Antimicrobial effect of human milk on Bordetella pertussis

FEMS MicrobiologyLetters 70 (1990)269-274 Published by Elsevier

269

FEMSLE04081

A n t i m i c r o b i a l effect of h u m a n milk o n Bordetella pertussis K. Redhead i T. Hill t and B. Mulloy 2 Divisions of t Bacteriology and 2 Chemistry, National Institute for Biological Standards and Control, Potters Bar. Hertfordshire. ILK.

Received 30 March 1990 Accepted 6 April 1990 Key words: Milk; Bordetellapert~sis; Bacteriostasis

1. SUMMARY It has been demonstrated that human milk, unlike bovine milk, can reduce the viability of Bordetella pertussis. This antibacterial activity was not due to the presence of antibiotics or antibodies in the human milk. Reducing the level of available iron or increasing the concentration of lysozyme in bovine milk did not induce anti-B. pertussis activity. Analysis of total fatty acids revealed that human milk contained significantly more linoleic acid than bovine milk. However, the addition of linoleic acid to bovine milk did not inhibit the growth of B. pertussis.

2. INTRODUCTION The bactericidal and bacteriostatic properties of milk have been of interest to clinicians and microbiologists for a long time. Secretory immunoglobulin A (slgA) is generally considered to be the main host defence component of human milk although several other milk constituents have

been discovered which possess antibacterial activities including lysozyme [1,2], lactoferrin [3] and the lactoperoxidase system [4-6]. Investigations of these non-antibody defences have concentrated on their contribution to the protection of the suckling infant gut and they have been shown to be active against a wide range of bacteria, including many enteric pathogens [7-9]. During suckling milk passes through a large portion of the infant upper respiratory system, and is frequendy inhaled into the main respiratory tract [10], where it can mix with the mucosal secretions. Bordetella pertussis is the causative agent of whooping-cough, a disease of all ages but predominantly the young, including the newborn. It exists normally as a non-invasive pathogen of the upper respiratory tract and may often confront an environment such as that described. In this paper we report on the ability of B. pertussis to grow in media containing human or bovine milk and examine the milks for factors which may inhibit its growth.

3. MATERIALS AND METHODS Correspondence to" K. Redhead, Divisionof Bacteriology, Na-

tional Institute for Biological Standards and Control, Blanche Lane, South Mimms, Potters Bar, Hertfordshire EN6 3QG, U.K.

3.1. Bacterial maintenance Bordetella pertussis strain Wellcome 28 OV28),

obtained from Dr. P. Novotny (Wellcome Re-

0378-1097/90/$03.50 © 1990 Federation of EuropeanMicrobiologicalSocieties

270 search Laboratory, Beckenham, Kent), was maintained in 5~ (v/v)giycerol-l~ (w/v) casamino acids in ampoule~ held at - 7 0 °C. Cells from ampeules were grown in 250-ml Erlenmeyer flasks with 100 ml of Stainer and Scholte medium [11] per flask, agitated at 100 rpm on an orbital incubator at 36°C for 2 days.

treated as previously described and added to SS at a concentration of 25~. (iii) Linoleic acid (Aldrich Chemical Co.) was dissolved in ethanol (99.8~ v/v) to a concentration of 100 mg/ml, diluted 2000-fold in SS and filter-sterilized (millipore filter, 0.22/~m pore size). This solution was added to SS and SS plus 25~ milk to produce the desired concentration of linoleic acid.

3.2. Preparation of milks Raw milk from cows receiving no antibiotic treatment was kindly supplied by Dr. A. Madell (Royal Veterinary College, Boltons Park, Herts.). Untreated human milks from donors, receiving no antibiotic treatment, who had negligible antibody titres to B. pertussis antigens when previously tested [12] were kindly provided by Dr. M. Thomas (St. Georges Hospital, London). The milks were stored a t 4 ° C for up to 24 h before treatment. To remove the lipid portion of the milks all samples were centrifuged at 12000 x g for 30 rain at 4°C. The aqueous portion was separated and further centrifuged at 130000 × g for 30 rain then the upper cleat layer was removed and filter sterilized (Nalgene filter unit, 0.45 /tin pore size). Finally the preparations were heated at 560C for 30 rain and stored at - 2 0 °C until used.

3.3. Cultioation conditions A set volume (25 ml) of sterile treated bovine or human milk was added to 75 ml of Stainer and Scholte medium (SS) in a 250 ml Erlenmeyer flask. Washed bacteria from 48 h cultures of B. pertussis were used to inoculate flasks to an initial viable count of approximately 5 x 10s/mi. Flasks con~dning 100 ml of SS medium only were also inoculated and included as controls in each cultivation experiment. The cultures were incubated at 36 ° C for up to 48 h on an orbital shaking incubator (100 rpm) and samples aseptically removed at intervals. The number of viable bacteria present in samples was estimated as pre.~ously described [13]. Variations were made in the culture media as follows: (i) the iron content of the SS was lowered from 18 /zM to 4 pM, prior to the addition of treated milk, by reducing the amount of FeSO4 included in the supplement. (ii) Lysozyme was added to untreated bovine milk to raise the f'mal concentration to 100 /zg/ml. The milk was then

