Immunology of Breast Milk

Marietta Xanthou

Neonatal Intensive Care Unit, Aghia Sophia Children's Hospital, Athens 11587, Greece

A most important adaptation of the gastrointestinal (GI) tract of the newborn to the extrauterine environment is the development of a mucosal barrier against the pénétration of bacteria, viruses, and other environmental antigens, i.e., food pro teins. The mechanisms of host défense at mucosal sites, both immunologie and nonimmunologic, are underdeveloped during the neonatal period and more so in prématurés and small for date neonates (1). The intestinal epithelial cells lack a well-defined brush border and retain a prim itive transport mechanism for endocytosis of large molécules (2). It has been proven that the neonatal intestine may absorb larger quantities of ingested antigenic proteins and proliferating microorganisms than the adult intestine (3). The number of the Peyer's patches is small in mammals during the neonatal period and although récent studies hâve described T-helper and T-suppressor lym phocytes, B-lymphocytes, and macrophages on their lymphoid follicles (4), the functional ability of the cells seems to be immature until about 1 month of âge (5). Furthermore, plasma cells, which are normally abundant in the mucosal lam ina propria, are not detected in utero or during the immédiate neonatal period (6). Taking into considération ail of thèse immaturities and the absence of mucosal antigenic stimulation during intrauterine life, it is no wonder that human and other mammalian neonates exhibit characteristic absence of secretory IgA (slgA) during the early postnatal period. Shortly after birth, however, the ingestion of food antigens and the colonization of the gut with various microorganisms provide enough antigenic stimulation to evoke a secretory immune response. By 1 month of âge, approximately ail infants will hâve some détectable slgA (7). The availability of immunological factors in the sécrétions of the mammary gland may represent an evolutionary event to compensate through breast feeding for the transient deficiency of mucosal immunity in the neonate. Indeed human colostrum and milk bear many similarities to the external body fluids bathing the mucosal surfaces of the GI and respiratory tracts. Evidence so far also suggests that the immunoglobulin-producing cells observed in the mammary tissue during lactogenesis and in the colostrum and milk are of mucosal origin (8). Fresh human milk contains many components that provide spécifie as well as nonspecific défenses against infectious agents or other macromolecules (9-11). 101 102 IMMUNOLOGY OF BREAST MILK

The contribution of colostrum to neonatal immunity involves both humoral and cellular factors.

HUMAN MILK ANTIBODIES

The most important of the humoral immunological factors in human milk are its antibodies, the principal immunoglobulin being the 11S secretory IgA. The slgA exists as a dimer of two 7S IgA molécules linked together by a polypeptide chain, called J-chain, and associated with a nonimmunoglobulin protein referred to as the secretory component. The IgA dimers produced by the plasma cells of the mammary gland are transported to specialized columnar epithelial cells where they unité with the secretory component (SC) and are discharged into the alveolar space (Fig. 1) (12,13). Linkage of SC to the IgA dimer provides most mammals with antibody molé cules that can coexist with the proteolytic enzymes of intestinal sécrétions (14). Thus, about 70% to 75% of the ingested colostral IgA has been found to survive passage through the gut of the newborn and is excreted in the fèces (15). The slgA constitutes about 20% of the total protein content of milk (16). Apart from slgA, other immunoglobulins such as 7S IgA, IgG, IgM, IgD, and IgE hâve been found in human colostrum and milk but in much smaller quantities (17). The highest levels of IgA and IgM are présent during the first 5 days of lactation (Fig. 2) (18) with levels of IgA ranging fourfold to fivefold higher than IgM, 20- to 30- fold higher than IgG, and sixfold higher than sérum IgA. As lactation progresses slgA as well as IgM décline, whereas IgG levels do not change significantly. After their rapid décline during the first days of life, IgA levels in milk décline more slowly until 3 months postdelivery. From then on and up to 12 months, the IgA levels are maintained at a stable level (19). However, when estimating the amount of immunoglobulins a breast-fed baby receives, one should also consider the in- creasing amounts of milk the baby is taking when getting older. It has been esti- mated that the breast-fed infant receives on average about 1 g of IgA each day and approximately 10 mg of IgM and IgG (20). Some mammals, e.g., piglets and calves, acquire ail the antibodies necessary for their protection from colostrum and milk by absorption through the gut wall during the first few hours of life. When deprived of milk, they die from septicemia (21). In man, passive immunity is derived almost entirely from the intrauterine trans port of maternai antibodies (2). Until receritly it was believed that in human neonates immunoglobulins of breast milk could not cross the intestinal wall. How ever, Ogra et al. (15) hâve shown that small quantities of milk IgA antibodies can be absorbed immediately after birth by human neonates. IgA immunoglobulin contains antibodies against many pathogens and potential pathogen viruses and bacteria, including several Enterobacteriaceae. Some spé cifie antibodies discovered in human milk are listed in Table 1. IgA antibodies found in milk possess specificity for infective agents endémie to IMMUNOLOGY OF BREAST MILK 103

Intestinal Lumen

y il lus Tip

Bosement Membrane

Secretory Component

DIMERIC \O/P0LYMERIC MONOMERIC IqA vwyIgM IgE /MONOMERIC IqG Blood Vessel

FIG. 1. Transport mechanisms for intestinal antibodies. (From Walker and Isselbacher, réf. 13.)

