Milk microbiome transplantation: recolonizing donor milk with mother's own milk microbiota

Donor human milk (DHM) provides myriad nutritional and immunological benefits for preterm and low birthweight infants. However, pasteurization leaves DHM devoid of potentially beneficial milk microbiota. In the present study, we performed milk microbiome transplantation from freshly collected mother's own milk (MOM) into pasteurized DHM. Small volumes of MOM (5%, 10%, or 30% v/v) were inoculated into pasteurized DHM and incubated at 37 °C for up to 8 h. Further, we compared microbiome recolonization in UV-C-treated and Holder-pasteurized DHM, as UV-C treatment has been shown to conserve important biochemical components of DHM that are lost during Holder pasteurization. Bacterial culture and viability-coupled metataxonomic sequencing were employed to assess the effectiveness of milk microbiome transplantation. Growth of transplanted MOM bacteria occurred rapidly in recolonized DHM samples; however, a greater level of growth was observed in Holder-pasteurized DHM compared to UV-C-treated DHM, potentially due to the conserved antimicrobial properties in UV-C-treated DHM. Viability-coupled metataxonomic analysis demonstrated similarity between recolonized DHM samples and fresh MOM samples, suggesting that the milk microbiome can be successfully transplanted into pasteurized DHM. These results highlight the potential of MOM microbiota transplantation to restore the microbial composition of UV-C-treated and Holder-pasteurized DHM and enhance the nutritional and immunological benefits of DHM for preterm and vulnerable infants. KEY POINTS: • Mother's own milk microbiome can be successfully transplanted into donor human milk. • Recolonization is equally successful in UV-C-treated and Holder-pasteurized milk. • Recolonization time should be restricted due to rapid bacterial growth.

Maternal dietary intervention during lactation impacts the maternal faecal and human milk microbiota

AIMS

To determine the effect of a two-week reduced fat and sugar and increased fibre maternal dietary intervention on the maternal faecal and human milk (HM) microbiomes.

METHODS AND RESULTS

Faecal swabs and HM samples were collected from mothers (n = 11) immediately pre-intervention, immediately post-intervention, and 4- and 8-weeks post-intervention, and were analysed using full-length 16S rRNA gene sequencing. Maternal macronutrient intake was assessed at baseline and during the intervention. Maternal fat and sugar intake during the intervention were significantly lower than pre-intervention (P=<0.001, 0.005, respectively). Significant changes in the bacterial composition of maternal faeces were detected after the dietary intervention, with decreases in the relative abundance of Bacteroides caccae (P=<0.001) and increases in the relative abundance of Faecalibacillus intestinalis (P = 0.006). In HM, the diet resulted in a significant increase in Cutibacterium acnes (P = 0.001) and a decrease in Haemophilus parainfluenzae (P=<0.001). The effect of the diet continued after the intervention, with faecal swabs and HM samples taken 4- and 8-weeks after the diet showing significant differences compared to baseline.

CONCLUSION

This pilot study demonstrates that short-term changes in maternal diet during lactation can alter the bacterial composition of the maternal faeces and HM. Clinical trial registration: Australian New Zealand Clinical Trials Registry (ACTRN12619000606189). Website: https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=377188&isReview=true.

Effects of Different Thawing and Warming Processes on Human Milk Composition

The composition of human milk is influenced by storage and processing practices. The effects of thawing and warming practices on human milk composition remain poorly studied despite their prevalence in home, research, and donor milk bank settings. This review comprehensively examines the impact of different thawing and warming methods on nutritional and bioactive human milk components. While some components such as carbohydrates and minerals remain stable under most typical thawing and warming conditions, others, such as fat, immune proteins, bacterial and human cells, and peptide amine hormones, are sensitive to warming. This review has identified that the data on the effects of milk thawing and warming is limited and often contradictory. Given that numerous important components of milk are diminished during cold storage, it is important that thawing and warming practices do not lead to further loss of or alterations to beneficial milk components. Further work in this field will facilitate greater standardization of thawing methods among researchers and underpin recommendations for thawing and warming of expressed milk for parents.

From hype to hope: Considerations in conducting robust microbiome science

Microbiome science has been one of the most exciting and rapidly evolving research fields in the past two decades. Breakthroughs in technologies including DNA sequencing have meant that the trillions of microbes (particularly bacteria) inhabiting human biological niches (particularly the gut) can be profiled and analysed in exquisite detail. This microbiome profiling has profound impacts across many fields of research, especially biomedical science, with implications for how we understand and ultimately treat a wide range of human disorders. However, like many great scientific frontiers in human history, the pioneering nature of microbiome research comes with a multitude of challenges and potential pitfalls. These include the reproducibility and robustness of microbiome science, especially in its applications to human health outcomes. In this article, we address the enormous promise of microbiome science and its many challenges, proposing constructive solutions to enhance the reproducibility and robustness of research in this nascent field. The optimisation of microbiome science spans research design, implementation and analysis, and we discuss specific aspects such as the importance of ecological principals and functionality, challenges with microbiome-modulating therapies and the consideration of confounding, alternative options for microbiome sequencing, and the potential of machine learning and computational science to advance the field. The power of microbiome science promises to revolutionise our understanding of many diseases and provide new approaches to prevention, early diagnosis, and treatment.

