Nonspecific Antibacterial Factors in Milk from Cows Immunized with Human Oral Bacterial Pathogens' N. TAKAHASHI,* G. EISENHUTH, 1. LEE, and C. SCHACHTELE3 Department of Oral Sciences Clinical Research Center for Periodontal Diseases School of Dentistry University of Minnesota Minneapolis 55455 N. LAIBLE and S. BlNlON Procor Technologies, Inc.

and Land O'Lakes, Inc. Arden Hills, MN 55112

alone showed similar responses. Per unit volume, the milk contained approximately 150 times less lysozyme than whole human saliva obtained from six subjects but higher concentrations of lactoperoxidase and iron-binding components. Purified bovine nonspecific factors prevented the growth of the bacteria used for immunization when bacteria were tested at concentrations similar to those found in saliva and milk. Because bovine nonspecific antibacterial factors could influence both the pathogenic target bacte ria and the indigenous microflora in oral passive immunization studies with bovine immunoglobulins, the presence of these proteins should be considered. (Key words: antibacterial factors, bovine milk, oral bacteria)

ABSTRACT

Both the immunoglobulins and nonspecific antibacterial factors in milk from cows immunized with pathogenic oral bacteria have the potential to influence the oral microflora during passive immunization studies. The first six milks after calving were collected from 2 cows immunized with adjuvant and from 14 cows immunized with adjuvant and heatkilled strains of periodontopathic Actinomyces, Porphyromonas, Prevotella, and Fusobacterium. Analysis of the products from the first to the sixth milks revealed that the protein and lysozyme content decreased approximately 66 and 72%, respectively; the mean specific activity of the enzyme remained relatively constant. In contrast, the mean lactoperoxidase activity increased 2.3-fold in the second milking and increased further in the fourth and sixth milkings. The mean iron-binding activity increased 1.2-fold from the first to the second milkings and then decreased 3.6-fold through the sixth milking. Cows immunized with adjuvant

INTRODUCTION

The concept of passive immunotherapy with bovine colostrum antibodies was first presented by Peterson and Campbell (32) in 1955, and, since that time, bovine colostrum or milk immunoglobulins from immunized cows have been used in several systems to reduce microbial or viral infections of humans. For example, diarrhea in human volunteers induced by Received December 2, 1991. Escherichia coli was prevented in bacterial Accepted March 10, 1992. 'This research was supported in part by Grant 1 B O challenge studies (44) by ingestion of milk DE98489 from the National Institute of Dental Research, immunoglobulins from cows immunized with National Institutes of Health,Bethesda, MD 20892. specific strains of the bacterium. In addition, 2Depmbmmt of Oral Biochemistry, Tohoku University, gastroenteritis in human infants caused by Sendai, Japan. 3T0 whom correspondence and reprint requests should rotavirus infection was treated with bovine immunoglobulins directed against the virus (4,7). be addressed. 1992 J Dairy Sci 75:181&1820