500 Γ

Q- ε

FIG. 2. Géométrie mean levels of IgG, IgM, and IgA in colostrum and milk of 200 female subjects at various intervais after 5 30 60 (80 the onset of lactation. (Adapted from Ogra Days Post Parfum and Ogra, réf. 18.) 104 1MMUNOLOGY OF BREAST MILK

TABLE 1. Spécifie antibody reactivity in human colostrum and milk

Bacteria Vi ruses Other

Escherichia coli (Ο + Κ Rotavirus Candida albicans antigens and enterotoxin) Streptococcus pneumoniae Polio virus types 1,2,3 Giardia Clostridium tetani Rubella virus Chlamydia Bordetella pertussis Echo viruses Food proteins Salmonella Coxsackie viruses A,B Shigella Respiratory syncytial virus Vibrio cholerae Cytomegalovirus Bacteroides fragilis Influenza A virus Streptococcus Mutans Herpès simplex Hemophilus influenzae type Arboviruses Β Semliki Forest Ross River Japanese Β Dengue or pathogenic for the intestinal and respiratory tracts. They may be présent in the milk and absent in the sérum. Pregnant women, given Escherichia coli 083 orally, had IgA antibodies in their milk but not in their sérum (22). Other investigators hâve observed similar responses using intrabronchial immunization in rabbits (23). Thus antibody response in the milk appears to be the characteristic feature of antigenic exposure in the gut associated lymphoid tissue or the bronchoepithelium associated lymphoid tissue, and strongly supports the concept of an enteromammary as well as a bronchomammary axis of immunologie reactivity in the breast. The more probable explanation for the appearance of milk antibodies after antigenic exposure in various mucosal sites in humans is that spécifie IgA-producing cells home to the mammary gland after they hâve been exposed to the antigen at thèse sites as for example in the Peyer's patches. There are also indications that the homing to the mammary gland is dépendent on hormonal factors (24). Other information indicates that homing may be further directed by glandular cell Ia- antigens coded for by the major histocompatibility complexes (25). As breast-fed babies receive considérable amounts of slgA antibodies directed against a variety of microorganisms that dominate in their surroundings, milk anti bodies can be used as an epidemiological tool.

What Is the Functional Rôle of Milk Antibodies? A number of action mechanisms hâve been suggested for secretory milk anti bodies such as opsonization and, in particular, ability to promote phagocytosis by macrophages, complément fixation, bactericidal, antitoxic, and antiviral activities as well as aggregation of antigens (17,26). However, there is now substantial evi- MMUNOLOGY OF BREAST MILK 105 dence that a major function of intestinal antibodies is the process of immune exclu sion at the mucosal surface (2), thus preventing infections and allergies. Finally, it has been shown that immunization or vaccination of the mother im- proves the passive immunity transferred to the baby via breast milk as antibodies (27,28).

CELLULAR IMMUNE FACTORS

The cells of human colostrum and milk hâve been studied relatively little although they were discovered more than a century ago. The cells of human colos trum were apparently first observed by the pioneer biologist-microscopist-photog- rapher Alexandre Donné who obtained their first photographie portrait in 1839 (29). He called them corpuscles and was unable to recognize their cellular nature. The corpuscles of Donné were found to be cells 20 years later, in 1868, when Beigel examined stained préparations of colostrum microscopically (30). How- ever, in those early days the functions of thèse cells were not examined and re- mained virtually unknown to immunologists. In 1966 Smith, in Goldman's laboratory, examining the "débris" that were ob tained after centrifuging colostrum, was surprised to find what Donné had ob served 122 years before. Phase microscopy revealed that colostrum was packed with living motile cells (31). Large numbers of leukocytes are présent in human milk, particularly during the first weeks of lactation. Early colostrum contains the highest concentration of cells, about 1-3 Χ 106-8 cells/ml (17). By the end of the first week of lactation cell concentration is of the order of 105 cells/ml but because of the increase in the quantity of milk offered, the total number of cells available to the neonate remains large (24). The leukocytes seen in colostrum are usually macrophages (40-50%), polymorphonuclear neutrophils (40-50%), and lymphocytes (5-10%). Numerous epithelial cells and epithelial cell-fragments are also seen in human colostrum. Later on, ma ture human milk cells consist of 85% to 90% macrophages, the remainder being mainly lymphocytes and very few epithelial cells or polymorphonuclear neutro phils. The proportion of the various white cell types does not change during breast feeding (32). Morphologically, the neutrophils are similar to those seen in blood except that they contain numerous fat globules and fewer granules. Macrophages also contain many fat globules and other granular material (Fig. 3). The typical colostral cor- puscle of Donné is a very large, lipid-laden macrophage. Lymphocytes appear similar to those found in peripheral blood.

Macrophages

Milk macrophages can be identified as being 20 μ to 40 μ in diameter. They are motile cells that adhère to glass. Histochemically, the milk macrophages differ 106 MMUNOLOGY OF BREAST MILK

FIG. 3. Typical human milk macrophages. from the blood monocyte by having increased numbers of lysosomes and somehow decreased peroxidase staining. Furthermore, their cytoplasm contains significantly higher amounts of IgA (33). The origin of milk macrophages has not been clearly defined yet. They could corne from blood monocytes that begin to home to the mammary gland during the last months of pregnancy and differentiate there to activated-appearing macro phages (9). There is little information about the function and rôle of the macrophages of breast milk. It has been suggested that macrophages in colostrum and milk are presumably involved in the défense against infections both in the breast and in the neonatal gut. In view of the extremely high proportion of macrophages in colos trum, it is tempting to postulate that thèse cells may be more important to the neo natal intestine than to the mammary gland. Because their cytoplasm contains large amounts of slgA antibodies, it has been suggested that thèse cells can function as a vehicle for transport or storage of immunoglobulins in breast milk (34). In rats, Pitt et al. (35) hâve shown that a Klebsiella-'mduced necrotizing enterocolitis could be prevented by feeding rat colostral macrophages but not by feeding soluble colostral factors. Thèse findings are in agreement with the observation that necrotizing enterocolitis is a rare disease in infants fed with untreated breast milk (36). Milk macrophages hâve been shown to possess Fc and C» receptors (35). Two récent studies performed to evaluate in vitro the ability of milk phagocytes (macro phages and polymorphonuclear neutrophils) to phagocytose and kill pathogenic organisms hâve shown good phagocytic activity against Staphylococcus aureus, E. coli and Candida albicans; however, their killing ability against E. coli has been found good in one (37) but quite poor in the other study (38). Milk macrophages hâve also been shown to participate in antibody-dependent cell-mediated cytotoxicity of Herpès simplex type 1 virus infected tissue culture 1MMUNOLOGY OF BREAST MILK 107