3.4. Analysis of fatty acids in milks The total fatty acids present, as free fatty acids and triglycerides, in the treated human and bovine milks were determined by gas-fiquid chromatography of their methyl esters. Samples (one nil) of the milks were freeze-dried then treated according to the method of Wait and Hudson [14]. The resulting fatty acid methyl esters were separated on a 10 m capillary column (BPI from SGE Ltd., Milton Keynes) using helium as the carrier gas at 0.6 ml/min, with a temperature program rising from 180°C to 230°C at 2° C/rain, the upper temperature being held for 10 min. Margaric acid (Sigma) was used as an internal standard (50 lag/ determination) and a calibration curve prepared for linoleic acid over the range 5 to 25 lag/ determination. Free fatty acids were determined by HPLC of their 3-nitrophenylhydrazides, prepared according to the method of Miwa et at. [15], on a 100 x 3 mm cartridge of Chromspher C18 (Chrompack, London, UK) using as eluant a gradient of 65 to 90~ acetonitrile in 0.1 M acetate buffer (pH 4.6) over 25 rain. Detection was by absorbance at 400 mnL

4. RESULTS AND DISCUSSION

4.1. Viability of B. pertussis in the presence of milks The effect of bovine and human milk preparations on viable counts of R pertussis is shown in Table 1. The pattern of growth of B. permssis in the presence of 25~ bovine milk was almost identical to that seen in SS medium alone, with approximately 20-fold increases in the number of viable bacteria after 24 h and a further doubling after 48 h incubation. However, in the presence of 25~ human milk the opposite effec~ was observed

271 with the viability dropping by about 20-fold after 24 h incubation and by a further 10-fold after 48 h incubation. The effect of the human milk appears to be bacteriostatic rather than bacteriocidal as approximately 0.5% of the initial bacterial inoculum was still viable after 48 h. These results ,/ere unexpected, as an earlier :eport suggested that bacteria which were susceptible to bacteriostatic effects of human milk were similarly affected by bovine milk [16].

ble 1) and the bacteria grew just as well in such a medium as in SS only. In addition to diff:rences in their lysozyme contents, bovine and human milks also possess different levels of lactoperoxidase activity. However, all reports (see ref. 19) so far show that there is more activity associated with bovine rather than human milk. Therefore it is unlikely that the lactoperoxidase system could account for the loss of viability in the presence of human milk.

4.2. Effect of variations in culture media on viability The lowered iron content of 4 p M in the SS medium was still sufficient for B. pertussis growth (Table 1). However, at this level most, if not all, of the iron present was bound to lactoferrin following the addition of the milk preparations and the bacteria were therefore in an essentially iron-restricted environment. B. pertussis has been previously shown to grow in such an environment [17] and as expected it grew just as well in the presence of bovine milk at either level of iron availability (Table 1). The reduction in iron levels also had no significant effect on the lo.~s of viability in the presence of human milk. The concentration of lysozyme in human milk is relatively high at approximately 100 p g / m l [18] whereas there is very little, usually less than 0.5 p g / m l [7], in bovine milk. Bovine milk, to which had been added lysozyme to bring it up to the same concentration as in human milk, still had no adverse effect on the viability of B. pertussis (Ta-

4.3. Lipid composition of milks and effect on viability HPLC analysis showed no qualitative difference in free fatty acids between bovine and human milks. Free linoletc acid was present in quantities below 2 p g / m l (the lower limit of quantification of the method) in both. However. analysis of the total fatty acids, including triglycerides, showed that all the bovine milks examined contained less than 5 ~tg/nfl of iinoleic acid whereas the human milks contained 15-24 pg/nd. It is possible that free fatty acids could be released from triglyceride stores by endogenous lipase activity. It has been reported that unsaturated fatty acids, including linoleic, are responsible for the killing of the parasite Giardia lamblia by human, but not bovine, milk [20]. It has also been shown that B. permssis is very sensitive to the inhibitory action of certain unsaturated fatty acids also in-

Table 1 Effect of milkson viabilityof B. pertussis HM, humanmilk; BM, bovinemilk; -Fe. iron content loweredto 4 pM: Lys. Lysoz)me. Medium

Numberof viablebacteria/ml " Incubationtime(h) 0

24

48

SS b SS+ HM b SS+ BM b SS-Fe+ HM c SS-Fe+ BM "

4.1 (1.3-7.0) 4.9 (2.2-7.2) 6.3 (2.2--11.0) 7.2 (5.3-9.1) 6.0 (3.8-8.2)

77 (54-117) 0.18 (0.I-0.25) 143 (62--220) 0.25 (0.14-0.35) 145 (69-220)

160 (60-190) 0.015 (0.01-0.021) 330 (140-810) 0.021 (0.008-0.033) 180 (170--190) 200 (170-230)

SS+BM+Lys

¢

7.8(6.0-9.6)

• Results are means of replicate ~ t ~ b Means from three replicateexperiments. c Means from duplicate experiments.

171

(146-195)

x l08. figures in paremhe*es = range of values.

