The prenatal gut microbiome: Are we colonized with bacteria in utero?

Abstract

The colonization of the gut with microbes in early life is critical to the developing newborn immune system, metabolic function and potentially future health. Maternal microbes are transmitted to offspring during childbirth, representing a key step in the colonization of the infant gut. Studies of infant meconium suggest that bacteria are present in the fetal gut prior to birth, meaning that colonization could occur prenatally. Animal studies have shown that prenatal transmission of microbes to the fetus is possible and physiological changes observed in pregnant mothers indicate that in utero transfer is likely in humans as well. However, direct evidence of in utero transfer of bacteria in humans is lacking. Understanding the timing and mechanisms involved in the first colonization of the human gut is critical to a comprehensive understanding of the early life gut microbiome. This review will discuss the evidence supporting in utero transmission of microbes from mother to infants. We also review sources of transferred bacteria, physiological mechanisms of transfer and modifiers of maternal microbiomes and their potential role in early life infant health. Well-designed longitudinal birth studies that account for established modifiers of the gut microbiome are challenging, but will be necessary to confirm in utero transfer and further our knowledge of the prenatal microbiome.

Introduction

The human microbiome is the collective genome of the microbes (microorganisms including bacteria, fungi, viruses, eukaryotes and archaea) inhabiting different sites of the body. The gut microbiome exceeds the size of all other microbial communities and is predominantly composed of trillions of bacteria, often representing thousands of species¹. The gut microbiome is a diverse and specialized ecosystem that can be altered by many factors, including diet, age and antibiotic use^2–5. It is comprised of distinct communities of bacteria that are generally stable during adulthood and mostly mutualistic and/or commensal in their relationships with the human host⁶. It is becoming increasingly evident that the bacteria that populate our gut play a role in health and disease. For example, a low diversity (the overall variety of different types of bacteria) of the gut microbiome has been associated with obesity^7,8 and inflammatory bowel disease⁹, while a decrease in the abundance (the relative proportions) of some bacteria has been associated with type 2 diabetes¹⁰. We are beginning to understand how perturbations to the microbiota impact human physiology and health, even in early life. But, the early establishment and subsequent development of the gut microbiome is less studied and knowledge of how these periods may or may not impact future health is incomplete^11–14.

As infants are born, they are exposed to microbes from their mother and the surrounding environment. A main contributor of this exposure is the birth process itself. Infants born vaginally acquire bacteria resembling the maternal vaginal microbiome (predominantly Lactobacillus and Prevotella), whereas infants born via Cesarean (C-section) acquire bacteria resembling the skin microbiome (predominantly Staphylococcus)^15–19. The microbes that infants are first exposed to at birth are thought to play a role in the subsequent maturation of microbial communities, specifically in the gut²⁰. Bacterial diversity increases with age and composition gradually resembles that of adults^21,22, achieving measurable stability early [by ~3 years] in life and consisting of a “core” set of inhabitants that generally persist^22–25. The proportion of certain bacterial groups that compose the gut microbiome and the rate at which the community stabilizes have been associated with future health outcomes²⁶. Recent research has focused on identifying factors responsible for shaping the early infant gut microbiome and understanding how the succession of the newborn gut perhaps sets the stage for future health^{3,21,22,27–29}. Development of the newborn gut microbiome is influenced not just by the mode of birth, whether it be a vaginal or C-section^30,31, but also by several other factors such as duration of gestation, antibiotic exposure, nutrition and genetics^32–38. These exposures may influence the bacteria that infants are first exposed to, which are themselves involved in the early development of the immune system, metabolic programming, neurodevelopment and risk for future disease^{11,12,14,39,40}. Hence, it is thought that there is a critical window in early infancy, and possibly in utero, during which cross talk between the gut microbiome and biological systems (most notably immune and metabolic) mediates development and sets the stage for future health^1,20,41.