Impact of breastfeeding and other early-life factors on the development of the oral microbiome

The oral cavity is home to the second most diverse microbiome in the human body. This community contributes to both oral and systemic health. Acquisition and development of the oral microbiome is a dynamic process that occurs over early life; however, data regarding longitudinal assembly of the infant oral microbiome is scarce. While numerous factors have been associated with the composition of the infant oral microbiome, early feeding practices (breastfeeding and the introduction of solids) appear to be the strongest determinants of the infant oral microbiome. In the present review, we draw together data on the maternal, infant, and environmental factors linked to the composition of the infant oral microbiome, with a focus on early nutrition. Given evidence that breastfeeding powerfully shapes the infant oral microbiome, the review explores potential mechanisms through which human milk components, including microbes, metabolites, oligosaccharides, and antimicrobial proteins, may interact with and shape the infant oral microbiome. Infancy is a unique period for the oral microbiome. By enhancing our understanding of oral microbiome assembly in early life, we may better support both oral and systemic health throughout the lifespan.

Characterisation of Mid-Gestation Amniotic Fluid Cytokine and Bacterial DNA Profiles in Relation to Pregnancy Outcome in a Small Australian Cohort

A well-established association exists between intrauterine bacteria and preterm birth. This study aimed to explore this further through documenting bacterial and cytokine profiles in Australian mid-gestation amniotic fluid samples from preterm and term births. Samples were collected during amniocenteses. DNA was extracted and the full-length 16S rRNA gene was amplified and sequenced. Levels of the cytokines IL-1β, IL-6, IL-10, TNF-α and MCP-1 were determined using the Milliplex MAGPIX system. Bacterial DNA profiles were low in diversity and richness, with no significant differences observed between term and preterm samples. No differences in the relative abundance of individual OTUs between samples were identified. IL-1β and TNF-α levels were significantly higher in samples containing reads mapping to Sphingomonas sp.; however, this result should be interpreted with caution as similar reads were also identified in extraction controls. IL-6 levels were significantly increased in samples with reads that mapped to Pelomonas sp., whilst TNF-α levels were elevated in fluid samples from pregnancies that subsequently delivered preterm. Bacterial DNA unlikely to have originated from extraction controls was identified in 20/31 (64.5%) mid-gestation amniotic fluid samples. Bacterial DNA profiles, however, were not predictive of preterm birth, and although cytokine levels were elevated in the presence of certain genera, the biological relevance of this remains unknown.

Effect of Holder pasteurisation and UV-C irradiation on bacteriophage titres in human milk

Human milk is the optimal nutrition source for infants and contains a complex mix of bioactive compounds and microorganisms. When unavailable, pasteurised donor milk may be provided, particularly to preterm infants. Holder pasteurisation (HP) is typically implemented in human milk banks to prevent pathogen transmission. Given the impact of heat on milk bioactives, ultraviolet-C irradiation (UV-C) is an alternative being explored and has demonstrated effective bactericidal activity. In addition to bacteria, milk contains viruses including primarily bacteriophages (phages) and which likely influence the developing bacterial microbiome of infants. However, the effect of pasteurisation on human milk phages is unknown. This study assessed the effect of HP and UV-C on titres of exogenous bacteriophages inoculated into human milk. Ten donor human milk samples were tested in parallel with water controls. Milk samples or water controls were inoculated to a final concentration of 1 × 104 PFU/mL (±1 log) each of a thermotolerant Escherichia coli phage (T4) and a thermosensitive Staphylococcus aureus phage (BYJ20) and subjected to HP and UV-C treatments. UV-C inactivated both phages within milk and water controls, however, HP was ineffective against the thermotolerant T4 phages. Initial data suggests that UV-C treatment may eliminate phage with potential to affect preterm infant gut colonization. Further studies should extend this to other phages.

Human Milk Lipids and Small Metabolites: Maternal and Microbial Origins

Although there has been limited application in the field to date, human milk omics research continues to gain traction. Human milk lipidomics and metabolomics research is particularly important, given the significance of milk lipids and metabolites for infant health. For researchers conducting compositional milk analyses, it is important to consider the origins of these compounds. The current review aims to provide a summary of the existing evidence on the sources of human milk lipids and small metabolites. Here, we describe five major sources of milk lipids and metabolites: de novo synthesis from mammary cells, production by the milk microbiota, dietary consumption, release from non-mammary tissue, and production by the gut microbiota. We synthesize the literature to provide evidence and understanding of these pathways in the context of mammary gland biology. We recommend future research focus areas to elucidate milk lipid and small metabolite synthesis and transport pathways. Better understanding of the origins of human milk lipids and metabolites is important to improve translation of milk omics research, particularly regarding the modulation of these important milk components to improve infant health outcomes.

