faster than those of multiparous cows. The irritation reactions decreased more rapidly in older than in primiparous cows. In an experimental E. coli mastitis model, the clinical response to the challenge varied markedly among individual cows. In general, systemic signs disappeared within 2-3 days, and local signs within one week. Differences in systemic and local clinical signs between Lf- and enrofloxacin-treated cows were not significant. Bacterial counts in milk decreased faster in enrofloxacin- than Lftreated cows, the difference almost reaching significance. Bacteria were eliminated from the challenged quarters of the enrofloxacin-treated cows within 3.8 days and Lf-treated cows within 5.8 days on average. Differences in somatic cell counts between treatments were not significant. However, NAGase activity remained high for a longer time in milk of cows treated with Lf than with enrofloxacin. Daily and quarter milk yield profiles over the experimental period were similar in both treatment groups.
ABBREVIATIONS aa Amino acid Apo-Lf Iron-free lactoferrin CFU Colony-forming units CMT California Mastitis Test CNS Coagulase-negative staphylococci DELFIA Dissociation-enhanced lanthanide fluoroimmunoassay E. coli Escherichia coli IL-1 Interleukin-1 IL-2 Interleukin-2 IL-8 Interleukin-8 ISB Iso Sensitest Broth i.v. Intravenous K. pneumoniae Klebsiella pneumoniae LAL Limulus amebocyte lysate Lf Lactoferrin LPS Lipopolysaccharide MIC Minimum inhibitory concentration NAGase N-acetyl-ß-D-glucosaminidase 14 P. aeruginosa Pseudomonas aeruginosa PC Post-challenge PMN Polymorphonuclear neutrophils S. aureus Staphylococcus aureus s.c. Subcutaneous SCC Somatic cell count Str. agalactiae Streptococcus agalactiae Str. dysgalactiae Streptococcus dysgalactiae Str. uberis Streptococcus uberis TNF- Tumour necrosis factor alpha
LIST OF ORIGINAL PAPERS This thesis is based on the following original papers referred to in the text by Roman numerals I-IV: I II III IV 1. INTRODUCTION Lactoferrin (Lf) is a component of the natural protection systems of humans and animals. Non-specific, multifunctional Lf is present in milk and such external secretions of the body as saliva, bile, tears and sperm. It is released by secondary granules of neutrophils and epithelial cells in high concentrations in response to inflammatory stimuli (Baggiolini et al. 1970, Harmon & Newbould 1980, Brock 1995, Lönnerdal & Iyer 1995, Nuijens et al. 1996, Aguila & Brock 2001, Plaffl et al. 2003). Lf, the red protein, was first identified in 1939 in milk whey (Soerensen & Soerensen 1939). Since the 1960s, scientists have studied its structure, functions and potential applications. Today, scientists are still searching for new ways to use Lf in the treatment of bacterial, viral and fungal infections, sepsis, cancer and tumours and immunosupressory illnesses both in human and veterinary medicine. Lf also has an application as a supplement in functional foods for humans and animals like calves (Joslin et al. 2002, Roblee et al. 2003). In 1995, Lohuis and co-workers suggested that Lf might have potential in the treatment of bovine mastitis. Four years later, the Department of Process and Environmental Engineering at the University of Oulu, Finland, started to develop the absorption chromatography method for large-scale, reasonably priced Lf purification from cheese whey. Production of bovine Lf had until this point been expensive. In autumn 1999, our group began to study the potential role of Lf in the treatment of bovine mastitis. In economic terms, mastitis is generally considered to be the most serious disease of dairy cows. Practical measures for controlling intramammary infections caused by Staphylococcus aureus, Streptococcus agalactiae and, to a lesser extent, other streptococci have been developed and are widely applied (Erskine et al. 1987, Lam et al. 1997). These measures are less effective against mastitis, which is caused by Gram-negative bacteria belonging to the normal flora of the gastrointestinal tract. Teat dipping and dry-cow therapies do not help in the control of coliform mastitis, and clinical coliform mastitis remains a problem. The trends towards loose housing, increased herd size and crowding of cows tend to increase the coliform load in the environment as well as the likelihood of infections (Hogan et al. 1989, Schukken et al. 1990, Lam et al. 1997, Peeler et al. 2002). The treatment of coliform mastitis is problematic because signs are due to the response of bacterial endotoxin, but the effect of antimicrobial treatment is questionable (Jones & Ward 1990, Pyörälä et al. 1994). In addition, the use of broad-spectrum antimicrobial drugs, such as fluoroquinolones, increases selection pressure for development of resistant bacterial strains. Thus, antimicrobial therapy is a dual-edged sword, and new treatment approaches are needed. We decided to concentrate on Lf alone, without combining it with antimicrobial drugs. There were several reasons for taking this approach. Public concern over food safety is increasing, and people are more aware of the risk of using antimicrobial drugs. Political and regulatory pressure against the wide use of broad-spectrum antimicrobial drugs for dairy cattle has also increased. Lf as a natural protein sidesteps these concerns. Before taking Lf on board as a therapeutic agent, preliminary studies should be undertaken. Improved knowledge of the physiological concentrations of Lf at drying-off is necessary prior to Lf being used as a dry-cow product. The in vitro efficacy of Lf against udder pathogens also requires additional studies to those available (Oram & Reiter 1968, Arnold et al. 1980, Reiter 1985, Lönnerdal & Iyer 1995). The disposition kinetic character of exogenic Lf in the udder remains unknown. The nature of Lf and its potential use in the treatment of mastitis, especially coliform mastitis, are discussed in the following review of the literature. The structure and general mechanisms of action of Lf are briefly reviewed, and the antibacterial properties are presented. Lf concentrations in milk of healthy and mastitic cows are described. Lf as an alternative to antimicrobial and preventive dry-cow therapies is discussed. Finally, a general review of severe Escherichia coli mastitis and problems associated with its treatment are provided.
2. REVIEW OF THE LITERATURE 2.1 Structure and mechanisms of action of lactoferrin Bovine Lf is a monomeric metal-binding glycoprotein synthesized as a 708-amino acid (aa) protein with a 19-aa signal peptide (Mead & Tweadie 1990, Pierce et al. 1991). Lf consists of two globular lobes, each of which contains one iron-binding site. These lobes, designated as the N- and C-lobes, represent the N- and Cterminal halves of the polypeptide (Moore et al. 1997). The two metal-binding sites of Lf that lie between the two domains of each lobe are highly similar to each other (Moore et al. 1997) and to the corresponding sites in human Lf (Anderson et al. 1989). Iron- binding occurs concomitantly with the binding of two bicarbonate anions that appear to play a prominent structural role (Anderson et al. 1989, Iyer & Lönnerdal 1993, Moore et al. 1997). Lf serves an important function in the mammalian host defence mechanism (Smith & Oliver 1981, Brock 1995, Nuijens et al. 1996) (Figure 1). It occurs in external secretions, the secondary granules of neutrophils and epithelial cells (Baggiolini et al. 1970, Harmon & Newbouls 1980, Brock 1995, Lönnerdal & Iyer 1995, Nuijens et al. 1996, Aguila & Brock 2001), and is released by these cells in higher concentrations in response to inflammatory stimuli. The bacteriostatic activity of Lf depends on factors affecting iron binding, including pH, certain ions and iron-sequestering mechanisms of bacteria. The effect of Lf could be eliminated by saturating it with iron (Bullen et al. 1972), and at lower pH values, the iron-binding capacity of Lf is lost (Mazurier & Spik 1980). Besides bacteriostatic and bactericidal activities, Lf is a potent activator and regulator of various immunological functions such as granulopoiesis (Sawatzki & Rich 1989), cytokine production, antibody synthesis in vitro, natural killer cell cytotoxicity, lymphocyte proliferation, complement activation and production of interleukin (IL) -1, IL-2 and TNF- (Sanchez et al. 1992, Brock 1995, Kimber et al. 2002) (Figure 1). Lf may also participate in antitumour defence (Tsuda et al. 2002) and has been shown to have antiviral (Yasumura et al. 2002), antifungal (Abe et al. 2002) and antiparasitic activities (Omata et al. 2001) (Figure 1). Baker and co-workers (2002) suggested that Lf is a moonlighting protein, i.e. a protein that has several distinct biological activities that manifest in different environments.
