Protective
Factors
in Milk and the Development
of the Immune
System
Lars
A.
Hanson,
MD, PhD,
Staffan
Ahlstedt,
PhD,
Bengt
Andersson,
MD, Barbro
Carlsson,
PhD,
Sven
P. F#{228}llstr#{246}m,
MD,
Lotta
Mellander,
MD, Oscar
Porras,
MD,
Tommy S#{246}derstr#{246}m,
MD, PhD,
and
Catharina
Svanborg
Eden,
MD, PhD
From the Departments of Clinical Immunology and Pediatrics, University of GOteborg, GOteborg, Sweden
ABSTRACT. The neonate is immature in certain
immu-nologic functions. The slow development of secretory immunoglobin A (IgA) seems to be compensated by Se-lective transfer of secretory 1gM into exocrine secretions on mucous membranes during the first few months of life. Secretory IgA and secretory 1gM antibodies against
Escherichia coli and poliovirus are already found in the neonate, possibly in response to the maternal
anti-idi-otypic IgG antibodies transplacentally exposing the fetus. Via such a mechanism, food antibodies could occur before direct food exposure in the infant. Human milk provides large amounts of antibodies (as a crude comparison, about 50 times the amount of antibodies given to a patient with hypogammaglobulinemia). The milk antibodies, domi-nated by secretory IgA, protect especially against intes-tinal infections. The milk also contains oligosaccharide analogues to epithelial receptors for bacteria. They, as
we!! as a number of milk components such as lactoferrin
and lysozyme, may contribute to host defense. The food antibodies in human milk may influence the infant’s immune response to foreign food proteins introduced during weaning. (Pediatrics 1985;75(suppl):172-176; mu-cosal immunity, secretory IgA, secretory 1gM, milk-medi-ated protection, food antibodies.
The neonate is immunologically immature, and
this immaturity includes deficiencies in certain T-and B-lymphocyte functions.’ Immunogbobulin M (1gM) antibodies are efficiently formed, however, and serum levels increase rapidly during the first few weeks of life.2 The prolonged 1gM responses seen in the neonatal period3 may be caused by
Read before the Workshop on Current Issues in Feeding the Normal Infant, Palm Springs, CA, April 8-11, 1984.
Reprint requests to (L.A.H.) Department of Clinical Immunol-ogy, University of GOteborg, Goteborg, Sweden.
PEDIATRICS (ISSN 0031 4005). Copyright ( 1985 by the American Academy of Pediatrics.
deficient T-hebper cell function. The slow appear-ance of IgG2 and IgG4 and also of IgG1 and IgG3 antibodies during the first 2 years of life can be related to a paucity of B lymphocytes capable of producing these subclasses.4 The active transport via the placenta of maternal IgG1-4 richly
compen-sates this IgG deficiency of the young infant. In contrast, the IgA antibodies in the form of serum IgA, as well as secretory IgA (SIgA), are slow to appear in the infant.5 Mother’s milk is, however, a very rich source of SIgA.6 This report will briefly review the functional aspects of breast milk im-munity and aspects of the developing immune sys-tem with regard to defense against infections and allergies in the light of recent data.
ONTOGENY OF MUCOSAL IMMUNITY
Most infections and allergies occur after exposure of the mucosal membranes to noxious agents. The efficacy of factors providing mucosal defense in the
young is therefore of great interest. Studies of the ontogeny of SIgA, presumably important in the protection of mucosae, have given quite variable results (see Hanson et a15 for a recent review). In a
study of the development of salivary SIgA anti-bodies against Escherichia coli 0 antigens, Melban-den et al7 found such antibodies even during the
first day of life. They slowly increased, approaching adult levels toward the end of the first year of life. Infants hospitalized for surgical diseases often had
1gM antibodies, efficiently produced by the young infant, were transferred into the secretion via the transport mechanism normally used by SIgA, and this transfer included the secretory component pro-vided by the exocnine glandular epithelium. Thus, when IgA is low or absent, the transport mechanism becomes available for 1gM instead. In the IgA-deficient patient a similar compensatory appear-ance of secretory 1gM occurs.8 Despite their poorer stability compared with SIgA, it is possible that these secretory 1gM antibodies can play a role in mucosab defense until SIgA takes oven after the first few months of life.