Characterisation of human milk bacterial DNA profiles in a small cohort of Australian women in relation to infant and maternal factors

Human milk is composed of complex microbial and non-microbial components that shape the infant gut microbiome. Although several maternal and infant factors have been associated with human milk microbiota, no study has investigated this in an Australian population. Therefore, we aimed to investigate associations between human milk bacterial composition of Australian women and maternal factors (body mass index (BMI), mode of delivery, breast pump use, allergy, parity) and infant factors (sex, mode of feeding, pacifier use, and introduction of solids). Full-length 16S rRNA gene sequencing was used to characterise milk bacterial DNA profiles. Milk from mothers with a normal BMI had a higher relative abundance of Streptococcus australis than that of underweight mothers, while milk from overweight mothers had a higher relative abundance of Streptococcus salivarius compared with underweight and obese mothers. Mothers who delivered vaginally had a higher relative abundance of Streptococcus mitis in their milk compared to those who delivered via emergency caesarean section. Milk of mothers who used a breast pump had a higher relative abundance of Staphylococcus epidermidis and Streptococcus parasanguinis. Milk of mothers whose infants used a pacifier had a higher relative abundance of S. australis and Streptococcus gwangjuense. Maternal BMI, mode of delivery, breast pump use, and infant pacifier use are associated with the bacterial composition of human milk in an Australian cohort. The data from this pilot study suggests that both mother and infant can contribute to the human milk microbiome.

Methodological approaches for studying the human milk microbiome

Human milk contains a low-biomass, low-diversity microbiome, consisting largely of bacteria. This community is of great research interest in the context of infant health and maternal and mammary health. However, this sample type presents many unique methodological challenges. In particular, there are numerous technical considerations relating to sample collection and storage, DNA extraction and sequencing, viability, and contamination. Failure to properly address these challenges may lead to distortion of bacterial DNA profiles generated from human milk samples, ultimately leading to spurious conclusions. Here, these technical challenges are discussed, and various methodological approaches used to address them are analyzed. Data were collected from studies in which a breadth of methodological approaches were used, and recommendations for robust and reproducible analysis of the human milk microbiome are proposed. Such methods will ensure high-quality data are produced in this field, ultimately supporting better research outcomes for mothers and infants.

Effect of a reduced fat and sugar maternal dietary intervention during lactation on the infant gut microbiome

Objective

A growing body of literature has shown that maternal diet during pregnancy is associated with infant gut bacterial composition. However, whether maternal diet during lactation affects the exclusively breastfed infant gut microbiome remains understudied. This study sets out to determine whether a two-week of a reduced fat and sugar maternal dietary intervention during lactation is associated with changes in the infant gut microbiome composition and function.

Design

Stool samples were collected from four female and six male (n = 10) infants immediately before and after the intervention. Maternal baseline diet from healthy mothers aged 22–37 was assessed using 24-h dietary recall. During the 2-week dietary intervention, mothers were provided with meals and their dietary intake was calculated using FoodWorks 10 Software. Shotgun metagenomic sequencing was used to characterize the infant gut microbiome composition and function.

Results

In all but one participant, maternal fat and sugar intake during the intervention were significantly lower than at baseline. The functional capacity of the infant gut microbiome was significantly altered by the intervention, with increased levels of genes associated with 28 bacterial metabolic pathways involved in biosynthesis of vitamins (p = 0.003), amino acids (p = 0.005), carbohydrates (p = 0.01), and fatty acids and lipids (p = 0.01). Although the dietary intervention did not affect the bacterial composition of the infant gut microbiome, relative difference in maternal fiber intake was positively associated with increased abundance of genes involved in biosynthesis of storage compounds (p = 0.016), such as cyanophycin. Relative difference in maternal protein intake was negatively associated with Veillonella parvula (p = 0.006), while positively associated with Klebsiella michiganensis (p = 0.047). Relative difference in maternal sugar intake was positively associated with Lactobacillus paracasei (p = 0.022). Relative difference in maternal fat intake was positively associated with genes involved in the biosynthesis of storage compounds (p = 0.015), fatty acid and lipid (p = 0.039), and metabolic regulator (p = 0.038) metabolic pathways.

Conclusion

This pilot study demonstrates that a short-term maternal dietary intervention during lactation can significantly alter the functional potential, but not bacterial taxonomy, of the breastfed infant gut microbiome. While the overall diet itself was not able to change the composition of the infant gut microbiome, changes in intakes of maternal protein and sugar during lactation were correlated with changes in the relative abundances of certain bacterial species.

Clinical trial registration: Australian New Zealand Clinical Trials Registry (ACTRN12619000606189).

Environmental determinants of human milk composition in relation to health outcomes

Humans are exposed to environmental factors at every stage of life including infancy. The aim of this mini-review was to present a narrative of environmental factors influencing human milk composition. Current literature shows lactation is a dynamic process and is responsive to multiple environmental challenges including geographical location, lifestyle, persistent pollutants and maternal factors (ethnicity, diet, stress, allergy and adiposity) that may influence human milk composition in a synergistic manner and should be considered in order to improve infant and maternal outcomes on a populations scale. Further interventional studies on larger international cohorts are needed to elucidate these complex relationships. Lactating women should aim for a healthy lifestyle and maintain a healthy body composition prior to and throughout the reproductive period, including during lactation.

