Contributions of the maternal oral and gut microbiome to placental microbial colonization in overweight and obese pregnant women

Abstract

A distinct bacterial signature of the placenta was reported, providing evidence that the fetus does not develop in a sterile environment. The oral microbiome was suggested as a possible source of the bacterial DNA present in the placenta based on similarities to the oral non-pregnant microbiome. Here, the possible origin of the placental microbiome was assessed, examining the gut, oral and placental microbiomes from the same pregnant women. Microbiome profiles from 37 overweight and obese pregnant women were examined by 16SrRNA sequencing. Fecal and oral contributions to the establishment of the placental microbiome were evaluated. Core phylotypes between body sites and metagenome predictive functionality were determined. The placental microbiome showed a higher resemblance and phylogenetic proximity with the pregnant oral microbiome. However, similarity decreased at lower taxonomic levels and microbiomes clustered based on tissue origin. Core genera: Prevotella, Streptococcus and Veillonella were shared between all body compartments. Pathways encoding tryptophan, fatty-acid metabolism and benzoate degradation were highly enriched specifically in the placenta. Findings demonstrate that the placental microbiome exhibits a higher resemblance with the pregnant oral microbiome. Both oral and gut microbiomes contribute to the microbial seeding of the placenta, suggesting that placental colonization may have multiple niche sources.

Full text

1
Scientific RepoRts | 7: 2860 | DOI:10.1038/s41598-017-03066-4
www.nature.com/scientificreports
Contributions of the maternal
oral and gut microbiome to
placental microbial colonization
in overweight and obese pregnant
women
Luisa F. Gomez-Arango1,2, Helen. L. Barrett
1,2,3, H. David McIntyre1,4, Leonie K. Callaway1,2,3,
Mark Morrison2, 5, 6 & Marloes Dekker Nitert
2,6
A distinct bacterial signature of the placenta was reported, providing evidence that the fetus does
not develop in a sterile environment. The oral microbiome was suggested as a possible source of the
bacterial DNA present in the placenta based on similarities to the oral non-pregnant microbiome.
Here, the possible origin of the placental microbiome was assessed, examining the gut, oral and
placental microbiomes from the same pregnant women. Microbiome proles from 37 overweight
and obese pregnant women were examined by 16SrRNA sequencing. Fecal and oral contributions
to the establishment of the placental microbiome were evaluated. Core phylotypes between body
sites and metagenome predictive functionality were determined. The placental microbiome showed
a higher resemblance and phylogenetic proximity with the pregnant oral microbiome. However,
similarity decreased at lower taxonomic levels and microbiomes clustered based on tissue origin.
Core genera: Prevotella, Streptococcus and Veillonella were shared between all body compartments.
Pathways encoding tryptophan, fatty-acid metabolism and benzoate degradation were highly enriched
specically in the placenta. Findings demonstrate that the placental microbiome exhibits a higher
resemblance with the pregnant oral microbiome. Both oral and gut microbiomes contribute to the
microbial seeding of the placenta, suggesting that placental colonization may have multiple niche
sources.
e establishment of the early microbiome in neonates inuences infant growth and immune function. Recently,
the maternal microbiome has been shown to prepare the newborn for host-microbial symbiosis, driving postnatal
innate immune development1. Perturbations of infant microbial colonization have been associated with increased
risk of asthma and obesity2–4. It has recently been suggested that intestinal microbial colonization may be initiated
in utero, possibly by distinct bacteria present in the placenta and amniotic uid5. Mounting evidence supports
the presence of bacterial DNA in the placenta, raising questions on the potential role of intrauterine bacteria in
placental function and fetal development.
Diverse hypotheses have been proposed for the mechanism of placental colonization: Vertical translocation
from the vagina6, or hematogenous spread from the gut7 and the oral cavity8. Hematogenous spread from the
oral cavity has received a wider attention due to the known association of periodontal disease with preterm
birth9. e largest metagenomic study characterizing placental microbial communities demonstrated that the
placental microbiome appears to be dierent from those of other body sites, albeit with some similarities to
the oral non-pregnant microbiome10. However, the composition of the placental microbiome has not yet been
1Faculty of Medicine, The University of Queensland, Brisbane, Australia. 2UQ Centre for Clinical Research, The
University of Queensland, Brisbane, Australia. 3Obstetric Medicine, Royal Brisbane and Women’s Hospital, Brisbane,
Australia. 4Mater Research, The University of Queensland, Brisbane, Australia. 5Diamantina Institute, Faculty
of Medicine and Biomedical Sciences, The University of Queensland, Brisbane, Australia. 6School of Chemistry
and Molecular Biosciences, The University of Queensland, Brisbane, Australia. Correspondence and requests for
materials should be addressed to L.F.G. (email: luisa.gomezarango@uq.net.au)
Received: 8 February 2017
Accepted: 21 April 2017
Published: xx xx xxxx
OPEN

www.nature.com/scientificreports/
2
Scientific RepoRts | 7: 2860 | DOI:10.1038/s41598-017-03066-4
prospectively compared within the same pregnant woman, during the same pregnancy, and with samples of the
microbiomes at other body sites. is research is essential, to assess how the placental microbiome might be
initiated and develop.
e purpose of this study was to assess similarities of the placental microbiome with other maternal micro-
biomes within the same individual. To our knowledge, this is the rst study to address the origin of the placental
microbiome by examining oral, gut and placental samples from the same pregnant women. Microbiome proles
from these three dierent tissue compartments were collected from women enrolled in the SPRING cohort11.
Results
Study population. Maternal characteristics from the 37 overweight (n = 13) and obese mothers (n = 24)
included in this substudy are presented in Table1. By design, women who delivered preterm, developed gesta-
tional diabetes mellitus or preeclampsia were excluded from this substudy. All women were of Caucasian ethnic-
ity and the majority delivered vaginally.
Source-tracking and taxonomical analyses of the maternal gut, oral and placental microbiome.