Some studies have described the early development of the infant gut microbiome^{22,28,38,42,43}, but it remains unclear how and when the gut is initially colonized with bacteria. Evidence continues to mount suggesting that the colonization of the infant gut begins prenatally, in utero, challenging the widely accepted notion that this process only begins at or after birth⁴⁴. However, the sources of bacteria and means by which they access the fetal gut remain largely unknown. Prenatal development is a crucial period in which the microbiome may already interact with biological systems and maternal microbial transmission may help explain how disruptive events during pregnancy impact fetal programming⁴⁵. Thus, determining if the first colonizers of the gut microbiome are via environmental acquisition or transfer prior to birth is crucial to expanding our understanding of the earliest periods of gut microbiome development. The purpose of this paper is to 1) review the evidence suggesting that in utero transmission of microbes from mother to infant in humans occurs, 2) describe potential sources of transferred bacteria, 3) discuss potential physiological mechanisms of transfer and 4) discuss modifiers of the maternal microbiome and how they may influence early life infant health.

The fetal gut is not sterile

The long-standing hypothesis that a fetus develops in a sterile environment and, thus, is born with a sterile gut is being challenged^44,46. Until recently, bacterial colonization was most commonly observed in the context of infections of the fetal membranes, amniotic fluid and/or preterm delivery, whereas there was little evidence for bacteria at these sites in healthy pregnancies^47–49. Recent advances in DNA sequencing technology have improved our ability to detect bacteria in maternal tissues beyond what was possible with culture-based methods alone (Table 1). For example, in a study of 23 healthy newborns reduced diversity of bacteria in meconium (first stool after birth) was observed compared to adult stool samples, with reductions in Bacteroidetes and Firmicutes (bacteria common in the adult, but not infant gut microbiome)⁵⁰. The bacterial composition of meconium, considered a proxy for the in utero gut microbiome, did not differ based upon vaginal or C-section birth. This is consistent with birth mode exerting influence on the gut microbiome after actual birth, implying that colonization of the gut occurs prior to delivery, independently of delivery mode^31,51. The meconium from 21 healthy neonates contained species from Staphylococcus and Bifidobacterium, commonly found in the human gut, and meconium from 52 pre and full-term infants was characterized by a generally low diversity of bacteria from dominant genera (Firmicutes and Proteobacteria), suggesting a gut founder population^52,53. More than 50% of bacteria in the meconium also colonized amniotic fluid, whereas there was almost no overlap with vaginal and oral populations⁵³. Thus, the infant gut may be populated in utero, prior to delivery, possibly in part through the consumption of amniotic fluid.

Table 1.

Summary of literature identifying maternal bacteria as potential sources of infant transfer in humans.