(39). We hâve also shown them in vitro to elicit antibody-dependent cytotoxicity using erythrocytes Ai as targets. However, their cytotoxic ability was inferior when compared to that elicited from adult or neonatal peripheral blood monocytes when the same target cells were used (40). The latter could be owing to the exten- sive fat ingestion by thèse cells. In order to further investigate their killing ability we also used the chemiluminescence assay and investigate their oxidative metabolic responsiveness during phagocytosis of opsonized zymosan particles. Human colostral phagocytes possess a definite oxidative metabolic activity that is intimately involved with the bactericidal function of thèse cells; however, their activity was significantly lower than that elicited by peripheral blood monocytes of healthy adults (41). From the above we may conclude that human milk macrophages possess both phagocytic and killing abilities; however, how much they use their abilities in vivo and in what respects remains to be elucidated. Human milk macrophages are also involved in a variety of biosynthetic and ex- cretory activities. Thus they produce lactoferrin, lysozyme (9,42), components of complément, properdin, factor B, and epithelial growth factor (17). Finally, some data indicate that milk macrophages may be important in the rég ulation of T-cell function by direct cellular coopération or by processing the antigen (33).

Lymphocytes

The colostral lymphocytes comprise about 10% of the total cells. Of thèse, in colostrum, 50% are T-lymphocytes, but as lactation progresses, along with the drop of total lymphocyte-numbers, the percentage of T-cells falls to around 28% (43). Colostral and milk lymphocytes hâve been found to respond in vitro to a number of viral and microbial antigens (31,44). Several studies hâve shown, however, that colostral and milk lymphocytes often respond to différent antigens than those to which peripheral blood lymphocytes of the same persons do (43,45). Thus, while rubella virus stimulâtes T-lymphocytes in secretory sites and milk as well as in systemic sites, E. coli K, antigen stimulâtes only milk lymphocytes. Thèse studies support the concept of the enteromammary and bronchomammary axis reactivity in the breast that we discussed concerning slgA. As Parmely and Béer (46) hâve stated, "cells in colostrum and milk do not represent a random collection of leukocytes expelled as a casual by-product of exosecretion but rather give every indication of being a selected population of cells accumulated by spé cifie, active biological processes." Milk lymphocytes display a decreased response to mitogens such as PHA (43). This decreased reactivity has been attributed to various causes such as a relative deficiency of a certain population of T-cells (45), to suppressive factors found in the milk (24), or to déficient macrophage-T-cell interactions (34). 70« IMMUNOLOGY OF BREAST MILK

The majority of colostral T-lymphocytes are fully mature as they stain with monoclonal antibody to OKT3 cell surface markers but do not possess OKT6 sur face markers, which are only found on immature cells (47). The T-lymphocyte population in human milk was also found to contain both OKT4 (helper/inducer phenotype) and OKT8 (suppressor/inducer phenotype) sub sets. The relative ratio of OKT4-to-OKT8 positive T-cell subsets in milk was found to be higher when compared with that found in peripheral blood T-cells (48). Colostral lymphocytes hâve been found to be as reactive in the mixed lympho cyte reaction as their blood counterparts. However, the cytotoxic potential of hu man colostral lymphocytes is selectively déficient and its expression is dépendent on the particular stimulating or target antigens presented (49). This finding suggests that some, but not ail, cytotoxic T-cell precursors réside in colostrum. In rats leukocytes may be transmitted through the milk from the mother's bloodstream to the suckling's bloodstream; this may be bénéficiai, causing adoptive immunity, or harmful, causing graft-versus-host disease (50). We hâve no évidence thus far that human milk leukocytes can cross the human intestinal wall and enter the neonatal blood circulation. However, the transfer of maternai T-cell reactivity to tuberculin protein (PPD) from the mother to the hu man neonate has been observed through breast feeding (51). It has also been suggested that partly digested cells or smaller molécules transported by the T-cells, such as transfer factor or migration inhibitory factor can be transferred to the breast-fed baby (52). More studies are needed to elucidate the immunological benefits for the newborn transmitted through the immunocompetent cells in human milk. Most B-lymphocytes in colostrum or milk bear surface IgA (52). There are con- flicting results regarding the production of IgA antibodies by human milk lympho cytes. Some investigators think they do (53), whereas others found no active synthesis of IgA (34). At the moment it is not clear whether IgA antibodies présent in milk are only a product of cells secreting IgA in the mammary gland paren- chyma or also of cells présent in the milk itself. Some investigators think that the immunoglobulins in milk originate mainly from colostral macrophages that trans port it passively (34). There are also some soluble factors in human milk that promote IgA synthesis. Pittard and Bill (33) hâve recently shown that cell-free média from colostral cells stimulate IgA synthesis in peripheral blood cells, but IgG and IgM synthesis is not affected. So far the rôle of the ingested lymphocytes in the intestine of the newborn needs much further investigation.

ADDITIONAL PROTECTIVE FACTORS

In addition to human milk antibodies and its cellular immunological factors, several other substances hâve been demonstrated to provide additional protection against systemic and gastrointestinal infection in newborn infants. IMMUNOLOGY OF BREAST MILK 109

Human colostrum and milk contain ail the components of the complément Sys tem. Active C3 production has been reported in vitro in breast milk cell cultures (54). Lactoferrin and other iron-binding proteins hâve been shown to possess inhibiting effects on the growth of E. coli, S. aureus, and C. albicans by withdrawing the iron that thèse organisms need for their metabolism (55). Lysozyme is found in human milk in amounts 300 times greater than in cow's milk. Lysozyme is known to possess powerful bacteriolytic properties (56). Interferon, migration inhibiting factor, alpha-fetoprotein, and a number of other substances are also présent although their rôles hâve not yet been fully determined (9). Antistaphylococcal factors appear to be active against expérimental staphylococ- cal infections and may be important for local mammary gland protection (57). Growth factors that promote the growth of Lactobacillus bifidus are valuable because they inhibit growth of enteric gram-negative bacilli (9). Antiviral factors hâve also been demonstrated in lipid and aqueous phases of human and animal milk in vitro; their in vivo rôles remain to be elucidated (10). Récent studies hâve also demonstrated the présence of other substances in hu man milk that promote growth and maturation of intestinal epithelial tissue (58). The accélération of the maturation of epithelial cells will shorten the period of vul- nerability in the neonate and lessen the susceptibility to disease. However, in discussing antibacterial factors in human milk one should not for- get the interactions between them: Although complément lysis of E. coli did not occur in the présence of secretory IgA alone, it occurred when, in addition to com plément and IgA, lysozyme was présent as well (56). As pointed out by Welsch and May (10), although it is often difficult to assign a précise clinical significance to each factor, it is hard to avoid the conclusion that the sum of the factors is responsible, at least partially, for the différences in the morbidity and mortality of breast-fed versus non-breast-fed infants. Finally, HLA antigens and other antigens that hâve been associated with cell- to-cell récognition events hâve recently been found in fat globules and epithelial cells of human milk (59). Their rôle in cellular homing mechanisms or immune responsiveness remains to be clarified.