To investigate the contribution of the maternal oral and gut microbiome to the establishment of the placen-
tal microbiome, 16S rRNA sequencing was performed in 98 samples from 37 pregnant women. e placental
microbiome is relatively limited in abundance when compared to the maternal oral and gut microbiomes. On
average, placental samples yielded less than a third of bacterial reads present in the maternal gut and oral microbi-
omes (Supplementary Figure4). To distinguish between placental samples and contamination introduced during
DNA extraction, purication and amplication, unsupervised ordination methods showed a separate clustering
between pooled negative control and placental samples (R = 0.995, p = 0.021) (Supplementary Figure5). Four
main bacterial phyla (Firmicutes, Bacteroidetes, Actinobacteria and Proteobacteria) were identied in all micro-
biomes (Supplementary Figure3). Firmicutes and Proteobacteria were highly abundant in all placental samples,
representing nearly 80% of total microbial abundance. In the Firmicutes phylum, Streptococcus, Lactobacillus
and Veillonella were the most predominant and Pseudomonas, Haemophilus and Acinetobacter dominated in the
Proteobacteria phylum. e possible origin of the placental microbiome was determined by SourceTracker anal-
yses at all taxonomic levels (Phylum, Class, Order, Family, Genus, and OTUs). At phylum level, each placenta
shared phyla with both the maternal oral and gut microbiome (Fig.1a). At lower taxonomic levels, the placental
prole showed higher resemblance with the maternal oral than gut microbiome, however the level of similarity
declined with each taxonomic level. e overlap between the placental and gut microbiomes rapidly decreases to
<10% from the Class level onward (Fig.1a).
Phylogenetic distances revealed a signicant proximity of the maternal oral with the placental microbiome
(weighted: placenta-oral: 0.14 (0.11–0.16) vs. placental-gut: 0.20 (0.18–0.24) and unweighted: placental-oral: 0.72
(0.62–0.79) vs. placental-gut: 0.89 (0.85–0.90) (Fig.1b). Despite these similarities between the placenta and the
maternal oral microbiome, the overall microbial composition at dierent taxonomic levels of the placental micro-
biome (visualized through PCoA plots and measured by the Anosim test) showed a unique and highly variable
phylogenetic clustering (p < 0.001) clearly separating the samples based on tissue origin (Fig.2). PCoA plots and
clustering signicance of all three body sites are included in Supplementary Figure6.
The placental core microbiome shares phylotypes with the maternal oral and gut microbiome.
e core microbiome of each compartment (as dened in the Materials and Methods) were compared. We
chose to focus on what we dene as a core microbiome, to provide a more conservative interpretation of the
Mother (n = 37)
Age (years) 34.5 (30.3–36.8)
BMI (kg/m2)31.5 (27.9–35.8)
Blood pressure (mmHg)
Systolic 108.0 (105.0–115.0)
Diastolic 69.0 (62.0–73.5)
Gestational age at delivery (wks) 39.4 (38.6–40.3)
Mode of delivery
Vag inal 54.1%
Cesarean 45.9%
Antibiotic at delivery
Yes 48.6%
No 51.4%
Gestational weight gain (kg) 8.3 (5.8–12.3)
Birth weight (g)*3589 (3210–3946)
Gender*
Male 67.7%
Female 32.3%
Table 1. Maternal clinical characteristics. Clinical characteristics of mother-baby dyads. All data is presented as
median with 25–75th interquartile range. *Data available from only 31 infants.

www.nature.com/scientificreports/
3
Scientific RepoRts | 7: 2860 | DOI:10.1038/s41598-017-03066-4
consequences of the interindividual variations we report here. e list of all phylotypes at both family and genus
level between the dierent sample types are listed in Supplementary Figure7 and Supplementary Table3. All
phylotypes assigned to the core placental microbiome were also present within the maternal oral and gut micro-
biomes. At the family- and genus-levels of classication, the placental microbiomes did possess phylotypes not
detected in the other microbiomes, but these phylotypes were not part of the core placental microbiome. e pro-
portion of families that were unique to the core microbiome of the oral cavity was 0.36 whereas it was 0.30 for the
gut samples. At genus level, 41% of all genera were unique for the oral and 43% of genera unique for the gut core
microbiome respectively. e core families: Actinomycetaceae, Micrococcaceae, Oxalobacteraceae, Neisseriaceae,
Pasteurellaceae and Pseudomonadaceae and genera: Actinomyces, Rothia, Haemophilus, Pseudomonas were shared
between the oral and the placental microbiomes. No families or genera were exclusively shared by the gut and the
Figure 1. Maternal oral and gut microbial inuences on the placental microbiome (a). Bayesian source-
tracking results for placental samples at dierent taxonomic levels. Proportions of the maternal oral (blue),
gut (yellow) and unknown source of environment (grey) on the placental microbiome. Placental samples
showed a greater degree of similarity with the maternal oral microbiome. (b) Boxplots showing distances from
unweighted and weighted Unifrac distances between the maternal oral (blue) and gut (yellow) with respect
to the placental microbiome and between the oral and gut (white) microbiomes (permutations = 999). Each
boxplot shows the median, lower and upper quartiles of the Unifrac distances. A lower Unifrac distance shows
a greater resemblance between the two microbial communities. Pair-wise comparison were done by Mann-
Whitney U tests and annotated as ****p < 0.0001 and NS: not signicant.

www.nature.com/scientificreports/
4
Scientific RepoRts | 7: 2860 | DOI:10.1038/s41598-017-03066-4
placental microbiome. All three microbiomes shared the families Prevotellaceae, Streptococcaceae, Veillonellaceae
and Enterobacteriaceae, representing 8.5% of all core families (Supplementary Figure7). e abundance of these
four families in all placental samples represented 33.3% of total bacterial abundance. Similarly, at genus level,
Streptococcus, Veillonella and Prevotella were shared between all body compartments (Fig.3a). Streptococcus had
the highest abundance (0.43 (0.18–0.62)) followed by Veillonella (0.07 (0.02–0.22)) and Prevotella (0.06 (0.01–
0.26)) (Fig.3b) in placental samples and their abundances diered signicantly (p < 0.001) across the three dif-
ferent body sites. A total of 38 OTUs belonging to Streptococcus, Veillonella and Prevotella were identied in all
body compartments (Supplementary Figure8). Nine OTUs were shared between the three body sites, however
only Veillonella dispar was taxonomically assigned to species-level.
