Our understanding of the significance of microbes for human health has expanded over the last decades from viewing bacteria solely as potential pathogens to a dawning appreciation of the complexity of host-microbe interactions in health and disease. The early triumphs in clinical microbiology revealed the causative role of specific microbes in classical infectious diseases such as cholera, dysentery or diphtheria. Subsequent research has elucidated multifactorial causal networks involving potential pathogens, as well as a myriad of environmental and host susceptibility factors in the development of conditions also classified as infectious diseases such as community-acquired pneumonia or tuberculosis. More recently, we have begun to grasp the significance of environmental and indigenous microbes for host physiology and the development of non-infectious disease. Microbes have a profound effect on maintaining homeostasis in key physiological processes including immune, neural and metabolic functions. Perhaps even more importantly, intimate contact with the microbial world is necessary for normal development in early life.
Microbes and host immune maturation
Our current knowledge regarding the developmental significance of early microbial contact stems from experimental studies using animals raised in controlled microbial environments the most revealing of which is total lack of microbial contact (‘germ-free’ animals). Hallmark animal studies have demonstrated that animals devoid of normal host-microbe interaction in early life exhibit defective maturation of the alimentary and immune systems and appear to be particularly prone to allergic-type adaptive immune responses.Reference Sudo, Sawamura and Tanaka 1 , Reference Mazmanian, Liu, Tzianabos and Kasper 2 It is of particular note that introduction of a single bifidobacterium strain in the neonatal period is reportedly sufficient for inducing at least some degree of immune maturation.Reference Sudo, Sawamura and Tanaka 1 More recently, even intestinal contact with a specific microbial surface polysaccharide alone has been reported to restore a more physiological immune phenotype in germ-free mice.Reference Mazmanian, Liu, Tzianabos and Kasper 2 On the other hand, certain disease states may be induced in experimental animals merely by colonizing them with a microbiome from an affected individual. In a beautiful series of experiments by Ridaura et al.Reference Ridaura, Faith and Rey 3 germ-free mice colonized with fecal microbes obtained from an obese human exhibited excessive weight and fat mass gain whereas mice whose gut microbiome originated from a lean individual remained lean. These animal studies clearly demonstrate that intestinal microbial colonization and contact with indigenous microbes in the neonatal period is crucial for normal immune and metabolic development.
An association between the composition of the intestinal microbiome and chronic non-infectious disease appears evident based on accumulating data from human studies. Aberrant gut colonization patterns have been reported in a wide variety of human diseases including atopic diseases,Reference Kalliomäki, Kirjavainen and Eerola 4 – Reference Abrahamsson, Jakobsson and Andersson 7 inflammatory bowel disease (IBD),Reference Schwiertz, Jacobi and Frick 8 , Reference Gevers, Kugathasan and Denson 9 necrotizing enterocolitis (NEC) in preterm infants,Reference Mshvildadze, Neu and Shuster 10 , Reference Mai, Young and Ukhanova 11 diabetes mellitus,Reference Dunne, Triplett and Gevers 12 , Reference Kostic, Gevers and Siljander 13 obesityReference Turnbaugh, Hamady and Yatsunenko 14 , Reference Kalliomäki, Collado and Salminen 15 and neurological conditions ranging from autistic spectrum disordersReference De Angelis, Piccolo and Vannini 16 to Parkinson’s disease.Reference Scheperjans, Aho and Pereira 17 The pathogenesis of most if not all of the chronic disorders in which indigenous microbes are thought to play a role involves immune-mediated or inflammatory processes and it is therefore plausible to hypothesize that pathological host-microbe interaction may have a causal role in their development.
Most reports showing a link between disease and gut microbiome composition are cross-sectional by design and therefore inherently incapable of establishing causality. Data from a limited number of longitudinal studies indicate, however, that the development of eczemaReference Abrahamsson, Jakobsson and Andersson 6 and obesityReference Kalliomäki, Collado and Salminen 15 in childhood may be preceded by aberrant gut colonization patterns in early infancy. Further corroboration for the hypothesized causal role for deviant early microbial contact in disease development has been provided by experimental animal models showing that the disease phenotype may be transferred with the gut microbiome at least in the case of obesity.Reference Ridaura, Faith and Rey 3 , Reference Koren, Goodrich and Cullender 18 , Reference Cox, Yamanishi and Sohn 19 Causality may also be indirectly inferred from epidemiological studies assessing the impact of factors known to perturb early gut colonization on later health.