1810

ANTIBACTERIAL FACTORS IN IMhlUNE BOVINE MILg

The E. coli and rotavirus studies showed that antigen-antibody interactions occurring in the human bowel lumen or mucosa were protective. Several attempts have been made to use passively supplied antibcdies to alter bacterial colonization of the oral cavity. Lehner et al. (17) utilized monoclonal antibodies directed against a cell surface protein of Srreptococcus mutans to reduce the concentrations of this cariogenic bacterium in the mouth of rhesus monkeys. In that study, the antibodies were repeatedly applied to the teeth, and dental caries development was suppressed. Similar antibody preparations were used in a studies with human volunteers (22) in which application to the teeth decreased S. mutans in dental plaque. Recently, egg yolk antibodies from chickens immunized with S. mutans were used to reduce caries formation dependent on S. mutans in pathogen-free rats (30). Bovine milk antibodies prepared from cows immunized with S. mutans were used as a part of the diet to prevent colonization of the oral cavity of gnotobiotic rats with a consequent reduction in caries development (27). Utilization of antibody preparations, such as oral rinses in human volunteers, reduced the bacterium concentrations in samples of dental plaque (5, 6). All of the studies with S. mutans suggest that the antibodies in the rinses prevented adherence of the bacteria to the teeth. Human periodontal disease results from entrance into the mouth and colonization of oral tissues by specific pathogenic bacteria (41, 42). The development of a complex pathogenic microbiota by these bacteria can lead to disease development (3, 46). Interestingly, passive immunization against experimental oral infection of hamsters with the periodontopathic bacterium Porphyromonas gingivalis was achieved with rabbit antiserum (28). As part of a systematic approach to evaluation of the potential for passive oral immunization against periodontopathic bacteria, we immunized cows with specific bacteria and collected immune colostrum and milk. Because milk from cows contains various nonspecific antibacterial factors (34). it was necessary to evaluate the quantities and activities of these entities prior to the immunization studies. Although many studies have been conducted on lysozyme (10, 11, 18, 19, 20), lactoperoxidase

1811

(33, 3 3 , and iron-binding proteins (9, 29, 39) from bovine mi& studies have not followed the concentration of protein and each of these factors in milkings from specifically immunized animals nor compared the results with concentrations found in human saliva. In this communication, we present results on the concentrationsof lysozyme, lactoperoxidase, and iron-binding protein in milk obtained from immunized cows following calving and compare this with concentrations found in whole human saliva. In addition, the effect of each nonspecific factor on the viability of the bacteria used for immunization was evaluated. MATERIALS AND METHODS

Bacterial Strains Actinomyces viscosus "VU627 and Actinomyces naeslundii WVU398A were obtained from William Liljemark (University of Minnesota, Minneapolis, MN). Actinomyces naeslundii B120 was obtained from George Bowden (University of Manitoba, Winnipeg, MB, Can.). Fusobacterium nucleatum ATCC 25586 and ATCC 10953, Porp. (Bacteroides) gingivalis ATCC 33277, and Prevotella intermeda (Bacteroides intermedius) ATCC 25611 and ATCC 25261 were obtained from the American Type culture Collection (Rockvue, MD). Growth Conditions

Acrinomyces were grown in the chemically defined FMC medium of Terleckyj et al. (48) supplemented with 10 pg/ml of myo-inositol, 20 pVml of 1 M potassium phosphate buffer (pH 7.2), 10 pl/ml of 1 M sodium carbonate, and 20 Nrnl of 1 M sodium bicarbonate. Fusobacterium nucleatum was grown in FMC medium supplemented with 100 pg/ml of calcium chloride and 100 pg/ml of potassium nitrate. Porphyromonas gingivalis and Prev. intermedia were grown in the medium of Socransky et al. (43) modified by the addition of 1% ( M o l ) trypticase (BBL, Cockeysville, MD), .5 pg/ml of menadione, and substitution of 1.0 pg/ml of biotin, 1.0 pg/ml of choline chloride, 1.0 pg/ml of folic acid, 2.0 pg/ml of myo-inositol, 1.0 pg/ml of niacinamide, 1.0 Journal of Dairy Science Vol. 75, No. 7, 1992

1812

TAKAHASHI ET AL.

pg/ml of pantothenate, 1.0 pg/d of pyridoxal.Hc1, .1 pg/ml of riboflavin, and 1.0 pg/ml of thiamineHCl in place of the vitamin mixture. All cultures were grown anaerobically at 37'C in a glove box with an atmosphere of 80% N2,10% H2, and lW0 C02. For preparation of large quantities of cells, the Actinomyces and Furobacterium cultures were incubated for 18 h, and the Porphyromonas and Prevotella were incubated for 24 h.

soluble fraction was separated from the fat, which was discarded. Collection of Human Mixed Saliva