To exclude that this eect was driven by cross contamination at delivery, for instance by fecal contamination
of placental tissue, dierences in the placental core microbiome between cesarean (n = 17) and vaginal delivered
Figure 2. Dierences in microbiome composition among the placental and maternal oral populations. PCoA
plots for placental (red) and maternal oral (blue) microbiome at dierent taxonomic levels. Dierences in
microbial composition between the placenta and oral samples were determined by Anosim statistic test. An
R value close to 1.0 indicates total dissimilarity between the two groups. Signicant dierences between the
placental and oral microbiome were reported at all taxa levels (p < 0.001).

www.nature.com/scientificreports/
5
Scientific RepoRts | 7: 2860 | DOI:10.1038/s41598-017-03066-4
(n = 20) placentas were explored. No phylogenetic signicant clustering was evident between cesarean vs. vaginal
deliveries at either family (R = 0.019, p = 0.259) or genus level (R = 0.016, p = 0.272) nor were dierences in the
number of dierent taxa (p > 0.5) and richness (p > 0.5) present (Supplementary Figure9). Moreover, none of the
taxa (at family and genus level) mentioned previously were found to have dierent abundance between vaginally
delivered and cesarean placentas. Detailed FDR values are listed in Supplementary Figure10. Remarkably, no dif-
ferences in family Lactobacillaceae (p = 0.20) and genus Lactobacillus (p = 0.32), which are known to be abundant
in the vaginal tract, were noted between cesarean and vaginally delivered placentas.
The placental microbiome has a distinct functional prole. Using PICRUSt, the predicted core
functions of the placental microbiome were compared to those of the core maternal oral and gut microbiome.
Pathways encoding for carbohydrate metabolism, bacterial structure and vitamin-related pathways were over-
represented in the oral and gut microbiome (Fig.4). In contrast, microbial communities in the placenta appear
to be enriched with genes related to tryptophan, fatty acid metabolism and benzoate degradation (Fig.4 and
Supplementary Figure11 (LEfSe analysis)).
Discussion
e purpose of this study was to investigate possible seeding sources for the bacterial DNA detected in the pla-
centa, which is currently unclear. Similarities between the oral non-pregnant and placental microbiome have been
previously suggested10. Distinct functional proles and microbial populations are detected in the placental micro-
biome. e present study is the rst to show that the placental bacterial prole is most similar to the pregnant oral
microbiome and less alike the maternal gut microbiome in the same individual. However, the core microbiomes
of the three body sites share some distinct taxa.
irty-seven term placentas of overweight and obese pregnancies not complicated by preterm birth, GDM or
preeclampsia were selected for this substudy, as these factors are associated with an altered placental microbiome
composition12–15. All placentas displayed the presence of low abundance of non-pathogenic bacteria. Comparable
results have previously been reported10, 12, 14. However, a recent study reported no dierences between placental
samples and contamination controls16. Based on placental samples obtained from 6 uncomplicated pregnancies,
only a small number of sequences was detected in placental samples and no clear clustering was observed between
samples and environmental and reagent controls. In contrast, the 37 placental samples in this study yielded on
average 4.4 times more sequences than the negative control and show a clear clustering that is distinct from
the pooled reagent and PCR negative controls. e dierences between the studies may be a result of the use of
dierent regions of the 16S rRNA gene. Lauder et al. used V1-V2 primers, which may have a predilection for
environmental and ecological niche contamination and thus biased their study towards detecting contaminants
rather than non-contaminants. Here, the V6–V8 region of the 16S rRNA gene was targeted, covering a slightly
larger amplicon size. However the main taxa detected in contamination controls by Lauder et al.16 including
Bradyrhizobiaceae, Methylobacterium, Comamonadaceae, Propionibacterium and Sphingobacterium were also
detected in the negative controls of this study. ese sequences may reect contamination of reagents in the kits
used for extraction and sequence generation.
Figure 3. Core shared and distinct genera between the maternal gut, oral and placental microbiome. (a) Core
microbiomes consisting of genera detected in >50% of samples from each body site were obtained and plotted
in a Venn diagram. ree shared genera: Prevotella, Streptococcus and Veillonella were present in all samples.
No unique core genera was detected in the placenta, all were shared between the two maternal microbiomes.
Detailed genera among the three body sites are listed in Supplementary Table3. (b) Relative abundances of
genera: Prevotella, Streptococcus and Veillonella in all three maternal compartments. Relative abundances among
body sites and genera were signicantly dierent (p < 0.001).

www.nature.com/scientificreports/
6
Scientific RepoRts | 7: 2860 | DOI:10.1038/s41598-017-03066-4
is study is unique in determining the possible origins of the placental microbiome by comparing it with the
oral and gut microbiomes at all taxonomic levels within the same individual. e placental microbiome shows the
greatest resemblance to the oral microbiome especially at high taxonomic levels (e.g. phylum, class, order) but the
overlap decreases signicantly at lower taxonomic levels. Aagaard et al.10 reported a higher similarity of the pla-
cental microbiome with the oral microbiome from unrelated non-pregnant subjects from the US human micro-
biome project (HMP), although comparisons were only reported at phylum level. Phylogenetic analyses showed
that the placental microbiome clusters independently from the maternal oral microbiome at all taxonomic levels
emphasizing a distinct placental microbial community. is could be a result of the tissue-specic environmental
conditions that enable the colonization by specic bacteria belonging to the same higher taxonomic category e.g.
dierent families belonging to the order or class of bacteria.