2.2 Antimicrobial activity of lactoferrin Lf has broad-spectrum antimicrobial activity in vitro as well as in vivo against E. coli, S. aureus, K..pneumoniae, Bacillus subtilis, Streptococcus mutans and Candida albicans, among others (Oram & Reiter 1968, Arnold et al. 1980, Reiter 1985, Zagulski et al. 1989, Lönnerdal & Iyer 1995, Bhimani et al. 1999). However, the bacteriostatic activity against udder pathogens Streptococcus uberis and Streptococcus agalactiae was unaffected (Rainard 1986, Fang et al. 2000). Fang and co-workers (1999, 2000) postulated that binding to Lf could be one of the mechanisms that virulent Str. uberis exploits to increase its pathogenicity in intramammary infections since adherence to mammary epithelial cells was greatly increased by addition of Lf in in vitro tests: Lf may act as a bridging molecule between bacteria and mammary epithelial cells during Str. uberis mastitis. However, the involvement of milk in Str. uberis adherence to bovine epithelial cells may be complicated by other milk components that affect bacterial interaction with epithelial cells (Fang et al. 1999). Immunomodulator Antiparasitic Antifungal Antiviral Antibacterial Activity Protease Protease Inhibitor Ribonuclease Procoagulant Cationic Peptides Autoantibodies Granulopoiesis Transcription factor Iron Absorbtion Hypoferraemia Anti-Inflammatory Antitumour Lactoferrin Figure 1. Proposed functions of lactoferrin (Brock 2002).
The mechanism by which Lf inhibits bacterial growth has not been fully elucidated. Its antimicrobial and anti-inflammatory effects are believed to be a result of the powerful iron-chelating ability that makes iron unavailable to bacteria (Reiter & Oram 1967, Weinberg 1978), thereby causing nutritional deprivation in susceptible organisms. This argument is supported by the findings that selected, highly virulent pathogens have evolved techniques to circumvent this effect, either by secreting high-affinity low-molecular-weight iron chelators (siderophores) that can compete with Lf, or by expressing Lf receptors, which are highly species-specific (Ellison 1994, Brock 1995, Lin et al. 1998). Siderophores are synthesized by many micro-organisms grown under iron-limited conditions. They bind iron in the extracellular environment, then the iron-siderophore complexes are actively bound by outer membrane receptors of bacteria, and the iron is internalized (Neilands 1984, Neilands et al. 1985). Gram-negative bacteria have as many as four high-affinity iron acquisition systems: aerobactin, enterobactin, citrate and ferrichrome. The enterobactin iron acquisition system has the highest affinity of all siderophores (Bagg & Neilands 1987. Lin et al. 1998, 1999); this iron acquisition system is composed of various enzymes that are required for the synthesis and secretion of the siderophore enterobactin (Neilands et al. 1985, Lin et al. 1998, 1999). The enterobactin receptor FebA participates in transporting and deferrating ferric enterobactin. S. aureus bacteria also use siderophores staphyloferrin A and B and aureochelin under iron-restricted growth conditions (Modum et al. 2000, Diarra et al. 2002b). S. aureus isolated from humans use transferrin, haemoglobin, ferroxamines and even Lf as sources of iron in iron-restricted environment (Triever & Courcol 1996). Evidence from a number of studies indicates that the antimicrobial activity of Lf is more complex than simple Fe chelation. Lf has direct bactericidal activity and can kill susceptible bacteria by a mechanism distinct from sequestering of Fe, which is mediated through peptides known as lactoferricins (Arnold et al. 1980, Dalmastri et al. 1988, Ellison et al. 1988, 1990, Brock 1995). The smaller size of lactoferricins renders access to target sites on microbial surfaces easier than it is for native Lf, which may explain the superior bactericidal effectiveness of lactoferricin as compared with native Lf. By comparing minimal inhibitory concentration (MIC) values on a molar basis, bovine lactoferricin was estimated to be about 12-fold as effective against E. coli as undigested bovine Lf (Bellamy et al. 1992). Bellamy and co-workers (1992) also found that the antimicrobial domain is near the N-terminus of Lf, comprising parts of the N-terminal residues 1-48 of bovine Lf (Hoek et al. 1997), in a region distinct from Lf's iron-binding sites. Apo-Lf (iron-free form of Lf) was shown to alter bacterial cell membrane function, or integrity and permeability, directly damaging the outer membrane of Gram-negative bacteria (Ellison et al. 1988, 1990, Bellamy et al. 1992, Yamauchi et al. 1993, Appelmelk et al. 1994), by binding to lipopolysaccharide endotoxin (LPS) and blocking its detrimental effects (Appelmelk et al. 1994). The bacterial killing happened through osmotic damage, and the destabilization of the bacterial cell wall is thought to occur through a non-specific binding of calcium and magnesium to Lf (Ellison et al. 1988). LPS release was blocked by iron saturation of Lf, and exogenous calcium and magnesium blocked the activity of Lf as well (Ellison et al. 1988, 1990). The bacteriostatic form of Lf is either apo-Lf or partly saturated Lf, while the fully Fe3+ -saturated form has no antimicrobial activity (Bishop et al. 1976, Arnold et al. 1982). Approximately 6-8% of Lf in milk is iron-saturated, which means that most of the Lf in milk is present in the apo-form (Goldsmith et al. 1982). Lf receptors have been identified on the surface of such bacteria as S. aureus and CNS (Levay & Viljoen 1995). Lf has been suggested to bind to Lf receptors and promote the activation of bovine complement via an alternative pathway, resulting in a decrease in the number of staphylococci in the udder after infusion of Lf (Kai et al. 2002b). 2.3 Lactoferrin concentrations in bovine milk and udder secretions 2.3.1 Lactating and non-lactating healthy cows Lf is present in the colostrum and normal milk of most mammals (Masson & Heremans 1971). Compared with other species, bovine milk and colostrum have relatively low concentrations of Lf; the concentration varies from 0.02 to 0.35 mg/ml in the milk of healthy lactating cows (Welty et al. 1976) (Table 1). Lf produced by the mammary gland may serve a dual role, protecting both the mammary gland and the neonatal intestine from infection. Growth and development of the gastrointestinal tract in newborn animals fed their dam's milk are known to be more rapid than in those fed commercial formula (Berseth et al. 1983, Heird et al. 1984, Widdowson 1985). Previous studies have shown that Lf is a beneficial supplement in the diet of neonatal calves prior to weaning. Lf alters the microbial populations in the gut of these non-ruminants and increases preweaning weight gain (Joslin et al. 2002, Robblee et al. 2003). Two to four days after cessation of regular milking, the concentration of Lf has been shown to increase dramatically (Table 1). The linear increase continues at a rate of about 1.15 mg/ml per day as a result of an increased net synthesis of Lf during the first 14 to 21 days of involution. After three to four weeks of involution, the maximum Lf concentration (about 20 mg/ml) is achieved in the udder secretion (Welty et al. 