Author (reference)	Population	Sample Size	Method	Relevant Finding(s)	Public Data
Meconium
Hu, J et al. (50)	Neonates	23	16s rRNA sequencing	Not Sterile, not affected by delivery mode	NA
				High adundacne of Proteobacteria
				Low abundance of Bacteroidetes
Jimenez, E et al. (52)	Neonates	21	culture	Not Sterile	NA
			16s rRNA sequencing	E. fecalis, S. epidermidis and E. coli predominant species
Ardissone, AN et al. (53)	Neonates	52	16s rRNA sequencing	Majority of meconium not sterile	NA
				Genera found in meconium and amniotic fluid accounted for −61% of the relative abundance in infant meconium
Vaginal
Aagaard, K el al. (53)	Women	24 pregnant	16s rRNA sequencing	Microbiome is reduced in taxa during pregnancy (lower diversity) and structure of microbiome varies with gestational age	NA
		60 non-pregnant		Predominance of Lactobacillales. Clostridials, Bacteroidales. and Actinomycetales
DiGiulio DB et al. (57)	Pregnant Women	49	16s rRNA sequencing	Vaginal microbiome remains stable through pregnancy	SRA(SRP #288562)
				Dominance by Lactobacillus. Presence of Gardnerella associated with preterm birth
Gut
Koran, O et al. (62)	Pregnant Women	91	16s rRNA sequencing	Between-individual diversity of the gut microbiome increases during pregnancy	NA
				From trimester 1 to 3 abundances of Proteobacteria and Actinobacteria increase
Collado, MC et al.(63)	Pregnant Women	36 normal weight	qPCR	Increase in diversity of gut microbiome in pregnancy regardless of weight status	NA
		18 overweight	FCM-FISH	Overweight women had higher numbers of Bactaroidas and S. aureus
Amniotic Fluid
Collado, MC et al (44)	Mother-infant pairs	15	culture	Not Sterile. Low abundance, richness and diversity of bacteria	NA
			16s rRNA sequencing	Abundance of Proteobactaria (Entarobactar and Escherichia/Shigella), Propiotiibactarium and Enterobacteriaceae species
				Microbiota similar to meconium
Wang, X et al. (89)	Pregnant Women	36 preterm	culture	Most prevalent species detected were Ureaplasma parvum, Escherichia coli and Fusobacterium nucleatum	NA
		8 term	16s rRNA sequencing	High degree of bacteria overlap with cord blood.
Bearfield, C et al. (90)	Pregnant Women	48	qPCR	Streptococcus spp. and F. nucleatum detected in amniotic fluid and oral dental plaque	NA
DiGiulio DB et al. (88)	Pregnant Women	166	culture/real-time PCR bidirectional sequencing	Culture and PCR methods detected Mycoplasma hominis, Ureaplasma sp	NA
				Streptococcus agalactiae, Lactobacillus sp., Prevotella sp., and Fusobacterium nucteatum
Placenta
Prince, AL et al (49)	Pregnant Women	27 term	Shotgun sequencing	Not sterile	NA
		44 preterm		Term pregnancy placentas have a greater abundance of Enterobacter species (gammaproteobacteria) and Lactobacillus crispatus (prevalent vaginal species)
Stout, MJ et al. (61)	Pregnant Women	195	Histopathology	27% of basal plate samples contained Gram-positive and -negative intracellular bacteria	NA
			Histochemistry
Aagaard, K et al. (32)	Pregnant Women	320	16s rRNA sequencing	Most abundant species detected is E. coli. Placental microbiome resemhles oral microhiome	SRA(SRP #238913); dbGaP
			Shotgun sequencing	Several other non-pathogenic phyla detected including Firmicutes, Tenericutes, Proteobacteria, Bacteroidetes, and Fusobacteria.
Collado, MC et al. (44)	Mother-infant pairs	15	culture	Not Sterile. Low abundance, richness and diversity of bacteria	NA
			16s rRNA sequencing	Abundance of Proteobacteria (Enterobacter and Escherichia/Shigella), Propionibacterium and Enterobacteriaceae species
				Microbiota similar to meconium
Cord Blood
Wang, X et al. (89)	Pregnant Women	36 preterm	Culture	Bacteria detected in cord blood include Escherichia coli, Streptococcus agalactiae.	NA
		8term	16s rRN A sequencing

^*SRA=Sequence Read Archive

The maternal microbiome changes during pregnancy

In addition to known physiological changes, microbiome communities change during pregnancy as well. The vaginal microbiome has distinct features of microbial abundance and diversity when compared to other body sites that harbor microbial ecosystems⁶. 16S rRNA sequencing of the vaginal samples from 396 healthy women revealed that the primary bacteria present are Lactobacillus⁵⁴. Lactobacillus contribute to the regulation of vaginal pH and mediate infection via the production of lactic acid. Despite variation in the dominant species of vaginal Lactobacillus detected across women, the overall physiological function of Lactobacillus is conserved, explaining the overwhelming abundance of these bacteria⁵⁴. Lactobacillus dominance in the vaginal microbiome persists into pregnancy, which is characterized by low diversity and high stability^55–57. In a longitudinal study of the vaginal microbiota in healthy women with term deliveries, shifts within communities dominated by Lactobacillus, from one species to another, were observed with a heightened temporal stability of the vaginal microbiome during pregnancy^56,58,59. Lactobacillus dominance inhibits ascending infections which could expose the fetus to pathogenic microbes common to intra-amniotic infections and vaginosis^58,60. Stabilization of the vaginal microbiome in pregnancy, specifically Lactobacillus species dominance, may represent a protective evolutionary adaptation that favors colonization of offspring with beneficial microbes (i.e. Lactobacillus) during, and potentially before, delivery⁵⁹. In addition to vaginal ascension, it is believed that endometrial uterine tissue hosts microbes that incorporate into the placenta at the time of placental implantation during pregnancy⁶¹, suggesting that the fetus could be exposed to bacteria from the vaginal or uterine sources prior to delivery.

