Human Milk

Mother's milk delivered naturally through breast-feeding has been the sole source of infant nutrition in mammalian species for millions of years. Since human beings learned to domesticate cattle about 10,000 years ago, nonhuman mammalian milk also has been used to supplement or replace maternal milk in human infants. The development and widespread use of commercially prepared infant formula products have been phenomena of the 20th century and notably of the past 6 decades. Such products provide an alternative to breast-feeding that is useful in certain situations. Nonetheless, compelling evidence shows that breast-feeding is an ideal source of infant nutrition, whose use is associated with lower rates of postneonatal infant mortality in the United States and in other parts of the world [[1], [2], [3]]. Human milk helps to protect infants against a wide variety of infections; helps to reduce the risk for allergic and autoimmune diseases, the risk of obesity and its complications, and the risk for certain types of neoplasms later in life; and has been associated with slightly better performances on tests of cognitive development in some studies [3]. For these reasons, the American Academy of Pediatrics (AAP) and the World Health Organization (WHO) recommend that in the absence of specific contraindications (see “Benefits and Risks of Human Milk”), healthy term infants should be exclusively breast-fed or fed expressed breast milk beginning within the first hour after birth through 6 months of age [1,3].

Over the past few decades, the immune responses at intestinal and respiratory mucosal surfaces to local infections have been intensely studied. These investigations have led to the development of concepts of immunity on mucosal surfaces of gastrointestinal, respiratory, and genitourinary tracts and identification of mucosa-associated lymphoid tissue and local mechanisms of defense that are distinct from the internal (systemic) immune system. This chapter reviews existing information on major aspects of the physiologic, nutritional, immunologic, and anti-infective components of the products of lactation. Evidence on the contribution of human milk to the development of immunologic integrity in the infant and its influence on the outcome of infections and other host-antigen interactions is also discussed.

Physiology of lactation

Developmental anatomy of the mammary gland

The rudimentary mammary tissue undergoes several developmental changes during morphogenesis and lactogenesis: In the 4-mm human embryo, the breast tissue appears as a tiny mammary band on the chest wall [4,5]; by the 7-mm embryonic stage, the mammary band develops into the mammary line, along which eventually develops the true mammary anlage; by the 12-mm stage, a primitive epithelial nodule develops; by the 30-mm stage, the primitive mammary bud appears. These initial phases of development occur in both genders (Table 5–1 ). Further development in the male apparently is limited, however, by androgenic or other male-associated substances [6,7]. Castration in male rat embryos early in gestation leads to female breast development, whereas ovariectomy in female rat embryos does not alter the course of development of the mammary anlage. Toward the end of pregnancy, initial phases of fetal mammary differentiation seem to occur under the influence of placental and transplacentally acquired maternal hormones, with transient development of the excretory and lactiferous ductular systems. Such growth, differentiation, and secretory activities are transient and regress soon after birth [7,8].

TABLE 5–1.

Possible Endocrine Factors in Growth of Human Female Mammary Glands

Clinical State	Growth Characteristics	Maturational Hormones
Prenatal	Rudimentary	None
Infancy	Rudimentary	None
Puberty	Growth and budding of milk ducts	Growth hormone, prolactin-estrogen, corticosteroids, prolactin (high doses)
Pregnancy	Growth of acinar lobules and alveoli	Estrogen, progesterone, prolactin, growth hormone, corticosteroids
Parturition	Alveolar growth	Prolactin, corticosteroids
Lactational growth of tissue	None	None
Secretory products	Casein, α-lactalbumin	Prolactin, insulin, corticosteroids

At thelarche, and later on at menarche, true mammary growth and development begin in association with rapidly increasing levels of estrogens, progesterone, growth hormone, insulin, adrenocorticosteroids, and prolactin [8,9]. Estrogens seem to be important for the growth and development of the ductular system, and progestins apparently are important for lobuloalveolar development (see Table 5–1). Final differentiation of the breast associated with growth and proliferation of the acinar lobes and alveoli continues to be influenced by the levels of estrogen and progesterone. Other peptide hormones, such as prolactin, insulin, and placental chorionic somatomammotropin, appear to be far more important for the subsequent induction and maintenance of lactation (see Table 5–1).

Prolactin secretion from the pituitary gland seems to be under neural control, and the increasing innervation of the breast observed throughout pregnancy apparently is regulated by estrogens [9]. Intense neural input in virgin and parturient, but not in currently pregnant, mammals has been shown to result in lactation. Lactation in goats can be induced by milking maneuvers. Adoptive breast-feeding also is well documented in primitive human societies. Sudden and permanent cessation of suckling can result in the termination of milk secretion and involution of the breast to the prepregnant state as the concentrations of prolactin decline. Estrogen and progesterone also may amplify the direct effects of prolactin or may induce additional receptors for this peptide hormone on appropriate target tissues in the breast.

Endocrine control of mammary gland function

Breast tissue is responsive to hormones, even as a rudimentary structure, as illustrated by the secretion of “witch's milk” by male and female newborns in response to exposure to maternal secretion of placental lactogen, estrogens, and progesterone [5]. The secretion of this early milk ceases after exposure to maternal hormones has waned. Sexual differentiation, marked by puberty, is the next major stage in mammary development. As pointed out earlier, androgens inhibit the development of mammary tissue in males, whereas the development of mammary tissue in females depends on estrogen, progesterone, and pituitary hormones [10]. The postpubertal mammary gland undergoes cyclical changes in response to the release of hormones that occurs during the menstrual cycle. The last stage of development occurs during menopause, when the decline in estrogen secretion results in some atrophy of mammary tissue.

During the menstrual cycle, the mammary gland responds to the sequential release of estrogen and progesterone with a hyperplasia of the ductal system that continues through the secretory phase and declines with the onset of menstruation. The concentration of prolactin modestly increases during the follicular stage of the menstrual cycle, but remains constant during the secretory phase [11]. Prolactin secretion seems to be held in readiness for the induction and maintenance of lactation.

Initiation and Maintenance of Lactation

Pregnancy is marked by profound hormonal changes reflecting major secretory contributions from the placenta, the hypothalamus, and the pituitary gland, with contributions from many other endocrine glands (e.g., pancreas, thyroid, and parathyroid). Increased estrogen and progesterone levels during pregnancy stimulate secretion of prolactin from the pituitary, whereas placental lactogen seems to inhibit the release of a prolactin-inhibiting factor from the hypothalamus. Prolactin, lactogen, estrogen, and progesterone all aid in preparing the mammary gland for lactation. Initially in gestation, an increased growth of ductule and alveolobular tissue occurs in response to estrogen and progesterone. In the beginning of the second trimester, secretory material begins to appear in the luminal cells. By the middle of the second trimester, mammary development has proceeded sufficiently to permit lactation to occur should parturition take place.

When the infant is delivered, a major regulatory factor, the placenta, is lost, and new regulatory factors including the maternal-infant interaction and neuroendocrine regulation are gained for control of lactation. Loss of placental hormone secretion results in an endocrine hypothalamic stimulation of prolactin release from the anterior pituitary gland and neural stimulation of oxytocin from the posterior pituitary. The stimulation of the nipple by suckling activates a neural pathway that results in release of prolactin and oxytocin. Prolactin is responsible for stimulating milk production, whereas oxytocin stimulates milk ejection (the combination is known as the let-down reflex). Oxytocin also stimulates uterine contractions, which the mother may feel while she is breast-feeding; this response helps to restore the uterus to prepregnancy tone. Milk production and ejection depend on the complex interaction of stimulation by the infant's suckling, neural reflex of the hypothalamus to such stimulation, release of hormones from the anterior and posterior pituitary, and response of the mammary gland to these hormones to complete the cycle.

Milk Secretion

Milk is produced as the result of synthetic mechanisms within the mammary gland and the transport of components from blood. Milk-specific proteins are synthesized in the mammary secretory cells, packaged in secretory vesicles, and exocytosed into the alveolar lumen. Lactose is secreted into the milk in a similar manner, whereas many monovalent ions, such as sodium, potassium, and chloride, depend on active transport systems based on sodium-potassium adenosine triphosphatases (Na⁺,K⁺-ATPases). In some situations, the mammary epithelium, which may behave as a “mammary barrier” between interstitial fluid derived from blood and the milk because of the lack of space between these cells, may “leak,” permitting direct diffusion of components into the milk. This barrier results in the formation of different pools or compartments of milk components within the mammary gland and is responsible for maintaining gradients of these components from the blood to the milk.

Lipid droplets can be observed within the secretory cells of the mammary gland and are surrounded by a milk fat globule membrane. These fat droplets seem to fuse with the apical membrane of the secretory cells and to be exocytosed or “pinched off” into the milk [10]. Some whole cells also are found in milk, including leukocytes, macrophages, lymphocytes, and broken or shed mammary epithelial cells. The mechanisms by which these cells enter the milk are complex and include, among others, specific cellular receptor-mediated homing of antigen-specific lymphocytes.

As the structure of the mammary gland is compartmentalized, so is the structure of milk. The gross composition of milk consists of cytoplasm encased by cellular membranes in milk fat globule membranes (fat compartments composed of fat droplets), a soluble compartment containing water-soluble constituents, a casein-micelle compartment containing acid-precipitable proteins with calcium and lactose, and a cellular compartment. The relative amounts of these components change during the course of lactation, generally with less fat and more protein in early lactation than in late lactation. The infant consumes a dynamic complex solution that has physical properties permitting unique separation of different functional constituents from one another, presumably in forms that best support growth and development.

Lactation Performance

Successful lactation performance depends on continued effective contributions from the neural, endocrine, and maternal-infant interactions that were initiated at the time of delivery. The part of this complex behavior most liable to inhibition is the mother-child interaction. Early and frequent attachment of the infant to the breast is mandatory to stimulate the neural pathways essential to maintaining prolactin and oxytocin release.

A healthy newborn infant placed between the mother's breasts locates a nipple and begins to suck spontaneously within the first hour of birth [12]. This rapid attachment to the mother may reflect olfactory stimuli from the breast received by the infant at birth [13]. Frequent feedings are necessary for the mother to maintain an appropriate level of milk production for the infant's proper growth and development. Programs to support lactation performance must emphasize proper maternal-infant bonding, relaxation of the mother, support for the mother, technical assistance to initiate breast-feeding properly and to cope with problems, and reduction of environmental hindrances. Such hindrances may include lack of rooming-in in the hospital, use of supplemental formula feeds, and lack of convenient day care for working mothers.

Lactation ceases when suckling stops; any behavior that reduces the amount of suckling by the infant initiates weaning or the end of lactation. Introduction of water in bottles or of one or two bottles of formula a day may begin the weaning process regardless of the time after parturition, but can be most damaging to the process when the mother-infant dyad is first establishing lactation.

Secretory products of lactation: nutritional components of human colostrum and milk

Colostrum and milk contain a rich diversity of nutrients, including electrolytes, vitamins, minerals, and trace metals; nitrogenous products; enzymes; and immunologically specific cellular and soluble products. The distribution and relative content of various nutritional substances found in human milk are presented in Table 5–2 . The chemical composition often exhibits considerable variation among lactating women and in the same woman at different times of lactation [14], and between samples obtained from mothers of infants with low birth weight and from mothers of full-term infants [15,16]. Mature milk contains the following average amounts of major chemical constituents per deciliter: total solids, 11.3 g; fat, 3 g; protein, 0.9 g; whey protein nitrogen, 760 mg; casein nitrogen, 410 mg; α-lactalbumin, 150 mg; serum albumin, 50 mg; lactose, 7.2 g; lactoferrin, 150 mg; and lysozyme, 50 mg.

TABLE 5–2.

Distribution of Secretory Products in Human Colostrum and Milk^*

Water	86%-87.5%
Total Solids	11.5 g
Nutritional Components
Lactose	6.9-7.2 g
Fat	3-4.4 g
Protein	0.9-1.03 g
α-lactalbumin	150-170 mg
β-lactoglobulin	Trace
Serum albumin	50 mg
Electrolytes, Minerals, Trace Metals
Sodium	15-17.5 mg
Potassium	51-55 mg
Calcium	32-43 mg
Phosphorus	14-15 mg
Chloride	38-40 mg
Magnesium	3 mg
Iron	0.03 mg
Zinc	0.17 mg
Copper	15-105 μg
Iodine	4.5 μg
Manganese	1.5-2.4 μg
Fluoride	5-25 μg
Selenium	1.8-3.2 μg
Boron	8-10 μg
	Total 0.15-2 g
Nitrogen Products
Whey protein nitrogen	75-78 mg
Casein protein nitrogen	38-41 mg
Nonprotein nitrogen	25% of total nitrogen
Urea	0.027 g
Creatinine	0.021 g
Glucosamine	0.112 g
Vitamins
C	4.5-5.5 mg
Thiamine (B1)	12-15 μg
Niacin	183.7 μg
B6	11-14 μg
B12	<0.05 μg
Biotin	0.6-0.9 μg
Folic acid	4.1-5.2 μg
Choline	8-9 mg
Inositol	40-46 mg
Pantothenic acid	200-240 μg
A (retinol)	54-56 μg
D	<0.42 IU
E	0.56 μg
K	1.5 μg

^*Estimates based on amount per deciliter.

Human milk contains relatively low amounts of vitamins D and E (see Table 5–2) and little or no β-lactoglobulin (the major whey protein in bovine milk). The fat globule membrane seems to have a high content of oleic acid, linoleic acid, phosphatidylpeptides, and inositol [17]. In addition, a binding ligand that promotes absorption of zinc has been identified in human milk [18,19]. Temporal studies have indicated that concentrations of many chemical components, especially nitrogen, calcium, and sodium, decrease significantly as the duration of lactation increases [20,21]. Several components have been found to change in concentration as a function of water content, however, because their total daily output seems to be remarkably constant, at least during the first 8 weeks of lactation [22,23].

Milk production progresses through three distinct phases, characterized by the secretion of colostrum, transitional (early) milk, and mature milk. Colostrum comprises lactational products detected just before and for the first 3 to 4 days of lactation. It consists of yellowish, thick fluid, with a mean energy value of greater than 66 kcal/dL and contains high concentrations of immunoglobulin, protein, fat, fat-soluble vitamins, and ash. Transitional milk usually is produced between days 5 and 14 of lactation, and mature milk is produced thereafter. The concentrations of many nutritional components decline as milk production progresses to synthesis of mature milk. The content of fat-soluble vitamins and proteins decreases as the water content of milk increases. Conversely, levels of lactose, fat, and water-soluble vitamins and total caloric content have been shown to increase as lactation matures [24,25].