272 Table 2 Effect of fatty acids on viabilityof B. pertussis HM, human milk; BM, bovine milk; LA, linoleic acid in/~8/ml of medium. Medium

SS b SS+ HM b SS + BM b SS + LA2.5c SS+ LAS.0 c SS+ BM+ LA2.5 c SS+ BM+ LA5.0 c

Number of viablebacteria/nil • Incubation time (h) 0

24

48

2.3 (2 ? -: .5) 2.4 (1.2-3.5) 2.3 (1.8-2.9) 3.8 (2.9-4.7) 5.3 (2.2-8.3) 4.1 (3.2-5.0) 5.0(2.8-7.1)

77 (26-113) 0.104 (0.027-0.254) 96 (53--150) 58 (28-87) 18 (13-23) 60 (29-90) 93 (62-123)

235 (60-400~ 0.015 (0.007-0 b23) 190 (100-25C) 140 (100-180) 120 (110-130) 405 (310-500) 105 (90-120)

a Results are means of replicateexperiments x 10s; figures in parentheses= range of values. b Means from three replicate experiments. c Means from duplicate experiments.

chiding finoleic [21]. It therefore seemed likely that the higher levels of linoleic acid in h u m a n milk could account for its bacteriostatic action. The addition of linoleic acid, at concentrations of 2.5 or 5.0 t t g / m l , to SS plus 25% bovine milk raised the linoleic acid concentration to the same range as when h u m a n milk was used, i.e. approximately 15 to 25 ~tg/ml of milk. However, this had no greater inhibitory effect on the growth of B. pertussis (Table 2) than that seen when bovine milk alone was added. Addition of linoleic acid to SS medium alone also had n o detectable inhibitory effect (Table 2) so discounting the possibility that its action was being blocked by other components of bovine milk. It appears that h u m a n milk possesses an ability to inhibit R pertussis which is not shared by bovine milk and that this property is independent of the effect of lactoferrin, lysozyme levels, the lactoperoxidase system and differences in the fatty acid composition. As B. pertussis m a pathogen of the upper respiratory tract and as a suckling child can easily inspire milk [10] this property m a y offer some degree of protection against whooping cough between birth and immunization. This could be particularly significant for children who receive no maternal antibodies to B. pertussis through cord blood or breast milk. If possible, it would be interesting to examine the epidemiology of pertussis in preqmmunized children to ascertain whether

breast feeding shows an inverse .,:orrelation to infection independent of maternal antibody levels to B. pertussis.

REFERENCES [1] Fleming,A. (1922) Proc. R. Soc. Lond. Ser. B 93, 306-317. [2] Rosenthal, L and Lieberman, H. (1931)J. Infect. Dis. 18, 226-239. [3] Oram, J.D. and Reiter, B. (1968) Biochim. Biophys. Acta 170, 351-365. [4] Wright, R.C. and Tramer, :L (1957) J. Dairy Rcs. 24, 174-183. [5] Jago, G.R. and Morrison, M. (1962) Prec. Soc. Exp. Biol. Med. 111,585-588. [6] Reitcr, B., Picketing, A. and Oram, J.D. (1964) in 4th International Symposium on Food Microbiology(Molin, N., ed.), pp. 297-305, Almqvistand Wicksell, Uppsala. [7] Vakil, J.R., Chandan, R.C, Parry, R.M. and Shahani, K.M. (1969) J. Dairy Sci. 52, 1192-1197. [8] Arnold, R.R., Brewer. M. and (]anthier, J.J. (1950) Infect. lmman. 28, 893-898. [9] Reiter, B. and Harnulv, G. (1984) J. Food Protect. 47, 724-732. [10] Yamada, S. and Sakurai, N. (1988) Jap. J. Hyg, 43, 638-644. [11] Stainer, D.W. and Schohe, M.J. (1971) J. Gen. Microbiol. 63, 211-220. [12] Gearing, AJ.H., Bird, C.R., Redhead, K. and Thomas, M. (1989) FEMS Microbiol Immunol. 47, 205-212. [13l Redhead, K. (1985) J. Med. Microbiol. 19, 99-108. [14] Wait, 17,.and Hudson, M. (19b.~ Lett. Appl. Microbiol. 1, 95-99.

273

[15] Miwa, H., Yamamoto, M., Nishida, T., Nunoi, K. and Kikuchi, M. (1987) J. Chromatogr. 416, 237-245. [16] Boesman-Finkelstein, M. and Finkelstein, R.A. (1985) FEMS Mierobiol. Left. 27, 167-174. [17] Redhead. K., Hill, T. and Chart, H. (1987) J. Gan. Microbiol. 133, 891-898. [18] Goldman, A.S.. Garza, C.. Nichols. B.L. and Goldblam. R.M. (1982) J. Pediatr. 100, 563-567.

[19] Relier, B. (1984) in Human Milk Banking (Williams, A.F. and Baum, J.D., Eds.), pp, 29-53, Raven Press, New York. [201 Rohrer, L., Winterhaher, K.H., Eckert, J. and Kohler, P. (1986) Antimierob. Agents Chemother. 30, 254-257. [21] Field, L.H. and Parker, C.D. (1979) J. Clin. Microbiol. 9, 651-653.