Cord blood and saliva from the first few hours of life contained IgG antibodies, presumably tnanspba-centally obtained. In addition there were often bow levels of IgA and 1gM in the cord blood and SIgA and 1gM in the saliva with antibody activity against
E coli 0 antigens as well as poliovirus.8 The latter finding was remarkable as poliovirus virtually does not occur in Sweden because of efficient vaccina-tion programs using inactivated vaccine. Neither secretory IgA nor 1gM is known to pass from mother to fetus via the placenta. If the SIgA and 1gM polio antibodies are of fetal origin, how can they occur when the antigen is not available?7 We hypothesize that the maternal tnansplacentalby transferred IgG contains not only antipolio, but also anti-antipolio antibodies. Such anti-idiotype antibodies are as specific for the binding sites of the antipobio antibodies as is the poliovirus antigen. Such anti-idiotypic antibodies are able to induce production of the idiotypic antibodies, in this case antipolio antibodies. If this assumption is correct, an infant might also become sensitized to food proteins in fetal life via maternal anti-antifood antibodies. This might provide an explanation for the occasional occurrence of allergic reactions very early in life, without any known previous exposure to the allergen.
The capacity of anti-idiotypic antibodies to influ-ence the immune response of the neonate has been demonstrated in a mouse model. Newborn mice were given monocbonab anti-idiotypic antibodies against the E coli K13 polysacchanide capsule (anti-anticapsule). At the 50-ng bevel (given neonatabby), the anti-idiotypic antibodies enhanced the antibody response to immunization with a K13-polysaccha-ride at 12 weeks of age.’#{176}The response favored the cell clone producing the anti-K13 antibody, which carried the same idiotype against which the mono-cbonal anti-K13 antibody was directed, illustrating a new possibility of directing the character of the antibody response. In practice, it might become possible to select and enhance a clone that syn-thesizes antibodies efficiently protective against an infection, on possibly blocking an allergen.
S#{246}derstn#{246}met al” demonstrated that anti-idi-otypic antibodies reaching the mouse offspring via the milk also primed the homologous idiotype-pro-ducing clone. Thus, in addition to providing
nutni-ents and passive immunity, breast-feeding might
actively influence the immune system of the
off-spring.
HUMAN
MILK
AND
PROTECTIVE
FACTORS
Human milk contains some 1gM and little IgG, but SIgA is the dominant antibody component.6 Presumably, the SIgA antibodies are formed locally in the mammary glands; there seem to be sufficient numbers of IgA producing cells in the glands to account for their synthesis.’2 Experiments in mice have led some investigators to suggest that a pon-tion of milk IgA originated from serum’3; IgA
di-mers from serum would be transferred into milk by
binding to an available secretory component in the mammary gland. Studies in rats and mice, however, have not shown that milk IgA is derived from
serum. 14,14a
Some of the SIgA antibodies of human milk are directed against food proteins ofthe mother’s diet.’5 The higher her intake of a certain protein, the higher is her milk SIgA antibody against that pro-tein. The milk SIgA is also directed against the microorganisms exposing the mother, mainly via the gut,’6 but also via the lungs.’7 Experiments in animal models and man have shown that intestinal antigenic exposure can result in homing of lymph-oid cells from the Peyer’s patches to the various
sites of local SIgA production, including the mammary glands.’8”#{176} This phenomenon is cabled the enteromammanic link. As a consequence, the breast-fed infant receives milk SIgA antibodies against the microbes and food proteins to which the mother has been exposed and to which the infant will also be exposed.
The dimenic IgA antibody molecule is stabilized with covalent bonds to the secretory component. The resulting SIgA molecule resists proteolytic en-zymes better than serum antibodies.2#{176} This makes the SIgA molecule apt to function, eg, on the mu-cous membranes of the gut. SIgA antibodies do not activate complement, enhance phagocytosis, or in-duce inflammation. They only bind to the antigen, be it a bacterial, viral, or food constituent. In this way, the SIgA provides an “antigen avoidance sys-tem”2’ preventing contact between the gut mucosa and potentially pathogenic microorganisms or of-fensive food.
making it likely that the milk SIgA can be impon-tant in the defense against infection of the infant.
Breast-feeding protects against gastrointestinal infections according to a number of studies.2224 Protection against cholera in the breast-fed baby could be related to the level of the mother’s milk IgA antibodies against Vibrio cholerae bipopolysac-charide and enterotoxin.25
Experiments were performed with a rat model using intragastnicably administered E coli Ki with type 1 pili which cause bacteremia in neonatal rats. Immunization of rat dams against the type 1 pibus antigen protected the offspring. By switching the
offsprings between immunized and unimmunized
dams, it could be shown that protection was most efficiently mediated via the milk.26 This is of inter-est because we have seen that parenteral immuni-zation with inactivated cholera vaccine or macti-vated polio vaccine during lactation can increase not only the serum, but also the milk SIgA titers.27’28
This effect was obtained only in mothers who had been previously naturally exposed to these
infec-tious agents. For unknown reasons, the perorally administered live polio vaccine often decreased pre-vaccination milk SIgA polio antibody bevels.28’29
Breast-feeding may protect against respiratory tract infections and otitis media.’32 Pneumococci and Haemophilus influenzae together account for
about 70% of cases of otitis media in infants. Such bacteria adhere to human pharyngeal cells,33’34 but in the presence of human milk this attachment is completely prevented (B. Andersson, 0. Porras, L.