As the result of several manufacturing errors, the nutrient composition of infant formulas has been legislated [26], resulting in the paradoxical situation that human milk may not always meet the recommended standards for some nutrients, whereas infant formulas may exceed the recommendation. Human milk nutrient composition varies with time of lactation (colostrum versus early milk versus mature milk) and, to some extent, maternal nutritional status. The appropriate amounts of each nutrient must be considered within these constraints.

Minerals

The mineral content of human milk is low relative to that of infant formulas and very low compared with cow's milk, from which most formulas are prepared; although human milk is sufficient to support growth and development, it also represents a fairly low solute load to the developing kidney. The levels of major minerals tend to decline during lactation, with the exception of magnesium, but with considerable variability among women tested [27]. Sodium, potassium, chloride, calcium, zinc, and phosphorus all seem to be more bioavailable in human milk than in infant formulas, reflecting their lower concentrations in human milk. Iron is readily bioavailable to the infant from human milk, but may have to be supplemented later in lactation [28,29]. Preterm infants fed human milk may need supplements of calcium and sodium [30].

Vitamins

Human milk contains sufficient vitamins to maintain infant growth and development, with the caveat that water-soluble vitamins are particularly dependent on maternal intake of these nutrients [31]. Preterm infants may require supplements of vitamins D, E, and K when fed human milk [32,33]. The low content of vitamin D in human milk has been related to the development of rickets in a few breast-fed infants, as discussed later [34]. The AAP currently recommends that all breast-fed infants receive 200 IU of oral vitamin D drops daily beginning in the first 2 months of life and continuing until adequate amounts of vitamin D are provided through daily consumption by the infant of fortified formula or milk [3].

Carbohydrates and Energy

Lactose is the primary sugar found in human milk and usually is the carbohydrate chosen for the preparation of commercial formulas. Lactose supplies approximately half the energy (of a total 67 kcal/dL) taken in by the infant from human milk. Lactose (a disaccharide of glucose and galactose) also may be important to the neonate as a carrier of galactose, which may be more readily incorporated into gangliosides in the central nervous system than galactose derived from glucose in the neonate [35]. Also, glycogen may be synthesized more efficiently from galactose than from glucose in the neonate because of the relatively low activity of glucokinase in early development [36]. Human milk also contains other sugars, including glucose and galactose and more than 100 different oligosaccharides [37]. These oligosaccharides may have protective functions for the infant, especially with respect to their ability to bind to gastrointestinal pathogens [38].

Lipids

Fats provide almost half of the calories in human milk, primarily in the form of triacylglycerols (triglycerides) [33]. These lipids are supplied in the form of fat globules enclosed in plasma membranes derived from the mammary epithelial cells [39]. The essential fatty acid linoleic acid supplies about 10% of the calories derived from the lipid fraction. Triacylglycerols serve as precursors for prostaglandins, steroids, and phospholipids and as carriers for fat-soluble vitamins. The lipid profiles of human milk differ dramatically from the lipid profiles of commercial formulas, and despite considerable adaptation of such formulas, human milk lipids are absorbed more efficiently by the infant.

Cholesterol, an important lipid constituent of human milk (12 mg/dL), usually is found in only trace amounts in commercial formulas. It has been suggested that cholesterol may be an essential nutrient for the neonate [40]. A lack of cholesterol in early development may result in turning on of cholesterol-synthetic mechanisms that are difficult to turn off later in life, influencing induction of hypercholesterolemia [41]. Some studies suggest that breast-feeding of the neonate is associated with lowered adult serum cholesterol levels and reduced deaths from ischemic heart disease [42].

Interest has increased in the role that long-chain polyunsaturated fatty acids may play in human milk, especially docosahexaenoic acid and arachidonic acid. These long-chain polyunsaturated fatty acids are not found in unsupplemented infant formulas, but are present in human milk. They are structural components of brain and retinal membranes and may be important for cognitive and visual development. In addition, they may have a role in preventing atopy [43]. Numerous studies have found that infants fed formula without docosahexaenoic acid or arachidonic acid have reduced red blood cell amounts of these fatty acids [[44], [45], [46]]; however, findings in visual and cognitive functional studies in term infants have been inconsistent [47,48]. These studies have been complicated by the finding of slower growth in some preterm infants fed formulas supplemented with long-chain polyunsaturated fatty acids [45,46]. The inconsistent findings in infants fed supplemented formula may reflect the difficulty in determining the optimal amounts of docosahexaenoic acid and arachidonic acid and their precursors linoleic and linolenic acids in such supplemented formulas. These lipids seem to be best delivered from human milk.

Protein and Nonprotein Nitrogen

The exact protein content of mature human milk is variable, but is around 1 g/dL, in contrast to infant formulas, which usually contain 1.5 g/dL; the milk from mothers who deliver preterm infants may have slightly more protein [49]. The nutritionally available protein may be less than 1 g/dL—possibly 0.8 g/dL—as a result of the proportion of proteins that is used for non-nutritional purposes. In addition, human milk contains a considerably greater percentage of nonprotein nitrogen (25% of the total nitrogen) compared with formulas (5% of the total nitrogen) [50].

Human milk protein is primarily whey predominant (acid-soluble protein), whereas formulas prepared from bovine milk classically reflect the 18% whey–82% casein protein composition in that species. The whey-to-casein protein ratio in humans may change during lactation, with the whey component ranging from 90% (early milk) to 60% (mature milk) to 50% (late milk) [51]. Formulas for preterm infants have been reconstituted from bovine milk to provide 60% whey and 40% casein proteins; all major formulas for term infants in the United States are now bovine whey-protein–predominant preparations, in an attempt to make them closer to human milk in composition. These differences in protein quality are reflected by differences in the plasma and urine amino acid responses of infants fed human milk or formulas that are casein-protein–predominant or whey-protein–predominant [[52], [53], [54]]. Generally, term infants do not respond with the dramatic differences seen in preterm infants when fed formulas with different protein quality [[55], [56], [57], [58]].

The nonprotein nitrogen component of human milk contains various compounds that may be important to the development of the neonate: polyamines, nucleotides, creatinine, urea, free amino acids, carnitine, and taurine [59]. The significance of the presence of these components is not always clear, but when they are not fed, as in the case of infant formulas that contain little taurine [52] or of soy formulas that contain little carnitine [60], apparent deficiencies that may influence the development of the infant occur. Taurine is important for bile salt conjugation and for support of appropriate development of the brain and retina [40], whereas carnitine seems to be important for appropriate fatty acid metabolism [61].

Nucleotides, in particular, seem to bridge the gap between the nutritional and the immunologic roles of human milk components. Human milk contains most of these compounds in the form of polymeric nucleotides or nucleic acids [62,63], whereas formulas contain nucleotides (when they are supplemented) only in the monomeric forms (Table 5–3 ). Nucleotides seem to enhance intestinal development, promote iron absorption, and modify lipid metabolism in their nutritional role [64]. These compounds perform an immunologic function by promoting killer cell cytotoxicity and interleukin (IL)-2 production by stimulated mononuclear cells from infants either breast-fed or fed nucleotide-supplemented formulas [65].

TABLE 5–3.

Nucleotides in Human Milk and Supplemented Formula

	Human Milk*	Human Milk†	Formula*
Nucleic acid (%)	48	42	4
Nucleotides (%)	36	52	81
Nucleosides (%)	8	7	15
Total (μmol/L)	402	163	141

^*See reference 63.

^†See reference 62.

Nucleotide supplementation also has been reported to reduce the number of episodes of infant diarrhea in a group of infants of lower socioeconomic status in Chile, in a manner analogous to that for protection afforded by human milk [66]. In 1998, it was reported that nucleotide-supplemented formulas promoted the immune response of infants to Haemophilus influenzae type b polysaccharide immunization at 7 months of age, and a similar response was observed for diphtheria immunization [67]. Infants fed human milk for more than 6 months showed a similar response and exhibited an enhanced titer response to oral polio vaccine; this latter response was not observed in the group fed nucleotide-supplemented formula [67]. Nucleotides are emerging as nutritional and immunologic components of human milk.

Nutritional Proteins

As noted, the nutritional proteins in human milk are classified as either whey (acid-soluble) or casein (acid-precipitable). Within these two classes of proteins, several specific proteins are responsible for supporting the nutritional needs of the infant.

Human casein is composed primarily of β-casein and κ-casein, although the actual distribution of these two proteins is unclear [68]. By contrast, bovine milk contains α_s1-casein and α_s2-casein (neither of which is found in human milk), in addition to β-casein and κ-casein [69]. These two human milk casein proteins seem to account for approximately 30% of the protein found in human milk, in contrast to the earlier calculation of 40% (the amount commonly used to prepare reconstituted, so-called humanized formulas from bovine milk, which normally contains 82% casein proteins).

The whey-protein fraction contains all of the proposed functional proteins in human milk (immunoglobulins, lysozyme, lactoferrin, enzymes, cytokines, peptide hormones), in addition to the major nutritional protein, α-lactalbumin. Whey proteins compose approximately 70% of human milk proteins, in contrast to 18% in bovine milk. Although α-lactalbumin is the major whey protein in human milk, β-lactoglobulin is the major whey protein in bovine milk (and is not found in human milk) [50]. A consistent fraction of human milk whey protein is composed of serum albumin. Its source is unclear; some evidence indicates that it may be synthesized in the mammary gland [70]. Most of the serum albumin probably is synthesized outside the mammary gland, however.

Milk proteins are characterized by their site of synthesis and are species specific. Proteins such as α-lactalbumin and β-lactoglobulin are species and organ specific, whereas proteins such as serum albumin are species specific, but not organ specific [71]. The net result of these differences in proteins used for nutrition by the neonate is that different amounts of amino acids are ingested by the neonate, depending on the source of milk; even reconstitution of the whey and casein classes of proteins from one species in a ratio similar to that of another species does not result in an identical amino acid intake. These differences are reflected in plasma amino acid profiles of infants fed commercial milks versus human milk, regardless of the ratios of reconstitution [54].

Bioactive Proteins and Peptides

Although a major proportion of human milk protein is composed of the nutritional proteins just described, many of the remaining proteins subserve a variety of functions, other than or in addition to the nutritional support of the neonate. These proteins include carrier proteins, enzymes, hormones, growth factors, immunoglobulins, and cytokines (the latter two are discussed later in “Resistance to Infection”). Whether these proteins are still functional after they have been ingested by the neonate has not always been established, but it is clear that human milk supplies a mixture that is potentially far more complex than just nutritional substrate.

Carrier Proteins

Many nutrients are supplied to the neonate bound to proteins found in human milk. This binding may play an important role in making these nutrients bioavailable. Lactoferrin is an iron-binding protein (a property that also may play a role in its bacteriostatic action) that is apparently absorbed intact by the infant [72]. Lactoferrin may be important in the improved absorption of iron by the infant from human milk compared with iron from cow's milk preparations, which contain little lactoferrin [28]. Lactoferrin also may bind other minerals, including zinc and manganese, although the preferred mineral form seems to be the ferric ion.

Numerous other proteins seem to be important as carriers of vitamins and hormones. Folate-binding, vitamin B₁₂–binding, and vitamin D–binding proteins all have been identified in human milk. These proteins apparently have some resistance to proteolysis, especially when they are saturated with the appropriate vitamin ligand [73]. Serum albumin acts as a carrier of numerous ligands, whereas α-lactalbumin acts as a carrier for calcium. Finally, proteins that bind thyroid hormone and corticosteroids have been reported to be present in human milk [74,75], although serum albumin may partially fulfill this function.

Enzymes

The activity of more than 30 enzymes has been detected in human milk [76]. Most of these enzymes seem to originate from the blood, with a few originating from secretory epithelial cells of the mammary gland. Little is known about the role of these enzymes, other than lysozyme and the lipases, in human milk. The enzymes found in human milk range from ATPases to antioxidant enzymes, such as catalase, to phosphatases and glycolytic enzymes. Although these enzymes have important roles in normal body metabolism, it is unclear how many of them either function in the milk itself or survive ingestion by the neonate to function in the neonate.

Lysozyme seems to have a part in the antibacterial function of human milk, whereas the lipases have a more nutrient-related role in modulating fat metabolism for the neonate. Two lipases have been identified in human milk: a lipoprotein lipase and a bile salt–stimulated lipase [77]. Lipoprotein lipase seem to be involved in determining the pattern of lipids found in human milk by regulating uptake into milk at the level of the mammary gland. Human milk bile salt–stimulated lipase is an acid-stable protein that compensates for the low activity of lipases secreted into the digestive tract during early development [78]. These two enzymes regulate the amount and the pattern of lipid that appears in milk and the extremely efficient absorption of lipid by the infant. Human milk lipid is absorbed much more readily than lipid from commercial milk formulas despite the many adaptations that have been made to improve absorption, illustrating the effective mechanisms supported by the lipases.

Hormones and Growth Factors

Peptide, steroid hormones, and growth factors have been identified in trace amounts in human milk, although as with most enzymes, it is unclear to what degree they function in the neonate who has ingested the milk. As discussed previously, binding proteins for corticosteroids and thyroxine have been identified in milk and, by extrapolation from observations of other milk components, may play a role in making these bioactive compounds more readily available to the infant.

Among the hormones identified in human milk are insulin, oxytocin, calcitonin, and prolactin. Most of these hormones apparently are absorbed by the infant, but their role in in vivo function is unclear [79]. Breast-fed infants seem to have a different endocrine response from that of formula-fed infants, presumably reflecting the intake of hormones from human milk [80]. The advantages or disadvantages to the infant of these responses are unknown.

Human milk also contains a rich mixture of growth factors, including epidermal growth factor, nerve growth factor, and transforming growth factor-β [81]. In addition, various gastrointestinal peptides have been identified in human milk. Presumably, the supply of these various factors to the infant through milk compensates for their possible deficiency in the infant during early development.