1976; Schanbacher et al. 1993). Lf concentrations have increased, over 100-fold those in normal milk (Schanbacher et al. 1993), and throughout the non-lactating period Lf concentrations remain high (Sordillo et al. 1987). Welty and co-workers (1976) also found that in some cows the concentration of Lf was less than 10 mg/ml after 10 days of involution, whereas much higher Lf concentrations in milk, 75 to 100 mg/ml (Table 1), were measured in other cows. The dramatic increase in Lf concentrations after cessation of milking appears to be a manifestation of the process of involution, while during normal lactation Lf is a minor milk protein (Masson & Heremans 1971). The interaction of Lf with other proteins such as immunoglobulin G or casein, in nonlactating mammary secretions may affect the antimicrobial properties of Lf (Wang & Hurley 1998); the fully involuted udder is very resistant to coliform infections, mostly due to the high Lf content of the secretion (Oliver & Bushe 1987). In the mammary gland, efficacy of Lf is negatively correlated with citrate content of milk (Bishop et al. 1976); citrate concentration varies greatly depending on the time of lactation. It is at its highest after parturition and during mid-lactation (2.38 mg/ml), and at its lowest in the dry period (about 0.50 mg/ml) (Nonnecke and Smith 1984a). Bicarbonate, which is required for the binding of iron by Lf, mole per mole, can overcome the effect of citrate especially during the dry-cow period, when citrate is absorbed from the udder secretion to blood, while bicarbonate diffuses from blood to the udder secretion (Bishop et al. 1976, Griffith & Humphrey 1977, Reiter 1985). The molar ratio of citrate to apo-Lf decreases with udder involution and increases with lactogenesis (Smith et al. 1971, Peaker & Linzell 1975). The molar ratio has been found to be more important in the inhibition of the growth of Gram-negative bacteria in vitro than the absolute concentration of either component (Bishop et al. 1976; Nonnecke & Smith 1984a). A citrate to apo-Lf ratio of 75 results in 50% growth inhibition of coliform bacteria, whereas ratios of 300 and greater result in less than 10% growth inhibition. In other words, the smaller the citrate to Lf molar ratio, the more effective the inhibition of growth (Bishop et al. 1976). These results suggest that the ratio of citrate to Lf is important in evaluating Lf as a non-specific protective factor in bovine mammary secretions. Based on information available, the citrate to Lf molar ratio would be more than 1000 in normal milk and colostrum but only around 10 in the secretion of the fully involuted bovine mammary gland. Nonnecke & Smith (1984a) found the mean molar ratio one week before the dry period to be as high as 1154, declining to 28 two weeks after the last milking. 2.3.2 Mastitic cows The mean concentration of Lf in the milk of cows with subclinical mastitis (0.2- 1.2 mg/ml) has been shown to be higher than in milk of normal cows (Hagiwara et al. 2003) (Table 1). In addition, Lf concentration in subclinical mastitis might depend on the pathogenicity of each bacterial species; the mean Lf concentrations in milk from quarters infected with S. aureus or streptococci were significantly higher than in milk from quarters infected with coagulase-negative staphylococci (CNS) and Corynebacterium bovis (Hagiwara et al. 2003). In cows with clinical mastitis, Lf concentrations in milk can range from 0.3 to 2.3 mg/ml (Table 1); these concentrations are generally higher than in normal cows or those with subclinical mastitis (Kawai et al. 1999, Hagiwara et al. 2003). Kawai and co-workers (1999) reported that with peracute mastitis by E. coli, the Lf concentration in milk was significantly lower than that from cows with acute mastitis. Mean Lf concentration in milk from cows with peracute E. coli mastitis was initially significantly lower than that from cows infected with environmental streptococci and CNS, but was elevated after 46 h from diagnosis of mastitis, being higher than in mastitis caused by other pathogens (Harmon et al. 1975, Kawai et al. 1999). The significant differences in Lf concentrations in milk from normal cows and cows with subclinical and clinical mastitis may be associated with the severity of inflammation or may depend on the virulence of the bacteria involved. Kawai and co-workers (1999) also speculated that the lower Lf concentration in quarters infected with E. coli might be insufficient to inhibit bacterial growth because multiplication of E. coli in an inflamed udder quarter is fast, while Lf concentration increases slowly in the early phase of inflammation. 2.4 Lactoferrin alone or in combination with antimicrobial drugs in the treatment of bovine mastitis Researchers have since the 1960s suggested that Lf with its broad-spectrum antimicrobial effect would be a good candidate for a non-antibiotic treatment of infections (Masson et al. 1966). Lohuis and co-workers (1995) and Diarra and coworkers (2002a) later speculated that exogenous Lf infusion could potentially be useful in the treatment of bovine mastitis, and could partly replace the use of antimicrobials. Another application could be the prevention of mastitis at dryingoff (Kai et al. 2002b). Nickerson (1989) and Oliver & Sordillo (1989) pointed out that the early period of mammary involution coincides with a period of increased susceptibility to intramammary infection; endogenous Lf concentrations are low during this time. Clinical mastitis is, however, rare in the middle of the nonlactating period (Nonnecke & Smith 1984a), when the concentration of Lf in mammary gland secretion is at its highest (Welty et al. 1976). Kai and co-workers (2002b) suggested the same as speculated here: administration of exogenous Lf during early involution could help limit bacterial growth, not only for its bacteriostatic and bactericidal effect, but also for the priming effects on the innate immunity of the host. In addition, a high concentration of Lf promotes phagocytic activity in the mammary gland and the activation of bovine complement (Kai et al. 2002a). Lf could affect the disposition of apoptotic cells and bacteria during the early non-lactating period, and exogenous Lf could thus be used as a dry-cow therapy. Pyörälä & Pyörälä (1998) investigated the efficacy of antimicrobial agents in the treatment of clinical mastitis during lactation in a large retrospective study. They showed that S. aureus infection is rather resistant to antimicrobial treatment, especially when the isolates are penicillin-resistant and infections occur in older cows. Diarra and co-workers (2002a) demonstrated a synergistic effect between Lf and penicillin against three isolates of S. aureus from bovine mastitis, while Lf alone showed a weak inhibitory activity. The ultrastructural alterations caused by Lf to the bacteria enhance the activity of some antimicrobial agents (Sanchez & Watts 1999, Diarra et al. 2002a), and by combining Lf with antimicrobial drugs to treat infections, better results could be obtained than with antimicrobials or Lf alone.
Table 1. Concentration of lactoferrin in secretions of the bovine mammary gland in healthy and mastitic cows.