HUMAN MILK AND NEONATAL MORBIDITY AND MORTALITY

There is now growing information documenting that human breast milk is supe- rior to cow's milk in minimizing both the morbidity and mortality caused by a number of infective and noninfective disorders. Many epidemiologic studies support the idea that breast feeding can protect the infant against infections and death from diarrheal disease particularly in underdeveloped countries and disadvantaged communities (60-63). One should nonethe- less bear in mind that this is in part owing to lack of contamination of human milk. However, breast milk is also able to prevent the development of induced necrotiz- 110 IMMUNOLOGY OF BREAST M1LK ing enterocolitis in a rat model (35). Largua et al. (64) report from Argentina that only small doses of colostrum (5 mg/kg daily) reduce the incidence of endémie neonatal E. coli diarrhea in the nursery. However, even in well-developed countries investigators hâve not uniformly supported the belief that breast-fed babies hâve diarrhea less often than formula-fed ones (64,65). With respect to necrotizing enterocolitis it seems that frozen human milk does not protect prématurés from developing the disease (67), whereas untreated human milk with living cells does (36). The protective effect of human milk has been shown by some investigators to extend to other extraintestinal infections such as those caused by respiratory syncytial virus (68) and Rotavirus (69). Finally, although Chandra (70) has suggested a lower incidence of allergy in breast-fed infants, more studies are required to define the antiallergic rôle of breast feeding. From the data available thus far regarding the effect of human breast milk on the incidence and severity of neonatal infections, we can corne to the following conclusion. Newborns and more so prématurés are born with various immaturities, some of which concern mùcosal immunity. However, they cope quite well pro- vided they are not challenged with conditions of stress. However, when increased activity of their host-defense mechanisms is demanded, they are unable to over- come difficulties and severe generalized and often overwhelming infections may develop. Human milk with its many valuable immunological factors helps to increase neonatal résistance to infections. This is more obvious in disadvantaged communities where the babies meet with increased challenges. It will be interesting to see what the effect of untreated human milk in sick low- birth-weight babies nursed in intensive care units in developed countries is. Babies in thèse units are usually colonized with pathogenic antibiotic-resistant, gram- negative organisms (71). From preliminary studies by our group and other investigators (72), it seems that human milk—pasteurized, frozen, or untreated—does not quantitatively alter the bacterial flora. However, Gothefors et al. (72) found decreased virulence of the bacteria présent in stools and speculate that this may be one of the ways in which breast feeding prevents infections. This fascinating field is still widely open for research.

REFERENCES

1. Ogra PL. Ontogeny of the local immune System. Pediatrics 1979;64(suppl):765-74. 2. Walker WA, Isselbacher KJ. Uptake and transport of macromolecules by the intestine: possible rôle in clinical disorders. Gastroenterology 1975;67:531-50. 3. Moxey PC, Trier JS. Structural features of the mucosa of human fetal small intestine. Gastroen terology 1975;68:1002-7. 4. Richman LK, Graeff AS, Yorchoon R, et al. Simultaneous induction of antigen spécifie IgA- helper T-cells and IgG-suppressor T-cells in the murine Peyer's patch after protein feeding. / 1m- munol 1981;126:2079-87. IMMUNOLOGY OF BREAST MILK 111

