The extrauterine development of many components of the human immune system is delayed(114)(Table 1), and the delays partly explain why young infants are more susceptible to many types of infections and why the susceptibility increases with the degree of prematurity. Certain immunologic agents are transmitted either through amniotic fluid or via the placenta during fetal life. The risk of infections in newborn infants is, however, further lessened by human milk feeding(1517). Moreover, breast-feeding protects against gastrointestinal and respiratory infections well past the newborn period(1823).

Table 1 Examples of protective effects of some of the defense agents in human milk whose production is delayed in newborn infants

A host of investigations performed over the past 40 y indicate that the protection is mainly due to defense agents in human milk, many of which are developmentally delayed in the infant (Table 1). The defense system in human milk is comprised of antimicrobial, antiinflammatory, and immunomodulating agents that are adapted to mucosal sites, are often multifunctional, and are not well represented in other milks used in human infant feeding. The antimicrobial factors include secretory IgA, lactoferrin, lysozyme, glycoconjugates, oligosaccharides, and antiviral lipids generated by partial digestion of milk fat(2430). The antiinflammatory agents involve some of the antimicrobial factors, enzymes that degrade inflammatory mediators, cellular protective agents, epithelial growth factors, and antioxidants(3133). The immunomodulators include nucleotides, cytokines, and antiidiotypic antibodies(3439).

In addition to the soluble and compartmentalized immune agents, human milk, particularly early in lactation, contains many leukocytes (≈1-3 × 106/mL)(40). About 80% of those cells are neutrophils, 15% are macrophages, and 5-10% are lymphocytes(40). The vast majority of the lymphocytes are T cells(41). Furthermore, virtually all leukocytes in human milk are activated. In that respect, the neutrophils and macrophages have an increased expression of CD11b/CD18 and a decreased expression of L-selectin(42), the macrophages are more motile than blood monocytes(43), and the T cells display the memory phenotype, CD45RO, and other phenotypic markers of activation(41, 44). The fate of these cells in the recipient is uncertain, but there is evidence from experimental animal models that milk T cells enter tissues of the neonatal animal(4547). Furthermore, some observations suggest that cellular immunity to tuberculosis or to schistosomal antigens may be transferred to the infant by breast-feeding(48, 49). Thus, it is possible that some of these maternal cells function in the recipient infant to compensate for certain developmental delays in the T cell system(6, 12).

The discovery of reciprocal relationships between postnatal development of the human immune system and immune factors produced by the mammary gland led us to question whether the evolution of these two processes was associated. That question will be explored in the context of mammalian evolution.


The basic axioms of biologic evolution are 1) all species descended from a common origin; 2) as a result of multiplication, populations of a species increase; and 3) because of environmental pressures, biologic selection occurs(50, 51). Indeed, when the environment changes sufficiently, a species may no longer be well adapted as happened during dramatic disruptions in the Ordovician, Permian, Triassic, and Cretaceous periods(52). Individuals of a species who are genetically well adapted compete successfully for newly created environmental niches. Because of natural selection and saturation of environmental sites, the overall genomic composition of a particular species stabilizes. Thus, evolutionary success is determined by the degree of reproductive capacity and ability to cope with the environment long enough to reach the reproductive period and assure the survival of offspring.


Paleontologic evidence and comparisons of genetic material from living mammalian species (the molecular clock approach)(53, 54) suggest that vertebrates evolved from deuterostome ancestors approximately 500-600 million years ago (mya) and developed in a gradual manner, where many innovations were retained in their evolutionary descendants(55).