Unstimulated whole human saliva was collected from six adult subjects (available staff) in the morning (1000 to 1100 h). Oral consent was obtained under the guidelines of the University of Minnesota Committee on the Use of Human Subjects in Research. The saliva was centrifuged (12,000 x g, 10 min, 4T), and the Heat inactivation supernatant was filtered through a sterile membrane ( . 2 2 - p pore size) prior to storage at Cells were harvested by centrifugation (12,000 x g , 30 min, 4'C) and washed with -20°C. sterile phosphate-buffered saline (pH 7.0). After two additional washes, concentrated cell Protein and Enzyme Assays suspensions were heated at 65'C, and samples The Biuret method (15) was used to deterwere removed at various intervals, diluted, and mine protein with bovine serum albumin as the plated onto blood agar plates supplemented standard. with hemin and menadione (45). The number Lysozyme in the skim milk and the saliva of colony-forming units was determined fol- was determined by a lysoplate method with lowing anaerobic incubation at 37°C. Concen- plates prepared by mixing dried Micrococcus trated suspensions of nonviable cells were lysodeikticus cells (.14%, wt/vol; Sigma stored at -2o'C. Chemical Co., St. Louis, MO) with agarose (1.0 wt/vol) in phosphate-buffered saline, pH 6.5. Wells (4 mm) were lilled with 15 pl of Bovine immunization sample or a solution of chicken egg white Fourteen cows were immunized in duplicate lysozyme (.25 to 2.0 pg/ml of buffer). The with pooled strains in the genera Actinomyces, standards contained 1.0% (Mol) bovine sePrevotella, Porphyromonas, and Fusobac- rum albumin (31), and the plates were interium and individually with A. viscosus cubated for 18 h at 37°C in a plastic bag (38). w W 6 2 7 , F. nucleatwn ATCC 25586, Porp. The clearance zone was considered to be gingivalis ATCC 33277, and Prev. intermedia propoItional to the logarithm of the enzyme ATCC 25611. A total of 1 x 1Olo heat- concentration, and the results were taken from inactivated cells in 1 ml were mixed with 1 ml standard curves. of incomplete Freund's adjuvant (Difco, The concentration of lactoperoxidase was Detroit, MJ) immediately before immunization. estimated photometrically by the method of Pregnant Holstein cows were immunized sub- Mansson-Rahemtulla et al. (24). The reaction cutaneously two to three times at approxi- mixture contained 5 mM potassium thiocyamately 28-d intervals prior to calving. Control nate and 10 pA4 2-nitrobemic acid in potascolostrum was obtained from cows immunized sium phosphate buffer @H, 6.5), and the samwith 1 ml of sterile saline mixed with 1 ml of ple was diluted in the same buffer. Reactions incomplete Freund's adjuvant. were started by the addition of hydrogen peroxide at a final concentration of .1 M,and the decrease in absorbance at 412 nm was Collection of Colostrum and Milk followed at 25'C. The 2-nitrobenzoic acid soColostrum was collected in accordance with lution was prepared by adding 2generally accepted dairy practices from the mercaptoethanol to 50 pA4 5,5-dithio-bis-2first six milkings after calving. Aliquots of nitrobenzoic acid dissolved in potassium phoseach milking were frozen at -2o'C. Prior to phate buffer. Production of 1 mM analysis, the milk samples were thawed and hypothiocyanite in 60 s was defined as 1 unit centrifuged (12,000 x g, 10 min, 4'C), and the of activity. Journal of Dairy Science Vol. 75, No. 7, 1992

ANTIBACTERJAL FACIDRS IN IMMUNE BOVINE MILK Quanttfication of iron-Binding Capacity