5. Reade JC. The development of bactericidal activity in rat peritoneal macrophages. Aust J Exp Biol MedSci 1968;46:231-6. 6. Losonsky GA, Ogra PL. Mucosal immune System. In: Ogra PL, éd. Neonatal infections. New York: Grune and Stratton, 1984:51-65. 7. Roux ME, McWilliams M, Philips-Quagliate JM, et al. Origin of IgA secreting plasma cells in the mammary gland. J Exp Med 1977;146:1311-21. 8. Fishant JM, Murphy D, Neifert M, et al. The broncho-mammary axis in the immune response to respiratory syncytial virus. J Pediatr 1981;99:186-91. 9. Goldman AS, Smith CW. Host résistance factors in human milk. J Pediatr 1973;82:1082-90. 10. Welsch JK, May JT. Anti-infective properties of breast milk. J Pediatr 1979;94:1-9. 11. William WB. Breast milk immunology. Am J Dis Child 1979;133:83-7. 12. Tomasi TB, Grey HM. Structure and function of immunoglobulin A. Prog Allergy 1972;16:81- 213. 13. Walker WA, Isselbacher KJ. Ν Engl J Med 1979;297:767-73. 14. Lamm ME. Cellular aspects of immunoglobulin A. Adv Immunol 1976;22:223-9. 15. Ogra SS, Weintraub D, Ogra PL. Immunologie aspects of human colostrum and milk. III. Fate and absorption of cellular and soluble components in the gastrointestinal tract of the newbom. J Immunol 1977;119:245-8. 16. Hambraeus LB, Lônnerdal B, Forsum E, Gebre-Medhin M. Nitrogen and protein components of human milk. Acta Paediatr Scan 1978;67:561-5. 17. Ogra PL, Losonsky GA. Défense factors in products of lactation. In: Ogra PL, éd. Neonatal in fections. New York: Grune and Stratton, 1984:72. 18. Ogra SS, Ogra PL. Immunologie aspects of human colostrum and milk. I. Distribution characteristics and concentrations of immunoglobulins at différent times after onset of lactation. J Pediatr 1978;92:546-9. 19. Goldman AS, Garza C, Nichols BL, Goldman RM. Immunologie factors in human milk during the first year of lactation. J Pediatr 1982;100:563-7. 20. McClelland DBL, McGrath J, Samson RR. Antimicrobial factors in human milk. Studies of con centration and transfer to the infant during the early stages of lactation. Acta Paediatr Scand [Suppl] 1978;271:3-20. 21. Butler JE. Immunologie aspects of breast feeding. Anti-infectious activity of breast milk. Semin Perinatol 1979;3:255-69. 22. Goldblum RM, Ahltstedt S, Carlsson B, et al. Antibody forming cells in human colostrum after oral immunization. Nature 1975;257:797-9. 23. Montgomery PC, Rosner BR, Cohn J. The secretory antibody response. Anti-DNP antibodies induced by dinitrophenylated type III pneumococcus. Immunol Commun 1974;3:143-56. 24. Mellander L, Carlsson B, Dahlgren U, Hanson LA. Humoral and cellular immunities transmitted by breastmilk. In: Lambert HP, Wood CBS, eds. Immunological aspects of infection in the fétus and newborn. London: Académie Press, 1981:139-46. 25. Wiman K, Curman B, Forsum V, et al. Occurrence of Ia-antigen on tissues of non-lymphoid ori gin. Nature 1978:276;711-3. 26. Walker WA. The nature and rôle of mucosal surface immunity against infection in the newborn. In: Lambert HP, Wood CBS, eds. Immunological aspects of infection in the fétus and newborn. London: Académie Press, 1981:83-99. 27. Losonsky GA, Fishaut JS, Strussenberg J, Ogra PL. Effect of immunization against rubella on lactation products. I. Development and characterization of spécifie immunologie reactivity in breast milk. J Infect Dis 1982;145:654-60. 28. Svennerholm AM, Hanson LA, Holmgren J, Lindblad BS, Nilsson B, Quereshi F. Différent se cretory immunoglobulin A antibody responses to choiera vaccination in Swedish and Pakistan women. Infect Immun 1980;30:427-30. 29. Donné A. Cours de microscopie. Paris: Bailliere, 1844-45; vol 1 et Atlas no. 1. 30. Mayer G, Klein M. Histology and cytology of the mammary gland and its sécrétions. In: Kon SK, Growie AT, eds. Milk: the mammary gland and its sécrétion; vol. I. New York: Académie Press, 1961:47-126. 31. Smith CW, Goldman AS. The cells of human colostrum. I. In vitro studies of morphology and functions. Pediatr Res 1968;2:103-9. 32. Mandyla H, Xanthou M, Baum JB, Matsaniotis N. Changes of human milk leucocytes during feeding time, pasteurization and various processes of milk storage. In: Stern L, Bard H, Friis- Hansen B, eds. Intensive care in the newborn IV. New York: Masson, 1983:51-9. 33. Pittard WB, Bill K. Immunoregulation by breast milk cells. Cell Immunol 1979;42:437^U. 112 IMMUNOLOGY OF BREAST MILK

34. Pittard WB, Polmar SH, Fanaroff AA. The breast milk macrophage: a potential vehicle for immunoglobulin transport. J Reticuloendothel Soc 1977;22:597-603. 35. Pitt J, Barlow B, Hein WC. Protection against expérimental necrotizing enterocolitis by maternai milk. I. Rôle of milk leukocytes. Pediatr Res 1977;11:906-9. 36. Burnard ED, Thomas DB. Fresh human milk in the prévention of necrotizing enterocolitis in very low birth weight infants. In: Stem L, Salle B, Friis-Hansen B, eds. Intensive care of the newborn III. New York: Masson, 1981:151-8. 37. Robinson JE, Harvey BAM, Soothhill JF. Phagocytosis and killing of bacteria and yeast by hu man milk cells after opsonisation in aqueous phase of milk. Br Med J 1978;1:1443-5. 38. Ho PC, Lawton JWM. Human colostral cells: Phagocytosis and killing of E. coli and C. albicans. J Pediatr 1978;93:910-5. 39. Kohi S, Malloy MM, Pickering LK, et al. Human colostral antibody dépendent cellular cytotoxicity against herpès simplex virus-infected cells mediated by colostral cells. J Clin Lab Immunol 1978;1:221^*. 40. Xanthou M, Mandyla-Sfagou H, Maravelias C, Baum JD, Matsaniotis N. Phagocytes of human colostrum and their function. In: Stern L, Salle B, Friis-Hansen B, eds. Intensive care of the newborn III. New York: Masson, 1981:201-11. 41. Mandyla H. Luminol-dependent chemiluminescence of colostral and mature human milk leuco cytes. [PhD Thesis]. Athens, Greece: University of Athens, 1984. 42. Pitt J. The milk mononuclear phagocyte. Pediatrics 1979;64:745-9. 43. Ogra SS, Ogra PL. Immunologie aspects of human colostrum and milk. II. Characteristics of lymphocyte reactivity and distribution of Ε-rosette forming cells at différent times after the onset of lactation. J Pediatr 1978;92:550-5. 44. Ho FC, Wong RLC, Lawton YWM. Human colostral and breast milk cells. A light and électron microscopic study. Acta Paediatr Scand 1979;68:389-96. 45. Parmely MJ, Béer AE, Billingham RE. In vitro studies of the T-lymphocyte population of human milk. J Exp Med 1976;144:358-70. 46. Parmely MJ, Béer AE. Colostral cell-mediated immunity and the concept of a common secretory immune System. J Dairy Sci 1977;60:655-65. 47. Richie ER, Bass R, Meistrich ML et al. Distribution of T-lymphocyte subsets in human colos trum. J Immunol 1982;129:1116-9. 48. Bertotto A, Gentili F, Vaccaro R. Monoclonal antibody defined T-cell surface phenotypes in hu man breast milk. Boll Ist Sieroter Milan 1983;62:370-1. 49. Parmely MJ, Williams SB. Sélective expression of immunocompetence in human colostrum: preliminary évidence for the control of cytotoxic lymphocytes including those spécifie for paternal alloantigens. In: Ogra PL, Dayton DH, eds. Immunology of breast milk. New York: Raven Press, 1979:173-80. 50. Béer AE, Billingham RE. Immunologie benefits and hazards of milk in maternal-perinatal rela- tionships. Ann Intern Med 1975;83:865-71. 51. Schlesinger JJ, Coveli HD. Evidence for transmission of lymphocyte response to tuberculin by breast feeding. Lancet 1977;2:529-32. 52. Diaz-Jouanen E, Williams RC Jr. T- and B-lymphocytes in human colostrum. Clin Immunol Im- munopathol 1974;3:248-55. 53. Murillo GJ, Goldman AS. The cells of human colostrum. II. Synthesis IgA and Betaie. Pediatr Res 1970;4:71-5. 54. Nakujima S, Baba AS, Tamura N. Complément System in human colostrum. Int Arch Allergy Appl Immunol 1977;54:428-32. 55. Kirkpatrick CH, Green T, Rich RR, et al. Inhibition of growth of Candida albicans by iron unsat- urated lactoferrin. Relation to host défense mechanisms in chronic mucocutaneous candidiasis. J Infect Dis 1971;124:539-44. 56. Adinolfi M, Glynn A. The interactions of antibacterial factors in breast milk. Dev Med Child Neu- rol 1979;21:808-19. 57. Gyorgy P, Dhonomitta S, Steers E. Protective effects of human milk in expérimental staphylococcus infections. Science 1962;137:338-9. 58. Carpenter G. Epidermal growth factor is a major growth-promoting agent in human milk. Science 1980;210:198-9. 59. Newman RA, Ormerod MG, Geaves MF. The présence of HLA-DR antigens on lactating human breast epithelium and milk fat globule membranes. Clin Exp Immunol 1980;41:478-86. IMMUNOLOGY OF BREAST MILK 113