Due to the interindividual variability of bacterial communities, the core microbiome for all three environ-
ments was used to assess the potential source for the placental microbiome. Analyses at family and genus lev-
els show a greater share of phylotypes between the maternal oral and placenta microbiomes. Interestingly, four
families and three genera belonging to the same families are shared between all three maternal microbiomes:
Veillonella, Streptococcus and Prevotella, are commonly present in placenta, oral and gut microbiomes10, 17–19.
ese genera are frequently observed in oral and gut microbiome samples20, 21, especially in healthy subjects,
with all genera detected in both the oral and stool microbiomes of ~45% of the subjects examined as part of
the US Human Microbiome Project (HMP)22. It has also been proposed that the gastric cavity possesses a core
microbiome, comprised of phylotypes assigned to Prevotella, Streptococcus, Veillonella, Rothia and Haemophilus23.
Specic oral microorganisms can seed distal sites below the stomach, which may explain the abundance in the
pregnant women’s digestive tract and the selective enrichment of these genera in the placenta. ese genera were
also found in most of the placental samples examined in this study. In pregnancy, the gut barrier has increased
leakiness24 and pregnant women have increased prevalence of bleeding in the oral cavity25. ese changes are
associated with the hormonal and cardiovascular changes of pregnancy. It is therefore possible for Prevotella,
Streptococcus and Veillonella to enter the bloodstream from the oral cavity and/or the gut lumen to colonize the
placenta. In addition, these genera are successful colonizers and biolm producers26–28, facilitating their ability to
thrive within many dierent body sites.
In a mouse model of periodontal infections, injection of bacteria-including Veillonella, Streptococcus and
Prevotella spp—obtained from human saliva and subgingival plaques-- into the tail vein causes placental trans-
location of these bacteria8. However, not all bacteria present in the pooled oral samples were translocated to the
murine placenta, suggesting once more that the nutritional and physiological ecologies of the placenta imposes
selective pressures aecting microbial colonization and persistence. In that context, we identied Veillonella dis-
par as a common species between the three body sites. is species has been found in neonates, is associated with
an inherent maternal gut microbiome and persists at least throughout the rst year of life29. For growth, it utilizes
short-chain organic acids, particularly lactate30. Lactate is produced by the uteroplacental tissues, where it is used
for fetal energy production31. us, the placenta may be an optimal niche for V. d is pa r and could possibly have
a cooperative metabolic system with the placenta. Recent studies have also revealed by co-culturing of isolated
intestinal Streptococcus and Veillonella strains that these strains have combined immunomodulatory properties
Figure 4. Distinct predictive metabolic prole in the placental microbiome. Heat map demonstrating the
predictive functional proling of microbial communities in the placenta, gut and oral samples, using 16s rRNA
gene sequences. Stronger intensity of red indicates higher pathway activity and blue lower activity. Signicant
microbial functional pathways determined by the LEfSe algorithm are displayed in this heat map (LDA
score > 3.0). Bolded pathways were signicantly enriched in the placenta in comparison to the maternal oral
and gut microbiome.

www.nature.com/scientificreports/
7
Scientific RepoRts | 7: 2860 | DOI:10.1038/s41598-017-03066-4
that dier from those of the individual strains32. e presence of these genera in the placenta may therefore trig-
ger immunomodulatory mechanisms that may inuence placental immune homeostasis.
Even though key placental phylotypes are shared with the maternal oral and gut microbiomes, there is evi-
dence for a unique placental microbiome. e results of this study suggest that the placental microbiome does not
have only one specic and unique source of colonization but that selective bacteria translocated from dierent
maternal microbiomes could contribute to the assemblage of the placental microbiome. Even though murine
models have well documented the transmission of oral microorganisms in the placenta8, 9, 33, 34, increased gut bac-
terial translocation during pregnancy and lactation has been also been reported7. Maternal gut microorganisms
are transported to mesenteric lymph nodes and mammary gland by mononuclear cells during late pregnancy and
lactation in mice7. Moreover, specically labeled intestinal bacteria have been recovered from murine placentas35.
e results of this study indicate that bacteria from the maternal oral cavity and gut could employ specic trans-
location mechanisms to colonize the placenta.
Interestingly, the bacteria that make up the placental microbiome are predicted to have distinctive functional
capacities relative to the other microbiomes. While the role of oral and gut microbes is related to carbohydrate and
amino acid metabolism, as well as vitamin biosynthesis36, 37, the placental microbiomes are predicted to be enriched
with genes regulating tryptophan, fatty acid metabolism and benzoate degradation. Placental tryptophan metab-
olism is important for neurodevelopment in the fetus and perturbations of placental tryptophan metabolism have
been associated with altered neurodevelopmental processes in the fetus38. Catabolism of tryptophan in the placenta
is linked with the establishment and maintenance of the feto-maternal immune tolerance39, placental circulation
and growth40 and modulation of antimicrobial activity41 by inhibiting ascending infections from the vagina39. is
overrepresented pathway in placental microbes may indicate a selective mechanism of natural placental colonizers
to prevent colonization by foreign microorganisms. Veillonella species are equipped with the beta subunit respon-
sible for the synthesis of L-tryptophan, possibly explaining the enrichment of tryptophan metabolism pathway
in our dataset. Pathways encoding fatty acid metabolism were also enriched in placental bacteria. e placental
microbiota may aid in ecient extraction of energy from circulating fatty acids and play a crucial role in supplying
energy-yielding substrates to the fetus. Moreover, bacterial genes mapping to pathways involved in benzoate degra-
dation were also enriched. Aromatic compounds such as benzoate are used as a carbon source for many microor-
ganisms. Benzoate is commonly used as food preservative in many products including in carbonated beverages42.