The evolutionary innovation that defines mammals was the development of the mammary gland. It is likely that the mammary gland developed in certain reptilian insectivores about 190 mya (Fig. 1) as a consequence (in the context of natural selection) of their newborns obtaining some nutrients from secretions produced by epidermal glands located on the ventral part of the mother's thorax/abdomen(56). In such a scenario, the ventral location of the mammary gland would be favored by face to face interactions between the mother and the infant. Attractant pheromones may have also played a role in the process. The mammalian placenta evolved, but there is evidence that the mammary gland appeared before the placenta. Monotremes (Prototheria), the echidna (Tachyglossus aculeatus), and the duck-billed platypus (Ornithorhynchus anatinus), are mammals that display reptilian features including egg-laying, filiform sperm, microchromosomes, and bones in the pectoral girdle found only in the fossil remains of mammalian-like reptiles (theriodonts) from which mammals probably evolved (Fig. 1). These remnants of primitive mammals have well developed mammary glands, but no placenta(57).

Figure 1
figure 1

Relationship between the evolution of the mammary gland and its immunologic functions and postnatal delays in the production of those immunologic agents. Reptilian ancestors of mammals were probably theriodonts. Prototherian mammals are egg-laying monotremes such as the duck-billed platypus. Metatherian mammals (marsupials) have a choriovitelline type of placenta and often carry their newborn infants in a skin pouch. Eutherians are placental mammals.

Thus, immunologic activity may have been an early feature of the mammary gland. A primordial immunologic adaptation may have been to secrete antimicrobial agents such as fatty acids, oligosaccharides, lysozyme and iron-binding proteins to protect the recipient infant from the bacterial flora of the mother's skin. The fatty acids may have been similar to the those found in human sebum(58). In that respect, lysozyme, which is phylogenetically very ancient, is produced by mammalian apocrine glands(59), and melanotransferrin, an iron-binding protein found in sweat glands, has a 40% homology with lactoferrin(60). Milk produced by monotremes displays certain major features of the immune system found in milk from eutherian mammals including humans (Fig. 1). They include lysozyme(61), transferrins(61, 62), and difucosyllactose(63, 64). The oligosaccharide(63, 64) interferes with the attachment of Campylobacter jejuni(65) or Escherichia coli stable toxin(66) to epithelial cells. The protection by these and other immune components of milk against pathogens found on the nipples or areola is found in mammals as diverse as the marsupial quokka (Setonix branchyusus)(67) and humans living in primitive societies(19).

Primates appeared about 65-100 mya, the earliest hominids (the genus Australopithecus) diverged from other primates about 5-6 mya(68), our more direct Homo ancestors arose about 2.5 mya(69, 70), and H. sapiens emerged at least 100 000-200 000 y ago (Fig. 1)(7177). The evolutionary proximity of our species to other mammals has been reconstructed from phenotypic and genotypic analyses of existing mammalian species. The genomes of humans and chimpanzees (Pan troglodytes and Pan paniscus) are remarkably similar(73, 76, 7881), whereas far greater differences in genomic and phenotypic features are found between H. sapiens, other primates, or other mammalian orders(7881).

These evolutionary relationships hold for the composition of milks from various species. The composition of human milk is closest to the chimpanzee and gorilla, less similar to other primates, and even more different from milks produced by more distantly related mammals. The immune system in human milk, including the type of IgA, is most similar to that found in our most closely related living primate, the chimpanzee(82). The evolutionary proximity of antibodies in milk from humans and closely related primates is also demonstrated in an antigen-binding specificity of the antibodies. A human antibody against the mammalian α-galactosyl epitope cross-reacts with enteric bacteria that display similar epitopes. This antibody specificity is common in apes and Old World monkeys, but is undetected in New World monkeys, prosimians, and other types of mammals(83) that are more distantly related to the genus Homo. IgA antibodies with that specificity are present in human milk(83). Thus, the evolutionary relationship of our species to other primates holds in respect to immune functions of the mammary gland.