60. Kansench H. The relationship of bottle feeding to malnutrition and gastroenteritis in a pre-indus- trial setting. J Trop Pediatr 1972;18:302-6. 61. Matta IM, Urrutia JJ, Lechtig A. Infection and nutrition of children of a low socio-economic rural community. Am J Clin Nutr 1971;24:249-54. 62. Yong HB, Buckley MA, Hamza B, Mandarano C. Milk and lactation: some social and develop- mental correlates among 1000 infants. Pediatrics 1982;69:169-75. 63. Narayanan I, Prakash K, Gujral VV. The value of human milk in the prévention of infection in the high-risk low-birthweight infants. J Pediatr 1981;99:496-8. 64. Largua AM, Urman J, Ceriani JM. Immunidad local en el recient nacido. Primer experencia con la administration de colostro humano a recien nacidos preterminos. Arch Pediatr Argent 1974;72:109-12. 65. Jelliffe DB, Jelliffee EFP. Human milk in the modem world. Oxford: Oxford University Press, 1980. 66. Cushing AH, Anderson L. Diarrhea in breast-fed and non breast-fed infants. Pediatrics 1982; 70:921-5. 67. France G, Marmes DJ, Steele RW. Breast-feeding and salmonella infection. Am J Dis Child 1980;134:147-52. 68. Downham MAP, Scott R, Sims DG, Webb JKG, Gardner PS. Breast feeding protects against respiratory syncytial virus infections. Br Med J 1976;ii:274-6. 69. Berger R, Hadziselimovic F, Just M, Reigel P. Effect of feeding human milk on nosocomial rotavirus infections in an infant ward. Dev Biol Stand 1983;53:219-28. 70. Chandra RK. Prospective studies of the effect of breast feeding on incidence of infection and allergy. Acta Paediatr Scand 1979;68:641-54. 71. Graham JM, Taylor J, Davies PA. Some aspects of bacterial colonization of ill, low birthweight and normal newborns. In: Stern L, Friis-Hansen B, Killeberg P, eds. Intensive care of the newborn. New York: Masson, 1976:59-72. 72. Gothefors L, Carlsson B, Ahlstedt S, Hanson LA, Winberg J. Influence of maternai gut flora and colostral and cord sérum antibodies on présence of E. Coli in faeces of the newbom infants. Acta Paediatr Scand 1976;65:225-32.

DISCUSSION

Dr. Rubaltelli: It is well known that necrotizing enterocolitis (NEC) is a multifactor disease and in many nurseries it is apparently related to the use of umbilical cathéters and to microbiological ecology, among many other things. Human milk, for instance, is not available in our unit, and we use a spécial preterm formula; however, since we substituted trans- cutaneous pC02 and p02 measurements for umbilical arterial catheterisms, we do not see NEC anymore, whereas in other nurseries where they are using raw human milk they still see the disease. I do not think I can support the affirmation that human milk prevents NEC. The incidence of NEC in a neonatal unit will dépend on many différent things, among which the microbiological ecology and the way the sick newborn is approached are probably the most important. Dr. Xanthou: You are right, there is a great debate regarding: (a) the etiology of NEC, and (b) the effect of human milk. As you mentioned, différent etiologies hâve been proposed. For instance, a substantial proportion of investigators hâve found it to be associated with épidémies of Klebsiella or Clostridium, and they believe that infective microorganisms at least contribute to the development of the disease. Now regarding the second argument, it seems that in ail the studies that came to the conclusion that human milk has no effect on NEC, either frozen or pasteurized human milk, which does not contain any living cells, was used. It seems that the living cells are really active. Eric Burnard in Australia did a very careful study with fresh human milk containing living cells and was able to reduce the incidence of NEC in his neonatal unit to almost nil (1). 114 IMMUNOLOGY OF BREAST MILK

