Analysis of the human breast milk microbiome and bacterial extracellular vesicles in healthy mothers

The human gut microbiome, comprising over 100 trillion microbes, plays a crucial role in host immune system development and overall health. Early gut microbiome acquisition in infancy significantly impacts lifelong health. Factors like delivery method, antibiotic use, environment, and nutrition influence this process. Human breast milk (HBM) is vital for infant nutrition and contains immune components such as antibodies, immune cells, antimicrobial proteins, cytokines, and human milk oligosaccharides. While previously considered contamination, HBM is now recognized to contain commensal bacteria influencing infant gut colonization. Extracellular vesicles (EVs), nanometer-sized membrane vesicles containing bioactive molecules, are released by bacteria. These bacterial EVs have been found in various body fluids and can affect host cells. This study aimed to characterize the HBM microbiota of healthy Korean mothers and quantify key bacteria affecting infant gut colonization by analyzing both the microbiota and bacterial EVs using culture-independent techniques.

Early studies using culture-dependent methods identified some bacteria in HBM. However, the advent of non-culture-based sequencing enabled more comprehensive investigation of HBM microbiota diversity. Research indicates that HBM contains abundant commensal bacteria affecting infant gut colonization. While the core HBM microbiome composition remains debated, several studies point to Firmicutes, Proteobacteria, Actinobacteria, and Bacteroidetes as the most frequently found phyla. Specific genera like *Staphylococcus*, *Streptococcus*, *Lactobacillus*, and *Bifidobacterium* are frequently identified, although the exact composition varies across studies. There is a lack of research analyzing bacteria-derived EVs in HBM, making the study of their potential impact on the infant microbiome crucial.

This IRB-approved study (IRB No.: 1810-004-309) enrolled 22 healthy lactating mothers (July 2017–June 2018), excluding those who delivered prematurely, were ill, or used antibiotics during lactation. HBM samples were collected 1–4 months postpartum, avoiding colostrum samples. EVs were isolated using differential centrifugation (10,000 *g* for 10 min at 4°C), followed by 0.22-µm filtration to remove bacteria and particles. DNA was extracted from both bacteria and EVs using a DNeasy PowerSoil Kit. Bacterial 16S rDNA V3-V4 hypervariable regions were amplified using specific primers and sequenced on an Illumina MiSeq instrument. Bioinformatic analysis involved read trimming using cutadapt, merging paired-end reads with CASPER, quality filtering, chimera detection with VSEARCH against the SILVA gold database, OTU clustering with VSEARCH, and taxonomic classification using the GREENGENES database with UCLUST. Alpha-diversity (Chao1, ACE, Shannon, Simpson, Fisher’s) and beta-diversity (PCA) analyses were performed using R version 3.3.2. Pearson correlation coefficients assessed correlations between bacteria and bacterial EVs. Subgroup analyses examined differences based on parity (primiparous vs. multiparous) and delivery method (vaginal vs. Cesarean section).

The study included 22 lactating mothers (mean age 33.1 ± 3.4 years); 16 were primiparous, and 6 were multiparous. 8 had vaginal deliveries, and 14 had Cesarean sections. Samples were collected 11.1 ± 2.4 days postpartum. Alpha-diversity analysis showed significantly higher richness and evenness in bacterial samples than in bacterial EV samples (p < 0.001 for Chao1 and ACE, p < 0.001 for Fisher's index). Beta-diversity analysis (PCA) showed significant differences between bacteria and EVs within the same sample. At the phylum level, Firmicutes was dominant (bacteria: 56.4% ± 19.4%; EVs: 35.8% ± 12.9%), followed by Proteobacteria, Bacteroidetes, and Actinobacteria. At the genus level, *Streptococcus* and *Staphylococcus* were predominant in bacterial samples, while *Bacteroides*, *Acinetobacter*, and *Lactobacillaceae* were prevalent in EV samples. Significant positive correlations were observed between several genera in bacteria and EV samples, notably *Acinetobacter*, *Ruminococcaceae*, *Bifidobacterium*, *Comamonadaceae*, *Rothia*, and *Clostridiaceae*. Subgroup analysis revealed differences in bacterial composition based on parity and delivery method, with some genera showing significant variations.

This study is the first to analyze both HBM microbiota and bacterial EVs using next-generation sequencing. The findings align with previous studies showing Firmicutes, Proteobacteria, Bacteroidetes, and Actinobacteria as dominant phyla. However, the significant differences between bacterial and EV communities highlight the existence of metabolically active bacteria not necessarily reflecting overall HBM abundance. The positive correlation between several genera in both bacterial and EV samples suggests potential functional relationships. The presence of gut-associated genera like *Bacteroides* and *Lactobacillaceae* in EVs supports the hypothesis of vertical transfer of maternal gut microbiota to infants through HBM. The strong correlation between some genera and their EVs suggests high metabolic activity, warranting further investigation into their roles in gut colonization and immune development. Differences observed based on parity and delivery mode suggest multiple factors influencing HBM microbiota.

This study reveals the diverse bacterial communities in HBM, including bacterial EVs, and their interrelationships. Key metabolically active bacteria and the potential role of their EVs in vertical transfer of gut microbiota are highlighted. Further research should confirm the impact of bacterial EVs on infant gut colonization and identify factors influencing EV release in HBM. Understanding this process could lead to strategies promoting beneficial bacteria colonization in infants.

The relatively small sample size limits the generalizability of the findings. The lack of simultaneous analysis of maternal and infant samples (mouth, skin, stool) hinders determination of precise bacterial origins in HBM. Furthermore, the study didn't directly demonstrate the mechanism of HBM microbiota and EV involvement in infant gut colonization. Future research should address these limitations.