The composition of human milk provides a complex and complete nutritional substrate to the neonate. Human milk supplies not only individual nutrients, but also enzymes involved in metabolism, carriers to improve absorption, and hormones that may regulate metabolic rates. Commercial formulas have not yet been developed to the point that they can provide an analogous complete nutritional system.

Special Considerations for the Premature Neonate

Many of the benefits of human milk evident in term neonates are also evident in preterm neonates with very low birth weight (<1500 g) fed unfortified human milk; the incidence of necrotizing enterocolitis also seems to be reduced [[82], [83], [84]]. Nonetheless, the content of calcium, phosphorus, protein, sodium, vitamins, and energy in unfortified human milk may be inadequate to meet the needs of the very low birth weight preterm neonate, and if used as an exclusive source of nutrients, unfortified human milk may be associated with impaired growth and nutrient deficiencies. Meta-analysis of studies comparing these premature infants fed unfortified or fortified human milk found that fortified milk was associated with greater increases in weight, length, and head circumference and better nitrogen balance and bone mineral content without an increase in feeding intolerance or complications [83,85].

The feeding of fortified human milk does not seem to be associated with a substantially increased rate of feeding intolerance or with a reduction in the beneficial effects of human milk on rates of infections or necrotizing enterocolitis [84]. The quality of the data regarding the utility of fortified human milk is still limited, however, and additional research is needed to define better means by which to provide the benefits of human milk to these preterm infants, while meeting their specific nutritional requirements.

Resistance to infection

Component mechanisms of defense: origin and distribution

Fresh human milk contains a wealth of components that provide specific and nonspecific defenses against infectious agents and environmental macromolecules (Table 5–4 ). These component factors include cells such as T and B lymphocytes, polymorphonuclear neutrophils (i.e., polymorphonuclear leukocytes), and macrophages; soluble products, especially immunoglobulins; secretory immunoglobulin A (sIgA); immunomodulatory cytokines and cytokine receptors; components of the complement system; several carrier proteins; enzymes; and numerous endocrine hormones or hormone-like substances. Additional soluble factors that are active against streptococci, staphylococci, and tumor viruses also have been identified [24,86]. Epidermal growth factor promotes growth of mucosal epithelium and maturation of intestinal brush border.

TABLE 5–4.

Immunologically and Pharmacologically Active Components and Hormones Observed in Human Colostrum and Milk

Soluble	Cellular	Hormones and Hormone-like Substances
Immunoglobulin sIgA (11S), 7S IgA, IgG, IgM, IgE, IgD, secretory component	T lymphocytes	Epidermal growth factors
Cytokines (see Table 5–8)	B lymphocytes	Prostaglandins
Histocompatibility antigens	Neutrophils	Neurotensin
Complement	Macrophages	Relaxin
Chemotactic factors	Epithelial cells	Somatostatin
Properdin		Bombesin
α-fetoprotein		Gonadotropins
Folate uptake enhancer		Ovarian steroids
Carrier proteins		Thyroid-releasing hormone
Lactoferrin		Thyroid-stimulating hormone
Transferrin		Thyroxine and triiodothyronine
B12-binding protein		Adrenocorticotropin
Lysozyme		Corticosteroids
Lipoprotein lipase		Corticoid-binding protein
Leukocyte enzymes		Insulin

The developmental characteristics of sIgA have been studied more extensively than those of other components [[87], [88], [89]]. On the basis of available information, most IgA-producing cells observed in milk have their origin in the precursor immunocompetent cells in the gut-associated lymphoid tissue (GALT) and bronchus-associated lymphoid tissue (BALT). Exposure of IgA precursor B lymphocytes in GALT or BALT to microbial and dietary antigens in the mucosal lumen is an important prerequisite for their initial activation and proliferation. Such antigen-sensitized cells eventually are transported through the systemic circulation to other mucosal surfaces, including the mammary glands, and, as plasma cells, initiate the synthesis of immunoglobulin against specific antigens previously experienced in the mucosa of the respiratory or alimentary tract [87]. It has been proposed that T cells observed in the milk also may be derived from GALT and BALT in a manner similar to that of IgA-producing cells. Little or no information is available regarding the site of origin of other cellular or soluble immunologic components normally present in human milk. Specific antibody and cellular immune reactivity against many respiratory and enteric bacterial and viral pathogens and ingested food proteins also is present in human breast milk (Table 5–5 ).

TABLE 5–5.

Specific Antibody or Cell-Mediated Immunologic Reactivity in Human Colostrum and Milk

Bacteria	Viruses	Other
Escherichia coli	Rotavirus	Candida albicans
Salmonella	Rubella virus	Giardia species
Shigella species	Poliovirus types 1, 2, 3	Entamoeba histolytica
Vibrio cholerae	Echoviruses	Food proteins
Bacteroides fragilis	Coxsackieviruses A and B
Streptococcus pneumoniae	Respiratory syncytial virus*
Bordetella pertussis	Cytomegalovirus*
Clostridium tetani and Clostridium difficile	Influenza A virus
Corynebacterium diphtheriae	Herpes simplex virus type 1
Streptococcus mutans	Arboviruses
Haemophilus influenzae type B	Semliki Forest virus
Mycobacterium tuberculosis*	Ross River virus
	Japanese B virus
	Dengue virus
	Human immunodeficiency virus
	Hepatitis A and B viruses

^*Evidence of reactivity for antibody and cellular immunity.

Soluble products

IgA

As observed in other peripheral mucosal sites, the major class of immunoglobulin in human colostrum and milk is 11S sIgA. Other isotypes—7S IgA, IgG, IgM, IgD, and IgE—also are present. The 11S IgA exists as a dimer of two 7S IgA molecules linked together by a polypeptide chain, the J-chain, and is associated with a nonimmunoglobulin protein referred to as the secretory component. The sIgA protein constitutes about 75% of the total nitrogen content of human milk. The IgA dimers produced by plasma cells at the basal surface of the mammary epithelium are bound by the polymeric immunoglobulin receptor on the basolateral surface of mammary epithelial cells, which transports them through these cells, where they are released into the alveolar spaces as an 11S IgA dimer associated with a portion of the polymeric immunoglobulin receptor referred to as the secretory component [90].

Sequential quantitation of class-specific immunoglobulin in human colostrum and milk has shown that the highest levels of sIgA and IgM are present during the first few days of lactation (Fig. 5–1 ). Levels of IgA are 4 to 5 times greater than levels of IgM, 20 to 30 times greater than levels of IgG, and 5 to 6 times greater than levels of serum IgA [89]. As lactation progresses, IgA declines to levels of 20 to 27 mg per gram of protein, and IgM levels decline to 3.5 to 4.1 mg/g protein. IgG levels do not show any significant change during early and late lactation and usually are maintained in the range of 1.4 to 4.9 mg/g protein (see Fig. 5–1). Although a dramatic and rapid decline in milk IgA and IgM occurs during the first week of life, this decrease is more than balanced by an increase in the volume of milk produced as the process of lactation becomes established (Table 5–6 ; see Fig. 5–1).

FIGURE 5–1.

Comparison of mean levels of IgG, IgA, and IgM in colostrum and milk at different intervals after onset of lactation in mothers who were breast-feeding.

(Data from Ogra SS, Ogra PL. Immunologic aspects of human colostrum and milk, II: characteristics of lymphocyte reactivity and distribution of E-rosette forming cells at different times after the onset of lactation. J Pediatr 92:550-555, 1978.)

TABLE 5–6.

Level of Immunoglobulins in Colostrum and Milk and Estimates of Delivery of Lactational Immunoglobulins to the Breast-Feeding Neonate^*

Day Postpartum	Percentage of Total Proteins Represented by Immunoglobulin			Output of Immunoglobulin (mg/24 hr)
	IgG	IgM	IgA	IgG	IgM	IgA
1	7	3	80	80	120	11,000
3	10	45	45	50	40	2000
7	1-2	4	20	25	10	1000
7-28	1-2	2	10-15	10	10	1000
<50	1-2	0.5-1	10-15	10	10	1000

^*Estimates based on the available data for total immunoglobulin and daily protein synthesis (see references 8, 87, and 88).

IgA antibodies found in milk possess specificity for infectious agents endemic to or pathogenic for the intestinal and respiratory tracts (see Table 5–4). These antibodies may be present in the milk in the absence of specific circulating IgA. In a study in which pregnant women were given oral feedings of Escherichia coli O83, development of IgA antibody in human milk was evident in the absence of detectable serum antibody-specific responses [91]. In another study, investigators observed similar responses in animal models using intrabronchial immunization with Streptococcus pneumoniae. These and other studies [[92], [93], [94], [95]] have strongly supported the concept of a bronchomammary, and enteromammary, axis of immunologic reactivity in the breast.

Despite the elegance of studies that have defined the mechanisms of IgA cell trafficking from GALT and BALT to the mammary glands, the actual number of B cells or IgA plasma cells in the mammary glands is sparse. Colostrum and milk may contain large amounts of IgA (11 g in colostrum and 1 to 3 g per day in later milk), as shown in Table 5–6. The reasons underlying the apparent disparity between the content of immunoglobulin-producing cells and concentrations of immunoglobulin are unknown. The disparity may be related to the unique hormonal environment of the mammary glands. The hormones that have been consistently observed in human milk are listed in Table 5–4.

The effects of pregnancy-related and lactation-related hormones on regulation of immunologic reactivity present in the resting and lactating breast have been examined [96]. In a study of immunoglobulin production in the nonlactating human breast, several interesting findings were noted [97]. Few mononuclear cells were present in the nonlactating breast of nulliparous and of parous women, although IgA-containing cells predominated. Synthesis of IgA seemed to be slightly increased in the parous women. IgA was found in the mammary tissues during the proliferative stage of the menstrual cycle in the nulliparous women and during the luteal phase in the parous women. The number of IgA-producing cells in the nonlactating breast was observed to increase with parity. These findings suggest that the immunologic makeup of the nonlactating and the lactating breast may be significantly influenced by the hormonal milieu.

In another study of virgin mice given exogenously administered hormones [98], an extended exposure to estrogen, progesterone, and prolactin was necessary for maximal increments in IgA-producing plasma cells in the breast. Similarly, castrated males exposed to these hormones became moderately receptive to mammary gland homing of cells specific for IgA synthesis. As would be expected, testosterone eliminated female breast receptivity to these cells. These studies suggest the existence of a hormonally determined homing mechanism in the mammary gland for class-specific, immunoglobulin-producing cells.

More recent studies have proposed another possible influence of lactational hormones on immunocompetent cells. In limited observations, combinations of prolactin with estrogen and progesterone (in concentrations observed normally at the beginning of parturition) seemed to have an amplifying effect on the synthesis and secretion of IgA from peripheral blood lymphocytes [99]. This observation raises the possibility that the high levels of sIgA observed in colostrum and milk may be partly the result of selective, hormonally mediated proliferation of antigen-sensitized IgA cells in the peripheral blood. The immunoglobulin could acquire a secretory component during its passage through the mammary epithelium and eventually appear in the colostrum or milk as mature sIgA. Although the appearance of sIgA antibody in milk characteristically follows antigenic exposure in GALT or BALT, the precise nature of the IgA content in milk seems to be determined by various other factors operating in the mucosal lymphoid tissue. These factors include the dendritic cell and regulatory T-cell network in GALT and possibly in BALT [100,101], the nature of antigens (soluble proteins versus particulate microbial agents) [102], and the route of primary versus secondary antigenic exposure [103].

It has been estimated that the breast-fed infant may consistently receive about 1 g of IgA each day, and approximately 1% of this amount of IgM and IgG [104,105]. The estimates of lactational immunoglobulin delivered to the breast-fed infant at different periods of lactation are presented in Table 5–6. Most ingested IgA is eliminated in the feces, although 10% may be absorbed from the intestine into the circulation within the first 18 to 24 hours after birth [106]. Feces of breast-fed infants contain functional antibodies present in the ingested milk [107]. Other studies support the finding of prolonged survival of milk IgA in the gastrointestinal tract. Infants fed human milk have shown the presence of all immunoglobulin classes in the feces. Fecal IgA content was three to four times greater than that of IgM after human milk feeding. Comparative studies on survival of human milk IgA and bovine IgG in the neonatal intestinal tract have suggested that the fecal content of IgA may be 14 to 20 times greater after human milk feeding than that of bovine IgG after feeding of bovine immunoglobulin [108].

Endogenous production of sIgA by the infant's mucosal immune system increases progressively in the postnatal period [109]. Nonetheless, breast-fed infants have substantially greater concentrations of fecal sIgA than formula-fed infants. Prentice and colleagues [110] found approximately 10-fold and approximately 4-fold higher concentrations of sIgA in the stools of breast-fed infants compared with formula-fed infants at 6 and 12 weeks of postnatal age, even though only approximately 15% to 20% of ingested sIgA appeared in the feces.

Direct information about the role of milk IgA in antimicrobial defense is available in several studies. sIgA interferes with bacterial adherence to cell surfaces [111]. Colostrum and milk can inhibit the activity of E. coli and Vibrio cholerae enterotoxins in experimental settings [112]. The antitoxic activity of human milk seems to correlate well with its IgA content, but not with its IgM and IgG content. Precoating of V. cholerae with specific sIgA protects infant mice from disease [113]. Similar results have been obtained by using specific purified milk sIgA in preventing E. coli–induced and Shigella dysenteriae–induced disease in rabbits [114]. Less definite, but suggestive, is a study conducted with human milk feeding relative to the intestinal replication of orally administered live poliovirus vaccine [115]. This study found that breast-feeding may reduce the degree of seroconversion for poliovirus antibody in the vaccinated infants. Because antipolio IgA is present in human milk and colostrum, the investigators concluded that specific IgA may bind poliovirus and influence viral replication in the intestinal mucosa. Extensive experience with oral polio immunization worldwide has not found a convincing association, however, between breast-feeding and live vaccine failures. Other studies have shown that the magnitude of poliovirus replication in the intestine is determined by the presence and level of preexisting sIgA antibody. With high levels of intestinal IgA antibody, little or no replication of vaccine virus was observed in the gut. With lower levels, varying degrees of viral replication could be shown [116].

Epidemiologic studies strongly support the notion that breast-feeding protects the infant against infectious diseases (see “Benefits and Risks of Human Milk”). It is impossible in these studies, however, to dissect the relative contribution of sIgA from the contributions of other soluble or cellular components present in colostrum and milk.