Dr. Stem: Dr. Rubaltelli, your suggestion is that the disease is always a phenomenon of vascular compromise from the cathéter. Pathologically it looks like an ischémie enteropathy that could be cathéter related or attributed to hypoxia. However, there is also concern that the soluté load of the formula, rather than the nature of the formula, may be responsible because a partially compromised bowel just can not deal with a high soluté load and gives way quicker. It may therefore not be so much the constituent of the milk but the actual soluté concentration. Dr. Lozano: I wonder if you hâve any précise information regarding how the humoral and the cellular response will be modified by previous infections in the mother or by labor and delivery. Dr. Xanthou: We do not hâve a lot of information about the milk of sick mothers because in those circumstances they usually do not breast-feed. Infections will clearly cause a rise in antibody titers in the milk. When the mother has mastitis, then increased numbers of polymorphoneutrophils may appear in the milk due to the local infection. I do not think information is available on milk from mothers with, let us say, neoplastic diseases of the breast. Furthermore, I am not aware of studies regarding the influence of the mode of deliv ery and so on, on the composition of the milk. Dr. Goyens: Which mechanism could explain the antiallergic activity of human milk? Could it be that human milk has an antiallergic potential by itself, or is the incidence of allergy lower in breast-fed infants because the first contact with allergens is delayed, allow- ing the biological clock to go on tuming and the immune System to mature? I do not know if there are any data with regard to this in humans. Ann Fergusson showed that when mice are exposed perinatally to ovalbumin, they develop a very heavy allergie reaction, whereas when they are exposed for the first tirne 24 hr later, they become tolérant. Dr. Xanthou: There is still much work to be done in order to make sure that breast milk has an antiallergic activity. One of the mechanisms that is proposed dépends on the prés ence of slgA in human milk. Thèse slgA prevent the entrance of allergenic substances, e.g., cow's milk proteins, into the intestinal wall and from thèse into the systemic circula tion, which triggers the allergie reaction in the newbom (2). This sort of immune exclusion is very important during the neonatal period owing to the immaturity of the intestinal tract. The intestinal wall is much more perméable at that time than later in life. Dr. Heim: Could you tell us how allergy would be if you give an infant breast milk from another mother. How dangerous is that in respect to later immunological reactions, allergy, and so forth? Dr.Xanthou: Most human milk banks take milk from various mothers and usually mix ail samples together, so that the babies are not receiving milk from their own mothers. It would be préférable to hâve the baby taking his/her own mother's milk because it contains antibodies against microorganisms that are présent in her environment and her gastrointesti nal tract, which are clearly more useful for the baby. However, if that is impossible, the next best thing to do is to give another mother's milk. I am not worried about graft-versus- host disease (GVHD) at ail; it has not been shown to occur in humans. Although we do a lot of exchange transfusions, even intrauterine transfusions very early in gestation (3), very rarely do cases of GVHD occur, and my suspicion is that thèse only probably occur in extremely immunodeficient babies. Dr. Heim: I am very happy to hear this because on the North American continent, espe- cially in the United States, the idea is growing among people that we hâve to avoid giving foster mothers' milk or wet nurses'milk to prématuré infants. We hâve been doing it in IMMUNOLOGY OF BREAST MILK 115

Toronto for many years, giving fresh prématuré breast milk to prématuré infants, within 24 hr after collection, not pasteurized, and no reaction has been observed. Dr. Xanthou: Human milk banks hâve been working very successfully in Europe for the last 10 years. So far no complications such as GVHD hâve been reported. Many of thèse human milk banks try to give untreated human milk, kept at 4°C not only for 24 hr, but even for 72 hr. This storage does not increase the number of pathogenic microorganisms. The number of living cells in human milk decreases slightly but their function remains in tact, even after 72 hr (4). Dr. Heim: I wonder if other colleagues from the United States could comment on it. Dr. Stern: There is a long expérience with that. In fact in the United States alone, there were 80,000 prématurés raised in Martin Cooney's incubator exhibitions by a séries of wet nurses who breast-fed ail the infants. That began in 1900 and went on until the late 1930s. The interesting thing is that Cooney insisted that the same nurse always wetnurse the same infant. Thèse were not the natural mothers, but women who were brought in to breast-feed. It is interesting that nobody ever studied that population. As far as I know, nobody ever suggested that there was anything wrong with any of this. I do not know that it is such a new phenomenon; it was quite common in Germany, too, at the turn of the century. Dr. Xanthou: It is certainly better to hâve the baby breast-fed by the same person than given banked human milk. Indeed, it has been shown that when the sucking baby is colonized this may induce production by the foster mother of antibodies that the baby gets back through the milk. The milk contains the antibodies that défend the baby against the potential pathogenic organisms that colonize him. Dr. Mariai: When we give raw human milk in the neonatal intensive care unit, we are faced with a dilemma. On the one hand, we know that in some way we may prevent infec tion, and on the other, we know that many viruses are transferred to the baby through hu man milk: CMV, for example, and, in southern Europe, hepatitis B. We hâve had cases of CMV pneumonia in babies receiving raw human milk from other mothers. I suggest that we always encourage mothers to give their babies their own milk. I am not really in favor of giving raw human milk to ail preterm babies, because of the risk of CMV infection. Dr Xanthou: This is true. However, in Alabama, Stagno (5) studied this particular problem and found that indeed babies do get CMV through breast milk but at the same time, they get a lot of antibodies. About 50% excreted the virus in the urine, but the morbidity was almost nil. He followed them until the âge of 6 or 7 years, and observed that their growth and psychomotor development was normal. He therefore came to the conclusion that this was a very good way to immunize thèse babies against CMV. Dr. Marini: Of course, not ail the babies receiving CMV will develop the infection but when you deal with very small preterm babies, they will suffer a severe infection and some will die. I am well acquainted with Stagno's studies (6) concerning excrétion of CMV but others hâve reported contradictory findings. Of course, the problem is not only restricted to the milk; blood transfusions and so on also need to be considered. So, my conclusion is that there is a risk. I think, however, that hepatitis Β représente a greater risk in the south of Europe. When we know that the mother is Β positive, newborn transmission can be prevented by giving immunoglobulins and then vaccinating the baby. We do not hâve the serology of ail the mothers who give milk; consequently, we hâve to vaccinate ail the in fants against hepatitis B, otherwise we run the risk of rapid dissémination of the disease. Dr. Xanthou: The idéal situation would be to test ail mothers who donate milk for hepati tis B, but unfortunately this cannot be done in ail nurseries. 116 IMMUNOLOGY OF BREAST MILK