Mode of delivery
Newborn infants acquire their indigenous microbiome primarily from the mother. It is well established that a substantial inoculum is provided by maternal intestinal and vaginal microbes during delivery and from an evolutionary point of view it is likely that this serves a specific purpose by guiding neonatal colonization. It is also well known that infants born by cesarean section delivery exhibit increased neonatal morbidity including respiratory problems and hypoglycemia compared with vaginally delivered infants. These issues arising from delayed adaptation to extrauterine life have been attributed to the lack of stress signals provided by the mother during vaginal delivery. Interestingly, birth by cesarean section is also associated with increased risk of chronic immune-mediated or inflammatory conditions such as asthma,Reference Thavagnanam, Fleming, Bromley, Shields and Cardwell 20 type I diabetesReference Cardwell, Stene and Joner 21 and obesity.Reference Barros, Matijasevich and Hallal 22 , Reference Darmasseelane, Hyde, Santhakumaran, Gale and Modi 23 According to a recent Danish epidemiological study comprizing of 1.9 million subjects, the risk of asthma or IBD is ~1.2-fold in infants born by cesarean section and smaller but statistically significant increases in occurrence of juvenile arthritis, immune deficiencies and leukemia were also observed.Reference Sevelsted, Stokholm, Bønnelykke and Bisgaard 24 It is likely that the increased morbidity in infants born by cesarean section is at least partially attributable to altered early gut colonization. In line with this notion, the highest risk of obesity has been observed in individuals born by cesarean section and exposed to antibiotics, which also perturb microbial colonization.Reference Mueller, Whyatt and Hoepner 25 The early gut microbiome during the 1st days of life in infants born vaginally shares features with the vaginal microbial community whereas in infants born by cesarean section the early gut microbiome resembles that of the maternal skin.Reference Dominguez-Bello, Costello and Contreras 26 , Reference Matsumiya, Kato, Watanabe and Kato 27 The vaginal lactobacilli in the infant gut are rapidly replaced by microbes from other sourcesReference Matsumiya, Kato, Watanabe and Kato 27 and most gut microbes in vaginally delivered infants originate in the maternal gut. The significance of delivery mode to intestinal colonization is demonstrated by the study by Jakobsson et al. Reference Jakobsson, Abrahamsson and Jenmalm 28 according to which the intestinal microbial diversity is lower throughout the first 2 years of life in individuals delivered by cesarean section. The differences in the gut microbiome are still detectable between cesarean and vaginally delivered infants at the age of 7 years.Reference Salminen, Gibson, McCartney and Isolauri 29 These differences are also reflected in infant immune development up to the age of 12 months.Reference Huurre, Kalliomäki and Rautava 30
Early antibiotic exposure
Exposure to antibiotics is known to exert a devastating effect on the gut microbiome. The intestinal microbial balance has been reported to recover after antibiotic exposure but there are also data to suggest that repeated antibiotic perturbation may result in incomplete microbial recovery.Reference Dethlefsen, Huse, Sogin and Relman 31 – Reference Fouhy, Guinane and Hussey 33 In the neonatal period, empirical antibiotic treatment has been associated with increased occurrence of NECReference Cotten, Taylor and Stoll 34 and late-onset sepsisReference Shah, Nathan, Doherty and Patole 35 and short-term gut microbiota perturbations including lower diversityReference Greenwood, Morrow and Lagomarcino 36 and particularly reduced colonization with bifidobacteriaReference Fouhy, Guinane and Hussey 33 , Reference Tanaka, Kobayashi and Songjinda 37 and bacteroides.Reference Arboleya, Sánchez and Milani 38 Alarmingly, perinatal antibiotic exposure has been reported to result in long-term metabolic consequences and obesity in an experimental animal model even when the alterations in gut microbiota composition caused by antibiotics are reversible.Reference Cox, Yamanishi and Sohn 19 Accumulating epidemiological data indicate that antibiotic perturbation of microbial contact during the critical developmental window in early life may have long-term health implications also in humans.
Early antibiotic use is associated with increased risk for a variety of chronic non-infectious diseases in later life. Antibiotic exposure during the 1st year of life is significantly more common in children who later develop IBDReference Shaw, Blanchard and Bernstein 39 , Reference Virta, Auvinen, Helenius, Huovinen and Kolho 40 and exposure to antibiotics during the first 6 months of life has been reported to be associated with an increase in body mass up to the age of 3 years.Reference Trasande, Blustein and Liu 41 The risk of developing asthma at school age was independently increased by antibiotic exposure in the 1st week of life in a birth cohort study of >5000 children in Sweden.Reference Goksör, Alm and Pettersson 42 This association is plausible in the light of data indicating that neonatal antibiotic exposure results in decreased early gut microbiome diversity,Reference Greenwood, Morrow and Lagomarcino 36 which in turn is reportedly associated with the development of asthma in later life.Reference Abrahamsson, Jakobsson and Andersson 7 It is of note that also maternal antibiotic use during pregnancy appears to be associated with increased risk of asthma in childhood.Reference Stensballe, Simonsen, Jensen, Bønnelykke and Bisgaard 43 These epidemiological associations must be interpreted with caution since they may be explained by an underlying immunological factor predisposing to both infections and immune-mediated or inflammatory disease. The infection for which antibiotics had been administered may also play a role in disease pathogenesis. Nonetheless, careful examination of the available data reveals important time and dose-dependent effects,Reference Virta, Auvinen, Helenius, Huovinen and Kolho 40 which may be interpreted to corroborate the hypothesis of long-term consequences as a result of microbial perturbation during a critical developmental period.