IgG and IgM

Normal neonates exhibit characteristic paucity or lack of serum IgA and sIgA during the first 7 to 10 days after birth. At that time, the presence of IgM and IgG in milk may be important to compensate for immunologic functions not present in the mucosal sites. IgG and IgM participate in complement fixation and specific bactericidal activity, functions not associated with IgA. Studies done after oral feeding of immune serum globulin (mostly IgG) suggested that IgG may survive in the gastrointestinal tract of infants with low birth weight [117]. Other immunoglobulin isotypes in milk also may be able to serve as effective substitutes for IgA in the neonates of IgA-deficient mothers in prevention of infection with enteric or respiratory pathogens.

IgE and IgD

Investigations have failed to show local synthesis of IgE in the breast [[118], [119], [120]]. Although IgE may be detected in 40% of colostrum and milk samples, the concentrations are extremely low, and many samples of colostrum and milk contain no IgE activity when paired samples of serum contain high IgE levels. IgD has been detected in most colostrum and milk samples. It has been suggested that nursing women with high serum IgD levels are more likely to have high IgD concentrations in their milk. The possibility of some local production of IgE and IgD cannot be ruled out [120].

Cellular elements

Human colostrum and milk contain lymphocytes, monocytes-macrophages, neutrophils, and epithelial cells [121]. Early colostrum contains the highest concentration of cells, approximately 1 × 10⁶ to 3 × 10⁶ cells/mL. By the end of the first week of lactation, cell concentration is of the order of 10⁵ cells/mL. Total cell numbers delivered to the newborn throughout lactation may remain constant, however, when adjustments are made for the increase in volume of milk produced [89]. The two major cell populations in human milk—macrophages and neutrophils—are difficult to distinguish by common staining methods because of numerous intracytoplasmic inclusions. More accurate estimates made by flow cytometry analysis suggest that the relative percentages of neutrophils, macrophages, and lymphocytes in early milk samples are approximately 80%, 15%, and 4% [122,123]. The remaining cells are present in smaller amounts, especially in the absence of active suckling, engorgement, or local breast infection.

Macrophages

Histochemically, the milk macrophage differs from the blood monocyte in showing decreased peroxidase staining, with increased lysosomes and significant amounts of immunoglobulin, especially IgA, in the cytoplasm [[124], [125], [126]]. The intracellular immunoglobulin in macrophages represents 10% of milk IgA [127]. Kinetic studies on the release of IgA by human milk macrophages suggest that immunoglobulin release by macrophages, in contrast to immunoglobulin release by other phagocytic cells, is a time-dependent phenomenon and is not significantly influenced by the use of secretagogues or stimulants, such as phorbol myristate acetate [127]. Active phagocytosis is associated, however, with significant increase in release of IgA [128].

The precise functions of macrophages in colostrum or milk have not been fully explored. These cells have been suggested as potential transport vehicles for IgA [126,129]. Milk macrophages possess phagocytic activity against Staphylococcus aureus, E. coli, and Candida albicans, with possible cytocidal activity against the first two organisms [130]. Milk macrophages participate in antibody-dependent, cell-mediated cytotoxicity for herpes simplex virus type 1–infected cells [131]. Infection of milk macrophages by respiratory syncytial virus results in the production of the proinflammatory cytokines IL-1β, IL-6, and tumor necrosis factor (TNF)-α [132]. These cells also are involved in various other biosynthetic and excretory activities, including production of lactoferrin, lysozyme [133], components of complement [134], properdin factor B, epithelial growth factors, and T lymphocyte–suppressive factors [87]. Milk macrophages also have been suggested to be important in regulation of T-cell function [135,136].

Lymphocytes

Milk contains a few lymphocytes; 80% are T cells, and 4% to 6% are B cells [123]. The small number of B cells reflects the sessile nature of these cells, which enter the lamina propria of the mammary gland to transform into plasma cells. Although several investigators have been unable to show in vitro antibody synthesis by milk lymphocytes, studies performed with colostral B cells transformed by Epstein-Barr virus showed production of IgG and J-chain–containing IgM and IgA [137]. A small population of CD16⁺ natural killer (NK) cells also can be identified in most milk samples, but cannot be accurately quantitated [123]. In functional studies, colostral cells exhibit NK cytotoxicity, however, which is enhanced by interferon (IFN) and IL-2. Colostral cells also elicit antibody-dependent and lectin-dependent cellular cytotoxic responses. The NK and the antibody-dependent and lectin-dependent responses in colostral cells have been observed to be significantly lower than those of autologous peripheral blood cells. Reduced cellular cytotoxicity of colostral cells also has been observed against virus-infected targets and certain bacteria. With several specific virus-infected targets, colostrum and milk cells conspicuously lack cellular cytotoxicity compared with autologous peripheral blood cells. There is also an apparent exclusion of cytolytic T cells in the milk that are specific for certain human leukocyte antigens (HLA) [138,139].

Most T lymphocytes in colostrum and milk are mature CD3⁺ cells. CD4⁺ (helper) and CD8⁺ (cytotoxic) populations are present in human milk, with a proportion of CD8⁺ T cells higher than that found in human blood (Table 5–7 ). The CD4⁺/CD8⁺ ratio in milk is significantly lower than the ratio observed in peripheral blood and is not due to an increase of CD8⁺ cells in the peripheral blood of women during the postpartum period. Colostral and milk T lymphocytes manifest in vitro proliferative responses on stimulation with numerous mitogens and antigens. Several studies have shown a selectivity in lymphocyte stimulation responses in colostral and milk lymphocytes to various antigens compared with peripheral blood lymphocyte responses [140,141]. Antigens such as rubella virus stimulate T lymphocytes in secretory sites and milk and in systemic sites [140]. By contrast, E. coli K1 antigen, whose exposure is limited to mucosal sites, produces stimulation of lymphoproliferative responses only in milk lymphocytes. These findings support the concept of select T-cell populations in the mammary gland.

TABLE 5–7.

Lymphocyte Subpopulations in Human Milk and Autologous Blood^*

Lymphocyte Subpopulation	Human Milk	Blood
CD3+†	83 ± 11	75 ± 7
CD3+ CD4+†	36 ± 13	44 ± 6
CD3+ CD8+†	43 ± 12	27 ± 4
CD4+/CD8+‡	0.88 ± 0.35	1.70 ± 0.45
CD19+†	6 ± 4	14 ± 5

^*Expressed as mean ± standard deviation (SD).

^†Expressed as percentage of total lymphocytes.

^‡Ratio of CD3⁺/CD4⁺ to CD3⁺/CD8⁺ lymphocytes.

Adapted from Wirt DP, et al. Activated-memory T lymphocytes in human milk. Cytometry 13:282-290, 1992.

In addition to antigen selectivity, a general hyporesponsiveness to mitogenic stimulation of milk lymphocytes relative to peripheral blood lymphocytes has been observed [136,140]. The decreased reactivity of milk lymphocytes to phytohemagglutinin may be partly the result of a relative deficiency of certain populations of T cells in milk. Macrophage–T cell interactions also have been postulated to be responsible for this relative hyporesponsiveness [140], although it is unknown whether the effects are the result of decreased helper or increased regulatory function. Studies have shown that milk lymphocytes exhibit reduced responses to allogeneic cells, but display good ability to stimulate alloreactivity [138]. Generally, the T-cell proliferative responses to phytohemagglutinin and tetanus toxoid in breast-fed infants seem to be significantly higher than the responses in bottle-fed infants, possibly secondary to the presence of maternally derived cell growth factors and other lymphokines in human milk [136,142].

Virtually all CD4⁺ and CD8⁺ T cells in milk bear the CD45 isoform CD45RO that is associated with immunologic memory [123,143]. In addition, the proportion of T cells that display other phenotypic markers of activation, including CD25 (IL-2R) and HLA-DR, is much greater than that in blood [123,144]. Consistent with their memory phenotype, T cells in human milk produce IFN-γ [143]. A significantly greater number of CD4⁺ T cells in colostrum express the CD40 ligand (CD40L), which helps these cells provide help for B-cell antibody production and macrophage activation (see Chapter 4 for more details) compared with autologous or heterologous blood T cells [145]. The function of these memory T cells in the recipient human infant is currently unknown, however. Mucous membrane sites in the alimentary or respiratory tract, or both, of the recipient infant would seem to be potential entry sites for human milk leukocytes. Very few memory T cells are detected in blood in infancy [146]. It is possible that maternal memory T cells in milk may help compensate for the developmental delay in their production in the infant.

The proportion of T lymphocytes bearing the γδ T-cell receptor (γδ-TCR) is approximately two times greater in colostrum than in blood [147,148]. Human γδ-TCR⁺ cells populate organized lymphoid tissues and represent half of the intraepithelial lymphocytes in the gut [149]. The intestinal epithelia may have a selective affinity for γδ-TCR⁺ cells and provide a favorable environment for maternal T cells in milk to be transferred to the breast-fed infant. Evidence from experimental animal studies indicates that milk lymphocytes enter tissues of the neonate [[150], [151], [152], [153]], but this has not been shown in humans.

In humans, possible transfer of maternal T-cell reactivity to tuberculin protein from the mother to the neonate through the process of breast-feeding has been suggested [106,154,155]. The implications of these observations are that maternal cellular products or soluble mediators of cellular reactivity may be transferred passively to the neonate through breast-feeding. The occurrence of such phenomena in humans has not been studied carefully, and there is no direct evidence that milk T cells, either αβ-TCR⁺ or γδ-TCR⁺, play a role in the transfer of adoptive immunoprotection to the recipient infant. Similarly, there is at present no evidence to suggest that there is any immunologic risk to the human neonate of maternal T cells ingested through breast-feeding.

Neutrophils

Milk contains numerous neutrophils. Although the absolute counts in actively nursing mothers exhibit considerable variability among different samples, highest numbers are generally observed during the first 3 to 4 days of lactation. The numbers of neutrophils decrease significantly after 3 to 4 weeks of lactation, and only rare neutrophils are observed in samples collected after 60 to 80 days postpartum. Leukocytes in human milk seem to be metabolically activated. Although the neutrophils are phagocytic and produce toxic oxygen radicals, they do not respond well to chemoattractants by increasing their adherence, polarity, or directed migration in in vitro systems [156]. This diminished response was found to be due to prior activation in that the neutrophils in milk displayed a phenotypic pattern that is typical of activated neutrophils. The expression of CD11b, the α chain subunit of Mac-1, was increased, and the expression of L-selectin was decreased [122].

Epithelial Cells

On the basis of their anatomic distribution, epithelial cells in the human mammary gland can be classified into two main types: myoepithelial and luminal. Epithelial cells of both types seem to be more heterogeneous, however, on histologic and physicochemical testing [138,157,158]. They include secretory cells, which contain abundant rough endoplasmic reticulum, lipid droplets, and Golgi apparatus. The secretory cells apparently produce casein micelle. The squamous epithelial cells usually are seen in the regions of the cutaneous junction of the nipples, especially near the galactophores. The ductal or luminal cells, which exist in clusters, have many short microvilli, tight junctions, and remnants of desmosomes [157,158].

In human milk, relatively few epithelial cells are observed in the early phases of lactation. Most epithelial cells appear after 2 to 3 weeks and are seen in appreciable numbers, even 180 to 200 days after the onset of lactation. With the possible exception of the synthesis of secretory component and casein and possibly other products, with which secretory epithelial cells have been associated in the stroma of the mammary gland, the role of epithelial cells in the milk remains to be defined.

Possible Functions of Cellular Elements

The information reviewed so far provides strong evidence for the existence of numerous dynamic cellular reactions in the mammary gland, colostrum, and milk. The specific functional role, collectively or individually, for the epithelial cells, monocytes, neutrophils, or lymphocytes in the mammary gland or the milk remains to be defined. In view of the high degree of selectivity and the differences in the quantitative and functional distribution of cellular elements, it is suggested that the mammary gland, similar to mucosal surfaces, may function partitioned from the cellular elements in peripheral blood, in a manner similar to that for other peripheral sites (e.g., the genital tract) of the mucosal immune system. It is unknown, however, whether the characteristic proportions of macrophages, T lymphocytes, other cytotoxic cells, or epithelial cells are designed for any specific functions localized to the mammary gland in the lactating mother or to epithelium or lumen of the intestinal or respiratory mucosa of the breast-feeding infant, or both.

The observations on the transfer of delayed hypersensitivity reactions in human neonates and of graft-versus-host reactivity in the rat raise the possibility that milk cells may function as important vehicles in transfer of maternal immunity to neonates. The potential beneficial and harmful roles of such cell-mediated transfer through the mucosal routes need to be investigated further. The paucity of NK and other cytotoxic cells in the colostrum may have a role for the breast-feeding neonate, especially in influencing the antigen processing and uptake of replicating microorganisms and their immune response at systemic or mucosal levels or both. Although further elucidation of the functions of these cellular elements in the mammary glands and the suckling neonate is needed, it is likely that their presence in the milk represents a highly selective phenomenon and not a mere contamination with peripheral blood cells.

Other defense factors

Direct-Acting Antimicrobial Agents

General Features

The defense agents in human milk, although biochemically diverse, share the following features: (1) They usually are common to mucosal sites. (2) They are adapted to resist digestion in the gastrointestinal tract of the recipient infant. (3) They protect by noninflammatory mechanisms. (4) They act synergistically with each other or with factors produced by the infant. (5) Most components of the immune system in human milk are produced throughout lactation and during gradual weaning. (6) There is often an inverse relationship between the production of these factors in the mammary gland and their production by the infant during the same time frames of lactation and postnatal development. As lactation proceeds, the concentration of many factors in human milk declines. Concomitantly, the mucosal production of these factors increases in the developing infant. Whether the inverse relationship between these processes is due to feedback mechanisms, or the processes are independent, is unclear.

Carbohydrate Components

Human milk contains several oligosaccharides and glycoconjugates, including monosialogangliosides that are receptor analogues for heat-labile toxins produced by V. cholerae and E. coli [159]; fucose-containing oligosaccharides that inhibit the hemagglutinin activity of the classic strain of V. cholerae [160]; fucosylated oligosaccharides that protect against heat-stable enterotoxin of E. coli [161]; mannose-containing, high-molecular-weight glycoproteins that block the binding of the El Tor strain of V. cholerae [159]; and glycoproteins and glycolipids that interfere with the binding of colonization factor (CFA/II) fimbriae on enterotoxigenic E. coli [162]. The inhibition of toxin binding is associated with acidic glycolipids containing sialic acid (gangliosides). Although the quantities of total gangliosides in human milk and in bovine milk are similar, the relative frequencies of each type of ganglioside in milk from these two species are distinct.