Effect of Cold Storage on the Viable and Total Bacterial Populations in Human Milk

Expression and cold storage of human milk is a common practice. Current guidelines for cold storage of expressed milk do not take into account the impact on the milk microbiome. Here, we investigated the impact of cold storage on viable bacterial populations in human milk. Freshly expressed milk samples (n = 10) were collected and analysed immediately, stored at 4 °C for four days, −20 °C for 2.25 months and 6 months, and −80 °C for 6 months. Samples were analysed using propidium monoazide (PMA; a cell viability dye) coupled with full-length 16S rRNA gene. An aliquot of each sample was additionally analysed without PMA to assess the impact of cold storage on the total DNA profile of human milk. Cold storage significantly altered the composition of both the viable microbiome and total bacterial DNA profile, with differences in the relative abundance of several OTUs observed across each storage condition. However, cold storage did not affect the richness nor diversity of the samples (PERMANOVA all p > 0.2). Storage of human milk under typical and recommended conditions results in alterations to the profile of viable bacteria, with potential implications for infant gut colonisation and infant health.

The Viable Microbiome of Human Milk Differs from the Metataxonomic Profile

Bacteria in human milk contribute to the establishment of the infant gut microbiome. As such, numerous studies have characterized the human milk microbiome using DNA sequencing technologies, particularly 16S rRNA gene sequencing. However, such methods are not able to differentiate between DNA from viable and non-viable bacteria. The extent to which bacterial DNA detected in human milk represents living, biologically active cells is therefore unclear. Here, we characterized both the viable bacterial content and the total bacterial DNA content (derived from viable and non-viable cells) of fresh human milk (n = 10). In order to differentiate the living from the dead, a combination of propidium monoazide (PMA) and full-length 16S rRNA gene sequencing was used. Our results demonstrate that the majority of OTUs recovered from fresh human milk samples (67.3%) reflected DNA from non-viable organisms. PMA-treated samples differed significantly in their bacterial composition compared to untreated samples (PERMANOVA p < 0.0001). Additionally, an OTU mapping to Cutibacterium acnes had a significantly higher relative abundance in PMA-treated (viable) samples. These results demonstrate that the total bacterial DNA content of human milk is not representative of the viable human milk microbiome. Our findings raise questions about the validity of conclusions drawn from previous studies in which viability testing was not used, and have broad implications for the design of future work in this field.

Human Milk Lactose, Insulin, and Glucose Relative to Infant Body Composition during Exclusive Breastfeeding

Human milk (HM) components may influence infant growth and development. This study aimed to investigate relationships between infant body composition (BC) and HM lactose, insulin, and glucose (concentrations and calculated daily intakes (CDI)) as well as 24-h milk intake and maternal BC at 3 months postpartum. HM samples were collected at 2 months postpartum. Infant and maternal BC was assessed with bioimpedance spectroscopy. Statistical analysis used linear regression accounting for infant birth weight. 24-h milk intake and CDI of lactose were positively associated with infant anthropometry, lean body mass and adiposity. Higher maternal BC measures were associated with lower infant anthropometry, z-scores, lean body mass, and adiposity. Maternal characteristics including BC and age were associated with concentrations and CDI of HM components, and 24-h milk intake. In conclusion, 24-h intake of HM and lactose as well as maternal adiposity are related to development of infant BC.

The human milk microbiome: who, what, when, where, why, and how?

Human milk (HM) contains an incredible array of microorganisms. These likely contribute to the seeding of the infant gastrointestinal microbiome, thereby influencing infant immune and metabolic development and later-life health. Given the importance of the HM microbiota in this context, there has been an increase in research efforts to characterize this in different populations and in relation to different maternal and infant characteristics. However, despite a decade of intensive research, there remain several unanswered questions in this field. In this review, the "5 W+H" approach (who, what, when, where, why, and how) is used to comprehensively describe the composition, function, and origin of the HM microbiome. Here, existing evidence will be drawn together and critically appraised to highlight avenues for further research, both basic and applied. Perhaps the most interesting of these is the potential to modulate the HM microbiome using pre/probiotics or dietary interventions. Another exciting possibility is the personalization of donor milk for women with insufficient supply. By gaining a deeper understanding of the HM microbiome, opportunities to intervene to optimize infant and lifelong health may be identified.

Author Correction: Centrifugation does not remove bacteria from the fat fraction of human milk

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Can we modulate the breastfed infant gut microbiota through maternal diet?

Initial colonisation of the infant gut is robustly influenced by regular ingestion of human milk, a substance that contains microbes, microbial metabolites, immune proteins, and oligosaccharides. Numerous factors have been identified as potential determinants of the human milk and infant gut microbiota, including maternal diet; however, there is limited data on the influence of maternal diet during lactation on either of these. Here, we review the processes thought to contribute to human milk and infant gut bacterial colonisation and provide a basis for considering the role of maternal dietary patterns during lactation in shaping infant gut microbial composition and function. Although only one observational study has directly investigated the influence of maternal diet during lactation on the infant gut microbiome, data from animal studies suggests that modulation of the maternal gut microbiota, via diet or probiotics, may influence the mammary or milk microbiota. Additionally, evidence from human studies suggests that the maternal diet during pregnancy may affect the gut microbiota of the breastfed infant. Together, there is a plausible hypothesis that maternal diet during lactation may influence the infant gut microbiota. If substantiated in further studies, this may present a potential window of opportunity for modulating the infant gut microbiome in early life.