In contrast to the vaginal microbiome, the gut microbiome is characterized by a greater abundance and more diverse bacteria composition⁶. Yet, the structure of the gut microbiome also fluctuates throughout pregnancy^62,63. In a study of 91 pregnant women the gut microbiome had a significantly (P < 0.001) reduced overall richness (a diversity metric) with significantly increased abundance of Actinobacteria (5.1% vs. 9.3%, respectively; P = 0.003) and Proteobacteria (0.73% vs. 3.2%, respectively; P = 0.0004) from the first to the third trimester. Actinobacteria and Proteobacteria are commonly found in the adult gut, but in much less abundance than Bacteroidetes and Firmicutes. Additionally, with advancing gestational age, diversity in the gut microbiome between mothers grew, while total bacterial load increased and the abundance of bacterial groups within a given person (alpha diversity) decreased⁶². This suggests gestational duration may be important in maternal transfer, which is supported by guts of preterm infants showing lower bacterial diversity than those of term infants^64,65. Comparing fecal sample microbiota from 21 pregnant women one week before and one year post-delivery revealed that postnatal samples had higher bacterial richness and unique taxonomic clustering compared to prenatal samples, illustrating an impact of pregnancy on the gut. Slightly lower functional richness in the prenatal samples suggest that the gut microbiome changes prior to delivery for an unknown reason³⁸. Other studies of the gut microbiome during pregnancy show stability throughout gestation. Fecal samples from seven pregnant women showed communities remained colonized by bacteria typically found in healthy adults (Firmicutes, Bacteroidetes and Actinobacteria), indicating no prepartum changes in composition or diversity⁶⁶. In a similar study, in which 40 pregnant women were repeatedly sampled during gestation, diversity measures of the gut microbiome did not significantly change during pregnancy⁵⁷. These contrasting findings may be explained by heterogeneity in factors that independently influence the maternal gut microbiome, such as maternal insulin resistance and BMI⁶⁷. Larger studies designed to carefully measure factors such as diet, health, antibiotic use, group B Streptococcus status, and age will be necessary to further clarify the role of maternal microbiome changes during pregnancy and their influence on the developing fetus.

Bacterial transfer from mother to offspring

Analyses of meconium have provided evidence against the hypothesis that the womb is a sterile environment. Yet, direct evidence of bacterial colonization in the human fetal gut is lacking and the potential mechanism(s) for colonization of the gut in utero is(are) unclear. It has been postulated that the maternal transmission of microbes may in fact be a universally shared phenomenon and colonization of the infant gut in utero could be the result of a beneficial evolutionary process, even in humans (Table 1)⁶⁸.

Despite a lack of evidence in humans, there is support for vertical transfer (from mother to offspring) in animal models^68–70. For example, significant reductions in Lactobacillus (P = 0.02), a dominant component of the vaginal microbiome, were observed in the guts of mice vaginally born to stress-exposed mothers versus controls and maternal vaginal Lactobacillus abundance was positively correlated (r²=0.74, P < 0.001) with offspring’s gut Lactobacillus abundance⁷¹. Similar results were obtained when 14 pregnant monkeys were exposed to stress for six weeks during pregnancy⁷². Offspring born to mothers exposed to stress, particularly late in pregnancy, had significantly reduced Lactobacillus and Bifidobacterium (P ≤ 0.05) at day 2 post-delivery compared to controls. In another mouse study, eight female mice were fed plain milk or milk containing genetically labeled bacteria from conception until term at which time fetuses were obtained via sterile C-section for stool sampling⁵². The samples of offspring born to mothers inoculated with a genetically labeled Enterococcus species were found to contain the identical genetic label, which was not detected in control mothers or offspring.

In human studies the evidence is less conclusive. The fecal samples of 17 healthy mother-infant pairs were found to contain strains belonging to Bifidobacterium species⁷³. This was observed in 11 of 12 of vaginal deliveries, but not in C-section deliveries, indicating that delivery method may impact transmission of maternal intestinal bacteria. This is consistent with other work showing an increased likelihood (OR=19 (5.16–69.96); P < 0.001) of infants having detectable Bifidobacterium species at one month of age if the same species was detected in maternal samples collected during pregnancy⁷⁴. In a more recent work that sampled 15 mother-infant pairs, infant meconium contained bacteria distinct from the maternal gut microbiome that were detected in placental and amniotic fluid⁴⁴. Several of the predicted gene functions⁷⁵ of bacteria found in placenta and amniotic fluid were also found in meconium (i.e. energy metabolism and membrane transport), despite the overall composition and function of meconium being unique. Intrauterine sites containing specific, low diversity microbiota in pregnancy, therefore, may be a source of transferred bacteria and function⁴⁴. These findings, although not definitive, indicate that the maternal gut and urogenital microbiomes play a role in shaping the infant gut microbiome. However, the sources of the maternal bacteria and the processes by which they pass to the fetus or neonate are not well understood. Emerging work posits that the colonization of the gut microbiome begins prenatally via transfer of bacteria (Figure 1) through the placental barrier or through ingestion of amniotic fluid. This contrasts with prior assumptions that ascension of bacteria from the vagina was the only plausible method of bacterial colonization in intrauterine tissues^57,76,77.