It crosses the placenta and elevated serum concentrations of benzoate have been associated with neurologic distur-
bances43. We also emphasize that our considerations here are based on a predictive analysis, and await conrmation
using shotgun metagenome sequencing or other more functional approaches, coupled perhaps with dietary analysis.
e absence of vaginal swabs is a limitation of this study and restricts the analysis of the placental micro-
bial colonization pattern only to oral and gut microbes. In European ethnic groups, the vaginal microbiome is
enriched in Lactobacillus spp.; and while members of this genus can be readily found in the gut microbiome none
were unique for the placental core microbiome. is might indicate that if the vaginal microbiome contributes
to the seeding of the placental microbiome, it is not with bacteria that are unique to the vaginal microbiome. To
that end, Veillonella spp. were essentially absent from the HMP vaginal samples19, 22, suggesting that the origin
of these phylotypes in our placental samples were unlikely to have originated from the vagina. Furthermore,
analysis of the placental microbiome of vaginal deliveries compared with those delivered by Cesarean section,
showed no signicant dierences on microbial diversity and taxa. Taken together, we conclude our results are true
representations of the placental microbiome and are not aected by contaminations of the placenta by the vaginal
microbiome or maternal feces during the delivery process. e sequencing strategy was based on the 16S rRNA
gene amplication restraining the identication of some taxa at species level. is explains the reason why most
microbial associations were based on genus level and metagenomic pathways were only inferred. In addition, the
disappearance of overlap between the placenta and other maternal sites from Class level and below might be due
to primer sets used and depth of sequencing. e phyla and taxa reported in this study parallels that of Aagaard
et al.10 using whole genome shotgun sequencing but are fairly dierent from that of Collado et al.5. Dierent
regions of the 16S rRNA gene can return dierent results and therefore studies might not be directly comparable.
e use of whole genome sequencing provides better resolution and reduces amplicon bias. is study cannot
provide evidence whether the bacteria detected in all placental samples are alive and metabolically active. In
addition, appropriate environmental controls would be benecial to exclude all potential contamination in the
present placental samples. Moreover, as the gut microbiome undergoes structural changes throughout pregnancy
becoming less diverse in each individual but more variable between individuals, fecal samples obtained during
the third trimester may provide further insights into the shared placental-gut phylotypes. e use of core micro-
biomes including only those bacteria present in ≥50% of participants reduces the risk for over interpretation of
consequences of interindividual variation. And since it is not yet clear when in pregnancy the placental microbi-
ome is seeded, comparisons with the gut microbiome in early second trimester may even be more appropriate.
Conclusion
is study provides evidence that both the maternal oral and gut microbiome may contribute to the seeding of
the placental microbiome. e placental microbial communities show a higher similarity with the bacteria found
in the maternal oral microbiome than the gut microbiome especially at higher taxonomic levels. e presence of
shared phylotypes between all three compartments suggests that placental bacteria sequences may have multiple
niche sources. Given that the gut microbiome undergoes structural changes throughout pregnancy44 becoming
less diverse in each individual but more variable between individuals, stool samples obtained during the third
trimester may provide further insights into shared placental-gut phylotypes. However, since it is not yet clear when
in pregnancy the placental microbiome is seeded, comparisons with the gut microbiome in early second trimester
are important and may even be more appropriate. Further studies into the signicance of intrauterine microbial
populations and dynamics could identify their role in feto-placental communication, fetal development and health.

www.nature.com/scientificreports/
8
Scientific RepoRts | 7: 2860 | DOI:10.1038/s41598-017-03066-4
Materials and Methods
Ethics, consent and permission. This study was approved by the Human Research Ethics commit-
tees of the Royal Brisbane and Women’s Hospital (HREC/RBWH/11/467) and e University of Queensland
(2012000080). Written informed consent was obtained from all participants prior to enrolment in the trial. All
experiments were performed in accordance with relevant guidelines and regulations.
Study subjects and sample collection. Overweight and obese pregnant women included in this study
were participants in the double-blind randomized controlled trial: SPRING (Study of Probiotics IN the preven-
tion of Gestational diabetes) (ANZCTR 12611001208998)11. In total 37 women were selected for having a com-
plete matched sample set of placenta (n = 37) and maternal oral (n = 37) with 24 mothers also having maternal
fecal samples available. Detailed clinical characteristics of these women are presented in Table1. Maternal feces,
oral swabs and placental samples were collected at separate times antepartum and postpartum: Maternal fecal
samples were self-collected at 16 weeks gestation, refrigerated and stored at −80 °C within 24 hours of collection.
At 36 weeks gestation, maternal oral swabs were collected by sterile dry swab (Copan Diagnostics, Murrieta,
CA) and stored immediately at −20 °C prior to transfer to −80 °C. Term placentas were collected by clinical
practitioners. Within 1 hour of delivery each placenta was transported to the laboratory for processing. Trained
researchers were provided with sterile supplies and stringent instructions to excise cuboidal 1 cm3 sections from
the fetal side of the placenta. Excisions were placed in autoclaved 2 mL tubes and immediately placed in liquid
nitrogen and stored at −80 °C until processing.
Sample processing and microbiome sequencing. Genomic DNA was isolated from maternal feces,
placental tissue and oral swabs. A total of 0.25 grams of stool sample was extracted using the repeated bead beat-
ing and column (RBB + C) method followed by the Qiagen AllPrep DNA extraction kit as previously detailed45, 46.
DNA extraction and recovery from oral and placental tissue (10 mg) was achieved by the automated Maxwell
16 system following mechanical disruption using sterile zirconia beads (0.1 and 0.05 mm diameter) in 300 μL
lysis buer (NaCl 0.5 mol/L, Tris–HCl 50 mmol/L, pH 8.0, EDTA 50 mmol/L and SDS 4% w/v). e Maxwell
16 Buccal Swab LEV DNA Purication kit and Maxwell 16 Tissue DNA Purication kit (Promega, Madison,
WI, USA) were used following the manufacturer’s recommendations. Negative controls consisting of extraction
reagents and PCR amplication were added to each set of fecal, oral and placental samples and included in the
sequencing reaction. Puried DNA was quantied by the Nanodrop ND 1000 spectrophotometer (Nanodrop
Technologies). Oral, fecal and placental DNA extractions were performed separately to avoid cross contamination
between samples.
