Human milk-associated bacterial communities associate with the infant gut microbiome over the first year of life

Abstract

Introduction: Microbial communities inhabiting the human infant gut are important for immune system development and lifelong health. One critical exposure affecting the bacterial colonization of the infant gut is consumption of human milk, which contains diverse microbial communities and prebiotics. We hypothesized that human milk-associated microbial profiles are associated with those of the infant gut.

Methods: Maternal-infant dyads enrolled in the New Hampshire Birth Cohort Study (n = 189 dyads) contributed breast milk and infant stool samples collected approximately at 6 weeks, 4 months, 6 months, 9 months, and 12 months postpartum (n = 572 samples). Microbial DNA was extracted from milk and stool and the V4-V5 region of the bacterial 16S rRNA gene was sequenced.

Results: Clustering analysis identified three breast milk microbiome types (BMTs), characterized by differences in Streptococcus, Staphylococcus, Pseudomonas, Acinetobacter, and microbial diversity. Four 6-week infant gut microbiome types (6wIGMTs) were identified, differing in abundances of Bifidobacterium, Bacteroides, Clostridium, Streptococcus, and Escherichia/Shigella, while two 12-month IGMTs (12mIGMTs) differed primarily by Bacteroides presence. At 6 weeks, BMT was associated with 6wIGMT (Fisher's exact test value of p = 0.039); this association was strongest among infants delivered by Cesarean section (Fisher's exact test value of p = 0.0028). The strongest correlations between overall breast milk and infant stool microbial community structures were observed when comparing breast milk samples to infant stool samples collected at a subsequent time point, e.g., the 6-week breast milk microbiome associated with the 6-month infant gut microbiome (Mantel test Z-statistic = 0.53, value of p = 0.001). Streptoccous and Veillonella species abundance were correlated in 6-week milk and infant stool, and 4- and 6-month milk Pantoea species were associated with infant stool Lachnospiraceae genera at 9 and 12 months.

Discussion: We identified clusters of human milk and infant stool microbial communities that were associated in maternal-infant dyads at 6 weeks of life and found that milk microbial communities were more strongly associated with infant gut microbial communities in infants delivered operatively and after a lag period. These results suggest that milk microbial communities have a long-term effect on the infant gut microbiome both through sharing of microbes and other molecular mechanisms.

Keywords: breast milk microbiome; breastfeeding; cesarean delivery; infant gut microbiome; microbial co-occurrence.

Figure 1

Top 25 most abundant taxa on average by sample type and timepoint for (A) breast milk and (B) infant stool. Colors indicate the average relative abundance of each taxon. The y-axis indicates the sample collection timepoint. For the x-axis, G indicates the genus while F indicates the family.

Figure 2

Microbial alpha diversity in the infant gut and human milk over the first year of life. Boxplots showing Simpson’s diversity index in (A) infant stool and (B) breast milk, Shannon diversity index in (C) infant stool and (D) breast milk, and observed ASVs in (E) infant stool and (F) breast milk from 6-weeks to 12-months of age.

Figure 3

Proportion of infant gut reads from ASVs in paired milk over time. The median (IQR) of the total number of infant gut microbial reads from ASVs that also occurred in breast milk from an infant’s mother divided by the total number of microbial reads in a sample for (A) all infants (Kolmogorov–Smirnov test value of p = 0.0058 compared to random dyads), (B) infants delivered by Cesarean section (Kolmogorov–Smirnov test value of p = 0.054 compared to random dyads), and (C) infants delivered vaginally (Kolmogorov–Smirnov value of p = 0.49 compared to random dyads).

Figure 4

Infant gut and human milk microbiome clusters. (A) Barplot of the average relative abundance of microbial taxa overall in 6-week infant stool and by 6-week infant gut microbiome type (IGMT). (B) Boxplot of logit(Simpson’s diversity index) by 6-week IGMT (Kruskal-Wallis rank sum test, value of p = 1.6 × 10⁻⁵). (C) Barplot of the average relative abundance of microbial taxa overall in 12-month infant stool and by 12-month IGMT. (D) Boxplot of logit(Simpson’s diversity index) by 12-month IGMT (Kruskal-Wallis rank sum test, value of p = 0.0056). (E) Barplot of the average relative abundance of microbial taxa overall in 6-week breast milk and by 6-week breast milk microbiome type (BMT). (F) Boxplot of logit(Simpson’s diversity index) by 6wBMT (Kruskal-Wallis value of p = 3.3 × 10⁻¹⁶).

Figure 5

Breast milk microbiome type is associated with infant gut microbiome type at 6-weeks of age. (A) All dyads (n = 144, Fisher’s exact value of p = 0.039), (B) dyads where infants were delivered by Cesarean section (n = 40, Fisher’s exact value of p = 0.0028), (C) infants delivered vaginally For 6wBMT vs. 6wIGMT (n = 104, Fisher’s exact value of p = 0.29), (D) dyads with no formula exposure (n = 108, Fisher’s exact value of p = 0.080). (E) There was no association between 6wBMT and 12mIGMT (n = 108, Fisher’s exact value of p = 0.51).

Figure 6

Correlations between breast milk and infant gut ASVs occur between both the same and different ASVs, and are stronger overall longitudinally. (A) Spearman correlation between 6-week breast milk microbial taxa and 6-week infant gut microbial taxa. (B) Spearman correlation between 6-week breast milk microbial taxa and 12-month infant gut microbial taxa. (C) Correlation plot showing the overall magnitude of microbial taxa correlation for each breast milk–infant gut timepoint combination.