Figure 1. Maternal sites that may contribute to offspring gut microbiome.

During pregnancy there are changes to maternal microbiomes. Translocation of bacteria from the oral and gut microbiomes of mothers during pregnancy, in addition to ascension of bacteria from the vaginal microbiome, may explain the presence of non-pathogenic bacteria in intrauterine locations. Maternal-derived bacteria detected in neonatal meconium, a proxy for the in utero gut microbiome, is suggestive of the prenatal transfer of bacteria from mother to infant.

Pathogenic species of bacteria have been detected in intrauterine locations for some time in the context of infections, complications or preterm delivery^78–80. The types and origins of bacteria existing in these sites, in the context of healthy pregnancy, are relatively unknown. Current studies have assessed placental microbiome in term and preterm pregnancies to address this knowledge gap. A recent study of 55 pregnant women with pre-eclampsia compared placental bacteria with that of 55 matched normotensive pregnant controls. While no bacteria were found in controls, 13% of the women with pre-eclampsia harbored bacteria known to be associated with gastrointestinal infection and periodontitis⁸¹. This is intriguing, as it suggests an oral-placental microbiome relationship. Bacterial infection of the placenta or fetal membranes is viewed as a risk for preterm delivery, consistent with the observation that pathogenic bacteria are found in greater abundance in placentas from preterm deliveries⁸². A majority of placentas from preterm deliveries contain bacteria, but common, non-pathogenic species belonging to Bifidobacterium and Lactobacillus are also found, albeit at lower abundances, in term placentas^82–84. Forty eight placental samples from healthy pregnancies were recently found to contain a unique and diverse microbiome predominately composed of commensal bacteria. When compared with gut, oral, skin, nasal, and vaginal microbiomes obtained from non-pregnant controls, the placental microbiome did not share similarities with the gut or vaginal microbiome, but resembled the oral microbiome^6,32. Whole genome sequencing of bacteria at the different sites demonstrated distinct metabolic gene functions of the placenta and vaginal microbiome (ex. vitamin metabolism) compared to other body sites (ex. carbohydrate and amino acid metabolism), but meconium was not available for comparison³². Interestingly, associations of Lactobacillus in human placentas with immune responses in the developing fetal gut suggest that fetal exposure to bacteria in utero plays an important role in immune development prior to birth^85,86. In humans, oral-placental transfer and feto-placental interface are difficult to study requiring animal models to elucidate these mechanisms. When pooled samples of human saliva and subgingival plaque were injected into ten pregnant mice, all of the placentas contained one or more species of bacteria present in the human oral samples⁸⁷. This suggests that bacteria in the blood caused by periodontal infection, or other infections associated with pregnancy complications⁴⁹, can be a source of hematogenous bacterial translocation to the placenta.

In addition to placental tissue, bacteria are also present in amniotic fluid and cord blood, two additional potential sources of mother-to-infant transmission. Multiple bacterial species were identified using culture and DNA-based analyses of amniotic fluid from 166 preterm deliveries, in which 15% of samples contained bacteria⁸⁸. A larger proportion of mothers with bacteria in their amniotic fluid delivered prematurely, consistent with prior literature associating the presence of bacteria in the amniotic cavity with preterm delivery^79,88. Similarly, paired amniotic and cord blood samples from 36 pregnancies with preterm complications shared bacteria, specifically, F. nucleatum (species of oral origin also found in placenta), suggesting a common source via oral hematogenous translocation⁸⁹ (Figure 2). Importantly, bacteria are also found in amniotic fluid of healthy, term deliveries. Nearly 70% of amniotic fluid collected via amniocentesis prior to elected C-sections contained bacteria also detected in paired oral samples from 48 women⁹⁰. This observation mirrors the oral bacteria presence in placental tissue and supports hematogenous translocation of oral bacteria to intrauterine locations (Figure 1). Moreover, bacteria in amniotic fluid are not always pathogenic and appear to harbor a more diverse microbial population than previously thought.