A barcoded primer set based on universal primers 926F (5′-TCG TCG GCA GCG TCA GAT GTG TAT AAG
AGA CAG AAA CTY AAA KGA ATT GRC GG -3′) and 1392R (5′ -GTC TCG TGG GCT CGG AGA TGT GTA
TAA GAG ACA GAC GGG CGG TGW GTR C-3′) was used to amplify 500 bps of the hypervariable V6–V8
region of the 16S rRNA gene. Specicity and amplicon size were veried by gel electrophoresis. PCR products
were cleaned with AMPure XP beads. Amplicons were barcoded using the Nextera XT Index Kit set A and set B.
Amplicons were puried with the Promega Wizard Gel Extraction kit, followed by a second AMPure XP beads
cleaning to reduce potential contamination with human DNA. Fecal, oral and placental tagged-amplicons were
quantied, normalized and pooled. e pooled libraries were sequenced using the Illumina MiSeq platform and
workows established at the University of Queensland’s Australian Center for Ecogenomics (www.ecogenomic.
org). Sequences were deposited in the NCBI database (SUB 2559729). Sequences were joined, demultiplexed and
quality ltered using QIIME (Quantitative Insights Into Microbial Ecology v 1.9.1)47. An open reference OTU
picking method using 97% identity to the Greengenes 13_8 database48 was selected. OTUs with a relative fre-
quency below 0.01 were removed. Resultant data demonstrated a total median sequenced reads of 44259.5 (IQR:
38477–53917), 76707 (IQR: 58486–125071) and 60478 (IQR: 13965–121911) for fecal, oral and placental samples
respectively. A total of 34, 38 and 26 dierent genera were identied in fecal, oral and placental samples respec-
tively. e alpha diversity curve for all the sample types determined by Chao 1 is provided in Supplementary
Figure1.To evaluate the potential impact of contamination in each set of samples, a set of reagent controls, to
which no additional tissue or DNA was added were included. A DNA extraction control and PCR amplication
control were processed in an identical manner to the rest of the samples, starting from the lysis step. From each
set of samples, the DNA extraction control and the PCR amplication control were pooled together and sent for
sequencing. Quality ltering and OTUs detected in the negative controls were deleted from the generated OTU
tables (Supplementary Table1). Pooled placental negative controls yielded 13840 sequenced reads. e OTUs
detected in negative controls are listed in Supplementary Table2, representing 13 dierent genera. To examine if
the sequence reads in the negative control were distinct from those reported in placental samples, unsupervised
ordination methods (PCoA, PCA and NMDS) were used to identify placental and pooled negative control clus-
tering. Due to the dierent origin of the samples, OTU tables at each taxonomy level (phylum, class, order, family,
genus and OTUs) were normalized to relative abundance using the cumulative sum scaling (CSS) normalization
method aer deletion of the sequences present in the negative samples49. Relative abundances at phylum level
aer and before normalization are reported in Supplementary Figures2 and 3.
Microbial diversity and statistical analysis. To investigate the seeding role of the maternal oral and gut
microbiome to the placental microbiome, bacterial source tracking analyses were performed using SourceTracker
v 1.150. Oral and gut samples were designated as sources and the placental sample of the corresponding mother
was selected as sink. Alpha and beta diversity were calculated on normalized OTUs tables. Alpha diversity was
measured by the Chao 1 and Shannon indices, representing the number, richness and distribution of taxa. Beta
diversity was calculated using both phylogenetic (Unifrac distance) and non-phylogenetic (Bray-Curtis) distances
matrices and visualized through PCoA. Anosim testing was used to conrm signicant dierences in microbial

www.nature.com/scientificreports/
9
Scientific RepoRts | 7: 2860 | DOI:10.1038/s41598-017-03066-4
community composition. Microbial diversity analyses were performed within QIIME and with the Calypso so-
ware tool (http://bioinfo.qimr.edu.au/calypso/). Core microbiomes consisting of OTUs detected in 50% of sam-
ples from each dataset were obtained. Placental, gut and oral phylotypes at genus and family level were plotted
using Venny 2.151. Clinical metadata is presented as median with interquartile range in all instances (Table1).
In silico metagenomics using PICRUSt. Functionality of the dierent metagenomes, grouped by oral, gut
and placenta were predicted using the soware PICRUSt 1.1.052. is tool predicts the functional composition
from the 16S rRNA gene data based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) Orthologs clas-
sication. Signicant dierences in microbial functional pathways were tested by the LEfSe (Linear Discriminant
Analysis Eect Size) algorithm (http://huttenhower.sph.harvard.edu)53.
Data availability statement. Sequences were deposited in the NCBI database (SUB 2559729).
References
1. Gomez de Aguero, M. et al. e maternal microbiota drives early postnatal innate immune development. Science (New Yor, N.Y.)
351, 1296–1302, doi:10.1126/science.aad2571 (2016).
2. Fujimura, . E. et al. Neonatal gut microbiota associates with childhood multisensitized atopy and T cell dierentiation. 22,
1187–1191, doi:10.1038/nm.4176 (2016).
3. Ajslev, T. A., Andersen, C. S., Gamborg, M., Sorensen, T. I. & Jess, T. Childhood overweight aer establishment of the gut microbiota:
the role of delivery mode, pre-pregnancy weight and early administration of antibiotics. International journal of obesity (2005) 35,
522–529, doi:10.1038/ijo.2011.27 (2011).
4. Cox, L. M. et al. Altering the intestinal microbiota during a critical developmental window has lasting metabolic consequences. Cell
158, 705–721, doi:10.1016/j.cell.2014.05.052 (2014).
5. Collado, M. C., autava, S., Aao, J., Isolauri, E. & Salminen, S. Human gut colonisation may be initiated in utero by distinct
microbial communities in the placenta and amniotic uid. Scientic reports 6, 23129, doi:10.1038/srep23129 (2016).
