Placental Microbiome and Its Role in Preterm Birth

1. The placenta is one of the least understood human organs, especially when considering its effect on fetal outcomes.

2. Recent work has revealed that the placenta is not sterile.

3. Many unexplored questions remain regarding microbial colonization in the placenta and the effect on pregnancy and fetal outcomes.

After completing this article, readers should be able to:

1. Summarize the types of microbes found in placentas.

2. Describe the effect of vaginal microbiome on pregnancy outcomes.

3. Recognize the effect and association of maternal-fetal microbiomes on long-term fetal outcomes.

The conventional paradigm is that the placenta is a sterile organ and any intrauterine infection would be caused by ascending infection from the lower genital tract. (1) Indeed, there is abundant evidence that links infection and inflammation with preterm birth (PTB), including associations with subclinical intrauterine infections, intra-amniotic infection, and extrauterine maternal infections, such as pyelonephritis and periodontal disease. (2)(3)(4)(5)(6) The most common microorganisms associated with PTB are genital Mycoplasma species, in particular Ureaplasma urealyticum and Ureaplasma parvum. (7) Others include Leptotrichia/Sneathia, Atopobium vaginae, and bacterial vaginosis (BV)–associated bacteria types 1, 2, and 3, which are bacteria associated with BV, a condition in which the normal vaginal flora of lactobacilli are replaced by other low- and high-grade pathogens. Other studies have identified the intrauterine bacterial species using culture-dependent or culture-independent methods. (8)(9)(10)(11)(12) We refer the reader to an excellent review by Jefferson (2) for a full description of the many microbial species associated with PTB. PTB-causing infections have been postulated to originate in the reproductive or genitourinary tract, ascend upward through the cervix, and possibly breach the placental barrier. However, although there are undoubtedly strong associations, these species have not been definitely found to be causative agents of PTB, suggesting that there was more to the findings than hitherto understood.

The Human Microbiome Project consortium characterized the microbiome in multiple human organs and helped establish the link between changes in the microbiome and human health. (13) The advent of culture-independent microbiome analyses and the pioneering investigations of the Human Microbiome Project have provided important insights into human physiology and pathophysiology in multiple fields of medicine, such as gastrointestinal disorders, endocrine diseases, and cardiovascular disorders, to name a few. Traditional consideration of microbes as pathogens has shifted instead to an understanding of the complex and necessary interactions between the human host and the microbial communities in various body habitats. Of importance, microbial communities demonstrate relationships of mutualism with the human host, providing resistance to acquisition of pathologic infections, immune system education, nutrient breakdown in the gut, and many other processes that are beneficial to the human host.

In line with this, bacterial colonization of the human placenta has been found not only in patients with clinical infections or in preterm births but also in normal pregnancy and term placentas. Our group provided morphologic evidence that both gram-positive and gram-negative bacteria of diverse morphologic origins were present in a third of all placentas from preterm and term pregnancies. (12) Of interest, these microbes were found to be intracellular, within extravillous trophoblast (EVT) cells in the maternal decidua (Figure). EVT cells are fetally derived and directly contact maternal stromal and immune cells. EVT cells have unique immune-privileged status because they are coated with self-antigens, including the major histocompatibility antigen HLA-G, which is implicated in the maintenance of immune tolerance of the fetus. (14) A new study supports the existence of such a placental microbiome. Aagaard et al (15) recently identified a low-abundance but metabolically rich microbiome in the human placenta using culture-independent whole-genome shotgun technology on genomic DNA isolated from human placentas (preterm and term). Of particular interest was the cataloguing of commensal bacterial species in the normal and term placenta, including Escherichia coli, Prevotella tannerae, Bacteriodes species, Fusobacterium species, and Neisseria lactamica among others. Among the bacterial species found in placenta, E coli was surprisingly identified as the species of highest abundance.