Figure 2. Proposed mechanisms of maternal transfer of bacteria to the fetus in utero.

Gut: The lumen of the maternal distal gut is lined with enterocytes that under normal conditions form a cellular and mucosal barrier to gut microbes (yellow spheres). Diet, stress, antibiotic exposure, disease and pregnancy may alter the thickness of the mucosal layer and the integrity of the enterocyte border. Gaps in this layer (intestinal permeability) allow bacteria to cross the intestinal barrier into blood or lymphatic vessels where they translocate to other body sites. Similarly, resident dendritic cells (DC) probe the lumen and transport bacteria across the gut border in an immune stimulating process. Oral: Dental injury or surgery and oral conditions that cause inflammation (gingivitis) allow oral bacteria contained in salivary and subgingival microbiome communities exposure to the circulatory system. Placenta: Bacteria already present in the endometrial lining or urogenital regions may be incorporated into the developing placental decidua. Bacteria transferred in the blood from other maternal microbiomes to the placenta may populate the decidua, fetal membranes and sinuses and transfer to the developing fetus in utero via amniotic fluid and cord blood.

The maternal gut microbiota is another potential source of transferred bacteria. The intestine has barrier properties that limit the passage of harmful substances across the intestinal epithelium⁹¹. Increased intestinal permeability is characterized by enhanced passage of luminal contents across the mucosa, which may stimulate immune system processes^68,92,93. Bacteria that cross the epithelial barrier may access the circulatory or lymphatic systems and translocate to other sites (Figure 2) where they may impact immune development and response⁹². Excess permeability has been associated with obesity⁹⁴, gastrointestinal diseases⁹⁵, and liver disease⁹⁶ and can result in translocation of bacterial lipopolysaccharide (LPS), endotoxins and actual bacteria into circulation and the periphery. Bacterial translocation has been found to be increased during pregnancy and lactation in mice and bacteria can be found in low amounts in the blood of healthy pregnant women^97,98. Maternal bacteria, therefore, may translocate to intrauterine sites and take part in the vertical transfer of microbes to the fetal gut in utero¹⁴. Alterations to the endothelial integrity of the placenta in pregnancy could allow a formerly impermeable barrier to “leak” bacteria and LPS into the cord blood and amnion³². This hypothesis is supported by the observation of LPS in the cord blood of healthy and preterm preganacies⁹⁹. Similarly, cord blood obtained during 20 healthy, elected C-section deliveries showed 45% of blood samples to contain species also identified in the oral cavity⁸⁹. In pregnant mice inoculated with a genetically labeled strain of bacteria isolated from human breast milk, the identical strain was detected in the amniotic fluid of the inoculated mice, but not in a placebo group¹⁰⁰. Despite detecting bacteria in the placenta, it is difficult to determine their origin. The processes facilitating translocation of bacteria to the placenta, crossing of the placental barrier and contact with the developing fetus may be a ubiquitous aspect of pregnancy. Maternal transfer of bacteria to the fetus could occur for physiological reasons, such as fetal intestinal immune development^85,92. More animal model and clinical research is warranted to clearly define potential maternal sources of bacteria (oral, gut, vaginal, amniotic fluid and placenta) and to define mechanisms underpinning vertical transmission of bacteria to the infant gut in utero.