More than 50 types of monosialylated oligosaccharides have been identified in human milk, and new types are still being recognized [163]. Monosialoganglioside 3 constitutes about 74% of total gangliosides in human milk, but the percentage is much lower in bovine milk [164,165]. Also, the level of the enterotoxin receptor ganglioside G_M1 is 10 times greater in human milk than in bovine milk [165]. This difference may be clinically important because G_M1 inhibits enterotoxins of E. coli and V. cholerae [166]. Also, intact human milk fat globules and the mucin from the membranes of these structures inhibit the binding of S-fimbriated E. coli to human buccal epithelial cells [167].

Oligosaccharides in human milk also interfere with the attachment of H. influenzae and S. pneumoniae [168]. In this regard, N-acetylglucosamine (G1cNAc) (1-3) Gal-disaccharide subunits block the attachment of S. pneumoniae to respiratory epithelium. Human milk can interfere with the binding of human immunodeficiency virus (HIV) envelope antigen gp120 to CD4 molecules on T cells [169]. Some evidence from animal models suggests that the oligosaccharides and glycoconjugates in human milk protect in vivo [[170], [171], [172]], but relevant clinical data are scarce [173] and transmission of HIV through breast-feeding is a well-established means for mother-to-infant transmission of HIV (see “Benefits and Risks of Human Milk”).

In addition to the direct antimicrobial effects of the carbohydrates in human milk, nitrogen-containing oligosaccharides, glycoproteins, and glycopeptides are growth promoters for Lactobacillus bifidus var. pennsylvanicus [174,175,176]. The ability of human milk to promote the growth of L. bifidus may reside in the oligosaccharide moiety [177] and peptides [178] of caseins. It seems that these factors are responsible to a great extent, however, for the predominance of Lactobacillus species in the bacterial flora of the large intestine of the breast-fed infant. These bacteria produce large amounts of acetic acid, which aids in suppressing the multiplication of enteropathogens. It also has been reported that Lactobacillus species strain GG aids in the recovery from acute rotavirus infections [179] and may enhance the formation of circulating cells that produce specific antibodies of the IgG, IgA, and IgM isotypes and serum levels of those antibodies [180].

Generation of Antiviral, Antiparasitic Lipids from Substrata in Human Milk

Human milk supplies defense agents from fat as it is partially digested in the recipient's alimentary tract. Fatty acids and monoglycerides produced from milk fats by bile salt–stimulated lipase or lipoprotein lipase in human milk [78], lingual/gastric lipase from the neonate from birth, or pancreatic lipase after a few weeks of age are able to disrupt enveloped viruses [181,182]. These antiviral lipids may aid in preventing coronavirus infections of the intestinal tract [183] and may defend against intestinal parasites such as Giardia lamblia and Entamoeba histolytica [184,185].

Proteins

The principal proteins in human milk that have direct antimicrobial properties include the following.

α-Lactalbumin

α-Lactalbumin is a major component of the milk proteins and may possess some important functions of immunologic defense. This protein appears as large complexes of several α-lactalbumin molecules, which can induce apoptosis in transformed embryonic and lymphoid cell lines.

Lactoferrin

Lactoferrin, the dominant whey protein in human milk, is a single-chain glycoprotein with two globular lobes, both of which display a site that binds ferric iron [186]. More than 90% of the lactoferrin in human milk is in the form of apolactoferrin (i.e., it does not contain ferric iron) [187], which competes with siderophilic bacteria and fungi for ferric iron [[188], [189], [190], [191], [192]] and disrupts the proliferation of these microbial pathogens. The epithelial growth-promoting activities of lactoferrin in human milk also may aid in the defense of the recipient infant [193]. The mean concentration of lactoferrin in human colostrum is 5 to 6 mg/mL [194]. As the volume of milk production increases, the concentration decreases to about 1 mg/mL at 2 to 3 months of lactation [194,195].

Because of its resistance to proteolysis [[196], [197], [198]], the excretion of lactoferrin in stool is higher in infants fed human milk than in infants fed cow's milk [72,[199], [200], [201]]. The mean intake of milk lactoferrin per day in healthy breast-fed, full-term infants is about 260 mg/kg at 1 month of lactation and 125 mg/kg by 4 months [199]. The quantity of lactoferrin excreted in the stools of low-birth-weight infants fed human milk is approximately 185 times that excreted in stools of infants fed a cow's milk formula [202]. That estimate may be too high, however, because of the presence of immunoreactive fragments of lactoferrin in the stools of infants fed human milk [203]. Consistent with this suggestion, Prentice and colleagues [110] found that only approximately 1% of lactoferrin ingested in breast milk was excreted intact in stool by 6 weeks of postnatal age.

In addition, a significant increment in the urinary excretion of intact and fragmented lactoferrin occurs as a result of human milk feedings [110,203,204]. Stable isotope studies suggest that the increments in urinary lactoferrin and its fragments are principally from ingested human milk lactoferrin [205].

Lysozyme

Relatively high concentrations of lysozyme single-chain protein are present in human milk [194,195,206]. This 15-kDa agent lyses susceptible bacteria by hydrolyzing β-1,4 linkages between N-acetylmuramic acid and 2-acetylamino-2-deoxy-d-glucose residues in cell walls [207]. Lysozyme is relatively resistant to digestion by trypsin or denaturation owing to acid. The mean concentration of lysozyme is about 70 μg/mL in colostrum [194], about 20 μg/mL at 1 month of lactation, and 250 μg/mL by 6 months [195]. The approximate mean daily intake of milk lysozyme in healthy, full-term, completely breast-fed infants is 3 to 4 mg/kg at 1 month of lactation and 6 mg/kg by 4 months [199].

Few studies have been conducted to examine the fate of human milk lysozyme ingested by the infant. The amount of lysozyme excreted in the stools of low-birth-weight infants fed human milk is approximately eight times that of the amount excreted in the stools of infants fed a cow's milk formula [202], but the urinary excretion of this protein does not increase as a result of human milk feedings.

Fibronectin

Fibronectin, a high-molecular-weight protein that facilitates the uptake of many types of particulates by mononuclear phagocytic cells, is present in human milk (mean concentration in colostrum is 13.4 mg/L) [208]. The in vivo effects and fate of this broad-spectrum opsonin in human milk are unknown.

Complement Components

The components of the classic and alternative pathways of complement are present in human milk, but the concentrations of these components except C3 are exceptionally low [209,210].

Anti-inflammatory Agents

Although a direct anti-inflammatory effect of human milk has not been shown in vivo, numerous clinical observations suggest that breast-feeding protects the recipient infant from injury to the intestinal or respiratory mucosa [211,212]. This protection may be partly due to the more rapid elimination or neutralization of microbial pathogens in the lumen of the gastrointestinal tract by specific or broad-spectrum defense agents from human milk, but other features of human milk suggest that this is not the sole explanation. Proinflammatory mediators are poorly represented in human milk [213]. By contrast, human milk contains a host of anti-inflammatory agents [214], including a heterogeneous group of growth factors with cytoprotective and trophic activity for the mucosal epithelium, antioxidants, antiproteases, cytokines and cytokine receptors and antagonists, and other bioactive agents that inhibit inflammatory mediators or block the selected activation of leukocytes. Similar to the antimicrobial factors, some of these factors are well adapted to operate in the hostile environment of the recipient's alimentary tract.

Growth factors in human milk include epidermal growth factor [215,216], transforming growth factor-α and transforming growth factor-β [217,218], lactoferrin [193], and polyamines [219,220]. These and a host of hormones [221], including insulin-like growth factor, vascular endothelial growth factor, growth hormone–releasing factor, hepatocyte growth factor, prolactin, leptin, and cortisol [222], may affect the growth and maturation of epithelial barriers, limit the penetration of pathogenic microorganisms and free antigens, and prevent allergic sensitization. Corticosterone, a glucocorticoid that is present in high concentrations in rat milk, speeds gut closure in the neonatal rat [223]. Although macromolecular absorption does not seem to be as marked in the human neonate [[224], [225], [226]], the function of the mucosal barrier system in early infancy is important to host defense, and this system may be affected by factors in human milk. In this regard, the maturation of the intestinal tract as measured by mucosal mass, DNA, and protein content of the small intestinal tract seems to be influenced by milk, particularly early milk secretions [227].

Antioxidant activity in colostrum has been shown to be associated with an ascorbate compound and uric acid [228]. In addition, two other antioxidants present in human milk, α-tocopherol [229,230] and β-carotene [230], are absorbed into the circulation by the recipient gastrointestinal mucosa. Serum vitamin E concentrations increase in breast-fed infants from a mean of 0.3 mg/mL at birth to approximately 0.9 mg/mL on day 4 of life [229].

Very high concentrations of the pleiotropic anti-inflammatory and immunoregulatory cytokine IL-10, which attenuates dendritic cell, macrophage, T-cell, and NK-cell function, have been shown in samples of human milk collected during the first 80 hours of lactation [231]. IL-10 is present not only in the aqueous phase of the milk, but also in the lipid layer. Its bioactive properties were confirmed by the finding that human milk samples inhibited blood lymphocyte proliferation and that this property was greatly reduced by treatment with anti–IL-10 antibody. Mice with a targeted disruption in the IL-10 gene, when raised under conventional housing conditions, spontaneously develop a generalized enterocolitis that becomes apparent at the age of 4 to 8 weeks (time of weaning) [232]. These observations suggest that IL-10 in rat milk, and perhaps in human milk, may play a crucial role in the homeostasis of the immature intestinal barrier by regulating aberrant immune responses to foreign antigens.

Soluble cytokine receptors and cytokine receptor antagonists also are potent anti-inflammatory agents. Human colostrum and mature milk have been shown to contain biologically active levels of IL-1 receptor antagonist (IL-1Ra) and soluble TNF-α receptors I and II (sTNF-αRI and sTNF-αRII) [233]. The in vivo relevance of these observations also has been shown in a chemically induced colitis model of rats. Animals with colitis fed human milk had significantly lower neutrophilic inflammation than animals fed either chow or infant formula [234]. Similar “protective” effects were seen in rats with colitis fed an infant formula supplemented with IL-1Ra [234], suggesting that this anti-inflammatory agent present in milk may contribute to the broad protection against different injuries provided by human milk feeding.

The presence in human milk of platelet-activating factor acetylhydrolase (PAF-AH), the enzyme that catalyzes the degradation and inactivation of PAF, is intriguing [235]. Elevated serum concentrations of PAF have been found in rat and human neonates with necrotizing enterocolitis, whereas the concentrations of PAF-AH were found to be significantly lower than in control (unaffected) neonates [236,237]. Serum concentrations of PAF-AH at birth are less than concentrations in adults and gradually increase [237]. The enzyme is actively transferred from the mucosal to the serosal fluid in intestine of neonatal rats, particularly in the earliest postnatal period [238]. Other anti-inflammatory factors present in human milk include an IgE-binding factor, related antigenically to FcεRII (the lower affinity receptor for IgE), that suppresses the in vitro synthesis of human IgE [239] and the glycophosphoinositol-containing molecule protectin (CD59) that inhibits insertion of the complement membrane attack complex to cell targets [240]. The in vivo fate and effects of these anti-inflammatory factors in human milk are still poorly understood.

Modulators of the Immune System

Several seemingly unrelated types of observations suggest that breast-feeding modulates the development of the immune system of the recipient infant.

• •Prospective and retrospective epidemiologic studies have shown that breast-fed infants are at less risk for development of certain chronic immunologically mediated disorders later in childhood, including allergic diseases [241], Crohn disease [242], ulcerative colitis [243], insulin-dependent diabetes mellitus [244], and some lymphomas [245].
• •Humoral and cellular immune responses to specific antigens (i.e., vaccines) given during the first year of life seem to develop differently in breast-fed and in formula-fed infants. Several studies have reported increased serum antibody titers to H. influenzae type b polysaccharide [246], oral poliovirus [247], tetanus [248], and diphtheria toxoid [249] immunizations in breast-fed infants. In regard to cell-mediated immunity, breast-fed infants given bacille Calmette-Guérin vaccine at birth or later show a significantly higher lymphocyte transformation response to purified protein derivative than infants who were never breast-fed [249]. Maternal renal allografts survive better in individuals who were breast-fed than in individuals who were not [[250], [251], [252]]. In this respect, the in vitro allogeneic responses between the blood lymphocytes of mothers (stimulating cells) and their infants (responding cells), as measured by an analysis of the frequencies of cytotoxic T lymphocyte (CTL) precursors directed against HLA alloantigens (CTL allorepertoire), are low in breast-fed infants [253].
• •Increased levels of certain immune factors in breast-fed infants, which could not be explained simply by passive transfer of those substances, also suggest immunomodulatory activity of human milk. Breast-fed infants produce higher blood levels of IFN in response to respiratory syncytial virus infection [254]. It also was found that the increments in blood levels of fibronectin that were achieved by breast-feeding could not be due to the amounts of that protein in human milk [208]. In addition, it was found that human milk feeding led to a more rapid development in the appearance of sIgA in external secretions [202,204,110,248,255], some of which, such as urine, are far removed anatomically from the route of ingestion [204,110].
• •Breast milk inhibits the response of human adult and fetal intestinal epithelial cells, dendritic cells, and monocytes to ligand-induced activation by toll-like receptor 2 and TLR3, but augments activation via TLR4 and TLR5 [256]. Inhibition of activation via TLR2 seemed to be mediated by a soluble form of this receptor [257], whereas augmentation of TLR4 signaling was associated with an as yet uncharacterized protein. These immunomodulatory activities were not present in infant formula. These results suggest that breast milk can modulate the gut innate immune recognition of and response to microbes, which may affect the nature of the gut microbial flora (e.g., microbiome) and risk for disease in early and later life [258,259].