Phagocytosis by polymorphonuclear neutrophils (PMNs) and macrophages is an important part of the host defence mechanism against pathogenic bacteria (Vanfurth & van Zwet 1986). Many obligate and facultative intracellular bacteria have developed sophisticated systems to avoid phagocytosis and intracellular killing by phagocytic cells. In S. aureus, capsular polysaccharides contribute to virulence by interfering with phagocytosis and intracellular killing by PMNs. For these pathogens, the ability to survive in phagocytes and mammary epithelial cells results in the persistence of infection and the failure of antimicrobial treatments (Diarra et al. 2003). Ideally, synergism should exist between antimicrobial agents and bactericidal activity of phagocytes to eliminate pathogens. Miyauchi and coworkers (1998) reported that Lf enhanced phagocytosis by human neutrophils in vitro. Diarra and co-workers (2003) suggested that if a similar effect exists in dairy cows, Lf could be used to treat mastitis, especially during periods such as early lactation, when the function of PMNs is depressed. 2.5 Treatment of coliform mastitis Coliform mastitis, mostly caused by E. coli, is sometimes associated with severe or even fatal systemic and local signs, and often causes considerably diminished milk production (Golodetz 1985, Lohuis et al. 1990, Rantala et al. 2002); only one-third of cows return to full lactation in the affected udder quarter (Golodetz 1985, Shpigel et al. 1997). McDonald & Anderson (1981) found that inactive neutrophils were related to a high incidence of severe mastitis. The speed and extent of the host response and the host's resistance to infection largely determine the prognosis of coliform mastitis, which in turn determines the recovery to normal function (Paape et al. 2002, Burvenich et al. 2003). During the puerperal period the mobilization and function of leucocytes, which are important factors affecting the pathogenesis of coliform mastitis, may be suppressed (Kehrli et al. 1989, Shuster et al. 1996). Compromised host response in some cows can result in an exponential growth of E. coli bacteria and in development of severe mastitis (Hirvonen et al. 1999, Burvenich et al. 2003). The severity of mastitis has been demonstrated to be significantly related to bacterial count in milk (Shuster et al. 1996, Hirvonen et al. 1999). The higher the number of viable bacteria in milk 10-12 hours after infection, which is when bacterial numbers peak, the more severe the mastitis (Hill et al. 1984). Pyörälä & Pyörälä (1998) reported that the mean bacteriological cure rate for E. coli mastitis was 71% and that spontaneous elimination of bacteria was typical. However, this retrospective study contained all cases from mild to severe mastitis during the lactation period, not only infections just after parturition. The treatment of peracute and acute E. coli mastitis using antimicrobial agents is problematic, and treatment results have been controversial (Jones & Ward 1990, Katholm & Andersen 1992, Pyörälä et al. 1994a, Erskine et al. 2002), partly because signs arise in response to LPS. In some experimental studies, antimicrobial therapy was no more effective than no treatment at all (Jones & Ward 1990, Pyörälä et al. 1994a). In other studies, however, antimicrobial treatment in serious cases has resulted in faster elimination of bacteria, increased survival rate of cows and reduction in the loss of milk production (Shpigel et al. 1997, Dosogne et al. 2002, Rantala et al. 2002). Antimicrobials continued to be used to treat coliform mastitis despite the lack of convincing evidence for their therapeutic value. Only a few antimicrobial substances are pharmacologically suitable to treat coliform mastitis and are approved for use in lactating cattle. Therefore, many broad-spectrum antimicrobials are used extra-labelly, which increases the risk of drug residues both in milk and meat. The use of broadspectrum antimicrobials also increases selection pressure for development of resistant bacterial strains. In humans, the use of bactericidal drugs, such as gentamicin, to treat sepsis due to Gram-negative bacteria may result in bacterial lysis, LPS release and shock (Goto & Nakamura 1980, Shenep & Mogan 1984, Periti & Mazzei 1999). In bovine mastitis, antimicrobial treatments of coliform mastitis have mostly failed to show effective killing of bacteria (Jones & Ward 1990, Pyörälä et al. 1994a), and thus, the risk for release of substantial amounts of LPS may be low. In a recent experimental E. coli mastitis study, Dosogne & co-workers (2002) showed that treatment with enrofloxacin was not associated with enhanced release of endotoxin. For treatment of coliform mastitis, in particular, new approaches are needed. Lf is perhaps one of the most promising candidates. It could partly replace the use of antimicrobials, and in addition, it has an LPS blocking effect. Zhang and coworkers (1999) suggested that Lf could potentially be used to treat endotoxininduced septic shock in humans. Moreover, Zagulski and co-workers (1986) showed that Lf increased survival rates in an experimental E. coli mouse model; the mortality rate in the bovine Lf pretreated (24 hours before challenge) mice was seven times lower than in control mice. In addition, bovine Lf given intravenously (i.v.) to rabbits significantly prolonged survival time after systemic experimental infection with E. coli (Rainard 1986). While no studies have yet been carried out using Lf to treat coliform mastitis in dairy cows, the results from the above-mentioned experimental models are encouraging.
3. AIMS OF THE STUDY The main hypothesis and specific objectives of this study are presented in Figure 2. LACTOFERRIN HAS THERAPEUTIC POTENTIAL IN MASTITIS Study I To describe the normal concentrations of Lf and citrate in the early dry secretion of healthy cows Study II To determine in vitro the antibacterial activity of Lf against bovine udder pathogens Study III To study the disposition kinetics of Lf in milk as well as local and systemic tolerance of Lf in lactating dairy cows Study IV To investigate the efficacy of LF in treating cows with experimentally induce E. coli mastitis Figure 2. The main hypothesis and the four objectives of the study.
4. MATERIALS AND METHODS 4.1 Animals (I, III, IV) Seventy clinically healthy Finnish Ayrshire and Friesian cows and one Finnish landrace cow from the research farms of the University of Helsinki and MTT Agrifood Research were used. Table 2 shows the number and types of the cows in different experiments. In Study III, preliminary trials were conducted before the final experiments. . 4.2 Lactoferrin and citrate determinations (I, II) Lf concentrations in the samples were determined using DELFIA (dissociationenhanced lanthanide fluoroimmunoassay) at the University of Oulu, Finland (Isomäki 1999). Concentration of citrate in the samples was measured by photometry (Mutzelburg, 1979) using a commercial kit (Boehringer Mannheim GmbH, Mannheim, Germany).
Table 2. Number and types of cows used in the experiments
The molar ratios were calculated using the following formula: Molar ratio= (average citrate concentration) / (molecular weight of citrate) (average Lf concentration) / (molecular weight of Lf) Molecular weight of citrate is 192 g/mol and that of Lf 77 000 g/mol. 4.3 Lactoferrin purification (II-IV) Lf was purified from cheese whey or concentrated cheese whey at the University of Oulu, Finland, using an expanded bed adsorption chromatography method developed by Isomäki (1999), and freeze-dried Lf was stored at 21°C. The amount of endotoxin in Lf was tested with LAL (Limulus amebocyte lysate) using the kinetic BioWhittaker-QCL method (Walkersville, MD, USA); it did not exceed 1.7 ng/ml. Before udder infusion, Lf was sterile-filtered (32 mm Acrodisc PT Syringe Filters 0.8/0.2 µm, Gelman Laboratory REF: 4658). 4.4 Bacteria (II) Five isolates of E. coli, S. aureus and CNS, and two isolates of P. aeruginosa and K. pneumoniae, originally isolated from subclinical or clinical cases of bovine mastitis, were used. These isolates were received from the mastitis laboratory of the Faculty of Veterinary Medicine and from the National Veterinary and Food Research Institute, Helsinki. One of the S. aureus isolates was reference isolate M60 kindly provided by Dr. A. J. Guidry (Immunology and Disease Resistance Laboratory USDA, Beltsville, MD, USA). During the experiment bacteria were maintained on blood agar plates at 8°C. To adapt the bacterial isolates to grow in whey, they were incubated overnight at 37°C in a growth medium consisting of 2/3 Iso Sensitest Broth (ISB CM473, Oxoid Ltd., Basingstoke, Hampshire, England) and 1/3 sterile whey. The cultures were tested using Gram-staining for purity. Bacteria were harvested by centrifugation (5000 g for 10 min) and washed twice between centrifugations using sterile saline (0.9% NaCl, 20°C). A suspension containing approximately 109 colony-forming units (CFU) in 0.9% NaCl was prepared, according to the McFarland standard (bioMérieux sa, 69280 Marcy I“Etoile, France), by spectrophotometry (550 nm, Hitatchi U-2000, Hitachi, Ltd., Tokyo, Japan). Bacterial suspension was diluted to the final concentration of 1.5x103 CFU/ml used in each well. Two growth media were used for bacterial cultures: ISB and whey. Whey was prepared from three lits of fresh raw milk obtained from the university dairy herd by high-speed centrifugation of defatted milk (32600 g for 60 min at 4°C). Whey was sterile-filtered and frozen in 40-ml aliquots immediately after preparation for later use. 4.5 Analysis of bacterial growth by turbidometry (II) The antibacterial effect of Lf was tested by using turbidometry (Bioscreen instrument, Labsystems, Helsinki, Finland). The instrument is a fully automated analysing system for measuring bacterial growth using the vertical light path with the wide-band absorption principle; 200 samples can be measured simultaneously. Each well contained 100 µl of ISB broth or 150 µl of whey as the growth medium, 50 µl of bacterial suspension and 50 µl of Lf concentrate. Physiological saline was added to bring the final volume to 300 µl: 50 µl and 100 µl in ISB and whey wells, respectively. Tested amounts of Lf were 200 µg (final Lf concentration in the well was 0.67 mg/ml), 500 µg (1.67 mg/ml) and 800 µg (2.67 mg/ml). Control wells contained all components except Lf, which was replaced by 50 µl of 0.9% NaCl. Five parallel wells were used. Wells between whey and ISB were filled with 0.9% NaCl to prevent cross-contamination. The wells on two 100-well plates were covered and preincubated in the Bioscreen instrument for 30 min at 37°C. The change in turbidity was monitored automatically every hour for 20 h at 37°C. The plates were shaken 10 min before each measurement. Lag time (time from beginning of incubation until the time-point when absorbance began to increase), slope (slope of the growth curve in the logarithmic growth phase) and maximum absorbance (highest absorbance value measured during the 20-h period) were used as variables describing the bacterial growth. After the 20-h incubation period, bacterial survival and bactericidal effect of Lf in the wells were confirmed by culturing aliquots of 10 µl on blood agar plates and incubating the plates overnight at 37°C.
