Introduction
Breast milk (BM) is the first food for newborn infants and is recommended by the World Health Organization as the “exclusive diet” for the first 6 months of life [
1,
2]. BM contains a unique and optimal combination of nutrients and bioactive components, including immunoglobulins and cytokines, bioactive lipids, human milk oligosaccharides (HMOs), microRNAs, hormones, and microorganisms, among others [
3]. This unique composition of BM adapts to the need of the offspring and exhibits variations that extend across individuals, lactational stages, daily fluctuations, and even between feeding sessions [
4]. The concentrations of these diverse components are also contingent upon factors such as diet, maternal genetic makeup, gestational age, and the health status of the mother [
3].
It was once believed that the BM microbes were a form of extrinsic contamination and that human BM was a nearly sterile fluid, however this theory has now been rejected [
5]. To date, many studies have concluded that BM is home to its own unique microbiome, including beneficial, commensal, and potentially probiotic bacteria [
6‐
8]. The intricate and ever-evolving process through which the BM microbiota is introduced remains a subject of complexity, with facets yet to be fully understood.
Two conceivable mechanisms have been proposed to elucidate the introduction of milk microbiota. Firstly, the notion of "retrograde transfer" involves the external influx of bacteria, sourced from the areola skin and the oral cavity of the infant. The second mechanism, known as the "entero-mammary pathway," encompasses the migration of bacterial species emanating from the maternal gut to the mammary glands [
3,
9].
The application of culture-independent molecular techniques, and particularly those based on 16S rRNA genes, allowed a complementary biodiversity assessment of the human milk microbiome [
10]. Pioneering studies indicated a high complexity and inter- individual variability in the milk microbial communities with few genera (
Streptococcus,
Staphylococcus,
Propionibacterium,
Corynebacteria,
Pseudomonas,
Ralstonia,
Serratia,
Sphingomonas, and
Bradyrhizobiaceae) representing approximately half of the bacterial community abundance [
11]. Nonetheless, the relative proportional representation of these genera exhibited substantial variations across different subjects [
11]. Other studies such as the MAMI study and CHILD cohort study also identified that the BM microbiota composition is diverse and mostly dominated by the “core genera” including
Staphylococcus and
Streptococcus species [
12,
13].
The content of BM undergoes dynamic shifts during nursing to cater to the evolving needs of the developing newborn across various stages [
14]. Around mid- pregnancy, the synthesis of colostrum commences and extends for approximately five days postpartum, followed by a gradual transition to transitional BM, which persists for around two weeks [
14]. By the fourth week after childbirth, BM is fully maturate, maintaining relatively consistent composition throughout the remainder of the lactation period [
14]. Previous studies have reported changes in BM microbiota over the course of lactation, for example colostrum samples were dominated by
Weissella,
Leuconostoc,
Staphylococcus,
Streptococcus, and
Lactococcus [
15]. In contrast, at one and six month postpartum, BM samples were enriched by representatives of the oral cavity such as
Veillonella,
Leptotrichia, and
Prevotella [
15].
Our current understanding of the BM microbiome predominantly stems from studies involving mothers with healthy pregnancies, a perspective that may not readily extend to mothers that develop pregnancy complications like preterm birth (PTB). Notably, components of BM beyond the microbiome (e.g., macronutrients, bioactives etc.) differ between mothers with uncomplicated pregnancies and those facing pregnancy-related complications [
16,
17]. Given this, it is reasonable to assume similar disparities could manifest within the BM microbial communities.
Currently, only few studies have invesigated the BM microbiota in PTB [
18‐
20]; however, most of these studies have limitations as highlighted in a recent publication by Asbury et al. [
21]. Furthermore, these investigations have primarily concentrated on dissecting the microbial composition of preterm BM samples without comparing them to term birth controls. Thus, it is important to include a matched case–control cohort of women with term and preterm BM samples and study if pregnancy related complications can result in dysbiosis of BM microbiome and potentially impact the colonization of the infant gut microbiome and the developmental trajectory of their immune system.