These and other observations suggest that the ability of human milk to modulate the development of the infant's own mucosal and systemic immune systems may be associated with immunoregulatory factors present in colostrum and in more mature milk. Several different types of immunomodulatory agents can be identified in human milk [214]. Among the numerous substances with proven or potential ability to modulate the infant immune response are prolactin [260], α-tocopherol [229], lactoferrin [261], nucleotides [67], anti-idiotypic sIgA [262], and cytokines [263]. It is evident that many of these factors in milk have other primary biologic functions, as in the case of hormones or growth factors, and that their potential as immunoregulatory agents overlaps with their antimicrobial or anti-inflammatory properties [214].

Cytokines in Human Milk

In the 1990s, several cytokines, chemokines, and growth factors that mediate the effector phases of natural and specific immunity were discovered in human milk, and many more have been identified subsequently (Table 5–8 ). Human milk displays numerous biologic activities characteristic of cytokines, including the stimulation of growth, differentiation of immunoglobulin production by B cells [125,264,265], enhancement of thymocyte proliferation [266], inhibition of IL-2 production by T cells [267], and suppression of IgE production [239]. IL-1β [268] and TNF-α [269] were the first two cytokines quantified in human milk. In colostrum, TNF-α is present mainly in fractions of molecular weight of 80 to 195 kDa, probably bound to its soluble receptors [233]. Milk TNF-α is secreted by milk macrophages [269,270] and by the mammary epithelium [271].

TABLE 5–8.

Cytokines, Chemokines, and Colony-Stimulating Factors in Human Milk

Cytokines	Chemokines	Colony-Stimulating Factors
IL-1β	CXCL1	G-CSF
IL-2	CXCL8 (IL-8)	GM-CSF
IL-4	CXCL9	M-CSF
IL-6	CXCL10	Erythropoietin
IL-7	CCL2
IL-8	CCL5	Interferons
IL-10	CCL11
IL-13
IL-15	TGF-β
IFN-γ
TNF-α

G-CSF, granulocyte colony-stimulating factor; GM-CSF, granulocyte-macrophage colony-stimulating factor; IL, interleukin; M-CSF, monocyte colony-stimulating factor; TGF-β, transforming growth factor-β; TNF-α, tumor necrosis factor-α.

Adapted from references 277 and 278 and other sources.

IL-6 was first shown in human milk by a specific bioassay [272]. In this study, anti–IL-6–neutralizing antibodies inhibited IgA production by colostrum mononuclear cells, suggesting that IL-6 may be involved in the production of IgA in the mammary gland. The presence of IL-6 in milk also has been shown by immunoassays [269,271,273,274]. Similarly, IL-6 is localized in high-molecular-weight fractions of human milk [273]. The association of IL-6 with its own receptor has not been studied in milk, although the expression of IL-6 receptor by the mammary epithelium [271] and in secreted form in the milk [233] may explain the high molecular weight of this cytokine in human milk. The expression of IL-6 messenger ribonucleic acid (mRNA) and protein in milk cells and in the mammary gland epithelium suggests that milk mononuclear cells and the mammary gland are likely major sources of this cytokine [270,271,275].

The presence of IFN-γ in human milk also has been reported [148,271,274], although some investigators have found significant levels of IFN-γ only in milk samples obtained from mothers whose infants had been delivered by cesarean section. The significance of this observation is unclear at present. IFN-γ bioactivity and its association with specific subsets of milk T cells also remains to be determined [148]. (The presence and possible function of IL-10 in human milk are discussed in “Anti-inflammatory Agents.”)

Chemokines are a novel class of small cytokines with discrete target cell selectivity that are able to recruit and activate different populations of leukocytes (see Chapter 4). Chemokines are grouped into families, which are defined by the spacing between cysteine residues (see Chapter 4). CXC chemokines, in which cysteine pairs are separated by one amino acid, and CC chemokines, in which paired cysteines are adjacent to each other, are found in human milk. Certain CXC chemokines that contain an ELR motif, including CXCL8 (also known as IL-8) and CXCL1 (GRO-α), predominantly attract neutrophils, whereas basophils, eosinophils, dendritic cells, monocytes, and specific subsets of T and B lymphocytes are attracted by specific CC chemokines and non-ELR CXC chemokines. The presence of many CXC and CC chemokines has been described in human milk (see Table 5–8) [[275], [276], [277], [278]].

Colony-stimulating factors—highly specific protein factors that regulate cell proliferation and differentiation in the process of hematopoiesis—are also present in human milk [148,[279], [280], [281], [282]]. The concentrations of monocyte colony-stimulating factor in particular seem to be 10-fold to 100-fold higher in human milk than in serum, and monocyte colony-stimulating factor evidently is produced by epithelial cells of the ducts and alveoli of the mammary gland under the regulatory activity of female sex hormones [281].

Although it is tempting to speculate that cytokines present in milk may be able to interact with mucosal tissues in the respiratory and alimentary tracts of the recipient infant, the functional expression of specific receptors for cytokines on epithelial or lymphoid cells in the airway and gastrointestinal mucosa has not been fully explored [214]. A receptor-independent mechanism of cytokine uptake by the gastrointestinal mucosa during the neonatal period has not been shown to date.

Whether and to what extent cytokines in breast milk contribute to the beneficial effects of human milk in the gut and elsewhere is largely unknown. Indirect evidence suggests that IL-7, which is a growth factor for T-cell progenitors (and for memory T cells), in breast milk may support thymic growth. Thymus size was found to be larger in breast-fed than in formula-fed 4-month-old infants in Denmark [283], and thymus size and IL-7 content of breast milk were directly correlated with each other in exclusively breast-fed infants in the Gambia [284]. In the latter study, reduced breast milk content of IL-7 was observed in the “hungry season” in association with reduced thymus size and thymic production of T cells; however, whether IL-7 in breast milk was absorbed intact and causally related to greater thymus size or was merely a surrogate for other factors cannot be determined from this study.

Milk and altered pregnancy

Several investigators have examined the effects of prematurity, early weaning, galactorrhea, and maternal malnutrition on the process of lactation. The immunologic aspects of these studies have focused largely on evaluation of the total content of sIgA and specific antibody activity. As described previously, the mammary secretions of the nonlactating breast contain sIgA, although the amount seems to be much lower than in the lactating breast [285]. Mammary secretions of patients with galactorrhea seem to contain sIgA in concentrations similar to those of normal postpartum colostrum [286]. Although malnutrition has been associated with reduced secretory antibody response in other external secretions, maternal malnutrition does not seem to affect the total sIgA concentration or antimicrobial-specific antibody activity in the milk [287]. By contrast, as noted in the preceding paragraph, poor maternal nutrition in the “hungry season” in the Gambia was associated with decreased thymus size [284].

The nutritional and immunologic composition of milk from mothers of premature infants seems to be significantly different from that of milk from mothers of infants born at term [16,49,195,288]. Comparative studies conducted during the first 12 weeks of lactation suggest that the mean concentrations of lactoferrin and lysozyme are higher in preterm than in term milk. sIgA is the predominant immunoglobulin in preterm and in term milk, although the sIgA concentration seems to be significantly higher in preterm milk collected during the first 8 to 12 weeks of lactation. sIgA antibody activity against certain organisms (E. coli somatic antigen) in the preterm milk was observed to be less than, or at best similar to, that found in term milk. In addition, the number of lymphocytes and macrophages in milk seems to be lower at 2 weeks, but significantly higher at 12 weeks in milk from mothers with preterm (born at 34 to 38 weeks of gestational age) infants than in milk from mothers with full-term infants [288]. The authors of these investigations have proposed that some of the observed changes may reflect the lower volume of milk produced by mothers delivered of preterm infants. The possibility remains that changes in the immunologic profile of preterm milk may be a consequence of inadequate stimulation by the preterm infant, alterations in the maternal hormonal milieu, or other factors underlying premature delivery itself.

Benefits and risks of human milk

Benefits

Gastrointestinal Homeostasis and Prevention of Diarrhea

Development of mucosal integrity in the gut seems to depend on maturation of the mucosal tissue itself and the establishment of a normal gut flora. The former represents an anatomic and enzymatic blockage to invasion of microorganisms and antigens, and the latter represents an inhibition of colonization by pathogenic bacteria. Although permeability of the neonatal gut to immunoglobulin is short-lived, damaged neonatal gut is permeable to a host of other proteins and macromolecules for several weeks or longer. Large milk protein peptides and bovine serum albumin have been shown to enter the circulation and to induce a circulating antibody response. The inflamed or ischemic gut is even more porous to antigens and pathogens. Various proven and presumed mechanisms for the role of sIgA and the normal flora have been proposed to compensate for these temporary inadequacies.

Breast-feeding has been strongly implicated in supporting gastrointestinal homeostasis in the neonate and in establishing normal gut flora. Observations have shown a reduced rate of diarrheal disease in breast-fed infants, even in the face of contamination of the fed milk with E. coli and Shigella species [289]. A preventive and therapeutic role for breast-feeding also has been suggested in nursery outbreaks of diarrheal disease caused by enteropathogenic strains of E. coli [290] and rotavirus [291]. Breast-feeding plays an inhibitory role in the appearance of E. coli O83 agglutinins found in the feces of colonized infants. In addition, the diverse serotypes of aerobic, gram-negative bacilli present in the oropharynx and the gastrointestinal tract of the neonate may serve as a source of antigen to boost the presensitized mammary glands, leading to a further modulation of specific bacterial growth in the mucosa [292]. The precise role of antibody that blocks adherence of these pathogens to the gut and the effects of other factors, such as lactoperoxidase, lactoferrin, lysozyme, and the commensal flora, in those situations are uncertain.

Extensive epidemiologic evidence supports the “prophylactic value” of breast-feeding, particularly exclusive breast-feeding. in the first 6 months of life with the addition of complementary feeding thereafter, in the prevention or amelioration of diarrheal disease in infants and young children in developed and developing nations and is summarized in several reviews [1,3,86,87,293]. Ample experimental animal data on the value of specific colostral antibody in preventing diarrheal illness are available from studies of colostral deprivation. These include colibacteriosis associated with E. coli K88 in swine; rotaviral gastroenteritis in cattle, swine, and sheep; and diarrheal illness associated with transmissible gastroenteritis of swine [294]. In humans, cholera is rare in infancy, especially in endemic areas where the prevalence of breast-feeding is high. The experience with an outbreak of cholera in the Persian Gulf lends support to the possibility that the absence of breast-feeding is an important variable in increasing the risk of cholera in infancy.

A few reports have claimed that nursery outbreaks of diarrhea associated with enteropathogenic strains of E. coli can be interrupted by use of breast milk. Conflicting data exist regarding prevention of human rotaviral infection and disease. Evaluation of nursery outbreaks of rotaviral disease has suggested that the incidence of infection and of illness was lower in breast-fed infants, but the incidence of symptoms in formula-fed infants also was very low. Studies done in Japan have noted a fivefold decrease in incidence of rotaviral infection among breast-fed infants younger than 6 months. Most rotavirus infections in neonates are asymptomatic, regardless of breast-feeding or bottle-feeding [[295], [296], [297]]. On the basis of careful clinical observations, Bishop and coworkers[298] in Australia first questioned the positive effects of breast-feeding in rotavirus infection.

More recent case-control studies of enteric viral infections in breast-fed infants have suggested that breast-feeding may protect infants from hospitalization rather than from infection itself [299,300]. Longitudinal follow-up of a large cohort of infants during a community outbreak of rotavirus has shown that attack rates of rotavirus infection were similar in breast-fed and in bottle-fed infants. The frequency of clinical disease with diarrhea seemed to be significantly lower in breast-fed infants, however. The protection observed in these patients was more a reflection of altered microbial flora from breast-feeding than of specific immunologic protection against rotavirus. Breast milk contains sIgA directed against rotavirus, with the greatest concentrations in colostrum and lower amounts in mature milk; neutralization of rotavirus by breast milk correlates imperfectly with antibody concentrations measured by enzyme immunoassay, suggesting that other factors present in milk contribute to rotavirus neutralization [301]. It seems that breast-feeding provides significant protection against diarrheal disease, although the mechanisms of such protection remain to be more fully defined [299,300].

Necrotizing Enterocolitis

Necrotizing enterocolitis is a complex illness of the stressed premature infant, often associated with hypoxia, gut mucosal ischemia, and necrolysis and death [302,303]. Clinical manifestations have occasionally been associated with bacteremia and invasion by gram-negative bacilli, particularly Klebsiella pneumoniae, into the intestinal submucosa. Clinical manifestations include abdominal distention, gastric retention, and bloody diarrhea. Classic radiographic findings include air in the bowel wall (pneumatosis intestinalis), air in the portal system, and free infradiaphragmatic air (signifying perforation). Treatment involves decompression, systemic antibiotics, and, often, surgery [[304], [305], [306]].

Numerous studies have suggested a beneficial role of breast milk in preventing or modifying the development of necrotizing enterocolitis in high-risk human infants. Some pediatric centers have claimed virtual absence of necrotizing enterocolitis in breast-fed infants; however, many instances of the failure of milk feeding to prevent human necrotizing enterocolitis also have been reported [307]. Outbreaks of necrotizing enterocolitis related to Klebsiella and Salmonella species secondary to banked human milk feedings have been documented [124,308,309]. In an asphyxiated neonatal rat model of necrotizing enterocolitis, the entire syndrome could be prevented with feeding of maternal milk. The crucial factor in the milk seemed to be the cells, probably the macrophages [124]. It also is possible that antibody and nonspecific factors play a role, as does establishment of a gut flora. Prophylactic oral administration of immunoglobulin has been found to have a profound influence on the outcome of necrotizing enterocolitis in well-controlled studies [309]. Penetration of the gut by pathogens and antigens is increased with ischemic damage, and noncellular elements of milk may aid in blockage of this transit [310].

Necrotizing enterocolitis is a complex disease entity whose pathogenesis and cause remain to be defined. Although breast-feeding may be protective, many other factors are related to the mechanism of mucosal injury and the pathogenesis of this syndrome.