4.6 Experimental designs (I, III, IV) Study I, milk and dry udder secretion samples for the determination of Lf and citrate concentrations were collected on the last day of the drying-off process, and 2 and 6 days later. The drying-off process lasted 2 weeks and was initiated approximately 8 weeks before parturition. During this period the cows were milked only once a day during the first week and once every second day during the second week. The somatic cell count (SCC) was less than 250 000 cells/ml milk in all except four cows, which had earlier had subclinical mastitis. Aseptic milk samples cultured had less than five colonies of minor pathogens (mostly Corynebacterium sp.) in some cows. Because these pathogens are considered to be of minimal importance in mastitis (Oliver & Juneja 1990), we included these cows in the experiment. The samples were frozen at -20°C, and Lf and citrate concentrations were measured 4 to 7 months later as described earlier. Study III, a 1-g dose of Lf was infused into one udder quarter of six cows after the morning milking. Before udder infusion, Lf was dissolved in water and sterilefiltered (32 mm Acrodisc PT Syringe Filters 0.8/0.2 µm, Gelman Laboratory REF: 4658). Infusion volume was 91 ml, and the final Lf concentration 11 mg/ml. Aseptic milk samples were taken before infusion and at 1, 2, 4, 8, 24 and 48 h after infusion from test and contralateral control quarters. Milk was cultured for bacterial growth at 24 and 48 h. The milk samples were frozen at -20°C, and Lf concentrations were measured 2 weeks later. An aliquot of milk was separated and used for determination of N-acetyl-ß-D-glucosaminidase (NAGase) activity. NAGase was used as an indicator for inflammation in the milk. Pharmacokinetic parameters of Lf were calculated using a commercial software package (PK Solution 2.0.4, Non-compartmental Pharmacokinetic Data Analysis, SUMMIT Research Services, Ashland, USA). In Study IV, one udder quarter of each cow was infected via the teat canal with approximately 1500 CFU of E. coli strain FT238 from clinical mastitis suspended in 10 ml of pyrogen-free saline 4 h after evening milking. The strain was nonhaemolytic, intermediately serum-resistant and sensitive to enrofloxacin in vitro (MIC <0.25 mg/ml) (Pyörälä et al. 1994a, Rantala et al. 2002). The strain was also included in the in vitro testing for Lf (Study II) and found to be fully susceptible to Lf. Treatment was started at 12 h post-challenge (PC), after the morning milking, when clinical signs became visible. Three of the cows received Lf three times (Table 3). The remaining three cows received enrofloxacin (Baytril®, Bayer AG, Leverkusen, Germany) three times (Table 3). Flunixin meglumine (Finadyne®, Schering-Plough, Farum, Denmark) was administered intravenously (i.v.) to all cows (Table 3). A cross-over design was used so that each cow served as its own control (Hirvonen et al. 1999), and a contralateral udder quarter of the same cow was inoculated at the second challenge. 4.7 Clinical observations and collection of blood and milk samples (III, IV) In Study III, cows were carefully monitored post-infusion for possible adverse reactions to the Lf infusion. The udder reactions were recorded on a four-point scale (0 = no pain, no oedema; 1 = slight soreness, slight oedema; 2 = moderate soreness, moderate oedema; 3 = severe pain, severe oedema, marked increase in size). Rectal temperature was measured and possible systemic signs, such as reduced appetite or depression, and the appearance of milk were recorded at the same time-points. In Study IV,
systemic and local clinical signs of the cows were monitored throughout the
experiment; during the first 2 days every 4 h, and thereafter, every time the
cows were milked. Heart rate, rectal temperature, rumen motility, appetite and
general attitude were recorded. The udder was palpated for soreness, swelling
and hardness, and quarter milk samples were evaluated visually for
Table 3. Six dairy cows were divided in two treatment groups A and B. Number 1 refers to the first challenge in one udder quarter and number 2 to the second challenge of the other udder quarter three weeks later
In Study IV, blood samples were drawn from the jugular vein of each cow. Blood was collected at 12 h before challenge and again at 12, 16, 20, 36, 60, 84 and 180 h PC. Serum was separated and samples were frozen at -70°C for later use. Aseptic milk samples were collected from the challenged and the contralateral quarters 5 days and 12 h before the challenge and at 12, 16, 20, 36, 40, 44, 60, 84, 108, 132 and 156 h and 1, 2 and 3 weeks PC for bacteriology, SCC and NAGase determination. 4.8 Determination of indicators of inflammation in milk (I, III, IV) The California Mastitis Test (CMT) and determination of milk NAGase activity were used as indicators of inflammation in the infused quarters. In the CMT, the Scandinavian scoring system was used (Klastrup 1975). NAGase activity was analysed using a microplate modification developed by Mattila & Sandholm (1985) from the fluorogenic method of Kitchen and co-workers (1978). SCCs in milk were measured using the Coulter Counter method (Mattila 1985). 4.9 Other methods (IV) In the experimental challenge study (Study IV), the E. coli strain FT238 was used for the challenge. The bacterial strain, which had been kept frozen, was cultured and grown on blood agar plates overnight. A few colonies were subcultured in ISB broth (Iso Sensitest Broth, Oxoid, Basingstoke, Hampshire, England) and incubated for 18 h at 37°C. The broth culture was centrifuged at 4000 rpm (2800 x g) for 10 min, and the pellet was resuspended in sterile saline. The bacteria were washed three times in saline and diluted to an estimated concentration of 108 CFU/ml. The final concentration was determined by measuring turbidity at 620 nm. The inoculation dose of approximately 1500 CFU was suspended in 10 ml of pyrogen-free saline. After the challenge, bacterial counts in milk were determined by preparation of a 10-fold dilution series of milk in sterile saline. Bacteria were then cultured on blood agar, incubated at 37°C for 24 h and counted. The amount of endotoxin (LPS) in the milk samples of Study IV at 20 h PC was measured using a LAL test as described in section 4.3. The ratio between the amount of endotoxin and the bacterial count was calculated for both treatment groups. The concentration of Lf in the milk of infected udder quarters was not measured because the consistency of the milk changed considerably after acute mastitis had developed, and clots and flakes in milk have a marked effect on results. Furthermore, separating endogenous and exogenous Lf is impossible. 4.10 Statistical methods (I-IV) In Study I, temporal patterns in Lf and citrate concentrations, and differences between the two breeds and between parities were tested by repeated measurement analysis of variance. Each udder quarter was considered to be an independent sampling unit. The cows were grouped into three parities: primiparous cows, cows that had calved twice and cows that had calved three or more times. As data on the molar ratio of individual quarters of an udder were skewed, median values were used in the statistical analysis. Significance of within-subject effects were evaluated by Greenhouse-Geisser adjusted p-values, and p-values less than 0.05 were considered significant. Friedman“s nonparametric analysis of variance was used when testing changes in molar ratios during the experiment. In Study II, the effect of different Lf concentrations on lag time, slope and maximum absorbance was tested by repeated measures analysis of variance with concentration as a within-factor. The significance of concentration was evaluated by Greenhouse-Geisser adjusted p-values. Concentrations of 0.67, 1.67 and 2.67 mg/ml were further compared with a negative control. In Study IV, repeated measures analysis of variance with treatment and sampling time as within-factors was used to test differences in variables (SCC, NAGase activity, bacterial counts, systemic and local clinical signs, and total and quarter milk yields) between the two treatments. The significance of effects was evaluated by using Greenhouse-Geisser adjusted p-values. Possible carry-over effects were assessed by T-test procedures (Jones & Kenward 1989). Statistical significance of the ratio of endotoxin to bacterial count was determined using the Wilcoxon Signed Ranks Test.