6. Goldenberg, . L., Culhane, J. F., Iams, J. D. & omero, . Epidemiology and causes of preterm birth. Lancet 371, 75–84, doi:10.1016/
s0140-6736(08)60074-4 (2008).
7. Perez, P. F. et al. Bacterial imprinting of the neonatal immune system: lessons from maternal cells? Pediatrics 119, e724–732,
doi:10.1542/peds.2006-1649 (2007).
8. Fardini, Y., Chung, P., Dumm, ., Joshi, N. & Han, Y. W. Transmission of diverse oral bacteria to murine placenta: evidence for the
oral microbiome as a potential source of intrauterine infection. Infection and immunity 78, 1789–1796, doi:10.1128/iai.01395-09
(2010).
9. Han, Y. W. et al. Fusobacterium nucleatum induces premature and term stillbirths in pregnant mice: implication of oral bacteria in
preterm birth. Infection and immunity 72, 2272–2279 (2004).
10. Aagaard, . et al. The placenta harbors a unique microbiome. Science translational medicine 6, 237ra265, doi:10.1126/
scitranslmed.3008599 (2014).
11. Nitert, M. D. et al. SPING: an CT study of probiotics in the prevention of gestational diabetes mellitus in overweight and obese
women. BMC pregnancy and childbirth 13, 50, doi:10.1186/1471-2393-13-50 (2013).
12. Prince, A. L. et al. The placental microbiome is altered among subjects with spontaneous preterm birth with and without
chorioamnionitis. American journal of obstetrics and gynecology, doi:10.1016/j.ajog.2016.01.193 (2016).
13. Amaraseara, ., Jayaseara, . W., Senanayae, H. & Dissanayae, V. H. Microbiome of the placenta in pre-eclampsia supports the
role of bacteria in the multifactorial cause of pre-eclampsia. The journal of obstetrics and gynaecology research 41, 662–669,
doi:10.1111/jog.12619 (2015).
14. Antony, . M. et al. e preterm placental microbiome varies in association with excess maternal gestational weight gain. American
journal of obstetrics and gynecology 212, 653.e651–616, doi:10.1016/j.ajog.2014.12.041 (2015).
15. Bassol s , J. et al. Gestational Diabetes is associated with changes in Placental Microbiota and Microbiome. Pediatric research,
doi:10.1038/pr.2016.155 (2016).
16. Lauder, A. P. et al. Comparison of placenta samples with contamination controls does not provide evidence for a distinct placenta
microbiota. Microbiome 4, 29, doi:10.1186/s40168-016-0172-3 (2016).
17. Nasidze, I., Li, J., Quinque, D., Tang, . & Stoneing, M. Global diversity in the human salivary microbiome. Genome research 19,
636–643, doi:10.1101/gr.084616.108 (2009).
18. Arumugam, M. et al. Enterotypes of the human gut microbiome. Nature 473, 174–180, doi:10.1038/nature09944 (2011).
19. Consortium, H. M. P. Structure, function and diversity of the healthy human microbiome. Nature 486, 207–214, doi:10.1038/
nature11234 (2012).
20. Wang, M., Ahrne, S., Jeppsson, B. & Molin, G. Comparison of bacterial diversity along the human intestinal tract by direct cloning
and sequencing of 16S rNA genes. FEMS microbiology ecology 54, 219–231, doi:10.1016/j.femsec.2005.03.012 (2005).
21. Hayashi, H., Taahashi, ., Nishi, T., Saamoto, M. & Benno, Y. Molecular analysis of jejunal, ileal, caecal and recto-sigmoidal
human colonic microbiota using 16S rNA gene libraries and terminal restriction fragment length polymorphism. Journal of
medical microbiology 54, 1093–1101, doi:10.1099/jmm.0.45935-0 (2005).
22. Segata, N. et al. Composition of the adult digestive tract bacterial microbiome based on seven mouth surfaces, tonsils, throat and
stool samples. Genome biology 13, 42, doi:10.1186/gb-2012-13-6-r42 (2012).
23. Nardone, G. & Compare, D. e human gastric microbiota: Is it time to rethin the pathogenesis of stomach diseases? United
European gastroenterology journal 3, 255–260, doi:10.1177/2050640614566846 (2015).
24. err, C. A. et al. Early life events inuence whole-of-life metabolic health via gut microora and gut permeability. Critical reviews in
microbiology 41, 326–340, doi:10.3109/1040841x.2013.837863 (2015).
25. Boutigny, H. et al. Oral Infections and Pregnancy: nowledge of Gynecologists/Obstetricians, Midwives and Dentists. Oral health
& preventive dentistry 14, 41–47, doi:10.3290/j.ohpd.a34376 (2016).
26. Mashima, I. & Naazawa, F. e interaction between Streptococcus spp. and Veillonella tobetsuensis in the early stages of oral
biolm formation. Journal of bacteriology. doi:10.1128/jb.02512-14 (2015).
27. Nobbs, A. H., Lamont, . J. & Jeninson, H. F. Streptococcus adherence and colonization. Microbiology and molecular biology
reviews: MMB 73, 407–450, Table of Contents, doi:10.1128/mmbr.00014-09 (2009).
28. Ouda, T. . E., awana, T., Saito, A., Ouda, . & Ishihara, . Synergy in biolm formation between Fusobacterium nucleatum and
Prevotella species. Anaerobe 18, 110–1116 (2012).
29. Bached, F. et al. Dynamics and Stabilization of the Human Gut Microbiome during the First Year of Life. Cell host & microbe 17,
690–703, doi:10.1016/j.chom.2015.04.004 (2015).
30. Do, T., Sheehy, E. C., Mulli, T., Hughes, F. & Beighton, D. Transcriptomic analysis of three Veillonella spp. present in carious dentine
and in the saliva of caries-free individuals. Frontiers in cellular and infection microbiology 5, 25, doi:10.3389/fcimb.2015.00025
(2015).