Impact of expression mode and timing of sample collection, relative to milk ejection, on human milk bacterial DNA profiles

AIM

To investigate the impact of expression mode: electric breast pump or hand expression, and timing of sample collection: pre- and post-milk ejection on human milk (HM) bacterial DNA profiles.

METHODS AND RESULTS

Three HM samples from the same breast were collected from 30 breastfeeding mothers: a pre-milk ejection pump-expressed sample (pre-pump), a post-milk ejection pump-expressed sample (post-pump) and a post-milk ejection hand-expressed sample (post-hand). Full-length 16S rRNA gene sequencing was used to assess milk bacterial DNA profiles. Bacterial profiles did not differ significantly based on mode of expression nor timing of sample collection. No significant differences were detected in the relative abundance of any OTUs based on expression condition (pre-pump/ post-pump and post-pump/post-hand) with univariate linear mixed-effects regression analyses (all P-values > 0·01; α = 0·01). Similarly, no difference in richness was observed between sample types (number of observed OTUs: post-pump/post-hand P = 0·13; pre-pump/post-pump P = 0. 45).

CONCLUSION

Bacterial DNA profiles of HM did not differ according to either expression method or timing of sample collection.

SIGNIFICANCE AND IMPACT OF THE STUDY

Hand or pump expression can be utilized to collect samples for microbiome studies. This has implications for the design of future HM microbiome studies.

DNA extraction method influences human milk bacterial profiles

AIMS

To evaluate four DNA extraction methods to elucidate the most effective method for bacterial DNA recovery from human milk (HM).

METHODS AND RESULTS

Human milk DNA was extracted using the following methods: (i) Qiagen MagAttract Microbial DNA Isolation Kit (kit QM), (ii) Norgen Milk Bacterial DNA Isolation Kit (kit NM), (iii) Qiagen MagAttract Microbiome DNA/RNA Isolation Kit (kit MM) and (iv) TRIzol LS Reagent (method LS). The full-length 16S rRNA gene was sequenced. Kits MM and method LS were unable to extract detectable levels of DNA in 9/11 samples. Detectable levels of DNA were recovered from all samples using kits NM (mean = 0·68 ng μl^-1 ) and QM (mean = 0·55 ng μl^-1 ). For kits NM and QM, the greatest number of reads were associated with Staphylococcus epidermidis, Streptococcus vestibularis, Propionibacterium acnes, Veillonella dispar and Rothia mucilaginosa. Contamination profiles varied substantially between kits, with one bacterial species detected in negative extraction controls generated with kit QM and six with kit NM.

CONCLUSIONS

Kit QM is the most suitable of the kits tested for the extraction of bacterial DNA from human milk.

SIGNIFICANCE AND IMPACT OF THE STUDY

Choice of extraction method impacts the efficiency of bacterial DNA extraction from human milk and the resultant bacterial community profiles generated from these samples.

Human Milk From Atopic Mothers Has Lower Levels of Short Chain Fatty Acids

Short chain fatty acids (SFCAs) are microbial metabolites produced in the gut upon fermentation of dietary fiber. These metabolites interact with the host immune system and can elicit epigenetic effects. There is evidence to suggest that SCFAs may play a role in the developmental programming of immune disorders and obesity, though evidence in humans remains sparse. Here we have quantified human milk (HM) SCFA levels in an international cohort of atopic and non-atopic mothers (n = 109). Our results demonstrate that human milk contains detectable levels of the SCFAs acetate, butyrate, and formate. Samples from atopic mothers had significantly lower concentrations of acetate and butyrate than those of non-atopic mothers. HM SCFA levels in atopic and non-atopic women also varied based on maternal country of residence (Australia, Japan, Norway, South Africa, USA). Reduced exposure to HM SCFA in early life may program atopy or overweight risk in breastfed infants.

Infection-mediated preterm birth: Bacterial origins and avenues for intervention

Preterm birth (PTB) is globally the leading cause of death and disability in children under five years of age. Intra-amniotic infection is well recognised as a major cause of PTB. Importantly, it is the most common cause of extreme PTB (birth prior to 28 weeks gestation), which is frequently associated with a wide range of serious neonatal morbidities. Recent developments in next generation sequencing technologies, combined with many years of culture-based microbiological data have allowed us to gain a deeper understanding of the pathogenesis of infection-mediated PTB. In particular, studies have revealed numerous potential routes to intra-amniotic infection beyond the classically described ascending vaginal route. Currently, antibiotic therapy is standard treatment for suspected or confirmed intra-amniotic infection, although its use in this context has had mixed success due to problems ranging from inappropriate antibiotic selection in relation to the target organism/s, to poor transplacental drug passage. In this review, we will draw together evidence from animal models and human studies to characterise pathways to intra-amniotic infection. We will then thoroughly outline current therapeutic protocols for cases of intra-amniotic infection and suggest potential new avenues for treatment.

Placental and intra-amniotic inflammation are associated with altered fetal immune responses at birth

INTRODUCTION

High-grade placental inflammation is associated with preterm birth and poor neonatal outcomes. Recent reports suggest that low-grade placental inflammation is common in uncomplicated pregnancies. The relationship between placental inflammation and innate immune anti-microbial responses is unknown. In this study we sought to identify any association between placental inflammation and fetal immune responses.