Pregnancy is a unique state with profound physiologic changes. During pregnancy, the body undergoes substantial changes in anatomy, immunology, endocrinology, and metabolism, which can lead to fluctuations of physical function in different organs. Not surprisingly, the placental microbiome studies support that such temporal alterations are accompanied by remodeling of this microbiome during pregnancy. For example, the study by Stout et al found that the number of placentas harboring intracellular bacteria increased by half in placentas before 28 weeks of gestation. Moreover, when the study by Aagaard et al compared the placental microbiome of term relative to preterm placentas, they found preterm placentas had changes in abundance of certain bacteria taxa, such as Burkholderia species, whereas term placenta had an increase in paenibacilli. Whether the alterations in microbial patterns are a feature of gestational age or indicative of specific functions or physiologic stages of placental and fetal development remains unknown.

The bacterial type, functions, and host responses at maternal-fetal interface may also collectively determine the outcome of the presence of a placental microbiome. The aforementioned EVT cells have been found to be preferred sites of entry for placental pathogens, such as Listeria monocytogenes (16) and Toxoplasma gondii (17), as well as E coli (18), in primary human placenta cultures, suggesting that EVTs embedded in the maternal basal plate may serve as a protective environment for placental microbes. Whether this is true of commensal organisms has not yet been tested. Stout et al (19) recently found that placentas had an increase in HLA-G positivity in preterm births compared with those from term births, implying the contribution of maternal (host) factors in etiology of preterm birth. Whether and how changes in HLA-G levels affect the placental microbiome remains to be determined.

Traditional teaching is that normal vaginal ecology is dominated by lactobacilli and the local production of lactic acid contributes to an environment that restricts survival of pathogenic bacteria. BV has been broadly defined as alterations in bacterial composition of the vagina and more specifically as a relative decrease in the number of lactobacilli. Such shifts away from the normal vaginal flora have been linked to an increased risk of sexually transmitted infections, suggesting that the normal vaginal bacterial communities serve important functions in preventing survival of pathogenic species. (20)(21) Culture-independent sequencing techniques have refined this traditional understanding of the normal vaginal microbiome. A large study of asymptomatic, nonpregnant, reproductive age women found that the vaginal microbiota could be categorized into 5 discrete groups based on the type and abundance of microbial communities present. (22) Four of the 5 communities had the Lactobacillus genus as the predominant community present (iners, cispatus, jensenii, and gasseri species), but one community type was not dominated by Lactobacillus. This unique heterogeneous community had increased evenness (ecologically defined as multiple species present in relatively equal proportions in contrast to Lactobacillus dominance) with communities of anaerobic bacteria, such as Prevotella, Atopobium, Gardnerella, and others. The findings of Ravel et al suggest that normal vaginal flora is not required to have Lactobacillus predominance, there may be racial differences in normal vaginal microbiota, and the traditional diagnosis of BV may not be consistent with current culture-independent evidence of vaginal bacterial ecology.

There is ample evidence that abnormal vaginal flora (albeit broadly and heterogeneously defined) is associated with PTB. (23)(24)(25)(26)(27)(28)(29) Despite the clear correlation, well-designed randomized trials that treated BV have not realized a significant reduction in PTB. (30)(31) Molecular-based investigations have examined not only who is present in the vaginal microbiota but also the metabolic functions of normal vaginal communities and how this information will shed light on understanding the pathogenesis of PTB. One could speculate that immune system and hormonal changes that occur in pregnancy might alter vaginal microbiome communities compared with the nonpregnant state. Using 16S microbial census analysis, Aagaard and colleagues (32) examined the vaginal microbiome community in pregnant patients (n = 24) compared with a nonpregnant population (n = 301) and found that pregnancy confers a unique microbiome signature compared with the nonpregnant state. The pregnant microbiome was characterized by decreased species richness and diversity. Furthermore, richness decreased with increasing proximity to the cervix. Hyman et al (33) hypothesized that differences in the vaginal microbiome may be associated with PTB. In a prospective cohort study, these investigators sampled the posterior fornix in 46 patients classified as high risk for preterm birth and 42 low-risk controls and then followed the pregnancy forward for gestational age delivery outcome. The Shannon Diversity Index (a measure of number and abundance of species present) was decreased in pregnancies that ultimately resulted in a preterm birth. Taken together, Aagaard et al describe depressed species richness and diversity of the pregnancy-related vaginal microbiome compared with the nonpregnant state, and the findings of Hyman et al suggest that oversuppression of the vaginal microbial communities may be associated with adverse obstetric outcomes. Romero and colleagues (34) sampled the vaginal microbiome every 2 to 4 weeks in a prospective cohort of 22 women who delivered full term without complication. Consistent with the findings by Ravel et al in nonpregnant patients, they document that there is no single core vaginal microbiome in pregnancy. Rather, multiple community states are dominated by Lactobacillus species. In addition, in pregnant women who delivered at term, a heterogeneous non–Lactobacillus-dominated community is found rarely. Using measures of community stability, the researchers examined whether pregnant women change from one community type to another. Their findings suggest that vaginal bacterial communities are significantly more stable (fewer community changes over time) in pregnancy than in the nonpregnant state. In pregnant women who change community states, the change most commonly occurs from one Lactobacillus-dominant community to another Lactobacillus-dominant community and rarely to the heterogeneous community type of microbiota.