Microbial transfer from the mother
Microbes constituting the indigenous human gut microbiome are not frequently found in the environment, which implies that infants acquire their gut microbes from other humans. As alluded to above, it is thought that the mother is the most important source of colonizing bacteria and, in line with this notion, infants have been shown to receive a significant bacterial inoculum during vaginal delivery.Reference Dominguez-Bello, Costello and Contreras 26 , Reference Makino, Kushiro and Ishikawa 44 Less is known about other routes of microbial transfer.
Microbial contact during fetal life
It was long assumed that the fetus develops in a sterile environment without microbial contact except in cases of chorionic and fetal infection, which often lead to premature birth or fetal demise. Epidemiological and experimental studies have shown, however, that the maternal microbial environment has a significant impact on the risk of immune-mediated disease later in childhood. Douwes et al. Reference Douwes, Cheng and Travier 45 have reported that children whose mothers have lived in a farming environment and hence presumably exposed to microbes more than their more urban-dwelling counterparts, display a reduced risk for asthma. The investigators later isolated a bacterium, Acinetobacter lwoffii, from cow sheds and demonstrated that exposing pregnant mice to it protected the offspring from asthma in an experimental animal model.Reference Conrad, Ferstl and Teich 46 The precise mechanism of this indirect in utero microbial contact remains unknown. Interestingly, however, it is becoming evident that the healthy fetus may also be directly exposed to microbes.
The presence of microbes or their components in the uterus during pregnancyReference Stout, Conlon and Landeau 47 and also during the non-pregnant stateReference Andrews, Goldenberg and Hauth 48 appears to be a physiological phenomenon. Microbes or microbial DNA have been detected in the placentaReference Satokari, Grönroos, Laitinen, Isolauri and Salminen 49 , Reference Aagaard, Ma and Antony 50 and the umbilical cordReference Jiménez, Fernández and Marin 51 in healthy pregnancy as well as in infant meconium, which formed during fetal life.Reference Mshvildadze, Neu and Shuster 10 , Reference Dominguez-Bello, Costello and Contreras 26 , Reference Gosalbes, Llop and Vallés 52 It has recently been suggested that the human placenta harbors a distinct microbiome predominated by Proteobacteria.Reference Aagaard, Ma and Antony 50 The origin and physiological role of the hypothesized in utero microbiome remains unknown. However, there are data to suggest that the placenta microbiome is modulated by excessive maternal weight gain during pregnancy.Reference Antony, Ma and Mitchell 53 Microbes in the placenta have also been reported to be associated with placenta innate immune function.Reference Rautava, Collado and Salminen 54 Cross analyses of separate data sets have suggested that the placenta microbiome shares features with the microbial community detected in the oral cavity,Reference Aagaard, Ma and Antony 50 which is particularly interesting given the suggested association between poor dental health and complications of pregnancy including prematurity.Reference Zi, Longo, Bueno-Silva and Mayer 55 On the other hand, experiments conducted in a murine model have shown that labeled bacteria introduced to the intestine of pregnant mice may be detected in the placenta and in meconium harvested after sterile cesarean section delivery,Reference Jiménez, Fernández and Marin 51 , Reference Jiménez, Marín and Martín 56 which suggests a link between the microbes in the maternal gut, the placenta and the fetal gut.