Certain factors in mammalian milks are remarkably conserved in many mammalian species. They include antigenic similarities between the carboxy-terminal, intracytoplasmic regions of the Muc-1 mucins found in milks from humans and many other mammalian species(84) and structural similarities in lysozyme and α-lactalbumin in all mammalian species that have been investigated(85). In contrast, some immunologic functions of the mammary gland are not as highly conserved in certain mammalian species. Because of exposures to dissimilar environmental microorganisms, one might anticipate that certain immune responses by different orders of mammals diverged as a consequence of natural selection. Furthermore, because of major distinctions in the rate and degree of motor development, the type and extent of exposure to microorganisms by young infants of those species are not entirely alike. Thus, it would be predicted that immunologic adaptations to these different environments are also reflected in the immunologic composition of milk produced by those species. A striking example is the qualitative and quantitative immunologic differences between human and cow's milk (Table 2). The quantities and types of antimicrobial oligosaccharides and glycoconjugates(27) and the concentrations of Ig isotypes(2426, 56, 86), lactoferrin(2426), lysozyme(2426), lactoperoxidase(87, 88), and complement components(31) are very different in milks from those species.

Table 2 A comparison of the relative expression of certain major antimicrobial proteins in human and bovine milk


The immune status of the newborn infant is an outcome of the development of the immune system and the maternal contributions that are received during fetal life. The production of certain components of the fetal immune system is delayed throughout pregnancy. There are two potential evolutionary reasons for these physiologic delays. The first is that the fetus is shielded from most microbial pathogens and therefore does not require the full-fledged defense system required for independent survival. Therefore, energy and nutritional factors can be directed toward developing other organ systems that will be required for immediate extrauterine life. The second reason is that the delays are adaptations to avoid untoward immunologic reactions to maternal tissues. Although the phenomenon is not unique to our species, the prevention of allograft rejection by the fetus against the mother or vice versa permits the prolonged gestation found in H. sapiens. The lag in the fetal immune system sets the stage for delays in the immunologic system that are manifest at birth.

Postnatal developmental delays in the immune system are found in every mammal that has been investigated. If such delays occurred in the earliest mammals before compensatory immune agents were present in milk, the progeny would have been at a survival disadvantage. It is more likely that the immune activities of the mammary gland permitted newborns of primitive mammals, who had evolved a slowed rate of immune development, to survive. Further development of the immune system of newborns of the immediate precursors of early mammals may have been delayed, but those developmental delays were compensated for by certain immune factors obtained from the egg yolk. This is suggested by the recent discovery of lysozyme in tortoise eggs(89). Thus, part of the evolutionary pattern may have been a switch from the immune factors provided in eggs to the immune system produced by the mammary gland.

In keeping with that contention, mirror-image patterns of immunologic development and immune factors in milk are found not only in eutherians(placentarians) but also in metatherians (marsupials) (Fig. 1) that are profoundly dissimilar from eutherians because of their choriovitelline placentas and the prolonged postnatal attachment of their atricial newborns (essentially fetuses) to the nipple of the mammary gland. In that respect, newborn infants of two species of the primitive marsupial family Didelphis, the North American Virginia opossum (Didelphis virginiana) and the Brazilian gray, short-tailed opossum(Monodelphis domestica) are immunologically immature(90, 91) and acquire passive immunity via maternal antibodies in milk(91, 92).

The isotypes of Igs in human and bovine milk (Table 2) and how they relate to the immunologic development of the off-spring are striking examples of different evolutionary adaptations. The fetal or newborn calf produces very little IgG(93, 94), and very little of that Ig is transmitted across the bovine epichordial placenta(93). Consequently, the concentration of IgG in extracellular fluids in the newborn calf is very low(93, 94). IgG is, however, the dominant Ig in bovine colostrum (Table 2)(86). This ingested IgG is intestinally absorbed during the immediate newborn period. As a result, adult concentrations of IgG are achieved in the extracellular fluids of the calf after the first nursing. The human fetus also does not produce IgG, but adult levels of serum IgG are attained at birth via transport across the placenta(11). In contrast, little IgG is found in human milk(28). The dominant Ig isotype in human milk is secretory IgA, a dimeric form of IgA bound to a portion of the polymeric Ig receptor, secretory component (Table 2)(24, 26, 28). Secretory IgA antibodies act principally to protect mucosal sites by neutralizing toxins and interfering with adherence of bacterial enteric and respiratory pathogens including Vibrio cholerae, Shigella species, E. coli, and Streptococcus pneumoniae to epithelial cells(28).