Maternal environment and the infant microbiome

If maternal transfer of the microbiome to the fetus does occur under normal conditions in humans the process is likely impacted by maternal environmental factors known to influence the microbiome, including diet, antibiotic exposure, stress and health status. Western diets are known to alter the gut microbiome^101,102 and are associated with increased intestinal permeability in human and animal studies^103,104. Pregnant mice fed a high fat Western diet demonstrated alterations to their gut microbiome and their offspring had significantly lower overall bacterial abundance, a higher ratio of Firmicutes to Bacteroidetes and altered immune development compared to a normal diet group¹⁰⁵. In pregnant monkeys fed a high fat or control diet during and after gestation a maternal high fat diet altered both the maternal gut microbiome as well as that of their offspring. Offspring were maintained on a high fat or control diet after birth and those born to mothers on a high fat diet had a reduced representation of Campylobacter, thought to mediate gastrointestinal outcomes such as IBD and gastritis. Offspring maintained microbiome changes resultant from the maternal high fat exposure during pregnancy despite consuming a low-fat diet after weaning. In both studies, maternal diet altered offspring gut microbiota, which may be resultant from bacterial translocation and/or immune mediated responses in the gut. In humans evidence is not as abundant. In a longitudinal study of 81 mother-infant dyads, meconium from infants born to a subset of mothers whose diets differed from the mean (n=26) was compared. The overall microbiome of the meconium significantly differed (P = 0.001) as a function of maternal gestational diet (high fat vs. a balanced diet), with a significant depletion of Bacteroides (P < 0.05) in the infants exposed to maternal high-fat diet.¹⁰⁶ In both humans and primates, these transgenerational effects were independent of maternal obesity status, suggesting that diet has a dominant influence on the gut microbiome¹⁰². Although these findings need further, large scale replication, data points to maternal diet imparting in utero effects on the fetal microbiome.

Antibiotic exposure influences the gut microbiome^4,107,108 and administration during pregnancy alters vaginal microbiology prior to birth, with potential long-term effects on the early colonization of the neonate¹⁰⁹. Exposure to antibiotics prenatally and in infancy increases risk of overweight and asthma in childhood^{110–112,122}. This is consistent with animal studies demonstrating that early life antibiotic exposure can impart long-term consequences on immunity and metabolism through altering the gut microbiota⁴ In a study of nearly 10,000 children, prenatal exposure to antibiotics was associated with an increased prevalence of overweight (adjusted prevalence ratio 1.29 [95%CI: 1.1–1.45]) and obesity (1.27 [95%CI: 1.03–1.62]) at age 7–12 years that varied by sex and birth weight¹¹³. Similarly, children whose mothers were administered antibiotics in the 2^nd or 3^rd trimester had an increased risk of obesity at 7 years (relative risk 1.84 [95%CI: 1.33–2.54])¹¹⁴. Of note, antibiotic resistant genes present in maternal gut bacteria are detectable in samples from perinatal and early life infant fecal samples^115,116. Thus, bacteria as well as their specific functions are transmitted trans-generationally in processes that likely occur prior to delivery. Prenatal stress and maternal health may also impact maternal microbiomes. A study of healthy infants revealed that maternal prenatal stress was associated with a higher abundance of pathogenic members of Proteobacteria and lower abundance of two bacteria generally associated with positive health outcomes, Bifidobacteria and Lactobacillus¹¹⁷. The infants of mothers reporting high maternal stress levels also have higher rates of allergy and gastrointestinal disorders, suggesting a stress-altered maternal microbiome may mediate health outcomes. The gut microbiome profiles of obese adults bear similarities to the gut bacteria in children born to mothers with a high BMI. Maternal BMI is associated with decreases in Bifidobacterium and increases in Bacteroides bacteria in offspring¹¹⁸ and significantly correlates with infant gut Bacteroides (r=−0.4, P = 0.007) and Staphylococcus (r=0.77, P < 0.0001) at six months of life¹¹⁹. Additionally, maternal obesity results in a significantly reduced diversity of gut bacteria in offspring, even at age two¹²⁰. These patterns are also observed in the gut microbiome of obese adults.