Neonatal Sepsis

The incidence of bacteremia among premature infants fed breast milk has been suggested to be significantly lower than that among premature infants receiving formula feedings or no feeding [[311], [312], [313]]. A decrease in the incidence of neonatal sepsis, including sepsis associated with gram-negative bacilli and E. coli serotype K1, also has been linked to breast-feeding [[314], [315], [316]]; antibody and compartmentalized cellular reactivity to this serotype have been shown in human colostrum. Other studies have failed, however, to show clear evidence of protection against systemic infection in breast-fed infants [[317], [318], [319]]. A review of the evidence for protection of very low birth weight neonates from late-onset sepsis through the use of human milk suggested that the quality of the current evidence was insufficient to show a beneficial effect [320], although other authors have reached different conclusions [84,321].

Prevention of Atopy and Asthma

One of the most challenging developments in human milk research has been the demonstration in breast-fed infants of a reduced incidence of diseases with autoregulated or dysregulated immunity, long after the termination of breast-feeding [[241], [242], [243], [244], [245]]. Since the first report in 1936 [322], numerous published studies have addressed the effect of infant feeding on the development of atopic disease and asthma. Beneficial results of breast-feeding as prophylaxis against atopy have been observed in most of the studies; however, in others, beneficial effects were reported only in infants with a genetically determined risk for atopic disease. Finally, no beneficial effect at all or even an increased risk has been suggested in some breast-fed infants.

Kramer [323], in an extensive meta-analysis of 50 studies published before 1986 that focused on infant feeding and atopic disease, attempted to shed some light on the controversy. Of the 13 studies on asthma included in this analysis, 7 claimed a protective effect of breast-feeding, whereas 6 claimed no protection. Several serious methodologic drawbacks have been noted in this analysis, however. In many of the studies analyzed, early infant feeding history was obtained months or years after the feeding period, ascertainment of the infant feeding history was obtained by interviewers who were aware of the disease outcome, or insufficient duration and exclusivity of breast-feeding were documented; all were confounding variables that considered inappropriate “exposure standards.” Nonblind ascertainment of disease outcome was found to be the most common violation of the “outcome standards.”

Kramer's analysis also found that failure to control for confounding variables was a common violation in “statistical analysis standards” identified in several studies. The effect of infant feeding on subsequent asthma may be confounded by other variables that are associated with infant feeding and with unique investigational conditions. Factors that seem to have the greatest potential for confounding effects include family history of atopic disease, socioeconomic status, and parental cigarette smoking. Only 1 of 13 studies on asthma included in the meta-analysis adequately controlled for these confounding factors. Three of the studies that did not show a protective effect of breast-feeding on asthma had inadequate statistical power. The effect of infant feeding on the severity of outcome and on the age at onset of the disease was virtually ignored in most of the studies [323].

Although this extensive meta-analysis may suggest some uncertainty about the prophylactic benefit of breast-feeding, other studies strongly support a positive effect of breast-feeding on the development of atopic disease and asthma. The first study [241] consisted of prospective, long-term evaluation from infancy until age 17 years; the prevalence of atopy was significantly higher in infants with short-duration (<1 month) or no breast-feeding, which increased to a demonstrable difference by age 17 years, than in infants with intermediate-duration (1 to 6 months) or prolonged (>6 months) breast-feeding. The differences in the prevalence of atopy persisted when the groups were divided according to positive or negative atopic heredity. The atopy manifestations in the different infant feeding groups did not remain constant with age. In particular, respiratory allergy, including asthma, increased greatly in prevalence up to age 17 years, with a prevalence of 64% in the group with short-duration or no breast-feeding [241].

In the second study, a prospective, longitudinal study of the prevalence and risk factors for acute and chronic respiratory illness in childhood, the investigators examined the relationship of infant feeding to recurrent wheezing at age 6 years and the association with lower respiratory tract illnesses with wheezing early in life [324]. Children who were never breast-fed had significantly higher rates of recurrent wheezing at 6 years of age. Increasing duration of breast-feeding beyond 1 month was not associated with significantly lower rates of recurrent wheezing. The effect of breast-feeding was apparent for children with and without wheezing lower respiratory tract illnesses in the first 6 months of life. In contrast with the findings of the first study, however, the effect of breast-feeding was significant only among nonatopic children [324].

The exact mechanisms by which breast-feeding seems to confer long-lasting protection against allergic sensitization are poorly understood. It is likely, however, that multiple synergistic mechanisms may be responsible for this effect, including (1) maturation of the recipient gastrointestinal and airway mucosa, promoted by growth factors present in human milk [[216], [217], [218]]; (2) inhibition of antigen absorption by milk sIgA [325]; (3) reduced incidence of mucosal infections and consequent sensitization to bystander antigens [326]; (4) changes in the microbial flora of the intestine of breast-fed infants [302]; and (5) direct immunomodulatory activity of human milk components on the recipient infant [214].

Numerous earlier and more recent studies have greatly contributed to the understanding of macromolecular transport across the immature gut and its consequences in terms of the generation of circulating antibody or immune complexes, the processes that are blocked predominantly by sIgA, the glycocalyx, and the intestinal enzymes. These mucosal immunologic events have been the basis for the concept of immune exclusion. Immune exclusion is not absolute, however, because uptake of some antigens across the gut may be enhanced rather than blocked by interaction with antibody at the mucosal surface. Beginning with the observations of IgA-deficient patients, it has become clear that the absence of the IgA barrier in the gut is associated with an increased incidence of circulating antibodies directed against many food antigens and an increased occurrence of atopic-allergic diseases [325]. Some studies have noted complement activation in serum after feeding of bovine milk to children with cow's milk allergy. The neonate is similar in some respects to the IgA-deficient patient [327], and increased transintestinal uptake of food antigen with consequent circulating antibody formation in the premature infant has been reported [328]. Other studies have suggested that early breast-feeding, even of short duration, is associated with a decreased serum antibody response to cow's milk proteins [226]. Prolonged breast-feeding not only may partially exclude foreign antigens through immune exclusion, but may also, because the mother's milk is the infant's sole food, prevent their ingestion [329]. Intact bovine milk proteins and other food antigens and antibodies have been observed, however, in samples of colostrum and milk [8].

Other Benefits

Epidemiologic evidence suggests that bacterial and viral respiratory infections are less frequent and less severe among breast-fed infants in various cultures and socioeconomic settings [330,331]. Antibodies and immunologic reactivity directed against herpes simplex virus, respiratory syncytial virus, and other infectious agents [92,102,164,331,332] and that protect against enterovirus infections [333] have been quantitated in colostrum and milk. Adoptive experiments in suckling ferrets have shown that protection of the young against respiratory syncytial virus can be transferred in colostrum containing specific antibody. The neonatal ferret gut is quite permeable, however, to macromolecules and permits passage of large quantities of virus-specific IgG. In the absence of either documented antibody or cellular transfer in the human neonate across the mucosa, any mechanisms of protection against respiratory syncytial virus and other respiratory pathogens remain obscure.

Data are lacking in humans regarding passive protection on other mucosal surfaces, such as the eye, ear, or genitourinary tract. Some epidemiologic evidence suggests that recurrence of otitis media with effusion is strongly associated with early bottle-feeding and that breast-feeding may confer protection against otitis media with effusion for the first 3 years of life [334]. Foster feeding–acquired antibody to herpes simplex virus has been found to result in significant protection against reinfection challenge in experimental animal studies [332].

Numerous other benefits have been associated with breast-feeding, including natural contraception during active nursing [335] and protection against sudden infant death syndrome [336], diabetes [337], obesity [338], and high cholesterol level and ischemic heart disease later in life [42]. Of particular more recent interest has been the association of breast-feeding with improved intellectual performance in older children. Several studies have shown enhanced cognitive outcome in breast-fed children, although controversy exists regarding the mechanisms by which such improved performance may occur [[339], [340], [341]]. Health benefits for the mother also may be associated with breast-feeding: A reduced incidence of breast cancer has been noted in women who have lactated [342].

Potential risks

Noninfectious Risks

Human milk is the optimal form of nutrition for healthy term infants in almost all situations. The failure to initiate lactation properly during early breast-feeding may present a risk of dehydration to the infant because insufficient fluids may be ingested. Inappropriate introduction of bottles and pacifiers also may interfere with proper induction of lactation. Later in lactation, introduction of bottles may induce premature weaning as the result of a reduction in the milk supply.

Some circumstances have been identified in which breast-feeding is contraindicated and others have been identified in which continued breast-feeding should be conducted with caution to protect the infant [3]. Infants with inherited metabolic diseases may require alternative forms of nutrition: neonates with classic galactosemia owing to deficiency of galactose-1-phosphate uridyltransferase should receive lactose-free milk (lactose is a glucose-galactose disaccharide); infants with phenylketonuria may receive some human milk to support their requirement for phenylalanine, but often may be better managed by use of specially prepared commercial milks. Mothers who have received radionuclides for diagnostic or therapeutic purposes should use alternative forms of nutrition for the days to weeks required for these compounds to be eliminated, as should mothers receiving certain chemotherapeutic and immunosuppressive agents and actively using drugs of abuse, including amphetamines, cocaine, heroin, and phencyclidine [343]. Low-level maternal exposure to environmental chemicals and tobacco smoking should be avoided as much as possible, but are not a contraindication to breast-feeding.

Antimicrobial agents taken by mothers only rarely represent a contraindication to breast-feeding. As first principles, antimicrobials that may be safely given to infants may be safely given to their lactating mothers, and blood concentrations that may be achieved through breast milk ingestion are lower than therapeutic doses used in infants [344]. Breast-feeding by mothers receiving chloramphenicol is contraindicated because its use may be associated with fatal complications in newborn infants. The effects of metronidazole are uncertain, but to minimize exposure to this drug, which is mutagenic in bacteria, mothers receiving single-dose therapy should discontinue breast-feeding for 12 to 24 hours [343]. Excretion of antibiotics in human milk is also discussed in Chapter 37. For up-to-date information regarding drugs and lactation, the reader should consult http://www.toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen?LACT.

Several instances of specific nutrient deficiencies in breast-fed infants have been described, specifically related to lack of vitamin K, vitamin D, vitamin B₁₂, folic acid, vitamin C, and carnitine. In each of these instances, several case reports have appeared warning against deficiencies that have resulted in clinical consequences to the neonate. Hemorrhagic disease reported in a few breast-fed infants was successfully treated with vitamin K [345]. These infants did not receive vitamin K at birth. Mothers who practice unusual dietary habits, such as strict vegetarianism, may have reduced levels of vitamin B₁₂ and folic acid in their milk, and deficiencies in breast-fed infants of such mothers have been reported [346,347]. Cases of rickets in breast-fed infants have been reported, particularly during winter among infants not exposed to the sun [34,348]. Deficiency of carnitine, a nutrient responsible for modulating fat absorption, also has been reported to result in clinical symptoms in breast-fed infants in mothers ingesting unusual diets [61,349]. These concerns can best be addressed in almost all cases by counseling mothers regarding nutritional practices and by the provision of supplemental vitamins and other micronutrients when appropriate; this is the case in the developed world and even more so in the developing world where the untoward consequences of not breast-feeding are particularly great [293].

Management of hyperbilirubinemia associated with breast-feeding has been an area of some controversy. Present recommendations are for continued breast-feeding with efforts to increase the volume of milk ingested, with the provision that with severe hyperbilirubinemia a brief interruption of breast-feeding might be appropriate [3].

Infectious Risks

Human milk may contain infectious agents that are secreted into the milk; enter milk during lactation; or are acquired when milk is improperly collected, stored, and later fed to the infant. Formal training and evaluation of breast-feeding practices by trained caregivers is the best way to reduce these risks; routine culture or heat treatment of a mother's milk even when it is stored and later used to feed her infant is not cost-effective [3].

Stored milk is now commonly used to feed infants when their mothers are unable to breast-feed directly because of work or travel constraints or when the infant is premature or otherwise unable to breast-feed effectively. Inadvertent feeding of stored milk from other than the birth mother has occurred in nurseries. If this occurs, the AAP recommends that this be handled in the same manner as if accidental exposure to blood or other body fluids has occurred [344] (see Chapter 35 for additional information).

In the United States, the Human Milk Banking Association of North America (http://www.hmbana.org/) collects human donor milk for the purpose of administration to infants whose mothers' milk is unavailable or inadequate. Members of this association follow guidelines formulated in consultation with the U.S. Food and Drug Administration (FDA) and Centers for Disease Control and Prevention. These guidelines help to ensure that donors are screened for transmissible infections and that the milk is carefully collected, processed, and stored. Using these practices, donor milk is collected and pooled and subjected to Holder pasteurization (62.5° C) for 30 minutes, which reliably kills bacteria and inactivates HIV and cytomegalovirus (CMV) and eliminates or substantially reduces the amounts of other viruses. The pooled milk is tested to ensure that it meets standards and is frozen for later distribution and use.

Bacterial Infections

Transmission of bacterial pathogens, including S. aureus, group B streptococci, mycobacteria, and other species, may occur through breast-feeding (Table 5–9 ). Mastitis and breast abscesses may be associated with substantial concentrations of bacteria in the mother's milk. Generally, feeding an infant from a breast affected by an abscess is not recommended [344]. Infant feeding on the affected breast may be resumed, however, 24 to 48 hours after drainage and the initiation of appropriate antibiotic therapy. Mastitis usually resolves with appropriate antimicrobial therapy and with continued lactation, even if feeding from the affected breast is temporarily interrupted. In both of these conditions, feeding from the unaffected breast need not be interrupted.

TABLE 5–9.

Infectious Agents Transmitted through Breast-Feeding

Organism	Transmission	Disease	Intervention
Cytomegalovirus	+	VLBW infants	Consider risk/benefit
Hepatitis B virus	+	+	HBIG/hepatitis B vaccine
Hepatitis C virus	HIV+ mothers only	?	See text
Herpes simplex virus	+	+	See text
HIV	+	+	U.S.: Do not breastfeed*
HTLV-1	+	±	U.S.: Do not breastfeed*
HTLV-2	+	±	U.S.: Do not breastfeed*
Rubella virus	+	0	None
West Nile virus	±	±	None
Group B streptococci	+	±	See text
Staphylococcus aureus	+	±	See text
Mycobacterium tuberculosis	+	−	See text

HBIG, hepatitis B immunoglobulin; HIV, human immunodeficiency virus; HTLV, human T-lymphotropic virus; VLBW, very low birth weight.

^*In many other parts of the world, the benefits of breast-feeding often outweigh the risks of alternative methods of infant feeding. See text for discussion of risk versus benefit in other parts of the world.