Evolutionary adaptations also enhance the survival of defense factors in human milk in the recipient infant. The nature of defense agents secreted by the mammary gland, their physical state in milk, antiproteases in milk, and the diminished ability of the recipient infant to process the factors are responsible for the enhancement. 1) Defense factors in human milk such as lactoferrin, lysozyme, and secretory IgA are inherently resistant to digestion(9597), whereas others are compartmentalized and are thus shielded from digestive enzymes or denaturing conditions(98100). 2) Antiproteases in human milk such as α1-antichymotrypsin andα1-antitrypsin protect immune agents in human milk that are proteins from digestion(101). 3) In addition, ingested defense factors in human milk are protected during the first month of life because stimulated gastric production of HCl(102, 103), and pancreatic secretion of trypsin and chymotrypsin(104107) are very low at that time.


One advantage of a delayed development of the immune system is that less energy and nutrients are required to maintain and mobilize the immune system of the infant. Indeed, when the immune system is challenged, a great deal of cell division and differentiation occurs(108). Both processes increase nutritional demands. Spared energy/nutrients may be used for the growth and development of the CNS(109, 110) and of alveoli and vascular structures of the lung(111).

The postnatal developmental delays are evolutionary successes because of compensatory maternal defense factors that that are transmitted to the infant via breastfeeding. They include antimicrobial, antiinflammatory, and immunomodulating agents in human milk. The antimicrobial factors are common to mucosal sites and protect by noninflammatory mechanisms. The provision of a wide spectrum of agents throughout lactation effectively prevents infections that the nursling may encounter. Furthermore, it is nutritionally efficient that the bulk of antimicrobial agents as well as other defense agents in human milk are used to nurture the recipient infant.

There are two general evolutionary strategies regarding antimicrobial agents in human milk. The first is a constitutive production of protective factors that is independent of exposure to infectious or other foreign agents. They include lactoferrin, lysozyme, glucoconjugates, and oligosaccharides. The second strategy is an antigen-dependent mechanism that leads to the production of secretory IgA antibodies that protect against enteric and respiratory pathogens.

The origin, specificity, and spectrum of the secretory IgA antibodies in human milk are achieved as follows. Maternal IgM+ B cells in Peyer's patches of the small intestine or in the submucosa of the tracheobronchial tree are programmed to recognize microbial antigens by the antigen-binding specificity of their surface IgM antibodies. When B cells encounter those antigens, they switch their surface membrane Ig isotype to IgA and migrate sequentially to local lymphatics, regional lymph nodes, major lymphatic vessels, and the systemic circulation(28). Under the influence of hormones produced late in pregnancy and during lactation(112), IgA+ B cells home to the mammary gland where they transform into plasma cells and secrete large amounts of dimeric IgA antibodies that bind to the same microbial antigens encountered at the maternal mucosal sites. Thus, these antibodies protect against mucosal pathogens in the maternal environment(28). Because the production of secretory IgA is delayed in newborn infants, those who are breast-fed are more resistant to enteric and respiratory pathogens than are non-breast-fed ones.

The increasing development of the immune system, including the ability to mount mucosal secretory antibody responses and the expansion of the antigen binding repertoire of antibodies, coincides with the onset of weaning and widening exposures of the infant to environments outside of the maternal sphere. Although the mammary gland produces antimicrobial agents throughout lactation, the need for the agents slowly declines as the infant's immune system matures.

The infant is exposed to microbial immunogens while receiving protective agents in human milk. This is tantamount to an attenuated immunization, in that the pathogenicity of the microbial agent is reduced by accompanying immune factors. One example is the transmission of infectious cytomegalovirus by breast-feeding. Cytomegalovirus is commonly excreted in human milk produced by seropositive women(30). Antiviral factors in human milk(30, 113) are simultaneously transmitted to the recipient infant. The infant becomes infected, develops a systemic immune response including the formation of specific antibodies, but does not become diseased. A second possible immunization mechanism provided by breast-feeding is the transfer of antiidiotypic secretory IgA antibodies into human milk(39). In this respect, idiotypic antibodies directed against binding sites of antibodies that recognize epitopes of foreign antigens serve as surrogates for foreign immunogens such as polioviruses(39).