1813

for bovine recombinant lysozyme (Sigma) contained enzyme at a final concentration of 20 U/ ml (milk concentration) or 450 U/ml (saliva concentration) in sterile saline. The mixture for bovine lactofenin (Sigma) contained protein at a final concentration of 20 (saliva) or 250 (milk) pg/d in sterile saline. The mixture for bovine lactoperoxidase (Sigma) contained 5 U/ ml of lactoperoxidase, .05 [milk (36)] or 1 [saliva (47)] mM NaSCN, and .04 or .8 mM in sterile saline and was incubated at 37'C for 30 min to convert SCN- to OSCN-. The skim milk reaction mixture, which contained colostrum and 0 or .04 mM H202, was incubated at 37'C for 30 min. The reaction was started by addition of cells and incubation at 37'C for 120 min. Controls contained cells in sterile saline. After incubation for 120 min, the mixtures were serially diluted with sterile saline and plated onto brain-heart infusion agar (Difco) for Actinomyces and Streptococcus and on blood agar supplemented with hemin and menadione for the other strains. Colonies were enumerated after a 2-d incubation for the Actinomyces, Streptococcus, and Fusobacterium and 7 d incubation for the other strains.

The total iron-binding capacity in skim milk and saliva was estimated by titration with an iron chelate solution as described by Bates et al. (1). Milk samples were acidified with 6N HC1 to pH 4.7 and centrifuged (l0,OOO x g, 10 min, 4'C) to remove insoluble casein. The Supernatants were n e u t r a h d with 6N NaOH and then filtered through cellulose acetate membranes (.45-pm pore size). The saliva samples were treated in the same way. Samples were diluted to .6 ml with Tris buffer (pH 7.2, .1 M final concentration) in glass cuvettes. Tris buffer was placed into control cuvettes, and aliquots of iron chelate diluted in the same buffer were added to the sample and control cuvettes. Increases in absorbance at 470 nm were recorded in a double beam spectrophotometer (model ACTA Cm, Beckman Instruments Inc., Fullerton, CA). Because of formation of slight turbidity following addition of the iron chelate, simultaneous readings at 560 nm in the control cuvettes were used to correct the data. Unsaturated bovine or human lactoferrin was used as the standard and was prepared by dissolving the protein (Sigma) in .1 M sodium acetate buffer (pH 4.0) containing .2 M monosodium phosphate and .04 M diso- Statistical Analysis dium EDTA, followed by overnight dialysis at When appropriate, the mean values ob4'C against the same buffer, and followed by tained in the studies were compared using the overnight dialysis at 4'C against distilled waStudent's t test. ter. Effect of Antibacterial Factors on Bacterial Growth

Actinomyces viscosus wW627, A. naeslundii WW398A, and Streptococcus sanguis ATCC 10557 were grown in brain-heart infusion broth (Difco); Porp. gingivalis ATCC 33277, Prev. intermedia ATCC 25261, and F. nucleatum ATCC 25586 were grown in BM medium (40) at 37'C in a glove box (80% N2, 10% H2, 10% C o d . All test procedures were performed in the anaerobic box. The bacteria were harvested onto a sterile filter (.22-pm pore size) and washed twice by drawing sterile saline through the filter. The cells were collected with a spatula and suspended in saline. The quantities of the factors tested were based on an approximation of the quantities normally found in human saliva (Table 3) and first bovine milk (Table 2). The reaction mixture

RESULTS Heat lnactlvatlon of Bacteria

To preserve the integrity of antigens, the bacterial cells were inactivated at 65'C prior to immunization. Evaluation of the time course of killing revealed differences in the time required to reduce the colony-forming units to m undetectable concentration. Table 1 presents results obtained with the Actinomyces. Results were similar among the different strains of Prevotella and Porphyromonas; no colonyforming units were detected after 90 min of exposure to 65'C (data not shown). The two Fusobacterium strains were more heat resistant; 150 min of exposure were required to prevent the growth of colonies after incubation (data not shown). For bovine immunization, large batches of cells were prepared and heated Journal of Dahy Science Vol. 75, No. 7, 1992

1814

TAKAHASHI ET AL.

TABLE 1. Effect of treatment at 65'C on viability of oral Actinomyces.

Incubation 0 10 20 30 45 60 90 120

A, naeslundii W398A

1.4 x '0 1 2.9 x 108 2.9 x 1 6 8.0 x Id 5 .o