Mothers with active tuberculosis should refrain from breast-feeding for at least 2 weeks or longer after institution of appropriate treatment if they are considered contagious. This recommendation also applies to the uncommon situation where mastitis or breast abscess is caused by Mycobacterium tuberculosis [344].

Viral Infections

Viruses that have been detected in human milk include CMV, hepatitis B (HBV) and hepatitis C (HCV) viruses, herpes simplex virus, HIV-1, human T-lymphotropic virus 1 (HTLV-1) and HTLV-2, rubella virus, and West Nile virus (see Table 5–9) [344]. It is unknown whether varicella virus is secreted into human milk. Although some of these viruses do present a risk to the infant, for most, but not all, the benefits of breast-feeding to the infant are greater than the risk.

Cytomegalovirus Infection

CMV infection is a common perinatal infection. The virus is shed in the milk in about 25% of infected mothers. Although breast-feeding from infected mothers may result in seroconversion in 70% of breast-feeding neonates, the infection often is not associated with clinical symptoms of disease. Infants with very low birth weight (<1500 g) may exhibit evidence of clinical disease, however, with thrombocytopenia, neutropenia, or hepatosplenomegaly seen in 50% of very low birth weight infants infected through breast-feeding. The decision to breast-feed a premature infant by an infected mother should be based on weighing the potential benefits of human milk versus the risk of CMV transmission [344].

Hepatitis B Virus Infection

Hepatitis B surface antigen has been detected in milk of HBV-infected mothers. Nevertheless, breast-feeding does not increase the risk of HBV infection among these infants. Infants born to HBV-positive mothers should receive hepatitis B immunoglobulin and the recommended series of hepatitis B vaccine without any delay in the institution of breast-feeding [344].

Hepatitis C Virus Infection

The RNA of HCV and antibody to HCV have been detected in the milk from infected mothers. Transmission via breast-feeding has not been documented in anti-HCV–positive, anti-HIV–negative mothers, but is a theoretical possibility about which these mothers should be informed before deciding whether they will breast-feed. According to current guidelines, HCV infection does not contraindicate breast-feeding [344].

Herpes Simplex Virus Type 1

Herpes simplex virus transmission directly from maternal breast lesions to infants has been shown. Women with lesions on one breast may feed from the other unaffected breast, making sure that lesions on the other breast or on other parts of the body are covered and using careful hand hygiene [344].

Human Immunodeficiency Virus Type 1

Numerous studies have shown HIV in milk [[350], [351], [352], [353], [354]]. The findings include isolation of HIV from milk supernatants collected from symptom-free women and from cellular fractions of maternal milk, recovery of HIV virions in the histiocytes and cell-free extracts of milk by electron microscopy, and detection of viral DNA by polymerase chain reaction in greater than 70% of samples from HIV-seropositive lactating women.

Transmission of HIV through breast-feeding may account for one third to one half of all HIV infections globally, with risk of transmission being approximately 15% when breast-feeding continues beyond the first year of life [355,356]. The risk of postnatal HIV transmission seems to be constant throughout the first 18 months of life; risk is cumulative as duration of breast-feeding increases [357]. Risk of transmission via breast milk is greater when maternal HIV infection is acquired during lactation; when viral load is greater or maternal disease is more advanced; when infants are breast-fed and formula-fed; when the mother has bleeding or cracked nipples, mastitis, or a breast abscess; and when the infant has thrush or certain other coinfections (see Chapter 21 for more details).

Current recommendations from the AAP [344] and other authorities [358] state that in populations such as that of the United States, in which the risk of death from infectious diseases and malnutrition is low and in which safe and effective alternative sources of feeding are readily available, HIV-infected women should be counseled not to breast-feed their infants and not to donate milk. One report found that highly active antiretroviral therapy administered during pregnancy or postpartum suppresses HIV RNA, but not DNA in breast milk [359]. At present, the AAP recommends that infants of HIV-infected mothers in the United States receiving highly active antiretroviral therapy should not be breast-fed.

Despite the potential risk of HIV infection in infants of HIV-infected breast-feeding mothers, consideration of cessation of breast-feeding must be balanced against the other beneficial effects described in this chapter. In areas of the world where infectious diseases and malnutrition are important causes of death early in life, the beneficial effects of breast-feeding often outweigh the potential risk of HIV transmission through breast-feeding. Studies in such settings have shown that HIV-free survival at 7 months of age is similar in exposed infants that were breast-fed or formula-fed from birth [[360], [361], [362]]. Studies in Africa show that mothers who exclusively breast-feed in the first 6 months of life have reduced risk of HIV transmission compared with mothers who supplement breast-feeding with other foods and milk sources [360,361,363]. In areas of the world where the burden of infectious diseases and malnutrition is high and where alternatives to breast milk that provide adequate nutrition are unacceptable, affordable, feasible, and safe, exclusive breast-feeding is recommended by the WHO and other authorities for women whose HIV status is unknown and for women known to be HIV-infected [293,344,358,364]. The WHO policy also stresses the need for continued support for breast-feeding by mothers who are HIV-negative, improved access to HIV counseling and testing, and government efforts to ensure uninterrupted access to nutritionally adequate human milk substitutes [344].

Human T-Lymphotropic Viruses 1 and 2

HTLV-1 is endemic in Japan, the Caribbean, and parts of South America. This infection can be transmitted from mother to infant, and transmission occurs primarily through breast-feeding. HTLV-2 infection has been identified in some Native Americans and Native Alaskans and in some injection drug abusers in the United States and Europe. Mother-to-infant transmission of HTLV-2 has been shown, although the frequency with which this occurs and the route of transmission are uncertain. Women in the United States who are known to be seropositive for HTLV-1 or HTLV-2 should not breast-feed. Routine screening for HTLV-1 and HTLV-2 is not recommended [344].

Rubella

Rubella virus has been recovered from milk after natural and vaccine-associated infection. It has not been associated with significant disease in infants, although transient seroconversion has been frequently shown. No contraindication to breast-feeding exists in women recently immunized with currently licensed rubella vaccines.

West Nile Virus Infection

The RNA of West Nile virus has been detected in human milk, and seroconversion in breast-feeding infants also has been observed. Although West Nile virus can be transmitted in milk, the extent of transmission in humans remains to be determined. Most infants and children infected with the virus to date have been asymptomatic or have had minimal disease [344]. Because the risk is uncertain, the AAP recommends that women in endemic areas may continue to breast-feed.

Current trends in breast-feeding

International and national organizations have endorsed breast-feeding as the optimal means of feeding for healthy term infants [344]. The percentage of mothers initiating breast-feeding in developing countries generally is 80% or greater and often 90% or greater. The health and economic consequences for bottle-fed infants in these countries are severe, however. In the United States, at one point in the early 1970s, the rate of breast-feeding initiation was 25%. The rate of initiation since that time has increased but fluctuated over time. In a survey (IFPS II) conducted by the FDA in 2005-2007, 83% of respondents initiated breast-feeding [365].

In contrast to this favorable trend, the recommendation for exclusive breast-feeding for the first 6 months of life in the United States and internationally is uncommonly followed. In the United States, supplemental feeding was often initiated in the hospital in the immediate postpartum period, and by 3 months of age more than 61% of infants had received formula [365]. Similarly, although breast-feeding initiation approaches 100% in developing nations in Africa, Asia, and the Caribbean, the rates of exclusive breast-feeding for the first 6 months of life are approximately 50% and approximately 30% for infants less than 2 months and 2 to 5 months of age [1].

Within the United States, various demographic patterns seem to be associated with breast-feeding behavior. Older mothers, mothers with a college education, and mothers with higher incomes all are more likely to breast-feed. African American and Hispanic mothers, mothers of lower socioeconomic status who are participants in the Women, Infants, and Children (WIC) program of the U.S. Department of Health and Human Services, and mothers who live in the southern regions of the United States are much less likely to breast-feed. The low rate of breast-feeding for mothers enrolled in WIC is of particular concern because that agency has a specific policy to encourage breast-feeding. Many states now depend on formula manufacturer rebates to fund part of their WIC programs, however, creating a conflict of interest. The disturbing part of the demographic pattern of breast-feeding in the United States is that the infants of lower socioeconomic status mothers, who would accrue the greatest health and economic benefits from breast-feeding, are those least likely to be breast-fed.

Although demographic studies indicate who is breast-feeding, they do not explain the behavioral differences among groups of mothers. One of the more complete models designed to explain breast-feeding behavior includes components that address maternal attitudes and family, societal, cultural, and environmental variables [366]. Individual studies have shown that the maternal decision-making process is closely related to the social support and influence that come from the family members surrounding the mother [367]. The husband in particular seems to have a strong positive influence, whereas the mother's mother may have a negative influence on the breast-feeding decision. Social support seems to be different among ethnic groups, as are maternal attitudes; such differences may provide one explanation for differences in breast-feeding behavior among ethnic groups [368,369]. Higher rates of exclusive breast-feeding have been observed for infants born in U.S. hospitals who follow the practices recommended by the WHO/United Nations Children's Fund Baby-friendly Hospital Initiative, which emphasizes 10 steps that promote successful breast-feeding [370].

Summary and conclusions

Human milk contains a wide variety of soluble and cellular components with a diverse spectrum of biologic functions. The major milk components identified to date exhibit antimicrobial, anti-inflammatory, proinflammatory, and immunoregulatory functions; cytotoxicity for tumor cells; ability to repair tissue damage; receptor analogue functions; and other metabolic effects. The biologic activities of different milk components are summarized in Table 5–10 .

TABLE 5–10.

Possible Role of Soluble and Cellular Factors Identified in Human Milk

Factor	Antimicrobial	Anti-inflammatory	Proinflammatory	Immunoregulatory	Other
Immunoglobulin (sIgA)	+++	++	—	++	++
Other immunoglobulins	+++	+	++	+	—
T-lymphocyte products	+++	++	++	—	—
PMNs, macrophages	++	—	+	++	—
Lactoferrin	+++	+++	—	—	—
α-Lactalbumin	—	++	—	—	—
Carbohydrates
Oligosaccharides	++	++	—	—	++
Glycoconjugates	++	++	—	—	++
Glycolipids	—	—	—	—	—
Lipid and fat globules	++	—	—	—	—
Nucleotides	+	—	—	++	++
Defensins	+	—	—	+	—
Lysozymes	±	—	—	—	—
Cytokines, chemokines
TGF-β	—	++	++	++	—
IL-10	—	++	++	++	—
IL-1β	—	++	++	++	—
TNF-α	—	—	—	++	—
IL-6	—	—	—	++	—
IL-7	—	—	—	(prothymus)	—
Others	—	—	—	++	—
Prostaglandins	—	++	—	—	—
Leptin*	—	—	—	++	++
Antiproteases	—	++	—	—	—
Other growth factors	—	++	—	++	—
sTLR-2, sCD14	—	+++	—	++	—

+ to +++ = minimal to moderate effect; — = no known effect; ±= equivocal.

IL, interleukin; PMNs, polymorphonuclear neutrophils; sIgA, secretory IgA; TGF, transforming growth factor; TLR, toll-like receptor; TNF, tumor necrosis factor.

^*IL-1α, TNF-β, and IL-6 are associated with increased levels of leptin.

Polymorphonuclear leukocytes, macrophages, lymphocytes, and epithelial cells are observed in human milk, but their functions in milk are unknown. It is possible that their primary task is the antimicrobial defense of the mammary gland itself. These cells may help to promote tolerance to their mothers' HLA allotype, which may have implications regarding immune responsiveness and allograft rejection [371].

The bulk of the antimicrobial effects of human milk are associated with milk immunoglobulin, especially the sIgA isotype, which makes up to 80% of all immunoglobulins in the human body. Milk antibodies seem to provide protection against many intestinal pathogens, such as Campylobacter, Shigella, E. coli, V. cholerae, Giardia, and rotavirus, and against respiratory pathogens such as respiratory syncytial virus. Milk antibodies also effectively neutralize toxins and various human viruses. The role of small amounts of IgG and IgM in milk is uncertain. Milk IgA antibodies generally induce antimicrobial protection in the absence of any inflammation. Lactoferrin, lysozyme, α-lactalbumin, and other milk proteins and carbohydrates and lipids also contribute to the antimicrobial and immunomodulatory properties of human milk. Milk also contains numerous cytokines, chemokines, growth factors, soluble TLRs, and CD14, which modulate inflammatory and immunologic responses in the gut.

The passive transfer of the diversity of maternal biologic experiences to the neonate through the process of breast-feeding represents an essential component of the survival mechanism in the mammalian neonate. For millions of years, maternal products of lactation delivered through the process of breast-feeding have been the sole source of nutrition and immunity during the neonatal period and early infancy for all mammals, including the human infant. During the past 150 to 300 years, human societies have undergone remarkable changes, which have had a major impact on the basic mechanisms of maternal-neonatal interaction, breast-feeding, and the environment. Such changes include introduction of sanitation and nonhuman milk and formula feeds for neonatal nutrition, use of antimicrobial agents, introduction of processed foods, and exposure to newer environmental macromolecules and dietary antigens. The introduction of such human-made changes in the neonatal environment has had a profound impact on human homeostatic mechanisms, while allowing new insights into the role of breast-feeding in the developing human neonate.

Comparative analysis of natural (traditional) forms of breast-feeding and artificial feeding modalities has shown that natural breast-feeding is associated with significant reduction in infant mortality and morbidity; protection against acute infectious diseases; and possible protection against allergic disorders and autoimmune disease, acute and chronic inflammatory disorders, obesity, diabetes mellitus and other metabolic disorders, allograft rejection, and development of many malignant conditions in childhood or later in life. This information has been reviewed by Hanson in an elegant monograph [372] and by others [86]. Despite the overwhelmingly protective role attributed to natural breast-feeding and the evolutionary advantages related to the development of lactation, several infectious agents have acquired, during the course of evolution, the ability to evade immunologic factors in milk and to use milk as the vehicle for maternal-to-infant transmission. The potential for the acquisition of infections, including HIV, HTLV, CMV, and possibly other pathogens, highlights potential hazards of breast-feeding in some clinical situations. It is reasonable to conclude that the development of lactation, the hallmark of mammalian evolution, is designed to enhance the survival of the neonate of the species and that breast-feeding may have a remarkable spectrum of immediate and long-term protective functions.