A.
Hanson, et a!, unpublished data). This activityis mediated by the antibody-rich high-molecular-weight milk fraction, but does not correlate well with the titers of antibodies against phosphoryl choline, H influenzae capsule b, or outer membrane proteins as measured with enzyme-linked immu-nosorbent assay (ELISA). A bow-molecular-weight fraction without antibodies also prevented the at-tachment of pneumococci. This fraction is rich in oligosacchanides, including the disaccharide that is
part of the specific receptor for pneumococci iden-tified by Andensson et al.36 It is possible that such analogues of microbial epithebial receptors in milk are important for mucosab defense of the infant. Bovine milk and certain bovine milk-based formu-las showed a weak anti-attachment activity, which
is not seen with buffalo milk (B. Andersson et al, unpublished data).
Human milk contains numerous host defense factors in addition to antibodies and receptor ana-bogues. Lactofernin, lysozyme, and others are anti-bacterial, but their biologic significance has not been tested.
There are about 106 leukocytes per milliliter in colostrum and early human milk. Among these are
macnophages and granubocytes with phagocytic ca-pacity. Their robe is unknown. There are also T and B lymphocytes. According to some reports, these B lymphocytes do produce IgA antibodies37; according to others they do not.38 Recently, S#{246}denstr#{246}met ab” established in vitro cultures of human cobostrab cells
by transformation with Epstein-Barn virus.” Using immunofluorescence and the enzyme-linked im-munospot (ELISPOT) technique,39 antibody pro-duction could be demonstrated after 8 months of in vitro culture in about 80% of the cells with 1gM production dominating, followed by IgA and IgG. The dominance of 1gM producers may be explained by their more efficient stimulation due to higher numbers of Epstein-Barn virus receptors on these cells. So far only IgAl and no IgA2 synthesis has been demonstrated in these preliminary studies. The biologic role, if any, of the milk cells is still unknown.
HUMAN
MILK
AND
FOOD
ANTIBODIES
Food intolerance in infancy is common, and im-munobogic mechanisms may often be involved. Fast-appearing symptoms after challenge in pa-tients with bovine milk allergy may be mediated by IgE antibodies, but beyond that we are largely ignorant about possible immunologic components. Cell-mediated immunity, as well as immune corn-plex-induced inflammation may also be responsible for symptoms, especially in patients reacting more slowly to oral challenge. With immune complexes as a possible part of the pathogenesis, serum IgG
antibodies to food could indicate a capacity of the patient’s immune response to incite inflammation.
difference was enhanced if we used pepsin-degraded fl-lactogbobulin. It may be that the native proteins do not expose the determinants inducing untoward reactivity. Infants with rapid-onset symptoms to bovine milk exposure did not show any increased bevels of serum IgG on IgA antibodies to bovine milk proteins.
In a preliminary study, we saw significantly lower levels of serum IgG antibodies against bovine milk proteins in nonallergic infants weaned gradually oven a period of more than 3 weeks with concomi-tant breast-feeding compared with infants trans-fenred directly to a bovine milk formula.43 In an extended study, we have performed multiple regres-sion analyses relating the serum antibody response against bovine milk proteins to the age at weaning, at first bovine milk exposure, and at sampling. The serum IgG antibodies were rebated directly to the length of cow’s milk exposure after the cessation of breast-feeding and inversely to the age at first exposure to bovine milk. These data show an effect of breast-feeding on the immune response to bovine milk exposure, but the investigated variables ex-plained only 17% of the variation of the antibody levels.44 IgG antibodies against f.-bactogbobulin showed significant negative correlations to the age at weaning and the age at first bovine milk expo-sure, but only 8% of the variation was explained. One implication of these findings is that other, yet unknown factors are of greaten importance for the induction of the IgG food antibodies. “Oral tolen-ance” with decreasing serum antibodies after oral antigen exposures obviously does not occur regu-banly in the young infant.45
The appearance of food antibodies, including IgE during the first few months of life, before known food exposure,46 as suggested above, may result from exposure of the fetus to maternal transplacen-tally transported anti-idiotypic IgG antibodies. It
is also possible that food proteins can reach the fetus via the maternal circulation on after birth via the breast milk. As early as 1931, Gy#{246}ngysuggested that human milk could contain egg antigens.46a Recently, we have demonstrated soy protein in human milk. Kilshaw and Cant47 found fl-bactogbob-ulin, ovalbumin, and ovomucoid at the nano-pico-gram level. It is surprising that these food proteins are found in the milk at the same time as there are homologous antibodies. Possibly these antibodies are of low affinity; Kilshaw and Cant found no evidence of immune complexes in the milk.