METHODS

Cord blood samples collected from late preterm and full-term Caesarean section deliveries (n = 44) were exposed to various immune challenges (resiquimod, LPS, PGN, poly (I:C), cGAMP, and 5'ppp-dsRNA) and production of inflammatory mediators (G-CSF, IFN-γ, IL-1β, IL-6, IL-8, IL-10, and TNF-α) was measured by multiplex assay. Hospital histology reports were used to assess the extent of inflammation in the placenta.

RESULTS

Almost half (47.7%) of placentae examined here showed histological evidence of inflammation. Resiquimod, LPS, and PGN elicited strong inflammatory responses in neonatal cord blood, while poly (I:C), cGAMP, and 5'ppp-dsRNA elicited weaker responses. Fetuses with evidence of chorioamnionitis and fetal inflammatory reaction in their placentae had significantly increased immune responses to cGAMP and 5'ppp-dsRNA (ligands for STING and RIG-I, respectively) and significantly decreased immune responses to poly (I:C) (a TLR3 agonist). Interestingly, STING, RIG-I, and TLR3 are all involved in viral response pathways, suggesting that fetuses exposed to chorioamnionitis or fetal inflammatory reaction might respond differently to viruses postnatally.

CONCLUSION

Our data suggest that low-level placental inflammation is associated with altered innate cytokine responses at birth.

The Not-so-Sterile Womb: Evidence That the Human Fetus Is Exposed to Bacteria Prior to Birth

The human microbiome includes trillions of bacteria, many of which play a vital role in host physiology. Numerous studies have now detected bacterial DNA in first-pass meconium and amniotic fluid samples, suggesting that the human microbiome may commence in utero. However, these data have remained contentious due to underlying contamination issues. Here, we have used a previously described method for reducing contamination in microbiome workflows to determine if there is a fetal bacterial microbiome beyond the level of background contamination. We recruited 50 women undergoing non-emergency cesarean section deliveries with no evidence of intra-uterine infection and collected first-pass meconium and amniotic fluid samples. Full-length 16S rRNA gene sequencing was performed using PacBio SMRT cell technology, to allow high resolution profiling of the fetal gut and amniotic fluid bacterial microbiomes. Levels of inflammatory cytokines were measured in amniotic fluid, and levels of immunomodulatory short chain fatty acids (SCFAs) were quantified in meconium. All meconium samples and most amniotic fluid samples (36/43) contained bacterial DNA. The meconium microbiome was dominated by reads that mapped to Pelomonas puraquae. Aside from this species, the meconium microbiome was remarkably heterogeneous between patients. The amniotic fluid microbiome was more diverse and contained mainly reads that mapped to typical skin commensals, including Propionibacterium acnes and Staphylococcus spp. All meconium samples contained acetate and propionate, at ratios similar to those previously reported in infants. P. puraquae reads were inversely correlated with meconium propionate levels. Amniotic fluid cytokine levels were associated with the amniotic fluid microbiome. Our results demonstrate that bacterial DNA and SCFAs are present in utero, and have the potential to influence the developing fetal immune system.

Characterization of the bacterial microbiome in first-pass meconium using propidium monoazide (PMA) to exclude nonviable bacterial DNA

Numerous studies have reported bacterial DNA in first-pass meconium samples, suggesting that the human gut microbiome is seeded prior to birth. However, these studies have not been able to discriminate between DNA from living bacterial cells, DNA from dead bacterial cells or cell-free DNA. Here we have used propidium monoazide (PMA) together with 16S rRNA gene sequencing to determine whether there are intact bacterial cells in the fetal gut. DNA was extracted from first-pass meconium (n = 5) and subjected to 16S rRNA gene sequencing with/without PMA treatment. All meconium samples, regardless of PMA treatment, contained detectable levels of bacterial DNA; however, treatment with PMA prior to DNA extraction decreased the DNA yield by approximately 20%. PMA-treated meconium samples did not differ significantly from untreated samples in terms of observed number of OTUs (P = 0·945); although they did differ taxonomically, with around one quarter of OTUs identified in untreated samples only, suggesting that they have originated from cell-free/nonviable DNA. The mean Sørensen coefficient for treated vs untreated samples was 0·527. Our findings suggest that the fetal gut is seeded with intact bacterial cells prior to birth. This is an important finding, as exposure to live bacteria during gestation might have a significant impact on the developing fetus. SIGNIFICANCE AND IMPACT OF THE STUDY: DNA-based microbiome studies performed using 16S rRNA gene sequencing are limited by their inability to discriminate between live bacterial cells, dead bacterial cells and cell-free DNA. Here we use propidium monoazide (PMA) to exclude nonviable bacteria from microbiome analysis of first-pass meconium samples and thereby reveal that the majority of the purported fetal gut microbiome is from intact bacterial cells. This work demonstrates the importance of excluding nonviable bacteria when analysing the microbial community in low-biomass samples such as meconium.