The longer-term impact of perturbations to the early microbiome on disease risk in children are not yet well understood, but it is clear that maternal contributions to the infant microbiome before, during and after birth are involved. Some exposures during pregnancy alter the maternal microbiome and may enhance intestinal permeability and translocation, facilitating the entry of maternal bacteria to the uterine environment (Figure 1). The colonization of the fetal gut with the “right” bacteria could be a normal process, whereas acquisition of the “wrong” bacteria in utero, due to negative maternal exposures, may influence future disease risk in children. Thus, maternal transmission of microbes in utero could be viewed as a beneficial or detrimental process to the infant. Data suggest that early postnatal colonizers of the infant gut influence immune priming^97,121, preparation for breast milk^122,123, programming metabolism^124,125 and neurodevelopment^71,126. Some of these colonizers could be present in utero and it is not known if gut bacteria, acquired prenatally, have the same impact on infant development as those acquired during birth or after. This is an important distinction, as the keystone species of the gut are thought to influence the development of the gut microbiome^22,127,128. Importantly, the influences first colonizers impart on host health are likely mediated by external factors, yet the mechanisms and temporality underlying these associations remain elusive. Studies administering probiotic supplementation to mothers during pregnancy suggest that altering the maternal microbiome can alter the infant microbiome prior to delivery^85,129, but causally linking maternal factors in pregnancy to alterations in the microbiome that impact risk for obesity or type 2 diabetes, for example, is difficult. In a prospective study of 909 infants, colonization with Bacteroides fragilis in the first month of life was associated with significant increases in BMI up to age 10, however the relationship was mediated by fiber intake and comparisons with maternal samples were not studied¹³⁰. Similarly, children born to mothers receiving probiotic supplementation during pregnancy, compared to control, showed a 40% reduction in atopic dermatitis. More interestingly, atopic children within the treatment group had significantly higher levels of Bifidobacterium dentium compared to healthy children, as early as 10 days, indicating a differential outcome based on early colonizers of the gut¹³¹. This is consistent with the hypothesis that mothers pass microbes on to their children in the womb¹³² and the transmission of maternal microbes may be modifiable^133,134. Larger studies that confirm and identify novel environment-microbiome relationships in pregnant women will be critical to understanding how maternal environment in pregnancy impacts future infant health. Furthermore, characterization of the processes involved in in utero, peri- and postnatal transmission, with a specific focus on factors that are known to shape the early life microbiome, is needed before the contributions of each factor to overall microbiome composition and health outcomes can be understood.

Conclusion

The trans-generational transmission of microbes is only beginning to be understood⁷³. In animal models, there is ample evidence of maternal transmission in utero with accompanied mediation of disease risk and resilience in offspring¹³⁵. However, the process and sequence of events is not yet well understood in humans. In addition to alterations to the gut microbiome, pregnant women also undergo physiological changes that favor translocation of bacteria to intrauterine sites where gut and oral bacteria have been measured^32,90,100. This, combined with the detection of bacteria in the meconium of most infants, suggest that vertical transmission in utero may occur.

Studies targeting colonization of the fetal gut in utero will be paramount to furthering our understanding of the early life gut microbiome. However, establishing birth cohorts is challenging, costly and time consuming. Although difficult, the collection and analysis of multiple maternal intrauterine tissue microbiomes and comparison with other body sites is imperative to this effort. Longitudinal studies tracking pregnant women and their children will be needed to assess the consequences of maternal health and maternal microbiome during pregnancy on child health in early and later life. A more comprehensive functional understanding of the microbiome, whether from whole genome sequencing, inferred metagenomics or metabolomics, is currently lacking in the study of maternal microbiomes through pregnancy. Combined with compositional study of the maternal microbiome, a better functional understanding of these microbiomes will help reveal their potential influence on the infant. Additionally, studies must vigilantly account for known modifiers of the microbiome such as antibiotic exposure, diet, disease status and maternal exposures during pregnancy, among others¹³⁶. Many important intrauterine sites, such as placenta and amniotic fluid, contain relatively low abundances of bacteria compared to gut or oral samples, making analysis of these samples sensitive to contamination^137–139. Future studies must include adequate controls in analyses and implement strict, validated collection techniques in order to account for potential contamination¹⁴⁰. Current sequence-based technologies cannot discriminate between “live” and active versus “dead” and inactive bacterial DNA, thus sequences can arise from dead or fragmented bacterial DNA present in the samples sites⁷⁶. The field is beginning to address this, yet studies must currently consider this possibility, especially in the context of low biomass samples.

If indeed in utero transfer of maternal microbes to the fetus occurs in humans, the maternal microbiome during pregnancy could be a target for modification to optimize this process to support transfer of beneficial microbes and suppress the transfer of harmful or pathogenic bacteria to the infant. This could effectively open the opportunity for personalized treatments that embrace the microbiome¹⁴¹ to target early life infant health¹⁴², introducing a novel facet to the concept of heal the mother, heal the baby.