It should be born in mind that the discovery of the placenta microbiome was based on data obtained by sequencing of the 16S rRNA gene and therefore only demonstrates the presence of microbial DNA without direct evidence of viable bacteria. Nonetheless, the presence of microbial DNA in the intrauterine compartment suggests that the fetus may be in direct contact with microbial components during gestation. Meconium, which is formed during intrauterine life, also appears to harbor a unique microbial community.Reference Mshvildadze, Neu and Shuster 10 , Reference Dominguez-Bello, Costello and Contreras 26 , Reference Gosalbes, Llop and Vallés 52 The meconium microbiome is rich in bacilli and other Firmicutes and drastically differs in composition from the gut microbiome later in the neonatal period.Reference Moles, Gómez and Heilig 57 There is indirect evidence based on separate studies suggesting that the microbes in meconium originate from swallowed amniotic fluid.Reference Ardissone, de la Cruz and Davis-Richardson 58 Interestingly, the composition of the meconium microbiome is not only dependent on maternal health status but also seems to be associated with subsequent child health.Reference Gosalbes, Llop and Vallés 52
Breast milk as modulator of the infant gut microbiome and source of microbes
After birth, the most important determinant of infant gut colonization is breastfeeding. Breastfed infants harbor an intestinal microbiome dominated by Aktinobacteria and particularly bifidobacteria while the gut microbiome in formula-fed infants is more diverse and resembles that of older children.Reference Harmsen, Wildeboer-Veloo and Raangs 59 , Reference Roger, Costabile, Holland, Hoyles and McCartney 60 The predominance of bifidobacteria in infant feces is thought to result from factors in breast milk, which favor the growth of these bacteria. A considerable amount of energy in breast milk is provided in the form of indigestible oligosaccharides, which do not provide energy to the infant but may be utilized by intestinal microbes. The functions of human milk oligosaccharides include acting as decoy receptors for potential pathogens and thus protection from infectious disease, and promoting the growth of bifidobacteria (reviewed by BodeReference Bode 61 ). Bifidobacteria in infant feces are often considered to be ‘good’ or ‘health-promoting’ and, indeed, colonization with bifidobacteria has been associated with reduced risk for both allergic diseaseReference Björkstén, Sepp, Julge, Voor and Mikelsaar 5 and excessive weight gain.Reference Kalliomäki, Collado and Salminen 15 , Reference Dogra, Sakwinska and Soh 62 It is not clear to what extent the documented beneficial long-term health effects of breast milk including reducing the risk of obesity and immune-mediated disease (reviewed by Rautava and WalkerReference Rautava and Walker 63 ) are mediated via modulation of the gut microbiome. It is important to recognize that the composition of breast milk varies between mothers, which may at least partially explain the discrepant data concerning the relationship between breastfeeding and the risk of immune-mediated disease such as allergy. Furthermore, the mechanisms underlying the association between abundance of intestinal bifidobacteria and favorable health outcomes still remain unknown.
In addition to specifically promoting the growth of intestinal microbes thought to be beneficial to health, recent evidence indicates that breast milk may be a source of bacteria for the infant gut. Even non-lactating mammary gland tissue has been reported to contain both live bacteria and bacterial DNA.Reference Urbaniak, Cummins and Brackstone 64 The microbial community in breast milk characterized by 16S rRNA gene pyrosequencing appears to be unique with a composition clearly distinct from that observed at other mucosal sites including the skin, gut, mouth or vagina.Reference Cabrera-Rubio, Collado and Laitinen 65 The origin of the bacteria in breast milk is currently not known but there are circumstantial evidence suggesting a link between the maternal gut and mammary gland. Increased intestinal bacterial translocation has been reported during pregnancy and lactation in experimental animals.Reference Perez, Doré and Leclerc 66 Maternal intestinal microbes have been detected in immune cells circulating in peripheral blood and in breast milk in both mice and humans suggesting active and selective transport.Reference Perez, Doré and Leclerc 66 In a clinical trial, the probiotic bacterium Lactobacillus reuteri has been detected in breast milk after oral intake in lactating mothers.Reference Abrahamsson, Sinkiewicz, Jakobsson, Fredrikson and Björkstén 67
The teleological reason for microbes in breast milk and their role in infant gut colonization remain largely speculative. The same bifidobacterium species have been consistently found in maternal feces, breast milk and the infant gut but the study provided no mechanistic insight as to whether the bifidobacteria in infant gut originated from breast milk or were transferred from the mother via other routes.Reference Grönlund, Gueimonde and Laitinen 68 Martin et al. Reference Martín, Langa and Reviriego 69 studied breast milk, mammary areola and breast skin samples from eight breastfeeding mothers and oral swab and fecal samples from their infants. Identical Lactobacillus specimens were discovered from the breast milk, infant oral mucosa and infant feces in all mother–infant pairs. These data provide corroboration for the notion of bacterial transfer via breast milk but its significance for infant gut colonization is still unknown and it is likely that maternal and environmental transfer of bacteria takes primarily place by other routes.