As discussed earlier, several factors can influence the composition of the BM microbiota [
3,
22]. Previous studies such as INSPIRE have shown that BM microbiota vary among cohorts originating from different geographical regions [
6]. The landscape of advanced research often tilts towards high-income nations; thus, an empirical void emerges concerning investigations delving into the characterization of BM microbiota among mothers residing within resource-constrained contexts. This need is more prominent within Asian and marginalized refugee and migrant populations, wherein pregnancy-related complications carry profound implications for both maternal and infant well-being.
As part of our efforts to assess the molecular signature in pregnancy in mothers residing in low resources settings, we designed the MSP study [
23,
24] with an aim to characterize cross-omic trajectories in pregnant women with and without pregnancy-associated complications to improve our understanding of their role in maternal and neonatal outcomes. Thus longitudinal, high frequency sampling was conducted as the part of the study to characterize microbial composition in various anatomical sites in pregnant women including BM samples collected postpartum [
23,
24]. This is the first study to be conducted to characterize the BM microbiome in Karen and Burman women [
23,
25].
We hypothesize considerable differences in the composition of preterm and term BM samples. Since the vast majority of the neonates in our study population were exclusively breastfed, we anticipate this is as the important source of infant gut colonization. The impact of pregnancy related complications on the mother’s milk microbiota could translate to changes in the infant gut microbial colonization and long-term health outcomes of preterm infants, especially considering the high rates of morbidity and lack of resources in this vulnerable population.
Discussion
Traditionally, BM was believed to be sterile, however, recent research has shed light on its microbial diversity [
6,
11,
12,
21,
42‐
50], revealing a potential influence on both the early gut colonization of the neonates [
44] and the development of the immune system [
42]. The origin of the BM microbiota remains a subject of ongoing and sometimes conflicting debate. Among the numerous hypotheses, the enteromammary and retrograde pathways are extensively discussed. The former suggests the transfer of maternal gut microbes to the mammary glands [
9]. The enteromammary route requires transfer maternal/infant gut microbes to the mammary glands [
51]. The evidence to support this concept is provided by the migration of B-lymphocytes from the maternal gut to the mammary gland, where they differentiate into plasmacytes and produce specific IgA antibodies to protect the infant from pathogens [
52]. The retrograde pathway on the other hand involves transfer of infant oral microbiota during nursing or suckling, which in turn also leads to microbial colonization of the mammary ducts [
53]. Other proposed sources for the bacteria in BM include maternal skin, oral, use of breast pump and its plausible that several pathways contribute to the microbial content of BM.
Propionibacterium,
Staphylococcus, and
Corynebacterium are few of the typical inhabitants of the adult skin [
54] and are also found in BM [
6,
11,
43,
55], presenting a possibility that maternal areolar skin microbiota may also contribute to the composition of BM microbiota. However, a comparison of the bacterial communities found on the sebaceous skin (like the ones found on breast) and those detected in the BM samples indicates that although the two communities share common taxa, major differences also exist [
11,
56]. Among the universally predominant taxa in BM,
Staphylococcus and
Streptococcus are most frequent bacteria [
57], they are also referred to as the core genera of BM microbiota [
6]. Lackey et al., demonstrated that although the BM communities varied geographically, in samples collected from mothers across the USA, Spain, Ethiopia, Sweden, Gambia, Ghana, Kenya, and Peru, the BM core genera was universally composed of
Staphylococcus and
Streptococcus [
6]. Consistent with the previous studies, our data also unveiled a similar pattern in out cohort of mothers. In order to eliminate the potential influence of skin-related microbial contamination in BM samples during the collection process; clean (breast was cleansed with povidone solution prior to sample collection) and natural (samples collected in their natural state without cleaning the breast) BM samples were collected. No appreciable differences in diversity or relative abundances were found when the bacterial communities from the two sample types were compared, suggesting that the bacterial communities present in our BM samples were not attributed to skin contamination; rather, they appear to be intrinsic constituents of the BM microbiota.