5. RESULTS This section summarizes the main findings. For more detailed results, the reader should refer to the original papers located at the end of the thesis. 5.1 Lactoferrin and citrate concentrations in milk and dry cow secretion (I) Mean Lf concentration increased while mean citrate concentration decreased in milk and dry udder secretion during the sampling period. The concentrations of both were highly variable at each time-point. Mean lactoferrin concentration in milk increased from 5.29 mg/ml on the last day of the drying-off process to 8.09 and 11.26 mg/ml 2 and 6 days later, respectively, and citrate concentration decreased from 1.85 mg/ml to 1.54 and 1.09 mg/ml, respectively. No significant difference was found in Lf or citrate concentrations between the two breeds or between different parity groups. Median molar ratio of citrate to native Lf decreased from 153 to 86 and further to 44. This difference was statistically significant (p<0.01). 5.2 Inhibitory activity of lactoferrin against udder pathogens (II) The best inhibitory activity of Lf in the ISB was seen against E. coli and P. aeruginosa. The inhibitory effect of Lf for E. coli was concentration-dependent (Figure 3), and variation in the effect against the five isolates of E. coli was small. None of the E. coli isolates was totally resistant to Lf. The effect of Lf on the maximum absorbance and the slope of E. coli in ISB was significant (p=0.025 and p<0.001, respectively), whereas the effect on lag time was not (p = 0.065). The growth of two isolates of P. aeruginosa was clearly inhibited in ISB. By contrast, the growth of two K. pneumoniae isolates was hardly inhibited at all. In whey, K. pneumoniae showed variable susceptibility and the results were contradictory. The isolates of CNS and S. aureus in ISB showed more variation to Lf than did E. coli. Lf had significant effects on lag time (p=0.014) and maximum absorbance (p=0.014) of S. aureus, while no significant effects were seen for CNS. For to the slope, a significant difference was present between Lf treatments and the control in the growth of S. aureus (p=0.001) and CNS (p=0.002). The growth of four CNS isolates was somewhat inhibited by Lf, but one was totally resistant. Three isolates of S. aureus were more susceptible to Lf than the other two isolates at 0.67 mg/ml of Lf. Lf treatment had a significant inhibitory effect on lag time (p=0.015), maximum absorbance (p=0.011) and slope (p=0.027) of S. aureus in whey. The isolates of E. coli, P. aeruginosa and CNS did not grow sufficiently well in whey to draw any conclusions. We also conducted limited in vitro testing using Apo-Lf (data not shown) and the results were comparable with those of native Lf. 5.3 Disposition kinetics of lactoferrin in lactating dairy cows (III) Intramammary administration of Lf produced elevated Lf concentrations in milk for several hours (Figure 4). Mean elimination half-life of total Lf in milk was 2.2 (range 1.2-4.3) h. The mean maximum concentration of 6.3 (3.0-12.3) mg/ml was reached between 1 and 4 h post-infusion. The average Lf concentration in milk from six cows before the administration of Lf was 0.4 (0.1-1.1) mg/ml. Four hours after infusion, two different patterns were seen; cows 1-3 in first lactation had lower Lf concentrations than the older cows 4-6 (Figure 4). After 8 h of administration, the average Lf concentration had decreased to 0.8 (0.2-1.6) mg/ ml. Twenty-four hours post-infusion, Lf started to increase again, and 48 h after infusion, the mean Lf concentration was 1.5 (0.7-2.7) mg/ml. The increase was lower in primiparous than in multiparous cows.
Figure 3. Inhibitory activity of lactoferrin (Lf) against Escherichia coli was clearly seen.
Lf caused some local tissue irritation in the udder quarters (Figure 5), but systemic signs, such as fever or anorexia, were not observed. The udder quarters of primiparous cows seemed to react faster than those of multiparous cows. The severity of the reactions varied between cows and became visible 2 to 4 h after infusion. In all but one cow, the infused udder quarters became hard or swollen. In the same cows, the ipsilateral quarters swelled slightly. Four to eight hours after administration of Lf, the local reactions reached their peak. The reactions decreased more rapidly in older than in primiparous cows (Figure 5). At 24 h post-infusion, the udder quarters of multiparous cows were normal, while primiparous cows took 48 h to return to normal state. No bacteria were isolated from milk samples taken 24 and 48 h after administration of Lf. The CMT score, indicating the number of cells in the milk of the infused quarters, increased within 8 h of administration and then started to decrease 2 days postadministration. As estimated using CMT, SCC returned to normal levels within 4 days of the infusion. Milk NAGase activity increased after the administration of Lf, the highest value attained being 120 (from 48 to 120) arbitrary units/ml in one cow. NAGase activity of less than 18 is equivalent to SCC of less than 300 000 cells/ml (Myllys et al. 1998).
Figure 4. Lactoferrin (Lf) concentrations (mg/l) in milk after infusion of 1 g of Lf into one udder quarter of three primiparous (1, 2 and 3) and three multiparous (4, 5 and 6) cows.
5.4 Efficacy of lactoferrin in treatment of experimentally induced E. coli mastitis (IV) All cows became infected with E. coli, and all developed clinical mastitis within 12 h PC. The clinical response varied greatly among individual cows. Statistically significant carry-over effects from the first challenge to the second challenge were not detected in any of the variables investigated. According to clinical observations, the response of cows tended to be milder in the second challenge; five cows during the first challenge and one cow during the second challenge developed severe systemic signs of mastitis. Systemic signs generally disappeared within two to three days, and local signs within one week. Differences in systemic and local clinical signs (Figure 6) between Lf- and enrofloxacin-treated cows were not significant.
Figure 5. Irritation reactions in the udder after administration of lactoferrin (Lf). Cows 1,
Bacterial counts in the milk decreased faster in enrofloxacin- than Lf-treated cows, the difference almost reaching significance (p=0.054) (Figure 6). Bacteria were eliminated from the challenged quarters of enrofloxacin-treated cows on average in 3.8 (median 5) days and from Lf-treated cows in 5.8 (median 6) days. After the second challenge, the three cows in the enrofloxacin treatment group had very low bacterial counts in milk already at the first sampling at 12 h PC. Average bacterial count, average amount of endotoxin and the corresponding ratio of the amount of endotoxin to the number of bacteria in milk 20 hours PC are shown in Table 4. The corresponding ratios of the concentration of LPS to the number of bacteria in milk for the enrofloxacin- and Lf-treated cows were statistically significant (p = 0.028). Milk SCCs of the challenged quarters were high at 12 h PC, after which they decreased in all cows. After the first challenge, the SCCs of two cows, one from the enrofloxacin-treated group and the other from the Lf-treated group, approached pre-infection counts within 2 weeks PC. After the second challenge, the SCCs of all enrofloxacin-treated cows and those of one Lf-treated cow returned to pre-infection counts within 2 weeks. Differences in SCCs between treatments were not statistically significant. Milk NAGase activity remained high for a longer period in cows treated with Lf than in enrofloxacin-treated cows (p=0.046) (Figure 6). The mean decrease in total daily milk production at one day PC was similar for enrofloxacin- and Lf-treated cows: 23.8% and 24.1%, respectively. Daily and quarter milk yield profiles over the experimental period were also similar for both treatment groups (Figure 6).
Table 4. Average bacterial count, average amount of LPS and the corresponding ratio of the amount of LPS to the number of bacteria in milk for enrofloxacin- and Lf-treated cows with experimentally induced E. coli mastitis at 20 h PC.