The composition of the breast milk microbiome is affected by maternal factors. A significant change in the composition of the breast milk microbiome has been observed over the span of lactation.Reference Cabrera-Rubio, Collado and Laitinen 65 The most common genera in colostrum samples detected by 16S sequencing included Leuconostoc, Weissella, Staphylococcus, Streptococcus and Lactococcus according to analyses of breast milk samples from 18 mothers from Finland.Reference Cabrera-Rubio, Collado and Laitinen 65 A significant increase in Veillonella, Prevotella and Leptotrichia was seen 1 and 6 months after delivery. In a similar study conducted using quantitative polymerase chain reaction (qPCR) to identify bacteria, Lactobacillus, Enterococcus and Streptococcus were the most prominent bacteria in breast milk and the abundance of Bifidobacterium and Enterococcus species increased throughout the lactation period.Reference Khodayar-Pardo, Mira-Pascual, Collado and Martínez-Costa 70 In breast milk samples collected from African mothers and analyzed by qPCR, streptococci, staphylococci and lactobacilli were the most prevalent bacteria but the breast milk microbiome composition varied as a function of maternal HIV status.Reference González, Maldonado and Martín 71 The duration of gestation also seems to have an impact on the breast milk microbiome. Particularly the abundance of bifidobacterium species has been observed to be higher in breast milk from mothers who delivered at full term as compared with mothers of preterm infants.Reference Khodayar-Pardo, Mira-Pascual, Collado and Martínez-Costa 70 The significance of these differences remains to be determined.
Factors known to influence child health outcomes have also been reported to have an impact on the breast milk microbiome. As discussed above, delivery by cesarean section results in aberrant infant gut colonization and immune development and is associated with increased risk of obesity and chronic immune-mediated disease later in childhood. Interestingly, breast milk from mothers who had undergone cesarean section delivery exhibited a higher microbial diversity but reduced frequency of bifidobacteria according to a report based on milk samples obtained from 32 healthy mothers and analyzed by qPCR.Reference Khodayar-Pardo, Mira-Pascual, Collado and Martínez-Costa 70 Whether the reduced bifidobacterium content in breast milk contributes to the decreased colonization by bifidobacteria observed in infants born by cesarean section is unknown. The mechanisms by which mode of delivery affects the breast milk microbiome, also remain to be determined but the changes are detectable at least until 6 months after delivery.Reference Cabrera-Rubio, Collado and Laitinen 65 The contribution of intrapartum antibiotic prophylaxis often administered to mothers before cesarean section delivery to altered breast milk microbiome is not known, nor is the effect of antibiotics administered during vaginal delivery. Interestingly, however, the observed qualitative and quantitative differences in breast milk microbial communities appear to be more pronounced in women undergoing elective cesarean section while the impact of surgical delivery in women in labor is more moderate.Reference Cabrera-Rubio, Collado and Laitinen 65 This has been interpreted to suggest that the stress and/or hormonal signals related to labor have an impact on bacterial transfer to the mammary gland.Reference Cabrera-Rubio, Collado and Laitinen 65
The associations between breastfeeding and childhood obesityReference Yan, Liu, Zhu, Huang and Wang 72 on the one hand and early gut microbiome composition and weight gainReference Kalliomäki, Collado and Salminen 15 on the other set an interesting background to the observation according to which differences in the microbial composition of breast milk are detectable between overweight or normal weight mothers.Reference Cabrera-Rubio, Collado and Laitinen 65 Obese mothers exhibited higher total bacterial counts and particularly higher counts of staphylococci and lactobacilli in their breast milk while the number of bifidobacteria was higher in the breast milk of normal weight mothers. Changes related to excessive weight gain during pregnancy were similar to those seen in obese mothers. Whether the breast milk microbiome provides an explanatory link between breastfeeding duration, the gut microbiome and obesity is currently not understood. Interestingly, maternal overweight is also associated with decreased concentration of immunomodulatory factors in breast milk.Reference Collado, Laitinen, Salminen and Isolauri 73
The immunomodulatory potential of breast milk
Breast milk contains a vast array of immune factors such as immunoglobulins, oligosaccharides, innate immune receptors and antimicrobial proteins including lysozyme, lactoferrin and defensins, which provide protection against infectious disease during the vulnerable 1st months of life (reviewed by Rautava and WalkerReference Rautava and Walker 63 ). While the well-established protective effects of breastfeeding against early infections are likely to be mostly attributable to this passive immunoprotection (and probably also reduced exposure to potentially contaminated water), the role of breast milk components with active immunomodulatory potential remains less well elucidated. It is likely that the reduction in the occurrence of NEC associated with human milk feeding as compared with cow’s milk-based formula documented in preterm infantsReference Quigley and McGuire 74 is mediated by both breast milk antimicrobial proteinsReference Trend, Strunk and Hibbert 75 and the immunomodulatory and maturational properties of breast milk,Reference Rautava, Nanthakumar and Dubert-Ferrandon 76 , Reference Rautava, Lu, Nanthakumar, Dubert-Ferrandon and Walker 77 as well as the beneficial effects of breast milk on gut microbiota discussed above. Recently, oropharyngeal administration of minute amounts of colostrum has been shown to induce the production of secretory IgA antibodies and lactoferrin and reduce the incidence of clinical sepsis in extremely premature infants.Reference Lee, Kim and Jung 78 The intimate interplay between intestinal and breast milk microbes and the immunomodulatory factors in breast milk may also at least partially explain the long-term protective effects of breastfeeding against immune-mediated and inflammatory chronic disease.