More than 800 different bacterial species, mainly from four major phyla
Firmicutes,
Actinobacteria,
Bacteroidetes, and
Proteobacteria have been reported in the BM samples [
6,
11,
42,
43,
45‐
50,
56,
58,
59]. Among the top four phyla in our cohort,
Firmicutes dominated with 83% of the overall composition, followed OD1 (6.33%), Actinobacteria (5.45%) and Proteobacteria (3.99%). OD1 has been reported in BM samples by other studies as a minor phyla [
58], however, in our study it appeared as the second most abundant phylum. OD1, also known as
Parcubacteria, is a group of uncultured bacteria discovered in various terrestrial water environments, lakes, and wetlands [
60,
61]. These terrestrial and aquatics wetlands are common in both Thailand and Myanmar, and along the border area [
62,
63]. This suggests that the composition of BM microbiota could be influenced by the surrounding environment. Additionally, we identified other soil and water-related bacteria in our BM cohort, including
Unclassified Pedobacter, Unclassified Planctomyces, Unclassified Rheinheimera, Burkholderia gladioli, Rhizobium, Micrococcus, Unclassified Rubrobacter, Rhodobacter, Bradyrhizobium,
Novosphingobium,
Pseudomonas,
Sphingobium,
Sphingopyxis,
Sphingomonas or
Xanthomonas. This may indicate that leading a lifestyle in close contact with nature may possibly affect the enteromammary transmission of gut bacteria to the BM.
BM is divided into three distinct stages: colostrum, transitional milk, and mature milk, apparently adapting to the growing needs of the infant [
4]. Few studies have tracked the progression of microbial communities in human milk over time [
11,
64‐
66]. Cabrera-Rubio et al
. were first to define the microbial communities in BM samples from 18 mothers collected at 2 days, 1 month and 6 months of lactation using pyrosequencing and qPCR [
56]. They showed that BM undergoes considerable changes over time from colostrum to transitional and mature milk, including an increased abundance of typical oral occupant (e.g.,
Veillonella) in transitional and mature BM [
56]. Consistent with the study, our data showed a progressive increase in oral bacteria
Veillonella from colostrum to mature milk. This could be attributed to the increased interaction between BM and the infant's oral cavity as breastfeeding continues, potentially leading to the retrograde influence on the composition of BM's microbiota. A similar pattern emerged with other genera, such as
Lactobacillus,
Corynebacterium,
Propionibacterium as their proportions increased when the milk matures.
Lactobacillus have been reported to be more abundant in the gut of breast-fed neonates when compared with formula-fed babies [
67]. Together with other probiotic bacteria
Lactobacillus, have been shown to improve intestinal barrier functions in neonates by promoting mucosal barrier homeostasis, enhancing mucine production and reducing intestinal permeability [
68] ultimately leading to a healthy immune system in early and adult life [
69]. Additionally,
Lactobacillus, Propionibacterium, and
Veillonella are lactose fermenters that could prevent accumulation of lactate possibly neutralizing its unfavorable effects in infant gut [
70‐
72], the above facts suggests that BM favors the colonization of the selective bacteria in the gut of the neonates.
Contrary to other studies [
56,
66,
73], we observed an increase in diversity as lactation progresses, this phenomenon could potentially be attributed to the fluctuations in other biologically active constituents within milk throughout the breastfeeding period. Among these, Human Milk Oligosaccharides (HMOs), which function as metabolic substrates for specific intestinal microbes like
Lactobacillus and
Bifidobacterium, display varying concentrations across different stages of lactation [
74]. The increase in BM diversity possibly contributes to the progression of the infant gut microbiota's maturation, considering that the diversity of the infant gut microbiota generally increases [
75] in similar time intervals.
Previous studies have shown that prematurity impacts the other components of mother's milk: for instance protein content in preterm mother's milk is higher than in term mother's milk [
76,
77]. Similarly concentration of amino acids, including valine, threonine and arginine is higher in preterm mother's milk [
78]. Preterm BM appears also rich in sIgA but deficient in leptin [
79‐
81].