Re: "Amniotic fluid from healthy term pregnancies does not harbor a detectable microbial community" (2018) 6:87, https://doi.org/10.1186/s40168-018-0475-7

A recent publication by Lim et al. 2018 (Amniotic fluid from healthy term pregnancies does not harbor a detectable microbial community) strongly concluded that the microbiome of amniotic fluid primarily originates from reagent contamination. However, upon closer inspection of the methods used and data presented in this study, in particular the supplementary data, such conclusions do not appear to be supported by the results. We outline such methodological/data interpretation concerns and invite the authors to discuss these.

Identification and removal of contaminating microbial DNA from PCR reagents: impact on low-biomass microbiome analyses

Reagent-derived contamination can compromise the integrity of microbiome data, particularly in low microbial biomass samples. This contamination has recently been attributed to the 'kitome' (contamination introduced by the DNA extraction kit), prior to which attention was mostly paid to potential contamination introduced by PCR reagents. In this study, we assessed the proportion to which our DNA extraction kit and PCR master mix introduce contaminating microbial DNA to bacterial microbial profiles generated by 16S rRNA gene sequencing. Utilizing a commercial dsDNase treatment protocol to decontaminate the PCR master mix, we demonstrated that the vast majority of contaminating DNA was derived from the PCR master mix. Importantly, this contamination was almost completely eliminated using the simple dsDNase treatment, resulting in a 99% reduction in contaminating bacterial reads. We suggest that dsDNase treatment of PCR reagents should be explored as a simple and effective way of reducing contamination in low-biomass microbiome studies and producing more robust and reliable data. SIGNIFICANCE AND IMPACT OF THE STUDY: Reagent contamination with microbial DNA is a major problem in microbiome studies of low microbial biomass samples. Levels of such contaminating DNA often outweigh what is present in the sample and heavily confound subsequent data analysis. Previous studies have suggested this contamination is primarily derived from DNA extraction kits. Here, we identified the PCR master mix as the primary source of contamination, and showed that enzymatic removal of the contamination drastically reduced the blank signal and improved precision. Decontamination of PCR master mixes may have the potential to improve the sensitivity and accuracy of low-biomass microbiome studies.

A Critical Review of the Bacterial Baptism Hypothesis and the Impact of Cesarean Delivery on the Infant Microbiome

Numerous studies suggest that infants delivered by cesarean section are at a greater risk of non-communicable diseases than their vaginal counterparts. In particular, epidemiological studies have linked Cesarean delivery with increased rates of asthma, allergies, autoimmune disorders, and obesity. Mode of delivery has also been associated with differences in the infant microbiome. It has been suggested that these differences are attributable to the "bacterial baptism" of vaginal birth, which is bypassed in cesarean deliveries, and that the abnormal establishment of the early-life microbiome is the mediator of later-life adverse outcomes observed in cesarean delivered infants. This has led to the increasingly popular practice of "vaginal seeding": the iatrogenic transfer of vaginal microbiota to the neonate to promote establishment of a "normal" infant microbiome. In this review, we summarize and critically appraise the current evidence for a causal association between Cesarean delivery and neonatal dysbiosis. We suggest that, while Cesarean delivery is certainly associated with alterations in the infant microbiome, the lack of exposure to vaginal microbiota is unlikely to be a major contributing factor. Instead, it is likely that indication for Cesarean delivery, intrapartum antibiotic administration, absence of labor, differences in breastfeeding behaviors, maternal obesity, and gestational age are major drivers of the Cesarean delivery microbial phenotype. We, therefore, call into question the rationale for "vaginal seeding" and support calls for the halting of this practice until robust evidence of need, efficacy, and safety is available.

Comparison of Meconium DNA Extraction Methods for Use in Microbiome Studies

The establishment of human gut microbiota commences initially in utero. Meconium—the first fecal material passed after birth—can be used to study fetal gut contents; however, processing meconium samples for microbiome studies presents significant technical challenges. Meconium hosts a low biomass microbiome, is tar-like in texture and contains high concentrations of PCR inhibitors. This study aimed to evaluate four different DNA extraction methods to elucidate the most effective method for bacterial DNA recovery and sequencing analysis from first-pass meconium. Samples from five infants were collected and processed using the following extraction kits: (1) Qiagen QIAamp DNA Stool Mini (QS); (2) Qiagen QIAamp DNA Microbiome (QM); (3) MoBio PowerSoil (PS); (4) MoBio MagAttract PowerMicrobiome (PM). Additionally, Kit PM was employed with a double inhibitor removal treatment (IRT) step (PM2). Bacterial DNA recovery was assessed by qPCR. Any PCR inhibition in samples was measured by spiking DNA eluates with 0.1 ng of pure Streptococcus agalactiae (GBS) DNA followed by qPCR quantitation. Kit PM yielded the highest average total DNA yield (79.3 ng per gram of meconium). Samples extracted with kit PS had the highest detectable levels of 16S rRNA gene by qPCR. The ability of each kit to overcome PCR inhibition varied, with qPCR on GBS-spiked DNA from kits QS, QM, PS, and PM recovering 87.1, 91.0, 88.8, and 37.9% GBS DNA, respectively. Double IRT improved the performance of kit PM, increasing GBS recovery to 56.5%. However, once DNA yield was normalized to the level recovered with the other kits 100% of GBS DNA was detected, suggesting that levels of PCR inhibitors are related to DNA yield from kit PM. Ion Torrent 16S rRNA gene sequencing revealed a high level of inter-kit variation in meconium microbiome structure. In particular, kit QM showed a bias toward extracting Firmicute DNA, while the other kits extracted primarily Proteobacterial DNA. Choice of extraction kit greatly impacts on the ability to extract and detect bacterial DNA in meconium and on the microbiome community structure generated from these samples.