Breast milk as a vehicle for immune tolerance induction
Transforming growth factor (TGF)-β, one of the most thoroughly investigated immunomodulatory factors in breast milk, may be considered a paradigmatic mediator of host–microbe interaction in breast milk and the infant gut. The functions of TGF-β range from inducing the production of IgA class antibodies and direct anti-inflammatory and maturational effects on immune and intestinal epithelial cells to induction of regulatory T cells central to establishing and maintaining immune tolerance (reviewed by Rautava and WalkerReference Rautava and Walker 63 ). Of the three isoforms of TGF-β, TGF-β2 is the most abundant in breast milk while TGF-β1 is produced by immune cells in the intestine. Data from experimental animal models suggest that breast milk may be an important source of TGF-β to the developing intestine in early life when endogenous production of TGF-β by the developing immune system is low.Reference Penttila, van Spriel and Zhang 79 Breastfed infants have been observed to exhibit higher serum TGF-β2 concentrations throughout the 1st year of life when compared with formula-fed infants suggesting that factors in breast milk may also act as triggers for TGF-β production in the infant.Reference Kainonen, Rautava and Isolauri 80
TGF-β is a key mediator of active immune tolerance (Box 1).Reference Weiner 81 Studies conducted in a murine asthma model have elegantly demonstrated that TGF-β is necessary for induction of immune tolerance toward environmental antigens via breast milk.Reference Verhasselt, Milcent and Cazareth 82 Recently published data indicate that TGF-β2 regulates intestinal epithelial cell inflammatory responses against bacterial antigens.Reference Nguyen, Jiang and Jacobsen 83 In clinical studies, breast milk TGF-β concentrations have been found to be inversely associated with the risk of eczema and allergy-related outcomes in the child.Reference Kalliomäki, Ouwehand, Arvilommi, Kero and Isolauri 84 , Reference Oddy, Halonen and Martinez 85 The breast milk concentration of TGF-β2 is reportedly lower in atopic than healthy mothers,Reference Laiho, Lampi and Hämäläinen 86 and the contributions of hereditary risk and breast milk factors to the risk of disease in the offspring are not clear. Interestingly, however, there are data suggesting that breast milk TGF-β may also play a role in establishing tolerance toward indigenous microbes.
NEC is a devastating intestinal inflammatory condition, which afflicts primarily preterm neonates. The pathogenesis of NEC is complex and still largely unknown, but intestinal and immunologic immaturity and increased intestinal bacterial translocation are thought to result in an excessive inflammatory reaction against intestinal microbes leading to a vicious cycle of yet more impaired gut barrier and propagated inflammation (reviewed by Neu and WalkerReference Neu and Walker 87 ). It has recently been reported that preterm infants who develop NEC have lower serum TGF-β concentrations from birth.Reference Maheshwari, Schelonka and Dimmitt 88 The protective effect of human breast milk against the development of NEC has been convincingly demonstrated in clinical trialsReference Quigley and McGuire 74 but the mechanisms underlying the clinical efficacy are not fully established. Experimental studies using immature human gut models indicate that exposure to TGF-β2 in a concentration corresponding to that in breast milk exerts a direct anti-inflammatory effect in the immature intestineReference Rautava, Nanthakumar and Dubert-Ferrandon 76 and induces maturation in intestinal epithelial cells.Reference Rautava, Lu, Nanthakumar, Dubert-Ferrandon and Walker 77 Furthermore, TGF-β2 modulates intestinal epithelial cell innate immune responses to bacterial antigens and restores potentially detrimental stress responses elicit against lipopolysaccharide, a cell well component of gram-negative microbes.Reference Nguyen, Jiang and Jacobsen 83 These effects are reflected in the results of a clinical study, which suggest that TGF-β2 concentrations in the human milk received by very low birth weight preterm infants tend to correlate inversely with the incidence of NEC and feeding intolerance.Reference Frost, Jilling, Lapin, Maheshwari and Caplan 89 More insight into the effects of breast milk TGF-β in inducing tolerance toward microbes and protection against intestinal inflammation is provided by a series of experiments conducted in a porcine model. Piglets fed colostrum rich in TGF-β2 exhibited significantly lower intestinal inflammation when compared with those fed infant formula with a low concentration of TGF-β2 and TGF-β2 directly attenuated inflammatory responses toward bacterial antigens in porcine intestinal epithelial cells.Reference Nguyen, Sangild, Ostergaard, Bering and Chatterton 90 In a parallel manner, infant formula supplemented with TGF-β2 has been reported to attenuate intestinal inflammation in an experimental murine IBD model.Reference Oz, Ray, Chen and McClain 91
In addition to breast milk TGF-β2 modulating inflammatory responses elicited against microbial antigens, both the concentration of TGF-β2 in breast milk and neonatal responsiveness to TGF-β2 seem to be modulated by specific microbes. Maternal supplementation with the probiotic bacterium Lactobacillus rhamnosus GG has been shown to significantly increase the amount of TGF-β2 in breast milk.Reference Rautava, Kalliomäki and Isolauri 92 A similar increase has subsequently been reported after maternal supplementation with the combination of L. rhamnosus GG and Bifidobacterium lactis Bb-12,Reference Huurre, Laitinen, Rautava, Korkeamäki and Isolauri 93 while consumption of L. reuteri is associated with decreased breast milk TGF-β2 content.Reference Böttcher, Abrahamsson, Fredriksson, Jakobsson and Björkstén 94 The expression of SMAD7, a key negative regulator of TGF-β responsiveness, has been reported to be decreased in preterm infants after supplementation with the probiotic Bifidobacterium breve.Reference Fujii, Ohtsuka and Lee 95 Given the interactions between microbes in the maternal gut, breast milk and the fetal gut on the one hand and TGF-β expression and function on the other, it is intriguing to speculate that the function of microbes and microbial components in breast milk might be to introduce them to the infant gut in a tolerogenic immune milieu created by factors including TGF-β to induce tolerance to colonizing maternal microbes.