Streptococcus was the predominant genera in our preterm BM samples, whereas the abundance of
Staphylococcus was lower than previously reported [
21]. In a case–control study examining the gut microbiota of 121 mothers with vaginal deliveries, the mothers giving birth prematurely were found to have lower abundance of
Streptococcus, four days postpartum [
82], whereas another study reported a higher abundance of
Streptococcus in the gut microbiota of mothers who deliver preterm before delivery [
83]. Evidence also suggests that PTB is associated with maternal Group B
Streptococcus (GBS) colonization worldwide, previous work from SMRU suggests a low proportion (12%) of mothers carry Group B Streptococcus at birth [
84,
85]. Due to limitation in analysis, we were not able to resolve the genus
Streptococcus to species level, also we did not have the maternal gut microbiota samples from our cohort available for the present study. We also observed several gut commensals in our BM samples such as
Faecalibacterium, Prevotella, Clostridium, Bacteroides, Enterobacter which could represent the “enteromammary” pathway of translocated maternal gut bacteria. Interestingly these commensals were significantly enriched in preterm BM samples as opposed to term samples which could provide them a competitive advantage in the colonization of the preterm infant gut.
Faecalibacterium,
Prevotella,
Clostridium are major butyrate producers [
86,
87] butyrate support enterocyte proliferation, increase barrier function via induction of tight junction proteins, also have a range of antimicrobial and anti-inflammatory effects [
88] that could support the immature digestive and immune system of the preterm babies that have unique challenges at birth.
At species level,
Staphylococcus haemolyticus was more abundant in the preterm BM samples whereas
Staphylococcus epidermidis was enriched in the term BM samples. Previous studies have shown a high level of colonization of
Staphylococcus haemolyticus in the gut and skin of preterm infants [
89]. Whereas another study comparing bacterial diversity in the fecal samples of preterm and term infants showed lower levels of
Staphylococcus epidermidis in the fecal samples of preterm infants [
90]. This could provide indication of the vertical transmission of BM microbes from the mother to her infant, a process likely influenced by maternal health status. Preterm BM samples also demonstrated higher richness and diversity in terms of both core and rare taxa which could indicate an attempt to maximize ecosystem multifunctionality [
91].
Antibiotic exposure is known to be associated with disruption in the richness, diversity and metabolic pathways of the intestinal microbiota [
92]. Hence, it is conceivable that maternal antibiotic exposure may also perturb the BM microbiota. To reduce the risk of neonatal infections antibiotic treatment is often recommended in some cases [
93]. Antibiotic exposure in utero and during infancy has been associated with an increased risk for the same diseases [
94‐
96]. Recent studies have shown that intrapartum antibiotic exposure was significantly associated with changes in the milk microbial composition [
97]. In our study, we did not find any significant impact of antibiotics exposure over the course of pregnancy or close to delivery neither on the diversity nor on the composition of the overall BM microbiota. These inconsistencies may have resulted from variations in the type, dosage, and timing of antibiotic administration, as well as from other environmental and genetic factors which require further investigation using larger cohorts and more studies.
Overall, we found significant differences in BM microbial communities depending on the lactation stage and gestational age. BM microbiota of PTB mothers was highly individualized likely suitable for the preterm infants. The strength of the study relies in the fact that we had a matching case control cohort which essentially minimizes biases and the effect of confounding factors. While results from this study are promising and warrant more research, it is worth noting that our study has few limitations. Firstly, low sequencing accuracy and low coverage of terminal regions associated with 16 S rRNA gene sequencing can result in low taxonomic resolution, as seen in our data where we had limited resolution at the species level. Secondly, the number of subjects that developed PTB was lower than the rates reported internationally, and a larger figure would have been more desirable for analytical purposes. Eventually, a deeper understanding of the determinants and progression of BM microbiota can provide insights into how the microbiota can be manipulated to improve infant health. These crucial early life phases and their effect on health and disease need to be deeply examined in order to support optimal microbial immune homeostasis.