6. DISCUSSION Bovine Lf may have potential in the prevention of bovine mastitis at drying-off, as well as in the treatment of lactating cows. The idea for using Lf for dry cow therapy is to protect the udder with exogenous Lf against new infections during the involution period, when the cow's own Lf concentration is still low. During this period the total daily production of Lf by the mammary gland typically increases at least two-fold that of the day of drying-off (Welty et al. 1976). However, endogenous Lf concentrations are still quite low, and citrate concentrations high. During the first two weeks of the dry period, the risk for mastitis is many times higher than during the lactating period (Smith et al. 1985, Nickerson 1989, Oliver & Sordillo 1989), probably due to cessation of milking, which provides a better environment for bacteria to multiply in the milk compartment of the udder. In the first part of this thesis the natural concentrations of Lf and citrate were determined during the early drying-off period, because dry cow therapy was one of the suggested indications for Lf. Lf and citrate concentrations varied between cows, in agreement with the findings of Welty and co-workers (1976); neither breed nor parity affected Lf or citrate concentrations or changes in these concentrations over time. In addition, variation in Lf and citrate concentrations was high between udder quarters of individual cows. One explanation given for between-quarter variations is previous mastitis in some of the udder quarters, but we were unable to verify this. Another more plausible explanation is reduction in secretory activity of the mammary gland and decline in milk volume for reasons other than previous mastitis (Smith & Schanbacher 1977, Watson 1980). Reduction in milk production of individual cows, and between udder quarters, differs before drying-off. Tsuji and co-workers (1990) found out that 24 h after parturition colostrum from dairy breeds contained on average four times more Lf than colostrums from beef breeds, and in multiparous dairy cows Lf content in colostrum was two to three times higher than in primiparous cows. Another interesting finding was that the molar ratio of citrate to Lf, which has been found to be more important in the inhibition of growth of Gram-negative bacteria in vitro than the absolute concentration of either component alone (Bishop et al. 1976, Nonnecke & Smith 1984), decreased from a median of 153 on the day of drying-off to 44 on day 6 after drying-off. The inhibition of bacterial growth depends on the value of the molar ratio: the smaller the molar citrate-to-Lf ratio, the more effective the inhibition (Bishop et al. 1976). Thus, increasing the amount of Lf in milk during the early drying-off by using exogenous Lf could shift the ratio to be less favourable for bacterial growth. Kai and co-workers (2002b) have shown that Lf treatment for staphylococcal mastitis in the early non-lactating period results in higher cure rates than antimicrobial treatment. The cure rate reported was 92%, which was due to bacteriostatic or bactericidal effects as well as to the induction of the innate immunity of the host. This cure rate is in contrast to the 48% cured with antimicrobials over the same 7-day time frame. Cell population in the cured udder quarters contained mainly phagocytes, including PMNs and cells positive for CD11b, which is known to be a complement receptor. The authors speculated that locally infused Lf promoted the migration of PMNs by inducing the expression of such cytokines as IL-1ß and IL-8. We agree with their vision of clinical applications for Lf in mastitis during drying-off. Exogenous udder infusion of Lf as a dry-cow therapy in a long-acting formula, possible in combination with antimicrobial agents, may enhance the defence mechanisms of the cow and help limit bacterial growth in the udder. Another potential application for exogenous Lf is prevention of infections in the first days after drying-off, probably used without antimicrobial agents. Experiments for dose determination and pharmaceutical formulation of Lf for possible combination with antimicrobials or bicarbonate to overcome the effect of citrate are still needed. During the non-lactating period most iron in the bovine mammary gland is bound to Lf, and coliform mastitis rarely occurs in the fully involuted mammary gland. However, one problem is that some Gram-negative bacteria can overcome the iron-limiting effect of Lf by using their iron acquisition systems like enterobactin. The enterobactin receptor FepA acts in the transportation and deferration of ferric enterobactin (Lin et al. 1998). It can be hypothesized that if coliform mastitis occurs during the dry period, it may be due to strains which possess mechanism such as FepA and enterobactin. FepA has been studied for possible development of a specific vaccine against E. coli mastitis during the dry period (Lin et al. 1999). Lf had an in vitro antibacterial effect against some major udder pathogens, as also shown in many earlier studies (Nonnecke & Smith 1984b, Rainard 1986, Diarra et al. 2003). The best inhibitory activity was against E. coli and P. aeruginosa. Variable results were seen against the other important udder pathogens S. aureus, CNS and K. pneumoniae. The strong in vitro inhibitory activity of Lf against E. coli was in agreement with many previous studies (Nonnecke & Smith 1984b, Rainard 1986, Dionysius et al. 1993, Diarra et al. 2003), but contradictory results have also been reported (Sanchez & Watts 1999). Rainard (1986) tested the in vitro susceptibility of 35 E. coli isolates to Lf, which had originally been isolated from clinical or subclinical bovine mastitis. Most of the isolates were completely inhibited. A few isolates partially resisted the bacteriostatic action of Lf, but none was totally resistant. Nonnecke & Smith (1984b) reported bacteriostatic, but not bactericidal, activity of bovine Lf against mastitis-causing isolates of E. coli and K. pneumoniae. Dionysius et al. (1993) reported that an Lf concentration of 1.0 mg/ml inhibited growth of all 19 isolates of enterotoxigenic E. coli isolated from porcine enteritis. They found the degree of inhibition to be strain-dependent and bacterial killing to occur at relatively high initial concentrations of bacteria. We used 0.67, 1.67, and 2.67 mg/ml concentrations of Lf to compare the inhibitory activity of Lf, and also observed concentration- and strain-dependency. Despite extensive research on these antibacterial effects in vitro, their role in vivo remains controversial. In vitro experiments showing inhibition of growth through iron sequestration cannot mimic the complex interactions occurring during infection in vivo, when iron is available from a much wider range of sources, including haemoglobin, to which Lf cannot bind (Brock 2002). Diarra and co-workers (2003) tested inhibition of two clinical isolates of S. aureus from bovine mastitis with Lf in vitro and in vivo, alone and in combination with penicillin G. Their data both in vitro and in vivo in a mouse mastitis model suggested that Lf is as effective as penicillin G against S. aureus. We focused on native Lf because the cost of purifying apo-Lf would limit this agent's use in treating bovine mastitis. Whey, prepared from milk of healthy cows, was used as a growth medium in the in vitro studies since it simulates the environment of the milk compartment of the cow's udder. Unfortunately, E. coli was unable to grow in whey prepared from milk of healthy cows. This has also been described earlier (Maisi et al. 1984); whey is known to inhibit the growth of a number of bacterial species. Whey prepared from milk of mastitic cows may have been a better medium, but standardizing this medium would have been difficult. Based on the results of our in vitro studies, we decided to focus on E. coli mastitis, which still posses a major challenge for the dairy industry, while P. aeruginosa and K. pneumoniae have less prominent roles. Since E. coli infection remains in the milk compartment, it is a good target of intramammary treatment with Lf. S. aureus and perhaps some CNS bacteria cause deep infections in the udder gland, and it was difficult to predict how the exogenous Lf would spread in the udder. The internalization of S. aureus in neutrophils and epithelial cells would also have been problematic, despite Lf enhancing phagocytosis and intracellular killing (Diarra et al. 2003). Finally, E. coli was chosen because of the LPS-neutralizing capacity of Lf against Gram-negative bacteria. We found that a 1-g dose of Lf infused into the udder quarter increased Lf concentrations close to those found in mastitic milk; these concentrations are within the effective range, not too low or too high. Excessively low concentrations, e.g. 0.5 mg/ml, have no therapeutic effect. Excessively high concentrations, e.g. 12 mg/ml, cause moderate to severe irritation reactions in udder quarters and may even disturb or destroy the normal function of the epithelial cells of the udder. However, Lf concentrations are much higher during the middle of the non-lactating period than in the early non-lactating period, when the udder is very resistant to coliform infections (Smith et