Modulating child health outcomes through specific probiotic interventions on the pregnant and breastfeeding mother
After the significance of early microbial contact for the development of immune-mediated and inflammatory conditions was realized, considerable research interest has been focused on modulating host–microbe interactions during the vulnerable perinatal period and early infancy. Promotion of breastfeeding, prudent use of antibiotics and reassessing the benefits and potential harms of cesarean section deliveries are obvious and presumably effective means to reduce the burden of chronic disease on the population level. In addition, supplementation with specific microbes with desirable effects on the host offers itself as a means of directly influencing early host–microbe interaction. Probiotics have recently been defined as ‘live microorganisms that, when administered in adequate amounts, confer a health benefit on the host (Box 1).’Reference Hill, Guarner and Reid 96 Specific probiotics have been extensively studied and shown efficacy in reducing the risk of atopic disease and eczema.Reference Panduru, Panduru, Sălăvăstru and Tiplica 97 There are a number of clinical trials suggesting that probiotic bacteria may reduce the incidence of NEC in preterm infants and the authors of a recent meta-analysis of clinical studies conclude the probiotics are efficacious in prevention of NEC.Reference AlFaleh and Anabrees 98 However, probiotic effects are highly strain and species-specific in vitro and in vivo and it may not be appropriate to devise meta-analyses based on studies using different probiotic bacteria. A sufficiently powered clinical trial assessing the effect of specific probiotics on the incidence of NEC is therefore called for.Reference Abrahamsson, Rautava, Moore, Neu and Sherman 99 Specific probiotics have also been suggested to be of potential benefit in prevention or treatment of various immune-mediated or inflammatory conditions ranging from obesity and the metabolic syndrome to IBD and even certain neurological conditions. It is often assumed that probiotics exert their effects by modulating the composition of the gut microbiome. Interestingly, however, recent research advances indicate that specific probiotics may offer health benefits to the child indirectly via the pregnant or breastfeeding mother.
The first clinical trial assessing the impact of maternal probiotic supplementation on infant health was conducted in Finland.Reference Kalliomäki, Salminen and Arvilommi 100 The primary objective of the study was to reduce the occurrence of eczema in children at high risk due to family history of atopic disease. Maternal supplementation with the probiotic L. rhamnosus GG or placebo was commenced in 159 mothers 2–4 weeks before delivery and continued 6 months after delivery in a randomized, double-blind design. If breastfeeding was discontinued during this time period, the study product was given directly to the infant. The risk of developing eczema was significantly reduced by the probiotic intervention at the age of 2 years and the protective effect persisted to the ages of 4 and 7 years.Reference Kalliomäki, Salminen, Poussa, Arvilommi and Isolauri 101 , Reference Kalliomäki, Salminen, Poussa and Isolauri 102 Perhaps even more interestingly, a similar or more pronounced protective effect was seen in the subgroup of infants who received the probiotic intervention solely via the pregnant and breastfeeding mother,Reference Rautava, Kalliomäki and Isolauri 92 while little or no impact on infant gut microbiome composition and only transient colonization with L. rhamnosus GG was observed to result from the probiotic intervention.Reference Rinne, Kalliomäki, Salminen and Isolauri 103 , Reference Gueimonde, Kalliomäki, Isolauri and Salminen 104
The efficacy of maternal probiotic supplementation in prevention of eczema in high-risk infants was later confirmed in a randomized, placebo-controlled trial of 241 mother–infant pairs in whom maternal supplementation with either the probiotic combination of L. rhamnosus LPR and Bifidobacterium longum BL999 or L. paracasei ST11 and B. longum BL999 was commenced 2 months before delivery and continued for 2 months after birth during exclusive breastfeeding while none of the infants received probiotics directly.Reference Rautava, Kainonen, Salminen and Isolauri 105 A highly significant reduction in the risk of developing eczema was seen in both probiotic groups compared with placebo at the age of 2 years. In line with the results of these clinical trials, a recent meta-analysis of probiotic studies with different administration protocols concluded that both prenatal and postnatal probiotic intervention is needed to achieve reduction in eczema in the child.Reference Panduru, Panduru, Sălăvăstru and Tiplica 97 As discussed above, compiling data obtained from studies using different probiotic strains may not be appropriate. It is likely that even when effective, only certain probiotic bacteria elicit a beneficial effect against specific disease states while other probiotic strain may be ineffective or even have a detrimental effect. In addition, the mechanisms of the protective effects of maternal probiotic administration remain poorly elucidated.
