Human milk microbiota profiles in relation to birthing method, gestation and infant gender

This study revealed a range of bacterial genera in human milk, consistent with previous studies [20–22]. Interestingly, even when a baby was born extremely prematurely (subject S24), the mother’s milk was similar in composition to that of a woman giving birth at full term (subject S30). It can be speculated that this might be a fail-safe mechanism whereby the mother is “ready” to pass along her bacterial imprint irrespective of when the baby is born. If so, the microbiota would appear to be recalcitrant to late pregnancy hormonal and inflammatory changes, which could indicate an evolutionary pressure protecting this niche for the baby’s benefit. Further studies on this concept are warranted, and cases where the milk profiles are very different or the outcome for the baby negative could be particularly insightful.

It was not surprising that milk from emergency caesarean (C) section deliveries (i.e. non-elective) did not differ from women who gave birth vaginally, as this decision is made at the time of labour, when the hormones and timing for birth have been initiated. We also did not see differences between non-elective C section deliveries and vaginal births which is in contrast to a study published by Cabrera-Rubio et al. in which they concluded that the human milk microbiome is shaped by mode of delivery [23]. However, we do not believe that the authors adequately proved this claim. In their analysis, the milk samples collected were from both obese and normal weight women with no indication of the proportion in each group. Since it was shown in that same paper that body mass index influences the milk microbiota, the subject’s weight could have been a confounding factor responsible for the observed differences [23]. In addition, the author’s claim that the milk microbiota is influenced by mode of delivery was based solely on visual observations of bar plots (which were not very distinct between the two groups), with no principal coordinate analysis (PCoA) plots or statistical analyses to support this claim. These conflicting results between our group and that of Cabrera-Rubio highlight the need for future studies with larger sample sizes and inclusion of women from various demographics.

Because of the multivariate nature of the data, differences, if they are present, may have been masked by the different variables confounding each other. Thus, a larger sample size allowing for linear regression analysis would strengthen our claims. However, there is the possibility that no matter the sample size, differences will only be apparent when examined at the level of the individual. Schwarzberg et al. [24] showed that treatment for periodontitis had no effects on microbial profiles when compared to controls. However, when pre- and post-treatment effects for each individual were compared, significant changes in bacterial composition were observed. The authors thus emphasized the importance of comparing shifts from a personalized healthy state to a personalized disease state in order to truly understand biological changes [24].

Proteobacteria and Firmicutes were the dominant phyla, consistent with a previous high-throughput study performed in Switzerland and analyzed with a different sequencing platform (454 sequencing) [21].

The detection of Staphylococcus, as the most abundant organism, is likewise consistent with other studies [20–22]. Staphylococcus is more abundant in the gut of breast-fed newborns compared to formula-fed ones [25, 26], but their numbers start to decrease after the first week of life when oxygen has been consumed and an environment favourable for anaerobes is created [4, 27]. Unlike Staphylococcus, which is present in high abundance in at least the first week of life, Proteobacteria are never found in high numbers in the faeces of newborns, infants or Western adults [4, 28, 29]. Thus, an important question arises, as to what the evolutionary significance is of having such a diverse population of bacteria in milk, if only a select few seem to colonize during the development of the neonate. There are a few possibilities; firstly, persistent colonization is not always needed for beneficial effects, as transient exposure can be just as effective [30, 31]. Secondly, bacteria in milk may not have to be passed on from the mother to the child to exert their beneficial effects. Many of the protective factors in milk such as antibodies, immune cells, lactoferrin and beta defensins originate from the mammary gland and not from the blood [32]. With the ability of bacteria to regulate host gene expression, such as anti-microbials, and their ability to stimulate the immune system, the plethora of bacteria in breast milk could be inducing up-regulation of these protective factors in the breast that then get passed on in high concentrations to the neonate via milk.

From another perspective, there is the possibility that some of these milk microbes have either limited or no effects on the offspring but are present for the benefit of the mother, such as in the protection against mastitis. Mastitis is a painful inflammatory condition of the breast with the main causative agent being Staphylococcus aureus, and it has been shown that some milk commensals have the ability to prevent S .aureus growth and infection [33]. The same is true for a skin derived strain of Propionibacterium acnes, which prevents growth of S. aureus via its by-products of glycerol fermentation [34]. With the high abundance of glycerol present in milk, milk-derived strains of Propionibacterium may have similar growth-inhibiting properties, which would account for its presence in every milk sample collected.

Lactobacillus was present in high abundance in milk, so for those women who deliver by C section and thus do not transfer lactobacilli from the vagina to the infant, the milk could provide a means for these beneficial organisms to reach the infant gut.