al. 1985, Oliver & Bushe 1987). Kai and co-workers (2002a) have shown that a high concentration of Lf (1 mg/ml) promotes phagocytic activity in the mammary gland. The mean elimination half-life of exogenous Lf of 2.2 h was shorter than optimal; antibacterial effects might be enhanced with a half-life of 4 to 6 h. The udder size, which is related to the age of the cows, was found to influence Lf concentrations. Udder volume is normally larger and milk yield higher in older cows than in primiparous cows, and the dilution effect of exogenous Lf might be different in older cows. After reaching a peak, Lf concentrations decreased faster in younger cows. This could be due to the smaller udder size of younger cows, and thus, the smaller gland cistern and milk volume of the udder. Lf may have been more concentrated in the distal parts of the udder quarter, being removed while collecting milk samples. Differences of blood circulation in younger and older cows would also have a role in removing the exogenous Lf. Endogenous Lf in blood and released by neutrophils, and probably exogenous Lf infused into the udder quarters, is transported to the liver, where it is taken up by specific receptors and metabolized (McAbee & Esbensen 1993). Interestingly, Lf concentrations decreased to normal concentrations in four out of six cows within 8 h of Lf administration, but at 48 h, the concentrations in the older cows were again several-fold higher than in milk samples taken before administration. This rise is most likely due to endogenous Lf production, which results from the cow's foreign body reaction to the administered protein, thus stimulating the granules of PMN to release endogenous Lf into the udder. However, in the primiparous cows, Lf concentrations remained at normal concentrations 48 h after administration, which is difficult to explain. In milk samples taken at 1, 2 and 4 h after exogenous Lf infusion, measured Lf concentrations exceeded the infusion concentration. One explanation for this may be that as a water-soluble substance the exogenous Lf did not spread throughout the milk compartment of the infused udder quarter, but remained in the gland and teat cisterns. Formation of endogenous Lf may also play a role in the elevated Lf concentrations in older cows. Intramammary infusion of Lf caused local reactions in the udder quarters. This could be due to a foreign body reaction of the udder against exogenous protein. Another explanation for the udder irritation could be contamination of Lf with endotoxin, provoking an inflammatory reaction. The 1-g dose of Lf used here contained small quantities of endotoxin, which could partly be responsible for the mild local reactions. It is impossible to purify endotoxin-free Lf from whey. However, the low amounts of endotoxin and mild irritation reactions caused by exogenous Lf would have been irrelevant in severe forms of E. coli mastitis, when huge amounts of endotoxin are released. Based on the results of the disposition kinetic studies, we concluded that a 1-g dose of Lf is insufficient to examine its impact over a longer period. Thus, we decided to use a 1.5-g doses in the experimental E. coli mastitis study. Because of local irritation, increasing the dose further was not possible. We therefore chose to use three doses at 12-h intervals to extend the period of therapeutic Lf concentrations in milk so that Lf would have better clearance of bacteria and elimination of LPS. Since the degree of irritation caused by these three 1.5-g doses of Lf to the mastitic udder quarters was unclear, we did not use a longer treatment. While the approach not to conduct repeated-dose disposition kinetic studies could be criticized, carrying out these studies would have been economically impossible. Unfortunately, well-controlled in vivo experiments using physiologically relevant concentrations of Lf are lacking. No significant differences were observed in the outcome of mastitis between cows treated with Lf and those receiving enrofloxacin; however, bacteria were eliminated faster in the cows treated with enrofloxacin. The three cows treated with Lf at the first challenge had lower bacterial counts after the second challenge. The rapid immune response of the host probably eliminated bacteria from the udder before the treatment with enrofloxacin was started. One explanation for the rapid immune response may be longer persistence of bacteria in the first challenge or a possible immunostimulatory effect of Lf in these cows. Promptness and extent of response of individual cows to infection largely determines the prognosis and outcome of coliform mastitis (Shuster et al. 1996, Hirvonen et al. 1999, Burvenich et al. 2003). In addition, multiparous cows may be more susceptible to periparturient infectious diseases than primiparous ones (Mehrzad et al. 2002). Outcome of mastitis and the role of Lf treatment might have changed had we used multiparous cows in our experiment. The effect of Lf may differ in naturally infected cows diseased soon after parturition, and if four or more intramammary doses of Lf are used. Combining Lf with some carefully selected antimicrobial agents could limit the growth of E. coli bacteria better than Lf alone. The efficacy of Lf was thus not unequivocally confirmed in the present experiment, and the small size of the treatment groups could be one reason. Furthermore, our experimental mastitis was induced in vitro with one E. coli strain susceptible to Lf. In the field, E. coli mastitis can be caused by a variety of phenotypically or genotypically different strains, and no specific pathogenic mastitis-causing strains have been identified (Kaipainen et al. 2002). However, some strains of E. coli are known to circumvent the iron-limiting effect by using other sources of iron (Lin et al. 1998). For this reason, a 100% effect of Lf will never occur in spontaneous E. coli mastitis. Much of the LPS release transpires following bacterial phagocytosis and killing by neutrophils. Consequently, by the time treatment was initiated, maximal release of LPS is probably over. We agree with the suggestions by Erskine and coworkers (1991) that either bacterial growth must be inhibited to reduce the bacterial exposure of the udder quarter and the cow to LPS, or the effect of the LPS release must be neutralized to reduce the severity of acute coliform mastitis. Zhang and co-workers (1999) proposed a human Lf-derived 33-mer synthetic peptide (Lf-33) to treat endotoxin-induced septic shock. With both in vitro and in vivo mice models, they showed that survival rate was dramatically increased by injecting Lf-33 simultaneously with LPS; mortality was reduced from 93% (14/15 animals dead) to 6% (1/15 animals dead). In our experimental study, LPS concentrations in milk were about 1400 times lower in Lf- than in enrofloxacintreated cows, which may indicate the LPS neutralizing effect of Lf infused into the E. coli-affected udder quarters. Ziv & Schultze (1983) tested in an experimental endotoxin mastitis model polymyxin B, an antimicrobial drug with the capacity to neutralize LPS, but failed to find significant beneficial effects with its administration shortly after LPS. In some in vitro studies on mastitis pathogens, Lf has been combined with antimicrobial agents (Sanchez & Watts, 1999, Diarra et al. 2002a, 2003), and a synergistic effect has been found. Testing Lf alone avoided the problems related to the use of antimicrobial drugs and allowed the actual net effect of Lf against several bacterial species in vitro and against experimental E. coli mastitis in vivo to be determined. Only a few potentially effective antimicrobial drugs exist for the treatment of severe E. coli mastitis. Enrofloxacin is one of these, but it causes marked tissue irritation if not administered i.v. (Pyörälä et al. 1994b), as also seen in our study. More importantly, fluoroquinolones are valued drugs in human medicine, and their abundant use in food-producing animals could lead to a problematic increase in bacterial resistance. Exogenous Lf, possibly combined with a bicarbonate, would support clearance of bacteria from the udder while simultaneously blocking the detrimental effects of LPS. Upon determining the optimal dose, its potential value in treating coliform mastitis, especially severe E. coli mastitis, is immeasurable. Further pharmacokinetic studies and field trials are, however, needed to more fully elucidate its role in the treatment of severe coliform mastitis.
7. CONCLUSIONS 1. Lf concentrations gradually increased and citrate concentrations decreased in milk during the first days of the dry period. The Lf and citrate concentrations markedly varied between cows and even between udder quarters of the same cow; neither the breed nor the parity of cows affected the concentrations. 2. The in vitro inhibitory activity of Lf against E. coli and P. aeruginosa was good in the ISB broth. Some inhibitory activity was found also against S. aureus. The activity of Lf against CNS was variable and no significant effects were seen. 3. Intramammary administration of one gram of Lf produced elevated Lf concentrations in milk for several hours. Udder size, which is related to the age of cows, influenced the concentrations. 4. Exogenous Lf |