There are data to suggest that maternal intervention with specific probiotics during pregnancy may modulate innate immune function in the placenta and in the fetal gut.Reference Rautava, Collado and Salminen 54 Experimental studies have shown that fetal human gut models are responsive to bacteria and that the probiotic L. rhamnosus GG and its components exert an immunomodulatory effect in the human fetal intestinal epithelium.Reference Ganguli, Collado and Rautava 106 Still, it is not known whether particular probiotic bacteria administered to the mother are transferred to the amniotic cavity or the fetus despite experimental data suggesting that specific microbes introduced in the maternal gut during pregnancy may be recovered in the placenta in mice.Reference Jiménez, Fernández and Marin 51 After birth, there are convincing experimental and clinical data indicating that maternal gut bacteria are transferred to breast milk.Reference Perez, Doré and Leclerc 66 Specifically, the probiotic L. reuteri has been detected in breast milk after oral administration in lactating women.Reference Abrahamsson, Sinkiewicz, Jakobsson, Fredrikson and Björkstén 67 These data suggest that the breastfeeding infant may receive the probiotic via breast milk. In addition, the protective effects of defined probiotic interventions during breastfeeding may be partially explained by modulation of breast milk immune factors including TGF-β as discussed above.
Realization of the profound impact early-life microbial contact may have on human health has lead to a rapid increase in our understanding of factors which may disturb or support the establishment of the indigenous intestinal microbiome (Fig. 1). The impact of cesarean section delivery on the risk of developing immune-mediated and inflammatory conditions are convincingly documented by epidemiological studies and backed by microbiological and experimental data. In a similar fashion, early antibiotic exposure has been shown to have detrimental long-term consequences. These risks should be taken into consideration in clinical care and particularly when devizing practice guidelines on the population level.
Microbiota: the community of microorganisms residing in a distinct ecological niche or environment.
Microbiome: the genetic pool of the microbiota of a given location and their interaction with the environment.
N.B. the terms ‘microbiota’ and ‘microbiome’ are often used interchangeably as if they were synonymous.
Microbial contact: direct interaction with a microbe or microbial component that results in a host response (e.g. inflammation or tolerance).
Immune tolerance: any mechanism by which a potentially injurious immune response is prevented, suppressed or shifted to a non-injurious immune response.Reference Weiner 81
Probiotics – live microorganisms that, when administered in adequate amounts, confer a health benefit on the host.Reference Hill, Guarner and Reid 96
Investigations on microbial contact during fetal life and its significance for health are a new field of study, which may influence both our notion of developmental origins of disease and future care of the pregnant mother. After birth, the beneficial effects of breastfeeding are well known but further complemented with regard to both scope and mechanisms by the recent discovery of the breast milk microbiome and its connections with immune factors in breast milk.
Finally, the pregnant and breastfeeding mother is an emerging target for interventions aiming to support healthy microbial contact and to reduce the risk of chronic disease in the offspring. Maternal interventions with specific probiotics have proven to be efficacious in reducing the risk of eczema. No serious adverse effects have been reported in clinical trails but the long-term efficacy and safety of probiotic interventions need to be monitored. In the future, research efforts should be aimed to find optimal means to combat the epidemics of obesity and immune-mediated disease.
Professors Erika Isolauri and Seppo Salminen are acknowledged for valuable and insightful comments regarding the manuscript.
The author is funded by the Emil Aaltonen Foundation and the Collegium for Science and Medicine of the University of Turku, Finland.
Conflict of Interest