ACKNOWLEDGMENT
This study was supported by grants from the Medical Faculty of Goteborg University, the Swedish Medical Research Council (No 215), the Wennergren-Arla
Foun-dations, the First of May Flower Fund, and the Ellen,
Walter and Lennart Hesselman Foundation for Scientific
Research, Sweden.
REFERENCES
1. Hayward AR, Development of immunity mechanism in Soothill JF, Hayward AR, Wood CBS (eds): Paediatric Immunology. Oxford, England, Blackwell, 1983, p 48 2. Berg T: Immunoglobulin levels in infants with low birth
weights. Acta Paediatr Scand 1968;57:369
3. Smith MD, Evans J, Steel CM: Age-related variations in proportion of circulating T cells. Lancet 1974;2:922
4. Andersson U, Bird AG, Britton 5, et al: Humors! and cellular immunity in humans studied at the cell level from birth to two years of age. Immunol Rev 1981;57:5
5. Hanson LA, Carlsson B, Dahigren U, et al: The secretory IgA system in the neonatal period. Ciba Found Symp
1980;77:187
6. Hanson LA: The mammary gland as an immunological
organ. Immunol Today 1982;3:168
7. Mellander L, Carisson B, Hanson LA: The appearance of secretory 1gM and IgA antibodies to Escherichia coli in saliva
during early infancy and childhood. J Pediatr 1984;104-564 8. Brandtzaeg P: Human secretory immunoglobulin M: An
immunochemical and immunohistochemical study. Immu-nology 1975;29:559
9. Deleted in proof
10. S#{246}derstr#{246}mT, Stein K-E: Neonatal priming of mice with monoclonal idiotypic or anti-idiotypic antibodies and pro-tection against E coli K13 infection in mice. Presented at the Fifteenth Annual Meeting of the Scandinavian Society for Immunology, 1984
11. S#{246}derstr#{246}mT, S#{246}derstr#{246}mR, Hanson LA: The leucocytes of human breast milk. Proceedings of the Symposium on the Nutritional Aspects of Breast Milk, Oregrund, in press 1984 12. Brandtzaeg P: The secretory immune system of lactating
human mammary glands compared with other exocrine or-gans, in Ogra PL, Dayton DH (eds): Immunology of Breast Milk. New York, Raven Press, 1979, p 91
13. Halsey JF, Mitchell C, Meyer R, et al: Metabolism of im-munoglobulin A in lactating mice: Origins of immunoglob-ulin A in milk. Eur J Immunol 1982;12:107
14. Dahlgren U, Ahlstedt 5, Hedman L, et a!: Dimeric IgA in the rat is transferred from serum into the bile but not into milk. Scand J Immunol 1981;14:95
14a.Koertge TE: The quantitative capability of the enzyme-linked immunosorbent assay (ELISA) to study transport of antibody into secretions, dissertation. University of Iowa,
Ames, IA, 1984
15. Cruz JR, Garcia B, Urrutia JJ, et a!: Food antibodies in milk from Guatemalan women. J Pediatr 1981;99:600
16. Hanson LA, Carlsson B, Cruz JR, et a!: Immune response in the mammary gland, in Ogra PL, Dayton DH (eds):
Immunology of Breast Milk. New York, Raven Press, 1979,
p 145
17. Pen BA, Theodore CM, Losonsky GA, et a!: Antibody content of rabbit milk and serum following inhalation or ingestion of respiratory syncytia! virus and bovine serum albumin. Clin Exp Immunol 1982;48:91