Dr. Ràihà: In Sweden, first of ail, we screen ail the mothers in order to find out which mothers are CMV négative and which are CMV positive. We also screen ail the donors and collect milk from CMV positive and négative donors in separate pools. AH the preterm infants of CMV négative mothers primarily get their own mother's milk, but if they need substitution, then they only receive fresh, CMV négative milk. Dr. Stem: What do you do with CMV positive milk? Dr. Ràihà: CMV positive milk is usually pasteurized or it is also used for making human milk protein and human milk fat for fortification of pooled milk. Dr. Xanthou: I would prefer the second solution because pasteurization does not really kill viruses. Dr. Rubaltelli: Suspension of breast feeding will not prevent the transmission of hepatitis Β virus from the mother to the infant, because we know that transmission does not occur only through breast milk. I agrée therefore that we need to be careful in giving banked human milk to prématuré infants, and that we should look for CMV and hepatitis Β virus, give immunoglobulins, and vaccinate. I do not think, however, that one should discourage a hepatitis Β positive mother from breast-feeding her own infant, certainly not in those countries where we suspect hepatitis Β virus infection to be fairly common. Dr. Businco: First of ail, I would like to mention some récent data from Sweden (7), that shows that breast milk in vitro can prevent the attachment of some bacteria such as pneumococcus and hemophilus influenzae to the epithelial cells of the mucosa of the throat. I think this is very important because it can explain the lower morbidity due to otitis média in breast-fed infants. With regard to the risk of graft-versus-host reactions in breast milk, we are used to feeding infants who hâve severe combined immune deficiencies and severe T-cell defects with fresh untreated breast milk and hâve never seen graft-versus-host reac tions in thèse babies. The controversy in the literature concerning the protective effect of breast-feeding newboms at risk for atopy, I mean newborns of atopic parents, can be partly explained by the fact that the différent studies were not planned in the same way. Some were conducted prospectively and others retrospectively. We al way s need to be very cau- tious in interpreting the latter results. Since 1978, we hâve been conducting a program of atopy prévention in at-risk newborns; prolonged breast feeding is advised for ail thèse ba bies. Our study (8) totally confirms the results of other authors, demonstrating the prophylactic effect of prolonged breast feeding. However, one should insist that it is absolutely necessary to control the diet of the nursing mothers because some antigens clearly pass into the milk and can consequently sensitize the baby. Large quantities of milk or eggs, for example, should be excluded from the mother's diet. Furthermore, it appears that the prophylactic effect of breast feeding is nonspecific because, as we ail know, breast-fed babies expérience fewer infections than nonbreast-fed; there is also a strict corrélation between viral infections, entends, bronchiolitis, and atopic sensitization; several years ago Oscar Frick (9) demonstrated that the atopic sensitization in children is triggered by viral infec tions. Maybe one of the prophylactic effects of breast milk could be owing to the lower morbidity from infections in the breast-fed infants. Dr. Shmerling: Further to what has just been said on the prophylactic effect of breast milk against the development of gastrointestinal allergy to cow's milk, a few years ago we published an analysis of cases of pure gastrointestinal cow's milk allergy (10). We observed that those children who had been breast-fed for at least 3 to 4 weeks had a much milder course of the disease than those who had never been breast-fed. Another important observa tion (unpublished) was that children of atopic parents who developed allergies, not only gastrointestinal cow's milk protein induced allergy but also severe generalized anaphylactic IMMUNOLOGY OF BREAST MILK 117 shock reactions, after 3, 4, or 6 months of full breast feeding had ail received a bottle of cow's milk during the first day of life. Now, this situation may be the clinical correlate of the expérimental situation described by Fergusson and Strobel from Edinburgh, to which Dr. Goyens referred. The early progressive transfer of antigens through breast milk may be necessary to establish immunotolerance; whereas the massive ingestion of antigens, during the first or second day of life in thèse newborns at risk for atopy, may cause a breakthrough into the System of immunotolerance and trigger the mechanisms leading to immunointoler- ance.

REFERENCES

1. Burnard ED, Thomas DB. Fresh human milk in the prévention of necrotizing enterocolitis in very low birth weight infants. In: Stern L, Salle B, Friis-Hansen B, eds. Intensive care in the newborn III. New York: Masson, 1981;151-8. 2. Walker WA, Isselbacher KJ. Uptake and transport of macromolecules by the intestine: possible rôle in clinical disorders. Gastroenterology 1975;67:531-50. 3. Bowman JM, Friesen RF, Bowman WD, Mclnnis AC, Barnes PH, Grewar D. Fetal transfusion in severe Rh isoimmunization. JAMA 1969;207:1101-6. 4. Mandyla H, Xanthou M, Baum JB, Matsaniotis N. Changes of human milk leucocytes during feeding time, pasteurization and various processes of milk storage. In: Stern L, Bard H, Friis- Hansen B, eds. Intensive care in the newborn IV. New York: Masson, 1983:51-9. 5. Dworsky M, Yow M, Stagno S, Pass RF, Alford C. Cytomegalovirus infection of breast milk and transmission in infancy. Pediatrics 1983;72:295-9. 6. Stagno S, Brasfield DM, Brown MB, et al. Infant pneumonitis associated with cytomegalovirus, Chlamydia, Pneumocystis, and ureaplasma: a prospective study. Pediatrics 1981;68:322-9. 7. Hanson LA, Ahlstedt S, Andersson B. Protective factors in milk and the development of the immune System. Pediatrics 1985;75(suppl): 172-6. 8. Businco L, Marchetti F, Pellegrini G, Cantani A. Prévention of atopic diseases in "at risk new borns" by prolonged breast-feeding. Ann Allergy 1983;51:296-9. 9. Frick LO, Ferman DF, Mills S. Development of allergy in children. I. Association with virus infections. J Allergy Clin Immunol 1979;63:228-34. 10. Rein D, Shmerling DH. Der Einfluss vorangegangenes Stillens und der Friihdiagnose auf den Krankheitsverlauf bei Kuhmilch-Protein-Intoleranz. Helv Paediatr Acta 1979;34:29-40.