Planting the seed: Origins, composition, and postnatal health significance of the fetal gastrointestinal microbiota

It has long been assumed that establishment of the fetal microbiome commences with the birthing process. However, recent studies have found bacterial DNA in umbilical cord blood, placenta, amniotic fluid, meconium, and fetal membranes in healthy normal pregnancies, leading to suggestions that the seeding of the fetal microbiome may commence in utero long before delivery. The origins of the microbiota of the fetal gastrointestinal (GI) tract have not yet been conclusively determined, although bacterial translocation from the maternal circulation, or ascension from the vagina, are both likely to be contributing pathways. Mother-to-child efflux of bacteria during pregnancy has the potential to markedly influence postnatal health, as the composition of gut microbiota determines production of important metabolites which are absorbed systemically and which modify immune function and development. Hence, the importance of understanding the colonization of the fetal GI microbiome is becoming clear, although few studies have investigated the origins, dynamics, and timing of the fetal microbiome. This is the topic of this review. By gaining a deeper understanding of the mechanisms underpinning fetal microbiome seeding, strategies may be developed to optimize fetal immune development and reduce the risk of adverse health and developmental outcomes.

Preclinical evaluation of drugs to block inflammation-driven preterm birth

Intrauterine inflammation, the major cause of early preterm birth, can have microbial and sterile aetiologies. We assessed in a Transwell model the anti-inflammatory efficacies of five drugs on human extraplacental membranes delivered after preterm spontaneous labour (30-34 wk). Drugs [TPCA1 (IKKβ inhibitor), 5 z-7-oxozeaenol (OxZ, TAK1 inhibitor), inhibitor of NF-κB essential modulator binding domain (iNBD), SB239063 (p38 MAPK inhibitor) and N-acetyl cysteine (free radical scavenger free radicals)] were added after 12 h equilibration to the amniotic compartment. Concentrations of IL-6, TNF-α, MCP-1, IL-1β and PGE₂ in the media, and IL6, TNFA and PTGS2 mRNA expression levels in membranes, were determined after 12 h. Data were analysed using mixed models analyses. Thirteen of the 28 membranes had histological chorioamnionitis (HCA⁺); five were positive for bacterial culture and six for fetal inflammatory reaction. Baseline PGE₂ and cytokine production was similar between HCA^- and HCA⁺ membranes. Anti-inflammatory effects were also similar between HCA^- and HCA⁺ membranes. TPCA1 and OxZ were the most effective drugs; each inhibited amniotic secretion of 4/5 pro-inflammatory mediators and mRNA levels of 2/3, regardless of stimulus. We conclude that treatment with TPCA1 or OxZ, in combination with antibiotics, may minimise the adverse effects of intrauterine inflammation in pregnancy.

Effects of cytokine-suppressive anti-inflammatory drugs on inflammatory activation in ex vivo human and ovine fetal membranes

Intrauterine infection and inflammation are responsible for the majority of early (<32 weeks) spontaneous preterm births (PTBs). Anti-inflammatory agents, delivered intra-amniotically together with antibiotics, may be an effective strategy for preventing PTB. In this study, the effects of four cytokine-suppressive anti-inflammatory drugs (CSAIDs:N-acetyl cysteine (NAC), SB239063, TPCA-1 and NEMO binding domain inhibitor (NBDI)) were assessed on human and ovine gestational membrane inflammation. Full-thickness membranes were collected from healthy, term, human placentas delivered by Caesarean section (n=5). Using a Transwell model, they were stimulatedex vivowith γ-irradiation-killedEscherichia coliapplied to the amniotic face. Membranes from near-term, ovine placentas were stimulatedin uterowith lipopolysaccharide,Ureaplasma parvumor saline control and subjected to explant culture. The effects of treatment with CSAIDs or vehicle (1% DMSO) on accumulation of PGE2and cytokines (human interleukin 6 (IL6), IL10 and TNFα; ovine IL8 (oIL8)) were assessed in conditioned media at various time points (3–20 h). In human membranes, the IKKβ inhibitor TPCA-1 (7 μM) and p38 MAPK inhibitor SB239063 (20 μM) administered to the amniotic compartment were the most effective in inhibiting accumulation of cytokines and PGE2in the fetal compartment. NAC (10 mM) inhibited accumulation of PGE2and IL10 only; NBDI (10 μM) had no significant effect. In addition to the fetal compartment, SB239063 also exerted consistent and significant inhibitory effects in the maternal compartment. TPCA-1 and SB239063 suppressed oIL8 production, while all CSAIDs tested suppressed ovine PGE2production. These results support the further investigation of intra-amniotically delivered CSAIDs for the prevention of inflammation-mediated PTB.