18. Goldblum RM, Ahlstedt 5, Car!sson B, et a!: Antibody forming cells in human colostrum after oral immunization.
Nature 1975;257:797
19. Lamm ME, Weisz-Carrington P, Roux ME, et a!: Mode of induction of an IgA response in the breast and other secre-tory sites by oral antigen, in Ogra PL, Dayton DH (eds):
Immunology of Breast Milk New York, Raven Press, 1979, p 105
20. Lindh E: Increased resistance of immunoglobulin A dimers to proteolytic degradation after binding of secretory
corn-ponent. J Immunol 1975;114:284
21. Soothill JF: Immunological aspects of infant feeding, in
22. Chandra RK: Prospective studies of the effect of breast-feeding on incidence on infection and allergy. Acta Paediatr Scand 1981;68:691
23. Larsen SA, Homer DR: Relation of breast versus bottle feeding to hospitalization for gastroenteritis in a middle class US population. J Pediatr 1978;92:417
24. Cunningham 55: Morbidity in breast-fed and artificially fed infants. II. J Pediatr 1979;95:685
25. Glass RI, Svennerholm A-M, Stoll BJ, et a!: Protection against cholera in breast-fed children by antibodies in breast
milk.N EngI J Med 1983;308:1389
26. SOderstrOm T, Brinton CC, Cole MF, et a!: The rnurine immune response against Escherichia coli capsular polysac-charides and type 1 pili: Protection against invasive disease, dissertation. University of Goteborg, Goteborg, Sweden,
1984
27. Svennerho!m A-M, Hanson LA, Holmgren J, et a!: Different secretory IgA antibody response to cholera vaccination in Swedish and Pakistani women. Infect Immun 1980;30:427 28. Svennerholm A-M, Hanson LA, Holmgren J, et a!: Antibody
response to live and killed poliovirus vaccines in the milk of Pakistani and Swedish women. J Infect Dis 1981;143:707 29. Hanson LA, Car!sson B, Jalil F, et a!: Different secretory
IgA antibody responses after immunization with inactivated and live poliovaccines. Rev Infect Dis, in press 1984
30. Deleted in proof 31. Deleted in proof
32. Saarinen UM: Prolonged breast feeding as prophylaxis for recurrent otitis media. Acta Paediatr Scand 1982;71:567 33. Andersson B, Eriksson B, Falsen E, et a!: Adhesion of
Streptococcus pneumoniae to human pharyngeal epithelial cells in vitro: Differences in adhesive capacity among strains isolated from subjects with otitis media, septicemia, or
men-ingitis or from healthy carriers. Infect Immun 1981;32:311 34. Porras 0, Carlsson B, Hanson LA: Detection of soy protein
in soy lecithin, margarine, and, occasionally, soy oil.
mt
Arch Allerg AppI Immunol, in press, 1984 35. Deleted in proof
36. Andersson B, Dahrn#{233}n J, Frejd T, et a!: Identification of an
active disaccharide unit of a glycoconjugate receptor for pneumococci attaching to human pharyngeal epithelia! cells. J Exp Med 1983;158:559
37. Murillo GJ, Goldman AS: The cells of human colostrum: II. Synthesis of IgA and fllc. Pediatr Res 1970;4:71
38. Crago 55, Prince SJ, Putlow TG, et a!: Human colostral cells: I. Separation and characterization. Clin Exp Immunol 1979;38:585
39. Czerkinsky CC, Nilsson LA, Nygren H, et a!: A solid-phase enzyme-linked immunospot (ELISPOT) assay for
enumer-ation of specific antibody-secreting cells. J Immunol Meth-ods 1983;65:109
40. Lippard VW, Schloss OM, Johnson PA: Immune reactions induced in infants by intestinal absorption of incompletely
digested cow’s milk proteins. Am J Dis Child 1936;51:562
41. Kletter B, Gery I, Freier 5, et a!: Immune responses of normal infants to cow’s milk: I. Antibody type and kinetics
of production. Int Arch Allergy AppI Immunol 1971;40:656
42. F#{228}!!str#{246}mSP, Ahlstedt 5, Hanson LA: Specific antibodies
in infants with gastrointestinal intolerance to cow’s milk protein. Int Arch Allergy AppI Immunol 1978;56:97
43. Hanson LA, Ah!stedt 5, Carlsson B, et al: Secretory IgA
antibodies against cow’s milk protein in human milk and
their possible effect in mixed feeding. Int Arch Allergy Appl Immunol 1977;54:457
44. F#{228}!lstr#{246}mSP, Ahlstedt 5, Carlsson B, et a!: Influence of breast feeding on the development of cow’s milk protein
antibodies and the IgE level. Int Arch Allergy Appl Immunol 1984;5:87
45. Tomasi TB: Oral tolerance. Transplantation 1980;29:353
46. Hattevig G, Kjellman B, Johansson SGO, et a!: Clinical
symptoms and IgE responses to common food proteins in
atopic and healthy children. Clin Allergy 1984;71:8
46a.Gy#{246}rgy P: Eiklar-Antigen in der Fravenmilch. Z. Kinder-heilk 1931;132:1
47. Kilshaw PJ, Cant AJ: The passage of maternal dietary proteins into human breast milk. Int Arch Allergy, in press