In addition to the control over the bacterial enteric flora that is exerted by direct-acting antimicrobial agents in human milk, human milk provides growth factors that encourage the proliferation of a protective enteric flora. Certain glycoconjugates in human milk, particularly κ caseins(114), promote the growth of lactobacilli and bifidobacilli in the lower intestinal tract of the recipient. These bacteria produce large amounts of organic acids that inhibit the growth of bacterial enteropathogens such as Shigella and Salmonella sp. and E. coli.

Breast-feeding protects against infections without provoking inflammation(19, 22, 31). This may be explained by a paucity of inflammatory mediators in human milk and the presence of many antiinflammatory agents in milk(3133, 115). It is likely that the inhibition of inflammation by those agents spares the functions of mucosal sites of the recipient infant. For example, platelet-activating factor-acetylhydrolase(33) and IL-10(100) in human milk may compensate for developmental delay in the production of those antiinflammatory factors(5, 6), and that may account in part for the decreased risk to necrotizing enterocolitis of newborn infants fed human milk(116).

Breast-feeding also prepares the recipient to resist certain immune-mediated diseases, such as type I diabetes mellitus(117), Crohn's disease(118), and lymphomas(119) that emerge in late childhood. The mechanisms are undetermined, but it is possible that some of the protection may be due to immunomodulators in human milk. The evolutionary consequences would be to permit the child to mature to a sexually active adult who could better cope with the environment. Further research will be required to examine that possibility.

Finally, it should be realized that some defense agents may be created after partial digestion of human milk constituents in the gastrointestinal tract of the recipient. For example, antiviral fatty acids and monglycerides liberated from milk fat by in vivo lipolysis disrupt enveloped viruses(30), and lactoferricin created by partial proteolysis of lactoferrin kills Candida albicans(120), certain enteric bacteria(121), and Giardia(122) by damaging their outer cell membranes(120122).


Stabilization of these evolutionary processes is evidenced by a lack of geographic/ethnic variation in the immunologic development of the human infant and the composition of human milk as long as nutrient intake is not limiting(123). This suggests that those evolutionary outcomes are highly successful. Although no basic differences have been reported, subtle variations in the mammary gland immune system or the postnatal development of the immune system may have occurred in isolated populations. It is worthwhile to explore such possibilities by studying isoforms of immune agents in human milk and the definitive patterns of immune development. Such variations should be sought, but the evidence suggests that immune functions of the mammary gland and postnatal immune development are not influenced by ethnicity or geographic location.


The postnatal development of the human immune system and the immune functions of the mammary gland are related evolutionary events in all mammalian species. Consistent patterns of postnatal development of the immune system and immune functions of the mammary gland found in widely separated human populations suggest that this evolutionary embrace assisted in the dissemination of Homo sapiens from Africa to other environments(70, 71, 73, 124, 125) by enhancing infant survival. In that respect, breast-feeding protects against common infections in many populations.

Although the protection provided by breast-feeding is due mainly to defense agents in human milk that compensate for those that are not sufficiently produced by the infant, the relationships may be more complex. Delays in producing some agents may be linked to lags in appearance of others. That is suggested in a recent study that provided some evidence that the lag in IL-10 production may be secondary to a delayed TNF-α production and a reduced expression of TNF-α receptors(6). That is plausible because TNF-α is a potent stimulus for the production of IL-10(126). Furthermore, the rate of postnatal development of the production other agents such as antiadherent oligosaccharides and glycoconjugates that are well represented in human milk(27) is not established. Investigations will be needed to determine whether the production of those agents by the infant during the early postnatal period and by the mammary gland also evolved in a reciprocal pattern.