As current and future studies carefully classify the census of vaginal microbial communities and answer who are there, we must then turn our attention to questions regarding what they are doing and how do they do it. How the vaginal microbiome may affect metabolic functions, genomic redundancy, interactions with host, and beneficial or harmful pathways in pregnancy remains to be elucidated.

Epidemiologic studies have established a firm link among oral flora, periodontal disease, and PTB. This finding suggests that a bloodborne route is involved in microbial colonization of the placenta. (6)(35) Consistent with this idea, periodontal pathogens have been detected in placentas of women with preeclampsia and in the amniotic fluid of pregnant women with preterm labor. In support of this concept, state-of-the-art next-generation sequencing and metagenomic analyses conducted by Aagaard et al revealed that the placental microbiome is most homologous with oral flora and not vaginal flora. Future studies determining the extent that the oral flora community is capable of oral-uterine transmission during pregnancy are needed to provide new knowledge of the diversity of oral bacteria involved in PTB. Multiple other questions remain unanswered, such as how the oral flora evade systemic immune defenses as they possibly traverse the circulatory system to arrive at the uterine-placental interface. Deeper characterization of both microbial biomasses will determine whether the homology findings are a concrete observation or whether the placental microbes have originated from yet unknown habitats. Table 1 summarizes a few recent studies on placental, vaginal, and oral microbes that maybe associated with prematurity and which may influence the establishment of the fetal microbiome.

The existence of a placental microbiome has major implications for fetal colonization in utero and much earlier than has been suspected. The implications of prenatal microbial exposure are currently unknown, with outcomes ranging from immune tolerance to intestinal inflammation and uncontrolled systemic inflammatory response syndrome (36)(37) that may even contribute to PTB. (38) Evolving non–culture-based sequencing techniques have suggested the presence of bacteria in the amniotic fluid and in the absence of a history of rupture of membranes. This microbial colonization of the amniotic fluid may allow for prenatal development of a fetal microbiota and affect fetal outcomes. For example, Kacerovsky et al (39) evaluated microbial colonization of the amniotic cavity and found increased levels of fetal inflammatory cytokines in pregnancies complicated by premature rupture of membranes. Several studies from the Extremely Low Gestational Age Newborns study investigators revealed that fetoplacental inflammation is strongly associated with early PTB, suggesting that microorganisms colonizing the placenta might provoke distinctive newborn inflammatory responses, depending on the type of microbes.