www.nature.com/scientificreports/
10
Scientific RepoRts | 7: 2860 | DOI:10.1038/s41598-017-03066-4
31. Vaughan, O. ., Davies, . L., Ward, J. W., de Blasio, M. J. & Fowden, A. L. A physiological increase in maternal cortisol alters
uteroplacental metabolism in the pregnant ewe. e Journal of physiology 594, 6407–6418, doi:10.1113/jp272301 (2016).
32. van den Bogert, B., Meijerin, M., Zoetendal, E. G., Wells, J. M. & leerebezem, M. Immunomodulatory properties of Streptococcus
and Veillonella isolates from the human small intestine microbiota. PLoS One 9, e114277, doi:10.1371/journal.pone.0114277 (2014).
33. Han, Y. W. et al. Term stillbirth caused by oral Fusobacterium nucleatum. Obstetrics and gynecology 115, 442–445, doi:10.1097/
AOG.0b013e3181cb9955 (2010).
34. Han, Y. W. et al. Transmission of an uncultivated Bergeyella strain from the oral cavity to amniotic uid in a case of preterm birth.
Journal of clinical microbiology 44, 1475–1483, doi:10.1128/jcm.44.4.1475-1483.2006 (2006).
35. Jimenez, E. et al. Is meconium from healthy newborns actually sterile? esearch in microbiology 159, 187–193, doi:10.1016/j.
resmic.2007.12.007 (2008).
36. Qin, J. et al. A human gut microbial gene catalogue established by metagenomic sequencing. Nature 464, 59–65, doi:10.1038/
nature08821 (2010).
37. Belda-Ferre, P. et al. e oral metagenome in health and disease. e ISME journal 6, 46–56, doi:10.1038/ismej.2011.85 (2012).
38. Nic Goeden, J. C. V. a. A. B. Placental tryptophan metabolism as a potential novel pathway for the developmental origins of mental
diseaseS. Translational Developmental Psychiatry 1 (2013).
39. Sedlmayr, P., Blaschitz, A. & Stocer, . e role of placental tryptophan catabolism. Frontiers in immunology 5, 230, doi:10.3389/
mmu.2014.00230 (2014).
40. Blaschitz, A. et al. Vascular endothelial expression of indoleamine 2,3-dioxygenase 1 forms a positive gradient towards the feto-
maternal interface. PLoS One 6, e21774, doi:10.1371/journal.pone.0021774 (2011).
41. Sedlmayr, P. et al. Localization of indoleamine 2,3-dioxygenase in human female reproductive organs and the placenta. Molecular
human reproduction 8, 385–391 (2002).
42. Davis, B. A., aubertas, . F., Pearson, S. . & Bowen, W. H. e eects of benzoate and uoride on dental caries in intact and
desalivated rats. Caries rese arch 35, 331–337, doi:47471 (2001).
43. Aschenbrenner, D. S. J. V. Drug therapy in nursing. Lippincott Williams & Wilins, 1306 (2009).
44. oren, O. et al. Host emodeling of the Gut Microbiome and Metabolic Changes during Pregnancy. Cell 150, 470–480, doi:10.1016/j.
cell.2012.07.008 (2012).
45. Yu, Z. & Morrison, M. Improved extraction of PC-quality community DNA from digesta and fecal samples. BioTechniques 36,
808–812 (2004).
46. Gomez-Arango, L. F. et al. Connections Between the Gut Microbiome and Metabolic Hormones in Early Pregnancy in Overweight
and Obese Women. 65, 2214–2223, doi:10.2337/db16-0278 (2016).
47. Caporaso, J. G. et al. QIIME allows analysis of high-throughput community sequencing data. Nature methods 7, 335–336,
doi:10.1038/nmeth.f.303 (2010).
48. DeSantis, T. Z. et al. Greengenes, a chimera-checed 16S rNA gene database and worbench compatible with AB. Applied and
environmental microbiology 72, 5069–5072, doi:10.1128/aem.03006-05 (2006).
49. Paulson, J. N., Stine, O. C., Bravo, H. C. & Pop, M. Dierential abundance analysis for microbial marer-gene surveys. Nature
methods 10, 1200–1202, doi:10.1038/nmeth.2658 (2013).
50. nights, D. et al. Bayesian community-wide culture-independent microbial source tracing. Nature methods 8, 761–763,
doi:10.1038/nmeth.1650 (2011).
51. Oliveros, J. C. An interactive tool for comparing lists with Venn’s diagrams. (2015).
52. Langille, M. G. et al. Predictive functional proling of microbial communities using 16S rNA marer gene sequences. Nature
biotechnology 31, 814–821, doi:10.1038/nbt.2676 (2013).
53. Segata, N. et al. Metagenomic biomarer discovery and explanation. Genome biology 12, 60, doi:10.1186/gb-2011-12-6-r60 (2011).
Acknowledgements
e authors would like to sincerely thank the SPRING participants and the SPRING trial group members Anne
Tremellen, Katie Foxcro, Dr Jacinta Tobin, Dr Shelley Wilkinson, Dr Chris McSweeney, Prof Peter O’Rourke
and Dr Barbara Lingwood. Special thanks to Dr Erin Shanahan, Dr Naoki Fukuma for expert technical advice
and Nicola Angel from the Australian Centre for Ecogenomics. e SPRING study is supported by the National
Health & Medical Research Council (grant no. 1028575) and the Royal Brisbane and Women’s Hospital
Foundation.
Author Contributions
L.F.G.-A. provided intellectual input, researched data, performed analysis and wrote and edited the manuscript.
H.L.B., H.D.M. and L.K.C. provided intellectual input and edited the manuscript. M.M. provided intellectual
input and technical expertise and edited the manuscript. M.D.N. provided intellectual input, designed the
research and wrote and edited the manuscript.
Additional Information
Supplementary information accompanies this paper at doi:10.1038/s41598-017-03066-4
Competing Interests: e SPRING study has received probiotics and placebo products from Chr. Hansen A/S.
e authors have no further conicts of interest.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre-
ative Commons license, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not per-
mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© e Author(s) 2017