Many potential sources for a prenatal microbiome exist. A recent study, which analyzed the microbiota of meconium samples obtained within 48 hours of birth, compared organisms identified in the meconium with previously published literature on maternal microbiomes from various sites. Ardissone and colleagues (38) found that amniotic fluid flora accounted for a greater relative abundance of bacteria found in meconium than either the oral or vaginal cavities of pregnant women. A previous study was also able to isolate Enterococcus, Streptococcus, Staphylococcus, and Propionibacterium from umbilical cord blood, suggesting a hematogenous mechanism by which microbiomes from distal locations in the mother can be transmitted to the fetus. (40)

In normal gut development, the interaction between Toll-like receptors and commensal bacteria help to downregulate inflammatory cascade. However, the premature infant has a gastrointestinal tract with highly immunoreactive intestinal epithelial cells and submucosa. The delayed colonization of the preterm infant’s gut by commensal bacteria results in upregulation of genes involved in inflammation and other proinflammatory molecules that could potentially contribute to preterm birth. (37) The prevalence of several bacteria associated with inflammatory responses in prematurely born infants (including Enterococcus and Enterobacter) was greater in infants born at 33 weeks or less compared with infants born at more than 33 weeks of estimated gestational age. (38) The preterm gut also experiences abnormal bacterial colonization with a decreased rate of diversification (41) and altered microbiome composition with an increased number of pathogenic bacteria. (42)

Beyond the interaction with PTB, this fundamental alteration in the development and qualitative characteristics of the preterm gut microbiota (43) results in an increased risk of necrotizing enterocolitis and other gastrointestinal, as well as immunologic issues later in life. (44)(45) One recent study observed that low intestinal microbial diversity at age 1 month was associated with IgE-associated eczema. Furthermore, the gastrointestinal tracts of adults and children with Crohn disease have altered microbial composition. Other studies have noted an association with late-onset sepsis. (44)(46) Tarr and colleagues matched bloodstream pathogens from episodes of late-onset sepsis to pathogens identified in stools of 7 of 11 children.

Not only do these diseases have significant rates of mortality, but the growth and neurodevelopmental outcomes of these children are also affected. (47)(48)(49) Prenatal and neonatal infections have been associated with a variety of adverse neurologic outcomes in the literature. Lee et al (50) examined the association between perinatal infection and brain structure (injury, size, and white matter tract development) and neurobehavioral outcomes in premature infants born at less than 30 weeks’ gestation. They found an association between both maternal genitourinary tract infections and histologic chorioamnionitis and changes in neurobehavioral outcomes at age 2 years (cognitive and language abilities for maternal infection and language abilities for histological chorioamnionitis). Although maternal infection was not associated with changes in the frequency of brain injury, brain size, or white matter development, histologic chorioamnionitis was additionally associated with changes in white matter development in the corpus callosum. On the other hand, neonatal infections were associated with brain structural changes but not neurobehavioral differences at age 2 years. Not only do these diseases have significant rates of mortality, but infections affect growth and neurodevelopmental outcomes of these children. Thus, systemic inflammation adds considerably to the increased risk.

The long-term implications of neonatal microbial environments for patient management are enormous but challenging to quantify during a short-term evaluation. It is essential to understand the microbial diversity and composition at various epidermal and mucosal sites in the human newborn body. We have now added the placenta as the home for microbes that result in adverse neurologic outcomes during infancy and childhood. In many ways, the observation that the placenta harbors a microbiome is not surprising because it is the first organ to aid in nourishing and maintaining the fetus. Yet microbes in our adult microbiome exceed our human genome by 11-fold. However, we must be mindful that progress has allowed a better understanding of the metabolic effect of the gut microbiome on all aspects of human biology: from obesity, to diabetes, to cardiovascular diseases, to autism. These findings have resulted in new treatment strategies to prevent disease in adults. Understanding who the placental microbes are and how they may or may not be shaped by those that reside in the maternal genital tract or oral mucosa during pregnancy is essential for determining maternal and fetal outcomes. We need to know when a pregnant woman is considered to have an infection with inflammation and potential dire consequences, such as a PTB, vs a maternal-fetal environment that is simply colonized with harmless microbes. Ongoing research will undoubtedly begin to provide foundations to develop new clinical management strategies for microbe-associated PTB.

BV

bacterial vaginosis

EVT

extravillous trophoblast

PTB

preterm birth